fy15 lanl annual report 66bfa 2

Summary

FY15 ANNUAL PROGRESS REPORT Laboratory Directed Research and Development Los Alamos National Laboratory About the Cover Established in 2003, the Los Alamos Engineering Institute is an ongoing re- search and education collaboration between the Los Alamos National Labora- tory and the University of California San Diego’s Jacobs School of Engineering. It promotes mission-relevant engineering research collaborations with UCSD faculty and students and provides the Laboratory with a proactive approach…

Page 1

FY15 ANNUAL PROGRESS REPORT Laboratory Directed Research and Development Los Alamos National Laboratory

Page 2

About the Cover Established in 2003, the Los Alamos Engineering Institute is an ongoing re- search and education collaboration between the Los Alamos National Labora- tory and the University of California San Diego’s Jacobs School of Engineering. It promotes mission-relevant engineering research collaborations with UCSD faculty and students and provides the Laboratory with a proactive approach to recruiting, retention and revitalization of its technical staff. Several alumni of the Engineering Institute’s Dynamics Summer School were recruited to Los Alamos through LDRD Postdoctoral Research and Development projects; they continue on at the Laboratory today as engineering scientists, some leading LDRD Early Career projects of their own and some contributing to projects as co-Investigators. These early career researchers are featured on the cover. Disclaimer The Los Alamos National Laboratory strongly supports academic freedom and a researcher’s right to publish; therefore, the Laboratory as an institution does not endorse the viewpoint of a publication or guarantee its technical correctness. With respect to documents available from this server, neither the United States Government nor the Los Alamos National Security, LLC., nor any of their employees, makes any warranty, express or implied, including the warranties of merchantability and fitness for a particular purpose, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial products, process, or service by trade name, trademark, manufacturer, or otherwise, does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or the Los Alamos National Security, LLC. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or the Los Alamos National Security, LLC., and shall not be used for advertising or product endorsement purposes. Unless otherwise indicated, this information has been authored by an employee or employees of the Los Alamos National Security, LLC. (LANS), operator of the Los Alamos National Laboratory under Contract No. DE-AC52-06NA25396 with the U.S. Department of Energy. The U.S. Government has rights to use, reproduce, and distribute this information. The public may copy and use this information without charge, provided that this Notice and any statement of authorship are reproduced on all copies. Neither the Government nor LANS makes any warranty, express or implied, or assumes any liability or responsibility for the use of this information. Issued March 2016 LA-UR-16-21921

Page 3

Structure of this Report In accordance with U.S. Department of Energy Order (DOE) 413.2B, the Laboratory Directed Research and Development (LDRD) annual report for fiscal year 2015 (FY15) provides summaries of each LDRD-funded project for the fiscal year, as well as full final reports on completed projects. The report is organized as follows: Overview: An introduction to the LDRD Program at Los Alamos National Laboratory (LANL), the program’s structure and strategic value, the LDRD portfolio management process, and highlights of outstanding accomplishments by LDRD researchers. Project Summaries: The project summaries are organized first by Focus Areas: Complex Natural and Engineered Systems, Information Science and Technology, Materials for the Future, Nuclear and Particle Futures, and Science of Signatures. Within each category, summaries are organized by LDRD component in the following order: Directed Research (DR, Exploratory Research (ER), Early Career Research (ECR), and Postdoctoral Research and Development (PRD). Annual reports for ongoing projects appear first, followed by full final reports that ended in FY15. Projects are listed in numerical order according to their project identification number, which consists of three parts. The first is the fiscal year in which the project began; the second is a unique numerical identifier; and the third identifies the project component. At Los Alamos, postdocs are hired throughout the fiscal year, and at the time this report was published, some PRD projects did not have significant progress to report due to the fact that the postdoc had just been hired. In the first few weeks of a new PRD project, a postdoc will spend time in general employee training, as well as any additional training required for work in a specialized lab or facility. This was the case for the following PRD projects: 20140685PRD4 20150743PRD3 20150708PRD2 20150758PRD3 20150710PRD2 20150711PRD2 20150759PRD3 20150742PRD3 20150702PRD1 20150701PRD1 20150744PRD3 20150712PRD2 20150705PRD2 20150741PRD3 20150717PRD2 Acknowledgements Technical Review William Priedhorsky Jeanne Robinson Publication Designer Andrea Maestas Team Contributors Lisa Lujan Debbie Martinez Stephen Schultz Epolito Ulibarri Susan Whittington Peter Haase

Page 4

Table of Contents 12 Program Overview 66 Understanding The Catalytic Conversion of Oligosaccharides to Fuels and Chemical Complex Natural and Engineered Systems Feedstocks Andrew Sutton 25 Discovery Science of Hydraulic Fracturing: Innovative Working Fluids and Their Interactions 71 Methane Coupling Chemistry Promoted by with Rocks, Fractures, and High Value Hydro- Catalysts Containing Inexpensive Metals carbons John C. Gordon Hari S. Viswanathan 73 Fundamental Actinium Science In Search of 29 Combating Antibiotic Resistance: Targeting Efflux Radiotherapeutics Pump Systems at Multiple Scales Eva R. Birnbaum Sandrasegaram Gnanakaran 76 Label-Free Measurement of Single Cells by 32 Quantitative Biology: From Molecules to Cellular Impedance Cytometry in a Microfluidic Device Function Babetta L. Marrone Angel E. Garcia 80 A New Hypothesis to Explain the Variability of the 37 SHIELDS: Space Hazards Induced Near Earth Outer Radiation Belt: Can we Predict Post-storm by Large Dynamic Storms - Understanding, Fluxes of Energetic Electrons Based only on Pre- Modeling, Predicting storm Fluxes of the Lower-energy Population? Vania K. Jordanova Gregory S. Cunningham 40 Critical Watersheds: Climate Change, Tipping 82 Multidisciplinary Studies of Long Non-coding Points, and Water Security Impacts RNAs: Towards a Structural Basis for RNA in Richard S. Middleton Epigenetics Karissa Y. Sanbonmatsu 43 Using Microreactors for Efficient Plutonium Separations (U) 85 How Trees Die: Multi-scale Studies of Carbon Stephen L. Yarbro Starvation and Hydraulic Failure during Drought Sanna A. Sevanto 46 Maximizing Flux through Engineered Metabolic Pathways 92 Pyrocumulus Collapse: Unpredicted Wildfire Clifford J. Unkefer Dangers Young-Joon Kim 53 Deciphering Nature’s Chemical Toolbox: Decoding the Logic of Biosynthetic Assembly Lines 96 Biocatalysts for Remediation of Uranium Wastes Alexander Koglin Francisca Rein Rocha 55 Toward a Coupled Multi-physics Modeling 100 Structure Determination of Large and Membrane- Framework for Induced Seismicity Bound Proteins by Nuclear Magnetic Resonance Satish Karra (NMR) Spectroscopy Ryszard Michalczyk 57 The World’s First Drought and Insect Caused Global Tree Mortality Monitoring System 104 Redox active Catalysts for C-C Coupling Reactions Chonggang Xu Relevant to Renewable Energy John C. Gordon 61 From Troposphere to Ionosphere: How Much Do Thunderstorms Disturb the Total Electron 106 Novel Chemical Architectures for Supercapacitor Distribution? Electrolytes: Comparing In Situ Scattering Erin H. Lay Measurements to Theory and Simulation Cynthia F. Welch

Page 5

110 One-step Supercritical Fluid Extraction (SFE) and 139 Chemically Modifying the Uranyl Ion Separation of Rare Earths (RE) Jaqueline L. Kiplinger Stephen L. Yarbro 143 Catalytic Mechanism and Inhibition of Metallo- 114 Low-Grade Thermal Energy Recovery beta-lactamases (MBL), the Ultimate Threat Robert P. Currier Against Antibiotics Ryszard Michalczyk 118 Building a Foundation for Understanding How Pathogens Subvert the Host Immune System 146 Stimuli Responsive, Functional Biopolymers: Thomas C. Terwilliger Quinic Acid-Based Polymers and Their Assemblies Hsing-Lin Wang 120 Ultrafast Vacuum Ultraviolet Spectroscopy of Complex Materials 150 Single-Cell Genomics for Better Control of Plant Dmitry A. Yarotski Pathogens Shunsheng Han 123 Bayesian Information Gap Decision Analysis Velimir V. Vesselinov 152 Exploring Doubly Parasitic Radioisotope Production Via Secondary Neutron Fluence From 125 Tracking Microbial Activity to Predict the Impacts the 100 MeV IPF Irradiations of Climate Change on Ecosystem Function Eva R. Birnbaum Cheryl R. Kuske 155 Hybrid Nanostructures for Photoreduction of CO 2 127 Complexes Containing Redox-Active Ligands for to Hydrocarbons the Synthesis of Fuels from Readily-Available Hongwu Xu Carbon Sources John C. Gordon 159 Joint Inversions of Seismic and Gravity Data in Volcanic Areas to Advance Hazards Assessment: 128 Bottom-up Chemical Synthesis of Large, Well- A Focus on the Alaskan Subduction Zone and Defined, and Organo-Processable Nanographene- Kilauea, Hawaii based Triarylamine for Optoelectronic Monica Maceira Applications Hung-Ju Yen 162 Discovery of Novel Bioactive Natural Products Alexander Koglin 130 Petabyte-Scale Computational Analyses of Genomic Data to Elucidate Aging Mechanisms 165 From Food to Fuel: Making Ammonia Synthesis William S. Hlavacek Viable for Energy Storage Applications James M. Boncella 132 Access to Industrially Important Optically Active beta-X-alcohols via Direct Enantioselective Ester 168 Genetically Encoded Tools for Light-controlled Hydrogenation Molecular Assembly Pavel Dub Geoffrey S. Waldo 133 Synthesis and X-ray Spectroscopy of Actinide 173 Mesoscopic Lattice Boltzmann Modeling and Thiocyanates Investigation of Boiling Multiphase Flows Stosh A. Kozimor Qinjun Kang 134 Anaerobic, Solvothermal Synthesis of Lanthanide Information Science and Technology and Actinide Kagomé Antiferromagnets Stosh A. Kozimor 177 Information-Driven Materials Discovery and Design 135 A Physics-Based Numerical Model for Next- Turab Lookman Generation Laminar Flow Batteries Qinjun Kang 180 Next Generation Quantum Molecular Dynamics Anders M. Niklasson 137 Resolving Kinetic Scales in 3D Global Magnetosphere Simulations 183 Optimization and Control of Dynamic Networks William S. Daughton Angel E. Garcia

Page 6

187 Scalable Codesign Performance Prediction for 283 Exploring Mechanisms of Catalysis on Plutonium Computational Physics Surfaces (U) Stephan J. Eidenbenz Marianne P. Wilkerson 189 Cyberphysical Systems and Security 286 Mesoscale Materials Science of Ductile Damage in Scott N. Backhaus 4 Dimensions: Towards the Computational Design of Damage-Tolerant Materials 191 Disruptive Innovation in Numerical Ricardo A. Lebensohn Hydrodynamics Jacob I. Waltz 289 Aging in Delta Plutonium Alloys: A Fundamental Approach 195 Empowering the Expert: Machine Learning with Franz J. Freibert User Intelligence Reid B. Porter; Gowri Srinivasan 292 A New Approach to Mesoscale Functionality: Emergent Tunable Superlattices 204 Quantum Methods for Fast Signal Processing and Marc Janoschek Metrology Rolando D. Somma 295 Meso-Photonic Materials for Tailored Light- Matter Interactions 206 Stochastic Modeling of Phase Transitions in Houtong Chen Strongly Interacting Quantum Systems Christopher Ticknor 298 Fighting Carbon with Carbon: All-Carbon Nanomaterial Photovoltaics 236 A Computationally Efficient Model for Warm Stephen K. Doorn Dense Mixtures Didier Saumon 305 Design Principles for Materials with Magnetic Functionality 241 Software/Hardware Mapping for Data Locality Joe D. Thompson Optimization Hristo N. Djidjev 310 Non-Precious Metal Electrocatalysts for Clean Energy 247 Contextual Learning and Recognition Piotr Zelenay Alexei N. Skurikhin 316 Phase Stability of Multi-Component 252 A New Approach to Multiscale Plasma Physics Nanocomposites Under Irradiation Simulations Blas P. Uberuaga Gian L. Delzanno 323 Quantum Chemistry, Information, Materials and 258 Sparse, Distributed, and Robust Network Control Metrology Marian Anghel Angel E. Garcia 263 Integrated Photonics Pathfinder (IPP) 329 Non-Equilibrium Phenomena in Materials, Fluids, Kevin P. Mccabe; Ivan Christov; Wojciech H. Zurek and Climate Angel E. Garcia 270 A Quadrature Approach for Non-Gaussian Uncertainty Representation and Propagation 335 Attosecond Dynamics of Correlated Electrons in David Palmer f-Electron Materials Steve M. Gilbertson Materials for the Future 337 Controlling the Electronic Structure of 275 Photoactive Energetic Materials for Quantum Emerging Atomically Thin Materials Through Optical Initiation Heterostructuring Robert J. Scharff Jinkyoung Yoo 278 Multiferroic Response Engineering in Mesoscale 340 A Novel Crystal Plasticity Model that Explicitly Oxide Structures Accounts for Energy Storage and Dissipation at Dmitry A. Yarotski Material Interfaces Jason R. Mayeur

Page 7

342 Room Temperature Oxidation and Corrosion of 392 Unraveling Interfacial Charge and Energy Transfer Plutonium Processes in Single Layer 2D Transition Metal Alison L. Pugmire Dichalcogenides Aditya Mohite 347 Novel Mesoscale Modeling Approach for Investigating Energetically Driven Nanoscale 394 Microstructure Based Continuum Process Defect/Interface Interactions Modeling of Weapons Metals Abigail Hunter Rodney J. Mccabe 354 Magnetic Field Effects on Convection-Modified 397 Solute and Microstructure Prediction during Solid-Liquid Interfaces Processing (U) Amy J. Clarke Amy J. Clarke 359 Effects of Joining Processes on Bimetal Interface 399 In situ X-ray Imaging and Diffraction to Content and Radiation Damage Resistance Understand the Mechanics of Initiation John S. Carpenter Mechanisms in Explosive Single Crystals Kyle J. Ramos 365 Probing Interface Reactions of Calcite Nanocrystals at Elevated Temperatures and 402 Enabling Mesoscale Science: Nonlocal Pressures Dislocation-Flux Crystal Plasticity Under Shock Rex P. Hjelm Jr Loading Conditions Darby J. Luscher 369 Spin-state Transitions as a Route to Multifunctionality 404 Embedded Fiber Sensor Approach for Dynamic Vivien Zapf Pressure and Temperature Measurements in Explosives 372 Beyond the Chemical Reaction Zone: Detonation George Rodriguez Product Gases in the Warm Dense Regime Dana M. Dattelbaum 407 Thin-Film Heat Switch for Active Thermal Management of CubeSat Payloads 374 Topological Kondo Insulators Alexander H. Mueller Joe D. Thompson 409 Sub-Grid Meso-Scale Model for Twinning and Slip 376 Semiclassical Modeling of Non-adiabatic Processes Processes in Molecular Materials Curt A. Bronkhorst Dima V. Mozyrsky 412 Higher Order Spin Noise Spectroscopy: 378 Making nano-Mg a reality From Foundation of Quantum Mechanics to Irene J. Beyerlein Applications Nikolai Sinitsyn 382 Toward Tunable Functionalties Using Epitaxial Nanoscaffolding Films 414 Three-Dimensional Porous Nanographene for Quanxi Jia Highly Efficient Energy Storage Hsing-Lin Wang 385 Direct-gap Group-IV Nanocrystals:Cheap, Versatile Materials for Solar Cells 416 Controlled Helium Release from Composite Sergei A. Ivanov Plasma Facing Materials through Interface Design Yongqiang Wang 387 Metal and Semiconductor Nanocrystal Superlattices Under Pressure: Multiscale Tuning 419 Precision ‘Bottom-Up’ Fabrication of Non-classical of Structure and Function Photon Sources Jennifer A. Hollingsworth Jennifer A. Hollingsworth 390 Interactions of Electrons with Quantum- 422 Perovskite Solar Cells: The Next Frontier in Energy Confined Systems Probed by Scanning Tunneling Harvesting Spectroscopy Aditya Mohite Victor I. Klimov

Page 8

424 Defect-Induced Emergent Magnetism in 473 Additive Manufacturing of Mesoscale Energetic (Nonmagnetic) Complex Oxides and their Materials: Tailoring Explosive Response through Interfaces Controlled 3D Microstructure Scott A. Crooker Alexander H. Mueller 426 Energetic Materials Cocrystal Engineering: Toward 477 Efficient Carbon Nanotube Growth on Graphene- Superior Munitions Metal Surfaces Philip Leonard Enkeleda Dervishi 429 Majorana Fermions for Quantum Information 479 Understanding and Controlling Magnetism in Filip Ronning Multiferroics with THz Pulses Rohit P. Prasankumar 431 3-Dimensional Characterization of Nuclear Fuels: Microstructural Evolution under Representative 481 Design Principles for High Performance Organic Temperature and Thermal Gradients Photovoltaics Donald W. Brown Aditya Mohite 435 Very Low Temperature Scanning Point Contact 484 Synthesis of Novel Energetic Materials Spectroscopy Investigation of Inhomogeneous David E. Chavez States on the Nano-scale Roman Movshovich 486 Investigating Structure-Directing Agents in Nonconventional Nanowire Synthesis Using a 442 Excited State Quantum Interactions in Carbon Transmission-Electron-Microscope Flow-Cell Nanotubes Holder Stephen K. Doorn Jennifer A. Hollingsworth 447 Enhancing Thermoelectric Properties of 488 Quantum Control of Tailor-designed Photoactive Topological Insulators through Nanostructuring Energetic Materials Nikolai Sinitsyn Tammie R. Nelson 449 “Upscaling” Nanoscale Thermoelectrics: The 490 Ultrafast Carrier Dynamics in Novel Two- Meso-macroscale Design Challenge for Real- Dimensional Nanomaterials World Energy Needs Victor I. Klimov Jennifer A. Hollingsworth 492 New Room Temperature Multiferroic Thin Films 453 Giving Cold Atoms Weight: Creating Heavy Enabled by Strain Engineering Fermions in Optical Lattices Quanxi Jia Cristian D. Batista 494 Search for the Topological States in F-electron 456 Topology in Superposition: Quantum Systems Decoherence in Many-body Systems Tomasz Durakiewicz Wojciech H. Zurek 496 Rational Design of Multiferroics and Influence of 459 Accurate Interfacial Structures for Atomistic Cationic Disorder on Multiferroicity in Perovskites Simulations: Minimizing the Grand-Canonical Free Blas P. Uberuaga Energy Danny Perez 499 Studies on Functional Materials: Design and Optimization 463 Understanding and Controlling Magneto-Electric Turab Lookman Coupling in Multiferroic Materials Dmitry A. Yarotski 501 Probing and Controlling the Surface States of Topological Insulators 468 Understanding of Nanoscale Fracture and Scott A. Crooker Its Application in Developing High Fracture Toughness Nanoscale Composites 503 Three-Dimensional Nitrogen-Doped Porous Nan Li Nanographene for High-Performance Supercapacitor Hsing-Lin Wang

Page 9

505 Photophysical Properties of Self-Assembled 550 Probing New Sources of Time-Reversal Violation Nanoclusters with Neutron EDM Jennifer Martinez Takeyasu Ito 508 Dynamic Strength and Phase Transition Kinetics in 553 The Role of Short-lived Actinide Isomers in High Geophysical Materials Fluence Environments (U) Cynthia A. Bolme Marian Jandel 510 In-situ, 3D Characterization of Solidification in 556 Research Enabling a Next Generation Neutron Metals Lifetime Measurement Amy J. Clarke Steven Clayton 512 Dendritic Microstructure Selection in Cast 559 k_effective: First Measurement of a Nanosecond- Metallic Alloys Pulsed Neutron Diagnosed Subcritical Assembly Amy J. Clarke Anemarie Deyoung 514 Frustrated Materials 562 Multi-Scale Kinetics of Self-Regulating Nuclear Cristian D. Batista Reactors Venkateswara R. Dasari 517 Designing and Probing Novel Materials by Pressure Tuning of Nanocrystals 565 Next-Generation Double Beta Decay Experiment Hongwu Xu Steven R. Elliott 523 NMR Study of Quantum States of Matter 567 Cold Cathodes for Next Generation Electron Joe D. Thompson Accelerators: Methodologies for Radically Improving Performance and Robustness 526 Electronic and Photonic Transport in Chiral Nathan A. Moody Materials and Nanostructures Diego A. Dalvit 570 Nuclear Science for Signatures, Energy, Security, Environment 528 Alternating Positive-Negative Charge Systems: Albert Migliori New Compounds and Synthetic Routes David E. Chavez 575 Peta-scale Studies of Cosmic Explosions and Supernova Shock Breakout with Palomar 532 Microstructured Biohybrid Synthesis of Transient Factory Photosynthetic Assemblies Przemyslaw R. Wozniak Gabriel A. Montano 592 First Principle Study of Relativistic Beam and 536 Broken Symmetries in Superconductors Plasma Physics Enabled by Enhanced Particle-In- Albert Migliori Cell Capability Chengkun Huang 538 Hybrid Metal-Semiconductor Nanostructures for Optimized Photosynthetic Algal Growth 596 Answer to Heavy Element Production Puzzle by Jennifer A. Hollingsworth Measuring Neutron-induced Charged Particles at LANSCE 541 Ultrafast Measurements of Emergent Magnetism Hye Young Lee in New Complex Oxide Materials Scott A. Crooker 600 Effects and Mitigation of Hot Electrons in Direct Drive Implosions 543 Shock-Driven Material Dynamics Investigated by Natalia S. Vinyard Ultrafast X-ray Diffraction Cynthia A. Bolme 603 Hybrid Shock Ignition as an Alternate Concept for Fusion Energy Nuclear and Particle Futures Eric N. Loomis 546 Illuminating the Origin of the Nucleon Spin 606 Quantum Kinetics of Neutrinos in the Early Ivan M. Vitev Universe and Supernovae Vincenzo Cirigliano

Page 10

608 Designing the Next Generation Compton Light 644 Emittance-Reduction System for Future Source Accelerator Solutions Nikolai Yampolsky Kip A. Bishofberger 610 Combined Klystron and Linac (Klynac) 646 Reactor Power for Large Displacement Bruce E. Carlsten Autonomous Underwater Vehicles Patrick R. Mcclure 611 Multi-GeV Electron Radiography Frank E. Merrill 650 Towards Generating Laboratory Gigagauss Magnetic Fields and Their Impact on ICF Dynamics 612 Photocathodes in Extremes: Understanding Kirk A. Flippo and Mitigating High Gradient Effects on Semiconductor Cathodes in X-FELs 668 3D Turbulent Magnetic Reconnection Nathan A. Moody Experiments and Simulations Scott C. Hsu 615 Superconducting Nuclear Recoil Sensor for Directional Dark Matter Detection 672 Boosting New Physics Discoveries with Jet Markus P. Hehlen Substructure Christopher Lee 618 Neutrinos and Fundamental Symmetries in Nuclei Stefano Gandolfi Science of Signatures 620 Assessing the Quantum Physics Impacts on Future 678 Optical and Laser Spectroscopy of Th-229 X-ray Free-electron lasers Electronic and Nuclear Transitions for the Mark J. Schmitt Development of Solid State Nuclear Quantum Sensors 622 Transport Properties of Magnetized High-Energy Xinxin Zhao Density Plasmas Jerome O. Daligault 681 Remote Raman-LIBS Spectroscopy (RLS) Signature Integration 624 Magnetic Rayleigh-Taylor Instability Samuel M. Clegg Daniel Livescu 683 Explosives Signatures for Detection: Nonlinear 627 Enhancing the Long-Baseline Neutrino GHz to THz Responses Experiment Oscillation Sensitivities with Neutron David S. Moore Measurements Keith R. Rielage 686 Chemical Signatures of Detonation Born From Extreme Conditions (U) 629 Direct Numerical Simulations of Magnetic David Podlesak Rayleigh-Taylor Instability Daniel Livescu 689 Integrated Biosurveillance Benjamin H. Mcmahon 631 Extreme-Scale Kinetic Plasma Modeling of Turbulence and Mix Using VPIC 692 Signatures of Change - Habitat Earth Brian J. Albright Reinhard H. Friedel 633 Ultra-Bright Electron Beam Acceleration in 694 High Performance Atom-Based Sensors for Fields Dielectric Wake Accelerators and Rotations Evgenya I. Simakov Malcolm G. Boshier 636 Beyond the Standard Halo 699 Battlefield MRI Michael S. Warren Michelle A. Espy 638 Coherent Diffractive Imaging of Ultrafast Ejecta 705 Laser-Driven Neutron Source for Detection of Processes Nuclear Material Cynthia A. Bolme Andrea Favalli 641 In Search of Light WIMPs Alexander Friedland

Page 11

708 Deployment and Installation Technologies 745 Practical Antennas from Disruptive Technology for Distributed Measurement Systems in John Singleton Inconvenient/Hazardous Environments David D. Mascarenas 748 Next Generation Earth Models Monica Maceira 711 Trojan Horse Drug Development Approach: Targeting Gene Dosage Control to Induce 753 Wide Field-of-View Plasma Spectrometer Bacterial Suicide Ruth M. Skoug Sofiya N. Micheva-Viteva 758 Phase Transitions at Extremes: Emergence of 713 Hand-held Laser-Ultrasound Two-Dimensional Topological Defects Scanner Vivien Zapf Eric B. Flynn 764 Magnetic Nanomarker Detection and Imaging 715 Remote Whispering Applying Time Reversal with SQUIDs Brian E. Anderson Andrei N. Matlashov 719 Time Resolved Phonon Spectroscopy for 770 Electron Capture Spectroscopy for Neutrino Mass: Cryogenic Bolometer Readout Isotopes, Science, and Technology Development John J. Goett III Michael W. Rabin 721 Measuring Winds in the Stratosphere Using 775 Micro-Mirror Full-Frame Programmable Spectral Passive Acoustic Sensors Filters for the Long-wave Infrared Omar E. Marcillo Steven P. Love 723 Matter Wave Circuits 780 Cryogenic Laser Refrigerator for Infrared Imaging Changhyun Ryu Markus P. Hehlen 725 Chemical Shift Signatures of Nuclear Material: 785 Agile Persistent SSA Surveillance Networks Using 235U and 239Pu NMR Spectroscopy Mobile Platforms Michael T. Janicke W T. Vestrand 728 Solid-State Gamma-Ray Detectors Based on 788 Ultrafast Nanocomposite Scintillators: Decay Rate Quantum Dots Enhancement by Electromagnetic Coupling to Jeffrey M. Pietryga Plasmon Resonances Richard C. Schirato 731 Signatures of Reactor Operations from Plutonium Production samples (U) 792 W-Band Synthetic Aperture Radar (SAR) Anna C. Hayes-Sterbenz Bruce E. Carlsten 733 Mapping Relativistic Electron Precipitation: 796 Feasibility Study of Novel Fabrication of Dielectric Where and When? Structures for W-Band Synthetic Aperture Radar Steven K. Morley for Satellite Deployment Bruce E. Carlsten 735 Exploiting Cross-sensitivity by Bayesian Decoding of Mixed Potential Sensor Arrays Rangachary Mukundan 738 Measurement of Extinct Radionuclides in Historic Nuclear Debris (U) Warren J. Oldham 741 Ultra-sensitive Parallel Micro-imaging with Atomic Magnetometer Igor M. Savukov 743 Segregated Fuel-Oxidizer Propulsion for CubeSat Deployment Bryce C. Tappan

Page 12

A Message From William Priedhorsky, Los Alamos LDRD Program Director The U.S. Department of Energy has charged the Laboratory Directed Research and Development (LDRD) program with supporting high-risk, potentially high-value research at the national laboratories. That LDRD is a proving ground for new concepts in research and development makes its many successes that much more remarkable. Investing in high-risk science and engineering means we may not succeed every time, and yet for more than 25 years, LDRD has supported some of the most impactful technologies to come out of the National Nuclear Security Administration (NNSA) laboratories. Currently, LDRD is the largest single source of capability investment at Los Alamos, and many key NNSA programs, as well as leading R&D scientists and engineers, trace their roots to research that began under LDRD sponsorship. For example, proton radiography (pRad), whose foundations were developed by LDRD over a decade ago, has made contributions to the nuclear weapons program through more than 500 dynamic experiments. Today pRad influences decisions regarding the reuse of pits from one weapon system to another, and it provides data to help the U.S. Army improve the penetration resistance of armor for our troops on the battlefield. In the words of Los Alamos Muon Tomography Team Leader Chris Morris, “Early on, LDRD provided the resources to develop the proof of principle that is foundational to pRad. Now a key capability for maintaining the nation’s nuclear stockpile, pRad is the direct result of the synergy between the Laboratory’s defense mission and basic R&D scientists.” To be sure, there are many technical staff members at the Laboratory who credit LDRD for creating a pathway for their innovative ideas to impact national security missions. In fact, LDRD is a major vehicle for attracting, training, and retaining new technical staff, thus filling the talent pipeline to support the broad generational turnover of national security staff currently underway. From reducing global nuclear dangers, to improving our energy security, to protecting our service men and women in the field, to assuring the security of our most precious cyber assets, LDRD has made seminal contributions to every facet of national security. 12

Page 13

Overview Directed Research is organized around Focus Areas that define key areas of science, technology, and engineering in support of Los Alamos missions and that map directly Laboratory Directed Research and Development is the to the four Los Alamos science pillars, plus an additional most prestigious source of research and development multi-disciplinary Focus Area that is not captured by the funding at the Los Alamos National Laboratory. It follows pillars. Between them, they capture the capabilities that a strategic guidance derived from the missions of the U.S. are essential to our Laboratory missions in the long term Department of Energy, the National Nuclear Security Ad- (3-15 years). For each Focus Area, coordinators led a ministration, and the Laboratory. To execute that strategy, process to engage broadly with the Lab to set investment the Los Alamos LDRD program creates a free market for priorities for the FY15 Strategic Investment Plan, published ideas that draws upon the bottom-up creativity of the labwide. Laboratory’s best and brightest researchers. The combina- tion of strategic guidance and free-market competition provides a continual stream of capabilities that position the Laboratory to accomplish its missions. The LDRD program provides the Laboratory Director with the opportunity to strategically invest in forward-thinking, potentially high-payoff research that strengthens the Labo- ratory’s capabilities for national problems. Funded in FY15 with approximately 5.5 percent of the Laboratory’s overall budget, the LDRD program makes it possible for research- ers to pursue cutting-edge research and development. This in turn enables the Laboratory to anticipate, innovate, and deliver world-class science, technology, and engineering. Program Structure The Los Alamos LDRD program is organized into four program components with distinct institutional objectives: Directed Research (DR), flagship investments in mission solutions; Exploratory Research (ER), smaller projects that invest in people and skills that underpin key Laboratory Exploratory Research capabilities; Early Career Research (ECR), supporting the The ER component is focused on developing and maintain- development of early-career researchers; and Postdoc- ing technical staff competencies in key strategic disciplines toral Research and Development (PRD), recruiting bright, that form the foundation of the Laboratory’s readiness qualified, early-career scientists and engineers. In FY15, for future national missions. Largely focused on a single the LDRD program funded 278 projects with total costs discipline, ER projects explore highly innovative ideas that of $121.5million.TheseprojectswereselectedthroughaunderpinLaboratoryprograms.InFY15,LDRDfunded127rigorousandhighlycompetitivepeerreviewprocessandERprojects,whichrepresentsapproximately34arereviewedformallyandinformallythroughoutthefiscalprogram’sresearchfunds.ExploratoryResearchprojectsyear.TheLDRDProgramOfficeholdsareserveeachyeartoarefundeduptoanaveragemaximumof$350K per year make modest investments that address new opportunities. for three years. In FY15, the reserve budget was approximately $1.0M.UnlikeDRproposals,divisionendorsementsarenotDirectedResearchrequiredforERproposals;instead,thiscomponentoftheTheDRcomponentmakeslong-rangeinvestmentsinLDRDprogramisoperatedasanopenandcompetitivemultidisciplinaryscientificprojectsinkeycompetencyorpathforeverystaffmembertopursuefundingforhis/technology-developmentareasvitaltoLDRD’slong-termhergreatidea.TheERcomponentisacriticalchannelforabilitytoexecuteLaboratorymissions.InFY15,LDRDpurelybottom-upcreativityattheLaboratory.Nonethe-funded46DRprojects,whichrepresentsapproximatelyless,itisstronglydrivenbymissionneedsviathedefinition54ofthe12ERresearchcategories,andtheassignmentofprojectsaretypicallyfundeduptoamaximumof$1.8M investment between them. per year for three years. 13

Page 14

Directed Research Focus Areas Mission Impact Advance theory, algorithms, and high-performance computing Information Science and Technology to accelerate the integrative and predictive capability of the scientific method. Rapidly meet mission needs based on a thorough knowledge Materials for the Future of materials properties and interactions in relation to composition, structure, and scale. Apply science and technology tools to extremely complex Science of Signatures problems in signature, identification, and characterization, understanding, control or mitigation. Advance fundamental and applied nuclear science, including Nuclear and Particle Futures accelerator science and technology, in support of all Laboratory missions. Understand, predict, integrate, design, engineer, and/or control complex systems that significantly impact national Complex Natural and Engineered Systems security, particularly those involving energy, infrastructure, or societal stainability. Exploratory Research Technical Categories Capability Biological, Biochemical, and Cognitive Sciences Biosciences Chemistry and Chemical Sciences Chemistry Information and knowledge sciences, computer and Computational and Numerical Methods computational sciences High-performance computing, data analysis, and data-driven Computer Science, Mathematics, and Data Science science Theoretical, computation and modeling, and experimental Defects and Interfaces in Materials methods to understand defects and interfaces in materials Earth and Environmental Sciences and Space Physics Earth and space sciences Weapons science and engineering, advanced manufacturing, Engineering Applications sensors, and remote sensing Theory, computation and modelling, and experimental Emergent Phenomena in Materials Functionality methods to understand behavior of materials High-energy Density, Plasma, and Fluid Physics High-energy density plasmas and fluids and beams Measurement Science, Instrumentation, and Diagnostics Measurement methods that enable new scientific discovery Nuclear and Particle Physics, Astrophysics, and Cosmology Nuclear physics, astrophysics, and cosmology Fundamental interactions and excitations in atomic, optical, Quantum and Optical Science and molecular systems 14

Page 15

Early Career Research candidate, with his or her specialty a secondary factor. In The ECR component of the LDRD program is designed FY15, LDRD funded 76 PRD projects, which represents 7% to strengthen the Laboratory’s scientific workforce by of the program’s research funds. These postdocs are sup- providing support to exceptional staff members during ported full-time for two years. their crucial early career years. The intent is to aid in the sometimes challenging transition from postdoc to In addition to approximately 62 Director’s Postdocs, the full-time staff member, and to stimulate research in LDRD program supported 14 distinguished postdoctoral disciplines supported by the LDRD program. In FY15, the fellows at a higher salary and for a three-year term. Dis- LDRD program funded 29 ECR projects, which represents tinguished postdoctoral fellow candidates typically show approximately 3% of the program’s research funds. Early evidence of solving a major problem or providing a new Career Research projects are funded up to $225K per year approach or insight to a major problem and show evidence for two years, and only up to 60% of their overall funding of having a major impact in their research field. To recog- can be from the LDRD program. nize their role as future science and technology leaders, these appointments are named after some of the greatest leaders of the Laboratory’s past. Postdoc Research and Development The PRD component ensures the vitality of the Laboratory More postdocs are hired through DR and ER projects than by recruiting outstanding researchers. Through this invest- directly through PRD appointments. Counting both ave- ment, the LDRD program funds postdoctoral fellows to nues, in FY15 the LDRD program supported 55% of the 488 work under the mentorship of PIs on high-quality projects. postdocs at the Laboratory. The primary criterion for selection of LDRD-supported postdocs is the raw scientific and technical talent of the Christopher Lee Sarah Hernandez 2015 Early Career Research Program Award 2015 Best Poster on Plutonium Materials Department of Energy Office of Science American Nuclear Society’s Plutonium Futures • Joined the Lab in 2012 as a scientist • Graduate research assistant and PhD candidate at the University of Texas at • 2012 Early Career Researcher project (PI) Arlington • 2014, 2016 Postdoc R&D project (PI) • Co-investigator on 2014 Directed Research project • Served as chair of the Exploratory Research NPAC review team in FY15 • Lead author of article published in Journal of Physics: Condensed Matter 15

Page 16

Project Selection fiscal year. The primary objective of a project appraisal is to assess progress and provide peer input to help PIs maintain the highest quality of work. The appraisals also The LDRD program is the vehicle by which the Labora- help the LDRD Program Office monitor and manage the tory harvests the ideas of some of our best and brightest program portfolio. scientists and engineers to execute DOE/NNSA missions. This bottom-up approach is balanced by a program man- Continuing DR projects are appraised every year of the life agement strategy in which Senior Laboratory leadership of the project, with external reviewers playing an impor- sets science and technology priorities, then opens an LDRD tant role in the review that takes place in the project’s competition for ideas across the breadth of the Laboratory. second year. The internal-external review is open to all Panels formed from the Laboratory’s intellectual leaders Laboratory staff and leaders. Four project appraisers – two rigorously review proposals. Conflict of interest is carefully internal and two external – are nominated by the PI and regulated, and evaluation criteria include innovation and approved by the LDRD Program Director. When possible, creativity, potential scientific impact, viability of the re- the appraisal is held as part of a broader workshop hosted search approach, qualifications of the team and leadership, by the Laboratory. The Chair of the project appraisal panel and potential impact on Laboratory missions. The selection is responsible for writing a formal report a formal report of processes are modeled on best practices established by the review that details how well a project is addressing and the National Science Foundation (NSF) and National Insti- meeting its goals, as well as documenting any weaknesses tutes of Health (NIH). the panel may have observed. The PI is then required to respond to the concerns documented in the report with a To guarantee fairness and transparency, and to ensure that revised project plan. the strongest proposals are funded, the selection panels include managers and technical staff drawn from the full Written appraisals, held in the LDRD archives, address: range of technical divisions. Serving on an LDRD selection (1) Brief summary of accomplishments; (2) Assessment panel is often a starting point on the path to leadership of quality of science and technology, relevance to Labo- roles in the scientific community. Past LDRD panelists have ratory and national missions, progress toward goals and gone on to be Laboratory Fellows, division leaders, pro- milestones, project leadership, and the degree to which gram directors, association Fellows, and chief scientists, the project may establish or sustain a position of scientific while others have become leaders in academia. leadership for the Laboratory; and (3) Recommendations by the committee for changes in the scope or approach Benefits of Serving on LDRD Panels of the project. The criteria for the most important point – The mission of the Laboratory is to solve the nation’s most number (2) above – are derived from criteria developed difficult national security problems. By their nature, these by the National Academy of Science to assess all federally problems lack a well-defined path to solution. In fact, the sponsored research. path is often completely unknown. It is rare that such creative work is done alone; the ideas and results from In addition to formal project appraisals, which are con- many colleagues are needed, often drawn out in confer- ducted annually, the LDRD Program Director and Deputy ences, hallway conversations, journals, and seminars. LDRD Program Director meet informally with PIs in their labs at is an internal arena in which Laboratory staff serve as peer least once a year to discuss their projects. The purpose reviewers and play a key role of interaction in the scientific of these one-on-one meetings is to give PIs individualized process. Proposal selection panelists are chosen for their assistance and to determine what the LDRD Program Of- subject-matter expertise, and the discussions in which fice can do to positively impact the success of the project. they engage are not only critical to the LDRD process, but Every DR project has also been assigned a Program Devel- they also provide an opportunity for panelists to educate opment Mentor to assist the transition of LDRD successes themselves on the latest results and practices, and expose to mission. themselves to opportunities for collaboration. As noted in an evaluation of peer review conducted by the UK House Continuing ER and ECR projects are appraised in their first of Commons, “Peer review is regarded as an integral part and second years. The LDRD Deputy Program Director col- of a researcher’s professional activity; it helps them be- laborates with the technical divisions to conduct project come part of the research community.” appraisals. Like DRs, the projects are appraised according to the Federal criteria of quality, performance, leadership, Annual Project Appraisals and relevance. In FY15, the LDRD Program Office conducted an appraisal of every ongoing project it intended to fund in the next 16

Page 17

Directed Research Project Appraisals in Depth The DOE Order (413.2b, CRD) regarding LDRD requires careful evaluation of ongoing LDRD projects in order to assess their scientific quality and mission relevance. The Los Alamos LDRD program doesn’t just meet this requirement, it turns the evaluation process into an opportunity to steer its R&D towards ever-increasing excellence and mission impact. For this reason, flagship investments in DR undergo especially rigorous evaluations. High-quality Reviewers Ensure a Technically Sound DR Project Appraisal When selecting members of an appraisal panel, a Principal Investigator must seek out individuals with independence, expertise, and stature. All panel members must be independent of the project, without management or financial con- nections, in adherence with the published LDRD Conflict of Interest Policy. Between them, the panelists must have the technical expertise required to span the breadth of the project with sound technical review. And finally, while panels often contain members with a mix of seniority, there must be one or two members with stature. In the context of a DR appraisal panel, stature is recognized leadership in their field, evidenced by strong professional credentials such as an institute leader, named professorship, society fellowships, and national and/or international prizes. The LDRD Program Director approves all panels well in advance of the appraisal. 17

Page 18

Mission Relevance project established the scientific foundation for acceler- ated aging of plutonium with experiments that allowed Los Alamos and Lawrence Livermore national laboratories Mission relevance is one of the most important criteria to produce the equivalent of 60-year-old plutonium in a in the evaluation of a potential LDRD project; it is care- period of only four years. The resulting data were used as fully considered in project selection and tracked annually a foundation for the 2006 Pit Lifetime Assessment for the through the data sheet process. Many of the technologies nation. The fundamental understanding of plutonium aging that put Los Alamos on the map have deep roots in LDRD continues as a priority for our LDRD portfolio. and are valuable to DOE/NNSA mission areas of nuclear security, energy security, environmental remediation, and “Over the years LDRD helped develop resonant ultrasound scientific discovery and innovation. LDRD work also ben- spectroscopy from a lab curiosity into a powerful tool for efits the national security missions of the Department of important measurements in condensed matter physics,” Homeland Security, the Department of Defense, and Other said Albert Migliori, director of the Los Alamos Seaborg Federal Agencies. As a result, the scientific advances and Institute. “In fact, it is the only tool that can track and mea- technology innovations from LDRD provide multiple ben- sure the aging of plutonium in real time.” efits to all Los Alamos stakeholders, consistent with Con- gressional intent and the Laboratory’s scientific strategy. Today this capability enables new studies of aging in delta- plutonium alloys; advanced experiments to watch aging on Enduring Impact on Stockpile Stewardship a daily basis will form the basis for pit lifetime estimates A key responsibility of the Stockpile Stewardship program that are physically sound and advance the understanding is to assess aging of the nation’s stockpile. A 1997 LDRD of fundamental radiogenic processes in delta-plutonium. Mission Impact of FY15 LDRD Portfolio ($M) First and foremost, Los Alamos LDRD projects are required to address one or more DOE/NNSA mission areas. Due to the nature of basic R&D, the work may also benefit the mission challenges of other federal agencies. The multi-mission impact of LDRD projects is captured in the chart above, which is why the total expected benefit is approximately double actual costs of the program in FY15. 18

Page 19

In partnership with the LDRD program, the Los Alamos Engineering Institute makes investments in multidisciplinary engineering research that integrates advanced modeling and simulations, novel sensing systems, and new developments in information technology. Investments in Engineering Research and Development: Capabilities for Mission, Winning Technologies, and Top Talent Established in 2003, the Engineering Institute is an ongoing research and education collaboration between the Los Ala- mos National Laboratory and the University of California San Diego’s Jacobs School of Engineering. It promotes mission- relevant engineering research collaborations with UCSD faculty and students and provides the Laboratory with a proac- tive approach to recruiting, retention and revitalization of its technical staff. The Los Alamos Dynamics Summer School is a very selective summer school in which upper-level US-citizen undergradu- ate students from universities around the nation attend lectures and work in teams of three with a Los Alamos mentor on research projects related to the Engineering Institute’s technology focus. David Mascarenas, Stuart Taylor, and Eric Flynn are examples of top-notch alumni who were recruited to the Laboratory with support from LDRD (through Post- doctoral Research and Development projects) and converted to staff. Their accomplishments are remarkable. Mascarenas Taylor joined Flynn was became an Los Alamos converted R&D Engineer as an R&D to an R&D at Los Alamos Engineer in Engineer at in 2012 after 2013. As a Los Alamos a Director’s postdoc he in 2013 Funded fel- designed after joining lowship. He and field- the Lab as currently co- tested sen- Director’s directs the Dy- sor nodes for Funded namic Summer School and focuses his structural health monitoring (SHM) postdoc in 2011. His research focuses research on investigating the applica- on wind turbine rotor blades as part on nondestructive testing, signal pro- tion of compressive sensing techniques of a 2010 LDRD Directed Research cessing, ultrasonics, applied statistics, to structural health monitoring, the de- project. Today he applies his expertise optimization and structural dynamics. ployment of wireless sensor networks in SHM on a project that resulted in Flynn won a 2014 R&D 100 Award for from aerial robots, standoff experimen- a 2015 R&D 100 Award. The award- the Acoustic Wavenumber Spectrom- tal mechanics, and the development of winning technology, SHMTools, is eter (supported by LDRD), leads a techniques to interface humans to data software that facilitates continuous 2015 Early Career Project, was re- using vibro-tactile interfaces. He led a embedded monitoring and damage cently honored with the Achenbach 2015 Early Career Research project, and detection for new and aging infra- medal for his contributions in the recently received a Presidential Early structure. It has aerospace, civil, and field of SHM, and was on the team Career Award. mechanical infrastructure applica- that developed SHMTools. tions. 19

Page 20

Performance Metrics The LDRD program is a key resource for addressing the long-term science and technology goals of the Laboratory, as well as enhancing the scientific capabilities of Laboratory staff. Through careful investment of LDRD funds, the Labora- tory builds its reputation, recruits and retains excellent scientists and engineers, and prepares to meet evolving national needs. The impacts of the LDRD program are particularly evident in the number of publications and citations resulting from LDRD-funded research, the number of postdoctoral candidates supported and converted by the program, and the number of awards LDRD researchers received. The following performance metrics are updated annually to reflect the most current data available as more complete information often becomes available well into the next fiscal year. Intellectual Property U.S. Patents Issued Invention Disclosures FY12 FY13 FY14 FY15 FY12 FY13 FY14 FY15 LANL Patents 72 77 51 50 LANL Disclosures 129 103 71 71 LDRD Supported 11 32 14 10 LDRD Supported 28 34 12 16 % due to LDRD 15% 42% 27% 20% % due to LDRD 22% 33% 17% 23% Science and Engineering Talent Pipeline Postdoc Support Postdoc Conversions FY12 FY13 FY14 FY15 FY12 FY13 FY14 FY15 LANL Postdocs 581 596 508 488 LANL Conversions 41 57 50 55 LDRD Supported 349 367 266 266 LDRD Supported 21 26 29 24 % due to LDRD 60% 61% 52% 55% % due to LDRD 51% 46% 58% 44% Peer-reviewed Publications and Citations Publications Citations FY12 FY13 FY14 FY15 FY12 FY13 FY14 FY15 LANL Pubs 2119 2082 2215 1872 LANL Citations 33321 22021 9914 3243 LDRD Supported 458 534 475 392 LDRD Supported 8146 7620 2182 1175 % due to LDRD 22% 24% 21% 21% % due to LDRD 24% 35% 22% 36% External Collaborations (FY15) University of California New Mexico Universities International All 48 30 214 902 20

Page 21

On the Cover and Beyond The numerous publications made possible with LDRD funding help Los Alamos maintain a strong presence and scientific reputation in the broader scientific community. Not only does the program support a significant fraction of the Laboratory’s publications, it also supports much of the research featured on the covers of peer-reviewed journals. These are just a few examples of the highly visible research and development by LDRD researchers in 2015. The cover of the first issue of the Journal of Dynamic Behavior of Materials featured a series of proton radiographs of disks taken by the Los Alamos Proton Radiography Team. The images reveal the internal structure of explosively shocked aluminum, copper, tantalum, and tin. Los Alamos is an international center of excellence for research on the dynamic behavior of materials and materials in extremes. Proton radiography, and much of the R&D that grows out of enduring capability, has roots in the Los Alamos LDRD program. An article by Los Alamos scientists and collaborators in the journal Applied Spectroscopy describes the feasibility of adding Raman spectrometry to the ChemCam Laser Induced Breakdown Spectrom- eter (LIBS) instrument used with such great efficacy on NASA’s Mars Curiosity Rover. Raman spectroscopy and LIBS are highly synergistic analytical techniques. Raman spectroscopy is sen- sitive to the sample’s molecular structure from which mineralogy is directly determined, and LIBS is an elemental analysis technique that can detect all elements above the detection limit independent of the elemental mass. The new work shows that an integrated Raman spectros- copy and laser-induced breakdown spectroscopy instrument would be a valuable geoanalyti- cal tool for future planetary missions to Mars, Venus, and elsewhere. The cover highlights the many ChemCam LIBS analyses around the “Windjana” drill hole created by the NASA Mars rover, Curiosity. LIBS was developed with support from the Los Alamos LDRD program and has multiple applications, including detecting nuclear and other hazard materials, verifying construction materials, and studying cave environ- ments. In its “backpack” format, LIBS inexpensively takes atomic emission analysis from a traditional laboratory setting into the field. Los Alamos researchers Carolos Tomé and Ricardo Lebensohn (pictured left to right) recently saw their viscoplastic self-consistent code (in background) research reach 1,000 citations in Google Scholar. Their co-authored 1993 Acta Metallurgica et Materialia paper, “A self-consis- tent anisotropic approach for the simulation of plastic deformation and texture development of polycrystals: Application to zirconium alloys” was based on their joint work on a computational code to reliably simulate materials behavior. Their viscoplastic self-consistent (VPSC) code has been distributed free of charge to more than 300 external users. Lebensohn currently leads an LDRD Directed Research project titled “Mesoscale Materials Science of Ductile Damage in 4 Dimensions: Towards the Computational Design of Damage-Tolerant Materials.” The project addresses one of the most difficult outstanding problems in materials science: the development of a predictive, microstructure-sensitive ductile failure model. Lebensohn has also contributed to the LDRD peer-review processes for DR and ER proposals. 21

Page 22

Awards and Recognitions Awards and Recognitions In its more than 25 years of existence, the Los Alamos LDRD program has made many significant impacts on national security missions, steadily helping the Laboratory anticipate, innovate, and deliver solutions to some of the nation’s toughest challenges. The driving force behind each accomplishment has been the focused initiative of many talented scientists and engineers who choose to apply their knowledge and expertise in service to the nation. The LDRD program is proud to support the work of some of the Laboratory’s most accomplished researchers, who in FY15 received many prestigious awards, honors, and recognitions. Wolfram-Prandl Prize Marc Janoschek received the 2014 Wolfram-Prandl Prize during the German Confer- ence for Research with Synchrotron Radiation, Neutrons and Ion Beams at Large Facilities in Bonn, Germany. He was honored for “his pioneering studies of the spin dynamics in chiral helimagnets and the development of a cryogen-free apparatus for spherical neutron polarimetry.” Janoschek’s work on helical magnets has potential for novel memory, computing, and sensing applications continues in a 2015 LDRD Directed Research project. He also serves as co-chair of the Defects and Interfaces in Materials panel for the Exploratory Research peer-review process. APS Woman Physicist of the Month The American Physical Society (APS) named Kathy Prestridge as the Woman Physicist of the Month for July 2015. Prestridge studies the behavior of materials under high strain conditions, shock-driven instabilities, mixing and turbulence at high resolu- tion. She and her team’s research has been highlighted on the cover of Journal of Fluid Mechanics. She leads the Professional Skills Development Workshop sessions at various APS meetings, as is serving her second term as Chair of the APS Committee on the Status of Women in Physics. Prestridge has been co-investigator on LDRD projects and currently serves as Chair of the High-Energy Density, Plasma, and Fluid Physics panel for the Exploratory Research peer-review process. ASME Fellow The American Society of Mechanical Engineers (ASME) named Daniel Livescu a Fellow. The ASME Committee of Past Presidents confers the Fellow grade of mem- bership on worthy candidates to recognize their outstanding engineering achieve- ments. Livescu is an authority in the field of fluid mechanics and has made significant contributions to the LANL/DOE stewardship mission as a Principal Investigator for the NNSA Defense Science Programs. His research focuses on direct-numerical simulation of turbulence and large-scale flow computations is supported by a 2015 LDRD Explor- atory Research project. He has also served on the Computational Co-design panel for the Exploratory Research peer-review process. 22

Page 23

E.O. Lawrence Award Two Los Alamos LDRD researchers received Ernest Orlando Lawrence awards from the Department of Energy. The honor is conferred for their contributions in research and development that supports DOE’s science, energy and national security missions. Since 1959, the Lawrence Award has recognized mid-career scientists and engineers in the United States who have advanced new research and scientific discovery in the chemical, biological, environmental and computer sciences; condensed matter and materials; fusion and plasma sciences; high energy and nuclear physics; and national security and nonproliferation. Eric Dors was recognized for his technical leadership and systems engineering inte- gration of next generation satellite-based nuclear explosion sensing and detection systems, and for its impact to the nonproliferation mission. His research focuses on the development of a new generation of exo-atmospheric radiation sensors used to fulfill a critical mission need for satellite-based nuclear explosion monitoring crucial to DOE’s nonproliferation mission of global nuclear detonation monitoring and verifica- tion of the Limited Test Ban Treaty. Dors received support from LDRD through a 2001 Exploratory Research project. He also serves on the Engineering Applications panel for the Exploratory Research peer-review process. Currently, Dors is program manager for Department of Defense and Intelligence Community space programs within the Lab’s Emerging Threats Program Office. Christopher Fryer was recognized for making seminal advances in theory and modeling answering fundamental questions in astrophysics, for achievement in computational mul- tiphysics, and for contributions impacting high-energy density science. He was honored specifically for his major advances addressing fundamental questions in astrophysics, computational multiphysics, and high-energy density science, and more specifically, for supernova core collapse work using 3-dimensional modeling assimilation to model, explain, and predict astrophysical observations (e.g. from NASA’s Swift mission) and phe- nomena. Fryer’s work has deep roots in the LDRD program, most recently through a 2011 Directed Research project focused on building detailed models of supernova. Fryer is an American Physical Society fellow, a former Feynman fellow, and a Los Alamos National Laboratory fellow. GSA Vice President/President Elect The Geological Society of America (GSA) elected Claudia Mora as vice president/presi- dent elect. Mora is a stable-isotope geochemist whose research spans the traditional fields of geology, soil science and climate science. At Los Alamos, she heads the Earth and Environmental Sciences Division’s (EES) largest group, Earth System Observations. This group’s research is broad and far-reaching, intersecting geology, ecology and atmospheric sciences. Mora received support from LDRD through a 2009 Postdoctoral Research and Development project, and most recently, an Exploratory Research project. Her LDRD work has been published in the Journal of Applied Meteorology and Climatol- ogy, and in the International Journal of Wildland Fire. 23

Page 24

Complex Natural & Engineered Systems

Page 25

Complex Natural & Engineered Systems Directed Research Continuing Project Discovery Science of Hydraulic Fracturing: Innovative Working Fluids and Their Interactions with Rocks, Fractures, and High Value Hydro-carbons Hari S. Viswanathan 20140002DR Introduction with aqueous hydraulic fracturing. The proposed work Shale gas is an unconventional fossil energy resource could shift the momentum toward greater large scale that is already having a profound impact on US energy industry interest in “greener” fracturing fluids leading to independence and is projected to last for at least 100 greater public acceptance of fracking. Discoveries in fluid years. Production of methane and other hydrocarbons properties, rock properties, and their integrated interac- from low permeability shale involves hydrofracturing of tions will be required. The proposed study brings togeth- rock, establishing fracture connectivity, and multiphase er leading scientists from C, EES, MPA and T divisions to fluid-flow and reaction processes all of which are poorly develop and apply new experimental methods in observ- understood. The result is inefficient extraction with ing rock fracturing to efficiently extract hydrocarbons many environmental concerns. These phenomena are in combination with novel, benchmarked models that part of a broader class of problems involving coupled will enhance US national and energy security. Success fluid flow and fractures that are critical to other energy in the proposed work will position LANL at the fore- security areas such as shale oil, geothermal, carbon se- front of shale-gas technology, creating opportunities for questration, and nuclear waste disposal as well as crack significant industrial partnerships, and a leadership role propagation in weapons applications. A science-based in DOE programs in shale gas. This work also maintains capability is required to quantify the governing meso- capability for test containment and leakage prediction in scale fluid-solid interactions, including microstructural the unlikely case of a future US nuclear test, or the more control of fracture patterns and the interaction of engi- likely case of foreign testing. If the need for resumed neered fluids with hydrocarbon flow. These interactions testing should ever arise, the capability for understand- depend on coupled thermo-hydro-mechanical-chemical ing underground transport will just as critical as it once (THMC) processes over scales from microns to tens of was. meters. Determining the key mechanisms in subsurface THMC systems has been impeded due to the lack of Progress sophisticated experimental methods to measure frac- The project received a combination of outstanding and ture aperture and connectivity, multiphase permeability, excellent during the midterm review. A summary of the and chemical exchange capacities at the high tempera- metrics includes: ture, pressure, and stresses present in the subsurface. • 18 publication in high impact peer-reviewed jour- Our goal is to use unique LANL microfluidic and triaxial nals, five submitted, and more in preparation core flood experiments integrated with state-of-the-art • Organized session/conferences including AGU ses- numerical simulation to reveal the fundamental dynam- sions December 2014, ARMA session June 2015, and ics of fracture-fluid interactions to transform fracking a CNLS conference in September 2015 from ad hoc to safe and predictable approaches that are • More than 10 invited talks given by the DR team based on solid scientific understanding. We will develop • Several follow on projects have been funded includ- CO2-based fracturing fluids and fracturing techniques to ing $1.4Mrepresentingthreesuccessfulproposalsenhanceproduction,greatlyreducewastewater,whileforunconventionalfossilfuels,$200K for a project simultaneously sequestering CO2. with Apache and Texas tech to maximize liquid oil production from shale oil and gas, and $600K from Benefit to National Security Missions UNESE to simulate gas migration from fractures from Significant R&D is required to increase shale gas produc- clandestine nuclear tests (UNESE) tion while reducing environmental impacts associated 25

Page 26

Numerous key capabilities became fully operational in the microfluidic experiments. last year, including: • Triaxial coreflood with in situ tomagraphy capable of Task 4: In the next FY, we focus on 1) comparing lattice shear, compressive and hydraulic fractures Boltzmann simulations to the microfluidics experiments • Microfluidics with real rock and high temperature and of task 3. In addition, we will continue to study flow and pressure transport in the shale matrix that is not currently observ- • Fracture propagation simulation with fluid capability in able with our experimental program. two dimensions • High performance computing implementation of dis- Task 5: We plan to study multiple sites in the next FY. In crete fracture network model to simulate hydrocarbon addition, since we can run larger fracture networks with production our high performance computing implementation, we • Pore scale models of shale matrix will study infracture variability and reactive transport in a discrete fracture network. The midterm review resulted in the DR team coming up with a detailed plan of the remaining 18 months that ex- Conclusion ists in the form of a presentation. This presentation was Fracking phenomena involve fluid-solid interactions em- submitted to Q. bedded within coupled thermo-hydro-mechanical-chemi- cal processes over scales from microns to tens of meters. A quick summary of the overall project is given by: 1) 18 The proposed work is part of a broader class of complex publications in high impact journals, 2) 5 follow on projects systems involving coupled fluid flow and fractures that are funded by DOE, 3) unique LANL experimental and model- critical to energy security, such as shale oil, geothermal, ing capability aligned with the DOE big ideas SubTER which carbon sequestration, and nuclear waste disposal, as well is focusing controlling fracture propagation and fluid flow as, crack propagation in weapons and materials applica- in the subsurface and is expected to fully roll out in FY17. tions. Predicting and controlling fracture propagation due to fluid-solid interaction would be transformative, with Future Work significant impact beyond the hydraulic fracturing of rock. Task 1: We will continue to conduct triaxial coreflood ex- periments to characterize fracture patterns and apertures Publications under different stress conditions as well as using different Aldrich, G., J. D. Hyman, S. Karra, C. W. Gable, N. working fluids (e.g. water and CO2). Have move the triaxial Makedonska, H. S. Viswanathan, J. Woodring, and B. apparatus to AET and have successfully measurement Hamann. Analysis and visualization of discrete fracture fracture properties under in situ conditions. We have also networks using a flow topology graph. To appear in modified our apparatus to create hydraulic fractures in IEEE T Vis Comput Gr.. addition to shear fractures. In FY16, we plan a systematic Bazant, Z., M. Salviato, V. T. Chau, H. S. Viswanathan, and study of fracture-permeability in shale using these newly A. Zubelewicz. Why fracking works. 2014. Journal of developed capabilities. Applied Mechanics. 81 (10): 101010. Task 2: We will continue to develop models of fracture Birdsell, D., H. Rajaram, D. Dempsey, and H. Viswanathan. propagation modeling work that are validated by task 1. Hydraulic fracturing fluid migration in the subsurface: The focus of fy16 work will be to test our new 3d simula- A review and modeling results. 2015. Water Resources tion capability against experiments. We will also continue Research. : 1. to develop our integrated solid-fluid solver so that frac- Carey, J. W., H. Mori, D. Brown, and R. Pawar. ture propagation due to fluid pressure can be accurately Geomechanical behavior of caprock and cement: simulated. The end goal is quantitative simulation of the Plasticity in hydrodynamic seals. 2014. Energy hydraulic fracture experiments of Task 1. Procedia. : 5671. Task 3: We will continue to conduct microfluidic experi- Carey, J. W., Z. Lei, E. Rougier, H. Mori, and H. S. ments of sweep efficiency. We have succeeded in conduct- Viswanathan. Fracture-permeability behavior of shale. 2015. Journal of Unconventional Oil and Gas ing microfluidic experiments with real rock (e.g. shale and Resources. : 27. cement) at high pressure. In the next FY, we will demon- strate how matrix-fracture interactions change simplistic Chen, L., J. D. Hyman, L. Zhou, T. Min, Q. Kang, E. Rougier, models of hydrocarbon extraction using these new micro- and H. Viswanathan. Effect of fracture density on fluidic capabilities and comparing them to standard glass effective permeability of matrix-fracture system in 26

Page 27

shale formations. AGU Books. Hyman, J. D., S. L. Painter, H. S. Viswanathan, N. Makedonska, and S. Karra. Influence of Injection Mode Chen, L., L. Zhang, Q. Kang, H. Viswanathan, J. Yao, and in Kilometer Scale Three Dimensional Discrete Fracture W. Tao. Nanoscale simulation of shale transport Networks. 2015. Water Resources Research. : 1. properties using the lattice Boltzmann method: permeability and diffusivity. 2015. Scientific Reports. : Jackson, R. B., A. Vengosh, J. W. Carey, R. Davies, F.

1. O’Sullivan, and G. Petron. The environmental costs and benefits of fracking. 2014. Annual Review of Chen, L., Q. Kang, B. Carey, and W. Tao. Pore-scale study Environment and Resources. : 1. of diffusion–reaction processes involving dissolution and precipitation using the lattice Boltzmann method. Jimenez-Martinez, J., M. L. Porter, J. D. Hyman, J. W. Carey,
2. International Journal of Heat and Mass Transfer. and H. S. Viswanathan. Mixing in a three-phase system: 75: 483. Enhanced production of oil-wet reservoirs by CO2 injection. 2016. Geophysical Research Letters.. : 1. Chen, L., Q. Kang, H. Deng, J. W. Carey, and W. Tao. Mesoscopic study of the formation of pseudomorphs Kang, Q., L. Chen, A. J. Valocchi, and H. S. Viswanathan. with presence of chemical fluids. 2014. Geosciences Pore-scale study of dissolution-induced changes in Journal. : 1. permeability and porosity of porous media. 2014. Journal of Hydrology. 517: 1049. Chen, L., Q. Kang, H. S. Viswanathan, and W. Tao. Pore- scale study of dissolution-induced changes in Karra, S., N. Makedonska, H. S. Viswanathan, S. Painter, hydrologic properties of rocks with binary minerals. and J. Hyman. Effect of advective flow in fractures
3. Water Resources Research. (50): 1. and matrix diffusion on natural gas production. 2015. Water Resources Research. : 1. Chen, L., Q. Kang, Y. Mu, Y. He, and W. Tao. A critical review of the pseudopotential multiphase lattice Boltzmann Kelkar, S., K. Lewis, S. Karra, G. Zyvoloski, S. Rapaka, H. S. model: Methods and applications. 2014. International Viswanathan, P. K. Mishra, S. Chu, D. Coblentz, and Journal of Heat and Mass Transfer. 76: 210. R. Pawar. A simulator for modeling coupled thermo- hydro-mechanical processes in subsurface geological Chen, L., Q. Kang, Z. Dai, H. S. Viswanathan, and W. Tao. media. 2014. International Journal of Rock Mechanics Permeability prediction of shale matrix reconstructed and Mining Sciences. : 569. using the elementary building block model. 2015. Fuel. : 346. Lei, Z., E. Rougier, E. E. Knight, and A. Munjiza. A framework for grand scale parallelization of the Chen, L., W. Fang, Q. Kang, J. Hyman, H. S. Viswanathan, combined finite discrete element method in 2D. 2014. and W. Tao. A generalized lattice Boltzmann model for Computational Particle Mechanics. : 307. flow through tight porous media with Klinkenberg’s effect. 2015. Physical Review E. : 1. Lei, Z., E. Rougier, E. Knight, A. Munjiza, and H. Viswanathan. A generalized anisotropic deformation Chen, Y., Q. Kang, Q. Cai, M. Wang, and D. Zhang. formulation for geomaterials. 2015. Computational Lattice Boltzmann Simulation of Particle Motion in Particle Mechanics. : 1. Binary Immiscible Fluids. 2015. Communications in Computational Physics. : 757. Lei, Z., E. Rougier, E. Knight, L. Frash, J. W. Carey, and H. Viswanathan. A non-locking composite tetrahedron Frash, L., J. W. Carey, E. Rougier, Z. Lei, and H. Viswanathan. element for the combined finite discrete element Method for in-situ tensile-hydraulic-fracturing in method. To appear in Engineering Computations.. triaxial-coreflood systems. International Journal of Rock Mechanics and Mining Sciences.. Liu, H., A. J. Valocchi, C. Werth, Q. Kang, and M. Oostrom. Pore-scale simulation of liquid CO< sub> 2</sub> Hyman, J. D., A. Guadagnini, and C. L. Winter. Statistical displacement of water using a two-phase lattice Scaling of Geometric Characteristics in Stochastically Boltzmann model. 2014. Advances in Water Resources. Generated Pore Microstructures. 2015. Computational 73: 144. Geosciences. : 1. Liu, H., Q. Kang, C. R. Leonardi, B. D. Jones, S. Schmieschek, Hyman, J. D., S. Karra, N. Makedonska, C. W. Gable, S. L. A. Narváez, J. R. Williams, A. J. Valocchi, and J. Harting. Painter, and H. S. Viswanathan. dfnWorks: A discrete Multiphase lattice Boltzmann simulations for porous fracture network framework for modeling subsurface media applications—a review. 2015. Computational flow and transport. 2015. Computers and Geosciences. Geosciences. : 1. 84: 10. Makedonska, N., S. L. Painter, Q. M. Bui, C. W. Gable, and 27

Page 28

S. Karra. Particle tracking approach for transport in Science China Technological Sciences. : 1375. three-dimensional discrete fracture networks. 2015. Computational Geoscience. : 1. Zhang, L., Q. Kang, L. Chen, and J. Yao. Simulation of flow in multi-scale porous media using the lattice Middleton, R. S., J. S. Levine, J. L. Bielicki, H. S. Boltzmann method on quadtree grids. To appear in Viswanathan, J. W. Carey, and P. Stauffer. Jumpstarting Communications in Computational Physics.. commercial-scale CO2 capture and storage with ethylene production and enhanced oil recovery in Zubelewicz, A., E. Rougier, M. Ostoja-Starewski, E. the US Gulf. 2015. Greenhouse Gases: Science and E. Knight, C. Bradley, and H. S. Viswanathan. A Technology. : 241. mechanisms-based model for dynamic behavior and fracture of geomaterials. 2014. International Journal of Middleton, R. S., J. W. Carey, R. Currier, J. D. Hyman, Q. Rock Mechanics and Mining Sciences. : 277. Kang, J. Jimenez-Martinez, M. L. Porter, and H. S. Viswanathan. Shale gas and non-aqueous fracturing fluids: opportunities and challenges for supercritical CO2. 2015. Applied Energy. : 500. Middleton, R., H. Viswanathan, R. Currier, and R. Gupta. CO2 as a fracturing fluid. 2015. Energy Procedia. : 7780. O’Malley, D., S. Karra, R. P. Currier, N. Makedonska, J. D. Hyman, and H. Viswanathan. Where does water go during hydraulic fracturing?. 2015. Groundwater. : 1. Oostrom, M., Y. Mehmani, P. Romero-Gomez, Y. Tang, H. Liu, H. Yoon, Q. Kang, V. Joekar-Niasar, M. Balhoff, and T. Dewers. Pore-scale and continuum simulations of solute transport micromodel benchmark experiments. 2014. Computational Geosciences. 4-0: 1. Padrino, J. C., B. VanderHeyden, X. Ma, and D. Z. Zhang. A separate phase drag model and a surrogate approximation for simulation of the steam assisted gravity grainage process. To appear in SPE. Porter, M., J. Jimenez-Martinez, R. Martinez, Q. McCulloch, J. W. Carey, and H. Viswanathan. Geo-material microfluidics at reservoir conditions for subsurface energy resource applications. 2015. Lab on a Chip. : 1. Summerscales, O. T., B. L. Scott, H. S. Viswanathan, and A. D. Sutton. Synthesis and reactivity of cis-FeH2(dcpe)2 (dcpe = 1,2-bis(dicyclohexylphosphino)ethane). 2016. Inorganic Chemistry Communications. 63: 57. Viswanathan, H., J. D. Hyman, S. Karra, J. W. Carey, M. L. Porter, E. Rougier, R. P. Currier, Q. Kang, L. Zhou, J. Jimenenz, N. Makedonska, L. Chen, and R. S. Middleton. Using discovery science to increase the efficiency of hydraulic fracturing while reducing water usage. 2015. In Hydraulic Fracturing: Environmental Issues, ACS Symposium Series 1216.. Edited by Drogos, D.. Vol. 1216, p. 71. Laramie, Wyoming: Oxford University Press. Zhang, L., Q. Kang, J. Yao, Y. Gao, Z. Sun, H. Liu, and A. Valocchi. Pore scale simulation of liquid and gas two- phase flow based on digital core technology. 2015. 28

Page 29

Complex Natural & Engineered Systems Directed Research Continuing Project Combating Antibiotic Resistance: Targeting Efflux Pump Systems at Multiple Scales Sandrasegaram Gnanakaran 20140121DR Introduction ing gained in this project should lead to novel strategies Multi-drug resistant bacteria such as Staphylococcus of antibiotic therapy. Because we chose to focus on a aureus and Mycobacterium tuberculosis are emerging biothreat agent (which is also of global public health at an alarming rate, yet new antibiotics are few and far concern), Burkholderia pseudomallei, this project will be between. The scarcity of new antibiotics is in part due to of interest to the biodefense mission of DOE. The multi- a lack of understanding of the mechanisms of multi-drug scale nature of this work, coupling genetic, regulatory, resistance. Antibiotic efflux is one of the most important and biomolecular structure / function aspects in one mechanisms of bacterial multi-drug resistance. Antibiot- study will contribute to advances in basic energy science ics are pumped out of the cell by efflux pumps contain- (biofuels) and basic biomedical research. ing three protein components. We propose to study the highly active efflux pumps in Burkholderia pseudomal- Progress lei, a high-priority bio-threat agent. X-ray crystallography and molecular dynamics simulation will be valuable for Structural and mechanistic studies of efflux pumps. We understanding how the efflux pump protein complexes have obtained synthetic genes for 9 efflux pump proteins are formed spanning the inner-membrane, periplasm, from Burkholderia pseudomallei for structure determi- and outer-membrane, and how the opening-closure nation. We have purified 5 proteins to high purity and dynamics of the complex determines the rate of efflux begun crystallization trials for three proteins. Crystals of of an antibiotic. Gene and protein expression studies will one outer membrane protein have been obtained. Dur- be essential to identify how the presence of an antibiotic ing the next year we plan to optimize the crystallization turns the gene circuits on or off to regulate the expres- conditions for the outer membrane protein for which we sion of the active efflux pumps, thereby determining the already have crystals and to determine the structure of net efflux. The same efflux pumps are also responsible this protein, we plan to carry out extensive screening of for releasing small metabolites, such as quorum-sensing crystallization conditions for and attempt to determine molecules, which act as global regulators of genes the structures of the other two proteins for which we belonging to pathways critical to bacterial growth, vi- have highly purified protein, and we plan to obtain pure ability, and morphology. Therefore, gene expression and protein for an additional two proteins. pathways analyses are essential to determine how the release of these metabolites offers additional fitness to To date, the high-resolution structure of an entire efflux the bacterium in terms of enhanced growth and viability. pump complex has remained elusive. Our efforts have Integrated modeling of structural, genetic, and cellular focused on filling this gap by combining data from vari- processes will provide valuable insight into how these ous experimental techniques with the available crystal processes are coupled and how this coupling imparts structure and cryo-EM data to develop computational overall bacterial fitness under the selection pressure methods to construct atomic-level resolution models from antibiotics. of the full efflux pump complex. We have used these models to probe the mechanistic details of drug efflux. Benefit to National Security Missions Future studies will integrate these molecular level find- Countermeasure development for treating pathogen in- ings into a mechanistic kinetic model of antibiotic efflux. fection is central to our project, as efflux pumps are the predominant form of multi-drug efflux, and understand- Gene expression, biochemical, and cellular analysis of 29

Page 30

efflux pumps. We first studied the effect of a P. aeruginosa The plan going forward is to develop further the mechanis- efflux pump, mexAB-oprM, on biofilm formation and re- tic model, to integrate further with cell states and struc- lated virulence especially in the presence of ciprofloxacin. ture. The model development will focus on incorporating We performed cellular assays to show that mexAB-oprM new antibiotic accumulation data into the P. aeruginosa facilitates biofilm formation thereby rendering P. aerugino- model, characterizing kinetic parameters of efflux for mu- sa impermeable to ciprofolaxin and increasing its survival tants, and extending the model to B. thailandensis. We will in the presence of the drug. Second, we performed gene also initiate collaborations to model antibiotic uptake data. expression by qPCR to identify the genes responsible for The integration will focus on iterating with the cellular biofilm formation and virulence. Finally, we demonstrated task to refine the cell state model, adding known feedback by neutrophil lysis assay that efflux-mediated expression loops, and coupling to the efflux kinetics predicted by the of the virulence factors abrogated host innate immune de- structural task. fense in the presence of ciprofloxacin, which is an addition- al resistance. These studies revealed that mexAB-oprM Future Work is more than a pump. It offers additional resistance by During the next year, structural biology studies will op- facilitating biofilm formation. We also performed genome- timize the crystallization conditions for the outer mem- wide expression studies on P. aeruginosa in the planktonic brane protein for which we already have crystals and to state by RNA-seq. Preliminary analysis revealed that the determine the structure of this protein, we plan to carry removal of mexAB-oprM induces an adaptive mechanism out extensive screening of crystallization conditions for in the presence of ciprofloxacin, which causes DNA dam- and attempt to determine the structures of the other two age leading to an SOS repair response, which is again proteins for which we have highly purified protein, and we associated with the release of pyocin and phage proteins plan to obtain pure protein for an additional two proteins. that are able to protect the bacteria. Future theoretical studies will integrate molecular level findings from molecular dynamics simulations study into a Future studies will determine the effect of the B. thali- mechanistic kinetic model of antibiotic efflux. andensis efflux pumps in developing resistance against ciprofloxacin both in the planktonic and biofilm stages Future genetic and cellular studies will determine the by combining gene expression, biochemical, and cellular effect of the B. thaliandensis efflux pumps in developing studies. We will apply a reporter assay to measure the resistance against ciprofloxacin both in the planktonic and release of quorum sensing molecules and correlate biofilm biofilm stages by combining gene expression, biochemi- formation vis-a-vis drug resistance. We will initiate collabo- cal, and cellular studies. We will apply a reporter assay to ration to carry out gene expression and functional studies measure the release of quorum sensing molecules and on BSL-2/3 B. pseudomallei. Finally, we will perform gene correlate biofilm formation vis-a-vis drug resistance. We expression and cellular studies to determine the effect of will initiate collaboration to carry out gene expression and abrogating host innate immune defense. functional studies on BSL-2/3 B. pseudomallei. Finally, we will perform gene expression and cellular studies to deter- Integrated mathematical models of efflux pump medi- mine the effect of abrogating host innate immune defense. ated drug resistance. A minimal model of P. aeruginosa response to cipro has been developed. The level of detail The plan going forward for the integrated mathematical of this model is precisely tuned to capture quantitatively framework is to develop further the mechanistic model, to the pulsed behavior seen in the measurements. The model integrate further with cell states and structure. The model was used to predict membrane permeability and efflux development will focus on incorporating new antibiotic ac- from the experimental antibiotic accumulation data. Sensi- cumulation data into the P. aeruginosa model, characteriz- tivity analysis was performed, indicating all parameters of ing kinetic parameters of efflux for mutants, and extending the model contribute substantially to the behavior. A more the model to B. thailandensis. We will also initiate collabo- detailed mechanistic model was developed, which indi- rations to model antibiotic update data. The integration cates the pulsed behavior can be explained by a delayed will focus on iterating with the cellular task to refine the influx of antibiotic across the inner membrane plus a cyto- cell state model, adding known feedback loops, and cou- plasmic regulation of efflux. Preliminary integration of the pling to the efflux kinetics predicted by the structural task. model with a model of cellular states was performed. The integrated model indicates that biofilm and efflux synergis- Note: Scope of work has changed due to reduced budget. tically increase resistance, and that the synergy increases at high membrane permeability. 30

Page 31

Conclusion We will characterize three coupled processes: expulsion of antibiotics from the cell by efflux pumps, stimulation of ef- flux pump production by antibiotics, and efflux pump-me- diated biofilm formation that shields the bacterium from antibiotics. We will combine experimental and theoretical approaches to develop the first predictive model of this system. This model will integrate these molecular, genetic, and cellular processes and will identify critical elements of each process that can inactivate drug resistance. The capability for integrated modeling of a complex adaptive molecular transport system will have broad applications in biosecurity, biofuel production, and microbial clearance of toxic material. Publications Ganguly, K., M. Wren, J. Phillips, P. Pardington, H. Schweizer, M. Wall, S. Gnanakaran , and G. Gupta. Interdependence of multi drug efflux pumps and quorum sensing system in Pseudomonas. 2014. In International Congress on Infectious Diseases. (Cape Town, South Africa, 2-6 April). Vol. 21, Suppl 1, Edition, p. 92. International Journal on Infectious Diseases: —. Hung, L. W., H. B. Kim, S. Murakami, G. Gupta, C. Y. Kim, and T. C. Terwilliger. Crystal structure of AcrB complexed with linezolid at 3.5 Å resolution. . 2014. Journal of Structural and Functional Genomics. . 14 (2): 71. Phillips, J. L., Ganguly, Wren, Gupta, M. E. Wall, and Gnanakaran. Systems Level Study of Bacterial Multi-Drug Resistance from Efflux Machinery. 2014. BIOPHYSICAL JOURNAL. 106 (2): 791A. Phillips, J., and S. Gnanakaran. A data-driven approach to modeling the tripartite structure of multidrug resistance efflux pumps. . 2015. Proteins: Structure, Function, and Bioinformatics.. 83: 43. Zgurskaya, H., S. Gnanakaran, and C. Lopez. The permeability barrier of Gram-negative cell envelopes and approaches to bypass it. To appear in ACS Infectious Diseases. 31

Page 32

Complex Natural & Engineered Systems Directed Research Continuing Project Quantitative Biology: From Molecules to Cellular Function Angel E. Garcia 20140566DR Introduction Progress Biology is continuously transforming from a qualita- Biochemical conversion of cellulosic biomass to biofuels tive science involving observations and associations to can mitigate the negative environmental impacts of the one where quantitative models are necessary for both burning of fossil fuels. Thermochemical pretreatments fundamental understanding and predictive capability. improve enzyme accessibility to embedded cellulose These theoretical and computational models build on fibers which are then hydrolyzed by enzymes to glucose methods and concepts in physics, mathematics, and and converted to biofuels by fermentation or catalysis. chemistry and are becoming increasingly influential The efficiency of these processes is key to the economic in many areas of biology. LANL has a long tradition in viability of cellulosic biofuels and, although several theoretical biology dating back to the 1970s and more physicochemical aspects of cellulose are readily captured recently CNLS has helped establish an international pres- by different experimental approaches, other properties ence in quantitative biology through the q-bio summer remain elusive and poorly understood. Computational school and annual conference and through its continu- methods offer a complementary approach using, for ing emphasis on applying quantitative methodology to example, Molecular dynamics (MD) have provided high- challenging biological problems ranging from the struc- resolution data regarding structural, thermodynamic, ture and dynamics of individual macromolecules, e.g., and mechanical properties of model crystalline cellulose DNA, RNA, proteins, etc., to the interacting biochemical fibers and surfaces. Recent computational studies focus- reactions that drive cellular function. Our research will ing on cellulose have provided valuable insights into the build bridges at the mesoscale between macromolecular construction of coarse-grained (CG) models from MD scales and complex cellular function by taking advantage simulations. In principle, CG models preserve the physi- of coarse graining (averaging) over degrees of freedom cal and chemical properties of a determined molecular at microscales that retain the impact on large-scale (cel- system after averaging key aspects of the atomistic de- lular) activity. By taking advantage of the quantitative tails. Different CG approaches for the simulation of col- expertise in CNLS and at LANL, we continue to empha- loidal states are available which allow us access to large- size opportunities with applications to infectious disease scale biomass systems over extended simulation times. mitigation and bio-surveillance, to developing effective In this study, we introduced a CG model for crystalline algal and cellulosic biofuels, and to develop fundamental cellulose whose parametrization rules are consistent understanding of cellular function over the scales from with one modeling framework. We tested the efficacy molecules to cells. and versatility of our model on two important cellulosic systems relevant to the biofuel/bioenergy research Benefit to National Security Missions field. This study is part of a general effort to extend the implementation of the coarse-graining approach to the The fundamental discovery science associated with the modeling of large crystalline bio systems. CNLS biology effort has potential applications in energy security via algal and cellulosic biofuels modeling, in Bacterial pathogens are developing increasing resistance biosecurity through understanding and modeling of in- to antibiotic treatment, motivating a search for next- fectious disease epidemiology and cellular mechanisms, generation approaches for treating bacterial infections. and in health related areas such as the ability to design An Edisonian approach is inadequate implying that a de- effective vaccines using biological models of cellular tailed understanding of the molecular basis for antibiotic function. resistance is needed. Investigations have identified a pri- 32

Page 33

mary resistance mechanism in bacteria: the existence and soluble second messenger protein. We find that inclusion operation of efflux pumps, three part protein complexes of a positive feedback loop gives rise to a bistable switch, that bridge the inner and outer membranes of many bacte- ensuring robust responses to stimulation above a thresh- ria. We developed a fast consistent method for construct- old level. ing three-part structures that uses existing data-driven models and provides molecular modeling approaches for Future Work constructing the structure of efflux pumps. We discov- Our innovative approach to improve cellulose hydrolysis ered that conformational changes in the inner membrane relies on using liquid ammonia to catalyze a structural component responsible for drug movement have limited crossover between different forms of cellulose where the impact on conformations of other pump components, resultant structure enhances enzymatic hydrolysis rates by and that two distinct models derived from conflicting a factor of five. We will provide a theoretical rationale for experimental data are consistent with currently available the ongoing work on designer cellulases, improve current measurements. These simulations suggest mechanisms for kinetic models for enzyme-cellulose interactions, and de- both smaller and larger drugs entering the pump, leading velop new predictive models based on the ongoing experi- to a better understanding of the role of the efflux pumps in mental studies. Our results will help improve the efficiency antibiotic drug resistance. of converting cellulose to biofuels. Mitochondria undergo constant remodeling through HIV evolves rapidly and can escape antibodies produced by regulation of two opposing processes, mitochondrial fis- the immune system. One prevention strategy is to develop sion and fusion. Although some aspects controlling these a vaccination schedule that elicits broad antibody response processes have been identified, the role of mitochondrial that is effective against multiple HIV strains. By building morphology is not determined. We investigated how mor- a model of within-host dynamics that captures the emer- phological features extracted from time-lapse live-cell im- gence of strain-specific antibody response, we will investi- ages of mitochondria could be used to predict mitochon- gate whether sequential vaccination or single multi-strain drial evolution. A random model was trained to classify vaccination induces stronger and/or broader response. We mitochondria poised for either fission or fusion based on a will also address the evolution of drug resistance in HCV. series of morphological and positional features. Mitochon- drial perimeter positively correlated with a mitochondrial Cellular function is built on molecular motion at the small- fission event. Similarly mitochondrial solidity (compact est scale where full-atom molecular dynamics is an effec- shape) positively correlated with mitochondria fusion. Our tive tool for characterizing and understanding the physics results indicate that fission and fusion are can be predicted and biology of those dynamics. Full simulation of even based upon mitochondrial morphological features, imply- molecular complexes is, however, far beyond current com- ing that mitochondrial fission and fusion may be influenced putational capabilities. Thus, one needs a scheme for sav- by mechanical properties of mitochondrial membranes. ing the important degrees of freedom while averaging over other less important motions, a process known as coarse Antigen receptors play a central role in adaptive immune graining. We will develop effective coarse graining strate- responses. Although the molecular networks associated gies for modeling macromolecular complexes in bacteria with these receptors have been extensively studied, we and introduce advances from computer science to attain lack a systems-level understanding of how combinations more efficient simulations that will provide new insights of interactions and modifications are regulated during into mesoscale cellular function. signaling. Thus, we need to piece together information about individual molecular mechanisms to form large- Conclusion scale computational models of signaling networks. To this end, we developed an interaction library for signaling by a We will combine experimental probes with simulation/ high-affinity receptor. The library, consisting of executable modeling to characterize biomolecular environments that rules for protein–protein and protein–lipid interactions, will allow evaluation of potential drug design. We will extends earlier models for receptor signaling and intro- develop mesoscopic models that bridge scales in biofuel- duces new interactions not previously considered. This relevant algae systems and provide a molecular level interaction library is a toolkit with which existing models understanding and design principles for engineering more can be expanded and from which new models can be built. effective cellulosic biomass. We will develop multiscale As an example, we present models of branching pathways models of pathogenesis in bacteria. We will use novel from an adaptor protein, which influence production of mathematical methods to improve rule-based methodol- a phospholipid protein at the plasma membrane and a ogy applied to bio-chemical reaction networks and em- 33

Page 34

ploy them to understand and characterize specific cellular Matthews, Clark, Hurst, Mithen, R. G. Bristow, P. mechanisms. We will construct high-level models of infec- C. Boutros, Fraser, Cooke, Raine, Jones, Menzies, tious diseases including HIV and HCV based on systems- Stebbings, J. o. n. Hinton, J. o. n. Teague, McLaren, Mudie, Hardy, Anderson, Joseph, Goody, B. e. n. level understanding of immune response. Robinson, Maddison, Gamble, Greenman, D. a. n. Berney, Hazell, Livni, Fisher, Ogden, Kumar, Thompson, Publications Woodhouse, Nicol, Mayer, T. i. m. Dudderidge, N. Abbink, , L. F. Maxfield, Ng’ang’a, E. N. Borducchi, M. C. Shah, Gnanapragasam, Voet, Campbell, Futreal, J. Iampietro, C. A. Bricault, J. E. Teigler, Blackmore, Easton, A. Y. Warren, C. S. Foster, M. R. Stratton, Parenteau, Wagh, S. A. Handley, Zhao, H. W. Virgin, H. C. Whitaker, McDermott, D. S. Brewer, and D. E. Korber, and D. H. Barouch. Construction and Evaluation Neal. Analysis of the genetic phylogeny of multifocal of Novel Rhesus Monkey Adenovirus Vaccine Vectors. prostate cancer identifies multiple independent clonal 2015. JOURNAL OF VIROLOGY. 89 (3): 1512. expansions in neoplastic and morphologically normal prostate tissue (vol 47, pg 367, 2015). 2015. NATURE Chylek, L. A., Akimov, Dengjel, K. T. G. Rigbolt, B. i. n. Hu, GENETICS. 47 (6): 689. W. S. Hlavacek, and Blagoev. Phosphorylation Site Dynamics of Early T-cell Receptor Signaling. 2014. PLOS Cooper, C. S., Eeles, D. C. Wedge, Van Loo, Gundem, L. B. ONE. 9 (8). Alexandrov, Kremeyer, Butler, A. G. Lynch, Camacho, C. E. Massie, Kay, H. J. Luxton, Edwards, Kote-Jarai, Chylek, L. A., B. S. Wilson, and W. S. Hlavacek. Modeling Dennis, S. u. e. Merson, Leongamornlert, Zamora, Biomolecular Site Dynamics in Immunoreceptor Corbishley, Thomas, Nik-Zainal, O’Meara, Matthews, Signaling Systems. 2014. SYSTEMS BIOLOGY APPROACH Clark, Hurst, Mithen, R. G. Bristow, P. C. Boutros, TO BLOOD. 844: 245. Fraser, Cooke, Raine, Jones, Menzies, Stebbings, J. o. n. Hinton, J. o. n. Teague, McLaren, Mudie, Chylek, L. A., D. A. Holowka, B. A. Baird, and W. S. Hlavacek. Hardy, Anderson, Joseph, Goody, B. e. n. Robinson, An Interaction Library for the FcεRI Signaling Network. Maddison, Gamble, Greenman, D. a. n. Berney, Hazell, 2014. Frontiers in Immunology. 5: 172. Livni, Fisher, Ogden, Kumar, Thompson, Woodhouse, Chylek, L. A., D. A. Holowka, B. A. Baird, and W. S. Hlavacek. Nicol, Mayer, T. i. m. Dudderidge, N. C. Shah, An interaction library for Vie Fc epsilon RI signaling Gnanapragasam, Voet, Campbell, Futreal, Easton, A. network. 2014. FRONTIERS IN IMMUNOLOGY. 5. Y. Warren, C. S. Foster, M. R. Stratton, H. C. Whitaker, McDermott, D. S. Brewer, and D. E. Neal. Analysis of Chylek, L. A., L. A. Harris, C. Tung, J. R. Faeder, C. F. the genetic phylogeny of multifocal prostate cancer Lopez, and W. S. Hlavacek. Rule-based modeling: a identifies multiple independent clonal expansions in computational approach for studying biomolecular neoplastic and morphologically normal prostate tissue. site dynamics in cell signaling systems. 2014. Wiley 2015. NATURE GENETICS. 47 (4): 367. Interdisciplinary Reviews: Systems Biology and Medicine. 6 (1): 13. Giorgi, E. E., D. O. Stram, D. Taverna, S. D. Turner, F. Schumacher, C. A. Haiman, A. Lum-Jones, M. Chylek, L. A., L. A. Harris, J. R. Faeder, and W. S. Hlavacek. Tirikainen, C. Caberto, D. Duggan, B. E. Henderson, Modeling for (physical) biologists: an introduction to L. Le Marchand, and I. Cheng. Fine-Mapping IGF1 the rule-based approach. 2015. PHYSICAL BIOLOGY. 12 and Prostate Cancer Risk in African Americans: (4). The Multiethnic Cohort Study. 2014. Cancer Epidemiology, Biomarkers & Prevention: A Publication Chylek, L. A., L. A. Harris, Tung, J. R. Faeder, C. F. Lopez, and of the American Association for Cancer Research, W. S. Hlavacek. Rule-based modeling: a computational Cosponsored by the American Society of Preventive approach for studying biomolecular site dynamics in Oncology. 23 (9): 1928. cell signaling systems. 2014. WILEY INTERDISCIPLINARY REVIEWS-SYSTEMS BIOLOGY AND MEDICINE. 6 (1): 13. Gottardo, R., R. T. Baller, B. T. Korber, S. Gnanakaran, J. Phillips, X. Shen, G. D. Tomaras, E. Turk, G. Imholte, Chylek, L. A., V. Akimov, J. Dengiel, K. T. G. Rigbolt, B. Hu, L. Eckler, H. Wenschuh, J. Zerweck, K. Greene, H. W. S. Hlavacek, and B. Blagoev. Phosphorylation site Gao, P. W. Berman, D. Francis, F. Sinangil, C. Lee, S. dynamics of early T-cell receptor signaling. 2014. PLoS Nitayaphan, S. Rerks-Ngarm, J. Kaewjungawal, P. One. 9: e104240. Pitsuttithum, J. Tartaglia, M. L. Robb, N. L. Michael, J. H. Kim, S. Zolla-Pazner, B. F. Haynes, J. R. Mascola, Cooper, C. S., Eeles, D. C. Wedge, Van Loo, Gundem, L. B. S. Self, P. Gilbert, and D. C. Montefiori. Plasma IgG Alexandrov, Kremeyer, Butler, A. G. Lynch, Camacho, to linear epitopes in the V2 and V3 regions of HIV-1 C. E. Massie, Kay, H. J. Lmcton, Edwards, Kote-Jarai, gp120 correlate with a reduced risk of infection in the Dennis, S. u. e. Merson, Leongamornlert, Zamora, RV144 vaccine efficacy trial. 2013. PLoS One. : e75665. Corbishley, Thomas, Nik-Zainal, Ramakrishna, O’Meara, 34

Page 35

Gundem, , Van Loo, Kremeyer, L. B. Alexandrov, J. M. C. B. E. Dale, S. J. Marrink, and Gnanakaran. MARTINI Tubio, Papaemmanuil, D. S. Brewer, H. M. L. Kallio, Coarse-Grained Model for Crystalline Cellulose Hoegnas, Annala, Kivinummi, Goody, Latimer, Microfibers. 2015. JOURNAL OF PHYSICAL CHEMISTRY O’Meara, K. J. Dawson, Isaacs, M. R. Emmert-Buck, B. 119 (2): 465. Nykter, Foster, Kote-Jarai, Easton, H. C. Whitaker, D. E. Neal, C. S. Cooper, R. A. Eeles, Visakorpi, P. J. Lopez, C. A., Sethi, Goldstein, B. S. Wilson, and Campbell, McDermott, D. C. Wedge, and G. S. Bova. Gnanakaran. Membrane-Mediated Regulation of the The evolutionary history of lethal metastatic prostate Intrinsically Disordered CD3 epsilon Cytoplasmic Tail of cancer. 2015. NATURE. 520 (7547): 353. the TCR. 2015. BIOPHYSICAL JOURNAL. 108 (10): 2481. Hong, M. K. H., Macintyre, D. C. Wedge, Van Loo, Lunke, Martincorena, , Roshan, Gerstung, Ellis, Van Loo, McLaren, L. B. Alexandrov, Sloggett, Cmero, Mangiola, Lonie, D. C. Wedge, Fullam, L. B. Alexandrov, J. M. Tubio, Naeem, Sapre, Phal, Chin, Kerger, J. S. Pedersen, Stebbings, Menzies, Widaa, M. R. Stratton, P. H. Jones, Ryan, Waring, Haviv, A. J. Costello, N. M. Corcoran, and P. J. Campbell. High burden and pervasive positive and C. M. Hovens. Unexpected complexity in the selection of somatic mutations in normal human skin. origins of prostate cancer metastases. 2015. BJU 2015. SCIENCE. 348 (6237): 880. INTERNATIONAL. 115: 92. Mayes, H. B., Tian, M. W. Nolte, B. H. Shanks, G. T. Hong, M. K. H., Macintyre, D. C. Wedge, Van Loo, Patel, Beckham, Gnanakaran, and L. J. Broadbelt. Sodium Lunke, L. B. Alexandrov, Sloggett, Cmero, Marass, Tsui, Ion Interactions with Aqueous Glucose: Insights Mangiola, Lonie, Naeem, Sapre, P. M. Phal, Kurganovs, from Quantum Mechanics, Molecular Dynamics, and Chin, Kerger, A. Y. Warren, Neal, Gnanapragasam, Experiment. 2014. JOURNAL OF PHYSICAL CHEMISTRY Rosenfeld, J. S. Pedersen, Ryan, Haviv, A. J. Costello, N. B. 118 (8): 1990. M. Corcoran, and C. M. Hovens. Tracking the origins Nemenman, , J. R. Faeder, Gnanakaran, W. S. Hlavacek, and drivers of subclonal metastatic expansion in Munsky, M. E. Wall, and Y. i. Jiang. The Seventh q-bio prostate cancer. 2015. NATURE COMMUNICATIONS. 6. Conference: meeting report and preface. 2014. Hovens, , Hong, Macintyre, Wedge, Van Loo, Lunke, PHYSICAL BIOLOGY. 11 (4). Alexandrov, Slogget, Cmero, Mangiola, Lonie, Naeem, Phillips, J. L., and Gnanakaran. A data-driven approach Sapre, Phal, Kerger, Pedersen, Ryan, Haviv, Costello, to modeling the tripartite structure of multidrug and Corcoran. Tracking clonal diversity in metastatic resistance efflux pumps. 2015. PROTEINS-STRUCTURE prostate cancer progression. 2015. JOURNAL OF FUNCTION AND BIOINFORMATICS. 83 (1): 46. CLINICAL ONCOLOGY. 33 (7). Russell, C. A., P. M. Kasson, R. O. Donis, Riley, Dunbar, Kong, R. u. i., M. K. Louder, Wagh, R. T. Bailer, deCamp, Rambaut, Asher, Burke, C. T. Davis, R. J. Garten, Greene, Gao, J. D. Taft, Gazumyan, Liu, M. C. Gnanakaran, S. I. Hay, Herfst, N. S. Lewis, J. O. Lloyd- Nussenzweig, Korber, D. C. Montefiori, and J. R. Smith, C. A. Macken, Maurer-Stroh, Neuhaus, C. R. Mascola. Improving Neutralization Potency and Parrish, K. M. Pepin, S. S. Shepard, D. L. Smith, D. L. Breadth by Combining Broadly Reactive HIV-1 Suarez, S. C. Trock, Widdowson, D. B. George, Lipsitch, Antibodies Targeting Major Neutralization Epitopes. and J. D. Bloom. Improving pandemic influenza risk 2015. JOURNAL OF VIROLOGY. 89 (5): 2659. assessment. 2014. ELIFE. 3. Kozer, , Barua, Henderson, E. C. Nice, A. W. Burgess, W. Schulze, , Imbeaud, Letouze, L. B. Alexandrov, Calderaro, S. Hlavacek, and A. H. A. Clayton. Recruitment of Rebouissou, Couchy, Meiller, Shinde, Soysouvanh, the Adaptor Protein Grb2 to EGFR Tetramers. 2014. Calatayud, Pinyol, Pelletier, Balabaud, Laurent, Blanc, BIOCHEMISTRY. 53 (16): 2594. Mazzaferro, Calvo, Villanueva, Nault, Bioulac-Sage, Langan, , Petridis, H. M. O’Neill, S. V. Pingali, Foston, M. R. Stratton, J. M. Llovet, and Zucman-Rossi. Exome Nishiyama, Schulz, Lindner, B. L. Hanson, Harton, W. T. sequencing of hepatocellular carcinomas identifies Heller, Urban, B. R. Evans, Gnanakaran, A. J. Ragauskas, new mutational signatures and potential therapeutic J. C. Smith, and B. H. Davison. Common processes drive targets. 2015. NATURE GENETICS. 47 (5): 505. the thermochemical pretreatment of lignocellulosic Shlien, , B. B. Campbell, de Borja, L. B. Alexandrov, D. M. biomass. 2014. GREEN CHEMISTRY. 16 (1): 63. Merino, Remke, Bakry, Dirks, Huang, R. G. Grundy, Lopatina, L. M., and J. V. Selinger. Polymer-disordered Durno, Aronson, M. D. Taylor, Z. F. Pursell, C. E. liquid crystals: Susceptibility to an electric field. 2013. Pearson, Malkin, Bouffet, Hawkins, P. J. Campbell, and Physical Review E. 88: 062510. U. r. i. Tabori. COMBINED HEREDITARY AND SOMATIC MUTATIONS OF REPLICATION ERROR REPAIR GENES Lopez, C. A., Bellesia, Redondo, Langan, S. P. S. Chundawat, RESULT IN RAPID ONSET OF ULTRA-HYPERMUTATED 35

Page 36

MALIGNANT BRAIN TUMORS IN CHILDREN. 2015. NEURO-ONCOLOGY. 17: 9. Shlien, , B. B. Campbell, de Borja, L. B. Alexandrov, Merico, Wedge, Van Loo, P. S. Tarpey, Coupland, S. a. m. Behjati, Pollett, Lipman, Heidari, Deshmukh, Avitzur, Meier, Gerstung, Y. e. Hong, D. M. Merino, Ramakrishna, Remke, Arnold, G. B. Panigrahi, N. P. Thakkar, K. P. Hodel, E. E. Henninger, A. Y. Goeksenin, Bakry, G. S. Charames, Druker, Lerner-Ellis, Mistry, Dvir, Grant, Elhasid, Farah, G. P. Taylor, P. C. Nathan, Alexander, Ben-Shachar, S. C. Ling, Gallinger, Constantini, Dirks, Huang, S. W. Scherer, R. G. Grundy, Durno, Aronson, Gartner, M. S. Meyn, M. D. Taylor, Z. F. Pursell, C. E. Pearson, Malkin, P. A. Futreal, M. R. Stratton, Bouffet, Hawkins, P. J. Campbell, and U. r. i. Tabori. Combined hereditary and somatic mutations of replication error repair genes result in rapid onset of ultra-hypermutated cancers. 2015. NATURE GENETICS. 47 (3): 257. Stieh, D. J., J. L. Phillips, P. M. Rogers, D. F. King, G. C. Cianci, S. A. Jeffs, S. Gnanakaran, and R. J. Shattock. Dynamic electrophoretic fingerprinting of the HIV-1 envelope glycoprotein. 2013. Retrovirology. 10 (1): 33. Veelen, V., Matthijs, S. Luo, and B. Simon. A Simple Model of Group Selection that Cannot Be Analyzed with Inclusive Fitness. 2014. Social Science Research Network. Wagener, , L. B. Alexandrov, Montesinos-Rongen, Schlesner, Haake, H. G. Drexler, Richter, G. R. Bignell, McDermott, and Siebert. Analysis of mutational signatures in exomes from B-cell lymphoma cell lines suggest APOBEC3 family members to be involved in the pathogenesis of primary effusion lymphoma. 2015. LEUKEMIA. 29 (7): 1612. Westrate, L. M., J. A. Drocco, K. R. Martin, W. S. Hlavacek, and J. P. MacKeigan. Mitochondrial Morphological Features are Associated with Fission and Fusion Events. 2014. PLoS ONE. 9 (4): e95265. Westrate, L. M., J. A. Drocco, K. R. Martin, W. S. Hlavacek, and J. P. MacKeigan. Mitochondrial Morphological Features Are Associated with Fission and Fusion Events. 2014. PLOS ONE. 9 (4). Yates, L. R., Gerstung, Knappskog, Desmedt, Gundem, Van Loo, Aas, L. B. Alexandrov, Larsimont, Davies, Li, Y. S. Ju, Ramakrishna, H. K. Haugland, P. K. Lilleng, Nik-Zainal, McLaren, Butler, Martin, Glodzik, Menzies, Raine, Hinton, Jones, L. J. Mudie, Jiang, Vincent, Greene- Colozzi, Adnet, Fatima, Maetens, Ignatiadis, M. R. Stratton, Sotiriou, A. L. Richardson, P. E. Lonning, D. C. Wedge, and P. J. Campbell. Subclonal diversification of primary breast cancer revealed by multiregion sequencing. 2015. NATURE MEDICINE. 21 (7): 751. 36

Page 37

Complex Natural & Engineered Systems Directed Research Continuing Project SHIELDS: Space Hazards Induced Near Earth by Large Dynamic Storms - Understanding, Modeling, Predicting Vania K. Jordanova 20150033DR Introduction instruments on Van Allen Probes and geosynchronous Our society is increasingly dependent on technologies satellites will be applied for the first time to the global susceptible to harmful conditions in space, e.g., Galaxy model. 15, a $250Mtelecommunicationsatelliteingeosyn-BenefittoNationalSecurityMissionschronousorbit,failedtooperatein2010duetoaspaceOurnationaleffortsinnuclearnonproliferationdependstorm,anditsrecoveryoperationcostwas$3.5M. in many ways on our satellite sensing systems. Many of Predicting space weather hazards remains a big space the nation’s civilian and military space assets operate physics challenge due to the complex multi-scale nature in the inner magnetosphere, an extremely hazardous of the magnetosphere. This project will address this region of space causing satellite failures and anomalies. challenge using a cross-Laboratory team, LANL state-of- The ability to reliably distinguish between various modes the-art models, computational facilities, and data from of failure is very important in anomaly resolution and national security payloads. forensics and may be used in decision-making exercises at the highest levels. The crucial national security need This project will develop a new capability, the SHIELDS to maximize the safety of civilian and military satellites framework, to understand, model, and predict one of requires a capability to predict dynamic conditions in the most important space weather hazards, the space- space, as called out in national studies supporting the craft Surface Charging Environment (SCE), i.e. the hot National Space Weather Program by DOE, DOD, NOAA, (less than few 100 keV) electron fluxes whose enhance- NASA, and NSF. This is especially needed during solar ment can lead to satellite failures during disturbed maximum (in ~2015) with its expected high storm occur- conditions. This project aims at addressing two mecha- rence. nisms that may be crucial for the increase of the SCE, but are not included in any existing global model: a) Our SHIELDS framework will address the ubiquitous particle injection during geomagnetic substorms (mag- threats posed by surface charging and will specify the netospheric reconfiguration events), and b) scattering by fast-changing Surface Charging Environment (SCE) which plasma waves. Substorms may contribute significantly to represents a seed population for the harmful radiation the plasma pressure increase, causing sudden enhance- belts. SHIELDS utilizes unique data from DOE/LANL geo- ments of the ring current (keV) fluxes. synchronous spacecraft, which are operational national security payloads that additionally provide scientific data Plasma waves redistribute energy throughout the col- at critical locations in the magnetosphere. SHIELDS is lisionless magnetospheric environment; wave-particle directly relevant to the Lab’s Global Security Space Situ- interactions can serve as a dominant mechanism for ational Awareness (SSA) and Energy Security missions energy diffusion and particle loss. Including these pro- and the “Complex Natural and Engineered Systems” cesses requires models that are targeted at key regions/ priority area to understand key drivers of the radia- physics regimes, however, the coupling of these models tion belts (e.g., seed populations, plasma waves) and across multiple spatial and temporal scales remains an advance toward prediction of the space environment. extreme challenge. Guided by theory and observations, Being sought both nationally and internationally, such the goal of this project is to bridge macro- and micro- unique LANL capability will position us at the forefront of scopic computational models, allowing for the first time space science exploration. to study the SCE consequences on a global scale. New data assimilation techniques employing data from LANL 37

Page 38

Progress fluxes from the Van-Allen Probe B valid for the substorm The particle tracing model (PTM) has been ported to both date. The results showed a significant improvement in the C and FORTRAN 90 and the field interpolation algorithms estimation of the ring current pressure, in agreement with have been tested and updated to produce more physical observations from the Van Allen Probe A. These results results (preserving higher order spatial derivatives across have been written in a paper submitted to the Geophysical grid cells). The new code has been coupled with the global Research Letters. We have also worked on estimating ap- MHD BATS-R-US model from the Space Weather Modeling propriate spatial correlations within RAM-SCB to regularize Framework (SWMF) and we have begun tracing equatori- the covariance matrix, which is an important element for ally mirroring particles in these MHD fields. We have also the success of the assimilation. begun modifying the coupling between BATS-R-US and the inner magnetospheric RAM-SCB module. Currently, the The first requisite of efficient interfacing between the electric field is communicated to RAM-SCB from the rest of multiple disparate codes that this project involves was the the SWMF by mapping only the potential electric field out migration of existing projects to a unified yet distributed from the ionosphere into the RAM-SCB domain. While this version control system. This was achieved in part through accounts for much of the convective electric fields associ- the creation and integration into a GIT repository. This ated with storms, it neglects the inductive electric fields facilitated efficient contributions of our external collabora- produced during very dynamic events like substorms. The tors who subsequently brought the BATS-R-US/RAM-SCB electric field coupling has been modified so that the induc- coupled code into parity with the latest SWMF release. tive fields can be communicated directly with RAM-SCB. This parity can be maintained with little overhead moving forward, allowing SHIELDS to benefit from the latest SWMF The coupling between the MHD code BATS-R-US and the optimizations and features. The performance of various Particle-in-Cell (PIC) code iPIC3D has been completely re- components of the code has been examined and we found written. The original coupling was through flat files, which that the PIC codes are the most significant computational was simple to implement but was only intended as a tem- components. We are addressing this for the Curvilinear porary solution. Once the MHD-EPIC algorithm was proven PIC code by identifying the key computational kernels and to work, a general parallel coupler was implemented since regions with potential for shared memory parallelization; the existing SWMF couplers were not suitable for passing these options are currently being explored. large amount of data between massively parallel models. The new coupler keeps the grid description and interpola- Finally, we organized a SHIELDS kick-off workshop in Febru- tion methods private for the models. This eliminates the ary 23-25, 2015, attended by both external and internal need to describe the grid in an abstract manner and pass LANL collaborators. The workshop consisted of introduc- this information to the SWMF or the coupled model. The tory talks given by SHIELDS team members, followed by coupling between BATS-R-US and iPIC3D now works in 3D, general discussions, and provided a broad forum for inter- and the two grids don’t need to be aligned. We also allow actions on space weather topics. multiple PIC regions to be used. This new efficient and flexible coupler was used to model Ganymede’s magneto- Future Work sphere in 3D with 4 PIC regions. The results gave excellent • Expand the coupling of the particle tracing model agreement with Galileo measurements, much better than (PTM) with BATS-R-US and RAM-SCB to off-equatorially with Hall MHD, showing that the MHD-EPIC algorithm pro- mirroring particles and investigate the formation of vides a better description of reconnection than the fluid injection boundaries. Complete the electric field cou- models can. The results of this work are being prepared pling so that inductive fields are passed to RAM-SCB. for publication in the Journal of Geophysical Research. Perform testing and validation, comparing results with Van Allen Probes and geosynchronous satellite obser- We have successfully implemented a data assimilation vations. method into RAM-SCB to properly simulate particle injec- • Develop further the coupling of BATS-R-US and the tions. The assimilation method applied was the localized iPIC3D code by extending the coupling for electron ensemble Kalman filter, which uses an ensemble of model pressure and multi-ion plasma, and developing direct simulations, along with observational data, to calibrate coupling between separate iPIC3D regions within the the model. The assimilation is able to nudge the RAM- SWMF. Improve the efficiency of the coupling and ap- SCB electron flux so that it’s in better agreement with the ply it to simulate the magnetosphere of the Earth and observed flux. In particular, we successfully simulated the Mercury. effects of particle injections during an isolated substorm • Implement a novel data assimilation scheme, based that took place on July 18 2013 by assimilating electron on the ensemble Kalman filter, to better nudge the 38

Page 39

coupled RAM-SCB and BATS-R-US system into repro- Daldorff, L., G. Toth, T. I. Gombosy, G. Lapenta, J. Amaya, ducing substorm injections. We have developed this S. Markidis, and J. U. Brackbill. Two-way coupling of scheme theoretically, showing that it captures the a global Hall magnetohydrodynamics model with a local implicit Particle-in-Cell model. 2014. Journal of most relevant correlations between the models, which Computational Physics. 268: 236. is crucial for the propagation of information in the coupled system. Denton, M. H., M. F. Thomsen, V. K. Jordanova, M. G. • Incorporate wave-particle interactions in RAM-SCB Henderson, J. E. Borovsky, J. S. Denton, Pitchford, and using quasi-linear theory and event-specific diffusion D. P. Hartley. An empirical model of electron and ion coefficients. To calculate the necessary cold plasma fluxes derived from observations at geosynchronous density, couple RAM-SCB with a 2D plasmasphere orbit. 2015. SPACE WEATHER-THE INTERNATIONAL model. Perform initial calculations of transport coeffi- JOURNAL OF RESEARCH AND APPLICATIONS. 13 (4): cients with the wave PIC code using RAM-SCB input. 233. • Develop further the SHIELDS distributed version con- Godinez, H. C., Y. Yu, M. G. Henderson, B. Larsen, and V. trol system, incorporating dependency relationships K. Jordanova. Ring current pressure estimation with between the components to facilitate large scale de- RAM-SCB using data assimilation and Van Allen Probe velopment and build towards production use. Incorpo- flux data. Geophysical Research Letters. rate unit tests and begin the development of workflow tools that allow coherent execution of the many suc- Jordanova, V. K.. Global modeling of wave generation cessive components of the SHIELDS framework. processes in the inner magnetosphere. To appear in Magnetosphere-Ionosphere Coupling in the Solar • Upgrade the Curvilinear PIC code (CPIC) from single- System. Edited by Chappell, R., and R. Thorne. block grid geometry to multi-block pursuing three approaches: (1) overlapping grid patches, (2) non- Peng, I. B., S. Markidis, E. Laure, A. Johlander, A. Vaivads, overlapping grid patches, and (3) the immersed Y. Khotyaintsev, P. Henri, and G. Lapenta. Kinetic boundary method. Upgrade the particle mover and the structures of quasi-perpendicular shocks in global field solver to the new grid configuration. Evaluate the particle-in-cell simulations. 2015. Physics of Plasmas. robustness, simplicity and computational performance doi: 10.1063/1.4930212 (22): 1. of these approaches. Welling, D. T., V. K. Jordanova, Glocer, Toth, M. W. Liemohn, and D. R. Weimer. The two-way relationship between Conclusion ionospheric outflow and the ring current. 2015. Using a new, system-level approach, we will provide the JOURNAL OF GEOPHYSICAL RESEARCH-SPACE PHYSICS. transformative understanding of mechanisms driving dis- 120 (6): 4338. turbed geomagnetic conditions, critically needed for their Yu, Y., V. Jordanova, D. Welling, Larsen, S. G. Claudepierre, accurate prediction and to prevent damages to technologi- and Kletzing. The role of ring current particle cal systems in space. We will obtain, for the first time, a injections: Global simulations and Van Allen Probes realistic description of the magnetosphere, resolving short observations during 17 March 2013 storm. 2014. timescales of substorm dynamics and enabling a better Geophysical Research Letters. 41 (4): 1126. prediction of the spacecraft Surface Charging Environment (SCE). We will examine, for the first time self-consistently Yu, Y., V. K. Jordanova, S. Zou, R. Heelis, M. Ruohoniemi, on a global scale, the fundamental generation of plasma and J. Wygant. Modeling subauroral polarization waves and their feedback on particle dynamics. We expect streams (SAPS) during the March 17, 2013 storm. this to lead to new discoveries and high-impact innovative 2015. Journal of Geophysical Research - Space Physics. publications. doi:10.1002/2014JA020371 (3): 1738. Zhao, L., Y. Yu, G. Delzanno, and V. Jordanova. Bounce- and Publications MLT-averaged diffusion coefficients in a physics-based Birn, J., A. Runov, and M. Hesse. Energetic electrons magnetic field geometry obtained from RAM-SCB for in dipolarization events: Spatial properties and the March 17 2013 storm. 2015. Journal of Geophysical anisotropy. 2014. Journal of Geophysical Research - Research - Space Physics. doi:10.1002/2014JA020858 Space Physics. 119 (5): 3604. (4): 2616. Birn, J., A. Runov, and M. Hesse. Energetic ions in dipolarization events. 2015. Journal of Geophysical Research - Space Physics. doi:10.1002/2015JA021372 (120): 1. 39

Page 40

Complex Natural & Engineered Systems Directed Research Continuing Project Critical Watersheds: Climate Change, Tipping Points, and Water Security Impacts Richard S. Middleton 20150397DR Introduction framework to directly address interactions and feed- Water is critical for domestic health, wealth, and se- backs between climate, terrestrial ecosystems, wildfire, curity. Our changing climate will have a significant and and hydrology. detrimental impact on terrestrial ecosystems and subse- quent water supply for the energy-water nexus (EWN). Benefit to National Security Missions Currently, we do not understand the interactions and The project directly addresses the energy security and feedbacks between climate, forest mortality, wildfire, environment missions, with anticipated substantial and hydrology. Without this understanding, we cannot contributions to the DOE applied energy and science predict and plan for climate-induced impacts on society programs. and the EWN. Environment mission: The project focuses on climate This project develops the necessary basic science and and energy impacts, specifically exploring climate modeling capabilities to (1) predict, (2) quantify, and (3) change/extremes and water supply for the energy-water mitigate climate-related impacts on our natural ecosys- nexus (EWN). This includes all energy production/ex- tems and the EWN. The project focuses on the novel traction that requires water including coal-fired and gas concept of “critical watersheds”, watersheds that are plants, nuclear power plants, renewable energy (e.g., both critical to society and vulnerable to climate change, hydropower), and conventional and unconventional and climate-induced extreme events. The project is oil and gas production. The stretch goal of the project advancing the predictive understanding of integrated is to develop quantitative mitigation plans—including watershed hydrology and the role of disruptive events in remediation and restoration—that increase ecosystem ecosystems. The project focuses on processes that oper- resiliency (e.g., fire suppression, hydrologic stability) to ate and are manifested on multiple scales, requiring mul- changing climate and extremes. Science and modeling tiple study sites to capture the range of climate-ecosys- capabilities developed in this project directly address tem-hydrologic interactions. Our study sites range from challenges set out by the DOE-SC including integrat- small-scale ecosystem study plots and hillslopes (< 0.1 ing fire behavior, drought mortality, and hydrology into km2) and small watersheds (<2 km2) up to mesoscale standalone watershed frameworks as well as providing watersheds and the entire Colorado River basin. The key input for DOE-SC’s Accelerated Climate Model for broad project focus will is on the US Southwest though Energy (ACME). the science, tools, and understanding developed will be broadly applicable to critical domestic and global water- Energy security mission: the EWN is a key focus of the sheds. The project will use combination of existing data project, understanding how nuclear energy, fossil en- from LANL and multiple collaborators, new modeling ap- ergy, and renewable energy/power will be impacted by proaches, new model-based data analyses and synthe- climate change/extremes, particularly in critical water- ses, and new experiments. The project will uniquely po- sheds such as the Colorado River. sition LANL to predict, quantify, and mitigate the impacts of climate change on water supply to municipalities and Agency relevance: Project outcomes are designed to energy producers, with multiple potential sponsors in directly address challenges and needs set out by DOE-SC DOE, state and federal agencies, and private industry. (see above) and DOD (e.g., SERDP). The project ad- dresses applications critical to DOE-EERE (e.g., biomass The project will develop a first-of-a-kind fully-integrated 40

Page 41

and hydropower programs). Long-term, the project could stration, approximately two months ahead of anticipated position LANL for non-traditional federal agencies includ- schedule. ing USDA (food security), EPA (environmental impacts), and Department of State (geopolitically critical watersheds Task 2 involved development of the ATS framework and such as the Nile River basin). incorporating key processes critical for understanding interaction between climate, vegetation, fire, and hydrol- ogy. Specifically, we have focused on understanding the Progress role of ground cover (e.g., litter, duff layer) in both fire and The project has made significant advances in each area hydrology and have successfully integrated a unique, novel described in Year 1 planning. This includes successfully representation of groundcover into ATS. We expect this to addressing and on-time progress for accomplishments in have a transformational impact on ecohydrologic modeling each FY15 goal, task, and sub-task. and understanding. We have developed a scientific manuscript (soon-to-be- Task 3 has involved collecting and collating large volumes submitted) that details challenges and opportunities to un- of data to support our science and modeling work at derstand the interaction between climate change, extreme multiple scales and for multiple sites. This includes regional events, and climate-induced disturbances. This manuscript climate data, vegetation disturbances (e.g., infestation, directly addresses our first year goals. The challenges and historic wildfire impacts), and historic hydrology. This data opportunities in the paper also will guide key scientific collection is currently supporting ongoing research and questions and approaches in Year 2 as well as provide the serve as a critical platform for Years 2 and 3. basis for focuses discussion with collaborators and the wider community. And in the longer term (e.g., post-proj- Specific accomplishments: ect), we anticipate that the paper will serve as a scientific roadmap and lay the groundwork for a national workshop, The preliminary QUICFIRE modeling framework has already potentially funded by DOE Office of Science, that brings advanced to a stage where potential sponsors (e.g., DHS) together key researchers in this area. are preparing to send seed money to develop the technol- ogy in new direction (i.e., urban fire modeling) not part of We have developed a preliminary framework that incorpo- this project. We have been selected as primary conveners rates climate drivers, vegetation responses, offline wildfire for an upcoming targeted special session (American Geo- behavior, and hydrology. The Advanced Terrestrial Simula- physical Union, 2015) that will bring together key domestic tor (ATS) framework will undergo testing, validation, and and international researchers to better understand critical calibration in the remaining four months of the project. We watersheds in transition. We will us this opportunity as anticipate that ATS will be applied to a key small watershed an ad hoc workshop on climate and ecohydrology interac- in the Valles Caldera National Preserve (VCNP) before the tions. end of Year 1, an activity that was not anticipated until Year 2 in original scoping and planning. Science and modeling developed in Year 1 has enabled multiple proposals, within and beyond DOE, and collabora- In Task 1 we developed specific modeling and experimen- tions that are advancing the concept of critical watersheds tal research plans for Years 2 and 3 including developing and the impacts of climate change on natural ecosystems research access and collaboration with the VCNP and and water for the energy-water nexus. identifying an illustrative watershed (Jaramillo River), ac- cess to multiple field sites and data with the University of The research team presented multiple talks (5-10) at New Mexico (UNM), New Mexico State University (NMSU), several national and international conferences. Multiple Oregon State University (OSU), and Bandelier National manuscripts are published, in review, and preparation Monument (BNM). We have developed a new approach to highlighting new science developed in first eight months of integrating dynamic vegetation responses and succession the project, including a significant “roadmapping” manu- into the ATS framework and, ahead of schedule, expect script outlining the key scientific challenges and opportuni- preliminary outcomes by the end of Year 1 (contributing ties for understanding critical watersheds. to FY15 Task 2 below). In conjunction with the VCNP, we have identified specific field sites and data to validate and We formed a reciprocal relationship with USDA’s Southern calibrate vegetation dynamics and ATS. We have identified Research Station (SRS) to develop the QUICFIRE model. a new approach to upscaling fine-physics wildfire modeling into a reduced model of wildfire ecology called QUICFIRE. The preliminary QUICFIRE model is approaching demon- 41

Page 42

Future Work Middleton, R. S., J. S. Levine, J. M. Bielicki, H. S. Viswanathan, J. W. Carey, and P. H. Stauffer. Goal 1. Understand climate-ecohydrologic interactions/ Jumpstarting commercial-scale CO2 capture and processes at small scales (<5 km2). We will incorporate storage with ethylene production and enhanced oil physical representations of vegetation, wildfire, and recovery in the US Gulf. 2015. GREENHOUSE GASES- hydrology into the Advanced Terrestrial Simulator (ATS) SCIENCE AND TECHNOLOGY. 5 (3): 241. ecohydrologic framework. We will move ATS beyond preliminary development (Year 1 goal) and demonstrate Middleton, R. S., J. W. Carey, R. P. Currier, J. D. Hyman, ATS on the Jaramillo River watershed (~3 km2) in the Valles Kang, Karra, Jimenez-Martinez, M. L. Porter, and H. S. Caldera National Preserve (VCNP). Compare and validate Viswanathan. Shale gas and non-aqueous fracturing ATS with an existing ecohydrologic model (PARFLOW). fluids: Opportunities and challenges for supercritical CO2. 2015. APPLIED ENERGY. 147: 500. Goal 2. Understand climate-ecohydrology interactions at regional scale (1000s km2), and develop a fully calibrated hydrologic model of the Colorado River based on high- resolution multispectral remote sensing, historic natural flows, regional climate drivers, drought mortality, wildfire, extreme climate events, and climate-driven vegetation changes. Analyze multiple coupled climate-ecosystem and energy-water nexus (EWN) scenarios. Compare/validate outputs with existing hydrologic models. Goal 3. We will deliver 3-5 publications demonstrating the basic science needs and project advancements linking cli- mate change, extreme events, dynamic ecosystems, wild- fire, and hydrology. Bring multiple key researchers together for a workshop/special session at a national/international conference. Conclusion The overarching goal of this project is to develop the sci- ence and modeling capabilities to predict and quantify climate impacts on critical watersheds and water supply. This will have a potentially transformative impact on our understanding of climate change and the energy-water nexus (EWN) and our ability to mitigate and adapt to cli- mate change. Specifically, we are developing a new under- standing of the interaction and feedbacks between climate change and extreme events, climate-driven disturbances such as wildfire, drought and forest mortality, hydrology, and water for the EWN. Publications Allen, C. D., D. D. Breshears, and N. G. McDowell. On underestimation of global vulnerability to tree mortality and forest die-off from hotter drought in the Anthropocene. 2015. ECOSPHERE. 6 (8). Fisher, R., S. Muszala, M. Verteinstein, P. Lawrence, C. Xu, N. McDowell, R. Knox, C. Koven, J. Holm, B. Rogers, and others. Taking off the training wheels: the properties of a dynamic vegetation model without climate envelopes. 2015. Geoscientific Model Development Discussions. 8 (4). 42

Page 43

Complex Natural & Engineered Systems Directed Research Final Report Using Microreactors for Efficient Plutonium Separations (U) Stephen L. Yarbro 20130003DR Abstract constraints make actinide processing operations an ex- Chemical separations, especially difficult separations cellent candidate for microfluidics. Microfluidic systems depend on many factors. Fundamentally, reactions and and microreactors have high mass and heat transfer rates are controlled by the mass and heat transfer char- rates, which reduce stage volume and increase effi- acteristics of the equipment. In equipment where the ciency. Given the exacting requirements for processing length scales are very large with respect to the boundary actinides, these features are very attractive for reducing layers and turbulent eddies are formed, mass and heat footprint and associated handling and disposal costs. transfer are controlled by the contact area and mixing Also, actinide separations continue to be important to energy per unit volume. In equipment such as microre- the US for national security and energy programs. In this actors, where the length scales are small with respect to work, mass transfer rates for several design concepts the boundary layer and the flow is predominantly lami- were verified with experimental testing, supported by nar, heat and mass transfer are controlled by different multiscale modeling including continuum and Lattice- mechanisms. In this work, plutonium mass transfer was Boltzmann. Fabrication and testing of the equipment characterized with both neutral and anionic extractants under various operating conditions occurred in paral- and in channels with diameters from 0.1 mm (micro) to lel with refinement of the chemical systems chosen for 5 mm (milli). The mass transfer rates were measured study. The microfluidic devices were tested with plutoni- and mass transfer models were developed to predict be- um solutions and parameters were developed for simple havior. Also, several different slug-flow and membrane- mixing, separation and reduction-oxidation methods. based microreactor designs were successfully evaluated. This work resulted in the design and fabrication of more It was demonstrated that microchannel contactors are complex devices based on the selected chemical system, very efficient for Pu separations, with characteristic equi- Aliquat-336 in hydrochloric acid, and scale-up of the librium times of <0.5 seconds per stage. Therefore, auto- preliminary contacting equipment. Hydrochloric acid mated micro- and milli-reactor/contactor systems can be was selected as the aqueous system for its relevance to used to reduce footprint and increase the efficiency of pyrochemical residue treatment, and to reduce the com- plutonium separation processes. plexity of stripping plutonium from the organic phase for purification and waste management. Background and Research Objectives All industrially significant actinides have hazards associ- The goals were the following: ated with their storage and handling. All are radioactive, and fissile actinide isotopes require mass and volume Objective 1: demonstrate that a microreactor/contactor- limits to prevent a criticality accident. Typical separation based liquid-liquid extraction system can provide the operations use hazardous chemicals, such as concentrat- necessary mass transfer rates for a compact and efficient ed nitric or hydrochloric acid and various organic extract- separation process; ants for separation and purification. To protect workers from the radiation and direct exposure to hazardous Objective 2: develop the engineering approach to dem- materials, processing is done in remote-handling facili- onstrate a workable separation apparatus; ties or within engineered barriers like gloveboxes. Such Objective 3: demonstrate that the precise control facilities are expensive to build, operate and maintain. achievable in a microreactor can improve efficiency, Actinides also have high waste-handling, storage and dis- selectivity, and control in nuclear material processing. posal costs due to their radiotoxicity. These processing 43

Page 44

Scientific Approach and Accomplishments ics and mass transfer behavior in complex systems. System Enhanced mass transfer is important to improving the effi- variables such as contact angle, which is a measure of wet- ciency of actinide separation processes. It enables the con- tability, can be examined to predict changes in conversion tacting stage volume and flows to be minimized, reducing rates. While model sensitivities may be explored in 2-D, a the associated processing costs. Therefore, the approach full 3-D model is needed to capture conversion rates that focused on developing milli- and microfluidic contactors of were observed in the experiment, due to the additional various dimensions and phase contacting methods, then flow pattern induced by the wall. characterizing the mass transfer rate and operation char- Also, due to the solvent ability of the organics, it is very acteristics of each. Mass transfer and operating data was important to select chemically-resistant polymer materials used to develop models for future optimization. Specifi- for the contactors, to ensure that contaminants and leaks cally, three main areas of research were pursued. do not affect the operations. At these very small volumes,

1. Developing experimental and analytical protocols for trace contaminants can influence the operation by altering characterizing the mass transfer performance of vari- the wettability or by stripping or plugging the channels. ous contactor configurations. Overall, the project met the objectives with the various ac-
2. Designing and developing fabrication methods for complishments outlined below. contactor configurations. Two approaches were se- Objective 1 lected for development. In one approach, contactors were tested that used a micro-mixing “tee”-joint which • Successfully measured mass transfer coefficients for produced interspersed droplets (slugs) of alternating uranium and plutonium in a single stage with both organic and aqueous solutions in a narrow channel, tributyl phosphate (TBP) and Aliquat-336 (A336) in a and then separated the two immiscible phases. In the hydrochloric acid system, and in multiple stages with second approach, contactors were tested that used A336 in a hydrochloric acid system. different configurations of membranes at the interface between two serpentine microchannels in which the • Successfully measured mass transfer rates for uranium organic and aqueous streams flowed countercurrent to and plutonium in both TBP and A336. each other. Various membrane combinations (thick- ness, porosity and surface wetting) were tested, and Objective 2 very thin graphene membranes with precise porosity • Fabricated and successfully tested serpentine mem- characteristics were developed. brane countercurrent contactors with 60-cm path
3. Modeling the mass transfer of the various contact- lengths and ten contact stages. ing approaches using both continuum and Lattice- • Fabricated and successfully tested slug-flow contactors Boltzmann methods. with phase-separators incorporated into the contactor. In developing microfluidic contactors for actinide separa- • Fabricated graphene membranes on a glass substrate tions, there are several important aspects that impact the with a microchannel design, fabrication and operation. For example, surface roughness and wettability are very important to controlling • Implemented a successful moving mesh 2-D model of the droplet generation, separation and flow in membrane micro-channel membrane counter-current flow mass contactors. Depending on the direction of mass transfer, a transfer with reaction. thin film adhering to the surface in a slug-flow contactor can improve the mass transfer in the device. Also, the wet- • Initiated Lattice-Boltzmann models and incorporated ting characteristics of the construction materials influence mass transfer effects at the phase boundary. how well the phases are separated in a slug-flow device, and how well they absorb into the membranes for the Objective 3 membrane contactors. • Successfully demonstrated microfluidics for glovebox For modeling, resolving the interface and ensuring con- operations with plutonium. servation of mass is difficult and can prove troublesome to obtaining accurate models. To overcome this problem, a moving-mesh method was used. The moving-mesh method can be effectively applied to model hydrodynam- 44

Page 45

Patent disclosures Chamberlin, R. M., S. L. Yarbro, N. Xu, D. L. Decker, and P. T. Martinez. Microfluidic separations for actinide process- • 13-00193 - Using microchannels to simulate counter- ing and analysis . Invited presentation at 249th Ameri- current contacting columns can Chemical Society National Meeting. (Denver, CO, 22-26 March, 2015). • 14-00165 - Laminated thermoplastic film microfluidic device Chamberlin, R. M., S. L. Yarbro, N. Xu, K. J. Spencer, and A. Castro. (U) Using microreactors for efficient plutonium Impact on National Missions separations: FY13 annual progress report. 2014. Report LA-CP-14-00131. The U.S. has an ongoing need for efficient actinide separa- tion methods, specifically, developing the ability to deploy Dervishi, E.. Graphene as a potential membrane material. advanced separation methods for weapons materials puri- 2015. Los Alamos National Laboratory unclassified re- fication, advanced nuclear systems, nuclear forensics, and lease (in preparation). environmental management. It expands US capabilities McCulloch, Q., S. L. Yarbro, R. M. Chamberlin, K. R. Weis- in detection, measurement, and analysis of nuclear and brod, and E. Dervishi. Fabrication and testing of coun- radiological materials. In addition, microfluidics provide an tercurrent and quasi-countercurrent microfluidic sys- ability to conduct small-scale research and development tems. 2014. Report LA-UR-14-28228. with scarce and unique materials. Multi-stage systems have been developed and demonstrated based on a new Roberts, R. M.. FY14 report: Lattice Boltzmann modeling of understanding of the mass and heat transfer effects at microreactor systems. 2014. Report LA-UR-14-28228. small length and time scales, enabling a robust capability Weisbrod, K. R., R. M. Roberts, R. M. Chamberlin, and S. to carry out difficult separations. This results in a process- L. Yarbro. Simulation of mass transfer in a microfluidic ing capability that can meet new programmatic needs in a experiment using the moving mesh method. 2014. In variety of national security, research, space and nonprolif- COMSOL User’s Conference. (Boston, MA, 8-10 Octo- eration programs. ber 2014). , p. 1. Boston : COMSOL. Publications Xu, N., A. Castro, J. Gao, B. Manard, L. Meyers, R. M. Cham- Boland, K. S., R. M. Chamberlin, G. S. Goff, and S. L. Yarbro. berlin, and L. Tandon. Chemical separation and mea- Physical properties of Aliquat 336 loaded with metal surement tool box for bulk SNM analysis. Presented (Fe, U) chlorides. Presented at 39th Annual Actinide at DHS-DNDO-NTNFC Annual Review. (Chicago IL, 4-7 Separations Conference. (Salt Lake City, UT, 18-21 May, August 2014). 2015). Chamberlin, R. M., B. T. Manard, K. J. Spencer, R. E. Keller, D. R. Porterfield, D. L. Decker, and S. L. Yarbro. (U) Plu- tonium separations by liquid-liquid extraction in micro- fluidic cells. 2015. Los Alamos Controlled Publication report in preparation. Chamberlin, R. M., D. L. Decker, D. R. Porterfield, P. T. Mar- tinez, and S. L. Yarbro. Kinetic factors in Aliquat-336 extraction of Pu(IV) from hydrochloric acid solution. 2015. Manuscript in preparation. Chamberlin, R. M., K. J. Spencer, D. R. Porterfield, R. C. Keller, and S. L. Yarbro. (U) New solvent system for a plutonium separation. 2015. Los Alamos Controlled Publication report in preparation. Chamberlin, R. M., S. L. Yarbro, A. P. Borrego, D. L. Decker, C. C. Finstad, J. Gao, P. Martinez, Q. McCulloch, D. R. Porterfield, J. Rowley, K. R. Weisbrod, and N. Xu. Micro- fluidic and micro-engineered separations for actinide processing and analysis. Presented at 38th Actinide Separations Conference. (Albuquerque NM, 20-22 May, 2014). 45

Page 46

Complex Natural & Engineered Systems Directed Research Final Report Maximizing Flux through Engineered Metabolic Pathways Clifford J. Unkefer 20130091DR Abstract Background and Research Objectives The long-term goal of synthetic biology is to engineer Maximizing metabolic flux requires finely tuned regula- microorganisms to produce high-value products for tion of individual steps. The long-term goal of synthetic biotechnological applications such as, the carbon neutral biology is to engineer microorganisms to produce production of biofuels or chemicals. To achieve this goal, high-value products for medical, pharmaceutical, and it will be necessary to engineer functional multi step biotechnological applications. To achieve this goal, it will metabolic pathways in organisms, which will require a be necessary to engineer functional metabolic path- new approach to independently regulate the expres- ways in organisms. Maximizing flux through metabolic sion of each of the enzymes in the pathway in order to pathways is not simply a matter of over expressing all of optimize the flux through the pathway. Until recently, the enzymes in the pathway. In living systems, control of engineering strategies for altering gene expression have function occurs at the cellular (transcription to RNA and focused on transcription control using strong inducible translation to protein) and molecular levels (allostery promoters. We have developed a new approach for – feedback inhibition or activation). These controls are translational control of gene expression involving a class implemented by regulating the concentrations of the of modular synthetic riboregulator that has the potential species taking part in biochemical reactions, including to finely tune protein expression and independently con- the enzymes (e), substrates (s), products (p) and regula- trol the concentration of each enzyme in an engineered tory molecules (r), such that the rate of an enzymatic metabolic pathway. This development is important be- reaction can be generally expressed as v = v(ce, cs, cp, cause the most straightforward approach to altering the cr). De Novo design of optimized pathways is not pos- flux through a particular metabolic step is to increase or sible because even in model organisms, the cellular decrease the concentration of the enzyme. Our design concentrations and kinetic coefficients necessary to includes a cis-repressor at the 5’ end of the mRNA that design a regulated pathway are not known. In addition, suppresses translation by forming a stem-loop helix cellular metabolism is a complex network of branched occluding the ribosomal binding sequence. A trans- and interconnecting pathways. Metabolites may serve expressed activating-RNA frees the ribosomal-binding several functions. For example, many intermediates sequence, turning on translation. Using Watson-Crick in the central metabolism that many organisms use to base-pairing stability, we designed a cis-repressor that generate energy also serve as substrates for the biosyn- completely shuts off translation of the target gene and a thesis of essential cellular components, e.g., amino acids trans-activator that restores translation demonstrating (protein synthesis) and nucleotides (DNA synthesis). it is possible to achieve translational control of gene ex- Because of this complex metabolic network, engineering pression over a wide dynamic range. We have also found an organism to direct a particular metabolite to produce that a targeting sequence can be modified to develop ri- a desired product could have an unanticipated deleteri- boregulators that can independently regulate translation ous effect on growth, rendering the de novo design of of many genes. Finally we demonstrated that by subtly regulated metabolic pathways extremely complex. altering the sequence of the trans-activator, it is possible Riboregulators can be used to finely tune gene expres- to alter the ratio of the repressed and activated states sion. Because the output of a metabolic reaction is di- and achieve intermediate translational control. These rectly proportional to the enzyme concentration (d[cp]/ riboregulators will find broad application in achieving dt ~ kcat[ce]), by far the most straightforward approach DOE’s goal of carbon neutral renewable production of to altering the flux through a particular metabolic step transportation fuels. 46

Page 47

is to increase or decrease the concentration of the en- ies of the same strong constitutive promoter. To achieve zyme catalyst. Current strategies for engineering gene our overall goal of a wide dynamic range of riboregulator expression are limited to increasing the expression of a control, it was necessary to perform the following tasks: particular gene by placing it under the control of a stronger 1) create a modular system with the capacity to regulate transcriptional promoter or using one of several strategies multiple genes independently and consistently; 2) design to knock down or knock out a wasteful gene. Our overall crRNA regulatory sequences that sequester the RBS and goal was to develop the molecular biology tools necessary completely block translation (crRNA OFF); 3) identify taR- to finely tune the expression of individual enzymes in an NA regulatory elements that free the RBS from the crRNA engineered metabolic pathway. Specifically our goals were OFF state and turn back on translation to the fullest extent

1. to develop a family of riboregulators that can be used possible; 4) demonstrate that we can design orthogonal for translational regulation of the expression of each of the targeting sequences that can be used to direct a taRNA to enzymes in a complete metabolic pathway and 2) demon- its specific partner crRNA; 5) demonstrate intermediate strate a strategy for the simultaneous directed evolution of levels of translation by altering the regulatory sequence in riboregulators to optimize the expression of the enzymes the taRNA. to maximize flux through a metabolic pathway. Scientific Approach and Accomplishments Modular design of riboregulators. In our riboregulator sys- tem (Fig. 1), both the cis-repressor (crRNA) and the trans- activator RNAs (taRNA) have modular structures composed of targeting (black) and regulatory (red) sequences. The crRNA is within the 5’ untranslated region (UTR) of the mRNA and naturally folds to a structure that sequesters the ribosomal binding sequence (RBS), preventing transla- tion of the downstream gene (Fig. 1a). The taRNA is tran- scribed in trans and the binding and subsequent structural transition between these two regulatory RNA elements dictates whether or not the transcribed mRNA will be translated into the protein product (Fig. 1b). The target- ing sequences are complementary and target a particular taRNA to a specific crRNA (Fig 1c). The stability of the taRNA regulatory element is dictated by the sequence in its variable region. The all ON version of the taRNA has little potential to form secondary structures in its regulatory helix as depicted in Fig. 1d. Depending on the relative sta- bility of the structural elements in the regulatory regions of the crRNA and the taRNA, the complex can rearrange to form the translation-activating extended duplex, freeing the RBS to turn on translation (Fig. 1d). Since the extended duplex between the elements has a much greater potential stability, expression tuning is accomplished by altering the availability of the sequence within the taRNA. Availability is determined by weakening the intramolecular base- pairing of the taRNA helix and adjusting the intermolecular Figure 1. (a) The crRNA has a modular design that includes a interactions in a common spacer region (Fig.1d). By tuning targeting sequence (black) and a regulatory sequence (red) that the relative stability between the regulatory helices of the forms a stem-loop structure at the 5’-end of the gene, seques- crRNA and the taRNA and that of the extended crRNA/ tering the Ribosomal Binding Sequence (RBS) and preventing taRNA duplex, it is possible to alter the ratio between translation of the mRNA; (b) the taRNA is transcribed in trans repressed (Fig. 1c) and activated (Fig. 1d) states and to and contains modules that are complementary to the targeting achieve intermediate translational control (Fig.1e). and variable regulatory sequences that produce stem-loop struc- tures of differential stability. The all ON version of the taRNA has In our systems, translation of both the crRNA:gene fusion little potential to form secondary structures as depicted in (d). (c) and the taRNA are driven independently by separate cop- The targeting sequence binding event results in the formation of 47

Page 48

a taRNA:crRNA/mRNA complex. In this complex, translation of (CR-3 and CR-4). Based on Watson-Crick base-pairing the target gene is either repressed or (d) the complex rearranges stabilities, we estimated the thermodynamic stability of to free the RBS and activate translation. (e) It is possible to tune each of these stem loop structures as follows: CR-1, -34.3 the relative stability of the secondary structural elements in the kcal/mol; CR-2, -40.0 kcal/mol; CR-3, -45.9 kcal/mol; CR-4, crRNA and taRNA to obtain intermediate levels of translation. -57.5 kcal/mol. As expected, the translation was depen- dent on the thermodynamic stability of the RBS-occluding helix. Complete occlusion of RBS and minimal leakage of translation were possible only when the cis-repressor was designed to form a stem loop that had a nearly perfect double helical structure (CR-4, Fig. 2b). Significantly, subtle changes in the sequence near the RBS, with single base pair mismatches, resulted in increased translation of the cat gene (as seen for CR-2 and CR-3), indicating that even minimal structural breathing around the RBS allowed ribo- some access and translation of cat and imparted chloram- phenicol resistance. Riboregulated translation. We chose CR-4 as the regula- tory element for further experiments as this sequence/ structure represented the maximal OFF riboregulated state (crRNA OFF) and provided minimal leakage of transla- tion and protein expression. Our next goal was to identify taRNA regulatory elements that free the RBS from the crRNA OFF and turn back on the translation to the fullest extent possible. The taRNA contained a targeting se- quence (TS1) that could recognize a specific cis-repressor RNA through Watson-Crick base pairing and a regulatory element that could hybridize with the crRNA regulatory element in a fashion that could expose the RBS, leading to Figure 2. (a) crRNA sequences that occlude the RBS with stem- translation (Fig. 1). To test the taRNA, we designed the four loop structures of increasing stability. Potential Watson-Crick (−) constructs diagrammed in Figure 3a. In the Maximum ON and wobble base pairs are indicated. (b) Plot of the growth rate construct, expression of cat is driven by the pT7A1 with no of transformants carrying a plasmid encoding a crRNA/ (chlor- translational control. The transcription of the crRNA:cat amphenicol resistance) fusion. In every case, translation is driven fusion construct is also driven by pT7A1; however, as constitutively by the T7A1 promoter/T7 terminator pair and described above, translation of cat is fully attenuated by a translation is controlled by CR-1 (O), CR-2 (Δ), CR-3 (), or CR-4. crRNA with the CR-4 regulatory element. In the Regulated ON construct, transcription of the crRNA:cat fusion and The cis-Repressor. We initially optimized the cis-repressor the taRNA are driven independently by identical pT7A1 element in the absence of a taRNA to determine the se- promoters; taRNA is designed to hybridize with the crRNA quence and structure of the cis-repressor element required in a fashion to free the RBS and modulate the translation for complete repression of mRNA translation. We trans- of cat (Fig. 1c). The sequence of the taRNA-ON variant formed E. coli with a plasmid encoding the chlorampheni- provides for minimal structure, facilitating the formation of col antibiotic resistance gene under control CR-1 (Fig 2a) an extended duplex with the crRNA. In addition, the spacer and measured the bacterial growth of the transformants is fully complementary to the crRNA, facilitating taRNA using the antibiotic chloramphenicol as a sensitive, titrat- strand invasion and RBS release (Fig. 1c). In the Regulated able marker for translational leakage (Fig. 2b). Under our OFF construct, the stability of the secondary structure of experimental conditions, we found that the growth of wild the taRNA and lack of complementary spacer nucleotides type E. coli was completely inhibited in a culture medium prevents rearrangement of the regulatory complex (Fig. containing 30 μg/mL chloramphenicol. Because our initial 1b), leaving the RBS occluded and blocking translation. sequence, CR-1, showed a significant growth phenotype to chloramphenicol concentrations of 120 μg/ml, we devel- Cells transformed with the Regulated ON construct grew at oped cis regulatory elements with the helix extended by chloramphenicol concentrations up to 750 μg/mL, indicat- eight base pairs (CR-2, Fig. 2a) and then further increased ing that regulated ON taRNA restored translation to a very its stability by systematically removing mismatched bulges 48

Page 49

significant extent (Fig 3b). The chloramphenicol-resistant Summary of riboregulator properties. To optimize engi- phenotype caused by expression of the Regulated ON neered metabolic pathways, it is necessary to tune the taRNA, demonstrated that the taRNA is capable of freeing level of expression of each enzyme catalyst independently the RBS on the cis-repressed mRNA, resulting in translation in the pathway. We have shown that riboregulators can of cat. By contrast, expression of the Regulated OFF taRNA achieve translational control of gene expression over a did not activate translation and its phenotype was essen- wide dynamic range. In our system, the crRNA is encoded tially identical to the Maximum OFF state, indicating that in the 5’-untranslated region of the mRNA and folds to the structure within the taRNA and mismatches in the four block the ribosome-binding site. We have demonstrated nucleotide spacers effectively prevented the formation of a crRNA regulatory sequence that can completely shut off a translation-activating extended duplex as diagrammed in translation (CR-4, Fig. 2). Expression of a trans-activator Figure 1d. RNA (taRNA) initiates a targeted binding event, followed by structural rearrangements between crRNA and taRNA, which dictates whether or not the transcribed mRNA will be translated into the protein product. We have designed a regulatory sequence in the taRNA that can interact with the CR-4 regulatory sequence to turn the translation back on to a high level. In our riboregulator system (Fig. 1) both the cis-repressor and trans-activator RNAs are composed of targeting and regulatory sequences. The targeting sequences on a crRNA/taRNA pair are complementary and formation of the targeting helix is required for taRNA to modulate translation of a crRNA-regulated gene. While we have tested initially only two targeting sequences, our design with 34 nucleotides, in principle, provides ample sequence diversity to design multiple orthogonal target- ing sequences. In our modular design, only the targeting sequence is changed to develop riboregulators that can independently regulate translation of many genes. By subtly altering the taRNA’s regulatory helix and spacer, we can alter the equilibrium between repressed and activated states and achieve intermediate translational control. We introduced sequence diversity at only five positions in the regulatory region of the taRNA and were able to identify taRNAs that produced intermediate levels of cat expres- sion. We have demonstrated the riboregulators work when engineered cis-RNA regulated genes are encoded in the bacterial chromosome and the taRNA is plasmid encoded. More recently we used have demonstrated the riboregu- lation of translation of chromosome encoded antibiotic Figure 3. (a) Schematic of the plasmid encoded expression cas- resistance genes in the photosynthetic bacterium Syn- settes used to test the efficacy of the riboregulators. In the Maxi- mum ON construct, expression of is driven by the pT7A1 with no echococcus elongatus. translational control. The Maximum OFF construct transcription of the crRNA: fusion is driven by pT7A1; translation of is attenu- Impact on National Missions ated by the crRNA. In the Regulated ON construct, transcription The riboregulators described in this report will be valu- of the taRNA is driven independently by pT7A1; taRNA modulates able for optimizing metabolic flux through engineered the translation of . In the Regulated OFF construct the stability metabolic pathways using directed evolution approaches. of the secondary structure of the taRNA prevents rearrangement We envision chromosomal genes encoding each enzyme of the regulatory complex (see ); translation remains blocked. (b) in the pathway to be placed under independent transcrip- Plot of growth rate of transformants as a function of chloram- tion control using the same strong promoter/terminator phenicol concentration. The were transformed with the various pair for each gene. These genes will be translationally expression vectors and are plotted as follows: Maximum ON (), Maximum OFF (), Regulated ON (O), Regulated OFF () and wild- repressed by a crRNA consisting of a unique targeting type. sequence and the Maximum OFF regulatory helix CR-4. Targeting sequences will specifically pair a single crRNA 49

Page 50

controlled gene with its taRNA partner. The pathway dehydrogenase complexed with oxaloacetate and ade- taRNAs will be plasmid encoded making it straightforward nosine-5-diphosphoribose. 2015. Protein Data Bank. to introduce diversity and generate a library of taRNAs Hall, R. S., R. Martí-Arbona, S. P. Hennelly, T. S. Maity, F. that cover the range of translational activation for each Mu, J. M. Dunbar, C. J. Unkefer, and P. J. Unkefer. In- gene. Transformants would include all combinations of TA vitro Characterization of an L-Kynurenine-Responsive regulatory sequences, creating a cell population expected Transcription Regulator of the Oxidative Tryptophan to have a range of metabolic fluxes. It will be necessary to Degradation Pathway in Burkholderia xenovorans. develop high-throughput screening or selection methods 2013. Journal of Molecular Biology Research. 3 (1): 55. to identify cells in this population with optimal metabolic flux. Riboregulators will provide the metabolic engineer a Hayes, R. L., J. K. Noel, A. Mandic, P. C. Whitford, K. Y. San- bonmatsu, U. Mohanty, and J. N. Onuchic. General- new tool to tune the level of expression of each enzyme ized Manning Condensation Model Captures the RNA independently in a pathway. The synthetic biology capa- Ion Atmosphere. 2015. Physical Review Letters. 114: bility developed in this project and more specifically the 258015. riboregulators developed here will play an essential role in engineering organisms optimized for biofuel production Hayes, R. L., J. K. Noel, U. Mohanty, P. C. Whitford, S. P. and for carbon neutral production of commodity chemical Hennelly, and K. Y. Sanbonmatsu. Molecular simula- feedstocks that drive U.S. Industry. Developing these syn- tion captures free energy stabilization of RNA by Mg2+. thetic biology approaches will be essential to achieving the 2014. Biophysical Journal. 106 (7): 1508. Department of Energy’s goals of replacing petrochemicals Hennelly, S. P., I. V. Novikova, and K. Y. Sanbonmatsu. The with renewable biological feedstocks. expression platform and the aptamer: cooperativity between Mg2+ and ligand in the SAM-I riboswitch. Publications 2013. NUCLEIC ACIDS RESEARCH. 41 (3): 1922. Fox, D. T.. Maximization of Isoprenoid Production and Ex- traction in a Model Organism. Presented at Algae Bio- Hennelly, S. P., I. V. Novikova, and K. Y. Sanbonmatsu. The mass Summit. (Orlando, FL, 1-3 Oct. 2013). expression platform and the aptamer: cooperativity between Mg2+ and ligand in the SAM-I riboswitch. Fox, D. T.. Biocatalytic Production of Commodity Chemicals. 2013. Nucleic Acids Research. 41 (3): 1922. Presented at American Society of Microbiology. (Bos- ton, MA, 17-20 May 2014). Jarchow-Choy, S. K., A. T. Koppisch, and D. T. Fox. Synthetic Routes to Methylerythritol Phosphate Pathway Inter- Fox, D. T., M. Krishnamurthy, T. Kern, M. K. DeAguero, S. mediates and Downstream Isoprenoids. 2014. CUR- P. Hennelly, T. Dale, S. N. Twary, S. R. Starkenburg, RENT ORGANIC CHEMISTRY. 18 (8): 1050. R. Marti-Arbona, K. Sanbonmatsu, and C. J. Unkefer. Riboregulation of gene expression in Synechococcus Jarchow-Choy, S., D. Fox, E. Schmidt, A. Koppisch, R. Marti- elongatus PCC7942. Presented at Pacificahem 2015. Arbona, and C. Unkefer. Monitoring the distribution of (Honolulu HI, 15-20 Dec 2015). isoprenoid products from the MEP pathway in algae and cyanobacteria. 2013. ABSTRACTS OF PAPERS OF Gonzalez, J. M., R. Marti-Arbona, and C. J. Unkefer. Crystal THE AMERICAN CHEMICAL SOCIETY. 245: -. structure of phosphoenolpyruvate carboxylase from Methylobacterium extorquens. Presented at American Jha, R. K., T. L. Kern, D. T. Fox, and C. E. M. Strauss. Engi- Crystallographic Association Annual Meeting. (Albu- neering an Acinetobacter regulon for biosensing and querque, NM, 24-28 May 2014). high-throughput enzyme screening in E-coli via flow cytometry. 2014. NUCLEIC ACIDS RESEARCH. 42 (12): Gonzalez, J. M., R. Marti-Arbona, and C. J. Unkefer. Crystal 8150. structure of Enolase from Synechococcus elongatus. 2014. Protein Data Bank. Kirmizialtin, S., S. P. Hennelly, A. Schug, J. N. Onuchic, and K. Y. Sanbonmatsu. Integrating Molecular Dynamics Gonzalez, J. M., R. Marti-Arbona, and C. J. Unkefer. Crystal Simulations with Chemical Probing Experiments Using structure of Malyl-CoA lyase from Methylobacterium SHAPE-FIT. 2014. Methods in Enzymology. 553: 215. extorquens. 2015. Protein Data Bank. Krishnamurthy, M., S. P. Hennelly , K. Y. Sanbonmatsu , and Gonzalez, J. M., R. Marti-Arbona, and C. J. Unkefer. Crystal C J. Unkefer . Riboregulators as tunable gene switches structure of Malyl-CoA lyase from Methylobacterium for post-transcriptional control of gene expression. extorquens. 2015. Protein Data Bank. 2015. In 249th ACS National Meeting. (Denver, CO, March 22-26). , p. BIOL. Abstracts of Papers 249 ACS Gonzalez, J. M., R. Marti-Arbona, and C. J. Unkefer. Crystal National Meeting: ACS. structure of Methylobacterium extorquens malate 50

Page 51

Krishnamurthy, M., S. P. Hennelly, K. Y. Sanbonmatsu, Novikova, I. V., S. P. Hennelly, and K. Y. Sanbonmatsu. . Ex- and C. J. Unkefer. Riboregulators as gene switches for perimentally determined secondary structures of can- post-transcriptional regulation of gene expression. Pre- cer-related long non-coding RNAs. 013. In 245th ACS sented at Precision Genome Engineering and Synthetic National Meeting & Exposition. (New Orleans, 7-11 Biology, Keystone Symposia Conference,. (Big Sky, Mo- April, 2013). , p. PHYS. Abstracts of Papers: ACS. nanta, 11-16 Jan 2014). Novikova, I. V., S. P. Hennelly, and K. Y. Sanbonmatsu. Tack- Krishnamurthy, M., S. P. Hennelly, R. Marti-Arbona, J. M. ling Structures of Long Noncoding RNAs. 2013. INTER- Gonzalez, K. Y. Sanbonmatsu, and C. J. Unkefer. Prog- NATIONAL JOURNAL OF MOLECULAR SCIENCES. 14 ress toward tunable riboregulator switches for post- (12): 23672. transcriptional control of gene expression and the development of a synthetic pathway for acetyl-CoA Novikova, I. V., S. P. Hennely, C. S. Tung, and K. Y. San- biosynthesis . Invited presentation at Asia-Oceania Al- bonmatsu. Rise of the RNA machines: Exploring the gae Innovation Summit. (Deajeon, South Korea, 11-17 structure of long non-coding RNAs. 2013. Journal of Nov 2014). Molecular Biology. 425: 3731. Krishnamurthy, M., S. P. Hennelly, T. Dale, S. R. Starken- Sanbonmatsu, K. Y.. Rise of the RNA Machines: Long Non- burg, R. Marti-Arbona, S. N. Twary, K. Y. Sanbonmatsu, Coding RNAs. Invited presentation at 10th Horizons D. T. Fox, and C. J. Unkefer. Tunable riboregulator in Molecular Biology Conference International . (Max switches for post-transcriptional control of gene ex- Planck Institute, Gottingen, Germany, Sept 9-12 2013). pression. 2015. ACS Synthetic Biology. : 1. Sanbonmatsu, K. Y.. Integrating simulations and experi- Marti-Arbona, R.. New Advances Toward the Understand- ments of the SAM-I riboswitch. Invited presentation at ing and Manipulation of Metabolic Regulation. Invited RNA Dynamics. (Telluride, Colorado, 22-26 July, 2013). presentation at Department of Chemistry, Rice Univer- Sanbonmatsu, K. Y.. Large-scale simulations of biomolecu- sity. (Houston, Texas, 13 April 2014). lar machines. Invited presentation at Biosupercomput- Marti-Arbona, R.. New Advances Toward the Understand- ing Symposium. (Tokyo Japan, December 2012). ing and Manipulation of Metabolic Regulation. Invited Sanbonmatsu, K. Y.. Understanding the atomistic mecha- presentation at University of Puerto Rico, Rio Piedras nism of magnesium effects of RNA dynamics. Invited Campus. (San Juan, Puerto Rico, 30 Oct 2015). presentation at Mini-symposium on Modeling, Simula- Marti-Arbona, R., F. Mu, F. Nowak-Lovato, M. S. Wren, P. J. tion and Function of Biomolecular Assemblies. (Tokyo Unkefer, and C. J. Unkefer. • “Automated Genomic University, Tokyo Japan, November 2012). Context Analysis and Experimental Validation Platform Sanbonmatsu, K. Y.. Computational & Experimental Stud- for Discovery of Prokaryote Transcriptional Regulator ies: Ribosomes, Riboswitches & Long non-coding RNAs. Functions” . 2014. BMC Genomics. 15: 1. Invited presentation at Fourth Annual Summer Sym- Martí-Arbona, R., T. S. Maity, J. M. Dunbar, C. J. Unkefer, posium on Cellular Dynamics of Macromolecular Com- and P. J. Unkefer. Discovery of a Choline-Responsive plexes. (Montreal, Canada, 3-8 June 2014). Transcriptional Regulator in Burkholderia xenovorans. Sanbonmatsu, K. Y.. Structural Studies of Intact LncRNAs in 2013. Journal of Molecular Biology Research . 3 (1): 91. Plants and Mammals. Invited presentation at Keystone Nguyen, T. T., R. Martí-Arbona, R. S. Hall, T. Maity, Y. E. Val- Symposium: Long non-coding RNAs: Marching towards dez, J. M. Dunbar, C. J. Unkefer, and P. J. Unkefer. Iden- Mechanism. (Santa Fe, NM, 2-6 March 2014). tification and In-vivo Characterization of a Novel OhrR Sanbonmatsu, K. Y.. Dynamics of riboswitches: Molecular Transcriptional Regulator in Burkholderia xenovorans simulation. 2014. Biochimica et Biophysica Acta: Gene LB400. 2013. Journal of Molecular Biology Research. 3 Regulatory Mechanisms. 1839: 1046. (1): 37. Sanbonmatsu, K. Y.. Dynamics of riboswitches. Invited Novikova, I. V., A. Dharap, S. P. Hennelly, and K. Y. Sanbon- presentation at RNA Society. (Madison, WI, 26-31 May matsu. Shotgun secondary structure determination of 2015). long non-coding RNAs. 2013. Methods. 63: 170. Sanbonmatsu, K. Y.. Structural architecture of lncRNAs in Novikova, I. V., S. P. Hennelly, and K. Y. Sanbonmatsu. Sizing plants and mammals. Invited presentation at The Royal up long non-coding RNAs: Do lncRNAs have secondary Society: Long non-coding RNAs: evolution of new epi- and tertiary structure? . 2012. BioArchitecture. 2 (6): genetic and post-transcriptional functions. (Kavli Royal 189. Society Centre, Buckinghamshire, 28-29 Sept 2015). 51

Page 52

Sanbonmatsu, K. Y.. Integrating Molecular Dynamics Simu- lations with Chemical Probing Experiments. Invited presentation at Telluride Science Research Center: RNA Dynamics. (Telluride, CO, 20-24 July 2015). Sanbonmatsu, K. Y.. Dynamics of riboswitches: Molecular simulations. Invited presentation at Wellcome Trust Conference of Computational RNA Biology. (Cambridge UK, 11-13 Nov 2014). Twary, S. N., S. R. Starkenburg, C. J. Unkefer, and M. A. Al- varez. Comparative transcriptomics across algal species after induced lipid accumulation. Presented at Interna- tional Conference on Algal Biomass, Biofuels and Bio- products. (Santa Fe, NM, 15-18 June 2014). Unkefer, C. J.. Development of a synthetic pathway for acetyl-CoA biosynthesis. Invited presentation at Special Microalgae Workshop at the Korea Advanced Institute of Science and Technology. (Deajeon, South Korea, 21 Nov 2014). Unkefer, C. J., S. N. Twary, S. R. Starkenburg, M. Teshima, M. A. Alvarez, and P. J. Unkefer. Characterization of Lipid Metabolism in Nannochloropsis salina 1776: De- velopment of a Systems Biology Pipeline and Transcrip- tome Analysis. Invited presentation at Asia-Oceania Algae Innovation Summit. (Deajeon, South Korea, 17- 20 Nov 2014). 52

Page 53

Complex Natural & Engineered Systems Early Career Research Continuing Project Deciphering Nature’s Chemical Toolbox: Decoding the Logic of Biosynthetic Assembly Lines Alexander Koglin 20140624ECR Introduction gens and its impact on lethality rates is widely discussed The overarching goal of this project is to support the by several funding agencies (incl. HHS, DARPA, BARDA, biosecurity mission of LANL in defeating biothreats due DTRA, DOD), in scientific press releases and among gen- to drug-resistant bacteria and viruses and confront the eral public press outlets. accelerating development of antibiotic resistance by identifying novel bioactive compounds. This project This project is directly related to LANL’s missions in focuses on the identification of new biosynthetic clusters Global Security, Biothreat Reduction, and in alignment that produce novel natural products with new chemical with the National Security goal of LANL’s Bioscience properties (chemotypes) to create a compound database Division. It addresses Priority Area 1 of the Complex for drug development and to decipher the logic of natu- Natural and Engineered Systems focus area. Specifically, ral product biosynthesis. Those compounds have the po- the ability to decode the biosynthetic logic will acceler- tential to be developed into highly effective antibacterial ate the field of synthetic and systems biology and allow treatments while being immune to current resistance for efficient engineering of drug candidates. This project mechanisms, since pathogenic bacterial strains have not supports LANL public health mission (Homeland Security been exposed to those compounds. Act 2002, DOE Strategic Plan 2011, Public Health Secu- rity and Bioterrorism Preparedness and Response Act Genetic and enzymatic studies provide insight of isolated 2002) as well as its mission in global threat reduction. It enzymatic functions in natural product biosynthesis, but also has direct appeal to WFO sponsors, as expressed in have not led to new strategies to discover novel natural ongoing interests by DTRA to develop countermeasures products biosynthesized by complex enzyme assem- for natural and man-made biothreats. It will both lever- blies. Bioinformatics and structural studies are provid- age and strengthen LANL’s capabilities in computing, ge- ing insight of three-dimensional folding and selective nomic and biosynthetic research. We will keep program protein complex formation processes of key enzymes managers in Global Security (Science of Signatures), essential for the formation of biosynthetic clusters that DTRA and DOE BES programs appraised of the results of produce medically relevant compounds. The unique our research so that we can develop future funding for and not-yet understood spontaneous self-assembly of this effort. these enzymes into clusters defines the very selective biosynthesis of all secondary metabolites of which only Progress an estimated 0.1% have been identified so far. Once For the bioinformatic part of the project, it was pro- understood, known biosynthetic complexes can then be posed to use and develop code that identifies as many engineered to produce new drugs or lead to an efficient new gene clusters encoding for biosynthetic machineries identification of new enzymatic assembly lines in vast for natural product production as possible in the vast genome data and isolation of their products that are genome space of bacteria, fungi and plants. Within the novel drug leads. analyzed 3,000+ genomes of bacterial and fungal strains, more than 16,000 new gene clusters were found. From Benefit to National Security Missions this vast amount of available strains, five were chosen New antibacterial and antiviral treatment options are ur- (Myxococcus xanthus DK 1622, Corallococcus coralloi- gently needed to reduce the threat to public health. The des DSM 2259, Streptosporangium roseum DSM 43021, accelerating development of drug resistance in patho- Catenulispora acidiphila DSM 44928, Herpetosiphon au- 53

Page 54

rantiacus DSM 785) in which the biosynthetic clusters are natural products in plants. We plan to complete the SQL unusual, i.e. not following the classical linear, multi-modu- database by the end of September. lar organization (Koglin, Walsh, Nat Prod Rep., 2009, 101; Strieker et al., Curr. Opin. Struct. Biol. 2010, 20, 234). This Future Work approach significantly reduces the probability of isolating Natural products represent the class of small bioactive compounds similar to known or identified natural prod- molecules that are widely applied as antibacterial, antivi- ucts. Within these five strains a total of 124 gene clusters ral, antitumor, antifungal, immunosuppressant and anal- have been identified and further characterized. gesic drugs. This project focuses on methods to identify genes coding for new biosynthetic enzyme clusters that Initially, the linear EMBL file format was chosen as the data produce bioactive metabolites with new chemical proper- output format. The advantage of this file format is that it is ties (chemotypes) in genetic information of bacteria, fungi compatible with a wide range of software for data analysis and plants, and to predict the chemical composition or and for genome, gene and peptide sequence alignment. chemical structure of these biosynthesized natural prod- The disadvantages of a linear data depository became ucts based on their encoding. This knowledge will allow (1) obvious as soon as the amount of produced data increased identification of new drug classes with new physiological significantly. While the manual comparison of 124 lin- properties, and (2) manipulation of these assembly lines ear data sets can be accomplished, it becomes increas- to produce specific drugs and/or efficiently manufacture ingly harder if not impossible as the number of identified novel and unique metabolites. biosynthetic systems reached 2,000 and more. The EMBL file format deposits all identified information for one full In the next funding period, we will continue analyzing vast genome in a single text file, including all DNA and peptide bacterial and fungal genome data in an automated search sequences of all protein with all predicted enzymatic func- algorithm we developed. We will identify and characterize tions and substrates. In cases a strain contains more than novel natural product assembly lines, continue to com- 30 biosynthetic systems, the file size is significantly larger plete the database of biosynthetic clusters to allow the than 2MB (output files with up to 35MB have been gener- identification of common recognition motifs (jigsaw puzzle ated) and cannot be utilized for comparative analysis on links). We will continue updating and improving the search gene, peptide or protein function level. algorithm using the new data. As a consequence, we decided to change the layout of our We will also expand the search for relevant biosynthetic data depository and the design of our database significant- clusters and natural products to plant genomes and in- ly and currently create an SQL-based database that allows clude those in our created libraries. to deposit each biosynthetic system separated in genes, peptide sequence, protein domains, domain and protein We will begin isolation of first, predicted natural products function and produced secondary metabolite and inde- and confirmation of their structure/composition. pendent from the strain. This way, we now enable com- parative analyses within one strain family and between all Conclusion sequenced bacterial strains simultaneously. At the same This project leads to the identification of biosynthetic time, it reduced depository files to manageable sizes and clusters in diverse genomes and the informed isolation of increases the speed for comparative analyses. We can now novel natural products from strains known to synthesize further include information from existing databases con- potent bioactive compounds. We collect and characterize taining information of known bioactive natural products data in a library of diverse biosynthetic clusters and me- and increase the learning base for the machine learning dicinally relevant natural products. This information is the efforts to reliably predict the chemical composition and basis to decipher the mechanism of complex self-assembly structure of a novel secondary metabolite from bacteria in natural product biosynthesis, which will reflect in the de- and fungi before the compound was first isolated. This SQL velopment of algorithms to predict the chemical structure database also allows the integration of plant genomes and of natural products based on their genetic encoding. database that include information of natural products from plants. For instance, with the help of the Bitter Database of the Weizman Institute that connects the chemical struc- ture of bitter tasting natural products with the genome of the producing plant, we can now identify the genes and enzymes that are responsible for the production of this compound. This will help us to gain information to predict 54

Page 55

Complex Natural & Engineered Systems Early Career Research Continuing Project Toward a Coupled Multi-physics Modeling Framework for Induced Seismicity Satish Karra 20150693ECR Introduction disposal. The coupled modeling framework developed The recent shale-gas revolution, has been accompanied in this work will form the basis for future forecasting by a five-fold increase in small to moderate magnitude tools. This will have a direct and immediate impact on (greater than 3) seismicity in the central and eastern US. policy-making as well as on the DOE’s Offices of Fossil It is likely that many of these earthquakes are induced by Energy (FE) and Energy Efficiency & Renewable Energy industry practice of disposing of fracking fluids into deep (EERE) programs in carbon storage, unconventional fossil formations. The issue of induced seismicity presents a fuels, and geothermal energy along with State agencies significant threat to US energy security, as regulatory with regard to waste water disposal. Our research will barriers could curb the expansion of unconventional also place LANL at the forefront when the DOE initiatives shale-gas operations as well as deployment of renewable such as SubTER and Energy-Water Nexus commence. energy technologies such as carbon sequestration and enhanced geothermal systems. Progress The project had started on June 15, 2015 and there are An induced earthquake occurs when pressurized fluid no accomplishments to report for this period. enters a fault and causes frictional failure. However, induced seismicity does not simply manifest as one Future Work single earthquake, but rather as a migrating, sporadic se- The main goal of this project is to develop the INDUS quence of events. Modeling these sequences requires an framework that is aimed to be built on top of two HPC understanding of the feedbacks between pressurization codes – PFLOTRAN (developed at LANL and other DOE and fault rupture, and the ability to capture both small labs) for flow and PyLith for earthquake dynamics. The and large space and time-scales. coupler developed in FY15 will be used in the exchange of fluid pressure and earthquake slip information This project addresses the technology gaps in model- between PyLith and PFLOTRAN sub-simulations, time- ing induced seismicity. We propose a high-performance stepping control, and mesh interpolation. Once testing computing coupled multi-physics framework in which on benchmark problems is complete, the coupler will be earthquakes alter fluid flow through fault networks, used to perform simulations on real datasets. thereby influencing pressure migration and the location, timing and magnitude of subsequent earthquakes. Conclusion Our primary goal is development of a new physics-based Benefit to National Security Missions framework for induced seismicity that couples the tradi- Shale gas, geothermal energy, carbon sequestration, and tionally separate fields of geomechanics (earthquakes) enhanced oil and gas recovery have a clear role in the and hydrology (fluid injection). The coupling will be U.S. energy policy, both in securing cost-effective energy achieved through the use of innovative new fracture- and reducing atmospheric CO2. Injection-induced seis- permeability damage relationships that describe flow micity poses a significant mitigation challenge to the US changes by as much four orders of magnitude in the energy sector, as it is intrinsically linked to the shale-gas fractured zone around a fault rupture. The framework fracking revolution, renewable energy technologies such leverages recent advancements in two high-performance as Carbon Capture and Storage and Enhanced Geo- computing open-source simulators – PFLOTRAN for thermal Systems, as well as underground waste water subsurface flow and PyLith for fault dynamics. The meth- 55

Page 56

odology will have wide utility for industry, academic and government partners in carbon sequestration, waste water injection and enhanced geothermal systems. Publications Chang, J., S. Karra, and K. B. Nakshatrala. Large-scale optimization-based non-negative computational framework for diffusion equations: parallel implementation and performance studies. Journal of Scientific Computing. Grasinger, M., D. O’Malley, V. Vesselinov, and S. Karra. Decision analysis for robust CO2 injection: application of Bayesian-information-gap decision theory. To appear in International Journal of Greenhouse Gas Control. 56

Page 57

Complex Natural & Engineered Systems Early Career Research Final Report The World’s First Drought and Insect Caused Global Tree Mortality Monitoring System Chonggang Xu 20130733ECR Abstract est inventory and remote sensing products. The fusion The critical urgency of forecasting climate impacts of different sources of information makes it feasible to and feedbacks makes understanding, quantifying, and accurately quantify tree mortality from MODIS imag- predicting terrestrial carbon balance and subsequent ery, which has never been possible from image analysis climate impacts one of the greatest science challenges alone. The ED model tracks the growth and mortality for currently facing the world. The real-time monitoring cohorts of plants with similar sizes and we have incor- systems for dominant types of disturbances will provide porated the world-leading theory and data on mortality a key foundation for our understanding of global carbon mechanisms into the model, which provides mechanistic balance. We currently already have a fire-burned area accuracy unparalleled in other global vegetation models. monitoring system and comprehensive land-use change The FRT model is a leading model that simulates light database; however, there is no global monitoring system reflectance and transmittance for a very heterogeneous of drought/insect-caused vegetation mortality, which canopy represented by different tree cohorts, the same could be at similar magnitude of fire-caused tree mortal- canopy structure representation in the ED model. An ity. Armed with the world’s leading capability in dynamic automated fitting algorithm based on Ensemble Kalman vegetation modeling and tree mortality research, we filter (EnKF) was used to tune the tree mortality param- developed the world’s first automated global drought/ eter in the ED model so that the simulated light reflec- insect-caused tree mortality monitoring system. The tance in the FRT model tracks real-time observation mortality monitoring system was developed based on from MODIS imagery. Our ultimate goal is to develop the fusion of different sources of information includ- an unprecedented capability at LANL to simultaneously ing real-time mortality signals from remote sensing monitor global vegetation dynamics and estimate their imagery, vegetation change information simulated from consequences for global carbon balance and climate a vegetation dynamics model, radiative transfer and feedback; however it will take a very large infrastruc- reflectance information from a forest reflectance model, ture development in terms of computing and basic data and different sources of background information from handling for real-time global tree mortality monitoring. forest inventory and remote sensing products. We have Thus, although the developed ED-FRT-MODIS system is a successfully tested this system locally at Los Alamos and global tree mortality monitoring system, we only applied regionally in NM State. the system to areas where we have already classified high and medium resolution images in New Mexico. Background and Research Objectives Scientific Approach and Accomplishments Armed with the world’s leading research in vegetation mortality and state-of-the-art capability in dynamics veg- With the support of this project, we have successfully etation modeling and vegetation anomaly detection, we developed a tree mortality detection, attribution and developed the world’s first global monitoring system for quantification system. Our forest disturbance detection drought and insect-caused tree mortality. The developed and attribution system used anomaly detection of time monitoring system fused real-time mortality signals from series stacks of satellite imagery (Landsat or MODIS) to MODIS imagery, predicted vegetation dynamics informa- identify forest mortality events and classify them as fire-, tion from an ecosystem demography (ED) model, radia- drought-, or insect-caused. The system is divided into tive transfer and reflectance information from a forest three components: Pixel Processing, Disturbance Detec- canopy reflectance and transmittance (FRT) model, and tion, and Attribution and Evaluations. There are two different sources of background information from for- phases of pixel processing, image selection and calcula- 57

Page 58

tion of vegetation indices (VIs). Image selection requires Our ED-FRT-MODIS tree quantification system is built on the identification of one image for each year within a the fusion of vegetation dynamics from ED and the remote specific phonological window that has the least amount of sensing signals. In order to account for the prior knowl- cloud or snow coverage. The ideal time window for coni- edge on specie drought tolerance, we have developed a fers in New Mexico is late October/early November. Pixels mechanistic hydraulic failure model (Fig. 2 a) and the initial that are contaminated by clouds or snow can be filled from runs showed that it performed relatively well at the site adjacent time windows if a suitable clear image is unavail- level (Fig. 2b). A sensitivity analysis has been conducted able. Images are then opened sequentially, appropriate VIs to identify key model parameters for the prediction of are calculated, and saved as time series stacks. We use five hydraulic failure, which is critical for the success of the lit- VIs (Normalized Difference Vegetation Index (NDVI), Nor- erature survey and laboratory measurements targeted for malized Burn Ratio (NBR), Brightness, Greenness, and Wet- the global mortality monitoring. This new model has been ness) that have been proven to be useful for identifying used to predict the tree mortality in SW USA for pinion changes in vegetation. We perform disturbance detection pine (Fig. 3). by examining the time series for each pixel independently for dates when the trend deviates by a predetermined amount (2-3σ) for each VI. Specifically, pixels and dates are flagged for potential mortality when NDVI, Greenness, or Wetness decrease or NBR or Brightness increase. Finally, mortality events are identified and attribution is pro- posed according to combinations of flags and then growth through other pixels for contagious disturbances like fire and insects. Specifically, fire occurs during severe events (NDVI, NBR, Greenness, and Brightness are flagged regard- less of Wetness), and grows into adjacent pixels with lower disturbance severity; drought is indicated by increases in Brightness and losses of Wetness in the remaining un- burned pixels (this potentially misses drought killed pixels that burn in the same year); finally, pixels with insect mor- tality are those that have less disturbance (primarily those with a loss of greenness or NDVI). Our system has been tested for the LANDSAT images in Los Alamos (Fig. 1). Figure 2. Plant hydraulic failure and evaluation at a LANL mea- surement site (Sevilleta, NM). Figure 3. Predictions of piñon pine mortality events using our hydraulic-faiure model in ED using the CMIP5 predictions by CESM under RCP 8.5 (McDowell, William and Xu et al In Review). Figure 1. Detection of fire- (1993-2012; shades of gray), With the new hydraulic failure model, we have first tested drought- (2002-2004; shades of red and 2011-2012; shades of orange), and insect-induced (1993-2012; shades of blue) forest our ED-FRT system at a 3km by 3 km location in Los Ala- mortality in the Los Alamos/Santa Fe vegetated region (green) mos. It shows that the ED-FRT system is able to track the using our methodology. estimation of mortality from the high resolution Quickbird image (Fig. 4). With the local model evaluations, we have 58

Page 59

implemented our tree mortality detection, attribution and Phase II ); PI: Jeff Chamber (LBL) quantification system for the NM state (Fig. 5). • Global Tree Mortality Prediction Based on Hydraulic With the support of this project, we have published 6 Function Failure, 2015-2017, Role: PI . papers, with 1 paper in review and 3 papers in preparation. Funding: LANL LDRD ER, $1.2million;CXu(PI),SWehavealsopresentedourworkatAGU,ESAandIUROSevanto(co-PI)meetingsin2013and2014.•CriticalWatersheds:ClimateChange,TippingPoints,andWaterSecurityImpacts,2015-2017,Role:co-I.Funding:LANLLDRDER,$2.7 million; R. Middleton (PI), C. Xu (co-I), R. Linn, N. McDowell Publications Adam, H., P. William, C. Xu, S. Rausher, and N. McDowell. Empirical and process-based approaches to climate- induced forest mortality models. 2013. Frontiers in Functional Plant Ecology. : 438. McDowell, N. G., N. C. Coops, P. S. A. Beck, J. Q. Chambers, Gangodagamage, J. A. Hicke, Huang, Kennedy, D. J. Figure 4. ED-FRT Model Evaluation for the means and individual pixels for an 3km by 3km area in Los Alamos. Krofcheck, Litvak, A. J. H. Meddens, Muss, Negron- Juarez, Peng, A. M. Schwantes, J. J. Swenson, L. J. Ver- non, A. P. Williams, Xu, Zhao, S. W. Running, and C. D. Allen. Global satellite monitoring of climate-induced vegetation disturbances. 2015. TRENDS IN PLANT SCI- ENCE. 20 (2): 114. McDowell, N. G., R. A. Fisher, Xu, J. C. Domec, Holtta, D. S. Mackay, J. S. Sperry, Boutz, L. e. e. Dickman, Gehres, J. M. Limousin, Macalady, Martinez-Vilalta, Mencuccini, J. A. Plaut, Ogee, R. E. Pangle, D. P. Rasse, M. G. Ryan, Sevanto, R. H. Waring, A. P. Williams, E. A. Yepez, and W. T. Pockman. Evaluating theories of drought-induced vegetation mortality using a multimodel-experiment framework. 2013. NEW PHYTOLOGIST. 200 (2): 304. Figure 5. Mortality detection and attribution for fire, drought and insect for NM state. Muss, J., and C. Xu. Remote Identification and Attribution of Tree Mortality in Northern New Mexico. Remote Impact on National Missions Sensing of Environment. The development of ED-FRT-MODIS system represented an enormous step forward in our understanding of terrestrial XU, C., J. Muss, and N. McDowell. A Tree Mortality Quanti- carbon feedback to atmosphere, which is a key area of fication system based on the integration of a dynamic vegetation model and a radiative transfer model. Re- climatic change research in LANL’s mission to mitigate the mote Sensing of Environment. impacts of global energy demand. With the support of this project, the PI has successfully secured the current follow- XU, C., J. Muss, and N. McDowell. A global sensitivity analy- on findings: sis for process-based tree mortality prediction under future climatic change. Ecological Modelling. • Developing Global Simulation of Drought-associated Vegetation Mortality within the Community Land Xu, C.. Decoupling correlated and uncorrelated uncertainty Model, 2015-2017, Role: Co-I. Funding: DOE office of contributions for nonlinear models. 2013. Applied science, $3million;N.McDowell(PI),C.XU(co-I),S.MathematicalModelling.37:9950–9969.Sevanto,MCaiXu,C.,J.Muss,andN.McDowell.UncertaintyandSensitiv-ityAnalysisforatreemortalityquantificationsystem.•NextGenerationEcosystemExperimentsintheTropics,PresentedatIUFRO2014Worldcongress.(SaltLake2014-2024:Role:DeputyPOC.Funding:DOEofficeofCity,5-11Oct,2014).science,$15million (5 million per year for 3 years for Phase I and potential extension with another 7 years of Xu, C., R. Fisher, J. Muss, and N. McDowell. Uncertainty 59

Page 60

and Sensitivity Analysis for Process-based Tree Mortal- ity Modelling. Invited presentation at ECOLOGICAL AS- SOCIATION OF AMERICA 2013. (Minneapolis, 8-13 AGU 2013). Xu, C., R. Fisher, N. McDowell, and S. Sevanto. Our Lim- ited Ability to Predict Vegetation Responses to Water Stress. 2013. New Phytologist. 200 (2): 8. Xu, C., S. Sevanto, N. McDowell, and J. Muss. Challenges and Solutions for Simulating Drought-Caused Tree Mortality in Earth System Models. Presented at 2014 AGU. (San Francisco, 5-9 Dec, 2014). Xu, C., and M. Chen. Application of Remote Sensing in Ecosystem and Landscape Modeling. 2013. In Remote Sensing of Natural Resources. Edited by Wang, G.. , p. 173. Florida: CRC Press. 60

Page 61

Complex Natural & Engineered Systems Early Career Research Final Report From Troposphere to Ionosphere: How Much Do Thunderstorms Disturb the Total Electron Distribution? Erin H. Lay 20130737ECR Abstract This work addressed the emerging field of coupling between tropospheric and ionospheric phenomena. It was largely an empirical study that bounded the mag- nitudes and extent of ionospheric disturbances caused by underlying thunderstorms. It also focused on under- standing the coupling of energy from the thunderstorm, through the bottomside of the ionosphere, and into the bulk ionospheric plasma, called the F-region. The project was extremely successful with two publications and one submitted paper, several follow-on funding opportuni- ties, including a funded Intelligence Community Postdoc proposal that brought a new postdoc to LANL. The tools developed in this work for analysis of ionospheric data will continue to be used at LANL for many future proj- ects. Background and Research Objectives The ionosphere is a high-density layer of plasma through Figure 1. Illustration of the thunderstorm/ionosphere system which all satellite-detected natural and man-made and altitudes of importance at which ionospheric fluctuations electromagnetic signals from the Earth’s surface must have been detected. pass. Similar to dispersion of light by transparent sol- Shorter period (1 – 4.5 min; 3.7 – 16.7 mHz) wave varia- ids, significant perturbations in the ionospheric plasma tions associated with severe thunderstorms have also electrons disrupt electromagnetic signals. Variation in been observed in the F-region [3,9,10] and D-layer [11]. ionospheric plasma has traditionally been attributed Model results of acoustic wave propagation are consis- to changes in geomagnetic activity (space weather). tent with observations, and indicate that such waves Only recently, the ionospheric community has begun to originating from thunderstorm heights (< 12 km) should realize that tropospheric weather (below ~12 km) could be able to reach 250-350 km altitude within ~250 km have a significant effect on the ionospheric plasma [1,2]. horizontally of the source [12,13]. Although propagation Thunderstorm-originated atmospheric gravity waves and ray tracing models indicate that acoustic waves can (AGW) with periods longer than ~5 minutes have been reach ionospheric heights, the source for such waves is observed in the ionospheric F-region (200-400 km) [3,4] still unclear. While models typically use thunderstorm and at lower D-layer altitudes (80-100 km) [5-8]. The convection as the source, such a mechanism has not source of these gravity waves is typically thought to be been confirmed. In addition, the atmospheric cutoff at the pressure force of overshooting of convective thun- lower frequencies makes it unclear whether the source derstorm plumes [1]. Figure 1 illustrates the thunder- is a narrowband emitter, or whether it is a broadband storm/ionosphere system and altitudes of importance at emitter that is filtered as it propagates to the iono- which we have detected ionospheric fluctuations. sphere. 61

Page 62

As these buoyancy and acoustic waves propagate into the project based on the frequent observation of highly the plasma of the ionosphere, the motion of the neutral coherent acoustic waves that were closely localized over molecules drags the ionospheric ions. Charge separation, extremely large thunderstorms. This new objective was in turn, generates electric fields that cause motion of the to utilize these acoustic signatures to determine a source electrons. Because electrons are bound to the magnetic region in the storm for these waves. We were able to field lines of the Earth’s magnetic field, field-aligned ir- develop a technique that yielded two probable source regularities can occur. While the general physics behind regions in the thunderstorms: the convective core and the this coupling is understood, there have been very few mea- stratiform region. This work will soon be submitted soon to surements done to allow an understanding of the expected Nature Geoscience. The location technique is also poten- magnitude and spatial size of the ionospheric response to tially useful in locating other tropospheric sources such as thunderstorm-produced buoyancy and acoustic waves. In explosions, tsunamis, and earthquakes. We have proposed addition, the static and electromagnetic field generated by to several new sources of funding (Underground Nuclear electric charges in the thunderstorm itself also can intro- Explosion Signatures Experiment, Space Nuclear Detection duce instabilities and electron density fluctuations. Program, LANL LDRD) to aid in locating such events with this technique. The research objectives of this project were 1) to char- acterize spatial size, magnitude, and temporal evolution Scientific Approach and Accomplishments of ionospheric perturbations related to nearby thunder- The scientific approach of this study was to correlate storms on a regional scale, and how they vary with thun- ionospheric fluctuations with thunderstorm activity on a derstorm properties (i.e. size, height, lightning activity); regional and global scale to better understand thunder- and 2) to characterize the magnitude and spatial extent storm influence on the ionospheric plasma. of troposphere/ionosphere coupling on a global scale as a function of geomagnetic latitude. Understanding the bottomside ionosphere was important for generating an understanding of coupling processes Objective #1 was addressed in several papers: Lay et al. from below. We used a technique developed previously at [3] presented preliminary case studies of ionospheric LANL to probe the lower ionospheric electron density and response in the bulk of the plasma. Lay et al. [14] was a fluctuations [6,7,16]. Using this technique, we found that statistical study that characterizes, for the first time, the the background quiet-time bottomside electron density spatial extent and magnitude of ionospheric buoyancy and profile increased significantly faster in altitude than previ- acoustic waves response versus thunderstorm size. Lay ously thought. Based on a case study over a thunderstorm, et al. [15] characterized the bottom of the ionosphere to we found that a nighttime profile over a thunderstorm understand how waves couple from below into the main becomes much less steep than in a quiet region, becoming plasma layer, and it found that the background bottomside more like a daytime profile. These results were published ionosphere has a much steeper increase in electron den- in Lay et al., [15]. Figure 2 shows our experimentally-de- sity with altitude than used in most previous models. This termined nighttime profile as a solid line, which increases has significant implications in future modeling efforts. in electron density with altitude much faster than that of the previously used community model (red dashed line). Objective #2 was addressed by searching for similar Above thunderstorms the electron density does not in- correlations between thunderstorms and ionospheric crease as rapidly with altitude (Figure 2, black dashed line). disturbances in lower latitudes. However, we found that ionospheric dynamics in lower latitudes is dominated by Next we looked at case studies of F-region (where the processes other than thunderstorms (equatorial plasma main ionospheric plasma resides) perturbations above drifts, solar/geomagnetic storms) so much so that effects thunderstorms. We found very good spatial and temporal due to thunderstorms are indistinguishable from the noise. correlation between thunderstorms and ionospheric grav- This finding is important in understanding the relative ity waves (periods 6 - 20 minutes) and ionospheric acoustic effect of thunderstorms on overall ionospheric fluctua- waves (periods 2 – 4 minutes). Figure 3 shows an example tions globally. Because the finding was a null result, this of the results from this study. Two nights with thunder- part of the work was not published. However, this work storm activity are plotted on the two rows. The underlying lead to some interesting follow-on proposals addressing blue contours show lightning density in the thunderstorm. ionospheric scintillation at low latitudes (the degradation The left column shows the acoustic waves over the thun- of an electromagnetic signal as it propagates through a derstorm in red. The right column shows the gravity waves disturbed plasma). in red. Both the left and right columns show that with the larger thunderstorm in the top row, there are more acous- An additional objective developed during the course of 62

Page 63

tic and gravity waves. The smaller thunderstorm still has row, there are more acoustic and gravity waves. The smaller visible waves, but to a smaller extent. thunderstorm in the bottom row still has visible waves, but to a smaller extent. These results were published in Lay et al., [3] and submit- With the case studies motivating further investigation, ted for publication to Journal of Geophysical Research we performed a statistical study of data from May – July – Space Physics [14]. In order to analyze these F-region 2005 over the U.S. Great Plains. This analysis compared data, we developed data analysis tools for freely-available thunderstorm size with the associated area covered by measurements of total electron content made by Global ionospheric perturbations, as well as the magnitude of Positioning System ground receivers. These tools will be the perturbations. We found that the ionospheric area useful to the lab for any future ionospheric projects involv- covered by gravity waves increased with respect to thun- ing total electron content measurements. In fact, they are derstorm size, typically covering 10 times the area of the already being used for follow-on projects and proposals. thunderstorm (Figure 4c). The ionospheric area covered by acoustic waves (Figure 4a) and the maximum acoustic wave amplitude (Figure 4b) increase as thunderstorm area increases. However, there seems to be a minimum effect detectable only when thunderstorm size exceeds a thresh- old of ~30,000 km2. These results have been submitted for publication to Journal of Geophysical Research – Space Physics [14]. Figure 2. Electron density profile versus altitude. Solid black line shows our experimentally-determined nighttime profile, which increases in electron density with altitude much faster than that of the previously used community model (red dashed line). Above thunderstorms the electron density does not increase as rapidly with altitude (black dashed line). Figure 4. Results of statistical study of data from May – July 2005 over the U.S. Great Plains comparing thunderstorm size with the area of ionospheric acoustic waves (a), the maximum magnitude of the acoustic waves (b), the area of the ionospheric gravity waves (c), and the maximum magnitude of gravity waves (d). In addition to the peer-reviewed publications above, sever- al notable follow-on projects developed from this project: The finding of coherent acoustic waves above thunder- storms led to the development of a phase-differential method to trace those waves back to the source. Locat- ing the source can give an understanding of the source mechanism. Preliminary findings suggested an acoustic source from the collapse of a high convective core, as well as a source in the stratiform region of the storm associated with large horizontal “spider” lightning and ionospheric electrical discharges called sprites. These findings will soon be submitted to Nature Geoscience. Further development of this method and investigation of source coupling phys- Figure 3. Ionospheric gravity waves (red, right column) and ics has been submitted to LANL DR (20160065DR). This ionospheric acoustic waves (red, left column) over thunderstorms method can also be used for locating explosions via their (blue contours) on June 9-10, 2005 (top row) and June 16-17, ionospheric signatures. Using this method the West, TX 2005 (bottom row). Above the larger thunderstorm in the top 63

Page 64

fertilizer plant explosion in 2013 was located amid a front This work also brought Dr. Robert Haaser to our team as of thunderstorms. A proposal to the Underground Nuclear part of the Intelligence Community Postdoc Program. He is Explosions Signature Experiment was submitted to exploit studying ionospheric effects of meteors and bolides using this signature, and additional related proposals are in the tools developed from this work. The IC postdoc spon- preparation. sor, Peter Bythrow, will also be visiting LANL soon as part of the project, and will be renewing contacts in the areas of A notable cluster of acoustic waves in the statistical data Space Nuclear Power and National Security. was not associated with any thunderstorms or geomag- netic storms, yet it was associated with a time period The results of this study directly impact LANL’s on-orbit with meteor activity. The hypothesis that the cluster was and ground-based nuclear detection (NUDET) program. associated with meteor activity led to a funded Intelligence The newly-developed understanding of characteristics Community Postdoc proposal. We now have a postdoc, Dr. (magnitude, area) of ionospheric perturbations associated Robert Haaser, working in this position under the mentor with thunderstorms will impact our interpretation of on- supervision of Drs. Erin Lay and William Junor. orbit detected signals that passed over thunderstorms. The improved understanding of the bottomside ionosphere will The development and availability of ionospheric GPS tools aid in ground-based NUDET detection. Ground-based very- created the opportunity for several new proposals. We are low-frequency (VLF) signals travel in the Earth-Ionosphere currently submitting proposals to address ionospheric scin- waveguide, of which the bottomside ionosphere forms the tillation (degradation of radio signals as they pass through upper boundary. Thus, the VLF wave attenuation is very a plasma) using GPS measurements. We have submitted sensitive to the bottomside electron density profile, and an to LANL ER (20160231ER) and plan to submit to the NASA accurate location and characterization of an event is there- ROSES call. fore also sensitive to that profile. Finally, the results of this study are of interest to external funding agencies that sup- Impact on National Missions port our work in that understanding thunderstorm-gener- This study focused on understanding the magnitude and ated wave characteristics helps us to distill thunderstorm- extent of ionospheric perturbations, which is important driven waves from other phenomena, such as meteors and for predicting ionospheric disruption of satellite commu- explosions. This, in turn, can allow us to better understand nications, GPS geolocation signals, and electromagnetic ionospheric fluctuations from those other sources. signals such as EMP that are an important part of the Lab’s national security mission. The empirical results from this References work gave an estimate of the ionospheric area affected by

1. Vadas, S. L., and D. C. Fritts. Thermospheric responses thunderstorm perturbations and associated magnitudes to gravity waves arising from mesoscale convective in mid-latitude regions. These findings also indicated that complexes. 2004. JOURNAL OF ATMOSPHERIC AND primary disruptions in the equatorial region are not likely SOLAR-TERRESTRIAL PHYSICS. 66 (6-9): 781. to be associated with thunderstorms, but instead with scintillation. Proposed follow-on work will use the tools 2. Immel, T. J., Sagawa, S. L. England, S. B. Henderson, developed here to further understand scintillation disrup- M. E. Hagan, S. B. Mende, H. U. Frey, C. M. Swenson, tions. and L. J. Paxton. Control of equatorial ionospheric morphology by atmospheric tides. 2006. GEOPHYSICAL The empirical findings from this study will be used to test RESEARCH LETTERS. 33 (15). and evaluate existing atmospheric gravity wave models, enabling us to distill and discover the physical processes 3. Lay, E. H., Shao, and C. S. Carrano. Variation in total responsible for energy and mass transport between the electron content above large thunderstorms. 2013. bottomside and main plasma layer, and to better nowcast GEOPHYSICAL RESEARCH LETTERS. 40 (10): 1945. and forecast ionospheric disturbances. More in-depth
2. Vadas, S. L., and H. -. Liu. Numerical modeling of the modeling is required to fully understand atmospheric/ion- large-scale neutral and plasma responses to the body ospheric coupling for forecasting capabilities. The neces- forces created by the dissipation of gravity waves sary work has been proposed as a LDRD-DR. from 6 h of deep convection in Brazil. 2013. JOURNAL The proposed follow-on work for ionospheric signatures OF GEOPHYSICAL RESEARCH-SPACE PHYSICS. 118 (5): detection of explosives directly impacts work on detection 2593. and characterization of explosive events. These signatures
3. Yue, J. i. a., S. L. Vadas, She, Nakamura, S. C. Reising, could be used as a complimentary technique to determin- Liu, Stamus, D. A. Krueger, Lyons, and T. a. o. Li. Con- ing location and yield of explosive events. 64

Page 65

centric gravity waves in the mesosphere generated by Geophysical Research: Space Physics. 119 (6): 4916. deep convective plumes in the lower atmosphere near 16. Shao, , E. H. Lay, and A. R. Jacobson. Reduction of Fort Collins, Colorado. 2009. JOURNAL OF GEOPHYSI- electron density in the night-time lower ionosphere in CAL RESEARCH-ATMOSPHERES. 114. response to a thunderstorm. 2013. NATURE GEOSCI- 6. Lay, E. H., and X. -. Shao. Multi-station probing of ENCE. 6 (1): 29. thunderstorm-generated D-layer fluctuations by using time-domain lightning waveforms. 2011. GEOPHYSICAL Publications RESEARCH LETTERS. 38. Lay, E. H., X. M. Shao, A. K. Kendrick, and C. S. Carrano. Ionospheric acoustic and gravity waves associated with 7. Lay, E. H., and Shao. High temporal and spatial- thunderstorms. 2015. In press, J. of Geophys. Res. – resolution detection of D-layer fluctuations by using Space Physics, July 2015, LA-UR-15-22667. time-domain lightning waveforms. 2011. JOURNAL OF Lay, E. H., X. M. Shao, and A. R. Jacobson. D region electron GEOPHYSICAL RESEARCH-SPACE PHYSICS. 116. profiles observed with substantial spatial and temporal change near thunderstorms. 2014. Journal of Geophys- 8. Yue, J. i. a., Hoffmann, and M. J. Alexander. Simultane- ical Research: Space Physics. 119 (6): 4916. ous observations of convective gravity waves from a ground-based airglow imager and the AIRS satellite Lay, E. H., X. M. Shao, and C. S. Carrano. Variation in total experiment. 2013. JOURNAL OF GEOPHYSICAL RE- electron content above large thunderstorms. 2013. SEARCH-ATMOSPHERES. 118 (8): 3178. Geophysical Research Letters. 40: 1945. 9. GEORGES, T. M.. HF DOPPLER STUDIES OF TRAVELING Shao, X., E. H. Lay, and A. R. Jacobson. Ionospheric varia- IONOSPHERIC DISTURBANCES. 1968. JOURNAL OF AT- tions in response to lightning discharges and their MOSPHERIC AND TERRESTRIAL PHYSICS. 30 (5): 735. parental thunderstorms. Presented at International Conference on Atmospheric Electricity. (Norman, OK, 10. BAKER, D. M., and K. DAVIES. F2-REGION ACOUSTIC USA, 15-20 June 2014). WAVES FROM SEVERE WEATHER. 1969. JOURNAL OF ATMOSPHERIC AND TERRESTRIAL PHYSICS. 31 (11): 1345. 11. Marshall, R. A., and J. B. Snively. Very low frequency subionospheric remote sensing of thunderstorm- driven acoustic waves in the lower ionosphere. 2014. JOURNAL OF GEOPHYSICAL RESEARCH-ATMOSPHERES. 119 (9): 5037. 12. Walterscheid, R. L., G. Schubert, and D. G. Brinkman. Acoustic waves in the upper mesosphere and lower thermosphere generated by deep tropical convection. 2003. JOURNAL OF GEOPHYSICAL RESEARCH-SPACE PHYSICS. 108 (A11). 13. Zettergren, M. D., and J. B. Snively. Ionospheric signa- tures of acoustic waves generated by transient tropo- spheric forcing. 2013. GEOPHYSICAL RESEARCH LET- TERS. 40 (20): 5345. 14. Lay, E. H., X. M. Shao, A. K. Kendrick, and C. S. Carrano. Ionospheric acoustic and gravity waves associated with thunderstorms. 2015. Submitted to J. of Geophys. Res. – Space Physics, April 2015, LA-UR-15-22667. 15. Lay, E. H., X. M. Shao, and A. R. Jacobson. D region electron profiles observed with substantial spatial and temporal change near thunderstorms. 2014. Journal of 65

Page 66

Complex Natural & Engineered Systems Early Career Research Final Report Understanding The Catalytic Conversion of Oligosaccharides to Fuels and Chemical Feedstocks Andrew Sutton 20130757ECR Abstract other biomass derived products which are inherently Our recent work has developed a procedure to produce over-functionalized for use as chemical feedstocks or fu- branched alkanes containing twelve carbon atoms from els. Carbohydrates represent an attractive carbon source starch in two steps using mild acidic conditions with lan- for conversion to hydrocarbons and chemical feedstocks. thanide and palladium catalysts. The reaction proceeds Glucose is the most abundant monosaccharide on the well at 200 oC and 200 psi H2 but the mechanism and planet and its use as a fuel and chemical feedstock reaction intermediates are unknown. If we can fully precursor represents a worthwhile target, if cost effec- understand the steps required for this reaction we can tive chemical transformations can be developed. In develop routes to obviate the need for expensive cata- contrast to introducing functional groups to crude oil lysts and move to ambient reaction conditions. derived substrates, biomass conversion to fuels requires removal of functional groups that represents a complete Background and Research Objectives paradigm shift from traditional catalysis science. Bio- We have clearly demonstrated the feasibility of convert- mass is composed of carbohydrates and aromatics, both ing oligosaccharides to branched alkanes which helps of which are chemically linked into large polymers and LANL position itself as a leader in alternative energy fuel chemically “over-functionalized”. The primary challenges production. However, the process uses palladium (Pd) in biomass conversion are thus depolymerization to and Lewis acidic lanthanide (Ln) compounds as catalysts. monomers and as the resulting intermediates possess an Palladium is an expensive metal (~ $5000/100g) and abundance of oxygen, deoxygenation. These two chemi- rare-earth metals have been highlighted by the Depart- cal transformations both require the catalytic cleavage ment of Energy (DOE) as “Critical Materials” (i.e. materi- of carbon-oxygen bonds, and are the primary thrusts of als that are essential to many clean energy technologies catalysis science in biomass conversion. This new field but also at significant risk of supply disruption). We presents great opportunities to redefine some areas of are currently unaware as to the mechanism or the key modern catalysis science. intermediates involved in this transformation and if Recently, organocatalyzed aldol chemistry has been we could identify these and develop a stepwise under- developed at LANL to extend the chain length of furan standing of the process, then we believe that we could aldehydes (5-hydroxymethylfurfural (HMF) R = OH or develop non-precious metal catalysts (NPMC) that could furfural (FF) R = H) derived from glucose (Figure 1). This perform both the hydrogenation (palladium catalyzed method of chain extension (required for resultant fuel reaction) and the hydrodeoxygenation reaction (HDO, energy density) can be performed with a variety of catalyzed by Pd and Ln). A deeper understanding of how donor molecules to allow molecules with between 8 and the reaction proceeds could also enable us to perform 16 carbon atoms to be synthesized. These molecules these transformations under ambient temperatures and share common functional group combinations com- pressures which would lead to significant cost benefits prised of furans, olefins, carbonyls, and hydroxyl groups. on subsequent scale-up of this process to produce fuel Our recent catalytic HDO approach uses palladium (Pd) or feedstocks on a commercial scale. with Lewis acidic lanthanide triflate salts (Ln(OTf)3). The challenge to perform controlled low temperature Sequential hydrogenation of exocyclic substrate unsatu- dehydration and deoxygenation of multifunctional sub- ration, then furan ring-opening followed by HDO of the strates is significant and any advances or insight into this resultant polyketone species produces linear alkanes chemistry can be readily applied to polyols, sugars and (Figure 2). Direct hydrogenation of the furan to produce 66

Page 67

a tetrahydrofuran (THF) adduct requires a significantly more energy intensive route to ring-open so is a less desirable route. To eliminate the conversion of glucose to HMF of FF and use glucose or starch directly we used an adaptation of Garcia-Gonzalez chemistry. In our approach we converted starch to cyclized tetrahydrofuranylfuran derivatives and even performed the same reaction with starch extracted from a potato. These species share the same functional groups as the aldol products shown in Fig- ure 1, and thus similar reaction conditions can be used to convert these molecules into branched alkanes (Figure 3). Although this conversion of starch to alkanes works well, we currently have no details on the intermediates formed or the mechanism by which alkanes are produced Figure 3. H NMR following the La(OTf) catalyzed dehydration of 3-pentanol in acetic acid. Substrate key: 3-pentanol = (red circle), 3-acetoxypentane = (pink circle), 2-acetoxypentane = (orange circle), 2-pentene = (green diamond). Scientific Approach and Accomplishments Figure 1. Temperature-dependent product distributions ob- Non-food based carbohydrates are attractive renewable served when 3-pentanone (A), 3-pentanol (B), and 2-pentene (C) starting materials for conversion into fuels and feedstocks. are subjected to standard HDO conditions for 15 hours ([Sub- Glucose and xylose, as the main building blocks of lignocel- strate] = 400 mM, [La(OTf)] = 60 mM, 10 wt% Pd/C relative to lulosic biomass, are the most abundant monosaccharides substrate, 2.5 mL acetic acid, 200 psi H). Substrate key: 3-penta- on the planet and their use as precursors for these applica- none = (black square), 3-pentanol = (red circle), 3-acetoxypen- tions are a worthwhile target. Conversion of these sugars tane = (pink circle), 2-acetoxypentane = (orange circle), 2-pen- into platform chemicals such as furfural or 5-hydroxymeth- tene = (green diamond), pentane = (blue triangle). ylfurfural (HMF) provides useful precursor molecules for further chain extension or derivatization in order to add value or energy density. In order for these molecules to be converted into direct drop-in fuel replacements, the abundant functional groups associated with carbohydrates (i.e. oxygen atoms) need to be removed. Recently we developed a route for selective chain extension of furan aldehydes (furfural and HMF) to produce isolable higher order furfuraldehydes containing between 8 and 16 carbon atoms. These can then be selectively converted into linear alkanes in excellent yields using a standard set of catalysts and reaction conditions. The chain extended furfuralde- hydes are converted into polyketones via hydrogenation of their exocyclic double bond followed by subsequent acid- catalyzed ring opening. However, the following hydrodeoxygenation HDO reac- tion to convert the polyketones into hydrocarbons requires Figure 2. Temperature-dependent product distributions ob- more forcing conditions (200 °C and 200 psi H2). Reducing served during the HDO reaction with La(OTf) absent for 15 hours. the energy required for this final step (i.e. via lower tem- ([Substrate] = 1.2 M, 10 wt% Pd/C relative to substrate, 2.5 peratures and pressures) should allow for a more eco- mL acetic acid, 200 psi H). Substrate key: 3-pentanone = (black nomically viable process for potential commercialization. square), 3-pentanol = (red circle), 3-acetoxypentane = (pink Studying this reaction and understanding the stepwise circle), 2-acetoxypentane = (orange circle). mechanism would allow us to improve the HDO reaction 67

Page 68

by specifically addressing the highest energy transforma- Employing Pd/C and La(OTf)3 independently rather than tions. Herein we report the use of a simple model system in tandem allowed us to isolate intermediates in the HDO to probe a stepwise HDO reaction applicable to bio-derived process. Subjecting 3-pentanone to HDO conditions (200 molecules. o C) in the absence of La(OTf)3 yields only 3-acetoxypen- tane (Figure 2), confirming that the role of La(OTf)3 is to Initial mechanistic studies utilizing polyketone substrates facilitate C-O bond cleavage. Following this reaction as a were complicated by the presence of multiple reactive function of temperatures shows reduction to the alcohol at centers within the substrates. For example, 2,5,8-nonane- 100 o C followed, at higher temperatures, by a conversion trione could potentially yield 39 possible intermediates. To into the acetoxypentanes with nearly complete conversion overcome this complication, 3-pentanone was chosen as a at 180 o C. As expected, heating 3-pentanol in the pres- model substrate containing a single functional group that ence of La(OTf)3 in acetic acid results in efficient produc- was anticipated to proceed through a pathway that includ- tion of 3-acetoxypentane at 100 ° C and 2-pentene at 150 ° ed 3-pentanol and 2-pentene. This allows for rapid and C; additionally, the 2-acetoxypentane intermediate is also easy identification of intermediates and gives us a model observed at temperatures greater than 100 ° C (Figure 3). system where we can effectively screen catalysts and reac- tion conditions to elicit the desired transformations. To test The individual steps comprising the net dehydration this hypothesis, we subjected each of these compounds reaction were also studied. The esterification of 3-penta- to our typical HDO conditions (Pd/C, La(OTf)¬3, acetic nol proceeds to completion at 100 °C in acetic acid, and acid, 200 psi H¬2, 200 °C, 15 hours) and observed pentane heating 3-acetoxypentane with La(OTf)3 results in nearly formation in all cases. We also obtained ethyl acetate as a complete conversion into 2-pentene at temperatures side product from acetic acid reduction under these condi- greater than 150 oC (Figure 4). The reverse reaction (i.e. tions. the acetoxylation of 2-pentene) does not proceed at 150 oC in acetic acid alone, although a small amount of both In order to observe the intermediates, we performed the 2- and 3-acetoxypentane are observed in the presence of HDO reaction of each substrate as a function of tempera- La(OTf)3 and prolonged heating. 2-acetoxypentane can be ture. Our standard reactions involved combining the reac- substituted for 3-acetoxypentane with no change in reac- tants in a stainless steel reactor and heating for 15 hours tivity observed. at temperatures between 25 and 200 oC (with 200 psi H2 if required). Following cooling of the reactor, the crude reaction mixtures were filtered and yields were obtained by 1H NMR relative to an internal standard. For the HDO reactions of 3-pentanone, 3-pentanol and 2-pentene, the product distributions are shown in Figure 1. At temperatures above 175 ° C, 3-pentanone forms pen- tane as sole product at up to 85% yield. At temperatures below 175 ° C, 3-pentanol and the intermediates 2- and 3-acetoxypentane are observed (Figure 1A). The ace- toxypentanes prove to be an important intermediate in our process and their importance will be discussed later. Decreasing the temperature results in lower conversions of 3-pentanone and the expected 2-pentene intermediate is not observed under these conditions. The conversion of 3-pentanol into pentane goes to comple- tion at temperatures above 175 ° C with both 2- and 3-ace- toxypentane observed at below this temperature. Below 100 °C, 3-acetoxypentane was the major observed product (Figure 1B) and as the reaction temperature is decreased, Figure 4. Temperature-dependent product distributions ob- lower conversion of 3-pentanol is seen. The lack of observ- served during the C-O bond cleavage reaction with La(OTf) in able 2-pentene indicates that the alkene intermediate is acetic acid. 15 hours. ([Substrate] = 1.2 M, [La(OTf)] = 40 mM, readily hydrogenated under these reaction conditions; ac- 2.5 mL acetic acid). Substrate key: 3-acetoxypentane = (pink cordingly, hydrogenation of 2-pentene proved to be facile circle), pentane = (blue triangle). at all temperatures investigated (Figure 1C). 68

Page 69

In order to determine the importance of acetoxypentane as detailed in the supporting information. The formation and whether this is a key intermediate that facilitates facile of the acetoxypentane is favored by 0.31 kcal/mol. This C-O cleavage, we attempted the dehydration of 3-penta- step is followed by the release of HOAc and formation nol and 3-acetoxypentane in sulfolane. In the presence of of pentene and is exergonic, releasing ΔG=6.9 kcal/mol. La(OTf)¬3 excluding acetic acid, 3-pentanol did not under- Again, depending on the different models we studied for go C-O cleavage reactions. However, subjecting 3-acetoxy- coordination of HOAc in solution, the reaction free energy pentane to identical conditions yielded complete conver- can vary by 0.2 kcal/mol, and in the crudest model by 2.6 sion into the alkene at 150 oC. This result indicates that kcal/mol. However this does not change the conclusion the acetate moieties are important in enabling more facile of the favorability of this step. Finally, reaction with H2 C-O bond cleavage and subsequent alkene formation. yields pentane in a highly exergonic reaction, with a ΔG=- 17.5kcal/mol which supports the experimental data. This reaction pathway is summarized in Figure 5; palladium reduced the ketone to the alcohol which is converted to The results of this study have important implications for the acetoxypentane by acetic acid. The C-O bond cleavage the design of catalysts for the HDO of biomass derived is facilitated by the La(OTf)3 resulting in the alkene which molecules into fuels, especially those that facilitate the is readily hydrogenated to the alkene. Using acetic acid as removal of a ketone fragment. Our standard HDO reaction a solvent also has the benefit of being able to solubilize all conditions permits both ketone reduction and C-O bond the intermediates which prevents precipitation and char cleavage. The latter transformation is a multi-step process formation upon heating. that is aided by the formation of the corresponding ace- toxy compound. Cleavage of the C-O bond in this interme- diate requires lower temperatures than the parent alcohol under the conditions we employ. While a number of both precious- and non-precious metal catalysts for ketone hydrogenation at temperatures as low as 25 °C are known, to the best of our knowledge very few are stable under the acidic conditions employed in this study. To that end, we are currently developing more effec- tive ketone reduction approaches into our HDO process. By breaking down the HDO reaction into several reaction steps that are easy to study, we can rapidly screen cata- lysts and conditions in order to develop faster and lower energy routes to bio-derived replacements for fuels and chemical feedstocks. Additionally, by studying this reaction Figure 5. Reaction Pathway and Intermediates Formed in Ace- in depth, we have uncovered the importance of acetate toxy Mediated Hydrodeoxygenation. esters as intermediates in the cleavage of C-O bonds within our HDO process. This can be applied to other approaches We also analyzed this reaction pathway via computational that have also adopted our approach of utilizing acetic acid studies applied to reaction free energies for all the in- and could give insight into the mechanism involved in the termediate steps in Figure 5. The computed free energy use of other acid catalysts and the selection of acid used landscape indicates that once the first step (pentanone to for this purpose. We have achieved this by using a simple pentanol) is accomplished, the rest of the reaction follows commercially available surrogate for this reaction that can in an exergonic fashion (i.e. downhill). be easily and rapidly studied and can be used effectively to screen new catalysts and reaction conditions in the future. Our calculations did not include the modeling of the Pd/C By doing this we have revealed an additional component of catalysis within ketone reduction, however we find this our initial HDO chemistry using acetic acid for us to con- reaction to be uphill by only ΔG=+0.4 kcal/mol. From the sider as we strive to minimize required energy inputs for point of view of the accuracy of Density Functional Theory, the conversion of biomass derived substrates into drop-in these two species are isoenergetic (pentanone+H2 and fuels and feedstocks. Future work is focused on optimizing pentanol). However, Pd/C lowers the H2 dissociation bar- performance through the use of flow-reactors to capitalize rier, thus accelerating this reaction. Step 2 consists of the on the insight presented in this manuscript and applying acetoxylation of pentanol. As in all the remaining steps, this stepwise approach to polyketone substrates. the reaction free energy can be affected by side reactions in solution; therefore, several possibilities were considered 69

Page 70

Impact on National Missions The conversion of biomass into fuels and feedstocks is an area of huge growth and has seen considerable investment in the past decade. Success in this arena has the poten- tial to secure a domestic supply of energy (in the form of fossil fuel replacements) and could also provide important routes to chemical feedstocks. LANL has begun to develop a capability to convert biomass into value added chemicals through external research collaborations and internal LDRD funding. In order to continue to be competitive in this bio- energy arena, continued investment is required and this will not only be of great value to the DOE complex but will also allow LANL to become a leader in a developing and promising research area. Publications King, A. E., T. J. Brooks, Y. Tian, E. R. Batista, and A. D. Sut- ton. Understanding Ketone Hydrodeoxygenation for the Production of Fuels and feedstocks From Biomass. 2015. ACS Catalysis. 5 (2): 1223. 70

Page 71

Complex Natural & Engineered Systems Exploratory Research Continuing Project Methane Coupling Chemistry Promoted by Catalysts Containing Inexpensive Metals John C. Gordon 20150454ER Introduction LANL’s considerable expertise in catalysis using earth Current methodologies for the conversion of methane abundant metals. Success will competitively position (CH4) into commodity chemicals or fuels more amenable LANL for funding from DOE OBES, DOE-FE, ARPA-E, and to transportation, storage and use either depend upon industry. high energy input (Fischer-Tropsch) or on toxic/rare metals (e.g. Ir, Hg). The emerging importance of CH4 Progress with respect to the Nation’s energy portfolio makes The original proposal highlighted the potential of this an unsatisfactory scenario. To have any hope of low-valent Mn compounds supported by phosphine wide-spread implementation, any prospective CH4 ligands to affect the desired CH4 chemistry. Several functionalization catalyst must not require extensive methodologies have been employed to access various energy input and must not be constructed of elements phosphine containing scaffolds, including chelating such that it is prohibitively expensive and/or impossible phosphines as well as monodentate phosphines to obtain in sufficient quantities for Global utilization. containing sterically encumbered groups. Our specific goal in this project is to prepare molecular We have had limited success in placing a chelating systems comprised of inexpensive first row transition phosphine onto divalent Mn as a result of competitive metals that can facilitate both C-H and C-C bond binding of the Lewis basic solvent in which the coupling chemistries utilizing CH4 to produce ethane initially chosen Mn starting materials have to be (C2H6). C2H6 is itself a precursor to ethylene (C2H4), dissolved. We will pursue the use of alternative metal a valuable chemical feedstock used in the production starting materials that have higher solubilities in non- of important commodity chemicals that include coordinating organic solvents. polyethylene, surfactants, detergents, alcohols and others. Annual worldwide production of C2H4 was 120 x One alternative to phosphine based chemistry of Mn 106 tons in 2008. is the use of imines as ligands in order to more closely match the polarizability of the ligand to the metal. Benefit to National Security Missions A second feasible approach is to use charged (i.e. In direct line with the Chemistry & Chemical Sciences ER anionic) ligands. These approaches were not explored category, this project directly addresses how to advance at first because imines can be more prone to unwanted our understanding of the chemistry and chemical reactivity and charged ligands also necessitate lower methods that directly bear upon energy security, the electron density at the metal. Using DFT calculations, impact of energy production, and mitigating that impact. we have found that the most promising combination for In this case, the chemistry focuses on how to effectively CH4 activation at Mn (in terms of both thermodynamics upgrade an inert, abundant and cheap molecule in and kinetics) occurs when an amide-imine platform with the form of CH4 into a useful C2 chemical feedstock. an ethylene bridge is used. Based on these results, a Moving away from precious metals to first row transition ligand of this type has been synthesized, with current metals will potentially enable us to develop a catalytic experiments focusing on obtaining good purity and yield system for C-H bond activation (a necessary step in so that the metal chemistry can commence in earnest. functionalizing CH4) and C-C coupling chemistries that are economically scalable. This effort will capitalize on Another interest lies in low valent, low coordinate 71

Page 72

Fe complexes. Unlike Mn to date, we have found that Fe has a propensity to bind phosphines under all metal oxidation states of interest. Towards the pursuit of sterically encumbered neutral phosphine ligands that will enable Fe to bind and activate methane towards ethane formation, the syntheses of phosphobenzene ligands has also been pursued. The use of very large substituents on the benzene ring appears to complicate the isolation of the desired products. Unsubstituted phenyl substituents have been successfully made, however. These less sterically demanding ligands will be explored in terms of their chemistry with iron and the resulting complexes derived from these ligands will be explored with respect to their propensity to promote both the C-H and C-C activation chemistries that will be required to convert methane into ethane. Future Work We will begin the next FY by synthesizing and spectroscopically characterizing initial target phosphine complexes. The structural and magnetic properties of some of these complexes will also be used to benchmark DFT functionals and basis sets in order to determine the combination that most accurately describes the physical and electronic structure of complexes being used in this work without undue computational cost or compromising numerical accuracy. Conclusion The primary result from successful completion of this project will be experimental demonstration of each of the steps along the catalyzed pathway for conversion of CH4 into C2H6. While our ultimate goal is catalytic C2H6 production, even stoichiometric conversion of CH4 to C2H6 would be a significant accomplishment. 72

Page 73

Complex Natural & Engineered Systems Exploratory Research Continuing Project Fundamental Actinium Science In Search of Radiotherapeutics Eva R. Birnbaum 20150575ER Introduction required to produce large amounts of 225Ac ensures We propose a multi-disciplinary effort to enable the that only domestic facilities currently operated by rational design of actinium-225 complexes for targeted the Department of Energy Office of Science have the radiotherapy treatment of cancer. Actinium-225 is one capability to meet anticipated research demand for the of the most promising isotopes identified for use in isotope, if it can be shown functional in viable clinical targeted alpha therapy of cancer, a treatment strategy trials. Among the National Institutes of Health’s primary that utilizes high-energy alpha particles to selectively mission is the fight against cancer, and 225Ac’s potential eradicate malignant tissue. The greater widespread to incite a paradigm shift in treatment of this destructive clinical utility of 225Ac is hindered by radiotoxic side disease in many forms is well recognized by the effects that result from detachment of the radiometal scientific community. We anticipate that our research from its biological delivery vector. The design of ligands into fundamental principles that govern the chemical that effectively sequester actinium under biological bonding of actinium to biological vectors will catalyze conditions, however, has been impeded by a lack of further biological research, clinical trials supported by fundamental knowledge of this rare, highly-radioactive DOE isotope production infrastructure, and ultimately element. The proposed work will launch, for the first widespread, curative use of 225Ac in patients. time, a comprehensive investigation of the chemical and electronic properties of the element actinium in support Progress of rational design of actinium ligands for medical Efforts are underway to understand the chemical use. The chemical hardness, coordination number, properties of actinium (Ac) employing the long-lived and covalency of actinium ions in solution will be isotope, 227Ac (t1/2 = 21.77 years). This knowledge is determined. Preliminary data obtained with the stable necessary to design novel ligands that can bind actinium analog lanthanum and a heavier actinide (plutonium) to biological agents, allowing 225Ac to selectively reveal that fluorescence spectroscopy and grazing target diseases such as cancer and infectious pathogens incidence X-ray absorption spectroscopy can probe after injection into the patient. For this research, these important chemical properties while exhibiting we have utilized a supply of 10 mCi of 227Ac, which high sensitivity for the detection of the extremely low corresponds to only 0.61 μmol or 140 μg of actinium. concentrations of actinium available for investigation. The small quantities available, combined with challenges In addition, density functional theory (DFT) calculations associated with handling this highly radioactive will be employed to interpret the spectroscopic data. element, necessitate a method of chemical analysis With understanding of the electronic structure and basic that is sensitive to low concentrations. Fluorescence chemical properties of actinium, rational ligand design spectroscopy, a technique commonly used for the will be pursued in the service of clinical use of 225Ac- detection of sub-micromolar concentrations of analyte, containing compounds, which we hope can save lives. has been explored as a method for probing Ac3+ chemistry. Because Ac3+ is spectroscopically silent, an Benefit to National Security Missions organic fluorophore with a metal ion-binding unit is An improved understanding of the fundamental needed to relay information. Furthermore, to extract chemistry of actinium is essential to the clinical use useful data regarding the Ac3+ ion, a fluorescent sensor of 225Ac for the therapeutic treatment of disease, in is required whose response varies along with the bound particular malignant disease. The nature of the facilities metal ion. These efforts led to the design and synthesis 73

Page 74

of a new ligand, Ds-DOTAM. This ligand bears a macrocyclic Currently, we are working in close collaboration with metal ion-binding unit and a fluorescent dansyl group. The theorists conducting DFT and molecular dynamics oxygen atom of the sulfone group of dansyl can interact calculations to confirm this EXAFS interpretation of novel with a metal ion, perturbing its photophysical properties. Ac chemical structure. These X-ray absorption spectra of The properties of the ligand and baseline fluorescence actinium will enable, for the first time, the determination measurements were first collected using lanthanum (La) of Ac–ligand interatomic distances and coordination as a surrogate. Indeed, addition of La(NO3)3 to Ds-DOTAM numbers in the solution phase. Our results will find their leads to a red-shift in the absorbance maximum of the way into textbooks, as they provide information useful in ligand and a quenching of the dansyl emission. Density developing actinide separations schemes and biological Functional Theory (DFT) calculations were performed delivery agents for alpha radiotherapy. to understand the molecular orbital energies as well as the effect of metal binding. These results indicated that Future Work this red-shift arose from a stabilization of the lowest In year two, we will start the second phase of our study unoccupied molecular orbital (LUMO), primarily due to an using organic fluorophores to bind actinium. We will electrostatic interaction of the M3+ ion with the dansyl use the photophysical response of a new set of Ac- sulfone group. The addition of Ac(NO3)3 resulted in similar chelating ligands, lariat crown ethers, to investigate but not identical spectroscopic changes as observed Ac’s preferred electronic structure. These cyclic ligands with La. The UV-vis absorbance spectrum is also shifted, contain an additional side group (the lariat) to assist with although interpretation is complicated by signal from cation binding. They will be additionally modified with an excess of NO3- ion carried over from radiochemical a fluorophore-containing group to give an appropriate separation procedures. These data, interpreted with the spectroscopic signal as a probe. Stable, surrogate help of theoretical calculations, suggest that the actinium lanthanum compounds will provide initial insight ion is slightly softer than its lanthanum counterpart and into analogous lanthanide electronic structure and may prefer a slightly different coordination environment. experimental method development using the necessarily small concentrations of sample Ac. Titrations will be Although the fluorescence spectroscopic results revealed performed to determine the relative thermodynamic that both La3+ and Ac3+ interact with Ds-DOTAM, the stability of novel La and 227Ac compounds we consider precise nature of that interaction will certainly be different for this work. Promising compounds will be modeled for the two ions, which possess substantially different with DFT and compared to spectroscopic data, allowing ionic radii. To ascertain structural aspects of actinium confirmation of electronic structure information, and chemistry, we are also performing X-ray absorption near ultimately to select ligand systems that are most likely edge spectroscopy (XANES) and extended X-ray absorption to provide clear information on fundamental chemical fine structure (EXAFS) measurements using high intensity properties of Ac. synchrotron sources coupled to fluorescence detection methods. The use of 2 mCi of 227Ac in 0.45 mL (270 μM) We will also continue to expand the XANES and EXAFS gave a spectrum with a sufficiently large signal-to-noise data sets for other simple actinium complexes. We have ratio for the XANES and EXAFS measurements. For this already validated our ability to analyze extremely low particular sample, Ac3+ was dissolved in 0.1 M ammonium solution concentrations of actinium. Building on the initial acetate pH 5 buffer. These conditions mimic those used acetate data set, we will define the coordination chemistry to create 225Ac-labeled antibodies for medical use and of 227Ac with foundational ligand sets (e.g. H2O, Cl, Br, I). can therefore shed light on the speciation and nature These measurements will provide a basis for the analyzing of Ac3+ prior to conjugation. Preliminary analysis of more complicated actinium compounds ligated by the solutions showed the inflection point for the Ac LIII-edge lariat crown ethers or other more complex ligands. at ~15,870 eV, which is the first XANES measurement made on the element actinium. Future XANES experiments Conclusion are underway using Ac’s LIII-edge to learn about Ac’s The proposed work will elucidate fundamental chemical electronic structure and chemical bonding tendencies. It is properties of the element actinium to enable the rational possible that substantial edge-shifts (of 2 to 3 eV) may be design of actinium-225 complexes for targeted alpha observable and used to quantify orbital mixing in the Ac– therapy. A better understanding of the element, actinium, ligand interaction. The first ever actinium EXAFS spectrum will allow application of the specific therapeutic isotope, was also collected using our small stock of 227Ac. actinium-225, in medicine for treatment of cancer. Preliminary analysis of those data suggested that the major species present in solution was Ac(O2CMe)2(H2O)3. 74

Page 75

Publications Wilson, J. J., E. R. Birnbaum, E. R. Batista, R. L. Martin, and K. D. John. Synthesis and Characterization of Nitrogen- Rich Macrocyclic Ligands and an Investigation of Their Coordination Chemistry with Lanthanum(III). 2015. IN- ORGANIC CHEMISTRY. 54 (1): 97. Wilson, J. J., Ferrier, Radchenko, J. R. Maassen, J. W. Engle, E. R. Batista, R. L. Martin, F. M. Nortier, M. E. Fass- bender, K. D. John, and E. R. Birnbaum. Evaluation of nitrogen-rich macrocyclic ligands for the chelation of therapeutic bismuth radioisotopes. 2015. NUCLEAR MEDICINE AND BIOLOGY. 42 (5): 428. 75

Page 76

Complex Natural & Engineered Systems Exploratory Research Final Report Label-Free Measurement of Single Cells by Impedance Cytometry in a Microfluidic Device Babetta L. Marrone 20130239ER Abstract two electrodes. It has been demonstrated for cell count- Microfluidic-based devices have important applications ing, and for discriminating cells on the basis of size, and in biomedical diagnostics and bioprocess monitoring, differences in cell morphology [2]. Cell physical param- as well as in biotechnology research such as sequenc- eters like membrane capacitance can be calculated from ing and protein engineering. The main objective of this impedance measurements. By varying frequencies, infor- project was to design, build, and test a microfluidic mation about cell volume, and cytoplasm conductivity sensor device for counting and measuring cells. Cells and permittivity can be obtained. Single cell impedance will be manipulated in the device using Surface Acoustic analysis has been applied to measurement of cell vi- Wave (SAW) technology. Sensing will be based on single ability, and cell stage analysis [2-4]. It has also been used cell impedance cytometry, an emerging technology for measuring heterogeneity in bacterial populations that simply measures the current displaced by a cell as applied to bioprocess monitoring [5] and to differenti- it flows between two electrodes. The measurements ate between normal red blood cells (RBCs) and RBCs are label-free, and provide accurate information about infected with a parasite [6] using phase differences. The cell size and direct dielectric properties that are directly impedance response of cancer cells and comparable nor- related to functional state, such as lipid production or mal cells were found to be different in magnitude and proliferation stage. As part of this research we will de- phase [7] and three white blood cell populations were termine the underlying cell properties that contribute to discriminated based on their impedance responses[8]. In acoustic contrast between the cell and medium; and the some examples, measured differences in impedance re- underlying cell properties that contribute to differences sponse have been directly traced to re-structuring of the in impedance signals between cells. Model biological cytoskeleton, a process that is associated with disease systems to be studied will include: sensing and separat- states [9-16]. ing red and white blood cells and sensing of algae cells Impedance analysis is relatively easy to implement on a based on lipid content. The measurements made using microfluidic device, using integrated digital electrodes the sensing devices will be validated using standard flow [1-3]. However, accurate measurement of cell proper- cytometry methods. The primary application that will be ties in a microfluidic channel depends on consistent pursued in the future using this technology is monitor- positioning of the particle within the channel detection ing cell growth and lipid content in microalgae during volume [17]. Traditional methods of focusing single cells cultivation. The technology will have primary impact on in a microchannel use hydrodynamic focusing to confine the development of advanced biofuels and bioproducts and direct a slower flowing (sample) stream by a faster as supported by the DOE, Office of Energy Efficiency and flowing sheath stream [18]. A newly demonstrated ap- Renewable Energy, Bioenergy Technologies Office. proach to cell focusing in a microfluidic channel is using Surface Acoustic Wave technology (SAW), which propa- Background and Research Objectives gates a sound wave across the surface of a microchan- Single cell impedance cytometry is emerging as a leading nel. The SAW effect is an immediate and gentle means of edge approach for non-destructive, label-free analysis of manipulating cells to focus them into the measurement informative cell features because of its technical sim- volume or to deflect them for sorting. Shi et al. [19] plicity [1] as well as its versatility, i.e. its potential to be reported using standing SAW as a focusing mechanism in applied to a wide range of particle types and a broad set a microfluidic device, using interdigital transducers (IDTs) of applications. Impedance cytometry simply measures patterned on a piezoelectric substrate to generate the the current displaced by a single cell as it flows between 76

Page 77

acoustic field. In a recent Review by Xuan et al. [20] on dif- impedance measurement. (d) Top/bottom electrode arrange- ferent methods of particle focusing in microfluidic devices, ment for accurate impedance measurements. (e) Time resolved the primary advantage of SAW to manipulate particles was differential impedance measurement by this team of a single par- ticle at a single frequency. that the channel size can be flexible to accommodate a wide range of particle sizes from 1µm (bacteria) to large particles (800µm droplets) or cell clusters (e.g. circulat- Technical milestones ing tumor cell aggregates) or larger organisms in an algae Lab-on-a-Chip Device Development/Demonstration culture, such as rotifers. Another advantage is that by using SAW, sheath fluid is eliminated, thus reducing consumable • Successfully made microfluidics devices using a soft li- cost and waste from the device. Cheung et al. [4] report thography approach with Interdigital Electrodes (IE) via 11 papers on microfluidic particle focusing for impedance metal deposition and masking, in collaboration with sensing flow cytometry, but none use SAWs. This indicates LANL’s Center for Integrated NanoTechnology (CINT). that there is a significant opportunity to optimize micro- We finally settled on a device manufactured commer- fluidic impedance cytometry by combining this technology cially and made to our specifications, from Micronit with SAW technology for enhanced particle focusing and Microfluidics company (http://www.micronit.com/), sorting. Figures 2 and 3. Scientific Approach and Accomplishments • Used IEs to generate Surface Acoustic Waves (SAWs) There are three parts to the R&D approach in this project: for particle/biological cell separation and translocation. • Design/simulate and build a microfluidic device com- • Measured differential impedance particle/cell signals bining SAW (for focusing and sorting) with impedance with and without acoustic focusing for detection based sensing of cell properties. on size. • Use the device to study and validate the impedance • Measured impedance spectroscopy for particle/cell sensing and develop signatures by attributing the mea- property assessment with and without acoustic focus- surements to specific cell properties. ing. • Demonstrate the approach through application to a • Integrated the microfluidic device with the CCD camera relevant model system; we focused on counting and for full confirmation of measurement integrity. measuring microalgae cells during cultivation. Figure 1 shows a composite of the project, including the concept of impedance sensing, the expected signals from impedance sensing, the original concept of the impedance sensor combined with acoustic separation in the same microfluidic device; and an impedance spectrum from a single particle in our device. Figure 2. The microchannel device with impedance electrodes integrated and fabricated by Micronit using our design. Figure 1. (a) Differential impedance circuit concept, and (b) impedance spectrum [21] (Gawad et al, Lab on a Chip 1 (2001) 76). (c) LANL’s SAW concept for cell focusing and separation with 77

Page 78

an input two sets of training data: one with fluorescent labels for a chemical of interest (protein, RNA, lipids, etc.) and one without any labels. The script then searches for a linear or quadratic combination of the label free measure- ments that can reproduce the intensity distribution for the labeled target in the labeled cells. Once the script has learned the combination of features that predict the target features for the unlabeled cells, this prediction power is tested on additional data sets that the user can provide. The analysis to discover signatures in labels-free measure- ments is based upon simple and well known linear regres- Figure 3. Close-up of an impedance electrode array. The chan- sion approaches. The script can utilize a Genetic Algorithm nel is running horizontally in this image. There are 4 electrodes; (provided by MathWork’s MATLAB) to automatically the two shown above are on top of the channel. The two below remove features that are not informative or misleading for are on the bottom surface of the channel. Both sets extend the prediction of the target feature in the unlabeled cells. across the channel. The reference electrode is further upstream The script provides a Graphical User Interface for the user (out of vision). to manually add and remove features to be used in the Data Analysis label-free parameter identification. The script allows the user to change the file name for the inputs to the training • Developed a MATLAB code to identify single cell spec- and testing data. The script works in its most basic func- troscopic absorption/scattering measurements from an tionality, and an example data set is included which utilizes imaging flow cytometer as candidates for eliminating algae cells that increase their lipid content during nitrogen the use of chemical labels and tags in measurement of starvation. The script is not stand-alone. It requires Math- algae lipid accumulation. work’s Matlab and several routines in the Matlab Global Search Toolbox, Parallel Processing Toolbox, and Optimiza- • Developed MATLAB and LabView codes to measure tion Toolbox. differential impedance amplitude at single (and mul- tiple) frequencies, to establish the basis for impedance The code is highly preliminary, and a users manual and particle/cell size and property measurements in our several other functions need to be added. microfluidic devices. The outcome of this project is highly relevant to both the • Modeling: Constructed a COMSOL model of the electri- Energy Security and National Security missions of the cal field in a fluid-filled channel between two elec- Laboratory. We developed a new measurement modality trodes. and new set of signatures for investigating and identifying biological particles in real time for high-impact applications Applications in biofuel production, as well as medical diagnostics, drug discovery and screening, toxicology testing, and biomedi- • Demonstrated 1-D particle/cell focusing in a micro- cal research. The greatest impact of this simple, label-free, channel by surface acoustic waves generated from nondestructive approach is that it can be implemented on straight interdigital transducers (SIDT) small, low cost platforms, thus enabling introduction of • Demonstrated bacteria/blood cell separation in a mi- real-time sensing and measurement technologies into new crofluidic device by surface acoustic waves generated fields of bioenergy and medical diagnostics. from SIDTs. The research supports LANL’s Signature of Science pillar • Developed particle/biological cell detection and size by developing new signatures of cell phenotypes based on measurements by single-frequency impedance mea- their intrinsic physical properties. The project has helped surement in a microfluidics device. to build a new capability at LANL in Bioengineering and Microengineering. The capability to design and test lab-on- • Investigated impedance spectroscopy (multi-frequen- chip devices is crucial for increasing our competitiveness cy) on a particle/cell in a microfluidic device. for external funding opportunities from DOE, Bioenergy Technologies Office (for algae cultivation monitoring); and Impact on National Missions the National Institutes of Health and the Department of This script imports flow cytometry data and searches this Defense, who are actively seeking to develop miniaturized data to discover label-free signatures. The script takes as sensor devices for biomedical and biosecurity applications 78

Page 79

in support of National and Global Security mission chal- 13. Rosenbluth, M. J.. Analyzing cell mechanics in hema- lenges. tologic diseases with microfluidic biophysical flow cytometry. 2008. Lab on a Chip. 8: 1062. The team has relevant intellectual property under devel- opment. We will actively pursue technology transfer to 14. Moutzouri, A. G.. Severe sepsis and diabetes mellitus industry, working with the Feynman Center for Innovation have additive effects on red blood cell deformability. and directed to both the biofuel and medical diagnostic 2008. J. Infection. 57: 147. industries. 15. Khismatullin, D. B.. The cytoskeleton and deformability References of white blood cells. 2009. Current Topics in Mem- branes. 64: 47.

1. Capua, R. Di. Towards the realization of label-free bio- sensors through impedance spectroscopy integrated
2. Diez-Silva, M.. Shape and biomechanical characteristics with IDES technology. 2012. Eur. Biophys. J.. 41: 249. of human red blood cells in health and disease. 2010. MRS Bulletin. 35: 382.
3. Sun, T., and H. Morgan. Single-cell microfluidic imped- ance cytometry: A review. 2010. Microfluid. Nano-
4. Spencer, D., and H. Morgan. Positional dependence of fluid.. 8: 423. particles in microfluidic impedance cytometry. 2011. Lab on a Chip. 11: 1234.
5. Gou, H. L.. Label-free electrical discrimination of cells at normal, apoptotic, and necrotic status with a micro-
6. Golden, J. P.. Hydrodynamic focusing-A versatile tool. fluidic device. 2011. J Chromatography A . 1218: 5727.
7. Anal. Bioanal. Chem.. 402: 325.
8. Cheung, K.. Microfluidic impedance-based flow cytom-
9. Shi, J. D.. Acoustic tweezers: Patterning cells and etry. 2010. Cytometry Part A . 77A: 648. microparticles using standing surface acoustic waves (SSAW). 2009. Lab on a Chip. 9: 2890.
10. David, F.. Viability and membrane potential analysis of Bacillus megaterium cells by impedance flow cytome-
11. Xuan, X.. Particle focusing in microfluidic devices. 2010. try. 2011. Biotechnology and Bioengineering . 10: 483. Microfluid. Nanofluid.. 9: 1.
12. Kuttel, C.. Label-free detection of Babesia bovis infect- Publications ed red blood cells using impedance spectroscopy on Ai, Y. e., and B. L. Marrone. Droplet translocation by fo- a microfabricated flow cytometer. 2007. Acta Tropica. cused surface acoustic waves. 2012. MICROFLUIDICS 102: 63. AND NANOFLUIDICS. 13 (5): 715.
13. Han, A.. Quantification of the heterogeneity in breast Ai, Y. e., and B. L. Marrone. Separation of Biological Cells cancer cell lines using whole-cell impedance spectros- in a Microfluidic Device Using Surface Acoustic Waves copy. 2007. Clin. Cancer Res.. 13: 139. (SAWs). 2014. MICROFLUIDICS, BIOMEMS, AND MEDI- CAL MICROSYSTEMS XII. 8976.
14. Holmes, D.. Leukocyte analysis and differentiation using high speed microfluidic single cell impedance Ai, Y., C. K. Sanders, and B. L. Marrone. Separation of Esch- cytometry. 2009. Lab on a Chip. 9: 2881. erichia coli bacteria from Peripheral Blood Mononucle- ar Cells (PBMC) using Standing Surface Acoustic Waves
15. Lincoln, B.. Deformability-based flow cytometry. 2004. (SSAW). 2013. Analytical Chemistry. 85 (19): 9126. Cytometry Part A . 59A: 203. Marrone, B. L.. Flow cytometry in algae biofuels and by-
16. Suresh, S.. Connections between single-cell biome- products research. Invited presentation at 4th Inter- chanics and human disease states: Gastrointestinal national Conference on Algal Biomass, Biofuels and Bioproducts. (Santa Fe, NM, 15-18 June, 2014). cancer and malaria. 2005. Acta Biomaterialia . 1: 15. Marrone, B. L., Y. Ai, and J. E. Coons. Cyto-AMP Cell Acous-
17. Suresh, S.. Mechanical response of human red blood tic Manipulation Platform. 2014. R&D 100 Award, 2013 cells in health and disease: Some structure-property- Nomination LA-UR-14-21909. function relationships. 2006. J. Mater. Res. . 21: 1871. Munsky, B., S. Sellars, et.al. Automated label free classifica-
18. Suresh, S.. Biomechanics and biophysics of cancer tion of algal lipid accumulation. Presented at 4th In- cells. 2007. Acta Biomaterialia . 3: 413. ternational Conference on Algal Biomass, Biofuels and Bioproducts . (Santa Fe, NM, 15-18, June 2014). 79

Page 80

Complex Natural & Engineered Systems Exploratory Research Final Report A New Hypothesis to Explain the Variability of the Outer Radiation Belt: Can we Predict Post-storm Fluxes of Energetic Electrons Based only on Pre-storm Fluxes of the Lower-energy Population? Gregory S. Cunningham 20130297ER Abstract Background and Research Objectives Earth is surrounded by two radiation belts containing The radiation belt modeling community has a qualita- MeV ‘killer’ electrons that are trapped in earth’s geo- tive understanding of phenomena that can affect flux magnetic field. MeV electrons can deposit dose in sensi- levels of MeV electrons in the outer radiation belt. Solar tive satellite instruments, rendering those instruments output ultimately delivers the energy needed to cause or even entire satellites inoperable. Fluxes of MeV variability of MeV electron fluxes in the outer radiation electrons in the outer radiation belt can vary by orders belt, but the details are still being integrated into quan- of magnitude during geomagnetic events driven by the titative models. For many years, it was thought that sun, and the scientific community has a poor ability to inward radial diffusion of lower-energy electrons was predict the changes that are observed. An ability to reli- the dominant energization mechanism that creates MeV ably predict fluxes in the outer radiation belt driven by electrons, and indeed the equilibrium two-belt structure changes in the sun’s output would be of great utility to of inner magnetospheric electrons was well-described satellite operators, who could anticipate and accurately by radial diffusion in the 1970’s. However, our team at attribute anomalies to space weather as opposed to a LANL argued in recent years that electrons in the outer hostile attack by an adversary. belt must be accelerated locally by wave-particle interac- tions in order to produce the observed peaks in phase- In our LDRD ER project, we tested a hypothesis that ex- space density (PSD) versus radial distance from earth, plains the variability of outer radiation belt MeV electron since radial diffusion can only produce a monotonically fluxes. The hypothesis was that outer-belt fluxes of MeV increasing PSD versus radial distance from earth [3]. electrons are determined by the source population of Other groups had shown prior to our LDRD project that lower-energy electrons that are convected earthward chorus waves were capable of accelerating electrons, by electric fields, and diffused upward in energy (ac- and appeared to be the only effective mechanism for celerated) through interaction with electromagnetic the degree of acceleration that is needed to produce the waves. Given the source population and intensity of peaks in phase space density; however, it had not yet electromagnetic waves, we believed that we could been proven with quantitative calculations that chorus quantitatively predict the acceleration and hence the waves were responsible for recovery of fluxes after drop- fluxes of MeV electrons that would result. Underlying as- outs during geomagnetic events. This is the hypothesis sumptions in the hypothesis were that 1) MeV electrons that we tested in the project. are first wiped out at the start of the event, and 2) the effect of the electromagnetic waves that accelerate the Scientific Approach and Accomplishments lower-energy source population to higher energies can We were able to test the hypothesis that chorus waves be predicted using quasilinear theory for specific events are responsible for creating the MeV electron popula- using an empirical model. The project was successful in tion in the outer belt during geomagnetic events using 1) showing that if the low-energy source population and event-specific measurements of the low-energy source intensity of electromagnetic waves is known, then the population (100 keV electrons) and 2) event-specific acceleration of the source population to MeV ener- measurements of the intensity of the chorus waves, to gies can be predicted for a specific event. Because the drive a LANL computational model called DREAM3D, process of acceleration takes hours to days, this enables which computes the temporal evolution of the phase forecasting of the hazardous MeV population hours to space density of electrons in response to interactions of days in advance, an important capability. the electrons with ultra low-frequency (ULF) and very 80

Page 81

low-frequency (VLF) waves that interact resonantly with models in DREAM3D using the Van Allen Probes. 2014. the periodic motion of the electrons as they gyrate around Geophysical Research Letters. 41 (5): 1359. field lines and drift around earth. The time-varying source 2. Chen, Y., G. Reeves, R. Friedel, and G. S. Cunningham. population was measured by the MagEIS instrument on Global time-dependent chorus maps from low-Earth- the Van Allen Probes mission (nee Radiation Belt Storm orbit electron precipitation and Van Allen Probes data. Probes), which was launched in August 2012 just prior to 2014. Geophysical Research Letters. 41 (5): 755. the start of the LDRD project. The measured source popu- lation was used as a time-varying low-energy boundary 3. Tu, W., G. S. Cunningham, Y. Chen, M. G. Henderson, condition in the DREAM3D diffusion code. The event-spe- E. Camporeale, and G. D. Reeves. Modeling radiation cific intensity of chorus waves was inferred indirectly from belt electron dynamics during GEM challenge intervals measurements of precipitating low-energy electrons at low with the DREAM3D diffusion model. 2013. Journal of altitude provided by the Polar-Orbiting Environmental Sat- Geophysical Research-Space Physics. 118 (10): 6197. ellite (POES), normalized to actual electromagnetic wave measurements at high altitude by the Van Allen Probes [2]. Publications The wave intensities were extrapolated to all latitudes us- Chen, Y., G. D. Reeves, G. S. Cunningham, M. G. Henderson, ing statistical models binned by geomagnetic activity. Using and R. J. Redmon. Forecasting and Remote Sensing the event-specific source population and wave intensities, Outer-belt Relativistic Electrons from Low-Earth-Orbits. we were able to show that the enhancement of MeV elec- Geophysical Research Letters. trons that was observed by the Van Allen Probes during the October 2012 storm was due to acceleration by chorus Chen, Y., G. Reeves, R. Friedel, and G. S. Cunningham. waves, thus confirming our hypothesis [1]. We also showed Global time-dependent chorus maps from low-Earth- orbit electron precipitation and Van Allen Probes data. that using a purely statistical model for either the source 2014. Geophysical Research Letters. 41 (5): 755. population or the waves does not produce the enhance- ment, making event-specific models critical to the success- Cunningham, G. S.. Radial diffusion of radiation belt par- ful test of the hypothesis. ticles in non-dipolar magnetic fields. Journal of Geo- physical Research-Space Physics. Impact on National Missions The work done in the LDRD ER project was an outgrowth Ripoll, J. F., J. M. Albert, and G. S. Cunningham. Electron lifetimes from narrowband wave-particle interactions of the Dynamic Radiation Environment Assimilation Model within the plasmasphere. 2014. Journal of Geophysical (DREAM) project, which began as an LDRD/DR ~10 years Research-Space Physics. 119 (11): 8858. ago. DREAM has spun off several projects, including a $6M 18-month study of HANE impact on low-earth orbit- Tu, W., G. S. Cunningham, Y. Chen, M. G. Henderson, E. ing (LEO) assets that supported the development of the Camporeale, and G. D. Reeves. Modeling radiation belt 3D diffusion code, DREAM3D, that was used in this LDRD electron dynamics during GEM challenge intervals with ER project. The fact that, in this LDRD ER project, we the DREAM3D diffusion model. 2013. Journal of Geo- were able to show that enhancement of MeV electrons physical Research-Space Physics. 118 (10): 6197. following a geomagnetic storm is due to chorus waves, Tu, W., G. S. Cunningham, Y. Chen, S. K. Morley, G. D. and that we can quantitatively predict the enhancement Reeves, J. B. Blake, D. N. Baker, and H. Spence. Event- given event-specific information on the low-energy source specific chorus wave and electron seed population population and chorus wave intensities, means that it may models in DREAM3D using the Van Allen Probes. 2014. be possible to forecast the buildup of MeV electrons in the Geophysical Research Letters. 41 (5): 1359. outer radiation belt a few hours to days in advance. Given LANL’s heritage in space environment instrumentation on Zheng, L., A. A. Chan, T. P. O’Brien, W. Tu, G. S. Cunning- geosynchronous, GPS, and scientific satellites, the poten- ham, J. M. Albert, and S. R. Elkington. Effects of mag- netic drift shell splitting on electron diffusion in the tial to use the data from these satellites to build a space radiation belts. Geophysical Research Letters. weather prediction capability with DREAM3D would seem to be at hand, and we are actively pursuing funding to build this capability. References

1. Tu, W., G. S. Cunningham, Y. Chen, S. K. Morley, G. D. Reeves, J. B. Blake, D. N. Baker, and H. Spence. Event- specific chorus wave and electron seed population 81

Page 82

Complex Natural & Engineered Systems Exploratory Research Final Report Multidisciplinary Studies of Long Non-coding RNAs: Towards a Structural Basis for RNA in Epigenetics Karissa Y. Sanbonmatsu 20130319ER Abstract single-molecule co-localization imaging) will help us lead Epigenetics focuses on changes in gene expression, this field, laying the structural foundation for lncRNA which are passed on to subsequent generations of cells epigenetics. without changing DNA sequence. The field has revolu- Background and Research Objectives tionized traditional notions of genetics and inheritance, producing new insight into environmental response, RNA (ribonucleic acid),DNA’s molecular cousin, is simi- cell differentiation, and hereditary disease. Two major lar to DNA in composition; however, it often performs modification events, site- specific DNA methylation a function in the cell, unlike DNA, which is considered and histone modification on chromatin, alter the gene solely an information carrier. In all living systems, includ- expression profiles. Nevertheless, the specifics of how ing humans and bacteria, the information for the blue DNA and histone-modifying enzymes find these unique print of the organism resides in DNA. The blue print locations on chromatin remain a mystery. Long non-cod- is implemented by transferring the genetic informa- ing RNAs (LncRNAs or lincRNAs, pronounced ‘link RNAs’) tion from DNA to RNA. Next, the genetic information is have emerged in the past 2 years as major players in converted from the RNA into proteins, which make up epigenetics, recruiting many factors to chromatin. These the structures of the cells and perform the biochemistry. RNAs are interesting in their own right: they are as big RNA has long been thought to mainly serve as a transfer- as the ribosome, have unknown structure, and are mis- ring medium, moving information from DNA to proteins. sion critical for a large number of biological processes. However, in the past five years, it has been found that The number of lncRNAs rivals the number of proteins, the vast majority of the human genome (more than conjuring an image of a nucleus teaming with RNA- 90%) is converted into RNA, but never gets converted based molecular machines. Harvard and MIT have led into proteins. It has been found that much of this RNA the recent explosion in cell-biology studies of lncRNAs; is in the form of long non-coding RNA (lncRNA, or ‘link’ however, to date, there are no structural studies. With RNA). These gigantic RNAs do not ‘code’ for protein, but our decade of experience in structural studies of the ri- instead are thought to have a function themselves: turn- bosome and state-of- the-art expertise in RNA biochem- ing on and off other genes. The RNAs are often associ- istry, LANL is in a prime position to lead this new, high ated with epigenetics. Epigenetics has exploded in the impact field. We have recently published a large amount past five years and is closely related to stem cell pro- of preliminary data on lncRNA experiments taking the gramming. In epigenetics, Larmarkian-like effects occur, first steps in lncRNA structural studies. In this project, where the environment modifies, but does not mutate we will focus on four high impact systems and address DNA. These modifications are passed down to future fundamental unanswered questions: generations of organisms. For example, babies undergo- ing extreme stress shortly after birth have an altered (1) Are lncRNAs structured?, (2) Do they have sub- stress response, which is passed down to the grand- domains?, (3) Where on lncRNAs do epigenetic proteins children. While many long RNAs have been found to be bind?, and (4) Do the binding domains contain a tertiary critical for cancer, hereditary disease, brain function and core? development, the mechanism of long RNA action is not understood. We have successfully performed and pub- An arsenal of innovative experimental and computa- lished the first ever structural study of long RNAs and tional techniques (including SHOT-GUN chemical prob- laid the foundation for mechanistic studies. ing, structure-based molecular simulation, and in vivo 82

Page 83

Scientific Approach and Accomplishments large systems, 2-D structures are impossible to accurately We accomplished our goals. We were able to complete the determine computationally due to the large number of first secondary structure of an intact long non-coding RNA, permutations and possible structures. We have devised a resulting in several publications and invited talks. We em- novel experimental technique called SHOT-GUN structure ployed our novel shotgun secondary structure (3S) meth- determination. We used this to obtain the 2-D structure od, which did garner significant attention in the RNA com- experimentally. munity (especially at the recent RNA 2013 conference). Impact on National Missions In this method we perform several rounds of structural chemical probing. In chemical probing, a special reagent We have produced preliminary data for NIH calls in cancer- is added that only reacts with highly mobile nucleotides, related non-coding allowing us to map which nucleotides are base paired in RNAs and several calls for NIH’s epigenomics program for double helices and which nucleotides are single stranded. epigenetic mark discover, including the role of noncod- Thus, we perform 1 round of chemical probing on the ing RNAs in regulation of transcription (Jerome Garcia). entire long non-coding RNA. We then perform successive The project is directly related to DARPAs CLIO Memory rounds on fragments of the RNA. If we obtain a match thrust, including “basic research towards use of epigenetic between signals for the full RNA and its fragments, this systems to report environmental events” (Cathy Cleland). It demonstrates that the fragments have a modular fold and is indirectly related to DOE BER Biological Systems Science comprise modular subdomains. By probing smaller and “Low dose radiobiology effects on epigenetic regulation”. smaller fragments, we are able to determine the complete The ER project culminated with an invited talk at the Royal fold of the long noncoding RNAs, a hard problem, which Society in England given September 29, 2015. In addition, has been long sought after. To verify the fold, we per- a TEDxABQ talk was given for the general public in Septem- formed comparative studies. ber 2014 which currently has ~500,000 views on YouTube. We applied our structure technique (shotgun secondary The talk was directly related to this ER project. These two structure) to several more long non-coding RNA systems. presentations have helped raise the visibility of LANL in the The first system is COOLAIR, a canonical epigenetic switch life sciences. in plants that allows plants to flower only after prolonged Publications exposure to cold. The second system is Braveheart, a long non-coding RNA that plays a key role in heart cell Computational & Experimental Studies: Ribosomes, Ribo- switches & Long non-coding RNAs. Invited presenta- development. The third system is Gas5, which is critical tion at Fourth Annual Summer Symposium on Cellular for hormone signaling. To accomplish this, we transcribed Dynamics of Macromolecular Complexes. (Montreal, the RNA for each system, denatured the RNA, folded the May 2014). RNA and chemically probed the RNA. In chemical prob- ing, a special reagent is added that only reacts with highly Structural Studies of Intact LncRNAs in Plants and Mam- mobile nucleotides, allowing us to map which nucleotides mals. Invited presentation at Long non-coding RNAs: are base paired in double helices and which nucleotides Marching towards Mechanism Keystone meeting. are single stranded. In this way, we determined candidate (Santa Fe, March 2014). secondary structures of the long non-coding RNAs. Hayes, R. L., J. K. Noel, A. n. a. Mandic, P. C. Whitford, K. Y. Sanbonmatsu, Mohanty, and J. N. Onuchic. General- In the past year, we finalized the secondary structures ized Manning Condensation Model Captures the RNA of COOLAIR, Braveheart, and Gas5, which is critical for Ion Atmosphere. 2015. PHYSICAL REVIEW LETTERS. hormone signaling. We validated these folds in vivo in 114 (25). collaboration with Laurie Boyer (MIT) and Caroline Dean (John Innes Centre). We completed a study on a new sys- Hayes, R. L., J. K. Noel, P. C. Whitford, Mohanty, K. Y. San- tem called HULC that is a long non-coding RNA thought to bonmatsu, and J. N. Onuchic. Reduced Model Captures interact with microRNAs. The three manuscripts for the 3 Mg2+ -RNA Interaction Free Energy of Riboswitches. systems are almost ready for submission. 2014. BIOPHYSICAL JOURNAL. 106 (7): 1508. In conclusion, we have performed the first structural Hayes, R. L., J. K. Noel, P. C. Whitford, Mohanty, K. Y. San- bonmatsu, and J. N. Onuchic. Reduced Model Captures study of long non-coding RNAs (ribonucleic acids). The Mg2+ -RNA Interaction Free Energy of Riboswitches. only example of a large RNA complex whose 3-D structure 2014. BIOPHYSICAL JOURNAL. 106 (7): 1508. has been solved is the ribosome, which took 3 decades of work. We took the first step in structural studies of long Hennelly, S. P., I. V. Novikova, and K. Y. Sanbonmatsu. The non-coding RNAs by determining the 2-D structure. For 83

Page 84

expression platform and the aptamer: cooperativity Sanbonmatsu, K. Y.. Towards structural classification of between Mg2+ and ligand in the SAM-I riboswitch. long non-coding RNAs. To appear in Biochimica et Bio- 2013. NUCLEIC ACIDS RESEARCH. 41 (3): 1922. physica Acta (BBA)-Gene Regulatory Mechanisms. 1 (1): 1. Novikova, I. V., A. Dharap, S. P. Hennelly, and K. Y. Sanbon- matsu. Shotgun secondary structure determination of Sanbonmatsu, K. Y.. Structural architecture of long non- long non-coding RNAs. 2013. Methods. 63 (170): 1. coding RNAs in plants and animals. Invited presen- tation at Long non-coding RNAs: evolution of new Novikova, I. V., S. P. Hennelly, C. S. Tung, and K. Y. San- epigenetic and post-transcriptional functions. (Royal bonmatsu. Rise of the RNA machines: Exploring the Society, UK, 28-29 Sept. 2015). structure of long non-coding RNAs. 2013. Journal of Molecular Biology. 425 (19): 3731. Novikova, I. V., S. P. Hennelly, and K. Y. Sanbonmatsu. Do lncRNAs have secondary and tertiary structure?. 2012. RNA BIOLOGY. 9 (12): 1398. Novikova, I. V., S. P. Hennelly, and K. Y. Sanbonmatsu. Struc- tural architecture of the human long non-coding RNA, steroid receptor RNA activator. 2012. NUCLEIC ACIDS RESEARCH. 40 (11): 5034. Novikova, I. V., S. P. Hennelly, and K. Y. Sanbonmatsu. Sizing up long non-coding RNAs: Do lncRNAs have secondary and tertiary structure?. 2012. Bioarchitecture. 2 (6): 189. Novikova, I. V., S. P. Hennelly, and K. Y. Sanbonmatsu. Experimentally determined secondary structures of cancer-related long non-coding RNAs. Invited presenta- tion at 245th ACS National Meeting & Exposition. (New Orleans, 7-11 April, 2013). Novikova, I. V., S. P. Hennelly, and K. Y. Sanbonmatsu. Tack- ling Structures of Long Noncoding RNAs. 2013. INTER- NATIONAL JOURNAL OF MOLECULAR SCIENCES. 14 (12): 23672. Sanbonmatsu, K. Y.. Rise of the RNA Machines: Long Non- Coding RNAs. Invited presentation at 10th Horizons in Molecular Biology Conference International. (Max Planck Institute, Gottingen, Germany, 9-12 Sept. 2013). Sanbonmatsu, K. Y.. Integrating simulations and experi- ments of the SAM-I riboswitch. Invited presentation at RNA Dynamics. (Telluride, CO, 22-26, July 2013). Sanbonmatsu, K. Y.. Large-scale simulations of biomolecu- lar machines. Invited presentation at Biosupercomput- ing. (Tokyo, Japan, Dec. 2012). Sanbonmatsu, K. Y.. Understanding the atomistic mecha- nism of magnesium effects of RNA dynamics. Invited presentation at Mini-symposium on Modeling, Simula- tion and Function of Biomolecular Assemblies.. (Uni- versity of Tokyo, Tokyo, Japan, Nov. 2012). Sanbonmatsu, K. Y.. Dynamics of riboswitches: Molecular simulations. 2014. BIOCHIMICA ET BIOPHYSICA ACTA- GENE REGULATORY MECHANISMS. 1839 (10): 1046. 84

Page 85

Complex Natural & Engineered Systems Exploratory Research Final Report How Trees Die: Multi-scale Studies of Carbon Starvation and Hydraulic Failure during Drought Sanna A. Sevanto 20130442ER Abstract Our species, adapted to low and unpredictable precipita- In this project we have developed a portable Nuclear tion, do not use this extra reserve, but conserve water Magnetic Resonance (NMR) system for measuring whenever possible. changes in tree water content in field conditions, and Background and Research Objectives used the system to monitor water dynamics in living trees during drought and a leaf senescence-bud break Vegetation mortality events are observed and predicted cycle. The system uses light and inexpensive electro- to worsen with future climate [1], but our ability to magnets instead of permanent magnets and operates predict where and when these could occur is rudimen- at ultra-low magnetic fields (<µT). These features allow tary [2]. At the heart of this incapability is a lack of basic operation in the natural setting and environment of understanding of fundamental physiological processes trees, and use of the system with other instrumentation relating water and carbon use and transport to plant without interference. We have developed a temperature function and mortality [3] due largely to a lack of tools correction protocol for the NMR signal to allow opera- to monitor these processes in intact plants [3, 4]. The tion without external temperature control, and success- key metabolic and hydraulic failure mechanisms leading fully used the system for measuring plant water uptake to plant mortality have been identified theoretically [3, simultaneously with neutron radiography visualization 4], but tests have been impossible due to a lack of non- of water movement in trees, as well as leaf gas exchange invasive measurement techniques. Revealing the mecha- and diurnal stem diameter variation measurement. The nisms and irreversible processes leading to plant mortal- visualization experiments confirmed that the majority ity requires a non-invasive method that simultaneously of our NMR signal results from water moving fast in the detects changes in water and carbohydrate content conduits of the xylem, the tissue specialized in water of the key tissues, namely the xylem (water transport transport. These experiments also led to identification tissue) and the phloem (sugar transport tissue). These of a new parameter, hydraulic conductivity between the tissues strongly interact in a functional plant, but they xylem and phloem (tissue responsible for carbohydrate operate under opposite pressure gradients and invasive transport), that seems to be a key to identifying the measurements alter their state and operation, affecting most likely tree mortality mechanism during drought. the results obtained. Mortality mechanisms influence the expected lifetime NMR and MRI are ubiquitous tools for non-invasive stud- under stress, and therefore identifying them improves ies of soft-tissue anatomy and function in medicine and predictions of forest mortality. Our drought experiments biology [5]. Atomic nuclei possessing an odd number show that the NMR system can detect gas-filled water of nucleons (e.g. 1H or 13C) can be magnetized (polar- conduits in the xylem, but the arid-zone tree species ized) in an externally applied field, and will emit and that we used do not allow gas bubble formation in their absorb electromagnetic radiation at a frequency specific xylem on a regular basis at an observable scale. Gas to the isotope and the strength of the applied magnetic bubble formation in the water conduits is a hypothesized field. By manipulating this magnetization, one can non- mechanism for tree mortality during drought. When invasively detect and analyze the chemical structure of excessive it could lead to a catastrophic failure of the materials. The frequency of the NMR signal depends water conducting system and consequent mortality of on magnetic field strength. Thus a magnetic field gradi- trees. Species at moist environments, however, can also ent can be used to spatially encode the NMR signal and use it as an extra water reserve to boost photosynthesis produce structural images (MRI) or track the move- in a controlled manner during short drought episodes. 85

Page 86

ment of materials. However, because the strength of the verse proportionality of the NMR signal strength to sample NMR signal is proportional to the applied magnetic field, temperature in the natural temperature range [11]. Using typically NMR is performed in very high (> 1T) magnetic this temperature correction we are now able to separate fields requiring specialized permanent or superconduct- the plant physiological signal from the temperature signal ing magnets housed in buildings rendering the technique and detect stomatal closure in plants under drought (Fig 2) somewhat restrictive for in vivo applications with plants. as long as the air temperature is measured simultaneously In laboratory conditions NMR and MRI have been used with NMR. to non-invasively study the distribution and transport of water in different plant tissues [6, 7]. There has also been growing interest in specially designed “plant focused” sys- tems that adapt to the plant, maintaining a vertical geom- etry and not inhibiting the plant’s access to light [8, 9] but all these systems still require rather large and exotic high strength permanent magnets (to increase signal) that make it difficult to measure many plants (or locations on a plant) simultaneously for high throughput studies, or studies in the greenhouse or field. Our goal was 1) to develop the Ultra-Low-Field (ULF) NMR technique so that we can accurately quantify water and carbon dynamics from tissue level to whole-plant scales Figure 1. A schematic presentation and a photograph of our in the greenhouse and in the field over time scales from low field NMR system for measuring water content changes in minutes to months; and 2) to use this tool together with trees shows (a) the 41 cm diameter Helmholtz coils which pro- traditional plant physiological measurements to test duce the magnetic field; b) copper shielding used to block radio hypotheses about water allocation during drought and frequency (rf) noise with one panel removed for clarity; c) the progress of plant mortality. solenoid which both excites the 1H nuclei and detects the rf NMR signal their relaxation produces. The solenoid is wrapped around Scientific Approach and Accomplishments a former on the tree. To reach our goals we employed a three-stage process. At the first stage we developed the special tree NMR system (Fig 1), calibrated it against known water content changes in stems of different tree species, and developed a tem- perature correction model for the signal [10]. The temperature of the magnet as well as the sample affects the NMR signal strength. Therefore, NMR measure- ments are usually done in temperature-controlled environ- ments where the magnet is cooled actively. We aimed at building a portable NMR system for field use, and a cooling Figure 2. The average diel NMR signal without temperature system would have been impractical adding to the power correction (a) and with temperature correction (b) for a well demands in the field. Natural temperature cycles thus watered (black line) and water limited (red line) period in our affected our signal. These cycles correlate with any plant greenhouse experiments. Without temperature correction, the physiological responses, such as stomatal opening and water-limited signal resembles well-watered signal contradicting transpiration rate that depend on temperature or incoming the information on water transport and use rates measured with radiation making separating the temperature-induced part other plant physiological method. Once temperature correction of the NMR signal from the plant physiological response is applied, the water-limited period clearly shows no changes essential if NMR is to be used for plant physiological stud- in water content supporting the data obtained with other plant physiological measurements. ies. To characterize the temperature responses of our system, we attached a second solenoid near the detector At the second stage, we focused on improving our system and applied a constant AC signal to it. By observing the to provide information on the contributions of xylem and variation in this test signal over one week, we created phloem tissues, and water bound to the cell walls vs. water a temperature correction function that adds the linear moving freely in the conduits. Water, first allocated to temperature responses of our electronics to the known in- phloem tissue and then taken out to support xylem func- 86

Page 87

tion as drought progresses, could be an indicator of when sue anatomy of these plants [16]. the tree becomes dysfunctional, even before visible signs Combined with x-ray tomography analysis of sample of degradation, such as leaf browning or wilting, appear anatomy, performed at Paul Scherrer Institute, these [12]. Changes in proportions of bound and free water in experiments showed that the majority of our NMR signal the xylem could be indicators of bubble formation in the comes from water in the xylem (Fig 3). They also indicated water conduits. Excessive bubble formation during drought that, as predicted by our theoretical work on hydraulic is one of the hypothesized mortality mechanisms of plants, coupling between the xylem and the phloem [16], pinon but currently there are no methods for detecting this non- pine has a higher hydraulic conductivity between the destructively in field conditions. During drought, bubble xylem and the phloem than one-seed juniper (Fig 3). This is formation in xylem conduits (cavitation) could become in line with the stomatal control strategy of transpiration of excessive if soil water availability does not match transpira- these species. One-seed juniper that allows xylem dehy- tion rate. The bubbles could block the water pathway from dration during drought protects the living phloem cells the soil to the leaves and lead to a hydraulic failure and from turgor loss with a low hydraulic conductivity between mortality of the plant [13]. the xylem and the phloem [16]. These results also show In NMR the relaxation of magnetization after excitation that the hydraulic conductivity between the xylem and can be described by a two-component exponential decay the phloem is one of the key parameters to understanding function with two different time constants. The shorter the interaction of carbon and water transport in plants, time constant (T1) is related to the proportion of water and the susceptibility of the plant to different mechanisms bound to the cell walls or living cells, the longer time of mortality during drought [16]. Our approach, which constant (T2) reflects the amount of freely moving water combined NMR and neutron radiography measurements, in the conduits [14]. Changes in these time constants thus allows, for the first time, direct, in-vivo measurements of contain information on water distribution in the sample. this parameter, which is very difficult to measure with tra- They also could contain signals of bubble formation in the ditional methods [17, 18]. Our tests for detecting bubble xylem, which would both reduce the signal strength (fewer formation in the branches were inconclusive. We managed hydrogen atoms in the sample) and reduce the amount of to produce a decline in the NMR signal from a branch using free water in the sample. the surfactant once, but that signal was not strong enough to be reflected in the neutron radiography images, and To measure these time constants, we improved our system despite several re-trials, we did not manage to reproduce by increasing the magnetic field strength so that the time the result in NMR. constants could be determined separately. We also built a replica of the system for short experiments, while the first system was used for long-term drought experiments in a greenhouse and in the field (see stage 3 below). In the short-term experiments we measured water move- ment in branches of pinon pine and one-seed juniper simultaneously with our NMR system, leaf gas exchange measurements and neutron radiography at Paul Scher- rer Institute in Switzerland and at LANSCE in Los Alamos. To calibrate our detection limits for bubble formation, we Figure 3. The NMR (black) and neutron (green measured also used a surfactant to promote bubble formation in below and red above the NMR detector coil; see Figure 4) signal intensity for juniper (a) and pinon (b) branches. In both cases the the conduits [15]. In the neutron radiography measure- signal intensity declined when D2O replaced H2O in the tissues. ments we used heavy water (D2O) as a tracer that both The NMR signal saturated at the low level (47% of the original in NMR and neutron signal intensity react to. Pinon pine and juniper and 27% in pinon) in both species while the neutron sig- one-seed juniper were selected for these experiments nal did not saturate in pinon even after 14 hours of experiment. because they are two co-existing species that have a very The neutron signal is directly proportional to sample thickness different stomatal response strategy to drought. Pinon pine whereas the NMR signal is focused on the location of maximum closes stomata very early during drought, allowing very magnetic field strength. The difference in saturation of these little variation in xylem hydration state. One-seed juniper, signals indicates that the majority of the NMR signal originates on the other hand, keeps the stomata open longer during at the water conductive tissue (xylem), while the neutron signal drought and allows its xylem to dry out and theoretically sees the living tissue (bark and phloem) surrounding the xylem. Two different slopes representing two compartments of water risks more bubble formation in its conduits [13]. These can be identified in the decline of the neutron signal for both spe- strategies should be reflected in water distribution and tis- 87

Page 88

cies. The first one (slope I) represents fast filling compartment in coils with silicone grease. During these measurements the the xylem; the second one slow filling in the living cells. In juniper tree was also tied with guy-wires and ropes to eliminate the first slope is much steeper than in pinon suggesting lower noise caused by wind shaking the tree. Traditional plant hydraulic conductivity between xylem and phloem tissues. physiological parameters, such as leaf gas exchange, leaf water potential and soil moisture content were measured To further study the capabilities of NMR for detecting regularly during these experiments to allow for linking the bubble formation in the xylem we performed experiments changes in NMR signal with changes in plant physiology. with saturated and unsaturated cellulose sponges with We also erected a small micrometeorological weather known water content using a powerful commercial NMR station in the vicinity of these experiments to measure system (NMR-MOUSE®, Magritek, San Diego, CA) to al- variation in environmental drivers of plant function such as low for higher resolution in time constant determination. photosynthetically active solar radiation, air temperature These experiments revealed that separating the effects relative humidity, and precipitation. of bubble formation from the effects of elastic shrinkage in the NMR signal amplitude or time constants requires The drought experiment consisted of observing plant very large scale bubble formation to occur. Tree stems responses to several degrees of drought severity from are elastic and shrink and swell diurnally with changing well-watered to a girdled tree to speed up xylem dehydra- transpiration rate [19]. This serves as elastic water reserve, tion in the greenhouse. In the field the drought period was balancing any mismatch of transpirational demand and maintained until the tree lost all its leaves. After that wa- water availability from the soil. If this reserve is consumed, tering was resumed and NMR and stem diameter signals and the soil water availability does not match the transpi- observed until new leaves were fully grown. Losing leaves ration rate, trees risk hydraulic failure (cavitation in xylem and producing new ones induces a major carbon and water conduits). On the other hand, some bubble formation, if reallocation event where resources are transported from controlled and repairable, can act as an extra water source storage to the actively growing parts. These experiments for the plant and boost photosynthetic productivity during thus served both as tests of NMR signal responses to short periods of drought [20]. The balance between sever- changing water content and reallocated carbon. ity of, and benefit from, bubble formation is one of the hot debate topics in plant physiology and vegetation mortality Our results from the greenhouse show ceasing of the [4]. Unfortunately, our experiments showed that studying diurnal use of the elastic storage of water (shrinking and the balance with ULF-NMR is not straight forward, and if swelling) once watering is stopped, which is consistent anything is detected, bubble formation has to be so exces- with stomatal closure of the plant (Fig 4). Surprisingly, the sive that the tree is very close to death already. overall water content of the trees declined very slowly, and only after the bark was removed close to the NMR system. At the third stage, we conducted long-term experiments This indicates that aspen trees are very well sealed against in a greenhouse and in the field to test our system and water loss when the stomata are closed, and large-scale demonstrate the capability for long-term measurements in bubble formation occurs in these trees only when their natural environments. We performed drought-rewatering metabolic processes have already ceased. experiments on aspen trees to investigate indicators of physiological changes in water and carbon allocation in the NMR signal. We also continued tackling the question whether bubble formation could be detected with our system by combining NMR with continuous stem diam- eter variation measurements. Stem diameter variations are known to follow a diurnal cycle and correlate strongly with transpiration and sapflow rates [21] indicating the magnitude and consumption of the elastic storage capac- ity [22]. If measured simultaneously on bark and on the xylem, stem diameter variations can be used for calculating the osmotic pressure in the phloem tissue. This pressure fluctuates with carbohydrate availability [23], and loss of it predicts tree mortality relatively accurately [12]. The low Figure 4. Time series of FID amplitude of the NMR signal and magnetic field of our NMR system allowed for uninter- water use of an aspen tree in our greenhouse experiment start- rupted measurements with these two systems simulta- ing July 11, 2013. Water use was measured by placing the tree neously for several months. For field measurements the on a balance. The temperature corrected FID amplitude (blue) NMR system was protected from moisture by covering the is sifted down by 0.1 to allow for clarity. To encourage water 88

Page 89

loss the tree was manipulated in several ways: A) watering was For these species, in this environment, bubble formation stopped; C) tree crown was cut 1m above the NMR coil; D) tree is thus rather a threat than a means to boost productiv- was cut 10cm above NMR coil, and E) tree was girdled above ity. The observation of water loss from the xylem during and below the NMR coil. B marks a gap in data resulting from leaf loss was unexpected and emphasizes how little we magnet power supply reset. The reversible diurnal fluctuation in actually know about how trees cope with environmental water content ceases in a few days once watering is stopped, but stress, and how new measurement methods lead to new to detect real overall water loss (overall declining trend in the discoveries. In this project we developed new methods for end), the tree had to be cut and girdled. in-vivo, non-destructive measurements of water content Interestingly, leaf loss resulted in a clear decline in the fluctuations and imaging water movement in woody plants NMR signal indicating water loss, but this was not reflected to the level at which these methods can be applied to in the xylem diameter variations as shrinkage of the tissue further studies of plant responses to environmental stress. (Fig 5). Xylem tissue shrank earlier during the drought pe- With these methods we identified one new parameter that riod indicating depletion of the xylem elastic water storage should be considered when modeling tree mortality during capacity. The sharp decline in NMR signal in the absence drought, and observed a new possible strategy of plants to of decline in xylem diameter during leaf loss could thus be control water stress during environmental changes. due to emptying of xylem conduits. The NMR and xylem diameter signals did not resume their original level 20 days after watering resumed, but the whole stem diameter did recover completely. This indicates that the tree refilled its water reserves in the phloem and bark tissues rapidly after watering was resumed, but it did not completely refill the empty xylem conduits. After new leaves flushed, the diur- nal variation indicative of elastic storage use resumed both in the NMR and stem diameter signals, but with a lower amplitude indicating changes in water-carbon balance of bark and phloem tissues. The sharp decline in whole stem diameter during leaf loss shows that the water lost from xylem did not end up in the phloem and bark storage tis- sue. Losing water from the xylem conduits to the benefit of e.g. roots during leaf loss could be beneficial for plant sur- vival during drought to reduce osmoregulation demands or ensure maintenance of soil-root connection. As long as the conduits can be refilled, or new ones built to resume the Figure 5. Time series of simultaneous measurements of NMR water transport capacity to support new leaves, no per- and tree stem diameter variations in an aspen tree in the field. manent harm would be done. Our observation of potential Time is adjusted to zero when the tree lost all its leaves under gas bubble formation in the xylem during leaf loss is the imposed drought. The vertical lines mark moments when water- first of its kind and replication of the experiment is needed ing was stopped (a), resumed (b) and new fully grown leaves to confirm the result, and if confirmed, to investigate the observed (c). The NMR signal and diameter measured on bark abundance of this resource allocation strategy in trees. (whole stem) increase slightly and the xylem stops shrinking right after leaf fall as a response to an increase in water availability af- Our measurements thus show that in the arid-zone species ter a thunderstorm. After that NMR and whole stem signals drop such as aspen, pinon pine and one-seed juniper, bubble dramatically suggesting water lost from the stem. The xylem formation rate is so low that we cannot distinguish it from diameter does not shrink at this stage indicating that the change the use of elastic water stores at daily scale, but we can de- in the NMR signal could be due to bubble formation in the xylem tect it during drought or major reallocation events such as (most of NMR signal reflects xylem water). With production of new leaves, the diurnal fluctuations in all signals slowly resume. leaf loss. The low diurnal bubble formation rate is contra- dictory to many studies performed with acoustic emission detection methods that claim regular bubble formation in trees [24]. But it is in line with our own observations about Impact on National Missions stomatal control of transpiration in these arid-zone spe- Capability to predict vegetation responses to changing cli- cies. These species cannot rely on continuous water source mate is needed to support formulating policies and strate- to refill gas-filled conduits regularly and close their stomata gies for mitigating the global and national threat of climate well before bubble formation becomes detectable [25]. change. Vegetation regulates the terrestrial water and 89

Page 90

carbon balance, acting as a carbon sink and water redis- 2001. PLANT CELL AND ENVIRONMENT. 24 (5): 491. tributor within the soil as well as between the soil and the 8. Windt, C. W., Soltner, van Dusschoten, and Bluemler. atmosphere. The presence or absence of vegetation, and A portable Halbach magnet that can be opened and forests especially, influences the local climate and weather closed without force: The NMR-CUFF. 2011. JOURNAL and has a direct impact on natural resources as well as en- OF MAGNETIC RESONANCE. 208 (1): 27. ergy consumption and production. In this project we have developed new methods for measuring plant physiological 9. Jones, , P. S. Aptaker, Cox, B. A. Gardiner, and P. J. Mc- responses to drought. These methods have allowed us to Donald. A transportable magnetic resonance imaging identify new key parameters that could help in predict- system for in situ measurements of living trees: The ing plant responses to drought and their susceptibility to Tree Hugger. 2012. JOURNAL OF MAGNETIC RESO- mortality under droughts of different duration more accu- NANCE. 218: 133. rately. While more measurements of these parameters are needed to build a global picture of plant susceptibility to 10. Yoder, , M. W. Malone, M. A. Espy, and Sevanto. Low- drought for global scale vegetation modeling, this project field nuclear magnetic resonance for the in vivo study has allowed our method to reach the stage at which it can of water content in trees. 2014. REVIEW OF SCIENTIFIC be used in further studies that assist in developing mecha- INSTRUMENTS. 85 (9). nistic presentations of plant mortality for the models. 11. Malone, M. W., J. Yoder, J. F. Hunter, M. A. Espy, L. T. References Dickman, R. O. Nelson, S. C. Vogel, H. Sandin, and S.

1. Allen, C. D., A. K. Macalady, Chenchouni, Bachelet, Mc- Sevanto. In vivo observantion of drought response in a Dowell, Vennetier, Kitzberger, Rigling, D. D. Breshears, tree with low-field NMR and neutron imaging. Fron- E. H. (. Hogg, Gonzalez, R. o. d. Fensham, Zhang, Cas- tiers in Plant Science. tro, Demidova, Lim, Allard, S. W. Running, Semerci, and
2. Sevanto, , N. G. McDowell, L. T. Dickman, Pangle, and Cobb. A global overview of drought and heat-induced W. T. Pockman. How do trees die? A test of the hydrau- tree mortality reveals emerging climate change risks lic failure and carbon starvation hypotheses. 2014. for forests. 2010. FOREST ECOLOGY AND MANAGE- PLANT CELL AND ENVIRONMENT. 37 (1): 153. MENT. 259 (4): 660.
3. McDowell, , W. T. Pockman, C. D. Allen, D. D. Breshears,
4. Xu, , N. G. McDowell, Sevanto, and R. A. Fisher. Our Cobb, Kolb, Plaut, Sperry, West, D. G. Williams, and E. limited ability to predict vegetation dynamics under A. Yepez. Mechanisms of plant survival and mortality water stress. 2013. NEW PHYTOLOGIST. 200 (2): 298. during drought: why do some plants survive while oth-
5. McDowell, N. G., and Sevanto. The mechanisms of ers succumb to drought?. 2008. NEW PHYTOLOGIST. carbon starvation: how, when, or does it even occur at 178 (4): 719. all?. 2010. NEW PHYTOLOGIST. 186 (2): 264.
6. Casieri, C., L. Senni, M. Romagnoli, U. Santamaria, and
7. McDowell, N. G.. Mechanisms Linking Drought, Hy- F. De Luca. Determination of moisture fraction in wood draulics, Carbon Metabolism, and Vegetation Mortal- by mobile NMR device. 2004. JOURNAL OF MAGNETIC ity. 2011. PLANT PHYSIOLOGY. 155 (3): 1051. RESONANCE. 171 (2): 364.
8. GOLDMAN, M., M. CHAPELLI, V. H. CHAU, and A. ABRA- 15. Holtta, , Juurola, Lindfors, and Porcar-Castell. Cavi- GAM. PRINCIPLES OF NUCLEAR MAGNETIC-ORDERING. tation induced by a surfactant leads to a transient
9. PHYSICAL REVIEW B. 10 (1): 226. release of water stress and subsequent ‘run away’ em- bolism in Scots pine (Pinus sylvestris) seedlings. 2012.
10. Windt, C. W., E. d. o. Gerkema, and Van As. Most Wa- JOURNAL OF EXPERIMENTAL BOTANY. 63 (2): 1057. ter in the Tomato Truss Is Imported through the Xylem, Not the Phloem: A Nuclear Magnetic Resonance Flow 16. Sevanto, . Phloem transport and drought. 2014. JOUR- Imaging Study. 2009. PLANT PHYSIOLOGY. 151 (2): 830. NAL OF EXPERIMENTAL BOTANY. 65 (7): 1751.
11. Peuke, A. D., M. Rokitta, U. Zimmermann, L. Schreiber, 17. Sevanto, , Holtta, and N. M. Holbrook. Effects of the and A. Haase. Simultaneous measurement of water hydraulic coupling between xylem and phloem on diur- flow velocity and solute transport in xylem and phloem nal phloem diameter variation. 2011. PLANT CELL AND of adult plants of Ricinus communis over a daily time ENVIRONMENT. 34 (4): 690. course by nuclear magnetic resonance spectrometry.
12. Savage, J. A., M. J. Clearwater, D. F. Haines, T. Klein, M. 90

Page 91

Mencuccini, S. Sevanto, R. Turgeon, and C. Zhang. Allo- Hart, D. J., M. W. Malone, S. Sevanto, M. A. Espy, and N. cation, stress tolerance and carbon transport in plants: Cruz. Measuring teh water content in trees using NMR. How does phloem physiology affect plant ecology?. To Presented at 14th Annual Student Symposium. (Los Alamos, NM, 5 Aug, 2014). appear in Plant, Cell and Environment. Malone, M. W., J. Yoder, M. A. Espy, and S. Sevanto. Disen- 19. Peramaki, M., E. Nikinmaa, S. Sevanto, H. Ilvesniemi, E. tangling biology and physics to measure water in trees. Siivola, P. Hari, and T. Vesala. Tree stem diameter varia- Presented at Plant Biology. (Portland, OR, 12-16 Jul. tions and transpiration in Scots pine: an analysis using 2014). a dynamic sap flow model. 2001. TREE PHYSIOLOGY. 21 (12-13): 889. Malone, M. W., J. Yoder, M. A. Espy, and S. Sevanto. Dis- entangling biology and physics to measure wayer in 20. Holtta, , Cochard, Nikinmaa, and Mencuccini. Capaci- trees. Presented at NMR^2. (Albuquerque, NM, 3 May, tive effect of cavitation in xylem conduits: results from 2014). a dynamic model. 2009. PLANT CELL AND ENVIRON- Malone, M. W., J. Yoder, M. A. Espy, and S. Sevanto. Disen- MENT. 32 (1): 10. tangling biology and physics to measure water in trees. 21. Sevanto, , Nikinmaa, A. n. u. Riikonen, Daley, J. C. Pet- Presented at LANL Postdoc Research Day. (Los Alamos, NM, 11 Jun, 2014). tijohn, T. N. Mikkelsen, Phillips, and N. M. Holbrook. Linking xylem diameter variations with sap flow mea- Sevanto, S.. Phloem transport and drought. 2013. In 3rd surements. 2008. PLANT AND SOIL. 305 (1-2): 77. International Conference on Plant Vascular Biology. (Helsinki, Finland, 26-30 Aug, 2013). , p. 56. Helsinki: 22. Sevanto, S., T. Holtta, T. Markkanen, M. Peramaki, E. Helsinki University Press. Nikinmaa, and T. Vesala. Relationships between diurnal xylem diameter variation and environmental factors in Sevanto, S.. Phloem transport and drought. 2014. Journal Scots pine. 2005. BOREAL ENVIRONMENT RESEARCH. of Experimental Botany. : doi: 10.1093/jxb/ert467. 10 (5): 447. Yoder, J.. Measuring absolute water content of trees in vivo 23. Mencuccini, , Holtta, Sevanto, and Nikinmaa. Concur- with low field NMR. Invited presentation at Physics Division Summer Seminar PostDoc Series. (Los Alamos, rent measurements of change in the bark and xylem NM, June 26, 2013). diameters of trees reveal a phloem-generated turgor signal. 2013. NEW PHYTOLOGIST. 198 (4): 1143. Yoder, J., M. A. Espy, M. W. Malone, and S. Sevanto. In vivo observation of drought response in a tree with low- 24. Holtta, T., T. Vesala, E. Nikinmaa, M. Peramaki, E. field NMR. 2014. Review of Scientific Instruments. 85: Siivola, and M. Mencuccini. Field measurements of 095110. ultrasonic acoustic emissions and stem diameter varia- tions. New insight into the relationship between xylem Yoder, J., M. Espy, N. McDowell, J. Resnick, and S. Sevanto. tensions and embolism. 2005. TREE PHYSIOLOGY. 25 Measuring absolute water content of trees in vivo (2): 237. by low field NMR in an uncontrolled environment. Presented at 54th Experimental Nuclear Magentic 25. Garcia-Forner, N., H. D. Adams, S. Sevanto, A. D. Col- Resonance Conference. (Pacific Grove, CA, 14-19 April, lins, L. T. Dickman, P. J. Hudson, M. Zeppel, J. Martinez- 2013). Vilalta, and N. G. McDowell. Responses of two semi- arid confer tree species to reduced precipitation and warming reveal new perspectives for stomatal regula- tion. To appear in Plant, Cell and Environment. Publications Cruz, N., J. Yoder, and M. Espy. Active Q-switching in NMR applications. Presented at 13th Annual Student Sym- posium, Los Alamos National Laboratory . (Los Alamos, NM, July 26-27). Cruz, N., M. W. Malone, D. J. Hart, S. Sevanto, and M. A. Espy. Tracking water flow in trees using NMR. Present- ed at 14th Annual Student Symposium. (Los Alamos, NM, 5 Aug, 2014). 91

Page 92

Complex Natural & Engineered Systems Exploratory Research Final Report Pyrocumulus Collapse: Unpredicted Wildfire Dangers Young-Joon Kim 20130487ER Abstract regions with complex topography. We also performed This project aims at understanding the physical mecha- high-resolution simulations over bell-shaped barrier nisms behind the Las Conchas Fire, which occurred in similar to the topography of the fire area under various Santa Fe National Forest near Los Alamos, New Mexico, atmospheric conditions using HIGRAD and found that a on June 26, 2011. The Fire surprised everyone when it localized wildfire-like heat source is largely affected by unexpectedly burned approximately 35,000 acres in less topographically generated gusty winds. Our investiga- than 7 hours during its first night even though it was tion sheds light on the interaction of mountain waves burning downhill in sparse vegetation and under milder with wildfire to modelers and managers of wildfire. wind conditions than had been present on that after- Observed nominal atmospheric conditions from the first noon. night of the fire are likely to occur in a vast set of fire- We considered two potential mechanisms for this baf- susceptible communities bordering mountains across fling fire behavior, especially its unexpected nighttime the country. This research addresses important implica- blowup; (1) downdrafts associated with the soot-laden tions for the management as well as research of wildfire pyro-cumulus column (pyro-cu) that towered above the in that, in order to prepare for potential wildfire, such fire, causing a sustained density current carrying fire at environmental conditions should be taken into account. high speed, and (2) downslope windstorms due to the Background and Research Objectives breaking of large-amplitude mountain waves developed over Jemez Mountains near Los Alamos. We discussed Between 8 pm local time (MDT) on June 26 and 3 am some insights on these mechanisms and explored their on June 27, the Las Conchas Fire (hereafter referred to possible effects on wildfire behavior dynamics. as the LC Fire) grew from 8,000 to 43,000 acres, spread- ing downhill in sparse fuels and lighter winds than were We first performed simulations using the LANL-de- present during the first several hours of the fire. Fire veloped large-eddy simulation model (HIGRAD; High- behavior experts and fire management officers expected GRADient model) that includes a multi-component the fire to reach 9,000 to 12,000 acres by sunrise due continuum model to scope the potential dynamics that to the anticipated burning conditions, but it surprisingly could result when a soot and water vapor laden column, increased 440% in size before 3 am, causing significant such as a pyro-cu, looses its support from underneath. threat to life and property and eventually forcing the The simulations illustrated that a rapid descent of evacuation of nearby towns. heavier-than-air gas mixture due to its own weight could occur under certain atmospheric and wildfire conditions It is important that the fire community deciphers the although the mechanisms to sustain the process were conditions that increase probability of such events and unclear. the phenomenology that controls their intensity. An initial hypothesis for the peculiar behavior in the LC Fire We then performed simulations using a mesoscale atmo- is that a particulate-laden portion of the fire-induced spheric model (WRF; Weather Research and Forecasting “pyro-cumulus (pyro-cu)” column lost buoyancy and model) and found suggestive results that account for the produced strong downdrafts, and unleashed a symbi- coupled topography-atmosphere interaction and result- otic density current that increased fire spread rate and ing influences on fire, which may be required to fully intensity. Another hypothesis is that the interactions understand and prepare for atypical wildfire behavior in between the stable nocturnal atmosphere and the local 92

Page 93

topography influenced the fire behavior or pyro-cu column side of topographic feature on the landscape, i.e., create stability. We investigated the viability of such hypotheses a downslope windstorm. We investigated the potential with numerical simulations. that MWs could act as a catalyst for rapid intensification or ‘blowup” of a wildfire, i.e., a pronounced increase in the Some wildfire anomalies have occurred in the early to late energy output of a wildfire event over a relatively short evening period. During typical wildfire seasons (mid-Spring period of time (say, on the order of an hour). to early Fall), the atmospheric stability profile transitions from an unstable convective boundary layer profile during Our research was inspired by the desire to explore basic the mid and late afternoon times to a stably stratified pro- physical mechanisms relevant to the first night of the LC file in the evening. Most often the intensity of a wildfire Fire and identify what combined fire/atmosphere dynam- will decrease with the onset of evening as surface winds in ics could have led to the unexpected observed fire behav- general decrease and relative humidity increases. Wildfire ior. Our primary research objective is to investigate and management activities are often designed in anticipation understand the interaction between the meteorology, of this type of behavior and when the fire unexpectedly dynamic fire intensity, plume dynamics and topography changes to the contrary and increases its activity, practi- that would lead to significant downslope acceleration tioners can be surprised. There is a need to understand of the LC Fire during the night when the ambient wind what combination of aspects of the evening atmospheric slowed down and fuels became less continuous. Our future conditions and transitioning fire behavior set the state for investigation will ultimately involve detailed explorations of such surprises. We focused on phenomenology that might the interactions of winds with a heat source, identification exacerbate surface winds and thereby lead to the blowup of potential causes of strong downdrafts and subsequent of a wildfire such as the LC Fire. downslope density currents as well as developing a better understanding of the interactions between a density cur- In the early years of wildfire behavior research, several rent and a fire front. commonalities in meteorology, such as the vertical profiles of wind speed and direction, were noted during several Scientific Approach and Accomplishments blowup fires. It was hypothesized that each of the anomaly Our scientific approach is to investigate available observa- fires was influenced by a vertical wind profile in which tions as well as perform and validate numerical simula- substantially higher wind speeds were found near the tions. We first investigated the likelihood of MWs in the LC surface than aloft. One way to achieve such conditions Fire area by analyzing available observation and analysis/ locally pertains to the events occurring when a portion forecast data, e.g., using selected LANL meteorological of a pyro-cu column does not have sufficient buoyancy to tower data. We then performed scoping numerical simu- stay aloft. The potential energy of a particle-laden plume lations using the Weather Research and Forecast model that is heavier than the surrounding air can be converted (WRF) and main simulations using our large-eddy-resolving into kinetic energy as the large air mass begins to descend, model, HIGRAD, in order to identify the conceivable magni- just as in the vicinity of some thunderstorms. The inter- tude of the wind changes due to rapid decent of a portion section of the descending winds and the surface can lead of the pyro-cu column under the topographic setting of the to a strong torrent of horizontal surface winds, a.k.a. a LC Fire. Through numerical simulations, we investigated “density current”, accelerating and intensifying fires in its the interaction of an isolated heat source with an atmo- path. Collapsing pyro-cu’s have been implicated in several sphere conducive to the presence of MWs. We report on devastating historical fire events, including the Dude Fire, our qualitative characterization of the manner in which the Chisholm Fire, the Canberra Fires, and the Brasstown MWs interact with a heat source plume. Fire; nevertheless, the phenomenology and warning signs for such an event are not known. First, HIGRAD was modified to include the multi-phase property of fire-contaminated air. With a multi-component Another set of phenomena resulting in localized gusty continuum model included, simulations were performed winds or “downslope windstorms” are topographically gen- to scope the potential dynamics that could result when erated gravity waves, i.e., mountain waves (hereafter re- a soot and water vapor laden column, such as a pyro-cu, ferred to as MWs). Stable stratification of the atmosphere looses its support from underneath. The loss of such sup- can lead to generation of MWs. In fact, strong (especially, port could for instance occur due to shifting wind direction nocturnal) winds are often found near steep topography at various altitudes diminishing fire intensity. An idealized when wildfires grow unexpectedly (e.g., 22 mph for the column of a gaseous mixture containing heat, dry air, water 2013 Yarnell Hill Fire in Arizona). When these waves are vapor, and fullerene was initialized over the LC Fire area of sufficient magnitude, they can overturn and produce topography where the fire occurred, where the gas col- an enhancement of near surface wind speeds on the lee- 93

Page 94

umn represented an idealized fire plume. The simulation region of strong surface winds. This phenomenon could illustrated the expected rapid descent heavier-than-air gas reverse the flow in the vicinity of wildfire or change it er- mixture due to its own weight under certain atmospheric ratically, putting firefighters at unpredicted risk, i.e., fire and wildfire conditions. However, the proposed physical can unexpectedly strike them from behind. These types mechanism did not fully explain how the vertical motion is of erratic wildfire behavior due to mountains should be sustained for the duration of the fire blowup. investigated systematically and applied to saving precious lives of firefighters. The area surrounding the LC Fire is highly mountainous and strong surface winds are often observed in the area. In this work, we demonstrated that the high-resolution In order to validate the likelihood of the topographic influ- MW dynamics may be crucial to explain the Las Conchas ence on the fire, a series of MW simulations using WRF Fire surrounded by complex terrain. This finding has were performed and the presence of MW activity around important implications for fire management community the LC Fire period was investigated. The simulations were and can provide knowledge to better prepare for future initialized at 0000 UTC 26 June 2011 (or 1800 MDT 25 June wildfires. The physical interpretation of Las Conchas Fire in 2011) before the fire started (at 1300 MDT 26 June 2011) the context of MW simulation in eddy-resolving scale is not using the local atmospheric conditions before and during found in the literature. The results out of this project are the fire together with the area topography. When fully original and of high quality as was confirmed during the developed (at 09 UTC or 06 MDT 27 June 2011), MWs were 2014 Fall AGU Meeting where the work was first present- found very nonlinear in the entire area of the cross section ed. This work also drew some public attention through a and a large portion of the low-level flow over mountains media interview and a subsequent magazine article [The reached a “hydraulic jump” state. The low-level flow was Earth Magazine, May 2015]. separated from the flow above and became well mixed due to the turbulence created by wave breaking with flow Impact on National Missions reversal (associated with “rotors”) in the downstream. LANL’s capacity to perform its primary missions was threat- ened by the Las Conchas Fire, an unexpectedly grown HIGRAD was further modified to represent the buoyancy wildfire near high complex topography. According to our source-sink term in the prognostic subgrid-scale turbulent investigation, this wildfire developed into much large scale kinetic energy equation, which is critical to the formation with much greater magnitude possibly due to the genera- of MWs. Monin-Obukhov formulations for both surface tion of strong density currents associated with a collapse heat flux and momentum flux were added, respectively, of pyro-cumulus and/or wind accelerations by the breaking to characterize a stably stratified flow environment and of MWs, which was channeled by street canyons, thereby to provide a mechanism for momentum sink at the sur- focusing the kinetic and thermal energy and potentially in- face. Also implemented are grid-dependent and dynamic creasing damage by an order of magnitude. It is essential stability-aware prescription of an appropriate subgrid- that this kind of potential threat be understood in order to scale turbulence length, and stability-dependent turbulent save lives and safeguard critical facilities of LANL and other Prandtl number. We successfully simulated both linear national laboratories under similar environments. Dis- and nonlinear MWs over idealized bell-shaped ridge-type cerning this type of erratic wildfire behavior is crucial for topography, designed to mimic the Las Conchas Fire area. elucidating the fire-related effects from an urban nuclear In nonlinear case (Fr-1=1.2), downstream turbulence and event (on US soil or elsewhere). gusty winds due to MW breaking were clearly simulated. The simulations were continued with a heat source added LANL is involved in activities supporting DHS (Department in order to study the interaction of nonlinear MWs with of Homeland Security), STRATCOM (Strategic Command), heat source crudely representing wildfire. The turbulent and DTRA (Defense Threat Reduction Agency) to under- wakes and gusty winds generated by breaking of non- stand potential damage and lives at risk from an urban linear MWs sweep the fire represented by heat source conflagration that could likely result from this type of downstream. The fire has spread downstream along the disasters. It will enhance LANL’s unique capacity to study downslope faster and more widely than without MW influ- coupled fire/atmosphere phenomena in realistically com- ence. Further downstream with turbulent wakes/rotors, plex environments and therefore will help secure LANL’s the winds were weaker and reversed, which could poten- role supporting the USDA (US Department of Agriculture) tially cause unexpected fire behavior. Forest Service wildfire research and management missions, and position LANL to serve a larger research role and to The rotors of turbulent air along the downstream of the assume direct relevance to the fire management sector mountain can complicate the behavior of wildfire because of the USDA Forest Service. More generally, this work will the wind directions under the rotors are reversed near the 94

Page 95

help establish the capability to enable wildfire and CBRN havior trends. 2014. PhD Dissertation, Florida State (Chemical, Biological, Radiological, and Nuclear) material University / LANL LA-UR-14-22933. 28 Apr. 2014. dispersion risk assessment across the DOE complex. Canfield, J., R. Linn, J. Sauer, M. Finney, and J. Forthofer. A numerical investigation of the interplay between fire- References line length, geometry, and rate of spread. 2014. Agri-

1. Sauer, J., J. Canfield, R. Linn, E. Koo, J. Winterkamp, and cultural and Forest Meteorology. 189 (190): 48. Y. -J. Kim. Assessing fire behavior with fundamental fluid dynamics theory. 2014. In International Associa- Kim, Y. -J., J. Sauer, R. Linn, J. Canfield, K. Costigan, D. tion for Wildland Fire, Large Wildland Fires Confer- Munoz-Esparza, and S. Goodrick. Identification of ence. (Missoula, 19-23 May 2014). , p. 00. Missoula, Potential Mechanistic Drivers in Anomalous Behav- ior of Large Wildland Fire: The 2011 Las Conchas Fire MT: International Association of Wildland Fire. Blowup. International Journal of Wildland Fire (To be
2. Kim, Y. -J., R. Linn, J. Sauer, J. Canfield, K. Costigan, and submitted by 10/16/15). D. Munoz-Esparza. Enhanced Risk of Wildfire Result- Kim, Y. -J., R. Linn, J. Sauer, J. Canfield, and K. Costigan. En- ing from the Interactions between Pyro-Cumulus and hanced Risk of Wildfire Resulting from the Interactions Mountain Waves: Implications for Fire Research and between Pyro-Cumulus and Mountain Waves: Impli- Management. 2014. In American Geophysical Union cations for Fire Research and Management. 2014. In Fall Meeting. (San Francisco, CA, 15-19 December American Geophysical Union Fall Meeting. (San Fran- 2014). , p. 00. Washington, DC: American Geophysical cisco, 15-19 Dec. 2014). , p. 00. San Francisco: Ameri- Union. can Geophysical Union.
3. The Las Conchas Fire, the largest wildfire in New Mex- Linn, R., J. Winterkamp, J. Canfield, J. Sauer, J. L. Dupuy, M. ico history, threatened the town of Los Alamos and Finney, C. Hoffman, R. Parsons, P. Fancois, C. sieg, and J. Forthofer. An Emerging Role for Numerical Modelling Los Alamos National Laboratory in June and July 2011. in Wildfire Behavior Research: Explorations, Explana-
4. The Earth Magazine http://www.earthmagazine. tions, and Hypothesis Development. 2014. In American org/article/fire-driven-clouds-and-swirling-winds- Geophysical Union Fall Meeting. (San Francisco, 15-19 whipped-record-setting-new-mexico-blaze. Dec. 2014). , p. 00. San Francisco: American Geophysi- cal Union.
5. Sauer, J., D. Munoz-Esparza, J. Canfield, K. Costigan, R. Linn, and Y. -J. Kim. New Insights into Boundary Layer Sauer, J.. Towards improved capability and confidence in Influence on Stratified Atmospheric Flows over Terrain coupled atmospheric and wildland fire modeling. 2013. through Large-Eddy Simulations. Journal of the Atmo- PhD Dissertation, Florida State University, 12 Nov. spheric Sciences. 2013.
6. Kim, Y. -J., J. Sauer, R. Linn, J. Canfield, K. Costigan, and Sauer, J., D. Munoz-Esparza, J. Canfield, K. Costigan, R. D. Munoz-Esparza. Identification of Potential Mecha- Linn, and Y. -J. Kim. New Insights into Boundary Layer Influence on Stratified Atmospheric Flows over Terrain nistic Drivers in Anomalous Behavior of Large Wildland through Large-Eddy Simulations. Journal of the Atmo- Fire: The 2011 Las Conchas Fire Blowup. Interna- spheric Sciences. tional Journal of Wildland Fire (To be submitted by 10/16/2015). others, D. Munoz-Esparza and. Winds, Mountains, and Wildland Fire: Improved Understanding of Coupled Publications Atmosphere-Topography-Fire Interactions Through The Las Conchas Fire, the largest wildfire in New Mexico Large-Eddy Simulation. To appear in American Geo- history, threatened the town of Los Alamos and Los Al- physical Union Fall Meeting. (San Francisco, CA, 14-18 amos National Laboratory in June and July 2011. 2015. Dec. 2015). The Earth Magazine http://www.earthmagazine.org/ article/fire-driven-clouds-and-swirling-winds-whipped- record-setting-new-mexico-blaze. The Role of Mesoscale Parameterized Turbulence in De- velopment of Trapped-Lee and Breaking Waves over Complex Terrain. Journal of the Atmospheric Sciences (To be submitted by 10/31/15). Canfield, J.. Wildfire dynamics: Understanding some be- 95

Page 96

Complex Natural & Engineered Systems Exploratory Research Final Report Biocatalysts for Remediation of Uranium Wastes Francisca Rein Rocha 20130590ER Abstract with toxic and radioactive materials (mostly subsurface) One of the strategic priorities of the U. S. Department of accumulated over 50 years of nuclear activities during Energy has been the cleanup of numerous sites across the Cold War. Contamination by uranium is especially the Nation that contain subsurface contamination with problematic. Given the large volumes of soil and ground- radionuclides from legacy nuclear waste. Given the water contamination with uranium and other heavy toxic daunting volume scales of soil and groundwater con- metals at low concentrations, bioremediation has been tamination with uranium and toxic heavy metals at low regarded as more cost-effective than other methods concentrations, in-situ bioremediation has been regard- involving immobilization or preconcentration of contami- ed as a promising approach to cost-effective environ- nants for ex-situ treatment. Despite significant advances mental decontamination. A biological approach is based in this area, the development and practical realization on the fact that some bacteria contain specific multi- of viable technologies for remediation at such daunt- heme cytochromes that can catalyze the reduction of ing scales still rely on research at the fundamental (bio) uranium from its soluble uranyl form (UO2 2+) to the less molecular level. mobile state as uraninite (UO2). However, catalysis by Unlike organic contaminants, radionuclides and toxic microorganisms faces drawbacks owing to slow reactions heavy metals cannot be biologically “degraded” but can leading to incomplete reductions and release of con- undergo redox transformations that change their mobil- taminants as well as the need for continuous supply of ity. In natural environments, the transport of uranium is sacrificial chemical reductants, which is impractical and mostly determined by its ability to dissolve and dis- promotes deactivation due to competing re-oxidations sipate in water flows; this ability, in turn, is ultimately by the reversible activity of reductants. Moreover, the determined by its redox states. The dramatic decrease difficult molecular manipulation for improvement of key in solubility accompanying the reduction of U(VI) ions to structural properties and functional components in such produce the insoluble mineral UO2 has been viewed as a living systems is a bottleneck for further advances in the path of great potential to the sequestration and trapping derivation of bioremediation technologies. To address of this environmental contaminant in the soil under- this grand challenge, the goal of this ER project has been ground, thereby preventing migration of contaminated the development of an abiotic approach to the effec- plumes into aquifers and natural groundwater resources. tive reduction of hexavalent uranium. In this proof-of- Although the hexavalent state is the most stable for ura- concept, the electrocatalytically active reductase found nium in aqueous medium (where the dominant species in uranium-reducing microorganisms was produced as is the uranyl ion, UO22+, free or in some form of com- artificial biocatalyst films onto electrode surfaces for plexation), the redox behavior of uranium is remarkably successful demonstration of the UO2 2+ → UO2 immo- rich and often very complicated due to a range of acces- bilization under small electrochemical potentials. The sible oxidation states and sensitivity of their thermody- achieved results represent an important step toward namic potentials to medium effects and coordination viable strategies for practical remediation of nuclear environments around the metal [1]. waste, in addition to creating capabilities of relevance to Laboratory missions in this direction. Sulfate-reducing bacteria of the genus Desulfovibrio are anaerobes that derive energy from the dissimilatory re- Background and Research Objectives duction of sulfate coupled with oxidation of dihydrogen Within the U. S. Department of Energy (DOE) complex or organic substrates [2]. Some of these bacteria are also alone, thousands of discrete sites are contaminated 96

Page 97

capable of reducing toxic metals, which is a metabolism proton-coupled redox behavior [8]. Following optimization relevant to bioremediation applications. The overall reduc- of conditions and electronic/structural characterizations, tion is a process involving two electrons, but the mecha- we have tackled the interfacial molecular engineering of nism by which toxic metals are reduced in D. desulfuricans Cyt c3 electrodes for the electrocatalytic reduction and im- has not yet been established, despite its great potential for mobilization of uranium [9,10]. utilization in bioremediation of hexavalent uranium and An important milestone in the first phase of the project chromium [3,4]. Studies have confirmed that the tetra- was the successful implementation of the new prepara- heme cytochrome c3 present in such bacteria is specifi- tive methods and bio-analytical techniques needed for the cally involved in the reductive reaction with U(VI) [5]. Of expression, isolation, purification, and structural charac- particular interest here is the metal reductase activity of terization of our uranium-reducing biocatalyst. Producing cytochrome c3 from strain G20 (hereafter, denoted Cyt c3). the tetraheme reductase Cyt c3 at the target qualitative This Cyt c3 produced by D. desulfuricans G20 contains four and quantitative levels was a particularly significant accom- unequal c-type hemes that are each attached to a peptide plishment because multi-heme cytochromes are generally by two thioether bridges with bis-histidinyl axial ligation more difficult to express correctly, since co-factors are [6] (Fig. 1). Cyt c3 is negatively charged at neutral pH (with needed to enable the proteins to obtain their native fold. a low isoelectric point of 5.8) and, as a result, it has a In Cyt c3, four hemes must be attached via thioether link- negative redox potential. Its molecular weight is only 14.4 ages to the polypeptide chain at the CXX(XX)CH binding kDa, which is similar to some monoheme cytochromes. sites, and the correct distal axial ligands must be con- The four hemes in Cyt c3 are situated relatively close to nected to the metal center (trivalent iron) in the nascent each other and feature intramolecular heme-heme interac- protein. In our approaches involving E. coli recombinant tions [7], which are crucial to cooperative electron-transfer expression, specific assembly proteins were required for mechanisms for multi-electron catalysis. Most importantly, these critical interactions to form properly. Among the in-vitro tests with purified Cyt c3 showing that the fully achieved improvements in the expression of Cyt c3, the reduced species is oxidized upon reduction of metals have pC3 plasmid carrying the Cyt c3 gene was co-expressed provided encouraging evidence toward our goal to develop with the pEC86 plasmid carrying genes of helper matu- cell-free systems for applications in both in situ and ex situ ration proteins. Other systematic changes have also led reduction of hexavalent uranium. to further optimizations towards high yields, especially strategies involving auto-induction and temperature ef- fects. We have found that auto-induction processes allow for control over amounts of media and supplements for bacteria growth to high densities prior to protein expres- sion. Another key finding was that decreasing the typi- cal growth temperatures, which causes bacteria to grow slower, has allowed the protein more time to properly fold into its native form. Following expression, the protocols for extraction, purification, and analysis of Cyt c3 have also been implemented and applied, from which the products were purified/analyzed by ion-exchange chromatographic column (Sepharose), gel electrophoresis (SDS-PAGE), Coo- massie/heme staining, absorption spectroscopy (UV-Vis and IR), and electrochemistry (cyclic and pulse voltamme- tries). As our target application requires surface immobilization Figure 1. The reductase tetraheme cytochrome c3 (Cyt c3). of the biocatalyst, we have next produced Cyt c3 tagged with histidines (His-tag Cyt c3) for tethering onto the Scientific Approach and Accomplishments functionalized electrodes. To this end, a twelve-histidine Our approach has combined interfacial bioelectrochemis- residue was placed at the C-terminal of the protein, with a try, in-situ optical spectroscopies, and site-directed muta- cleavage site (TEV protease) introduced between protein genesis into the multi-electron redox catalysis of uranium and His-tag to provide the flexibility of cleaving the tags reduction by D. desulfuricans G20 cytochrome c3 (Cyt c3). and obtain the wild-type Cyt c3 from this same derivative. The investigations of Cyt c3 alone and its mutants allowed Following purification via Ni-NTA column, we have structur- for mapping out physical and chemical factors dictating the 97

Page 98

ally and functionally characterized this new His-tag Cyt c3 amide groups are predicted to enhance the metal-protein construct. Measurements by far-UV circular dichroism (CD) interaction. Therefore, we substituted the binding pocket showed that this construct is well-folded and has retained of Cyt c3 (in the regions of lysine 14 and 56) with gluta- its helical structure. Importantly, our spectroscopic and mate/aspartate residues to enhance specific binding and electrochemical characterizations in aqueous solutions as identify structural features that are essential to uranium well as immobilized as electrode films (covalently attached reduction. Because glutamate and aspartate have been monolayers or membrane-embedded multilayers) clearly shown to be good binding partners to uranyl(VI) owing indicated that His-tagging had no negative impact on ei- to the negative charges of their side chains, the following ther the structural stability or redox potentials of the heme mutations have been made to Cyt c3: Asn12 to Asp, Thr13 sites, consequently retaining the reductase activity of Cyt to Glu, and Phe54 to Glu. Systematic structure-function c3 [9]. relationships for these mutants (in comparison to the par- ent species above) are currently being finalized to provide The kinetic and thermodynamic properties of the Cyt c3 a better understanding of both the nature of biocatalyst- catalyzed electron-transfer reaction between uranium(VI) uranium binding and the mechanism for multi-electron and electrode (gold or glassy carbon) were studied by reductase activity. cyclic voltammetry, both for homogeneous solutions and electrode-attached films. By using in-situ spectroscopy to follow UV-Vis spectral changes associated with heme- based absorption bands, we observed that the heme- Fe(II) sites of the electrochemically reduced biocatalyst are oxidized to their heme-Fe(III) states concurrently with the reduction of U(VI) in solution into U(IV) in the form of uraninite precipitate. In the presence of a complexing agent, such as citrate, the uranyl ions alone (i.e. without the catalyst) can only undergo a one-electron reduction at relatively very negative potentials (below -0.7 V vs Ag/ AgCl). Consequently, an aqueous UO22+ solution contain- ing citrate is electrochemically inactive in the range of Figure 2. Electrocatalytic activity of uranium-reducing Cyt c3 potentials examined. In contrast, similar measurements for electrodes. the catalyst in the same medium show a reversible electro- chemical response at accessible potentials (above -0.5 V vs Impact on National Missions Ag/AgCl) for the redox processes associated with the four This research supports our Laboratory missions by provid- heme sites of Cyt c3 (Fig. 2). In the presence of uranyl(VI) ing advancements toward the permanent environmental ions added to the solution, the shape of the voltammo- decontamination of radionuclides and toxic metals in grams changes due to a cathodic current that develops complex subsurface environments. The project ties directly upon reduction of Cyt c3. Interestingly, the growth of this into the LANL/LDRD Grand Challenge “Complex Biologi- cathodic current follows the increase in the concentration cal Systems” (specifically “to resolve national challenges of U(VI). This behavior is therefore consistent with the in energy, health and environment” as well as “biological regeneration of oxidized Cyt c3 by the electrocatalytic cycle interactions and metabolic patterns that control ecosys- accompanying the U(VI) → U(IV) reduction (Fig. 2). The im- tem functions”). In addition to DOE, this research has high mobilization of Cyt c3 into a membrane or ion-exchanger relevance to NNSA, DoD, and DHHS (NIH) programs. With film has not significantly impacted the overall electrocata- underlying themes related to sensing/detection of radio- lytic activity [10]. nuclides or toxic substances, nuclear forensics, molecular bioelectronics, and bioenergy, this effort is conducive to With our methodologies established and successfully programs within DHS, Intelligence Agencies, DoD (DARPA, demonstrated for both wild-type and His-tagged Cyt c3, we DTRA, ONR), and DOE (BES, ARPA-E) among some of the then turned our studies to improved biocatalyst derivatives identified agencies/sponsors. by site-directed mutagenesis. The mobile uranyl dication (UO2 2+) is a linear dioxo species that prefers to coordi- References nate up to six donor ligands in the equatorial plan. Uranyl-

1. Morris, D. E.. Redox energetics and kinetics of uranyl protein motifs show that binding typically occurs through coordination complexes in aqueous solution. 2002. carboxyl groups such as in aspartate and glutamate. In Inorganic Chemistry. 41: 3542. addition, hydrogen bonds between uranyl and backbone 98

Page 99

2. Louro, R. O.. Proton thrusters: overview of the struc- c3 (in preparation). Environmental Science & Technol- tural and functional features of soluble tetrahaem ogy. cytochromes c3. 2007. Journal of Biological Inorganic Chemistry. 12: 1.
3. LLoyd, J. R.. Microbial reduction of metals and radionu- clides. 2003. Microbiology Reviews. 27: 411.
4. Renshaw, J. C., L. J. C. Butchins, F. R. Livens, I. May, J. M. Charnok, and J. R. LLoyd. Bioreduction of uranium: environmental implications of a pentavalent interme- diate. 2005. Environmental Science & Technology. 39:
6. Payne, R. B., L. Casalot, T. Rivere, J. H. Terry, and J. D. Wall. Interaction between uranium and the cyto- chrome c3 of Desulfovibrio desulfuricans strain G20.
7. Archives of Microbiology. 181: 398.
8. Pattarkine, M. V., J. J. Tanner, C. A. Bottoms, Y. H. Lee, and J. D. Wall. Desulfovibrio desulfuricans G20 tetra- heme cytochrome structure at 1.5 Å and cytochrome interaction with metal complexes. 2006. Journal of Molecular Biology. 358: 1314.
9. Akutsu, H., and Y. Takayama. Functional roles of the heme architecture and its environment in tetraheme cytochrome c. 2007. Accounts of Chemical Research. 40: 171.
10. Rein, F. N., D. M. Vu, and R. C. Rocha. Elucidation of multiple proton-coupled redox microstates in a tetra- heme reductase (in preparation). Bioelectrochemistry.
11. Rein, F. N., D. M. Vu, and R. C. Rocha. Direct electron transfer of tetraheme cytochromes at functionalized electrode interfaces (in preparation). Biosensors and Bioelectronics.
12. Rein, F. N., D. M. Vu, and R. C. Rocha. Multi-electron bioelectrocatalysis of uranium reduction by a cyto- chrome c3 (in preparation). Environmental Science & Technology. Publications Rein, F. N., D. M. Vu, and R. C. Rocha. Elucidation of mul- tiple proton-coupled redox microstates in a tetraheme reductase (in preparation). Bioelectrochemistry. Rein, F. N., D. M. Vu, and R. C. Rocha. Direct electron transfer of tetraheme cytochromes at functionalized electrode interfaces (in preparation). Biosensors and Bioelectronics. Rein, F. N., D. M. Vu, and R. C. Rocha. Multi-electron bio- electrocatalysis of uranium reduction by a cytochrome 99

Page 100

Complex Natural & Engineered Systems Exploratory Research Final Report Structure Determination of Large and Membrane-Bound Proteins by Nuclear Magnetic Resonance (NMR) Spectroscopy Ryszard Michalczyk 20130620ER Abstract here with LANL’s synthetic stable isotope capabilities, Knowledge of the structures and mechanisms of both will open funding possibilities from NIH, DTRA and DOE. large proteins and their complexes and membrane- Background and Research Objectives bound proteins is critical to our understanding of, and intervention in, processes as varied as antibiotic resis- Crystallography has been successful in structure de- tance and biosynthesis, biofuels production, and disease termination of several large proteins and protein com- development. Application of crystallography to these plexes. However, its application to flexible systems, such systems suffers from (1) an inability to crystallize large as described here, suffers from an inability to crystallize flexible structures; (2) the effects of crystal packing; and large flexible structures as well as the effects of crystal (3) the necessity to include lipid membranes in a full packing on flexible structures. Similarly, application of analysis of structure and function. In contrast, Nuclear crystallography to membrane proteins is also limited Magnetic Resonance (NMR) is a versatile technique that due to crystallization issues caused by hydrophobicity can provide structural, dynamic, and kinetic data on bio- and the necessity to include lipid membranes in a full logical systems and their interactions in solution. analysis of structure and function of these proteins. In contrast, solution NMR is a versatile technique that can Presently the size of biomolecules suitable for NMR provide structural, dynamic, and kinetic data on biologi- spectroscopy is limited to <50 kiloDaltons (kDa) for high cal systems and their interactions with each other or resolution structures. We propose to address this critical with ligands. However, interpretation of complex NMR shortcoming by introducing specific, site-selective iso- data of protein complexes or enzymes critically relies tope labeling of amino acids in proteins and optimization on the ability to uniquely assign spectral resonances to of NMR methods for this specific labeling scheme. distinct nuclei in the molecule. In large biomolecules and molecules with crowded and overlapping resonances, We will develop new 3D and 4D NMR experiments for unambiguous assignments or measurements of distance resonance assignments and unambiguous determina- constraints and subsequent structure determination is tion of constraints for structure calculation. The methods very challenging and existing methods are frequently will use optimized magnetization transfer pathways to inadequate [1]. minimize signal losses by utilizing new selectively stable isotope labeled amino acids. We will devise new label- Characterization of protein structures by NMR requires ing schemes for all twenty natural amino acids to enable unambiguous assignment of backbone and side chain these studies. The approach will be validated by applying resonances and determination of unequivocal distance it initially to structure determination of a large enzyme constraints from nuclear Overhauser effects (nOe) (34kDa) and finally to structural studies of a protein measured in NOESY experiments (Figure 1). These con- complex involved in the biosynthesis of enterobactin straints are used for structure calculations. With increas- (132kDa). ing protein size, the number of resonances increases dramatically leading to crowded or overlapping NMR The success of this project will open the structural biol- spectra. Current established strategies rely on proteins ogy field to the detailed mechanistic studies of other uniformly enriched with 13C and 15N and/or with selec- large protein complexes and membrane-bound proteins tive enrichment of the methyl groups of Leu, Val and Ile involved in many biological processes placing LANL at with 13C and full protonation (e.g. 13C1H3). The latter the forefront of the field. Leveraging the results obtained selectively enriched amino acids are introduced into 100

Page 101

perdeuterated proteins [2]. The biomolecule is typically data points to be recorded without loss of resolution uniformly enriched in 15N so that nOes involving those and sensitivity, shortening data acquisition time, and residual protons (amide and methyl groups) can be mea- dramatically improving sensitivity. sured and identified. This sparse labeling approach gener- • Develop new, selective stable isotope labeling patterns ates strong methyl-group nOes and allows their assign- for amino acids and incorporate them into target pro- ments but obviously reduces the total number of resulting teins and complexes. We devised new stable isotope distance constraints dramatically, leading to less well labeling schemes for selected amino acids to simplify defined structures. Further, faster relaxation in larger pro- NMR studies of large and membrane inserted biomol- teins leads to signal losses and decreased peak intensities. ecules. The isotopic labels were to be incorporated Hence, unambiguous data are not easily accessible and at specific positions to optimize NMR signal and limit structure determination is compromised. To allow for high- relaxation losses during experiments. Co-optimization resolution NMR studies of large proteins and complexes, of sample labeling and NMR methods development a specific combination of amino acid labeling patterns is will improve the spectral resolution and signal quality required to optimize magnetization transfer pathways in allowing very large and membrane-inserted proteins to NMR experiments (Figure 2). become accessible by NMR spectroscopy for the first time. We attempted to prepare the selectively labeled amino acids will by direct chemical synthesis and/or by using labeled metabolic precursors (glycerol, pyruvate, glycine, succinate) for in vivo protein production to al- low very specific and selective isotopic enrichment. • Validate the new NMR methods and labeling patterns Figure 1. Flow chart of solution structure determination from by applying them to structure determination of a NMR data. large protein complex. Little is known about the large enzymatic assembly lines responsible for the biosyn- thesis of bioactive compounds (antibiotics, anti-cancer drugs, virulence factors, toxins, etc.). We employed the stable isotope labeling schemes and the NMR methods developed above to solution NMR studies of a 34kDa enzyme and a large, 132kDa, protein complex involved in the biosynthesis of enterobactin. The complex will be the largest structure determined by solution NMR. Scientific Approach and Accomplishments Amino acid labeling for NMR structural studies of large systems. While synthetic schemes for many isotopically labeled amino acids have been fully developed, they Figure 2. Magnetization transfer pathways in uniformly labeled, have not been used to produce amino acids with the site glycerol enriched and our ‘designer’ amino acids. specific isotopic labels required to support our studies of In this project, we have set out the following goals: large biological systems using our new NMR methods. The critical point is that the spin systems (13C-1H, 15N-1H, • Develop NMR methods optimized for selectively 15N-13Cα) must be isolated to minimize relaxation losses labeled large proteins and complexes. We aimed to while still allowing for necessary magnetization transfer to develop new 3D and 4D NMR experiments for reso- make resonance assignments (Figures 2 and 3). For optimal nance assignments and unambiguous determination NMR performance, all amino acids have amino nitrogens of distance and orientational constraints with supe- labeled with N-15 and α-carbons labeled with C-13 and rior sensitivity for large proteins and complexes. The duterated. The carbonyl carbons are unlabeled and the methods are designed to take advantage of optimized side chains have C-13 labeled and protonated carbons magnetization transfer pathways to minimize signal alternating with unlabeled deuterated carbons. A common losses by utilizing new amino acids selectively labeled precursors to all amino acids is [2-13C, 15N, 2H]-glycine with stable isotopes. We also aimed to optimize non- attached to a chiral auxiliary synthon (2,10-camphorsul- uniform sampling schemes and data reconstruction tam, Figure 3). We have synthesized [2-13C, 15N]-glycine algorithms that require only a fraction of all potential 101

Page 102

from 2-13C-bromoacetic acid and 15N-potasium succin- tion, extraction) and/or by substituting reaction solvents imide. The reaction can be performed on 1 mole scale and with their deuterated versions. While increasing the cost of proceeds with good yield. The purified glycine was subse- synthesis, this led to highly deuterated products. quently protected by esterification with methanol on the All amino acid side chains were synthesized individually carboxyl group and with thiomethyl groups on the amino and attached to the glycinate sultam, cleaved and depro- functionality. Deuteration of amino- and carboxyl-protect- tected to yield 18 natural amino acids. Two amino acids, ed glycine proceeded at 25-40oC with catalytic amounts of cysteine and tryptophan, were obtained using enzymatic sodium metal in methanol-d and resulted in 98%+ dutera- reaction of tryptophan synthase, with labeled serine and tion and pure product with minimal workup (extraction). benzenethiol (cysteine) or indole (tryptophan) as sub- The deuterated protected glycine was attached to the strates. Overall, we successfully synthesized all 20 natural sultam and served as the initial building block for all amino amino acids. Our labeling pattern and synthesis, as well acids (Figure 3). as observations of isotopic dilution during deuteration are currently being written up for publication. In vivo selective protein labeling. For metabolic labeling of proteins, we have cloned the gene for the small protein, ubiquitin (76 amino acids) into E. coli hosts. Expression of this protein under various labeling conditions allowed us to monitor incorporation of isotopes at specific positions in amino acids, since NMR spectra for ubiquitin are well resolved and characterized. We have performed growth in the presence of labeled glycine (which we synthesized), py- ruvate or succinate. Analysis of produced ubiquitin showed Figure 3. General synthetic scheme for preparation of labeled significant scrambling of all isotopes leading us to conclude L-amino acids. The common precursor, glycinate sultam, is inside that metabolic labeling using these precursors will not be the blue rectangle. acceptable for NMR of large proteins. Analysis of isotope incorporation in the protein shows that for this approach Synthetic pathways for the production of isotopically to be successful we would need to judiciously incorporate enriched amino acids have been worked out over the last inhibitors of amino acid synthesis into growth media to ac- decades, with major contributions from the Stable Isotope complish desired labeling. Since the chemical synthesis can Resource (SIR) at LANL and Kainosho’s group at RIKEN, provide large quantities of amino acids with exact labeling Japan [3]. While we have the expertise and technology to patterns, the in vivo approach was abandoned. synthesize all natural amino acids with any site specific isotope labels to allow for the proposed high-resolution NMR methods development. For successful NMR studies NMR studies, the specific combinations of labeling pat- of large and complex systems it is critical to limit possible terns required to optimize magnetization transfer path- relaxation pathways, reduce the length of pulse sequences ways for studies on very large systems have not yet been to minimize signal losses, and reduce experimental time, done. The needed labeling pattern necessary poses its own which is usually very long (days to weeks). For NMR meth- unique problems, particularly with respect to deuterium ods development part of the project, we have modified incorporation. While 13C and 15N labels cannot undergo and tested recently developed 3D TROSY-hNCAnH and exchange to 12C and 14N, depending on the chemical 3D TROSY-hN(CO)CAnH NMR experiments [4] by convert- structure, some deuterium positions may be exchangeable ing them to double-TROSY versions for further increased with solvent leading to isotopic dilution. For preparation sensitivity in large proteins. The experiments were tested of good NMR samples, amino acids need to be >98% 13C, on a large, 34kDa protein diisopropyl fluorophosphatase >98% 15N and >95% D at selected positions. Early on we (DFPase) from Loligo vulgari, an enzyme that hydrolyses noticed that several standard synthetic steps lead to isoto- organophosphates. It is an excellent test system – large, pic dilution at deuterated positions. One example is silica very stable and with well-defined structure. gel purification during which acidity of the silica gel was sufficient to pull of the deuterium atom and exchange it For data acquisition and handling we have implemented with the hydrogen from the solvent system. We have also and tested on DFPase Non-Uniform Sampling (NUS) observed isotopic dilution when oxalyl chloride was used schemes based on Iterative Soft Threshold (IST) recon- in esterification reactions. We mitigated these problems struction algorithms [5] to determine minimum number by resorting to alternate purification methods (crystalliza- of data points in multidimensional experiments that allow 102

Page 103

for faithful reconstruction. Since the method allows for The project resulted in conversion of Alex Koglin from Op- collection of significantly reduced number of data points in penheimer Postdoctoral Fellow to Scientist 2. Alex brought the experiment, it allows us to reduce experimental time to LANL new capabilities in fast discovery of natural thera- 5-10-fold for 3-dimensional NMR experiments and 15-20- peutics using genome mining and analysis of gene cluster- fold for 4-dimensional experiments. ing, work that he is continuing as an independent scientist. NMR structure determination of large proteins. We have References performed all experiments necessary for assignments and

1. Hyberts, S. G., D. P. Frueh, Arthanari, and Wagner. FM structure calculations, including our new and improved reconstruction of non-uniformly sampled protein NMR NMR experiments, on a sample of DFPase. We have ob- data at higher dimensions and optimization by distilla- tained all assignments and distance constraints and are in tion. 2009. JOURNAL OF BIOMOLECULAR NMR. 45 (3): the final stages of structure validation for publication and

release. For the large enterobactin complex, isotopic dilu- tion issues during synthesis have delayed protein synthe- 2. Tugarinov, V., and L. E. Kay. Ile, Leu, and Val methyl as- sis. The protein complex is currently being prepared for signments of the 723-residue malate synthase G using NMR analysis. Structure calculation of this large complex a new labeling strategy and novel NMR methods. 2003. will follow and will be the largest de novo single structure JOURNAL OF THE AMERICAN CHEMICAL SOCIETY. 125 obtained by high-resolution NMR. (45): 13868. Summary. During the course of the project, we have syn- 3. Ohki, , and Kainosho. Stable isotope labeling methods thesized all twenty natural amino acids with well defined, for protein NMR spectroscopy. 2008. PROGRESS IN specific stable isotope labeling patterns. The expertise NUCLEAR MAGNETIC RESONANCE SPECTROSCOPY. 53 we gained in this area puts us at the forefront of protein (4): 208. labeling methodologies. We explored in vivo protein label- ing and decided that it was not selective enough for our 4. Frueh, D. P., Arthanari, Koglin, C. T. Walsh, and Wag- needs. We have improved sensitivity of the necessary 3D ner. A Double TROSY hNCAnH Experiment for Efficient NMR experiments and applied them to study of an impor- Assignment of Large and Challenging Proteins. 2009. tant enzyme, DFPase. We have also implemented non-uni- JOURNAL OF THE AMERICAN CHEMICAL SOCIETY. 131 form sampling methods for significantly faster NMR data (36): 12880. acquisition. We are in the final stages of DFPase structure 5. Hyberts, S. G., A. G. Milbradt, A. B. Wagner, Arthanari, validation and began work on the large protein complex, and Wagner. Application of iterative soft threshold- enterobactin synthase, which will be the largest structure ing for fast reconstruction of NMR data non-uniformly solved and will place us at the forefront of modern struc- sampled with multidimensional Poisson Gap schedul- tural biology. ing. 2012. JOURNAL OF BIOMOLECULAR NMR. 52 (4): 315. Impact on National Missions Knowledge of the structures and mechanisms of both large proteins and their complexes and membrane-bound proteins is critical to our understanding of and interven- tion in processes as varied as antibiotic resistance and biosynthesis, biofuels production, and disease develop- ment (Parkinson’s, Alzheimer’s, diabetes, etc.). This project is directly related to LANL’s missions in Global Security, Biothreat Reduction, and Energy Security and will open the structural biology field to the detailed mechanistic studies of large protein complexes and membrane-bound proteins involved in the above as well as many other processes placing LANL at the forefront of the field. Leveraging the results obtained here with LANL’s synthetic stable isotope capabilities, will open funding areas from NIH, DTRA and DOE. We will work closely with program managers in Global Security (Kirsten McCabe) and DOE (BES/BER) so that future funding for these efforts can be developed. 103

Page 104

Complex Natural & Engineered Systems Exploratory Research Final Report Redox active Catalysts for C-C Coupling Reactions Relevant to Renewable Energy John C. Gordon 20130672ER Abstract Chemicals from Biomass. To the best of our knowledge The majority of glycerol is currently sourced from animal there are currently no useful methods available to take fats and vegetable oils, with a small amount made syn- advantage of these kinds of glycerol derived synthons in thetically for ultrapure applications. Production is rapidly this regard. This is in spite of the low cost of glycerol due increasing due to biofuel directives being implemented. to its abundance as a by-product of biodiesel produc- Projections suggest that by 2020 the production of tion. As mentioned above, glycerol is readily converted glycerol will be six times that of the demand, as usage is into species such as acetol and acrolein, which contain limited to food additives, the pharmaceutical industries carbonyl (C=O) functionalities potentially amenable to and some chemical applications, including conversion to participation in Pinacol type self- and cross-coupling acrolein, acetol, ethanol and epichlorohydrin. Novel ap- reactions with other C3-C6 biomass derived molecules. proaches to sustainable fuel sources have been focused Current approaches for glycerol conversion are essential- on generation of bio-diesel from various feedstocks. As ly limited to use in the energy inefficient Fischer-Tropsch a major by-product of biodiesel production, glycerol has process. The Pinacol coupling reaction is a 2-electron found some use as a food additive, in coolant applica- process that can aid in forming C-C bonds, however this tions, and currently there is a great deal of interest in presents significant challenge for Fe as this metal more conversion of glycerol into other chemicals including readily facilitates 1-electron chemistry. Nonetheless, the propanediol, acrolein, and polyethers along with many goal of using an Fe-based catalyst for this reaction type other useful chemicals. Thus far there has been little is attractive due to the pairing of an inexpensive metal discussion of creating larger carbon chains derived from with potentially very large scale biomass-derived sub- glycerol or its derivatives, aside from Fisher-Tropsch strates chemisties. This proposal aims to generate a method Scientific Approach and Accomplishments based on iron catalysts to convert renewable glycerol de- rived C3 fragments into C6 species that can then be used This work has attempted to use redox active ligands in subsequent fuels and feedstocks chemistry. capable of conferring up to two-electron redox behavior. This is a possible way to promote bond-making reactions Background and Research Objectives and carbon chain extension reactions involving carbonyl The research in this project has targeted the develop- containing substrates. We began with the interrogation ment of inexpensive catalysts, comprised of earth of diimine ligands to try to facilitate the desired chem- abundant metals, which are capable of supporting mul- istry. A series of (diimine)Fe complexes were made and tiple electron reductions pertinent to biomass derived their efficacy towards carbonyl (C=O) containing sub- molecule upgrading. This chemistry was aimed towards strate coupling reactions was investigated. This provided the promotion of carbon-carbon bond forming reactions mixed results. In retrospect, we believe that this par- of ketonic and aldehydic moieties, with the ultimate ticular class of ligands was insufficently oxidizing with goal being the conversion of readily available three- respect to the iron centers to which they were bound. carbon fragments derived from glycerol (e.g. acrolein, When we used model substrate compounds such as ben- acetol (hydroxyacetone), etc.) into six (or higher) carbon zophenone, acetophenone, etc. to interrogate reactivity, containing fragments that are then potentially viable as instead of the desired coupling reactions, we primarily fuel precursors or can be further converted into higher observed substrate binding via an arene group, rather value added feedstock chemical feedstocks. A recent than through binding of the carbonyl group by the metal DOE report highlights a list of targeted Top Value-Added center. This led us to the conclusion that the Fe centers 104

Page 105

were less oxidized by these ligand systems (therefore lower valent and less oxophilic) than we had hoped for in order to facilitate the targeted Pinacol chemistry. These arene complexes are interesting in their own right and a publica- tion should result on their chemistries. From our initial results, the more a prospective complex can present itself as containing Fe(II), the better it should be able to facilitate the desired chemistry. This poses the question of how can the redox activity of the ligand be chosen such that we can exercise more granular control over the electronic state and hence the reactivity of the metal center? Two possible approaches to this question are the incorporation of an oxidant into the ligand and the creation of an internal barrier to ligand oxidation. Quinone and various derivatives thereof are well-known 2-electron organic oxidants, and as several 1,2-diimino quinones have been synthesized and, in one case, used as a ligand for Cu. We have considered this kind of ligand motif as one that can faciiltate the desired redox chemistry. We have most recently used computations to probe the redox behavior of potentially dianionic quinone supported catalysts and have set about synthetically accessing these. Impact on National Missions This work advances our understanding of the chemistry and chemical methods that directly bear upon energy security. Success in this area would seed the development of future approaches to more efficient chemistries, poten- tially providing a bridge from basic science to global scale energy solutions. This is a high profile area of renewable energy. Work in this area is relevant DOE Office of Basic En- ergy Sciences (Catalysis Science) and DOE Office of Energy Efficiency and Renewable Energy (BETO). 105

Page 106

Complex Natural & Engineered Systems Exploratory Research Final Report Novel Chemical Architectures for Supercapacitor Electrolytes: Comparing In Situ Scattering Measurements to Theory and Simulation Cynthia F. Welch 20130681ER Abstract charge/discharge rates, strong chemical durability for Many renewable energy sources produce only intermit- long device lifetimes, good thermal stability to provide tent electricity generation. Supercapacitors promise to function over a very broad range of temperatures, and be one leading scheme for storing that energy for stable low volatility and flammability for enhanced safety [1]. use. If their energy densities can be improved, they may Recently, Gogotsi [2] and coworkers demonstrated that find application in a range of devices since they can de- carefully matching the pore size in the electrode to the liver an order of magnitude more power than chemical electrolyte moieties enhances capacitance dramatically. batteries. Recent studies suggest that matching electro- They accomplished this via a clever fabrication route to lyte size to electrode pore geometry yields dramatic in- control the electrode pore structure. Little attention has creases in capacitance. Most efforts to improve superca- been paid to electrolyte size and shape. pacitor performance focus on optimizing the electrode. We propose that tailoring the size and shape of the In this study, we took the novel approach of optimizing electrolyte in supercapacitors to match the electrode the electrolyte by controlling its constituent size and surface geometry will lead to enhanced energy densities, shape distributions. In addition to our integrated ex- forming our central hypothesis. This hypothesis leads perimental and modeling tools, we developed a new us to consider dendrimer electrolytes as candidates for theoretical method to describe these complex systems; use in supercapacitors, as well as to study their behav- this method enabled linking specific chemistries to our ior near and in the electrode pores. The novelty of our multi-scale simulation models. The primary product of proposal lies in optimizing the electrolyte rather than this study is a predictive framework that ties electrolyte the electrode; this approach may prove more potent and chemistry to ultimate device performance for standard robust than modifying the electrode. Certain dendritic electrode configurations, as well as new fundamental in- chemistries possess all of the key properties needed in sight into the molecular-level behavior of these complex a good electrolyte, and several specific aspects of the systems. physical chemistry recommend them for application in supercapacitors. These characteristics include precisely Background and Research Objectives controlled size and shape, broad range of available sizes, Many processes involved in manufacturing and trans- high charge densities, low viscosities at relevant concen- porting goods suffer from inefficiencies due to wasted trations, and the ability to capture impurities. Many key energy. Similarly, many renewable energy sources (such questions about the behavior of the electrolytes near as wind and solar) yield only intermittent electricity electrodes must be answered for progress to be made. generation. We therefore require advanced devices By combining experiment, theory, and simulation in this that quickly capture and release stored energy. If their project, we provide insight into many of these questions. energy densities can be improved, given that superca- pacitors may deliver an order of magnitude more power Scientific Approach and Accomplishments than chemical batteries, they may dramatically improve Theoretical description of conditions needed for den- the performance of dynamic energy capture and re- drimer electrolyte adsorption to flat electrode surface use systems. This may be achieved by improvements Our envisioned application of dendrimers as supercapac- to either the electrode designs or to the electrolyte. itor electrolytes revolves around placing these molecules Nanoporous carbon in various forms often serves as the at and removing them from charged surfaces. Thus, we electrode material. Key properties for a successful elec- need a prescription for the conditions needed to re- trolyte material include high ionic conductivity for fast 106

Page 107

lease a charged dendrimer from an oppositely charged flat the dendrimers normal to the electrode. This gives infor- substrate. We developed this prescription by (a) identifying mation on the thickness of the dendrimer layer adsorbed an effective segment step length that reflects the intramo- to the electrode surface, as well as on how that thickness lecular repulsions due to excluded volume and electrostat- changes with applied voltage. Though our initial results ics, as well as the dendrimer’s branching, and (b) using suggested a confirmation of the theory and simulation that effective step length to derive an analytical equation results described in (2) above, we were unable to follow that describes the boundary between captured and free up with further experiments at SPEAR or any other neu- dendrimers. We validated this theoretical prediction by tron reflectivity beam line, due to budget cuts in both the performing Monte Carlo computer simulations of coarse- BES-supported Lujan Center at LANSCE and in the LDRD grained model dendrimers escaping from charged surfaces. program. The simulations considered generations 2 through 6 with Simulations to explore behavior of PAMAM dendrimer a range of lengths between the branch points, as well as a electrolyte at an oppositely charged flat electrode surface range of solution ionic strengths and surface charge densi- Molecular dynamics simulations were performed to ties. Full details can be found in our publication on this observe structural features at the electrode-electrolyte in- subject: J. Chem. Phys. 2013, 139, 164906. terface in order to aid in the interpretation of the NR data Scaling theory and simulations to describe interactions described in (3) above. These simulations were aimed at between concentrated dendrimer electrolyte solution specifically modeling PAMAM dendrimer chemistry in the and flat electrode surface presence of an oppositely charged carbon flat electrode. A single dendrimer electrolyte molecule can contain a To do so, we used two different united atom models: an large number of charged groups that are available to extended version of the TraPPE-UA force field [4] (Transfer- interact with an oppositely charged electrode surface. We able Potentials for Phase Equilibria – United Atom; param- need to understand how to maximize these interactions, eterized for PAMAM as part of this project) and the coarse- thereby maximizing capacitance. To aid in this effort, we grained MARTINI force field (parameterized for PAMAM by used the results of (1) above to develop a scaling theory Lee and Larson [5]). Various simulations were performed that describes the thickness of an adsorbed dendrimer to test the effectiveness of these models in simulating layer at an electrode surface, as well as the fraction of seg- PAMAM dendrimers. ments within the dendrimer that touch the surface, and Both models give reasonable results for the size of PAMAM the total number of dendrimers adsorbed as a function of dendrimers at different generations of growth and charge generation of growth (i.e., dendrimer size), surface charge densities. Both were also used to simulate the mobility of density, and dendrimer concentration. We demonstrate PAMAM dendrimers in the presence of electric fields and that these predictions agree well with extensive molecular to model dendrimer-dendrimer interactions in concentrat- dynamics simulations, which span a range of concentra- ed solutions. Overall, the TraPPE model was able to explore tions, electrode surface charges, and dendrimer electrolyte a larger range of electric field strengths and produce more size. Combined, the simulations and scaling argument realistic PAMAM dendrimer electrolyte / carbon electrode indicate that simultaneous adsorption and compression interactions in the presence of aqueous solvent molecules. at the interface take place. Full details can be found in our The fully charged PAMAM dendrimer does not adsorb to publication on this subject: ACS Macro Lett. 2014, 3, 180. the carbon electrode surface at zero field (as expected). In situ experiments to probe structure of dendrimer On the other hand, a high electric field causes the den- electrolyte-electrode interface drimer to adsorb and flatten on the surface, while the Building on (2) above, we conducted a series of experi- counterions move to the opposite surface. When the field ments on poly(amido amine) (PAMAM) dendrimer elec- is removed, both the dendrimer and counterions desorb. trolytes in combination with flat oppositely charged This behavior is consistent with the results obtained in (2) electrodes to better understand electrolyte-electrode above for the molecular dynamics simulations that did not interactions. For the experiments, we performed in situ account for specific chemical interactions. Further details neutron reflectivity (NR) experiments using an electro- on the chemistry-specific simulations will be described chemical cell [3] at the SPEAR beam line at the Los Ala- in manuscripts in preparation for submission to peer- mos Neutron Science Center (LANSCE). This cell features reviewed journals. In summary, the results suggest that a quartz substrate coated with palladium as the working PAMAM dendrimers will behave as good electrolytes in a electrode and a counter electrode of stainless steel coated supercapacitor with carbon electrodes. with gold. With this cell, we varied the voltage applied across the electrodes and then measured changes in the in-plane averaged scattering length density (SLD) profile of 107

Page 108

Simulations to evaluate stability of PAMAM dendrimers on the work described in (2) and (6) above to perform a against chemical degradation due to exposure to electric series of coarse-grained Brownian dynamics simulations fields on a set of model dendrimers, linear oligomers, and small We carried out first-principles molecular dynamics simula- electrolyte ions in the presence of a model porous elec- tions using the finite element density functional theory trode stack. Motivated by the computational results and code FEMTECK, which was developed in Japan by E. guided by all previous results obtained in (1) – (6) above, Tsuchida [6]. FEMTECK is an efficient simulation code we subsequently tested the performance of generation with the ability to apply an electric field. To gain access to 3 and generation 5 PAMAM dendrimer electrolytes in this unique capability, we developed a collaboration with supercapacitors composed of both flat carbon and porous Drs. Tsuchida and Choe at Japan’s National Institute of activated carbon electrodes. Those results are compared Advanced Industrial Science & Technology (AIST), and we against the performance of a typical supercapacitor elec- sent one of our postdoctoral researchers to AIST in Japan trolyte, sulfuric acid. We find that: a) there exists a mini- to learn how to use the code and port it to LANL’s Institu- mum charge density required for electrolyte molecules of tional Computing resources. We then carried out simula- different molecular sizes to enter the pores; b) the amount tions of solvated G0 PAMAM at high, neutral, and low pH of cationic electrolyte charge partitioned into the cathode (corresponding to uncharged, partially charged, and fully pores grows linearly with electrode charge density once charged PAMAM, respectively), with and without an ap- this minimum electrode charge has been met; c) the total plied electric field. We see more proton transfers between net charge on the electrodes depends on the molecular the dendrimer and the water for charged dendrimers with architecture; and d) dendrimers outperform a typical the electric field applied, but we do not see any evidence small molecule electrolyte at the same molarity of charge of carbon-carbon or carbon-nitrogen bonds breaking, sug- in experimental measurements of charge accumulation gesting that PAMAM dendrimers are stable against electro- at porous activated carbon cathodes, as predicted by the lytic degradation. simulations and supporting our proposition. We are in the process of submitting this work for publication in a peer- Scaling theory and device-level simulations to describe reviewed journal. capacitance of dendrimers Based on geometrical considerations of dendrimer electro- Impact on National Missions lytes, we developed a scaling theory to predict their capac- We expect agencies such as DOE EERE and ARPA-E, as itance at a flat electrode surface. This theory shows that well as the renewable energy industrial sector, to find the for electrode surfaces that are saturated with dendrimer results of this study a solid foundation for future funded electrolytes, capacitance should decrease with increas- programs. Therefore, we are preparing proposals aimed at ing dendrimer generation of growth (or size). We have these DOE agencies, and we are discussing our results with also examined the partitioning of dendrimer electrolytes LANL’s Feynman Center for Innovation to establish a path into pores that mimic those of porous electrodes. These forward for potential patents and collaborations with in- simulations show optimal behavior for capacitance when dustry. The collaboration with AIST (see (5) above) brought dendrimer size matches that of the pore. More specifically, a new theory and modeling capability (FEMTECK) to LANL. dendrimers whose size is smaller than the pore can easily This project also produced two peer-reviewed publications enter and exit the pore, while those that are too large tend (with 4 – 5 more expected submissions), as well as multiple to get stuck at the pore entry, thereby blocking the pore. presentations at local, national, and international venues. Dendrimers whose size matches the pore enter readily and Furthermore, it supported the hiring of two postdoctoral then remain inside the pore for extended periods of time. researchers, both of whom have brought new expertise These results are described in more detail in 1 – 2 manu- and capabilities to LANL. One of these postdocs has re- scripts being prepared for submission to peer-reviewed cently been awarded a Seaborg Institute Postdoctoral Fel- journals. lowship. In addition, this project provided partial support for 3 graduate student interns at LANL, 2 of whom were Simulations and experiments aimed at testing the central converted to LANL Technologists. hypothesis of this study, that the use of dendritic electro- lytes will lead to higher charge storage in supercapacitors References We propose that dendrimer polyelectrolytes may provide

1. Basic Research Needs for Electrical Energy Storage. enhanced supercapacitor performance relative to small
2. Office of Basic Energy Sciences, U.S. Department molecule electrolytes when paired with porous electrodes of Energy. with confining geometries commensurate with the size of the electrolyte molecules. To test this hypothesis, we built
3. Chmiola, , Yushin, Gogotsi, Portet, Simon, and P. L. 108

Page 109

Taberna. Anomalous increase in carbon capacitance at pore sizes less than 1 nanometer. 2006. SCIENCE. 313 (5794): 1760. 3. Zamlynny, V., I. Burgess, G. Szymanski, J. Lipkowski, J. Majewski, G. Smith, S. Satija, and R. Ivkov. Electro- chemical and neutron reflectivity studies of spontane- ously formed amphiphilic surfactant bilayers at the gold-solution interface. 2000. LANGMUIR. 16 (25): 9861. 4. Wick, C. D., J. M. Stubbs, N. Rai, and J. I. Siepmann. Transferable potentials for phase equilibria. 7. Primary, secondary, and tertiary amines, nitroalkanes and ni- trobenzene, nitriles, amides, pyridine, and pyrimidine. 2005. JOURNAL OF PHYSICAL CHEMISTRY B. 109 (40): 18974. 5. Lee, , and R. G. Larson. Effects of PEGylation on the Size and Internal Structure of Dendrimers: Self-Penetration of Long PEG Chains into the Dendrimer Core. 2011. MACROMOLECULES. 44 (7): 2291. 6. Tsuchida, E., and M. Tsukada. Adaptive finite-element method for electronic-structure calculations. 1996. PHYSICAL REVIEW B. 54 (11): 7602. Publications Maerzke, K. A., N. J. Henson, P. M. Welch, and C. F. Welch. Deformation of poly(amido amine) dendrimers at sur- faces. Presented at 249th American Chemical Society National Meeting. (Denver, CO, 22-26 Mar. 2015). Maerzke, K. A., N. J. Henson, P. M. Welch, and C. F. Welch. Deformation of poly(amido amine) dendrimers at charged surfaces. Invited presentation at National In- stitute of Advanced Industrial Science and Technology (AIST). (Tsukuba, Japan, Oct. 2014). Welch, C., P. Welch, N. Henson, V. Hartung, T. Lin, S. Ed- wards, and J. Stull. Novel chemical architectures for su- percapacitor electrolytes: Comparing in situ scattering measurements to theory and simulation . Invited pre- sentation at Los Alamos National Laboratory Materials for the Future Capability Review. (Los Alamos, NM, 4-7 May 2014). Welch, P. M.. The escape of a charged dendrimer from an oppositely charged planar surface. 2013. Journal of Chemical Physics. 139: 164906. Welch, P. M., C. F. Welch, and N. J. Henson. The flattening of dendrimers from solutions onto charged surfaces. 2014. ACS Macro Letters. 3: 180. 109

Page 110

Complex Natural & Engineered Systems Exploratory Research Final Report One-step Supercritical Fluid Extraction (SFE) and Separation of Rare Earths (RE) Stephen L. Yarbro 20140186ER Abstract In this case, the separation factor is calculated on the We have demonstrated a prototype system (Figure 1) basis of the heavier organic molecule and a separation that exploits the solvent power of supercritical carbon factor greater than one indicates the nozzle is separating dioxide (SCCO2) with co-solvents for lanthanide (rare the two components. Therefore, this approach has the earth) separation from oxides and potentially ore. Also, potential to simplify lanthanide processing and separa- we have demonstrated that the different lanthanide- tion. Essentially, in one-step, rare earth elements are organic species can be separated in a curved diverging extracted directly from oxide with the SCCO2/co-solvent nozzle while still in the supercritical gas phase. The key mixture (typically tri-butyl phosphate (TBP)) and sepa- results are: rated with an aerodynamic method. The aerodynamic method separates the heavier elements from the lighter • Essentially 100% extraction in 100 minutes elements while dis-solved in the supercritical fluid us- ing the centrifugal force. Current production methods • Nd Predicted separation factor – 1.3 Measured require complex equipment and many steps such as separation factor – 1.2 to 1.8 dissolution, flotation, and filtration, and washing, sepa- ration requiring hun¬dreds of separation stages, precipi- • Separation factor is the ratio of concentration ratios: tation, and calcination. Supercritical fluids are power- ([Nd]/[Yb])Output1¬ / ([Nd]/[Yb])Output2 ful sol¬vents. Coupled with the right co-solvents, the supercritical fluid can selectively remove the lan¬thanide from the ore, and then use an aerodynamic technique developed for isotope separation to separate the lan- thanides while still in the homogeneous supercritical fluid. After separation, reducing the pressure recovers the lanthanide solids thereby eliminating the majority of the current production steps along with the associated complexity, cost, and waste. Background and Research Objectives Rare earth elements (REEs) are a group of 15 chemical elements in the periodic table, specifically the lan- thanides. Although relatively abundant in the earth’s crust, REEs rarely occur in concentrated forms, making them economically challenging to obtain. These ele- ments constitute critical components of many important technologies and products, such as hybrid vehicles, wind turbines, and cell phones. Given this global demand for green and sustainable products in energy, military, and manufacturing industries, REE demand in the United States and throughout the world is projected to in- Figure 1. Prototype SCCO Lanthanide Extraction and Separa- crease. In recent years, China has been providing 95 to tion System. 97 percent of REEs worldwide. Because China has 110

Page 111

different stra-tegic materials, including uranium ores and potentially spent fuel. Our long-term goal is for Los Ala- mos to be a world leader in using novel methods, such as supercritical fluids for producing strategic metals that are critical to US defense and security needs. Scientific Approach and Accomplishments Experimental Accomplishments We used a commer¬cially available “off-the-shelf” (COTS) SuperCritical Fluid Technologies SFT-110 unit (Figure 2). We Table 1. Key results from extraction and separation tests collaborated with the LANL Center for Inte¬grated Nano- demonstrated its ability to control and limit REE exports, it technologies (CINT) and their mico-fabrication capa¬bility is crucial that the United States expand its ability to obtain to build simple nozzles for testing. Essentially, there were REE resources. Mining in the natural environment is the two phases to the R&D plan: (1) Measure lanthanide primary means of REE acquisition; however, it results in solubility’s from a well-characterized oxide substrate to a large quantity (greater than 90 percent) of excess and develop the procedures and analytical support, then (2) unused materials and other environmental impacts. If the Use the solubility data to design and fabricate a simple United States is to ensure a continuous supply of REEs, nozzle device and with select SCCO2 streams, measure responsible mining and extraction practices will need to separation factors and validate the CFD model. The nozzle be developed and enhanced. Additionally, effective recy- was designed based on the maximum pressure and flow- cling, recovery, and reuse of spent consumer and industrial rate through the SFT-110 unit. After the nozzle design was products may reduce the need to develop new mineral completed, monolithic stainless steel nozzle construction resource areas. was achieved via 3-D laser printing, and stainless tubing was brazed for installation (Figure 3). Once the unit was The increasing importance of REEs for the manufacture of installed a series of solubility tests at varying flowrates modern devices, upon which society has become reliant, and acid concentrations was completed to measure the along with uncertain supplies, is encouraging exploration average extraction efficiency and extraction rate. With this and development of new mining sites. While REEs are an data, a series of tests was completed to measure separa- important resource needed to sustain our modern tech- tion factors. nologies, the waste footprint and environmental impact from conventional REE mining and extraction operations is significant. The requirements, regulations, and financial obligations and assurances for a new mine and associated processing are usually complex and take years of plan- ning. Therefore, enhanced technology that reduces the complexity and waste generation with a high recovery and separation efficiency would have a large economic benefit and reduce the current environmental impact. We have measured the solubility and separation efficiency of selected strategic lanthanides (Nd, Yb) from oxide samples using SCCO2 and a co-solvent TBP. Our plan was: (1) based on the literature, choose successful co-solvent (TBP) and measure solubility at different conditions, and (2) develop a simple aerodynamic system based on the Becker nozzle for demonstration and measure separa- tion efficiencies at different conditions. We have demon- strated the technical feasibility and ability for reducing the Figure 2. SFT-110 COTS SCCOExtraction System without Separator com¬plexity and cost of production of strategic lanthanide metals. Long-term goal of this work: Ultimately, we intend for this work to provide the engineering basis for developing a program for using supercritical fluids to process a variety of 111

Page 112

cates that additional chemistry is changing the weight ratio and therefore the separation factor. Impact on National Missions In 2012, Congress commissioned a study on the defense implications of China’s control of the rare earth supply and the impact on US security [2]. Con¬gress is encouraging DOD to develop a long-term strategy to identify material weaknesses and vulnerabilities associ¬ated with RE’s and to protect long-term U.S. national se¬curity interests. In addition, The Department of Energy (DOE) Energy Efficien- cy and Renewable Energy (EERE) Advanced Manufacturing Office is providing $120M to fund a Critical Materials Strat- egy focusing on critical materi¬als supply issues (Figure 4) [3]. A portion of their mission is to enable new sources of Figure 3. Cutaway View of Separator Nozzle with Output critical materials not now commercially viable, improve the Streams Identified economics of processing existing sources, and identify new uses for co-products and by-products that do not currently Modeling contribute to the economics of materials production. This In the Becker process for uranium enrichment [1], a gas jet mission integrates well with LANL’s Materials for the Future of ~96% hydrogen and 4% UF6 expands through a narrow Science Pillar’s vision, “Pursuing the discovery sci¬ence slit. The gas moves at high speeds (comparable to those and engineering required to establish design principles, in a modern centrifuge) parallel to a semicircular wall of synthesis pathways, and manufac¬turing processes …”. very small radius of curvature (Figure 3). At gas velocities We envision demonstrating the process on lanthanides to approaching 400 m/s, with a radius of curvature of 0.1 build a relation¬ship with EERE and then extend the work mm, the centrifugal acceleration achieved is 1.6 x l09 m/ to actinides to build programs related to nuclear material s2 or 160 million times gravity. This acceleration exceeds processing issues. those achieved in centrifuges by a factor of a thousand or more, in an apparatus with no moving parts. The centrifu- gal force causes the heavier components to move closer to the curved wall than the lighter components as it flows around the semicircle. At the other side, where the gas has changed direction by 180°, a knife-edge separates the flow into a light fraction and a heavy fraction. We have demonstrated that lanthanides dissolved in the supercriti- cal fluid phase can be separated by a similar technique. As a part of the investigation of the solubility of lanthanides in supercritical carbon dioxide with co-solvents, we have developed a 2-D model of curved flow in a Becker nozzle to provide the basis for a preliminary design for testing. The model is based on the non-dimensionalized Navier-Stokes equations for incompressible flow in Cartesian coordinates. Figure 4. Relation of Supply Risk to Mission Critical Elements In order to calculate the pressure field, the fluid is assumed to be slightly compressible and a penalty method is used References along with the divergence of the velocity field. This model

1. Becker, E. W., K. Bier, W. Bier, R. Schütte, and D. Seidel. has been combined with a second model that calculates Separation of the isotopes of uranium by the separa- the separative performance of a single stage assuming that tion nozzle process. 1967. Angewandte Chemie Inter- sonic velocity and equilibrium distribution of the species national Edition in English. 6 (6): 507. has been achieved. The results are summarized below.
2. Rare earth elements in national defense: Background, The key result is that the predicted separation factor and oversight issues, and options for Congress. 2012. Con- the measured separation factor are close with the mea- gressional Research Service. sured results bracketing the predicted number. This indi- 112

Page 113

3. U.S. Department of Energy Critical Materials Strategy.
4. U.S. Department of Energy. Publications Yarbro, S. L., R. M. Chamberlin, P. Martinez, L. Meyers, and Q. McCulloch. Two-dimensional model of curved flow with supercritical carbon dioxide in a diverging nozzle.
5. Los Alamos National Laboratory unclassified re- lease. 113

Page 114

Complex Natural & Engineered Systems Exploratory Research Final Report Low-Grade Thermal Energy Recovery Robert P. Currier 20150303ER Abstract Background and Research Objectives Energy-intensive industrial processes, such as electric A common cooling task in many industrial processes is power production, refineries, chemical plants, and to transfer residual energy (in the form of sensible waste smelters, routinely discard enormous quantities of heat) to the environment. Water is typically the cooling low-grade thermal energy, known as waste heat, to medium of choice because of its availability, high spe- the environment. This lowers the net energy efficiency. cific heat capacity, and large heat of vaporization. Most Waste heat rejection usually involves evaporating water industrial operations use variations of two different wet in cooling towers, which in turn results in high water cooling technologies - once through and closed-loop demand. For example, US electric power plants use (cooling tower) cooling systems. A small portion of in- approximately 196 billion gallons per day (8587 cubic dustry does use dry cooling (convective heat transfer to meters/sec). This leaves electricity security dependent air rather than evaporation). However, once-through wet on the water supply. Cost-effective techniques that can cooling is involved in delivering almost 25% of the U.S. utilize even a portion of this waste heat will improve electricity. Closed-loop wet systems consumed about efficiency, reduce water consumption, and thus improve 88% of water used in electricity generation in 2011. It is energy security. Methods proposed to date have low ef- now recognized that U.S. thermoelectric power genera- ficiency, are expensive, and/or do not scale-up favorably. tion is vulnerable to disruptions in the water supply. This project evaluated an alternative thermal energy Recovery of waste heat (even partially) would clearly be recovery concept that uses only conventional, afford- of benefit to U.S. industry. The fundamental challenge in able, and scalable processing equipment such as pumps, effective waste heat recovery is associated with the fact heat exchangers, and turbines. The following steps, that the low temperatures involved leave little ther- executed in a circulating flow system, are the essence of modynamic driving force for conversion to more useful the concept: water and gas are used to form an incom- forms of energy. pressible mixture that entraps a gas; the mixture is then One low water consumption waste heat recovery option pumped to high pressure; waste heat is used to release involves solid-state thermoelectric generators (TEGs). the entrapped gas at elevated pressure; the liberated However, a major scale difference remains between the gas is further warmed by waste heat; then the warmed Rankine-gas turbine cycles used in thermoelectric power gas is expanded to recover usable energy (electricity plants (producing hundreds of megawatts of electricity) or shaft work). Some of the recovered energy is used and TEGs that individually generate only a few watts. to power the pump. Established engineering principles Retrofitting a large coal-fired power plant with a ther- were used to elucidate key design parameters, evaluate moelectric recovery system will undoubtedly require the best working fluids using performance calculations, significant capital investment [1]. Thus no operational and identify efficiency improvement and cost-saving TEG systems have been constructed at scale. Other opportunities. Assessments of three methods based waste heat conversion options use conventional process on this concept were conducted. Crystalline clathrates, equipment (compressors, heat exchangers, expand- reversible chemical reactions, and non-volatile solvents ers). For example, closed-loop Brayton cycles based on that can absorb large amounts of gas were evaluated. supercritical CO2 have been examined at the pilot scale The solvent approach using tailored ionic liquids appears [2]. However commercial-scale implementation remains to be the most promising. It may foster the elimination a challenge. Specifically, materials that can withstand the of some process equipment, resulting in reduced capital pressures and temperatures over economically mean- costs. 114

Page 115

ingfully product lifetimes are needed. Turbo-expanders needed to compress gas to the same pressure. It is impor- unique to supercritical CO2 power cycles still face signifi- tant to recognize that any alternative cycles and working cant design challenges, e.g., high fluid density near the fluids based on trapping gas in an incompressible form critical point leads to high wheel loading while material must satisfy certain operational constraints. For example, compatibility and operating temperatures impose a signifi- the thermo-physical and phase behavior of the fluids must cant limitation on seal and bearing design. Also, the heat match the temperature range of intended operation. Other exchanger requirements may dictate high-cost designs and constraints also exist, which may be either thermodynamic manufacturing. In addition, robust process control under or kinetic. To provide some perspective on these, consider supercritical conditions will require deeper knowledge of first the heat associated with formation/dissociation of the physical parameters near the critical point, and their fluc- trapped gas (conservation of energy dictates that the heat tuations. Kalina absorption cycles, which involve a binary of formation equals exactly the heat of dissociation). Some working fluid, are also being pursued [3,4]. The Kalina cycle insight into the role this heat plays in overall conversion ef- involves a mixture of two working fluids instead of a single ficiency can be obtained by comparing a power generating fluid as traditionally employed in conventional Rankine cycle involving trapped gas with a conventional refrigera- cycles. A mixture of two fluids with different boiling points tion cycle. results in a solution that boils over a range of temperatures For the refrigeration cycle, a fluid with a high heat of as its liquid composition changes. By appropriate choice vaporization is desirable. A large heat of vaporization of the ratio between the components of the solution, the minimizes the amount of fluid that must be pumped, or boiling point of the working solution can be adjusted to circulated through the cycle (that pumping involves energy suit the heat input temperature. Water and ammonia is consumption). The power generation situation is some- the most widely used combination, but other combina- what different. Here the overall efficiency is directly related tions are feasible. The change in boiling temperature can to the fraction of total heat supplied (QT + QM) that is suc- be exploited so that a larger average temperature differ- cessfully converted to electricity by the turbine/generator ence is achieved between the heat source and the work- assembly. Since QT is ultimately dumped to the low tem- ing fluid being vaporized. This results in more heat being perature sink (i.e. to the environment), it is advantageous extracted from the source than with a single component in a power generating cycle to minimize its magnitude. working fluid. Present-day Kalina cycles have relatively low The ideal working fluid (capable of trapping gas) must also efficiency and involve numerous thermal recuperators and possess physical properties that minimize parasitic losses heat exchangers [5]. The objective of this project was to and enable affordable operation. These properties include evaluate some conceptually simple and scalable alterna- not only the heat associated with trapping a gas molecule, tives that could offer improved conversion efficiency and/ but also the viscosity and density, both of which affect the or reduced process costs. pumping costs. The heat capacity and thermal conductivity In pursuit of viable alternatives, our focus was on the use must also be considered as these affect the heat transfer of the low-grade heat to generate gas suitable for driving coefficients and thus the required heat exchange area. a turbine. The work extractable from a turbine is, thermo- Rate parameters, including diffusion coefficients, mass dynamically speaking, just the difference between the inlet transfer coefficients, and possibly even chemical reaction gas enthalpy and the gas enthalpy at the exit (enthalpy rate parameters, must also be considered in selecting and is the internal energy of the gas, which is temperature designing a system. dependent, plus the pressure times the volume). Thus to maximize power generation, one would like to have a high Scientific Approach and Accomplishments enthalpy gas fed into the turbine. This generally corre- Crystalline clathrates, chemical reactions involving a sizable sponds to achieving the highest possible gas temperature shift the number of moles of gas, and solvents that display and pressure. At the same time, parasitic energy losses ultra-high gas solubility were all evaluated as possible op- associated with generating such gas must be minimized in tions. order to achieve high net efficiencies. For example, simply compressing low-pressure recycled gas to the inlet pres- Clathrates are structured inclusion compounds, in which sure will require a compressor that will consume most of one component forms a crystalline structure (the “host”) the turbine output, resulting in a low efficiency. However, that facilitates the inclusion another component (the if working fluids could be identified that can bind (or trap) “guest”). Interaction between the guest and the host gas in an incompressible form such parasitic losses might framework is through van der Waals forces, as opposed be greatly reduced. This is because far less work is re- to chemical bonds. Gas hydrates are sub-set of the clath- quired to pressurize a dense, incompressible fluid than is rates consisting of a water-based host framework. The guests are smaller gas molecules that are captured within 115

Page 116

polyhedral cages of hydrogen bonded water molecule. pressures. Also, in the high-pressure part of the process, Calculations were performed using CH2F2 as the candidate gas solubility in the liquid mixture increases with pressure, hydrate-forming gas. It was discovered that the relatively also limiting the amount of pressurized gas available. The large heat of formation for gas hydrate slurries (50-70 kJ/ basic difficulty is thermodynamic - the need for a small mole of gas) lead to net conversion efficiencies below heat of reaction limits attainable changes in the equilib- those reported for the Kalina cycle. This is due to the need rium constant with only a modest swing in temperature. to pass the heat of formation to the ambient sink. Since In a gas-liquid reaction whose stoichiometry involves a relatively high heat of formation is seen in all the water- large number of gas molecules, the reaction will be pushed based hydrates, focus shifted to other non-water based strongly towards the liquid (product) phase at the high- inclusion compounds. This included urea, thiourea, cy- pressure portion of the cycle, contrary to the objective. clodextrins, calixarene, Dianin compounds, crown ethers, Finding purely vapor phase reactions with a large change phenols, quinols, and hydroquinols as hosts. The list of in number of moles and just the optimal thermo-chemistry candidates was narrowed by first considering their stability was deemed to be improbable. range (temperature and pressure), heat of formation, the Another property that can be exploited is ultra-high gas number of gas molecules included per mole of host, gas solubility in a non-volatile liquid solvent that is for all solubility in the host liquid, viscosity of the host liquid, and practical purposes incompressible. The gas solubility must the anticipated kinetics. Phenol-based clathrates proved be very high, must vary significantly with temperature, and the most attractive as they form clathrates within the tem- the heat of absorption must not be too large. Ionic liquids perature range of interest and the heat of formation (20-30 (ILs) were examined for this purpose as they offer numer- kJ/mole) and most other physical properties were reason- ous potential advantages: (1) they match the temperature able. However, the gas molecule loading (on the kilogram ranges of interest; (2) they can absorb large quantities gas; gas per kilogram phenol basis) proved problematic. The (3) the enthalpy of absorption is favorable (typically 10–20 viscosity together with gas molecule loading in the slurry kJ/mole); (4) most have negligible vapor pressure; and (5) translated into high pumping costs (high in-process para- the physical properties (e.g. viscosity, thermal conductivity) sitic energy losses) and low net efficiency. can be modified by tuning the molecular structures. Heat Reversible chemical reactions also provide an opportunity and material balances were performed for a power cycle to accomplish low-grade heat recovery. For an exothermic involving absorption of ammonia in an IL (1-hexyl-3-methy- reaction, lowering the temperature drives the reaction limidazolium chloride) as a test case. Despite the relatively in the forward direction thereby increasing the moles of high viscosity, which increases the energy required for products at equilibrium. Likewise, raising the temperature pumping, the conversion efficiencies were comparable to drives the equilibrium back towards the reactants, increas- a Kalina cycle. However due to the low vapor pressure of ing their quantity. A properly chosen chemical reaction the ILs, other advantages were evident. Specifically, very would lower the number of moles of compressible gas at low pressures can be set on the downstream side of the the cool portion of the cycle and release gas upon heat- turbines. In comparison, downstream pressures must be ing. A thermo-chemical analysis was conducted to identify limited in a Kalina cycle to avoid condensation of the water attributes of the ideal reaction - gas molecules react with that is mixed with the ammonia gas. The heat remaining one liquid (the reactants) to form another incompressible in the water-ammonia mixture is partially recovered using liquid (the product). Following compression and warming, a costly rectifier and/or recuperator units in a Kalina cycle. the reaction equilibrium ideally shifts back towards reac- It was shown that use of an ammonia-ionic liquid mixture tants, liberating high-pressure gas. A large change in the as a power cycle working fluid has a conversion efficiency number of moles is sought, along with a small heat of reac- equivalent to a Kalina cycle. However, the ammonia- tion. A heat of reaction too large will again lower the net ionic liquid mixture offers practical and capital cost saving conversion efficiency since that heat is ultimately dumped advantages: 1) the cycle does not require a complicated to the ambient sink in the cool part of the cycle. Once the rectifying column to purify the ammonia, 2) turbine outlet ideal system was identified, synthesis of the necessary temperatures can be significantly lower in the ammonia- ingredients could follow. However, the analysis showed ionic liquid cycle compared to the Kalina cycle, and 3) the the small heat of reaction adversely affected the ability to ammonia liquid system does not require a large recupera- liberate gas upon modest warming. A small heat of reac- tor heat exchanger. Use of non-volatile ILs removes the tion corresponds to only a small change in the equilibrium need for these pieces of equipment for heat recovery thus constant. Attempts to liberate gas by warming increases reducing the capital investment. Since IL-based absorption the concentration of reactants, which pushes the reac- proved the most attractive option examined, a plan for tion equilibrium back towards the liquid product at high further development and testing was developed. The next 116

Page 117

step in process development should be design, construc- tion, and operation of an engineering test module demon- strating a fully integrated power production cycle. Impact on National Missions This work directly supports Energy Security by increasing the net efficiency of conventional electric power produc- tion. Also, reducing water consumption in such processes translates into more secure U.S. energy (less vulnerability to diminishing fresh water supplies). While electric power plants are the initial focus, similar waste heat streams ex- ist in energy-intensive industries such as cement, metals smelting, refining, chemicals, and steel production. References

1. Yazawa, K.. Optimization of thermoelectric topping combined steam turbine cycles for energy economy. Applied Energy . 109: 1.
2. Turchi, C.. 10 MW supercritical CO2 turbine. 2013. US Department of Energy.
3. Lolos, P., and E. Rogdakis. A Kalina power cycle driven by renewable energy sources. 2009. Energy. 34: 457.
4. Ibrahim, M. B., and R. M. Kovach. A Kalina cycle appli- cation for power generation. 1993. Energy. 18: 961.
5. Nasruddin, A., R. Usvika, M. Rifaldi, and A. Noor. En- ergy and exergy analysis of kalina cycle system (KCS) 34 with mass fraction ammonia-water mixture variation.
6. Journal of Mechanical Science and Technology. 23: 1871. 117

Page 118

Complex Natural & Engineered Systems Exploratory Research Final Report Building a Foundation for Understanding How Pathogens Subvert the Host Immune System Thomas C. Terwilliger 20150715ER Abstract interactions responsible for pathogen evasion and to This research brings together a team from three LANL target these interactions for developing new drugs. We Divisions to integrate LANL capabilities to understand will form a multi-disciplinary team to identify and model how pathogens subvert host immune defenses. Our different stages of autophagy and how they are inhib- short-term goal was to bring together key LANL teams ited by intracellular pathogens. This will be supported and develop a foundation for our long-term plan. Our by experimental measurements of transcriptome and demonstration system will be influenza virus escaping proteome levels in infected hosts corresponding to dif- autophagy. In this LDRD-Directors’ Reserve project we ferent stages of autophagy, and by structural analysis of developed a foundation for detailed modeling of how specific host-pathogen protein-protein complexes critical influenza interferes with autophagy and for making to pathogen evasion. experimental measurements to evaluate those models. Our short-term goal is to bring together key LANL teams Further we generated antibodies to influenza proteins and develop a foundation for our longterm plan. Our that we will be able to use to perturb the system and to demonstration system will be influenza virus escaping monitor the interactions between influenza proteins and autophagy. We will develop a foundation for detailed host proteins. modeling of how influenza interferes with autophagy and for making experimental measurements to evaluate Background and Research Objectives those models. Further we will generate antibodies to One of the key strategies used by the human immune influenza proteins that we will be able to use to perturb system to eliminate pathogens is to engulf and destroy the system and to monitor the interactions between the pathogen, a process termed “autophagy”. In re- influenza proteins and host proteins. sponse, many pathogens have developed counter-strat- egies in which they turn off the destruction step in au- Scientific Approach and Accomplishments tophagy, and some even co-opt the autophagy system to Development of models describing influenza interfer- provide themselves with a safe haven to reproduce. The ence with autophagy normal autophagy process occurs in several steps. The As proposed, we have extended our models of gene/ first is initiation of an intracellular vesicle that engulfs protein networks in autophagy to include the effects of the pathogen. This is followed by expansion, closure, and influenza proteins. These models will now allow us to maturation of the vesicle, and finally by digestion of the make predictions about the effects of interfering with pathogen inside the vesicle. Many of the host proteins specific steps in the autophagy process or in the interfer- involved in autophagy are known, and some pathogen ence of autophagy by influenza proteins. The extended proteins responsible for inhibiting autophagy have been models include interactions that have been identified in identified. However, neither the exact gene/protein the literature such as influenza proteins M2, H1, and NS1 networks that are induced during the various stages of and their interactions with host proteins LC3-II, Atg5, autophagy nor the specific mechanism of interference by Atg12, Atg16L, Beclin-1, Vps4, and p150. As described intracellular pathogen proteins is known. below, this theoretical framework is already a powerful Our long-term goal to develop a quantitative under- tool for integrating our experimental results with our standing of how an intracellular pathogen escapes and understanding of the system. takes advantage of host autophagy. Success in achieving Development of experimental approaches for testing our this goal will position us to identify critical host-pathogen 118

Page 119

models of autophagy and interference by influenza We carried out exploratory genome-wide transcriptome and proteome analyses on influenza-infected and unin- fected epithelial cells. These showed that we could identify many human genes and their expression levels over time after infection, an important prerequisite for further work on this project. We additionally carried out experiments to visualize autophagosomes inside uninfected and infected cells using stable epithelial cell lines expressing green fluorescent protein fused to host protein LC3. An exciting outcome of this part of the work is that our theoretical work (prior to this project) had suggested there may be situations where the levels of autophagy proteins might oscillate between low and high levels. Our experimental work has indicated the presence of such oscillations, sug- gesting that the theory may indeed be on the right track. Impact on National Missions The scientific impact of our current and future work on this project will be a quantitative understanding of how an intracellular pathogen evades autophagy. The project will provide crucial clues for identification of new therapeutic targets and discovery of new drugs. Our comprehensive approach to the problem will enhance our understand- ing of human proteomics (using antibody technology) and protein-protein interactions (using split-green fluorescent protein technology). Our approach will yield both new ap- proaches for network modeling and new network models, and it will identify the structural biology of key protein- protein interactions. 119

Page 120

Complex Natural & Engineered Systems Postdoctoral Research and Development Continuing Project Ultrafast Vacuum Ultraviolet Spectroscopy of Complex Materials Dmitry A. Yarotski 20130814PRD4 Introduction work is poised to make a broad impact on condensed Exotic properties of strongly correlated electron materi- matter physics and will open new directions in complex als, such as high-temperature superconductivity and materials research. multiferroicity, have inspired an extensive research effort to harness these properties for technological applica- Benefit to National Security Missions tions. However, the labyrinthine pattern of compet- Our work directly addresses the Grand Scientific Chal- ing interactions between charge, spin and lattice has lenges identified in the Basic Energy Sciences Advisory prevented the development of predictive theoretical Committee report, which are central to DOE’s missions in frameworks and establishment of basic material design energy, science, and security in general, and to the LANL principles. In general, it is a rather elaborate task for Nuclear Security and National Defense mission and the regular time-averaging probes to decipher the nature of Materials Pillar in particular. Our thrust to couple model- the quasiparticles and the interactions involved in collec- ing, complex materials synthesis, and ultrafast optical/X- tive material behavior because they give little informa- Ray science aligns well with MaRIE vision of “material co- tion on the specific degrees of freedom and underlying design” and further develops new capabilities for MaRIE microscopic physics. In contrast, ultrafast optical spec- science. During this project, we anticipate to access large troscopy offers unsurpassed ability to discriminate the BES X-Ray facilities at LCLS and APS in accord with institu- coupling between charge, spin, and lattice in the time tional priority in supporting national user facilities. The domain because their dynamics differ by several orders proposed experiments are critical to achieving design of magnitude (e.g., tens of picoseconds for phonons vs and synthesis of new materials with controlled func- hundreds of ps for spins). tionalities. Materials with tunable and novel functional- ity are an enabling component in the development of In this research, we will apply novel ultrafast vacuum next-generation devices for sensing, information storage, ultraviolet photon sources to perform time-resolved and spintronics applications. We believe that our inte- X-Ray magnetic linear (and circular) absorption dichro- grated capabilities in complex material design, synthesis ism spectroscopy of a number of strongly correlated and characterization will be of great interest to multiple electron materials and heterostructures where spin sponsors, including DOE-BES, DOD, IC, and industry. interactions with other degrees of freedom determine material functionality. These techniques will provide new Progress insight into the spin dynamics in these materials with The goal of this project is to probe ultrafast spin dynam- unprecedented time resolution, element specificity, and ics in strongly correlated materials using time-resolved sensitivity to antiferromagnetic ordering and dynamics X-ray magnetic linear absorption dichroism (XMLD) spec- near buried interfaces. The results of our advanced char- troscopy. The XMLD technique probes p->d transitions acterization effort will directly feedback to the theoreti- of magnetic ions and is valuable tool for understand- cal effort of our colleagues from Theoretical Division. We ing how the spin order is coupled to other degrees of strongly believe that such a combined ultrafast theory freedom (charge, orbital, lattice), because electrons on and characterization approach aimed at resolving spin, d-orbitals of transition metal ions are usually responsible charge and lattice interactions, as well as the associated for magnetism in complex materials. In particular, our quasiparticle dynamics, will ultimately result in a predic- recently developed ultrafast table top soft x-ray source is tive framework for understanding emergent phenomena capable of spanning the 3p-3d transitions (M-edges) of in strongly correlated electron materials. As such, this 120

Page 121

many transition metal components in strongly correlated creased through the AFM transition, the amplitude of the electron compounds. The dynamics of spin degrees of ultrafast transient decreases and oscillations (~100 GHz) freedom is studied by photo-exciting (pumping) a sample develop which correspond to coherent excitations in the with a femtosecond near-infrared laser pulse which takes material. The oscillation strength increases with increasing the material out of equilibrium and perturbs the spin align- temperature. This behavior clearly indicates that magnetic ment through transient modulation of charge density or order suppresses coherent excitation of charge density lattice distortions affecting exchange interactions. Subse- waves in Cr which implies strong coupling between these quently, a soft x-ray pulse probes the dynamics by monitor- two orders. We are working with theorists in both T-4 ing absorption dichroism (polarization dependence) of the and Uppsala University to understand the origins of such magnetic ion of interest. multiple-order coupling and its implications on the mate- rial functionality. The results of these experiments were In FY15, we focused on investigations of the mechanisms presented at the 2015 APS March Meeting, and we are of complex spin-charge interplay in chromium (Cr). Cr is a currently working on the manuscript that will summarize peculiar material which undergoes several phase transi- our findings and describe physical mechanisms responsible tions involving antiferromagnetic (AFM) spin alignment in for the observed effects. spin density waves and formation of charge density waves. Prior to obtaining a novel insight into the Cr non-equilibri- Future Work um properties, we had to overcome several technical chal- Exotic properties of complex materials emerge from the lenges related to unusually high surface sensitivity of M- strong interactions among spin, charge, and lattice degrees edge XMLD measurements. First, we discovered that even of freedom (DOF). Existing theories cannot predict the under high vacuum and the cryogenic temperatures, the emergent behavior, largely because available time-aver- thin water layer can contaminate the sample surface and aging experimental probes cannot decipher the interac- absorb large portion of extreme ultraviolet (XUV) probe tions between particular DOFs. In this regard, ultrafast pulse. This problem has been mitigated by introducing a optical spectroscopy (UOS) offers an unmatched ability to liquid nitrogen cooled shield around the sample that freez- separate various DOF and coupling among them in time es out any water remaining in the vacuum chamber before domain. Our goal is to apply advanced UOS techniques the sample is cooled. A second obstacle was a high sensi- to study time-dependent and element-specific processes tivity of XMLD signal to the sample roughness. By closely that are important for understanding the DOF interplay in monitoring and controlling the position of the beam on the complex materials. sample surface we could consistently measure the same region of the sample thus reducing signal variations cause In FY15, we applied our optimized table-top ultrafast X-Ray by inhomogeneous surface profile. The new experimental linear absorption dichroism probe to directly reveal the dy- system configuration now allows continuous measure- namics of spin and charge orders in chromium. In FY16 we ments of magnetic dynamics in a broad temperature range will complete the experimental data analysis and develop (5-450 K) in samples that move due to temperature depen- theoretical understanding of the strong charge-spin cou- dent cryostat arm expansion/contraction. pling and its signatures in ultrafast X-Ray dynamics. These results will be summarized in the manuscript and submit- Subsequent measurements of temperature-dependent ted for publication in peer-reviewed journal. We strongly XMLD dynamics in Cr crystals showed an increase in XMLD believe that these results will provide novel insight into signal at ~310K consistent with the onset of AFM spin element-specific magnetic dynamics in complex materi- ordering in this material. Interestingly, another transition als. In the FY16, we will also continue the ultrafast XMLD at ~120K was also observed, which corresponds to the experiments on multiferroic materials where magnetic spin flip from transverse to longitudinal spin density wave and ferroelectric polarizations are strongly coupled and order. Both of these observations clearly demonstrate high allow manipulation of electric polarization with magnetic sensitivity of our table-top XMLD method to AFM spin field and magnetization with electric field. Our focus will order variations which were only observed before at large be single-phase material BiFeO3 where optical pulses will synchrotron facilities. Moreover, our ability to resolve sub- be used to disturb the magnetic ordering and X-Ray pulses 50 fs transients in photoinduced XUV absorption allowed will be applied to probe the ensuing modulations in strain, us to follow the dynamics of spin and charge degrees charge or orbital orders. These experiments will shed a of freedom across the AFM phase transition. Below the new light on the coupling between charge/lattice and spin transition temperature, XMLD signals show a very fast (~1 systems responsible for emergence of magnetoelectric ps) amplitude change that relaxes on time scales of >10 coupling in multiferroics. In addition, we will start extend- ps following photoexcitation. As the temperature is in- ing our ultrafast measurements to ferromagnetic systems 121

Page 122

using the transverse magneto-optical Kerr effect or X-Ray magnetic circular absorption dichroism techniques. Conclusion We expect that experimental and theoretical insights developed here will lead to a leap in our comprehension of the competing interactions in complex materials. Our novel ultrafast coherent photon probes will allow direct visualiza- tion of the real-time formation of competing orders, and will ultimately enable manipulation and coherent control of materials functionality. Our approach will fill the ur- gent need in integrated modeling, synthesis, and ultrafast nanoscale characterization capabilities essential to address several challenges in materials science. The understanding of complex materials properties developed here will have a significant impact on wide ranging technologies from information storage to energy applications. Publications Ahmed, T., B. McFarland, A. Chen, Q. X. Jia, D. A. Yarotski, and J. X. Zhu. Site dislocation between Fe and Mn sites in double perovskite Bi2FeMnO6: A theoretical explanation for XMCD anomaly. Physical Review Letters. McFarland, B. K., J. X. Zhu , R. P. Prasankumar, G. Rodriguez, R. L. Sandberg, A. J. Taylor, S. A. Trugman, and D. A. Yarotski. Ultrafast X-Ray Probe of Dynamics in Chromium. Presented at 2015 CINT User Meeting. (Santa Fe, September 21-22, 2015). McFarland, B. K., J. X. Zhu, R. P. Prasankumar, G. Rodriguez, R. L. Sanberg, A. J. Taylor, S. A. Trugman, and D. A. Yarotski. Soft X-Ray Probe of Ultrafast Dynamics in Spin Ordered Chromium. Physical Review Letters. McFarland, B., J. X. Zhu, R. Prasankumar, G. Rodriguez, R. Sandberg, A. J. Taylor, and D. A. Yarotski. Ultrafast Dynamics near the M-edge in Chromium. Presented at American Physical Society March Meeting. (San Antonio, Texas, March 2–6, 2015). 122

Page 123

Complex Natural & Engineered Systems Postdoctoral Research and Development Continuing Project Bayesian Information Gap Decision Analysis Velimir V. Vesselinov 20140000PRD4 Introduction across the USA. Applications to other fields are easily Bayes theorem is a mathematical technique that pro- achieved due to the model independence. vides a way to determine the likelihood of different models given some observed data. Bayes theorem Benefit to National Security Missions is mathematically rigorous, but its application is not This research aligns with the LANL mission and the always rigorous for two reasons: (1) We can enumer- strategic EES research goals as it aims to provide optimal ate the possible outcomes of dice-rolling, but not the decisions for national problems related to energy and possible outcomes of groundwater remediation. (2) We environmental security; these national problems require can determine conditional probabilities for coin-tossing, development of strategic scientific capabilities to im- but substantial uncertainty surrounds the conditional prove decisions in energy and water production, carbon probabilities in groundwater remediation. We circum- capture and sequestration, waste disposition and stor- vent these issues with a three-layered-onion approach. age, and restoration of contaminated sites. In all these The middle layer employs Bayes theorem. The inner areas, real-world applications require robust and accu- and outer layers employ a non-probabilistic uncertainty rate methods for capturing the existing uncertainties in quantification (UQ) methodology called Information-Gap the model analyses and decision process. The methods to overcome (1) and (2), respectively. This Bayesian- and tools developed in this work will satisfy these needs Information-Gap (BIG) UQ approach provides robust and strengthen existing research at LANL. This work will decision support for selecting a remediation approach provide opportunity for future basic science and pro- after the site characterization been performed. grammatic research growth working with new sponsors (e.g., DoD, SERDP, NSF, industry) and new research areas This project will extend this methodology to provide that require decision analyses for complex national secu- decision support throughout the site characterization rity problems. process. A probabilistic model will be used to generate hypothetical data that might be obtained from a site Progress characterization effort. The BIG UQ framework will be The fundamentals of the Bayesian-Information-Gap used to arrive at a decision before and after hypothetical decision theory (BIG-DT) described in a published paper: site characterization. Characterization efforts that have O’Malley, D., Vesselinov, V.V., A combined probabilistic/ a high chance of altering the decision are more worth- non-probabilistic decision analysis for contaminant re- while than those that have a low chance of altering the mediation, Journal on Uncertainty Quantification, SIAM/ decision. This also informs the transition from uncer- ASA, doi: 10.1137/140965132, 2014. The paper summa- tainty reduction to remedy selection. If all site character- rizes the theory and applies the methodology to solve a ization options are unlikely to alter the remedy selection, synthetic contaminant remediation problem. it is time to suspend site characterization and begin remediation. Techniques for estimating the progress and Bayesian-Information-Gap decision theory applied to execution time of the analyses will also be developed. solve CO2 sequestration problem. A paper submitted: O’Malley, D., Vesselinov, V.V., Bayesian-Information-Gap The combination of flexible computational demands that decision theory with an application to CO2 sequestra- arises from model independence and open source soft- tion, Water Resources Research, submitted, 2015. ware means that this framework can also be used out at tens of thousands of contaminated groundwater sites 123

Page 124

An initial version of a code in Julia for Bayesian-Informa- for Identifying Diffusive Trajectories with Stochastic tion-Gap analysis developed. The code has been tested Models. 2014. JOURNAL OF STATISTICAL PHYSICS. 156 on 64 processor computational clusters. It needs to be (5): 896. extended to be executable on the LANL HPC distributed O’Malley, , and V. V. Vesselinov. Groundwater remediation clusters. using the information gap decision theory. 2014. WATER RESOURCES RESEARCH. 50 (1): 246. An initial version of a code in Julia for Bayesian-Informa- tion-Gap Optimal Experimental Design (BIG-OED) devel- O’Malley, , and V. V. Vesselinov. Analytical solutions for oped. The code has been tested on 64 processor compu- anomalous dispersion transport. 2014. ADVANCES IN tational clusters. It needs to be extended to be executable WATER RESOURCES. 68: 13. on the LANL HPC distributed clusters. The code has applied to solve a synthetic test problem. The initial results are encouraging. Future Work • Develop a probabilistic model that describes the pos- sible outcomes of site characterization efforts. Site characterization efforts that should be considered include pumping tests and groundwater sampling campaigns. • Develop a methodology that incorporates the site characterization model into the Bayesian-Information- Gap (BIG) uncertainty quantification (UQ) framework. • Demonstrate the methodology on a synthetic test problem. • Develop an approach for estimating progress and ex- ecution time of these analyses. • Write a paper describing the methodology. • Demonstrate the methodology on a real-world prob- lem. Most probably the real-world problem will be related to the LANL Chromium contamination site. Conclusion This project will provide a framework for making decisions related to complex problems with a focus on groundwater remediation. Groundwater contamination is an important national problem with over $100billiondollarsinliabilitiesincludingaround$20 billion within the DOE complex. A substantial challenge in groundwater remediation sce- narios is to select between a variety of site characteriza- tion efforts and remediation options. This framework will provide insight into the value of different characterization efforts, inform the transition from studying the problem to solving the problem, and help select the best solution to the problem. Publications O’Malley, , V. V. Vesselinov, and J. H. Cushman. Diffusive mixing and Tsallis entropy. 2015. PHYSICAL REVIEW E. 91 (4). O’Malley, , V. V. Vesselinov, and J. H. Cushman. A Method 124

Page 125

Complex Natural & Engineered Systems Postdoctoral Research and Development Continuing Project Tracking Microbial Activity to Predict the Impacts of Climate Change on Ecosystem Function Cheryl R. Kuske 20140662PRD1 Introduction tion gap in our knowledge of and ability to predict the Ecosystem processes such as nutrient cycling sustain life outcomes of processes that are mediated by complex on earth, and are controlled in large part by communi- microbial communities in soils. Soils are an enormous ties of fungi and bacteria (collectively termed microbes) reservoir of carbon (C) on Earth, and the concerted that reside in the soil. Global changes, such as nitrogen activities of soil microbes regulate the storage or release deposition, warming, elevated CO2, and altered precipi- of this C. Second, the project directly contributes to tation, alter the abundance and biodiversity of microbial DOE missions in biological systems science and climate communities and significantly alter ecosystem functions. change responses. It also advances metagenomics tech- Soil microbial communities largely determine whether nologies and basic environmental surveys that enable terrestrial ecosystems act as carbon sources or sinks more accurate, specific detection of target microorgan- under elevated CO2. However, biochemical processes isms in environmental samples, thus contributing to a mediated by complex microbial interactions in the soil technology base supporting biothreat detection. Third, remain a black box in terrestrial ecosystem models. A this project advances our understanding of fungal and major hurdle in linking complex soil communities to bacterial biomass and metabolic capabilities; these are major ecosystem processes is the ability to distinguish organisms from which we have historically derived most among the active and dormant members of the commu- of our antibiotics, as well as other pharmaceuticals and nity. Understanding how dormancy structures microbial industrial enzymes. Fourth, many of the fungi we will be communities is of particular concern because it is a studying also have potential in processing C for biofuels strategy for survival under unfavorable environmental applications. conditions, and could be a pathway for resiliency of microbial communities to climate changes. I will use a Progress combination of high-throughput DNA and RNA sequenc- A soil microcosm experiment that compares microbial ing with surveys of functional genes in two different community responses to increased moisture, increased field-collected soils in the laboratory to determine the nitrogen, and warming temperatures is complete. Mea- relative contributions of the active populations, and to sures of soil respiration have been collected over the elucidate the biological and ecological constraints on the time-course experiment. Extraction of total soil RNA and ability of microbial populations to enter and exit from DNA have been completed, and sequencing of fungal dormancy. Collectively, these studies will help us under- and bacterial communities from both RNA and DNA has stand how the microbial community regulates climate been initiated. In addition, using funds received through change responses and feedbacks in the soil. We have the LDRD program, we generated 24 metatranscrip- completed two soil microcosm experiments: one that tomes from a subset of the treatments, and analysis of compares microbial community responses to increased community shifts based on mRNA is currently underway. moisture, increased nitrogen, and warming tempera- tures, and another that examines microbial response to In addition to manipulating soils with the above addi- environmental stress, including temperature, desiccation tions, we used three different stresses in order to de- and starvation. termine which community members are able to survive harsh environmental conditions, including heating to Benefit to National Security Missions 65C, freezing at -80C and long-term desiccation. Mea- First, this project will help fill a significant informa- surements of soil respiration and community composi- 125

Page 126

tion are planned for the next two months. In addition, we nity to measures of soil respiration, as well as to quantify are in the process of constructing contrived communities shifts in the active community over time. These measures of known composition to link known functional genes to will allow us to determine the short-term community ribosomal genes to gain a greater understanding of using responses to predicted global changes, and to link those rRNA as an indicator of activity. Co-extraction of RNA and shifts with ecosystem function. DNA are complete, and sequencing of fungal and bacterial communities is ongoing. Conclusion The goals of this project are to understand the relative We expect to finish the community analysis of fungal and contributions of active soil bacterial and fungal popula- bacterial communities generated by high throughput tions, and constraints on their ability to enter and exit amplicon sequencing within the next two months. The from dormancy, to enable accurate predictions of C or N timeline to complete the statistical analysis of communi- fluxes from the soil. By quantifying microbial activity and ties from the two experiments is estimated at an additional dormancy under dynamic environmental conditions, my four months. proposed studies will determine how the microbial com- munity regulates climate change responses and feedbacks In terms of publications, we expect to generate two from in the soil, which are critical for development of soil pro- the initial experiments, one linking soil respiration with cess models and to define appropriate input variables in community composition of fungal and bacterial communi- regional climate models. ties, and another examining the dynamic rank abundance curves of active and dormant fungal and bacterial com- Publications munities, which will also incorporate the findings from Dunbar, , L. V. Gallegos-Graves, Steven, Mueller, Hesse, D. the metatranscriptome survey. We expect to generate an R. Zak, and C. R. Kuske. Surface soil fungal and bacterial additional manuscript contrasting stress responses in fungi communities in aspen stands are resilient to eleven and bacteria within the two different soil types to further years of elevated CO2 and O-3. 2014. SOIL BIOLOGY & explore the relative effects of anthropogenic perturbations BIOCHEMISTRY. 76: 227. on community shifts and ecosystem functions. Kuske, C. R., C. N. Hesse, J. F. Challacombe, Cullen, J. R. Herr, R. C. Mueller, Tsang, and Vilgalys. Prospects and Future Work challenges for fungal metatranscriptomics of complex Objective: To investigate how dormancy determines com- communities. 2015. FUNGAL ECOLOGY. 14: 133. munity response to global change and the implications of dormancy on soil processes Mueller, R. C., Belnap, and C. R. Kuske. Soil bacterial and fungal community responses to nitrogen addition Rationale: The major goal of the SFA is to understand how across soil depth and microhabitat in an arid shrubland. 2015. FRONTIERS IN MICROBIOLOGY. 6. global change, primarily nitrogen deposition, will affect microbial communities, and in turn, important ecosystem Mueller, R. C., L. Gallegos-Graves, D. R. Zak, and C. R. functions, such as carbon cycling. Understanding how Kuske. Assembly of active bacterial and fungal dormancy structures microbial communities is important communities along a natural environmental gradient. for understanding both the long and short-term impacts 2015. Microbial Ecology. : 1. of global change on microbial biodiversity because it is a strategy for survival under unfavorable environmental con- Mueller, R. C., M. M. Balasch, and C. R. Kuske. Contrasting soil fungal community responses to experimental ditions, and is a potential pathway for resiliency of micro- nitrogen addition using the large subunit rRNA bial communities to global change. taxonomic marker and cellobiohydrolase I functional marker. 2014. MOLECULAR ECOLOGY. 23 (17): 4406. Approach: We are analyzing controlled laboratory ap- proaches to study the response of natural microbial com- munities in two contrasting soils to a multiple environmen- tal manipulations, including increased water availability, increased temperature and increased nitrogen availability. The responses to these perturbations were tracked over time by quantifying soil respiration using gas chromatogra- phy. The active and total communities are currently being measured using molecular analysis of rRNA and rDNA in order to both link the composition of the active commu- 126

Page 127

Complex Natural & Engineered Systems Postdoctoral Research and Development Continuing Project Complexes Containing Redox-Active Ligands for the Synthesis of Fuels from Readily-Available Carbon Sources John C. Gordon 20140664PRD2 Introduction Future Work The proposed work will investigate the development of In FY16 we will work on the synthesis and characteriza- catalytic systems that convert readily available, highly tion of quinone based systems that can support two oxidized precursors such as carbon dioxide, acetic acid, electron redox chemistries. The spectral and structural and glycerol into higher-order compounds containing characteristics of these molecules will be used in con- more carbon atoms that can readily be converted into junction with electronic structure calculations to evalu- alkanes, which are the main components of transporta- ate their potential in promoting chemistries of interest. tion fuels. This work addresses both the synthesis of the We will also use electrochemistry to guide subsequent higher order fuel precursors and their defunctionaliza- synthetic efforts towards catalytic redox chemistries tion to produce hydrocarbons that would be suitable as and this will also be useful in identifying stoichiometric fuels. This has the potential to convert biorenewable and reaction conditions that facilitate carbon-carbon bond abundant molecules into molecules with potential use as coupling reactions and the reduction of highly oxygen- fuels. ated substrates. Benefit to National Security Missions Conclusion The goal of the project is the synthesis of molecules that The goal of the project is the synthesis of molecules can serve as fuels from readily available and sustain- suitable as fuels from readily available and sustainable able molecules such as CO2, acetic acid, and glycerol. molecules such as carbon dioxide (CO2), acetic acid, and These goals are directly in line with DOE BES’s interests glycerol. This has the potential to convert biorenewable in fundamental chemistry and catalysis science and DOE and abundant molecules into molecules with potential EERE’s interests in sustainable and biorenewable chem- use as hydrocarbon fuels. istry. This work also directly impacts the area of climate and energy impact vis-a-vis the conversion of biorenew- able and abundant small molecules into higher order carbon containing species with potential use as hydro- carbon fuels for transportation applications. Progress The synthesis of a significant number of metal com- plexes supported by redox active ligands, including those containing iron and nickel have been achieved. In FY15 these were screened towards a number of reac- tions, including carbon-carbon bond coupling reactions, hydroboration, etc. In particular, we prepared a nickel based system that appears to exhibit interesting and unexpected redox behavior upon binding a ketonic (C=O bond containing) substrate that may have implication for how redox active metal complexes behave. We are cur- rently working on a manuscript that describes this work. 127

Page 128

Complex Natural & Engineered Systems Postdoctoral Research and Development Continuing Project Bottom-up Chemical Synthesis of Large, Well-Defined, and Organo-Processable Nanographene-based Triarylamine for Optoelectronic Applications Hung-Ju Yen 20140666PRD2 Introduction strengthen our leadership role in NG research, which has The main focus of this project is to develop synthetic strong ties to Laboratory mission in the areas of exotic strategies for making triarylamine-based nano-graphene materials and energy security. (TAA-NG), a small piece of graphene with well-defined molecular structure. NG with flexible alkyl chains at- Progress tached to the periphery of NG renders solubility in The main accomplishments for FY15 were the synthesis organic solvents and allows self-assembly of NGs with of heteroatom (N, or B) doped nano-graphenes (NGs) phase separation into nanoscaled morphology in com- through a newly designed synthetic route. Such synthe- posite materials. More importantly, the incorporation sis includes Curtis rearrangement, Sonogashira coupling, of heteroatoms such as triarylamine (TAA) into the Suzuki coupling, Diels-Alder cycloaddition, and Ullmann graphene frameworks leads to a triangular shape and reactions. Nanographenes containing various functional charge transfer characteristics, which, to the best of our gropus such as hydroxy, methoxy, tert-butyl, fluorine, knowledge, have never been attempted. The challenge and bromine. These molecules have been thoroughly of this project lies in the fact that the organic synthetic characterized using a series of organic spectroscopy scheme for TAA-NG requires more than 20 steps. This is probes such as proton and C13 NMR, Infrared, matrix a costly and time-consuming process. Although synthe- assisted laser desorption ionization spectroscopy. All of sis of simple NGs have been demonstrated, synthesis of the spectroscopy results show perfect agreement with TAA-NG remains a challenge and the final yield of this the proposed molecular structure. compound may be low, which could impede us from developing applications using this materials. Therefore, In addition, we have found these molecules form hier- our challenges are not only in the synthesis of TAA-NG, archical structure through self-assembly. Such self-as- but also in finding ways to increase the yield of the final semblies have a d-spacing between nanographene that product. is functional group dependent. The average d-spacing of the NGs in the domains is around 3.95-4.36 Å, which We expect the as-prepared TAA-NG will reveal fine- is much larger than that of pristine graphite (3.34-3.37 tuned optoelectronic properties tailored for the applica- Å). NGs with hydrogen, methoxyl, tert-butyl, hydroxy, tions in electrochromic, memory device, LED, photovol- bromo, and fluoro functional groups attached to the NG taic cell, fuel cell, lithium battery, and super-capacitor. flakes have corresponding d-spacing of 3.95, 4.08, 4.36, Success in demonstrating derivatized NGs in the above 4.04, 4.30, and 4.0 Å respectively. It is notable that the d- devices could have intellectual property value and war- spacing between flakes is consistent with the size of the rants publications in high impact journals. functional groups (OMe-tBu > Br > OMe > OH > F > H. XRD of the NGs and is consistent with the TEM measured Benefit to National Security Missions d-spacing, although they all reveal overall amorphous This project aims to develop nanographenes (NGs) with structure, lacking of a sharp diffraction peak. We use the tunable electronic structures and optical properties. center of the amorphous peak as a rough estimate of the The integration of functional NGs into clean energy representative d-spacing, the t-BuOMe HBC appears to technologies could bridge the gap between basic re- has two peaks center at ~ degree, the low angle peak is search and commercialization of graphene-based energy corresponding to 4.29 A, consistent with the estimated devices. Further development of NG-based materials will 4.3 A. In addition, these is a trend of XRD peak shifting 128

Page 129

to a higher angles from t-BuOMe-, OMe-, H-, Br- F- and Responsive, Fluorescent Amino Acids. 2015. CHEMICAL OH-, consistent with the our hypothesis of using functional SCIENCE. 6: 1150. groups to control d-spacing and electron density which Kuo, C. Y., W. Nie, H. Tsai, H. J. Yen, A. D. Mohite, G. Gupta, allow us to ascertain and optimize Li incorporation and A. M. Dattelbaum, D. J. William, K. C. Cha, Y. Yang, L. diffusion. Wang, and H. L. Wang. Structural Design of Benzo[1,2- b:4,5-b ‘]dithiophene-Based 2D Conjugated Polymers Future Work with Bithienyl and Terthienyl Substituents toward In FY16, we will focus on the synthesis of heteroatom (N, Photovoltaic Applications. 2014. MACROMOLECULES. or B) doped nano-graphenes through a newly designed 47 (3): 1008. synthetic route. Such synthesis includes Curtis rearrange- Park, Y. I., O. Postupna, A. Zhugayevych, S. W. Kyu, , Y. ment, Sonogashira coupling, Suzuki coupling, Diels-Alder S. Park, B. Kim, H. J. Yen, P. Cheruku, J. S. Martinez, J. cycloaddition, and Ullmann reactions. The molecular Park, S. Tretiak, and H. L. Wang. A new pH sensitive structures and crystallinity of these nitrogen/boron doped fluorescent and white emissive material through nanographenes will be characterized by X-ray diffrac- controlled inter-molecular charge transfer. 2015. tion, scanning electron microscopy, scanning tunneling CHEMICAL SCIENCE. 6: 789. microscopy, atomic force microscopy, MALDI-TOF Mass Yen, H. J., H. Tsai, C. Y. Kuo, W. Nie, A. D. Mohite, G. Gupta, spectroscopy, infrared spectroscopy and nuclear magnetic J. Wang, J. H. Wu, G. S. Liou, and H. L. Wang. Flexible resonance spectroscopy. Electronic and optical properties memory devices with tunable electrical bistability of these nitrogen doped nanographene will be determined via controlled energetics in donor-donor and donor- by UV-Vis spectrometer and fluorescence spectrometer. acceptor conjugated polymers. 2014. JOURNAL OF We will measure the redox potentials of these nanogra- MATERIALS CHEMISTRY C. 2 (22): 4374. phenes by cyclic voltametry, coupled with the absorbance spectra to construct the band diagram of nanographenes. Yen, H. J., H. Tsai, M. Zhou, A. Chen, E. F. Holby, S. The incorporation of nitrogen/boron in the nanographene Choudhury, X. Wang, L. Zhu, H. Lin, L. Dai, G. Wu, is expected to increase the incorporation of metal atoms and H. L. Wang. Bottom-up Chemical Synthesis of Three-Dimensional Nanographene Assemblies with and rate of diffusion. Therefore, we will incorporate these Controlled d-Spacing and Enhanced Electron Density nitrogen-doped nanographenes as part of anode material for Efficient Li Storage. Journal. in lithium ion battery (LIB) and evaluate the performance characteristics in terms of capacity, stability through hun- Yen, H. J., P. W. Liang, C. C. Chueh, Z. Yang, H. L. Wang, and dreds of charging-recharging cycles. We will also evaluate A. K.-Y. Jen. Single Crystal Perovskite Powders for Large the use of these novel nanographenes as new electrochro- Grained Perovskite Solar Cells. J. Mater. Chem. A. mic material, and demonstrate the use of these materials for memory devices. Conclusion We expect to achieve synthesis of processable nanogra- phene (NG). Our proposed synthesis promises ways to control NGs with size-dependent band gap, optical absorp- tivity, and charge transfer functionality. If successful, this project will likely generate a new class of materials with emergent functionality previously not accessible through fabrication methods. The integration of functional NGs into clean energy technologies could bridge the gap between basic research and commercialization of graphene-based energy devices. Further developing NG-based materials will strengthen our leadership role in NG research, which has strong ties to Laboratory mission in the areas of exotic materials and energy security. Publications Cheruku, P., J. H. Huang, H. J. Yen, R. S. Iyer, K. D. Rector, J. S. Martinez, and H. L. Wang. Tyrosine-Derived Stimuli 129

Page 130

Complex Natural & Engineered Systems Postdoctoral Research and Development Continuing Project Petabyte-Scale Computational Analyses of Genomic Data to Elucidate Aging Mechanisms William S. Hlavacek 20140670PRD2 Introduction molecular differences associated with age. We expect to The project has two objectives: 1) analyses of data from be able to characterize the rates of several distinct muta- different species and 2) analyses of data from thou- tional processes occurring in 50 different tissue types. sands of individuals and multiple human tissue types.To achieve the first objective, we will use the reference ge- Benefit to National Security Missions nome data available from NCBI to perform comparative The project will entail petabyte-scale data science. The genomics analysis across 25 species. A preference will be methods used and lessons learned are likely to be rel- given to non-human primates and rodents because the evant to many DOE mission areas such as biothreats and species in each of these groups have similar genomes DOE Office of Science. The specific problem that will be but diverse lifespans. For example, the naked mole addressed is highly relevant to the public health mission rat lives up to 30 years, whereas its relative the house of NIH. By analyzing large amounts of genomic data, we mouse lives only about four years. Additionally, in-depth hope to elucidate molecular signatures and causes of analysis will be performed for different breeds of dogs aging. Aging has been identified by the World Health Or- due to the large amounts of available molecular data. ganization as one of the main healthcare and economic Although all dogs share essentially the same genetic challenges of the 21st century. Remarkably, our current material, different breeds exhibit up to 2-fold differences understanding of the mechanisms of aging is extremely in life expectancy. limited and as a 2012 review in Cell concluded, “The underlying cause of aging remains one of the central The main outcome of this part of the project will be mysteries of biology.” Two (not necessarily exclusive) the identification of homologous genes responsible for hypotheses have been put forward to explain aging. The lifespan differences across species and the genes af- first postulates that aging is due to the accumulation of fected by nucleotide changes in different breeds of dogs. damage, which can come both from the external envi- The second objective is focused on humans. We will use ronment and from within our bodies and cells. The sec- predominately the data generated by initiatives such as ond theory suggests that aging follows a predetermined the International Cancer Genome Consortium (ICGC). biological timetable in which certain genes get switched The primary goal of the ICGC is the identification of on and/or off to control lifespan. critical genes involved in oncogenesis, but a byproduct of this large-scale worldwide initiative is an exhaustive Progress genetic and epigenetic characterization of the normal In the past year, we were able to build a capability to tissues (usually blood) and cancers of 25,000 individual download large quantities of data from external sourc- patients. In a study recently reported in a Nature article, es. Currently, we have a working pipeline that allows it has been demonstrated that these abundant molecu- downloading ~30TB per day of external data. Note that lar data can be used to identity signatures of molecular the download capability of the pipeline is throttled to processes causing somatic mutations. Moreover, these adhere to LANL’s internal requirements and practically signatures can be mapped to molecular mechanisms. the capacity of the pipeline could be further increased. This study sheds new light on the origins of cancer but it In total, we have downloaded ~300TB of data to a stag- just scratches the surface of what is possible. To achieve ing area and are currently working on the most effective the second project objective, we will use the molecular strategy for transferring data from this staging area to data available for up to 25,000 individuals to identify 130

Page 131

the HPC resources, where the data will be processed and Patel, S. Lunke, L. B. Alexandrov, and e. t. al. Tracking analyzed. We envision having this workflow completed in the origins and drivers of subclonal metastatic the next few months and to able to process at least 2 PB of expansion in prostate cancer. 2015. Nature Communications. 6: 6605. data by the end of this calendar year. Overall, we plan to download the majority of available genomics/biomedical Schulze, K., S. Imbeaud, E. Letouze, L. B. Alexandrov, data from around the world by the end of this project. J. Calderaro, and e. t. al. Exome sequencing of hepatocellular carcinomas identifies new mutational In the past year, we were also able to perform big data signatures and potential therapeutic targets. 2015. analysis outside of the hitherto mentioned pipelines. The Nature Genetics. 47 (5): 505. analysis was performed in regards to cancer and human Wegener, R., L. B. Alexandrov, M. Montesinos-Rongen, aging, and more specifically to elucidate the activity of M. Schlesner, A. Haake, H. G. Drexler, J. Richter, G. mutational processes in neoplastic and aging cells. This R. Bignell, U. McDermott, and R. Siebert. Analysis of work resulted in significant scientific output in regards to mutational signatures in exomes from B-cell lymphoma high impact peer-reviewed publications as well as a patent cell lines suggest APOBEC3 family members to be application. Additionally, there are another seven manu- involved in the pathogenesis of primary effusion scripts that are currently under preparation and we expect lymphoma. 2015. Leukemia. 29 (7): 1612. them to be submitted (and some of them accepted) by the end of this calendar year. Yates, L. R., M. Gerstung, S. Knappskog, C. Desmedt, G. Gundem, P. Van Loo, T. Aas, L. B. Alexandrov, and e. t. al. Subclonal diversification of primary breast cancer Future Work revealed by multiregion sequencing. 2015. Nature We plan to complete and fully automate our pipelines for Medicine. 21: 751. downloading and processing large-scale genomics data. This will provide LANL with an automated workflow for processing Big Data from any external sources. From a scientific perspective, we plan to processes ~5 PB of data in the next fiscal year accounting for about ~20,000 individual pairs of cancer-normal genomes. These data will allow us to have the final roadmap on both mutational processes in human cancer as well as mutational processes that are part of normal ageing processes. We will attempt to associate the activity of these processes to gene expres- sion patterns, epigenetics profiles, and clinical outcome. We envision that this work will results in between 10 and 20 high impact scientific publications. Conclusion The expected outcome is an in-depth characterization of the genetic changes that occur with age and the muta- tional processes responsible for these changes, which will have far-reaching implications for human health. In over- coming the technical challenges of the proposed work, we will establish a Big Data Science capability in Los Alamos that could be applied to diverse problems. The planned analyses will encompass roughly 25 petabytes of data and involve the use of advanced methods of multivariate sta- tistics, novel method development, and new approaches to the use of computational resources to solve a Big Data problem. Publications Hong, M. K., G. Macintyre, D. C. Wedge, P. Van Loo, K. 131

Page 132

Complex Natural & Engineered Systems Postdoctoral Research and Development Continuing Project Access to Industrially Important Optically Active beta-X-alcohols via Direct Enantioselective Ester Hydrogenation Pavel Dub 20140672PRD2 Introduction in the hydrogenation of a variety of C=O bonds that are Contrary to ketones, limited progress in the hydrogena- relevant to catalytically relevant industrial processes. tion of less-electrophilic esters has been made during One paper has been accepted and we have several oth- the last few decades. Industrial homogeneous reduction ers being worked on for publication including an invited of esters still relies mostly on the use of metal hydride review article. We have also gone to quite some effort reagents such as LiAlH4 or NaBH4. Using these reagents, to protect the intellectual property surrounding this hydrogenation selectivities cannot be controlled and technology. This includes the filing of four provisional US such spent (waste) reductants require energetically patents and a recent non-provisional PCT application. intensive regeneration input or disposal. Also the use of heterogeneous catalysis dictates the use of high tem- Future Work peratures and pressures and again selectivities cannot In FY16 we will further develop our ligand chemistries be controlled. to support new and effective catalysts for the hydroge- nation of ketonic substrates. These new complexes will No reports describe detailed enantioselective ester potentially be used in other important catalytic transfor- hydrogenation despite the tremendous potential for syn- mations as well, (e.g hydrosilylations, hydroborations, thetic organic chemistry. Current “state-of-the-art” ca- transfer hydrogenations, allylic substitutions, Michael talysis in this area of chemistry has required complicated additions, Henry reactions, etc.), so if time permits we multi-step ligand synthesis as well and these ligands will examine some of these reactions. once synthesized often exhibit air-sensitivity. This work will thus focus on developing new well-defined chiral Conclusion bifunctional molecular catalysts incorporating air-stable In addition to the environmentally benign processes ligands for the enantioselective hydrogenations of esters outlined above, our goals will go further: currently, the via dynamic kinetic resolution. ruthenium (Ru) catalyzed asymmetric hydrogenation of ketones (R. Noyori, Nobel Prize 2001) is the main route Benefit to National Security Missions to produce optically active secondary alcohols. The utili- Success in this chemistry could have significant positive zation of expensive Ru catalysts however is unattractive impact for LANL, vis-à-vis the application of new ap- in terms of cost. It is envisioned that if the enantioselec- proaches to polymers and materials synthesis, in energy tive hydrogenation of ketones could be catalyzed by less- applications, and in health related (e.g. pharmaceutical) expensive base metal complexes. We will incorporate R&D. This work if successful will potentially lead to high newly developed and synthesized air-stable chiral ligands profile publications (Angew. Chemie., JACS, etc. ) for the onto cheap metals such as copper in order to perform Laboratory and will place LANL at the forefront of this catalytic asymmetric ketone hydrogenation. developing area of chemistry. Publications Progress Dub, P. A., B. L. Scott, and J. C. Gordon. “Air-Stable NNS In FY15, we have made very good progress during the (ENENES) Ligands and their Well Defined Ru and Ir last year, making a large variety of catalysts based on Complexes for Outer-Sphere Molecular Catalysis”. 2015. Organometallics. 44: 4464. ligands that are supported by NNS (nitrogen, nitrogen and sulfur) donors. These have demonstrated efficacy 132

Page 133

Complex Natural & Engineered Systems Postdoctoral Research and Development Continuing Project Synthesis and X-ray Spectroscopy of Actinide Thiocyanates Stosh A. Kozimor 20140677PRD3 Introduction preparing M(NCS)x(y-) compounds. This includes the As the need to develop a closed nuclear fuel cycle grows, structural characterization of over 40 compounds by there are a number of technological challenges that single crystal X-ray diffractometry. Additional, prelimi- need to be overcome. One of the most difficult chal- nary measurements have been carried out at SSRL and lenges is the separation of the minor actinides (Am3+ the ALS synchrotrons. and Cm3+) from the trivalent lanthanide fission products as they share many physical properties. A number of Future Work studies have shown that this separation is possible by Work in FY16 will focus on developing a synthetic route employing soft ligands (such as those containing S or to trivalent f-block SCN1- complexes, and (b) establishing N donors), which preferentially bind the actinides (An) the bonding character of these complexes through mul- over the lanthanides (Ln). One particularly promising tiple spectroscopic and theoretical techniques, primarily extractant is the SCN1- anion that provides enhanced S, N, and C K-edge XAS and TDDFT. separation of An over Ln. Although little is understood about why the SCN1- extractant is successful, it seems Conclusion likely to be associated with the propensity of NCS1- to Completion of this work will mark the first synthesis form more covalent An–NCS vs Ln–NCS bonds. We of trivalent An3+(NCS)xy- compounds and the first propose to characterize the M–NCS interactions by syn- transplutonium S, N, and C K-edge XAS experiments. thesizing An(NCS)xy- (An = Pu, Am, Cm) complexes and Developing an understanding of the bonding in these subsequently analyzing the An(NCS)xy- compounds using complexes will shed light on both the differences in Ln/ ligand K-edge X-ray absorption spectroscopy (XAS) and An–NCS interactions and highlight how An–NCS bonding DFT calculations. varies as the 5f series is traversed from Pu to Cm. Benefit to National Security Missions Publications This investigation will develop a fundamental under- Macor, J. A., J. L. Brown, J. N. Cross, S. R. Daly, A. standing of bonding for plutonium and the other J. Gaunt, G. S. Girolami, M. T. Janicke, S. A. actinide elements. The results will be applied to DOE’s Kozimor, M. P. Neu, A. C. Olson, S. D. Reilly, efforts to design practical and efficient separations of the and B. L. Scott. Coordination Chemistry of lanthanides and minor actinides. Moreover, the research 2,2′-Biphenylenedithiophosphinate and will provide discrete probes for quantifying covalency Diphenyldithiophosphinate with U, Np, and Pu. that will be used in future efforts to analyze strategi- 2015. DALTON TRANSACTIONS. ASAP (ASAP): NA. cally important actinide materials and aid the DOE in its mission to support the national nuclear agenda that includes expanding nuclear power for energy security, preventing the spread of nuclear weapons, securing nu- clear materials against theft, reducing the size of nuclear arsenals, and cleaning up the legacy of the Cold War. Progress Over FY15 Cross has made considerable advances in 133

Page 134

Complex Natural & Engineered Systems Postdoctoral Research and Development Continuing Project Anaerobic, Solvothermal Synthesis of Lanthanide and Actinide Kagomé Antiferromagnets Stosh A. Kozimor 20140681PRD4 Introduction nide/actinide analogs of the jarosite minerals (A1+) We propose to synthesize these 2D kagomÁ 4f and 5f (Fe3+)3(OH)6(SO42–)2. These materials have never materials for the first time. The lanthanide/actinide ana- before been prepared. In the first year we will focused logs of the jarosite minerals (A1+)(Fe3+)3(OH)6(SO42–)2 heavily on setting up the capability for the hydrothermal will be sought. The choice of monovalent cations (A1+; synthetic method. Now with this in place, we are work- Ag, Rb, Cs, etc) and the oxo dianions (MO42–; M = S, ing hard on developing methods to crystalize com- Se, Te, Mo, W, etc) will template the kagomÁ network pounds that contain Ce3+ (4f1) and Yb3+ (4f13) metals. and expand the lattice to incorporate the f elements. The compounds will be characterized by UV-Vis, XAFS, Initially, we will focus on Ce3+ (4f1) and Yb3+ (4f13) magetometry, and XRD. The synthetic method will serve materials and apply the developed methods across the as a basis for syntheses for other lanthanide elements lanthanide series. The 4f studies will provide a basis for and eventually be used to expand to plutonium and the expanding to plutonium and the early actinides. Finally, early actinides. the structural, magnetic, and electronic properties of the new materials will be characterized in collaboration with Conclusion Bauer, Batista, and Marten. A new two-year effort to synthesis 2D frameworks con- taining magnetically frustrated f elements is proposed. Benefit to National Security Missions The synthetic component of this project will establish This research complements LANL’s efforts associated new design principles for unconventional actinide/ with DOE’s Office of Basic Energy Science heavy element lanthanide materials that might provide access to new chemistry program. The work focuses on understanding types of actinide superconductors. Spectroscopic results fundamental concepts associated with plutonium bond- will be used in advancing f element theory. Overall, ing and electronic structure. Overall, these efforts will These efforts will provide a strong scientific foundation provide a strong scientific foundation for the DOE and for the DOE and its laboratories to directly support the its laboratories to directly support the national nuclear national nuclear agenda and the integrated Plutonium agenda and the Integrated Plutonium Strategy. Strategy. Progress In FY15, La Pierre has developed a new capabilities in solid state synthesis and in hydrothermal synthesis. A number of new teranary oxides were also synthesized. Additionally, La Peirre prepared for the first time (NMe4) NpCl6, which was fully characterized and is being written up for publication. These compounds have been charac- terized by X-ray powder diffraction. Additionally samples have been prepared for analysis at the SSRL synchrotron. Future Work In FY16 we will continue our synthetic efforts to generate methods for preparing a series of lantha- 134

Page 135

Complex Natural & Engineered Systems Postdoctoral Research and Development Continuing Project A Physics-Based Numerical Model for Next-Generation Laminar Flow Batteries Qinjun Kang 20150700PRD1 Introduction structure, wettability, porosity, as well as the overall cell Flow batteries have emerged as an efficient and eco- geometry and dimension will be optimized. nomical device for the storage of large quantities of energy. They have also received much attention for their Benefit to National Security Missions promising applications as power sources for vehicles and The successful accomplishment of the proposed re- portable electronics, due to its quick recharge character- search will (1) result in a comprehensive multiscale istics compared to lithium-ion batteries. Membranes are numerical model for coupled transport and electro- the most costly and unreliable component of the flow chemical processes in flow batteries; (2) provide optimal batteries. The membrane-less laminar flow battery can electrode characteristics, overall cell geometry, and eliminate the need for an ion-exchange membrane by operational conditions (e.g. flow rate) for improved cell relying on laminar flow and slow diffusion to separate performance. These results may lead to next-generation the reactants in anode and cathode. However, power membrane-less high power density flow batteries and densities as high as membrane-based counterparts have facilitate their applications in energy storage and power not been achieved. A recently developed membrane- supply. The capabilities to be developed in this project less hydrogen bromine laminar flow battery (HBLFB) can also have direct application to other earth and energy achieve a high power density of 0.795 Wcm-2 at room programs, including geological storage of nuclear waste temperature and atmospheric pressure. In this battery, and carbon dioxide, development of petroleum and the characteristics of the microporous anode (composi- geothermal reservoirs, fuel cells, and micro reactors. tion, structure, wettability, porosity) and the cathode This project will support the LANL Grand Challenge to (flow field design, surface structure) have not been opti- develop transformative new energy technologies and mized. Previous work has shown that these parameters to significantly enhance and extend the use of current can have an impact on the power density by nearly an technologies in a manner that is sustainable and that order of magnitude. mitigates negative environmental, social, and national security impacts. It will also support DOE goals in pro- In this project, we will develop a physics-based multi- moting energy security. scale numerical model to understand the coupled physi- cochemical processes in membrane-less flow batteries, Progress with the ultimate goal to optimize cell parameters to Since this project started in January 2015, Dr. Li Chen, achieve higher power density. We plan to approach this has made significant progress towards developing a mi- problem in two stages by: 1) developing a microscale croscale model accounting for the complex structures of model accounting for the complex structures of the the microporous electrodes and all the relevant physico- microporous anode and all the relevant physico-electro- electro-chemical and transport processes. Particularly, chemical and transport processes; and 2) developing a he has developed a numerical model for coupled diffu- macroscopic model considering the entire cell, using the sion, conduction, and electrochemical reaction. The gov- upscaled microscale model results as input. erning equations for proton transfer in electrolyte and reactant transport in porous electrodes are solved using Processes including two-phase distributions within the the lattice Boltzmann method (LBM), with the electro- porous anode, electrolyte flow between the anode and chemical reactions occurring at the electrolyte/electrode cathode, as well as the electrochemical reactions at interface treated in the source term. both electrodes, will be considered, and the electrode 135

Page 136

tional conditions (e.g. flow rate) for improved cell perfor- Dr. Chen is currently working on adding fluid flow and mance. These results may lead to next-generation mem- convection of reactants in this newly developed model, brane-less high power density flow batteries and facilitate and solving the coupled governing equations for fluid their applications in energy storage and power supply. flow (Navier-Stokes equation), ion transport (Nernst-Plank equation), and electronic potential (Poisson equation) Publications within the LBM framework in an iterated fashion. The Chen, L., G. Wu, E. F. Holby, P. Zelenay, W. Tao, and Q. Kang. resulting model will be the first physics-based, comprehen- Lattice Boltzmann Pore-Scale Investigation of Coupled sive pore-scale model for coupled transport and electro- Physical-electrochemical Processes in C/Pt and Non- chemical processes in the electrodes of flow batteries, Precious Metal Cathode Catalyst Layers in Proton taking into account electrolyte flow, advection, diffusion, Exchange Membrane Fuel Cells. 2015. Electrochimica electrochemical migration and reaction. The transport ActaElectrochimica Acta. 158: 175. model will be utilized to fully understand various physio- chemical phenomena involved in flow batteries to help minimize transport losses and facilitate optimized material design and architectures, leading to increased flow battery performance and reduced cost. Future Work The overarching goal of this project is to develop a novel, physics-based numerical model to simulate coupled physico-electro-chemical and transport processes in flow batteries, to optimize cell parameters for achieving high power densities. We planned to approach this problem in two stages by: 1) developing a microscale model account- ing for the complex structures of the microporous anode and all the relevant physico-electro-chemical and transport processes; and 2) developing a macroscopic model consid- ering the entire cell, using the upscaled microscale model results as input. In the past five months, we have made significant progress in stage 1 by developing a microscopic numerical model for coupled diffusion, conduction, and electrochemical reac- tion. In FY16, we will first finish developing the microscale model (stage 1) by incorporating fluid flow and convection of reactants in the newly developed model. We will then start developing the macroscopic model considering the entire cell, using the upscaled microscale model results as input. Finally, we will consider all involved processes in the flow battery, including two-phase distributions within the porous anode, electrolyte flow between the anode and cathode, as well as the electrochemical reactions at both electrodes. We will optimize the electrode structure, wet- tability, porosity, as well as the overall cell geometry and dimension for better flow battery performance. Conclusion This project will (1) result in a comprehensive multiscale numerical model for coupled transport and electrochemi- cal processes in flow batteries; and, (2) provide optimal electrode characteristics, overall cell geometry, and opera- 136

Page 137

Complex Natural & Engineered Systems Postdoctoral Research and Development Continuing Project Resolving Kinetic Scales in 3D Global Magnetosphere Simulations William S. Daughton 20150703PRD1 Introduction of scaling the calculations to large numbers of cores The magnetic field of the Earth protects our planet (~100,000) and in remote visualization and analysis of against the stream of energetic particles that are con- the results. The scientific results from this research may tinually emitted from the sun. This interaction leads to benefit current and upcoming NASA missions focused the formation of the magnetosphere, a bubble of hot on understanding the near-Earth space environment. In ionized gas (plasma) surrounding our planet. During particular, the upcoming Magnetospheric Multi-Scale large solar storms, eruptions from the sun compress the (MMS) mission will focus on the kinetic physics of mag- magnetosphere and give rise to the injection of ener- netic reconnection with the Earth’s magnetosphere. The getic particles that can damage satellites. It is a major simulations performed in this project will be among first challenge to understand and properly model how these to study the complex coupling between these kinetic energetic particles are able to penetrate into the mag- scales and the global dynamics, supporting LANL scientif- netosphere. This project will develop a new simulation ic discovery and innovation in space science and nuclear, approach to accurately describe the dynamics of this particle, cosmology, and astrophysics. complex system, and offer detailed predictions that can be compared with spacecraft observations. This new Progress simulation approach will be far more accurate than cur- Excellent progress has been made towards the goals of rent models of the magnetosphere, since it will solve this PRD project, which started only four months ago the most basic underlying equations for the ions in the (February, 2015). First, we have ported the plasma plasma using some of the largest supercomputers in the simulation code H3D to computer clusters at LANL in country. In addition, the electrons will be modeled with order to perform global modeling of the Earth’s mag- a new theoretical approach that has demonstrated im- netosphere. The H3D code has a rigorous description proved accuracy in small test problems, but has not yet of the ions in the plasma, but only an approximate been applied to large-scale modeling. With these inno- fluid description of the electrons. One of the primary vations, we anticipate this new kinetic model will leap- goals of this project is to implement an improved fluid frog beyond the present models and provide a wealth model for the electrons into the H3D code. During the of new scientific results and predictions. To test these past few months, we have already implemented a new new predictions, comparisons will be made with satellite equation-of-state that accounts for the effects of elec- observations that measure the energetic particles in the tron pressure anisotropy, which in previous work has Earth’s magnetosphere. been shown to be a critical feature. Next, we have tested this new closure relationship for the electrons using a Benefit to National Security Missions simple two-dimensional test problem involving magnetic This research will exploit the newest generation of pet- reconnection in a current sheet. Magnetic reconnection ascale supercomputers using a kinetic simulation code is a basic progress that converts magnetic energy into specially optimized for these machines. Some of these plasma kinetic energy, while giving rise to changes in calculations will be performed locally at Los Alamos connectivity of the magnetic field. This process is critical using NNSA computers, while other calculations will be to model accurately, since it gives rise to the entry of performed on flagship DOE supercomputers such as solar wind plasma into the Earth’s magnetosphere. Our Titan at Oak Ridge. These efforts are at the very fore- preliminary hybrid simulation results agree quite well front of high performance computing, both in terms with first-principles kinetic simulations performed with 137

Page 138

VPIC, the standard tool for these problems at LANL. These reconnection at the dayside magnetopause. Depending on results show the development of intense current sheets how far we get, we will also look for opportunities to com- in the reconnection outflow, which are expected to be key pare our initial results with spacecraft observations. observational signatures in spacecraft data. We are cur- rently investigating the influence of grid resolution and dis- Conclusion sipation on these results within the H3D code. Meanwhile, The fluid models currently used to model the magneto- we have begun performance testing the 3D version of H3D sphere are not fully justified, particularly in critical regions on LANL HPC machines using a new allocation of resources such as shock waves and thin boundary layers. While more that were awarded to this project under the Institutional accurate simulations have been performed in these local Computing Program. Small 3D reconnection simulations regions, the feedback on the global dynamics remains have been performed including the new equations of state. poorly understood. In this project, we will demonstrate the Working towards our goal of 3D global simulations of the feasibility of simulations that more accurately describe the Earth’s magnetosphere, our initial 3D test problem is to entire system. We anticipate this project may dramatically simulation the magnetosphere of Ganymede, one of Jupi- improve our ability to model the space weather environ- ter’s moons. This problem is computationally much easier ment surrounding the Earth, in regions where communica- than the Earth’s magnetosphere, which will allow us to tion satellites play a crucial role in modern society. more carefully test the H3D code. Preliminary simulations for Ganymede look very promising and a full-scale produc- tion run on a 512-cubed grid is currently being computed on the Mustang computer at LANL. Future Work The dynamical coupling between the solar wind and the Earth’s magnetosphere is crucially dependent on a number of key regions where kinetic effects are important. This in- cludes the bow shock that develops upstream of the Earth, the shocked solar wind plasma (called the magnetosheath) and a very thin boundary layer along the periphery of the magnetosphere called the magnetopause. Depend- ing upon the orientation of the interplanetary magnetic field (IMF), there are several key processes that can lead to the entry of solar wind plasma into the magnetosphere. This includes magnetic reconnection, which occurs due to changes in the connectivity of field lines across the magne- topause, and the Kelvin-Helmholtz instability, which leads to vortex formation and turbulent transport along the magnetopause boundary. The goal of this project is to de- scribe the entire system with a kinetic description, in order to better understand the entry of solar wind plasma across the magnetopause and into the magnetosphere. In FY 15, we ported this hybrid code to LANL computer clusters, and implemented a new closure model for the electrons that will more accurately describe this physics. In FY16, we will carefully test this new electron model for a series of test problems, and publish this comparison in a paper. In addition, during FY16 we will continue the small scale 3D simulations of Ganymede (one of Jupiter’s moons) that we recently started, and compare the results with recently published papers from fluid models. Finally, during FY16 we will begin performing global simulations of the magnetosphere, starting with the limit of southward IMF where the dominant entry process is due to magnetic 138

Page 139

Complex Natural & Engineered Systems Postdoctoral Research and Development Final Report Chemically Modifying the Uranyl Ion Jaqueline L. Kiplinger 20120750PRD2 Abstract Scientific Approach and Accomplishments This is the final report of a three-year Frederick Reines Thorium Starting Materials: (CpR)2ThCl2. In 2010, LANL Distinguished Postdoctoral Fellowship project (March reported the development of a convenient and safe solu- 2012 to June 2015) at Los Alamos National Laboratory. tion route to the thorium starting material ThCl4(DME)2 During this project, several different areas of actinide (DME = dimethoxyethane), using inexpensive commer- research were explored and they are summarized below. cially available reagents, and avoiding the use of thorium metal. During research studies on metallocene thorium Background and Research Objectives azides (see Metallocene Thorium Diazides, below), it Since the Manhattan Project, which sought to identify was discovered that adding dimethoxyethane (DME) to volatile compounds for the separation of fissionable the reaction mixture when ThCl4(DME)2 is the starting isotopes for nuclear applications, there has been sig- material had a significant effect on reaction times, condi- nificant interest in actinide chemistry. Understanding tions, and yields. This finding was then acutely dem- the fundamental chemistry and speciation of actinides onstrated using the syntheses of known metallocene is critical to properly handling and processing nuclear dichlorides (C5Me5)2ThCl2, (1,2,4-tBu3-C5H2)2ThCl2, materials and disposing of radioactive waste. For many and (C5Me4Et)2ThCl2 (Figure 1). years, the majority of studies on thorium and uranium involved aqueous systems of direct relevance to nuclear fuel processes. However, under these conditions, sys- tems are operating at, or close to, thermodynamic sinks (e.g., uranyl) and important chemistry of the actinides can be masked. Under non-aqueous conditions, chem- istry is performed in anhydrous organic solvents, giving Figure 1. General scheme depicting the synthesis of metallo- opportunities to prepare actinide complexes that would cene thorium dichlorides from ThCl(DME); shows that the ad- dition of DME to a reaction mixture improves synthetic efficacy not normally exist and thus gain glimpses of the “hid- of ThCl(DME). den” and “true” character of thorium and uranium. This permits the study of novel thorium–ligand and uranium– The original syntheses of these dichlorides had all ligand bonds and their reactivity, which enables us to involved harsh conditions with long reaction times from address fundamental questions regarding the extent and the now-outdated starting material ThCl4. These results nature (5f vs. 6d) of covalency in actinide chemical bond- were communicated in the peer-reviewed journal Inor- ing and how it impacts reactivity and physicochemical ganic Chemistry Communications in 2014. Importantly, properties. This knowledge could impact on waste reme- also in 2014, the new starting materials NpCl4(DME)2 diation where a key approach to minimize the volume and PuCl4(DME)2, which are the transuranic analogs of of radioactive waste, and recycle useful components, ThCl4(DME)2 and developed in collaboration with LANL involves separation technologies; these aim to deploy colleague Andrew Gaunt, were reported in the journal ligands that exploit the different levels of covalency in Dalton Transactions. This is significant because it shows the chemical bonding of the myriad of elements that are that findings in thorium chemistry can be applied to the present in radioactive waste. transuranics, meaning it is likely that the results from this project can be translated to the Np and Pu starting materials as well. 139

Page 140

Thorium Starting Materials: ThI4(DME)2 synthesized from the U(III) complex (C5Me5)2U(bipy) and Since the synthesis, solubility, structure, and reactivity of halide salts (Figure 3). This family of new uranium-centered actinide halide starting materials are strongly influenced complexes is interesting due to being ligated by the redox- by the nature of the halide, and also because there had active chelating ligand 2,2’-bipyridine (bipy), making the been limited progress in the field of thorium tetraiodide complexes electronically flexible and calling into question chemistry, we were motivated to target the synthesis of the oxidation states of the uranium center and the bipy the thorium tetraiodide complex, ThI4(DME)2. A synthetic ligand. This will be a high-impact contribution to the field route to ThI4(DME)2 from ThCl4(DME)2 was developed of actinide chemistry with a full study of the electronic and optimized. It was discovered that ThI4(DME)2 exhibits properties across the series of compounds using absorp- excellent thermal stability compared to ThI4(THF)4, which tion spectroscopy and electrochemistry. Of special note undergoes rapid ring-opening of THF at ambient tempera- is the fluoride example, due to (1) the rarity of uranium ture. In a test reaction, it was found that salt metathesis fluorides relative to the other halides, and (2) the fact that between ThI4(DME)2 and K(LMe) (LMe = (2,6-iPr2C6H3) the fluoride ligand is poised to react with silicon reagents, NC(Me)CHC(Me)N(2,6-iPr2C6H3)) cleanly gives (LMe) a relatively unexplored type of reactivity with the potential ThI2O(CH2)4I, which is a rare example of a thorium to yield highly sought-after synthetic targets. β-diketiminate complex. A demonstration of the synthetic utility of ThI4(DME)2 was completed by the preparation of thorium(IV) alkoxide, amide, and organometallic com- pounds. This work was published in 2012 in Dalton Trans- actions and was selected for the cover of the issue (Figure 2). Figure 3. General scheme for the synthesis of (CMe)UX(bipy). X-ray crystal structures of (CMe)UBr(bipy) (structures for X = F and Cl are analogous) and (CMe)U(N)(bipy). Futhermore, it was found that a different type of metal- locene uranium bipyridyl halide is formed when the same U(III) starting material, (C5Me5)2U(bipy), was treated with benzyl (bz) halides. Here, the halide oxidatively adds to the uranium while the bz group migrates to the 5-position on the bipy ligand, forming a new carbon-carbon bond and a new “bipy-bz” ligand (Figure 4). The next step is to use benzyl fluoride; cleaving the very strong C-F bond will be significant, and based on our results with the other halides it is expected this reaction will proceed in a facile manner. Figure 4. General reaction scheme for the synthesis of (CMe) U(bipy-bz). Representative X-ray crystal structure of (CMe) UBr(bipy-bz) (structure for X = Cl is analogous). Metallocene Thorium Diazides. The first molecular urani- um nitride was reported by LANL in 2010; it was generated by photolysis of a uranium azide complex. In general, ac- Figure 2. Cover of the issue of Dalton Transactions containing tinide nitride ceramics are of interest as alternative nuclear the ThI(DME) report. fuels, but very little is known about their chemical proper- ties or reaction behavior due to the challenges surrounding Metallocene Uranium Bipyridyl Halides the study of extended materials. An alternative approach A series of metallocene uranium bipyridyl halides was is to generate well-defined molecular analogs, which are 140

Page 141

simpler to study. Since the 2010 LANL report, there have Impact on National Missions been only two other molecular uranium nitrides reported, The DOE complex and LANL clearly benefit from explor- and no molecular nitrides of any other actinides have been atory research into the chemistry of the actinide elements, reported. because of their criticality to DOE nuclear weapons mis- sion areas. The synthesis of new actinide compounds not With the generation of molecular thorium nitrides as a only provides important information about metal-ligand long-term target, this project’s goal was to generate a se- bonding but also improves our ability to reliably predict ries of molecular thorium azides, since azides are a classic their chemical behavior in a variety of applications from entry into nitride chemistry via photolysis, thermolysis, or performance validation of waste repositories, to the design chemical reduction. There are only two molecular tho- of selective chemical separation methodologies, to mitiga- rium azides in the literature, and neither of them are ideal tion of corrosion processes in aging weapons components. candidates for conversion to nitrides due to the steric and The recent focus on energy security and the tacit acknowl- electronic profiles of their ancillary ligands. edgment that nuclear energy must be part of the mix to achieve energy independence and reduce greenhouse gas This work has resulted in a series of new metallocene emissions has served to heighten this interest even further. thorium diazide complexes synthesized using salt metath- esis between (CpR)2ThCl2 and NaN3 (Figure 5, left), all of Publications which have been fully characterized. It was discovered that their diverse solid-state structures were governed by the Monreal, M. J.. Metallocene Thorium Azide Complexes: size of the ancillary ligands (Figure 5, right). All new com- Diverse Solid-State Structures Governed by Ancillary pounds were characterized using elemental analysis, X-ray Ligand Steric Effects. Invited presentation at 249th crystallography, multinuclear nuclear magnetic resonance National Meeting of the American Chemical Society . (NMR) spectroscopy, infrared (IR) spectroscopy (solu- (Denver, CO, 22-26 March 2015). tion and solid), and for the first time for any structurally Monreal, M. J.. Establishing a Surrogate Pyrochemical Ca- characterized actinide azide, Raman spectroscopy. These pability at LANL (Round Table Discussion Leader). Invit- results will be submitted to the high-impact peer-reviewed ed presentation at Joint Working Group (JOWOG) 22/2 journal Angewandte Chemie International Edition. Work is Technical Exchange @ Lawrence Livermore National currently underway to convert these azide complexes to ni- Laboratory. (Livermore, CA, September 2015). trides; initial photolysis experiments have yielded exciting preliminary results, which inspired new actinide chemistry Monreal, M. J., B. L. Scott, and J. L. Kiplinger. Organometal- proposals. lic thorium azide complexes: using steric pressure to control structure . Invited presentation at LANL Post- doctoral Women’s Group Colloquium. (Los Alamos, February 26, 2014). Monreal, M. J., B. L. Scott, and J. L. Kiplinger. Organome- tallic thorium azide complexes: using steric pressure to control structure. Invited presentation at LANL Women’s History Month Panel and Poster Session. (Los Alamos, NM, March 31, 2014). Monreal, M. J., B. L. Scott, and J. L. Kiplinger. Organo- metallic thorium azides: metallocene sterics govern structure . Invited presentation at Laboratory Directed Research and Development Program Review. (Los Ala- mos, NM, ay 14, 2014). Monreal, M. J., B. L. Scott, and J. L. Kiplinger. Organometal- lic thorium azide complexes: steric pressure controls structure . Presented at LANL Postdoc Research Day. (Los Alamos, NM, June 11, 2014). Figure 5. Syntheses of metallocene thorium diazides with cor- Monreal, M. J., B. L. Scott, and J. L. Kiplinger. Organometal- responding X-ray crystal structures on the right. Trend: smaller lic thorium azide complexes . Invited presentation at ancillary ligands multinuclear complexes (), while larger ancillary 27th Rare Earth Research Conference. (Squaw Valley, ligands mononuclear complexes (). CA, 22-26 June 2014 ). 141

Page 142

Monreal, M. J., R. J. Wright, J. T. Golden, D. E. Morris, B. L. Scott, P. P. Power, and J. L. Kiplinger . Thorium(IV) and uranium(IV) halide complexes supported by bulky β-diketiminato ligands. 2013. Organometallics (Invited Article – Special Issue: Recent Advances in Organo-f- Element Chemistry). 32: 1423. Monreal, M. J., R. K. Thomson, B. L. Scott, and J. L. Kip- linger. Enhancing the synthetic efficacy of thorium tetrachloride bis(1,2-dimethoxyethane) with added 1,2-dimethoxyethane: Preparation of metallocene tho- rium dichlorides. 2014. INORGANIC CHEMISTRY COM- MUNICATIONS. 46: 51. Travia, N. E., M. J. Monreal, B. L. Scott , and J. L. Kiplinger. Thorium-mediated ring-opening of tetrahydrofuran and the development of a new thorium starting mate- rial: Preparation and chemistry of ThI4(DME)2. 2012. Dalton Transactions. 41: 14441. Travia, N. E., M. J. Monreal, B. L. Scott, D. E. Morris , and J. L. Kiplinger. Developing the next generation thorium iodide starting materials. Invited presentation at 243rd American Chemical Society National Meeting. (San Di- ego, California, 25-29 March 2012). Travia, N. E., M. J. Monreal, B. L. Scott, and J. L. Kiplinger. A new starting material for thorium chemistry: ThI4(DME)2. Presented at Gordon Research Confer- ence on Inorganic Chemistry. (University of New Eng- land, Biddeford, Maine, 17-22 June 2012). 142

Page 143

Complex Natural & Engineered Systems Postdoctoral Research and Development Final Report Catalytic Mechanism and Inhibition of Metallo-beta-lactamases (MBL), the Ultimate Threat Against Antibiotics Ryszard Michalczyk 20120776PRD4 Abstract ever-changing nature of MBLs ensure that this is not a A number of bacterial pathogens produce enzymes able long-term solution. to inactivate β-lactam antibiotics, which comprise more than 80% of the world’s market of antimicrobials. These enzymes, known as β-lactamases, rank among the most proficient enzymes known. At present, there are no clinically useful MBL (metallo-β-lactamases) inhibitors available. Our research efforts aimed to discover MBL inhibitors or novel antibiotics resistant to inactivation. Joint Neutron and X-ray crystallography is the only methodology providing details about H atom positions Figure 1. Carbapenems like imipenem (above) are readily inactivated by MBLs. in these enzymes and potential inhibitors, details that will be fundamental for future drug design efforts. Three At present, there are no clinically useful MBL inhibitors broad lines of research were explored for NDM-1 and available. Research efforts aim to discover MBL inhibi- BcII metallo-β-lactamases: (1) Water activation, (2) Sub- tors or novel antibiotics resistant to inactivation. How- strate binding, and (3) Inhibition. ever, our limited knowledge of the MBL’s structure-func- tion thwarted such efforts thus far. The MBL’s reaction This research has direct implications for understanding mechanism has been shown not to involve covalently of these enzymes’ reaction mechanism and enabling the bound intermediates, making unfeasible the classic “sui- design of clinically useful MBL inhibitors that will allow cide substrate” inhibition approach. re-introduction of many powerful antibiotics currently withdrawn because of their sensitivity to β-lactamases. Joint Neutron and X-ray crystallography is the only This has direct impact on the public health security, the methodology providing details about H atom positions well-being of millions of people and the world pharma- (i.e. water, H-bonds, protonation state of amino acids), ceutical industry. fundamental for future drug design efforts. During my time at LANL I characterized the enzyme NDM-1, an MBL Background and Research Objectives of major concern. Three broad lines of research were A number of bacterial pathogens produce enzymes able explored: (1) Water activation, (2) Substrate binding, and to inactivate β-lactam antibiotics, which comprise more (3) Inhibition. This work required all the Protein Crystal- than 80% of the world’s market of antimicrobials. These lography Station resources: the deuteration and protein enzymes, known as β-lactamases, rank among the most expression laboratory, protein purification and crystal- proficient enzymes known. Due to their readiness for lization laboratory as well as the neutron beamline and horizontal transfer and evolution, MBLs entail a serious X-ray diffraction suite. public health issue. A special group of Zn(II)-dependent β-lactamases, termed metallo-β-lactamases (MBLs), Scientific Approach and Accomplishments display a worrisome substrate promiscuity, hydrolyzing As outlined in the Research Proposal, the project com- penicillins, cefalosporins, and carbapenems, the latest prises three main objectives centered on the structure- generation of β-lactams (Figure 1). Although mono- function relationships of metallo-β-lactamases (MBLs): bactams are the only known poor MBL substrates, the (1) Mechanism of water activation, (2) Mechanism of 143

Page 144

substrate binding, and (3) Inhibition. The progress on each of these points is detailed below. Mechanism of water activation Two MBLs are currently being studied: NDM-1 from K. pneumoniae and BcII from B. cereus. Both wild-type en- zymes are being expressed under BSL-2 conditions under IBC-127-2013. This objective will be achieved through solving neutron structures, for which mm-size crystals are needed, which requires large amounts of pure protein. The wild-type NDM-1 enzyme proved to be difficult to express in E. coli BL21(DE3) in a native form due to prote- olysis. After several unsuccessful attempts to determine conditions in which NDM-1 would express in a native form (adjusting induction growth temperature and length, use of IPTG or autoinducing media, use of protease inhibitor cocktails) we decided to focus first on the other MBL, BcII (which expresses very well), in order to pursue the neutron Figure 2. X-ray data collection and refinement statistics. structure. It was determined that BcII does not grow mm-size crys- Inhibition of MBLs tals using the vapor-diffusion method in 9-well plates and This part of the project is focused on two families of po- 0.5–1 ml drops. Instead, large BcII crystals grow using dialy- tential inhibitors: sulfonamides and bisthiazolidines. The sis method, using commercially available dialysis buttons sulfonamides 3-[(2-oxo-1,3-oxazolidin-3-yl)carbonyl]-2-pyr- with 3,000-Da cutoff membrane. BcII crystals are prone idinesulfonamide (OCPS), 6-methoxy-3-nitro-2-pyridine- to degradation through oxidation of residue Cys221 upon sulfonamide (MNPS) and 3-(4-methyl-5-oxo-4,5-dihydro- long incubation times (weeks to months). Since neutron 1,3,4-oxadiazol-2-yl)-2-pyridinesulfonamide (DOPS) are beamtime is scarce, BcII crystals will need to be freshly commercially available, and were selected based on their grown close to the data collection time. I plan to grow central 2-pyridine sulfonamide moiety. These compounds mm-size BcII crystals for the upcoming neutron beam cycle have the potential to bind to the second-zinc of MBLs starting in October 2014, for which the corresponding thereby inhibiting the enzyme competitively (through beamtime request application has been submitted and is preventing substrate binding) or uncompetitively (through pending approval. removing the second-site zinc). As a proof-of-concept, the structure of BcII complexed with 2-pyridine sulfonamide Mechanism of substrate binding was solved at pH 6, showing the expected binding mode, For reasons outlined above, this research is also focused i.e. bidentate binding to the second-zinc site through nitro- on BcII. Two double mutants were designed, cloned, puri- gen atoms present in the probably deprotonated sulfon- fied, crystallized, and their structures were solved through amide –NH2 and the pyrimidine. Crystallization trials are in x-ray crystallography, namely, BcII-DDE (Cys221Asp, progress for BcII complexed with compounds OCPS, MNPS His263Glu) and BcII-DDQ (Cys221Asp, His263Gln). As and DOPS. These crystals will also be shipped and data col- expected, both display an empty second-zinc site and no lected in the upcoming trip to SSRL. β-lactamase activity. The next step is to co-crystallize these The second group of compounds, bisthiazolidines (BTZs), mutants with different metals and intact β-lactam antibiot- is bicyclic heterocyclic compounds with two fused thia- ics. Crystals have already been grown in the presence of zolidine rings, thereby emulating penicillins. These com- ampicillin and different concentrations of Zn2+ (as a nega- pounds were obtained through a collaborative work with tive control), Ca2+, Mg2+, Sr2+, Ba2+, Fe3+, Co2+, Mn2+, researchers from Uruguay and Argentina, who synthesized Ni2+, Cu2+, Al3+, and Sm3+, in order to screen conditions and isolated them. I have solved the x-ray structures of BcII in which these metals would enable substrate binding but complexed with three of them (CS319, GM256 and VC26), no catalysis. These crystals will be flash-frozen and shipped and measured their inhibition constants in vitro, which to the Stanford Synchrotron Radiation Lightsource (SSRL) resulted to be in the micromolar range. These structures for x-ray data collection, for which I have been awarded have shed light on possible modifications leading to im- beamtime on July 8-9th, proposal SSRL 4B25A. proved inhibiting properties in these compounds, particu- 144

Page 145

larly VC26, which are being introduced by our collaborators in Uruguay. Impact on National Missions The proposed work utilized an innovative research strategy that is LANL-unique for novel antibiotics. MBLs comprise a massive global human health issue causing many deaths and substantial economic losses around world, due to the innumerable complications associated with untreatable infections caused by some of the most dangerous bacteria present in clinical settings and intensive care units. This work, which can only be performed at LANL’s world-class facilities, will provide the first neutron structures for antibi- otic resistance enzymes found in multidrug resistant patho- gens, with direct implications for understanding of these enzymes’ reaction mechanism and enabling the design of clinically useful MBL inhibitors that will allow re-intro- duction of many powerful antibiotics currently withdrawn because of their sensitivity to β-lactamases. This has direct impact on the public health security, the well-being of mil- lions of people and the world pharmaceutical industry. Publications Gonzalez, J. M., M. M. Gonzalez, M. Kosmopoulou, C. Saiz, V. Castillo, G. Mahler, R. Bonomo, J. Spencer, and A. J. Vila. Inhibition of Metallo-β-Lactamases by Bisthiazoli- dines. . Presented at 2014 American Crystallographic Association Annual Meeting. (Albuquerque, 24-28 May 2014). John, F. J. St, D. Dietrich, C. Crooks, E. Pozharski, J. M . Gonzalez, E. Bales, K. Smith, and J. C. Hurlbert. A Novel Member of Glycoside Hydrolase Family 30 Subfamily 8 with Altered Substrate Specificity. To appear in Acta Crystallographica Section D. Marti-Arbona, R., J. M. Gonzalez, and C. J. Unkefer. Crystal Structure of Phosphoenolpyruvate Carboxylase from Methylobacterium extorquens. Presented at 2014 American Crystallographic Association Annual Meet- ing. (Albuquerque, 24-28 May 2014). Turra, G., C. Faugel, J. M. Gonzalez, D. Presello, C. Andreo, and V. Campos-Bermudez. La glioxalasa I de maíz re- gula niveles de metilglioxal, una molécula señal de estrés en Fusarium verticilloides?. Presented at XXX Argentine Meeting of Plant Physiology. (Mar del Plata, Argentina, 21-24 September 2014). Unkefer, C. J., J. Chen, S. Z. Fisher, J. M. Gonzalez, J. P. Bacik, and M. J. Waltman. Time-of-flight Laue Neutron Crystallography at the Protein Crystallography Station (PCS) at LANSCE. . Presented at 2014 American Crystal- lographic Association Annual Meeting. (Albuquerque, 24-28 May 2014). 145

Page 146

Complex Natural & Engineered Systems Postdoctoral Research and Development Final Report Stimuli Responsive, Functional Biopolymers: Quinic Acid-Based Polymers and Their Assemblies Hsing-Lin Wang 20130778PRD1 Abstract ment of functional biomolecules in which the existing This project aimed to develop functional biomolecules functionalities can be coupled with conjugated oligo- with emergent (mechanical, optical, and electronic) mer[3][4], dye or other complex biomacromolecule properties that are not accessible through conventional for sensing and electronic applications. The functional synthetic methods. The approach uses molecules that groups can interact or self-assemble in ways that domi- are known to form polymers compatible with biological nate the formation of hierarchical structure and their systems with side chains that are active to which func- corresponding properties. Understanding the source tional groups can be attached. Depending on the dis- of these interactions between specific components will tribution and density of these functional groups on the allow us to achieve functional materials with emergent side chains, a wide spectrum of mechanical and optical properties. properties, and electronic structures can be accessed. Specifically, we proposed to investigate the develop- Our goal is to exploit the emergent properties, and ment of functional biopolymers in which the existing through the analysis of structure-property relationships, functionalities can be coupled with oligomer, dye or obtain valuable insight regarding the other complex biomacromolecules for sensing and design of functional biomolecules with enhanced func- electronic applications. We also proposed to develop tionality that can be used toward sensing, bioimaging, biomolecules with build-in functionality resulting from drug delivery, optical, electronic and energy applications. its natural occurrence, such as chirality, which allows us Such an approach has not yet been studied in a system- to further incorporate functional groups with control atic manner mainly because multidisciplinary knowledge in spatial resolution. We planned to develop unnatural ranging from organic synthesis, self-assembly, to materi- amino acid biomolecules and their self-assemblies and als chemistry and biochemistry is needed. The challenge explore their properties against various stimuli such as is to develop synthetic chemistry routes that are facile temperature, pH etc., striving to synthesize a series of and effective. Further, we will demonstrate control of biomolecules belong to the category of unnatural amino functional groups on in order to render desired struc- acid and demonstrate their use for bioimaging and tures and properties. organic electronic devices. These biomolecules can act as spectroscopic reporters on the state of the polymer Background and Research Objectives conformation and on the state of protonation levels. Stimuli-responsive polymers are polymers that respond Conformational changes within the polymer trigger the with dramatic property changes to small changes in their arrangement/proximity of the fluorophores to put them environment such as chemicals, temperature, pH etc. in different environments and thus lead to a modulation Development of the materials based on stimuli respon- of fluorescence. Biomolecules with conjugated structure sive molecules represents an emerging and interdisci- and amine end group leads to a pH-dependent optical plinary topic at the border of life sciences,[1] material properties as they can turn on/off the intermolecular sciences[2] and nanotechnology[2] as these polymers charge transfer process by varying the pH value of the are ideal candidates for many applications in the sensor solution. systems, drug delivery, multivalent drug scaffolds, and tissue engineering. In this context, we would like to investigate the develop- 146

Page 147

Scientific Approach and Accomplishments confocal fluorescence microscopy. Synthesis and characterization of non-natural fluorescent Water-soluble biopolymers for bioimaging amino acid and their applications to bioimaging Water-soluble cationic conjugated polymers (WSCPs) [5] We have synthesized a series of fluorescent unnatural show a strong affinity toward the negatively charged cell amino acids (UAAs) bearing stilbene and meta-phenylen- membranes, staining cell surface using WSCPs is a rela- evinylene (m-PPV) backbone and their optical properties tively unexplored area due to the endocytosis problem. were studied using a suite of spectroscopy probes. These However, the ability to functionalize these polymers with novel amino acids were derived from protected diiodo- specific recognition ligands enables selective detection L-tyrosine using palladium-catalyzed heck couplings with among the various cell lines and different organelles. a series of styrene analogs. Unlike the other fluorescent To date, only one report has appeared in the literature UAAs, whose emissions are restricted to a narrow range disclosing the ability of WSCPs to be stable cell staining of wavelengths, these new amino acids display emission agents. Water-soluble polythiophene conjugated with lapa- peaks at a broad range of wavelengths (from 400 to 800 tinib, an anticancer tyrosine kinase inhibitor, can be used nm); including NIR with a quantum yield of 4% in HEEPS to induce the polymer interaction with transmembrane buffer. The incorporation of both pyridine and phenol func- proteins and thus achieve stable cell membrane staining. tional groups lead to distinct red, green, and blue (RGB) emission, in its basic, acidic and neutral states, see Fig. 1. More importantly, these amino acids showed reversible pH and redox response showing their promise as stimuli responsive fluorescent probes. Figure 2. WS-PPV staining of various cell types. (a) Fixed and stained mouse fibroblast cell line 3T3 cells were stained with WS- PPV and imaged using two different filter channels (red, 670 nm/ green, 519 nm). (b) Fixed 3T3 cells were stained with actin stain- ing reagent. (c) Fixed mouse myoblast cell line C2C12 cells were stained with WS-PPV (yellow) and DAPI (blue). The serial muscle cells composed differentiated myotubes were clearly observed via WS-PPV staining. (d) C2C12 cells were fixed and stained with anti-myosin heavy chain antibody (green) and nucleus staining Figure 1. Emission spectra of unnatural amino acid in neutral (blue). (e) Fixed and stained mouse motor neuron cell line (NSC- (blue), acidic (green) and basic (red) environment showing the 34). The extension of neurite network can be observed through RGB emission. the staining of WS-PPV polymer (yellow). Cell nucleuses were localized using DAPI staining (blue). (f) Mouse motor neuron cells To further demonstrate the utility of these UAAs in peptide (NSC-34) were fixed and stained with anti-beta III tubulin anti- synthesis, one of the amino acids in this series was incor- body (magenta) and nucleus staining (blue). All cells were stained porated into a cell penetrating peptide (CPP) sequence with WS-PPV (1 µM) for 10 min. Scale bar = 100 µm. through standard solid phase peptide synthesis. The To further evaluate the robustness of WS-PPV in cell general use of these UAAs for cell imaging has been dem- imaging, we extended our studies to a series of different onstrated by applying CPP to two different cell lines using cell lines. Mouse fibroblast (3T3) cells, mouse myoblast 147

Page 148

(C2C12) cells, mouse motor neuron cells (NSC-34) and tubes is commonly used methods to quantify in vitro myo- human alveolar basal epithelial cells (A549) were stained genic differentiation. Despite the selectivity and specificity independently using 0.5 µM WS-PPV for 10 min. As shown of immunostaining, this widely used approach suffers from in Fig. 4, different cells could be easily visualized after WS- time-consuming staining protocols and cost-effectiveness. PPV staining. Myotube staining by using other than fluorescent antibod- ies is hardly explored so far. Thus, we wanted to validate the ability of WS-PPV in myotube staining. It is particularly noteworthy that when C2C12 mouse muscle cells were incubated with WS-PPV, myotubes were stained within 10 min (Fig. 2c) as compared to immunostaining that takes 2 days to be visualized (Fig. 2d). In case of NSC-34 cells, neurites (cell body of a neuron) were stained (Fig. 2e) with WS-PPV in the same staining condition while standard immunocytochemistry staining of neuron cells (Fig. 2f) has longer staining time. Further- more, the visualization of cells cultured on 3-D substrates sometimes can be a big challenge. For example, the absorption of fluorescent molecules in the substrate (e.g. hydrogel, porous scaffold) can easily increase the high fluorescent background when 3-D substrate is observed through microscopy analysis. Here, we demonstrated the other capability of WS-PPV to stain the cells inside 3-D sub- strate. The key technique is staining the cells in advance Figure 3. (top) Synthetic scheme of unnatural amino acids, and using suspension staining. After staining, the WS-PPV (bottom) molecular structure of two conjugated oligomers and stained cells were spun down and then seeded in the lu- the SEM images of the hierarchical self-assembled structures. men of cellulose based hollow fiber. After 3 days of cultur- Actin based filaments (i.e. filopodia) have been shown to ing, A549 cells were still spotted inside hollow fiber. This play a critical role in controlling the cell migration, explor- advantage allows us to monitor a long-term cell culture in ing their environments and wound healing. However, to 3-D substrate without using any additional post staining observe the formation of filament network, the stain- which will significantly reduce the fluorescent background. ing method is usually restricted to immunostaining and fluorescently labeled phallotoxins. These methods have Cell membrane staining using WSCPs without any targeting limitations in terms of their cost-effectiveness and long ligands for bio recognition is not known. We have shown staining time, usually greater than 48 hours. Therefore, the very exciting results using cationic poly(p- phenylene discovery and evaluation of alternate staining agents is a vinylene) (PPV), a WSCP for cell imaging. P2 as bioimaging valuable addition to this field and to the best of our knowl- polymer exhibits stability allowing long-term cell tracking, edge, polymeric imaging agents for filaments staining are distinguishing live and dead cells, and multi- functional cell barely explored. staining. Overall, our results show that using WSCP is far While imaging fixed 3T3 cells, we found that WS-PPV poly- superior than state-of-the art staining commercial dyes as mer effectively stained the cytoskeleton and the obtained WSCP has a much faster staining time at a much lower cost images are almost comparable to that of standard stain- (100 times cheaper than existing commercial dye), while ing images. Interestingly, actin fiber and nucleus can be providing more details and imaging resolution as opposed stained at the same time in a single stain by WS-PPV and to the commercial dyes. Such unprecedented properties can be observed simply by using different microscope fil- suggest that certain WSCPs may serve as next generation ters (Fig. 2a), where as, in case of commercial actin staining imaging dyes for biomedical research. agents (Fig. 2b), a separate secondary staining is required Programmed self-assembly of unnatural amino acid to visualize the nucleus. This is particularly advantageous Initially we have studied the self-assembly of three newly as staining with WS-PPV serves as a dual staining agent and synthesized m-PPV analogs (structures A, B in Fig. 3) to lowers the cist and time required for the imaging. test our hypothesis. The only difference between A and The C2C12 myoblast cell line is a widely used model to B is that a pyridine ring in molecule B replaces the phe- study myogenesis in vitro. Immunostaining of the myo- nyl ring in A. The pyridine moiety in compound B brings 148

Page 149

in additional H-bonding source along with π-stacking. and white light emissive material through controlled inter- As anticipated, structure B self-assembles to form nano- molecular charge transfer. 2014. CHEMICAL SCIENCE. : DOI: structures due to the nicely positioned two pairs of H- 10.1039/C4SC01911C. bonding acceptors (Fig. 3_blue) and donors (Fig. 3_red). Gao, Y., C. C. Wang , L. Wang , and H. L. Wang. Conjugated Electron microcopy studies revealed that the compound polyelectrolytes with pH-sependent conformations and B showed the nanofiber formation. Further morphologi- optical properties. 2007. LANGMUIR. 23: 776. cal characterization with SEM, TEM and AFM revealed that the tiniest nanofiber (fibril) has a diameter of 20 nm Publications and several nanofibers can come together to give 60 -100 Cheruku, P., J. H. Huang , H. J. Yen, R. S. Iyer,, Kirk D. Rec- nm nanofiber (bundles). This is particularly interesting as tor, J. S. Martinez, and H. L. Wang. Tyrosine-derived the morphology of these fibers is in close agreement with stimuli responsive, fluorescent amino acids . 2015. the amyloid fibers.19-22 As shown in Fig. 3, it is evident CHEMICAL SCIENCE. 6: 1150. that the build in chiral center leads to helical nanofiber as compare to dotted morphology for molecule without Park, Y. I., O. Postupna, A. Zhugayevych, H. Shin, Y. -S. Park, the H-bonding acceptor, Fig, 3A. Our results reveal a novel B. Kim, H. -J. Yen, P. Cheruku, J. S. Martinez, J. W. Park, programmed self-assembly of chiral conjugated oligomer, S. Tretiak , and H. -L. Wang. A new pH sensitive fluo- an amino acid precursor, with enhanced optical proper- rescent and white light emissive material through con- ties. To the best of our knowledge, nanofibers comprised trolled intermolecular charge transfer. 2014. CHEMI- CAL SCIENCE. : DOI: 10.1039/C4SC01911C. of conjugated oligomers with enhanced optical properties have never been reported. Impact on National Missions These new classes of biomolecules and their self-assem- blies may exhibit high temperature thermal transitions and strong mechanical properties due to the high degree of hydrogen bonding and rigid carbocyclic backbone. The correlation between nanoassembly structures and their as- sociated properties yield insights on the interplay between structure, dynamics and functions of the materials. We also expect this research to feed into novel strategies for the design and synthesis of nanoassemblies with adap- tive control to enable the materials with technological relevance in areas of optics, electronics, photovoltaic, and biomedical devices that impact our missions in energy security and threat reduction. References Shim, M. S., and Y. J. Kwon. Stimuli-responsive polymers and nanomaterials for gene delivery and imaging applica- tions. 2012. Advanced Drug Delivery Reviews. 64: 1046. Stuart , M. A. C., W. T. S. Huck, , J. Genzer, M. Müller, C. Ober, M. Stamm, et.al. Minko. Emerging applications of stimuli-responsive polymer materials. 2010. NATURE MA- TERIALS. 9: 01. Cheruku, P., J. H. Huang , H. J. Yen, R. S. Iyer,, Kirk D. Rec- tor, J. S. Martinez, and H. L. Wang. Tyrosine-derived stimuli responsive, fluorescent amino acids . 2015. CHEMICAL SCIENCE. 6: 1150. Park, Y. I., O. Postupna, A. Zhugayevych, H. Shin, Y. -S. Park, B. Kim, H. -J. Yen, P. Cheruku, J. S. Martinez, J. W. Park, S. Tretiak , and H. -L. Wang. A new pH sensitive fluorescent 149

Page 150

Complex Natural & Engineered Systems Postdoctoral Research and Development Final Report Single-Cell Genomics for Better Control of Plant Pathogens Shunsheng Han 20130779PRD1 Abstract and investigate how these changes benefit the wild type Many bacteria establish long-term populations on their population. hosts or within specific environments. One strategy for Scientific Approach and Accomplishments bacterial persistence is the formation of biofilms, bac- terial aggregates embedded in a self-produced matrix. To identify mutational events during P.chlororaphis colo- Compared to free-living cells, the cells living in the nization, bacterial cells was grown in biofilm conditions. biofilm microenvironment may be exposed to smaller Small colony variants were identified by visual inspec- amounts of free iron, oxygen, and nutrients. In this tion of bacterial colonies on plates. 2. SCV genome was study, we examined the transcriptional and genomic sequenced using next sequencing technology in Dr. Cliff alterations associated with the phenotype changes in Han’s group. To identify beneficial mutations, resulted the beneficial root-colonizer P. chlororaphis 30-84. This genome sequences were compared with published work comprises a genome-wide evolutionary analysis of genome sequence in collaboration with Dr. Patrick Chain a phenotypic switching event in P. chlororaphis. Genes and Chien-chi Lo. A number of mutations were identified identified by genomic or transcriptomic approaches and experimentally verified to be responsible for pheno- will provide novel insights to better understand genetic typical changes. The results provide us with a complete adaptations during biofilm formation. set of mutations that are involved in host association. 3. To study the role of these genomics changes in bacterial Background and Research Objectives persistence, beneficial populations was compared with Microbes survive under unfavorable conditions by under- the wild type strain in their ability to produce fungal- going a number of morphological and genetic changes. inhibiting metabolites and form biofilms. The SCVs ex- Consequently, natural populations often contain variants hibited high resistance to antimicrobials and adherence with beneficial mutations that contribute to the public ability. 4. Genomic mutations often lead to changes in good. However, due to technical limitation, the identity gene expression. To understand the beneficial mutation and mechanism of environment-induced mutations at transcritpome level, RNA sequencing has been con- remains poorly understood. Genome sequencing allows ducted to identify differentially expressed genes be- us to sequence bacterial genomes and thus provide valu- tween wild type and beneficial mutant strains. We found able information on genomic variation in response to that genes involved in iron uptake and stress responses environmental changes. The proposed research used the were highly expressed. Together, these data give us a rhizosphere-colonizing bacterium Pseudomonas chloro- comprehensive understanding of cell differentiations in raphis strain 30-84. The strain was first identified due to response to environmental signals. 5. The above analysis its ability to inhibit fungal pathogens and has become has identified candidate genes involved in successful dis- a model for a beneficial commensal bacterium. Strain ease control. This data has been published in the journal 30-84 demonstrates phenotypic variations resulting from Applied Environmental Microbiology, which is the #1 spontaneous mutations during rhizosphere coloniza- cited journal in Microbiology. tion. These phenotypic variants (such as small colonies Impact on National Missions or pigment deficient mutants) are beneficial to the wild type population by increasing cell attachment, biofilm Our results have demonstrated that bacteria can adapt formation and thus biological control activity. The goal of to adverse environmental conditions by altering their this research was to identify signature genomic and tran- genome composition and global gene expression pat- scriptomic changes in response to environmental signals tern. This project is a model on future analysis of micro- 150

Page 151

bial evolution. This is particularly important for the study of clinical antibiotic-resistant bacteria e.g. how they alter their genomes and transcriptomes to overcome antibiotic killing. Deciphering the molecular diversity in vivo is a sig- nificant breakthrough in Microbiology and results in a com- prehensive understanding of molecular process involved in bacterial adaptation. Knowing how bacteria couple envi- ronmental inputs with genomic changes is important in the development of disease management strategies. Publications Tao, Y., F. Wang, D. Jia, J. Li, Y. Zhang, C. Jia, D. Wang, and H. Pan. Cloning and functional analysis of the promoter of a stress-inducible gene (ZmRXO1) in maize. 2014. Plant Molecular Biology Reporter. : 1. Wang, D., C. Han, A. Dichosa, C. D. Gleasner, S. Johnsons, H. E. Daligault, J. M. Yu, E. A. Pierson, and L. S. III Pier- son. Draft genome sequence of Pseudomonas putida strain S610, a seedborne bacterium of wheat. 2013. Genome Announcement. 1 (6): e01048. Wang, D., C. Han, A. Dichosa, C. D. Gleasner, S. Johnsons, H. E. Daligault, J. M. Yu, E. A. Pieson, and L. S. III Pieson. Draft genome sequence of Enterobacter cloacae strain S611 . 2014. Genome Announcement. 2 (6): 00710. Wang, D., R. J. Dorosky, C. Han, C. H. Lo, A. Dichosa, P. Chain, J. M. Yu, L. S. III Pierson, and E. A. Pierson. Ge- nomic adaptations of a small colony variant in the biofilm of Pseudomonas chlororaphis 30-84. 2015. Ap- plied Environmental Microbiology. 81 (3): 890. 151

Page 152

Complex Natural & Engineered Systems Postdoctoral Research and Development Final Report Exploring Doubly Parasitic Radioisotope Production Via Secondary Neutron Fluence From the 100 MeV IPF Irradiations Eva R. Birnbaum 20130782PRD2 Abstract energy cyclotrons around the nation, and energetically The spallation neutron flux produced from the standard distinct from reactor neutron fluences. The potential of proton irradiation of rubidium chloride and gallium this unique capability for research in novel methods of targets at the Los Alamos National Laboratory (LANL) isotope production and materials science is unexploited. Isotope Production Facility (IPF) was investigated as a Several radioisotopes named in the Isotopes Subcommit- tool for an innovative route for isotope production and tee’s recent report to the Nuclear Science Advisory Com- materials science studies. Routine irradiations have been mittee (NSAC) are potentially suitable for such neutron- found to produce neutron fluxes as high as 1012 n cm-2 based production. These radioisotopes are necessary s-1, with approximately 50% of the total flux having to a variety of scientific and medical fields, and in many energy in excess of 1 MeV. Experimental measurements cases their future, uninterrupted supply is uncertain. of radionuclide production using this flux were com- This research project intended to establish new produc- pared with the predicted radionuclide yield using nuclear tion routes for these important isotopes. excitation functions from MCNPX event generators, Scientific Approach and Accomplishments evaluated nuclear data, and the TALYS nuclear code via a collaboration with computational nuclear data experts Custom target assemblies were fabricated to isolate neu- in XCP-3. Practical application of the secondary neutron tron activation samples from cooling water at the rear of flux in the realm of radioisotope production is consid- the standard RbCl-RbCl-Ga IPF target stack. In this way, ered, and has been explored experimentally for multiple exposure of the foils to the neutron flux created from productions undertaken by the Isotope Program in the a typical production run at IPF was assured, changes to past two years. Several small-scale production experi- the proton targets’ configuration (and hence on routine ments have been successfully completed to produce production deliverables) were avoided completely, and tracer amounts of radionuclides from this unique neu- a geometry was established which could be utilized tron flux. Findings have further been used in support of again for future irradiations and potentially for isotope a DOE Early Career Award proposal application aimed at production. Activation studies of multiple sample types the characterization of unstudied nuclear reactions using were performed over a period of two years, with exten- quasimonoenergetic neutron fluxes at multiple interna- sive comparison of experimental yields (quantified by tional facilities. gamma spectroscopy and radiochemisty) and theoretical predictions in collaboration with XCP-3 (Figure 1). The Background and Research Objectives results of this work have been summarized in confer- The 100 MeV Isotope Production Facility (IPF) at Los ence proceedings [1] and in Nuclear Instrumentation and Alamos National Laboratory operates in conjunction with Methods in Physics Research [2], and have also resulted the Los Alamos Neutron Science Center (LANSCE), pro- in student presentations at a national meeting [3]. ducing 82Sr, 68Ge, and other isotopes under the purview of the Isotope Production Program. A typical target stack for the large-scale production of 82Sr and 68Ge includes two targets of Inconel-encapsulated rubidium chloride powder and a third target of liquid gallium metal. The IPF’s high proton beam current and lengthy irradia- tions produce a secondary neutron flux with a utilitar- ian scale that is beyond the reach of the many medical 152

Page 153

1000 100 10 1 0.1 0.01 0.00 27 1 Al(n 2 ,x 7 2)4 A N l(n a ,x2)2Na 59Co(n 5 , 9 x5)8 Co C ( o n 5 , 9 x5)7 Co C ( o n 59 ,x5) C 6 o C (n o 5 , 9 x5)4 Co M (n n , 5 x5 9 )2 C M o( n n,x5)9Fe 89 8 Y 9 (n Y ,x ( 8 n )8 ,x8) Y 7(m 8 + 9 g) Y Y (n 8 , 9 x8)6 Y( Y n, 8 x8 9 )4 Y R ( b n,x8)3R 19 b 7Au( 1 n 9 , 7 x1)9 A 6 u( A 1 n 9 u , 7 x1)9 A 5 u( A n 1 u ,x 9 1 7 )94 A A u( u n,x1)92I 2 r 09Bi(n 20 ,x 9 2)0 B 7 i( B n 2 i 0 ,x 9 2)0 B 6 i( B n 2 i 0 ,x 9 2)0 B 5 i( B n 2 i 0 ,x 9 2)0 B 4 i( B n 2 i , 0 x 9 2)03 Bi B (n i ,x2)02 n l a Ni(n n , a x 5)4 N M i(n n , n x a 5) 2 N M i( n n na ,x 5)9 Ni F (n e n , a x 5)5 N C i( o n n , a x 5)6 N C i( o n n , a x 5)7 N C i( o n n ,x a 5 )8 N C i( o n n , a x 5)6 N N i(n i na ,x 5)7 Ni N (n i ,x6)0Co na n(n na ,x 6)4 n C ( u n,x6)7Cu 153 )i(C �leiY e il uN lau i eR 3. Delorme, K. A., J. W. Engle, F. M. Nortier, and B. R. x erimen al Yiel Neu ron Yiel rom 1 �2A40 + beam a I F C B er M i 0 n 3 i .02 Kowash. Production potential of Sc-47 using spallation A Y N 2011 neutrons at the Los Alamos Isotope Production Facility. Presented at Society of Nuclear Medicine and Molecu- lar Imaging. (St. Louis, MO, 7-11 Aug. 2014). Publications Delorme, K. A., J. W. Engle, F. M. Nortier, E. R. Birnbaum, and B. R. Kowash. Production potential of Sc-47 using Nu learRea ion Figure 1. Experimental yields for isotopes generated in the IPF spallation neutrons at the Los Alamos Isotope Produc- secondary neutron flux as compared to yields predicted by three tion Facility. Presented at Society of Nuclear Medicine different modeling programs. and Molecular Imaging. (St. Louis, MO, 7-11 Aug. 2014). Impact on National Missions Engle, J. W., C. T. Kelsey, Bach, B. D. Ballard, M. E. Fass- Characterization of the secondary neutron flux by a bender, K. D. John, E. R. Birnbaum, and F. M. Nortier. combination of experiment and theoretical prediction has Preliminary Investigation of Parasitic Radioisotope Pro- already enabled the flux to be put to use producing small duction Using the LANL IPF Secondary Neutron Flux. quantities of radionuclides needed for neutrino mass mea- 2012. 14TH INTERNATIONAL WORKSHOP ON TARG- surement efforts, developmental radiochemistries, and ETRY AND TARGET CHEMISTRY. 1509: 171. to collect preliminary data for multiple additional funding proposals. As a result of this work, the Isotope Program Engle, J. W., C. T. Kelsey, H. Bach, B. D. Ballard, M. E. Fass- bender, K. D. John, E. R. Birnbaum, and F. M. Nortier. has at its disposal an accessible, cost-efficient, and parasit- Preliminary investigation of parasitic radioisotope ic means of radionuclide production. Accurate flux charac- production using the LANL IPF secondary neutron flux. terization has resulted in multiple expressions of interest 2012. In 14th International Workshop on Targetry and from materials science and fusion energy communities, Target Chemistry. (Cancun, Mexico, 26-29 Aug. 2012). and before the end of the 2014 calendar year, first ship- Vol. 1, 1 Edition, p. 1. New York: American Institute of ments of tracer radiolanthanides produced in the second- Physics. ary neutron flux will be shipped from the Isotope Program at LANL to university collaborators. As a preliminary means Engle, J. W., E. R. Birnbaum, F. M. Nortier, J. A. Rau, K. D. to study the potential of spallation energy neutrons for John, and H. R. Trellue. Purification of (PU)-P-242 by irradiation with thermal neutrons. 2013. NUCLEAR IN- radioisotope production, the secondary neutron flux at IPF STRUMENTS & METHODS IN PHYSICS RESEARCH SEC- has been responsible for the inception of multiple novel TION B-BEAM INTERACTIONS WITH MATERIALS AND isotope production schemes for which funding support has ATOMS. 298: 70. been requested from DOE sponsors. Engle, J. W., E. R. Birnbaum, F. M. Nortier, J. A. Rau, K. D. References John, and H. R. Trellue. Purification of (PU)-P-242 by

1. Engle, J. W., C. T. Kelsey, H. Bach, B. D. Ballard, M. E. irradiation with thermal neutrons. 2013. NUCLEAR IN- Fassbender, K. D. John, E. R. Birnbaum, and F. M. Nort- STRUMENTS & METHODS IN PHYSICS RESEARCH SEC- ier. Preliminary investigation of parasitic radioisotope TION B-BEAM INTERACTIONS WITH MATERIALS AND ATOMS. 298: 70. production using the LANL IPF secondary neutron flux.
2. In 14th International Workshop on Targetry and Engle, J. W., E. R. Birnbaum, H. R. Trellue, K. D. John, M. W. Target Chemistry. (Cancun, Mexico, 26-29 Aug. 2012). Rabin, and F. M. Nortier. Evaluation of Ho-163 produc- Vol. 1, 1 Edition, p. 1. New York: American Institute of tion options for neutrino mass measurements with Physics. microcalorimeter detectors. 2013. NUCLEAR INSTRU- MENTS & METHODS IN PHYSICS RESEARCH SECTION
3. Engle, J. W., M. R. James, S. G. Mashnik, C. T. Kelsey, B-BEAM INTERACTIONS WITH MATERIALS AND ATOMS. L. E. Wolfsberg, D. A. Reass, M. A. Connors, H. T. Bach, 311: 131. M. E. Fassbender, K. D. John, E. R. Birnbaum, and F. M. Nortier. MCNPX characterization of the secondary neu- Engle, J. W., E. R. Birnbaum, H. R. Trellue, K. D. John, M. W. Rabin, and F. M. Nortier. Evaluation of Ho-163 produc- tron flux at the Los Alamos Isotope Production Facility. tion options for neutrino mass measurements with
4. NUCLEAR INSTRUMENTS & METHODS IN PHYS- microcalorimeter detectors. 2013. NUCLEAR INSTRU- ICS RESEARCH SECTION A-ACCELERATORS SPECTROM- MENTS & METHODS IN PHYSICS RESEARCH SECTION ETERS DETECTORS AND ASSOCIATED EQUIPMENT. 754: B-BEAM INTERACTIONS WITH MATERIALS AND ATOMS.

Page 154

311: 131. radioactive strontium production QA/QC. 2014. JOUR- NAL OF RADIOANALYTICAL AND NUCLEAR CHEMISTRY. Engle, J. W., J. W. Weidner, B. D. Ballard, M. E. Fassbender, 302 (1): 347. L. A. Hudston, K. R. Jackman, D. E. Dry, L. E. Wolfsberg, L. J. Bitteker, J. L. Ullmann, M. S. Gulley, Pillai, Goff, E. R. Medvedev, D. G., J. W. Engle, F. M. Nortier, S. V. Smith, and Birnbaum, K. D. John, S. G. Mashnik, and F. M. Nortier. L. F. Mausner. Theoretical approach to the production Ac, La, and Ce radioimpurities in Ac-225 produced in of Ti-44 (T-1/2=59.1 years) using high energy protons: 40-200 MeV proton irradiations of thorium. 2014. RA- Expanding Sc-44 availability. 2013. ABSTRACTS OF PA- DIOCHIMICA ACTA. 102 (7): 569. PERS OF THE AMERICAN CHEMICAL SOCIETY. 245. Engle, J. W., M. R. James, S. G. Mashnik, C. T. Kelsey, L. E. Wolfsberg, D. A. Reass, M. A. Connors, H. T. Bach, M. E. Fassbender, K. D. John, E. R. Birnbaum, and F. M. Nortier. MCNPX characterization of the secondary neu- tron flux at the Los Alamos Isotope Production Facility. 2014. NUCLEAR INSTRUMENTS & METHODS IN PHYS- ICS RESEARCH SECTION A-ACCELERATORS SPECTROM- ETERS DETECTORS AND ASSOCIATED EQUIPMENT. 754: 71. Engle, J. W., S. G. Mashnik, Bach, Couture, Jackman, Gritzo, B. D. Ballard, Fassbender, D. M. Smith, L. J. Bitteker, J. L. Ullmann, M. S. Gulley, Pillai, K. D. John, E. R. Birnbaum, and F. M. Nortier. Cross sections from 800 MeV proton irradiation of terbium. 2012. NUCLEAR PHYSICS A. 893: 87. Engle, J. W., S. G. Mashnik, J. W. Weidner, L. E. Wolfsberg, M. E. Fassbender, Jackman, Couture, L. J. Bitteker, J. L. Ullmann, M. S. Gulley, Pillai, K. D. John, E. R. Birnbaum, and F. M. Nortier. Cross sections from proton irradia- tion of thorium at 800 MeV. 2013. PHYSICAL REVIEW C. 88 (1). Engle, J. W., S. G. Mashnik, J. W. Weidner, L. E. Wolfsberg, M. E. Fassbender, K. Jackman, A. Couture, L. J. Bitteker, J. L. Ullmann, M. S. Gulley, C. Pillai, K. D. John, E. R. Birnbaum, and F. M. Nortier. Cross sections from pro- ton irradiation of thorium at 800 MeV. 2013. PHYSICAL REVIEW C. 88 (1): -. Fassbender, M. E., Ballard, E. R. Birnbaum, J. W. Engle, K. D. John, J. R. Maassen, F. M. Nortier, J. W. Lenz, C. S. Cut- ler, A. R. Ketring, S. S. Jurisson, and D. S. Wilbur. Proton irradiation parameters and chemical separation pro- cedure for the bulk production of high-specific-activity Re-186g using WO3 targets. 2013. RADIOCHIMICA ACTA. 101 (5): 339. Gott, M. D., B. D. Ballard, L. N. Redman, J. R. Maassen, W. A. Taylor, J. W. Engle, F. M. Nortier, E. R. Birnbaum, K. D. John, D. S. Wilbur, C. S. Cutler, A. R. Ketring, S. S. Jurisson, and M. E. Fassbender. Radiochemical Study of Re/W Adsorption Behavior on a Strongly Basic Anion Exchange Resin. 2014. RADIOCHIMICA ACTA. 102 (4): 325. Jackman, K. R., J. W. Engle, F. M. Nortier, K. D. John, E. R. Birnbaum, and D. E. Norman. Synthetic spectra for 154

Page 155

Complex Natural & Engineered Systems Postdoctoral Research and Development Final Report Hybrid Nanostructures for Photoreduction of CO to Hydrocarbons 2 Hongwu Xu 20130787PRD2 Abstract To develop efficient photocatalytic systems for conver- High-quality ZnTe/CdS core/shell nanorods and ZnTe/ sion of CO2 into fuels, three major challenges need to be CdS-Pt hybrid nanostructures have been successfully addressed: 1) absorption of the visible part of the solar synthesized and characterized. The growth of CdS shell spectrum, 2) design of structures that facilitate transfer on ZnTe core forms a type-II alignment in terms of the processes of multiple electrons and hydrogen atom with relative conduction/valence band edge of the core/shell. sufficient energy for reduction of CO2 and oxidation of This also significantly enhances the stability of ZnTe H2O, and 3) design of structures that promote adsorp- nanorod cores, while maintaining their long charge car- tion of CO2 and water close to the source where the rier lifetimes, ultrafast charge separation dynamics, and photo-generated electrons and holes are formed. The extended absorption spectra. Our results support the objectives of this project are to design, synthesize and potentiality of utilizing ZnTe-based hybrid nanomaterials characterize multi-component, composite nanostruc- for photocatalysis and solar energy conversion applica- tures, particularly core/shell structures, with controlled tions including CO2 photoreduction. nanoparticle size, shape and composition, allowing exploration of their potential photocatalytic applications Background and Research Objectives including CO2 reduction. Photocatalytic conversion of CO2 into fuels or use- Scientific Approach and Accomplishments ful chemicals is a promising approach to both reduc- ing atmospheric CO2 level and fulfilling the increasing As a critical step towards photoreduction of CO2 to hy- energy demands for hydrocarbon-based fuels. Although drocarbons, we design/synthesize hybrid nanostructures extensive efforts have been made on development of of photoactive materials that have both effectively sepa- semiconductor-based nanomaterials for solar-to-fuel rated charges and catalytic sites for chemical conversion. conversion, the efficiencies achieved so far are too low Semiconductor core/shell nanostructures are used for for practical applications [1]. Previous studies suggest initial photo-induced charge separation, while the metal- that single-electron reduction of CO2 is thermodynami- lic components provide catalytic sites for subsequent cally unfavorable. While the proton-coupled, multiple- chemical conversion. Compared to many other semi- electron reduction is more favorable, it is a multi-step conductors, ZnTe has a more negative conduction band process requiring transfer of multiple electrons, which is (-1.9 V) while still having a suitable band gap for visible kinetically hindered. For example, conversion of CO2 to light absorption. Thus ZnTe has a larger driving force for CH4 needs four steps involving transfer of eight elec- electron transfer, which may facilitate photoreduction of trons, which leads to high-overpotential reduction of CO2 [3]. However, ZnTe nanoparticles, while stable un- CO2 to hydrocarbons and alcohols. There are also other der reducing conditions, often undergo light-induced an- factors that limit photocatalytic efficiencies, including odic oxidation in aqueous solutions. Hence, coating their inefficient absorption of solar energy, fast recombina- surfaces with shells of a more stable phase, e.g. CdS, is tion of photo-excited electron-hole pairs, and backward a plausible approach to protecting ZnTe nanoparticles redox reactions. A major challenge lies in the fact that from dissolution during the charge separation process. the performance of energy materials/systems depends Moreover, these core/shell nanostructures can facilitate on their multiple properties, such as redox activity, light charge separation by forming a barrier to control back absorption, mechanical property and stability. However, electron transfer. Adding a co-catalyst, such Pt to the it is difficult to find simple, single-phase materials that ZnTe/CdS core/shell, can further help accumulate the satisfy all the criteria [2]. photo-induced electrons and facilitate CO2 conversion. 155

Page 156

ZnTe nanorods were synthesized using an organic-solution ultraviolet-visible (UV-vis) absorption spectra of ZnTe/CdSe method [4]. In a typical synthesis, first, 0.13 g of zinc ace- nanorods and the corresponding ZnTe seeds are shown in tate was mixed with 10 mL of octadecene and 2 mL of oleic Figure 3. The bare ZnTe nanorods exhibit the first exciton acid in a 50 mL three-neck flask. The mixture was heated absorption peak at 500 nm. By contrast, ZnTe/CdSe core/ to 200 °C under a N2 gas flow through a Schlenk-line, form- shell rods have the corresponding absorption peak at 600 ing a clear Zn-containing solution. The temperature was nm. This red shift is consistent with the occurrence of CdS then decreased to 160 °C by cooling the flask. Second, a shell on ZnTe core, which form a type-II nanostructure. Te precursor was prepared by mixing Te-trioctylphosphine (1M for Te) solution (0.5 mL), 0.7 M superhydride solution (0.7 mL) and 2 mL of oleylamine in a vial in a glovebox. Third, this precursor was transferred into a 10 mL plastic syringe and was directly injected into the Zn solution. Then the temperature was increased to 300 °C with a rate of about 10 °C/min. The resulting product was separated by adding 10-20 mL of ethanol followed by centrifugation. The isolated nanorods were re-dispersed in 10 mL of hex- ane, producing ZnTe colloidal suspensions. Figure 1A is a representative transmission electron microscope (TEM) im- Figure 2. (A) HRTEM and (B) powder XRD pattern of ZnTe/CdS age of the sample, showing uniform nanorod morphology. core/shell nanorods. Green and red lines denote the reflections of wurzite ZnTe and CdS, respectively. Figure 1. TEM images of (A) ZnTe nanorods, (B) ZnTe/CdS core/ shell nanorods and (C) ZnTe/CdS-Pt hybrid nanostructures. ZnTe/CdS core/shell nanorods were synthesized via suc- cessive growth of CdS on ZnTe nanorods. In a typical synthesis, 5 mL of ZnTe nanorod colloidal solution, 10 mL of octadecene, 0.5 mL of sulfur-oleylamine (1M) and 0.2 mL of trioctylphosphine were mixed in a flask attached to the Schlenk-line and slowly heated to 170 °C under an Ar gas flow. 1 mL of Cd precursor solution (0.1 M) was added Figure 3. UV-vis absorption spectra of ZnTe nanorods and ZnTe/ dropwise into the flask during a period of 10 min. The CdS core/shell nanorods. reaction system was kept at 170 °C for an additional 10 min Transient absorption (TA) spectra of ZnTe/CdSe quantum allowing for CdS shell growth. The product was separated dots (QDs) at different delay times after excitation at 400 by using the same procedure as for bare ZnTe nanorods nm are shown in Figure 4A, and their energy level diagram and was re-dissolved in 10 mL hexane or chloroform. TEM is shown in Figure 4B. There are three possible electronic imaging of ZnTe/CdS core/shell nanorods reveals that after transitions: T1, T2 and T3 (Fig. 4B), which are responsible CdS coating, ZnTe cores maintain their original rod-like for absorption bands 1, 2 and 3 (Fig. 4A), respectively. T1 morphology (Fig. 1B). Because of the small lattice mis- is the transition from the lowest energy valence-band (VB) match between ZnTe and CdS (about 4.4%), the growth of 1sh level in ZnTe to a delocalized conduction-band (CB) CdS overlayer on ZnTe is considered to be epitaxial; high- electron level in ZnTe. T2 is the spatially indirect transition resolution TEM image (Fig. 2A) of ZnTe/CdS core/shell [see from the VB 1sh level in ZnTe to the lowest energy CB 1se powder X-ray diffraction (XRD) pattern in Fig. 2B] does not level in CdS. T3 is the transition from the lowest energy reveal noticeable interfaces between ZnTe and CdS. With delocalized VB hole level to the CB 1se level in CdS. The the addition of CdS shells, the stability of ZnTe was found correlation among the kinetics at T1, T2 and T3 transi- to be dramatically enhanced relative to that of bare ZnTe tions indicates the presence of an internal charge separa- nanorods, with a shelf life of at least several months. The tion process from the ZnTe core to the CdSe shell. The TA 156

Page 157

results indicate attractive properties of ZnTe/CdS core/shell results demonstrate that the CdSe core and ZnSe shell nanorods. First, the ZnTe/CdS nanorods maintain the ad- behave independently under high pressure, despite that vantageous properties of the type-II nanostructures. They stresses may exist at the interface due to the CdSe/ZnSe have a longer charge carrier lifetime compared with that of lattice mismatch. Namely, consistent with precious stud- bare ZnTe nanorods, ultrafast charge separation dynamics, ies [6], the wurtzite CdSe core transforms to the rock salt and red-extended absorption spectra. Second, the ZnTe/ structure above about 5.6 GPa and then to the zincblende CdS nanorods have a higher reductive potential compared structure upon decompression to ambient condition (Fig. to other type II heterostructures, making a promising 5). As expected for ZnSe, which has a phase transition at photocatalytic system for reduction of CO2 under photo- about 13 GPa [7], the shell is unaffected by our experiment illumination. (10.7 GPa). This post-synthesis integrity of core/shell CdSe/ ZnSe further supports our hypothesis that surface-related processes during shelling are key to effecting shell vs. alloy formation [8]. Figure 4. (A) TA spectra of ZnTe/CdS core/shell nanorods at different delay times after 400 nm excitation. B) The energy level diagram of ZnTe/CdS core/shell nanorods. The arrows indicate three possible electronic transitions (T1, T2 and T3). CB - conduc- tion band; VB - valence band. Previous studies have demonstrated that adding a metal tip to semiconductor nanocrystals will enhance charge separation and transfer properties at metal-semiconductor interfaces. In addition, the metal tip may also serve as a catalytic site for photoreactions. To enhance the photo- induced charge separation and transfer properties of ZnTe/ CdS nanorods, we synthesized Pt decorated ZnTe/CdS nanorods following a previously developed method [5]. A typical reaction is as follows: 0.015 g of Pt acetylacetonate was added to 5 mL of dichlorobenzene along with 0.5 mL of oleylamine, 0.5 mL of oleic acid and 20 mg of 1,2-hexa- decanediol. The mixture was heated at 70 °C for 10 min under an Ar flow, forming a Pt precursor solution. The Figure 5. XRD patterns of CdSe/ZnSe core/shell quantum dots freshly prepared Pt precursor was introduced into ZnTe/ under different pressures. Rock salt (200) reflection indicated by CdS nanorods colloidal solution in diphenyl ether (10 mL) asterisk. CdSe and ZnSe wurtzite reference patterns are green that was pre-heated to 200 °C. After several minutes, the and red, respectively. reaction was stopped by removal of the heat. The product was separated by adding excessive ethanol followed by Impact on National Missions centrifugation. TEM image shows that almost all ZnTe/CdS nanorods have been decorated with Pt nanoparticles (Fig. Through controllable synthesis and systematic character- 1C). ization of core/shell nanomaterials, we have shown poten- tial utilization of these materials for practical applications In addition to the ZnTe/CdS core/shell nanoparticles, we including efficient photocatalytic conversion of CO2 plus investigated the stability of CdSe/ZnSe core/shell QDs at H2O into hydrocarbon fuels and chemicals. The results will high pressures up to 10.7 GPa using in-situ synchrotron lay the foundation for the ultimate development of solar- XRD coupled with diamond-anvil-cell (DAC) technique. to-fuel conversion devices and have broad implications for Core/thick-shell giant quantum dots (gQDs) possessing rational design of functional nanomaterials. type II electronic structures exhibit suppressed blinking and diminished nonradiative Auger recombination. Our 157

Page 158

References

1. Roy, S. C., O. K. Varghese, Paulose, and C. A. Grimes. Toward Solar Fuels: Photocatalytic Conversion of Car- bon Dioxide to Hydrocarbons. 2010. ACS NANO. 4 (3):
3. Amirav, , and A. P. Alivisatos. Photocatalytic Hydrogen Production with Tunable Nanorod Heterostructures.
4. JOURNAL OF PHYSICAL CHEMISTRY LETTERS. 1 (7): 1051.
5. Jin, , J. u. n. Zhang, R. D. Schaller, Rajh, and G. P. Wiederrecht. Ultrafast Charge Separation from Highly Reductive ZnTe/CdSe Type II Quantum Dots. 2012. JOURNAL OF PHYSICAL CHEMISTRY LETTERS. 3 (15):
7. Zhang, J., S. Jin, H. C. Fry, S. Peng, E. Shevchenko, G. P. Wiederrecht, and T. Rajh. Synthesis and Character- ization of Wurtzite ZnTe Nanorods with Controllable Aspect Ratios. 2011. JOURNAL OF THE AMERICAN CHEMICAL SOCIETY. 133 (39): 15324.
8. Habas, S. E., Yang, and Mokari. Selective growth of metal and binary metal tips on CdS nanorods. 2008. JOURNAL OF THE AMERICAN CHEMICAL SOCIETY. 130 (11): 3294.
9. Alivisatos, A. P.. Perspectives on the physical chemistry of semiconductor nanocrystals. 1996. JOURNAL OF PHYSICAL CHEMISTRY. 100 (31): 13226.
10. KOFFERLEIN, M., H. KARZEL, W. POTZEL, W. SCHIESSL, M. STEINER, G. M. KALVIUS, D. W. MITCHELL, and T. P. DAS. HIGH-PRESSURE PHASE-TRANSITION IN ZNSE.
11. HYPERFINE INTERACTIONS. 93 (1-4): 1505.
12. Acharya, K. P., H. M. Nguyen , M. Paulite , A. Piryatin- ski, J. Zhang, J. L. Casson , H. Xu, H. Htoon, and J. A. Hollingsworth. Elucidation of two giants: challenges to thick-shell synthesis in CdSe/ZnSe and ZnSe/CdS core/ shell quantum dots . 2015. Journal of the American Chemical Society. : DOI: 10.1021/jacs.5b00313. Publications Acharya, K. P., H. M. Nguyen , M. Paulite , A. Piryatinski, J. Zhang, J. L. Casson , H. Xu, H. Htoon, and J. A. Holling- sworth. Elucidation of two giants: challenges to thick- shell synthesis in CdSe/ZnSe and ZnSe/CdS core/shell quantum dots . 2015. Journal of the American Chemi- cal Society. : DOI: 10.1021/jacs.5b00313. 158

Page 159

Complex Natural & Engineered Systems Postdoctoral Research and Development Final Report Joint Inversions of Seismic and Gravity Data in Volcanic Areas to Advance Hazards Assessment: A Focus on the Alaskan Subduction Zone and Kilauea, Hawaii Monica Maceira 20130807PRD3 Abstract nity to study one of Earth’s most active volcanoes, in a Alaska and Hawaii present significant seismic and location that has produced historical M>7 tsunamigenic volcanic hazards to the U.S. By studying the causes of earthquakes. While previous seismic studies have fo- these hazards and the environments in which they occur, cused on studying the near-surface beneath the summit we can work toward better characterizing the hazards caldera, portions of Kilauea’s rift zones, or the seismi- they present. Three-dimensional passive-source seismic cally active tsunamigenic decollement fault beneath the velocity tomography is a powerful imaging technique volcanic material and the underlying oceanic crust, the that utilizes seismic waves produced by earthquakes to extents of these studies are limited by the natural seis- image complex subsurface structures, such as subduct- mic distributions. By combining gravity and seismic data, ing slabs or magma distributions beneath volcanoes. The we can address questions such as 1) What volume of resolutions of these studies are generally limited by the magma is stored beneath the summit and rift zones, and natural distribution of earthquakes. Gravity data, how- has the potential to erupt? 2) Can the aseismic and seis- ever, provide alternate sources of information regarding mic portions of the decollement be imaged, constraining subsurface structures, and are available at a variety of the maximum magnitude of earthquake this fault can scales. Due to the inherent relationship between density produce? 3) How does neighboring Mauna Loa Volcano and seismic velocity, gravity and seismic data can be contribute to the structure of Kilauea? jointly inverted, leading to the characterization of Earth Scientific Approach and Accomplishments structures that is not limited by the distributions of seis- micity. Joint inversion also places stronger constraints During her Director’s Postdoctoral Fellowship, Ellen on the temperature, compositional, fluid, and magmatic focused on studying Akutan and Makushin volcanoes in distributions. Alaska, the Colombian subduction zone (as a new data- set became available), and Kilauea Volcano. Her work Background and Research Objectives focused in two areas: 1) algorithm and methodology The Alaskan subduction zone provides an ideal location improvement and 2) better understanding of the tecton- to explore the geophysical expressions of volcanism, ics of these regions. including regions of elevated temperature, fluid content, Algorithm improvements include the extension of the in- and melting in the earth’s mantle and crust. The cur- version method to Cartesian (as opposed to geographic) rent distribution of seismic stations is focused near the coordinates; the addition of multiple optional types of relatively two-dimensional volcanic arc located above regularization and smoothing; altering how gravity par- the subducting oceanic plate, limiting the ability to tial derivatives are treated; reducing the computational resolve these anomalies. The inclusion of gravity data time; and the inclusion of topography and its effects on in a joint inversion will markedly increase the ability to gravity predictions. Outside of the algorithm itself, El- resolve these features in three dimensions. In this study len developed an objective method of determining the we addressed questions regarding the sources of vol- optimal combination of regularization parameters and canism such as 1) How are anomalies associated with relative weights for different datasets. Several of these magma genesis spatially related to volcanic centers? 2) improvements are particularly important for applications Can the cause(s) of these anomalies be independently to more local-scale problems, which is a relatively new constrained? direction for this joint-inversion algorithm. Kilauea Volcano, Hawaii, presents a unique opportu- 159

Page 160

In her work with Akutan and Makushin Volcanoes, Ellen nique through comparison with independent constraints used body and surface waves to develop a 3D P-wave and on earthquake locations, as well as to collaborate more S-wave velocity model for the two volcanoes and improve closely with several members of the ground-based nuclear earthquake locations. The body wave measurements explosion monitoring team in EES-17. and continuous seismic data are from the Alaska Volcano Ellen has also been working with a fellow postdoc (now Observatory, and she developed the surface wave dataset staff member) at the Lab in his work to further improve from ambient noise analysis of the continuous seismic computational and regularization aspects of joint inver- data. The combined analysis of the velocity structure sions, specifically to improve the recovery of sharp struc- at each volcano and the distributions of seismicity have tural boundaries, which tend to be smoothed out in most provided insight to near-surface faulting, magma path- inversions. This work has been published in Geophysical ways to the surface, and regions of the long-term storage Journal International. She has also worked with two sum- of magma. In comparison to other nearby volcanoes, this mer students, one with using body-wave seismic data from work demonstrates the breadth of magmatic structures at Bhutan (summer 2014) and one with using surface-wave different volcanoes, even those that share many similari- seismic data from Antarctica (summer 2015). ties. This work was published in March 2015 in the Journal of Geophysical Research, and it has been presented at Outside of her own research, Ellen has participated in several meetings. several activities at the Lab and in the geosciences com- munity. In October 2015, she will give an invited talk at the Her work on Colombia demonstrates for the first time GeoPRISMS Theoretical and Experimental Institute for the the benefits of including gravity data in a joint inversion Subduction Cycles & Deformation Meeting on her work with body waves and surface waves. Using data from the at Akutan and Makushin Volcanoes. In September 2014, national Colombian seismic network, Ellen refined the she had an invited webinar for the Incorporated Research body wave dataset and measured surface waves from local Institutions for Seismology on seismic and geodynamic earthquakes. Gravity data are from a global satellite-based constraints on subduction zone structure and volcanism. model and constrain velocities (via densities) in the shal- In April 2014, she gave a similar invited colloquium talk at low subsurface. The structure of the Colombian subduction the University of Minnesota. Ellen convened a session with zone has been a longstanding topic of debate, with several two external collaborators at the American Geophysical proposed models. The results of her study provide evi- Union 2013 Fall Meeting, entitled ‘Geophysical Observa- dence for a tear in the subducting tectonic plate in central tions and Models of Subduction’, and she chaired a session Colombia, with a boundary between two neighboring at the 2014 Fall Meeting. Ellen has reviewed proposals for plates at the Bucaramanga nest, a small region of some of the National Science Foundation, as well as manuscripts the most intense seismicity globally. Ellen is currently fin- for Nature Geoscience, Geology, Journal of Geophysical ishing up a manuscript of this work for submission to the Research, and Geochemistry Geophysics Geosystems. In journal Geochemistry Geophysics Geosystems in the next March 2014, she participated in a discussion panel at the month or so. She has also presented this work at several Lab as a part of Women’s History Month, and she has been national and international meetings. the postdoc representative to the EES Worker Safety and Ellen’s work at Kilauea Volcano builds upon her previous Security Team since summer 2014. research in the area. Here, the joint inversion of body wave, surface wave, and gravity data is complicated by the Impact on National Missions significant topography in the area and its effects on gravity The research discussed under this project will enhance at the local scale. She worked to better understand the the capability of the Laboratory through unique expertise magma plumbing system of the volcano, particularly how in imaging Earth structure. This expertise has applications deeper magma reservoirs are connected to eruptions at to Nonproliferation R&D under Nuclear Nonproliferation, the surface at the summit caldera and Southwest and East where accurate Earth models are needed to locate, iden- Rift Zones. This volcano poses both eruptive and tsunami tify and determine yield for seismic events of interest such hazards to the area. as underground nuclear explosions. The relationship to assessing hazards in volcanic zones could be of interest to Starting this last 2015 summer, Ellen also began applying DHS/FEMA as they are responsible for domestic prepared- the joint inversion technique to Iran, in order to begin her ness and response regarding, among other things, natural transition from a Director’s Postdoc to a programmatic disasters, and perhaps to the Department of the Interior’s postdoc. This region is of interest both tectonically and for USGS. High-resolution Earth models are also required global security, and working on a velocity model for the for characterization and monitoring within several other region will allow her to validate the predictions of the tech- 160

Page 161

areas of LANL mission space including geothermal energy American Geophysical Union Fall Meeting. (San Fran- development (Applied Energy Programs Office) carbon cisco, 15-19 Dec., 2014). sequestration (Fossil Energy, within the Office of Science Syracuse, E. M., M. Maceira , G. A. Prieto , H. Zhang, and Programs), and used fuel disposition and salt repository C. J. Ammon . Joint inversion of seismic and gravity science (both within the office of Civilian Nuclear Pro- data for velocity structure and hypocentral locations of grams), as well as the new DOE initiative SubTER in which the Colombian subduction zone. Presented at Seismo- LANL – and in particular EES-17 – is participating. logical Society of America Annual Meeting. (Pasadena, 21-23 April, 2015). Publications Holt, R. A., M. K. Savage , J. Townend , E. M. Syracuse , Syracuse, E. M., M. Maceira , E. Bergman , W. S. Phillips, M. and C. H. Thurber . Crustal stress and fault strength Begnaud , and H. Zhang. Velocity structure of the Iran in the Canterbury Plains, New Zealand. Presented at region using seismic and gravity observations. Invited American Geophysical Union Fall Meeting. (San Fran- presentation at Comprehensive Test Ban Treaty: Sci- cisco , 9-13 Dec. 2013). ence and Technology 2015 Conference. (Vienna, Aus- tria, 22-26 June, 2015). Holt, R. A., M. K. Savage, Townend, E. M. Syracuse, and C. H. Thurber. Crustal stress and fault strength in the Can- Syracuse, E. M., M. Maceira, G. A. Prieto , H. Zhang, and C. terbury Plains, New Zealand. 2013. EARTH AND PLAN- J. Ammon . Joint inversion of seismic and gravity data ETARY SCIENCE LETTERS. 383: 173. for velocity structure and hypocentral locations of the Colombian subduction zone. Presented at International Lemon, S., C. H. Thurber, M. M. Haney, E. Syracuse, H. Union of Geodesy and Geophysics General Assembly. Zhang , X. Zeng , and S. G. Prejean. Ambient noise (Prague, Czech Republic, 22 June - 2 July, 2015). tomography of the Katmai Volcanic Complex, Alaska. Presented at Seismological Society of America Annual Syracuse, E. M., M. Maceira, G. A. Prieto, H. Zhang, and C. Meeting . (Anchorage, 30 April - 2 May, 2014). J. Ammon. Multiple plates subducting beneath Colom- bia: Results of the joint inversion of seismic and gravity Lin, Y., E. Syracuse, M. Maceira, C. Larmat, and H. Zhang. data. Geochemistry Geophysics Geosystems. Joint inversion for Kilauea volcano with an edge-pre- serving constraint. 2014. In Seismological Society of Syracuse, E., M. Maceira, H. Zhang, and C. Thurber. Seis- America Annual Meeting. (Anchorage, 29 April - 2 May, micity and structure of Akutan and Makushin Volca- 2014). Vol. 85, p. 502. El Cerrito: BSSA. noes, Alaska, using joint body- and surface-wave to- mography. 2015. Journal of Geophysical Research. 120 Lin, Y., E. Syracuse, M. Maceira, H. Zhang, and C. Larmat. (2): 1036. Double-difference travel-time tomography with edge- preserving regularization and a priori interfaces. 2015. Syracuse, E., M. Maceira, and H. Zhang. Joint inversion Geophysical Journal International. 201: 574. of seismic and gravity data for velocity structure and hypocentral locations at Akutan and Makushin volca- Segall, , A. L. Llenos, S. -. Yun, A. M. Bradley, and E. M. noes. 2014. In Seismological Society of America Annual Syracuse. Time-dependent dike propagation from joint Meeting. (Anchorage, 30 April - 2 May, 2014). Vol. 85, inversion of seismicity and deformation data. 2013. p. 500. El Cerrito, California: BSSA. JOURNAL OF GEOPHYSICAL RESEARCH-SOLID EARTH. 118 (11): 5785. Syracuse, E. M., J. D. Pesicek, H. Zhang , and C. H. Thurber . Exploring the connection between intermediate- depth seismicity, slab hydration, and dehydration. Pre- sented at American Geophysical Union Fall Meeting. (San Francisco, 9-13 Dec., 2013). Syracuse, E. M., M. Maceira , H. Zhang , C. Larmat, C. Ammon, and C. Chai. Advanced Multivariate Inversion Techniques. Invited presentation at Review of Monitor- ing Research. (Albuquerque, 17-19 June, 2014). Syracuse, E. M., M. Maceira , G. A. Prieto , H. Zhang , and C. J. Ammon . Joint inversion of seismic and grav- ity data for velocity structure and hypocentral loca- tions of the Colombian subduction zone. Presented at 161

Page 162

Complex Natural & Engineered Systems Postdoctoral Research and Development Final Report Discovery of Novel Bioactive Natural Products Alexander Koglin 20130815PRD4 Abstract All enzymes that are predicted to be in involved in the Natural evolution designed highly efficient enzymatic biosynthesis of these novel natural products have been systems with an incredible and often enough not re- cloned into E. coli expression systems. Among those producible chemical diversity. Secondary metabolites, novel compounds are two completely new siderophores. natural products with bioactive properties, represent the Siderophores are metal, in this case primarily iron, ion class of compounds with the highest diversity in their chelators produced for bacterial iron homeostasis and it chemical structures, functional groups and properties. has been also demonstrated that siderophores are viru- This class contains often heavily alternated non-ribosom- lence markers of pathogenic strains and can contribute al peptides; ketides or acyl-derived small molecules and to the path and severity of infections. combinations thereof. The diverse biological functions Background and Research Objectives include the vast majority of antibacterial, antifungal, an- tiviral, anti-cancer, immunosuppressant, analgesic, signal In this project, available genomic DNA data of soil and transducing, energy-storage compounds and toxins. A marine bacteria will be exploited in a multidisciplinary variety of tailoring enzymes associated with the biosyn- approach to allow fast natural product discovery and thetic assembly lines define the chemical structure of all characterization. The research employs five key steps. products and can tremendously modify these step-by- Genome Analysis step synthesized molecules. Bioinformatics analysis of genomes of selected bacte- As DNA sequences encode ribosomal peptides and ria (e.g. Myxococcus xanthus and Streptosporangium resulting protein functions, the order and assembly of roseum) will identify natural product biosynthetic ma- enzymes in biosynthetic clusters responsible for second- chineries and predict product building blocks (e.g. amino ary metabolite synthesis define the chemical structure acids, sugars). of secondary metabolites. We investigate the genetic Isolation of Predicted Natural Products encoding and organization of enzymatic functions in The bacterial strains will be cultivated, isotopically en- newly discovered assembly lines to predict the chemical riched (e.g. 13C, 15N, 1H or 3H, 14C) by feeding labeled composition of those natural products. We will isolate substrates, and the natural products extracted from the those compounds and characterize their bioactivity for culture. Incorporated isotopes will readily identify pro- the development of drug candidates that are urgently posed compounds by NMR and high-res mass spec (MS) needed to counter emerging threats to public health for stable isotopes, or by β- or γ-detection for radioac- caused by bacteria, viruses, fungi and human cell alter- tive isotopes, after chromatographic separation. The ing events. availability of high-field NMR and excellent MS resources This project is focused on the prediction, identifica- will uniquely enable the project at LANL. tion and characterization of a series of novel secondary Structure Elucidation metabolites. Cultures of strains listed below have been Once separation from the culture is achieved, large scale established in the laboratory to isolate some of these culturing without labels will follow. NMR, elemental predicted natural products. These strains include: Strep- analysis, high-res MS, and x-ray SAS will be applied to tosporangium roseum, Myxococcus xanthus, Chlami- elucidate structures of newly identified compounds. domonas acidophila, Herpetosiphon aurantiacus, Clos- tridium thermocellum and Desulvovibrio aerotolerans. 162

Page 163

Validation of the Biosynthetic Machinery found. From this vast amount of available strains, five were Isolated genes of putative clusters will be transformed into chosen (Myxococcus xanthus DK 1622, Corallococcus coral- E. coli strains, expressed and biochemically characterized loides DSM 2259, Streptosporangium roseum DSM 43021, in vitro to validate the underlying biosynthesis of newly Catenulispora acidiphila DSM 44928, Herpetosiphon isolated compounds. Also, the entire biosynthetic cluster aurantiacus DSM 785) in which the biosynthetic clusters can be cloned into an overproduction bacterial strain, are unusual, i.e. not following the classical linear, multi- where high expression of the biosynthetic machinery will modular organization2. This approach significantly reduces make the natural product a major component of the cells. the probability of isolating compounds similar to known or In case the transgenic natural products are toxic for E. coli, identified natural products. Within these five strains a total a cell-free translation/transcription system is available and of 124 gene clusters have been identified. Bioinformatics will be employed. analysis revealed that 25 of these clusters showed unusual organization and hence are likely to yield poteniallt novel Bioactivity Studies bioactive natural products. Subsequently, these clusters The bioactivity of newly isolated natural products and their were analyzed in-depth using additional bioinformatic potency will then be evaluated in vitro by incubation with tools, such as pfam3, nrpspredictor24 and 2metdb5 to either bacterial, fungal, human cell lines or viral systems. facilitate structure and building block prediction of the This will give insights into the natural product’s properties natural products. and potency. The practical goals were to establish growth of five differ- Meeting the great demand for novel bioactive agents ent bacterial strains and the isolation as well as structural requires improved discovery processes of bioactive com- characterization of three novel natural products produced pounds. It is clear that microbes continue as a major by these strains, facilitated by feeding isotopically en- source of new drug leads. With short evolutionary cycles riched, predicted building blocks. To meet these goals, the and natural pressure to protect their biological niches, five above mentioned strains were successfully cultured soil and marine microbes are prime targets for identifica- under laboratory conditions. In addition to the bioinfor- tion of novel compounds. This project couples advances matics prediction of the amino acid building blocks, it is in DNA sequencing and bioinformatics with biochemical important to biochemically correlate the identified gene and labeling techniques, structure elucidation methods clusters to the 25 predicted natural products by cloning and advanced knowledge in natural product biosynthesis. and expression of the amino acid building block activation High level expertise in these areas, combined with Stable enzymes. This gives valuable insights into natural product Isotope Resource, give LANL potential to be an outstanding assembly mechanisms and allows estimation of how the contributor to this field. This strategy will give further in- predictions match the actual substrate building blocks. In sights into nature’s way to produce bioactive compounds, total 197 genes and gene fragments within the 25 clusters significantly shorten the time span for natural product were annotated to be involved in building block (amino discovery, and consequently speed up and reduce costs for acid) activation. The 92 essential of these genes have drug development. It will guide future developments of been successfully cloned and the generated expression accurate prediction of natural product structures exclu- plasmids were used to transform E. coli BL21 (DE3*) cells sively based on genomic DNA information. This supports to yield expression strains for the building block activating the mission of Bioscience Division in biosecurity and LANL’s enzymes. In initial test expression studies using 23 of these mission in threat reduction and global security. heterologous strains protein production was observed. In the subsequent large scale expressions, soluble protein Scientific Approach and Accomplishments was obtained in sufficient amount for biochemical charac- For the bioinformatic part of the project, it was proposed terization. By carrying out photometric assays6 with the to identify at least 10 new gene clusters encoding for 20 natural amino acids for each individual building block biosynthetic machineries for natural product produc- activating enzyme, it was possible to confirm the activation tion. Instead, we used a more sophisticated approach by of the predicted building blocks in 90% of the cases. systematically analyzing all available microbial genomes deposited in the NCBI database. The analysis was carried The feeding experiments for fast natural products isola- out using Perl-scripts developed by Alex Koglin and was tion are currently limited to stable isotope labels (13C), a compared to an adapted and heavily modified stand-alone radiological permit for the faster 14C-feeding experiments version of the publically available bioinformatic tool, An- was not provided in time. Practically, this means the com- tiSMASH1. Within the analyzed 2,773 genomes of bacte- pounds within the cell/supernatant extracts are separated rial strains, more than 19,000 of these gene clusters were by liquid chromatography (HPLC), the individual fractions 163

Page 164

are freeze-dried, the yielded solids are analyzed by NMR References (Nuclear Magnetic Resonance) spectroscopy for 13C-car- 1. Blin, K. a. i., M. H. Medema, Kazempour, M. A. Fisch- bon enrichment and hence the proposed natural product. bach, Breitling, Takano, and Weber. antiSMASH 2.0-a For one bacterial strain this whole process can take up versatile platform for genome mining of secondary me- to four months for one novel natural product. Isotopic tabolite producers. 2013. NUCLEIC ACIDS RESEARCH. enriched (13C)-cysteine was fed to Streptosporangium 41 (W1): W204. roseum in order to facilitate isolation of a natural product that was predicted to be composed of three cyclized cys- 2. Strieker, , Tanovic, and M. A. Marahiel. Nonribosomal teines. From analysis of individual fractions by the proce- peptide synthetases: structures and dynamics. 2010. dure stated above, isolation conditions for that predicted CURRENT OPINION IN STRUCTURAL BIOLOGY. 20 (2): natural product were established, large scale isolation 234. without label was carried out and the structure was eluci- 3. Finn, R. D., Bateman, Clements, Coggill, R. Y. Eberhardt, dated by NMR. The isolated natural product showed three S. R. Eddy, Heger, Hetherington, Holm, Mistry, E. L. cyclized cysteines and an additional aromatic amino acid. L. Sonnhammer, Tate, and Punta. Pfam: the protein We predicted structural similarity to pyochelin, a thiazole- families database. 2014. NUCLEIC ACIDS RESEARCH. 42 siderphore that is produced by several pathogenic strains (D1): D222. including Pseudomonas syringae, Pseudomonas aerugino- sa, and several B. pseudomallei strains and yersiniabactin 4. Roettig, , M. H. Medema, K. a. i. Blin, Weber, Rausch, the siderophore and virulence factor produced by Yersinia and Kohlbacher. NRPSpredictor2-a web server for pestis. We demonstrated that the determined structure predicting NRPS adenylation domain specificity. 2011. of this secondary metabolite is as predicted. We further NUCLEIC ACIDS RESEARCH. 39: W362. identified and characterized all enzymes that are involved in the biosynthesis of this novel siderophore in vitro. Due 5. Bachmann, B. O., and Ravel. METHODS FOR IN SILICO to the delay in obtaining the radiological permit, only two PREDICTION OF MICROBIAL POLYKETIDE AND NONRI- of the targeted three natural products has been structur- BOSOMAL PEPTIDE BIOSYNTHETIC PATHWAYS FROM ally characterized. These entirely novel siderophores, or DNA SEQUENCE DATA. 2009. COMPLEX ENZYMES IN iron chelators, demonstrate the strength to predict chemi- MICROBIAL NATURAL PRODUCT BIOSYNTHESIS, PART cal structures of natural products for a directed search for A: OVERVIEW ARTICLES AND PEPTIDES. 458: 181. novel bioactive compounds that assemble the pool for our drugs, known toxins and virulence factors. Both new struc- 6. Wilson, D. J., and C. C. Aldrich. A continuous kinetic tures and their biochemical characterization have been assay for adenylation enzyme activity and inhibition. submitted for publication. 2010. ANALYTICAL BIOCHEMISTRY. 404 (1): 56. Impact on National Missions Publications This project supports the LANL public health mission Inglis, S. R., Strieker, A. M. Rydzik, Dessen, and C. J. Scho- (Homeland Security Act 2002, DOE Strategic Plan 2011, field. A boronic-acid-based probe for fluorescence polarization assays with penicillin binding proteins and Public Health Security and Bioterrorism Preparedness and beta-lactamases. 2012. ANALYTICAL BIOCHEMISTRY. Response Act 2002) as well as its mission in global threat 420 (1): 41. reduction. It also has direct appeal to our WFO spon- Koglin, A., and M. Strieker. Biosynthesis of two novel anti- sors and the base knowledge acquired can be applied to bacterial compounds. Nature Chemical Biology. increasing efficiency of novel product synthesis to make it more cost effective. It will both leverage and strengthen Strieker, M., J. Schmidt, and A. Koglin. Identification and our capabilities in bioinformatics computing, genomic/ characterization of an unusual siderophore in S. rose- meta-genomic analysis and isotopic enrichment in a multi- um. Natural Product Reports. disciplinary project utilizing Matthias’s skills to bridge the disciplines in support of the laboratory’s missions in renewable energy, biosurveillance and biosecurity and development of countermeasures for bacterial infections under the Advanced Therapeutics Program. 164

Page 165

Complex Natural & Engineered Systems Postdoctoral Research and Development Final Report From Food to Fuel: Making Ammonia Synthesis Viable for Energy Storage Applications James M. Boncella 20130817PRD4 Abstract low vapor pressures of ionic liquids, their high charge The use of ionic liquids as media for the homogeneous, mobilities and wide electrochemical window. Needed, catalytic reduction of nitrogen to ammonia was investi- then, were electrochemically pure ionic liquids, free of gated. Novel and reproducible methods for the synthe- the nearly ubiquitous trace contaminants that accom- sis of extremely pure ionic liquids were discovered. The pany commercially available ILs. Even in research labs, it reduction of nitrogen to ammonia in organic solvents was thought until recently that the pale yellow color of was reproduced as reported in the literature, but trans- many ILs – now known to result from impurities – was an ferring this methodology to various ionic liquids gave inherent property of the materials. This work expanded much poorer catalytic activity. It is likely that the differ- on previous LANL research and resulted in the devel- ing solubility of the reagents in ionic liquids as compared opment of protocols for generating large quantities of to conventional organic solvents was responsible for the electrochemically pure ionic liquids. diminished reactivity of these catalytic systems. The Scientific Approach and Accomplishments partial synthesis of a tridentate phosphine ligand that can be covalently bound to an electrode surface was In order to facilitate N¬2 reduction, Pyrrolidinium completed. and imidazolium cations were chosen for their relative inertness under reducing conditions, and the Background and Research Objectives fluorinated bis(trifluoromethylsulfonyl)amide (Tf2N–) Room temperature ionic liquids (ILs) are salts character- and trifluoromethanesulfonate (TfO–) anions (Figure 1) ized by a melting point below 100 °C. Whereas many were used due to their chemical stabilities and the noted inorganic salts can require temperatures many hundreds propensity of the former for forming ionic liquids that of degrees Celsius to acquire a molten state, the salt are immiscible with water. The details of the synthetic pairs in ILs are prevented from interacting strongly with procedures that resulted from our research will be one another through a variety of approaches, includ- outlined in a forthcoming publication. The course of our ing the incorporation of sterically demanding functional research identified the following factors as being crucial groups with little polarizability, the distribution of charge to the synthesis of electrochemically pure, as illustrated over a large area, and asymmeterization of functional in Figure 2. groups about a charged ion. All of these strategies work to prevent ready incorporation of the ions into a crystal • Repeated distillations of the starting materials lattice, which result in a salt with a very low temperature • Maintenance of the amine alkylation reaction melting point. The use of these various approaches has temperature at or below room temperature led researchers in this field to develop a large library of • Thermal and vacuum activation/purification of the ILs, which have been incorporated into a growing body charcoal used for decolorization of the ILs of research that takes advantage of the unique physical properties afforded by ILs when compared to traditional • Filtration through a 0.8 μm nylon filter organic solvents. • Repeated washings of the product with distilled water Our work in this area was driven by an interest in per- forming the electrocatalytic reduction of N2 to NH3. • Removal of residual water from the product with It was envisioned that ILs would provide a superior either ultra-low pressures (turbopump) or activated medium in which to perform this reduction due to the molecular sieves 165

Page 166

The use of the methods described above led to ionic liq- we first alkylated 2.0 equiv of tBu2PH with 1.0 equiv of uids that were suitably pure to be used in the study of the 2,6-bis(chloromethyl)pyridine, by stirring in MeOH at 45 °C gas absorption, conductivity, and electrochemical proper- for 48 h. After cooling to room temperature, the addition ties of these materials under various temperature and of excess Et3N caused a white solid to precipitate. The pressure regimes. The details of this work will be commu- solution was evacuated to dryness then the mixture was nicated in a manuscript currently under preparation by our extracted with Et2O to afford the product in high yields collaborators. A surprising discovery that resulted from and purities. This ligand was next metallated by heating this work was that several ionic liquids under consideration a mixture of the ligand with MoCl3(thf)3 in THF at 50 °C aggressively retained acetone within the material, obviat- for 16 h. Crystallization of the product from DCM/hexane ing extraction methods that used acetone. The physical yields orange-brown blocks of (PNP)MoCl3. origin of this effect is not obvious given the lack of coor- dination sites in the ionic liquid and the vast difference in boiling points between the two materials. Figure 3. (a) General reaction for the reduction of dinitrogen to ammonia. (b) Synthetic protocol used by Nishibayashi and co-workers to perform the homogeneous chemical reduction of dinitrogen under ambient conditions. Nishibayashi and co-workers described the reduction of (PNP)MoCl3 to Mo as being effected by 6 equiv of 0.5 wt Figure 1. Ionic liquids synthesized and purified for gas absorp- % Na/Hg amalgam. We found that the reduction could be tion and N reduction studies. performed with sodium metal, with or without a catalytic amount of naphthalene. In the absence of naphthalene, the reaction proceeds slowly, requiring up to four days for the reduction to complete. The addition of 10 mol % naphthalene allows the reaction to finish in ca. 24 h (Figure 4); lower naphthalene loadings negatively impacted the reaction time and higher loadings led to diminished product purities. Figure 2. General reaction scheme used for synthesizing elec- trochemically pure ionic liquids. Synthesis of Nishibayashi complex and N2 reduction to NH3 The low-valent dimolybdenum complex [(PNP) Mo(N2)2]2(µ-κ1,κ1-N2) (Mo, PNP = 2,6-bis(di-tert- butylphosphinomethyl)pyridine) has been reported by Nishibayashi and co-workers to perform the catalytic reduction of dinitrogen to ammonia under 1 atm of N2 and at 298 K, using lutidinium triflate ([LutH][OTf]) and cobaltocene [(C5H5)¬2Co] to deliver hydrogen equivalents to the coordinated dinitrogen ligands (Figure 3). Our interest lay in determining whether or not the unique Figure 4. (a) Synthesis of the neutral tridentate ligand PNP. (b) physical properties of ionic liquids could be used to Synthetic scheme describing the modified synthesis. enhance the catalytic output of this reaction. Despite our ability to generate the desired Mo complex Commercial sources of the PNP ligand were found to from this method, we encountered significant differences contain significant quantities of an unidentified impurity from the synthetic protocol reported by Nishibayashi and (ca. 15%). Clean PNP ligand needed to synthesized, co-workers. Our best rational of these differences was 166

Page 167

consistent with reversible N2 binding to the complex, described above. Starting with chelidamic acid, a para- including solvatochromism and pressure- and atmosphere- bromo functionality was installed prior to reduction of dependent color changes. Dinitrogen ligands are poor the 2,6-substituents to hydroxymethyl groups. Conver- σ-donors and moderate π-acceptors. In Mo complex, the sion of the alcohols into halides was accomplished with terminal N2 ligands have both cis- and trans-π-accepting either phosphorus tribromide (PBr3) or trichlorotriazene ligands. Such a coordination environment should limit (TCT). Subsequent attempts to alkylate tBu2PH with these back-bonding into the N2 π* orbitals, leading to a weak species were unsuccessful, leading to complex mixtures interaction between the terminal N2 ligands and the metal of products with no clear evidence for formation of the center. In the original report on this complex, Nishibayashi desired product. Changes to the reaction conditions and co-workers report that satisfactory elemental analyses (solvent, temperature, reaction time, base) similarly failed could only be obtained by drying a solution of Mo under to yield the desired product. Considering the high yield of a stream of nitrogen, consistent with the idea that N2 PNP when using the parent 2,6-bis(chloromethyl)pyridine, coordination is weak for this species. Under the lower we are left to conclude that the para-bromo substituent is atmospheric pressure of Los Alamos compared to Tokyo, deleterious to this reaction. it is conceivable that the chemistry of a complex with a weakly bound gaseous ligand could be altered from that at sea level. Despite the differences in the solution chemistry, we suc- cessfully accomplished N2 reduction in a manner similar to that reported previously. In toluene, the solvent used in the original report, we obtained 14 equivalents of NH3, a marginal increase over the 12 equivalents reported under analogous conditions. The use of ionic liquids as a Figure 5. Synthetic scheme used for the attempted synthesis of reaction medium was hypothesized to stabilize charged modified PNP ligands. intermediates, facilitating the catalytic process. A variety of reaction conditions, including the rate and order of addi- Impact on National Missions tion of hydrogen atom equivalents, the nature of the ionic This work has demonstrated that the use of homogeneous liquid, and the reaction stoichiometry were screened, but catalysts for the reduction of nitrogen in ionic liquids is only trace quantities of ammonia were detected. When viable, though not nearly as efficient as in organic solvents. performing the reaction in toluene, the acid (lutidinium The discovery of novel methods for the purification of ionic triflate) is largely insoluble. It may be that this insolubility liquids demonstrates that it is possible to take advantage is instrumental in controlling the acidity of the solution and of the novel properties of ionic liquids for electrochemistry, the stability of the metal complex. Paradoxically, it appears and by extension, electrochemical devices. This will have a that the insolubility of lutidinium triflate is responsible for positive impact on research that relates to energy security. the success of the reaction in toluene. This work has spawned an ongoing outside collaboration with Michigan State University on the electrochemical Attempted synthesis of modified PNP ligand synthesis of ammonia. Only limited mechanistic information has been presented in the literature to elucidate the manner in which Mo reduces dinitrogen to ammonia. We were thus interested in synthesizing a ligand that retained the coordination properties of PNP while offering a means of attaching the ligand to a surface through which the reaction details could be probed by in situ IR and/or UV-vis monitoring. Our ap- proach to this problem involved attempting to append a linker group in the 4 position of the pyridine ring. The link- er would then be used to attach the molybdenum complex to indium tin oxide, boron-doped diamond, or a similarly transparent, conductive surface. Such a surface bound complex could be studied using spectro-electrochemistry. The synthetic scheme in Figure 5 outlines our approach to synthesizing a ligand that would meet the demands 167

Page 168

Complex Natural & Engineered Systems Postdoctoral Research and Development Final Report Genetically Encoded Tools for Light-controlled Molecular Assembly Geoffrey S. Waldo 20140660PRD1 Abstract modification, or alternant conformation in response to a Tools to study protein interactions with high spatio- second protein. This concept has been illustrated with a temporal resolution would advance our understanding Rac-protein light switch phototropin 1 (LOV2 domain) fu- of biology. Attaching proteins whose assembly could sion protein that was able to control cell mobility when be controlled by light is one way to achieve this goal. activated by light [1]. Several photo switchable versions Photo-switchable phototropin 1 and light-oxygen-voltage of fluorescent proteins have also been investigated for sensor proteins (LOV) domains are large and perturb the their use in live-cell fluorescence microscopy because proteins they are attached to. GFP is a highly fluorescent multiple proteins could be labeled with one of the labels barrel-shaped molecule with 11 beta strands. To over- being switched on or off after exposure to blue light [2]. come the issue of bulky tags, our group has developed The disadvantage of these photo switchable proteins is two-part split GFPs. Strand 11, the ‘tag’, (15 amino acids) that the entire protein needs to be affixed onto a target is attached to other proteins. The rest of the GFP, the protein, and they are simply passive reporters, not as- ‘detector’, (strands 1-10) is added or expressed in the sembling or disassembling with light. The goal of this cell separately, spontaneously binding the strand 11 to project was to develop a three-part split green fluores- form fluorescent GFP. We developed a three-part split cent protein (GFP) technology that is light activated, i.e. GFP whose reassembly is not spontaneous. Strand 10 the GFP fragments could be driven apart with light and and strand 11 are attached to two different proteins. reassemble in the dark. The split GFP technique would Only if the proteins interact with each other, are the have an advantage over other photo switchable proteins small GFP strands near enough to be detected by GFP since the GFP tags are only 15 amino acids long. Waldo strands 1 to 9. This three-part split GFP is thus a protein- et al. previously developed a version of GFP that can be protein interaction detector. Recently Boxer showed that split at either strand 10 or strand 11 leading to a tag that violet light reversibly displaces strand 10 from the rest of can be placed on a target protein [3]. Once GFP strands our two-part GFP. In this project we proposed develop- 1-9 is added to the solution with the labeled target pro- ing a light-activated three-part split GFP system so that tein, the complete GFP is formed leading to a fluorescent the small target strands could be removed by light. This system. This system can then be applied to many differ- technique could then be applied to study spatio-tem- ent applications, including in vivo labeling of proteins, a poral details of protein-protein interactions in influenza tool to determine protein solubility, and most recently, a infections of human cells using light control. We have means to improve protein crystallography. shown the light exchange property is slow in our three- The purpose of this Director’s Fellowship was to modify part GFP, likely because the affinity of strand 10 is higher this existing system to incorporate a light activated for the rest of the GFP, compared to the two-part split component. Currently, the binding of GFP strands 1-9 + GFP. From this we have learned what is needed to make S10-target-S11, GFP strands 1-9 + target-S10-S11, or GFP robust light-activated three-part split GFP system. We strands 1-9 + targetA-S10 + targetB-S11 (when targetA expressed and purified human and influenza proteins and target interact) is nearly irreversible (Figure 1). Re- that interact for use with the new split GFP technology. search by the Boxer lab indicated that the dissociation of our GFP 11-1-9 + S10 could be sped up by the addition Background and Research Objectives of 405 nm light [4]. Then, by incorporating a mutation Optical control of proteins could allow unparalleled in S10 that changes the color from green to yellow, two control of protein-protein interactions. Many impor- different versions of S10 could be used to monitor when tant biological events rely on the protein interaction, 168

Page 169

photo-dissociation and reassembly occurred. I proposed than test protein-10-11 + GFP strands 1-9, we used this to using this new light activated system to look at proteins look for photo-exchange (Figure 2A). The 10-test protein- involved in the human-influenza interaction. A number of S11 with the addition of GFP 1-9 is green. A second target influenza proteins are known to interact transiently with protein with the higher affinity tag of GFP strand 10-11 human proteins, which aids in influenza’s ability to evade with the mutation T203Y (mutation that confers a yel- the human immune system. low fluorescent spectral shift) is added to the solution at a higher concentration. After exposing to 405 nm laser pointer, the green strands would be displaced and the higher affinity yellow strand 10-11 would bind. In concept this experiment is straightforward. Exchange was much slower than in the GFP 11-1-9 + S10 system, consistent with a higher affinity of S10 for the rest of the GFP in the three-part split GFP. Figure 1. An overview of current GFP split technology devel- oped in Geoff Waldo’s lab. A. A target protein is labeled with strand 10 of GFP on the N-terminus and strand 11 on C-terminus, when GFP1-9 is added to a solution containing the target protein GFP is able to form a complete barrel and fluoresces green. B. A second split protein configuration involves a target labeled with strands 10 and 11 on the C-terminus. When GFP1-9 is added to solution the missing strands 10-11 can combine with GFP1-9 and fluoresce green. C. The third split system labels one target pro- tein with strand 10 from GFP and a second target protein with Figure 2. Proposed light-activated GFP systems. A. A target strand 11 from GFP. After addition of GFP1-9, if the two target protein is labeled with S10 T203Y on one terminus and S11 on proteins interact green fluorescence will result. the other terminus and when combined with GFP1-9 fluoresces green. Exposure to 405 nm light causes S10 and S11 to dissoci- Scientific Approach and Accomplishments ate. Excess target protein labeled by GFP10-11 T203Y is present in solution and binds with a higher affinity than the S10_tar- The first step in this project was to develop a split GFP get_S11 version. This is reflected by exchange to yellow (T203Y) light activated system that could then be modified through fluorescence. B. Either a target protein is labeled with S10 or genetic engineering to be used in in vivo cell studies of alternatively, S10 is connected to GFP11-1-9 with a designed pro- human-influenza protein interactions. I started with the tease cut site. The protease cut site can be utilized to separate existing three-part split GFP system that had previously S10 from GFP11-1-9. After 405 nm light exposure, excess target been developed by the Waldo lab [3]. Since the binding protein with S10 T203Y (yellow) is added. S10 T203Y, which has rate of 10-test protein-11 + GFP strands 1-9 is much slower higher affinity, binds preferentially over S10 T203. 169

Page 170

After this we moved to a different light activated system, cal. The residue at position 65 cannot be the only residue utilizing only S10 as a leaving group (Figure 2A). A sample responsible for the different response to 405 nm light, as of target protein-S10 + GFP strands 11-1-9 was incubated, this residue was already investigated. Since strand 10 was resulting in green fluorescence. The sample was then extensively engineered to increase its affinity for the three- illuminated with 405 nm with excess target protein-S10 part split GFP, it is most likely that the difference in off rate T203Y (yellow) added. Exchange would result in a yel- is due to the three residues that differ in strand 10 from low solution. Slow exchange was observed, under room the three-part GFP versus the two-part GFP. There are two light for 48 hours (Figure 3). A different experiment was possible follow up experiments to identify the difference. designed utilizing a caged GFP cleavable fused peptide of Ideally, we would want a system that has the benefits of GFP strand 10 (Figure 2A). To test the effect of the mu- the GFP 1-9 optimized system (strong fluorescence, good tations in the three-part split GFP vs. the two-part split solubility, protein interaction dependence on assembly) GFP, a light activated experiment was designed with both but maintains the light exchange properties of the two- our GFP strands 11-1-9 version from the optimized GFP part GFP system. It is clear that a possible route is to start strands 1-9 complementation system (Figure 1) and a ver- with the two-part system and mix mutations in from the sion based on our two-part split GFP (Figure 4) [4]. In this GFP 1-9 three-part system, and throughout the entire experiment, both versions of GFP 11-1-9-(protease site)-10 selection process, select for variants that maintain or were expressed and purified. The protease sites allowed improve the ability to exchange with 405 nm light but have the strand 10 to be cut using tabacco etch virus protease, improved fluorescence and solubility. This new split sys- but the GFP would remain assembled in the dark. This cut tem could then be used for photoswitching experiments. protein was incubated with excess of sulfate reductase-S10 A second experiment could be designed to specifically look T203Y. Exposure to 405nm laser pointer for 15 minutes at the different affinities of strand 10. This would essen- resulted in significant exchange (520 nm excitation, 550 nm tially just require altering the protease constructs to have emission) for the GFP based on the two-part split construct the strand 10 of the corresponding two-part or three-part but none for the version based on the GFP1-9 optimized split GFP. It could be that the residues present in strand 11 system (Figure 5). could significantly alter the affinity of strand 10. Figure 3. Light exchange experiment with sulfate reductase with C-terminal GFP strand 10. An increase in yellow fluores- cence (520 nm excitation, 550 nm emission) is seen in light exposed samples with excess SR_10 T203Y as expected. There is Figure 4. Protein sequence alignment of GFP11-1-9 based on a further increase after exposure to room light over 48 hours. A the optimized GFP1-9 system developed in the Waldo lab and the higher rate of exchange was expected. version of GFP11-1-9 used by the Boxer lab for caged GFP cleav- able fused peptide sensor. The Blue highlighted regions is the tev This experiment finally established some rationale for why cut sites, red text is residues that differ, green highlighted sec- all the previous trials failed with GFP 1-9 system. As figure tions are strand 11, and yellow highlighted section are strand 10. 4 illustrates, the version based on the two-part GFP con- struct and the GFP 1-9 optimized version are 94% identi- 170

Page 171

In addition, I expressed, purified, and began crystalliza- for the delivery of caged peptides via GFP. (3) Once at the tion trials of human protein p85 and influenza protein NS1 desired location, the caged GFP would be photo dissoci- that are known to interact with each other. These proteins ated, releasing the caged peptide for activating desired would be ideal targets to use with the new light activated biological responses. (4) Further, this allows the selective split GFP system. The expression, co-purification, and detection of proteins tagged with S10 T203Y, by observing some crystallization conditions were identified for these the appearance of yellow fluorescence that can be specifi- two proteins. Once the light activated system with GFP cally excited using 520 nm light (this wavelength does not is optimized these two proteins could make an ideal test excite GFP based on S10 T203). (5) Thus for the first time, system. P85 could be labeled with S10, NS1 with S11, and it will be possible to monitor proteins in specific sites in when GFP1-9 is added to solution the light activated stud- living cells, sites which have been off-limits to conventional ies could be conducted. live-cell labeling experiments. (6) In future research direc- tions, the two-part split GFP and three-part split GFP can be tweaked by evolution to increase photo exchange rate and solubility. Impact on National Missions This work firmly establishes precedence for using green fluorescent proteins for light-activated assembly and disassembly. Since the split GFP fragments are small and non-perturbing, this technology is disruptive since tagged proteins can be studied with less perturbation compared to existing tags. Biosecurity Many pathogen/host interactions have been hampered by the bulky size of tags in the past. Up to now, GFPs Figure 5. Light exchange experiment with caged GFP 11-1-9- have been passive tags, simply fluorescing in response to tev-10 constructs. A. Excitation spectra of GFP11-1-9-tev-10 light. For the first time, this new light-activated split GFP Boxer version with controls and sample with excess sulfate re- ductase_10 T203Y (SR_10) added to solution. There is a marked provides a tool for controlling protein localization, ag- shift toward 510 nm as excitation max in 405 nm exposed gregation, and interactions with light. This tool has strong sample (purple trace). B. Excitation spectra of GFP11-1-9-tev-10 positive impact on our mission to study, understand, and based on GFP1-9 optimized with controls and samples with ultimately control host-pathogen biology. excess SR_10 T203Y. No shift in excitation sample was noted as compared to sample in dark or control GFP11-1-9-tev-10. C. Energy independence Emission spectra in light exchange experiment. Only the GFP11- The ability to control gene networks using light, using 1-9-tev-10 boxer construct exposed to 405nm light after excess the very stable and robust GFP scaffold, means we can SR_10 T203Y was added had an increase in emission over the more easily engineer organisms for synthetic biology. This other samples. D. Plate photographed after excitation at 488 impacts the production of biofuels and bio-derived natural nm, emission at 510 nm (exposure time 1/15 sec). Second im- products. age is the same plate photographed after excitation at 520 nm, emission at 550 nm (exposure time 4 sec). Wells correspond to: References A8 200 µL TNG buffer: A9 20 µL TNG, 180 µL SR_10 T203Y: B8 20 µL Boxer, 180 µL TNG: B9 20 µL 11-1-9-rev-10 1-9 opt, 180 µL 1. Cabantous, S., T. C. Terwilliger, and G. S. Waldo. Protein TNG: C8 20 µL Boxer, 180 µL SR_10 T203Y kept in dark: C9 20 µL tagging and detection with engineered self-assembling 11-1-9-rev-10 1-9 opt, 180 µL SR_10 T203Y kept in dark: D8 20 fragments of green fluorescent protein. 2004. Nature µL Boxer, 180 µL SR_10 T203Y after 15 minutes 405 nm exposure: Biotechnology. 23: 102. D9 20 µL 11-1-9-rev-10 1-9 opt, 180 µL SR_10 T203Y after 15 minutes 405 nm exposure. D8 has a marked increase in yellow 2. Cabantous, S., H. B. Nguyen, J. D. Pedelacq, F. Koraichi, fluorescence. A. Chaudhary, K. Ganguly, M. A. Lockard, G. Favre, T. C. Terwilliger, and G. S. Waldo. A new protein-protein Conclusions. Six major conclusions can be drawn. (1) The interaction sensor based on tripartite split-GFP associa- results show that the two-part split GFP can be photo- tion. 2013. Scientific Reports. 3: 1. dissociated, and that the fragments spontaneously reas- semble in the dark. (2) Since the assembled GFP can transit 3. Do, K., and S. G. Boxer. GFP variants with alternative through cell membranes, this offers an immediate tool beta strands and their application as light-driven pro- 171

Page 172

tease sensors. 2013. Journal of the American Chemical Society. 135: 10226. 4. Wu, Y.. A genetically encoded photoactivatable RAC controls the motility of living cells. 2009. Nature. 461: 104. 5. Zhou, X. X., and M. Z. Lin. Photoswitchable fluorescent proteins: ten years of colorful chemistry and exciting applications. 2013. Current Opinion Chemical Biology. 17: 682. Publications Leibly, D. J., M. A. Arbing, I. Pashkov, N. DeVore, G. S. Wal- do, T. C. Terwilliger, and T. O. Yeates. A suite of engi- neered GFP molecules for oligomeric scaffolding. 2015. Structure. 23 (9): 1754. 172

Page 173

Complex Natural & Engineered Systems Postdoctoral Research and Development Final Report Mesoscopic Lattice Boltzmann Modeling and Investigation of Boiling Multiphase Flows Qinjun Kang 20140669PRD2 Abstract tions. Novel methods are highly demanded for a better Boiling multiphase flow is pervasive in nature and in- understanding of boiling multiphase flows. In the past dustrial applications. However it is not well understood years, the mesoscopic LB method has been developed despite many years of research effort. Novel methods into an alternative approach for simulating multiphase are badly needed for a better understanding of boiling flows, in which the interfaces can arise, deform, and multiphase flows. The main objectives of this research migrate naturally as a result of particle interactions, were to develop a novel, physics-based thermal lattice without needing a cluster of marker points to track the Boltzmann (LB) model for modeling boiling multiphase interfaces or capturing the interfaces via the evolution flows and to obtain new insights into the mechanisms of an order parameter. Moreover, this method has been of boiling heat transfer. Scientific accomplishments of recently demonstrated to be capable of modeling phase this project include the development of a novel, physics- change without any assumptions, though the model is based thermal LB model for modeling multiphase flow limited to low density ratios. with phase change, as well as numerical investigation of The main objectives of this research were to develop boiling multiphase flow and self-propelled Leidenfrost a novel, physics-based thermal LB model for modeling droplets on ratchet surfaces. This project supports phys- boiling multiphase flows, and to obtain improved under- ics-based prediction models of departure from nucle- standing of boiling phenomena, e.g., new insights into ate boiling, one of the grand challenge problems of the the physics of critical heat flux at different scales through DOE Consortium for Advanced Simulation of Light Water mesoscopic modeling. Reactors. The approach will also have significant impli- cations in other research areas including spray cooling, Scientific Approach and Accomplishments nucleate boiling heat transfer of nanofluids, and cooling Novel, physics-based thermal LB model for modeling of micro-electronic devices. This project has resulted in boiling multiphase flows four publications in barely one year. We developed a hybrid thermal LB model for simulat- ing thermal multiphase flows, which employs an im- Background and Research Objectives proved pseudopotential LB model and a finite-difference Boiling multiphase flows are ubiquitous in nature as well scheme to simulate the fluid flow and the temperature as in many industrial applications. Nucleate boiling, a field, respectively. The pseudopotential LB model and well-known boiling phenomenon, has been recognized the finite-difference solver of temperature are coupled as one of the most effective heat transfer modes and via the equation of state. Numerical simulations of used in a wide field of high-tech devices and systems boiling heat transfer have been conducted using the such as nuclear reactors, heavy-vehicle engines, comput- model. The complete three boiling stages, nucleate er chips, and micro-electronic devices. However, despite boiling, transition boiling, and film boiling, as well as years of tremendous research effort, many aspects of the boiling curve were successfully reproduced in our boiling are still not well understood, e.g., the physical simulations with the bubble nucleation, growth, depar- mechanism causing the critical heat flux is not revealed. ture, and coalescence being well captured. Some basic This is due to (1) the correlations developed by experi- features of boiling heat transfer were clearly observed in ments rely heavily on empirical parameters which are the numerical results, such as the severe fluctuation of only valid for a narrow parameter range and (2) numeri- transient heat flux in the transition boiling regime, the cal simulations based on traditional numerical methods critical heat flux in the boiling curve, and the feature that usually involve many assumptions and empirical correla- 173

Page 174

the maximum heat transfer coefficient lies at a lower wall tio. Moreover, we have also studied the performances of superheat than that of the maximum (critical) heat flux. Leidenfrost droplets on inclined ratchet surfaces. Different Furthermore, the influences of the heating surface wetta- inclination angles have been investigated. It is found that bility on boiling heat transfer have also been studied. The the droplet moves downhill in the early stage due to the numerical results show that the critical heat flux decreases downhill acceleration caused by the gravitational force. when the contact angle increases, which also causes the Later, the droplet turns around at a certain time with the boiling process to enter the film boiling regime at a lower help of the uphill acceleration, which is generated by the wall superheat. Besides, increasing the contact angle will vapor flow beneath the droplet. The maximum inclination reduce the required wall superheat for the onset of boil- angle at which a Leidenfrost droplet can still climb uphill ing (Figure 1). All of these findings are consistent with the successfully is found to be related to the initial radius of experimental studies in the literature, demonstrating that the droplet. the proposed model is capable of reproducing the basic features and the fundamental characteristics of boiling heat transfer. Figure 2. The streamlines for the case of initial droplet radius 45 at dimensionless time 1.45. Impact on National Missions Figure 1. Boiling curves at different contact angles. The This project supports the DOE/NNSA missions in Energy squares, circles, and triangles represent the results of static (liq- Security, Environmental Quality, and the missions of the uid) contact angles around 44.5, 50, and 55.5 degrees. Office of Science by enhancing our understanding of boil- Numerical investigation of self-propelled Leidenfrost ing multiphase flow and heat transfer. The project supports droplets on ratchet surfaces physics-based prediction models of departure from nucle- Using the developed model, we have also investigated the ate boiling, one of the grand challenge problems of the self-propelled motion of Leidenfrost droplets on ratchet DOE Consortium for Advanced Simulation of Light Water surfaces, a phenomenon that liquid droplets perform self- Reactors. The approach will also have significant impli- propelled motion when placed in contact with hot surfaces cations in other research areas including spray cooling, with asymmetric textures. This phenomenon has impor- nucleate boiling heat transfer of nanofluids, and cooling of tant applications in cooling systems with no moving parts, micro-electronic devices. such as microprocessor cooling. Publications We have investigated the dynamic behavior of Leidenfrost Li, Q., K. Luo, Q. Kang, Y. He, Q. Chen, and Q. Liu. Lat- droplets on horizontal ratchet surfaces. Numerical results tice Boltzmann methods for multiphase flow and show that the self-propelled motion of Leidenfrost drop- phase-change heat transfer. 2015. arXiv preprint lets originates from the asymmetry of the ratchets and the arXiv:1508.00940arXiv preprint arXiv:1508.00940. vapor flows below the droplets (see streamlines in Figure Li, Q., K. Luo, Q. Kang, and Q. Chen. Contact angles in the 2). It is found that the Leidenfrost droplets move in the pseudopotential lattice Boltzmann modeling of wet- direction toward the slowly inclined side from the ratchet ting. 2014. Physical Review EPhysical Review E. 90 (5): peaks, which agrees with the experimental observation. 053301. The effects of the initial droplet radius and the influences of the ratchet aspect ratio have been studied. We have Li, Q., Q. Kang, M. Francois, Y. He, and K. Luo. Lattice found that there exists a critical value of the aspect ra- Boltzmann modeling of boiling heat transfer: The boil- 174

Page 175

ing curve and the effects of wettability. 2015. Interna- tional Journal of Heat and Mass TransferInternational Journal of Heat and Mass Transfer. 85: 787. Li, Q., Q. Kang, M. Francois, and A. Hu. Lattice Boltzmann modeling of self-propelled Leidenfrost drop- lets on ratchet surfaces. 2015. arXiv preprint arXiv:1506.00700arXiv preprint arXiv:1506.00700. 175

Page 176

Information Science & Technology

Page 177

Information Science & Technology Directed Research Continuing Project Information-Driven Materials Discovery and Design Turab Lookman 20140013DR Introduction development of information science predictive tools to The goal of this project is accelerated materials discov- enable accelerated materials discovery is also a fun- ery by advancing the state-of-the-art statistical inference damental aspect of innovation in the above agencies. and optimization tools to find materials with desired or These tools are based on machine learning methods and targeted properties. This data-centric approach nar- are widely applicable to fields such as pattern recogni- rows the possible search space for improved prediction tion, statistical analysis and operation research, areas by teasing out hidden information that is present in the that are also of interest to these agencies. Our focus in data for known materials. Until very recently, new mate- this proposal is to demonstrate the new capability on rials have almost exclusively been discovered by intuition the class of materials known as piezoelectrics and other and costly trial and error. However, literally overnight functional materials. These are of interest to DOE/SC as the use of simple statistical inference tools is just start- they find applications as sensors and energy harvesting ing to define the new field of materials informatics. Our devices. More broadly, we will pioneer a new approach key innovation is an integrated design loop - an excellent to materials discovery using information science predic- example of the paradigm of codesign. It is a departure tive tools that will impact all missions that depend on in how to think about the problem of Materials Discov- materials, including the weapons program. The design ery by using sophisticated information theoretic tools loop we propose is also an example of the paradigm of developed in various contexts including pattern recogni- codesign (which is gaining greater traction within DOE, tion, operations research, bioinformatics, but applying for example,) - where learning a “model” is performed them in a unique way to materials science issues. As an iteratively guided by experiments and computer calcula- example, we will focus on the class of functional materi- tions, which in turn are guided by statistical analysis. als known as piezoelectrics. These are materials which produce a large strain if an electric field is applied, or Progress a large polarization if stress is applied. The need is to • Prediction, synthesis and characterization of NiTi- find materials with a large response and high working based ultra low thermal dissipation alloys. Several temperatures. Such materials have numerous applica- new alloys have been discovered by implementing tions, from energy harvesting devices to sensors, and our design framework. The alloys belong to the N. America has been gradually and continuously losing class: NiTiCuFePd. One of these compounds has the market share in these markets. These predominantly smallest thermal dissipation of any known com- lead-based compounds are now banned in many parts of pound. Our adaptive design framework is general Europe. Our objective will be to search for new lead-free and can be applied to any materials class. piezoelectrics. However, the paradigm we will develop • Prediction, synthesis and characterization of new will also be applicable to other materials classes with BaTiO3 based Pb-free alloys that are easier to syn- their own targeted properties, and where possible, we thesize. Further characterization work is underway will examine within this project other related materials to determine the piezoelectric coefficients. classes. • Discovery of a new feature (Excess Born charge) that classifies AB solids with fewer misclassifications than Benefit to National Security Missions any other scheme (work to appear in Phys. Rev. B). Accelerated materials discovery is of strategic impor- This has been a classic problem studied over the last tance and relevance to DOE/SC initiatives. Moreover, the 50 years and may be considered to be the earliest 177

Page 178

example using materials design ideas. Conclusion • An approach to classification learning which uses If successful, we will establish a new paradigm for address- orbital radii to design new ductile intermettalic com- ing complex optimization problems that include multiple pounds. Our predictive models indicate that the com- sources and choices to incorporate domain knowledge. pounds SoCo, ScIr and YCd should be ductile, whereas Further, we will have developed tools and algorithms to each was previously predicted to be brittle (to appear deal with error estimation, with experimental design and in Nature Scientific Reports, 2015). employing materials domain knowledge in machine learn- • Discovery of new polymer based dielectrics with large ing. Our success will be measured by our discovery of new wide band gaps and dielectric constants using high materials with desired properties and greater insight into through put electronic structure calculations. the dominant mechanisms (structure-property relation- • Discovery of new water-splitting catalysis compounds ships) underlying materials phenomena. A main deliver- using high through put electronic structure calcula- able will be an Information Science & Technology based tions. materials discovery capability for accelerated materials • An understanding of how the adaptive design algo- discovery. rithms are forgiving of issues and drawbacks of the inference training model. Publications • In the design of Light emitting diodes (LEDs), industry Information Science for Materials Discovery and Design. standard simulation codes (e.g. APSYS) are used to pre- 2015. dict the internal quantum efficiency before fabrication. We have demonstrated how adaptive design (using a Balachandran, P. V., D. Xue, J. Theiler, J. E. Hogden, and T. Gaussian model coupled to efficient global optimiza- Lookman. Adaptive design strategies for materials de- tion) can be used to guide APSYS to optimally design sign. Scientific Reports (2015). LEDs with better performance than what has been Balachandran, P. V., Theiler, J. M. Rondinelli, and Lookman. done so far. Materials Prediction via Classification Learning. 2015. SCIENTIFIC REPORTS. 5. Future Work Gubernatis, J.. Data Visualization and Structure Identica- A key component of the project has been the adaptive tion. 2014. LA-UR-14-22628. design strategy to suggest the next experiments for find- ing materials with single targeted properties. This will be Gubernatis, J.. Data Visualization and Structure Identica- studied in greater detail in its own right and applied to a tion. 2015. In Information Science for Materials Discov- number of other materials data sets to test robustness. It ery and Design. Edited by Lookman, T., F. J. Alexander, includes a number of machine learning tasks (e.g. batch and K. Rajan. Vol. 225, First Edition, p. 1. Berlin: Spring- mode or multiple experiments other than Kriging believer er Verlag. algorithm, Bayesian approach), coupled with tight coupling Lookman, T.. Heterogeneities, The Mesoscale and Multi- to synthesis, characterization and DFT (electronic calcula- functional Materials Codesign: Insights and Challenges. tions) efforts. Specific materials classes include Heusler 2014. In Mesoscopic Phenomena in Multifunctional alloys and Bi based piezoelectrics. Materials. Edited by Saxena, A., and A. Planes. Vol. 198, First Edition, p. 57. Berlin: Springer. Our materials design has so far involved a single property or objective. Our goal will be to formulate a multi-objective Lookman, T., P. V. Balachandran, D. Xue, G. Pilania, T. Shear- optimization strategy so that we can optimize two or more man, J. Theiler, J. Hogden, J. E. Gubernatis, K. Barros, E. properties. This will be performed with data sets (Max Ben-Naim, and F. J. Alexander. A perspective on materi- als informatics: state-of-the-art and challenges. 2015. phases and Apatites) that will facilitate benchmarking, as In Information Science for Materials Discovery and De- well as the data set for NiTi based compounds for which sign. Edited by Lookman, T., F. J. Alexander, and K. Ra- we have successfully searched for new low thermal dissi- jan. Vol. 225, First Edition, p. 3. Berlin: Springer-Verlag. pation alloys. We will now also optimize for large transition temperatures and large strains. Mannodi-Kanakkithodi, A., G. Pilania, T. Doan Huan, T. Lookman, and R. Ramprasad. Informatics-Driven Strat- A number of our materials problems (e.g. Heuslers) egy for the Accelerated Design of Polymer Dielectrics. require addressing the classification problem in an adap- Scientific Reports (2015). tive, iterative manner (e.g. finding new phase transforming Pilania, G., J. E. Gubernatis, and T. Lookman. Classification Heuslers). We will therefore study a feedback loop for clas- of octet AB-type binary compounds using dynamical sification to guide experiments. 178

Page 179

charges: A materials informatics perspective. To appear in Scientific Reports (2015). Pilania, G., P. V. Balachandran, J. E. Gubernatis, and T. Lookman. Classification of ABO3 perovskite solids: a machine learning study. 2015. Acta Crystallographica, Section B: Structural Science, Crystal Engineering and Materials. 71 (5): 507. Pilania, G., T. Lookman, and J. Gubernatis. Structure clas- sification and melting temperature prediction in AB solids via machine learning. 2015. Physical Review B, Vol 91, Issue 21, 214302 (2015). Pilania, G., and T. Lookman. Electronic structure and biaxial strain in RbHgF 3 perovskite and hybrid improper fer- roelectricity in (Na,Rb)Hg 2F 6 and (K,Rb)Hg 2F 6 su- perlattices. 2014. Physical Review B (Condensed Mat- ter and Materials Physics) . 90 (11): 115121. Xue, D., P. V. Balachandran, J. E. Hogden, J. Theiler, and T. Lookman. Accelerated search for materials with tar- geted properties by adaptive design. Nature Communi- cations. 179

Page 180

Information Science & Technology Directed Research Continuing Project Next Generation Quantum Molecular Dynamics Anders M. Niklasson 20140074DR Introduction Benefit to National Security Missions Quantum mechanical Born-Oppenheimer molecular This project will provide a new capability that will (a) dynamics (QMD) provides a highly powerful, but com- benefit our mainstream Laboratory missions in weapons putationally expensive, multidisciplinary tool to predict, science and energy security by enabling the predictive understand and design materials and processes directly modeling of fundamentally quantum mechanical pro- from the fundamental principles of quantum physics. cesses occurring in energetic materials, nuclear fuels, Merging QMD with future extreme-scale computing actinide chemistry, and fuel cells; and (b) provide a therefore holds the promise of a major paradigm shift in competitive advantage to LANL as we target new oppor- materials science, chemistry, and biology. Unfortunately, tunities in emerging areas such as the Materials Genome this potentially revolutionary opportunity will never be Initiative, biofuels, and partnerships with industry. To realized without a radical re-design of current approach- give one example, equations-of-state and opacities in es to QMD, overcoming a number of interconnected fun- highly nonequilibrium conditions important to the weap- damental shortcomings. Based on several LANL-unique ons mission rely increasingly on quantum mechanical innovations, including a conceptually new graph-based molecular dynamics simulations, but are constrained by approach to quantum mechanics, we propose a major the small length and short time scales presently acces- coordinated interdisciplinary effort to develop a truly sible. A number of recent DOE workshops have highlight- transformative computational framework for a new ed “the need to advance theoretical models of physical generation QMD that is specifically tailored for extreme- and chemical objects and processes, and to develop scale computing on heterogeneous architectures and algorithms that effectively enable those models to be allows simulations for time and length scales that are simulated on modern computing systems”, in particular: multiple orders of magnitude beyond current capacity. 1) advancing the speed and accuracy for quantum chem- A key element of the proposed work is the development istry modeling on extreme-scale platforms, and 2) bridg- of a radical alternative to the traditional computational ing length and time scales to enable seamless multiscale algorithms that are unsuited to large-scale simulations. chemical reactivity modeling. Our new graph-based approach for solving the electronic structure problem will enable us to combine, for the first Progress time, high on-node computational efficiency with mas- We have made significant progress on all deliverables sive parallelism for QMD simulations, including on-the- and milestones over the last 12 months. Notable high- fly interrogation of the electronic degrees of freedom for lights include our project review, where all aspects of visualization or quantum response properties. We will our work and project were ranked highly, and we hired develop and demonstrate this capability on two test- an outstanding postdoctoral fellow who is making good bed applications (involving aspects of nuclear fuels, and progress. We are heavily involved in the 2015 IS&T sum- protein structures and dynamics) where classical MD mer school where 6 graduate students are exploring methods are inadequate or fail, and where QMD simu- graph partitioning approaches to electronic structure lations have thus far been too expensive. This unique theory, their implementation on modern computational computational capability will strengthen LANL’s leader- architectures, and application to problems in chemistry ship in numerous mainstream priority areas in materials and materials science. The project has already gener- science, chemistry and biology, and it will have a lasting ated several peer-reviewed publications, invited talks at multidisciplinary impact on diverse applications that international conferences, as well as strong representa- advance our missions. 180

Page 181

tion from team members at the CECAM workshop ‘Next diagonalization per time step, whereas regular Born- Generation Quantum Molecular Dynamics: Challenges Oppenheimer formulations exhibit unphysical fluctuations and Perspectives’ in Bremen, Germany. Various technical unless a sufficiently high degree of convergence is reached highlights and progress toward our deliverables are listed at each time step. The thermostated extended Lagrang- below: ian framework thus offers an accurate approach to sample processes in the canonical ensemble at a fraction of the A shared memory ELLPACK-R implementation of our work- computational cost of regular Born-Oppenheimer molecu- horse SP2 algorithm was shown to scale linearly on 1 to lar dynamics. The algorithm, as implemented in the LATTE 16 threads for poly(ethylene) and liquid water systems. A code, has been coupled to the accelerated molecular dy- poly(ethylene) molecule and solvated protein were shown namics framework, AMDF, providing a powerful tool to the to scale linearly on a single Intel MIC node and a Haswell study of rare event processes with quantum accuracy. node. The ELLPACK-R SP2 algorithm was developed and shown to outperform existing sparse matrix libraries, such Large-scale calculations at zero electronic temperature as MKL (CPU and MIC) and CuSparse (GPU). Additionally were demonstrated using the DFTB+ code, and we are the distributed multi-node/multi-core graph partitioning pursuing QMD simulations using the nonorthogonal SP2 approach was shown to exhibit good performance for method in LATTE using the 3ob parameter set from dftb. a large systems using structure-based partitioning. The be- org, which includes a parameterization for sulfur. Finally, ginnings of a single node/multi-core/GPU graph partition- we have demonstrated crystallographic refinement of the ing of the SP2 algorithm has shown improved single node charge density of urea computed using DFT (VASP) and performance over a CPU-only version. DFTB (DFTB+, LATTE) against high-resolution experimental data (0.35 Angstrom). The refinements yield a figure of We designed a novel merge-based sparse-matrix mul- merit comparable to traditional methods (SHELX), but with tiplication algorithm for use in the SP2 algorithm which substantially different hydrogen displacement parameters. guarantees stride 1 access to the working data and thus These results have important implications for interpreta- minimizes memory divergence. This is especially important tion of hydrogen displacement parameters in small-mole- for accelerated architectures with limited cache, and disor- cule crystal structure models. dered (large-bandwidth) matrices. An automated algorithm and code for the parameteriza- Accelerated versions of the merge-based SP2 algorithm tion and optimization of tight binding models was devel- were developed in OpenCL and CUDA. The CUDA version oped for both the Hamiltonian matrix elements and the was co-developed with Nvidia and results presented at the interatomic pair potentials. This code has enabled us to 2015 GPU Technology Conference. Notably, the merge- rapidly derive new parameterizations or improve existing based algorithm was shown to be more than 5x faster than models parameterization. We have also explored a novel Nvidia’s in-house cuSparse library routines for disordered use of the extended Lagrangian Born-Oppenheimer MD matrices, and Nvidia is interested in incorporating the formalism in simulations of metals. method into their packages. Future Work We are designing and implementing a basic matrix library • We shall complete the development of our graph- (bml) with several internal matrix representations (dense, based, parallel density matrix build in the stand-alone sparse, and graph-based). Through a common API, the PROGRESS library and link it with the LATTE quantum application code can seamlessly utilize and switch be- molecular dynamics code. Rapid, positive progress was tween any of the supported internal formats. We are also made in the last 12 months, supported in part by the addressing O(>N)bottlenecks in the LATTE code that will 2015 IS&T summer school. enables its use in simulations containing >100000 atoms. • Application of accelerated MD to organic systems (liq- uids and biological) based on both adaptive self-consis- We have investigated the use of the extended Lagrangian tent field calculations and parallelism. Publications in Born-Oppenheimer molecular dynamics formalism con- these areas are currently in preparation. stant temperature dynamics. Three different approaches • Investigation of the utility of TB+U in obtaining good were considered: the NosÁ and Andersen thermostats, and electronic densities of states for actinide-based materi- Langevin dynamics. We find that the simulations based on als. The continued development of automated, ma- the extended Lagrangian Born-Oppenheimer framework chine-based algorithms for parameterizing TB models provides accurate canonical distributions even under ap- (both bond integrals and the corresponding repulsive proximate SCF convergence, often requiring only a single terms). 181

Page 182

• Continued exploration of sparse matrix methods on Martinez, E., M. J. Cawkwell, A. F. Voter, and A. M. N. traditional architectures as well as graphics process- Niklasson. Thermostating extended Lagrangian Born- ing units (via the cuSPARSE libraries), and more exotic Oppenheimer molecular dynamics. 2015. JOURNAL OF CHEMICAL PHYSICS. 142 (15): 154120. devices such as MICs. • Further work on the computation of charge densities Mniszewski, S. M., M. J. Cawkwell, J. Mohd-Yussof, N. Bock, from density functional tight binding for the simulation T. C. Germann, and A. M. N. Niklasson. Efficient paral- of protein structures and associated X-ray diffraction lel linear scaling calculation of the density matrix for patterns. Born-Oppenheimer Molecular Dynamics. 2015. Journal • Elimination of all parts of the LATTE quantum MD of Chemical Theory and Computation. code that have worse than linear scaling in terms of Niklasson, A. M. N., M. J. Cawkwell, E. H. Rubensson, and performance and memory use. These steps will enable E. Rudberg. Canonical density matrix perturbation the- > 100000 atoms to be simulated using our existing ory. 2014. Submitted to Phys. Rev. E, under review. code base. Our development of the new, general ‘BML’ library and interface will facilitate this work. Niklasson, A. M. N., S. Mniszewski, C. F. A. Negre, M. J. • Publication and continued testing of our new, linear Cawkwell, P. Swart, J. MohdYusof, T. C. Germann, M. scaling algorithm for obtaining the inverse factorization Wall, N. Bock, E. H. Rubensson, and H. Djidjev. Fermi of the overlap matrix in electronic structure theory. operator expansion using graph partitioning. 2015. To be submitted to Phys. Rev. Lett.. Conclusion Niklasson, A. M. N., and M. J. Cawkwell. Generalized ex- Our chief result will be a general, interdisciplinary, com- tended Lagrangian Born-Oppenheimer molecular putational capability that will enable us to analyze, under- dynamics. 2014. Journal of Chemical Physics. 141 (16): stand, predict and design materials and processes directly 164123. from quantum theory, at time and length scales that are multiple orders of magnitude beyond current capacity. This will be delivered in the form of the portable and openly available program package that can be used with a broad class of electronic structure codes. The development will be guided through two testbed application systems of LANL mainstream mission relevance: 1) protein structures and 2) nuclear fuels. Publications Cawkwell, M. J.. Fast quantum molecular dynamics simula- tions of shock-induced chemistry in organic liquids. In- vited presentation at American Physical Society March Meeting 2014. (Denver, 3-7 March 2014). Cawkwell, M. J., A. M. N. Niklasson, and D. M. Dattelbaum. Extended Lagrangian Born-Oppenheimer molecular dy- namics simulations of the shock-induced chemistry of phenylacetylene. 2015. JOURNAL OF CHEMICAL PHYS- ICS. 142 (6). Cawkwell, M. J., J. D. Coe, S. K. Yadav, X. -. Liu, and A. M. N. Niklasson. Extended Lagrangian Formulation of Charge-Constrained Tight-Binding Molecular Dynamics. 2015. JOURNAL OF CHEMICAL THEORY AND COMPUTA- TION. 11 (6): 2697. Cawkwell, M. J., M. A. Wood, A. M. N. Niklasson, and S. M. Mniszewski. Computation of the Density Matrix in Electronic Structure Theory in Parallel on Multiple Graphics Processing Units. 2014. JOURNAL OF CHEMI- CAL THEORY AND COMPUTATION. 10 (12): 5391. 182

Page 183

Information Science & Technology Directed Research Continuing Project Optimization and Control of Dynamic Networks Angel E. Garcia 20140565DR Introduction Benefit to National Security Missions Network science research has been primarily focused Information Science and Technology is a Laboratory stra- on the simplification of dynamic graphs to static graphs tegic pillar that touches a broad spectrum of science at and their topological characteristics. But processes on LANL from discovery science to pivotal program needs. networks such as the spread of a virus on a social net- This proposal directly addresses Complex Networks work, traversal through a cyber network, and cascading capabilities through the modeling of dynamic networks failures in a power grid are critically dependent on the and with applications to cybersecurity and power sys- dynamic nature of the underlying network. tems (“smart grids’’). These dynamic graphs are at the heart of technologi- This proposal develops core information science and cal and scientific problems in electrical power grids. technology capabilities needed to address the open Deployment of new technologies, such as renewable questions of the Office of Electricity that intersect with generation and electric vehicles, is rapidly transforming the scientific goals of DOE’s Office of Advanced Scien- electrical power networks by coupling previously dis- tific Computing Research (ASCR) Applied Mathematics tinct spatiotemporal scales and challenging traditional Program such as uncertainty quantification in complex approaches for designing, analyzing, and operating engineered networks. power grids. The interactions of spatiotemporal scales in power systems are pushing the limits power engineering This project will innovate with basic research supporting best practices and we need to develop general complex LANL’s internal cybersecurity programs and Global Secu- system models at the appropriate level of network detail rity programs. Effective cyber defense of the weapons necessary to isolate and analyze the relevant static, dy- complex is essential to its security and effectiveness. namic, and stochastic phenomena. Progress Computer networks are also inherently dynamic and We developed methods for reconstruction of spread- non-stationary. Capturing the dynamics of user activ- ing couplings from partial observations with a dynamic ity through coarse-level and high-fidelity modeling is message-passing algorithm. The problem of network critical to understanding normal activity and detection reconstruction and parameters estimation from diffu- of anomalous activity. Only recently has a network-wide sion data is important in many applications including (graph theory) viewpoint been taken. understanding the structure of the epidemic spread of disease. A number of recent papers introduced efficient The CNLS brings a unique perspective in the integration algorithms, based on the maximization of the likelihood of interdisciplinary approaches and ideas to the subject of observed cascades, assuming that the full information of network science. The CNLS has been a leader and for all the nodes in the network is available. innovator in Information Science and Technology. In particular the CNLS has helped provide new approaches In this work, we focus on a more realistic and restricted in theory and modeling of networks for more than 10 scenario, in which only a partial information on the cas- years. The CNLS has helped nucleate efforts at LANL for cades is available, and show that a robust reconstruction applications of networks in neurocomputation, smart of spreading parameters can be achieved with an algo- grid, and cybersecurity. rithm based on recently introduced dynamic message- 183

Page 184

passing equations for the spreading processes. This work is ized computer authentication systems such as those found being prepared for publication. at most large corporate environments. he connectivity of the authentication systems creates risk that malicious Distributed stability and control algorithms were designed users can traverse the network to gain unauthorized access for large-scale dynamical networks. Using recent tools computers and databases. Using these models we can from semi-definite programming and sum-of-squares study the temporal structure to answer reachability ques- methods, we have been been designing computation- tions such as how long it would take, starting at a single ally efficient and parallel algorithms to certify stability computer node, to connect to most of the other comput- of interconnected systems. We have proposed two such ers in the network. We proved asymptotic theorems that algorithms --- 1) a decentralized “stability-ring algorithm” give lower and upper bound for these times. The work has to certify asymptotic stability, and 2) a comparison sys- been submitted as the chapter, “Temporal Reachability in tems based algorithm to certify exponential stability of dynamic networks.” large-scale systems. With these algorithms we designed distributed and local control policies. We are also looking Future Work at applying these stability and control algorithms to real- For FY16 we will focus on the following tasks. world systems, such as the power systems network includ- ing design of a distributed optimal reactive power dispatch • Complete results on reconstruction of spreading net- algorithm for an electrical distribution feeder, with applica- works and further development of dynamic message tion to photovoltaic inverters. passing techniques for this application. • Construct ODE/PDE models for power and gas systems This work is under review in the papers 1) S. Kundu and M. that incorporate the different devices and controls that Anghel, “An iterative expanding-contracting ring algorithm are making inroads into electrical distribution net- for decentralized stability analysis”, 54th IEEE Conference works, e.g., PV systems, electric vehicle charging, and on Decision and Control, 2015; and 2) S. Kundu and M. frequency responsive loads. The ODE/PDEs approach Anghel, “Computation of linear comparison equations for will address the interaction of emergent phenomena stability analysis of interconnected systems”, 54th IEEE arising from the interaction of new smart-grid compo- Conference on Decision and Control, 2015 nents. • Develop reduced models for natural gas pipe network Models and control algorithms were developed for flow flow along with strategies for optimization and control. in natural gas pipeline systems. This area of research is Seek theorems about the structure and optimization becoming urgent because the recent advances in hydro- of these networks and develop algorithms that exploit fracking for gas production have significantly increased the the theoretical results. utilization of gas pipeline systems and made the operation • Sponsor two workshops: “Physics informed machine and control of these systems much more challenging. We learning,” which will build a new community of re- constructed models for distributed network flow systems searchers aiming to exploit physics to create higher where flows on edges are driven by node potentials. performance algorithms for learning and classifica- These structure of these models is very interesting and can tion; and “Conference on Data Analysis (CODA). The be exploited to find optimal solutions. We showed that 3rd conference on big data and statistics will focus of the network flow equations are unique and can be equiva- research across the DOE complex. lently recast as the solution of a strictly convex optimiza- tion problem. We also analyzed the maximum throughput Conclusion problem to maximize the amount of flow that can be deliv- We expect to develop research results with three interlock- ered to the loads while satisfying bounds on the node po- ing themes: tentials. We illustrated application of these general results Dynamic Networks: develop models and algorithms for to balanced, i.e. steady, natural gas networks and validated temporal random graphs; seek theorems for connectivity the theoretical results with simulations on a test case. and reachability in the models; and explore the range of possible temporal correlation patterns. This work generated three papers currently under review. Power Systems: create comprehensive ODE/PDE models for power systems; develop reduced-order stochastic mod- We created new dynamic network models as “temporal els of network power-outage cascades; develop mathemat- random graphs” by generalizing two well known static ical tools to capture system energy and thermal states. random graph models. These models are relevant for Cybersecurity: model as dynamic networks; develop tech- studying questions about the dynamic structure of central- 184

Page 185

niques including hierarchical models to capture statistics of Kernel Graphs. 2015. PLOS ONE. 10 (9). bursty network communication patterns; create inference Jaworsky, , Turitsyn, and Backhaus. THE EFFECT OF FORE- methods and fit cyber authentication network data to the CASTING ACCURACY ON THE SIZING OF ENERGY STOR- temporal random graph models. AGE. 2014. In 7TH ANNUAL DYNAMIC SYSTEMS AND CONTROL CONFERENCE, 2014, VOL 2. Publications Bienstock, , Chertkov, and Harnett. Chance-Constrained Kundu, S., and I. A. Hiskens. Overvoltages due to Synchro- Optimal Power Flow: Risk-Aware Network Control un- nous Tripping of Plug-in Electric-Vehicle Chargers Fol- der Uncertainty. 2014. SIAM REVIEW. 56 (3): 461. lowing Voltage Dips. 2014. IEEE Transactions on Power Delivery. 29 (3): 1147. Catanach, T., and M. Garcia. Power System Dynamic State Estimation using a Layered Local-Global Method. IEEE Kundu, S., and I. Hiskens. Nonlinear dynamics of hysteresis- Transactions on Power Systems. based load controls. 2014. In The 19th World Congress of the International Federation of Automatic Control. Chernyak, V. Y., Chertkov, Bierkens, and H. J. Kappen. Sto- (Cape Town, South Africa, 24-29 Aug. 2014). , p. 5419. chastic optimal control as non-equilibrium statistical Cape Town: IFAC. mechanics: calculus of variations over density and cur- rent. 2014. JOURNAL OF PHYSICS A-MATHEMATICAL Lubin, M., and Y. Dvorkin. A Robust Approach to Chance AND THEORETICAL. 47 (2). Constrained Optimal Power Flow with Renewable Gen- eration. IEEE Transactions on Power Systems. Doerfler, , M. R. Jovanovic, Chertkov, and Bullo. Sparsity- Promoting Optimal Wide-Area Control of Power Mak, T., P. Van Hentenryck, A. Zlotnik, H. Hijazi, and R. Networks. 2014. IEEE TRANSACTIONS ON POWER SYS- Bent. Efficient Dynamic Compressor Optimization in TEMS. 29 (5): 2281. Natural Gas Transmission Systems. Invited presentation at American Control Conference 2016. (Boston, 6-8 Jul. Farrell, M., T. Goodrich, N. Lemons, F. Reidle, F. Villaamil, 2016). and B. Sullivan. Hyperbolicity, degeneracy, and expan- sion of random intersection graphs. 2014. arXiv. Mehta, , N. A. Sinitsyn, Backhaus, and B. C. Lesieutre. Safe control of thermostatically controlled loads with in- Frolov, , Backhaus, and Chertkov. Efficient algorithm for lo- stalled timers for demand side management. 2014. EN- cating and sizing series compensation devices in large ERGY CONVERSION AND MANAGEMENT. 86: 784. power transmission grids: I. Model implementation. 2014. NEW JOURNAL OF PHYSICS. 16. Misra, S., M. W. Fisher, S. Backhaus, R. Bent, M. Chertkov, and F. Pan. Optimal compression in natural gas net- Frolov, , Backhaus, and Chertkov. Efficient algorithm for lo- works: A geometric programming approach. 2015. IEEE cating and sizing series compensation devices in large Transactions on Control of Network Systems. 2 (1): 47. power transmission grids: II. Solutions and applica- tions. 2014. NEW JOURNAL OF PHYSICS. 16. Nagarajan, H., E. Yamangil, R. Bent, S. Backhaus, and P. Van Hentenryck. Optimal Resilient Transmission Grid De- Garcia, M., S. Backhaus, and R. Bent. Power Flow-Based sign. Invited presentation at Power Systems Computa- Adaptive Generator Controls. Invited presentation at tion Conference 2016. (Genoa, Italy, 20-24 Jun. 2016). 52nd Annual Allerton Conference on Communication, Control, and Computing. (Allerton, Illinois, 1-3 Oct. Nagarajan, H., and S. Rathinam. On maximizing algebraic 2014). connectivity of networks for various engineering ap- plications. Invited presentation at European Control Garcia, M., T. Catanach, S. V. Wiel, and R. Bent. Line Outage Conference (ECC). (Linz, Austria, 15-17 Jul. 2015). Localization using Phasor Measurement Data in Tran- sient State. IEEE Transactions on Power Systems. Stolbova, , Backhaus, and Chertkov. Fault-induced delayed voltage recovery in a long inhomogeneous power-dis- Generous, N., G. Fairchild, A. Deshpande, S. Y. Del Valle, tribution feeder. 2015. PHYSICAL REVIEW E. 91 (2). and R. Priedhorsky. Global disease monitoring and forecasting with Wikipedia. . 2014. arXiv. Sulc, , Backhaus, and Chertkov. Optimal Distributed Con- trol of Reactive Power Via the Alternating Direction Goddard, , Klose, and Backhaus. Model Development Method of Multipliers. 2014. IEEE TRANSACTIONS ON and Identification for Fast Demand Response in Com- ENERGY CONVERSION. 29 (4): 968. mercial HVAC Systems. 2014. IEEE TRANSACTIONS ON SMART GRID. 5 (4): 2084. Sundar, K., H. Nagarajan, M. Lubin, L. Roald, S. Misra, R. Bent, and D. Bienstock. Unit Commitment with N-1 Hagberg, , and Lemons. Fast Generation of Sparse Random Secuity and Wind Uncertainty. Invited presentation at 185

Page 186

Power Systems Computation Conference 2016. (Ge- noa, Italy, 20-24 Jun. 2016). Tapia, A., N. LaLone, E. MacDonald, R. Priedhorsky, and M. Hall. Crowdsourcing rare events: Using curiosity to draw participants into science and early warning systems. . 2014. In International Conference on Infor- mation Systems for Crisis Response and Management (ISCRAM), 2014. (University Park, Pennsylvania, 18-21 May, 2014). , p. 135. Pennsylvania: ISCRAM. Turcotte, M.. Detecting localised anomalous behaviour in a computer network . 2014. In Advances in Intelligent Data Analysis XIII, LNCS 8819 . (Leuven, Belgium, 30 Oct - 1 Nov, 2014). , p. 321. Belgium: Springer. Yamangil, E., R. Bent, and S. Backhaus. Designing Resilient Electrical Distribution Grids. 2014. arXiv:1409. Yamangil, E., R. Bent, and S. Backhaus. Designing Resilient Electrical Distribution Grids. Invited presentation at Proceedings of the 29th Conference on Artificial Intel- ligence. (Allerton, Illinois, 1-3 Oct. 2014). 186

Page 187

Information Science & Technology Directed Research Continuing Project Scalable Codesign Performance Prediction for Computational Physics Stephan J. Eidenbenz 20150098DR Introduction develop includes: Polyhedral optimization methods to We will develop a performance-prediction capability that minimize data motion, load-balancing schemes vs. data relies on modeling architectural and algorithmic details movement requirements, static graph-based automated of a computational-physics code running on a specified parallelization approach for innermost loop computa- high-performance computing (HPC) architecture. This tions of physics domain codes, Just-In-Time Paralleliza- capability will fundamentally benefit a large number tion schemes and alternatives to checkpoint-restart of HPC code projects by providing techniques for rapid paradigm for failure tolerance. evaluation of novel algorithmic ideas; rapid evaluation of middleware “mapping” concepts; optimizing method & Benefit to National Security Missions architecture parameters for performance; and perhaps Our future computational physics code base for hydro- most intriguing semi-automated code generation for cru- dynamics and radiation transport are of fundamental cial subroutines. After the project ends, the performance importance to DOE and LANL. Maintaining LANL’s leader- prediction capability will be useable with minimal ad- ship role in national security will depend on our ability ditional effort for other computational physics domains to consistently provide performance improvement in and architectures of the generations after next, and for our codes for emerging architectures. Additionally, these decision support in HPC platform procurement to test techniques will be applicable to other application do- vendor proposals against request specifications. mains, in particular climate simulations and large-scale infrastructure simulations. Previous work related to AMD Building the performance prediction capability is our resulted in a successful transition of that project to the first goal, which will have a focus on models for state-of- DOE Office of Science BES, which will now support the the-art and conjectured architectures. Our second goal implementation and physics discovery work. Similarly in is to use the capability to assess novel solutions to the this project: If we are successful at identifying algorith- key exa-scale challenges of the data motion/movement, mic variations with substantial predicted performance high-parallelism, and fault tolerance applied to radiation gains, programmatic sources, e.g., ASC, may be willing to transport and hydrodynamics methods. support follow-on code development made possible by a successful LDRD project. Our initial algorithmic ideas that we will develop, model and assess for the applications include: Determining Progress optimal ways to hide communication and/or memory The performance prediction project has structured its latencies on specific devices for discrete ordinates mesh work along three focus areas: i) Develop and validate sweeps and analyzing tradeoffs of non-sweep-based al- hardware models, ii) develop and validate software mod- gorithms that increase independent work units and data els, and iii) design and implement the required scalable movement for deterministic radiation transport. Models discrete event simulation framework. The main results of the random access of memory and stochastic work are as follows. generation of a Monte Carlo fixed source calculation and various approaches to task-based parallelism for sto- Detailed models were developed for the Cielo super- chastic radiation transport. Rapid analysis of tree-code computer architecture and are in the process of validat- algorithms on different architectures for hydrodynamics. ing these models on benchmark applications. We have developed a model for the upcoming Trinity super- An initial set of middleware technologies that we will 187

Page 188

computer as far as its specification is known and will run important step in our ability to demonstrate the value in predictive runs of benchmark software and also radiation rapid performance prediction using discrete event simula- transport models as next steps. (iii) We have also built and tion software. Task 2 may also include the development already partially validated a few desktop machine architec- of a semi-formal process for evaluating the computational tures. intensity of existing production codes for easier implemen- tation within the discrete event simulator framework. A model of the radiation transport mini-app SNAP that we call SNAPSim was build and validated. We have designed Conclusion a similar model for the SPH oct tree code. We will vali- Based on results for different application domains and on date this in a next step. We have also built models for the preliminary analysis for transport and hydrodynamics, we POlyBench benchmark suite. In addition, a large molecular anticipate to achieve at least one order of magnitude per- dynamics project called SpecTADSim has been built in our formance improvement for stochastic radiation transport Simian framework that is ready for publication codes as these codes have undergone few optimization efforts for current node designs of many cores and more We developed a Lua or Python-based parallel discrete complicated memory hierarchies. Deterministic radiation event simulation system called Simian that we have open transport codes and smoothed particle hydrodynamics sourced. Three publications are under review for it. codes have seen some optimization and we expect our models to predict another 2-3X improvement in perfor- Future Work mance. For FY16 (year two of the DR), we plan to continue to im- prove upon the work performed in FY15. We have created Publications a Python-based discrete event simulation system called Chapuis, G., S. Eidenbenz, N. Santhi, and E. J. Park. Simian Simian that will serve as our primary simulation tool going integrated framework for parallel discrete event simu- forward. We consider our work going forward as catego- lation on GPUs. To appear in WinterSim 2015. (Hun- rized into three broader tasks: tington Beach, California, 6-9 Dec. 2015). Chapuis, G., S. Eidenbenz, and N. Santhi. GPU Performance

1. Continue to add more features to the hardware mod- Prediction Through Parallel Discrete Event Simulation els that describe the function of relevant architectures, and Common Sense. To appear in Valuetools 2015. including internode data movement (interconnect (Berlin, Germany, 14-16 Dec. 2015). modeling), intranode data movement (memory hierar- chy and cache structure modeling), failure resilience, Park, E. J., S. Eidenbenz, N. Santhi, G. Chapuis, and B. and power consumption. Each new addition to the Settlemyer. Parameterized benchmarking of parallel hardware modeling framework will be generalized discrete event simulation systems: communication, and parameterized for simple extension to different computation, and memory. To appear in The Winter Simulation Conference. (Huntington Beach, CA, 6-9 architectures. Dec. 2015).
2. Expand our catalog of physics application domains to include stochastic radiative transport. Consider the Santhi, N., S. Eidenbenz, and J. Liu. The Simian Concept: inclusion of different computational kernels associ- Parallel Discrete Event Simulation with Interpreted ated with radiative transport and hydrodynamics Languages. To appear in 2015 Winter Simulation Con- applications to ensure we are capturing all the relevant ference. (Huntington Beach, CA, 6-9 Dec. 2015). computational challenges associated with these types Zamora, R. J., A. F. Voter, D. Perez, N. Santhi, S. M. Mnisze- of numerical simulations. Begin exploring parallel solu- wski, S. Thulasidasan, and S. Eidenbenz. Discrete Event tion algorithmic alternatives and their suitability for Performance Prediction of Speculatively Parallel TAD. our existing suite of hardware models. SIAM Journal on Scientific Computing.
3. Hardware and software model validation will be an important part of our second year activities. Ensuring that we have sound models will allow us to explore variations with confidence that alternatives are ana- lyzed fairly and warrant further examination. Validation in the sense of Task 3 refers to the comparison of relevant numerical simulation with an actual code on an actual cluster compared to the simulated result. It is an 188

Page 189

Information Science & Technology Directed Research Continuing Project Cyberphysical Systems and Security Scott N. Backhaus 20150215DR Introduction Benefit to National Security Missions Many critical infrastructures, e.g., electric power grids Cyberphysical systems are crucial to the reliable delivery and natural gas pipelines, are cyberphysical systems— of energy, water, and other resources, creating strong physical networks whose flows are governed by the connections to LANL’s National Security and Energy Se- law of physics but are regulated by a control system curity missions including those US Government agencies coupled to cyber networks that transmit the information that monitor, analyze, or develop technology for critical required to optimize and control the physical networks infrastructure, e.g. DOE Office of Electricity, DOE Energy for reliability and efficiency. Crucial to the defense of Efficiency and Renewable Energy, DOE Fossil Energy, and cyberphysical systems is the concept of intrusion detec- Department of Homeland Security. tion and localization, however, this is not sufficient. If a system is under attack, it cannot simply be brought These systems are composed of interdependent complex down to purge the attacker. Doing so would grant the networks described by disparate mathematical models attacker his objectives, i.e. wide spread denial of the creating scientific challenges that go well beyond the services from the cyberphysical network. A proportional modeling and analysis of the individual networks. Inte- response is required. gration of approaches from complex systems, statistical physics, machine learning, and optimization will create The overarching objective of this work is to develop ad- a holistic and uniquely LANL approach that is able to vanced approaches to cyberphysical defense, which can localize attacks and develop appropriately scaled pro- be broken down into the following goals: portional responses. Both DOE and DHS are creating and running new programs in cyberphysical systems a • Detect and localize an attacker within the cyber- creating significant programmatic pull. Many efforts in physical network without reference to a predefined these programs are based on best practices or standard attack vector. Attackers are creative and detection engineering analyses. Our approach based on the sci- methodologies cannot be based on scripted attack ence of complex networks and probabilistic methods will scenarios. advanced detection and defense capabilities to the US • Develop algorithms for proportional response and Government agencies identified above. Components of designing resilient cyberphysical networks—Pro- the NNSA Complex, including our own Laboratory, are portional responses to attacks should weigh the cyberphysical systems with complex security issues, and estimated network degradation from the attack and this work will build capabilities for protecting the Com- the degradation caused by the response. Resilient plex. cyberphysical designs will greatly improve this bal- ance in favor of the cyberphysical operators. Progress • Deploy, demonstrate and validate—The efficacy • Established automated process for collecting large of cyberphysical intrusion detection and response data sets and building a data base for collection of methodologies will improve at much faster rates physical data from the NSSB when their development and refinement is closely • In process and near completion: Established a com- coupled with real-world experimentation that vali- munications monitoring capability to detect com- dates strengths and reveals weaknesses. munication events on the control/serial network in the NSSB. The combination of this data and the 189

Page 190

data in #1 forms a unique data set for our research and the rest of the community investigating cyberphysical systems. • Exploring two signal processing/statistical methods for removing background signals and tends to extract fluctuations in the physical data time series. • Exploring two methods for reconstructing graphical models of the NSSB cyberphysical system from correla- tions in the physical data streams. Future Work Planned milestones by FY16 year end: • Begin intrusion detection experiments on NSSB • Develop machine learning algorithms for graphical models of continuous variables and apply to data from NSSB test bed. Extend methods for continuous vari- ables to discrete variables • Build an interdependent network model of the NSSB suitable for network optimization and optimal network defense task in BY2. Conclusion Information theory: We will develop algorithms for online learning of graphical models using operational time-series data collected from sparse measurements on cyberphysical systems. Network optimization: We will develop new algorithms for decomposing and efficiently solving difficult multi-level optimization problems that involve discrete and continu- ous variables. Experimental validation: We will develop a unique cyber- physical test bed where we will deploy and test the tools and methods developed in the information theoretic and optimization thrusts. 190

Page 191

Information Science & Technology Directed Research Final Report Disruptive Innovation in Numerical Hydrodynamics Jacob I. Waltz 20130005DR Abstract dynamics problems – often referred to as “hydro codes” We describe the accomplishments and impact of the –solve these equations numerically given some specific LDRD Directed Research Project “Disruptive Innovation problem description that includes geometry information, in Numerical Hydrodynamics,” executed over the time material densities, initial pressures and temperatures, frame FY2012-FY2015. The goal of this project was to etc. In this manner the governing equations are trans- develop a new class of numerical methods applicable formed from a continuous form with no closed-form to shock hydrodynamics phenomena and designed for analytic solution, to a discrete system that can be solved advanced computer architectures. The research led to numerically on a computer. The approach used to carry several key breakthroughs and conclusions, which are out this transformation is the numerical method. expected to influence a variety of LANL programs. Historically, these equations have been solved using a discrete representation of the problem based on quad- Background and Research Objectives rilateral (in 2D) or hexahedral (in 3D) cells: the spatial This project was fundamentally motivated by two domain of the problem is subdivided into a collection of observations: first, a decreased rate of progress in the non-overlapping cells (referred to in toto as the compu- development of modeling and simulation tools by the tational ”mesh” or “grid”), and the governing equations Advanced Simulation and Computing (ASC) program; are solved numerically for each cell or alternatively the and second, the imminent (at the time of the proposal) cell vertices. An additional choice is the reference frame large-scale transition in High Performance Comput- in which the governing equations are solved. In the ing (HPC) from conventional to so-called advanced so-called Eulerian description, the governing equations or emerging architectures. The high-level goal of this are written for and solved in a reference frame that is at project, therefore, was to develop a novel simulation rest relative to the material, i.e. the “Laboratory” frame. methodology that combined algorithmic advances with In the so-called Lagrangian description, the govern- effective utilization of new hardware systems. Taken to- ing equations are written for and solved in a reference gether, the combined capability was expected to enable frame that moves with the material, i.e. the “co-moving” a dramatic leap forward in predictive capability. frame. A third description, referred to as the Arbitrary The particular focus of this project was numerical meth- Lagrangian-Eulerian (ALE) approach, also is possible. In ods for the simulation of shock hydrodynamics prob- the ALE description the equations are written for and lems, which we define as the study of material response solved in a reference frame that moves arbitrarily rela- to shock waves and associated physical phenomena. tive to the laboratory frame and the material; the ALE This field of physics is central to a broad portion of the description therefore contains both the Eulerian and LANL mission space (e.g. Inertial Confinement Fusion, Lagrangian descriptions as limiting cases, and can be Stockpile Stewardship, and dynamic experiments) as viewed as the most general. well as numerous fundamental science and engineering ALE methods based on triangular (in 2D) or tetrahedral areas (e.g. astrophysics and materials science). Math- (in 3D) cells have been explored for LANL applications, ematically, this physics is described by a set of coupled, but had not been successfully developed for a number nonlinear partial differential equations. These equations of reasons; detailed technical explanations can be found are supplemented by constitutive models, such as an in both the original proposal and in two of the journal equation of state, that describe material-specific char- articles produced by the project: N Morgan et al, J Com- acteristics. Computer codes that simulate shock hydro- 191

Page 192

put Phys 281:614-652 (2015); and J Waltz et al Int J Nu- proach involved the extension of a proven Eulerian method mer Meth Fluids 76:129-146 (2014). However, numerical to the more general ALE description, and subsequently methods based on triangular and tetrahedral meshes have into a Lagrangian-like regime. The major uncertainty with some key advantages over quadrilateral and hexahedral this approach was whether or not the method would be approaches, the most important of which for this proj- susceptible to large errors when applied to shock hydro- ect are rapid problem definition using commodity tools; dynamics problems. The second approach involved the simpler and more easily optimized data structures; and extension of a modern Lagrangian method for quadrilat- potentially higher accuracy. For these and other reasons, eral cells to the more general case of tetrahedral cells, and tetrahedral-based approaches have gained widespread ac- subsequently into the more general ALE description. The ceptance in the Computational Fluid Dynamics (e.g. aircraft major uncertainty with this approach was whether or not aerodynamics) and Computational Electromagnetics com- the method could be successfully adapted to tetrahedral munities, but were not yet demonstrated to be suitable for cells. shock hydrodynamics applications. Ultimately, both approaches proved to be successful, and The primary technical goal, therefore, was to develop a together were described in a total of six journal articles. tetrahedral-based numerical method for shock hydrody- Three of these articles have been ranked among the top namic simulations that overcame the historical difficulties downloaded articles for their respective journals. More im- associated with such methods. However, the imminent portantly, the tetrahedral-based methods developed in this transition to advanced computer architectures created an project were demonstrated to be 1000-10000 times faster additional constraint, namely, that any new method should than the methods used in production ASC codes, when also be designed a priori for these new platforms. With- measuring time to solution for a given level of accuracy. out taking this constraint into account, any new method The primary focus in FY14 was effective utilization of might become a dead-end path on future HPC technology. advanced architectures. The original target architecture Likewise, deployment of existing algorithms on emerging described in the proposal was the Graphics Processing Unit architectures would not fundamentally change the predic- (GPU). However, the target architecture was changed to tive capability of those algorithms. With this perspective in the Intel Many-Integrated-Core (MIC) architecture based mind, a secondary technical goal was that the new method on the anticipated results (which were later confirmed) of be designed a priori for advanced computer architectures. the Trinity supercomputer procurement. The basic ap- Based on these goals and the intended future application proach for this work was to take proven techniques for of the research results, the original project proposal de- conventional computer architectures, and attempt to apply fined the following three milestones (one per fiscal year): them to the MIC-type systems. An initial performance comparison of different data structures used by the most FY2013: Demonstrate a tetrahedral-based numerical compute intensive portions of the numerical methods method for single-material shock hydrodynamic problems was first performed (L. D. Risinger et al, LA-UR-13-28518, in the ALE reference frame 2013). This was followed by a more detailed study of over- all application performance, which was eventually pub- FY2014: Demonstrate effective use of advanced computer lished in J. Waltz et al, Int J Numer Meth Fluids 77:319-333 architectures (2015). These studies led to two important conclusions: first, although the vector performance of the Intel MIC is FY2015: Extend the single-material tetrahedral-based ALE often cited as a key advantage over conventional archi- method to multiple materials with strength tectures, these studies demonstrated that effective use of The order of these milestones was based on the level of threads was key to maximizing performance on the Intel perceived risk associated with each task. In what follows MIC; second, data structures and optimization techniques we describe the R&D approach and the accomplishments originally developed for conventional architectures could for each of the above. make effective use of the Intel MIC with little to no modifi- cation. Scientific Approach and Accomplishments Due to a budget cut of approximately $200k in FY14, de- The task with the most perceived risk in this proposal in- tailed performance optimization studies originally planned volved the development of a tetrahedral-based numerical for FY15 were deferred so that hierarchical parallelism method for single-material shock hydrodynamic problems could be explored. Parallelism in shock hydrodynamic in the ALE reference frame. Two parallel R&D approaches codes is typically based on the concept of domain decom- with different risks were therefore pursued. The first ap- position: the spatial extent of the problem is decomposed 192

Page 193

into sub-domains, and the governing equations are solved accurate level set method for tetrahedral grids. Unfortu- in parallel on each sub-domain by separate compute nately, although the technical work was completed in FY15, units. Domain decomposition has been the standard ap- documentation of the work in journal articles was not yet proach for ASC codes since the advent of the ASCI era. An complete by the time of this writing; however, the ASC pro- alternative parallelization strategy is the shared memory gram has a high level of interest in this work. approach. In this method, a single memory image of the problem (or a subset of the problem) is shared by multiple Impact on National Missions compute units. Each unit then operates on a portion of The impact of this work on National Missions has yet to be this memory image using e.g. threads. Both approaches fully determined. The ASC Program Office declined to fund have their relative advantages and disadvantages, and are a proposal to develop a new production simulation tool appropriate for different computing regimes. The relevant based directly on the results of this research project, but consideration for this work however was that advanced has agreed to support a low-level of follow on research to computer architectures were expected to require a hybrid address certain remaining questions – in particular, how to approach that employed domain decomposition across simulate contact mechanics within the tetrahedral-based major compute units, and shared memory within the formulation. The outcome of these studies could potential- major compute units. For example, the Intel MIC architec- ly lead to a more direct adoption of this project’s research ture consists of 60-70 individual compute cores, each of results. The results of this project also are expected to im- which can further support up to four individual processes pact technical choices made within existing ASC projects. (threads). On a large system composed of Intel MIC chips For example, the new level set methods are a strong candi- (i.e. a Trinity-like system), one could envision using domain date for implementation into existing simulation tools, and decomposition to parallelize across the different MIC chips, the performance studies on the Intel MIC architecture are and using threads to parallelize within each MIC chip. Do- expected to influence how existing tools are ported to the main decomposition was therefore implemented into the Trinity Phase 2platform. computer software in use for this research, and a detailed performance study was undertaken to assess the perfor- The results of this project also have generated strong inter- mance characteristics of the hybrid approach. While the est in communities beyond ASC. These include the nuclear results of this analysis have not been published (again due detection and seismo-acoustic coupling community; the to resource limitations), they were described in various nuclear blast effects modeling community; the high-explo- presentations (e.g. LA-UR-15-21734). The important con- sives experimental community; and the DOE/DOD Joint clusion from these studies is that on the Intel MIC, hybrid Munitions Program. Discussions with all of these programs parallelization strategies were faster than a pure domain are ongoing and are expected to result in future mission decomposition approach. contributions. The final milestone, extension of the tetrahedral-based Publications method to multiple materials with strength, was consid- Canfield, T. R., M. R. J. Charest, N. R. Morgan, L. D. Risinger, ered from the beginning of the project to be lower risk J. Waltz, and J. G. Wohlbier. Simulation of multi-ma- than the previous two milestones. The approach used for terial flows using a finite element Riemann solver and adaptive unstructured grids. Presented at Conference this milestone was to extend techniques demonstrated on Numerical Methods for Multi-Material Flows. (San for Eulerian methods to the ALE reference frame. This Francisco, CA, 2-6 Sep. 2013). extension was successfully demonstrated without strength for some very basic initial test problems, but the work Canfield, T. R., N. R. Morgan, L. D. Risinger, J. Waltz, and was somewhat impacted by staff availability and a $200k J. G. Wohlbier. Manufactured solutions for the three- budget reduction in FY15. Progress on demonstrating the dimensional Euler equations with relevance to inertial methodology with material strength also was impacted by confinement fusion. Presented at American Society of staff availability. Mechanical Engineers Verification and Validation Sym- posium. (Las Vegas, NV, 22-24 May 2013). On the other hand, some unexpected research break- Long, A. R., R. G. McClarren, J. Waltz, and J. G. Wohlbier. throughs were accomplished in FY15 in the area of level Implicit Monte Carlo adaptions for tetrahedral meshes set methods. Level set methods are a class of numerical with node-based unknowns. Presented at American techniques used to compute the motion of surfaces and Nuclear Society Joint International Conference on interfaces. Examples relevant to this work include both Mathematics and Computation. (LaGrange Park, IL, material interfaces and high explosive burn fronts. This April 2015). project was able to demonstrate, for the first time, a fast, 193

Page 194

Long, A. R., T. R. Canfield, M. R. J. Charest, N. R. Morgan, J. Waltz, J., N. R. Morgan, T. R. Canfield, M. R. J. Charest, L. Waltz, and J. G. Wohlbier. Implicit Monte Carlo discreti- D. Risinger, and J. G. Wohlbier. A 3D finite element zations for tetrahedral meshes with node-based un- arbitrary Lagrangian-Eulerian method for shock hydro- knowns. Presented at Nuclear Explosives Code Devel- dynamics on unstructured grids. 2014. Computers & opers Conference. (Los Alamos, NM, 20-24 Oct. 2014). Fluids. 92: 172. Morgan, N. R., J. Waltz, D. E. Burton, M. R. J. Charest, T. Waltz, J., N. R. Morgan, T. R. Canfield, M. R. J. Charest, and R. Canfield, J. G. Wohlbier, J. Bakosi, and A. R. Long. A J. G. Wohlbier. A nodal Godunov method for Lagrang- 3D Arbitrary Lagrangian Eulerian (ALE) hydrodynamic ian shock hydrodynamics on unstructured tetrahedral approach for tetrahedral meshes. Presented at Confer- grids. 2014. International Journal for Numerical Meth- ence on Numerical Methods for Multi-Material Fluid ods in Fluids. 76: 129. Flow. (Wurzburg, Germany, 7-11 Sep 2015). Waltz, J., N. R. Morgan, T. R. Canfield, M. R. J. Charest, Morgan, N. R., J. Waltz, D. E. Burton, M. R. J. Charest, T. R. and J. G. Wohlbier. A Godunov ALE method for 3D un- Canfield, and J. G. Wohlbier. A Godunov-like point-cen- structured grids. Presented at Nuclear Explosives Code tered essentially Lagrangian hydrodynamic approach. Developers Conference. (Los Alamos, NM, 20-24 Oct. 2015. Journal of Computational Physics. 281: 614. 2014). Morgan, N. R., J. Waltz, D. E. Burton, M. R. J. Charest, T. R. Waltz, J., T. R. Canfield, M. R. J. Charest, N. R. Morgan, L. D. Canfield, and J. G. Wohlbier. A point-centered arbitrary Risinger, and J. G. Wohlbier. Operator splitting and time Lagrangian Eulerian hydrodynamic approach for tetra- accuracy in Lagrange plus remap methods. Presented hedral meshes. 2015. Journal of Computational Phys- at Conference on Numerical Methods for Multi-Materi- ics. 290: 239. al Fluid Flow. (San Francisco, CA, 2-6 Sep. 2013). Morgan, N. R., J. Waltz, D. E. Burton, T. R. Canfield, L. D. Waltz, J., T. R. Canfield, N. R. Morgan, L. D. Risinger, and J. Risinger, J. G. Wohlbier, and M. R. J. Charest. A Godun- G. Wohlbier. Verification of a three-dimensional un- ov-like point-centered ALE finite element hydrodynam- structured finite element method using analytic and ic approach. Presented at Conference on Numerical manufactured solutions. 2013. COMPUTERS & FLUIDS. Methods for Multi-Material Fluid Flow. (San Francisco, 81: 57. CA, 2-6 Sep. 2013). Waltz, J., T. R. Canfield, N. R. Morgan, L. D. Risinger, and Morgan, N. R., J. Waltz, D. E. Burton, T. R. Canfield, M. R. J. G. Wohlbier. Manufactured solutions for the three- J. Charest, and J. G. Wohlbier. Application of a multi- dimensional Euler equations with relevance to Inertial dimensional approximate Riemann solution to point Confinement Fusion. 2014. Journal of Computational centered hydrodynamics on tetrahedral meshes. Pre- Physics. 267: 196. sented at Nuclear Explosives Code Developers Confer- ence. (Los Alamos, NM, 20-24 Oct. 2014). Wohlbier, J. G., L. D. Risinger, T. R. Canfield, M. R. J. Charest, N. R. Morgan, and J. Waltz. Programming for modern Risinger, L. D., J. Waltz, T. R. Canfield, M. R. J. Charest, N. R. architectures in the CHICOMA hydrocode. Presented at Morgan, and J. G. Wohlbier. Threading of indirect gath- Conference on Numerical Methods for Multi-Material er-scatter algorithms for 3D unstructured flow solvers. Fluid Flow. (San Francisco, CA, 2-6 Sep. 2013). 2013. Internal LANL report LA-UR-13-28518. Waltz, J.. GPU and multicore acceleration of a 3D un- structured mesh Eulerian-AMR hydrocode. Presented at Nuclear Explosives Code Developers Conference. (Livermore, CA, 22-26 Oct. 2012). Waltz, J.. Multimaterial hydrodynamics simulation for advanced architectures. 2015. Presentation given to DOE/DOD Joint Munitions Program 10th Annual Users’ Forum, Arlington, VA (2015); LA-UR-15-21374. Waltz, J., J. G. Wohlbier, L. D. Risinger, T. R. Canfield, M. R. J. Charest, A. R. Long, and N. R. Morgan. Performance analysis of a 3D unstructured mesh hydrocode on multi- and many-core architectures. 2015. Internation- al Journal for Numerical Methods in Fluids. 77: 319. 194

Page 195

Information Science & Technology Directed Research Final Report Empowering the Expert: Machine Learning with User Intelligence Reid B. Porter 20130013DR Abstract derstanding and exploiting the specialized datasets with Experts are very good at understanding and exploit- which they work; the problem is that there is too much ing domain-specific data, but in nearly all science and of this data for them to look at. Data processing tools defense applications, there is too much data for experts help to clean up, filter and identify the most relevant to look at it all. Machine Learning provides tools that can subsets of data, but recent developments in machine clean up, filter, and identify the most relevant subsets learning indicate that there is an opportunity to enhance of data. However, because these tools lack domain- the efficiency and accuracy of this process, by involving specific input, they are not as accurate as they could be the user in a more interactive dialog. and experts do not fully trust them. Machine Learning Traditional machine learning employs a fixed set of with User Intelligence advances the field of Interactive labels, supplied up-front by the user, to optimize data Machine Learning (IML) to provide a middle ground that processing tools before they are applied to a larger data- combines the complementary strengths of experts and set or data archive. This approach has proven successful machines in data analysis. both in theory (rigorous proofs) and in practice (com- This project has made a wide range of contributions to mercial applications). Machine Learning with User Intelli- Interactive Machine Learning that span theory, algo- gence pushes the state of the art in two main directions: rithms and applications. These contributions have been From up-front learning to interactive learning: Tradition- recognized in the broader research community with al machine learning tools stop when the human steps multiple best paper and invited speaker awards, and in. This means the expert is still the bottleneck in data acceptance to the highest impact machine learning exploitation: spending too much time in tedious and conferences. The project’s contributions have also been repetitive post-processing tasks and not enough time recognized in programmatic activities, and a number of on validation and research. Interactive machine learning new projects are underway to transition the theoreti- starts when the human steps in, and tools are a force cal and algorithmic advances made by the project into multiplier, where the user’s intelligence is the force to be practical software and operational tools. multiplied, not an expense to be minimized. In science and defense, we have critical applications From label learning to relational learning: Machine where experts must understand and validate tool out- learning with user intelligence formalizes the interactive puts before making a decision. Fortunately, experts are dialog with relational objects, stated in terms of motifs, also highly motivated and willing to interact with the clusters, alignments, matches and other generalizations tools in a much more sophisticated way than current of standard labels. By providing additional interface systems allow. This project has shown that Los Alamos, “bandwidth” between the domain expert and the data where experts in machine learning can work side-by-side processing tool (and thus creating a better expression of with experts in science and defense, is well positioned the dialog in terms of Graphical Models), higher quality to exploit the more motivated and informed user base associations are produced, and more effective tools can and address the fundamental challenges holding back be built. machine learning in specialized domains. The project has three main objectives: Background and Research Objectives Scientists and intelligence analysts are very good at un- • Identify and formalize new classes of user-interac- 195

Page 196

tion that would benefit experts advance the theory of Belief Propagation. Belief Propaga- tion is one of the most important algorithms for Graphical • Develop learning algorithms that translate these inter- Models because it provides distributed, scalable prediction actions into tools based on message passing. As science and defense applica- tions grow in size, the need for message passing algorithms • Apply these tools to analysis bottlenecks in the ex- increases, and our work currently defines the largest prov- pert’s application able correct class of problem that can be predicted with Belief Propagation. Two papers on this topic were accepted Scientific Approach and Accomplishments at the high impact Neural Information Processing Systems Over the course of this project Interactive Machine Learn- (NIPS) conference [2][3]. ing has gained momentum within the broader research community, and due to the timing of this project, we have Prediction is a fundamental component of learning and been privileged to have played a role in this emerging field. therefore advances in prediction feed directly into ad- One of our first accomplishments was to publish a com- vances in learning. In addition, our team was inspired by prehensive review article in the IEEE, Computers in Science the specific theoretical approach used to advance Belief and Engineering [1]. We were invited to present this mate- Propagation, and this led to a number completely novel rial at one of the first workshops on Interactive Machine research directions for learning that we hope to continue Learning, which was held in the spring of 2013. In the in future work [4][5]. spring of 2014, we hosted our own workshop on Interac- tive Machine Learning at Los Alamos. This event attracted Supervised Semantics: Scientists and data analysts are very leading researchers from academia (Caltech, Cornell, CMU, good at recognizing patterns and identifying important Indiana and UCSC), industry (Microsoft, Google) and gov- relationships in the data they are analyzing, but generally ernment (JPL) and we continue to benefit from students lack mechanisms to annotate relationships in a way that and collaborations from this event. Interactive Machine machine learning can understand. Our accomplishments Learning now appears to be a regular workshop in main- move IML far beyond labels, and include: stream machine learning conferences, and our team • Semantic interactions (such as merging and splitting) presented at the most recent workshop on Interactive Data were formalized as prediction in a Graphical Model Exploration and Analysis at the Knowledge Discovery and and learning algorithms developed to enable a new Data Mining (KDD) conference. type of interactive image analysis tool [6]. These tools Our review article provided a unified view of Interac- have been applied to applications in nuclear material tive Machine Learning research and also a road-map for forensics, geo-spatial analysis and bio-medical image our project (and the research community). The roadmap analysis (see Figure 1 and 5), and provide significant articulates two main technical thrusts, which we call the benefit to a number of expert users. This is discussed ‘Training Dialog’ (moving from up-front learning to interac- in more detail under impact. tive learning) and the ‘Training Vocabulary’ (moving from • New connections were made between Graphical label learning to relational learning). Common to both of Models and a number of widely used hierarchical im- these thrusts is the need to balance theory and practice, age segmentation methods including the Watershed and our project contained a diverse range of staff (from algorithm [7]. These connections enable these popular statistical physicists to application specialists) in order to methods to be more easily adapted and evaluated for advance both the underlying mathematics, and the specific particular applications and a software library imple- tools used by end-users in particular applications. We now menting these ideas was developed. describe the projects main contributions in more detail under these two broad categories of work. • New mechanisms to inject an expert’s ‘relational ad- vice’ into Graphical Models were developed that help Training Vocabulary control what parts of the model are learnt from data, Graphical Models provide a general purpose framework for and what parts are based on prior knowledge. To dem- learning and representing dependencies among variables onstrate the generality of the approach we evaluated in multivariate data. Graphical Models are key to expand- it on a wide range of applications from image analysis, ing the training vocabulary because they provide a balance to autonomous driving. This work was accepted for of rigor and flexibility on the theory front and domain presentation at the high impact AAAI (Advances in Arti- adaptability on the interactivity front. ficial Intelligence) conference [8]. Theory: A significant theoretical accomplishment was to 196

Page 197

• A statistical framework for estimating the uncertainty dependencies that occur as experimental parameters associated with the Graphical Models was developed are varied than they are in the dependencies found in for materials image analysis. This framework contex- a single experiment. We developed new methods for tualizes the variability in feature patterns of materials this use-case and demonstrated its utility with experi- and the imaging process, provides greater insight into mental data that included co-collected time series and the performance of different algorithms, and is used to video data (see Figure 3) [12]. identify where user feedback can improve accuracy [9]. Figure 2. (a) Mars Rover mast camera image of rock with ChemCam sampling region in white. (b) Image from ChemCam Figure 1. Supervised Semantic tools learn more from experts microimager with sampling location indicated by the red arrow. than standard machine learning tools because they understand (c) Graphical Model of shot-to-shot depth profile (30 shots) is semantic interactions like splitting and merging. These interac- consistent with a dusty surface layer and increased mixing with tions are encoded into the Graphical Model and then propagated depth. across the larger data set or data archive. Model Discovery In supervised semantic tools, the approximate structure of the Graphical Model is known, and the expert interaction is encoded into parameters. However, in other applications the potential dependencies within data are unknown, and the expert’s goal is to identify these dependencies with the Figure 3. Left) Interface to enable users to explore and search help of data. In this type of application prediction accuracy for differences in dependencies in experimental datasets. Right) is often less important than the Graphical Model structure, Radio Frequency (RF) signal data and video associated with two since the structure can identify dependencies that the sci- settings of an experiment parameter. entist may not have thought of. Our accomplishments with respect to these types of tools included: Training Dialog • A new computationally efficient mechanism to incor- One of the most common interactions between humans porate expert knowledge into Graphical Model struc- and machine learning is data triage: machine learning tools ture learning was developed. It enables experts to are applied to large datasets to identify subsets of data, encode known dependencies, and known dependency which are then passed to the expert to discard or follow directions, and it won a best poster award at the Los up. Alamos Postdoc Research Day [10]. Theory • We worked with experts from the ChemCam science Another significant theoretical accomplishment formalized team to identify and analyze chemical depth trends anomaly detectors and provided quantitative methods for in laser-induced breakdown spectroscopy (LIBS) data. evaluating their performance. Broadly speaking, anomalies There have been over 300,000 observations to date, are rare and ‘interesting’, and this idea can be placed on and each observation has over 600 channels, making firmer footing by defining anomaly detection in terms of manual evaluation of depth trends almost impossible. characterizing the background. This approach illuminates Our model discovery tools were used to summarize how different anomaly detection methods relate, and pro- and make this data more interpretable to the expert vides tools to characterize the quality of anomaly detection which helped elucidate rock surface weathering (see in the absence of ground truth [13]. We were invited to Figure 2) [11]. present this material as a plenary talk at two high impact Hyperspectral Imaging conferences. • Experts are often more interested in the changes in 197

Page 198

Hemi-Supervised Learning: Anomaly detection provides the framework for a wide range of data triage applications, but does not provide a mechanism for incorporating user feedback (or triage results). Since ‘interesting’ is rare, this feedback corresponds (primarily) to experts labeling data as ‘not-interesting’ (note, when interesting data is found, experts typically move into more focused analysis where ‘Supervised Semantics’ and ‘Model Discovery’ tools can be more helpful). We developed a new algorithm for incorpo- rating one-sided negative feedback into anomaly detec- tion, called Hemi-Supervised Expectation Maximization Figure 5. This project has produced a set of general purpose, [14]. We also created and released a large open dataset to yet easily tailored tools, that can help experts automate time- better compare interactive anomaly detection methods, consuming analysis in a number of advanced material applica- and found our method was one of the few to consistently tions. beat random sampling over large numbers of problem instances (previous comparisons typically involved a small Impact on National Missions number of hand selected problems, and statistical signifi- Our accomplishments empower experts to exploit special- cance was hard to evaluate). ized data in less time, and with greater accuracy, both by automating tedious and repetitive tasks and by enabling User Sampling Bias them to focus on validation and higher-level objectives. Another significant accomplishment was motivated by Specific applications successfully transitioned to program- image analysis tools where we observed experts interact- matic funding during the course of the project include: ing with machine learning in a different way. Instead of the tool going first, the expert would go first. They would Nuclear Material Forensics label a small amount of ‘interesting data’ and then iterate The ‘supervised semantic’ tools developed in this project with the machine learning system to find more. We called are now in operational use within a number of nuclear this unexplored form of Interactive Machine Learning forensics laboratories throughout DOE, FBI and NIST. The ‘User Driven Sampling Bias’ and we showed how expert Domestic Nuclear Detection Office (DNDO) within the bias, interacting with small amounts of data, can consis- Department of Homeland Security (DHS) has been funding tently outperform automated methods with all of the data software maturation since the summer of 2014, and beta (see Figure 4). This work received a Best Paper award at versions of the tools were recently delivered to DNDO for the SPIE/IS&T Visualization and Data Analysis Conference external Independent Validation and Verification. LANL (VDA-2014) [15]. In related work, our team used Amazon now has an ongoing program with DNDO to support the Turk to investigate human cognitive bias in the expression image analysis needs of approximately 40 scientists in of numerical uncertainty. This led to an algorithm for find- research activities and forensic exercises. ing new linguistic hedges in unstructured text [16] and also to recommendations and methods for quantifying uncer- Human Guided Sense Making tainty in reporting qualitative analysis results for forensic The Defense Threat Reduction Agency (DTRA) has recog- evidence. nized semantic interactions as a key technology for multi- modal data enrichment. In the summer of 2015, they funded LANL to develop a prototype for the geospatial image analysis domain, as part of their Advanced Analytics Program to strengthen national capabilities in countering weapons of mass destruction (CWMD). Interactive Image Search and Triage Interaction was also recognized by intelligence community sponsors as a low-risk way to incorporate image search tools into operational databases. We were funded to pro- Figure 4. Left) The expert selects pixels to label (red or green) in vide a baseline capability for image search and our tools iteration 1 producing a prediction result with error 0.108 (red dot became operational in the summer of 2014. We are now in plot). Right) Over 9 iterations the experts adds more training responding to requests for enhancements and extensions data (totaling 1% of the data) and the machine learning system outperforms the machine learning system trained with all of the to meet the specific needs of particular datasets and end data (red dashed line). users. 198

Page 199

The project also explored a number of other mission rel- edge Based Probabilistic Logic Learning. 2015. (Austin, evant applications that we believe are high value targets Texas, 25-30 Jan. 2015). p. 83. Austin, Texas: AAAI. for follow on work. 9. Director, H., J. Gattiker, and R. Porter. A multi-scale noise model for estimating uncertainty in computer vi- Advanced Materials: Building on our success with interac- sion algorithms. Technometrics. tive image analysis tools for nuclear material forensics we worked with experts who are acquiring increasing volumes 10. Oyen, D., K. Sentz, B. Anderson, and C. Anderson-Cook. of image data for stockpile stewardship and the weapons Bayesian discovery of bayesian networks with order enterprise (Figure 5). Experts need these images to derive priors . (Phoenix, AZ, Feb). quantitative measurements, and define uncertainty, but 11. Oyen, D., R. Porter, and N. Lanza. Discovering Compo- there is also increasingly need for prediction, and anomaly sitional Trends in Mars Rock Targets from ChemCam detection, as part of process monitoring and certified ma- Spectroscopy and Remote Imaging. (Washington DC, terial production. The tools developed under this project 13-15 Oct ). provide a much needed middle ground between commer- cial tools, which experts find are not quite right for their 12. Oyen, D., R. Porter, and K. Sentz. Interactive compara- particular problem, and custom analysis tools, which must tive analysis for multi-modal data exploitation and fu- be developed at significant cost in time and resources. sion. (Washington DC, 13-15 Oct). 13. Zimmer, B., and R. Porter. Hemisupervised Expectation Cybersecurity: Improving anomaly detection accuracy and maximization for rare category detection . ACM Trans- maximizing the limited expert feedback has the potential actions on Knowledge Discovery From Data. of making a significant impact in cybersecurity applications and we hope to identify opportunities for future work in 14. Harvey, N., and R. Porter. User-Driven Sampling Strate- this area. gies in Image Exploitation. 2014. (San Francisco, 23 Dec). p. 90170B. San Francisco: SPIE. References 15. Sentz, K., and S. Ferson. Natural language of uncertain-

1. Porter, R., J. Theiler, and D. Hush. Interactive Machine ty: numeric hedge words . 2015. International Journal Learning in Data Exploitation. 2013. IEEE Computing in of Approximate Reasoning. 57 (Feb.): 19. Science and Engineering. 15 (5): 12. Publications
2. Ahn, S., S. Park , M. Chertkov, and J. Shin. Minimum weight perfect matching via blossom belief propaga- Ahn, S., S. Park , M. Chertkov, and J. Shin. Minimum weight perfect matching via blossom belief propagation. To tion. (Montreal Canada, 7-12 Dec). appear in Proceedings of Neural Information Process-
3. Shin, J., A. E. Gelfand, and M. Chertkov. A Graphi- ing Systems (NIPS) 2015. (Montreal Canada, 7-12 Dec). cal Transformation for Belief Propagation: Maximum Weight Matchings and Odd-Sized Cycles . 2013. (Lake Burr, T., and A. Skurikhin. Conditional Random Fields for inverse problems in structured data. 2015. Algorithms. Tahoe, 5-10 Dec). p. 2022. : . 8: 466.
4. Porter, R., and B. Zimmer. Links between binary clas- sification and the assignment problem in ordered Chertkov, M., A. E. Gelfand, and J. Shin. Loop Calculus and Bootstrap-Belief Propagation for Perfect Matchings on hypothesis machines . 2015. (San Francisco, 2015). p. Arbitrary Graphs. 2013. In International Meeting on doi:10.1117/12.2083994. San Francisco: SPIE/IS&T. ‘Inference, Computation, and Spin Glasses’. (Sapporo,
5. Johnson, J., D. Oyen, P. Netrapalli, and M. Chertkov. Japan, July). , p. . Sapporo, Japan: . Learning planar ising models . Machine Learning Re- Director, H., J. Gattiker, and R. Porter. A multi-scale noise search . model for estimating uncertainty in computer vision
6. Porter, R., S. Lundquist, and C. Ruggiero. Learning to algorithms. Technometrics. Merge: A New Tool for Interactive Mapping. 2013. (Baltimore, Maryland, 18 May, 2013). p. 87431F. Balti- Gelfand, A. E., J. Shin, and M. Chertkov. Belief Propagation for Linear Programming. 2013. In IEEE International more, Maryland: SPIE. Symposium on Information Theory (ISIT). (Istanbul,
7. Porter, R., D. Oyen, and B. Zimmer. Learning watershed Turkey, July). , p. 2249. Istanbul, Turkey: IEEE. cut energy functions . 2015. (Rejijeck, Iceland, May). p. Grosklos, G., and J. Theiler. Ellipsoids for anomaly detection
8. Rejijeck, Iceland: Springer. in remote sensing imagery . 2015. In Proc. SPIE 9472.
9. Odom, P., T. Khot, R. Porter, and S. Natarajan. Knowl- (Baltimore, March ). , p. 94720P. Baltimore: SPIE. 199

Page 200

Harvey, N., and R. Porter. User-Driven Sampling Strategies Porter, R., D. Oyen, and J. Theiler. Data adaptive affinity in Image Exploitation. 2014. In IS&T / SPIE Visualization functions in unsupervised segmentation . To appear and Data Analysis VDA’2014. (San Francisco, 23 Dec). , in Computational Imaging, IS&T Electronic Imaging. p. 90170B. San Francisco: SPIE. 2016.. (San Francisco, Jan). Harvey, N., and R. Porter. User-Driven Sampling Strategies Porter, R., J. Theiler, and D. Hush. Interactive Machine in Image Exploitation. 2014. Information Visualization. Learning in Data Exploitation. 2013. IEEE Computing in (Nov): 1. Science and Engineering. 15 (5): 12. Johnson, J., D. Oyen, P. Netrapalli, and M. Chertkov. Learn- Porter, R., N. Harvey, and C. Ruggiero. Investigation of ing planar ising models . Machine Learning Research . Segmentation Based Pooling for Image Quantification. 2014. In IS&T / SPIE Image Processing: Machine Vision Matteoli, S., M. Diani, and J. Theiler. An overview of back- Applications VII. (San Francisco, ). , p. 90240F. San ground modeling for detection of targets and anoma- Francisco: SPIE. lies in hyperspectral remotely sensed imagery. 2014. Selected Topics in Applied Earth Observations and Re- Porter, R., S. Lundquist, and C. Ruggiero. Learning to mote Sensing (JSTARS)Selected Topics in Applied Earth Merge: A New Tool for Interactive Mapping. 2013. In Observations and Remote Sensing (JSTARS). 7: 2317. Algorithms and Technologies for Multispectral, Hyper- spectral, and Ultraspectral Imagery XIX. (Baltimore, Odom, P., T. Khot, R. Porter, and S. Natarajan. Knowledge Maryland, 18 May, 2013). , p. 87431F. Baltimore, Mary- Based Probabilistic Logic Learning. 2015. In AAAI’15. land: SPIE. (Austin, Texas, 25-30 Jan. 2015). , p. 83. Austin, Texas: AAAI. Porter, R., and B. Zimmer. Links between binary clas- sification and the assignment problem in ordered Oyen, D., K. Sentz, B. Anderson, and C. Anderson-Cook. hypothesis machines . 2015. In Proceedings of IS&T/ Bayesian discovery of bayesian networks with order SPIE Electronic Imaging, Image Processing: Algo- priors . To appear in Advances in Artificial Intelligence, rithms and Systems XIII . (San Francisco, 2015). , p. AAAI’16.. (Phoenix, AZ, Feb). doi:10.1117/12.2083994. San Francisco: SPIE/IS&T. Oyen, D., R. Porter, and K. Sentz. Interactive comparative Porter, R., and C. Ruggiero. Data Integration and Entity Res- analysis for multi-modal data exploitation and fusion. olution: Challenges and Opportunities for Nonprolifer- To appear in 2015 IEEE Applied Imagery Pattern Rec- ation and Arms Control. 2014. In Information Analysis ognition workshop (AIPR2015). (Washington DC, 13-15 Technologies, Techniques and Methods for Safeguards, Oct). Nonproliferation and Arms Control Verification. (Port- land, Oregon, May). , p. . Portland, Oregon: INMM. Oyen, D., R. Porter, and N. Lanza. Discovering Composi- tional Trends in Mars Rock Targets from ChemCam Ruggiero, C., A. Ross, and R. Porter. Segmentation and Spectroscopy and Remote Imaging. To appear in 2015 learning in the quantitative analysis of microscopy im- IEEE Applied Imagery Pattern Recognition workshop ages. 2015. In Proceedings of IS&T/SPIE Electronic Im- (AIPR2015). . (Washington DC, 13-15 Oct ). aging, Image Processing: Machine Vision Applications VIII. (San Francisco , Jan. ). , p. 94050L. San Francisco: Oyen, D., and T. Lane. Bayesian structure discovery of mul- SPIE/IS&T. titask Bayesian networks. 2015. Knowledge and Infor- mation Systems. 43 (1): 1. Sentz, K., and F. Hemez. The Future of Intelligent Systems for Safeguards, Nonproliferation, and Arms Control Oyen, D., and T. Lane. Bayesian discovery of multiple Verification. 2014. In INMM Information Analysis Tech- Bayesian networks via transfer learning. 2013. In IEEE nologies, Techniques and Methods for Safeguards, International Conference on Data Mining. ( Dallas, Nonproliferation and Arms Control Verification. (Port- Texas, 7-10 Dec). , p. 577. Dallas, Texas: IEEE. land, OR , May). , p. . Portland, OR : INMM. Oyen, D., and T. Lane. Interactive Exploration of Compara- Sentz, K., and S. Ferson. Natural language of uncertainty: tive Dependency Network Learning . 2014. In Work- numeric hedge words . 2015. International Journal of shop on Interactive Data Exploration and Analysis at Approximate Reasoning. 57 (Feb.): 19. KDD, August 2014.. (New York, 24-27 August). , p. 88. New York: ACM. Shin, J., A. E. Gelfand, and M. Chertkov. A Graphical Trans- formation for Belief Propagation: Maximum Weight Porter, R., D. Oyen, and B. Zimmer. Learning watershed cut Matchings and Odd-Sized Cycles . 2013. In Advances in energy functions . 2015. In International Symposium Neural Information Processing Systems 26 (NIPS-2013). on Mathematical Morphology. (Rejijeck, Iceland, May). (Lake Tahoe, 5-10 Dec). , p. 2022. : . , p. 497. Rejijeck, Iceland: Springer. 200

Page 201

Shin, J., A. Gelfand, and M. Chertkov. A graphical transfor- mation for belief propagation. IEEE Transactions on Information Theory . Skurikhin, A. N.. Learning tree-structured approximations for Conditional Random Fields. 2014. In IEEE Applied Imagery Pattern Recognition Workshop. (Virginia, 14- 16 Oct. 2014). , p. 1. Virginia: IEEE. Theiler, J.. Transductive and Matched-Pair Machine Learn- ing for Difficult Target Detection Problems. 2014. In Proc. SPIE 9088, Algorithms and Technologies for Mul- tispectral, Hyperspectral, and Ultraspectral Imagery XX. (Baltimore, Maryland, 13 June, 2014). , p. 90880E . Baltimore, Maryland: SPIE. Theiler, J.. Spatio-spectral anomalous change detection in hyperspectral imagery. 2013. In Global Conference on Signal and Information Processing. ( Austin,Texas, 3-5 Dec, 2013). , p. 953. Austin, Texas: IEEE. Theiler, J.. Matched-pair machine learning . 2013. Techno- metrics. 55 (4): 536. Theiler, J.. By definition undefined: adventures in anomaly (and anomalous change) detection. To appear in Proc. 6th IEEE Workshop on Hyperspectral Image and Signal Processing: Evolution in Remote Sensing (WHISPERS 2014). ( , ). Theiler, J.. Anomalousness: how to measure what you can’t define. 2015. In Fourier Transform Spectroscopy and Hyperspectral Imaging and Sounding of the Environ- ment. : JT1A.2. Theiler, J., and G. Grosklos. Problematic projection to the in-sample subspace in a kernelized anomaly detector. Geosciences and Remote Sensing Letters. Wolpert, D., and D. Rajnarayan. Using Machine Learning to Improve Stochastic Optimization. 2013. In Late Break- ing Papers in the Twenty-Seventh Conference on Arti- ficial Intelligence (AAAI). (Bellevue, Washington, 14-18 July). , p. . Bellevue, Washington: AAAI. Zimmer, B., and R. Porter. Hemisupervised Expectation maximization for rare category detection . ACM Trans- actions on Knowledge Discovery From Data. 201

Page 202

Information Science & Technology Early Career Research and Development Continuing Project Reducing Data Dimensionality in Seismic Inversion Gowri Srinivasan 20150691ECR Introduction for large scale data analytics in a wide range of inversion Seismic inversion (SI) refers to the problem of using problems. seismic observations to estimate the geophysical prop- erties of the earth through a wave equation. Inversion Benefit to National Security Missions techniques are an integral part of DOE-critical mission This project will advance LANL’s Integrating Information, areas such as energy production (natural gas extraction) Science, and Technology for Prediction pillar by closely and national security (aspects of Comprehensive Test integrating new analytical and computational aspects to Ban Treaty verification) that rely on accurate estimations advance predictive capabilities for large scale problems. of geophysical properties. Errors in inversion are costly; Aspects of the inversion problem include characterizing the result is inefficient extraction of natural gas since the nature and source of seismic activity, which is a topic fractures cannot be located, or inability to ascertain if an very relevant to LANL’s Science of Signatures (SoS) pillar. explosion was nuclear due to incorrect yield determina- Results of the proposed project will directly provide new tion. computational tools that will support existing programs in extraction of subsurface resources (e.g., natural gas Matching 1000s of terabytes of seismogram data to esti- recovery) and CTBT verification. Aspects of the modeling mate a large number of nonlinearly related parameters framework developed here are relevant for any high- in the wave equation is challenging and often compu- dimensional nonlinear complex system, and will be of tationally prohibitive. Hence simplifying but inaccurate great interest to several DOE programs, positioning LANL assumptions of elasticity and isotropy are currently to compete for new program space. Methods developed made to make the problem tractable. However, applying here will be attractive to the new DOE SubTer initiative, methods that reduce the dimensionality of the prob- where the state of stress in the ground is a key quantity lem without simplifying the physics will also make the of interest. problem tractable and also greatly increase the accuracy of predictions. Currently used Dimension Reduction Progress Techniques (DRTs) assume linear correlations between The project started on June 15, 2015, which is a week parameters which is often not justifiable. The goal of into it at the time of this progress report. However, lead- this proposal is to develop DRTs that capture nonlinear ing up to now, we have demonstrated both linear and correlations between parameters allowing for accurate kernel PCA on simple data sets including the Swiss roll and efficient inversion. We will investigate two separate dataset. We are in the process of setting up quadratically approaches to accelerate the convergence of large scale constrained quadratic programming (QCQP) algorithms viscoelastic SI problems. We propose a kernel-based to investigate whether there is significant speedup of the method to account for nonlinear correlations between KPCA method using the three standard kernels. parameters known as Kernel Principal Component Analysis. We will also consider joint dimension reduc- Future Work tion and emulation, by learning quantitative relation- In FY16 we will use quadratically constrained quadratic ships between parameters that replace computationally programming methods (QCQP) and other computation- expensive simulators. We will validate our methods ally efficient methods investigated in FY15 to learn the using field-scale datasets, a challenging but necessary kernel for a seismic inversion problem. We will start with step since methods that work well on synthetic datasets the assumption of elastic, slightly homogeneous me- often do not scale well. Doing so will be transformational 202

Page 203

dium and gradually progress to visco-elastic medium with anisotropy. We will also attempt joint dynamic dimension reduction of basis and emulation using both linear PCA and kernel PCA methods. We will try this first on the elastic slightly homo- geneous medium modeled in FY15 using the kernel identi- fied from our work leading up to this point. Conclusion We expect that applying Kernel Principal Component Anal- ysis (KPCA) methods will lead to accurate and accelerated convergence of Seismic Inversion problems. We will verify our results against the high dimensional model to ensure that critical features are captured with far fewer param- eters. Multivariate emulation is also unmanageable when the number of parameters exceeds 10. We expect that our joint emulation and dimension reduction technique will yield results that are more efficient that using either only emulators or lower dimensional models. The key is in com- bining the advantages of the two methods. Publications Hickmann, K., J. Hyman, and G. Srinivasan. Robust classifi- cation through dimension reduction using kernel prin- cipal component analysis. Siam Journal on Scientific Computing. 203

Page 204

Information Science & Technology Early Career Research Final Report Quantum Methods for Fast Signal Processing and Metrology Rolando D. Somma 20130741ECR Abstract this project, quantum frFTs have never been proposed, The development of novel quantum methods or trans- constructed or applied. Such transformations have the formations for fast signal processing is paramount in sci- potential to revolutionize signal processing and provide ence and technology. A generalization of the well-known a fundamentally new tool for fast quantum information Fourier transform (FT) to design transformations that tasks. could be implemented on a quantum processor with The research objective is to fill this gap by construct- dramatic speedups is highly desirable. Such a quantum ing and applying quantum variants of the frFT that, like transformation could be applied to problems in signal the quantum FT, can be implemented efficiently on a analysis and quantum metrology. In particular, fractional quantum processor. The same transformation requires Fourier transforms (frFTs) play an important role in signal significantly more operations when implemented on a analysis. frFTs are more powerful than the FT because conventional computer. We seek advanced quantum they can eliminate noise and improve the signal to noise methods for solving signal processing and algorithmic ratio in more general scenarios. Quantum versions of problems faster than currently possible. frFTs suitable for exponentially fast implementation could then revolutionize signal processing. Scientific Approach and Accomplishments This project aims at developing quantum fractional We followed several approaches and developed a quan- Fourier transforms for fast signal processing. In particu- tum transformation that is an excellent approximation to lar, for a sample of size N, a quantum frFT that could be the desired quantum fractional Fourier transform. The implemented using subpolynomial in N operations on basic idea is to relate the fractional Fourier transform to a quantum information processor would be ideal. (The the evolution operator of the quantum harmonic oscil- same transformation requires significantly more opera- lator. A discrete quantum harmonic oscillator was then tions on a conventional computer.) A quantum frFT will proposed suitable for quantum computer implementa- be used to build methods for radar analysis and image tions, and Trotter-Suzuki approximations were used to processing that are exponentially faster than conven- implement such a transformation. The approximation tional ones. The potential of such transformation for error decreases very fast (exponentially fast) with the quantum metrology (high-precision sensing) and related dimension of the space of the problem, making it an ex- quantum algorithms is also of interest. cellent candidate for applications. The most remarkable property is that the implementation cost of our approxi- Background and Research Objectives mate quantum fractional Fourier is subpolynomial in the The Fourier transform (FT) is ubiquitous in signal pro- dimension, that is, significantly smaller (almost expo- cessing. Interestingly, it also occupies a hallowed place nentially smaller) than the implementation cost of the in quantum information: a quantum variant can be corresponding conventional fractional Fourier transform. implemented very rapidly on a quantum processor. Such Thus, the main objective of this research project was quantum FT yields a number of exceptional applications achieved. such as in the famous Shor’s algorithm for factoring. We applied our approximate quantum fractional Fourier There are also useful non-quantum generalizations of transform to different problems. In particular, for the the Fourier transform, including the fractional Fou- simulation of quantum systems and signal processing rier transform (frFT), which represent other powerful problems, attaining substantial speedups with respect tools for conventional signal processing. Previous to 204

Page 205

to the corresponding classical algorithms in both cases. All our results have been published [1] and presented in conferences, and the goals of this ECR project were suc- cessfully accomplished. Impact on National Missions This project constitutes a novel approach in quantum sci- ence with interdisciplinary impact that has already gener- ated new interdisciplinary collaborations. This research is a direct response to the FY13/FY14 IS&T Grand Challenge on quantum information science. The project also en- hances other capabilities at LANL, Sensing, Metrology, and Quantum Cryptography, and brings integration between the pillars of the LDRD “Quantum Initiative”. Previous LANL research on quantum information led the PI to start new collaborations with Sandia National Labs on build- ing a small-scale quantum computer, to obtain an NSF grant bringing summer students to LANL, and to obtain an AFOSR grant in collaboration with the University of Pitts- burgh. The current project has enhanced LANL’s capabili- ties in the quantum arena for a sustainable future. It also allowed the ECR investigator to start a new program in quantum methods to solve other important science prob- lems of mission relevance – a recently funded project by NRO in collaboration with JHU/APL and the Joint Quantum Institute in Maryland. References

1. Somma, R. D.. Quantum simulations of one dimension- al quantum systems. 2015. arXiv: 1503.06319. Publications Somma, R. D.. Quantum simulations of one dimensional quantum systems. 2015. arXiv:1503.06319. 205

Page 206

Information Science & Technology Early Career Research Final Report Stochastic Modeling of Phase Transitions in Strongly Interacting Quantum Systems Christopher Ticknor 20130749ECR Abstract interacting dipolar BECs, but these BECs are not very We studied the collective behavior of quantum sys- degenerate (T/Tc>0.25). There is a pressing need to de- tems and developed methods to understand quantum velop methods capable of studying strongly interacting, gases near a phase transition. Specifically, we studied a “warm” dipolar BECs. trapped two-component Bose Einstein Condensate and Scientific Approach and Accomplishments thoroughly explored and characterized the excitations. We used this knowledge to explore magnetic fluctua- We analyzed the excitation spectrum for a two-compo- tions that diverge near a phase transition. We found nent quasi-two-dimensional Bose-Einstein condensate. that experiments should be able to observe scaling We studied how excitations change character across the behavior. miscible to immiscible phase transition. We furthered the understanding of the two component Background and Research Objectives BEC by studying the dispersion relation and comparing Quantum systems of reduced dimensions exhibit en- the character of the excitations of a single- and two- hanced quantum fluctuations compared to 3D systems. component BEC. We studied the single-component In a homogeneous 2D system, a Bose Einstein Conden- dispersion for a finite BEC system and look at examples sate (BEC) will never form as quantum fluctuations pre- of quasiparticles to understand and characterize the vent this. But there is a Berezinskii-Kosterlitz-Thouless dispersion relation. Next we presented the dispersion (BKT) transition, a topological phase transition in 2D sys- relation for a two-component BEC in both the miscible tems, wherein the coherence of the gas changes decay and immiscible parameter regimes. Then we presented behavior from exponential to algebraic as the tempera- examples of the quasiparticles for both regimes. ture is lowered. (Phase coherence can be thought of as: if we know the phase here at x, then we know some- Finally, we studied fluctuations in these trapped sys- thing about the phase there at x’). The BKT transition is tems. We demonstrated that measurements of number physically motivated by vortex pairs unbinding due to fluctuations within finite cells provide a direct means to thermal energy. Once vortex pairs are unbound, they study susceptibility scaling in a trapped two-component destroy phase coherence as they move through the gas. Bose-Einstein condensate. This system supports a sec- The BKT transition occurs in a variety of 2 dimensional ond-order phase transition between miscible (co-spatial) systems. and immiscible (symmetry-broken) states that is driven by a diverging susceptibility to magnetic fluctuations. Theoretical modeling of temperature dependent dipolar gases has largely been neglected due to the numeri- This body of work provides a path forward to understand cal challenges. The first theory looking at temperature the quantum excitations in a multi-component BEC. dependence was several years ago, but neglected the thermal correlation function. Additionally, this work was Impact on National Missions restricted to Cr, the weakest dipolar gas. I completed Understanding finite quantum systems is essential to the first full treatment of non-zero temperature trapped engineering quantum materials. The study of phase tran- dipolar gases. This resulted in a theory that has been sitions in the presence of non-local interactions has wide used to study correlations in a single component dipolar reaching impacts. Additionally, computational meth- gases. More recently other groups have followed suit. ods developed here could be transferred to the study Now experiments with Dy and Er are producing strongly 206

Page 207

of warm dense matter via quantum molecular dynam- ics (QMD using a hybrid Kohn-Sham and Thomas-Fermi (orbital free) density function theory of the electronic structure. The numerical implementation is similar to that of the PGP method. This development is of great interest to the Laboratory because it presents a means to produce thermal conductivities at much higher temperatures than are currently possible. Publications Baillie, , R. N. Bisset, Ticknor, and P. B. Blakie. Number Fluc- tuations of a Dipolar Condensate: Anisotropy and Slow Approach to the Thermodynamic Regime. 2014. PHYSI- CAL REVIEW LETTERS. 113 (26). Bisset, R. N., R. M. Wilson, and Ticknor. Scaling of fluctua- tions in a trapped binary condensate. 2015. PHYSICAL REVIEW A. 91 (5). Bisset, R. N., Ticknor, and P. B. Blakie. Finite-resolution fluctuation measurements of a trapped Bose-Einstein condensate. 2013. PHYSICAL REVIEW A. 88 (6). Ticknor, C.. Excitations of a trapped two-component Bose- Einstein condensate. 2013. PHYSICAL REVIEW A. 88 (1): 013623. Ticknor, C.. Dispersion relation and excitation character of a two-component Bose-Einstein condensate. 2014. PHYSICAL REVIEW A. 89: 053601. Ticknor, C., and R. Bisset. Finite-resolution fluctuation measurements of a trapped Bose-Einstein condensate. 2013. PHYSICAL REVIEW A. 88: 063624. 207

Page 208

Information Science & Technology Exploratory Research Continuing Project From the Finite Element Method to the Virtual Element Method Gianmarco Manzini 20140270ER Introduction 3. a general stabilization technique for elliptic equa- This project aims at the development of a new family of tion to fix the hourglass instability. numerical methods on polygonal and polyhedral un- structured meshes for diffusion, reaction-diffusion and The summer student worked mainly on the topics of convection-diffusion problems, the Stokes equations and item (1). The results of his work are in a LANL techni- compressible and incompressible elasticity equations. cal report and will be the content for a future research paper. These methods are a generalization of the Finite Element Method (FEM) to polygonal and polyhedral meshes and In (1) Virtual Element Methods were developed for any are based on the very new paradigm of Virtual Element order of accuracy in the diffusive regime and stabiliza- Methods (VEMs). In the virtual framework, only the part tion techniques will be developed and investigated in of the finite element space that refers to polynomials is the convection dominated regime. Results from this constructed, while the behavior of the method on the work was presented in a international congress and are rest of the space is approximated following a stability submitted for publication to an international research criterion. This fact has a dramatic impact on the com- journal. putational complexity that is greatly simplified as the implementation of virtual elements does not require the In (2) we developed the non-conforming formulation for explicit construction of the shape functions anymore. the VEM, carry on the convergence analysis and estab- lished the connection with the non-conforming MFD method. This work was presented in an international Benefit to National Security Missions workshop and a paper is submitted for publication to an The success of this research will provide unique capa- international research journal. bilities in the DOE complex that combine world-class algorithms, their theoretical foundation and high-end In (3) we investigated how the classical stabilization “a la HPC technologies. This new class of numerical algo- Belytschko” for the hourglass instability can be restated rithms is outstanding for mathematical models used in in the virtual framework and a new family of stabiliza- Environment (Climate and Energy Impact), and Scientific tion techniques (including the one mentioned above) Discovery and Innovation (Nuclear, Particle, Cosmology, can be derived. The results of the work will be published Astrophysics, Basic Understanding and Modeling of Ma- by International Journal for Numerical Methods in Engi- terials, Information Science and Technology). neering. Progress Future Work The research activity of the first year of this project The research team will extend the development and mainly concerned mentoring a summer study, and the implementation of the virtual element method (VEM) development of: formulation for reaction-diffusion problems and con- vection-diffusion problems. This is a highly non-trivial

1. virtual element methods for diffusion problems process. with convection and reaction terms;
2. a non-conforming formulation suitable to 2D and 3D We will develop and implement the VEM formulation for diffusion problems; and 208

Page 209

the steady Stokes problem. The numerical approximations Analysis. will have the following features suitable for 2D applica- Veiga, L. Beirao da, and G. Manzini. RESIDUAL A POSTE- tions, and order of accuracy higher than one. RIORI ERROR ESTIMATION FOR THE VIRTUAL ELEMENT METHOD FOR ELLIPTIC PROBLEMS. 2015. ESAIM- Conclusion MATHEMATICAL MODELLING AND NUMERICAL ANAL- The expected results are the developments of new classes YSIS-MODELISATION MATHEMATIQUE ET ANALYSE of methods and their characterization with respect to ac- NUMERIQUE. 49 (2): 577. curacy, efficiency, and effectiveness for important applica- tions where the VEM can outperform the existing compu- tational technologies by providing high-order continuity and accuracy for the numerical approximation and im- proved properties of the scheme (e.g. dispersion reduction for wave propagation). The virtual element algorithms may impact a wide range of modeling multi-physics applica- tions, including climate and environmental modeling (e.g., climate, ocean and sea-ice modeling, ASCEM Project, Arctic Terrestrial Simulator), plasma physics (e.g., Fokker- Planck equation in inertial confinement fusion simulations) and other CFD-based applications. Publications Cangiani, , Manzini, Russo, and Sukumar. Hourglass stabi- lization and the virtual element method. 2015. INTER- NATIONAL JOURNAL FOR NUMERICAL METHODS IN ENGINEERING. 102 (3-4): 404. Cangiani, A., G. Manzini, and O. Sutton. Numerical results using the conforming VEM for the convection-diffu- sion-reaction equation with variable coefficients. 2014. Los Alamos National Laboratory LA-UR-14-27709. Dios, B. Ayuso de, K. Lipnikov, and G. Manzini. The non- conforming virtual element method. SIAM Journal of Numerical Analysis. Gyrya, V., G. Manzini, and D. McGregor. A non-comforming virtual element discretization for singularly perturbed advection di usion reaction equations. 2014. Los Ala- mos National Laboratory LA-UR-14-27995. Manzini, Gianmarco. The nonconforming virtual element method for the convection-reaction-diffusion equa- tion. 2015. Manzini, Gianmarco. The Virtual Element Method. 2015. Manzini, Gianmarco. The virtual element method for flow and transport in porous media. 2015. Manzini, Gianmarco, Andrea Cangiani, and Oliver J. Sut- ton. Conforming and nonconforming virtual element methods for elliptic problems. 2015. Mathematics of Computation. Manzini, Gianmarco, Andrea Cangiani, and Vitaliy Gyrya. The non-conforming Virtual Element Method for the Stokes equations. 2015. SIAM Journal on Numerical 209

Page 210

Information Science & Technology Exploratory Research Continuing Project Large Fluctuations in Stochastic Dynamical Systems Timothy C. Wallstrom 20140302ER Introduction Benefit to National Security Missions The goal of this project is to understand dynamical sys- Our work will contribute to the understanding of dy- tems in the presence of large random fluctuations. Such namical systems in the presence of large fluctuations, systems are common in economics, physics, and several which is an area of mathematics of importance to both areas of biology. The usual analyses of random systems the Office of Science (DOE/SC) and the NSF (Other Fed- make certain technical assumptions on the size of the eral Agencies). The particular application on which we fluctuations, which are invalid when the fluctuations are will focus on is evolutionary dynamics, which is of broad too large. For these types of random systems, the usual interest, and is a central concern of Fundamental Biosci- analyses are invalid, and new methods are required. ence. We believe our models will describe evolutionary dynamics in a large number of areas of biology. We will We focus on the dynamical systems that govern evolu- be focusing specifically, however, on the evolution of tion, and specifically on the area of population genetics, HIV in the infected patient. Such research is of direct which is the mathematical framework for the study of interest to the National Institutes of Health (NIH). The evolution. On the one hand, these systems are suffi- tools we develop will be valuable for future advances in ciently rich to exhibit a broad range of phenomena. On our understanding of HIV and other pathogens, such as the other hand, striking recent mathematical results influenza, which will have importance for Basic Health have led to a complete characterization of the solution Research. classes, under large fluctuations, of the equations of population genetics, in certain simple cases. Thus, these Progress systems should be tractable, but rich enough to provide Over the past year, we have made progress in three main general insights. areas. First, we have obtained extensive high-quality pathogen sequence data, and analyzed this data using Our objective is to illuminate the dynamics of specific models of large fluctuations. Second, we have estab- realistic dynamical systems by connecting such systems lished new mathematical results for models with large to the general theory. In short, we want to connect the fluctuations. Third, we have written computer simula- data to the theory. The value of such an analysis has tion programs for simulating models with large fluctua- been amply demonstrated in the case of small fluctua- tions. We discuss these advances in more detail. tions, and we expect similar dividends in the case of large fluctuations. Sequence data: One of the key goals of our project is to connect the abstract mathematical theory of large We will test our formalism by applying it to the under- fluctuations in population genetics with actual genomic standing of HIV evolution in the infected individual. data from pathogen populations. To this end, we have HIV is the Human Immunodeficiency Virus, which is the been successful in obtaining extensive, high-quality causative agent in AIDS. Existing models of HIV evolu- sequence data for studying the evolution of pathogens tion have led to a number of stubborn mysteries, which within an infected patient. We have obtained data for we trace to an inadequate modeling framework, which both the Human Immunodeficiency Virus (HIV), and for does not account for large fluctuations. There is empiri- the Hepatitis C virus (HCV), so that we can compare the cal evidence of large fluctuations in HIV, and modern similarities and differences between how these patho- sequencing technology will permit detailed comparison gens evolve within the patient. Both of these pathogens of our models with experimental data. 210

Page 211

are expected to exhibit large fluctuations. The sequences deficiency Virus (HIV) and Hepatitis C Virus (HCV) datasets are several thousand bases long, which helps with statisti- obtained over the past year, and publish this analysis. We cal significance, and were obtained using the technique of will also complete and publish the mathematical analysis single genome amplification (SGA), an expensive technique of models with large fluctuations and finite population that ensures that the sequences are highly accurate. Our sizes. datasets contain dozens of sequences taken at several times, spanning the period from acute to chronic infection A second goal will be to use our methods to try to obtain in multiple patients. clinically significant insights into the biology of pathogen infection. The genealogies of both HIV and HCV appear to We have used this data to demonstrate that the usual undergo a qualitative change a few years into infection, genealogical model, which assumes small fluctuations, a change which may correspond to disease progression. does not fit either HIV or HCV. We are in the process of fit- However, this apparent change has not yet been under- ting the data to models with large fluctuations and writing stood quantitatively. We hypothesize that this change up our results. The paper should be completed in a few may correspond to a change in the size of the fluctuations months. in the evolutionary process, and may be detectable as a change in the parameter quantifying these fluctuations. New mathematical results: Most of the existing results for models of large fluctuations assume, as a mathematical A third goal will be to develop our methods to accom- idealization, that the population size is infinite. This limit modate common complications that arise in the analysis leads to useful simplifications, since the probability of of real datasets: sequencing errors, and recombination. In certain types of genealogical events becomes so small that our analysis so far, we have had the benefit of very high these events can be ignored. Although this is a good start- quality sequence data. In many cases, we will want to ap- ing approximation, actual populations are finite, and for ply a similar analysis to poorer quality data, which is more this reason the actual probability of these events, though readily available because it is less expensive to obtain. small, is not zero. We would like to know how likely they We will study the robustness of our inference techniques are. We have used mathematical analysis to calculate these in the presence of the types of sequencing errors that probabilities in the case of a large but finite population. occur commonly in the usual sequencing technologies. A This is an important step in making these models more second complication is recombination. In recombination, realistic, and connecting them to the actual biological data. a virus contains genetic material from two parent viruses. A paper describing these results is in preparation. Recombination is much more prevalent in HIV than in HCV. Through a combination of computer simulation and com- New simulation codes: We have written computer code parative studies between HIV and HCV, we will assess the to simulate our models of large fluctuations. Much of the impact of recombination on our inference methods. behavior of our models cannot be obtained through math- ematical analysis, because the full models are too com- Conclusion plex. Thus, we are dependent on computer simulations This project will lead to methods for connecting realistic to determine many of the properties of these models. To stochastic dynamical systems to their solution classes, to gain the best understanding of our models, the simulation an understanding of how progressive adaptation takes results are used in conjunction with mathematical analy- place in a predator-prey environment, and to clarification sis. We have used these simulation codes to explore the of the appropriate framework for modeling HIV evolution range of different genealogies that arise from our models, within the host. and to evaluate how well we can determine model param- eters from a sample of sequences. The simulation codes Publications complement our mathematical results, as described in the Bhattacharya, T., E. Giorgi, T. Wallstrom, P. Hraber, G. Learn, previous paragraph, in that they help us to understand the T. Kreider, Y. Li, F. Gao, B. Korber, B. Hahn, and G. Shaw. behavior of these models for finite populations. Detecting Selective Sweeps in HIV. 2014. Presentation to Scientific Leadership Group of Center for HIV/AIDS Vaccine Immunology (CHAVI), July 30, 2014. Chosen for Future Work presentation to full CHAVI collaboration meeting, and In the upcoming year, the third and last of this project, we presented there on September 29, 2014.. will complete work in several areas. First, we will complete the analysis of the Human Immuno- 211

Page 212

Information Science & Technology Exploratory Research Continuing Project Accelerating Time Integration for Multi-scale Simulations Shengtai Li 20140323ER Introduction in many areas of science and engineering, including en- One of the most difficult challenges in multiscale simula- vironmental and geosciences, climate, materials, com- tions is to bridge the gap between different time scales. bustion, high energy density physics, fusion, bioscience, The time step for the fast timescale variables tends to be chemistry, power grids and information networks. In the so small that the slow-scale variables have to wait a long past 20 years, numerical techniques, such as domain de- time until the micro-step simulation finishes a macro- composition and adaptive mesh refinement (AMR), have step. To overcome this limitation in time scales is becom- been developed to bridge different length scales. How- ing the key for the practical use of the multiscale simula- ever, the total elapsed time of the simulation is limited tions.The primary goal of this proposal is to develop an by the fastest timescale in the system, because it is often innovative numerical framework to accelerate the time necessary to use the same small time step for the whole integration for the multi-time-scale problems. To achieve system to avoid instability and unphysical artifact. this goal, we identify two major objectives. First, we will construct low dimensional, coarse-grained, effective We are developing a reduced order model (ROM) and reduced order models (ROM) via analytical and numeri- parallel in time method to accelerate the time integra- cal approximations. The ROM serves two purpose in our tion in multiscale simulations, and to resolve the paral- simulations: either the cost of each time step of simula- lel saturation issue for future exa-scale computing. Our tion is sharply reduced, or a large time step is allowed by numerical techniques can be applied to many multiscale removing the dynamics of fast components, or both. problems, including but not limited to defect (e.g., crack) formation and propagation in solid material, magnetic Second, we will accelerate the time integration via paral- reconnection layer in plasma physics, interface or front lel-in-time (PIT) strategy. By construction, the ROMs only propagation in combustion, interface formation and capture the original full microscopic dynamics approxi- propagation during phase transition in multi-phase flow mately. In many cases, the microscopic simulations with simulations, and the dust and gas coupling in disk-planet small time step is needed over the entire time domain to simulations. capture long time dynamics correctly, or at least in some small regions. The PIT strategy proposes to break the Progress global problem of time evolution into a series of inde- The research team developed a new reduced model pendent evolution problems on smaller intervals. Initial for the coupled dynamics of dust and gas. Our previous states for these problems are supplied by a less accurate reduced order model (ROM) uses a bi-fluid model and but fast sequential time integrator, for example by using treats the dust as a pressureless fluid. This has a draw- ROM. The smaller independent evolution problems can back that the stiff source term (drag force between dust then be solved in parallel using more expensive and and gas) appears in both the gas and dust equations. more accurate integrators. An iterative procedure is used The partial equilibrium state must be reached in both to ensure that the algorithm converges to the solution equations, which increases the complexity to solve the one would have obtained using a purely sequential fine equilibrium state. That is why in the old model, we do propagator over the entire time-domain. not implement the dust feedback to the gas. In the new ROM, we combine the gas and dust as one fluid. Only Benefit to National Security Missions one equation in the one-fluid model contains the stiff The need for multiscale simulation capability is pervasive source term, which is the velocity difference between 212

Page 213

dust and gas. Therefore the equilibrium state is easy to is no clear scale separation. The conventional multi- solve. The dust feedback to the gas is easier to implement scale technique does not work well for these type of than the old model. problems. We will develop an adaptive algorithm to identity when and where the reduced order model can We also developed a new Riemann solver for the new be safely used, when and where the parallel-in-time ROM and a new second-order method to solve the one- method must be done to improve the accuracy, and fluid model. The method has been verified and tested. when and where the conventional model can be used. We will develop techniques to bridge the gap between A parallel-in-time algorithm was implemented by embed- different regions and methods. ding our algorithm and code into a general framework • Develop a new algorithm to improve the accuracy of XBRAID to achieve the parallel in time capability. The our operator-splitting approach. For multi-physics and framework XBRAID greatly reduced our code development multi-scale problems, we will combine the spectral time and effort. deferred correction (SDC) algorithm with our second- order operator-splitting approach to improve the Our simulation package was enhanced with additional accuracy. We will also implement this algorithm to the capabilities. We extended our method to multiple planets, XBRAID parallel-in-time framework. which is necessary for simulating the newly-observed HL • Apply the new algorithm to the dusty disk simulations Tau proto-planetary disks and its surrounding multiple to simulate more observed proto-planetary disks, and gaps. We also allow the initial disk to be exponential de- compare the simulation results with the observations. caying at the outer-disk so as to avoid the gravitational in- stability issues. We have developed a non-uniform log-grid Conclusion method so that we can simulate a much larger disk with The primary goal of this project is to develop an innovative relatively small number of grid points. We have coupled numerical framework to accelerate the time integration the coagulation code from Til Bernstil of Harvard University for the multiscale problems. By using reduced order model with our disk simulations code. and parallel in time strategy, we expect to speed up our current simulation of planet formation in dusty disks by at The research team also adapted a radiative transfer code least two order of magnitude. Our proposed space-time RADMC-3D for our use and generated observations images parallelism can also solve spatial parallelization satura- from our numerical simulation results. tion issue: using more processors above a critical number will slow down the simulation due to the communication The newly (Dec 2014) observed HL Tau disk was stimu- overhead. Our proposed algorithm can be easily applied to lated using our coupled dust and gas dynamics with our other multiscale problems to accelerate the simulation by super-fast disk self-gravity solver. The observation images orders of magnitude. from our simulated data match quite well with those from the ALMA telescope. Our results have been presented on Publications the international conference and have gained world-wide Deng, W., H. Li, B. Zhang, and S. Li. Relativistic MHD Simu- recognition. lations of Collision-induced Magnetic Dissipation in Poynting-flux-dominated Jets/outflows. 2015. The As- Through simulations, we find it is essential to simulate the trophysical Journal. 805 (2): 1. dust and gas together to obtain accurate dust distribute in a disk. It is also important to include the disk self-gravity Fu, W., H. Li, S. Li, and S. Lubow. Effects of dust feedback in the simulations to obtain a clear picture of the gap on vortices in protoplanetary disks . 2014. Astrophysi- cal Journal Letter. 795 (2): L39. formation by the planet in the HL Tau disk. We also obtain upper and lower limit on the total disk mass for the future Fu, W., H. Li, S. Lubow, and S. Li. LONG-TERM EVOLUTION observations. OF PLANET-INDUCED VORTICES IN PROTOPLANETARY DISKS . 2014. The Astrophysical Journal Letters. 788 Future Work (2): L41. The following tasks are planned for the next fiscal year: Guan, X., H. Li, and S. Li. Relativistic MHD Simulations of Poynting Flux-Driven Jets . 2013. The Astrophysical • Develop an adaptive algorithm to achieve better ac- Journal 12/2013; 781(1). . 781 (1): 1. curacy and efficiency for the coupled dust and gas dynamics. We have found for many problems the time Li, S.. Modified FARGO Algorithm and its Combination with scales vary continuously over a wide range and there Adaptive Mesh Refinement. Journal of Computational 213

Page 214

and Applied Mathematics. Li, S., and H. Li. Numerical treatment of dust diffusion in dusty proto-planetary disks. 2014. In NUMERI- CAL MODELING OF SPACE PLASMA FLOWS: ASTRO- NUM-2013. (Biarritz, France, 1-5 Jul 2013). Vol. 488, 1 Edition, p. 96. San Francisco: Astronomical Society of Pacific Conference Series. Li, S., and H. Li. Long-Time Sustainability of Rossby Wave Instability in Protoplanetary Disks with Dead Zone. 2015. In Numerical Modeling of Space Plasma Flows: ASTRONUM-2014. (Long Beach, CA, 23-27 Jun. 2014). Vol. 498, 1 Edition, p. 92. San Francisco: Astronomical Society of Pacific Conference Series. Su, H., and S. Li. Structure-Preserving Numerical Methods for Infinite-Dimensional Birkhoffian Systems. 2015. Journal of Scientific Computing. 46 (1): 196. Su, H., and S. Li. Energy/Dissipation-Preserving Birkhoffian Symplectic Methods for Maxwell’s Equations with Dis- sipation Terms. Journal of Computational Physics. Zhai, X., H. Li, P. Bellen, and S. Li. Three-Dimensional MHD Simulation of Caltech Plasma Jet Experiment: First Re- sults . 2014. The Astrophysical Journal . 791 (1): 1. Zhang, X., H. Li, S. Li, and D. Lin. RESONANCES OF MUL- TIPLE EXOPLANETS AND IMPLICATIONS FOR THEIR FORMATION . 2014. The Astrophysical Journal Letters . 789 (1): L23. 214

Page 215

Information Science & Technology Exploratory Research Continuing Project Automated Identification and Reverse Engineering of Malware Christine M. Anderson-Cook 20140355ER Introduction Benefit to National Security Missions Malware (i.e., malicious program) analysis is a multi- Malware analysis is a multi-Billion dollar industry. Billion dollar industry. Beyond that, malware analysis is Beyond that malware analysis is critical to our national critical to our national security both to protect national security both to protect national secrets and to pre- secrets and to prevent corporate espionage. Almost all vent corporate espionage. Almost all large government large government facilities and corporations employ facilities and corporations employ trained cyber profes- trained cyber professionals to reverse engineer (RE) sus- sionals to reverse engineer suspected malware. This pected malware so that they can respond appropriately. work will allow them to not only investigate more of A highly trained professional can spend days to weeks the large amount of suspected malware discovered on uncovering the purpose of the malware, in an effort to their systems, but also allow a swifter response to these discover who sent it, and the extent of the attack. An threats. The preliminary classification work is already organization cannot adequately respond to a malware mature enough to be implemented on any enterprise infection until they understand what it does. network just as it already has been at LANL in Code Vision. We intend to ultimately implement the new re- The main goal of this project is to develop an automated verse engineering methodology developed in this work framework for the reverse engineering of malware. We in a tool suitable for use by cyber professionals. This will propose to develop a procedure that will take a program allow the methodology to be transitioned into practice trace (static or dynamic), and classify the subroutines as at other government agencies and corporations. As the a particular functionality (e.g., disk I/O, network, GUI, NNSA complex is a prime target for cyberattack, cyberse- registry, exploit, etc). The results will be a list of likely curity is an enabling technology for the nuclear weapons classifications of a subroutine’s functionality, with cor- mission. responding probabilities. This is precisely what a reverse engineer would do in a more qualitative manner when Progress they are unraveling the purpose of a program. How- New methodology for identification and classification ever, classifying subroutines manually is an incredibly of malware was developed. The current focus is on time consuming task. Thus, this work will substantially Advanced Persistent Threat (APT) malware. One new speed up RE times, providing a swifter response and approach uses the static trace of the program. A static thus reducing the expense and possibly the severity of a instruction trace can be viewed as a graph with nodes malware infection. representing subroutines and directed edges represent- ing calls between subroutines. Thus, techniques which A second important consequence of this work is that involve comparing graphs by checking for node and the task classification results can be used to automati- structure similarity may be useful. We expect static trac- cally assess the high-level functionality of a program as es from related programs to have similar static traces, a whole with far more accuracy than in previous work so they should have several common subroutines and (i.e., trojan, adware, IRC, advanced persistent threat, much of the graphs should be isomorphic. The research etc.). There has been no attempt until now to automati- looks at malware at the program level and classifies cally arrange and classify the various tasks of a program, programs based on their similarity with known malware hence, the potential for a large impact due to this work. subroutines. Malware and benign programs can share a substantial amount of code, implying that classifica- 215

Page 216

tion should be based on malicious subroutines that occur sample to ensure good performance. We further expect infrequently, or not at all in benign programs. Various ap- the subroutine classification results to provide extremely proaches to accomplishing this task are investigated, and informative features for an overall program classification a particularly simple approach appears the most effective. (benign, trojan, adware, IRC, advance persistent threat) This approach simply computes the fraction of subrou- routine. This overall classification routine will also be thor- tines of a program that are similar to malware subroutines oughly validated on a large dataset. whose likes have not been found in a larger benign set. Publications The second area of research considers the problem of Anderson, B., C. Storlie, M. Yates, and A. McPhall. Auto- identifying the function of different subroutines. Our goal mating Reverse Engineering with Machine Learning is to automate the task of identifying the general func- Techniques. 2014. In The 7th ACM Workshop on Secu- tion of the subroutines in the function call graph of the rity and Artificial Intelligence. (Scottsdale, 7 Nov. 2014). program to aid the reverse engineers in timely triage of , p. 103. Scottsdale: ACM Digital Library. where to focus their efforts. Two approaches to model Anderson, B., T. Lane, and C. Hash. Malware Phylogenet- the subroutine labels are investigated, a multiclass Gauss- ics Based on the Multiview Graphical Lasso. 2014. In ian process and a multiclass support vector machine. The The Thirteenth International Symposium on Intelligent output of these methods is the probability that the subrou- Data Analysis. (Leuven, Belgium, 29-31 Oct. 2014). , p. tine belongs to a certain class of functionality (e.g., file I/O, 1. Leuven, Belgium: Springer. exploit, etc.). Promising initial results illustrate the efficacy of this method based on evaluation based on a sample of Bolton, A.. Malware static trace analysis through bigrams 201 subroutines taken from two malicious families. and graph edit distance. 2015. LANL Report. Kao, Y., B. Reich, C. Storlie, and B. Anderson. Malware de- Another area of research is to examine methods for iden- tection using nonparametrcis Bayesian clustering and tifying the family of malware from which newly identi- classification techniques. To appear in Technometrics. fied programs are associated. Again an early diagnosis of malware phylogeny can help with efficiently determining if Neil, J., C. Hash, A. Brugh, M. Fisk, and C. Storlie. Scan sta- new programs represent new families of APT, or if the com- tistics for the online detection of locally anomalous monalities between existing programs for which reverse subgraphs. 2013. Technometrics. 55 (4): 403. engineering has been performed can be leveraged. Oyen, D., C. Anderson-Cook, K. Sentz, and B. Anderson. Bayesian discovery of Bayesian networks with order Future Work priors: A new approach to combining knowledge and The research team will continue to explore the local sub- statistical data. To appear in The Thirteenth AAAI Con- routine approach to classification of APT malware that has ference . (Phoenix, 12-17 February 2016). shown promise. We will take advantage of the fact that Oyen, D., K. Sentz, B. Anderson, and C. Anderson-Cook. only a handful of subroutines tend to make a program look Application of Bayesian discovery with order priors to malicious, to produce a subroutine importance measure. combining knowledge and statistical data. 2015. LANL This could be displayed graphically to provide a useful Report. reversing tool. We also intend to leverage the graphical nature of “malicious” subroutines to provide a more infor- Sexton, J.. Anomaly detection of multistage network intru- mative indicator of maliciousness if certain subroutines are sions. 2015. LANL report. connected than if the subroutines were not calling each Sexton, J., C. Storlie, and B. Anderson. Subroutine based other. We will finish the papers in development and start detection of APT malware. To appear in Journal of writing up the new results described above for scholarly Computer Virology and Hacking Techniques. journals. We will also continue to present this work at professional meetings. Sexton, J., C. Storlie, and J. Neil. Attack chain detection. To appear in Statistical Analysis and Data Mining . Conclusion Sexton, J., and C. Storlie. Internal stepping stone detection. This work is expected to result in a transformational ap- To appear in Technometrics. proach to the reverse engineering of malware. We expect to understand the limits of the methodology in terms of Storlie, C., B. Anderson, S. Vander Wiel, D. Quist, C. Hash, how well it can provide likely purposes of each subroutine and N. Brown. Stochastic Identification of Malware and program. We will validate the classification out-of- with Dynamic Traces. 2014. Annals of Applied Statis- tics. 8 (1): 1. 216

Page 217

Information Science & Technology Exploratory Research Continuing Project Temporal Graphs Aric A. Hagberg 20140389ER Introduction The DOE ASC program sponsors research in the area of Graphs are increasingly being used to model complex complex networks related to big data challenges. Our techno-social networks such as the power grid, the research develops new algorithms to efficiently generate internet, as well as computer and social networks. Un- large instances of random graphs for modeling and test- derstanding how these networks function and behave ing of high-performance computing applications. under changing circumstances is vital to several areas of national security. Progress We continued to refine our approach to the modeling This project will focus on developing temporal graph and analysis of the graph structure of computer authen- models for certain classes of complex networks such as tication events. These models are helping understand computer and social networks. The scientific research the risk of using centralized authentication systems such will elucidate the interplay between the structural evolu- as those found in most large organizations such as Los tion of the network, temporal properties of the network, Alamos National Laboratory. Our article “Connected and dynamic processes, such as the spread of a virus or components and credential hopping in authentication information, occurring on the network. The research graphs,” published as part of the International Confer- will examine how sensitive networks are to perturba- ence on Signal-Image Technology & Internet-Based tion, how to slow the speed of a virus and how to Systems (SITIS) in November 2014 described our first increase cybersecurity in the face of persistent attacks. results. We extended that research with a longer paper We will develop efficient algorithms to generate random “The structure of authentication networks”, currently instances of these models as well as algorithms to fit in review, that describes in detail the changes in degree data to the models. This project develops methods and (connectivity) of authentication networks over time and techniques for network science. Network science is one experiments in restricting the length of time credentials of three cross-cutting themes in the Integrating Informa- are valid. Along with these papers we released a com- tion, Science, and Technology for Prediction pillar at Los panion public data set on which we based our modeling Alamos. and analysis results. Benefit to National Security Missions The research team investigated metrics for connectiv- Network science is one of three cross-cutting themes ity in temporal graphs that are relevant for real systems in the Integrating Information, Science, and Technology such as the authentication event data. For static graphs for Prediction pillar at Los Alamos. Our work develops the connectivity of a node has a single simple definition new models, techniques and theorems for dynamically of the set of other nodes that you can reach by crossing changing graphs which are the underpinnings for many links. We developed a new definition of “reachability” applications of network science. For example temporal that captures the dynamics of unauthorized credential graphs are a major research interest in cybersecurity usage in a computer network with centralized authen- and fill a missing gap in random graph models. Prog- tication. This definition provides a measure of how long ress in modeling temporal graphs will provide capabil- it takes, starting at a given computer, to traverse the ity relevant to DOD missions in cybersecurity anomaly credential network and crossing as many edges as are detection. available at the current time. 217

Page 218

Two new random temporal graph models were construct- Publications ed based on the well-known static models, the Erdos-Renyi Farrell, M., T. Goodrich, N. Lemons, F. Reidl, F. S. Villaamil, and Chung-Lu random graphs. The models allow an adjust- and B. D. Sullivan. Hyperbolicity, degeneracy, and ex- able rate of change for dynamic edges with preservation of pansion of random intersection graphs. 2015. In to ap- the original static random graph structure. For these two pear in the 12th Workshop on algorithms and models for the web graph. models we proved theorems for the reachability time and the mixing time (the time it takes to thoroughly random- Hagberg, A., N. Lemons, A. Kent, and J. Neil. Connected ize any initial graph correlations). These theorems will Components and Credential Hopping in Authentication appear in the chapter “Temporal reachability in dynamic Graphs. 2014. In Signal-Image Technology and Inter- networks” of the book “Dynamics of Networks and Cyber- net-Based Systems (SITIS), 2014 Tenth International security”, Imperial College Press in 2015. Conference on. , p. 416. Hagberg, A., N. Lemons, and S. Misra. Temporal reachabil- We submitted for publication our fast algorithm for gen- ity in dynamic networks. In Dynamics of Networks and erating large-scale random kernel graphs with specified Cybersecurity. Imperial College Press. small subgraph properties. This new algorithm dramati- cally speeds up the time it takes to generate kernel graphs Hagberg, A., and N. Lemons. Fast generation of sparse ran- from quadratic scaling to linear scaling in the number of dom kernel graphs. 2015. PloS one. 10 (9): e0135177. graph nodes. In addition to the manuscript we released a Python software implementation of the algorithm to show Keszegh, B., N. Lemons, and D. P’alvolgyi. Online and Qua- the practical performance and simplicity. We have new si-online Colorings of Wedges and Intervals. Order. : 1. preliminary further results on this class of kernel models that appear to make it practical to generate networks with specified distributions of triangle, square, and higher order subgraphs. Future Work For FY16 we will continue with our three-part approach of data, algorithms, and theorems for the dynamics of cybersecurity network systems. We will continue to seek definitions and theorems for random temporal graphs. We will apply our definition and knowledge of reachability to analyze authentication networks. We will seek to develop and publish new cybersecurity graph data sets for model- ing and analysis, and we will develop models and infer- ence methods to fit cyber authentication network data. Our results will be published in leading journals and top conferences in mathematics, statistics and cybersecurity. The algorithms for generating models will be tested and published in the NetworkX Python software package. Conclusion We target a fundamental challenge in network science, namely modeling the time-dependent changes in complex networks. Networks targeted in this work are cyber and social networks. We anticipate that this project will have a broad impact in the network science communities: we will contribute new insights and basic methodologies for understanding network evolution affects both the network topology, and those dynamical processes occurring on the network. 218

Page 219

Information Science & Technology Exploratory Research Continuing Project Efficient Method for Large Scale Simulations of Fermionic Gases Interacting with Classical Fields Kipton M. Barros 20140458ER Introduction Benefit to National Security Missions We will develop a tool to simulate a broad range of We will develop a tool that will create fundamen- quantum mechanical phenomena that arise from tally new capabilities for bridging two LANL grand strongly correlated electrons in actinide, lanthanide, and challenges: “Information, Science and Technology” transition metal based compounds. These compounds (IS&T) and “Materials: Discovery Science to Strategic are of special interest due to the formation of meso- Applications”(MDSSA). This research is a direct response scale super-structures that produce novel macroscopic to FY13 IS&T Grand Challenge priorities: development functionality. Interacting quantum mechanical systems of methods for inference and prediction of large-scale are notoriously difficult to simulate; existing numeri- complex systems and design of algorithms to efficiently cal methods cannot span the gap between atomic and extract information from massive amounts of data. Our mesoscale. Our new algorithm improves state of the art project also responds specifically to the LDRD “Materials efficiency by orders of magnitude and makes the meso- for the Future” focused area and addresses the priori- scale accessible. ties of realizing design principles towards controlled functionality and developing multifunctional materials It applies to condensed matter systems of electrons in- to transform structural and functional performance and teracting with a classical field that might represent local integration, and tunable and emergent properties. In magnetic moments, charge density or a superconducting the BES report on Basic Research Needs, multifunctional order parameter. We focus on the Kondo lattice model materials are emphasized as solutions to a range of (KLM), where itinerant electrons interact with localized energy problems. The main bottleneck for understanding classical magnetic moments. Over the past decade, the the complex behavior of interacting quantum systems is KLM has been widely used to study giant and colossal the lack of efficient algorithms for simulating the corre- magneto-resistance. In recent years, KLM studies have sponding models. Developing this capability is crucial for focused on a variety of new phases that have special addressing future Laboratory mission challenges such as transport properties. These new phases arise due to the inference and prediction of large-scale complex systems, complicated, effective many-body interactions of the and prediction and control of emergent phenomena classical field, obtained by “integrating out” the conduc- in complex materials. The innovative codes that will tion electrons. Because electrons are quantum mechani- be developed under this project will be applied to the cal in nature, tracking the effective interactions for an simulation of real compounds that are being investigated evolving classical field requires repeated diagonalization in our experimental groups. These codes will support of an NxN matrix, where N is the number of atoms in a new program developments for modeling and simulating finite lattice. A direct approach fails for lattices beyond complex materials. This challenging task will have a ma- N=400 lattice atoms. In our algorithm, the computation- jor worldwide impact on the fundamental understand- al cost scales linearly with system size and preliminary ing of unconventional states of matter. Such states have studies of a triangular lattice of N=40,000 atoms lead to great application potential due to their unusual physical a variety of phenomena consistent with experiments: responses. chiral vortices and domains, skyrmions, bound hedge- hog dipoles, and interacting vortices. We will extend our Progress method to study a broad variety of 2D and 3D models The research team refactored and generalized our KPM relevant to real materials and implement a version opti- code to facilitate new application areas. For example, mized for parallel execution. 219

Page 220

adding support for a new Kondo Lattice Model now typi- includes effective dynamical waves, which appear to de- cally involves writing a couple new functions to specify the crease decorrelations times for equilibrium measurements, matrix hopping elements. and (3) SLL dynamics is realistic, enabling us to perform non-equilibrium and dynamical magnet studies. Various aspects of the code were optimized to improve builds times of the Hamiltonian matrix. In particular, we Interaction with local quantum chemistry experts has sug- now calculate Hamiltonian matrix elements in a multi- gested fruitful new research directions. For example, after threaded way using Intel’s Thread Building Blocks (TBB) discussions with Anders Niklasson (T-1), we have learned library. We also use TBB’s parallel-sort capability to speed that the Extended Lagrangian formalism could be very use- up construction of the Compressed Sparse Row (CSR) ful for solving dynamical self-consistency constraints that matrix format. Finally, using Cuda 6.5, our optimized code appear in condensed matter physics, such as strong mag- can now use Block-compressed Sparse Row (BSR) format, nets. These discussions have spun off a new ER proposal which takes advantage of the block matrix structure typi- currently under consideration by the CNM committee. cally seen in multi-orbital Hamiltonians. The BSR format eliminates storage of most sparse matrix indices, thereby We applied the above-described algorithms to study new reducing GPU bandwidth loads, and enabling an effective physical problems, including: speed-up of a factor of 2 for typical Hamiltonians. We are still working on the multi-GPU, MPI version of our code. Using our recent numerical developments to include cor- relations in quantum molecular dynamics via the so-called We also discovered a surprising and beautiful mathemati- Gutzwiller approximation. We have produced the first cal identity that enables much better scaling with the prototype for modeling single-orbital systems (like Hydro- number of random vectors. As discussed in our report last gen). By using this approach, we have computed the self- year, we now benefit from the decay of the density ma- diffusion coefficient as a function of the strength of the trix by using “correlated” random vectors. This year, we intra-orbital Coulomb interaction. Interestingly enough, we analyzed the “energy matrix”, which can be thought of as found that the self-diffusion coefficient exhibits a rather the anti-derivative of density matrix (i.e. the anti-derivative sharp maximum near the Mott transition. We are currently of the Fermi function, applied to the Hamiltonian). It turns finishing a manuscript on this topic. out that the energy matrix has better decay properties than the density matrix. By using automatic differentiation, We have found that crystals of merons and antimerons we can stochastically estimate the density matrix while are solutions of Kondo Lattice Models on high-symmetry still benefiting from the superior decay properties of the lattices. This important result shows that the Ruderman- energy matrix. In the naive KPM method (our original ER Kittel-Kasuya-Yosida interaction is not enough to character- proposal) the stochastic error scales like ‘s^\{-1/2\}’, in the ize the magnetic ordering of frustrated itinerant magnets. number of random vectors s. With our improvements, the By combining our novel numerical approach with analytical error decreases exponentially with ‘s’ for insulators. For treatments, only valid in the weak-coupling regime, we metals, the error now scales like s^\{-1\} and s^\{-5/6\} for 2d have discovered that exotic crystals of topological defects and 3d systems respectively. Our code is likely the only one can also be stabilized in itinerant magnets. We are cur- capable of simulating the dynamics of metals in a linear rently starting to write a manuscript on this topic. scaling way. Future Work Last summer, our student Julien Roussel successfully imple- The research team will develop a truly high-performance mented a Monte Carlo approach for the graph color prob- C++/MPI version of our code, which will be scaled to lem necessary to effectively choose correlations among hundreds of GPUs in order to reach tens or hundreds of KPM random vectors. thousands of atoms. The research team’s experience with Molecular Dynamics We will also use the code developed in 1. to simulate the simulations has helped us to improve our methodology propagation of topological defects in atomic crystals. also for dynamical simulations of magnetic spin systems. In particular, we have implemented a predictor-corrector Finally, the team will start simulations of Kondo Lattice scheme for time integration of “stochastic Landau-Lifshitz” Models (KLM) with small Fermi surface pockets. This is a (SLL) dynamics. This has had three benefits: (1) integration precondition for stabilizing the desired emergent meso- error now scales O(dt^2), making finite-temperature simu- scale structures. lations much more accurate, (2) this richer dynamics now 220

Page 221

Conclusion The algorithm and codes to be developed in this project will allow us to study phases that were previously inac- cessible. We will study the emergence of complex me- soscale magnetic patterns that we know should occur in low-density electron gases interacting with magnetic ions. Our codes will include a tool for incorporating the band structure parameters of the itinerant electrons which are computed via a tight-binding parameterization of the conduction bands obtained from first principle calculations [Local Density Approximation (LDA)]). We will deliver a high-performance implementation of our algorithm, and many associated analysis tools, to the LANL community. Publications Barros, , Sinkovits, and Luijten. Efficient and accurate simu- lation of dynamic dielectric objects. 2014. JOURNAL OF CHEMICAL PHYSICS. 140 (6). Barros, K., F. Venderbos, G. Chern, and C. D. Batista. Novel magnetic orderings in the kagome Kondo-lattice mod- el. 2014. Physical Review B. 90: 245199. Barros, K., and E. Luijten. Dielectric Effects in the Self-As- sembly of Binary Colloidal Aggregates. 2014. Physical Review Letters. 113: 017801. Gan, , Wu, Barros, Xu, and Luijten. Comparison of efficient techniques for the simulation of dielectric objects in electrolytes. 2015. JOURNAL OF COMPUTATIONAL PHYSICS. 291: 317. Lin, , C. D. Batista, Reichhardt, and Saxena. ac Current Gen- eration in Chiral Magnetic Insulators and Skyrmion Mo- tion induced by the Spin Seebeck Effect. 2014. PHYSI- CAL REVIEW LETTERS. 112 (18). Lin, , C. D. Batista, and Saxena. Internal modes of a sky- rmion in the ferromagnetic state of chiral magnets. 2014. PHYSICAL REVIEW B. 89 (2). Lin, S., C. Reichhardt, C. D. Batista, and A. Saxena. Dynam- ics of skyrmions in chiral magnets: dynamic phase transitions and equation of motion. 2014. Journal of Applied Physics. 115: 17D109. Lin, S., X. Wang, Y. Kamiya, G. Chern, F. Fan, D. Fan, B. Ca- sas, Y. Liu, V. Kiryukhin , W. Zurek, C. D. Batista, and S. Cheong. Topological defects as relics of emergent con- tinuous symmetry and Higgs condensation of disorder in ferroelectrics. 2014. Nature Physics. 10: 970–977. 221

Page 222

Information Science & Technology Exploratory Research Continuing Project Coupled ALE-AMR for 3D Unstructured Grids Jacob I. Waltz 20150414ER Introduction Progress This project researches and develops novel methods for The primary goal of the project for FY15 was to develop the numerical modeling of high-speed material flows the theory for and implement mathematical tech- that are applicable to a wide range of scientific and niques to compute physics-based motion of the mov- national security problems. The basic concepts under ing reference frame. Both the theoretical development exploration involve the combination of historically and the initial implementation have been successfully proven but heretofore independent techniques. The completed. Initial verification of the technique also has primary scientific challenge for the project lies in the been completed, and its application to more complex development of a mathematical approach that couples problems is in progress. A presentation on this topic and the different techniques in a consistent manner. Such the overall project will be given at a leading domestic approaches have not been previously developed either conference in July 2015; a journal article is expected in at Los Alamos National Laboratory or within the broader late FY15 or early FY16. computational science community. The new methods are designed to enhance fidelity and computational ef- Future Work ficiency relative to existing methods. The research team will extend the FY15 work on moving reference frame techniques to include mesh adaption. Benefit to National Security Missions Specific tasks include: We expect our research to lead to significant improve- ments in fidelity and computational efficiency for work • implement mathematical techniques to couple mesh related to NNSA Defense Programs, Nonproliferation, adaption to a moving reference frame. and Science Campaigns. The impacts of these improve- • test the techniques on standard problems involving ments will include faster responses to programmatic high-speed material flows questions; increased population sizes for Uncertainty • document the research results in journal articles Quantification and other sensitivity studies; greater and/or conference presentations detail in discovery-scale simulations; and an enhanced ability to model realistic 3D features. Conclusion The goal of this project is to develop a novel method for Future Mission: The jump in simulation capability that the numerical modeling of high-speed material flows. results from our research will enable the solution of The impacts of this work will include significant en- entirely new classes of problems and therefore has hancements in detail and accuracy; reduced uncertainty the potential to significantly expand the scope of the in simulation-based responses to programmatic issues; Laboratory’s simulation tools. New application areas and advances in scientific understanding in the broader might include design of blast mitigation structures for numerical modeling community. Given the unique na- urban environments, energetic disablement calculations ture of the work, it also will establish Los Alamos Nation- of Improvised Explosive Devices, anti-personnel and al Laboratory as an international leader in this area. anti-structural analysis, high-resolution studies of mix and ignition in Inertial Confinement Fusion targets, and Publications astrophysics. Waltz, J., J. Bakosi, and N. R. Morgan. A coupled ALE- AMR approach for shock hydrodynamics on tetra- 222

Page 223

hedral grids. Presented at United States National Con- gress on Computational Mechanics. (San Diego, CA, 27-31 Jul 2015). 223

Page 224

Information Science & Technology Exploratory Research Continuing Project Globally Optimal Sparse Representations Brendt E. Wohlberg 20150467ER Introduction Benefit to National Security Missions One of the major computational challenges in the mod- The project will develop a general data modeling tech- ern world is the processing and analysis of the growing nique, primarily with relevance to data in the form of deluge of data from such diverse sources as automated sampled signals, images, or video. Given the generality telescopes collecting astronomy data, physics data from of the method, it has potential relevance to any of the particle accelerators, and intelligence data from remote areas above that require processing or analysis of these sensing satellites, to name but a few. Many of these types of data. As such, the primary mission relevance is processing and analysis tasks require an effective model to Information Science and Technology. Nevertheless, for the data at hand. Such a model is provided by sparse specific potential application areas that have already representations, which have found an extremely wide been identified include analysis of radio frequency data range of applications, often with state-of-the-art results. for non-proliferation monitoring (relevant to agency These representations are mathematically simple to de- DOE/NNSA and mission Remote Sensing for Nuclear fine, but usually involve relatively expensive optimization Nonproliferation), analysis of remote sensing data for algorithms to compute them. For practical reasons, this non-proliferation and intelligence acquisition (relevant type of model is usually applied independently to small to DOE/NNSA and Intelligence Agencies, and mission Re- blocks of signal or image data. An alternative form of this mote Sensing for Nuclear Nonproliferation), modeling of representation that jointly models an entire signal or im- material properties and analysis of multiple modalities age exists, and has numerous advantages, but is not very of materials science data (relevant to Basic Understand- widely applied, primarily due to the great computational ing of Materials mission), structural health monitoring cost. for machinery, including vehicles (relevant to agency DOT and Commerce and Transportation mission), and Motivated by the recent development of significantly analysis of astronomical and sky-survey data (relevant to for efficient algorithms for these globally optimal sparse agency NASA and mission Nuclear, Particle, Cosmology, representations, this project will explore and develop and Astrophysics). the mathematical theory, algorithms, and applications of this approach. While there are numerous clear advan- Progress tages of this form of sparse representation, there are a Substantial progress has been made in the seven months number of technical challenges in fully exploiting them. since the start of the project. To summarize: Potential applications include processing of and feature detection in hyperspectral remote sensing imagery, A new method has been developed to improve algo- change detection in wide-field automated sky surveys, rithm convergence, reducing the time required to com- structural health monitoring for critical machinery and pute the data representation. This method has potential buildings, and detection and classification of features in benefits for a general class of optimization algorithms, of radio frequency signals for non-proliferation monitoring. which the problem at the core of this project is just one This work also has the potential for significant impact in example. A journal paper has been submitted to IEEE the fields of signal and image processing, where sparse Transactions on Signal Processing, and is currently under representations have become a major subject of re- review. search. The main algorithm has been refined and improved in a 224

Page 225

number of ways, and a careful comparison has been made tal work is required to resolve this question. Progress was with competing approaches, showing that it provides made on this aspect during the first fiscal year of the proj- substantial computational advantages in most problem ect, but a proper understanding of the optimal approach is regimes. A journal paper describing this work has been still to be developed. submitted to IEEE Transactions on Image Processing, and is currently under review. Dictionary learning Even in the case of standard sparse representations, the The general method has been extended to additional im- problem of learning a “dictionary” (the set of basic compo- age processing applications, and work is ongoing in im- nents from which the linear representation is assembled) proving application performance. A journal paper describ- is still rapidly developing, and is far from being a solved ing this work is in preparation. problem. Learning effective dictionaries in the convolution- al setting presents an additional set of challenges related In line with the commitment to open source expressed in to the vastly increased redundancy of the representation. the project proposal, an initial version of an open source The possibility of constructing multiscale dictionaries is software library has been completed and publicly released. one of the major advantages of this form, but also provides a further set of technical challenges. Some progress was Two difference collaborations have been established with made on this aspect during the first year, but significant external groups (at University of Rochester and at Aca- work remains to be done. demia Sinicia in Taiwan) working on audio signal process- ing problems. Thus far, these collaborations have led to a Algorithm development conference paper presented at ICASSP 2014, and another Efficient algorithms have recently been developed, but conference paper recently submitted to MLSP 2015. further increases in efficiency would both aid the com- putational experiments required for other aspects of the Ongoing collaboration on applications of sparse represen- project, and expand the range of potential applications. tation and convolutional sparse representation methods to Considerable progress has been made in this area, but anomaly detection problems, with collaborators at Politec- there are some possible improvements remaining to be nico di Milano and Tampere University of Technology, have explored. led to two conference papers that have already been pre- sented, and an additional one accepted for presentation. Conclusion The major technical goals are the development of theory, Joint work with a collaborator at Pontifical Catholic Uni- algorithms, and applications of the globally optimal sparse versity of Peru (with a joint UNM appointment) has led to representations. This work will primarily involve the devel- a conference paper accepted for presentation and a joint opment of signal processing theory and methods, together UNM/LANL patent on a novel video background model- with associated algorithms. If successful, these develop- ing method that exploits ideas from convolutional sparse ments have the potential to change the standard practice representations. in the application of sparse representation methods to a wide variety of problems, together with improved perfor- Invited research seminars have been presented at the mance and new capabilities of these methods. An open- UCLA Math department and at Microsft Research at Red- source library will be released to make these techniques mond. widely available to researchers in the field, as well as to practitioners in application areas. Future Work The second fiscal year of the project will focus on the fol- Publications lowing main aspects of the project: Boracchi, Giacomo, Diego Carrera, and Brendt Wohlberg. Novelty Detection in Images by Sparse Representa- Signal representation tions. 2014. In Proceedings of the IEEE Symposium Se- Initial experiments, and some initial reports in the litera- ries on Computational Intelligence (IEEE SSCI). , p. 47. ture, indicate that convolutional sparse representations Carrera, Diego, Giacomo Boracchi, Alessandro Foi, and are most effective when applied to signals and images Brendt Wohlberg. Detecting Anomalous Structures by that have been pre-processed to remove smoothly vary- Convolutional Sparse Models. 2015. In The Internation- ing components. The most effective choice of this pre- al Joint Conference on Neural Networks (IJCNN). processing, which can have a major performance impact, is far from clear, and substantial theoretical and experimen- Cogliati, Andrea, Zhiyao Duan, and Brendt Wohlberg. Piano 225

Page 226

Music Transcription With Fast Convolutional Sparse Coding. 2015. In IEEE International Workshop on Ma- chine Learning for Signal Processing. Jao, Ping-Keng, Yi-Hsuan Yang, and Brendt Wohlberg. In- formed Monaural Source Separation of Music based on Convolutional Sparse Coding. 2015. In Proceedings of IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP). , p. 236. Rodriguez, P., and Brendt Wohlberg. Incremental Principal Component Pursuit for Video Background Modeling. To appear in Journal of Mathematical Imaging and Vision. Rodriguez, Paul, and Brendt Wohlberg. Translational and Rotational Jitter Invariant Incremental Principal Com- ponent Pursuit for Video Background Modeling. 2015. In Proceedings of IEEE International Conference on Im- age Processing (ICIP). Wohlberg, Brendt. Endogenous Convolutional Sparse Rep- resentations for Translation Invariant Image Subspace Models. 2014. In Proceedings of IEEE International Conference on Image Processing (ICIP). , p. 2859. Wohlberg, Brendt. Efficient Algorithms for Convolutional Sparse Representations. To appear in IEEE Transactions on Image Processing. 226

Page 227

Information Science & Technology Exploratory Research Continuing Project Enabling Automatic Parallelism and Transparent Fault Tolerance Marion K. Davis 20150485ER Introduction Benefit to National Security Missions One of the three major challenges in harnessing the High performance computing is fundamental to the ad- potential of an extreme-scale computing system, as vancement of many areas of science. While innovations identified by the 2010 Summary Report of the DOE Of- in hardware technologies continue to allow us to build fice of Science Advanced Scientific Computing Advisory ever more powerful and complex computing systems, Committee (ASCAC), is exploiting massive parallelism. In our modes of programming them have advanced con- particular, they state that software implementations will servatively. Both DOE and DARPA have recognized the need new programming paradigms to make effective use need to re-examine the entire software stack that ranges of unprecedented levels of concurrency. Yet more recent from operating systems, runtime systems, programming is the 2011 DOE Advanced Scientific Computing Research systems, and applications. At the level of programming (ASCR) Exascale Programming Challenges Report, specifi- systems, the need to re-explore functional language con- cally calling out the need to revisit functional-language cepts has been explicitly identified. This project directly concepts. DOE and DARPA have also recognized the addresses this need, as well as its corresponding runtime potential of domain-specific languages (DSLs) for raising system. Breakthroughs in this area will benefit high per- the level of abstraction of programming. In the broader formance scientific computing at large, and all scientific scientific computing community it is becoming more disciplines that use it. widely understood that the only programming para- digm that can so scale will be data-dependence-based Progress asynchronous scheduling of purely functional compute The research team determined that the appropriate tasks. In turn, this paradigm enables a fault-tolerance break-out point from GHC (Glasgow Haskell Compiler) is mechanism much more light-weight than checkpoint/ at the STG (spineless tagless G-machine) intermediate restart because compute tasks, being pure (side-effect representation. Our system then will compile STG, with free), can be safely restarted if their failure is detected. our non-standard semantics, to executable code. The This project will develop and evaluate a robust and greater part of the overall effort is the runtime system, performant implementation of this paradigm in its pur- initially supporting serial execution and garbage collec- est form wherein compute tasks are both internally and tion, later parallel execution, and then some degree of externally (observationally) pure. While such systems parallel garbage collection. have been proposed, and their theoretical properties are well-understood, no non-trivial implementation exists We have established that we can modify and build a with an evaluation mechanism that is strict by default, recent version of GHC from source, thus allowing the i.e., exhibiting the behavior most commonly expected by possibility of not needing to parse GHC’s STG dump. programmers and required for high performance. By its However, given that STG is itself a complete higher-order, nature such a system will be able to automatically extract polymorphic, pure functional programming language in maximal parallelism from a program, and in the presence its own right, we regard the bridging of the gap between of uncommitted compute resources allow speculative the GHC specifics and STG as reported in the literature to evaluation of results that may or may not be needed in be of lesser immediate priority than realizing the missing the future. Further, it enables transparent (to the pro- parts of the toolchain—the compilation of STG to ma- grammer) fault tolerance mechanism as previously de- chine code and the creation of the supporting runtime scribed, and a type system that supports efficient deep system, which are the core functionality needed for the embedding of DSLs. 227

Page 228

overall goals of the project. Future Work Our goals for FY16 are as follows: Our initial proof of concept was the modification of an existing STG interpreter to provide variable-strictness • Largely complete serial implementation semantics: from the provided “lazy” (call-by-need, or • Begin implementation of runtime support for parallel non-strict with sharing) to, e.g., strictness in function argu- evaluation ments, constructors, or “let” expressions, while preserving the sharing. This provides our reference semantics for a The corresponding tasks are as follows: compiler-based implementation. Serial implementation We defined an abstract machine that is implemented by 1. Implement variably-sized heap and stack objects the code generator. Analogously to the simulator, it is a 2. Generalize support for unboxed types modification of the STG machine as reported in the litera- 3. Re-implement garbage collector for (1) and (2) ture, using variants of the “eval/apply” evaluation strategy 4. Investigate performance implications of using native appropriate for the specified degree of strictness, and later tail calls parallelization. 5. Generally improve code generation 6. Add explicit compiler flags for specifying strictness This latter is rather more general than originally envi- properties sioned: as proposed only fully strict semantics would be supported. We determined, however, that to be “currying Parallel evaluation friendly,” that is, to efficiently support closures and partial 1. Design and implementation of thread pool function application with proper scoping and lifetimes, re- 2. Design and implementation of multiple-heap mecha- quired machinery that readily generalizes to the STG model nism and so implemented it in its full generality. 3. Add par language construct for explicit task parallelism Critical to such an implementation is the ability to realize Conclusion efficient tail calls (interprocedural “jumps”). Our original First is the strict pure functional language implementation proposal was to directly generate LLVM (Low Level Virtual itself, as open source—the first of its kind, even disregard- Machine) code. However, given the unusual nature of the ing the automatic parallelization, explicit speculative control structure it was more expedient to implement a evaluation, and resilience mechanism that this system will subset of the portable machine language C— that is compil- implement. Second, the practical validity of arguments for able with a C compiler, including GCC and and Clang/LLVM. and against the strict but pure functional paradigm, in the Serendipitously we now have two independent mecha- context of scientific computing, while maintaining purity, nisms for tail calls: via C—, and directly from the C compil- will be testable. Third is the demonstration of utility of an ers given mild constraints on the C— implementation, and environment in which embedded DSLs may be developed still benefit from the LLVM (or GCC back-end) optimization. and their space/time performance meaningfully evaluated. We set up a distributed revision control system (Git), an automated building/testing system (Jenkins), project wiki, etc., to manage our project documents and code, and as a template for the eventual open-sourcing of the software. The research team also implemented a parser, a constraint- based Hindley-Milner-Damas type inferencer, a compiler (serial code only at this point) from STG to our implemen- tation of C—, a supporting runtime system, and a small test suite. The runtime system includes a 2-space copying garbage collector. In other words, we have an operational, demonstrable compiler toolchain. As such we are right on target with the Gannt progress chart in the original pro- posal. 228

Page 229

Information Science & Technology Exploratory Research Continuing Project Inserting Nonlinear N-Material Coupling PDF Information into Turbulent Mixing Models Jozsef Bakosi 20150498ER Introduction which has never been done. The project develops new mathematical models for turbulent mixing. We target applications in which the The project entails mathematics, theory, computation, detailed simulation of turbulence is intractable and the and validation, resulting in implementation of new mod- purpose of “turbulence models” is to predict the statis- els in a LANL production code. tics of turbulent mixing which is far more tractable. We address the mixing problem of multiple materials with Benefit to National Security Missions very different densities. In such problems the velocity Multi-component variable-density turbulent material affects the material fields and more importantly, the mixing is a central element in many LANL programs. material fields affect the velocity field. This results in Material fields are required for a predictive science of strong two-way coupling between the fields that is not turbulent mixing, reaction, and opacity. This work will accounted for in current statistical computations. directly and immediately impact modeling efforts and advance the predictive science in ASC (XCP) and several The work introduces a new principle: Any model for campaigns. The work will result in engineering models, the statistics must be derived from a relevant underly- directly relevant to LANL’s large physics codes, in which ing probability density function (PDF), a higher order the resolution of all scales are intractable and statistical mathematical object that is consistent with the physics methods are the only practical approach. of coupled hydrodynamics and material mixing and con- tains information on all statistics. This ensures physical The results of the work will be translated in the 3rd year correctness at the mathematical and theoretical levels into programmatic engineering efforts by implementa- and has never been done. tion of the developed equations into LANL’s RAGE pro- duction code. To this end we are in contact with Robert We synthesize two evolution equations governing mass Gore (XTD-IDA), project lead of two campaigns. transport in two realms describing turbulent mixing: the mass fractions transport in physical space, and the Progress Master Equation describing the transport of statistical Part of the first FY was spent exploring different prob- information in probability state space. We also apply two ability density functions (PDFs) to see which provide averaging methods: Reynolds averaging in physical space signature behavior of high-order statistical information and sample space averaging. This allows instantaneous seen in high-density-difference turbulent material mix- constraints to be inserted into statistically averaged ing. The developed necessary theoretical principles we equations. This synthesis has never been done. have (so far) focused on binary mixing, which then will be extended to the mathematics of multiple materials Also, we bridge the multi-fluid and single-fluid descrip- mixing models now and next FY. Based on data of high- tions of turbulent mixing. The Master Equation allows fidelity direct numerical simulations (DNS) of variable- accounting for widely different underlying PDFs of the density turbulent mixing between fluids of very differ- state space, representing very different non-equilibrium ent densities, we selected a beta-distributed stochastic stages of the mixing process. This two-pronged strategy process to describe the time-evolution of the binary allows representing both sharp and diffusive material materials PDF due to its favorable characteristics. interfaces within the same method and allows a con- tinuous bridging between multi and single fluid models, 229

Page 230

We have compared to data how the stochastic beta pro- fied and validated large eddy simulation (LES) code for a cess mixes the density,the mass fractions, and the volume number of simulation cases that will be used to produce fractions, in order to choose the correct variable for a data currently non-existent in the literature. Specifically, mixing process. We have studied the three mixing pro- we compute two different types of variable-density turbu- cesses to see which Master Equation (ME) mixing process lent binary mixing flow for validation: a spatially homoge- best represents material mixing physics, specifically, which neous but non-stationary and a spatially inhomogeneous produces a unique non-monotonic behavior of skewness but statistically stationary flow — both with various initial only seen variable-density mixing. We derived multiple configurations not computed previously. The LES results hierarchies of moment equations for statistics describ- together with the existing DNS data will be used to validate ing variable-density mixing to study which has the proper and guide our final specification of the binary material mix mathematical structure. A stochastic beta process for the model as well as the work in the next FYs on the multi- mass fractions (not density or volume fractions) is the only component model. process capable of correctly emulating observations in variable-density turbulent mixing. Besides the new and unexpected Monte Carlo simulations and the LES simulations for validation, we are also plan- The ME for the stochastic beta process for variable-density ning to validate our moment equation binary mix model mixing has three unknown parameters (b, kappa, S) that in an XTD division research code solving for our validation depend on the statistics of the mixing process. We derived cases. This step will help evaluate (1) the principles in our mathematical forms of b, kappa, and S, that ensure the binary moment equation mix model before proceeding to correct asymptotic behavior of the mass fractions PDF and the theory of the multi-material problem, and (2) how our its statistics. These results constitute an important step work fits into the existing moment method (BHR) currently toward the mathematical model not only for the ME but used in XTD. also for the moment equations derived from the ME. This will lead to engineering models that are applicable and can Future Work be inserted into moment closures for binary mixing. These Going forward, the research temam will connect the first theoretical results for the binary process provide a crucial year’s work on binary mixing to the existing moment step towards treating the multi-material problem and also method (BHR) currently used in XTD division. In particular, has some important and direct practical applications for we investigate how some of the statistics computed by binary mix models of programmatic interest (see below). BHR can be specified based on our model. This identifies differences between the models and sheds light on how We have also developed a Monte Carlo simulation code BHR can be improved. to explore the mathematical behavior of the stochastic processes described by Master Equations. This is an unex- Theoretical work on multi-component (ternary) variable- pected development of a novel methodology that was not density mixing will begin. We derive moment equation part of the original proposal and which produces several hierarchies from the Master Equation (ME) and Navier- unanticipated and important results. For example: one is Stokes. We work on specifying approximations of the faced with developing closure approximations for terms unknown terms in our model. This will yield a testable and in the moment equations derived from the ME and from directly useful engineering method for multi-component Navier-Stokes. Using this methodology, we are now in the mixing. position to test the validity of the modeled mixing clo- sures using Monte Carlo simulations of a realizable mixing Our Monte Carlo simulation code will continute to be used process. The importance of this technique becomes clear to validate closure approximation hypotheses but now for when one considers the established practice in turbulence the multi-component case with ternary mixing. The goals modeling of mixing processes. In established practice a here are similar to that of the binary case: using a physi- closure for an unknown unclosed term is first proposed cally realizable mixing process we test the validity of the then tested against data of a quantity in the physical equa- modeled mixing closures as well as the correct asymptotic tion with the unknown term. Using a Monte Carlo simula- behaviors. tion we can directly test the closure approximation for the unknown term independent of other physical effects and We continue the large eddy simulations to produce valida- their approximations. tion data with various initial conditions. Specifically, we compute two different types of variable-density turbulent We have also started work on performing simulations for binary mixing flow for validation: a spatially homogeneous model validation purposes. We will use an existing veri- but non-stationary and a spatially inhomogeneous but sta- 230

Page 231

tistically stationary flow — both with various initial configu- rations not computed previously. We will also extend the capability of the LES code to compute ternary mixing. Also, the research team will run various test cases with an XTD division research code that solves our simpler valida- tion cases using ordinary differential equations governing spatially homogeneous mixing. This is a crucial interme- diate step in model validation as well as comparing our model to BHR and is used for testing only parts of the modeled process independent of spatial transport. Conclusion The new turbulent material mix model will correctly ac- count for the nonlinear coupling and constrained statis- tics in the mixing of multiple materials that do not exist in current approaches and will avoid the conventional passive-scalar approximations that are inadequate for LANL problems. The new model will be implemented in a production code for immediate impact on programmatic efforts. We expect sizable improvements over the currently used passive-scalar approximations for multi-material mix- ing problems and highly visible breakthrough-level results in the area of material mixing to be published in the open literature. 231

Page 232

Information Science & Technology Exploratory Research Continuing Project Long-time Dynamics using Trajectory Splicing Arthur F. Voter 20150557ER Introduction two pillars: the basic understanding of materials, and Molecular dynamics (MD), the simulation of the evolu- IS&T. Indeed, our goal is to develop a novel capability tion of individual atoms in a material, is an essential that would enable the direct simulation of materials at tool to further our understanding of materials at the the atomic scale over extremely long timescales. Such a nanoscale. Due to its inherently serial nature, how- capability is invaluable to investigate the microstructural ever, MD is strongly limited in the timescales it can evolution of materials and understand their perfor- reach, a problem that becomes even more crippling on mance in operation conditions. Applications to materials large-scale parallel computers. There is thus a press- in extreme condition are directly relevant to the lab’s ing mission need for massively-parallel long-time MD mission, for example in support of the MaRIE effort. Ap- simulation techniques. Building on our many years of plication to nuclear materials are also directly relevant experience addressing this issue, we are introducing to BES and OFES. A second crucial aspect of the project a new paradigm in long-timescale simulation: parallel is to develop a novel simulation paradigm that exploits trajectory splicing (ParSplice). In this ParSplice approach, state-of-the-art concepts in computer science. We are a long trajectory of arbitrary accuracy is assembled from especially interested in the scalability of simulations up many short, independent segments generated efficiently to the peta and exa-scales. Doing so requires the devel- in parallel. Through the use of simulation simulators, opment of architecture-aware algorithms and codes that i.e., simple models of the simulation framework, we will are scalable and fault-tolerant. Our project contains an systematically explore the large design space of Par- important performance prediction component that will Splice methods to identify the best performing methods be used to predict and optimize the performance of the on existing and expected next-generation (or exascale) code on various architecture. These activities directly computer architectures, taking into account massive address challenges identified by ASCR. core counts, shorter mean times between failures, and energy costs dominated by data-motion. We will dem- Progress onstrate the efficiency of the approach in production During the first FY of the project, we have analyzed the conditions by investigating problems important to the “certification rules” (which enforce that the result is free mission of the lab such as void and bubble evolution, an- from bias) and determined that they could be safely re- nealing of defects at grain boundaries, and ionic trans- laxed from what we had proposed in the initial proposal. port in complex oxides. For systems where the trajectory This flexibility improves the performance of ParSplice is trapped in small regions of configuration space, we and simplifies the implementation. We have also ex- expect efficiency gains of several orders of magnitude at plored “placement rules”, i.e., the assignation of work on extreme computational scales compared to current ap- the different replicas. We have compared two strategies: proaches. ParSplice will revolutionize the way large-scale one based on the current end point of the trajectory, long-time MD simulations are performed. and one on its projected end point once the work unit that is about to be scheduled is completed. The latter Benefit to National Security Missions strategy proved to be extremely efficient, but it can be The project directly addresses fundamental challenges computationally expensive compared to the former. The of direct relevance to the lab and to DOE/SC/BES and simple strategy however was surprisingly efficient as DOE/SC/ASCR. Our project supports the Scientific compared to the previous state of the art, and could be Discovery and Innovation mission of the lab through leveraged at larger computational scales, when minimiz- 232

Page 233

ing the scheduling effort will be paramount. We have developed and implemented a first production version of ParSplice within LANL’s AMDF code. We have validated it using a simple defect, a trimer on a silver surface. We have assessed its scalability up to hundreds of cores and compared its performance to previous ap- proaches. The initial results are extremely encouraging and show that improvements of up to an order of magnitude are possible with the current code. We are currently apply- ing the code to more complex systems and initial results are very positive. We are now working on a manuscript that reports on our new method. Future Work During the next fiscal year, we will concentrate on improv- ing the scalability of ParSplice in order to be able to exploit peta-scale computational resources, and beyond. This will require designing tailored certification and placement rules that require minimal computational work, little or no syn- chronization, and built-in fault tolerance. It will also require the development of scalable software components (such as segment and state databases) that are able to handle large throughputs. This will likely require the development of a hierarchical management structure, in contrast to the current master/slave implementation. We will de- velop such scaling strategies and demonstrate them on IC resources or on application simulators for extrapolation to extremely large scales. Concomitantly, we will apply our current code at interme- diate scales (~1000 cores) in order to further demonstrate its practical uses in research settings. We will identify systems for which current methods are inadequate and demonstrate how ParSplice can overcome current limita- tions. Possible application areas include complex defects in metals and nanoscale clusters. Conclusion The research team will develop the ParSplice method and deliver an implementation that is nominally exascale ready. This state of readiness will be validated through extensive simulations using our performance prediction approach. The design and parameter spaces will be thoroughly explored so as to be ready to adjust the implementation novel architectures emerge. We will also deliver a thor- ough formal understanding of the method. Based on this analysis, the class of systems that would optimally benefit from ParSplice will be identified. The power of the ap- proach will be demonstrated on leadership-class comput- ers on physical problems of interest to the lab. 233

Page 234

Information Science & Technology Exploratory Research Continuing Project Spatial and Extreme Value Processes for Bridging Micro- and Macro-Scales in Materials Scott A. Vander Wiel 20150594ER Introduction in how we represent such processes. The programs here Advances in computing power have made it possible at LANL with interest and need in these technological to simulate very small physical systems in remarkable areas are ASC PEM Materials Modeling, and several cam- detail, often making use of “first principles,” or at least paigns. With increased interest in advanced and additive very reliable mathematical models for system evolution. manufacturing processes, a rigorous statistical linkage to However, bridging information from simulations at very the material microstructure as we are proposing to do small scales to the macro-scale required for modeling here will be important. In addition, the Office of Science large-scale systems (e.g. the global ocean, a nuclear is interested in funding work which offers the potential reactor, an implosion) remains a challenging and open to promote energy efficiency. By improving our ability to problem. In materials science, advances in bridging predict damage and failure events in metallic materials, scales are needed to develop more predictive models transportation sector industries can make more efficient of material behavior given its micro-scale composition, use of materials and reduce the weight of vehicles and enabling more reliable predictions of material properties reduce manufacturing energy intensity. in new, extreme environments, and enabling the design of new materials. Progress Flyer plate experiments on tantalum samples in combi- This project focuses on a basic problem in material sci- nation with macro-scale damage model calculations of ence: How do heterogeneous, micro-scale descriptions these experiments have provided quantitative informa- of a polycrystal affect macro-scale mechanical response tion on stress conditions at the estimated point when properties? Here the seemingly random spatial distribu- material voids begin to grow. These boundary condi- tion of micro-scale material properties often results in tions are used as inputs to meso-scale polycrystalline spatially coherent stress fields (under loading), leading models to interrogate the behavior of interacting tanta- to damage or failure in the material. This project will lum grains. focus on developing statistical theory, models, and tools to advance our ability to estimate macro-scale mechani- Finite element models of tantalum microstructure have cal response properties from heterogeneous, random, been created with a 3D voxel structure that matches microscale material descriptions. distributions of grain size, shape and crystal orientations. We can provide an endless supply of statistically equiva- Benefit to National Security Missions lent microstructure realizations for further modeling There is a demonstrated need for advancement in efforts. computational prediction of damage and failure in poly- crystalline metallic materials, particularly for our weap- We apply external stresses with time, representative of ons calculations. This is largely a result of inadequate those in flyer plate experiments to the polycrystalline physics and material statistics representation. Although finite element meshes to produce stress fields through- the computation mechanics and mechanics of materials out a representative volume of tantalum grains. These communities continue to make advances in this arena, calculations show that stress tends to elevate at interfac- much of the work is lacking proper rigor in how extreme es between two grains and more-so where three grains value physical processes are represented mathematical- meet. ly. This project is designed to address this shortcoming 234

Page 235

We have developed a statistical model to represent the tions for the polycrystal, as well as under different stress field in a multi-grain volume. The model incorpo- dynamic loads. rates several features observed in meso-scale computa- • Develop statistical models of the spatial dependence tions: (i) each grain has its own nominal stress level, (ii) and strength anisotropies describing the polycrystal. stresses tend to elevate at some, but not all, grain bound- • Develop statistical models of the resulting stress field, aries, (iii) boundary locations with elevated stress are characterizing the extremes of the field, as well as correlated across grains. Fitting of the statistical model is their spatial dependence. challenging because meshes are on the order of half a mil- • Search out and catalog physical data on polycrystal lion elements and this translates into vary large covariance configurations that are relevant to this effort. matrices that represent spatial dependence in stress fields. Conclusion Implementation of the statistical model is nearly com- This effort will demonstrate a viable approach for statisti- plete. We represent stress fields as integrals of Gaussian cally characterizing micro-scale properties of a polycrys- Markov random fields. The structure of this model allows tal, and using these properties to predict the material’s large-scale correlated fields to be computed blockwise in a strength and damage properties at a macro-scale, under Monte Carlo Markov chain (MCMC) calculation. The result some rather simple loading scenarios. We also expect this is a posterior distribution of parameters that control corre- effort will illuminate promising new directions for analysis lation length scales and distances that elevated boundary approaches that can bridge the micro- and macro-scales. stresses penetrate into grain bodies. In particular, we will seek new statistically rigorous repre- sentations of damage nucleation in polycrystalline materi- This is a totally new approach to characterizing stress fields als. As new experimental facilities come online, this work in polycrystalline materials with the potential to statisti- will serve to get some of the necessary analytical tools in cally bridge the gap between meso-scale and macro-scale place for gaining understanding from such experiments. models describing and predicting how materials fail. Writ- ing the initial implementation of the statistical model has Publications been challenging, but we are just now getting results and Bronkhorst, C. A., N. Bourne, G. T. Gray III, V. Livescu, C. B. enhancements to the model form will be much easier to Storlie, S. A. Vander Wiel, E. K. Cerreta, D. J. Luscher, implement than creation of the initial code from whole M. Ardeljan, and M. Knezevic. Deformation Induced cloth. Porosity Nucleation Mechanisms in Polycrystalline Me- tallic Materials (keynote lecture). Invited presentation Sparse matrix derivations of likelihood calculations re- at International Symposium on Plasticity 2015. (Mon- quired for MCMC implementation have been written up tego Bay, Jamaica, 4-8 January, 2015). and will form the core technical portion of a paper to be Bronkhorst, C. A., N. Bourne, G. T. Gray III, V. Livescu, C. B. submitted to a statistics journal. Storlie, S. A. Vander Wiel, E. K. Cerreta, D. J. Luscher, M. Ardeljan, and M. Knezevic. Deformation Induced Curt Bronkhorst presented some of this work in invited Porosity Nucleation Mechanisms in Polycrystalline Me- presentations at three conferences in 2015: (i) Mach Con- tallic Materials. Presented at Mach Conference 2015. ference for Multiscale Materials; (ii) TMS Annual Meeting; (Montego Bay, Jamaica, 8-10 April, 2015). (iii) International Symposium on Plasticity. Bronkhorst, C. A., N. Bourne, G. T. Gray III, V. Livescu, C. B. Storlie, S. A. Vander Wiel, F. L. Addessio, D. J. Luscher, Future Work M. Ardeljan, E. K. Cerreta, and M. Knezevic. Porosity This effort will be conducted in two main phases. In the Based Damage and Failure in Polycrystalline Tantalum first phase, statistical theory and methodology will be de- – Structural Linkages (invited). Invited presentation at veloped in concert with a simulation campaign focused on TMS 2015. (Orlando, Florida, 15-19 March, 2015). simulated materials. Once methodology has been suffi- ciently refined, and kinks and problems have been worked out in the synthetic setting, we expect to move focus to actual materials. Over the next fiscal year, we will: • Design and carry out dynamic loading simulations of polycrystalline structures, under different initial condi- 235

Page 236

Information Science & Technology Exploratory Research Final Report A Computationally Efficient Model for Warm Dense Mixtures Didier Saumon 20130244ER Abstract are elusive. It is a frontier of HEDP of great practical Warm dense matter (WDM) encompasses ionized fluids importance. at the confluence of condensed matter physics, plasma WDM mixtures lie at the heart of problems in planetary physics, and dense fluids and cannot be described ad- science and stellar astrophysics. Examples include the equately by the theories of these well-established fields. phase diagram of hydrogen and helium in Jupiter and It is a frontier of high energy density physics (HEDP) that Saturn and the gravitational stratification of elements presents considerable challenges to theory and experi- in white dwarf stars. WDM mixtures also occur during ments. the implosion phase of inertial confinement fusion (ICF) We developed theoretical and computational tools to capsules. model mixtures in the WDM regime. This is evolving WDM can be modeled successfully with large scale ab into a new, state-of-the-art capability to model extreme initio simulations, primarily with quantum molecular states of matter at LANL that will support future experi- dynamics (QMD) [1]. However, this method becomes ments at HEDP facilities and numerous programmatic prohibitively expensive at temperatures above 10eV for applications involving the hydrodynamics of dense most systems. At higher temperatures, a simpler ver- plasma mixtures. sion that uses the Thomas-Fermi model of the electron Our approach is based on a suite of approximate models - rather than the physically more accurate Schrödinger that account for the most important physics of warm, equation - can be applied. Nonetheless, this orbital-free dense plasmas while remaining computationally eco- molecular dynamics (OFMD) method remains computa- nomical. It is ideally suited to produce large tables of tionally expensive [2]. constitutive physical properties of materials such as Our main goal is to develop a new approach to model the equation of state (EOS) and transport coefficients WDM that is computationally much more efficient than for a wide range of applications. Most notably, we have QMD and OFMD with a more approximate yet realistic developed tools of remarkable computational efficiency description of the physics of dense plasmas. It would and very good accuracy to generate tables of diffusion then become practical to generate large tables of WDM coefficient for WDM mixtures. We have demonstrated properties for many different applications. In this the validity of our models by extensive comparisons project, we extended our single-component model to with results obtained with state-of-the-art ab initio mixtures and computed EOS and diffusion coefficients. simulations methods. We present applications to dense We have made several important advances and we have plasma mixtures encountered in white dwarf stars. been very successful in reaching our goal. Background and Research Objectives Scientific Approach and Accomplishments Warm dense matter conditions typically occur at tem- This LDRD project is an extension to mixtures of our peratures of 1eV to several 100eV and densities from research on dense, single-component plasmas which is 0.1 to 100 times normal solid densities [1]. WDM is funded separately. Thus the following exposition neces- too hot to be treated by the standard methods of sarily involves some discussion of our single-component condensed-matter and liquid-state physics, but too cool work. to be modeled as a weakly coupled plasma. It lies in an intermediate region where simplifying approximations 236

Page 237

Formalism for mixtures We have validated the underlying PYRRHO model of mix- We formulated a multiple-component mixture theory tures against published QMD simulations of a plasma of by expanding that of our single-component microscopic carbon and hydrogen relevant to ICF capsules [4]. The pair model of dense plasmas called Quantum Hyper-Netted distribution functions of ions in the C-H plasma are in ex- Chain (QHNC) [3]. We arrived at a much simpler form than cellent agreement at high temperatures and high densities. we originally anticipated. This set of equations is not only Deviations arise and grow steadily at lower temperatures valid for binary mixtures but is readily extended to mix- and densities because of the formation of C-C chemi- tures of an arbitrary number of ion species and quantum cal bonds in the plasma, something that PYRRHO cannot mechanical electrons [4]. describe. This pointed out a fundamental limitation of our pseudo-atom model for those material that are not fully Pseudo-atom model dissociated in the WDM regime. Nonetheless, we were The microscopic model for the bound states of the plasma able to validate PAMD for mixtures against OFMD simula- ions, the electronic screening, and the fluid structure of tions of the C-H mixture for the same conditions since the plasma is no longer the QHNC model, but evolved chemical bonds cannot form in the OF approximation. into an original variation that has proved far superior for modeling WDM. This new model, called PYRRHO, assumes We have compared our ionic pair distribution functions that the plasma is made of identical pseudo-atoms. The with published OFMD simulations for a mixture of copper- pseudo-atom is composed of a nucleus and its surround- deuterium and for a carbon-oxygen mixture computed ing spherically symmetric electron cloud of bound and with our own OFMD code, all with excellent agreement. screening electrons [5]. The structure of the pseudo-atom We also found excellent agreement with the OFMD EOS of is obtained with an average atom model in which the a Fe-He mixture (Figure 1) [15]. electrons can be described either with a full quantum mechanical (Kohn-Sham, or KS) model or in the orbital-free (OF) approximation. The pseudo-atom is combined with the integral equations of fluid theory for a complete, self- consistent model of the plasma. The solution of the set of equations gives a detailed microscopic description of the plasma without free parameters. The PYRRHO model is far more stable numerically than the original QHNC. The KS mode of PYRRHO runs well up to several 100eV, retaining a full quantum mechanical description of the plasma over the full regime of WDM, which is generally not possible with QMD. The PYRRHO model was generalized to mix- tures within the formalism described above. The PYRRHO model has been extensively tested against single-component warm dense plasmas and turns out to be very successful [4-7]. We have demonstrated the ability to solve the system of equations for up to seven-compo- Figure 1. Excess pressure (contributions other than non-inter- nent mixtures of dissimilar elements and mixtures where acting gas) of mixtures of iron (Fe) and helium (He) at a tempera- one component is a trace species, as well as ions with very ture of 50eV and density of 10g/cm3. Our OF-PAMD calculations are shown by open circles which are compared to state-of-the-art different charges and/or masses. Such plasmas are difficult OFMD simulations [15]. The agreement is better than 1.5% for to model with ab initio methods. all Fe:He mixing ratios, which validates the OF-PAMD moled for a highly asymmetric plasma mixture in the WDM regime. From [7]. The ion pair distribution function is the quantity predicted by PYRRHO that is most readily compared with other calcu- lations. We have made such comparisons in the WDM re- Pseudo-atom molecular dynamics gime with pair distribution functions from QMD and OFMD The PYRRHO model is described by a set of algebraic equa- simulations of elements as varied as hydrogen, carbon, tions and is static by construction. However, it generates beryllium, aluminum, iron and tungsten for temperatures ion-ion potentials for all components pairs in the mixture. of a few to hundreds of eV, all with excellent agreement [5, Those potentials can be used in a classical molecular 6]. dynamics simulation from which we can extract quantities that are intrinsically dynamic in character, i.e. those that 237

Page 238

involve particle collisions. The combination of PYRRHO uncertainty in the KS-PAMD result) as long as the interactions and classical molecular dynamics, which we call Pseudo- are not too strong (T>5eV at this density). We obtain very similar Atom Molecular Dynamics (PAMD) is a significant advance results at densities of 2.7 and 27 g/cm3. Adapted from [7]. and has become the central tool of our program to model A main goal of our project is to compute ion diffusion WDM [7]. coefficients, which are a much more sensitive test of the model’s ion-ion effective potential than pair distribution PAMD is similar to QMD and OFMD: it combines a quan- functions. The OF-PAMD diffusion coefficients for deuteri- tum mechanical (or orbital free) description of the elec- um, boron, iron and copper, over a wide range physical re- trons with the classical motion of the ions. With PAMD gimes all fall within the uncertainty of the published OFMD we can calculate the same plasma quantities as ab initio values [8, 9]. Figure 2 shows the diffusion coefficient of simulations but 10-100 times faster with MD simulations aluminum computed with OF-PAMD compared to our that typically contain 10-100 times more particles and run OFMD. Except for the lowest temperatures of 1-2eV, they for 10 times more time steps. These larger and longer agree within 10%, which is on par with the state-of-the-art. simulations reduce the statistical noise due to random fluc- Figure 2 also shows the quantum mechanical KS-PAMD tuations in the system. QMD and OFMD simulations, being results which differ from the OFMD result for T<100eV limited to small numbers of particles, typically give noisy or because of their more accurate treatment of the electrons. inaccurate results for transport coefficients. Furthermore, KS-PAMD and OF-PAMD agree very well at higher tempera- PAMD allows the modeling of mixtures where one species tures, as expected on physical grounds. This demonstrates is a trace component (1:100 or smaller mixing ratio) which that the PAMD model works very well over a wide range of is currently not possible with ab initio methods. Its relative conditions for a quantity as sensitive as the self-diffusion speed makes it possible to calculate large tables of physi- coefficient, it can also be run in the full quantum mechani- cal properties, something that still requires a rather heroic cal mode to very high temperatures, thus providing a con- effort with ab initio methods. tinuous description of the physics with a single model over a wide range of conditions, unlike QMD/OFMD. Effective potential theory In collaboration with Scott Baalrud (U. of Iowa), we devel- oped a new approach to compute ionic diffusion coeffi- cients in dense plasmas. In dilute plasmas, diffusion can be treated by considering only binary collisions between ions through the Chapman-Enskog formalism. This approach is known to fail dramatically as the interactions in the plasma grow and multiple collisions dominate. Introducing an effective ion-ion potential that accounts for ion correla- tions, the “effective potential theory” (EPT) extends the domain of applicability of the Chapman-Enskog formalism to moderately coupled plasmas [10, 11]. This is extremely fast computationally. Since the PYRRHO model generates the effective potential that is the input to the EPT model, we had an exciting opportunity to apply the EPT model to WDM and extend it to mixtures. The power of the EPT model is demonstrated in Figure 2 Figure 2. Coefficient of self-diffusion of aluminum at 8.1g/cm3 and temperatures from 1 to 500eV calculated with several mod- where it agrees with the KS-PAMD values to better than els. The excellent agreement between the reference calculation 2% over nearly the full range of conditions shown. At the of OFMD (green dots) with our OF-PAMD (blue dots) validates lowest temperatures, deviations rise rapidly as the plasma latter. The full quantum mechanical calculation KS-PAMD (black interactions become very strong and the EPT model fails. dots) diverges from the two OF calculations because it is a physi- Figure 2 summarizes our accomplishments in our ability cally more realistic description of the electrons. This is clearly to calculate diffusion coefficients in WDM: 1) OF-PAMD important at temperatures below 100eV. KS-PAMD merges gives very nearly the same values as the reference OFMD smoothly with OF-PAMD at higher temperatures, as expected. calculation, 2) the more physically accurate KS-PAMD can The red curve shows the results of the effective potential theory be run over the full range of density and temperature and (EPT) using the same ion-ion effective potential as in KS-PAMD. recovers the OF-PAMD results at high temperatures, and 3) The agreement with KS-PAMD is essentially perfect (within the 238

Page 239

the EPT model agrees perfectly with KS-PAMD over most of inter-diffusion over a wide range of densities, temperature the WDM regime. and mixing ratios. We estimate that a table of about 1000 points is necessary. No one would consider doing such a This is a remarkable result. We have demonstrated that calculation with ab initio methods but it is manageable we can compute the diffusion coefficient by combining with our new tools. We have computed the PYRRHO po- the PYRRHO calculation of the effective ion pair potential tentials for 700 points in the parameter space of interest. with the EPT model over much of the range of WDM condi- The completed table will be made available for inclusion in tions. There is no need to resort to noisy and lengthy MD white dwarf evolution codes. simulations, except in the strong coupling cases where EPT fails. This results in enormous savings in computing time. It is a transformative development for the computation of transport properties. Diffusion in white dwarf stars The calculation of ionic inter-diffusion coefficients in white dwarf demonstrates the new capability that we have de- veloped for mixtures. The very large surface gravity of white dwarf stars causes heavier elements to sink below the surface. This results in a layered structure with a very thin surface layer often composed of pure hydrogen, with an underlying layer of pure helium, on top of a core of carbon. The hydrogen layer is missing in about 25% of white dwarfs. Many white dwarf stars show traces of heavier elements (Mg, Ca, Si, Fe, etc.) at the surface that are thought to be of planetary origin, such as a dust disk or an asteroid belt. Those ele- ments diffuse downward in the star below the observable surface. The exciting astrophysical implication is that by measuring the amount of these accreted elements at the Figure 3. Coefficient of diffusion of a trace of Si in a dense He plasma. The values of temperature and density follow those surface, and with a theoretical knowledge of how fast they along the bottom of the convection zone in a sequence of cooling diffuse downward, we can infer the composition of the models of white dwarf stars, with hotter/younger white dwarfs orbiting planetary material which is otherwise invisible. on the left and cooler/older white dwarfs on the right. Tempera- ture varies along the density axis. Three calculations of the coef- In collaboration with Gilles Fontaine (U. de Montréal) we ficient of inter-diffusion of Si-He are shown. Paquette et al. (1986) have calculated the coefficient of diffusion of traces of Ca [12] is the theory of diffusion used in all white dwarf models that and Si for a range of white dwarf models with both H and we seek to update. KS-PAMD is our model for dense plasmas He surface layers. The results for the diffusion of Si in He coupled with molecular dynamics. The error bars and the scatter are shown in Figure 3. The agreement between KS-PAMD of those points reflect the statistical uncertainty in the calcula- and EPT is quite good but not perfect. The reason is that tion of the diffusion coefficient in a mixture containing a trace Si is at a trace concentration (1:200), which is challenging element (Si:He = 1:200 here). The green curve show the EPT for any MD simulation (and impossible to do with ab initio result, which is not only very smooth but is also computationally simulations) giving noisy KS-PAMD values. The smooth- much faster. Our more sophisticated model for dense plasmas lead to diffusion coefficients that are consistently 25-100% larger ness of the EPT result illustrates that our novel approach than those of Paquette et al. [12]. can handle the diffusion of traces without difficulty. We find that the diffusion coefficients that are universally used Impact on National Missions in white dwarf models [12] are usually within 50% of our NASA and NSF: The primary applications thrust of our more accurate calculation, but are up to a factor of two research is astrophysical – with an initial focus on under- lower at the highest density. standing heavy-element diffusion in white dwarf stars. In Another class of white dwarfs shows traces of carbon at a broader sense, this research created a robust capability the surface [13]. The carbon comes from the core and is for scientists to develop accurate equations of state and brought up to the surface by a convective/diffusive pro- transport coefficients for wide varieties of astrophysical cess [14] whose modeling requires the coefficient of C-He phenomena. 239

Page 240

DOE/Understanding of Materials: Warm dense matter is a vergence of the self-diffusion coefficient and viscosity frontier of materials physics. Because it describes a region obtained with Thomas-Fermi-Dirac molecular dynam- of temperature and density space devoid of simplifying ap- ics. 2012. PHYSICAL REVIEW E. 85 (6). proximations, all of the physics matters. It is also a region 10. Baalrud, S. D., and Daligault. Effective Potential Theory difficult to probe experimentally. The capability to model for Transport Coefficients across Coupling Regimes. mixtures in this regime advances our basic knowledge of 2013. PHYSICAL REVIEW LETTERS. 110 (23). materials behavior in extremes and our ability to design and analyze HEDP experiments. 11. Baalrud, S. D., and Daligault. Modified Enskog kinetic theory for strongly coupled plasmas. 2015. PHYSICAL NNSA/DP & Nuclear Weapons: During a nuclear explosion, REVIEW E. 91 (6). and in inertial confinement fusion, mixtures of warm dense matter can be created. This research potentially enhances 12. PAQUETTE, C., C. PELLETIER, G. FONTAINE, and G. our capability of modeling materials as they pass though MICHAUD. DIFFUSION-COEFFICIENTS FOR STELLAR this regime. PLASMAS. 1986. ASTROPHYSICAL JOURNAL SUPPLE- MENT SERIES. 61 (1): 177. NNSA/Non-proliferation: The ability to make accurate equations of state for a wide variety of warm, dense, 13. Dufour, P., P. Bergeron, and G. Fontaine. Detailed mixed materials lends itself to better anticipating and spectroscopic and photometric analysis of DQ white ameliorating the effects of rogue WMDs. A calculation for a dwarfs. 2005. ASTROPHYSICAL JOURNAL. 627 (1): 404. mixture of programmatic interest was initiated in FY15. 14. PELLETIER, C., G. FONTAINE, F. WESEMAEL, G. MI- References CHAUD, and G. WEGNER. CARBON POLLUTION IN

1. Desjarlais, M. P., J. D. Kress, and L. A. Collins. Electrical HELIUM-RICH WHITE-DWARF-ATMOSPHERES - TIME- conductivity for warm, dense aluminum plasmas and DEPENDENT CALCULATIONS OF THE DREDGE-UP PRO- liquids. 2002. PHYSICAL REVIEW E. 66 (2). CESS. 1986. ASTROPHYSICAL JOURNAL. 307 (1): 242.
2. Lambert, F., J. Clerouin, and G. Zerah. Very-high-tem-
3. Danel, , Kazandjian, and Zerah. Orbital-free molecular perature molecular dynamics. 2006. PHYSICAL REVIEW dynamics simulations of a warm dense mixture: Ex- E. 73 (1). amination of the excess-pressure matching rule. 2009. PHYSICAL REVIEW E. 79 (6).
4. CHIHARA, J.. UNIFIED DESCRIPTION OF METALLIC AND NEUTRAL LIQUIDS AND PLASMAS. 1991. JOURNAL OF Publications PHYSICS-CONDENSED MATTER. 3 (44): 8715. Saumon, D., C. E. Starrett, and J. Daligault. Diffusion coef- ficients in white dwarfs. 2015. In 19th European White
5. Starrett, C. E., Saumon, Daligault, and Hamel. Integral Dwarfs Workshop. (Montreal, Canada, 11-15 Aug. equation model for warm and hot dense mixtures. 2014). Vol. 493, p. 419. San Francisco: Astronomical
6. PHYSICAL REVIEW E. 90 (3). Society of the Pacific.
7. Starrett, C. E., and Saumon. A simple method for deter- Starrett, C. E., J. Daligault, and D. Saumon. Pseudoatom mining the ionic structure of warm dense matter. 2014. molecular dynamics. 2015. PHYSICAL REVIEW E. 91: HIGH ENERGY DENSITY PHYSICS. 10: 35. 013104 (5).
8. Starrett, C. E., and Saumon. Electronic and ionic struc- Starrett, C. E., Saumon, Daligault, and Hamel. Integral tures of warm and hot dense matter. 2013. PHYSICAL equation model for warm and hot dense mixtures. REVIEW E. 87 (1). 2014. PHYSICAL REVIEW E. 90 (3): 033110 (8).
9. Starrett, C. E., Daligault, and Saumon. Pseudoatom molecular dynamics. 2015. PHYSICAL REVIEW E. 91 (1).
10. Lambert, , Clerouin, and Mazevet. Structural and dy- namical properties of hot dense matter by a Thomas- Fermi-Dirac molecular dynamics. 2006. EUROPHYSICS LETTERS. 75 (5): 681.
11. Danel, J. -. F., Kazandjian, and Zerah. Numerical con- 240

Page 241

Information Science & Technology Exploratory Research Final Report Software/Hardware Mapping for Data Locality Optimization Hristo N. Djidjev 20130252ER Abstract Background and Research Objectives This project addresses the problem of software-hard- A new approach to addressing the challenges of solving ware codesign for next-generation computing systems computational problems at the exascale is to apply a as an optimization problem. Compute-intensive por- software-hardware codesign approach, in which the soft- tions of an application software and a hardware archi- ware and the corresponding hardware for solving a class tecture are parameterized by a set tunable of software of application problems are jointly co-developed and and hardware parameters and the problem is to find co-optimized. In contrast, the current approach to high a combination of these parameters that optimizes the performance computing (HPC) consists of two separate software-hardware pair with respect to performance and independent phases, first, hardware is designed and characteristics. A key step of such an approach is to be delivered to users; second, software for the application able to model the energy consumption and the running of interest is developed or adapted to run on that new time, key components of any cost function, as analytical hardware. However, in the overwhelming majority of functions of the software and hardware parameters, only cases, it is realized that, because of mismatch between having the abstract hardware model available (and not the data communication patterns and the hardware ar- the real hardware). chitecture, the available computing resources cannot be efficiently exploited, and the originally promised perfor- In order to develop an energy model, we assume that mance is hard or impossible to achieve. the code portion of interest is a stencil kernel and that it has already been optimized in a way that is beneficial In order to develop a methodology for the codesign of for any modern architecture, namely, into a form that software and hardware, we first look at properties that minimizes the number of data transfer operations. For are desirable for most software applications, regardless this purpose, we have used existing tools based on the of the specifics of the hardware architecture. One of the polyhedral method that divide the set of all operations most desired properties of a good mapping is the locality into blocks of code called tiles that can be executed of the data access. Modern computer processors such independently of each other. Then our energy model as Graphics Processor Unit (GPU) and Cell have limited evaluates the energy needed to execute each tile, and amount of local memory and the time of data transfer represents the total energy as a sum of the energies for between that memory and the global memory is huge the individual tiles. A similar approach is used to model compared to the cost of computation or the cost of local the time. access. Hence, it is important to develop a method that maps the operations of a software code onto the hard- Having analytical formulas for the time and energy as ware in a way that maximizes locality and data reuse, functions of the tunable parameters makes it possible i.e., such that as large as possible amount of computa- to run very efficient optimizations in order to find the tions are performed on the data in the local memory best choice of parameters for minimizing the time, the before that data is replaced with new data. In addition, energy, or a combination of both. Our validation analysis one wants to preserve the semantics of the original on a Graphics Processing Unit (GPU) architecture shows code and hence guarantee that the mapping will not that both models are very accurate and that the associ- affect the code correctness. The corresponding mapping ated optimization problem can determine optimal tile optimization problem is very difficult to solve efficiently sizes with respect to the energy consumption with aver- as it is in the NP-hard class and because it needs to be age error less than one percent. run multiple times, since it is a step of the software/ 241

Page 242

hardware space exploration loop that is invoked iteratively. the consumed energy of a software implementation. While Hence, we will look at a subclass of all software codes, energy and, especially, time are easy to measure, our task namely nested-loops, and at a subclass of all transforma- is much more difficult, as we need to estimate these pa- tions, namely affine transformations. Previous research rameters on abstract representations of the hardware and has shown that those subclasses capture the most relevant software. types of codes and transformations from a practical point of view (as loops are usually responsible for most of the Scientific Approach and Accomplishments total computation time) and, on the other hand, result in Approach overview optimization problems of manageable computational com- One of the most challenging issues of the proposed op- plexity. Our goal is to address the algorithmic challenges timization framework is searching the space of all pos- and to design mapping algorithms that are both accurate sible software-hardware pairs and finding the one that and scalable and to test their performance on a prototype is optimal with respect to the chosen cost function. The mapping tool. difficulty of this problem is that the space is huge and it will not be possible to exhaustively look at and evaluate We focus in our analysis on a subclass of the tiled nested all of its members. On the other hand, it is not clear how loop codes called dense stencils, which occur frequently in can one explore the structure of this space. One of the the numerical solution of PDEs and in many other contexts important observations made in this project is that the such as high-end graphics, signal and image processing, computational complexity of the optimization procedure numerical simulation, scientific computing, and bioinfor- can be significantly reduced, compared with the procedure matics. We chose stencils for our case studies since that originally described in the proposal, by eliminating the would allow us to model an entire class of algorithms in a need to search the space of all hardware parameters for an hierarchical way with a single generic model representing optimal solution. Previously, for each set of hardware pa- the whole class, while model parameters that are stencil- rameters (each point in the search space), one had to run dependent have to be separately specified for each stencil a problem for optimally mapping the software representa- of interest to complete its model. (However, the approach tion onto the corresponding hardware. The new approach is applicable to any other class of nested-loop codes that directly finds the sets of the optimal hardware parameters allows tiling.) We completely develop and validate the by including them as variables in the mapping optimization detailed models (including the stencil-dependent param- problem. For this to work, we need to be able to model as eters) of three specific stencils. Models for other stencils an analytical function the energy and the running time of can be developed in a similar way. a software implementation of a parameterized algorithm on a parameterized hardware, in particular, a parameter- In order to increase the amount of parallelism and im- ized model of a Graphics Processing Unit (GPU). We next prove the locality of memory accesses, it is often useful describe the development of such models. to divide the set of all computations of a code into groups called tiles [3]. Tiles are considered atomic, i.e., all the data Energy and time modeling needed to perform the computations in a tile can be made The input to our model is a tiled nested-loop code that available before the computation begins and there are no is optimized for execution time. Therefore, we assume cyclic data dependencies between tiles (tiling is legal). We the code is free of shared memory bank conflicts and the also note that in order to obtain the best performance, til- global memory accesses are coalesced. Then, we model ing is often applied hierarchically at multiple levels. energy of the code as an analytical function of software pa- rameters and GPU hardware parameters. For example: tile Unfortunately, in most cases tiling is only possible after sizes, program input parameters, the number of operations the code is transformed into equivalent, but easier to tile in the loop body are software parameters and the energy code. There are many transformations that can result in consumed by an operation (i.e., addition, multiplication, tillable codes, and they can be further distinguished by the etc.) is a hardware parameter. The list of all parameter amount of data communication and the amount of paral- used is given in Figure 1. lelism they offer. One can find an optimal transformation using the polyhedral method and some of the publicly We chose to model energy since energy efficiency has available tools based on it such as PLUTO [1]. been recognized as one of the biggest challenges in the roadmap to higher performance (exascale) systems for a Finally, in order to choose an optimal software/hardware number of reasons including cost of usage, reliability, en- combination, one needs an objective (cost) function that ergy conservation, and environmental impact [7]. The most can be used to rank the candidate solutions. Typical com- powerful computers today consume megawatts of power, ponents of such cost functions are the running time and 242

Page 243

enough to power small towns, and at cost of millions per always the case as the execution of some operations can year. overlap in time due to parallelism. Another important fea- ture is that the value of the predicted time is used as a pa- rameter in the energy model. The reason is that energy has two main components: static energy, which is the energy consumed when the processor is idle, and the dynamic energy, which is the energy needed to do the computation. Time is needed to evaluate the static energy portion, which is a product of the run time and the static power. One of the advantages of our modeling approach is that our models predict energy and time parameters of a soft- ware implementation on a given hardware by analyzing source code only, unlike other approaches [4], [5] that rely on parameters computed by running benchmarks for each individual code. We do use micro-benchmarks, but they are needed to characterize hardware, rather than codes. Validation: We validated our energy model on three stencil codes, Smith Waterman, Jacobi 1D and Jacobi 2D. Given two molecular sequences (like genomes), the Smith- Waterman algorithm (SW) solves the problem of finding an optimal alignment of the sequences that maximizes the overlap or other similarity measure. Using the dynamic programming methodology, SW iteratively constructs a ta- ble that identifies the optimum alignment. This algorithm for the iterative construction of the table has the struc- Figure 1. List of the parameters. S/H/P stand for software/ ture of a stencil code. The second algorithm, Jacobi 1D hardware/problem type, D/I stand for problem dependent/in- (1-dimensional 3-point stencil), is defined by a recurrence dependent, and E/C—for elementary/composite; MA stands for where the value at each point at a given time instance is multiply-add operation. is measured in 4-byte words. evaluated by taking the average of its neighboring ele- ments along with itself from the previous time step. Jacobi We compute the hardware parameters using microbench- 2D is the two-dimensional 5-point analog of Jacobi 1D. marks as well as regression. Then we apply the analytical model to compute total energy consumption of a program These stencils were implemented on a GPU. A GPU has an for a given set of software parameters without having to array of Streaming Multiprocessors (SMs) each one with really run the program. We report the error of our model a set of SIMD (Single-Instruction Multiple Data) Vector by comparing the modeled energy against the actual Processors. The number of such processors can vary from (measured) energy consumed by the GPU. Finally, we use 8 to 192 depending on the GPU. All SMs share an off-chip our energy model to select the optimal tile size for energy global memory and have a private memory and a register efficiency and report the number of non-optimal tile size file. The programming model consists of a grid of thread selections and hence the error in energy due to the selec- blocks. The threads of a thread block execute concur- tion of non-optimal tile sizes. rently on one SM and multiple thread blocks may execute concurrently on one SM (depending on available shared We model the time using a similar approach and using memory and register resources). As thread blocks termi- parameters that correspond to the ones of the energy nate, new thread blocks are launched [8]. The computa- model. For instance one energy parameter characterixing tions in a thread block are expected to be independent of the hardware is the energy consumed for one multiply-add those in other thread blocks. Hence, the execution order of operation; for the time model the corresponding param- the thread blocks in a grid can be arbitrary. Within a thread eter is the time needed for one such operation. In general, block, all the threads can access shared memory collabora- modeling time is more difficult than modeling the energy. tively and explicitly synchronize to have a uniform view of The reason is that, while for the energy the total amount shared memory data among threads. for running an algorithm is the sum of the energies spent performing individual operations, for the time this is not We used NVIDIA K20c GPU [9] to experimentally validate 243

Page 244

our models. Since NVIDIA has not disclosed energy or results are shown on Figure 3. The errors in all cases are power data for memory and compute operations for K20c, less than 1%, with average error 0.2%. we needed to obtain these parameters experimentally. We used two methods: benchmarking and regression. Micro- benchmarks are short pieces of code, each of which stress- es an operation of interest. We used the NVIDIA NVML library to measure instantaneous power at any time of the execution of a micro-benchmark. The total consumed energy is then the product of the average measured power and the execution time of the micro-benchmark. We designed micro-benchmarks for each parameter of interest to determine the value of that parameter experi- mentally. The micro-benchmarks were implemented in such a way that the execution time and total energy con- sumption is dominated by the operation of interest. The second method we used to determine the value of the hardware energy parameters is based on nonlinear Figure 2. Table of the energy parameter values from the micro- regression, as described next. We create a set of combina- benchmark approach and the regression approach. tions of software parameters and, for each such combina- tion, experimentally measure the energy consumed by the corresponding code. We then plug in the values of the known parameters into the formula for the energy derived in the previous section and obtain, for each experiment, an equation with the values of the elementary hardware parameters as unknowns. The left-hand side of the result- Figure 3. Energy optimization error. SW, J1D, and J2D stand for ing system of equations contains the energy values pre- Smith-Waterman, Jacobi 1D, and Jacobi 2D, respectively. dicted by the model and the right-hand side contains the Impact on National Missions corresponding measured values. We then use an optimi- zation method to find values of the unknown parameters Co-design has been recognized as one of the high-priority that minimize the error defined as a norm of the difference areas of research for Information Science and Technology between the left and right-hand sides of the equations. at LANL. This project focuses on a key step of the software/ hardware co-design process – mapping the software rep- The values of the parameters and the corresponding mod- resentation onto the hardware model in order to maximize eling errors are given in Figure 2 [6]. The errors are quite data reuse and locality and increase the amount of paral- low, less than 18% for the micro-benchmark method and lelism. The optimization framework and the energy and less than 9% for the regression method. time models developed in this project make it possible to find optimal sets of software and hardware parameters Using the models for energy optimization: Having the with respect to cost function of the energy consumption analytical models, we can use them for optimization, for and execution time. The project helped establish collabora- instance, for finding a combination of software parameters tions with Colorado State University (Prof. Sanjay Rajopad- that minimizes ]energy consumption. We did such an op- hye and his group) and the University of Rennes at France timization in order to find a combination of tile sizes that (Prof. Rumen Andonov) and led to the hiring of Guillaume minimize the energy consumption. We solve an optimiza- Chapuis (University of Rennes) as a postdoc at LANL. tion problem, where the objective function is that analyti- cal function we found during the energy modeling and the References free variables are the tile sizes. This is a nonlinear optimiza-

1. Bondhugula, U., A. Hartono, J. Ramanujam, and P. tion problem, but of relatively small size, and we were able Sadayappan. PLuTo: A Practical and Fully Automatic to solve it using the nonlinear solver Knitro [2] in a matter Polyhedral Program Optimization System. 2008. In of seconds. We validated the results of the optimization by ACM Conference on Programming Language Design comparing the energy consumption by codes using opti- and Implementation. (Tuscon, AZ, 2008). , p. 101. mal sets of parameters found through measurements with Tuscon, AZ: ACM. ones found through solving the optimization problem. The 244

Page 245

2. Byrd, R. H., J. Nocedal, and R. A. Waltz. Knitro : An Efficient Multi-GPU Computation of All-Pairs Shortest Integrated Package for Nonlinear Optimization. 2006. Paths. 2014. 2014 IEEE 28TH INTERNATIONAL PARAL- Energy. 83: 1. LEL AND DISTRIBUTED PROCESSING SYMPOSIUM. Djidjev, H., S. Thulasidasan, G. Chapuis, R. Andonov, and D.
3. Grosser, T., A. Cohen, Holewinski, Justin and Lavenier. Efficient multi-GPU computation of all-pairs Sadayappan, , and S. Verdoolaege. Hybrid Hexagonal/ shortest paths. 2014. In IEEE International Parallel & Classical Tiling for GPUs. 2014. In CGO. , p. 66. Distributed Processing Symposium (IPDPS) . (Phoenix, May 19-23, 2014). , p. 360. Phoenix: IEEE.
4. Leng, J., T. Hetherington, A. ElTantawy, S. Gilani, N. S. Kim, T. M. Aamodt, and V. J. Reddi. GPUWattch: Djidjev, H., S. Thulasidasan, G. Chapuis, R. Antonov, and D. Enabling Energy Optimizations in GPGPUs. 2013. In Lavenier. Accelerating graph algorithms using graph- Proceedings of the 40th Annual International Sympo- ics processors: shortest paths for planar graphs. 2014. sium on Computer Architecture. , p. 487. ACM. ADTSC Science Highlights.
5. Nagasaka, H., N. Maruyama, A. Nukada, T. Endo, and S. Djidjev, H., and M. Onus. Using graph partitioning for ef- Matsuoka. Statistical power modeling of GPU kernels ficient network modularity optimization. 2013. GRAPH using performance counters. 2010. In Green Comput- PARTITIONING AND GRAPH CLUSTERING. 588: 103. ing Conference, 2010 International. , p. 115. Djidjev, Hristo N.. Parallel seed-based approach to protein structure similarity detection. 2013. Los Alamos Na-
6. Ranasinghe, W., S. Rajopadhye, H. Djidjev, R. Andonov, tional Laboratory, LA-UR-13-27448. V. Krishna Tandrapati, and N. Prajapati. Energy Mod- eling and Optimization for Tiled Nested-Loop Codes. Djidjev, Hristo N.. Efficient Shortest Path Computations on
7. In 11th Workshop on High-Performance, Power- Multi-GPU Platforms. 2013. Los Alamos National Labo- Aware Computing. IEEE Computer Society. ratory, LA-UR-13-26585.
8. Rodero, I., and M. Parashar. Energy Efficiency in HPC Djidjev, Hristo N., Guillaume Chapuis, Sunil Thulasidasan, Systems . 2012. In Energy-Efficient Distributed Com- and Rumen Andonov. On Solving the Shortest Path puting Systems . By Zomaya, A. Y., and Y. C. Lee. , p. Problem for Planar Graphs using Graphics Proces- sors. 2012. Los Alamos National Laboratory, LA- 81–108. Hoboken, NJ, USA: John Wiley & Sons. UR-12-25700.
9. CUDA C Programming Guide v6.0. 2014. NVIDIA. Djidjev, Hristo N., Rumen Andonov, Sanjay Rajopadhye, and Vamsi Tandrapati. Co-Design Optimization of Jacobi
10. NVIDIA Next Generation CUDA Compute Architecture: Kernel for GPU Architectures. 2013. Los Alamos Na- Kepler GK110. 2012. tional Laboratory, LA-UR-13-28110. Publications Djidjev, Hristo N., Sunil Thulasidasan, Rumen Andonov, Aleksandrov, L., H. Djidjev, A. Maheshwari, and J. R. Sack. Guillaume Chapuis, and Dominique Lavenier. Ef- An Approximation Algorithm for Computing Shortest ficient Multi-GPU Computation of All-Pairs Shortest Paths in Weighted 3-d Domains. 2013. DISCRETE & Paths. 2013. Los Alamos National Laboratory, LA- COMPUTATIONAL GEOMETRY. 50 (1): 124. UR-13-28111. Chapuis, , Le Boudic-Jamin, Andonov, Djidjev, and Lavenier. Djidjev, Hristo, Guillaume Chapuis, Rumen Andonov, Sunil Parallel Seed-Based Approach to Protein Structure Thulasidasan, and Dominique Lavenier. All-Pairs Short- Similarity Detection. 2014. PARALLEL PROCESSING AND est Path algorithms for planar graph for GPU-acceler- APPLIED MATHEMATICS (PPAM 2013), PT II. 8385: 278. ated clusters. 2015. Journal of Parallel and Distributed Computing. : -. Chapuis, , Le Boudic-Jamin, Andonov, Djidjev, and Lave- nier. Parallel Seed-Based Approach to Multiple Protein Ranasinghe, W., S. Rajopadhye, H. Djidjev, R. Andonov, V. Structure Similarities Detection. 2015. SCIENTIFIC PRO- Krishna Tandrapati, and N. Prajapati. Energy Modeling GRAMMING. and Optimization for Tiled Nested-Loop Codes. 2015. In 11th Workshop on High-Performance, Power-Aware Chapuis, G., and H. Djidjev. Shortest-Path Queries in Planar Computing. IEEE Computer Society. Graphs on GPU-Accelerated Architectures. 2015. In Large-Scale Scientific Computing (LSSC). Vol. 9374, p. Wohlers, , Le Boudic-Jamin, Djidjev, G. W. Klau, and An-
11. Springer-Verlag. donov. Exact Protein Structure Classification Using the Maximum Contact Map Overlap Metric. 2014. ALGO- Djidjev, , Thulasidasan, Chapuis, Andonov, and Lavenier. RITHMS FOR COMPUTATIONAL BIOLOGY. 8542: 262. 245

Page 246

Wohlers, I., M. Le Boudic-Jamin, H. Djidjev, G. Klau, and R. Antonov. Exact Protein Structure Classification Using the Maximum Contact Map Overlap Metric, , 2014. Presented at International Conference on Algorithms for Computational Biology (AlCoB). (Tarragona, Spain, July 1-3, 2014). 246

Page 247

Information Science & Technology Exploratory Research Final Report Contextual Learning and Recognition Alexei N. Skurikhin 20130265ER Abstract Background and Research Objectives The potential of big data holds a unique opportunity for Many real-world machine learning problems require global transparency, geographic profiling of activities of recognizing objects whose interpretation depends on interest and many other problem domains. At the same the context that is represented in the form of interde- time, in spite of growing quantities of data, object rec- pendencies among potential data interpretations. The ognition and data interpretation often remains ill-posed. problem is known as context aware structured machine This problem of object disambiguation is extremely learning. This stands in contrast to the conventional common in a broad spectrum of applications, including approaches of machine learning, which map input ob- computer vision, image analysis, social network analysis, served instances to output variables, such as classifica- bioinformatics and text analytics. The object interpreta- tion labels, independently of each other, and thus often tion ambiguity can be addressed by exploiting the fact cannot address the problem of object disambiguation that objects rarely occur in isolation; they tend to co- (Figure 1). occur and co-vary, and this correlation structure, known as context, provides important information for the disambiguation of the object. However, context-aware machine learning typically leads to a significant increase in the computational complexity of the data interpreta- tion, known as inference process. Modeling contextual information and scaling context aware machine learn- ing methods to large problems are among the most challenging issues in developing the next generation of information extraction approaches. We developed a computationally tractable machine learning approach based on conditional random fields (CRFs) for context aware object recognition in imagery data. Learning and inference in the developed approach Figure 1. Illustration of context aware machine learning. is computationally efficient. This efficiency is achieved (a) Conventional machine learning approaches classify input observed entities described with feature vectors x(i) indepen- due to the approximation of the original intractable CRF dently of each other. (b) Our project focused on context aware probabilistic graph model with an ensemble of com- structured machine learning approach based on conditional putationally tractable spanning tree-structured graph random fields (CRFs). CRF-based approach classifies input ob- models. The developed approximation based technique served entities jointly (collectively) by exploiting dependencies demonstrated competitive performance on publicly y(i)-y(j) in the output domain, i.e. labeling of objects impacts available datasets used in computer vision and machine each other interpretation. Context aware recognition of the learning areas. The approach can be used as a founda- central object y(2) is primed by how the adjacent objects y(1) tion for learning more accurate approximations based on and y(3) are recognized. y(2) is most likely to be “B” when its graphs that are more structurally complex than spanning neighbors are recognized as “letters”. y(2) is most likely to be trees. “13” when its neighbors are “numbers.” 247

Page 248

CRFs offer a mathematically consistent approach for range interactions between the elements across multiple reasoning under uncertainty and modeling contextual spatial scales, using descriptors based on edge orientation dependencies. CRFs are probabilistic graphical models that histograms, phase congruency, differential invariants, and model conditional probabilistic distributions and support scale invariant interest points [5, 6]. These descriptors are probabilistic inference to answer queries about the output used to characterize pixel patches and their relationships variables y(i) (Figure 1). A graph consists of a set of nodes (e.g., similarities), while patches themselves constitute corresponding to random variables y(i) and a set of undi- complex objects which are modeled as flexible constella- rected edges encoding independence relations among the tions of parts augmented with interpart contextual rela- random variables (Figure 1). Given the evidence, i.e. the tions. This representation provides a context modeling observed values of x(i) in Figure 1, probabilistic inference framework with statistical summary of the whole image, aims to find labeling y_i corresponding to the observed spatial layout properties within local image regions sur- evidence. In the CRF framework, classification labeling of rounding pixel patches. This also reduces the graph size, objects acts as context for interlinked objects by constrain- as nodes of the graph correspond to the partitions of the ing potential interpretation (labeling) of each other. The image, which are larger than individual pixels. If, prior to challenge arises due to the computational intractability graph modeling, image is partitioned into pixel patches, of probabilistic inference (of classification labeling) over this drastically reduces the number nodes in the graph. general graphs; this is known to be NP-hard. As a result, computational complexity of the inference in the graph that that is based on pixel patches (instead of There are several classes of CRF graphs, and probabilistic individual pixels) is reduced. graphical models in general, where inference is tractable. They include tree structured graphs [1], graphs with low Initially we investigated approximate inference based on treewidth [2], and binary pairwise Markov Random Fields pseudo-likelihood (PL) approach done over original graph (MRFs) with submodular energy constraints [3]. For a model. Instead of calculating the probability of individual model satisfying the sub-modularity constraints the MAP variables at sites in MRF/CRF given all other variables at labeling can be found in polynomial time using graph cuts different sites, the PL approach approximates the joint approach [3]. Graph cuts, however, do not scale up well probability of the variables at nodes in MRF/CRF as a to large problems and do not provide the probabilities of product of individual conditional probabilities over each individual variables assignment associated with a solution. variable, where the conditioning is on the neighbors of the variable. This property makes it a very efficient pa- Due to the intractability of the exact inference in most rameter learning method with some loss of accuracy, for it graphical models, there has been a growing research effort avoids exactly calculating the likelihood partition function. to develop tractable approaches to approximate inference. The PL-based approximation has been shown to produce Most known and often used approaches to approximate consistent estimates in the large lattice limit and demon- inference for general graphs are loopy belief propagation strated reduced performance level otherwise [6, 7]. Finally, (LBP) [1] and generalized belief propagation (GBP) [4]. when the data and/or parameter dimension is extremely However, while belief propagation inference on trees is large, particularly whenever the likelihood involves sum- exact and computationally efficient, for general graphs LBP ming probabilities over many unobserved states such as is not guaranteed to converge and when it converges the genealogies in biology, and, for example, in applications in result is not guaranteed to be a unique fixed point (a global astronomy and cosmology, and corresponding models do optimum). For GBP there is neither guarantee of conver- not have an analytical likelihood function, we developed gence nor guarantees on the accuracy of the estimated and investigated an approach to choose summary statistics probabilities when inference does converge. for approximate Bayesian computation to enable inference without explicit likelihood [8, 9]. While, particularly, the PL Our objective was to develop methods for developing com- approach can be considered as an approximation of the putationally efficient algorithms for context aware learning original graph using ‘simple’ composite blocks covering and inference in CRF-based probabilistic graphical models. nearest neighbors of each graph node, the quality of such This broadly entailed development of (1) the approxima- an approximation often needs to be more accurate for tion of an arbitrary graph by computationally tractable real-world applications. subgraphs, and (2) learning with approximate subgraphs. We posed the problem of intractable inference over Scientific Approach and Accomplishments general CRF graph as computationally efficient infer- To characterize contextual cues used in CRF models, we ence over tractable approximations of the original graph developed algorithms to estimate similarities of image ele- model. A key challenge to this approximation is a choice ments (such as image pixel patches) by considering longer 248

Page 249

of subgraphs in the original graph such that the quality of approximation based on generated spanning trees. First, approximation of the original model can be adjusted, while we developed ensemble based spanning tree CRF ap- inference over subgraphs should be tractable. proximations [11]. Spanning trees are randomly generated and are then trained independently on the whole training To build approximations to the original graph model, we dataset. Classification is done by majority voting over the have chosen a set of spanning trees. Spanning trees are ensemble of these trained trees. Second, we developed graph ultra-sparsifiers of treewidth=1. Inference in tree- a deep architecture structured (hierarchical) approxima- structured graphs can be implemented efficiently using tion consisting of multiple layers [12]. Each layer of this belief propagation that is an iterative message passing al- architecture contains one randomly generated spanning gorithm for computing approximate marginal distributions tree whose structure varies across the layers. Learning of of each variable in a graphical model [1]. The algorithm is this approximate model begins with one layer architecture based on the idea of message passing along the edges of containing one generated spanning tree that serves as the the graph, where the message indicates how much each initial approximation of the original graphical model. Add- node should update its belief based on the neighboring ing new spanning tree layers continues until the classifica- node’s current information. In tree-structured graphs the tion error reduction becomes acceptably small. Classifica- message passing schedule can be considered as a two-pass tion is produced by the spanning tree on the top layer. In schedule, with forward message propagation followed by both approaches, generated trees are computationally backward propagation. The algorithm begins by defining tractable acyclic sub-graphs which allow efficient and exact one of the variables (nodes) as the root. In the forward inference. Our approaches, results of their evaluations and sweep the algorithm sends messages from leaves towards opportunities for further development of our approaches the root. In the backward sweep the messages are propa- are summarized in [11, 12, 13, 14, 15]. Results of evalu- gated in the opposite direction, from the root through tree ations of our CRF-based ensemble approach is shown in edges to all leaves. After the completion of the backward Figure 2. sweep, the full set of messages over every edge in both directions is estimated. This allows one to compute all the node and edge (joint) marginals. In general loopy graphs, neither convergence nor correctness of this procedure is guaranteed as the message of one node can pass back to itself through the loops. In tree-structured graphs, both convergence and exactness of the inference process is guaranteed. The important question is how to choose trees to ap- proximate the original graphical model. The number of spanning trees of the graph is exponential in the size of the graph. The exact number of spanning trees can be estimat- Figure 2. Results of evaluations of CRF-based ensemble ap- ed using the matrix-tree theorem by evaluating the deter- proach. minant of the combinatorial Laplacian of the graph. We took an approach that assumes a lack of prior information Impact on National Missions on how subgraph should be structured. We also assume This project contributes to the development of mathemat- that objects which have to be recognized can be present ics and computer science required to extract knowledge in any location inside the image. Therefore, we randomly from massive quantities of data. Current approaches to sample spanning trees corresponding to the original graph object recognition are overwhelmed by the sheer volume model. Each of sampled trees captures some aspect of the of data (e.g., the increasing Earth observation image data original model. The use of randomly sampled trees allows volume is going beyond zettabytes, 10^21 bytes), and the us to compensate to some extent for weakness of each of diversity of features associated with objects of interest. the individual spanning trees. Spanning trees are generat- Additionally, in a variety of applied domains, e.g., broad ed uniformly at random from among all spanning trees of area search, increasing efforts at concealment and decep- the original graph using the algorithm, introduced in [10], tion make finding and correctly characterizing activities that simulates a loop-erased random walk. The complex- of interest very challenging. The developed CRF based ity of this generation is the mean hitting time of a graph, approach can facilitate object interpretation disambigua- which can be much smaller than the cover time of a graph. tion and provide an ability to focus human analysts on “hot spots” in the data. The approach can also be expanded to We developed two approaches to the CRF general graph 249

Page 250

the analysis and interpretation of multi-source data, e.g., DC, USA, 14-16 Oct. 2014). p. 8 pp.. Piscataway, NJ, data from optical, hyperspectral, SAR and LiDAR sensors. USA: IEEE. References 12. Skurikhin, A. N.. Hierarchical spanning tree-structured approximation for conditional random fields: an em-

1. Pearl, J.. Probabilistic reasoning in intelligent systems: pirical study. 2014. (Las Vegas, NV, 8-10 Dec. 2014). networks of plausible inference. 1998. Vol. 8888p. 85. BERLIN, GERMANY: SPRINGER-VERLAG.
2. Bach, F. R., and M. I. Jordan. Thin junction trees. 2002.
3. Skurikhin, A. N.. Hidden conditional random fields for ADVANCES IN NEURAL INFORMATION PROCESSING land-use classification. 2015. (Milan, Italy, 26-31 July, SYSTEMS 14, VOLS 1 AND 2. 14: 569. 2015). p. 4 pp.. Piscataway, NJ, USA: IEEE.
4. Kolmogorov, V., and R. Zabih. What energy functions
5. Burr, T., and A. N. Skurikhin. Conditional random fields can be minimized via graph cuts?. 2004. IEEE TRANS- for pattern recognition applied to structured data. ACTIONS ON PATTERN ANALYSIS AND MACHINE INTEL-
6. Algorithms. 8: 466. LIGENCE. 26 (2): 147.
7. Burr, T., and A. N. Skurikhin. Conditional random fields
8. Yedidia, J. S., W. T. Freeman, and Y. Weiss. Generalized for modeling structured data. 2015. In Encyclopedia of belief propagation. 2000. In Advances in Neural Infor- information science and technology. Edited by Khos- mation Processing Systems. (Denver, CO, USA, 28–30 row-Pour, M.. Third Editionp. 6167. Hershey, PA, USA: November 2000). , p. 7 pp.. Denver, CO, USA: NIPS. IGI Global.
9. Skurikhin, A. N., S. R. Garrity, N. G. McDowell, and D. M. Cai. Automated tree crown detection and size Publications estimation using multi-scale analysis of high resolution Burr, T., and A. N. Skurikhin. Approximate inference options satellite imagery. 2013. Remote Sensing Letters. 4 (5): in data models for image analysis. 2013. presented at the Statistical Image Analysis Workshop, Santa Fe, NM,

LAUR 13-21603. 6. Burr, T., and A. N. Skurikhin. Approximate inference op- Burr, T., and A. N. Skurikhin. Conditional random fields for tions in data models for image analysis. 2013. present- modeling structured data. 2015. In Encyclopedia of in- ed at the Statistical Image Analysis Workshop, Santa formation science and technology. Edited by Khosrow- Fe, NM, LAUR 13-21603. Pour, M.. , Third Edition, p. 6167. Hershey, PA, USA: IGI Global. 7. Burr, T., and A. Skurikhin. Pseudo-likelihood inference for Gaussian Markov random fields. 2013. Statistics Burr, T., and A. N. Skurikhin. Algorithms for approximate Research Letters. 2 (3): 63. Bayesian computation. 2015. In Encyclopedia of infor- mation science and technology. Edited by Khosrow- 8. Burr, T., and A. Skurikhin. Selecting summary statistics Pour, M.. , Third Edition, p. 1559. Hershey, PA, USA: IGI in approximate Bayesian computation for calibrating Global . stochastic models. 2013. BioMed Research Interna- Burr, T., and A. N. Skurikhin. Conditional random fields for tional. 2013: 10 pages. pattern recognition applied to structured data. 2015. 9. Burr, T., and A. N. Skurikhin. Algorithms for approxi- Algorithms. 8: 466. mate Bayesian computation. 2015. In Encyclopedia of Burr, T., and A. Skurikhin. Pseudo-likelihood inference for information science and technology. Edited by Khos- Gaussian Markov random fields. 2013. Statistics Re- row-Pour, M.. Third Editionp. 1559. Hershey, PA, USA: search Letters. 2 (3): 63. IGI Global . Burr, T., and A. Skurikhin. Selecting summary statistics in 10. Wilson, D. B.. Generating random spanning trees more approximate Bayesian computation for calibrating sto- quickly than the cover time. 1996. In Annual ACM Sym- chastic models. 2013. BioMed Research International. posium on Theory of Computing (STOC). (Philadelphia, 2013: 10 pages. Pennsylvania, USA, 22-24 May, 1996). , p. 296. New Skurikhin, A. N.. Learning tree-structured approximations York, NY, USA: ACM. for conditional random fields. 2014. In 2014 IEEE Ap- plied Imagery Pattern Recognition Workshop (AIPR). 11. Skurikhin, A. N.. Learning tree-structured approxima- (Washington, DC, USA, 14-16 Oct. 2014). , p. 8 pp.. Pis- tions for conditional random fields. 2014. (Washington, cataway, NJ, USA: IEEE. 250

Page 251

Skurikhin, A. N.. Hierarchical spanning tree-structured ap- proximation for conditional random fields: an empirical study. 2014. In Advances in Visual Computing (ISVC 2014), Lecture Notes in Computer Science. (Las Vegas, NV, 8-10 Dec. 2014). Vol. 8888, p. 85. BERLIN, GERMA- NY: SPRINGER-VERLAG. Skurikhin, A. N.. Hidden conditional random fields for land- use classification. 2015. In the International Geosci- ence and Remote Sensing Symposium (IGARSS 2015). (Milan, Italy, 26-31 July, 2015). , p. 4 pp.. Piscataway, NJ, USA: IEEE. Skurikhin, A. N., S. R. Garrity, N. G. McDowell, and D. M. Cai. Automated tree crown detection and size estima- tion using multi-scale analysis of high resolution satel- lite imagery. 2013. Remote Sensing Letters. 4 (5): 465. 251

Page 252

Information Science & Technology Exploratory Research Final Report A New Approach to Multiscale Plasma Physics Simulations Gian L. Delzanno 20130334ER Abstract tosphere or laboratory experiments aimed at achieving Collision-less magnetized plasmas offer a formidable sustained nuclear fusion energy. In many applications challenge to the numerical modeler because of the high- the collision frequency is much smaller than the typical dimensionality of the model (the plasma distribution scales of interest, implying that a collision-less kinetic function lives in a six dimensional space), the high level description of the plasma is necessary (as opposed to of non-linearity and anisotropy that governs the dynam- a fluid/macroscopic description characterized by mac- ics of the plasma, and the wide variety of spatial and roscopic fluid moments like density, momentum and temporal scales that have to be bridged between the energy). In addition, the characteristic spatial scales of characteristic plasma scales and the system size. the plasma (such as the Debye length, namely the char- acteristic length on which electric fields are shielded by The most widely used numerical approach to collision- the plasma, or the plasma gyro-radii) can be several or- less magnetized plasmas is the Particle-In-Cell (PIC) ders of magnitude smaller than the characteristic system technique, where the plasma distribution function is size. Similar considerations apply to the temporal scales, discretized by macro-particles that move in a compu- implying that plasmas are very multi-scale. Indeed, the tational mesh under the action of the self-consistent highly non-linear, strongly anisotropic behavior of mag- electromagnetic field. PIC is relatively simple, extremely netized plasmas and their multi-scale nature, combined robust and fairly easily parallelized, but suffers from the with the high dimensionality of the kinetic description inherent statistical noise associated with the macro-par- (the plasma distribution function, namely the plasma ticles. This implies that PIC becomes extremely resource- density in phase space, lives in a six dimensional space intensive for problems that require a high signal-to-noise consisting of three spatial coordinates and three velocity ratio. coordinates), make the numerical modeling of collision- less magnetized plasmas a formidable challenge. In this project we have developed a new spectral capa- bility for the numerical modeling of collision-less mag- The most widely used approach to the numerical solu- netized plasmas. We have combined spectral discretiza- tion of the collision-less magnetized plasma model, i.e. tions in velocity space (based on Hermite functions and the Vlasov-Maxwell equations, is the Particle-In-Cell Legendre polynomials) and in physical space (either (PIC) method [2]. In PIC, the plasma distribution func- global Fourier basis or local discontinuous-Garlekin tion is discretized with macro-particles that move in a discretization). We have successfully formulated, imple- computational mesh due to the electromagnetic field mented, and tested these methods, and studied their generated self-consistently. PIC owes its popularity to its numerical stability and conservation properties. A com- simplicity, robustness and the relative ease for parallel- parison against PIC shows that the spectral method can ization that allows one to tap into the impressive recent be several orders of magnitude faster or more accurate advances in computing power. The major shortcoming than PIC, at least for the standard benchmark problems of PIC, however, is due to the statistical noise associated considered. with the macro-particles, which has a rather unfavor- able scaling proportional to the inverse of the square Background and Research Objectives root of the number of particles per cell [2]. This implies Magnetized plasmas are at the cornerstone of many that problems requiring a large signal-to-noise ratio also interesting phenomena in astro, space and laboratory require a lot of computational resources. As a figure of physics [1]. Important examples are the Earth’s magne- merit of the resource-intensive nature of PIC, a single 252

Page 253

run using the global gyrokinetic PIC code developed within scheme. The discretized set of non-linear equations is the SciDAC Gyrokinetic Particle Simulation Center for in- solved with a Jacobian-Free Newton-Krylov (JFNK) tech- vestigating the long-time evolution of turbulent transport nique using the GMRES solver for the inner linear itera- in nuclear fusion devices was estimated, in 2008, to take tions and with the possibility of applying a preconditioner approximately 25 million CPU hours, on 100,000+ cores to speed-up the convergence cycle. These techniques have [3]. Interestingly, the monetary cost of such simulation can been implemented in a Fortran code, which has been suc- be estimated at about $1,000,000 [4]. cessfully tested against a number of benchmark problems. The code is interfaced to the PETSc library in order to take A second class of numerical methods for the Vlasov- advantage of PETSc routines for efficient parallelization Maxwell equations is called Eulerian-Vlasov and discretizes both on shared- and distributed-memory architectures. phase space by introducing a six dimensional computa- The code is open source (LA-CC-14-021). Below we high- tional mesh [5]. As such, Eulerian-Vlasov methods do not light the main accomplishments of our study. suffer from statistical noise from particles (there are no particles) but require significant amount of memory. For Successful comparison against PIC techniques. We have instance, storing a field in double precision on a mesh with performed several standard tests (Langmuir waves, Landau one trillion cells requires ~8 Tb of memory. Memory is damping, two-stream instability, ion acoustic waves) where probably the main reason for the somewhat limited diffu- we have compared our Fourier-Hermite (FH) capability sion of Eulerian-Vlasov methods (at least relative to PIC), against PIC in terms of accuracy, computational time of the although other issues such as lack of discrete conservation simulation and efficacy. Efficacy combines both numerical laws or lack of positivity of the plasma distribution func- error and computational time into a single measure of the tion are also important. cost-effectiveness of an algorithm. Naturally algorithms with higher efficacy are better. These tests have been con- This project deals with the development of spectral ducted on the Vlasov-Poisson equations, i.e. the electro- methods for the numerical solution of the Vlasov-Maxwell static limit of the Vlasov-Maxwell equations, in one dimen- equations that are alternative to PIC and Eulerian-Vlasov sion. The comparisons have been performed both for an methods. In particular, the velocity part of the plasma explicit time discretization of the FH algorithm (mainly distribution function is expanded into basis functions, because the overwhelming majority of PIC techniques are resulting in a system of moment equations for the coef- explicit) and for the implicit Crank-Nicholson time discreti- ficients (moments) of the expansion. A spectral method is zation. In all cases FH has outperformed PIC by producing clearly immune to particle noise, but the most important results that are either orders of magnitude more accurate aspect of this approach is that with a suitable basis the in the same computational time or orders of magnitude low-order moments of the expansion correspond to the faster for the same accuracy. Consistently, the efficacy is typical moments of a fluid. This allows a smooth transition orders of magnitude higher for FH than for PIC, as exempli- from a fluid to a kinetic description of the plasma by simply fied in Figure 1 where results of our efficacy study for the increasing the number of moments retained. This feature two-stream instability problem are shown. In addition, we of the spectral method is critical in approaching multi-scale have performed the comparison for the echo problem. In problems and assessing the importance of kinetic effects, this case a wave is excited at the beginning of the simula- and is not available in PIC or Eulerian-Vlasov methods that tion and decays. A second wave is excited at a later time are instead forced to treat the full distribution function. and also decays. The non-linear interaction between the two waves creates a third wave (the echo) at a later time. Scientific Approach and Accomplishments This test requires high resolution and has been inaccessible Capability development. We have developed a new capa- to PIC because of the statistical noise mentioned above, bility for the solution of the Vlasov-Maxwell equations for despite using an astonishing 2,000,000 particles per cell a collision-less magnetized plasma. The velocity part of the in our implementation (PIC simulations normally use a plasma distribution function is expanded in Hermite basis few hundred particles per cell as a rule of thumb). FH, on functions [we have tried both the symmetrically- (SW) and the other hand, could easily handle the echo problem and asymmetrically-weighted (AW) basis] or Legendre polyno- reproduce the known theoretical results. This is shown in mials. The resulting set of partial differential equations is Figure 2, where the first three Fourier modes of the elec- discretized in the spatial coordinate by a spectral discreti- tric field are shown (the echo is in E1, while E2 and E3 cor- zation using a (global) Fourier decomposition or a (local) respond to the excited waves). These results are in Refs. discontinuous-Galerkin (DG) discretization. The final set of [6,7]. An example of another test problem, the interaction ordinary differential equations is discretized in time with between a plasma and a beam, is shown in Figure 3, where a second-order accurate, fully implicit Crank-Nicholson the nonlinear modifications of the beam are captured very 253

Page 254

well by the spectral method. (E3) and decays. The two waves interact non-linearly to produce a third wave (the echo, E1) at time t=3. The FH methods clearly shows the echo and reproduces its known theoretical aspects. PIC, on the other hand, cannot reproduce the echo because of the statistical noise due to the particles (despite 2,000,000 par- ticles per cell were used in this simulation, while typical PIC runs normally use few hundreds particles per cell). Figure 1. A comparison between the implicit Fourier-Hermite (FH) spectral method and the implicit Particle-In-Cell (PIC) tech- nique on a two-stream instability problem, showing the com- putational time to complete the simulation versus the efficacy (defined as the inverse of the numerical error times the com- putational time). On this example FH has orders of magnitude Figure 3. The interaction of a weak beam with a background higher efficacy than PIC, namely it is a much more cost-effective plasma studied with the FH method. The picture shows part algorithm. of the phase space during the saturation phase. The spectral method captures accurately the nonlinear distortion of the beam after the development of a beam-plasma instability. Formulation of spectral algorithms with stability properties and exact conservation laws in discrete form. A lot of effort has been devoted to devise numerical algorithms with certain important properties. For the Hermite expansion, it is possible to prove the numerical stability of the SW algorithm in a finite time step. This is not guaranteed for AW Hermite and one has to resort to an artificial collisional operator to stabilize the method. On the other hand, it is possible to prove that AW Hermite features exact conser- vation laws in discrete form for total mass, momentum and energy, while these are valid for SW Hermite only in certain limits. The Legendre expansion, on the other hand, can be numerically stable and can feature exact conservation laws in discrete form. For comparison, this is something that has never been achieved in PIC, where either total mass or momentum or total mass and energy can be conserved simultaneously. All these considerations have been proven theoretically and confirmed by numerical experiments. Most of the work has been performed on the Vlasov- Poisson equations in one dimension, with a Fourier or DG Figure 2. Comparison between the spectral Fourier-Hermite spatial discretization. Exact conservations laws in discrete (FH) method and the Particle-In-Cell (PIC) method on the echo form have been established for AW Hermite in multi- problem. The picture shows the first three Fourier modes of the dimensions and for the full Vlasov-Maxwell equations (i.e. electric field versus time. At time t=0 a wave is initialized (E2) in the electromagnetic case). These results can be found in the system and decays. At time t=1 a second wave is initialized 254

Page 255

Refs. [7,8,9,10]. has already been exploited in the FH method where we have changed the number of moments dynamically in the Optimization of the spectral basis. As we have argued simulation. The Fourier basis, however, limits the appli- above, the main advantage of spectral methods is their cability of this approach because one is forced to add or potential to bridge between a fluid/macroscopic and a eliminate a moment everywhere in the physical domain. kinetic/microscopic description of the plasma by retaining In many problems of interest, however, the plasma could more moments. This is true for both AW Hermite and Leg- behave as a fluid in part of the domain while a kinetic endre discretizations. We have exploited this feature for description is required elsewhere. A necessary step to the FH method, for which we have developed a technique address this type of problem was to develop a spectral where the number of moments is changed dynamically discretization based on a local expansion in physical space. during the simulation. This is achieved by a double-thresh- We have accomplished this task by using the DG discreti- old strategy: if the last moment of the simulation is below zation. We have successfully formulated, implemented a certain threshold, it is removed from the simulation; if and tested a DG spatial discretization combined with both it is above a (higher) threshold, we add a moment and Hermite and Legendre expansions in velocity space. The its equation to the code. We have successfully tested this stability of the resulting method and the conservation laws technique on Landau damping and two-stream instability have been studied theoretically and with numerical experi- problems showing that a sizeable gain of performance can ments. The results of our study are presented in Ref. [10]. be achieved in this way. These results are published in Ref. [11]. For the Hermite discretization, another form optimi- Impact on National Missions zation can be obtained acting on the basis itself. This is Magnetized plasmas lie at the heart of many problems in because the Hermite basis contains two free parameters global and energy security that are critical to the nation. (the velocity shift and scaling) that, if adjusted properly, Examples includes geomagnetic storms that develop near can significantly increase the rate of convergence of the Earth and can damage infrastructure on the ground and in series. We have formulated some physics-based criteria for space (the potential economic damage from a strong geo- dynamically changing the free parameters during the simu- magnetic storm has been estimated to $1-2T) or the quest lation. The preliminary results are encouraging and show a of (magnetic fusion) laboratory experiments to provide a significant improvement in the simulation. sustainable and virtually infinite source of energy to free the nation from its dependence on oil. We have developed Formulation, implementation and testing of a spectral a new capability for the numerical modeling of magnetized method based on a Legendre expansion in velocity space. plasmas that can be much more computationally efficient/ A spectral method based on a Hermite discretization is performing than traditional techniques. Although the very suitable when the plasma distribution function is project has focused on the formulation, implementation nearly Maxwellian (this is because the basis functions and testing of this new capability mostly on toy-problems, contain a Maxwellian weight function). In cases of strong it positions us uniquely towards future opportunities in non-Maxwellian behavior, the effectiveness of the Hermite the areas mentioned above. This obviously also includes expansion is tightly coupled to a robust optimization of the exascale initiative signed by President Obama. We are the velocity shift and scaling of the basis. An alternative constantly in contact with Dr. Pieter Swart (T-5, LANL point approach, which bypasses the optimization problem com- of contact for OASCR programs) and Dr. Luis Chacon (T-5, pletely, consists of using a Legendre-polynomial expansion LANL point of contact for exascale computing) to capitalize where free parameters are not necessary. We have been on potential opportunities. We are also establishing new able to formulate this method by using the Simultaneous collaborations with the international community (in partic- Approximation Term technique and deliver an algorithm ular, Prof. Jack Scudder of U. Iowa, Dr. Vadim Roytershteyn that is numerically stable and features exact conservation of the Space Science Institute, Prof. Stefano Markidis of the laws in discrete form. This had never been done before in KTH Royal Institute of Technology of Stockholm), where the context of the Vlasov-Poisson equations and its formu- collaborators are interested in applying our new capabil- lation, properties and successful testing can be found in ity to problems that could not be previously solved. This is Ref. [9]. also expected to lead to new funding opportunities. Formulation, implementation and testing of a spectral References method based on a local (DG) basis in physical space. A crucial aspect of the spectral expansion in velocity space 1. Goldston, R. J., and P. H. Rutherford. Introduction to is the ability to optimize the expansion basis, namely to Plasma Physics. 1995. retain only the moments that are necessary for an accurate 2. Birdsall, C. K., and A. B. Langdon. Plasma Physics via description (fluid or kinetic) of the dynamics. This feature 255

Page 256

Computer Simulation. 1991. pear in Computer Physics Communications. 3. Tang, W. M.. Scientific and computational challenges Camporeale, E., G. L. Delzanno, B. Bergen, and J. D. of the fusion simulation project (FSP). 2008. Journal of Moulton. On the velocity space discretization for the Vlasov-Poisson system: comparison between Hermite Physics: Conference Series. 125: 012047. spectral and Particle-in-Cell methods. Part 1: semi- 4. Walker, E.. The real cost of a CPU hour. 2009. Com- implicit scheme. 2013. arXiv:1311.2098. puter. 42 (4): 35. Delzanno, G. L.. Multi-dimensional, fully-implicit, spectral method for the Vlasov-Maxwell equations with exact 5. Cheng, C. Z., and G. Knorr. The integration of the Vla- conservation laws in discrete form. 2015. Journal of sov equation in configuration space. 1976. Journal of Computational Physics. 301: 338. Computational Physics. 22: 330. Delzanno, G. L., B. Bergen, J. D. Moulton, B. Srinivasan, and 6. Camporeale, E., G. L. Delzanno, B. Bergen, and J. D. E. Camporeale. Spectral method for the solution of the Moulton. On the velocity space discretization for the Vlasov equation based on Hermite polynomials. Pre- Vlasov-Poisson system: comparison between Hermite sented at American Physical Society Division of Plasma spectral and Particle-in-Cell methods. Part 1: semi- Physics Meeting. (New Orleans, 27-31 Oct 2014). implicit scheme. 2013. arXiv:1311.2098. Delzanno, G. L., E. Camporeale, B. Bergen, J. D. Moulton, 7. Camporeale, E., G. L. Delzanno, B. Bergen, and J. D. B. Srinivasan, and G. Manzini. Fully implicit spectral Moulton. On the velocity space discretization for the method for the solution of the Vlasov equation based on Hermite polynomials. Invited presentation at 2014 Vlasov-Poisson system: comparison between implicit US-Japan JIFT Workshop on Progress in kinetic plasma Hermite spectral and Particle-in-Cell methods. To ap- simulations. (New Orleans, 31 Oct-1 Nov 2014). pear in Computer Physics Communications. Delzanno, G. L., E. Camporeale, B. Bergen, and J. D. 8. Delzanno, G. L.. Multi-dimensional, fully-implicit, spec- Moulton. Fully-implicit Fourier-Hermite spectral tral method for the Vlasov-Maxwell equations with method for the Vlasov equation. Presented at Vlasovia exact conservation laws in discrete form. 2015. Journal 2013. (Nancy, France, 25-28 Nov 2013). of Computational Physics. 301: 338. Delzanno, G. L., J. Vencels, G. Manzini, S. Markidis, and 9. Manzini, G., G. L. Delzanno, J. Vencels, and S. Marki- E. Camporeale. Spectral method for the solution of dis. A Legendre-Fourier spectral method with exact the Vlasov-Maxwell equations. Invited presentation conservation laws for the Vlasov-Poisson system. LA- at Workshop on Magnetic Reconnection in Plasmas. UR-15-27359. (Stockholm, Sweden, 10-14 Aug. 2015). Manzini, G., G. L. Delzanno, J. Vencels, and S. Markidis. 10. Manzini, G., J. Vencels, G. L. Delzanno, and S. Markidis. A Legendre-Fourier spectral method with exact con- Spectral based-discontinuous Galerkin discretizations servation laws for the Vlasov-Poisson system. LA- of the Vlasov-Poisson system. LA-UR-15-27420. UR-15-27359. 11. Vencels, J., G. L. Delzanno, A. Johnson, I. Bo Peng, E. Manzini, G., G. L. Delzanno, and J. D. Moulton. Spectral/ Laure, and S. Markidis. Spectral Solver for Multi-scale Discontinuous Galerkin discretization of the Vlasov- Plasma Physics Simulations with Dynamically Adaptive Poisson system. Invited presentation at European Number of Moments. 2015. Procedia Computer Sci- Conference on Numerical Mathematics and Advanced ence. 51: 1148. Applications. (Ankara, Turkey, 14-18 Sept. 2015). Publications Manzini, G., J. Vencels, G. L. Delzanno, and S. Markidis. Spectral based-discontinuous Galerkin discretizations Camporeale, E., G. L. Delzanno, B. Bergen, and J. D. of the Vlasov-Poisson system. LA-UR-15-27420. Moulton. Fourier-Hermite spectral method for the Vlasov equation. Presented at 55th Annual Meeting of Vencels, J., G. L. Delzanno, A. Johnson, I. Bo Peng, E. Laure, the APS Division of Plasma Physics. (Denver, 11-15 Nov and S. Markidis. Spectral Solver for Multi-scale Plasma 2013). Physics Simulations with Dynamically Adaptive Num- ber of Moments. 2015. Procedia Computer Science. Camporeale, E., G. L. Delzanno, B. Bergen, and J. D. 51: 1148. Moulton. On the velocity space discretization for the Vlasov-Poisson system: comparison between implicit Vencels, J., G. L. Delzanno, A. Johnson, I. Bo Peng, E. Laure, Hermite spectral and Particle-in-Cell methods. To ap- and S. Markidis. Spectral Solver for Multi-scale Plasma 256

Page 257

Physics Simulations with Dynamically Adaptive Num- ber of Moments. Invited presentation at ICCS 2015 International Conference on Computational Science. (Reykjavík, Iceland, 1-3 June 2015). Vencels, J., G. L. Delzanno, G. Manzini, V. Roytershteyn, and S. Markidis. Weak turbulence simulations with the Hermite-Fourier spectral method. Presented at 57th Annual Meeting of the APS Division of Plasma Physics. (Savannah, Georgia, 16-20 Nov. 2015). Vencels, J., G. L. Delzanno, G. Manzini, and S. Markidis. Spectral methods for the solution of the Vlasov-Max- well equations. Invited presentation at ICNSP 2015 International Conference on Numerical Simulation of Plasmas. (Golden, Colorado, 12-14 Aug. 2015). 257

Page 258

Information Science & Technology Exploratory Research Final Report Sparse, Distributed, and Robust Network Control Marian Anghel 20130558ER Abstract Background and Research Objectives We developed new methods for the algorithmic con- The analysis and control of complex interconnected sys- struction of Lyapunov functions for the transient stability tems is a great engineering and operational challenge. analysis of dynamic networks including, in particular, Although emerging network control applications include power system networks. The proposed methodology large mechanical structures, distributed sensing, military uses recent advances in the theory of positive polyno- reconnaissance, and interacting mobile robot agents, the mials, semidefinite programming, and sum of squares focus of this project is the design of distributed control- decomposition, which have provided powerful nonlinear lers for large-scale power grid systems. tools for the analysis of systems with polynomial vec- Paradoxically, for a system which is arguably the most tor fields. In order to apply these techniques to power influential machine of the 20th century, the modern system networks described by trigonometric nonlineari- power systems have not yet fully embraced the technical ties we used an algebraic reformulation technique to opportunities of the 21st century (distributed sensors, recast the system’s dynamics into a set of polynomial computers, and advanced communication networks). At differential algebraic equations. We demonstrated the the same time, the power systems face the simultaneous application of these techniques to the transient stability convergence of several independent difficulties (inter- analysis of power systems by estimating the region of mittent renewable energy sources, frequent transmis- attraction of the stable operating point. sion congestions, relentless market pressure ) which the However, such methods scale very poorly with the size existing centralized control infrastructure will not be able of the system. For this reason we developed a decen- to solve. A paradigm shift based on modern distributed tralized approach to stability analysis of interconnected measurements, control, and management concepts is systems which does not require computing a Lyapunov necessary to modernize the grid. function for the full system. In this project we have In response to these challenges, this project seeks to developed a number of scalable and parallel iterative construct an algorithmic synthesis of nonlinear and dis- algorithms that decompose a system into a network tributed control techniques for large networked power of interacting subsystems and certify the asymptotic systems. More specifically, this project developed a stability of the interconnected systems by using only decentralized approach that exploits separability and the subsystems’ Lyapunov functions. When this stability decomposition of the corresponding control problem cannot be certified we design local and minimal control into nonlinear sub-problems that can be solved locally laws that guarantee asymptotic stability. We have ap- and efficiently. plied these algorithms to a number of dynamic networks including a network of interacting Van der Pol oscilla- Scientific Approach and Accomplishments tors and a network preserving power system model that includes dynamic loads in addition to the usual swing A traditional approach to transient stability analysis dynamics of generators. The analysis and control design involves the numerical integration of the nonlinear dif- methods developed in this project can be extended to ferential equations describing the system’s dynamics. more general, possibly more complex, networks includ- This method provides an accurate description of tran- ing, in the case of power systems, voltage dynamics and sient phenomena but has a number of serious disadvan- more complicated generator and load models. tages. First, its computational cost prevents time-domain simulations from providing real-time transient stability 258

Page 259

assessments and significantly constraints the number of cal energy functions (an alternative to Lyapunov functions cases that can be analyzed. Second, numerical simulations that is used in power system transient stability analysis) alone cannot quantitatively estimate the degree of system do not exist for these systems and the proposed SOS stability or produce actionable information to guide pre- analysis provides for the first time a constructive approach ventive control measures. for computing analytical Lyapunov functions for realistic power system models with transfer conductances (dissipa- For these reasons alternative approaches to transient sta- tive power lines). As shown in Figure 2 the proposed SOS bility analysis have been intensively explored. Among the method provides better estimates of the ROA when com- methods proposed, the Lyapunov function (LF) method can pared to various estimates based on the energy function be used to prove the stability of an equilibrium point and method that are either very conservative or provide only avoids the expensive time-domain integration of the sys- local estimates of the ROA. tem dynamics. The Lyapunov function is a generalization of the energy function for dissipative mechanical systems. As described in Figure 1, a Lyapunov function is positive in a bounded domain around the equilibrium point, except at the equilibrium point where it is zero, and decreases along every trajectory of the dynamical system. As a result the Lyapunov function decreases in time to zero, thus proving that the system’s trajectory converges to the equilibrium point (the only point where the Lyapunov function is zero). The largest level set of the Lyapunov function inside the bounded domain is an invariant of the system dynam- ics (every trajectory starting inside this level set remains inside the level set) and provides an approximation of the region of attraction of the equilibrium point. Unfortunate- ly, there is no general approach for constructing Lyapunov functions for a given dynamical system. Nevertheless, for polynomial dynamical systems, if we restrict the search for Lyapunov function in the space of sum-of-squares (SOS) Figure 1. An example of stability analysis using a Lyapunov function. A simple two-dimensional system has a global equilib- polynomials (polynomial that can be written as the sum of rium point at the origin. A quadratic Lyapunov function (LF) that squared polynomials and that are, therefore, positive) al- is positive everywhere in the phase space of the system is shown gorithmic approaches for constructing Lyapunov functions in green. The Lyapunov function is zero only at the equilibrium have been introduced more than a decade ago. Indeed, point. After a disturbance the trajectory of the system is shown testing that a SOS polynomial is positive reduces to solving in red in the two dimensional phase plane. The evolution of the a linear matrix inequality for which efficient semidefinite LF itself along the dynamics of the system is also shown in red programming methods are available. along the LF surface. The LF decreases along the trajectory and converges to zero. Since the LF is only zero at the equilibrium One of the first contributions of this project was to use SOS point, the decrease of the LF to zero certifies the stability of the techniques in order to develop an algorithm for the con- equilibrium point. Since this system is globally stable the whole struction of Lyapunov functions for classical power system phase space belong to the region of attraction. For this reason a models. We demonstrated the application of these tech- LF whose time derivative along the system’s dynamics is negative niques to the transient stability analysis of power systems everywhere in phase space can be easily found. Nonlinear sys- tems are generally not globally stable and the search for LFs has by estimating the region of attraction of the stable operat- to be performed in a local domain around the equilibrium point. ing point [1]. The application of SOS methods to power sys- The analysis in this case is very difficult but sum-of-squares tem models was innovative for two reasons. First, power methods can provide efficient algorithm to searching for LFs. system networks are not polynomial dynamical systems because their dynamics is described by trigonometric non- For large-scale dynamical systems the SOS method be- linearities. Nevertheless, using an algebraic reformulation comes inapplicable due to the fast growing computational technique we were able to recast the system’s dynamics complexity of the semidefinite programs that need to be into a set of polynomial differential algebraic equations. solved. For this reason, this project further developed a Second, we have also shown that systems with transfer decentralized approach to stability analysis of intercon- conductances can be analyzed as well without any concep- nected systems that does not require computing a Ly- tual difficulties. This is a significant result because analyti- apunov function for the full system. In particular we have 259

Page 260

developed a number of scalable and parallel iterative methodology for computing the comparison system per- algorithms that decompose a system into a network of in- forms better than the traditional approaches found in the teracting subsystems and certify the asymptotic stability of literature. In the same papers we show how the compari- the interconnected systems by using only the subsystems’ son equation approach can be used to design decentralized Lyapunov functions. control laws when the comparison approach cannot certify global system stability. This is an application of the com- parison equation approach that has not been discussed before in the control literature. The approach presented is parallel and scalable. The method can also be applied to complex real world systems, such as power systems, as we show in a paper that is currently in preparation. Moreover, we have generalized this approach and introduced a new algorithm that uses multiple comparison equations to fur- ther improve the performance of the comparison equation approach [7]. Figure 2. The region of attraction for the equilibrium point located at the origin, projected in the angle space of a simple power system model, is shown in thin black line connecting the unstable equilibria on the boundary of the region of attraction (ROA). The thick black lines show two different approximations to the ROA constructed using energy function methods. The dark gray area shows the best estimate of the ROA according to the proposed sum-of-squares (SOS) algorithm. Unlike the energy function approach, the SOS approach provides a globally optimal Figure 3. A network of nine Van der Pol oscillators along with estimate. the regions of attraction (ROA) of the isolated subsystem in red lines. We assume an extreme decomposition of the network in One algorithm uses the subsystems’ Lyapunov functions as which each oscillator (node) describes a subsystem. This restric- state variables and invokes a comparison principle to bind tion can be relaxed. Estimates of the ROA of the isolated subsys- their dynamics by the evolution of a new linear dynami- tem are in blue lines. The thick red lines defines the neighbors of cal system. If this linear system is stable and converges to each subsystem. the zero, then each component of the vector LF converges to zero. As a result, each subsystem converges to its zero Another suite of algorithms that we have developed has no state (equilibrium state) thus guaranteeing the global counterpart in the literature [2,3,4]. They use SOS methods convergence of the full, interconnected system. How- and input to state stability analysis to make local stability ever computing these comparison equations, for a given estimate for each subsystem under disturbances intro- interconnected system, still remains a challenge. We have duced by the interaction with its neighbors. This approach shown that SOS methods can be used to study the stabil- provides stability estimates that guarantee a suite of ity of an interconnected system by computing the vector contractions of the subsystem’s dynamics to a decreasing Lyapunov functions as well as the comparison equation sequence of local level sets defined by the LF of each sub- [5, 6]. In Figure 3 we show a network of nine Van der Pol system as presented in Figure 4. This approach is iterative oscillators along with the regions of attraction (ROA) of the and each subsystem updates its stability estimates based isolated subsystem in red lines. For this system we assume on new information received from its neighbors. If for each an extreme decomposition of the network in which each subsystem this sequence decreases to zero, the algorithm oscillator (node) describes a subsystem, but this restriction proves the global stability of the fully interconnected can be relaxed. In [5] and [6] we show that our proposed system. When the algorithms fail to provide stability guar- 260

Page 261

antees, they provide locally computable control laws that to guarantees the convergence of the subsystem trajectory to a restore the asymptotic stability of the full system under a lower level set. The control is local and it adapts to the subsystem given disturbance. The controls are distributive, adaptive, state and the disturbance information received from its neigh- bors. and minimal, because they are applied to a small number of subsystems and for a limited duration. This algorithm is applied in [3] to a network of Van der Pol oscillators and in [4] to a power system model of three generators and six load nodes as shown in Figure 5. For the power system model we perform an overlapping decom- position in which the dynamics of the reference node (gen- erator node 3) is shared with all the subsystems. The same iterative convergence analysis described above is applied to this system. The use of overlapping decompositions, in which some subsystems share some degrees of freedom, is very significant. Indeed, if a decomposition fails to have stable isolated subsystems, then the definition of the un- stable subsystems can be enlarged to include states from their neighbors until each subsystem become stable. Figure 5. Network of three generators and six load nodes. We perform an overlapping decomposition in which the dynamics of the reference node (generator node 3) is shared with all the sub- systems. This results in the following overlapping subsystems: S1 =\{1, 3\}, S2 = \{4, 3\}, S3 = \{5, 3\}, S4 = \{7, 3\}, S5 = \{2, 3\}, S6 = \{6, 3\}, S7 = \{8, 3\}, S8 = \{9, 3\}. The same iterative convergence analysis described in Figure 4 is applied to this system. The use of overlap- ping decompositions, in which some subsystems share some de- Figure 4. Distributed coordinated sequential stability certifica- grees of freedom, is very significant. Indeed, if a decomposition tion. The central subsystem interacts with three other subsys- fails to have stable isolated subsystems, then the definition of the tems. At the end of a disturbance the state of each subsystem unstable subsystems can be enlarged to include states from their moves away from the equilibrium point at zero. The neighbors neighbors until each subsystem become stable. communicate to the central subsystem their position in phase space by providing the value of their Lyapunov function at their All the algorithms developed are parallel and scalable and current state. Based on this information the central subsystems use limited communications between each subsystem and determines if its Lyapunov function will decrease in time under its neighbors. The analysis and control design methods the action of the local dynamics AND the influence received from developed in this project can be extended to more general, the neighbors. These disturbances are defined by the hashed possibly more complex, networks including, in the case of domains in the figure. This problem is solved locally by each subsystem using sum-of-squares (SOS) methods. If the answer power systems, voltage dynamics and more complicated is “yes” then all trajectories of the subsystem (red lines) starting generator and load models. from a ring (gray area) whose outer bound is the current level set will converge to a domain defined by a smaller level set. The Impact on National Missions central subsystem then communicates this new level set estimate This project is well aligned with LANL’s efforts in energy to its neighbors. The same analysis is performed in parallel by all security. This project addresses some of the challenges of subsystems. As the dynamics evolves each subsystem computes grid modernization efforts, and also addresses the com- a decreasing sequence of level sets that bound the the global putational needs for the next generation electric grid as dynamics. If all sequences converge to zero, then the full system expressed in recent workshops organized by the Depart- is asymptotically stable. If for some subsystems this stability can- not be guaranteed - the answer to the above stability question is ment of Energy (DOE) in partnership with Office of Electric- “no” - a controller is designed using again SOS methods in order ity Delivery and Energy Reliability (OE). 261

Page 262

More specifically, the report on “Advanced Control Meth- trol of large-scale nonlinear networks. To appear in 5th ods” conducted by the National Energy Technology Labo- IFAC Workshop on Distributed Estimation and Control ratory for the DOE in 2007 has identified the need for the in Networked Systems. (Philadelphia, 10-11 Sep. 2015). development of distributed intelligent agents that respond 6. Kundu, S., and M. Anghel. Computation of linear com- rapidly at the local level to unburden centralized control parison equations for stability analysis of intercon- systems. This project directly addresses this problem by nected systems. To appear in The 54th IEEE Conference developing a distributed control methodology that is ro- on Decision and Control . (Osaka, Japan, 15-18 Dec. bust with respect to system disturbances. 2015). Furthermore, the Smart Grid Research and Development 7. Kundu, S., and M. Anghel. A multiple-comparison- Program of the DOE/OE, identifies in its multi-year pro- systems method for distributed stability analysis of gram plan the development of advanced control methods. large-scale systems. Automatica. As the smart grid evolves, this program recognizes that more complex automated control systems will be neces- Publications sary to maintain the grid’s optimum operation. Control Anghel, M., F. Milano, and A. Papachristodoulou. Algo- algorithms investigated in this project directly address the rithmic Construction of Lyapunov Functions for Power fundamental conceptual challenges posed by such auto- System Stability Analysis. 2013. IEEE Transactions on mated control systems. Circuits and Systems I: Regular Papers. 60 (9): 2533. Finally, the April 2011 DOE workshop, “Computational Anghel, M., J. Anderson, and A. Papachristodoulou. Stabil- Needs for the Next Generation Electric Grid Proceedings,” ity Analysis of Power Systems using Network Decom- highlights a class of mathematical and computational position and Local Gain Analysis. 2013. In Bulk Power problems relevant for potential power systems research. System Dynamics and Control Symposium.. (Crete, It identifies, in particular, the difficulties currently encoun- Greece, 25-30 Aug. 2013). , p. 1. NA: IEEE Explore. tered by direct transient stability analysis methods. Again, Kundu, S., and M. Anghel. Sum-of-Squares Approach to the this project directly addresses this problem and our prelim- Stability Analysis of Inter-connected Systems. Present- inary results have already offered significant breakthrough ed at SCONES 2014, IEEE Symposium on the COntrol of results based on novel algebraic control techniques. NEtwork Systems.. (Boston, 27-28 Oct. 2014). References Kundu, S., and M. Anghel. A sum-of-squares approach to

1. Anghel, M., F. Milano, and A. Papachristodoulou. Algo- the stability and control of interconnected systems rithmic construction of Lyapunov functions for power using vector Lyapunov functions. 2015. In The 2015 system stability analysis. 2013. IEEE Transactions on American Control Conference. (Chicago, 1-3 July). , p.
2. Explore Digital Library: IEEE . Circuits and Systems I: Regular Papers.. 60 (9): 2533. Kundu, S., and M. Anghel. A sum-of-squares approach to
3. Anghel, M., J. Anderson, and A. Papachristodoulou. the stability and control of power systems via vec- Stability analysis of power systems using network tor Lyapunov function. To appear in European Control decomposition and local gain analysis. 2013. In Bulk Conference. (Linz, Austria, 15-17 July 2015). Power System Dynamics and Control Symposium. (Crete, Greece, 25-30 Aug. 2013). , p. 1. Explore Digital Kundu, S., and M. Anghel. Computation of linear compari- Library: IEEE . son equations for stability analysis of interconnected systems. To appear in 54th IEEE Conference on Deci-
4. Kundu, S., and M. Anghel. A sum-of-squares approach sion and Control. (Osaka, Japan, 15-18 December, to the stability and control of interconnected systems 2015). using vector Lyapunov functions. 2015. In The 2015 Kundu, S., and M. Anghel. Distributed coordinated control American Control Conference. (Chicago`, 1-3 July, of large-scale nonlinear networks. To appear in 5th 2015). , p. 5022. Explore Digital Library: IEEE. IFAC Workshop on Distributed Estimation and Control in Networked Systems. (Philadelphia, 10-11 Septem-
5. Kundu, S., and M. Anghel. A sum-of-squares approach ber, 2015). to the stability and control of power systems via vec- tor Lyapunov functions. To appear in The 2015 Europe- Motter, A. E., S. A. Myers, M. Anghel, and T. Nishikawa. an Control Conference. (Linz, Austria, 15-17 July, 2015). Spontaneous synchrony in power-grid networks. 2013. NATURE PHYSICS. 9 (3): 191.
6. Kundu, S., and M. Anghel. Distributed coordinated con- 262

Page 263

Information Science & Technology Exploratory Research Final Report Integrated Photonics Pathfinder (IPP) Kevin P. Mccabe 20140507ER Abstract within cryptographic algorithms. Single-photon, quantum communications (QC) offers Single-photon quantum communications (QC) has the “future proof”, lightweight cyber security solutions for potential to finally give defenders dominance over at- networks. The rapidly evolving network environment tackers in cyber space. QC has “future-proof” security (cloud computing, handheld devices, SmartGrid and rooted in laws of physics: today’s quantum-secured com- other critical infrastructure) presents cyber challenges munications cannot be compromised by unanticipated such as low latency one time signatures that are difficult future technological advances. But QC today only exists or impossible to meet with conventional cryptography. in point-to-point instantiations that have limited ability In addition to the myriad technical challenges, there is a to address the cyber security challenges of our increas- huge cost to continuously upgrade network infrastruc- ingly networked world. ture based on the larger and larger key sizes needed for the temporary security offered by our current public The IPP project, by use of emerging Si integrated electro- key infrastructure. The value proposition of quantum photonics has the overarching goal to reduce the QC communications drives the QC client (transmitter) to be transmitter size and cost by 100x each, and improve manufacturable and lightweight in terms of computa- controllable attenuation by 4 orders of magnitude. By tion, size and power. By leveraging U.S. investment in design, fab and testing of photonic die we seek to vali- emerging integrated photonics processes it is possible date performance models of optical components neces- to replace the current state of the art large and expen- sary for the coexistence of quantum communications sive, hand-built, optical and electrical transmitters with (QC) with conventional optical network communications. a monolithic integrated device and thus make quantum While LANL has demonstrated QC using modular inte- secured networks commercially viable. This research grated photonics, QC has never before been performed extends monolithically integrated photonics into an 11 with monolithically integrated photonics. Our research orders-of-magnitude smaller optical power (quantum) leverages the US investment in the underlying technol- regime to address key risks. ogy of optical telecommunications infrastructure and applies it to the quantum regime. Background and Research Objectives Communications networks have revolutionized the way Scientific Approach and Accomplishments we work, live, operate computer systems and run our Previous experience with integrated optics for QC has national infrastructure. But hardly a day goes by without shown that there are several areas of risk that will have a report highlighting the pressing need for improved to be managed in order for Si photonic systems to be cyber security technologies to protect our economic applicable to QC. The first is the modulation scheme. We and National security. The widespread deployment have found that photon polarization qubits are superior of cryptography provides confidentiality, authentica- to phase-based qubits even for optical fiber systems that tion, identification and other security functions for the require polarization correction. However, polarization protection of private data in motion as well as ensuring based transmitters and receivers must support multiple that control, access and network management can only polarization states. This means that integrated photon- be performed by authorized parties. Critical to these ics waveguides must work with both orthogonal po- functions is the secure distribution of the secret ran- larizations within a polarization basis. In prior research dom number sequences known as cryptographic keys to our Gen1 modular QC transmitter uses a GaAs phase authorized users [6]: secret keys are used as parameters 263

Page 264

modulator whose waveguide supports propagation of be a subject of future research. In working with research- both transverse electric (TE) and transverse magnetic(TM) ers at OpSIS and Acacia no fundamental problems were modes and modulates only the TE mode. That way we identified so we have gained insight and confidence in our can produce the minimum desired number of polariza- research approach. tion states, four, with a single modulator. However in Si photonics, the waveguides do not support both TE and TM Impact on National Missions polarizations simultaneously and the modulation methods, Cyber-security is a top national security threat. QC can injection of charge carriers into the waveguide, do not address long-duration/high-value security needs within selectively delay one polarization over the other. There- many network environments, such as: highly secure DoD fore we must split the single, TE-mode polarization into and DOE complexes; between government agencies in two waveguides, modulate one and recombine the two the Washington, DC area; financial networks; supervisory, waveguides with one of them rotated by 90° relative to control and data acquisition (SCADA) systems for critical the other one. We need to experimentally test combining infrastructure (power, water and SmartGrid) networks; or the two polarizations in a Si oxide waveguide on the same within a US Embassy compound. Further applications will substrate. The other area of risk is in the isolation needed extend security to constrained network environments. to protect our quantum channel and detectors from the Examples include: warhead verification and IAEA treaty conventional communication channel that must operate monitoring. Miniaturized QC systems will enable a hand- over the same optical communication fiber but be split and held quantum secured devices to be used for identifica- isolated in the photonics. In general the highest available tion, authentication, access control and secure telephone controllable extinction ratio in directional couplers is 60 calls. An aircraft or satellite acting as a trusted node could dB, but we require 110 dB. We believe that Si photonic establish ad hoc secure networks of ground, sea and air- couplers, unlike other directional couplers, can achieve based users on a continental or even a global scale. higher isolation ratios by cascading several components on the same substrate because there is little scattering or Quantum secured communications technology at the mode mixing to reduce the extinction ratio as is the case Laboratory has been licensed to Whitewood Encryption with discretely coupled components. Systems, Inc. a wholly owned subsidiary of Allied Minds. In year 1 of our 2 year research plan we collaborated with the University of Delaware that had a program called OpSIS for Optoelectronic Systems Integrated in Silicon. OpSIS separates design from fabrication and combines many de- signs on a single wafer to provide low cost access to cutting edge foundries. Through training and dialog with OpSIS we drafted specifications for the optical components of a monolithic quantum communications module. A photonic conceptual level design showing feasibility was completed. As we were preparing for a fabrication run of our first test chips, OpSIS lost its primary funding source and fabrication services ceased. LANL continued the research by collabo- rating with a senior researcher at Acacia Communications Inc., a recognized leader in optical transceivers used for the internet backbone. Physical designs are process depen- dent so the LANL developed physical designs for OpSIS had to be adapted by OpSIS for their process. This had some advantages because the Acacia process is more optimal for our application than the OpSIS was. Two fabrication runs were planned. The first run was with simple test struc- tures: a tunable splitter and phase shifters. This fabrica- tion run had a yield failure unrelated to our research. The second run combines some of the structures from the first run with cascaded splitters used as an attenuator and added an amplitude modulator. The second run did yield and characterization of the test structures on those die will 264

Page 265

Information Science & Technology Postdoctoral Research and Development Continuing Project Mixing and Diffusion in Granular Flows Ivan Christov 20130792PRD2 Introduction flow of granular materials. The ability to improve model This project aims to understand the complexity of force- mixing of granular matter will impact energy applications carrying contacts in a granular medium, which can often including storage and transportation of energy-related lead to its destabilization (due to external forcing) along materials (e.g., coal and biofuels). Granular materials unexpected fault lines, as commonly observed in geo- tend to segregate according to size, a property that can physical situations including earthquakes, mudslides, be an obstacle in manufacturing and transportation pro- and avalanches. The approach will be twofold. First, a cesses. This project addresses precise mixing character- model of how granular materials flows will be developed istics of granular materials and could lead to the design for the types of conditions encountered in geophysical of optimized container design for improved mixing and processes. This is a high-potential scientific problem be- more efficient transportation. The capabilities developed cause, in over three decades of research, this most basic in this project will also be relevant for other DOE pro- question regarding granular materials remains unsettled. grams and needs including waste storage, waste han- Second, the force-chain network structure of granular dling and transportation, high explosives, and handling materials will be obtained from model experiments, of materials produced in powder form. being conducted at Los Alamos, of a granular medium between sheared plates responding to the stimulus of Progress their relative motion. The mathematical techniques of The scientific accomplishments of the project in FY15 network analysis will be used to better understand the include the construction of a flow regimes diagram for dynamic behavior of the medium under such loads. This injection of a buoyant liquid into a confined porous is a cutting-edge endeavor because few such laboratory medium. Although the dynamics of granular media scale experiments and analysis have been carried out under various forcing and excitations are of interest on previously. Additionally, new algorithms for simulating their own, a common geophysical problem involves the universal emergent behaviors (clustering, ordering, penetration of a liquid into the pore space between topology) arising from the interplay between discontinu- static grains. Specifically, such fluid injection into a ous granular mixing dynamics and microscopic physics porous medium occurs in a large variety of geophysi- will be developed as part of this project. This is a high- cal and industrial processes such as seawater intrusion risk undertaking because scaling dynamics of granular into coastal aquifers, geological sequestration of carbon materials in geophysical situations to laboratory scale dioxide, waste fluid disposal, and oil and gas recovery. experiments has yet to be demonstrated. Success in the When the injected and displaced fluids are immiscible, project would imply the ability to extend the scaling con- a free boundary problem can be formulated to de- cepts that have been so successful in fluid dynamics and scribe the time evolution of the fluid-fluid interface. It aerodynamics to a new area: solid mechanics. is of interest to know how this interface evolves over time in order to understand the sweep efficiency of the Benefit to National Security Missions injection process. Work performed under this project Granular materials are ubiquitous and are found in a in FY15 focused on the case of two-dimensional (Car- host of applied programs. Hence, understanding mix- tesian) vertically confined porous media. Assuming the ing and flow of granular materials has impacts on many injected fluid is denser than the displaced fluid, then applied programs. This project aims to develop a model- the injected fluid only attaches to the bottom bound- ing and simulation capability for describing mixing and ary at early times for point source injection. As time 265

Page 266

progresses, the injected fluid contacts the top boundary, Future Work and the fluid flow becomes confined. Results obtained in In FY16, we will perform theoretical modeling of the this project establish that during the propagation process, experiments being conducted at Los Alamos on sheared because of the change of the horizontal and vertical length granular media confined between two moving plates. In scales, the interface exhibits a transition from an early time FY16, we will further expand our stated focus on geophys- unconfined behavior to a late time confined behavior, and ics, connecting to ongoing work at LANL on subsurface we derived a nonlinear advection-diffusion equation to transport of particulate materials with applications to describe the time evolution of the fluid- fluid interface. We hydraulic fracturing. The latter systems consist of granular analyzed this partial differential equation in two distin- materials undergoing (potentially large) deformation and guished asymptotic limits (early and late time) and found shear. A new aspect of this problem, then, is the possibility exact solutions in each case, resolving the structure of the of comminution (or grinding, breakage, etc). Furthering the moving front. Through detailed numerical simulations, we ongoing theoretical modeling we are performing, we will were then able to construct a phase diagram depending on incorporate comminution into the models. The models will two dimensionless quantities: the viscosity ratio of the dis- be analyzed to produce predictions that will be compared placed fluid to the injected fluid, and a dimensionless time. against observational data and benchmarked numerical In doing so, the transition processes between the early and simulations being performed at Los Alamos. Additionally, late time asymptotic limits were elucidated and the rel- in FY16, the theory will be extended to account for non- evance of the early time and late time exact solutions were uniformities of the flow conduits due to fracturing, com- demonstrated for the first time in the literature. minution and other physical effects present in such high- pressure situations. We will also study the effects of grain Another scientific accomplishment of this project in FY15 fragmentation on granular flow, a problem with relevance is the Hamiltonian formulation of the equations for wave to geophysical granular flows and earthquake physics. propagation in continua with microstructure such as a granular material. Although a granular materials’ mi- Conclusion crostructure is quite complex in practice, the collective The outcome of the first part of this work will be improved (continuum) behavior has been observed and well docu- prediction of rare and destructive seismic events, which mented. The simplest way to study wave propagation in have significant implications for society at large. The such a context is to derive a generic (universal) model, outcome of the second part of this work will be improved based on extended thermodynamics, for a continuum understanding of how localized processes affect structure with one inherent length scale (e.g., the grain size). Then, formation, which can lead to better prediction of dynamics by standard techniques, one can obtain weakly nonlin- as diverse as percolation in porous medium and spread of ear dispersive evolution equations for wave propagation infections on networks. through this continuum. What had not been appreciated in previous discussion of such equations in the literature is Publications that they possess Hamiltonian structure, endowing them Christov, I. C.. On a hierarchy of nonlinearly dispersive gen- with a number of desirable physical properties. We con- eralized KdV equations. 2015. PROCEEDINGS OF THE structed the Hamiltonian formulation of such nonlinearly ESTONIAN ACADEMY OF SCIENCES. 64 (3): 2012. dispersive wave equations for the first time, showing there is in fact a hierarchy of such nonlinearly dispersive evolu- Christov, I. C.. Comment on “The velocity field due to an tion equations. Additionally, as another novel feature of oscillating plate in an Oldroyd-B fluid” by C.C. Hopkins our work, we derived exact solitary wave solutions of these and J.R. de Bruyn [Can. J. Phys. 92, 533 (2014)]. 2015. equations with curious properties: namely, waves that are CANADIAN JOURNAL OF PHYSICS. 93 (12): 1651. compact and peaked. Christov, I. C., R. M. Lueptow, J. M. Ottino, and R. Sturman. A Study in Three-Dimensional Chaotic Dynamics: Gran- In FY15, this project has yielded over a dozen invited ular Flow and Transport in a Bi-Axial Spherical Tumbler. presentations and two contributed presentations (inter- 2014. SIAM JOURNAL ON APPLIED DYNAMICAL SYS- national and regional conferences such as the American TEMS. 13 (2): 901. Physical Society’s Division of Fluid Mechanics Annual Meet- ing). In addition, two journal publications have appeared Christov, I. C., R. M. Lueptow, J. M. Ottino, and R. Sturman. in top disciplinary and multidisciplinary journals (Journal of A Parametric Study of Mixing in a Granular Flow a Bi- Axial Spherical Tumbler. Dynamical Systems - Model- Fluid Mechanics, Physical Review E), three more have been ling. Edited by Awrejcewicz, J.. submitted for review, and a number are in preparation. Christov, I. C., and H. A. Stone. Shear dispersion in dense 266

Page 267

granular flows. 2014. GRANULAR MATTER. 16 (4): 509. Christov, I. C., and P. M. Jordan. On an instability exhibited by the ballistic-diffusive heat conduction model of Xu and Hu. 2014. ROCEEDINGS OF THE ROYAL SOCIETY A-MATHEMATICAL PHYSICAL AND ENGINEERING SCI- ENCES. 470 (2161): 20130557. Christov, I. C., and P. M. Jordan. Corrections to Morse and Ingard’s variational-based treatment of weakly-nonlin- ear acoustics in lossless gases. 2015. JOURNAL OF THE ACOUSTICAL SOCIETY OF AMERICA. 138 (1): 361. Christov, I. C., and P. M. Jordan. Maxwell’s “other” equa- tions. 2015. The Royal Society Publishing Blog. Khare, A., I. C. Christov, and A. Saxena. Successive phase transitions and kink solutions in phi(8), phi(10), and phi(12) field theories. 2014. PHYSICAL REVIEW E. 90 (2): 023208. Zheng, Z., B. Guo, I. C. Christov, M. A. Celia, and H. A. Stone. Flow regimes for fluid injection into a confined porous medium. 2015. JOURNAL OF FLUID MECHAN- ICS. 767: 881. Zheng, Z., I. C. Christov, and H. A. Stone. Influence of het- erogeneity on second-kind self-similar solutions for viscous gravity currents. 2014. JOURNAL OF FLUID ME- CHANICS. 747: 218. 267

Page 268

Information Science & Technology Postdoctoral Research and Development Continuing Project Thermodynamics and Information Processing at the Nanoscale Wojciech H. Zurek 20140667PRD2 Introduction course of the engine cycle can be comparable to ther- Sebastian Deffner is an emerging leader in the field of modynamic entropy, and its efficient processing using nonequilibrium statistical mechanics (as recognized by available hardware (which includes quantum informa- the community, and current leaders in the field). As a tion processing) is important. Information processing, postdoctoral fellow, he will bring his cutting-edge ex- including its quantum aspects, is of interest to LANL and pertise in nonequilbrium processes to LANL with direct DOE. relevance across disciplines. His research program is directly aligned with one of the Los Alamos mission to Progress foster energy security. In the last fiscal year we have significantly advanced our understanding of quantum thermodynamic devices and Nanodevices running thermodynamic cycles---Deffner’s of optimal quantum processes. Our progress is reflected principal area of research---provides optimal ways of in five publications and two preprints. converting different types of energy (e.g. thermal into mechanical) and is important for a variety of applica- Together with our collaborator Haitao Quan (Peking tions. University, formerly LANL director funded postdoc) we computed the work distribution for quantum pistons The engineering of highly efficient quantum engines and operating with non-interacting bosons and fermions. thermodynamic cycles is a low risk, cutting edge area of This work has been published in Phys. Rev. E 90, 062121 research with the potential to advance efficient energy (2014). conversion techniques and to directly impact ongoing experimental progress around the world. Similar results have also been obtained for quantum processes in diatomic molecules. Chem. Phys. 446, 18 There is already significant related interest at LANL. In (2015) reports the analytical treatment of the quantum addition to ongoing research in nanotechnology, cur- work statistics in driven Morse oscillators. rent efforts at P-21 (Boshiers’s lab), are devoted to the implementation of a superadiabatic quantum piston [A. In Phys. Rev. Lett. 114, 150601 (2015) we generalized del Campo, M. G. Boshier, Sci. Rep. 2, 648 (2012)], one the quantum Jarzynski equality to non-linear quantum of the key components to realize the engineered energy systems. In particular, we showed how quantum ther- systems proposed by Deffner for optimizing energy con- modynamic relations generalize to PT-symmetric quan- version and use. tum mechanics, i.e., a non-hermitian quantum theory. These results will also be present in an upcoming CNLS Benefit to National Security Missions workshop on PT-symmetry and a CNLS workshop on Deffner will explore laws of thermodynamics as they related topics for Feb 2016 is in preparation. apply on a nanoscale. At that level thermal and quantum fluctuations become important, and modify optimal Our insight in the foundations of statistical mechanics strategies for extraction of useful work. Efficient acquisi- attracted the attentions of Nature Physics. In an invited tion of energy is one of the missions of DOE, and of keen News and Views article Deffner summarized the state- interest to LANL. Moreover, at the level at which Seba- of-art and future perspectives of the role played by tian will explore thermodynamics, information gained in quantum entanglement in the foundations of statistical 268

Page 269

mechanics. Nature Physics 11, 383 (2015) In a new collaboration with the experimental group of Luis Orozco we have built expertise in optimizing quantum pro- In particular, we were able to show that the foundations cesses in cavity QED. In a next step we will be aiming at the of statistical physics can be built from purely quantum development of an ‘optimal’ quantum heat engine based symmetries — entanglement assisted envariance: arX- on this novel technology. iv:1504.02797 Conclusion We also furthered our understanding of optimal quantum We expect our results to have impact on the design of processes. A recent PRL reports the experimental study of quantum systems in the development of quantum comput- environment assisted speed-ups in cavity QED. This experi- ers, and in the fundamental understanding of nanotechnol- ment has been performed in the group of Luis Orozco at ogy experiencing quantum effects. Our research will open the University of Maryland: arXiv:1503.02591 new avenues in the fundamental understanding of time- dependent quantum systems operating arbitrarily far from Finally, in arXiv:1503.03455 Gardas and Deffner showed thermal equilibrium, and thereby process information. that no quantum heat engine can be more efficient than a Highly efficient quantum cycles will be designed. Carnot engine. This result settled a discussion brewing in the literature for over ten years. The main novel insight is Publications how to properly define heat and work for quantum sys- Acconcia, T. V., M. V. S. Bonança, and S. Deffner. Shortcuts tems with non-negligible coupling to the environment. to adiabaticity from linear response theory. To appear in PHYSICAL REVIEW E. Future Work Cimmarusti, A. D., Z. Yan, B. D. Patterson, L. P. Corcos, L. A. The main objective of this project research is the develop- Orozco, and S. Deffner. Environment assisted speed-up ment of a coherent framework for the understanding, the of the field evolution in cavity quantum electrodynam- design, and the theoretical foundations of nanothermody- ics. 2015. PHYSICAL REVIEW LETTERS. 114: 233602. namic devices, whose modes of operation allow the imple- mentation of information processing, and which generi- Deffner, . TEN YEARS OF NATURE PHYSICS From spooky cally operate far from thermal equilibrium. In collaboration foundations. 2015. NATURE PHYSICS. 11 (5): 383. with the experimental group of Ferdinand Schmidt-Kaler, Deffner, , and Saxena. Jarzynski Equality in PT -Symmetric we developed only recently the smallest possible quantum Quantum Mechanics. 2015. PHYSICAL REVIEW LET- heat engine, whose working medium consists of a single TERS. 114 (15). trapped ion. Deffner, S., and A. Saxena. Quantum work statistics of The experimental principle can be extended in two direc- charged Dirac particles in time-dependent fields. 2015. tions: (i) To date, we only studied a single ion trapped in PHYSICAL REVIEW E. 92: 032137. a one-dimensional harmonic oscillator. Optical lattices Gardas, B., and S. Deffner. Thermodynamic universality of are, to very good approximation, a network of harmonic quantum Carnot engines. 2015. PHYSICAL REVIEW E. oscillators. Thus, the working principles of the one-ion 92: 042126. heat engine can be readily generalized to multiple particles evolving in an optical lattice. Such a device would enable Gong, , Deffner, and H. T. Quan. Interference of identical the experimental study of the thermodynamic properties particles and the quantum work distribution. 2014. of fully controllable multi-particle quantum systems. (ii) PHYSICAL REVIEW E. 90 (6). In previous research, we also have been interested in the Leonard, , and Deffner. Quantum work distribution for a interplay of (quantum) information and thermodynamics. driven diatomic molecule. 2015. CHEMICAL PHYSICS. However, how to describe a quantum information reser- 446: 18. voir, i.e. a fully quantum mechanical hard disk, and its ther- modynamic properties is to date still an open question. To this end, we will be aiming at simple generalizations of the on-ion heat engine towards a minimal quantum Maxwell demon - a perfect device to study quantum thermodynam- ics and information theory. We are also planning to inves- tigate fluctuation relations generalizing the statements of the second law of thermodynamics and their consequenc- es on the dynamics of nanodevices. 269

Page 270

Information Science & Technology Postdoctoral Research and Development Final Report A Quadrature Approach for Non-Gaussian Uncertainty Representation and Propagation David Palmer 20130784PRD2 Abstract impossible to maintain persistent surveillance result- If the position and velocity of an object is known, and ing in large observation gaps. This inherent latency in with forces affecting its motion, good predictions as to the catalog information results in sparse observations where it will be in the future can be made. However, in and large propagation intervals between measurements the real world, the quantities are only approximately and close approaches. The large propagation intervals known, and only estimates of the region where the coupled with nonlinear RSO dynamics results in highly object might be found in the future can be made. This non-Gaussian probability distribution functions (pdfs). project looked at predicting satellite positions, but the In particular, low-Earth orbiting (LEO) satellites are heav- techniques being developed are more general and can ily influenced by atmospheric drag, which is very difficult be applied to a number of practical problems. to model. Uncertainties in atmospheric influences must be folded into estimation models to accurately represent Current standard prediction techniques simplify the the position uncertainties in calculating impact probabili- uncertainty to regions that are described by spheroids ties. This process then separates naturally into a predic- that may be stretched like a football, but not bent like a tion and correction cycle, where estimates are used to banana, coiled like a spring, or split into multiple re- predict the orbital position at a future time and observa- gions. These more complicated uncertainty regions can tions are used to improve or correct these predictions be explored by simulations. while decreasing uncertainty. The difficulty in this pro- cess lies in representing the non-Gaussian uncertainty In more technical terms; this work sought to improve and accurately propagating the uncertainty through the on the traditional sigma point approaches to nonlinear orbital dynamics model. Accurate assessment of confi- filtering by considering a more general formulation, and dence in position knowledge will be a significant im- applying the new sigma point filtering method to highly provement, particularly for the space situational aware- relevant problems. Our study was focused on space ness (SSA) community and the warfighter. situational awareness, which is a key topic in the Grand Challenge for Sensing and Measurement Science for Scientific Approach and Accomplishments Global Security. However, more generally, it covers many This postdoctoral fellowship work focused on the de- fields of simulation. velopment of the next generation of sigma point filters; improving on the traditional approaches by consider- Background and Research Objectives ing a more general formulation, and applying the new Recent events in space, including the collision of Rus- method to the highly relevant problem of SSA. The new sia’s Cosmos 2251 satellite with Iridium 33 and China’s approach to uncertainty propagation was developed by Feng Yun 1C anti-satellite demonstration, have stressed using the theory of Gauss-Hermite Quadrature (GHQ) the capabilities of Space Surveillance Network (SSN) integration and extending GHQ using Generalized Gauss- and its ability to provide accurate and actionable impact ian Cubature (GGC). The approach was applied to a probability estimates. The SSN network has the unique number of problems to study the convergence proper- challenge of tracking more than 18,000 resident space ties, robustness, and accuracy of the proposed method. objects (RSOs) and providing critical collision avoidance warnings to military, NASA, and commercial systems. Our new approach, named GGC filter, is a novel method However, due to the large number of RSOs and the for nonlinear filtering based on a generalized Gaussian limited number of sensors available to track them, it is cubature approach. Specifically, it is a new point-based 270

Page 271

nonlinear filter that is not based on one-dimensional UKF has the worst performance while the GGC method quadrature rules, but rather uses multi-dimensional cuba- does a better job at tracking the truth. All filters diverge ture rules for Gaussian distributions. The new generalized initially, but the UKF has the largest divergence. Also the Gaussian cubature filter is not in general limited to odd- UKF takes the longest amount of time to converge to the order degrees of accuracy, and provides a wider range of true state. order of accuracy. The method requires the solution of a set of nonlinear equations for finding optimal cubature points, but these equations need only be solved once for each state dimension and order of accuracy. This rule can also be extended to anisotropic cases where the order of accuracy is not isotropic in dimension. This method allows for tuning of the cubature rules to develop problem spe- cific rules that are optimal for the given problem. The gen- eralized Gaussian cubature filter is applied to benchmark problems in astrodynamics, and it is compared against existing nonlinear filtering methods. To test the performance of the new GGC rules for propa- gation of uncertainty through a dynamic system, a simple nonlinear two-dimensional example is used. This example models a nonlinear spring-mass system with a nonlinear friction term. The system dynamics for this simple example used a cubed term for the spring force. The simulation Figure 1. Errors in a model estimator at different Gaussian Cuba- time considered is 200 seconds with a sampling interval ture orders. The modeled system is a mass on a non-linear spring 0.1 seconds. Monte Carlo simulations were conducted to (force proportional to the cube of the displacement) with non- test the GGC approaches for different orders. The Monte linear friction. The fit improves rapidly as the order increases. Carlo samples are taken from the initial distribution, and 1,000 samples are used in the comparisons. The errors in the means of the system’s state is computed for the different order of the GGC method and compared to the means calculated from the Monte Carlo samples at each time step. These errors are shown in Figure 1 for the system state mean calculated with GGC. From this figure it can seen that as the order increases, the mean is more accurately estimated. During the initial portion of the simulation the errors are large for all the models since the nonlinear spring is fluxing greatly but the higher-order GGC still has better performance. As the energy dissipates the system fluctuation is lower but the higher-order models still show better performance. A second example is considered, a nonlinear spring-mass model state estimation example with nonlinear measure- Figure 2. The state space (position and velocity) trajectory of ments. The nonlinear spring-mass model is the same as the model system as in Figure 1, and some estimator outputs. In the one used in the first example to test the uncertainty this case, the estimator input was the exponential of the position, propagation performance of the GGC approach. Measure- plus noise. Because the Unscented Kalman Filter (UKF, red) was ments are simulated assuming a measurement model that developed for simple conditions, it has trouble with the non- consists of the exponential of the first state and additive linear measurements (which lead to non-Gaussian noise in the random noise. Measurements are sampled at 0.1 second derived position) showing much more erratic variation than the Gaussian Cubature models. intervals for a total of 200 second simulation time. The final example considered is an attitude estimation The state estimates for UKF, GGC order 3, and GGC 5 are example satellite orientation estimation. The performance shown in Figure 2. From Figure 2 it can be seen that the 271

Page 272

of the GGC filter, UKF, and the EKF are compared. The GGC method than it does with competing approaches. uses the approach for representing the pdf as sigma points discussed earlier. Large initial errors are considered. It is important to note that with large initial errors the attitude distribution becomes highly nonlinear. Attitude errors of -50, 50, and 160 degrees for each axis, respectively, are added into the initial-condition attitude estimate. The initial attitude standard deviation is set to (50 degrees) for each attitude component. Gyro biases are also estimated. The initial bias standard deviation is set to (20 deg/hr) for each axis. A plot of the norm of the attitude errors for this simulation case is shown in Figure 3. The EKF does not converge for this case since the first-order approxima- tion cannot adequately capture the large initial condition errors. The UKF does have better convergence properties than the EKF for this case, however the GGC provides the best convergence performance. Both the EKF and the UKF are more sensitive to the initial conditions than the GGC. The performance of the GGC is checked by running the fil- Figure 4. Comparison of growth of computational requirements ter 100 times. It always converges and the statistics of the vs order for different analysis schemes. Note the extreme loga- estimation results in the 100 runs are almost identical. rithmic scale, indicating that for 14th order accuracy, a full-grid approach would need millions of times more calculations than the Cubature approach. Even the more sophisticated previous approaches require 100-1000x the calculations. Impact on National Missions This research represents a novel and ground-breaking ap- proach to uncertainty quantification that has application across a number of areas in LANL’s portfolio. In particular, the development of accurate and consistent methods of characterizing uncertainty is highly relevant for the space situational awareness, which is key to insuring the security of national space systems critical to nonproliferation and other security challenges. This research offered an exciting opportunity to further solidify the theoretical underpin- ning of powerful and well-known methods in nonlinear/ non-Gaussian estimation while developing a new and computational efficient technique. Figure 3. Attitude estimation of a satellite. Given an initially This work investigated a technique called Generalized large uncertainty in the attitude, the Extended Kalman Filter Gaussian Cubature, where the rules are constructed to in- (EKF) never develops a good approximation to the satellite tegrate a set of functions exactly. Hermite and Sparse Ten- attitude. An Unscented Kalman Filter (UKF) does discover the sor products approaches for multi-dimensional integrals attitude and gradually improve its estimate, while the new algo- were discussed. The generalized cubature approach was rithm gives a large improvement in both speed and accuracy. used to solve multi-dimensional integrals of Gaussian dis- tributions. The solution was simplified greatly by using the For a general nonlinear problem, the Gaussian cubature generalized cubature technique along with properties of method requires less function evaluations as compared to Gaussian distributions. Additionally, the solution of nodes existing approaches to achieve equivalent accuracy. Figure and weights for the generalized cubature of a Gaussian 4 shows a comparison of the number of terms (function distribution was also greatly simplified by using the multi- evaluations) as a function of order required by the GGC dimensional Hermite polynomials. Cubatures were derived method vs the number required by current existing ap- for a number of orders and dimensions and used for both proaches. From this figure it is clear that the number of uncertainty propagation and sequential state estimation. function evaluations required grows slower with GGC 272

Page 273

Good results were shown with the new methods with and J. Koller. Modeling Satellite Drag Coefficients with three example problems. The authors are very optimistic Response Surfaces. 2014. Advances in Space Research. about the approach and believe that it can greatly improve (Oct): 00. the uncertainty quantification results for a large number of practical problems. References

1. Linares, R., and I. Impact-Team. A Framework for Inte- grated Modeling of Perturbations in Atmospheres for Conjunction Tracking (IMPACT). 2014. (Pasadena, CA, 2014). p. 0. Pasadena, CA: AIAA.
2. Linares, R., W. Andrew, M. A. Shoemaker, and J. L. Crassidis. Characterization of Inactive Rocket Bodies Via Non-Resolved Photometric Data. 2014. (Maui, Hawaii, 2014). p. 0. Maui, Hawaii: AMOS.
3. Klimenko, A., A. Brennan, H. Godinez, D. Higdon, J. Koller, E. Lawrence, R. Linares, D. M. Palmer, M. Shoe- maker, D. Thompson, A. Walker, B. Wohlberg, M. K. Jah, E. Suton, T. Kelecy, A. Ridley, and C. McLaughlin. A Framework for Integrated Modeling of Perturbations in Atmospheres for Conjunction Tracking (IMPACT). 2015. In AIAA Book on Space Operations, 2015. By AIAA, . p.
4. Reston, VA, USA: AIAA.
5. Mehta, P. M., A. Walker, E. Lawrence, R. Linares, D. Hig- don, and J. Koller. Modeling Satellite Drag Coefficients with Response Surfaces. 2014. Advances in Space Research. (Oct): 00. Publications Klimenko, A., A. Brennan, H. Godinez, D. Higdon, J. Koller, E. Lawrence, R. Linares, D. M. Palmer, M. Shoemaker, D. Thompson, A. Walker, B. Wohlberg, M. K. Jah, E. Suton, T. Kelecy, A. Ridley, and C. McLaughlin. A Frame- work for Integrated Modeling of Perturbations in At- mospheres for Conjunction Tracking (IMPACT). 2015. In AIAA Book on Space Operations, 2015. By AIAA, . , p. 0. Reston, VA, USA: AIAA. Linares, R., W. Andrew, M. A. Shoemaker, and J. L. Crassidis. Characterization of Inactive Rocket Bodies Via Non- Resolved Photometric Data. 2014. In Advanced Maui Optical and Space Surveillance Technologies (AMOS) Conference. (Maui, Hawaii, 2014). , p. 0. Maui, Hawaii: AMOS. Linares, R., and I. Impact-Team. A Framework for Inte- grated Modeling of Perturbations in Atmospheres for Conjunction Tracking (IMPACT). 2014. In AIAA - Space Operations Conference. (Pasadena, CA, 2014). , p. 0. Pasadena, CA: AIAA. Mehta, P. M., A. Walker, E. Lawrence, R. Linares, D. Higdon, 273

Page 274

Materials for the Future

Page 275

Materials for the Future Directed Research Continuing Project Photoactive Energetic Materials for Quantum Optical Initiation Robert J. Scharff 20140005DR Introduction the Weapons Program, but will also provide significant This project provides an innovative approach to establish advancements towards utilizing laser pulses as power- design principles and a scientific methodology for gener- ful sources for manipulating and controlling atomic and ating photoactive energetic materials with controllable molecular processes on unprecedented time and length optical functionality. The project aims to increase the scales. controllability of chemical dynamics in novel photoactive Benefit to National Security Missions materials through a concerted experimental and theo- retical effort that characterizes the dynamics of energy The overarching goal of this project – to achieve quan- localization, chemical bond activation, and chemical tum control of explosive initiation – utilizes the theory- reactions. This approach will provide critical insight on guided, materials-by-design approach outlined in the how to manipulate electronic structure through synthe- Materials Pillar Extreme Environments theme of “pre- sis in order to generate the desired material response to dicting and controlling functionality of materials in these overcome mechanisms that limit controllability. extremes”. Critical to the Materials Pillar mission is the establishment of design principles and scientific proto- The authors propose a three-part solution for the suc- cols for manufacturing advanced materials to control cessful design of photoactive energetic materials for functionality for predefined performance criteria central quantum optical initiation. The first part involves the to LANL mission needs. DOE programmatic drivers exist development of new explosives derivatized with an opti- for the development of photoactive energetic materi- cal chromophore. The optimization of optical response als that can address future milestones for enhanced properties (e.g. 1-photon and 2-photon absorption safety, security, and use-control of the nuclear stockpile. cross-sections and resonant wavelengths) of the photon Moreover, there are DoD directives for the develop- absorbing chromophore will enable efficient coupling of ment of new energetic materials that are insensitive to the laser field to the explosive and facilitate control of unplanned stimuli (e.g., spark, impact, and friction). The excited state dynamics. The second part addresses the development of photoactive energetic materials that op- need for the elucidation of electron-vibrational dynamics tically initiate through quantum controlled photochem- following photoexcitation. Our feasibility studies support istry would yield a “quantum lock-and-key” capability our main idea that exothermic NO2 formation could be valuable not only in applications for improved detonator achieved photochemically: following absorption, photo- optical isolation, but also for enhancement of use- chemistry can be accomplished through non-adiabatic control protections, explosive optical logic, and other conversion of the excess electronic energy into specific dual-use technologies to support initiatives concerning vibrational degrees of freedom on the explosive and DoD insensitive munitions or planned Life Extension controlling the onset of exothermic chemical reactions in Programs. the surrounding material. The third part of the proposal uses fs laser pulse shaping quantum control to optimize Progress photodissociation quantum yields and photoinitiation In our first 17 months, we have established a very exothermic chemistry. interdisciplinary research program aimed at understand- ing the photochemical reaction dynamics of several The achievement of optically driven control over the tetrazine-derivatized energetic molecules. Tetrazine is chemical reaction dynamics of energetic materials will a promising high nitrogen energetic chromophore that not only have an immediate and profound impact on has a large heat of formation and versatile chemistry 275

Page 276

for attachment to a number of ligands. Petrin tetrazine achieved with common nanosecond lasers. We were not chloride was the initial focus of detailed study, since petrin seeing signs of NO2 elimination predicted by the NA-ESMD is the trinitrate analog of PETN, the current DOE detonator predictions, but could have been missing small changes in explosive. Replacing PETN with a PETN derivatized with an the FTIR. Since many of the products of energetic material energetic chromophore was a logical first step. Tetrazine dissociation are expected to be small gas phase molecules, was also derivatized with 2 PETN molecules, with dinitroaz- we designed an experiment to measure these products as etidine (DNAZ), 2 DNAZ molecules, and nitroglycerin. More well. complex molecules with 2 DNAZ and tetrazines linked We built a photochemical mass spectrometry experiment through an azo group were tested as a means to achieve to measure relative yields of gas phase products. This extended conjugation and increase the two photon ab- experiment is essentially a windowed vacuum cell coupled sorption. High nitrogen tetrazine metal complexes were to a rough pump and turbopump with a commercial produced to enhance absorption through charge transfer. quadrupole mass spectrometer attachment. We adapted Several triazine-based energetics were also formed, which an existing vacuum cell to hold 4 KBr pellets that could be are accessible to irradiation in the near ultraviolet. irradiated individually in vacuum while pumping and moni- A number of new experimental capabilities were de- toring the mass spectrum. The new capability to measure veloped to enhance the level of detail and the speed of gas phase photoproducts proved decisive in determining characterization of the photoactive energetic materials. when the chemistry was localized on the chromophore and One of the initial objectives was the capability to screen when energy transfer led to chemistry on the conventional the numerous materials the synthesis team had produced energetic moiety. to determine which were worth scaling up to larger quanti- ties. We determined that photochemical bleaching through Future Work laser irradiation in an FTIR would allow the use of 10-20 Synthesis: Synthesis efforts will focus on pi extended mg of material to answer many of the relevant questions: conjugated systems in an effort to increase cross sectional How photoactive are the HEs? What is the chemistry areas of the molecule. Efforts will also focus on polycyclic produced? What type of laser irradiation will they be ac- materials in an effort to increase polarizability. We will also cessible to? Do the properties determined match those begin to scale-up molecules we have created in FY14. expected by the synthesis team or predicted by the theory team? Theory: Currently we are benchmarking the non-adiabatic dynamics with solvent using well studied model systems. Initial studies of FTIR photobleaching used CW lasers at In the next year this means we will be able to perform non- nonhazardous levels within the energetic material syn- adiabatic dynamics simulations of our target molecules thesis area at TA-9. While we were able to obtain pho- using solvent environment. tobleaching data, sometimes days of irradiation were necessary. Since this was not a laser laboratory, the use of Characterization: We will perform CW laser photo degrada- pulsed or high intensity lasers was not allowed. The sim- tion, fluorescence quantum yield and spectral measure- plest solution was to purchase a small FTIR and integrate it ments, and measure the optical absorption of the remain- into the laser laboratory at TA-40. This allowed us to port ing newly synthesized photoactive energetic materials to high power CW, pulsed nanosecond, and pulsed femtosec- identify the one photon active excited states and use FTIR ond lasers of various wavelengths into the FTIR. We wrote to measure the photo decomposition chemistry. We will software to quantitatively analyze the hundreds of photo- begin time resolved photochemical measurements (FSRS) bleaching spectra that could now be acquired in a day. We on the novel photoactive HE. We will begin to measure now have the capability to determine the photochemical two photon absorption spectra of photoactive HE in solu- quantum yield and basic photochemistry through FTIR of tion for comparison and validation to theory. novel photoactive explosives excited at a broad range of Conclusion wavelength and intensity conditions. As the primary goal is the design and development new The next stage of experiments focused on nanosecond energetic materials with controllable photochemical frequency-doubled Nd:YAG excitation at 532 nm that reactions, the expected outcome is a transformational would be relatively straightforward to transition to engi- and universal paradigm for the development of advanced neering applications, while also taking advantage of the future materials with structure-function relationships lead- large absorption of the tetrazine complexes at 532 nm. We ing to applications based on quantum control of chemistry. quickly determined that the quantum yield was very low, The development of photoactive energetic materials that but could be increased dramatically with intensities easily initiate through the quantum control of photochemis- 276

Page 277

try addresses future surety themes outlined in the FY12 Stockpile Stewardship and Management Plan and provides an enabling technology for future Directed Stockpile Work. The project will pave the way for new explosives that per- form like RDX yet are insensitive to unplanned stimuli. Publications Bjorgaard, J. A., K. A. Velizhanin, and Tretiak. Solvent effects in time-dependent self-consistent field methods. II. Variational formulations and analytical gradients. 2015. JOURNAL OF CHEMICAL PHYSICS. 143 (5). Bjorgaard, J. A., Kuzmenko, K. A. Velizhanin, and Tretiak. Solvent effects in time-dependent self-consistent field methods. I. Optical response calculations. 2015. JOURNAL OF CHEMICAL PHYSICS. 142 (4). Bjorgaard, J. A., Nelson, Kalinin, Kuzmenko, K. A. Velizhanin, and Tretiak. Simulations of fluorescence solvatochromism in substituted PPV oligomers from excited state molecular dynamics with implicit solvent. 2015. CHEMICAL PHYSICS LETTERS. 631: 66. Greenfield, M. T., S. D. McGrane, C. A. Bolme, J. A. Bjorgaard, T. R. Nelson, Tretiak, and R. J. Scharff. Photoactive High Explosives: Linear and Nonlinear Photochemistry of Petrin Tetrazine Chloride. 2015. JOURNAL OF PHYSICAL CHEMISTRY A. 119 (20): 4846. Hernandez, Alfonso, Nelson, Tretiak, and Fernandez- Alberti. Photoexcited Energy Transfer in a Weakly Coupled Dimer. 2015. JOURNAL OF PHYSICAL CHEMISTRY B. 119 (24): 7242. Kilina, , Kilin, and Tretiak. Light-Driven and Phonon- Assisted Dynamics in Organic and Semiconductor Nanostructures. 2015. CHEMICAL REVIEWS. 115 (12): 5929. White, A. J., V. N. Gorshkov, Tretiak, and Mozyrsky. Non-adiabatic molecular dynamics by accelerated semiclassical Monte Carlo. 2015. JOURNAL OF CHEMICAL PHYSICS. 143 (1). White, A. J., V. N. Gorshkov, Wang, Tretiak, and Mozyrsky. Semiclassical Monte Carlo: A first principles approach to non-adiabatic molecular dynamics. 2014. JOURNAL OF CHEMICAL PHYSICS. 141 (18). 277

Page 278

Materials for the Future Directed Research Continuing Project Multiferroic Response Engineering in Mesoscale Oxide Structures Dmitry A. Yarotski 20140025DR Introduction Benefit to National Security Missions Multiferroic (MF) materials exhibit strongly coupled This project aims to develop basic principles to control magnetic and ferroelectric (FE) orders and promise macroscale multiferroic (MF) functionality in meso- transformative technologies in energy, security, and scale composites through tuning the interactions on information processing. However, the inability to synthe- the nanoscale. This research directly supports all three size epitaxial materials with the desired three-dimension central themes of the Materials for the Future pillar in (3D) structure and to probe the emergent properties harnessing defects, interfaces and electromagnetic field resulting from the atomic-to-mesoscale evolution poses extremes to control collective behavior in MF materi- significant challenges to modern condensed matter als. Our approach also addresses LANL’s priorities in theory in providing a predictive description of the strong ‘meso’-science development and practical realization of coupling among spin, charge, orbital, and lattice degrees design principles towards both tunable and controlled of freedom. As a result, new materials discovery has functionality of complex materials. The proposed effort often relied on serendipitous findings rather than on aligns well with the MaRIE vision of ‘material co-design’ scientific principles underpinning the magnetoelectric and develops new integrated capabilities in synthesis, (ME) functionality. We target to use a systematic co-de- theory, and ultrafast x-ray characterization for MaRIE. sign approach to the discovery of MFs based on a closed This project further leverages the user programs at LCLS, synthesis-characterization-modeling loop. We emphasize APS, CINT, NHMFL, and the Lujan Center in accord with the rapid feedback from the experimental validation to the LANL institutional priority in supporting national user theoretical prediction and vice versa. Our approach is facilities. Materials with tunable functionality are en- enabled by a combination of recent LANL breakthroughs abling components in energy, sensing, and information in first-principles modeling of complex electron correla- technologies. Therefore, we expect this work to have tion phenomena, controllable synthesis of 3D mesoscale a direct and significant impact on near- and long-term films, and novel coherent photon probes of intrinsic LANL programs in Advanced Materials, Global Security, dynamics of competing orders in MF composites. Information Science and Technology, and Clean Energy. Built on our team’s unique expertise in theoretical Progress simulation of multifunctional metal-oxide materials, In the second year of the project we have continued a the effort starts with ab-initio identification of promi- synergistic theoretical and experimental effort to under- nent composite structures having high ME response. stand and control the magneto-electric (ME) coupling in Experimental results is then used to refine model single-phase and heterostructured multiferroic systems. Hamiltonians and to provide a better estimate for the In particular, we used first-principles density functional geometry and constituent materials, quite possibly theory (DFT) to reveal the mechanisms of magnetic cou- beyond the initial compounds and structures, for the pling across the interface between multiferroic (ferro- next optimization iteration. For a particular choice of electric (FE) and antiferromagnetic (AFM)) BiFeO3 (BFO) materials (BiFeO3:CoFe2O4 or La0.7Sr0.3MnO3:BiFeO3), and ferromagnetic (FM) La1-xSrxMnO3 (LSMO). In order we synthesize these epitaxial 3D composite films and to understand the implications of BFO pillar structure in systematically characterize their structural and physical LSMO matrix on the ME coupling, we calculated the net properties. magnetization induced in BFO and LSMO in the proxim- ity of interface. Using initially AFM ordered BFO and FM ordered LSMO, we electronically relaxed both system 278

Page 279

and found that collinear alignment of magnetization axis the magnetic, electronic and optical properties and their in both LSMO and BFO is more favorable than their per- interplay in double perovskite Bi2FeMnO6 (BFMO) by us- pendicular mutual orientation. These results explained the ing both the first-principles computation and experiment. dependence of induced magnetization in BFO on the LSMO Our ab initio calculations captured the complex exchange termination plane at the interface that was observed in interactions between Fe and Mn sites which cause ferro- our neutron scattering experiments. Moreover, this effect magnetic (anti-parallel) spin alignment on these ions. The should also lead to reduced magnetization in LSMO near predicted magnetic moment qualitatively agrees well with the interface due to spin disorder induced by the exchange Goodenough-Kanamori theory but our calculations reveal bias, which was initially confirmed in our ultrafast optical the importance of onsite (on both Mn and Co) correla- spectroscopy measurements on BFO/LSMO superlattices tions that should be accounted for using Hubbard model. and, recently in our X-Ray magnetic circular dichroism These correlations open up insulating band gap providing (XMCD) experiments on BFO/LSMO nanopillars samples. the possibility for simultaneous existence of ferrimagnetic Dependence of this effect on the pillar density and struc- and FE ordering in BFMO. Initial modeling results disagreed ture will provide essential insight for engineering ME with our XMCD spectra and measurements of magnetic coupling through manipulation of interfacial exchange bias moments of Fe and Mn ions which provided immediate in multiferroic composites. feedback for improving our first-principle computation routines. Better match to experimental data could be We have also studied the problem of strain and electronic achieved by accounting for complex interplay between the localization/delocalization in disordered ferromagnetic distortions and the disorder of the cation (Fe, Mn) sites as systems. We constructed a numerical algorithm to calcu- well as the magnetic structure. Improved accuracy of the late the conductance within the transfer matrix method. calculations will enable better understanding and design By targeting the structural disorder relevant to BFO or MgO of interfaced systems: recently, we used it to demonstrate pillar in LSMO matrix structure, we have found that with dependence of net magnetization on strain in BFMO a given strength of an impurity barrier, there is a critical nanostructures which can be tested by varying substrate thickness, below which the low-energy electronic wave materials in XMCD and magnetization measurements. functions become localized. This work gives a correspond explanation to the experiment which measure the trans- Future Work port property for the LSMO system with a critical thick- Theory: We will continue to use density functional the- ness. To test this prediction, we created a series of LSMO ory (DFT) to advance our microscopic understanding of films of varying thickness and nanopillar samples where magneto-electric coupling mechanisms in multiferroic effective “thickness” of LSMO was continuously varied by materials. In particular, we will use DFT to calculate the changing LSMO fraction in LSMO/MgO composite from XMCD response as a function of strain and validate them nanoscale pillar to mesoscale matrix. Magnetization and by comparison with experiments. With our first-principles XMCD (using synchrotron facilities) measurements have approach, we will perform ionic relaxation in BFMO and clearly shown continuous degradation of conductivity and calculate ferroelectric moment using state-of-the-art FM response with increased strain in LSMO. These results Berry’s phase approach. We will continue investigating the should aid in optimization of multiferroic functionality of mechanisms and optimal conditions for coupling FE/AFM/ the nanocomposites by controlling strain field distribu- FM orders across the interfaces. Depending on the mis- tion across the interfaces. Time-resolved second harmonic match of the lattice parameters across the interface, we spectroscopy has also demonstrated charge accumulation will employ either a coherent DFT framework or incoher- at the interfaces of BFO/LSMO nanopillars which leads to ent molecular dynamical simulations to study the interface enhancement of FE response in BFO. This effect is thought structure-function relationship. Finally, we will study the to be relevant to the predicted electron localization and multiferroic functionality in composites within our effec- will be more thoroughly investigated in the following ex- tive model aimed at finding design principles for raising the perimental/modeling effort. figure of merit. In parallel to optimizing composite multiferroic architec- Synthesis: We will use laser molecular beam epitaxy to tures, we are continuously searching for new single-phase synthesize both 3D multiferroic architectures with de- materials with enhanced magnetization, polarization, sired structural properties. Vertically aligned epitaxial and ME coupling that can be used in nanostructure fab- BiFeO3:CoFe2O4, BaTiO3:CoFe2O4 and BiFeO3/La(1-x) rication. For example, double perovskite Bi2FeMnO6 Sr(x)MnO3 films with appropriate volume ratios as de- (BFMO) is a potential candidate for the highly sought termined by the theoretical modeling will be extensively room-temperature multiferroic system. We have studied investigated. To establish the baseline, we will also grow 279

Page 280

and study the layered structures and single layer films, e.g. [(SnSe)1.15]m(VSe2)1, m = 1, 2, 3, and 4. 2014. BFMO. To systematically manipulate the strain state in the Chemistry of Materials. 26: 2862. nanocompistes, we will further vary the film thickness and Chen, , Weigand, Bi, Zhang, Lu, Dowden, J. L. MacManus- substrate materials. Driscoll, Wang, and Jia. Evolution of microstructure, strain and physical properties in oxide nanocomposite Characterization: The films synthesized will be systemati- films. 2014. SCIENTIFIC REPORTS. 4. cally studied not only by the conventional characterization techniques but also by advanced probing tools such as Chen, A.. Role of vertical strain and microstructure neutron diffraction and magnetoelectric measurements on magnetoresistance in vertically aligned at high magnetic fields. We will apply time-integrated and nanocomposites. Invited presentation at Materials time-resolved optical tools, covering the terahertz (THz) Science &Technology 2014. (Pittsburgh, PA, October through the x-ray frequencies, to investigate the mecha- 12-16, 2014). nisms underlying magnetoelectric functionality in these Chen, A. P., W. Zhang, J. L. MacManus-Driscoll, H. Wang, materials. We will also perform static x-ray magnetic circu- M. R. Fitzsimmons, and Q. X. Jia. Role of interfaces lar/linear dichroism (XMCD/XMLD) experiments in order to on competing interactions of ferroic films. Invited directly probe spin alignment on particular magnetic ions presentation at 2015 MRS Fall meeting. (Boston, as a function of composition. Massachusetts, 29 Nov - 4 Dec, 2015). Conclusion Chikara, S., J. Singleton, N. Lee, H. Y. Choi, Y. J. Choi, and V. We expect to identify an optimal geometry of 3D com- S. Zapf. Electric polarization observed in single crystals of multiferroic Lu2CoMnO6. Applied Physics Letters. posites and develop initial concepts underpinning their functionality. Continuous iterations through co-design loop Choi, E. M., T. Fix, A. Kursumovic, C. J. Kinane, D. Arene, will result in better understanding of the ME functionality S. L. Sahonta, Z. Bi, J. Xiong, L. Yan, J. S. Lee, H. in 3D structures. We anticipate to delivering a complete Wang, S. Lamgridge, Y. M. Kim, A. Y. Borisevich, I. theoretical suite for reliable prediction of MF functionality MacLaren, Q. M. Ramasse, M. G. Blamire, Q. X. Jia, in any possible 3D geometry and materials combination. and J. L. MacManus-Driscoll. Room temperature When combined with refined 3D synthetic and charac- ferrimagnetism and ferroelectricity in strained, thin terization tools, we expect to develop a complete set of films of BiFe0.5Mn0.5O3. 2014. Advanced Functional capabilities for accelerated materials discovery that will be Materials. 24 (47): 7478. applicable to a broad class of correlated electron materials, Dowden, P. C., Z. Bi, and Q. X. Jia. Method for controlling well beyond the MF heterostructures. energy density for reliable pulsed laser deposition of thin films. 2014. Rev. Sci. Instrum.. 85: 025111. Publications Ahmed, , La-o-Vorakiat, Salim, Y. M. Lam, E. E. M. Chia, and Fang, , Tai, Deng, Wu, Ding, J. u. n. Sun, and E. K. H. Salje. Zhu. Optical properties of organometallic perovskite: Fe-vacancy ordering in superconducting K1-xFe2- An ab initio study using relativistic GW correction and ySe2: first-principles calculations and Monte Carlo Bethe-Salpeter equation. 2014. EPL. 108 (6). simulations. 2015. SUPERCONDUCTOR SCIENCE & TECHNOLOGY. 28 (9). Ahmed, T., A. Chen, D. A . Yarotski, S. A. Trugman, Q. Jia, and J. X, Zhu. Magnetic, Electronic and Optical Haeusler, , Atkins, Falmbigl, S. P. Rudin, Neumann, and Properties of Double Perovskite Bi2FeMnO6. To appear D. C. Johnson. Insights from STEM and NBED studies in Scientific Reports. into the local structure and growth mechanism of misfit layered compounds prepared using modulated Ahmed, T., A. Chen, Q. X. Jia, D. A . Yarotski, and J. X . Zhu. reactants. 2015. ZEITSCHRIFT FUR KRISTALLOGRAPHIE. Site dislocation between Fe and Mn sites in double 230 (1): 45. perovskite Bi2FeMnO6: A theoretical explanation for XMCD anomaly. Physical Review Letters. Haeusler, , Atkins, Falmbigl, S. P. Rudin, Neumann, and D. C. Johnson. Insights from STEM and NBED studies Ahmed, T., Q. X. Jia, M. R. Fitzsimmons, Z. Bi, S. Rudin, and into the local structure and growth mechanism of J. X. Zhu. First-principles Study of Magnetic Coupling misfit layered compounds prepared using modulated at the BiFeO3-La(1-x)SrxMnO3 Hetero-structured reactants. 2015. ZEITSCHRIFT FUR KRISTALLOGRAPHIE. Interface. Physical Review B. 230 (1): 45. Atkins, R., M. Dolgos, A. Fiedler, C. Grosse, S. Fischer, S. P. Jain, , Wang, Roldan, Glavic, Lauter, Urban, Bi, Ahmed, Rudin, and D. C. Johnson. Synthesis and Systematic Zhu, Varela, Q. X. Jia, and M. R. Fitzsimmons. Synthetic Trends in Structure and Electrical Properties of 280

Page 281

magnetoelectric coupling in a nanocomposite Bi2FeMnO6. Invited presentation at Electronic multiferroic. 2015. SCIENTIFIC REPORTS. 5. Structure Approaches and Applications to Quantum Matter. (Santa Fe, 20 May, 2015). Jia, Q. X.. Effect of Interfaces on Ferroic Properties of Composite Films. Invited presentation at International Rudin, S. P., and D. C. Johnson. Density functional theory Symp. on Emerging Multifunctional & Bio-Directed calculations of the turbostratically disordered Mater.. (San Antonio, TX, 10 - 111, Oct.). compound (SnSe)(1+y)(VSe2)(n). 2015. PHYSICAL REVIEW B. 91 (14): 144203. Kogoj, J., L. Vidmar, M. Mierzejewski, S. A. Trugman, and J. Bonca. Thermalization after photoexcitation from the Sandberg, R., B. Mafarland, R. Prasankumar, G. Rodriguez, perspective of optical spectroscopy. Physical Review A. Chen, Q. X. Jia, A. Taylor, D. Yarotski, A. Ried, and H. Letters. Ohldag. Probing multiferroic materials with soft X-ray spectroscopy and imaging. Invited presentation at Lee, , S. A. Trugman, C. L. Zhang, Talbayev, X. S. Xu, SLAC User Workshop. (Menlo Park, CA, Oct. 10, 2014). S. -. Cheong, D. A. Yarotski, A. J. Taylor, and R. P. Prasankumar. The influence of charge and magnetic Sheu, Y. M., S. A. Trugman, L. Yan, J. Qi, Q. X. Jia, and A. order on polaron and acoustic phonon dynamics in J. Taylor. Polaronic transport induced by competing LuFe2O4. 2015. APPLIED PHYSICS LETTERS. 107 (4). interfacial magnetic order in a La0.7Ca0.3MnO3/ BiFeO3 heterostructure. 2014. Phys. Rev. X. 4: 021001. Lee, , Sangle, Lu, Chen, Zhang, J. S. Lee, Wang, Jia, and J. L. MacManus-Driscoll. Novel Electroforming-Free Sheu, Y. M., S. A. Trugman, L. Yan, Q. X. Jia, A. J. Taylor, and Nanoscaffold Memristor with Very High Uniformity, R. P. Prasankumar. Ultrafast optical manipulation of Tunability, and Density. 2014. ADVANCED MATERIALS. magnetoelectric coupling at a multiferroic interface. 26 (36): 6284. 2014. Nature Communications. 5: 5832. Li, , R. u. n. Zhao, Tang, Chen, Zhang, X. i. n. Lu, Wang, Sing, S., J. H. Haraldsen, J. Xiong, E. M. Choi, P. Lu, D. Yi, X. and H. a. o. Yang. Vertical-Interface-Manipulated D. Wen, J. Liu, H. Wang, Z. Bi, P. Yu, M. R. Fitzsimmons, Conduction Behavior in Nanocomposite Oxide Thin J. L. MacManus-Driscoll, R. Ramesh, A. V. Balatsky, Films. 2014. ACS APPLIED MATERIALS & INTERFACES. 6 J. X. Zhu, and Q. X. Jia. Induced magnetization in (8): 5356. La0.7Sr0.3MnO3/BiFeO3 superlattices. 2014. Phys. Rev. Lett.. 113: 047204. Li, , W. e. i. Zhang, L. e. Wang, Gu, Chen, R. u. n. Zhao, Y. a. n. Liang, Guo, Tang, Wang, Jin, Wang, and H. a. o. Tai, , C. -. J. Wang, M. J. Graf, Zhu, and C. S. Ting. Emergent Yang. Vertical Interface Induced Dielectric Relaxation in topological mirror insulator in t(2g)-orbital systems. Nanocomposite (BaTiO3)(1-x):(Sm2O3)(x) Thin Films. 2015. PHYSICAL REVIEW B. 91 (4): 041111. 2015. SCIENTIFIC REPORTS. 5. Tai, Y. Y., H. Choi, T. Ahmed, C. S. Ting, and J. X. Zhu. Li, B. o., Pan, Tai, M. J. Graf, Zhu, K. E. Bassler, and C. Edge states and local electronic structure around an S. Ting. Unified description of superconducting adsorbed impurity in a topological superconductor. pairing symmetry in electron-doped Fe-based-122 Nature Physics. compounds. 2015. PHYSICAL REVIEW B. 91 (22). Talbayev, , Lee, S. A. Trugman, C. L. Zhang, S. -. Cheong, Mun, E. D., Chern, Pardo, Rivadulla, Sinclair, H. D. Zhou, R. D. Averitt, A. J. Taylor, and R. P. Prasankumar. V. S. Zapf, and C. D. Batista. Magnetic Field Induced Spin-dependent polaron formation dynamics in Transition in Vanadium Spinels. 2014. PHYSICAL Eu0.75Y0.25MnO3 probed by femtosecond pump- REVIEW LETTERS. 112 (1). probe spectroscopy. 2015. PHYSICAL REVIEW B. 91 (6). Mun, E., D. F. Weickert, J. Kim, B. L. Scott, C. Miclea, Talbayev, , Lee, S. A. Trugman, C. L. Zhang, S. -. Cheong, R. Movshovich, J. Wilcox, J. Manson, and V. S. R. D. Averitt, A. J. Taylor, and R. P. Prasankumar. Zapf. Partially disordered antiferromagnetism and Spin-dependent polaron formation dynamics in multiferroic behavior in a frustrated Ising system, Eu0.75Y0.25MnO3 probed by femtosecond pump- CoCl2-2SC(NH2)2. To appear in Physical Review B. probe spectroscopy. 2015. PHYSICAL REVIEW B. 91 (6). Pan, , Li, Tai, M. J. Graf, Zhu, and C. S. Ting. Evolution of Trugman, S. A .. High temperature transport in the quasiparticle states with and without a Zn impurity in Holstein model. Invited presentation at Conference on doped 122 iron pnictides. 2014. PHYSICAL REVIEW B. Dynamics of Quantum Many Body Systems far from 90 (13). Equilibrium. (Krvavec, Slovenija, 15 December, 2014). Rudin, S.. Magnetic Structure, Cation Disorder, Octahedral Xiong, J., V. Matias, B. W. Tao, Y. R. Li, and Q. X. Jia. Rotations, and Antiferroelectric Distortions in Ferroelectric and ferromagnetic properties if epitaxial 281

Page 282

BiFeO3-BiMnO3 films on ion-beam-assisted deposited TiN buffered flexible Hastelloy. 2014. J. Appl. Phys.. 115: 17D913. Zapf, V.. Multiferroic behavior investigated at high magnetic fields: CdV2O4. Invited presentation at Telluride Workshop on oxide materials. (Telluride, CO, June, 2015). Zhang, , Chen, J. i. e. Jian, Zhu, L. i. Chen, Lu, Jia, J. L. MacManus-Driscoll, Zhang, and Wang. Strong perpendicular exchange bias in epitaxial La0.7Sr0.3MnO3:BiFeO3 nanocomposite films through vertical interfacial coupling. 2015. NANOSCALE. 7 (33): 13808. Zhang, , Li, Lu, Fan, Su, Khatkhatay, Chen, Jia, Zhang, J. L. MacManus-Driscoll, and Wang. Perpendicular Exchange-Biased Magnetotransport at the Vertical Heterointerfaces in La0.7Sr0.3MnO3:NiO Nanocomposites. 2015. ACS APPLIED MATERIALS & INTERFACES. 7 (39): 21646. Zhang, W., A. Chen, Z. Bi, Q. X. Jia, J. L. MacManus-Driscoll, and H. Wang. Interfacial coupling in heteroepitaxial vertically aligned nanocomposite thin films: from lateral to vertical control. 2014. Curr. Opin. Solid State Mater. Sci. . 18: 6. Zhang, W., J. Jian, A. Chen, L. Jiao, F. Khatkhatay, L. Li, F. Chu, Q. X. Jia, J. L. MacManus-Driscoll, and H. Wang. Strain relaxation and enhanced perpendicular magnetic anisotropy in BiFeO3:CoFe2O4 vertically aligned nanocomposite thin films. 2014. Appl. Phys. Lett.. 104: 062402. Zhao, , Tai, and C. S. Ting. Phase diagram of the isovalent phosphorous-substituted 122-type iron pnictides. 2015. PHYSICAL REVIEW B. 91 (20): 205110. 282

Page 283

Materials for the Future Directed Research Continuing Project Exploring Mechanisms of Catalysis on Plutonium Surfaces (U) Marianne P. Wilkerson 20140051DR Introduction layer and the interaction of those defects with bound Characterization of the molecular rearrangements that intermediates. occur on metal surfaces can provide insight into impor- Experiments on surrogate materials with the FTIR reflec- tant catalytic processes. Uncovering the role a metal tance accessory revealed that Polarization Modulation surface plays in catalyzing molecular transformations Infrared Refection Adsorption Spectroscopy (PM-IRRAS) is critical to obtaining a thorough understanding of its is preferred to properly investigate the surface chemistry performance in a given environment. High-resolution of actinide surfaces. The system is now fully configured imaging (Atomic Force Microscopy/Scanning Tunneling to run in this mode. A plutonium readiness review with Microscopy), Secondary Ionization Mass Spectrometry, the Radiation Protection group was completed in June, and Fourier Transform Infrared Absorption Spectroscopy and the instrument is undergoing final testing with sur- are being applied to characterize molecular and mate- rogates prior to accepting the first plutonium sample, rials states on Pu surfaces, providing unprecedented scheduled for the month of July. information about their roles in materials/environment interactions. We are using electronic structure and Various aspects of the ultra-high vacuum scanning atomistic methods to model reaction mechanisms and tunneling microscopy (UHV-STM) have been tested, thermodynamic stability of various materials. Further- modified, repaired, or replaced as needed. A service more, a comprehensive organometallic chemistry study technician is scheduled to visit LANL the week of June is being conducted to provide complementary informa- 22nd to address a lingering noise problem. Once the tion about metal-molecule binding and factors control- final issues have been addressed, initial experiments on ling reactivity. surrogates will begin in the 3rd quarter of FY15 and will move quickly on to thin film Pu oxides and, ultimately, Benefit to National Security Missions plutonium metal. The response of Pu materials to molecular environments is a key component of laboratory missions related to Final facility tie-in of the glovebox that supports the Stockpile Stewardship. Furthermore, the development Atomic Force Microscopy (AFM) is scheduled for comple- of the materials science and chemistry of the actinides, tion in June. The glovebox is now leak tight, and boasts particularly Pu, will advance and strengthen our position an oxygen and water concentration measured at the as the national Plutonium Center of Excellence. ppm level. During the final preparation phase, extensive analysis of surrogate films have aided in the selection Progress of the best AFM tip for plutonium. The first ever AFM The fiscal year was marked by several major project images of Pu metal surfaces will be analyzed in FY15 and accomplishments. A series of 7 at.% Ga delta-stabilized published the following year. plutonium samples, specifically designed to accommo- date each analytical system, have been cut and prepared Experiments with the Time-of-Flight Secondary Ion Mass in PF-4. All experimental systems are advancing towards Spectrometer have identified several interesting surface plutonium readiness, while others are awaiting receipt species following gas exposures on plutonium metal of the first Pu samples. The modeling efforts have made samples. As this work marks the first known application substantial gains as well, with the advancement of DFT of ToF-SIMS to a plutonium metal surface, time has been calculations that explore the role of defects in the oxide spent developing peak identification protocols. Isotopi- cally labeled gas exposures will commence this year to 283

Page 284

compliment the infrared work. ments of UO2 and U3O8 PAD films, images of PuO2 PAD films will be completed, and additional Pu coupons will be The goal of the synthetic chemistry work is to develop imaged, both before and subsequent to various exposures. molecular complexes that exhibit the relevant structural characteristics, react them with small molecules, and Surface measurements on Pu6Fe material will be con- determine the molecular transformations that occur us- ducted using the Secondary Ion Mass Spectrometry (SIMS). ing molecular and spectroscopic techniques. To this end, Experiments will include exposures to isotopically labelled the synthetic chemistry work has successfully prepared gases to identify dynamic reaction products. organometallic cerium complexes and exposed them to Polarization Modulation Infrared Reflection Absorption gases of interest. Solid state crystal structures of these Spectroscopy Synthesis work will be accelerated and reac- molecules have been obtained to confirm atom connectiv- tions to prove the ability of cerium to facilitate the reverse ity and infrared (IR) spectra have been recorded. These water-gas shift reaction will be carried out to help deter- IR spectra show good agreement to the theoretical spec- mine the mechanism of this reaction. New complexes tra provided by T-1. Current experiments are focused and intermediates will be synthesized and reaction with on translating both the experimental and theoretical IR appropriate gases will be performed. spectroscopy to well defined metallic surfaces and translat- ing the molecular synthesis to plutonium to determine the Modeling work will focus on further validation of model- effect on the IR stretching frequency for further spectro- ing methods via the use of all-electron based codes that scopic comparison. include relativistic effects and hybrid exchange-correlation functionals. Vibrational spectra will be calculated which The theory and modeling team has made progress in will allow our findings to be directly compared with experi- understanding the structure of the Pu oxide surface as well mental data, aiding in model validation and experimental as its interactions with molecules of interest. In particu- interpretation of data. Other surface defect structures lar, we have identified that high concentrations of oxygen (step edges, kinks, etc.) will also be considered. monovacancies are thermodynamically stable under UHV experimental condition on the <110> surface, and must Conclusion be considered as possible actives sites for catalysis. Vibra- tional calculations of various moieties are complete, and We will gain a deeper understanding of the fundamental will be compared with IR spectroscopy measurements. connections that exist between the composition, valence We are working to explore other possible reaction sites/ state and structural properties of a given defect center and surfaces including Pu6Fe, <111> PuO2 surfaces with and its propensity for catalysis through the following areas: without defects, Pu2O3 surfaces, and Fe/C defected PuO2 • Reaction mechanisms of Pu-catalyzed molecular trans- <110> surfaces. Identification of these sites is a key step formations. for making further progress in understanding surface • Thermodynamic models reactions. Finally, great progress has been made in using • High-resolution imaging of Pu surfaces. Density Functional Theory (DFT) modeling to interface syn- • Surrogate Pu coordination complexes with hitherto thesized molecular structures with bound intermediates to unexplored properties. their spectroscopic signatures for understanding relevant reactions on molecular sites. Synthetic cerium compounds and their IR spectra have been identified and anticipated Publications shifts when moving to Pu are in the process of being deter- Beaux, M. F., T. Durakiewicz, K. S. Graham, J. N. Mitchell, mined. Such findings and their agreements with experi- S. Richmond, E. D. Bauer, D. P. Moore, F. J. Freibert, P. mental spectra will aid in the interpretation of future data H. Tobash, and J. A. Kennison. Electronic structure of and give confidence in DFT predictions for molecular site/ polycrystalline d-Pu metal: a review of photoemission adsorbate interactions. spectra interpretations. 2014. In Pu Futures - The ‘Science’ 2014. (Las Vegas, NV, 7-12 Sept. 2014). , p. Future Work CMP.9. Las Vegas NV: American Nuclear Society. The Scanning Tunneling Microscope (STM) system will Hernandez, S. C., D. S. Schwartz, C. D. Taylor, and A. K. commissioned for plutonium during FY15. The first ever Ray. Ab initio study of Ga-Stabilitzed d-Pu bulk and STM images of Pu metal surfaces will be acquired. An surfaces. 2014. In Pu Futures - The ‘Science’ 2014. (Las additional gas manifold system will be constructed and Vegas, NV, 7-12 Sept. 2014). , p. M&MS.15. Las Vegas integrated into the system for gas exposure experiments. NV: American Nuclear Society. Building on Atomic Force Microscope (AFM) measure- 284

Page 285

Hernandez, S. C., D. S. Schwartz, C. D. Taylor, and A. K. Ray. time-of-flight SIMS. 2014. In Pu Futures - The ‘Science’ Ab initio study of gallium stabilized d-plutonium alloys 2014. (Las Vegas, NV, 7-12 Sept. 2014). , p. SS&C.8. Las and hydrogen-vacancy complexes. 2014. Journal of Vegas NV: American Nuclear Society. Physics: Condensed Matter. 26 (3): 235501/1. Wagner, G. L., M. T. Paffett, K. D. Rector, B. L. Scott, and Hernandez, S. C., M. P. Wilkerson, and M. N. Huda. M. P. Wilkerson. Characterization of products from Understanding oxygen adsorption on 9.375 at.% Ga- hydrolysis of UF6. 2014. In Pu Futures - The ‘Science’ stabilized delta-Pu(111) surface: A DFT study. 2015. 2014. (Las Vegas, NV, 7-12 Sept. 2014). , p. CCCC.4. Las Journal of Alloys and Compounds. 653: 411. Vegas NV: American Nuclear Society. Hernandez, S. C., T. J. Venhaus, and M. N. Huda. Atomic oxygen adsorption on 3.125 at.% Ga stabilized delta-Pu(111) surface. 2015. Journal of Alloys and Compounds. 643: 253. Hernandez, S. C., and C. D. Taylor. Density functional theory studies on atomic adsorptions on Ga stabilized d-Pu (111) surfaces. 2014. In Materials Research Society 2014 Spring Meeting and Exhibit. (San Franscisco, CA, 21-25 April 2014). , p. S6.05. San Franscisco CA: Materials Research Society. Holby, E. F.. Oxygen vacancies in PuO2 (110) surfaces via density functional theory. 2014. In Pu Futures - The ‘Science’ 2014. (Las Vegas, NV, 7-12 Sept. 2014). , p. III. Las Vegas NV: American Nuclear Society. Pugmire, A. L., C. H. Booth, T. J. Venhaus, and D. L. Pugmire. Structural insights into the oxide formed during the room temperature corrosion of Plutonium. 2014. In Pu Futures - The ‘Science’ 2014. (Las Vegas, NV, 7-12 Sept. 2014). , p. SS&C.7. Las Vegas NV: American Nuclear Society. Richmond, S., P. H. Tobash, and D. Schwartz. The synthesis of Pu6Fe from plutonium deuteride and iron powders. 2014. In Pu Futures - The ‘Science’ 2014. (Las Vegas, NV, 7-12 Sept. 2014). , p. M&MS.13. Las Vegas NV: American Nuclear Society. Schwartz, D. S., S. Richmond, C. D. Taylor, A. I. Smith, and A. L. Pugmire. Hydrogen effects in Pu-Ga alloys: defects and thermodynamics. 2014. In Pu Futures - The ‘Science’ 2014. (Las Vegas, NV, 7-12 Sept. 2014). , p. II. Las Vegas NV: American Nuclear Society. Smith, A. I., K. L. Page, S. Richmond, J. Siewenie, T. A. Saleh, M. Ramos, and D. S. Schwartz. Local structural investigation of the Pu-7at%Ga using neutron total scattering. 2014. In Pu Futures - The ‘Science’ 2014. (Las Vegas, NV, 7-12 Sept. 2014). , p. M&MS.7. Las Vegas NV: American Nuclear Society. Sutton, A. D., and M. P. Wilkerson. Catalysis with cerium organometallic complexes. 2014. In Pu Futures - The ‘Science’ 2014. (Las Vegas, NV, 7-12 September 2014). , p. CCCC.11. Las Vegas NV: American Nuclear Society. Venhaus, T. J., and D. P. Moore. Analysis of Pu surfaces with 285

Page 286

Materials for the Future Directed Research Continuing Project Mesoscale Materials Science of Ductile Damage in 4 Dimensions: Towards the Computational Design of Damage-Tolerant Materials Ricardo A. Lebensohn 20140114DR Introduction structure and ductile damage in selectively prepared The failure of structural materials has a significant polycrystalline aggregates. We will follow the volumetric impact on vast sectors of the economy, including en- deformation through time, i.e. in four dimensions. The ergy, transportation and defense. The costs arise both proposed integration will allow us to achieve the full from rare catastrophic events to the more mundane inversion of this 4-D data, and extract the controlling expenses of over-engineering or preventive part replace- aspects of ductile damage. ment. Consequently, there has been enormous effort in Benefit to National Security Missions the past to gain fundamental understanding of failure mechanisms and thus enable the development of more DOE’s Office of Basic Energy Sciences (OBES) has recently reliable components. highlighted the area of mesoscale science as the next grand-challenge in Materials Science. The two top capa- Most structural materials are polycrystalline aggregates, bility gaps for this were defined as the seamless integra- in which the constituent crystals are irregular in shape, tion of theory, modeling and simulation with synthesis have anisotropic mechanical properties, and contain a and characterization, and the dynamic characterization variety of defects. The deformation of these heteroge- of mesoscale phenomena. This project directly address- neous materials results in very dynamic and complicated es both topics. Success of this paradigm will establish responses. While centuries of metallurgical experience LANL’s leadership in this growing initiative, expected to and post-failure analysis have given us insight into gener- be central to future DOE programs. al aspects relating material processing and performance, our failure models remain empirically calibrated because The development of improved predictive damage mod- we have yet to achieve a thorough understanding of els that incorporate mesostructural sensitivity will be of the controlling processes at the scale of the materials’ interest to several Defense Programs (DP) stakeholders. heterogeneity, i.e. the mesoscale. Extension to dynamic regimes will especially extend its application to DP programs. The MaRIE campaigns will This project addresses one of the most difficult outstand- be a beneficiary of experience in the development of in- ing problems in Materials Science: the development situ diffraction and associated data processing. Applica- of a predictive, microstructure-sensitive ductile failure tion of the integrated approach to material development model. This will be achieved by integrating time-resolved should attract interest from advanced manufacturing three-dimensional (3-D) microstructural characterization initiatives. of polycrystalline materials with state-of-the-art model- ing and advanced data analysis tools, to ensure maximal Progress and optimal data extraction. Increasingly complex model Experimental materials will be used to develop this integrated capabil- • Using powder metallurgy, we have manufactured ity, which will be then tested and applied to the predic- Cu-W and Cu-Nb plates with fully dense structure tive design of damage-tolerant microstructures. [1]. We intend to draw upon existing High Energy Diffraction • We are executing a General User Proposal (GUP) in Microscopy (HEDM) techniques, integrating and combin- Advanced Photon Source (APS) and requested six ing them with appropriate modeling and data analysis days for the fall 2015 cycle. formulations, to discover relationships between micro- 286

Page 287

• We have submitted and executed a beam-time propos- Future Work al in Cornell High Energy Synchrotron Source (CHESS) HEDM measurements on Cu-W samples performed during to collect tomographic, nf- and ff-HEDM data in Cu-W FY15 are being analyzed and modeled and new experi- (December 2014). We are scheduled to perform similar ments on pure Cu and Cu-Nb will be conducted in APS and measurements on Cu-Nb samples in July 2015. CHESS, respectively, to measure damage evolution in poly- crystalline aggregates synthesized to maximize the extrac- Data Analysis tion of relevant microstructural information. The measured • We demonstrated the feasibility and accomplished the microstructures are being used as input for our improved registration of orientation field and tomography in a spectral model, whose predictions and underlying physics post-shocked Cu sample [9]. will be tested against the measured damage evolution. • Compressive sensing techniques were used to solve The major bottleneck of the proposed work, which lies on efficiently the inverse lattice orientation problem in the difficulty in inverting the low-quality diffraction data far-field high-energy diffraction microscopy. We are arising from heavily deformed and damaged materials, will extending this approach to solve both orientation and continue to be addressed. All four proposed strategies we spatial grain information [7]. be continued: 1) collecting the synchrotron data simulta- neously in tomography, orientation and stress mapping • C++/MPI framework is in development to estimate modes; 2) using correlations between multiple snapshots model parameters directly from measurements. during the deformation process; 3) employing direct simu- Strategies to compute reduced order models are being lations; and 4) using compressive sensing and hierarchical implemented, yielding encouraging results in terms of sparse representations. small approximation error and computation time [8]. We will capture microstructural and micromechanical in- • 100TB data storage was acquired, and we created the formation correlated with ductile damage, obtaining HEDM “meso4d” Unix group in CCS’s cluster Darwin to store snapshots during quasi-static tensile tests performed on experimental raw data, install and run data reduction notched bars. The measurements will be performed on a and simulation codes. All data reduction codes IceN- volume of ~0.5 mm3 in the notch region of the samples. In ine (nf-HEDM), HEXRD (ff-HEDM), FABLE (box-beam the third year, we will study polycrystalline Cu-Nb compos- HEDM) and Recon (tomo) have being successfully ites (in CHESS), and we will repeat HEDM measurements in installed in Darwin. pure Cu (in APS). Modelling 3-D tomographic reconstructions will allow us to detect • The feasibility of using the viscoplastic FFT model void formation at early stages of deformation. The grain (VPFFT) with direct input from HEDM data [2] was orientation data will inform us of plastic deformation at the demonstrated. grain scale, in particular near voids, while the local stress • The dilatational viscoplastic FFT model (DVPFFT) was mapping will provide information on stress concentrations, extended to consider void growth in polycrystalline which presumably drive void nucleation, and stress relax- materials and applied to the interpretation of micro- ations once voids have formed. structural effects on porosity evolution of shocked Cu. Conclusion • A moment-based method for calculating interface Our overarching goal is to be able to formulate a quantita- normals in voxelized 3D images was developed. The tive, microstructure-sensitive, experimentally-validated normal determination is required to investigate micro- mesoscale model of ductile damage, paving the way to- structural effects on void nucleation [3]. Simulations of wards a truly predictive failure model. This will also result existing HEDM experiments in post-shocked Cu are in in the advancement of novel experimental and analytical progress [9]. techniques that can then be extended and applied to other problems of strategic interest for the Laboratory and the • The FFT-based formulation was extended to consider Nation. microstructural effects on stress concentration promot- ing crack propagation [4, 5], and we have implemented Publications a non-local EVPFFT formulation [6]. Anglin, B. S., R. A. Lebensohn, and A. D. Rollett. Validation of a numerical method based on Fast Fourier Transforms for heterogeneous thermoelastic materials by comparison with analytical solutions. 2014. 287

Page 288

Computational Materials Science. 87: 209. M. Suter, and A. D. Rollett. In-situ observation of bulk 3-D microstructure evolution of polycrystalline copper Bingert, J. F., R. M. Suter, J. Lind, S. F. Li, R. Pokharel, and using synchrotron radiation. 2015. International C. P. Trujillo. High-Energy diffraction microscopy Journal of Plasticity. 67: 217. characterization of spall damage. 2014. In Dynamic Behavior of Materials, Volume 1. , p. 397. : Springer Rovinelli, A., R. A. Lebensohn, and M. D. Sangid. Influence International Publishing. of microstructure variability on short crack growth behavior. 2015. Engineering Fracture Mechanics. 138: Chen, C. F., D. Dombrowski, R. A. Lebensohn, B. Clausen, R. 265. Forsyth, and R. Pokharel. Processing and consolidation of copper for mesoscale materials science of ductile Sidky, E. Y., R. Chartrand, J. M. Boone , and X. Pan. damage in four dimensions. 2014. In 2014 World Constrained TpV minimization for enhanced Congress on Powder Metallurgy & Particulate exploitation of gradient sparsity: application to Materials (PM 2014) . (Orlando, FL, 18-22 May 2014). , CT image reconstruction. 2014. IEEE Journal of p. 0946. Orlando, FL: APMI International. Translational Engineering in Health and Medicine. 2: 1. Landecker, W., R. Chartrand, and S. DeDeo. Robust sparse Stein, C. A., A. Cerrone, T. Ozturk, S. Lee, P. Kenesei, H. coding and compressed sensing with the difference Tucker, R. Pokharel, J. Lind, C. Hefferan, R. M. Suter, A. Map. 2014. In European Conference on Computer R. Ingraffea, and A. D. Rollett. Fatigue crack initiation, Vision. (Zurich, 6-12 September, 2014). Vol. 8691, slip localization and twin boundaries in a nickel-based p. 315. Zurich: Springer International Publishing superalloy. 2014. Current Opinion in Solid State and Switzerland. Materials Science. 18 (4): 244. Lebensohn, R. A., J. P. Escobedo, E. K. Cerreta, Dennis- Subedi, S., R. Pokharel, and A. D. Rollett. Orientation Koller, C. A. Bronkhorst, and J. F. Bingert. Modeling gradients in relation to grain boundaries at varying void growth in polycrystalline materials. 2013. Acta strain level and spatial resolution. 2015. Materials Materialia. 61 (18): 6918. Science and Engineering A. 638: 348. Lebensohn, R. A., and A. Needleman. Numerical implementation of non-local polycrystal plasticity using Fast Fourier Transforms. Journal of the Mechanics and Physics of Solids. Lebensohn, R. A., and R. Pokharel. Interpretation of microstructural effects on porosity evolution using a combined dilatational/crystal plasticity computational approach. 2014. Journal of the Minerals, Metals and Materials Society (JOM). 66 (3): 437. Lieberman, E., A. D. Rollett, R. A. Lebensohn, and E. M. Kober. Calculation of grain boundary normals directly from 3-D microstructural images. 2015. Modelling and Simulation in Materials Science and Engineering. 23: 035005. Ozturk, T., C. Stein, R. Pokharel, C. Hefferan, H. Tucker, S. Jha, R. John, R. A. Lebensohn, P. Kenesei, R. M. Suter, and A. D. Rollett. Simulation domain size requirements for elastic response of 3-D polycrystalline materials. To appear in Modelling and Simulation in Materials Science and Engineering. Pokharel, R., J. Lind, A. K. Kanjarla, R. A. Lebensohn, S. F. Li, Kenesei, R. M. Suter, and A. D. Rollett. Polycrystal Plasticity: Comparison Between Grain-Scale Observations of Deformation and Simulations. 2014. Annual Review of Condensed Matter Physics. 5: 317. Pokharel, R., J. Lind, S. F. Li, P. Kenesei, R. A. Lebensohn, R. 288

Page 289

Materials for the Future Directed Research Continuing Project Aging in Delta Plutonium Alloys: A Fundamental Approach Franz J. Freibert 20150057DR Introduction of radiogenic damage process in δ-Pu exploring materi- The mission of LANLs’ Stockpile Stewardship program als of controlled thermal histories, Ga compositions and is to assess aging of our nation’s stockpile materials. alpha-decay dose. Both α-decay and thermodynamics are powerful physi- Benefit to National Security Missions cal processes in δ-Pu, with potential to couple strongly to phase changes in δ-Pu. Understanding the relative The single most important mission of LANLs’ Stockpile contributions of these processes is necessary to identify Stewardship program is to assess aging of our nation’s important aging mechanisms and accurately extrapolate stockpile; nonetheless, there are certain fundamental induced effects on δ-Pu properties and performance science questions that are not addressed in the pro- over decades. This project, founded in Defects and gram. Direct reuse of δ-Pu components expected to be Interfaces theme of the Materials for the Future Focus in service for many decades requires that physics and Area, is focused on the development of a mechanistic engineering assessments have a firm scientific basis. multi-scale understanding and control of defects and Both a-decay and thermodynamics are powerful physi- He bubbles, intrinsic and enhanced, across significant cal processes in δ-Pu, with potential to couple strongly length, time and temperature scales of δ-Pu aging. to phase changes in δ-Pu. Understanding the relative contributions of these processes and their coupling are The goal is to quantify and understand the radiogenic necessary to identify the important aging mechanisms changes in δ-Pu induced by the drivers of He ingrowth, and to accurately extrapolate induced effects on δ-Pu defect accumulation and δ-Pu phase instability as properties and performance margins over decadal time determined by consensus of state-of-the-art experimen- scales. Current physics-based lifetime estimates for US tal, computational and modeling tools. We will work pits are approximately 100yr. These estimates were to understand radiogenic defect accumulation as the developed within a Quantifying Margins and Uncertain- mechanism of ambient temperature lattice swelling over ties (QMU) formalism for primary physics performance. a 1-2yr. transient ingrowth period; to control numbers However, the physics models for the QMU determina- and configurations of various defects separately from He tions do not contain fundamental physics and therefore bubbles and determine the influence radiogenic defects are not predictive. The scientific work of understanding have on electronic and structural properties from nano- the effects of Pu aging must continue; otherwise today’s meter to bulk spatial scale correlated with δ-Pu phase solution to a Life Extension Program (LEP) problem might instability; and to understand defect accumulation in- induce two or more unforeseen problems in the future. teraction with He ingrowth as functions of experimental variables sufficient to induce and control void swelling. Progress The project team now includes a joint LANL/LBNL post- This work will lay the fundamental groundwork for a Pu doctoral researcher in support of collaborations at the aging strategy by revealing the importance, rate, and Stanford Linear Accelerator. scale of radiogenic changes in δ-Pu with exceptionally sensitive measurements, capable of showing aging ef- Local Structure and Nascent State Defect Evolution fects in real time over a wide temperature range, with The project team has made progress towards the goal of computational approaches describing defect structures using extended X-ray absorption fine structure (EXAFS) and their relative stability. This approach supports spectroscopy, as a probe of the local structure around unprecedented advances in understanding the physics defect sites. A proposal has been written for synchrotron 289

Page 290

time to conduct extended aging and annealing experi- products, the thermodynamic stability of Pu-Ga phases, ments, and was favorably received by the peer review and the introduction of radiation damage. By changing committee. Anticipate receiving time to conduct these temperature (4K-700K), Ga concentration, sorting out the experiments on the next available scheduling period. To relative importance of the three effects, and the tempera- further leverage the unique capabilities of the synchrotron tures at which they are most active. Expect that at tem- experiments, are redesigning sample holders to make re- perature ranges greater than 450K the face centered cubic sistivity measurements simultaneously with X-ray measure- of lattice Pu-Ga alloys can no longer support lattice dam- ments. Design, prototyping, testing, and safety approval age and defect accumulation. This assertion is based upon processes are moving forward on schedule. Anticipate be- thermal expansion measurements, which exhibit the final ing able to make a full-scale test of the system in advance stages of radiogenic defect annihilation with temperature. of next synchrotron experiment. Progress has been made Expect to validate this assertion within the next month of in the data analysis and modeling of EXAFS data. A new continuous elastic modulus monitoring at increasingly high manuscript is in preparation, using previously unpublished temperatures. data, detailing a method to quantify the amount of radia- Modeling and Theory tion damage present in a material relative to it’s undam- Development of new molecular dynamics models that aged state. Expect this to be submitted to a peer-reviewed account for known aging behavior of plutonium alloys. journal for publication by the end of the summer, and to Revised an existing meso-scale (100-10,000 nm) model use the developments therein to analyze the results of to account for radiation-damage-suppression terms from future experiments. a reaction-rate model. Estimation of interstitial defect Thermally Activated Defect Evolution and Kinetic transport rates as a function of gallium concentration in Behavior plutonium alloys. Developed a global damage-to-defect- Physical property measurements and characterization to-effects pathway mapping as a means to semi-quantify measurements have been made on delta Pu alloys. These the balance between damage accumulation and damage include low temperature heat capacity and electrical resis- annihilation, thus implement a means to discriminate the tivity to determine cryogenic aging behavior of the materi- strength of the various radiogenic/phase stability mecha- als. This currently involves allowing the samples to accu- nisms including daughter product ingrowth, irradiation mulate damage at 4 K for a short time and then heating damage and phase stability. Also, of concern is sustaining a the samples in-situ in the cryostat at various rates in order LANL molecular dynamics modeling capability given limited to investigate the annealing out of damage. Preparing for resources, so training early career staff in radiation damage the upcoming beamline experiments, new equipment is molecular dynamics modeling. being assembled in order to perform isochronal annealing experiments. A high resolution resistance bridge and pre- Future Work amp will be used in conjunction with the physical property The goal of this project is to quantify and understand the measurement system. System will be sent to Stanford for radiogenic changes in δ-Pu induced by the drivers of He in- the fall beamline experiments in order to measure elec- growth, defect accumulation and δ-Pu phase instability as trical resistivity while performing EXAFS. Furthermore, determined by consensus of state-of-the-art experimental, characterization of the samples before and after various computational and modeling tools. This project, founded treatments, using X-ray diffraction technique (phase identi- in the Defects and Interfaces theme of the Materials for fication, any changes in peak position, shape and intensity, the Future Focus Area for LDRD Strategic Investment, is identification of any additional phases and/or impurities). focused on the development of a mechanistic multi-scale understanding and control of inhomogeneities (i.e., defects Time Dependent/Defect Induced Thermodynamic and He bubbles), intrinsic and enhanced, across significant Properties Changes length, time and temperature scales that govern δ-Pu ag- Using high-resolution, high temperature resonant ultra- ing. sound spectroscopy (RUS), measuring in real time at the 10^-7 level changes in the elastic moduli of unalloyed Pu We propose implementing the following objectives by test- and delta Pu alloys to understand the relative contribu- ing the associated hypotheses. tions of radiogenic daughter product ingrowth, phase stability, and radiation damage. With the ability to observe Objective #1: Understand radiogenic defect accumulation changes in elastic moduli to 100 parts per billion, observ- as the mechanism of ambient temperature lattice swelling ing changes in the elastic moduli of Pu and Pu-Ga alloys in over a 1-2yr. transient ingrowth period. real time. Changes due to ingrowth of radioactive decay Hypothesis #1: The transport mechanisms observed dur- 290

Page 291

ing ambient temperature initial swelling transient are the effects of δ-phase Pu-Ga alloys . 2015. In NUCAR 2015 . same mechanisms that eventually will lead to void swell- (Mumbai, India, 9-13 February 2015). , p. 42. Mumbai: ing. Evolution and accumulation of ambient temperature Bhabha Atomic Research Centre. defects are controlled by thermal vacancy migration and Freibert, F. J., J. N. Mitchell, D. S. Schwartz, and A. Migliori. recovery processes occur as annealing of a defected state Radiogenic-thermally coupled lifetimes for defects of against defect stabilizing factors. aged δ-phase Pu-Ga alloys. 2015. In Plutonium Futures 2014, The Science. (Las Vegas, NV, 7-12 September Objective #2: Understand defect accumulation interaction 2014). , p. 220. La Grange Park, IL: American Nuclear with He ingrowth as functions of experimental variables Society. sufficient to induce and control void swelling. Maiarov, B. A., J. B. Betts, F. J. Freibert, and A. Migliori. Hypothesis #2: Point defect (i.e., interstitials, vacancies, Real-time aging studies and some high temperature clusters, etc.) structures anneal and thermally activate dis- measurements of Ga-stabilized δ-phase 239Pu . 2015. tinctly and separately from those associated with He (i.e., FY15 Annual Report. bubbles and voids). However, with experimentally provid- Mitchell, J. N., F. J. Freibert, T. E. Mitchell, and D. S. ed vacancies to be absorbed by bubbles, void swelling can Schwartz. Aging and the δ → α´ transformation in Pu- be induced in δ-Pu. Ga alloys . To appear in PTM: International Conference on Solid-Solid Phase Transformations in Inorganic Conclusion Materials. (Whistler, Canada, 28 June-3 July 2015). This project will form the basis for pit lifetime estimates that are physically sound and advance the understanding of fundamental radiogenic processes in δ-Pu. Because the focus of Defense Science Campaigns is the development of science impacting stockpile performance, an understand- ing of radiogenic effects in δ-Pu will impact experiment implementation. Because the focus of Directed Stockpile Work is the function of stockpile technology, aging indica- tors of performance impact could influence programmatic decisions on Pit Reuse and Lifetime Extension Programs and bound thermo-mechanical processing and supporting technological development for better definition of perfor- mance margins and uncertainties Publications Conradson, S. D., Bock, J. M. Castro, D. R. Conradson, L. E. Cox, Dmowski, D. E. Dooley, Egami, F. J. Espinosa- Faller, F. J. Freibert, A. J. Garcia-Adeva, N. J. Hess, E. K. Holmstroem, R. C. Howell, Katz, J. C. Lashley, R. J. Martinez, D. P. Moore, L. A. Morales, J. D. Olivas, R. A. Pereyra, Ramos, S. P. Rudin, and P. M. Villella. Nanoscale heterogeneity, premartensitic nucleation, and a new plutonium structure in metastable delta fcc Pu-Ga alloys. 2014. PHYSICAL REVIEW B. 89 (22). Conradson, S. D., Bock, J. M. Castro, D. R. Conradson, L. E. Cox, Dmowski, D. E. Dooley, Egami, F. J. Espinosa- Faller, F. J. Freibert, A. J. Garcia-Adeva, N. J. Hess, Holmstroem, R. C. Howell, B. A. Katz, J. C. Lashley, R. J. Martinez, D. P. Moore, L. A. Morales, J. D. Olivas, R. A. Pereyra, Ramos, J. H. Terry, and P. M. Villella. Intrinsic Nanoscience of delta Pu-Ga Alloys: Local Structure and Speciation, Collective Behavior, Nanoscale Heterogeneity, and Aging Mechanisms. 2014. JOURNAL OF PHYSICAL CHEMISTRY C. 118 (16): 8541. Freibert, F. J.. Current research into the self-irradiation 291

Page 292

Materials for the Future Directed Research Continuing Project A New Approach to Mesoscale Functionality: Emergent Tunable Superlattices Marc Janoschek 20150082DR Introduction controlled functionality and developing multi-functional The future availability of multifunctional materials is materials to transform structural and functional per- essential for the country to remain globally competitive formance, as well as tunable and emergent properties. and to ensure our energy and defense security. We must In particular, the mesoscale magnetic whirls (so called move beyond the current age of ‘informed serendip- “skyrmions”) that will be investigated in this proposal ity’ towards a new era of ‘materials by design’ enabling hold great potential for multi-functional remote sens- controlled functionality that would allow rapid creation ing, memory storage applications, and new computing of novel applications required for an ever-changing technologies. In each case future devices using this tech- world. Important technological functionalities are a nology will have ultra-low power consumption. Those consequence of mesoscale objects that respond to an new sensing, storage and computing devices are directly external stimulus; e.g. magnetic memories and sensors relevant for a new generation defense and prevention are built upon the ability to switch mesoscale domains technologies, as well as improving information science with magnetic fields. Here the mesoscale describes ob- and technology. The development of new mesoscale jects that have typical size in between the atomic-scale computational capabilities within this project that will (i.e. sub-nanometer) and the macroscale (~1 mm). The be used to design these materials will additionally en- critical step for mastering the designed functionality will able molecular dynamics calculations on the mesoscale, be to achieve control of materials directly on this meso- which is of broad relevance to the lab’s material’s strat- scale. This is the potential promised by the recent dis- egy, and stockpile stewardship , weapon’s research, and covery of a mesoscale object, a magnetic vortex called the study of matter at extremes in general. The meso- a “skyrmion” that emerges in magnetic materials . The scale imaging and synthesis methods developed in this stability of these topological objects makes them behave program will be crucial to study further functional mate- as particles when driven by external stimuli. Put differ- rials directly on the mesoscale as required for stockpile ently, skyrmions are the new atoms of the mesoscale stewardship, weapons research, and new materials for world. Similarly, just as different combination of atoms example required for the automotive, renewable energy, produce materials with different properties, different and defense industries. arrangements of skyrmions lead to distinct functional- Progress ities through coupling to electrical currents, magnetic and electrical fields, and temperature gradients. The Obtaining an improved understanding of the functional- resulting mesoscale architecture build from skyrmions ity that derives from magnetic mesoscale architecture is moreover clean, tunable, and movable, qualities that based on skyrmion building blocks represents a multi- are vital for achieving enhanced macroscale functional- scale material science problem. Notably the competition ity. This project will exploit the promising properties of of two or more atomic-scale magnetic interactions leads those mesoscale magnetic whirls to find design prin- to the emergence of skyrmions, and controls their size ciples for new functional materials, such as ultra-low- and arrangement on the mesoscale. Those mesoscale power high-density memory materials. parameters, in turn dictate the functionality of this assembly of skyrmions on the macroscale. In order to Benefit to National Security Missions tackle this problem our team has tailored and improved Our project responds specifically to the over-arching our computing, synthesis and imaging capabilities to grand challenge of LANL’s materials strategy. It addresses investigate these materials on each of those three length the priorities of realizing design principles towards scales in the first year of this project. These capabilities 292

Page 293

have been used to initiate a tightly integrated modeling- rently, this efforts yields 1-2 materials every week, and making-measuring loop that will be used through the has recently lead to the discovery of skyrmions in a so far entire project to create design principles for mesoscale unreported cobalt-zinc-manganese alloy at a temperature magnetic mesoscale architecture based on skyrmions. of 260 K (30 K below room temperature), making this an ideal material for applications (manuscript in preparation). To make progress on the understanding of the atomic- Moreover, promising preliminary data on a related mate- scale interactions that control the emergence of magnetic rial with a different cobalt-zinc-manganese concentra- mesoscale architecture based on skyrmions, we have car- tion suggests that skyrmions may even exist above room ried out first-principle electronic structure calculations for temperature. This new family of cobalt-zinc-manganese prototypical skyrmion materials from which the strength compounds thus represents a prototypical test bed for es- of these interactions may be determined theoretically. tablishing design principles and functionality for skyrmion Simultaneously, we have used neutron spectroscopy on materials for the second year of our project. the prototypical skyrmion material manganese silicide to measure the strength of these interactions in order to In summary, our project is making outstanding progress in verify our calculations (manuscript submitted and under all efforts (modeling-making-measuring) on all the relevant review at Phys. Rev. Lett.). This effort is currently applied to length scales that define the properties of skyrmion archi- more materials. tecture. Our team is therefore well setup for the second year of the project that will verify basic design principles We have also used and expanded a new algorithm devel- on the choice of magnetic ions, local environments and oped at Los Alamos to predict the emergence of various overall crystallographic symmetry that yield skyrmion lat- arrangements of skyrmion architectures directly on the tices for distinct functionalities. mesoscale based on atomic-scale interactions in a given material. This effort has already led to novel and exciting Future Work theoretical results that suggests a new mechanism for the Understanding the functionality of skyrmions represents a emergence of skyrmion architecture based on a type of multi-scale material science problem that stretches from atomic-scale magnetic interactions that typically exist in the atomic- over the meso- to the macroscale. During the rare earth and actinide materials, which are of interest to first year, our team has tailored and improved our com- Los Alamos. In addition, this type of new magnetic calcula- puting, synthesis and imaging capabilities to investigate tion has shown that magnetic anisotropy that can be tuned these skyrmion materials on each of those three length either via crystal chemistry or external parameters such as scales. Using these tools we have already identified a new strain, controls the symmetry of the arrangement of sky- class of skyrmion cobalt-zinc-manganese alloys that show rmions in materials (manuscript published: S.-Z. Lin et al., skyrmions near room temperature and thus represents Phys. Rev. B 91, 224407 (2015)). To test this experimentally, an optimal test-bed for establishing design principles and we have built a prototype of a strain cell that can be used functionality for skyrmion materials for the second year in combination with small angle neutron scattering (SANS). of our project. In the second year, we will use these new SANS is able to probe the symmetry of a skyrmion assem- materials to establish and test basic principles on choice bly directly on the mesoscale. First tests with the strain cell of magnetic ions, local environments and overall crystallo- prototype have shown how to improve it for experiments graphic symmetry that yield skyrmion lattices with dif- with skyrmion materials and a full-blown strain cell is cur- ferent sizes and symmetries that can be used for distinct rently being assembled for measurements in the second functionalities. From our modeling efforts, we know that year of this project. in particular magnetic anisotropy is a useful parameter to tune skrymions. Magnetic anisotropy can be tuned via To address the challenge of functionality on the mac- strain. Using a strain cell for small angle neutron scattering roscale, we are currently testing real time imaging of sky- that we have developed in the first year, we will test the rmion movement and rotation using transmission electron design rules postulated by theory. Our new magnetic me- microscopy. The expected velocity of skyrmions under the soscale algorithm developed at LANL has further suggests application of external stimuli such as electrical current or a new mechanism for the emergence of skyrmion architec- temperature and magnetic field gradients is also simulated ture based on a type of atomic-scale magnetic interactions using so-called micro-magnetic calculations that have the that typically exist in rare earth and actinide materials, skyrmion size as input parameters. which are of interest to Los Alamos. Our synthesis effort will therefore make materials to test these design rules for Finally, the first year has also been used to ramp up our rare earth and actinide based skyrmions. Finally, we will efforts in synthesizing known skyrmion materials as well use real time imaging of skyrmion movement and rotation as candidate materials identified through theory. Cur- 293

Page 294

as function of external stimuli such as electrical current or and C. D. Batista. Vortex crystals with chiral stripes in temperature and magnetic field gradients using transmis- itinerant magnets. Physical Review Letters. sion electron microscopy. The results will be compared Saxena, A., and S. Z. Lin. Skyrmions in Functional Materials. with micro-magnetic calculations that have the skyrmion To appear in Integrated Ferroelectrics. size as input parameters. Conclusion In pursuing our objective, we will not only pioneer a new direction for material science at LANL, as well as for the international community, and respond directly to national, Laboratory and DOE priorities, but have the real potential of transforming the research on functional materials. Using our fully-integrated modelling-making-measuring loop we will find “design principles” for various mesoscale sky- rmion architectures, which will immediately yield designed functionality via their unique properties. In particular, we expect to identify magnetic mesoscale architecture opti- mized for new memory or sensing applications. Publications Choi, H., S. Z. Lin, and J. X. Zhu. Density functional theory study of skyrmion pinning by atomic defects in MnSi. Physical Review B. Kugler, M., G. Brandl, J. Waizner, M. Janoschek, R. Georgii, A. Bauer, K. Seemann, A. Rosch, C. Pfleiderer, P. Böni, and M. Garst. Band structure of helimagnons in MnSi resolved by inelastic neutron scattering. 2015. Physical Review Letters. 115 (9): 097203. Lin, S. Z.. Generation of skyrmions by chopping magnetic chiral stripe domains with an electric current. Physical Review Letters. Lin, S. Z., A. Saxena, and C. D. Batista. Skyrmion fractionalization and merons in chiral magnets with easy-plane anisotropy. 2015. Physical Review B. 91: 224407. Lin, S. Z., and A. Saxena. Non-circular skyrmion and its anisotropic response in thin films of chiral magnets under tilted magnetic field. 2015. Physical Review B. 92: 180401. Lin, S. Z., and A. Saxena. Dynamics of Dirac strings and monopole-like excitations in chiral magnets under a current drive. Physical Review Letters. Lin, S. Z., and J. Zang. Skyrmion dynamics in chiral magnets. Skyrmion: the topological structures, magnetic and transport properties, and potential applications. Lin, S. Z., and S. Hayami. Ginzburg-Landau theory for skyrmions in inversion-symmetric magnets with competing interaction. Physical Review B. Ozawa, R., S. Hayami, K. Barros, G. W. Chern, Y. Motome, 294

Page 295

Materials for the Future Directed Research Continuing Project Meso-Photonic Materials for Tailored Light-Matter Interactions Houtong Chen 20150109DR Introduction questions regarding key technological gaps in photonics. Building on our internationally recognized expertise in The development of compact, lightweight, flexible, and metamaterials, we propose to tackle key science issues integrated optical elements and optoelectronic devices underpinning some “grand challenge” scale problems will impact Threat Reduction and Global Security appli- in photonics by developing transformational meso-pho- cations, such as flat lens antennas and focal plane array tonic materials based on two core innovations: a meta- detectors for communications, imaging, and sensing, molecule concept and the integration of functional ma- such as effluent detection, particularly for space and terials into meso-photonic structures. Meta-molecules satellite sensing of nuclear nonproliferation. Indeed, in are a collection of meta-atoms (sub-wavelength metal/ a recent presentation DoD has listed metamaterials and dielectric resonators) with tailored internal interactions, plasmonics as one of its six high priority S&T areas. The enabling designer light coupling and propagation. Inte- proof-of-concept meso-photonic solar thermophotovol- gration of functional materials (semiconductors, complex taics will create a pathway that greatly impacts renew- oxides, graphene, and transition metal dichalcogenides) able solar energy harvesting for our national energy results in tunable and reconfigurable photonic function- security, a core mission of LANL and DOE. The exquisite alities, enables novel optoelectronic architectures, and control of photonic and electronic density of states in provides a means to exploit the greatly enhanced light- meso-photonic cavities will be of great interest to BES matter interactions to study a wealth of basic physics for fundamental studies of light-matter interactions phenomena. enabling emergent meso-scale material properties and functionalities. The proposed research aligns with New In our joint experiment-theory effort, we will demon- Mexico’s Technology21 Roadmap for Science and Tech- strate the unusual power of this meso-photonic plat- nology, and contributes to the potential New Mexico form in two connected focus areas: (a) Photonic phe- Advanced Photonics Hub. It will also prepare us for the nomena and functionalities enabled by meta-molecule Mesoscale Science, Advanced Manufacturing, and Na- structures. We will demonstrate a host of photonic tional Photonics Initiatives. functionalities, including complete polarization control and arbitrary wavefront shaping based on the meta- Progress molecule concept, and develop an efficient proof-of- • A broadband omnidirectional absorber was dem- concept solar thermophotovoltaics device for energy onstrated based on a metallic metasurface archi- harvesting; (b) Enhanced meso-photonic functionalities tecture, which accomplished greater than 90% through integration of functional materials into meta- absorptance in the visible and near-infrared range of molecule structures. We will realize real-time control of the solar spectrum, and exhibited low emissivity to photonic functionalities, develop meso-photonic cavities prevent thermal radiation losses at mid- and far-in- for greatly improved optoelectronics, and explore the frared wavelengths. A manuscript has been submit- fundamental physics of strong light-matter interactions ted to Nano Letters. in low dimensional quantum materials. • New metasurfaces were designed, fabricated, and Benefit to National Security Missions characterized for optical antireflection coatings op- Our proposed work supports the Emergent Phenomena erating at terahertz and mid-infrared frequencies. A central theme in the Materials for the Future science paper has been published in Applied Physics Letters. pillar. This project addresses some grand challenge 295

Page 296

• We finished the design and numerical simulations of ture) of solar thermophovoltaics that can operate at metasurface structures for a flat lens operating at THz elevated temperatures. frequencies. • Build the characterization capability for metasurface • A plasmonic metasurface structure was demonstrated; intermediate structures to be used in solar thermo- it accomplished the simultaneous manipulation of photovoltaics. polarization and phase of the transmitted light, which • Explore new metasurface structures that enable multi- was further used to demonstrate a broadband near- band or broadband antireflection performance. perfect anomalous refraction and the generation of a radially polarized beam. A paper has been published in • Fabricate and characterize the performance of a meta- Advanced Functional Materials. surface flat lens to focus broadband terahertz radia- tion. • An ultra-broadband free-space terahertz modulator based on a semiconductor-integrated metasurface • Integrate semiconductors 2D Dirac materials to enable was demonstrated. Experimental validations confirm a polarization switching/modulation. bandwidth of at least 100%, spanning 0.5-1.5 THz with -10 dB modulation depth. A paper has been published • Fabricate and characterize graphene-enabled indepen- in Applied Physics Letters. dent tuning of a dual-band metasurface absorber. • Through numerical simulations, we demonstrated that • Model heat transfer and structural disorder of the the integration of graphene into metasurface struc- metasurface absorber/emitter within the intermediate tures enabled independent tuning of two absorption structure. bands of metasurface structure by application of volt- age biases to tune the graphene Fermi level. A manu- • Integrate 2D materials for enhanced photodetection. script has been submitted to Scientific Reports. • Model light-matter interactions of 2D Dirac materials • We computed the radiative heat transfer between two within metasurface cavities. sheets of 2D Dirac materials, within the framework of the local approximation for the optical response of Conclusion these materials. A paper has been published in Journal The anticipated deliverables are: a) Novel photonic func- of Physics: Condensed Matter. tionalities from meso-photonic structures; b) Proof-of-con- cept highly efficient meso-photonic intermediate structure • The spontaneous emission rate of a two-level quantum for solar thermophotovoltaic energy harvesting; c) Devel- emitter near a graphene-coated substrate under the opment of modeling and simulation capabilities for design- influence of an external magnetic field was investi- ing and understanding light-matter interactions in meso- gated. We demonstrated that the application of the photonic materials; d) Enhanced photonic functionalities magnetic field can substantially increase or decrease enabled by functional materials integration; and e) Funda- the decay rate. Our findings strongly suggest that a mental studies of strong light-matter interaction between magnetic field could act as an efficient agent for on- meta-molecules and two dimensional electronic quantum demand, active control of light-matter interactions materials. Our proposed work supports the Emergent in graphene at the quantum level. A paper has been Phenomena central theme in the Materials for the Future submitted to Physical Review Letters and deposited to science pillar, and will have great impact on renewable arXiv: 1506:02176. energy, communications, imaging and sensing. • We are processing the requirement of a solar simula- Publications tor, a vacuum chamber, and a temperature control Azad, A. K., W. J. M. Kort-Kamp, M. Sykora, N. R. Weisse- system for the characterization of the metasurface Bernstein, T. S. Luk, A. J. Taylor, D. A. R. Dalvit, and H. T. intermediate structure for solar thermophotovoltaics. Chen. Broadband metasurface absorber – Toward solar thermophotovoltaics. Scientific Reports. Future Work • Fabricate new metasurface structures for broadband Bartolo, N., F. Intravaia, D. A. R. Dalvit, and R. Messina. meso-photonic absorbers and narrowband emitters, Non-Equilibrium Casimir-Polder plasmonic interactions. Physical Review A. and explore the use of refractory plasmonic materials to construct the key component (intermediate struc- Chen, H. T.. Semiconductor activated terahertz 296

Page 297

metamaterials. 2015. Frontiers of Optoelectronics. 8 Yang, L., N. A. Sinitsyn, W. Chen, J. Yuan, J. Zhang, J. (1): 27. Lou, and S. A. Crooker. Long-lived nanosecond spin relaxation and spin coherence of electrons in Heyes, J. E., W. Withayachumnankul, N. K. Grady, D. monolayer MoS2 and WS2. 2015. Nature Physics. 11: R. Chowdhury, A. K. Azad, and H. T. Chen. Hybrid 830–834. metasurface for ultra-broadband terahertz modulation. 2014. Applied Physics Letters. 105 (18): Zhang, B., J. Hendrickson, N. Nader, H. T. Chen, and J. 181108. Guo. Metasurface optical antireflection coating. 2014. Applied Physics Letters. 105 (24): 241113. Intravaia, F., V. E. Mkrtchian, S. Y. Buhmann, S. Scheel, D. A. R. Dalvit, and C. Henkel. Friction forces on atoms Zhang, Y., T. Li, B. Zeng, H. Zhang, H. Lv, X. Huang, W. Zhang, after acceleration. 2015. Journal of Physics-Condensed and A. K. Azad. A graphene based tunable terahertz Matter. 27 (21): 214020. sensor with double Fano resonances. 2015. Nanoscale. 7: 12682. Kort-Kamp, W. J. M., B. Amorim, G. Bastos, F. A. Pinheiro, F. S. S. Rosa, N. M. R. Peres, and C. Farina. Active Zhang, Y., T. Li, Q. Chen, H. Zhang, J. F. O’Hara, E. Abele, magneto-optical control of spontaneous emission in A. J. Taylor, H. T. Chen, and A. K. Azad. Independently graphene. 2015. Physical Review B. 92: 205415 . tunable dual-band perfect absorber based on graphene at mid-infrared frequencies. Scientific Kort-Kamp, W. J. M., N. A. Sinitsyn, and D. A. R. Dalvit. Reports. Quantized beam shifts in graphene. Physical Review B, Rapid Communications. Li, J. X., S. Q. Chen, H. F. Yang, J. J. Li, P. Yu, H. Cheng, C. Z. Gu, H. T. Chen, and J. G. Tian. Simultaneous control of light polarization and phase distributions using plasmonic metasurfaces. 2015. Advanced Functional Materials. 25 (5): 704. Liang, L., M. Qi, J. Yang, X. Shen, J. Zhai, W. Xu, B. Jin, W. Liu, Y. Feng, C. Zhang, H. Lu, H. T. Chen, L. Kang, W. Xu, J. Chen, T. J. Cui, P. Wu, and S. Liu. Anomalous terahertz reflection and scattering by flexible and conformal coding metamaterials. To appear in Advanced Optical Materials. Rodriguez-Lopez, P., W. K. Tse, and D. A. R. Dalvit. Radiative heat transfer in 2D Dirac materials. 2015. J. Phys.: Condens. Matter. 27: 214019. Sinitsyn, N. A.. Exact results for models of multichannel quantum nonadiabatic transitions. 2014. Physical Review A. 90: 062509 . Sun, C., and N. A. Sinitsyn. Exact transition probabilities for a linear sweep through a Kramers-Kronig resonance. 2015. J. Phys. A: Math. Theor.. 48: 505202. Woods, L. M., D. A. R. Dalvit, A. Tkatchenko, P. Rodriguez- Lopez, A. W. Rodriguez, and P. Podgornik. A materials perspective on Casimir and van der Waals interactions. To appear in Review of Modern Physics. Wu, L., W. K. Tse, M. Brahlek, C. M. Morris, R. V. Aguilar, N. Koirala, S. Oh, and N. P. Armitage. High-Resolution Faraday Rotation and Electron-Phonon Coupling in Surface States of the Bulk-Insulating Topological Insulator Cu0.02Bi2Se3. 2015. Physical Review Letters. 115: 217602. 297

Page 298

Materials for the Future Directed Research Final Report Fighting Carbon with Carbon: All-Carbon Nanomaterial Photovoltaics Stephen K. Doorn 20130026DR Abstract Background and Research Objectives This DR effort focused on studies related to development Pioneering developments on nanographenes and com- of carbon nanomaterials (CNM) as active light harvesting positionally-defined carbon nanotubes (CNTs) indicate components in solar photovoltaics (PV). Goals were to these materials possess many properties of ideal light demonstrate the benefits these materials can bring to harvesters for thin-film photovoltaic (PV) applications. enhancing performance of solar cells and understanding These benefits include more efficient light absorption, system behaviors to enhance future performance. Our tunability of optical and electrochemical properties to areas of focus included development of two comple- match the solar emission spectrum and allow optimiza- mentary PV device architectures, supporting materials tion of interfacial charge flow between other device development (synthesis, purification, processing), fun- materials, and extremely efficient photogeneration of damental studies of materials photophysics and electro- long-lived free charges available for collection as electri- chemistry as a basis for tailoring energy and charge flow cal power [1]. A compelling opportunity thus exists to in our complex systems, and establishing and applying create new PV device types based on these materials theory to understand and guide the design of our un- as active light-harvesting components for overcoming derlying processes. A few highlights of notable results performance bottlenecks imposed by current light- include demonstration of best-in-class PV performance harvesting chromophores. By project-end we aimed to in our devices. We also provide clear direction for future harness these properties within functioning PV-devices efficiency enhancements and new roles our materials based on two platforms suitable for follow-on develop- can play in other hybrid PV systems. Our development ment. In establishing these prototype platforms, our of ultra-efficient, one-step purification processes will en- primary goals were: hance many new applications of CNM. At a fundamental • Demonstrate the benefits of carbon nanomaterials level, we have provided a first determination of critical (CNM) as active light harvesting components and energetics that drive photogeneration and collection of roles they may play in enhancing performance of electrons. We also established a first understanding of solar cells. dopant-induced emitting states behind unprecedented new photoluminescence behaviors in carbon nanotubes. • Probe and understand system characteristics from Surprising discoveries are leading to unforeseen devel- the individual materials level to integrated devices opment of electrochromic devices and novel photonic to establish new design principles for future perfor- materials for quantum information processing. The mance improvements. project has thus resulted in major leaps in scientific understanding of CNM, produced unexpected discov- Towards these goals we: Probed the fundamental eries, and new opportunities for follow-on work. The photophysical and electrochemical behaviors of CNM effort has been extremely productive, being on target for chromophores and the means to control and tune the generating ~50 publications, responsible for over 50 in- associated energy and charge flows. Established CNM vited talks, and resulting in 4 patent disclosures. Towards materials integration into functioning photovoltaic future opportunities, our work will have significant devices. Pursued these objectives within two promising impact in areas including energy harvesting, storage, and device architectures for realizing all-CNM photovoltaics: efficiency; development of new photonic and optoelec- a layered thin-film platform and as dye-sensitized solar tronic materials and devices; and for growing interest in cells, with CNM as active dye. photon-based quantum information processing. 298

Page 299

Our end results demonstrated best-in-class devices en- abled by improved materials processing. We revealed key behaviors not previously established and used the new understanding to establish design principles for further device improvements. Finally, we produced surprising discoveries and new capability that establish significant opportunity for future advances in energy harvesting, stor- age, and efficiency; thin film optoelectronics; and quantum information processing. Scientific Approach and Accomplishments Materials Development Graphene oxide was studied as an initial chromophoric material, but dropped in favor of our other candidates. However, with it we established for the first time how charge transport and dynamics in this material change with surface composition thru the first electrostatic force microscopy [2] and correlated photocurrent imaging on it. Figure 1. Illustration of effectiveness of one and two-step Direct photoluminescence microscopy also demonstrated separations. Inset shows one-step isolation of (6,5) CNTs (bottom dependence of its optical properties on surface composi- phase, purple color) from starting material. Absorption spectra tion. of (6,5) and (7,5) CNTs isolated in two steps. High spectral purity demonstrated.Color variation is shown as applied voltage is Isolating highly enriched semiconducting CNTs was critical varied. to our PV device development. We developed a single-step aqueous two-phase (ATP) separation for isolating (6,5) or (7,5) nanotubes (Figure 1) that gave dramatic improve- ments in speed, efficiency, reproducibility, and purity [3]. We also demonstrated the approach as a rapid method for benchtop removal of unwanted nanotube bundles (elimi- nating the need for expensive ultracentrufugation steps). The approach is also capable of producing long (microns) tubes essential for improving transport of energy and charge in our devices. Most importantly the ATP process- ing was shown to be compatible with our PV film develop- ment. We also established a breakthrough understand- ing of the ATP separations mechanism. Using chemical modifications and innovative direct photoluminescence imaging of separated tubes, we showed the mechanism works on diameter-dependent changes in surface composi- tion, rather than through direct interactions between the nanotube surface and separation medium (as previously Figure 2. Current-voltage curve showing 4X improvement in thought). (6,5) device performance. Thin Film Advances power conversion efficiency over past reports. That ef- Thin film device advances required both chromophore ficiencies are still in the 1% range, however, shows they are development and integration with other device materials. limited by the narrow CNT wavelength response. We next Through improved processing, film deposition, optimizing demonstrated this limitation could be overcome by gen- interlayer interactions, and tailoring of film component erating the first multi-chiral devices in which performance thicknesses, we produced CNT-based PV devices yielding was shown to be the sum of response from each chiral best in class performance for monochiral structures de- component, while extending wavelength coverage. signed from (6,5) or (7,5) nanotubes (Figure 2). The (7,5) devices showed a 4-fold improvement in performance, By making hybrid devices with other chromophores, we while the (6,5) devices yielded 4-fold improvements in also showed CNTs can act in tandem to create charge 299

Page 300

and energy cascades. This important result shows CNTs 3). These could be used to enhance charge generation have potential for enhancing performance of other PV in doped PV devices. As a result of our efforts, we now device architectures in hybrid assemblies. This finding will understand for the first time the structural and dopant form the basis for expanding the CNT role to harvest light dependences of the lifetimes and environmental depen- wavelengths at which only limited response occurs in other dences of dopant-state emission, which will guide dopant materials. development for applications [6]. Most significantly, we discovered that the dopant states can be utilized as single We have made significant advances in understanding of photon emitters at telecom wavelengths at room tempera- CNT films towards improving design principles in the areas ture [7] (Figure 3). Overcoming this long-standing chal- of effectively expanding to multiple CNT structures, film lenge in quantum optics places CNTs for the first time as morphology, and the function of other component materi- compelling sources for photonic-based quantum informa- als. In particular, rather than acting merely as a suspension tion processing. agent, we have discovered that the polymer PFO plays an active role in photogeneration of charges; a surprising find- ing that opens new routes to enhancing performance. In contrast, PFO-bpy can play a limiting role. We have shown via photoluminescence imaging of CNT surface structures that these contrasting behaviors originate in part from formation of very different surface structures. PFO forms an open structure that enhances charge transport and intertube interaction, while a more closed structure for PFO-bpy blocks such behavior. This important finding will drive new design in film morphology and has motivated our pursuit of other active polymers that can harness the full range of nanotube structures we can now access via ATP separations. Fundamental Photophysics and Lifetime Engineering This effort has yielded significant opportunity for prob- ing new exciton physics in CNTs and for developing novel techniques as probes of newly revealed behaviors. Some brief examples include Raman probing of intertube interac- tions in defined-composition CNT bundles. Our discovery of ultranarrow Raman responses superimposed on normal Figure 3. Illustration of single photon emission behavior excitation profile behavior in these bundles reveals delo- arising from carbon nanotubes doped in a solid-state process calized exciton states, useful for directing energy flow and developed in the project. Exciton localization at dopant trap sites opening a new area of study in defined structural interac- yields extended emission lifetimes and complete antibunching at tions. Single-tube optical measurements were developed room-temperature as signature of single photon emitter. for probing exciton transport in such bundles. Results have Theory shown intriguing coupling between exciton-exciton annihi- Theory has been an integral part of all the above stud- lation processes and photon emission statistics [4]. ies. In addition to developing tight-binding approaches to Our most significant fundamental breakthroughs have understanding intertube interactions, efforts at quantum come from studies of covalently doped tubes as routes to chemical modeling of our various systems have ranged improving efficiencies via extending excitation lifetimes by from enhancing LANL-developed codes for modeling using a localization strategy. Exciton localization at dopant excited-state dynamics and testing on model conjugated sites is an emerging area of CNT studies with great prom- ring systems (Figure 4), to understanding the dynamics of ise for improved quantum yields and new functionality, localized charging events. These results are essential for yet little is known about the new optical states that arise. understanding energy and charge flow in CNM devices. Our studies have revealed for the first time the electronic Capability for such studies was also greatly strengthened structure and chemical functionality of novel dopant sites through efforts at quantum-chemical modeling of the [5], definitively show exciton localization at these sites, and doped nanotubes described above [5]. Density Functional show one consequence of localization to be an unprec- Theory elucidated dopant structure, provided the basis edented 10-fold increase in excited-state lifetimes (Figure to understand the interplay of excitonic states and spa- 300

Page 301

tially localized dopant states, how these define extended mine quantum confinement effects in optical and elec- lifetimes, and provided a basis for understanding thermal trochemical properties, providing critical information for stability of these exciton traps. Deep insights into photo- tuning materials properties. These results also allowed the excited dynamics of chemically functionalized CNM thus first determination of exciton binding energies in nanog- resulted. Theory continues to provide a guide to ongoing raphenes. Finally, access to these unique hybrid materials nanographene synthesis and dopant design efforts. allowed us to make the first ever excited-state dynamics measurements of charge and energy flow in these systems, which are leading directly to new DSSC design principles. In particular, inefficient charge injection into TiO2, in competition with fast optical relaxation, was revealed as the primary performance bottleneck in these systems. The results point towards development of improved adsorp- tion chemistry, introduction of fixed dipole moments and reducing exciton binding energies as key synthesis targets for improvements. As a surprise development, our spectroelectrochemical studies resulted in the discovery of electrochromism in these materials [9] (Figure 5): the ability to change color under applied voltage. Electrochromic materials are of significant interest for color display technology and as a basis for smart windows for enabling advances in energy Figure 4. Quantum chemical calculations of ground (S1), and efficiency. Nanographene size tunability provides a full excited state (S2, S3, and S1*) exciton wavefunctions for various color palette to the effect. Our first generation device per- model cycloparaphenylene [CPP] ring structures as CNT surro- gates. Localization in S1* yields high quantum yields for photolu- formance in areas such as coloration efficiency, contrast, minescence. switching time and cost are as good or better than for existing devices based on inorganic materials. DSSC Cells Nanographene molecules were studied as light harvesting chromophores in dye-sensitized solar cells (DSSCs). Our success in this area followed directly from the enabling breakthrough development of an in-situ synthesis ap- proach that overcame severe barriers to integration of these chromophores into the parent TiO2 charge-accepting structures [8]. The key innovation was to only perform the final synthesis step after adsorption of precursors to the TiO2 support. This strategy completely overcame previous loading limits to give a 10-fold improvement in PV re- sponse for these systems. More importantly, this break- through also led to our ability to probe key behaviors of nanographenes towards developing strategies for further improvements. Figure 5. Absorbance vs. time-plot demonstrating breakthrough discovery of electro-chromism in nanographenes. Color variation In situ synthesis was extended to nanographene deposition is shown as applied voltage is varied. on transparent electrode surfaces. This allowed us to de- termine for the first time the critical values of their oxida- In support of DSSC efforts, effective solid-state electrolytes tion and reduction potentials [9]. These values are essen- were generated via incorporation of ionic liquids into sol- tial for engineering appropriate driving forces for charge gels via a microwave driven process. Anticipate replace- transport in PV cells and in developing potential battery ment of liquid cells will enhance DSSC lifetimes. Exciting applications as well. These results were produced from an side-development showed solid-state electrolyte can form innovative spectroelectrochemical technique developed basis for graphene supercapacitors, showing highest en- within the project that overcame lack of energy resolution ergy storage densities to date. plaguing previously existing techniques. Size control over the nanographene synthesis allowed us to directly deter- 301

Page 302

Summary program). In particular, our efforts in doped nanotubes are In summary, this project was highly successful in producing drawing in large numbers of new users. breakthrough discoveries and new understanding across areas spanning CNM materials development, unveiling of References new photophysical behaviors, and exciting new applica- 1. Arnold, M. S., J. L. Blackburn, J. J. Crochet, S. K. Doorn, tions unforeseen at project start. The primary efforts on PV J. G. Duque, Mohite, and Telg. Recent developments in device development produced best-in-class performance the photophysics of single-walled carbon nanotubes and clear strategies for future improvements and integra- for their use as active and passive material elements tion. Our efforts also generated state-of-the-art capabilities in thin film photovoltaics. 2013. PHYSICAL CHEMISTRY for thin film processing, optoelectronic device develop- CHEMICAL PHYSICS. 15 (36): 14896. ment, optical probes, and CNM chemistry as just a few 2. Yalcin, S. E., Galande, Kappera, Yamaguchi, Martinez, K. examples. Our outstanding productivity resulted in 25 cur- A. Velizhanin, S. K. Doorn, A. M. Dattelbaum, Chhow- rently published papers, three submitted, and 20 addition- alla, P. M. Ajayan, Gupta, and A. D. Mohite. Direct al papers in preparation. The work is also responsible for Imaging of Charge Transport in Progressively Reduced over 50 invited talks as well as three invention disclosures. Graphene Oxide Using Electrostatic Force Microscopy. 2015. ACS NANO. 9 (3): 2981. Impact on National Missions The results of this effort enhance LANL’s portfolio in renew- 3. Subbaiyan, N. K., Cambre, A. N. G. Parra-Vasquez, E. H. able energy research, which must continue to grow and Haroz, S. K. Doorn, and J. G. Duque. Role of Surfactants strengthen to meet the critical challenges facing the nation and Salt in Aqueous Two-Phase Separation of Carbon in areas of climate disruption and energy security. While Nanotubes toward Simple Chirality Isolation. 2014. ACS our CNM device efficiencies remain at the 1-2% level, the NANO. 8 (2): 1619. understanding developed in this project point towards 4. Ma, , Roslyak, J. G. Duque, Pang, S. K. Doorn, Piryat- compelling roles CNM can play as components for enhanc- inski, D. H. Dunlap, and H. a. n. Htoon. Influences of ing performance of hybrid PV systems. The capabilities Exciton Diffusion and Exciton-Exciton Annihilation on and expertise developed in this project are also translating Photon Emission Statistics of Carbon Nanotubes. 2015. directly into rapid LANL development of perovskite materi- PHYSICAL REVIEW LETTERS. 115 (1). als that are pushing PV efficiency limits into commercially interesting terrain. Our nanographenes as promising elec- 5. Ma, , Adamska, Yamaguchi, S. E. Yalcin, Tretiak, S. K. trochromic materials and novel solid-state electrolytes for Doorn, and H. a. n. Htoon. Electronic Structure and high-energy density supercapacitors will play future roles Chemical Nature of Oxygen Dopant States in Carbon in energy efficiency and storage. Nanotubes. 2014. ACS NANO. 8 (10): 10782. 6. Hartmann, N. F., S. E. Yalcin, L. Adamska, E. H. Haroz, X. Beyond energy-related applications, advances from our Ma, S. Tretiak, H. Htoon, and S. K. Doorn. Photolumi- effort will also impact threat reduction needs through nescence imaging of solitary dopant sites in covalently detector and sensors development and photonics applica- doped single wall carbon nanotubes. Nanoscale. tions, all of which rely on similar photophysical processes as were studied here. Hybrid nanotube/graphene photo- 7. Ma, , N. F. Hartmann, J. K. S. Baldwin, S. K. Doorn, and detectors demonstrated in the project have the potential H. a. n. Htoon. Room-temperature single-photon gen- for enhancing spatial resolution of near-IR imaging ar- eration from solitary dopants of carbon nanotubes. rays. Newly established capabilities for optoelectronic 2015. NATURE NANOTECHNOLOGY. 10 (8): 671. device development are also enhancing LANL studies of 8. Ji, , Wu, Adamska, K. A. Velizhanin, S. K. Doorn, and 2D materials including technologically important transi- Sykora. In Situ Synthesis of Graphene Molecules on tion metal dichalcogenides. The discovery of doped CNTs TiO2: Application in Sensitized Solar Cells. 2014. ACS as room-T single photon sources at telecom wavelengths APPLIED MATERIALS & INTERFACES. 6 (22): 20473. forms a strong basis for their pursuit as photonic materials for quantum information (QI) needs. This is a growing area 9. Ji, , S. K. Doorn, and Sykora. Electrochromic Graphene of interest for global security applications and there is sig- Molecules. 2015. ACS NANO. 9 (4): 4043. nificant opportunity to grow the role of photonic based QI Publications through these materials. Strengthened capabilities for opti- Adamska, L., G. V. Nazin, S. K. Doorn, and S. Tretiak. Self- cal probes and theory developments are also enhancing trapping of charge carriers in semiconducting carbon the user program at the Center for Integrated Nanotech- nanotubes: Structural analysis. 2015. Journal of Physi- nologies (CINT, part of the DOE nanoscience user facility cal Chemistry letters. 6: 3873. 302

Page 303

Adamska, L., I. Nayyar, H. Chen, A. K. Swan, N. Oldani, S. CAL CHEMISTRY CHEMICAL PHYSICS. 17 (22): 14613. Fernandez-Alberti, S. K. Doorn, and S. Tretiak. Self- trapping of excitons, violation of the Condon approxi- Ma, , N. F. Hartmann, J. K. S. Baldwin, S. K. Doorn, and H. a. mation and efficient fluorescence in conjugated cyclo- n. Htoon. Room-temperature single-photon generation paraphenylenes. 2014. Nano Letters. 14: 6539. from solitary dopants of carbon nanotubes. 2015. NA- TURE NANOTECHNOLOGY. 10 (8): 671. Arnold, M. S., J. L. Blackburn, J. J. Crochet, S. K. Doorn, J. G. Duque, A. Mohite, and H. Telg. Recent developments Ma, , and H. a. n. Htoon. Tailoring the photophysical prop- in the photophysics of single-walled carbon nanotubes erties of carbon nanotubes by photonic nanostruc- for their use as active and passive material elements tures. 2015. MODERN PHYSICS LETTERS B. 29 (17). in thin film photovoltaics. 2013. Physical Chemistry Ma, X., J. K. S. Baldwin, N. F. Hartmann, S. K. Doorn, and H. Chemical Physics. 15: 14896. Htoon. Solid-state approach for fabrication of photo- Crochet, J. J., J. G. Duque, J. H. Werner, B. Lounis, L. Cog- stable oxygen-doped carbon nanotubes. To appear in net, and S. K. Doorn. Disorder limited exciton transport Advanced Functional Materials. in colloidal single-wall carbon nanotubes. 2012. Nano Ma, X., L. Adamska, H. Yamaguchi, S. E. Yalcin, S. Tretiak, Letters. 12: 5091. S. K. Doorn, and H. Htoon. Electronic structure and Duque, J. G., L. Oudjedi, J. J. Crochet, S. Tretiak, B. Lounis, chemical nature of oxygen dopant states in carbon S. K. Doorn, and L. Cognet. Mechanism of electrolyte- nanotubes. 2014. ACS Nano. 8: 10782. induced brightening in single-wall carbon nanotubes. Ma, X., O. Roslyak, J. G. Duque, S. K. Doorn, A. Piryatinski, 2013. Journal of the American Chemical Society. 135: D. H. Dunlap, and H. Htoon. Influences of exciton dif- 3379. fusion and exciton-exciton annihilation on photon Fagan, J. A., E. H. Haroz, Ihly, H. u. i. Gui, J. L. Blackburn, emission statistics of carbon nanotubes. 2015. Physical J. R. Simpson, Lam, A. R. H. Walker, S. K. Doorn, and Review Letters. 115: 017401. Zheng. Isolation of > 1 nm Diameter Single-Wall Car- Ma, X., S. Cambre, S. K. Doorn, and H. Htoon. Temperature- bon Nanotube Species Using Aqueous Two-Phase Ex- dependent photoluminescence of water-filled carbon traction. 2015. ACS NANO. 9 (5): 5377. nanotubes: Thermal expansion of a single chain of wa- Galande, , W. e. i. Gao, Mathkar, A. M. Dattelbaum, T. N. ter molecules. Nano Letters. Narayanan, A. D. Mohite, and P. M. Ajayan. Science Martin, E. J. J., Berube, Provencher, Cote, Silva, S. K. Doorn, and Engineering of Graphene Oxide. 2014. PARTICLE & and J. K. Grey. Resonance Raman spectroscopy and PARTICLE SYSTEMS CHARACTERIZATION. 31 (6): 619. imaging of push-pull conjugated polymer-fullerene Hartmann, N. F., S. E. Yalcin, L. Adamska, E. H. Haroz, X. blends. 2015. JOURNAL OF MATERIALS CHEMISTRY C. 3 Ma, S. Tretiak, H. Htoon, and S. K. Doorn. Photolumi- (23): 6058. nescence imaging of solitary dopant sites in covalently Sau, J. D., J. J. Crochet, S. K. Doorn, and M. L. Cohen. Mul- doped single-wall carbon nanotubes. Nanoscale. tiparticle exciton ionization in shallow doped carbon Ji, , S. K. Doorn, and Sykora. Electrochromic Graphene Mol- nanotubes. 2013. Physical Chemistry Letters. 4: 982. ecules. 2015. ACS NANO. 9 (4): 4043. Solenov, D., C. Junkermeier, R. L. Reinecke, and K. A. Vel- Ji, Z., R. Wu, L. Adamska, K. Velizhanin, S. K. Doorn, and M. izhanin. Tunable adsorbate-adsorbate interactions on Sykora. In-situ synthesis of graphene quantum dots on graphene. 2013. Physical Review Letters. 111: 115502. TiO2: Applications in sensitized solar cells. 2014. ACS Subbaiyan, N. K., Cambre, A. N. G. Parra-Vasquez, E. H. Applied Materials and Interfaces. 6: 20473. Haroz, S. K. Doorn, and J. G. Duque. Role of Surfactants Kilina, , Kilin, and Tretiak. Light-Driven and Phonon-Assisted and Salt in Aqueous Two-Phase Separation of Carbon Dynamics in Organic and Semiconductor Nanostruc- Nanotubes toward Simple Chirality Isolation. 2014. ACS tures. 2015. CHEMICAL REVIEWS. 115 (12): 5929. NANO. 8 (2): 1619. Lawrence, D., C. Tran, A. T. Mallajosyula, S. K. Doorn, A. Subbaiyan, N. K., and S. K. Doorn. Nanotube micellar sur- Mohite, G. Gupta, and V. Kalra. High-energy density face chemistry: Surfactant surface structure, modifi- nanofiber-based solid state supercapacitors. To appear cation, and application. 2015. In Handbook of carbon in Journal of Materials Chemistry A. nanomaterials. Edited by D’Souza, F., and K. M. Kadish. Vol. 8, p. 1. New Jersey: World Scientific. Liu, J. i. n., Adamska, S. K. Doorn, and Tretiak. Singlet and triplet excitons and charge polarons in cycloparaphen- Subbaiyan, N., A. N. G. Parra-Vasquez, S. Cambre, M. ylenes: a density functional theory study. 2015. PHYSI- A. Santiago Cordoba, S. E. Yalcin, C. E. Hamilton, N. 303

Page 304

H. Mack, J. Blackburn, S. K. Doorn, and J. G. Duque. Benchtop extraction of isolated individual single- walled carbon nanotubes. 2015. Nanoscale. 9: 5377. Velizhanin, K. A., Dandu, and Solenov. Electromigration of bivalent functional groups on graphene. 2014. PHYSI- CAL REVIEW B. 89 (15). Yalcin, S. E., Galande, Kappera, Yamaguchi, Martinez, K. A. Velizhanin, S. K. Doorn, A. M. Dattelbaum, Chhowalla, P. M. Ajayan, Gupta, and A. D. Mohite. Direct Imaging of Charge Transport in Progressively Reduced Gra- phene Oxide Using Electrostatic Force Microscopy. 2015. ACS NANO. 9 (3): 2981. Yamaguchi, H., S. Ogawa, L. Adamska, A. M. Dattelbaum, G. Gupta, S. K. Doorn, H. Htoon, M. Chhowalla, K. A. Velizhanin, A. Mohite, and Y. Takakuwa. Valence band electronic structure evolution of graphene oxide upon thermal annealing fo optoelectronics. Journal of Physi- cal Chemistry Letters. 304

Page 305

Materials for the Future Directed Research Final Report Design Principles for Materials with Magnetic Functionality Joe D. Thompson 20130052DR Abstract netism also underpins many energy-efficient technolo- A significant challenge to materials physics and chemis- gies, notably wind turbines and ‘green’ automobiles, but try is the designed control of materials functionalities, magnetism is an intrinsically highly complex and delicate a challenge that must be met to satisfy ever increas- collective-electron function that is a far more challeng- ing demands for advanced materials that underpin the ing problem. Nation’s energy security and competitiveness. Strongly All magnets in wind turbines and other green technolo- magnetic materials, whose functionality results from gies presently rely on a material with a 3d-element, like a delicate balance of complex interactions among Fe or Co, and a rare-earth element Nd or Sm. The 3d- electrons, are key to numerous energy-efficient tech- element provides a large saturated magnetic moment nologies, but their useful function presently relies on and the 4f-element is the source of atomically derived rare and expensive elements, almost always imported. magnetic anisotropy that hinders magnetic moments The goal of this project has been to discover scientific from reversing their orientation as a magnetic field principles that, for the first time, allow designed control changes sign and is essential for ‘strong’ magnetic prop- of materials that are paradoxically both strongly mag- erties. The Department of Energy has recognized the netic and free of expensive rare elements. Substantial high demand and high supply risk of critical rare-earths progress in meeting our goal has allowed first-principles in its 2011 ‘Critical Materials Strategy’ that calls specifi- calculation of desirable magnetic properties in a strong cally for the long-term need for research aimed at creat- magnet devoid of rare elements and the identification of ing rare-earth-free strong magnets. Despite decades of essential design principles for discovering new magnetic making-and-measuring to find such ferromagnets that materials with technologically useful functionality. are free of 4f-electron elements, this approach has not been successful. Progress requires a strategy based on Background and Research Objectives scientific principles that ultimately will allow the design Advanced materials have been and will continue to be of a suitable magnet with magnetic anisotropy even in essential for the country to remain globally competitive the absence of a 4f-element. Though some magnetic and to ensure our energy and defense security. Indeed, anisotropy can be derived from d-electron elements, in recent national Materials Genome and Computational virtually all cases the coupling between electron spin Materials Science and Chemistry Initiatives are motivat- and orbital moments (spin-orbit coupling) that gives ed by the need to develop design principles that acceler- rise to magnetic anisotropy is quenched by the metallic ate the discovery and application of advanced materials crystal-chemical environment in which the d-element re- with useful function — a radical departure from tradi- sides. A successful strategy must understand conditions tional discovery of materials and functionality by ‘in- from atomic to macro scales that enable retention of formed serendipity’. Initial success in these initiatives is d-electron magnetic anisotropy, and this is the strategy promising. For example, first-principles calculations and that this project has taken. computational algorithms have reached the point where it is possible to predict useful properties of compounds Scientific Approach and Accomplishments relevant for photovoltaic, battery and thermoelectric Our computationally motivated, experimentally validat- applications. These materials properties and their appli- ed approach to this problem used two closely coupled cations are dominantly independent-electron function- loops – a relatively quick turn-around loop of computa- alities, which are relatively straightforward to calculate. tions, make and measure fed lessons learned into a sec- Like these independent-electron functionalities, mag- 305

Page 306

ond slower loop of in-depth theoretical and experimental a calculation was the first of its kind. These calculations understanding. Iteratively closing these loops has produced showed that the MAE density is double-valued as a func- a set of experimentally validated design principles for tion of the Coulomb energy U, but our photoemission and strong magnets without rare-earth elements. From work specific heat measurements constrained the value of U discussed below, these design principles are: (1) the crystal to be about 3.5 eV, which from our calculations gave the structure should be lower than cubic, ideally tetragonal or measured MAE density. This large U predicted theoretically hexagonal, and contain a high density of magnetic 3d ele- implies that electronic correlations are important, a fact ments; (2) strong electron-electron (Coulomb) interactions that we also have demonstrated by comparing the mea- are essential and become prominent when the electronic sured electronic specific heat with calculations that ignore bandwidth of d-elements is comparable to the Coulomb these correlations. This is a significant new design principle interaction energy, which tends to be realized in ternary that had not been appreciated previously by anyone. It is compounds that include Si, Ge, or P; (3) including a heavy important to note that the energy of the critical property, 4d- or 5d element in the material can be useful for promot- MAE, is more than three orders of magnitude smaller than ing magnetic anisotropy on the 3d element. U and much smaller than any other energy scale in the problem, highlighting the difficulty of accurately predicting Our initial hypothesis was that a family of rare-earth-free the MAE from first principles as we have done so success- materials, with compositions T23-xT’xZ6 (where T is a fully. These calculations further predicted that the MAE of 3d-element Cr, Mn, Fe, Co and Ni, T’ is a heavier 4d- or 5d- YCo5 might increase by a factor of six if we could tune the element and Z=B, C, or Al) was a promising starting point value of U to half its intrinsic value and also predicted with for exploring design principles. Density functional electron- remarkable accuracy (and with no adjustable parameters) ic structure and ordered magnetic moment were calcu- the orbital moment on Co atoms that we confirmed by x- lated for 35 new variants of this family, and based on those ray magnetic circular dichroism measurements. calculations, the most promising materials were made and their magnetic properties were measured. Trends in These LDA+DMFT calculations are extremely demanding magnetic properties predicted from calculations were sup- and time consuming, prohibiting their use for exploratory ported qualitatively, and often quantitatively, by experi- searches. For this, we turned to computational expertise ments. Though many of these materials had large satu- in the first loop. Using full-potential linearized augmented rated magnetic moments and high ferromagnetic ordering plane-wave (FP-LAPW) density functional (LDA) calcula- temperatures, which are desirable characteristics, none tions of electronic structure, we correctly predicted trends of these materials had substantial magnetic anisotropy. in the MAEs of XCo5, X=\{Th, Y, La, Ce\} and the XCo4B, X=\{Y, As learned by closing the two loops, the relatively strong La\} series of compounds, many of which we made and hybridization between electrons of the T and T’ elements measured. (As an experimental aside, we note that single and the absence of strong crystalline electric fields in the crystals are needed to determine the MAE, but finding cubic structure of these materials caused the spin-orbit experimental conditions under which these materials coupling, and hence magnetic anisotropy, to be negligible. might be prepared as single crystals also was prohibitively This family was not pursued further but did establish the time consuming. We found, however, a simple solution to design principle that structural anisotropy, for example in this obstacle. Because these materials are magnetically tetragonal or hexagonal structure types, was important. anisotropic, we prepared them in polycrystal form, ground the polycrystals into a powder, and aligned the powder In parallel, substantial effort, particularly in the second embedded in pre-hardened epoxy and subject to a mag- loop, was devoted to understanding why the rare-earth netic field while the epoxy hardened. From several tests, free magnet YCo5 had a magnetic anisotropy and associ- we showed that the aligned powders accurately repro- ated magnetic anisotropy energy (MAE) that is compa- duced the MAE measured in single crystals. This ‘make’ rable to that of the most widely used strong ferromagnet process took about a day instead of weeks to months to Nd2Fe14B. Though hexagonal YCo5 has been known for grow single crystals of a given material.) Despite predict- decades to have a large MAE, there was no microscopic un- ing correct trends, the quantitative accuracy of computed derstanding, in spite of numerous theoretical attempts, of MAEs compared to experimental values was poor, as precisely why this was the case. To address this problem, shown in the left panel of Figure 1; however, the accuracy we developed a theoretical framework (Local Density Ap- of the magnetic moments was generally superior, which is proximation + Dynamical Mean Field Theory – LDA+DMFT) useful because the anisotropy of the orbital moment μL is that included realistic electron-electron (Coulomb), related to the MAE of a material. From a large number of crystal-field and spin-orbit interactions on all five d-orbitals computational variations, we verified that anisotropy of μL of Co in its two inequivalent crystal structure sites. Such mimics that of the MAEs (right panel of Figure 1). Because 306

Page 307

these calculations converge very quickly, we may have not sufficiently optimized to produce significant magnetic identified a proxy for MAEs that is less computationally anisotropy. In contrast, we found that Coulomb interac- demanding, an important step in implementing our design tions were very strong in Fe3GeTe2, giving an MAE about principles in practice. 35% of that in YCo5 and that is among the highest of any rare-earth free ferromagnet, and single crystals of it were studied in detail as were variations on its chemical compo- sition. We note that the heavy 4d-element Te, like other 4d-and 5d-elements at the end of their row in the periodic table, has relatively large atomically derived spin-orbit coupling, which can be transferred to Fe and thus enhance Fe’s orbital moment. From x-ray circular dichroism, we de- termined the orbital and spin moments on Fe atoms, from photoemission found that the electronic bandwidth was quite narrow, consistent with strong electron-electron in- teractions, and imaged magnetic domains using magnetic force microscopy. Besides initial calculations that identified this material as promising, we applied LDA+DMFT meth- ods to calculate the photoemission spectra and spin and orbital moments. These calculations compared favorably with experiments but still could not account quantitatively for all measured atomic properties. We do not understand the origin of these quantitative differences but note that Fe3GeTe2 is an important validation of our design prin- Figure 1. Magnetization in units of Bohr magnetons per 3d ciples. element versus applied magnetic field. Open symbols corre- spond to field applied along the easy magnetic axis, where the magnetization saturates rapidly, and closed symbols are for field perpendicular to the easy axis. Fe5PB2 has the largest saturated magnetization, but it is only weakly anisotropic and consequently has a modest MAE, which is related to the area bounded by the easy and hard axis magnetization curves for this compound. Of the three ferromagnets, YCo5 has the smallest saturated magnetization but is highly anisotropic and has the largest MAE. Fe3GeTe2 has an intermediate MAE that is very large for an Fe- Figure 2. Comparison of calculated and measured magnetic based ferromagnet. properties of the series XCo5, where X is Th, Y, La and Ce, and Guided by initial design principles, the first loop performed of related compounds XCo4B. Symbols correspond to calcula- tions using different numbers of k (momentum) points in a three a total of nearly 1000 density functional calculations on po- dimensional grid used to calculated the LDA electronic structure. tentially magnetically anisotropic compounds, and about The number of k-points is given in the legends, and the larger 90 of the more promising materials were made, many as this number, the longer calculations take to converge but they aligned powders, and measured for comparison to pre- become more accurate. The left panel compares magnetic an- dictions. From those calculations and measurements, we isotropy energies (MAEs) and the right panel shows the relation identified two particularly promising materials Fe5PB2 and between calculated orbital moment anisotropy and measured Fe3GeTe2 whose magnetic anisotropy is compared to YCo5 MAE. in Figure 2. These ternary compounds contain elements P and Ge, which, like Si, do not ‘dissolve’ Fe, and this is a As mentioned earlier LDA+DMFT calculations, though condition that favors retention of Fe’s orbital moment. We quantitatively accurate, are extremely time consuming, found that the MAE of Fe5PB2 was relatively high, about and, therefore, we wanted to develop a much faster alter- 8% of that of YCo5, and at low temperatures exceeded that native that also captured strong electron-electron correla- of SrFe12O19 and BaFe12O19, which have applications tions, which LDA does not. To this end, substantial effort in microwave and small motor technologies but are not went into developing a general code that incorporates cor- useful for green technologies. Consequently, except for a relations within the Gutzwiller approximation for treating few substitution studies, Fe5PB2 was not pursued in-depth multi-orbital correlated electron systems, like the 3d-based because we showed that Coulomb interactions still were materials of interest for our project. This code includes the 307

Page 308

automatic generation of variational parameters based on organizing an international workshop on Electronic Struc- the crystal symmetry of the compound. In particular, we ture Approaches and Applications to Quantum Matter, applied our code to YCo5 based on a tight-binding pa- which included advances in the theory of strong magnets, rametrization, which we computed for this purpose. The and another team member serving as a guest editor of the tight-binding parametrization turned out to be challenging special issue of the Journal of Physics: Condensed Matter because of the hybridization of 3d-orbitals with s-character that was devoted to rare-earth replacement magnets. electrons. We also faced convergence problems when applying the Gutzwiller approximation to YCo5 because of Impact on National Missions the very large set of variational parameters (about 7000) By implementing Materials Genome initiative concepts, imposed by the rather low crystal symmetry of this com- this project of computationally motived, experimentally pound. It is important to emphasize, however, that our validated science has developed a set of principles for Gutzwiller+tight-binding code represents the first attempt the design of new materials with a specific functionality. to use this approach for computing the MAE of correlated Besides realizing a priority of the Laboratory’s strategy of electron materials. In spite of these challenges, we have Materials for the Future, success in this project has been demonstrated that this is a promising approach towards a a first step in realizing the Department of Energy’s Critical reliable, accurate computation of MAE in correlated elec- Materials Strategy of designing rare-earth-free strong mag- tron materials. Preliminary Gutzwiller-based calculations nets. Methodologies developed in this project are broadly based on a simpler band structure of YCo5 reproduced applicable to Laboratory materials-by-design problems in the experimental value of the MAE, and importantly these which electron-electron interactions are a fundamental calculations are over an order of magnitude faster than component of materials’ functionalities, ranging from new LDA+DMFT. types of sensing and near lossless transmission of energy to the response of Pu to environmental changes. In the A spin-off of our Guzwiller+tigh-binding code was the course of this project, a project-supported postdoc was implementation of the first quantum molecular dynamics converted to staff in T-1. algorithm, which includes correlations in the Gutzwiller approximation (QMD+GW). We wrote a code for a single- Publications orbital model and as a first test studied the change of Brito, N. L., E. D. Bauer, F. Ronning, J. D. Thompson, and the transport properties of a liquid metal as a function of R. Movshovich. Study of magnetic microstructure of increasing Coulomb interaction. In particular, we calculated Fe3GeTe2 using magnetic force microscopy. Applied the electronic conductivity and the ionic self-diffusion con- Physics Letters. stant when the liquid metal undergoes a Mott-Anderson Brito, N. L., E. D. Bauer, F. Ronning, J. D. Thompson, and R. transition. Movshovich. Surface magnetometry of Fe3GeTe2 via magnetic force microscopy. Journal of Physics: Con- Besides demonstrating success in meeting our scientific densed Matter. goal, this project also pursued programmatic funding opportunities and reached out to the broader scientific Chern, G. W., K. Barros, C. D. Batista, J. D. Kress, and G. Kot- community. In response to the DOE call for a Critical liar. Mott transition in a metallic liquid - Gutzwiller mo- Materials Hub proposal, we participated actively in the lecular dynamic simulations. Physical Review Letters. Laboratory’s response to that call. Though the proposal was among the three finalists, it was not selected. Addi- Janoscheck, M., J. C. Cesar, F. Ronning, E. D. Bauer, and J. D. Thompson. Physical properties of the ferromagnet tionally, project members were chosen to submit propos- Fe3GeTe2. Journal of Physics:Condensed Matter. als or contribute to proposals in response to DOE calls for Computational Materials Science and Chemistry and an Leiding, J., J. K. Ellis, and R. L. Martin. FP-LAPW/LDA cal- Energy Frontier Research Center, with emphasis in each culations of magnetic properties of XCo5, X=\{Th, Y, La, case on our approach and successes in developing design Ce\}, and XCo4B, X=\{Y. La\} magnets. Physical Review B. principles. Again, these proposals were highly ranked, but unfortunately not funded. In coordination with Laboratory Ronning, F., and J. L. Sarrao. Material’s prediction scores a hit. 2013. Physics . 6: 109. ARPA-E program managers, the concept of this project was discussed with two APRA-E program offices, the Rare-Earth Ronning, F., and S. Bader. Rare earth replacement mag- Alternatives in Critical Technologies and the Vehicle Tech- nets. 2014. Journal of Physics: Condensed Matter . 26: nology office. Both considered our concept attractive, but 060301. both also judged that it was premature to invest in it at this time. Out-reach activities included one team member co- Sarrao, J. L., F. Ronning, E. D. Bauer, C. D. Batista, J. X. Zhu, and J. D. Thompson. Building blocks for correlated su- 308

Page 309

perconductors and magnets. 2015. APL Materials. 3: 041512. Zhu, J. X., M. Janoschek, R. Rosenberg, F. Ronning, J. D. Thompson, M. A. Torrez, E. D. Bauer, and C. D. Batista. LDA+DMFT approach to magnetocrystalline anisotropy of strong magnets. 2014. Physical Review X. 4: 021027. Zhu, J. X., T. Durakiewicz, F. Ronning, M. Janoschek, E. D. Bauer, and J. D. Thompson. Electronic correlations and magnetism in the Hund ferromagnetic material Fe3GeTe2. Physical Review B. 309

Page 310

Materials for the Future Directed Research Final Report Non-Precious Metal Electrocatalysts for Clean Energy Piotr Zelenay 20130065DR Abstract Background and Research Objectives In order to fully realize the clean energy conversion and The oxygen reduction (ORR) and evolution (OER) reac- storage technologies (fuel cells, batteries, and water tions are crucial to a variety of electrochemical energy electrolyzers), development of highly-active, durable, storage and conversion technologies. The lack of under- and inexpensive electrocatalysts for the oxygen reduc- standing and control of ORR and OER electrocatalysis is tion and evolution reactions (ORR/OER) is necessary. a critical barrier to low-temperature fuel cells, metal-air This project has responded to that grand challenge by batteries, and water electrolyzers for hydrogen genera- combining experimental and multi-scale modeling ap- tion. High ORR and OER overpotentials are the main proaches for the purpose of (1) elucidating the nature of obstacle to making these technologies viable. Pt- and ORR and OER activity, (2) probing the interaction be- Ir-based catalysts are state-of-the-art in terms of activity tween meso- and nanoscale controls on electrocatalyst and durability for the ORR and OER, respectively. How- performance, and (3) designing and synthesizing non- ever, the prohibitive cost and scarcity of these precious precious metal catalysts (NPMCs) via a rational design. metals have limited their widespread implementation in clean energy applications. To fully realize these technolo- Active site characterization from density functional the- gies, highly active, durable, and inexpensive catalysts, ory (DFT) and model system experiments have provided based solely on earth-abundant elements, are desper- insights into the nature of the active site and critical role ately needed for oxygen electrodes. of water for the NPMC development graphitic NPMCs. Mesoscale/multiphysics studies using lattice Boltzmann Research objectives of this project have been: method (LBM) has developed the first physics-based (1) Molecular Level Insights into the Nature of Non-Pre- model to account for coupled fluid flow, ion transport, cious Metal Active ORR and OER Sites. Two-dimensional and electrochemical reactions for applications in energy carbon systems, such as graphene and graphene oxide conversion/storage technologies. Catalyst synthesis of (GO), can be readily modified and described theoreti- next-generation NPMC has produced the highest report- cally, making them ideal for elucidating electrocatalytic ed activity from two-dimensional graphene oxide (GO) active sites. Model electrocatalyst systems based on ORR electrocatalysts in acid electrolytes obtained from nitrogen-doped graphene and GO functionalized with highly ordered “dry” GO precursors. Further, a bifunc- nonprecious metals have been synthesized and charac- tional ORR/OER NPMC with low ORR-OER potential gap terized using spectroscopic and imaging tools to identify (0.6 V, ca. 0.1 V less than the lowest value reported to most active reaction sites. These graphene systems have date) has been developed from earth-abundant La-Sr- been modeled using density functional theory (DFT) Co-Fe-C catalyst. calculations to assign experimental spectra and inter- Through advancing the fundamentals of oxygen electro- rogate active-site geometry and electronic structure. catalysis and developing novel materials and concepts Hence, the scientific framework critical for optimization for energy applications this project has directly tackled of non-precious metal ORR, OER, and bi-functional ORR/ two top research priorities in the LDRD Energy & Earth OER electrocatalysts for clean energy production/storage Systems Challenge, including the need to utilize earth- have been developed in this project. abundant elements in place of precious metals. This (2) Theory and Optimization of the Mesoscale Catalyst work also positions LANL to target major new initiatives Structure. The electrochemical activity of NPMCs is in clean-energy catalysts in the future. 310

Page 311

highly coupled to the mesoscale structure of the catalyst: the interlayer graphene spacing from 3.3 Å to 10.8 Å due unlike precious metal catalysts there is no distinction to the formation of oxygen-containing functional groups between catalyst and support in NPMCs. Catalytic activity on the graphene basal plane and simultaneous intercala- is multiscale and requires theory to bridge the nano- and tion of water. Effective removal of intercalated water was mesoscales. Fundamental kinetic quantities for active sites, obtained by rinsing GO sheets with solvents of similar δd-, based on DFT, have been reconciled with mass-transport but lesser δp- and δH-character than water (e.g. ethanol phenomena via the LBM method. The resulting models and diethyl ether), leading to partially reduced GO materi- have provided an understanding of the linkage between als (Figure 1a). The effect of solvent rinsing on the removal three-dimensional mesoscale catalyst structure, mass- of intercalated water revealed a decrease in d-spacing transport effects, and active-site kinetics. from 10.8 Å to 7.5 Å, along with an increase in graphitic carbon content causing a confinement of water as ob- (3) Rational Design and Synthesis of Next-Generation served from the sharpening of the O-H stretch vibrational NPMCs. Based on the rational design derived from studies modes from FTIR (Figure 1c). of model graphene electrocatalysts, novel non-precious metal electrocatalysts have been synthesized. Carbon and a few oxide templates have been investigated for the stabilization and promotion of active sites. Such synthesis- based approach for new catalysis optimization, as well as a focus on the OER represents a major extension of NPMC research to date, both at LANL and elsewhere. While prac- tical NPMCs are more complex than the model systems, theoretical calculations have helped elucidate the catalytic processes observed spectroscopically, as in the case of the model graphene electrocatalytic system. This work creates a foundation for next-generation NPMC systems as a basis of a large electrochemical energy con- version and storage program at LANL. Figure 1. (a) Design principle of GO catalysts. (b) Electron micrographs of initial 2-D carbon GO. (c) Various characterization Scientific Approach and Accomplishments techniques employed for the monitoring of intercalated water. Model Systems (d) Resulting GO NPMCs. In this project, simpler carbons were utilized to better understand the process required for the formation of ac- Molecular dynamics models were used to find relaxed GO tive NPMCs. Typically, NPMCs are formed when nitrogen structures with intercalated water and various solvents heteroatoms and non-precious metals that, once incorpo- in order to aid the understanding of solvent effects on rated into the carbon lattice, can form a metal-nitrogen- multi-layer GO materials. Calculated averaged d-spacing carbon complex. State-of-the-art NPMCs are commonly between GO sheets showed values in good agreement synthesized from complex carbon precursors; by reducing with the decreasing d-spacing trend observed from experi- the complexity of such systems utilizing two-dimensional mental XRD data, supporting the hypothesis that solvents carbons such as graphene, graphene oxide (GO), and can substantially alter the structure of water-intercalated graphite we were able to identify the critical role of water GO. Thus, certain solvents can pull GO layers together in the formation of active NPMCs. close enough that they would react with residual alkali and transition metal ad-atoms leftover from the GO synthesis In the GO studies, we developed a simple solvent-drying process. Upon nitrogen incorporation via high temperature technique based on Hansen’s solubility parameters (disper- treatment in ammonia, these domains impede the egress sion, δd, hydrogen bonding, δH, and polarity, δp) to obtain of intercalated water molecules, leading to the forma- NPMC precursors that resulted in electrocatalysts with tion of active GO catalysts with increased ORR activity. high degree of ORR catalytic activity (the highest reported Overall, the two-dimensional sheet-like structure of GO to date for GO NPMC in acid electrolytes). The process in- is preserved, but large (5-20 µm) holes are formed in the volves monitoring of intercalated water, which exists abun- graphitic sheets after nitrogen incorporation (Figure 1d). dantly within GO sheets, using several characterization Such holey sheets address a significant challenge in using techniques: X-ray diffraction (XRD), X-ray photoelectron 2-D materials for catalysis, i.e., inaccessibility of active sites spectroscopy (XPS), and infrared spectroscopy (FTIR). Ex- deep within the layered material. These findings are cur- foliation and functionalization of graphite to GO increases 311

Page 312

rently under review in the sister journal of Science maga- mation from water, significantly modifying the active site zine, Science Advances. structure and ORR activity. The fourth motif, spontaneous *OH ligand modification of sites, has been shown to apply Model catalyst systems were also developed from gra- to FeN4 structures (both bi-vacancy and porphyrin). This phene, the simplest carbon structure. In this work, a novel understanding from modeling helps guide interpretation systematic bottom-up approach for the investigation of the of experiments, as in-situ and ex-situ observations of same role of each component in NPMCs was developed via nitro- initial structures are likely to be different (changed bond gen the incorporation on graphene structures using ener- lengths, planarity, etc.) due to this ligand modification. getic neutral atom beam (ENABLE), a unique novel tech- nique exclusive to LANL. ENABLE offers the ability to utilize These stability studies point to a novel active-site structure high kinetic energy reactive neutral atoms to overcome [FeCoN5(*OH) on a graphene zig-zag edge, Figure 3]. In- thermal barriers, allowing for direct scission and formation terestingly, this site was determined based on maximizing of chemical bonds without disrupting graphene’s pristine stability but was found to also have the highest calculated lattice structure. The correlation of chemical-structure to thermodynamic limiting potential (Ul) reported to date for ORR activity of these model nitrogen-doped systems was NPMC active sites. (Ul serves as a computationally acces- investigated using ultrathin film electrodes. These findings sible measure of ORR activity.) Thus, increasing the density revealed significant control towards the incorporation of of these active sites is predicted to significantly improve different N species into the graphene structure based on ORR catalyst activity. The capabilities and structures devel- substrate temperature or sequential atom doping (Figure oped are already aiding in follow-on projects for relating 2). structure to experimental signatures via computation as well as guiding further improvement in activity via struc- tural fine tuning. Figure 2. Nitrogen doping of graphene structures via ENABLE- assisted surface chemistry. Effect of substrate temperature and sequential doping on nitrogen speciation as observed from XPS characterization. Theory Figure 3. Free energy ORR (associative) pathway as a function The quantum-chemical modeling component of this of reaction coordinate on the FeCoN(*OH) structure. project was successful in understanding the atomic-scale behavior of probable ORR active site structures. The goal A pore-scale model based on the lattice Boltzmann meth- of linking atomic scale non-precious metal active site od (LBM) was developed for the coupled physico-electro- structures to ORR activity was achieved and, for the first chemical processes in both carbon/platinum (C/Pt) and time, coupled with structural stability considerations. NPMC cathode catalyst layers (CLs) for proton exchange Unlike previous studies, structural stability motifs were membrane fuel cells (PEMFCs). The model was applied to established and used to identify target sites for activity CLs to evaluate effects of nanoscale structural character- studies. Four structural motifs were discovered: (1) Fe is istics on the macroscopic transport properties, improving most stable in graphene with N nearest neighbors between fundamental understanding of reactive transport process- the Fe and C atoms, (2) resulting Fe-N structures are more es in nanostructured CLs. Application of the model to reac- stable embedded at graphene edges as compared to the tive transport processes is the first such study on NPMC CL bulk, (3) edge Fe-N structures are thermodynamically systems. driven to form small edge clusters (multi-metal atom sites), and (4) in-situ conditions (with applied potential in an High-resolution CL porous structures for PEMFCs were aqueous environment) can lead to spontaneous *OH for- reconstructed using the quartet structure generation set, 312

Page 313

which follows the experimental fabrication process in- Catalyst Synthesis cluding carbon seed generation, carbon phase growth, Pt Bifunctional OER/ORR catalysts were developed from deposition and ionomer coverage. Detailed distributions lanthanum-strontium-cobalt perovskites, (La0.85Sr0.15) of Pt, carbon, ionomer and void space were obtained from CoO3-δ (LSC). LSC demonstrates high OER activity com- the reconstruction process (Figure 4, left). It was found pared to state-of-the-art precious metal catalyst, IrO2, that the detailed layer structures significantly affect the however, its ORR activity is negligible (Figure 5a). Signifi- effective diffusivities. The tortuosity predicted is much cant improvement in ORR activity was obtained by modi- higher than that commonly used in the Bruggeman equa- fying LSC using a nitrogen precursor, cyanamide (CM), tion. Additionally, Knudsen diffusion plays a significant role and iron. Elemental energy-dispersive X-ray spectroscopy in oxygen diffusion and significantly reduces oxygen’s effec- (EDS) mapping (Figure 5b) and XRD analysis (not shown) tive diffusivity. The effective proton conductivity predicted revealed a phase separation of LSC into lanthanum oxides is close to the simulation results reported in literature and and a perovskite with more equal La:Sr ratio. Interest- much lower than experimental results, indicating other ingly, strontium was found atomically dispersed atop of proton transport mechanisms such as proton transport in graphene sheets throughout the catalyst as observed from liquid water. high-resolution transmission electron microscopy. Thus, catalytically active bifunctional OER/ORR catalysts were ob- tained when the initial perovskite structure (La0.85Sr0.15) CoO3-δ was converted into A-site defective perovskite. This A-site vacancy induces changes in the molecular orbit- al structure of the catalyst, favorably affecting OER activity. The catalyst activity observed from modified-LSC catalysts with nitrogen and iron demonstrate the lowest ORR-OER potential gap (0.6 V, ca. 0.1 V less than the lowest reported value) to date. Figure 4. Reconstructed 3D nanoscale structure of a cathode CL (left). Oxygen concentration in the NPMC CL obtained from the LBM simulations (right). Reactive transport processes in NPMC CLs and Pt CLs were numerically investigated at the nanoscale using the LBM (Figure 4, right). The simulation results show that although the total reaction surface area of NPMC CL considered in this study is larger than that of the Pt CL, the reaction rate within the NPMC CL is lower than that in Pt CL, due to a much smaller reaction rate coefficient. To further improve the performance of the NPMC CL, the reaction surface Figure 5. (a) ORR/OER activities of LSC and modified-LSC cata- area needs to be increased to overcome the slow reac- lysts: LSC+AB, LSC-AB-Fe-CM, and LSC-rGO-Fe-CM. (b) Elemental tion rate coefficient. In addition, the effects of micro- and EDS mapping from LSC-rGO-Fe-CM bi-functional catalyst. mesopores on mass transport were studied. Our results show that micropores (a few nm) contribute little to mass Lastly, based on our understanding gained of the critical transport. Mesopores (few tens of nm) or macropores role of water for the synthesis of graphitic NPMCs, a novel (larger than 50 nm) are required to increase the mass approach was developed to incorporate non-precious met- transport rate. Our study suggested that pore-scale model- als and nitrogen precursors into graphite using a water- ing can be used to optimize the CL microstructures for free process. Successful intercalation of dry metal and/or more desirable reaction surface area. This part of the work nitrogen precursors into graphite was obtained via high has resulted in six peer reviewed journal articles. Addition- temperature and pressure in an oxygen-free environment ally, significant progress was made in incorporating fluid obtaining a new catalyst precursor, graphite intercalated flow and ion transport into our model. The resulting model compound (N-GIC). Manuscripts for this process as well will be the first physics-based to account for coupled fluid as the study furthering the effect of solvent treatments in flow, ion transport, and electrochemical reactions and will alkaline environments are in preparation. These findings have applications in many other energy conversion/storage have been reported in a number of peer-reviewed jour- systems. nal articles, as well as conference presentations, many of which were invited contributions. 313

Page 314

Impact on National Missions lyst for oxygen reduction. 2013. Nature Communica- tions. 4: -. This research supported the Los Alamos LDRD Energy & Earth Systems Grand Challenge “concepts and materials for Elbaz, L., G. Wu, and P. Zelenay. Heat-treated non-precious- clean energy”. In particular, it tackled top research priori- metal-based catalysts for oxygen reduction. 2013. In ties of “new materials, for energy applications, containing Electrocatalysis in fuel cells: A non- and low-platinum earth-abundant elements that mimic properties of rare approach. Edited by Shao, M.. , p. 231. London: Spring- and expensive materials” and “energy generation and effi- er-Verlag. ciency”. This research will enable sustainable and low-cost Gao, W., G. Wu, M. T. Janicke, D. A. Cullen, R. Mukundan, energy devices with enhanced energy efficiency/capacities, J. K. Baldwin, E. L. Brosha, C. Galande, P. M. Ajayan, K. targeting new initiatives in mesoscale catalysis, which are L. More, A. M. Dattelbaum, and P. Zelenay. Ozonated scheduled to be announced by DOE-BES in the near future. graphene oxide film as a proton exchange membrane. Particularly, this research is relevant to the DOE-EERE mis- 2014. Angewandte Chemie International Edition. 53: sion of developing alternative energy sources, especially 3588. for transportation and stationary applications (fuel cells, hydrogen generation). By being directly relevant to renew- Higgins, D., G. Wu, H. T. Chung, U. Martinez, S. Ma, Z. Chen, able energy, this research earns the highest mark in the and P. Zelenay. Manganese-based non-precious metal energy security category. Finally, this effort has focused on catalysts for oxygen reduction in acidic media. 2014. ECS Transactions. 61 (31): 35. the fundamental understanding of materials for possibly the two most challenging reactions in electrochemistry, Holby, E. F., G. Wu, P. Zelenay, and C. D. Taylor. Modeling oxygen reduction and evolution, using a combination of non-precious metal matalyst mtructures and mheir modeling, experimentation with model systems, and cata- relationship to ORR activity. 2013. ECS Transactions. 58 lyst development making this project relevant to essential (1): 1859. understanding of materials and fundamental chemistry. Holby, E. F., and C. D. Taylor. Activity of N-coordinated Publications multi-metal-atom active site structures for Pt-free oxy- gen reduction reaction catalysis: Role of *OH ligands. Chen, L., G. Wu, E. F. Holby, P. Zelenay, W. Tao, and Q. Kang. 2015. Scientific Reports. 5: 9286. Lattice Boltzmann Pore-Scale Investigation of Coupled Holby, E. T., G. Wu, P. Zelenay, and C. D. Taylor. Structure of Physical-electrochemical Processes in C/Pt and Non- Fe-Nx-C defects in oxygen reduction reaction catalysts Precious Metal Cathode Catalyst Layers in Proton Ex- from first principles modeling. 2014. Journal of Physi- change Membrane Fuel Cells. 2015. 158: 175. cal Chemistry C. 118: 14388. Chen, L., Q. Kang, Q. Tang, B. A. Robinson, Y. He, and W. Hussey, D. S., D. L. Jacobson, D. Liu, B. Khaykovich, M. V. Tao. Pore-scale simulation of multicomponent multi- Gubarev, D. Spernjak, G. Wu, J. Fairweather, R. Mukun- phase reactive transport with dissolution and precipi- dan, R. Lujan, P. Zelenay, and R. L. Borup. Neutron im- tation. 2015. International Journal of Heat and Mass aging of water transport in polymer-electrolyte mem- Transfer. 85: 935. branes and membrane-electrode assemblies. 2013. Chen, L., Q. Kang, Y. L. He, and W. Tao. A critical review ECS Transactions. 58 (1): 293. of the pseudopotential multiphase lattice Boltzmann Li, Q., K. Luo, Q. Kang, Y. He, Q. Chen, and Q. Liu. Lattice model: methods and applications. 2014. International Boltzmann methods for multiphase flow and phase- Journal of Heat and Mass Transfer. 76: 210. change heat transfer. To appear in Progress in Energy Chen, L., Y. L. He, Q. Kang, and W. Tao. Coupled numeri- and Combustion Science. cal approach combining finite volume and lattice Liu, H., Q. Kang, C. R. Leonardi, B. D. Jones, S. Schmieschek, Boltzmann methods for multi-scale multi-physico- A. Narvaez, J. R. Williams, A. J. Valocchi, and J. Harting. chemical processes. 2013. Journal of Computational Multiphase lattice Boltzmann simulations for porous Physics. 255: 83. media applications - a review. 2014. Computational Chung, H. T., G. Wu, Q. Li, and P. Zelenay. Role of two car- Geosciences. eprint: 1404.7523. bon phases in oxygen reduction reaction on the Co- Martinez, U., G. M. Purdy, E. F. Holby, K. Artyushkova, J. PPy-C catalyst. 2014. International Journal of Hydrogen H. Dumont, A. Singh, N. H. Mack, P. Atanassov, D. A. Energy. 39: 15887. Cullen, K. More, M. Chhowalla, P. Zelenay, A. M. Dattel- Chung, H. T., J. H. Won, and P. Zelenay. Active and stable baum, A. Mohite, and G. Gupta. Critical Role of Inter- carbon nanotube/nanoparticle composite electrocata- calated Water for Electrocatalytically Active Nitrogen- 314

Page 315

Doped Graphitic Systems. Science Advances. Martinez, U., T. L. Williamson, K. Artyushkova, N. H. Mack, G. M. Purdy, J. H. Dumont, D. Kelly, W. Gao, A. M. Dat- telbaum, A. Mohite, G. Gupta, and P. Zelenay. Thin-film non-precious metal model catalysts for oxygen reduc- tion reaction. 2014. ECS Transactions. 64 (3): 293. Wu, G., and P. Zelenay. Nanostructured Nonprecious Metal Catalysts for Oxygen Reduction Reaction. 2013. Ac- counts of Chemical Research. 46 (8): 1878. Yalcin, S. E., C. Galande, R. Kappera, H. Yamaguchi, U. Martinez, K. A. Velizhanin, S. K. Doorn, A. M. Dattel- baum, M. Chhowalla, P. M. Ajayan, G. Gupta, and A. D. Mohite. Direct imaging of charge transport in progres- sively reduced graphene oxide using electrostatic force microscopy. 2015. ACS Nano. 9 (3): 2981–2988. Yamaguchi, H., J. Granstrom, W. Nie, H. Sojoudi, T. Fujita, D. Voiry, M. Chen, G. Gupta, A. D. Mohite, S. Graham, and M. Chhowalla. Reduced graphene oxide thin films as ultrabarriers for organic electronics. 2013. Advanced Energy Materials. : DOI: 10.1002/aenm.201300986. 315

Page 316

Materials for the Future Directed Research Final Report Phase Stability of Multi-Component Nanocomposites Under Irradiation Blas P. Uberuaga 20130118DR Abstract While these reports call for “radiation tolerant materi- The design of materials that are essentially “immune to als,” it is crucial to recognize that radiation damage is not radiation,” as called for by DOE, is one of the grand chal- a single problem; it is a generic umbrella that refers to a lenges in the materials science discipline. This is because host of very different problems. Broadly, one can catego- energetic radiation damages a material by a variety rize these into two fundamentally different effects, see of very different mechanisms, such as void swelling, Figure 1 [2-7]. On the left, void swelling, caused by clus- embrittlement, irradiation creep, and phase instabili- tering of radiation-induced point defects (vacancies and ties from radiation induced solute redistribution (RISR). interstitials), leads to void formation in materials result- Typically, the solution for one problem either has no ing in swelling and embrittlement. Grain boundaries and effect on or exacerbates the other damage mechanisms. interfaces are sinks for defects and can annihilate de- For example, using interfaces as sinks for vacancies and fects before they cluster, mitigating void swelling. On the interstitials can reduce void swelling but the same inter- right is a qualitatively different problem observed only in faces that mitigate void swelling can increase RISR. multi-component alloys, the current focus: radiation-in- duced solute redistribution (RISR). RISR leads to changes RISR causes phase instabilities that lead to catastrophic in the local chemical composition of the material, with failure through embrittlement, fracture and corrosion. enrichment of some and depletion of other solutes near We examine an innovative solution to the chronic prob- interfaces. As the composition profile becomes het- lem of radiation-induced phase instability: nanocompos- erogeneous near interfaces, new phases might form or ite materials design via interfaces that minimize defect enough segregation can occur to adversely affect prop- fluxes, which is the origin of RISR. We hypothesize sev- erties, leading to brittle intergranular fracture, increase eral factors can be used to control defect fluxes and sup- in the ductile-to-brittle transition temperature (DBTT), press RISR. Using a materials co-design R&D approach and accelerated stress corrosion cracking. Interfaces are where experiments and multi-scale modeling are closely the source of RISR. The end result of either void swell- integrated to test our hypotheses, we have gained ing or RISR is failure of the material, but the reasons are fundamental new insights of RISR to interfaces, and completely different. developed initial design principles for next-generation materials for nuclear energy applications that maintain Why does Cr deplete and Ni enrich at interfaces, as in chemical composition profiles that are critical for many Figure 1d? The literature provides the following insight materials properties in extreme environments. [5,8,9]: at high temperatures, radiation-induced vacan- cies and interstitials migrate to interface sinks. Depend- Background and Research Objectives ing on how solutes couple to these defects, they will Recent DOE workshops [1] have called for the develop- either be enriched or depleted at those interfaces. ment of materials that are “self-healing” or immune Different solutes couple differently with interstitial and to radiation to enable future advanced nuclear energy vacancy fluxes. Many years of experimental effort have systems. “Developing a first-principles description of generated a substantial knowledge database and phe- mechanical properties and phase stability in multi-com- nomenological models of RISR [10,11]. However, these ponent, multiphase systems under extreme conditions” models have little predictive power. Recent observations has been specifically called out as a grand challenge have shown that Cr enriches at GBs in ferritic-martensitic that involves many scientific issues, including [BES-ANES steels [12,13], the opposite behavior observed in fcc report, ref. 1]. steels (Figure 1d). Cr is a vacancy diffuser. The under- 316

Page 317

standing of interstitial solutes is poorer still, yet intersti- In what follows, we describe our efforts in examining these tial segregation can have a profound impact on interface hypotheses and the insight gained. We focus on Hypoth- brittleness. A key challenge is to understand how solutes eses #1 and #3. For several reasons, we were not able segregate to sinks to develop predictive models of their to pursue Hypothesis #2. However, during the course of behavior. the project, we developed a fourth hypothesis related to phase-stability under irradiation: Hypothesis #4: The phase stability of chemically complex precipitates under irradiation is dictated by, and can be controlled by, the minor alloying element chemistry. Scientific Approach and Accomplishments Hypothesis #1 Hypothesis #1 states that the atomic structure and chem- istry of interfaces will dictate RISR in a given materials sys- tem. To systematically study the role of interfaces on RISR, we developed a series of model metal/oxide two-dimen- sional bi-layer composite systems. We focused on metal/ oxide composite systems as these most closely mimic the nanostructured ferritic alloys (NFAs) [16] that are of inter- est for advanced nuclear energy systems. Using magnetron sputtering, we deposited a series of oxide films on Fe-Cr Figure 1. Two different aspects of radiation damage. : (a) sche- alloys [17, 18]. These samples were then irradiated with matic of radiation-induced defect agglomeration [2]; (b) swell- ion beams. Once irradiated, they were then character- ing of fuel rods in a reactor [3]; (c) void-induced embrittlement ized using advanced microscopy techniques (transmission of fuel cladding [4]. : (d) solute redistribution [5] at GBs in fcc electron microscopy (TEM), electron dispersion spectros- steel due to radiation leading to (e) brittle intergranular fracture copy (EDS), and electron energy-loss spectroscopy (EELS)). (irradiation-assisted stress corrosion cracking) [6] and (f) increase These techniques provide information about the structure in DBTT [7]. and chemical evolution induced by the irradiation. Three The state-of-the-art of research in radiation damage is systems were synthesized and analyzed: Fe+Cr/MgO, summarized in the BES-ANES report: Fe+Cr/TiO2, and Fe+Cr/Y2O3. In each case, the system was annealed as well as irradiated, to separate out the role of [I]t is often possible to empirically discover… materials temperature and radiation damage. that exhibit good resistance to a specific radiation damage process, but it is extremely challenging to develop a mate- Figure 2 summarizes the results of these experiments. Each rial that simultaneously exhibits high resistance to all the of the systems examined exhibits unique and surprising radiation damage mechanisms. [14] behavior. In the case of MgO, there is strong evidence of irradiation-assisted segregation of Cr to the interface, but As pointed out in Ref. [15], this is in large part due to RISR. no interdiffusion of Cr into the oxide. However, for both TiO2 and Y2O3, in addition to segregation from the Fe lay- Guided by three hypotheses, we have examined how er to the oxide, we see significant interdiffusion of Cr into interfaces influence RISR in an effort to develop design the oxide itself. In the case of TiO2, this occurs to a limited principles that can guide future materials development. degree during annealing, but is strongly enhanced by the Our hypotheses are: irradiation such that the Cr completely interpenetrates the oxide. In the case of Y2O3, only minimal interdiffusion oc- Hypothesis #1: RISR will depend on the atomic structure, curs during annealing. However, after irradiation, signifi- chemistry and energetics of interfaces. cant concentrations of Cr have diffused into the oxide. Hypothesis #2: RISR can be influenced by using intention- Thus, important conclusions from these experiments are ally added ‘sacrificial’ elements. that (a) all of the metal/oxide interfaces attract Cr, (b) this Hypothesis #3: For a given system and irradiation condi- attraction is strongly enhanced by the irradiation, and (c) tion, solute redistribution will depend on spacing between in many cases there is significant interpenetration of Cr interfaces. into the oxide. This interpenetration is often accompanied by phase changes. In the TiO2 and Y2O3 films, incorpo- 317

Page 318

ration of Cr leads to amorphization of the oxide. This is To test this hypothesis, we performed extensive kinetic particularly dramatic in the Y2O3 film, as bulk Y2O3 would Monte Carlo (KMC) simulations on the segregation of Cr to not normally be susceptible to amorphization. Cr signifi- ideal sinks during irradiation. We specifically focused on a cantly enhances the rate of failure of these oxides. In the low damage flux regime, as this means that defects will not case of MgO, even more complicated phase changes have accumulate quickly and thus the concentration of defects occurred, in which the Cr substitutes the Mg to form a will always be low, ideal conditions for the KMC simula- crystalline spinel-like phase (MgCr2O4). tions. The KMC model simulations the motion of vacancies and interstitials in a Fe matrix, near an ideal sink. Cr atoms interact with the vacancies and interstitials and are thus pushed and pulled accordingly. Once defects reach the sink, they are “perfectly annihilated” meaning that the sink eliminates them from the system. Defects are gener- ated at a given rate to mimic irradiation conditions in real experiments. Finally, the concentration of Cr and tempera- ture are varied to determine how the observed behavior depend on these variables. The results provided in Figure 3 are for a particular condition. Figure 2. Chemical profiles at metal/oxide interfaces after both annealing and irradiation. Of particular interest is the distribu- tion of Cr in the three sets of samples. The location of the inter- faces are indicated by the dashed lines. The behavior of Cr both after annealing and after irradiation depends significantly on the nature of the interface. Assisted by density functional theory (DFT) calculations, we have found that it is much easier for Cr to interdiffuse into these oxides than Fe. Thus, once the defects are cre- ated during the irradiation, Cr can diffuse into the oxide structures while Fe, even if it is forced into the oxide by the damage processes, will not mix with the oxide. Together, the experiments and theory indicate that cer- tain types of metal/oxide interfaces are much more stable against RISR than others. In particular, radiation-enhanced diffusion of Cr destabilizes many otherwise stabile inter- faces. These results confirm hypothesis #1, that RISR, and the consequences of RISR, are very sensitive to the nature of the interfaces involved. Hypothesis #3 Hypothesis #3 states that RISR can be controlled via the density of interfaces. This hypothesis is motivated by the fact that simple models found that, if the density of inter- Figure 3. Results of kinetic Monte Carlo (KMC) simulations faces is increased, the flux of radiation-induced defects examining the behavior of Cr near ideal sinks as a function of is reduced. Given that RISR depends on defect fluxes, we sink spacing, or sink density. As the sink density is increased, the hypothesized that the flux of alloying elements is also concentrations and fluxes of defects in the material decreases. influenced by interfacial density. This leads to lesser amounts of Cr segregation as the sink density is increased, mitigating RISR. 318

Page 319

As hypothesized, and as shown in Figure 3, increasing the of more personnel and we are close to a final model that density of interfaces does indeed decrease the amount of will be able to examine more complex materials than our Cr that segregates to those interfaces. As the interfacial KMC capability currently can. Figure 5 shows results from density increases, the concentrations and, more impor- the phase field simulations when ideal sinks are modeled. tantly, the flux of interstitials and vacancies correspond- These are sinks that perfectly annihilate any defect that ingly decrease. This decreases the overall flux and build-up reaches them. The phase field model shows that there is a of Cr to the interfaces. dependence on both the temperature and the Cr content for whether such sinks lead to Cr enhancement or deple- Hypothesis #4 tion at the sinks. This is in qualitative agreement with The last hypothesis developed during the course of the recent results from CEA [20]. As the model is continued to project and focused on the stability of pre-existing chemi- be developed, more realistic sink behavior will be incorpo- cal precipitates within the material under irradiation, rated. states that this stability is a consequence of the minor chemical species in the precipitate. This hypothesis grew out of our experiments on engineering alloys showing that, under irradiation, the major elements within the mate- rial remained essentially where they were, but it was the minor elements that redistributed significantly and led to the instability of the material. This is illustrated in Figure 4, in which the engineering material Rene N4 was irradiated to various doses up to 75 dpa (displacements per atom) at room temperature [19]. Rene N4 is a Ni-based superal- loy with many minor elements. In particular, the weight percent of those elements highlighted in Figure 4 are: 9.8% Al, 3.5% Ti, 7.5% Co, 9.8% Cr, and 1.5% Mo. Of these ele- ments, Mo exhibits the greatest amount of redistribution under irradiation. By 0.75 dpa, the chemical structure of the other elements is strongly intact, but the distribution of Mo has become much more diffuse, originally residing at the interphase boundaries between the cuboids forming Figure 4. Chemical distributions within Rene N4 after irradiation the structure and quickly becoming dispersed throughout to various doses. The distributions of the major elements remains the matrix. At the same time, structural ordering within mostly intact, even to the high doses, but the distribution of mi- the material is quickly lost during irradiation, with a con- nor elements such as Mo becomes quickly dispersed, even at the current drop in the mechanical hardness of the material. lowest level of damage considered. Given that the structural ordering is associated with the minor elemental species, this indicates that the mechanical properties of the material are very sensitive to the redis- tribution of these species under irradiation. Future work will work towards quantifying this effect and designing new alloying chemistries that mitigate radiation-induced disordering and dissolution of the material. Phase Field Development An important target of our research plan was the develop- ment of a phase field model, informed by atomistics, that could computationally probe the hypotheses to a level not possible with other computational approaches. In particular, because of the relative efficiency of the phase field approach, larger systems with more heterogeneities in terms of microstructure and chemistry could be consid- ered. Unfortunately, development of the phase field model progressed more slowly than anticipated, for a number Figure 5. Phase field simulations showing the behavior of Cr of reasons. Development accelerated with the addition near ideal sinks as a function of temperature and Cr concentra- 319

Page 320

tion, revealing that whether Cr is enriched or depleted at the Finally, a number of postdocs were converted to staff sinks depends on the irradiation conditions and the alloy content. as a consequence of this project. In particular, Nan Li The next step in developing this phase field model is to include (MPA-CINT), Osman Anderoglu (MST-8), Enrique Martinez more realistic properties for the sinks. (MST-8), and Samrat Choudhury (MST-8) were postdocs who were converted to staff positions at least in part as a Summary consequence of their involvement in this project. To summarize, we have proposed and tested several hypotheses that provide new insight into the design of References advanced materials for nuclear energy applications that

1. DOE-BES Workshop Reports: Advanced Nuclear Energy can mitigate radiation-induced segregation. In particular, Systems, Materials Under Extreme Environments; Sci- we have demonstrated that, indeed, RISR is sensitive to entific Grand Challenges in National Security: The Role the nature of the interfaces in the material, the density of of Computing at the Extreme Scale, DOE Workshop interfaces in the material, and the minor alloying chemistry Report. 2007. DOE-BES Workshop Reports. in the material
2. Ackland, G.. Controlling radiation damage. 2010. Sci- These insights provide new design principles for explor- ence. 327: 1587. ing new materials systems. In particular, nanostructured
3. Porollo, S. I., S. V. Shulepin, Y. V. Konobeev, and F. A. materials, already pursued to mitigate void swelling, may Garner. Influence of silicon on swelling and microstruc- be just as effective at mitigating RISR. However, care must ture in Russian austenitic stainless steel EI-847 irradi- be made in choosing the appropriate composite materi- ated to high neutron doses. 2008. Journal of Nuclear als. Some oxides, for example, will not only act as sinks for Materials. 378: 17. alloying elements, but will undergo potentially detrimental phase changes as a consequence of gettering those spe- 4. Garner, F. A., and D. L. Porter. Irradiation creep and cies. swelling of AISI 316 to exposures of 130 dpa at 385– 400 C. 1988. Journal of Nuclear Materials. 155: 1006. This project has laid the foundation for future efforts in developing advanced materials in which more realistic 5. Was, G. S.. Fundamentals of radiation materials sci- chemistries are considered. It has provided new funda- ence. 2007. mental insights into the role of interfaces in influencing
4. Busby, J. T., G. S. Was, and E. A. Kenik. Isolating the RISR that, in turn, provide new directions for developing effect of radiation-induced segregation in irradiation- materials that mitigate RISR. Finally, the capabilities devel- assisted stress corrosion cracking of austenitic stainless oped during the course of this project should find utility in steels. 2002. Journal of Nuclear Materials. 302: 20. other scientific efforts ongoing within the laboratory. We are currently pursuing such opportunities. 7. Lu, Z., R. G. Faulkner, and P. E. J. Flewitt. The role of irradiation-induced phosphorus segregation in the duc- Finally, as with all such projects, a number of other results tile-to-brittle transition temperature in ferritic steels. were also generated. Due to a lack of space, we cannot 2006. Materials Science and Engineering: A. 437: 306. highlight those here. However, they appear in our Refer-
5. Ardell, A. J.. Materials issues for generation IV systems. ence list [21-31].
6. NATO Science for Peace and Security Series B - Physics and Biophysics. Impact on National Missions As of now, while the work performed during the course 9. Wiedersich, H., P. R. Okamoto, and N. Q. Lam. A theory of this project has motivated and inspired several propos- of radiation-induced segregation in concentrated al- als to outside funding agencies, particularly the Office of loys. 1979. Journal of Nuclear Materials. 83: 98. Nuclear Energy, no such proposals have won funding. We
7. Was, G., and J. T. Busby. Role of irradiated microstruc- continue to be cognizant of funding opportunities that ture and microchemistry in irradiation-assisted stress would leverage our work and capabilities. We expect that, corrosion cracking. 2005. Philosophical Magazine. 85: once the phase field model has been completed, it will be

of interest to internal programs. Further, the work in this project was highlighted during a recent workshop on de- 11. Allen, T. R., and G. S. Was. Modeling radiation-induced veloping research directions for the Office of Environmen- segregation in austenitic Fe–Cr–Ni alloys. 1998. Acta tal Management. We thus expect that, once the results of Materialia. 46: 3679. that workshop have been evolved into a Funding Oppor- 12. Gupta, G., Z. Jiao, A. N. Ham, J. T. Busby, and G. S. Was. tunity, we will be very competitive because of the fruits of Microstructural evolution of proton irradiated T91. this project. 320

Page 321

2006. Journal of Nuclear Materials. 351: 162. 25. Watkins, E. B., Kashinath, Wang, J. K. Baldwin, Majews- ki, and M. J. Demkowicz. Characterization of a Fe/Y2O3
2007. Lu, Z., R. G. Faulkner, G. S. Was, and B. D. Wirth. Irradi- metal/oxide interface using neutron and x-ray scatter- ation-induced grain boundary chromium microchemis- ing. 2014. APPLIED PHYSICS LETTERS. 105 (4). try in high alloy ferritic steels. 2008. Scripta Materialia. 58: 878. 26. Martinez, , B. P. Uberuaga, and A. F. Voter. Sublattice parallel replica dynamics. 2014. PHYSICAL REVIEW E.
2008. Klueh, R. L.. Elevated temperature ferritic and martens- 89 (6). itic steels and their application to future nuclear reac- tors. 2005. International Materials Review. 50: 287. 27. Aguiar, J. A., O. Anderoglu, S. Choudhury, J. K. Baldwin, Y. Q. Wang, A. Misra, and B. P. Uberuaga. Nanoscale
2009. Zinkle, S. J.. Fusion materials science: Overview of chal- morphologies at alloyed and irradiated metal/oxide bi- lenges and recent progress. 2005. Physics of Plasmas. layers. 2015. Journal of Materials Science. 50: 2726. 12: 058101.
2010. Bach, H. T., O. Anderoglu, T. A. Saleh, T. J. Romero, C. T.
2011. Odette, G. R.. Recent progress in developing and quali- Kelsey, E. R. Olivas, B. H. Sencer, P. O. Dickerson, M. A. fying nanostructured ferritic alloys for advanced fission Connors, K. D. John, and S. A. Maloy. Proton irradiation and fusion applications. 2014. JOM. 66: 2427. damage of an annealed Alloy 718 beam window. 2015.
2012. Xu, Y., S. K. Yadav, J. A. Aguiar, O. Anderoglu, J. K. Bald- Journal of Nuclear Materials. 459: 103. win, Y. Q. Wang, J. A. Valdez, A. Misra, H. M. Luo, B. P.
2013. Senninger, , Martinez, Soisson, Nastar, and Brechet. Uberuaga, and N. Li. Solute redistribution and phase Atomistic simulations of the decomposition kinetics in stability at FeCr/TiO2-x interfaces under ion irradiation. Fe-Cr alloys: Influence of magnetism. 2014. ACTA MA-
2014. Acta Materialia. 89: 364. TERIALIA. 73: 97.
2015. Xu, Y., S. K. Yadav, J. A. Aguiar, O. Anderoglu, J. K. Bald-
2016. Sun, C., S. Zheng, C. C. Wei, Y. Wu, L. Shao, Y. Yang, K. T. win, Y. Q. Wang, A. Misra, H. M. Luo, B. P. Uberuaga, Hartwig, S. A. Maloy, S. J. Zinkle, T. R. Allen, H. Wang, and N. Li. Irradiation-induced formation of a spinel and X. Zhang. Superion radiation-resistant nanoengi- phase at the FeCr/MgO interface. 2015. Acta Materia- neered austenitic 304L stainless steel for applications lia. 93: 87. in extreme radiation environments. 2015. Scientific
2017. Sun, C., M. Kirk, M. Li, K. Hattar, Y. Wang, O. Anderoglu, Reports. 5: 7801. J. Valdez, B. P. Uberuaga, R. Dickerson, and S. A. Maloy.
2018. Yuryev, D. V., and M. J. Demkowicz. Computational Microstructure, chemistry and mechanical properties design of solid-state interfaces using O-lattice theory. of Ni-based superalloy Rene N4 under irradiation at
2019. Applied Physics Letters. 105: 221601. room temperature. 2015. Acta Materialia. 95: 357.
2020. Soisson, F.. Private communication. 2015. Private com- munication. Publications Aguiar, J. A., O. Anderoglu, S. Choudhury, J. K. Baldwin, Y.
2021. Choudhury, S., J. A. Aguiar, M. J. Fluss, L. L. Hsiung, A. Q. Wang, A. Misra, and B. P. Uberuaga. Nanoscale mor- Misra, and B. P. Uberuaga. Non-uniform solute segre- phologies at alloyed and irradiated metal/oxide bilay- gation at semi-coherent metal/oxide interfaces. 2015. ers. 2015. Journal of Materials Science. 50: 2726. Scientific Reports. 5: 13086. Bach, H. T., O. Anderoglu, T. A. Saleh, T. J. Romero, C. T.
2022. Lee, T., A. Caro, and M. J. Demkowicz. Atomistic model- Kelsey, E. R. Olivas, B. H. Sencer, P. O. Dickerson, M. A. ing of radiation-induced disordering and dissolution Connors, K. D. John, and S. A. Maloy. Proton irradiation at a Ni/Ni3Al interface. 2015. Journal of Materials Re- damage of an annealed Alloy 718 beam window. 2015. search. 30: 1456. Journal of Nuclear Materials. 459: 103.
2023. Martinez, E., and B. P. Uberuaga. Mobility and coales- Choudhury, S., J. A. Aguiar, M. J. Fluss, L. L. Hsiung, A. Mis- cence of small stacking fault tetrahedra in Cu. 2015. ra, and B. P. Uberuaga. Non-uniform solute segregation Scientific Reports. 5: 9084. at semi-coherent metal/oxide interfaces. 2015. Scien-
2024. Sun, C., E. Martinez, J. A. Aguiar, A. Caro, J. A. Valdez, K. tific Reports. 5: 13086. Baldwin, Y. Xu, B. P. Uberuaga, O. Anderoglu, and S. A. Klaver, T. P. C., E. del Rio, G. Bonny, and A. Caro . Inconsis- Maloy. Thermally induced interdiffusion and precipita- tencies in modelling interstitials in FeCr with empirical tion in a Ni/Ni3Al system. 2015. Materials Research potentials. Modelling and Simulation in Materials Sci- Letters. 3: 169. ence and Engineering. 321

Page 322

Lee, T., A. Caro, and M. J. Demkowicz. Atomistic modeling Yuryev, D. V., and M. J. Demkowicz. Computational design of radiation-induced disordering and dissolution at of solid-state interfaces using O-lattice theory. 2014. a Ni/Ni3Al interface. 2015. Journal of Materials Re- Applied Physics Letters. 105: 221601. search. 30: 1456. Martinez, , B. P. Uberuaga, and A. F. Voter. Sublattice paral- lel replica dynamics. 2014. PHYSICAL REVIEW E. 89 (6). Martinez, E., and B. P. Uberuaga. Mobility and coalescence of small stacking fault tetrahedra in Cu. 2015. Scientific Reports. 5: 9084. Senninger, , Martinez, Soisson, Nastar, and Brechet. Atom- istic simulations of the decomposition kinetics in Fe-Cr alloys: Influence of magnetism. 2014. ACTA MATERIA- LIA. 73: 97. Senninger, O., F. Soisson, E. Martinez, M. Nastar, C. C. Fu, and Y. Brechet. Modeling radiation induced segrega- tion in iron-chhromium alloys. Acta Materialia. Sun, C., E. Martinez, J. A. Aguiar, A. Caro, J. A. Valdez, K. Baldwin, Y. Xu, B. P. Uberuaga, O. Anderoglu, and S. A. Maloy. Thermally induced interdiffusion and precipita- tion in a Ni/Ni3Al system. 2015. Materials Research Letters. 3: 169. Sun, C., M. Kirk, M. Li, K. Hattar, Y. Wang, O. Anderoglu, J. Valdez, B. P. Uberuaga, R. Dickerson, and S. A. Maloy. Microstructure, chemistry and mechanical properties of Ni-based superalloy Rene N4 under irradiation at room temperature. 2015. Acta Materialia. 95: 357. Sun, C., S. Zheng, C. C. Wei, Y. Wu, L. Shao, Y. Yang, K. T. Hartwig, S. A. Maloy, S. J. Zinkle, T. R. Allen, H. Wang, and X. Zhang. Superion radiation-resistant nanoengi- neered austenitic 304L stainless steel for applications in extreme radiation environments. 2015. Scientific Reports. 5: 7801. Watkins, E. B., Kashinath, Wang, J. K. Baldwin, Majewski, and M. J. Demkowicz. Characterization of a Fe/Y2O3 metal/oxide interface using neutron and x-ray scatter- ing. 2014. APPLIED PHYSICS LETTERS. 105 (4). Xu, W., L. Li, J. A. Valdez, M. Saber, C. C. Koch, R. O. Scatter- good, and Y. Zhu. Effect of nano-oxide particle size on radiation resistance of iron-chromium alloys. Journal of Nuclear Materials. Xu, Y., S. K. Yadav, J. A. Aguiar, O. Anderoglu, J. K. Baldwin, Y. Q. Wang, A. Misra, H. M. Luo, B. P. Uberuaga, and N. Li. Irradiation-induced formation of a spinel phase at the FeCr/MgO interface. 2015. Acta Materialia. 93: 87. Xu, Y., S. K. Yadav, J. A. Aguiar, O. Anderoglu, J. K. Bald- win, Y. Q. Wang, J. A. Valdez, A. Misra, H. M. Luo, B. P. Uberuaga, and N. Li. Solute redistribution and phase stability at FeCr/TiO2-x interfaces under ion irradiation. 2015. Acta Materialia. 89: 364. 322

Page 323

Materials for the Future Directed Research Final Report Quantum Chemistry, Information, Materials and Metrology Angel E. Garcia 20130727DR Abstract functional and semi empirical methods, algorithms, and From quantum information to condensed matter ma- new computer codes. We also explored the intersection terials, the tools of quantum mechanics underpin the of quantum information, communications and comput- science and enable the exploration of new ideas and op- ing with new materials such as chiral superconductors portunities for applications. We proposed an integrated and topological insulators and also with novel detection and interdisciplinary set of problems including the quan- technologies such as ultra sensitive scanned approaches. tum chemistry of nanoscale materials, quantum commu- We studied how strong electronic correlations give rise nication, information and computing, quantum frustra- to new topological materials. These materials have pos- tion in materials, quantum magnets, cold atom physics, sible applications in photovoltaics, quantum computing, Casimir AMO physics, quantum phase transitions, and and in the development of advanced functional materi- quantum metrology. Our approach integrates these als. Another areas of interest are the physics of ultracold topics by looking for unexpected synergies. For example, atoms, Bose-Einstein condensates, and the possible use quantum information research provides new tools that of these systems for quantum computing. can be applied to condensed matter problems. Similarly, Scientific Approach and Accomplishments new materials such as topological insulators may provide an effective medium for implementing quantum memo- Our scientific approach is based on the use of quantum ry. We also explore fundamental topics such as quantum theory and statistical mechanics, applied mathematics, friction and novel materials with soft-matter-like phases and computational methods to study complex materials and investigate practical applications such as sensing and systems under extreme conditions (e.g., ultra cold device with ultra-high sensitivity. Examples of the latter atoms). These approaches are combined with experi- include scanned probes, cold atom magnetometry, and mental data to validate our models. The methods devel- nano-scale opto-mechanical devices. The integration of oped are applied to specific problems of interest, and, these topics will build capability in quantum science and through collaborations with experimental collaborators enable new exciting applications. we help develop new nanomaterials. Our work on this project resulted in 73 publications in refereed journals. Background and Research Objectives Here we present highlights of our accomplishments on Quantum mechanics is at the center of the properties of this project. matter at the microscopic level. These properties deter- The understanding of the electronic properties of mate- mine the macroscopic properties of materials. The goals rials is crucial for the design of new functional photoac- of this project are to develop scientific progress at the in- tive materials. When activated by light, organic materi- terfaces of quantum chemistry and materials, quantum als show dynamics that is often characterized by large information, and quantum optics. Of particular interest non-adiabatic couplings between multiple excited states. are the relations between spectroscopic and chemical As part of this project, we developed a non-adiabatic properties of organic nanomaterials. The tools of quan- excited state molecular dynamics (NA-ESMD) framework tum chemistry are used to understand how chemical that describes photoinduced phenomena in (conjugated) variations of nanomaterials change their properties. organic materials [1]. These methods have been used Spectroscopic properties of materials require the under- to describe the planarization of polyfluorenes, and the standing of quantum excited-states. For this purpose, energy transfer of poly phenylene vinulene (PPV), among our interdisciplinary team developed and employed new other applications [2-8]. The NA-ESMD framework pro- theoretical methods, including time-dependent density 323

Page 324

vides a computationally efficient and accurate description complementary approaches, such as those based on high of photoinduced dynamics that cannot be accomplished by throughput calculations or data driven discovery, together sophisticated ab initio methods. with the merits of machine-learning methods to accommo- date searches with multiple objectives or higher dimen- In condensed matter theory we have studied the proper- sional databases. ties of magnetic systems. Magnets are natural realizations of gases of interacting bosons whose relevant parameters, Impact on National Missions such as dimensionality, lattice geometry, amount of disor- The understanding of the electronic properties of materials der, nature of the interactions, and particle concentration is relevant to the properties of actinides and other f-elec- can vary widely between different compounds [9]. The tron systems. These materials are crucial for the laboratory particle concentration can be easily tuned by applying an and DOE mission on nuclear materials. The work on spec- external magnetic field, which plays the role of a chemical troscopic properties of organic nanomaterials is relevant to potential. This rich spectrum of realizations offers a unique the development of efficient photovoltaics, which in turn possibility for studying the different physical behaviors are important for developing renewable energy devices. that emerge in interacting Bose gases from the interplay The work on ultracold atoms is relevant for quantum between their relevant parameters. computing. We strongly impact DOE Office of Science mis- sion in the fundamental understanding of electronic and In this project we demonstrated that exotic magnetic magnetic materials with applications to energy, electron- vortex crystals can emerge near the magnetic field induced ics, and computing. The algorithms and codes generated to quantum critical point of frustrated magnets [10-11]. It is study excited states contribute to our national competitive- important to note that this was the first proposal for the ness in high performance computing. observation of magnetic vortex lattices in Mott insulators. In addition, we demonstrated that frustration can change References the nature of this quantum critical point in one-dimension-

1. Nelson, , Fernandez-Alberti, A. E. Roitberg, and Tre- al systems [12]. tiak. Nonadiabatic Excited-State Molecular Dynamics: Also, by combining numerical and field theory techniques, Modeling Photophysics in Organic Conjugated Materi- we demonstrated that an antiferromagnetic Ising model als. 2014. ACCOUNTS OF CHEMICAL RESEARCH. 47 (4): on stacked triangular layers exhibits a very exotic phase 1155. diagram with a number of Berezinskii-Kosterlitz-Thouless
2. Fernandez-Alberti, , V. D. Kleiman, Tretiak, and A. E. transitions (0, 2 or 4) that depends on the number of layers Roitberg. Nonadiabatic Molecular Dynamics Simula- [13]. In a similar context, we demonstrated the prolifera- tions of the Energy Transfer between Building Blocks in tion of Z6 vortices at the structural transition of a family of a Phenylene Ethynylene Dendrimer. 2009. JOURNAL OF multi-ferroic materials [14]. PHYSICAL CHEMISTRY A. 113 (26): 7535. An important part of this project is the organization of
3. Nelson, , Fernandez-Alberti, A. E. Roitberg, and Tre- workshops in the areas of quantum physics and its ap- tiak. Artifacts due to trivial unavoided crossings in the plication to nanomaterials and materials. In May 2013 we modeling of photoinduced energy transfer dynamics organized a workshop “Organic Solar Cells: Theory and Ex- in extended conjugated molecules. 2013. CHEMICAL periment, From Description to Prediction”. This workshop PHYSICS LETTERS. 590: 208. brought together the leading researchers in the field of or- ganic semiconductors with the primary focus on photovol-
4. Nelson, , Fernandez-Alberti, A. E. Roitberg, and Tretiak. taic applications. The emphasis was on deep understand- Nonadiabatic excited-state molecular dynamics: Treat- ing of fundamental processes occurring in organic solar ment of electronic decoherence. 2013. JOURNAL OF cells, as well as on improving practical tools for modeling/ CHEMICAL PHYSICS. 138 (22). characterization of existing and prediction/design of new photovoltaic materials and devices. 5. Nelson, , Fernandez-Alberti, A. E. Roitberg, and Tretiak. Conformational disorder in energy transfer: beyond In February 2014 we organized the workshop “Informa- Forster theory. 2013. PHYSICAL CHEMISTRY CHEMICAL tion Science for Materials Discovery and Design”. The aim PHYSICS. 15 (23): 9245. of this conference was to bring together the materials and information science communities to evaluate the oppor- 6. Adamska, , Nayyar, Chen, A. K. Swan, Oldani, Fernan- tunities and challenges in the emerging field of materi- dez-Alberti, M. R. Golder, Jasti, S. K. Doorn, and Tretiak. als informatics. Researchers presented contrasting but Self-Trapping of Excitons, Violation of Condon Ap- 324

Page 325

proximation, and Efficient Fluorescence in Conjugated coupled classical fields and fermions. 2013. Physical Cycloparaphenylenes. 2014. NANO LETTERS. 14 (11): Review B. 88: 235101. 6539. Bechtold, A., D. Rauch, F. Li, T. Simmet, P. Ardelt, A. Regler, K. Müller, N. A. Sinitsyn, and J. J. Finley. Three-stage 7. Ma, , Adamska, Yamaguchi, S. E. Yalcin, Tretiak, S. K. decoherence dynamics of an electron spin qubit in an Doorn, and H. a. n. Htoon. Electronic Structure and optically active quantum dot. 2015. Nature Physics. : 1. Chemical Nature of Oxygen Dopant States in Carbon Nanotubes. 2014. ACS NANO. 8 (10): 10782. Behunin, R. O., D. A. R. Dalvit, R. Decca, and C. C. Speake. Limits on the accuracy of isoelectronic gravity mea- 8. Liu, J. i. n., Adamska, S. K. Doorn, and Tretiak. Singlet surements at short separation due to patch effects. and triplet excitons and charge polarons in cyclopara- 2014. Physical Review D. 89: 051301. phenylenes: a density functional theory study. 2015. PHYSICAL CHEMISTRY CHEMICAL PHYSICS. 17 (22): Bel, G., and F. L. Brown. Theory of single molecule emission 14613. spectroscopy. 2015. The Journal of Chemical Physics. 142 (17): 174104. 9. Zapf, , Jaime, and C. D. Batista. Bose-Einstein con- Bisset, R. N., C. Ticknor, and P. B. Blakie. Finite-resolution densation in quantum magnets (vol 86, 2014). 2014. fluctuation measurements of a trapped Bose-Einstein REVIEWS OF MODERN PHYSICS. 86 (4). condensate. 2013. Physical Review A. 88 (6): 063624. 10. Kamiya, , and C. D. Batista. Magnetic Vortex Crystals in Bisset, R. N., R. M. Wilson, and C. Ticknor. Scaling of fluc- Frustrated Mott Insulator. 2014. PHYSICAL REVIEW X. tuations in a trapped binary condensate. 2015. Physical 4 (1). Review A. 91 (5): 053613. 11. Wang, , Kamiya, A. H. Nevidomskyy, and C. D. Batista. Bjorgaard, J. A., K. A. Velizhanin, and S. Tretiak. Solvent Three-Dimensional Crystallization of Vortex Strings in effects in time-dependent self-consistent field meth- Frustrated Quantum Magnets. 2015. PHYSICAL REVIEW ods. II. Variational formulations and analytical gradi- LETTERS. 115 (10). ents. 2015. The Journal of Chemical Physics. 143 (5): 054305. 12. Rahmani, , A. E. Feiguin, and C. D. Batista. Anyonic Liq- Bjorgaard, J. A., T. Nelson, K. Kalinin, V. Kuzmenko, K. A. uids in Nearly Saturated Spin Chains. 2014. PHYSICAL Velizhanin, and S. Tretiak. Simulations of fluorescence REVIEW LETTERS. 113 (26). solvatochromism in substituted PPV oligomers from excited state molecular dynamics with implicit solvent. 13. Lin, , Kamiya, Chern, and C. D. Batista. Stiffness from 2015. Chemical Physics Letters. 631–632: 66. Disorder in Triangular-Lattice Ising Thin Films. 2014. PHYSICAL REVIEW LETTERS. 112 (15). Bjorgaard, J. A., and M. E. Kose. Simulations of Singlet Ex- citon Diffusion in Organic Semiconductors: A Review. 14. Lin, , Wang, Kamiya, Chern, F. e. i. Fan, Fan, Casas, 2015. RSC Advances. 5: 8432. Y. u. e. Liu, Kiryukhin, W. H. Zurek, C. D. Batista, and Cheong. Topological defects as relics of emergent con- Campo, A. Del, and W. H. Zurek. Universality of phase tran- tinuous symmetry and Higgs condensation of disorder sition dynamics: Topological defects from symmetry in ferroelectrics. 2014. NATURE PHYSICS. 10 (12): 970. breaking. 2014. International Journal of Modern Phys- ics A. 29 (08): 1430018. Publications Campo, A. del. Shortcuts to Adiabaticity by Counterdia- Adamska, L., I. Nayyar, H. Chen, A. K. Swan, N. Oldani, S. batic Driving. 2013. Physical Review Letters. 111 (10): Fernandez-Alberti, M. R. Golder, R. Jasti, S. K. Doorn, 100502. and S. Tretiak. Self-trapping of excitons, violation of Condon approximation and efficient fluorescence in Campo, A. del, I. L. Egusquiza, M. B. Plenio, and S. F. Huel- conjugated cycloparaphenylenes. 2014. Nanotechnol- ga. Quantum Speed Limits in Open System Dynamics. ogy Letters. 14 (11): 6539. 2013. Physical Review Letters. 110 (5): 050403. Baratz, A., A. White, and M. Galperin. Effects of electro- Campo, A. del, J. Goold, and M. Pasternostro. More bang magnetic coupling on conductance switching of a gat- for your buck: Towards super-adiabatic quantum en- ed tunnel junction. 2014. Journal of Physical Chemistry gines. 2013. Scientific Reports. 4: 6208. Letters. 20: 3545. Campo, A. del., T. W. Kibble, and W. H. Zurek. Causality Barros, K., and Y. Kato. Efficient langevin simulation of and non-equilibrium second-order phase transitions 325

Page 326

in inhomogeneous systems. 2013. Journal of Physics: Kamiya, Y., and C. Batista. Magnetic vortex crystals in Condensed Matter. 25 (40): 325701. frustrated Mott insulator. 2014. Physical Review X. 4: 011023. Candia, R. D., J. S. Pedernales, A. d. Campo, E. Solano, and J. Casanova. Quantum Simulation of Dissipative Pro- Karzig, T., A. Rahmani, F. von Oppen, and G. Refael. Optimal cesses without Reservoir Engineering. 2015. Scientific control of Majorana zero modes. 2015. Physical Review Reports. 5. B. 91 (20): 201404. Chen, H., Y. Y. Tai, C. S. Ting, M. J. Graf, J. Dai, and J. X. Zhu. Kato, Y.. Multidiscontinuity algorithm for world-line Monte Disorder effects in multiorbital s±-wave superconduc- Carlo simulations. 2013. Physical Review E. 87 (1): tors: Implications for Zn-doped BaFe2As2 compounds. 013310. 2013. Physical Review B. 88 (18): 184509. Kilina, S., D. Kilin, and S. Tretiak. Light-Driven and Phonon- Chern, G. W., C. C. Chien, and M. Di Ventra. Dynamically Assisted Dynamics in Organic and Semiconductor generated flat-band phases in optical kagome lattices. Nanostructures. 2015. Chemical Reviews. 115 (12): 2014. Physical Review A. 90 (1): 013609. 5929. Chern, G., A. Rahmani, I. Martin, and C. Batista. Quantum Kim, J. W., Y. Kamiya, E. D. Mun, M. Jaime, N. Harrison, H. Hall Ice. 2012. PHYSICAL REVIEW B. Yi, Y. S. Oh, S. W. Cheong, C. D. Batista, and V. S. Zapf. Coexistence of Ising and Heisenberg moments in mul- Cimmarusti, A. ., Z. Yan, B. . Patterson, L. . Corcos, L. . tiferroic Ca3Co2-xMnxO6. 2014. Physical Review B. 89: Orozco, and S. Deffner. Environment-Assisted Speed- 060404. up of the Field Evolution in Cavity Quantum Electro- dynamics. 2015. Physical Review Letters. 114 (23): Kim, J. W., Y. Kamiya, E. D. Mun, M. Jaime, N. Harrison, J. D. 233602. Thompson, H. T. Yi, Y. S. Oh, S. W. Cheong, C. D. Batista, and V. S. Zapf. Multiferroicity with coexisting isotropic D’Alessio, L., and A. Rahmani. Thermally isolated Luttinger and anisotropic spins in Ca3Co2−xMnxO6. 2014. Physi- liquids with noisy Hamiltonians. 2013. Physical Review cal Review B. 89: 060404. B. 87 (17): 174301. Knolle, J., G. W. Chern, D. I. Kovrizhin, R. Moessner, and Deffner, S.. Ten years of Nature Physics: From spooky foun- N. B. Perkins. Raman Scattering Signatures of Kitaev dations. 2015. Nature Physics. 11 (5): 383. Spin Liquids in A2IrO3 Iridates with A=Na or Li. . 2014. Physical Review Letters. 113: 187201. Deffner, S., and A. Saxena. Jarzynski Equality in PT-Symmet- ric Quantum Mechanics. 2015. Physical Review Letters. Koutroulakis, G., T. Zhou, C. D. Batista, Y. Kamiya, J. D. 114 (15): 150601. Thompson, S. E. Brown, and H. D. Zhou. Magnetic Phases of the Spin-1/2 Triangular-Lattice Antiferromag- Delgado, F., C. D. Batista, and J. Fernando-Rosa. Local probe net Ba3CoSb2O9. 2015. Physical Review B. 91: 024410. of fractional edge states of S=1 Heisenberg spin chains. 2013. Physical Review Letters. 111 (16): 167201. Li, F., A. Saxena, D. Smith, and N. Sinitsyn. Higher-order spin noise statistics. 2013. New Journal of Physics. 15 Gan, Z., H. Wu, K. Barros, Z. Xu, and E. Luijten. Comparison (11): 113038. of efficient techniques for the simulation of dielectric objects in electrolytes. 2015. Journal of Computational Li, H., M. Catanzaro, S. Tretiak, and V. Chernyak. Excited- Physics. 291: 317. State Structure Modifications Due to Molecular Sub- stituents and Exciton Scattering in Conjugated Mol- Glasenapp, P., N. A. Sinitsyn, L. Yang, D. G. Rickel, D. Roy, ecules. 2014. Journal of Physical Chemistry Letters. 5 A. Grelich, M. Bayer, and S. A. Crooker. Spin noise (4): 641. spectroscopy beyond thermal equilibrium and linear response. 2014. Physical Review Letters. 113 (15): Li, H., S. Malinin, S. Tretiak, and V. Chernyak. Effective tight- 156601. binding models for excitons in branched conjugated molecules. 2013. The Journal of Chemical Physics. 139 Impens, F., R. Behunin, C. C. Ttira, and P. Maia. Neto. Non- (6): 064109. local double-path Casimir phase in atom interferom- eters. 2013. EPL (Europhysics Letters). 101 (6): 60006. Lin, S. Z., X. Wang, Y. Kamiya, and G. W. Chern. Topological defects as relics of emergent continuous symmetry and Intravaia, F., R. O. Behunin, and D. A. Dalvit. Quantum fric- Higgs condensation of disorder in ferroelectrics. 2014. tion and fluctuation theorems. 2014. Physical Review Nature Physics. 10: 970. A. 89: 050101. Lin, S., Y. Kamiya, G. W. Chern, and C. D. Batista. Stiffness 326

Page 327

from disorder in triangular-lattice ising thin films. 2014. Roy, D.. Cascaded two-photon nonlinearity in a one-di- Physical Review Letters. 112 (15): 155702. mensional waveguide with multiple two-level emitters. 2013. Scientific Reports. 3: 02337. Lin, S., and D. Roy. Role of kinetic inductance in transport properties of shunted superconducting nanowires. Roy, D.. Two-photon scattering of a tightly focused weak 2013. Journal of Physics: Condensed Matter. 25 (32): light beam from a small atomic ensemble: An optical 325701. probe to detect atomic level structures. 2013. Physical Review A. 87 (6): 063819. Liu, J., L. Adamska, S. K. Doorn, and S. Tretiak. Singlet and triplet excitons and charge polarons in cycloparaphen- Roy, D., L. Yang, S. Crooker, and N. Sinitsyn. Cross-correla- ylenes: a density functional theory study. 2015. Physi- tion spin noise spectroscopy of heterogeneous inter- cal Chemistry Chemical Physics. 17 (22): 14613. acting spin systems. 2014. Nature. Mun, E. D., G. W. Chern, V. Pardo, F. Rivadulla, R. Sinclair, V. Roy, D., L. Yang, S. Crooker, and N. Sinitsyn. Cross-correla- S. Zapf, and C. D. Batista. Magnetic field induced transi- tion spin noise spectroscopy of heterogeneous inter- tion in vanadium spinels. 2014. Physical Review Let- acting spin systems. 2015. Nature Scientific Reports. 5: ters. 112 (1): 017207. 9573. Muniz, R., Y. Kato, and C. D. Batista. Generalized spin-wave Roy, D., N. Bondyopadhaya, and S. Tewari. Topologically theory: application to the bilinear-biquadratic model. trivial zero-bias conductance peak in semiconductor 2014. PTEP. : 083I01. Majorana wires from boundary effects. 2013. Physical Review B. 88 (2): 020502. Nei, W., H. Tsai, R. Asadpour, J. C. Blancon, A. J. Neukirch, G. Gupta, J. Crochet, M. Chhowalla, S. Tretiak, M. Alam, Roy, D., Y. Li, A. Greilich, Y. Pershin, A. Saxena, and N. Sin- H. L. Wang, and A. Mohite. High efficiency solution- itsyn. Spin noise spectroscopy of quantum dot mol- processed Perovskite solar cells with millimeter-scale ecules. 2013. Physical Review B. 88 (4): 045320. grains. 2015. Science. 347 (6221): 522. Roy, D., and N. Bondyopadhaya. Statistics of scattered Nelson, T., S. Fernandez-Alberti, A. E. Roitberg, and S. Tre- photons from a driven three-level emitter in a one- tiak. Artifacts due to trivial unavoided crossings in the dimensional open space. 2014. Physical Review A. 89 modeling of photoinduced energy transfer dynamics in (4): 043806. extended conjugated molecules. 2013. Chemical Phys- ics Letters. 590: 208. Savukov, I., T. Karaulanov, and M. G. Boshier. Ultra-sensitive high-density Rb-87 radio-frequency magnetometer. Pershin, Y., V. Slipko, D. Roy, and N. Sinitsyn. Two-beam spin 2014. Applied Physics Letters. 104 (2): 023504. noise spectroscopy. 2013. Applied Physics Letters. 102 (20): 202405. Savukov, I., and T. Karaulanov. Magnetic-resonance imag- ing of the human brain with an atomic magnetometer. Pyka, K., J. Keller, H. L. Partner, R. Nigmatullin, T. Burger- 2013. Applied Physics Letters. 103 (4): 043703. meister, D. M. Meier, K. Kuhlmann, A. Retzker, M. B. Plenio, W. H. Zurek, A. del Campo, and T. E. Mehl- Savukov, I., and T. Karaulanov. Anatomical MRI with an stäubler. Topological defect formation and spontane- atomic magnetometer. 2013. Journal of Magnetic ous symmetry breaking in ion Coulomb crystals. 2013. Resonance. 231: 39. Nature Communications. 4: 3291. Savukov, I., and T. Karaulanov. Multi-flux-transformer MRI Rahmani, A.. Quantum dynamics with an ensemble of detection with an atomic magnetometer. 2014. Journal hamiltonians. 2013. Modern Physical Letters B. 27 (26): of Magnetic Resonance. 249: 49. 1330019. Shimada, H., J. L. Jacobsen, and Y. Kamiya. Phase diagram Rahmani, A., R. Muniz, and I. Martin. Anyons in Integer and the strong-coupling fixed point in the disordered Quantum Hall Magnets. 2013. Physical Review X. 3 (3): O(n) loop model. 2013. Journal of Physics A. 47: 031008. 122001. Rahmani, A., T. Kitagawa, E. Demler, and C. Chamon. Cool- Sonner, J., A. del Campo, and W. H. Zurek. Universal far- ing through optimal control of quantum evolution. from-equilibrium dynamics of a holographic supercon- 2013. Physical Review A. 87 (4): 043607. ductor. 2015. Nature Communications. 6: 7406. Rahmani, A., and G. Chern. Universal Rényi mutual infor- Torrontegui, E., S. Ibáñez, S. Martínez-Garaot, M. Mo- mation in classical systems: The case of kagome ice. dugno, A. del Campo, D. Guéry-Odelin, A. Ruschhaupt, 2013. Physical Review B. 88 (5): 054426. X. Chen, and J. G. Muga. Chapter 2 - Shortcuts to Adia- 327

Page 328

baticity. 2013. In Advances in Atomic, Molecular, and Optical Physics. Edited by Arimondo, E., P. R. Berman, and C. C. Lin. Vol. Volume 62, p. 117. -: Elsevier. Wang, Z., Y. Kamiya, A. H. Nevidomskyy, and C. D. Batista. Three-Dimensional Crystallization of Vortex Strings in Frustrated Quantum Magnets. 2015. Physical Review Letters. 115 (10): 107201. White, A. J., V. N. Gorshkov, S. Tretiak, and D. Mozyrsky. Non-adiabatic molecular dynamics by accelerated semiclassical Monte Carlo. 2015. The Journal of Chemi- cal Physics. 143 (1): 014115. White, A., S. Tretiak, and M. Galperin. Raman scattering in molecular junctions: a pseudoparticle formulation. 2014. Nano Letters. 14 (2): 699. White, C. E., N. J. Henson, L. L. Daemen, M. Hartl, and K. Page. Uncovering the True Atomic Structure of Disor- dered Materials: The Structure of a Hydrated Amor- phous Magnesium Carbonate (MgCO3·3D2O). 2014. Chemistry of Materials. 26 (8): 2693. Zhang, Z., K. Wierschem, I. Yap, Y. Kato, C. Batista, and P. Sengupta. Phase diagram and magnetic excitations of anisotropic spin-one magnets. 2013. Physical Review B. 87 (17): 174405. 328

Page 329

Materials for the Future Directed Research Final Report Non-Equilibrium Phenomena in Materials, Fluids, and Climate Angel E. Garcia 20130728DR Abstract Background and Research Objectives Materials and fluids are, on macroscopic scales, gov- Systems driven far from equilibrium constitute many of erned by continuous classical fields, often exhibiting the most difficult scientific challenges of the modern era. highly nonlinear phenomena under non-equilibrium con- Some prominent examples are materials strength, where ditions. On the microscopic scale, however, the granular- plastic (irreversible) motion involves the interactions of ity of the system can become important and interesting many dislocations as a material is strained past its elastic new behavior emerges from the particle discreteness. A limit, fluid turbulence – a venerable problem involved in variety of non-equilibrium processes that involve both transport of heat in oceans and atmospheres as well as regimes were investigated in this project. We used simu- in the sun and Earth’s interior, plasma instability associ- lations of atomic interactions on intermediate length ated with the interaction of the Earth’s magnetosphere scales to characterize the properties of geopolymer with the solar wind and with the generation of that wind concretes and compared with data from neutron scatter- through coronal mass ejection, the intrinsic behavior of ing studies, investigated a coupled granular continuum materials with unique properties arising from geometric model of earthquake dynamics, described complex frustration, and the behavior of fluid flow in porous me- interactions arising in colloidal particle suspensions, dia which is important in carbon sequestration scenari- developed new algorithms and mathematical approach- os, oil extraction methodologies, and improved hydraulic es to fluid flow in porous media to help assess carbon fracture technologies. Our research objectives including sequestration scenarios and enhanced hydraulic fracture developing novel algorithms to more effectively study methods, and probed the fundamental flow properties these difficult problems, providing unique quantitative of rotating and/or stratified fluids. We also explored understanding of mechanisms that form the basis for exciting areas of plasma physics including wave-particle practical applications, and discover new scientific princi- interactions in the Earth’s magnetosphere using kinetic pals underlying the physics of complex systems. simulations and magnetic reconnection in coupled fluid and magnetic field models of real plasmas. In all these Scientific Approach and Accomplishments areas, we focused on both science and application to This project has developed collaborative scientific solve national challenges in energy, climate, and environ- progress in materials, soft matter, fluid dynamics and ment. The approach was primarily theory and modeling plasma physics, all of which manifest important and with application to related experimental efforts involv- challenging aspects of systems driven out of equilibrium. ing neutron scattering, earthquake laboratory studies, Our scientific approach melds theory, modeling, and and physical realizations of rotating and/or stratified numerical simulation using tools of nonlinear science, fluid systems. Our interdisciplinary approach helped statistical physics, applied mathematics, and advanced bridge these problems and suggest multi-scale methods scientific algorithms. In some cases, we also collaborate to couple discreteness to continuum field models. Our with or directly perform experimental investigations that work was well aligned with LANL materials strategy, had complement the theory/modeling/simulation approach. strong components of information science and technol- Our work resulted in over 50 publications in refereed ogy to deal with data management and analysis, and journals so we present here some of the highlights of broadly supported capability in physics, applied math, our research efforts. materials science, and climate science. In this project, we focused on the structure and dynam- ics of a variety of materials of importance in funda- 329

Page 330

mental science and in applications. The manufacture and preparation of conventional concrete releases significant amounts of carbon dioxide, a greenhouse gas. Geopoly- mer concrete is a material that can substitute for conven- tional concrete but its properties are still being explored and discovered. Using a combination of molecular dynam- ics modeling and scattering experiments (neutrons and x-rays), we have advanced the structural characterization of these promising materials [1]. Another novel new mate- rial, that takes advantage of geometric frustration to create novel structural properties, is artificial spin ice for which much of the theory has been developed by CNLS postdocs. In this project, we proposed a new material structure by considering geometric frustration on an irregular lattice structure [2], providing a window into possible applications in which topological effects produce advanced functional materials. We also developed a new approach to creat- Figure 1. Geophysical rock domain with complex structure of ing novel structures using nanotechnology [3]. Our work rock fracture filled with fluid computed using a novel algorithm. on elastic-plastic deformation encompassed the coupling Pore pressure is shown as color coded with pressure difference of granular material with a continuum elastic material to vertical from bottom to top. replicate stick-slip motion in earthquakes, the characteriza- tion of interfacial defects [4], a possible explanation for Fluid dynamics underpins both the short and long term the anomalous behavior of solid helium, and the onset of evolution of the Earth’s oceans and atmospheres, i.e., irreversibility in amorphous solids under periodic shear. weather and climate, and controls, through stratification Using similar computational tools, we explored the clas- and rotation, the eruption of coronal mass ejections on the sical response to driving of topological arrangements of sun, the transport of heat in the core of the earth, and the spins called “skyrmions” [5]. mixing necessary for understanding supernova explosions. In this area, we studied how turbulent shear flows induce Solitons are a classical nonlinear solution to wave equa- fluid mixing using a combination of experimental measure- tions that arise in many systems including Bose-Einstein ments and novel data analysis [10]. Such flows develop condensates and optical fibers. In our work, we showed in key portions of the oceanic large-scale circulation that how these soliton systems could become unstable through controls the global distribution of heat. Parameterizing nonlinear interactions despite being seemingly quite stable the mixing properties of such flows is critical for model- from a linear (small amplitude) perspective [6]. We de- ing climate evolutions. Another accomplishment was scribed many other interesting properties of these nonlin- extending the domain of measurements of heat transport ear states. We also applied our methodology to traveling in thermal convection under rotation [11], a system highly waves in granular systems [7]. relevant to stellar convection and to Earth’s outer core transport. We also used equations for general rotating and Our research on fluid flows in porous media and fractured stratified flows to develop a novel new algorithm to speed rock sought to yield insights into the complicated nature of up computations in systems where there is a combination sedimentary formations under the Earth’s surface. In one of slow fluid modes and fast wave modes such as exist in approach, we investigated the flow regimes for fluid injec- the atmosphere and ocean [12]. We applied this in several tion into a confined porous medium [8]. In another, we de- contexts including a rotating and stratified atmosphere veloped an algorithm for explicitly generating pore struc- [13]. Rotation and stratification both drive fluid systems to tures (see Figure 1) similar to those created by hydraulic be more planar, i.e., they create flows that are anisotropic fracture [9]. We held a conference entitled “Grand Chal- and in certain limits they approach a quasi two-dimen- lenges in Geological Fluid Dynamics” in September 2015 sional fluids. We conducted experiments and analysis on a under the auspices of this project that brought together laboratory based two-dimensional fluid as an analog of the many world experts from academic, governmental and more complex geophysical system [14]. Our conference industrial institutions to explore novel theoretical, empiri- entitled “Ocean Turbulence” in 2013 brought together cal, numerical and experimental approaches to subsurface world experts in all aspects of turbulence in ocean flows, fluid processes. building on our continuing work in fluid dynamics and our interactions with climate modeling efforts at Los Alamos. 330

Page 331

When fluids are conducting, they can become unstable to to global security efforts. We sponsor postdoctoral fellows a combination of hydrodynamic and magnetic instabilities. working in areas relevant to the Laboratory mission. These Such fluids are known as “magnetohydrodyamic” and can postdoctoral fellows are candidates for future staff posi- support the spontaneous generation of a magnetic field tions at the Laboratory. through a dynamo effect and the reordering (or reconnec- tion) of magnetic field lines as those lines are acted upon References by hydrodynamic forces. Indeed, magnetic reconnec- 1. Provis, J., A. Hajimohammadi, C. White, S. Bernal, R. tion is one of the most important and poorly understood Myers, R. Winarski, V. Rose, T. Proffen, A. Llobet, and phenomena in astrophysics. In Figure 2, we illustrate a J. J. van Deventer. Nanostructural characterization of “tearing mode” instability from a 3D calculation of a two- geopolymers by advanced beamline techniques. 2013. fluid model of magnetic reconnection. We also suggested Cement and Concrete Composites. 36: 56. that conventional explanations of magnetic reconnection may be incorrect in not properly including the properties 2. Gilbert, I., G. W. Chern, S. Zhang, L. OBrien, B. Fore, C. of the turbulence [15]. Finally, we performed an analysis Nisoli, and P. Schiffer. Emergent ice rule and magnetic of the role of the magnetic field geometry in determining charge screening from vertex frustration in artificial the dynamic properties for a 2013 magnetic storm, where spin ice. 2014. Nature Physics. 10 (9): 670. the Van Allen Probes observed very rapid electron accel- 3. Chern, G. W., C. Reichhardt, and C. Nisoli. Realizing erations. Our work [16] shows that conventional dipolar three-dimensional artificial spin ice by stacking planar descriptions work well for higher energy electrons but that nano-arrays. 2014. Applied Physics Letters. 104 (1): details of the magnetic field determine the motions of 013101. lower energy electrons. 4. Wang, J., R. Zhang, C. Zhou, I. Beyerlein, and A. Misra. Characterizing interface dislocations by atomically informed Frank-Bilby theory. 2013. Journal of Materials Research. 28 (13): 1646. 5. Reichhardt, C., D. Ray, and C. J. Reichhardt. Collective Transport Properties of Driven Skyrmions with Random Disorder. 2015. Phys. Rev. Lett.. 114: 217202. 6. Kevrekidis, P. ., D. . Pelinovsky, and A. Saxena. When Linear Stability Does Not Exclude Nonlinear Instability. 2015. Physical Review Letters. 114 (21): 214101. 7. Xu, H., P. G. Kevrekidis, and A. Stefanov. Traveling waves and their tails in locally resonant granular Figure 2. Three dimensional realization of magnetic recon- systems. 2015. Journal of Physics A: Mathematical and nection illustrating tearing mode mechanism computed using a Theoretical. 48 (19): 195204. two-fluid plasma model in a triply-periodic domain. Lines denote magnetic field lines and color map shows charge density. 8. Zheng, Z., B. Guo, I. Christov, M. Celia, and H. Stone. Flow regimes for fluid injection into a confined porous Impact on National Missions medium. 2015. Journal of Fluid Mechanics. 767: 881. The materials research that we performed is in the meso- scopic regime, an important area for the MaRIE program 9. Hyman, J. D., and C. L. Winter. Stochastic generation and is applicable to recent DOE/BES priorities in funda- of explicit pore structures by thresholding Gaussian mental materials research. Our work on the fundamen- random fields. 2014. Journal of Computational Physics. tals of earthquakes may find application in the detection 277: 16. and triggering of earthquakes. The theoretical work 10. Odier, P., J. Chen, and R. Ecke. Entrainment and mix- on subsurface flows is relevant to energy recovery and ing in a laboratory model of oceanic overflow. 2014. carbon sequestration efforts in Applied Energy Programs. Journal of Fluid Mechanics. 746: 498. We continue to work on geophysically motivated prob- lems relevant to climate evolution and developed a novel 11. Ecke, R., and J. Niemela. Heat transport in the geo- parallel-in-time algorithm that may find application in strophic regime of rotating Rayleigh-Bernard convec- those efforts. The space plasma applications are relevant 331

Page 332

tion. 2014. Physics Review Letters. 113: 114301. and transport in a bi-axial spherical tumbler. 2014. SIAM Journal on Applied Dynamical Systems. 13 (2): 12. Haut, T., and B. Wingate. An asymptotic parallel-in- 901. time method for highly oscillatory PDEs. 2014. SIAM Christov, I., and H. A. Stone. Shear dispersion in dense Journal of Scientific Computing. 36 (2): A693. granular flows. 2014. Granular Matter. 16 (4): 509. 13. Whitehead, J. P., T. Haut, and B. Wingate. The effect of Christov, I., and P. M. Jordan. On an instability exhibited two distinct fast time scales in the rotating stratified by the ballistic-diffusive heat conduction model of Xu Boussinesq equations. 2015. IMA Journal of Numerical and Hu. 2014. Proceedings of the Royal Society A. 470 Analysis. : 1. (2161): 20130557. 14. Rivera, M., H. Aluie, and R. Ecke. The direct enstrophy Chu, H., C. Zhou, J. Wang, and I. J. Beyerlein. An Analyti- cascade of two-dimensional soap film flows. 2014. cal Model for the Critical Shell Thickness in Core/Shell Physics of Fluids. 26 (5): 055105. Nanowires Based on Crystallographic Slip. 2013. J. 61 (11): 2147. 15. Eyink, G., E. Vishniac, C. Lalescu, H. Aluie, K. Kanov, K. Bürger, R. Burns, C. Meneveau, and A. Szalay. Flux- Ecke, R. E.. Scaling of heat transport near onset in rapidly freezing breakdown in high-conductivity magneto- rotating convection. 2015. Physics Letters A. 379 (37): 2221. hydrodynamic turbulence. 2013. Nature. 497 (7450): 466. Ecke, R., and J. Niemela. Heat transport in the geostrophic regime of rotating Rayleigh-Bernard convection. 2014. 16. Zhao, L., Y. Yu, G. L. Delzanno, and V. K. Jordanova. Physics Review Letters. 113: 114301. Bounce averaged diffusion coefficients in a physics based magnetic field geometry from RAM-SCB. 2015. Eyink, G., E. Vishniac, C. Lalescu, H. Aluie, K. Kanov, K. Bürg- Journal of Geophysical Research: Space Physics. 120 er, R. Burns, C. Meneveau, and A. Szalay. Flux-freezing (4): 2616. breakdown in high-conductivity magnetohydrodynam- ic turbulence. 2013. Nature. 497 (7450): 466. Publications Gertjerenken, B., and P. G. Kevrekidis. Effects of interac- Aluie, H.. Scale decomposition in compressible turbulence. tions on the generalized Hong–Ou–Mandel effect. 2013. Physica D: Nonlinear Phenomena. 247 (1): 54. 2015. Physics Letters A. 379 (30–31): 1737. Barros, K., and W. Klein. Liquid to solid nucleation via onion Gilbert, I., G. W. Chern, S. Zhang, L. OBrien, B. Fore, C. structure droplets. 2013. Journal of Chemical Physics. Nisoli, and P. Schiffer. Emergent ice rule and magnetic 139 (17): 174505. charge screening from vertex frustration in artificial Beylkin, G., and T. S. Haut. Nonlinear approximations for spin ice. 2014. Nature Physics. 10 (9): 670. electronic structure calculations. 2013. Proceedings of Haim, L., A. Hagberg, and E. Meron. Non-monotonic reso- the Royal Society A. 469 (2158): 20130231. nance in a spatially forced Lengyel-Epstein model. Chern, C. W., M. Morrison, and C. Nisoli. Degeneracy and 2015. Chaos: An Interdisciplinary Journal of Nonlinear criticality from emergent frustration in artificial spin Science. 25 (6): 064307. ice. 2013. Physical Review Letters. 111 (17): 177201. Haut, T. S., T. Babb, P. G. Martinsson, and B. A. Wingate. Chern, G. W., C. Reichhardt, and C. J. Olson Reichhardt. A high-order scheme for solving wave propagation Avalanches and disorder-induced criticality in artificial problems via the direct construction of an approximate spin ices. 2014. New Journal of Physics. 16 (6): 063051. time-evolution operator. 2014. arXiv. Chern, G. W., C. Reichhardt, and C. Nisoli. Realizing three- Haut, T. S., T. Babb, P. G. Martinsson, and B. A. Wingate. dimensional artificial spin ice by stacking planar nano- A high-order time-parallel scheme for solving wave arrays. 2014. Applied Physics Letters. 104 (1): 013101. propagation problems via the direct construction of an approximate time-evolution operator. 2015. IMA Jour- Christov, I. C., and P. M. Jordan. Corrections to Morse and nal of Numerical Analysis. Ingard’s variational-based treatment of weakly-nonlin- ear acoustics in lossless gases. 2015. The Journal of the Haut, T., and B. Wingate. An asymptotic parallel-in-time Acoustical Society of America. 138 (1): 361. method for highly oscillatory PDEs. 2014. SIAM Journal of Scientific Computing. 36 (2): A693. Christov, I., R. Lueptow, J. Ottino, and R. Sturman. A study in three-dimensional chaotic dynamics: granular flow Hyman, J. D., and C. L. Winter. Stochastic generation of ex- 332

Page 333

plicit pore structures by thresholding Gaussian random ordered artificial spin ices: Avalanches and criticality fields. 2014. Journal of Computational Physics. 277: 16. (invited). 2015. Journal of Applied Physics. 117 (17): 172612. Karamatskos, E. T., J. Stockhofe, P. G. Kevrekidis, and P. Schmelcher. Stability and tunneling dynamics of a dark- Reichhardt, C., D. Ray, and C. J. Reichhardt. Magnus- bright soliton pair in a harmonic trap. 2015. Physical induced ratchet effects for skyrmions interacting with Review A. 91 (4): 043637. asymmetric substrates. 2015. New Journal of Physics. 17 (7): 073034. Kevrekidis, P. G., B. A. Malomed, A. Saxena, A. R. Bishop, and D. J. Frantzeskakis. Solitons and vortices in two- Reichhardt, C., D. Ray, and C. J. Reichhardt. Reversible dimensional discrete nonlinear Schrodinger systems ratchet effects for vortices in conformal pinning arrays. with spatially modulated nonlinearity. 2015. Physical 2015. Physical Review B. 91 (18): 184502. Review E. 91 (4): 043201. Reichhardt, C., D. Ray, and C. J. Reichhardt. Collective Kevrekidis, P. G., R. L. Horne, N. Whitaker, Q. E. Hoq, and Transport Properties of Driven Skyrmions with Random D. Kip. Bright discrete solitons in spatially modulated Disorder. 2015. Phys. Rev. Lett.. 114: 217202. DNLS systems. 2015. Journal of Physics A: Mathemati- cal and Theoretical. 48 (34): 345201. Reichhardt, C., J. Drocco, C. J. Reichhardt, and A. R. Bishop. Statics and Dynamics of Vortex Matter with Competing Kevrekidis, P. ., D. . Pelinovsky, and A. Saxena. When Lin- Repulsive and Attractive Interactions. 2013. Journal of ear Stability Does Not Exclude Nonlinear Instability. Superconductivity and Novel Magnetism. : 1. 2015. Physical Review Letters. 114 (21): 214101. Rivera, M., H. Aluie, and R. Ecke. The direct enstrophy cas- Khare, A., I. Christov, and A. Saxena. Successive phase tran- cade of two-dimensional soap film flows. 2014. Physics sitions and kink solutions in ϕ8, ϕ10, and ϕ12 field of Fluids. 26 (5): 055105. theories. 2014. Physical Review E. 90 (2): 023208. Wang, J., R. Zhang, C. Zhou, I. Beyerlein, and A. Misra. Middleton, R. S., J. W. Carey, R. P. Currier, J. D. Hyman, Q. Characterizing interface dislocations by atomically in- Kang, S. Karra, J. Jiménez-Martínez, M. L. Porter, and H. formed Frank-Bilby theory. 2013. Journal of Materials S. Viswanathan. Shale gas and non-aqueous fracturing Research. 28 (13): 1646. fluids: Opportunities and challenges for supercritical CO2. 2015. Applied Energy. 147: 500. Wang, X., and J. Whitehead. A bound on the vertical trans- port of heat in the ‘ultimate’ state of slippery convec- Nisoli, C., D. Abraham, T. Lookman, and A. Saxena. Quasi- tion at large Prandtl numbers. 2013. Journal of Fluid one-dimensional thermal breakage. 2013. Physical Re- Mechanics. 729: 103. view E. 88 (4): 042409. White, C., G. Kearley, J. Provis, and D. Riley. Structure of Nisoli, C., and A. R. Bishop. Attractive inverse square poten- kaolinite and influence of stacking faults: Reconciling tial, U(1) gauge, and winding transitions. 2014. Physical theory and experiment using inelastic neutron scat- Review Letters. 122 (7): 070401. tering analysis. 2013. The Journal of Chemical Physics. 138 (19): 4501. Odier, P., J. Chen, and R. Ecke. Entrainment and mixing in a laboratory model of oceanic overflow. 2014. Journal of White, C., J. Provis, B. Bloomer, N. Henson, and K. Page. In Fluid Mechanics. 746: 498. situ X-ray pair distribution function analysis of geopoly- mer gel nanostructure formation kinetics. 2013. Physi- Provis, J., A. Hajimohammadi, C. White, S. Bernal, R. Myers, cal Chemistry Chemical Physics. 15 (22): 8573. R. Winarski, V. Rose, T. Proffen, A. Llobet, and J. J. van Deventer. Nanostructural characterization of geopoly- White, C., K. Page, N. Henson, and J. Provis. In situ syn- mers by advanced beamline techniques. 2013. Cement chrotron X-ray pair distribution function analysis of the and Concrete Composites. 36: 56. early stages of gel formation in metakaolin-based geo- polymers. 2013. Applied Clay Science. 73: 17. Regev, I., T. Lookman, and C. Reichhardt. Onset of irrevers- ibility and chaos in amorphous solids under periodic Whitehead, J. P., T. Haut, and B. Wingate. The effect of shear. 2013. Physical Review E. 88 (6): 062401. two distinct fast time scales in the rotating stratified Boussinesq equations. 2015. IMA Journal of Numerical Regev, I., and C. Reichhardt. Rheology and shear band sup- Analysis. : 1. pression in particle and chain mixtures. 2013. Physical Review E. 87 (2): 201. Xu, H., P. G. Kevrekidis, and A. Stefanov. Traveling waves and their tails in locally resonant granular systems. Reichhardt, C. J., G. Chern, A. Libál, and C. Reichhardt. Dis- 2015. Journal of Physics A: Mathematical and Theoreti- 333

Page 334

cal. 48 (19): 195204. Zhao, L., Y. Yu, G. L. Delzanno, and V. K. Jordanova. Bounce averaged diffusion coefficients in a physics based mag- netic field geometry from RAM-SCB. 2015. Journal of Geophysical Research: Space Physics. 120 (4): 2616. Zheng, Z., B. Guo, I. Christov, M. Celia, and H. Stone. Flow regimes for fluid injection into a confined porous me- dium. 2015. Journal of Fluid Mechanics. 767: 881. Zheng, Z., I. Christov, and H. Stone. Influence of heteroge- neity on second-kind self-similar solutions for viscous gravity currents. 2014. Journal of Fluid Mechanics. 747: 218. Zhou, C., J. Su, M. Graf, C. Reichhardt, A. Balatsky, and I. Beyerlein. Plastic response of dislocation glide in solid helium under dc strain-rate loading. 2013. Physical Re- view B. 88 (2): 4513. 334

Page 335

Materials for the Future Early Career Research and Development Continuing Project Attosecond Dynamics of Correlated Electrons in f-Electron Materials Steve M. Gilbertson 20140622ECR Introduction The work will be of particular interest to DOE/SC and the The purpose of this project is to develop an experimen- scientific discovery and innovation missions. We will be tal setup capable of tracking electron dynamics in cor- looking at the fundamental time evolution of electron related electron systems. The time scales involved will dynamics within solid materials using newly developed be on the order of attosecond (10^-18 s). This will be the attosecond spectroscopy techniques. The questions first system of its type and what we are attempting to we can address involving correlated electrons in Mott measure will be the first observations of such dynamics. insulators and 4-f heavy fermion materials should greatly The goal of the project is to gain insight into quantum expand the laboratory’s knowledge base and will be coherent electron behavior in solid state materials. The compared to theoretical models that have been devel- majority of attosecond experiments thus far conducted oped at LANL. The materials chosen for this work were have concentrated on single atom dynamics in gases also specifically selected to address issues pertaining to and molecules, while applications to condensed mat- nuclear nonproliferation, and nuclear fuel research. The ter, especially coherent dynamics, are few. The project initial experiments will be performed on Indium-rich will focus on the measurement and first observations CeCoIn5 which shares similar properties to nuclear fuel of quantum multi-path interference in 5f Mott insula- compounds while secondary experiments will address tors and 4f heavy fermion materials. The materials will issues in UO2, a known nuclear fuel product, directly. be optically pumped with an infrared laser pulse and Finally, the work will address issues related to the MaRIE probed with a soft x-ray, attosecond laser pulse. This project. Our work will enable an important in situ char- will enable the transient recording of the excited state acterization capability for materials that can be studied dynamics. The experiments will require development of with the MaRIE facility in the future. Even more basi- an attosecond transient reflectivity apparatus - a unique cally, the attosecond source and spectrometer that will setup currently not in exsitence in the U.S. be developed can easily be applied for studying simpler atoms, molecules and possibly even biological samples. The system itself will be constructed using the most This will provide another tool for time resolved studies cutting-edge techniques in interferometer stabilization, in the areas of fundamental chemistry and biology. nonlinear optics, and state-of-the-art attosecond pulse generation methods. Such techniques have only been Progress demonstrated a few times worldwide. Once the experi- Over the past year we have met our goals. The atto- mental attosecond source setup is demonstrated, mate- second transient reflectivity setup was completed and rials of interest to the Department of Energy, namely f- tested yielding a very bright high harmonic extreme electron systems and systems with correlated electrons, ultraviolet light source. Additional chirped mirrors were will be tested. The expected observed phenomena will purchased to give more control over the few-cycle fem- be the first ever of such measurements. The data analy- tosecond laser pulses produced from the hollow core sis will be conducted with the expertise of Los Alamos fiber and a variety of laser pulses have been measured. material scientists. We have now characterized pulses from 30 femtosec- onds all the way down to 9.5 femtoseconds with >1.5 Benefit to National Security Missions mJ per pulse making our laser the shortest amplified The specific contributions that our work makes are system at LANL. We next implemented the double opti- almost entirely in the areas of the basic understanding cal gating technique after acquiring the necessary optics. of materials as well as overall scientific advancement. This was a major milestone for the project because we 335

Page 336

were able to measure the first attosecond laser pulses at Conclusion LANL. The attosecond pulses were spectrally characterized The ultimate goal of the project is to gain unprecedented under a variety of conditions and with different types of insight into quantum coherent electron behavior in solid generating gases giving estimates on total photon flux and state materials. We propose to exploit unique phenomena bandwidths. The interferometer was rearranged to allow in 5f Mott insulators and 4f heavy-fermion materials to for more flexibility in adding optics and both spatial and demonstrate quantum multi-path interference effects in temporal overlap were found again within a few femto- the attosecond response of these materials to an infrared seconds. More recently, the uranium dioxide sample was laser pulse. This will require development of an attosec- added to the setup and we are currently trying to get static ond transient reflectivity apparatus. We will show that the reflectivity data from the sample. Once this is accom- application of attosecond pulses to solid state materials plished, we will optically pump the sample to try and see can provide essential information about the dynamics and transient reflectivity. coherence of electrons. Along with these accomplishments, we also received sup- plemental funding which we are planning to use to imple- ment an attosecond time step piezo electric transducer (pzt) stage used for making fine temporal steps as well as locking the setup. When the funding was first awarded for this project, there were sufficient cuts that meant we had to reduce the total scope of our work. We decided at the time to use the attosecond pulses as a delta function probe after optical excitation allowing us the see dynamics that occur on a few femtosecond timescale. This meant we could save resources by not purchasing the closed loop pzt as well as reducing the strict requirements on inter- ferometer stabilization. However, with the supplemental funding, the pzt stage has been purchased along with several passive stabilization methods. We are hoping to get this implemented in the near term which may allow us to measure dynamics with attosecond precision before the experiment is over. In any event, the passive stabilization methods will help reduce temporal jitter and give us more temporal resolution in the data. Future Work In the upcoming fiscal year, we plan to begin the source development during the first couple of months followed by data acquisition. The plan of the previous year was modi- fied so that all equipment was purchased at once and up front so as to maximize the amount of time spent taking data in the second year. Since the hollow-core fiber as- sembly is complete as well as the upgrades for the laser system, the next step will be vacuum chamber construc- tion and alignment of the XUV optics. This will be followed by characterization of the attosecond laser pulses via re- flectivity measurements off of gold mirrors. Upon satisfac- tory completion of the attosecond pulse characterization, spatial and temporal overlap will be found at the location of the sample position. Once this is complete, samples that are already available will be installed for preliminary data acquisition. 336

Page 337

Materials for the Future Early Career Research and Development Continuing Project Controlling the Electronic Structure of Emerging Atomically Thin Materials Through Heterostructuring Jinkyoung Yoo 20150659ECR Introduction 3. Open novel opportunities of observation of emer- The overarching goal of the research is controlling the gent phenomena in the novel heterostructures. physical properties of emerging two-dimensional atomi- 4. Utilize the exotic physical phenomena (e.g. band cally thin materials (2D-ATMs, e.g. graphene, monolayer gap opening in atomically thin materials, tunneling- transition metal dichalcogenides) through heterostruc- mediated p-n junction properties) for semiconductor ture formation with conventional semiconductor (silicon, devices such as ultra-efficient photodetectors and germanium, gallium arsenide) growth on 2D-ATMs. artificial photosynthesis. Though the emerging 2D-ATMs have attracted great Benefit to National Security Missions attention in the last decade due to their exotic physical properties, there has not been methodology to control The goals of the LDRD project are closely related to the physical characteristics of 2D-ATMs since chemical semiconductor heterostructure formation and control doping and alloying of 2D-ATMs are not permanently of physical properties of emerging materials. Integrated stable. Applying bias to 2D-ATMs through semiconductor studies, from materials preparation to device demon- on 2D-ATM enables us to control the electronic proper- stration, offer benefits to basic science and applied ties of 2D-ATMs in precise and on-demand manner. The research. preliminary results of the team show that high-quality For basic science the project is directly relevant to the semiconductor layers and nanowires can be grown on grand challenges set by DOE SC, ‘control at the level of the 2D-ATMs through unconventional crystal growth electrons’ and ‘energy and information on the na- mechanism. The project can make the breakthrough noscale’. The project provides materials systems com- for 2D-ATM research field hindered by the absence of posed of ultimately (atomically) thin material and semi- reliable method of controlling physcial properties of conductors of which properties are precisely controlled 2D-ATMs and for conventional semiconductor research to tackle the grand challenges. field which has been suffered from the limit of inherent characteristics of conventional semiconductors. The project provides a reliable way to utilize the exotic properties such as abnormally large optical absorptance The synergistic output through a combination of low of two-dimensional atomically thin materials (2D-ATMs) dimensionality of 2D-ATMs and established science of which are useful to develop ultrafast and highly efficient semiconductors has broad impact in materials science detectors (especially for MaRIE) due to extremely small and device manufacturing. capacitance and ultimately thin heterojunction expedit- The research has four key goals: ing excited carrier separation and energy harvesting devices such as photovoltaic cells and water splitting

1. Provide a reliable and robust way of tuning proper- for hydrogen production (Energy security, Renewable ties of 2D-ATMs in precise manner since the main energy). Furthermore, the ultimate thinness of the approach of semiconductor growth on 2D-ATMs semiconductor/ 2D-ATMs makes it possible to transfer employs established semiconductor science and high-quality heterostructures onto various substrates technology. such as plastics and glasses without loss of physical char- acteristics. The transferrable heterostructures are the
2. Control carrier transport characteristics of 2D-ATMs key building blocks of embedded and flexible devices. through semiconductors grown on 2D-ATMs. 337

Page 338

Thus, the concepts explored in the project can also find Future Work diverse applications in electronic/photonic devices based In the next year, three milestones are set as the followings: on charged carrier management. Project success will pave the way for future research funded by Department of • Growth of semiconductors (silicon (Si), germanium Homeland Security, Department of Defense, Department (Ge)) thin films or nanowires on two-dimensional of Commerce, and pursuing initiatives in national nano- atomically thin materials (2D-ATMs, graphene and technology and manufacturing. monolayer molybednum disulfide (MoS2)) with con- trolled conductivity type of silicon and germanium. Progress Chemical vapor deposition in ultra-high/high vacuum Since the beginning of CY2015 (the project’s start date), 2 chamber will be employed for the growth. The target milestones have been achieved: 1) Growth of Ge(Si) thin of this goal is making both single crystalline semicon- films and nanowires on monolayer MoS2 and graphene, ductor thin film with grain size of >10 micrometer 2) Cross-sectional transmission electron microscopy (TEM) and single crystalline semiconductor nanowires with imaging of the interface between Ge and MoS2. Moreover, diameter and length of <100 nm and >1 micrometer three-terminal electronic device fabrication and theoretical on graphene and MoS2, respectively. N-type, intrinsic, calculation are in progress. The research results were pre- P-type semiconductors will be grown on 2D-ATMs via sented in 2015 Materials Research Society Spring Meeting introducing dopant precursors during the growth. in San Francisco in April. • Cross-sectional scanning transmission electron micros- In detail, a single crystalline Ge thin film was successfully copy studies of semiconductor/2D-ATMs to character- grown on monolayer and multi-layer MoS2 via chemical ize the structural properties of the heterointerfaces. vapor deposition. Single crystallinity of Ge thin film was The atom arrangements of grown semiconductor and confirmed by the cross-sectional TEM image. The same ap- 2D-ATM at the heterointerface will be investigated. proach was employed to grow Ge and Si thin films on gra- The detailed information obtained from the structural phene. Growth of Ge and Si nanowires was also achieved characterizations is employed to set the boundary con- by catalyst-assisted vapor-liquid-solid growth method in dition of electronic band structure calculation. a chemical vapor deposition chamber. The nanowires are • Theoretical calculation of electronic band structures also single crystalline and vertically aligned. Additionally, (mainly band gaps) of the heterostructures. Density- the positions of Ge(Si) nanowires on graphene and MoS2 Functional-Theory will be used for the calculation. were controlled by positioning catalysts. Conclusion The electrical properties of Ge(Si)/MoS2(graphene) hetero- structures are under investigation. For the research, multi- The achievements are closely related to fundamental terminal field effect transistor-type devices have been materials science from the perspectives of materials prepa- fabricated with the heterostructures. The multi-electrodes ration and control of properties, which can be directly on Ge(Si) thin film and on MoS2(graphene) layer enable us expanded to various device applications. to monitor electrical properties of semiconductor thin film The ideal outcome of the research is the con- grown on atomically thin materials (MoS2, graphene) and trol of the physical properties of 2D-ATMs through of atomically thin materials independently. Moreover, we semiconductor/2D-ATM heterointerfaces and the obser- are comparing the electrical properties of atomically thin vation of emergent phenomena in the heterostructures materials before and after Ge(Si) thin film growth. Accord- such as tunneling-mediated transport, band gap opening, ing to our recent observations, the electrical conductivity reversible strong localization of carriers, and other unex- type of MoS2 is converted from n-type to p-type by growth pected behaviors. of undoped Ge thin film. This behavior has been observed in more than 10 samples. The origin of the conductivity Additionally, the unconventional crystal growth mecha- type conversion is under investigation through theoretical nism, which is not based on chemical bonding between calculation (Density functional theory). Currently, charge substrate and grown material, provides novel insights of transfer between Ge and MoS2 generates the phenom- preparing high-quality heterostructures composed incom- enon. Since doping in atomically thin materials to control patible materials. conductivity type is a quite difficult issue, our observation provides a novel way and insight to control physical prop- erties of atomically thin materials. 338

Page 339

Publications Lin, Y. -C., J. Yoo, I. Bilgin, A. Mohite, S. Kar, and D. Pete. Growth and characterizations of germanium/mono- layer MoS2 heterostructures. 2015. 2015 Materials Research Society Spring Meeting. 339

Page 340

Materials for the Future Early Career Research and Development Continuing Project A Novel Crystal Plasticity Model that Explicitly Accounts for Energy Storage and Dissipation at Material Interfaces Jason R. Mayeur 20150696ECR Introduction Benefit to National Security Missions This project is focused on developing a novel mesoscale This project will deliver a simulation tool that addresses theory and computational framework that can be used relevant physics of mesocale deformation mechanisms to enhance our understanding of the role that interfaces in polycrystalline metals with an emphasis on the role play in dictating microstructural evolution in nanoscale played by material interfaces. This enhanced under- material systems. Recent work at LANL on fabricating standing will enable the design of materials with proper- bulk two-phase lamellar bimetallic nanocomposites has ties tailored for specific applications, which is directly shown that these material systems possess superior aligned with the Laboratory’s Materials for the Future strength and radiation damage resistance as compared Science Pillar. The proposed modeling framework is also the constituent metals and/or composites with more suited for bridging the mesoscale gap between atomistic conventional layer thicknesses, e.g. larger than one and macroscale continuum methods, and would comple- micron. Experiments show that these nanocomposites ment the modeling component of the Making, Measur- have highly oriented microstructures and that there is ing, and Modeling Materials (M4) facility as a part of the a predominant interface relationship between grains MarIE vision. along the bi-phase interfaces, which gives rise to their exceptional properties. The resulting microstructure in Progress these nanoscale composites was unexpected based on Since the funding for this project was appropriated in conventional understanding of the fabrication process Feb of 2015, the efforts of the PI and collaborators, in and active material deformation mechanisms. Earlier accordance with the project plan, have been directed conventional mesoscale and atomistic simulations towards: the simulation of size-dependent mechanical suggest that this unique microstructure arises due to a response of fcc nanocrystalline metals, the development competition between the energy stored at the bi-phase and implementation of a new constitutive model for bcc interfaces and the stability of certain crystallographic nanocrystalline metals into the existing computational orientations under the imposed deformation. In framework, and a continuing literature review and initial order to test this hypothesis, a mesoscale theory and efforts at theory development for the interface consti- computational framework that accounts for interface tutive model that is to be completed/implemented in energy, the relevant plastic deformation mechanisms, FY16. Some of the teams efforts have also been directed and the associated lattice reorientation is needed. Such at developing additional computational methods and a modeling framework does not currently exist, and capabilities to generate more realistic synthetic micro- the goal of this project is to develop a tool to bridge structures that are higher fidelity representations of the this gap in understanding. The proposed path of model actual nanocomposite microstructures of interest to this development would build upon an existing state-of-the- project. The PI was recently invited to and participated art nonlocal single crystal plasticity theory developed by in a workshop in Poitiers, France on “Interface-Mediated the principal investigator to include the interface energy Plasticity in Nanostructured and Architectured Materi- effects. The model would truly be one-of-kind and als”. At the workshop, the PI presented research results would uniquely position LANL to answer fundamental highlighting our efforts made towards understanding the questions related to the role of interfaces in nanoscale novel mechanical properties of these materials systems polycrystalline material systems. as well as our path forward for incorporating these es- sential interface physics that are currently missing from existing modeling frameworks. Our presentation was 340

Page 341

well-received and a potential collaboration with the group in France that is fabricating a complementary set of nano- composite materials remains a possibility. This collabora- tion could directly benefit from the computational model development being undertaken as part of this project. Future Work • Develop a single crystal constitutive law applicable for body-centered cubic (bcc) materials. • Implement the bcc single crystal constitutive law within the nonlocal finite element framework. • Develop the constitutive theory for the interface en- ergy terms. • Implement the interface energy constitutive theory within the nonlocal finite element framework. • Perform nonlocal finite element simulations of face- centered cubic (fcc) nanocrystalline materials systems without interface energy effects. • Perform nonlocal finite element simulations of bcc nanocrystalline materials systems without interface energy effects. • Perform nonlocal finite element simulations of nano- crystalline fcc/bcc multilayer materials systems with- out interface energy effects. Conclusion The primary goal of this research is to develop a nonlocal crystal plasticity model to study the competition between bulk-dominated and interface-dominated polycrystalline plasticity at the mesoscale. The model will be used to study mechanical response of nanocrystalline face-cen- tered cubic (fcc), body-centered cubic (bcc), and fcc/bcc lamellar composites. It is anticipated that the improved understanding of these nanoscale material systems ob- tained via simulation will facilitate next generation materi- als design by identifying relationships between process parameters and the resulting microstructure. 341

Page 342

Materials for the Future Early Career Research and Development Final Report Room Temperature Oxidation and Corrosion of Plutonium Alison L. Pugmire 20130738ECR Abstract surface chemistry and corrosion of plutonium, knowl- Previous work on plutonium oxidation and corrosion has edge of the surface oxide composition is paramount. primarily focused on ultra-high vacuum or accelerated The currently accepted description of the oxide layer is aging (e.g. high temperature) conditions. As a result, a shown in Figure 1. At ambient temperatures, the corro- significant gap exists in our understanding of the oxide sion products have been described as a thick PuO2 sur- layer formed at room temperature. This research has face layer over a thin Pu2O3 layer at the metal interface focused on performing a systematic study of plutonium (Figure 1, left). Under inert/non-oxidizing conditions, e.g. oxidation and corrosion in carefully controlled environ- ultra high vacuum (UHV), the PuO2 layer auto-reduces ments at various times throughout the corrosion pro- to form a thick Pu2O3 layer (Figure 1, right) [1]. cess. Samples of delta-phase plutonium alloyed with This generalized view of the oxide layer is based on years 1.9 at. % gallium were corroded for 1 month in dry air, of experiments encompassing a wide array of experi- humid air, and humid argon environments. After one mental conditions [2, 3]. While this description appears month, both plutonium and gallium oxides were ob- to be straightforward, several aspects regarding the served. After storage in a dry/inert atmosphere for six corrosion mechanism and products are still debated. and 16 months, further corrosion (and most likely radio- First, in an attempt to explain the increased corrosion genic aging) occurs to form non-crystalline plutonium rate in the presence of H2O, a hyperstoichiometric oxide oxides and gallium oxides. Based on the results of this (PuO2+x) has been proposed. However, this species has work and previous preliminary studies in our group, the only been observed by a few groups in extreme condi- plutonium oxide formed in the initial stages of oxidation tions (i.e. high temperatures, low pressures) [1,4] while appears to be disordered and non-crystalline (with no others debate its existence [5]. Second, the presence of long-range order), contrary to the historical depiction of the Pu2O3 layer at room temperature is questionable this oxide. The detection of gallium oxide has only rarely because it has not been experimentally observed but been observed, particularly for delta-phase plutonium. is only assumed based on its formation in very specific Its formation here has been attributed to the presence atmospheres (non-oxidizing/inert atmospheres) [1]. of an additional surface phase (alpha’-phase plutonium) Third, the initial weight gain studies in the 1950s demon- that was formed during sample preparation. Although strating the effects of alloying have yet to be explained. unexpected, this result shows that gallium can be oxi- In wet air, unalloyed, α-Pu corrodes 1.7 times faster than dized and may play a role in the oxidation. Overall, this alloyed, δ-Pu (3.3 at. % Ga), while in wet inert atmo- work has improved our understanding of the structural spheres, δ-Pu (3.3 at. % Ga) corrodes 6.5 times faster and chemical composition of the oxide layer formed un- than unalloyed, α-Pu [1,6]. Fourth, the surface oxide lay- der ambient conditions and has changed our model for er has been characterized as crystalline based on x-ray the oxide layer developed in conditions relevant to the diffraction (XRD) results. However, XRD is only sensitive safe storage and potential reuse of the material. to the crystalline form, and preliminary work from our group using EXAFS, spectroscopic ellipsometry (SE), Au- Background and Research Objectives gur electron (AES) and x-ray photoelectron spectrosco- The unpredictable behavior and reactivity of plutonium pies (XPS) indicate this layer is not as simplistic as once has led to a very poor understanding of its metallurgy thought, but may have amorphous Pu-O components and corrosion process. This not only poses a basic sci- [7,8]. Finally, because the plutonium oxidation process entific challenge, but directly affects the safe, long term is very slow at room temperature in O2 (~ 0.02 nm/hr) storage of this material. In an effort to understand the 342

Page 343

and is accelerated with elevated temperatures (> 25 °C) or duplicate sample placed in one of the exposure conditions a moist environment [5, 8], the vast majority of previous (wet air). The bulk chemical and structural components work has utilized accelerated growth conditions to produce of the oxide layer were investigated using EXAFS, and the oxide layers quickly for easier studies. This has resulted in a surface components were studied using XPS (when avail- significant gap in previous corrosion studies and leaves the able) and surface EXAFS (a synchrotron experiment new to room temperature growth of the oxide layer not well un- the PI). derstood. This regime is perhaps the most relevant regime After one month of exposure, the samples exposed to dry as it directly applies to the storage and potential reuse of air, wet air, and wet argon were photographed (Figure 2). this material in the nuclear stockpile. These photographs are the first recorded images of the Thus, many questions regarding the oxide layer remain, color changes that occur after plutonium has oxidized/ and the goal of this research has been to address several corroded at room temperature after one month. As seen key questions. What is the chemical composition of the in Figure 2, distinct color changes and oxide growth are corrosion products (PuO2, Pu2O3, PuO2+x, PuOx(OH)y)? observed [Manuscript in preparation]. Using previous work Is the surface layer oxide crystalline PuO2, an amorphous correlating the calculated spectral reflectivity (using SE) oxide, or a non-crystalline oxide? Is Pu2O3 present at and the color of the surface, the oxide layer thickness was the dioxide-metal interface? Why is a different rate ob- estimated [10, 11]. For dry air (A), the gold color is indica- served for unalloyed, alpha plutonium versus alloyed delta tive of a ~40 nm oxide layer. For wet air (B) and wet argon plutonium? The goal of this work was to answer these (B), the overall blue surface indicates the oxide layer is ~ 80 fundamental questions using a combination of surface - 100 nm, and the olive-green powder is characteristic of sensitive (SE, AES, XPS) and bulk techniques (XRD, EXAFS) PuO2 [12]. This visual inspection has allowed for an initial on a variety of sample types (i.e. unalloyed versus alloyed assessment regarding the extent of corrosion. The dry en- plutonium). However, in June 2013, the LANL director is- vironment is the least corrosive; the humid environments sued a pause of the facility where much of the work was are more corrosive, and the wet argon is most corrosive, to be performed (TA-55, PF-4). As a result, the work was producing the thickest oxide layer with the largest areas of refocused such that the goals of the project could still be green PuO2 powder. addressed using available materials and techniques. An EXAFS was used to investigate the bulk material from additional corrosion regime was added to the original both the gallium edge (providing structural information work scope (investigating the corrosion/radiogenic aging surrounding the gallium atoms) and the plutonium edge of materials as they were stored). The bulk and surface (which provides structural information surrounding the was characterized using EXAFS at an off-site location (the plutonium atoms). After 1 month, the gallium EXAFS Stanford Synchrotron Radiation Lightsource, SSRL) and the showed two metallic components corresponding to alpha’- surface was interrogated using XPS at another LANL facil- and delta-phase plutonium. The alpha’-phase commonly ity (TA-35, Bldg 0213, TFF, the Target Fabrication Facility). occurs after hand-polishing plutonium metal, but can be Most importantly, perhaps, a new plutonium capability removed with electropolishing. The electropolishing per- (SE) was developed and installed at TFF to complement the formed in this sample preparation clearly did not remove techniques used here. While XPS is sensitive to only the all of the stress-induced alpha’. Perhaps the most striking top 5 nm and EXAFS probes the bulk material up to 2. 5 result from the gallium EXAFS data was the presence of a microns, SE probes the bulk with extreme sensitivity to the gallium oxide. This has only rarely been observed in the ox- surface oxide layer. In this way a suite of techniques were ide layer, and it’s presence indicates that even though the used and developed to probe the oxide layer formed in plutonium alloy only contains 1.9 atomic percent gallium (2 ambient conditions. gallium atoms for every 50 plutonium atoms), the gallium Scientific Approach and Accomplishments is being oxidized. In the plutonium EXAFS data, two metal- lic phases were also observed (alpha’- and delta-phase The overall approach of this work was to study the am- plutonium). However, unlike the gallium data, the pluto- bient oxidation/corrosion of plutonium as a function of nium EXAFS data only showed indications of an oxide layer humidity, oxygen content, and exposure and study the (i.e. peaks with low amplitude that could not be modeled). structural and chemical components of the oxide layer us- Given the oxophilic nature of plutonium compared to ing multiple techniques (EXAFS, SE). This work investigated gallium and its visible presence (Figure 2), it likely follows the initial formation of the oxide layer on delta-phase plu- that if a gallium oxide is formed, a plutonium oxide would tonium alloyed with 1.9 at. % gallium exposed for 1 month be formed. Thus, the inability of EXAFS to detect the oxide to: 1) dry air (O2); 2) wet air (O2 + H2O); and 3) wet argon after 1 month indicates there is not a sufficient amount of (Ar + H2O). Experimental repeatability was tested using a 343

Page 344

plutonium oxide present, or it is very disordered and does method (PAD) is scheduled for January 2015. This first sam- not correspond to a crystalline plutonium oxide. ple is not only readily available, but is the most common room temperature form of a crystallographic plutonium ox- After 6 months of storage in a dry or argon (for the wet ide, providing a benchmark, or “standard” for PuO2 using argon sample only) environment, EXAFS and XPS data SE. A key feature that was included during assembly is the were acquired. Both the gallium and plutonium EXAFS data inclusion of gas sample bottles that will allow for introduc- showed a large increase in the amount of gallium and plu- tion of air or water, allowing for in-situ measurements of tonium oxide, in addition to small amounts of the original the oxidation and corrosion process. This capability has metallic surface. Both the gallium and plutonium oxides generated much interest in the weapons and dynamic test- were disordered in the EXAFS data and did not correspond ing community, and will hopefully result in future funding to crystalline oxides, as has been the historical interpre- for these studies. tation (Figure 1). The XPS data also corroborates these results and shows gallium and plutonium oxides. After 16 months, no observable metal fractions are observed in ei- ther the gallium or plutonium EXAFS, and only disordered oxides are observed. As in the 6 month data, only peaks associated with non-crystalline gallium and plutonium A. oxides were observed. Clearly, some local order is retained (up to 5 angstroms away from the plutonium/gallium atom). If the oxide was completely amorphous (no order), it would not be detectable by EXAFS. Finally, surface EXAFS (an experiment new to the PI) was performed at all time points (1 , 6, and 16 months). The preliminary results cor- roborated those detailed above, but highlighted the need for improved sample holder design and more calibrated B. experiments. C. Figure 2. Pu with 1.9 at. % Ga after one month of exposure to dry air (A); wet air (B), and wet argon (C). Overall, the results from this research indicate a non-crys- talline oxide layer (not c-Pu2O3, c-PuO2) is formed in the initial stages of corrosion. If removed from the corroding elements and left to corrode (and radiogenically age) for up to 16 months, this oxide will further corrode, but not Figure 1. Historical depiction of oxide products formed on delta- to a crystalline form of any known plutonium oxide (i.e. plutonium under ambient (left) and UHV (right) conditions (“” PuO2). This result is contrary to our historical view of the refers to the crystalline form). corrosion products, and has led to the modern depiction of the oxide products shown in Figure 3. Another interesting In order to compliment these studies, the last year of the result was the formation of a gallium oxide that has rarely project was focused on adding a new Pu capability (SE). been observed. These results, in addition to discussions The necessary equipment was acquired and assembled, in- with experts in the field, have led to the postulation that cluding all the necessary radiological safety components to the surface alpha’ layer is promoting gallium oxidation. allow for plutonium studies. Non-radiological testing was Gallium metal is not known to readily oxidize, and the performed on standards, and the first plutonium sample (a alpha’-phase is known to be a meta-stable phase in which PuO2 sample prepared by the Polymer Assisted Deposition 344

Page 345

gallium is trapped or ‘frozen’ in an alpha-phase plutonium emission of surface oxides and hydrides of delta pluto- lattice. If this is the surface that is readily available for nium. 2004. Surface Science. 571 (1-3): 74. oxidation, and gallium is intrinsically unstable in the alpha’- 3. Waber, J. T.. The corrosion and oxidation of metals. phase, the gallium will be thermodynamically favored to 1948. U.S. Atomic Energy Commision Report, LA-1381. oxidize. It is the PIs hope that programs will support future studies in this area so that the room temperature oxida- 4. Conradon, S. D., B. D. Begg, D. L. Clark, C. Den Auwer, F. tion and corrosion products and mechanism can be fully J. Espinosa-Faller, P. L. Gordon, N. J. Hess, R. Hess, D. W. understood. Two manuscripts are in preparation and will Keogh, L. A. Morales, M. P. Neu, W. Runde, C. D. Tait, be submitted in the upcoming months. D. K. Veirs, and P. M. Villella. Speciation and unusual reactivity in PuO2+x. 2003. Inorganic Chemistry. 42 x=0 (12): 3715. 5. Martin, P., S. Grandjean, M. Ripert, M. Freyss, P. Blanc, and T. Petit. Oxidation of plutonium dioxide: an x- PuO ray absorption spectroscopy study. 2003. Journal of 2-x Nuclear Materials. 320 (1-2): 138. 6. Raynor, J. B., and J. F. Sackman. Some observations on the oxidation of unalloyed and d-stabilized plutonium x=1 metal. 1965. In Plutonium 1965: Proceedings of the 3rd International Congress on Plutonium. ( , ). , p. 575. : Institute of Metals. δ-Pu 7. Costello, A. L., F. J. Espinosa-Faller, and S. D. Conrad- son. Corrosion in Ga-stabilized plutonium. Presented at Plutonium Futures – The Science 2012. (Cambridge, Figure 3. Modern depiction of the oxide layer formed on delta- UK, 15-20 July 2012 ). plutonium based on this research and recent work by our group [7,8]. 8. Pugmire, D. L., and H, G. G. Flores. The optical proper- ties of the Pu/Pu oxide thin film system. Presented at Impact on National Missions Plutonium Futures – The Science 2012. (Cambridge, Pu corrosion has been identified as a key issue in pit aging UK, 15-20 July 2012). and reuse, and this research has directly impacted our un- 9. Raynor, J. B., and J. F. Sackman. Oxidation of plutonium derstanding of the oxide layer that is formed in three ambi- in moist air and argon. 1963. Nature. 197 (4867): 587. ent environments relevant to the storage and reuse of this material. The corrosion process is more complicated than 10. Larson, D. T., and D. L. Cash. Plutonium oxide thickness originally thought. The oxide layer contains many non-

• color correlation. 1968. The Dow Chemical Company. crystalline components, particularly at the surface. This research has also allowed a new plutonium capability to be 11. Pugmire, D. L., F. J. Freibert, and J. P. Baiardo. The opti- developed that will be of interest to both the dynamic test- cal properties of the plutonium/plutonium oxide thin ing community and the weapons community in general. film system. 2013. Actinide Research Quarterly 2013. This research has fostered and strengthened LANL’s scien- August : 47. tific collaboration with another facility in the DOE complex, SSRL at the SLAC National Accelerator Laboratory in Menlo 12. Clark, D. L., S. S. Hecker, G. D. Jarvinen, and M. P. Neu. Park, CA, as well as with other experts in the EXAFS field at Plutonium. 2011. In The chemistry of the actinide Lawrence Berkeley National Laboratory. and transactinide elements. Edited by Morss, L. R., N. M. Edelstein, and J. Fuger. , p. 1. Dordrecht: Springer References Science+Business Media B.V..

1. Haschke, J. M., T. H. Allen, and L. A. Morales. Surface Publications and corrosion chemistry of plutonium. 2000. Los Ala- mos Science. 26: 252. Booth, C. H., Y. Jiang, S. A. Medling, D. L. Wang, A. L. Costello, D. S. Schwartz, J. N. Mitchell, P. H. Tobash, E.
2. Butterfield, M. T., T. Durakiewicz, J. J. Joyce, A. J. Arko, D. Bauer, S. K. McCall, M. A. Wall, and P. G. Allen. Self- K. S. Graham, D. P. Moore, and L. A. Morales. Photo- irradiation damage to the local structure of plutonium 345

Page 346

and plutonium intermetallics. 2013. JOURNAL OF AP- PLIED PHYSICS. 113 (9): -. Pugmire, A. L.. Spectroscopic Analysis of Actinyl Systems. 2013. Actinide Research Quarterly. (3): 18. Pugmire, A. L., C. H. Booth, T. J. Venhaus, and D. L. Pug- mire. Structural insights into the oxide formed during the room temperature corrosion of plutonium . Pre- sented at Pu Futures - The Science 2014. (Las Vegas, NV, 7-12 Sept 2014). Pugmire, A. L., C. H. Booth, T. J. Venhaus, and D. L. Pug- mire. Understanding the oxidation and corrosion of plutonium with EXAFS. Invited presentation at SSRL/ LCLS Users’ Meeting and Workshop. (Menlo Park, CA, 7 - 10 Oct. 2014). Pugmire, D. L., H. G. Garcia Flores, and A. L. Costello. The room temperature oxidation/corrosion of delta pluto- nium. Presented at XIII International Workshop: Fun- damental Properties of Plutonium. (Sarov, Russia, 8-13 Sept. 2013). Pugmire, D. L., H. Garcia Flores, A. L. Pugmire, and D. P. Moore. The room temperature oxidation/corrosion of delta-Pu: historical perspective vs. modern under- standing . Invited presentation at Pu Futures - The Sci- ence 2014. (Las Vegas, NV, 7-12 Sept 2014). 346

Page 347

Materials for the Future Early Career Research and Development Final Report Novel Mesoscale Modeling Approach for Investigating Energetically Driven Nanoscale Defect/Interface Interactions Abigail Hunter 20130745ECR Abstract in bulk and coarse-grained material counterparts, such The primary goal of this project is to investigate na- as deformation twinning [1-12]. As grain sizes reach noscale defect/interface interactions, such as nucle- ~30-50nm in face-centered cubic (fcc) materials, plastic- ation and growth of deformation twins, and dislocation ity becomes primarily driven by the motion and interac- dynamics as controlled by interfaces through develop- tion of partial (or extended) dislocations, which unlike ment of an innovative three-dimensional (3D) phase perfect dislocations leave behind stacking faults (SFs) field dislocation dynamics (PFDD) mesoscale model that in the crystal lattice [9, 13]. The increased number of links atomic-scale and nanoscale physics. Development individual partial dislocations and wide SFs are thought of this computational tool enables the study of deforma- to be responsible for the alternative deformation mecha- tion mechanisms at interfaces not possible with current nisms seen at the nanoscale [14-18]. These alternative mesoscale methods. deformation mechanisms, along with grain boundaries and bimetallic interfaces, can result in unique material This project proposed three key goals that all include in- behavior including high strength and improved ductility vestigation of a physical problem and novel extensions to [12, 19-23]. As applications for nano-devices increases, the PFDD model. These goals focus on (1) investigating there is not only a growing need to ensure reliability, but the formation of deformation twins and dependencies also the opportunity to engineer materials for specific with respect to temperature and grain size variations, applications, yet a predictive model to address all of (2) slip transmission of dislocations through bimetallic these issues still does not exist. interfaces, and (3) extending to deformation behavior in body-centered cubic (bcc) crystal structures. All goals Increased research efforts have been invested into were met successfully during the project; however, developing models that can accurately describe partial temperature dependence was not successfully included dislocation behavior. Many of the above observations in the PFDD framework. This project produced 8 jour- are from atomistic methods; however, these microscopic nal publications with 3 more in preparation that detail models are computationally intensive and limited in results including discovery of an unprecedented grain both time and length scales. Consequently, atomistic size effect and previously unreported twin formation attempts to address the issues of deformation twinning mechanism. This project has also fostered several col- and dislocation behavior at interfaces have only led to laborations between laboratory scientists and universi- controversy due to inappropriate time and length scales. ties including Purdue and Iowa State Universities. Three Conversely, classical continuum models focus on a mac- students were also funded at different times during the roscopic level accounting for dislocation densities rather duration of this project. than individual dislocations. While this is effective for modeling deformation processes in bulk material, on the Background and Research Objectives micron and sub-micron scales it is difficult to study dis- Plasticity in metals is mediated by the nucleation, mo- location interactions and many nanoscale deformation tion, and interaction of line defects in the crystalline behaviors fail to be predicted. The gap in this intermedi- lattice structure known as dislocations. In nanomaterials ate or mesoscale regime that lies between atomistic and (grain sizes less than 100nm), experimental studies and continuum scales has been recognized by Basic Energy atomistic simulations show that when the characteristic Sciences (BES), particularly in regards to the evolution length scale (ex: layer thickness, grain size) decreases and understanding of defect microstructure [24]. many deformation mechanisms arise that are not active 347

Page 348

Formulation of dislocation models, including discrete ture from the very beginning. In order to meet the other dislocation dynamics (DDD) and phase field models, aim project objectives, this was not completed. to fill this gap at the mesoscale. DDD models are a more Despite not including temperature dependence, this proj- developed approach for modeling deformation behavior ect still investigated important issues surrounding the for- at the mesoscale; however, they were originally developed mation of deformation twins in fcc metals. Specific focus for bulk crystals and have difficulty accounting for defor- was given to the impact of material properties and grain mation processes at interfaces and grain boundaries, and size variations on the formation of extended stacking faults deformation twinning [25]. DDD is also based on empirical (SFs) and two-layer deformation twins. While the primary rules, generally motivated by atomistic simulation. This focus of this goal changed slightly, valuable studies were basic framework impedes transformational changes, such completed resulting in several journal publications. The as accounting for temperature dependence and the effects remaining two goals focusing on fcc/fcc multilayers and bcc of surfaces and interfaces. On the other hand, phase field metals were fully met. models have been only recently applied to investigate dislocation behavior and interactions at the mesoscale Scientific Approach and Accomplishments [26, 27]. Phase field models are directly dependent on minimization of the total system energy, which is especially The 3D PFDD model is unique to Los Alamos National advantageous when accounting for partial dislocations, Laboratory (LANL), and has great potential to answer many deformation twinning, and interface interactions. Unlike questions about the effects of SFs, deformation twinning, DDD models, the phase field framework does not rely on and the impact of grain boundaries and heterophase rules or directly incorporate self-consistent atomic-scale interfaces on dislocation and material strength behavior. information to influence dislocation behavior. By tracking the motion of individual dislocations via a scalar-valued phase field variable defined on each active The goals of this project include innovative extensions of slip plane, phase field models can account for a number the 3D PFDD model that are not only of great interest to of dislocation behaviors, such as nucleation, annihilation, the scientific community, but also answer many questions and dislocation interaction with obstacles and externally about dislocation behavior on the nanoscale, particularly applied stresses [27-30]; however, at the beginning of the at interfaces. This project focuses on the following three project the PFDD framework was designed to study dislo- goals that each address (a) an interesting physical problem, cation characteristics of only fcc materials in bulk crystals and (b) a novel extension of the PFDD model: with a very limited description of some types of interfaces. The PFDD model does have a detailed energetic descrip- • Investigate the effects of grain size variation and tem- tion of the SF region between partial dislocations, which perature dependence on deformation twinning in fcc is in contrast to most other predictive dislocation models materials, such as copper, nickel, and silver. (b) Include that rely on only one parameter [31-35]. This energetic de- temperature dependence in the PFDD model. scription for the SF region is directly informed by atomis- tics [36, 37]. • Study dislocation behavior in fcc/fcc multilayers. (b) Implement a complex description of bimaterial inter- Deformation Twinning faces that includes both local interface effects (lattice As discussed above, temperature effects were not studied mismatch, residual Burgers vectors) and bulk effects as initially outlined by this project, but rather the effects (elastic moduli mismatch) in the PFDD model. of grain size variations and material dependence. Experi- mental results and numerical simulations demonstrate that • Account for dislocation behavior in bcc materials, such partial dislocations are the basic defect responsible for de- as tantalum and niobium. (b) Reformulate the PFDD formation twins, which occur when SFs form on adjacent model to account for the appropriate bcc slip systems, planes [1, 3, 19, 38-40]. Hence, to fully understand defor- Burgers vector geometry, and dislocation mobility. mation twin formation we also investigated extended SFs. Overall, all of these goals were met with one exception. Figure 1 shows a PFDD simulation of a leading partial dislo- Temperature dependence has still not been included in cation emerging from a grain boundary defect, propagating the PFDD framework. Preliminary simulations in which into the grain until the trailing dislocation in emitted in a the elastic moduli were changed to reflect temperature 31.87nm grain in nickel. The green region represents the changes did not impact results as expected. We concluded SF. Energetically, the SF will reach a maximum (or critical) that in order to accurately account for temperature the width at which it becomes more energetically favorable to fundamental formulation of the phase field framework emit the trailing partial rather than continue to expand the needed to be developed in a way that includes tempera- SF. 348

Page 349

ly high. If both conditions are fulfilled, leading partials on [121]% adjacent planes can propagate stable SFs across the entire grain cross-section without exceeding the energy penalty Leading# to provoke emission of a trailing partial. Achieving both cri- Par3al# Trailing# teria translates to a high propensity for twinning, which we w max# Par3al# Stacking#Fault# see most strongly in silver and copper. Overall, this study A#<#A max ## A#=#A max# Leading# resulted in 8 journal publications in collaboration with Dr. Par3al# Irene Beyerlein (LANL). 400#s# 450#s# 500#s# [101]% [112]% Figure 1. PFDD simulation of partial dislocation motion in a Cu! Ni! Cu! Ni! Cu! Ni! 31.87nm nickel grain. The leading partial emits from a grain boundary defect and extends into the grain until it becomes more energetically favorable to nucleate a trailing partial than to con- tinue to expand the SF (green region). The SF width at this point is the critical or maximum SF width. The trailing partial emits a! b! c! from the same grain boundary defect. [110]% We ran several simulations similar to Figure 1 for grain Figure 2. PFDD simulation of two-layer twin formation via the sizes varying from 10-120nm for several fcc metals includ- AE mechanism in a 65.8nm copper grain [44]. Frames (a) and (b) show the topmost glide plane, and frames (c) and (d) present ing nickel, aluminum, and copper [41]. Interestingly we cross-sectional views of the grain. The SFs are shown in green, discovered a previously unreported grain size effect in and hashed regions represent grain boundaries. Grain bound- which the critical SF width increases with increasing grain ary defects are on opposing grain boundaries and adjacent glide size [41, 42]. At first, this finding appeared to contradict planes. prior results from experiments [15] and atomistic simula- tions [17, 43], which generally report that SF widths are Fcc/fcc Multilayers wider in smaller grains. Our study shows that as the grain Thickness-dependent strengthening phenomena have size increases, the critical SF width also increases, but the been observed in bimetallic multilayers, and the interfaces fraction of the grain cross-sectional area taken up by the between the metals drastically impact the strengthen- critical SF region decreases. These grain size effects are ing mechanisms [21]. Furthermore, the strengthening intriguing and cannot be explained by any model to date. mechanisms, such as dislocation pile-ups, elastic moduli mismatch, and misfit dislocations, themselves are not fully Additionally, we found that monolayer SF formation (when understood and have been a recent topic of study and de- the SF covers the entire grain cross-section without trail bate [21]. Modeling dislocation behavior across nanolayer emission) is a necessary but insufficient condition for twin interfaces had never been attempted with a phase field formation [44]. This led to the discovery of another likely model prior to this project. elementary mechanism for twin formation called the alternate emission (AE) mechanism [44]. Many twinning The problem of heterophase interfaces is very complex. processes have been proposed most of which assume that In an effort to simplify the initial simulations only perfect the partial dislocations emit from the same grain bound- dislocations (no partial dislocations) were considered in ary [3, 45, 46]. Conversely, the AE mechanism considers a cube-on-cube orientation (no misorientation/rotation participation of two grain boundaries on opposite sides of between the fcc metals). Model extensions focused on ac- a nanoscale grain. Figure 2 shows the formation of a two- counting for bulk effects, such as elastic moduli mismatch. layer deformation twin in a 65.8nm grain of copper via the Furthermore, energetic expressions for local interface AE mechanism [44]. This mechanism may be energetically properties, such as lattice parameter mismatch and re- favorable over other proposed mechanisms because it ac- sidual Burgers vectors, were included in the PFDD model. commodates deformation without generating detrimental backstresses. The PFDD model can now model slip transmission of per- fect dislocations through fcc/fcc bimaterial interfaces for a This physical picture translates into two conditions that can variety of systems with cube-on-cube orientation. Figure help determine whether or not a two-layer twin will form 3 shows an example simulation for a perfect dislocation via AE. First, the material system should be sufficient to moving from copper to nickel. The green line represents accommodate a monolayer SF. Second, the energy re- the dislocation that approaches, meets, and then passes quired to nucleate a trailing partial must also be sufficient- through the interface (thin black line). The red point at the 349

Page 350

interface of Figure 3c represents the residual Burgers vec- system. We considered an infinite medium with a single tor in the interface. dislocation loop in niobium (Figure 4). The dislocation loop includes edge, screw, and mixed-type orientations. Since different parts of the loop have different mobilities [112]% the loop will not expand evenly across the silp plane, as Cu! Ni! Cu! Ni! Cu! Ni! the PFDD simulations show in Figure 4. The edge oriented dislocation lines move much faster creating an oval shape. For comparison, Figure 4 also shows results for the same configuration in copper, a fcc metal. In this case, the loop expands evenly because there is no notable different be- a! b! c! [110]% tween edge and screw mobilities. Niobium’ Figure 3. PFDD simulation of a perfect dislocation approach- ing, meeting, and transmitting through a copper-nickel interface. The dislocation line is green and the red point at the interface in frame (c) represents the residual Burgers vector. Similar simulations have been completed for various fcc/ 200 s 400 s 500 s fcc bimaterial systems. This work has led to the formu- [111]$Coppe{r}^{\prime }lationofananalyticalsolutionfortheminimumstressrequiredforsliptransmissiontooccur(i.e.thethresholdstress).Notably,wecananalyticallyinvestigatethedepen-denceofthethresholdstressonkeymaterialproperties.Thiscanalsobeusedforpredictivemodeling.Thiswork200s400swascompletedincollaborationwithProf.MarisolKo-[110]$ slowski and Yifei Zeng (Purdue University) and publications are in preparation. Future research includes extending to Figure 4. PFDD simulations of an expanding perfect disloca- partial dislocations, systems with misorientation, and sub- tion loop in niobium and copper. Green lines represent the sequent slip transmission studies. dislocation line, red indicates slipped regions, and blue regions are unslipped areas. In niobium the dislocation loop does not Bcc Materials expand evenly due to the difference in mobilities between edge Dislocation behavior in bcc materials had never been and screw oriented dislocation lines. There is no such difference studied using a phase field model prior to this project and in the edge and screw mobilities in fcc metals, hence the loop in was a major step for the PFDD framework. Accounting for copper expands evenly. bcc materials required reformulation of much of the PFDD This extension of the PFDD model to bcc metals creates algorithm. Most notable modifications include redefin- opportunities for many studies that are of interest to the ing dislocation and system geometry (Burgers vectors scientific community including investigation of kink and and slip systems), and describing the dislocation core in cross-slip mechanisms and fcc/bcc interfaces. This work bcc materials. In bcc materials, edge and screw oriented was completed in collaboration with Rebecca Runnels dislocations have very different core structures [47]. Edge (New Mexico Tech) and Dr. Irene Beyerlein (LANL) and pub- dislocations have planer core structures, similar to fcc lications are in preparation. materials, and move very quickly through the crystal lattice [47]. Conversely, screw dislocations have non-planar core Impact on National Missions structures, meaning that the core region can be spread A primary mission of the Advanced Simulations and Com- across non-parallel crystallographic planes [47]. This puting (ASC) Program is to provide simulation capabilities causes screw-type defects to be less mobile and produce required to maintain a safe, secure and reliable stockpile. larger amounts of inelastic strain making them key to the While LANL has advanced atomistic and continuum simu- dominant deformation mechanisms in bcc materials [47, lation tools, it is not clear how to connect information 48]. Clearly, there are fundamental differences between on these two very different length scales; hence, there line defects in fcc and bcc materials that cannot be ignored remains a computational gap at the mesoscale that trans- when modeling plasticity in bcc materials. lates into a gap in our understanding of material deforma- Initially, we approached a simplified problem that consid- tion. The PFDD model aims to fill this gap, and produce ered only perfect dislocations moving in a single active slip predictive multiscale simulations crucial for stockpile stew- 350

Page 351

ardship. Furthermore, understanding the effects of defects 10. Yamakov, V., D. Wolf, S. R. Phillpot, A. K. Mukherjee, on material behavior as materials age becomes increasing- and H. Gleiter. Deformation-mechanism map for nano- ly important for determining the performance, reliability, crystalline metals by molecular-dynamics simulation. and safety of weapon systems. The PI is intimately involved 2004. Nature Materials. 3 (1): 43. with developing and implementing material models for 11. Zhu, Y. T., X. Z. Liao, S. G. Srinivasan, Y. H. Zhao, M. Integrated Codes and Physics and Engineering Models. This I. Baskes, F. Zhou, and E. J. Lavernia. Nucleation and work is also in line with anticipated DOE/BES proposal calls growth of deformation twins in nanocrystalline alumi- on mesoscale modeling [24]. num. 2004. Applied Physics Letters. 85 (21): 5049. References 12. Zhao, Y. H., Y. T. Zhu, X. Z. Liao, Z. Horita, and T. G. Lang-

1. Chen, M. W., E. Ma, K. J. Hemker, H. W. Sheng, Y. M. don. Tailoring stacking fault energy for high ductility Wang, and X. M. Cheng. Deformation twinning in and high strength in ultrafine grained Cu and its alloy. nanocrystalline aluminum. 2003. Science. 300 (5623):
2. Applied Physics Letters. 89 (12): 121906.
4. Hirth, J. P., and J. Lothe. Theory of dislocations. 1982.
5. Farkas, D., S. Van Petegem, P. M. Derlet, and H. Van Swygenhoven. Dislocation activity and nano-void for-
6. Cockayne, D. J., M. L. Jenkins, and I. L. Ray. Measure- mation near crack tips in nanocrystalline Ni. 2005. Acta ment of Stacking-Fault Energies of Pure Face-Centred Materialia. 53 (11): 3115. Cubic Metals. 1971. Philosophical Magazine. 24 (192):
8. Liao, X. Z., F. Zhou, E. J. Lavernia, S. G. Srinivasan, M. I. Baskes, D. W. He, and Y. T. Zhu. Deformation mecha-
9. Liao, X. Z., S. G. Srinivasan, Y. H. Zhao, M. I. Baskes, Y. nism in nanocrystalline Al: Partial dislocation slip. T. Zhu, F. Zhou, E. J. Lavernia, and H. F. Xu. Formation
10. Applied Physics Letters. 83 (4): 632. mechanism of wide stacking faults in nanocrystalline Al. 2004. Applied Physics Letters. 84 (18): 3564.
11. Schiotz, J., and K. W. Jacobsen. A maximum in the strength of nanocrystalline copper. 2003. Science. 301
12. Schoeck, G.. The cross-slip energy unresolved. 2009. (5638): 1357. Philosophical Magazine Letters. 89 (8): 505.
13. Shan, Z. W., E. A. Stach, J. M. Wiezorek, J. A. Knapp, D.
14. Yamakov, V., D. Wolf, M. Salazar, S. R. Phillpot, and H. M. Follstaedt, and S. X. Mao. Grain boundary-mediated Gleiter. Length-scale effects in the nucleation of ex- plasticity in nanocrystalline nickel. 2004. Science. 305 tended dislocations in nanocrystalline Al by molecular- (5684): 654. dynamics simulation. 2001. Acta Materialia. 49 (14):
16. Swygenhoven, H. Van, P. M. Derlet, A. G. Froseth, S. Van Petegem, Z. Budrovic, and A. Hasnaoui. Nano-
17. Zhao, Y. H., X. Z. Liao, Y. T. Zhu, Z. Horita, and T. G. crystalline fcc metals: bridging experiments with Langdon. Influence of stacking fault energy on nano- simulations. 2004. Nanoscale Materials and Modeling- structure formation under high pressure torsion. 2005. Relations among Processing, Microstructure and Me- Materials Science and Engineering A-Structural Materi- chanical Properties. 821: 285. als Properties Microstructure and Processing. 410: 188.
18. Swygenhoven, H. Van, M. Spaczer, and A. Caro. Micro-
19. Yamakov, V., D. Wolf, S. R. Phillpot, A. K. Mukherjee, scopic description of plasticity in computer generated and H. Gleiter. Dislocation processes in the deforma- metallic nanophase samples: A comparison between tion of nanocrystalline aluminium by molecular-dy- Cu and Ni. 1999. Acta Materialia. 47 (10): 3117. namics simulation. 2002. Nature Materials. 1 (1): 45.
20. Swygenhoven, H. Van, M. Spaczer, A. Caro, and D.
21. Liao, X. Z., F. Zhou, E. J. Lavernia, D. W. He, and Y. T. Farkas. Competing plastic deformation mechanisms in Zhu. Deformation twins in nanocrystalline Al. 2003. Ap- nanophase metals. 1999. Physical Review B. 60 (1): 22. plied Physics Letters. 83 (24): 5062.
22. Vo, N. Q., R. S. Averback, P. Bellon, S. Odunuga, and A.
23. Liu, Y., D. Bufford, H. Wang, C. Sun, and X. Zhang. Me- Caro. Quantitative description of plastic deformation chanical properties of highly textured Cu/Ni multilay- in nanocrystalline Cu: Dislocation glide versus grain ers. 2011. Acta Materialia. 59 (5): 1924. boundary sliding. 2008. Physical Review B. 77 (13):
24. 22. Zhang, J. Y., G. Liu, X. Zhang, G. J. Zhang, J. Sun, and E. 351

Page 352

Ma. A maximum in ductility and fracture toughness in 35. Wang, Z. Q., and I. J. Beyerlein. Stress orientation and nanostructured Cu/Cr multilayer films. 2010. Scripta relativistic effects on the separation of moving screw Materialia. 62 (6): 333. dislocations. 2008. Physical Review B. 77 (18): 184112. 23. Zhang, J. Y., X. Zhang, G. Liu, G. J. Zhang, and J. Sun. 36. Hunter, A., I. J. Beyerlein, T. C. Germann, and M. Ko- Scaling of the ductility with yield strength in nanostruc- slowski. Influence of the stacking fault energy surface tured Cu/Cr multilayer films. 2010. Scripta Materialia. on partial dislocations in fcc metals with a three- 63 (1): 101. dimensional phase field dislocations dynamics model. 2011. Physical Review B. 84 (14): 144108. 24. From Quanta to the Continuum: Opportunities for Me- soscale Science. 2012. Basic Energy Science Advisory 37. Shen, C., and Y. Wang. Incorporation of gamma-surface Committee Mesoscale Science Subcommittee . to phase field model of dislocations: simulating dislo- cation dissociation in fcc crystals. 2004. Acta Materia- 25. El-Awady, J. A., S. B. Biner, and N. M. Ghoniem. A self- lia. 52 (3): 683. consistent boundary element, parametric dislocation dynamics formulation of plastic flow in finite volumes. 38. Ogata, S., J. Li, and S. Yip. Energy landscape of defor- 2008. Journal of the Mechanics and Physics of Solids. mation twinning in bcc and fcc metals. 2005. Physical 56 (5): 2019. Review B. 71 (22): 244102. 26. Wang, Y. U., Y. M. Jin, A. M. Cuitino, and A. G. Khach- 39. Venables, J. A.. Electron Microscopy of Deformation aturyan. Nanoscale phase field microelasticity theory Twinning. 1964. Journal of Physics and Chemistry of of dislocations: Model and 3D simulations. 2001. Acta Solids. 25 (7): 685. Materialia. 49 (10): 1847. 40. Christian, J. W., and S. Mahajan. Deformation Twin- 27. Koslowski, M., A. M. Cuitino, and M. Ortiz. A phase- ning. 1995. Progress in Materials Science. 39 (1-2): 1. field theory of dislocation dynamics, strain hardening 41. Hunter, A., and I. J. Beyerlein. Stacking fault emission and hysteresis in ductile single crystals. 2002. Journal from grain boundaries: Material dependencies and of the Mechanics and Physics of Solids. 50 (12): 2597. grain size effects. 2014. Materials Science and Engi- 28. Hunter, A., and M. Koslowski. Direct calculations of ma- neering A. 600: 200. terial parameters for gradient plasticity. 2008. Journal 42. Hunter, A., and I. J. Beyerlein. Unprecedented grain of the Mechanics and Physics of Solids. 56 (11): 3181. size effect on stacking fault width. 2013. Applied Phys- 29. Koslowski, M.. Scaling laws in plastic deformation. ics Letters Materials. 1: 032109. 2007. Philosophical Magazine. 87 (8-9): 1175. 43. Swygenhoven, H. Van, P. M. Derlet, and A. Hasnaoui. 30. Koslowski, M.. Effect of grain size distribution on Atomic mechanism for dislocation emission from nano- plastic strain recovery. 2010. Physical Review B. 82 (5): sized grain boundaries. 2002. Physical Review B. 66 (2): 054110. 024101. 31. Asaro, R. J., P. Krysl, and B. Kad. Deformation mecha- 44. Hunter, A., and I. J. Beyerlein. Predictions of an alter- nism transitions in nanoscale fcc metals. 2003. Philo- native pathway for grain-boundary driven twinning. sophical Magazine Letters. 83 (12): 733. 2014. Applied Physics Letters. 104 (23): 233112. 32. Douin, J., F. Pettinari-Sturmel, and A. Coujou. Dissoci- 45. Wu, X. L., X. Z. Liao, S. G. Srinivasan, F. Zhou, E. J. Laver- ated dislocations in confined plasticity. 2007. Acta nia, R. Z. Valiev, and Y. T. Zhu. New deformation twin- Materialia. 55 (19): 6453. ning mechanism generates zero macroscopic strain in nanocrystalline metals. 2008. Physical Review Letters. 33. Gu, P., B. K. Kad, and M. Dao. A modified model for 100 (9): 095701. deformation via partial dislocations and stacking faults at the nanoscale. 2010. Scripta Materialia. 62 (6): 361. 46. Yamakov, V., D. Wolf, S. R. Phillpot, and H. Gleiter. De- formation twinning in nanocrystalline Al by molecular 34. Martinez, E., J. Marian, A. Arsenlis, M. Victoria, and J. dynamics simulation. 2002. Acta Materialia. 50 (20): M. Perlado. Atomistically informed dislocation dynam- 5005. ics in fcc crystals. 2008. Journal of the Mechanics and Physics of Solids. 56 (3): 869. 47. Wang, Z. Q., and I. J. Beyerlein. An atomistically- 352

Page 353

informed dislocation dynamics model for the plastic anisotropy and tension-compression asymmetry of BCC metals. 2011. International Journal of Plasticity. 27 (10): 1471. 48. Zhang, R. F., J. Wang, I. J. Beyerlein, and T. C. Germann. Twinning in bcc metals under shock loading: a chal- lenge to empirical potentials. 2011. Philosophical Magazine Letters. 91 (12): 731. Publications Beyerlein, I. J., J. S. Carpenter, A. Hunter, L. S. Toth, and W. Skrotzki. Nano-enabled orientation alignment via ex- treme shear strains. To appear in Scripta Materialia. Cao, L., A. Hunter, I. J. Beyerlein, and M. Koslowski. The role of partial mediated slip during quasi-static de- formation of 3D nano crystalline metals. To appear in Journal of the Mechanics and Physics of Solids. Hunter, A., R. F. Zhang, I. J. Beyerlein, T. C. Germann, and M. Koslowski. Dependence of equilibrium stacking fault width in fcc metals on the gamma-surface. 2013. MODELLING AND SIMULATION IN MATERIALS SCIENCE AND ENGINEERING. 21 (2): -. Hunter, A., R. F. Zhang, and I. J. Beyerlein. The core struc- ture of dislocations and their relationship to the mate- rial g-surface. 2014. Journal of Applied Physics. 115: 134314:1. Hunter, A., and I. J. Beyerlein. Unprecedented grain size ef- fect on stacking fault width. 2013. Applied Physics Let- ters Materials. 1: 032109. Hunter, A., and I. J. Beyerlein. Stacking fault emission from grain boundaries: Material dependencies and grain size effects. 2014. Materials Science & Engineering A. 600: 200. Hunter, A., and I. J. Beyerlein. Predictions of an alternative pathway for grain-boundary driven twinning. 2014. Ap- plied Physics Letters. 104: 233112:1. Hunter, A., and I. J. Beyerlein. Relationship between mono- layer stacking fault and twins in nano crystals. To ap- pear in Acta Materialia. 353

Page 354

Materials for the Future Early Career Research and Development Final Report Magnetic Field Effects on Convection-Modified Solid-Liquid Interfaces Amy J. Clarke 20130755ECR Abstract solute, potentially leading to enhanced macro- and mi- In-situ characterization techniques are now affording crosegregation (i.e., chemical inhomogeneity), reduced direct interrogations of materials during synthesis and casting quality, and microstructural variations. processing, in part because of first experiments per- Specific types of micro- and macroscopic fluid flows formed by this team to monitor metal alloy melting and may be achieved during casting and solidification by solidification with x-rays and protons. X-rays permit controlling the geometry and heat flux. In addition to microstructural examinations of small fields-of-view in these variables, the application of a magnetic field may thin metal sections, whereas protons allow for imaging constitute a way to further influence alloy melt fluid of larger fields-of-view in thin and thick metal sections. flows and solidification structural development. For In this project, the influence of static magnetic field on example, Imhoff et al. have applied a 35 Tesla magnetic tin-bismuth alloys during solidification was explored, field to metal alloys during solidification, which resulted demonstrating a new approach to further control micro- in microstructures with enhanced crystallographic structural evolution and fluid flow. Dendritic pattern de- anisotropy and different phase morphologies compared velopment at the solid-liquid interface and macro-scale to those solidified without a magnetic field [1]. Weak fluid flow during solidification in a static magnetic field magnetic field has also been used to create zirconia- and were examined with real-time x-ray imaging at Argonne alumina-magnetite aligned solidification structures and National Laboratory’s Advanced Photon Source (APS) aligned titania-magnetite scaffolds that exhibit substan- and 800 MeV proton radiography (pRad) at Los Ala- tially higher strengths than similar (unaligned) structures mos National Laboratory, respectively. Proton imaging produced without a magnetic field [2]. In early work, revealed the macroscopic influences of magnetic field Li et al. reported increasing primary dendritic spacing on casting mold filling, whereas x-ray imaging captured with increasing magnetic field intensity from 0 to 14 early-stage microscopic dendritic growth. From this Tesla and larger dendritic spacing for a lower solid-liquid initial exploration of the effect of static magnetic field on interface velocity and a given magnetic field intensity metal alloy solidification, it was determined that pri- for an aluminum-4.5 wt.% copper alloy [3]. However, in mary dendritic spacing decreased with a static magnetic more recent work Li et al. reported decreasing primary field. Qualitative differences in casting mold filling and dendritic spacing with increasing magnetic field intensity alloy melt flow behaviors were also observed with a from 0 to 1 Tesla and larger primary dendritic spacing static magnetic field, compared to the results obtained for a higher solid-liquid interface velocity and a given without one. This work highlights the potential use of magnetic field intensity [4]. static magnetic field as an effective way to modify metal alloy solidification at the micro- and macroscopic length The effect of a weak (0.08 Tesla) transverse magnetic scales. field on convection and microstructural evolution in alu- minum-4 wt.% copper and aluminum-10 wt.% copper al- Background and Research Objectives loys was examined by Li et al. with in-situ x-ray imaging. Metal alloy melt flows interact with solid interfaces at a The observed microscopic flows were reportedly linked variety of length scales, from the microscopic scale den- to the microstructural evolution and the final micro- dritic tips that develop at the solid-liquid interface to the structure [4]. Apart from post-mortem microstructural macroscopic interactions with the casting mold during examinations and the qualitative in-situ x-ray imaging filling. The flows that arise in the alloy melt during cast- examinations described above, quantitative and in-situ ing and solidification will influence the distribution of 354

Page 355

characterizations of the effect of weak magnetic field on assessed. microstructural evolution are lacking. For both the x-ray and proton imaging experiments, At the macroscopic scale, electromagnetic breaking (EMBr) permanent neodymium-iron-boride magnets were used is used to control alloy melt fluid flow during continuous to provide a uniform static magnetic field. During casting casting, as flow control is critical for avoiding the entrap- and solidification, one magnet was placed on either side ment of inclusions and achieving the desired heat transfer of the metal alloy sample, as shown schematically Figure and solidification behavior. Visualizing the complex flows 1a. Ideally, the field strength between two the two large produced by EMBr is challenging, but large-scale eddy cur- parallel magnets would be constant, however the required rent simulations have been used to model these flows for spacing needed to achieve this would impede the ability continuous casting scenarios with EMBr and insulating or to conduct the experiments. Therefore, the field strength conducting walls [5]. Experimental observations of macro- between the two magnets varied with position, and the scopic fluid flow during metal casting are challenging and magnet strength and positioning were chosen after sys- are typically only simulated with water modeling. Liu et al. tematic calculations that directly aided in the experimental recently visualized macroscopic fluid flows during continu- design to provide the proper field strength profile. The ous casting without and with air gas injection by water result of these calculations for the casting experiments is modeling [6]. Although computational and experimental also shown in Figure 1a. A diminishing strength is observed simulations of macroscopic flows during metal casting have toward the center of the experimental set-up, and must been performed, direct observations of metal casting, es- be accounted for when selecting the experimental param- pecially those that include variables like magnetic field, are eters. nonexistent at scales relevant for industry. Turbulent flow and convection in the liquid may be altered by changing the effective liquid viscosity by the application of magnetic field [7], perhaps changing the local flow con- ditions at the solid-liquid interface to impact microstruc- tural evolution or the large-scale flows encountered during casting mold filling and subsequent solidification. We use x-ray and proton imaging to make movies during solidifica- (a) (b) tion to explore the potential of static magnetic field for the modification of alloy melt fluid flow and the forma- Figure 1. (a) The geometry used for the proton imaging experi- tion and growth of solid-liquid interface patterns (e.g., the ments of casting mold filling with and without the application formation and growth of dendrites). Understanding how of a static magnetic field and the magnetic field in the indicated static magnetic field may be used to control fluid flow and cross-section and (b) Grashof (Gr) and Hartmann (Ha) numbers microstructural development in metal alloys will enable indicating the regimes of unstable and stable flow with the ap- improved casting design and quality and the creation of plication of a static magnetic field. intended microstructures. For the experiments, the potential effect of field interac- Scientific Approach and Accomplishments tions on flow must be considered. For magneto-hydrody- The role of magnetic field on macro-scale alloy melt fluid namic flow, there are two dimensionless numbers we used flow and micro-scale solid-liquid interface pattern develop- to predict the magnetic damping of thermal convection ment has been examined using real-time x-ray and proton in field, namely the Hartmann and Grashof numbers. The imaging at Argonne and Los Alamos National Laboratories, Hartmann number represents the amount of magnetic respectively. X-rays permit microstructural examinations damping and is the ratio of the induced magnetic force of a small field-of-view in thin metal sections, whereas to the viscous force. In contrast the Grashof number is a protons allow for imaging of larger fields-of-view in thin predictor of the strength of the convection forces and is and thick metal sections. Proton imaging was used to calculated by taking the ratio of the thermosolutal buoy- measure the macro-scale influences of a magnetic field on ant force and the liquid’s viscous force. At high Hartmann casting mold filling and x-ray imaging was used to observe numbers, the thermosolutal convection is sufficiently initial microstructural development in the presence of a damped to have little direct influence on local structural magnetic field. These complementary experiments enable development; while at high Grashof numbers, there is suf- the relative effectiveness of the imposed magnetic field ficient buoyant driving force to transition to significant con- to modify fluid flow and structural evolution to be directly vection cell development. When both thermosolutal and magnetic forces are at play, the destabilization of the liquid 355

Page 356

(i.e. development of convection cells) can be predicted by age where a dendrite had traversed half the FOV. Second- balancing both effects. The calculated regions of stabil- ary dendrite spacings were measured near the tip of the ity and instability are plotted in Figure 1b. By imposing a primary dendrite arms. sufficiently large magnetic field on a sample, the Hartman number may be increased to eliminate undesired convec- tion in a system. For the present experiments, the size, strength and location of the magnets with respect to the area of interest for imaging were selected to sufficiently modify the flow behavior during metal alloy casting and solidification. For the relevant geometries, this required an effective magnetic field minimum of ~1200 gauss. The experiments were designed so that the saddle point in the field strength was no lower than this limit. The experi- ments presented here were performed using a tin-27 at.% bismuth alloy to take advantage of the high x-ray absorp- tion contrast between the tin-rich initial solid and the bismuth enriched liquid during solidification. (a) (b) Synchrotron x-ray imaging was performed at the Sector 32 Insertion Device beamline at Argonne National Labora- Fi gure 2. X-ray images from directional solidification of a tin-27 tory’s Advanced Photon Source. For imaging, a monochro- at.% bismuth alloy parallel to gravity with (a) no externally matic 28 keV x-ray beam was used. X-rays pass through applied magnetic field and (b) an 0.12 T static magnetic field a metal foil sample and impinged upon scintillator where orthogonal to the nominal dendritic growth direction. they were converted to visible light. The movie camera For the directional solidification experiments without captured 1024 x 1280 pixel images at an approximate and with the static magnetic field, a thermal gradient (G) frame rate of 0.85 Hz, with a pixel size of 1.3 microns. of 27 K/cm and a solid-liquid interface velocity (V) of 20 The imaging field of view (FOV) was ~1.4 x 1.7 mm. Each microns/s and a G of 27 K/cm and a V of 15 microns/s metal alloy sample was placed between two boron nitride were determined, respectively. In the presence of the crucible halves with a ~0.1 mm thick window and then static magnetic field, overall refinement of the dendritic inserted into a slotted steel rod with a 5 mm through hole structure was observed (i.e., a decrease in primary den- perpendicular to the flat plane of the crucible to accom- dritic spacing from 288 to 200 microns). Reduced den- modate the beam path. The images were post processed dritic spacing is essential for improving the properties and to remove the background and normalize the gray-scale performance of as-cast parts. The appearance of dendritic to a constant reference; all image analysis was performed side branches also changed with application of the static using ImageJ software. Primary dendrite tip velocities were magnetic field. The branches became more frequent, but measured using manual tracking in ImageJ, beginning with less prone to further growth, implying that the fluid flow the frame in which the tip of the first dendrite entered the and mixing adjacent to the dendrite arms was effectively FOV, and ending when the dendrite exited the FOV. suppressed. However, it must be recalled that this system Direct imaging of solidification with x-rays enabled the is not over damped. In continuous casting where EMBr measurement of branched dendritic structures that is employed, the forced convection velocities are high so formed as the first solid in the metal alloy. Figure 2 shows the effect of convection on the microstructure may be in representative images from the solidification sequence the form of remelting solidified regions as the liquid jets without an imposed magnetic field (Figure 2a) and in a impinge upon a cooler region. The effect here is less dra- 0.12 T static magnetic field orthogonal to the vertical axis matic, but has a strong effect on the local environments. of the image (Figure 2b). Individual dendrite velocities Reduction in liquid velocity requires that both solute and were measured multiple times and the average tip velocity thermal boundary layers near the growth front of the was determined. The primary and secondary dendrite arm dendrites will be more spread out than without magnetic spacing for the alloys was also determined. The center- damping. Most importantly, this means that the solute to-center distance between adjacent dendrite arms was is enriched near the dendrite tips, thereby changing the measured, and these values were averaged over multiple side branching environment and consequently the primary measurements to determine the overall spacing. The spacing. These findings fit into a larger overall picture of dendrite spacing measurements were performed on an im- convection effects on morphology development during solidification, supporting the trend reported by Li et al. of 356

Page 357

decreasing primary dendritic spacing with increasing mag- surements of the fluid stream and initial splashing during netic field intensity and larger primary dendritic spacing for the filling of the mold cavity to be made. Increased width a higher solid-liquid interface velocity and a given magnetic of the metal alloy stream was observed with the static field intensity [4]. magnetic field, suggesting that the velocity of the falling liquid was effectively decreased. In addition, less splashing For the pRad imaging experiments, a plate geometry was and droplet separation were observed during the introduc- selected for the casting mold filling experiments to limit tion of the metal alloy into the cavity, resulting in improved projection artifacts from overlapping features. Graphite filling behavior, as shown in Figure 4, and a decrease in the was used for the mold material; no coating was applied to susceptibility of the casting to surface defects that could the inside surfaces of the mold cavity. The casting mold influence the casting quality. Subtle changes in the liquid was directly filled from the downsprue, without a gating surface shape during filling were also observed with ap- system. A funnel shaped crucible was placed on top of the plication of the static magnetic field. graphite mold, and a pure bismuth plug was pre-melted in to the bottom of the crucible. A 40 g charge of a tin 27at.% bismuth alloy was loaded into the crucible. During heat- ing, the alloy charge melted first. When the bismuth plug finally melted, the alloy melt was released into the mold cavity. This enabled relatively repeatable pouring condi- tions, without mechanical control within the radiation enclosure of the proton beamline. A thinned window area was machined into the sides of the mold walls to facilitate the proton imaging. The mold stack was clamped in a water-cooled aluminum stand to develop a large gradient in the vertical axis of the mold and promote directional (a) (b) solidification. Each mold was instrumented with sheathed 0.5 mm diameter thermocouples to provide temperature Figure 3. Proton images of a casting mold filling of a tin-27 monitoring and feedback control. The thermocouples at.% bismuth alloy with (a) no externally applied magnetic were placed approximately 1.5 mm away from the plate field (20 Hz imaging rate) and (b) a 0.12 T static magnetic field aligned orthogonal to gravity (10 Hz imaging rate). cavity directly below the downsprue, near each corner of the plate. The crucible and top of the mold were heated using independently controlled induction heaters. Ad- ditional thermocouples were placed in the crucible above the mold cavity and adjacent to the downsprue to monitor the crucible heating. For the proton imaging of the casting mold filling, the identity lens configuration was used. This permits an imag- ing FOV with nominal dimensions of 120 × 120 mm and approximately 100 micron spatial resolution. Imaging rates were 10 and 20 Hz for the castings, with and without the imposed magnetic field, respectively. The entire mold cav- ity was visible within the FOV during the filling, while the base of the funnel was just visible in the images, as shown in Figure 3. The transmission radiographs were post processed to maximize local contrast and to normalize the gray-scale to a consistent reference. To remove artifacts in the proton images resulting from local varia tions in bright- ness due to deviations in beam intensity, the images were Figure 4. Normalized fraction filled measured from the proton divided by representative images of the bea m. images of casting mold fillings of a tin-27 at.% bismuth alloy with and without a 0.12 T static magnetic field. The proton imaging suggests that introduction of a static magnetic field across the casting mold cavity subtly, but Impact on National Missions significantly, modified the behavior of the liquid metal Convection of an alloy melt during casting mold filling plays alloy during filling. These images enabled the first mea- a critical role in the solute segregation that develops at 357

Page 358

the macro-scale, whereas localized convection cells that and bubble transport inside a slab continuous-casting develop during solidification drastically influence micro- mold. 2014. Metallurgical and Materials Transactions scale solute segregation and morphological and micro- B. 45: 675. structure evolution. Application of a static magnetic field 7. Easter, S., V. Bojarevics, and K. Pericleous. Numerical during solidification provides an opportunity to control modelling of liquid droplet dynamics in micrograv- alloy melt fluid flow characteristics and the interface shape ity. 2011. Journal of Physics: Conference Series. 327: and dynamic equilibrium at that interface. Materials for 012027. the future, including the creation of materials with appli- cation-tailored properties to achieve controlled functional- ity by directed synthesis and processing, is one of three science pillars at Los Alamos National Laboratory. Control- ling materials synthesis and processing and the design of materials functionality through the exploitation of defects and interfaces, understanding the dynamic evolution of materials, defect structures, and interfaces, and the use of novel characterization techniques, especially at extremes, are priorities listed in Los Alamos National Laboratory’s Di- rected Research and Development Defects and Interfaces in Materials (DIM) Exploratory Research Category. These priorities are also reflected in the laboratory’s Matter-Ra- diation Interactions in Extremes (MaRIE) Signature Facil- ity vision, and are deeply rooted in nationally recognized needs identified by the U.S. DOE Basic Energy Sciences, The National Academies, and the White House Office of Science and Technology Policy’s Materials Genome Initia- tive (MGI). References

1. Imhoff, S. D., T. J. Ott, T. J. Tucker, M. R. Katz, and J. C. Cooley. Primary phase alignment in the Mg-Sb system with a 35-T DC magnetic field. 2013. Emerging Materi- als Research. 2 (EMR2): 5.
2. Porter, M. M., J. Mckittrick, and M. A. Meyers. Biomi- metic materials by freeze casting. 2013. JOM. 65 (6):
4. Li, X., Y. Fautrelle, and Z. Ren. Influence of an axial high magnetic field on the liquid – solid transformation in Al – Cu hypoeutectic alloys and on the microstructure of the solid. 2007. Acta Materialia. 55: 1377.
5. Li, X., Y. Fautrelle, A. Gagnoud, D. Du, J. Wang, Z. Ren, H. Nguyen-Thi, and N. Mangelinck-Noel. Effect of a weak transverse magnetic field on solidification structure during directional solidification. 2013. Acta Materialia. 64: 367.
6. Singh, R., B. G. Thomas, and S. P. Vanka. Effects of a magnetic field on turbulent flow in the mold region of a steel caster. 2013. Metallurgical and Materials Trans- actions B. 44: 1201.
7. Liu, Z., B. Li, and M. Jiang. Transient asymmetric flow 358

Page 359

Materials for the Future Early Career Research and Development Final Report Effects of Joining Processes on Bimetal Interface Content and Radiation Damage Resistance John S. Carpenter 20130764ECR Abstract tion of experimental and modeling studies, were found This document is a final report on the LDRD-ECR award- to emanate primarily from the high density of interfaces ed in January 2013 entitled ‘Effects of Joining Processes and the specific atomic-level interfacial structures in the on Bimetal Interface Content and Radiation Damage Re- composites [5,8-11]. Various applications within the sistance’. The proposed work stemmed from the desire defense, transportation, and energy arenas were envi- to provide an industrially practical joining method for sioned for material exhibiting this collection of desir- bimetallic nanolayered composite materials that exhibit able characteristics [4,12]. However, the physical vapor highly desirable traits such as high strength, radiation deposition (PVD) process used to manufacture these damage resistance, and thermal stability. The key to nanoscale bimetal composites was too slow for scale maintaining these properties is control of the joining up due to its dependence on building up the structure process such that interface density (or layer thickness atom-by-atom. Within the past five years, work at LANL for a multilayered composite) and the atomic structure through an LDRD-DR grant has provided an industrially of the interfaces is maintained. The model system of scalable methodology for manufacturing bulk sheets of Cu-Nb was selected and the solid state joining process nanostructured bimetal composites through a top-down of friction stir welding (FSW) was chosen as the focus of methodology [13,14]. These bulk sheets were manufac- the proposed work. Successful friction stir welds were tured using accumulative roll bonding (ARB) and they ex- accomplished for the first time in a nanolayered bimetal- hibited microstructures, properties, and performance in lic composite material over the course of this study. In extreme conditions similar to the PVD material [15-18]. addition, through careful design of the FSW tooling and An example microstructure for the ARB material can be process parameter combinations, near full-strength butt seen in Figure 1. In order to realize the potential for this welds were achieved. These same high-strength butt class of materials in engineering applications, industrially welds also exhibited large scale stability of both interfa- feasible joining techniques are required. The planned cial density and interfacial structure. Although further objective for this work was to find a suitable industrially refinement is needed in FSW process parameters and feasible joining technique that maintained the nano- tool design, this study shows that the FSW methodology structure (both density and structure of interfaces). can be applied to join nanostructured metal composites A prior LDRD-DR at LANL focused on the fabrication of successfully. With industrially scalable processing tech- bulk bimetal layered composites using the ARB tech- niques available for the bulk nanostructured metals, an nique. Success was achieved for the Cu-Nb system and industrially practical joining technique was the remain- characterized material for this study was provided via ing obstacle to utilizing these synthetic materials in a this route. An initial attempt at joining was made utiliz- multitude of applications including those in the fields of ing electron beam welding but, as expected, the melting defense, transportation, and energy. of the material lead to a loss of both the density and structure of the interfaces. Through this seed study, it Background and Research Objectives was concluded that any joining methodology that relied During the late 90’s LANL scientists showed that as the on fully melting the parent material for joining purposes length scale of bimetal composites began to decrease, would destroy the desirable properties that make the novel increases in strength [1-3], radiation damage Cu-Nb material unique. The focus of this research then resistance [4,5] and thermal stability [6,7] were noted became solid-state joining methods, or methods that that far exceeded what was measured in bulk metals. did not require full melting of the material. The most The enhancements in properties, through a combina- 359

Page 360

promising of the solid-state joining methods was friction ing temperatures coupled with the high strength of this stir welding. material made for a difficult challenge without even taking into account the desire to maintain the material’s initial Friction stir welding (FSW) is a solid-state welding process nanostructure. Despite these challenges, an aggressive set that makes use of friction and deformational heating to of research objectives was outlined for this study: plastically deform and coalesce materials to create a joined structure [19,20]. This technique was initially developed Year One: Perform initial studies of FSW on ARB nanoscale for aluminum because the microstructural instabilities en- material to give fully dense, high quality welds. Deliver- countered during fusion welding required additional post- able: Publish results on FSW and fusion weld parameters weld heat treatments [19-21]. Figure 2 shows a schematic and the interfacial effects on weld quality, hardness, and of the FSW process in the inset [22]. In FSW, a rotating post-processing interfaces in a peer-reviewed journal. tool is inserted through the material thickness and travels Year Two: Continue to refine FSW joining technique param- down the joint line between two separate pieces of mate- eters to tailor strength and interfacial structure. Integrate rial [20,23]. A predetermined load maintains the shoulder radiation damage resistance studies on all samples, if contact as the rotating tool travels down the joint line and warranted. Deliverables: (A) Provide a butt-welded plate physically mixes the materials. with minimal reductions in strength and radiation damage resistance. (B) A set of optimized parameters for joining ARB Cu-Nb plates that maintains the desirable proper- ties. (C) Publish results on effects of weld parameters on microstructure, individual interface content and radiation damage resistance in peer-reviewed journal. Success would have a three-fold impact: • LANL would maintain its leadership status in cutting- edge research in the field of bimetallic nanolayered Figure 1. (a) Micrograph of ARB Cu-Nb with defined lamellar composite materials. structure at h = 10 [15]. (b) High resolution TEM micrograph of an interface at h = 18 nm. Interfacial structure consists of joined • The number and type of materials to which the FSW \{112\} planes with epitaxial orientation relationship [15]. method could be applied would be expanded. • A pathway for joining industrial-sized quantities of bimetal nanolayered composites would be realized al- lowing for a transition from a model material to a real engineering material. Scientific Approach and Accomplishments In this section, the scientific approach and accomplish- ments will be conveyed in the context of the two year set of research objectives seen in the previous section. In year 1, bead-on-plate FSWs were performed on na- noscale Cu-Nb layered composites using a tungsten alloy for tooling. Several sets of processing parameters were utilized in order to obtain a fully consolidated (i.e. defect- Figure 2. Diagram describing general FSW process with advanc- ing (AS) and retreating (RS) sides labeled [22]. free) weld zone. The tool geometry was simplistic, con- sisting of an unadorned cylindrical pin and a flat shoulder FSW, to this point, has been largely utilized on materials region. The best result obtained using this initial pin can comprising only a single metal or phase since the process be seen in Figure 3. A large defect on the advancing side is brings the material close to melting temperature. Cu and seen indicating that the pin failed to draw the material in Nb have substantially different melting temperatures that during the stir processing. In addition, it can be seen that add a complicating factor to any attempts at FSWing. FSW the microstructure within the weld nugget is nanocrystal- is also commonly applied to soft metals with low melt line and equiaxedl in grain shape with grain diameters of temperatures to minimize tool wear. The competing melt- ~7 nm. The measured strength, utilizing nanoindenta- 360

Page 361

tion methods, exhibited a near 250% increase in the weld Efforts were, therefore, made in the beginning of the year nugget when compared with the parent material. The to manufacture thicker Cu-Nb sheets and the outcome increase in strength observed in this region when com- was an increase from a 500 um overall thickness to a 3 mm pared with the parent material is a product of having 3D overall thickness. With this increase in sheet size, further nanograins as opposed to only a single nanoscale dimen- design space was opened up in FSW tooling design. A sion within the composite sheet structure. Although the single scroll, or channel, was carved into the shoulder in a strength was improved, the fine (~ 7 nm) grain diameters helical pattern in order to promote better welding. In ad- made it impossible to investigate interfacial structures. dition, flats were placed on the sides of the pin to create a No characterization method currently available would be more triangular shape which has been shown to promote able to provide orientation information from these small enhanced welding. Technical obstacles that needed to be volumes. These results have been communicated via a overcome during this part of the work included the small peer-reviewed journal publication [24]. scale machining of the tungsten alloy used in the tooling as well as modifying the ARB process to accommodate a thicker final sheet. Butt welding, or the joining of two separate plates, was performed while investigating the proper process param- eters to achieve a fully consolidated weld. Success in this endeavor was achieved through careful manipulation of process parameters in conjunction with the new tool de- sign. Two fully consolidated welds were achieved for two sets of process parameters with the material either under- going one pass or two passes to achieve effective joining. Images of the single pass weld can be seen in Figure 4. Through optical and electron microscopy it was noted that in the single pass weld, a layered microstructure was still visible throughout the weld nugget. In the two-pass weld, on the other hand, a nanocrystalline and equiaxed micro- Figure 3. Hardness versus interfacial spacing plot showing the structure was observed. large increase in hardness between the weld nugget and the parent material. Blue diamonds represent data points for refined ARB material without FSW. The inset electron microscopy images show the best case bead-on-plate weld and the equiaxed nano- crystalline structure within the weld nugget [24]. Previous work using high pressure torsion (HPT) [25] sug- gested that an ideal strain was present during rotational straining of the ARB Cu-Nb material. Similar to FSW, HPT involves a shoulder in contact with a material with the tooling having both a load and a rotational speed. Unlike FSW, HPT does not involve a translational movement of the shoulder contact nor include a pin. The HPT work showed that at high strains a nanocrystalline, equiaxed micro- Figure 4. Micrographs showing the fully consolidated one structure of Cu and Nb grains was created while at more pass butt weld (bottom) where red dashed line indicates where moderate strains, the lamellar structure was maintained. the material was joined. No defects on either the advancing or This indicated a need to reduce the level of strain dur- retreating side were noted. Detailed micrographs show that a ing FSW via process parameters such as rotational speed, layered structure is maintained during joining. downward load on the tool, and translational speed. In addition, the large defect seen in Figure 3 revealed a need Tensile testing was conducted on the one and two pass for a pin geometry redesign to pull more material in during welds, with the one-pass weld exhibiting a higher strength the stirring process. than the two-pass weld. In addition, the one-pass weld compared favorably with tensile test results for the parent At the beginning of year 2, initial focus was placed on material with no weld as seen in Figure 5. Failure of both redesigning the pin geometry. Pin size and geometry is the one and two-pass welds was a normal chisel-edge type restricted by the overall thickness of the Cu-Nb sheets. 361

Page 362

fracture surface (not shown in this report) that is indicative Impact on National Missions of a ductile failure. This is an important result because it The technical results outlined above have lead to new indicates that if this weld had been put into service, notice- collaborations with outside academic institutions, new able deformation would have occurred prior to failure. capabilities within the laboratory, and the introduction of new students and postdocs to the laboratory. The reaction of the Cu-Nb nanolamellar composite to FSW caught the attention of wear researchers from Clemson University. A collaboration with Professor Molly Kennedy and her research team has begun by studying the wear properties of the parent material with the eventual goal of exploring the wear properties of the equiaxed nanocrystalline mate- rial as described in Cobb et al. [24]. A second academic collaboration has also begun with the University of New Hampshire and is centered on understanding the mecha- nisms that lead to the plastic deformation seen during FSW in these composites. Initial characterization indicates that the lamellar composite geometry leads to distinct differ- Figure 5. Plot summarizing mechanical testing data for the ences in plastic flow. Professor Marko Knezevic is excited one-pass and two-pass welds. The welds are compared with the to apply his plasticity model and incorporate a second mechanical behavior of the parent material [26]. The one-pass phase in order to understand the physics of flow in the weld shows that it reaches more than 90% of the strength of confined layer geometry presented by the nanolamellar the parent material with limited decrease in ductility. The inset composites. shows the experimental setup which included dog-boned shaped tensile specimens. This work has enabled new capabilities within the lab. Through the fabrication of tooling for the FSWing, machin- Utilizing electron backscatter diffraction (EBSD) methods, ists here at the lab have become familiar with the response the atomic structure of the interface was inferred as well of W-La2O3 to milling and machining. This ultra-hard met- as the density of interfaces. Strain levels within a weld al can be used to manufacture FSW tools in the future and nugget or stir zone vary. The maximum value occurs near for other applications that require strength coupled with the leading edge of the shoulder’s interaction with the high temperature stability. Another capability developed material during tool rotation, also known as the advancing at the laboratory through this work has been the introduc- side. The minimum strain level is found on the opposite tion of a nanoindentation technique that can extract elastic side of the weld nugget from the advancing side. In Figure moduli for indents of limited (<50 nm) depth. Finally, this 4, the advancing or high strain side is on the left of the work has lead to the new capability of performing FSW on image while the retreating or low strain side of the weld bimetallic nanolamellar composite materials. nugget is seen on the right. The EBSD results (not shown in this report) indicated that the interface density and Shraddha Vacchani, a former graduate student from atomic structure of the interface were both maintained in Georgia Tech, assisted on this project during the summer the single pass weld on the retreating side. On the advanc- of 2013 by using the FSWed material in order to develop ing side, the layered composite microstructure was more the aforementioned nanoindentation methodology. She tortured making an assessment of the interfaces more has since started a post-doctoral researcher position here difficult. This result suggests that with further refine- at the laboratory due, in part, to the excellent experience ment of process parameters and a more complicated tool she had as a summer student. Josef Cobb of Mississippi design, butt welds could be made that provide little to no State University (MSU) was also involved in this project and change in the microstructural features responsible for the performed work in this area during academic years at MSU enhanced mechanical properties. and during summers at LANL. This research will form the core of his PhD thesis and the PI will serve on the disserta- Radiation damage tolerance testing is currently underway tion committee. in order to see if the resistance to radiation damage is maintained following the FSW process. References This work has directly led to three invited conference talks, 1. Misra, A., M. Verdier, Y. C. Lu, H. Kung, T. E. Mitchell, one invited seminar, one publication in print, and two pub- M. Nastasi, and J. D. Embury. Structure and mechanical lications in preparation. properties of Cu-X (X = Nb,Cr,Ni) nanolayered compos- 362

Page 363

ites. 1998. Scripta Materialia. 39: 555. 13. Carpenter, J. S., S. C. Vogel, J. E. LeDonne, D. L. Ham- mon, I. J. Beyerlein, and N. A. Mara. Bulk texture 2. Clemens, B. M., H. Kung, and S. A. Barnett. Structure evolution of Cu-Nb nanolamellar composites during and strength of multilayers. 1999. MRS Bulletin. 24: 20. accumulative roll bonding. 2012. Acta Materialia. 60 (4): 1576. 3. Mara, N. A., D. Bhattacharyya, P. Dickerson, R. G. Hoa- gland, and A. Misra. Deformability of ultrahigh strength 14. Carpenter, J. S., R. J. McCabe, S. J. Zheng, T. A. Wynn, 5 nm Cu/Nb nanolayered composites. 2009. Applied N. A. Mara, and I. J. Beyerlein. Processing Parameter Physics Letters. 92: 231901. Influence on Texture and Microstructural Evolution in Cu-Nb Multilayer Composites Fabricated via Accumula- 4. Zhang, X., N. Li, O. Anderoglu, H. Wang, J. G. Swadener, tive Roll Bonding. 2014. Metallurgical and Materials T. Hochbauer, A. Misra, and R. G. Hoagland. Nanostruc- Transactions A. 45A (4): 2192. tured Cu/Nb multilayers subjected to helium ion-irradi- ation. 2007. Nuclear Instruments & Methods In Physics 15. Han, W. Z., J. S. Carpenter, J. Wang, I. J. Beyerlein, and Research Section B-Beam Interactions With Materials N. A. Mara. Atomic-level study of twin nucleation from And Atoms. 261: 1129. face-centered-cubic/body-centered-cubic interfaces in nanolamellar composites. 2012. Applied Physics Let- 5. Misra, A., M. J. Demkowicz, X. Zhang, and R. G. Hoa- ters. 100 (1): 011911. gland. The radiation damage tolerance of ultra-high strength nanolayered composites. 2007. JOM. 59: 62. 16. Carpenter, J. S., S. J. Zheng, R. F. Zhang, S. C. Vogel, I. J. Beyerlein, and N. A. Mara. Thermal stability of Cu-Nb 6. Misra, A., R. G. Hoagland, and H. Kung. Thermal stabil- nanolamellar composites fabricated via accumulative ity of self-supported nanolayered Cu/Nb films. 2004. roll bonding. 2013. Philosophical Magazine. 93 (7): Philosophical Magazine. 84: 1021. 718. 7. Misra, A., and R. G. Hoagland. Effects of elevated 17. Han, W. Z., M. J. Demkowicz, N. A. Mara, E. G. Fu, S. temperature annealing on the structure and hardness Sinha, A. D. Rollett, Y. Q. Wang, J. S. Carpenter, I. J. of copper/niobium nanolayered films. 20. Journal of Beyerlein, and A. Misra. Design of Radiation Tolerant Materials Research. 2005: 2046. Materials Via Interface Engineering. 2013. Advanced Materials. 25 (48): 6975. 8. Demkowicz, M. J., Y. Q. Wang, R. G. Hoagland, and O. Anderoglu. Mechanisms of He escape during implan- 18. Han, W. Z., E. K. Cerreta, N. A. Mara, I. J. Beyerlein, J. S. tation in CuNb multilayer composites. 2007. Nuclear Carpenter, S. J. Zheng, C. P. Trujillo, P. O. Dickerson, and Instruments & Methods In Physics Research Section A. Misra. Deformation and failure of shocked bulk Cu- B-Beam Interactions With Materials And Atoms. 261: Nb nanolaminates. 2014. Acta Materialia. 63: 150. 524. 19. Dawes, C. J., and W. M. Thomas. Friction stir process 9. Demkowicz, M. J., R. G. Hoagland, and J. P. Hirth. Inter- welds aluminum alloys. 1996. Welding Journal. 75 (3): face structure and radiation damage resistance in Cu- 41. Nb multilayer nanocomposites. 2008. Physical Review Letters. 100: 136102. 20. Rhodes, C. G., M. W. Maloney, W. H. Bingel, R. A. Spurl- ing, and C. C. Bampton. Effects of friction stir welding 10. Demkowicz, M. J., D. Bhattacharyya, I. Usov, Y. Q. on microstructure of 7075 aluminum. 1997. Scripta Wang, M. Nastasi, and A. Misra. The effect of excess Materialia. 36 (1): 69. atomic volume on He bubble formation at fcc-bcc in- terfaces. 2010. Applied Physics Letters. 97: 161903. 21. Johnsen, M. R.. Friction stir welding shows great promise for joining of difficult-to-weld materials. 1997. 11. Mara, N. A., D. Bhattacharyya, J. P. Hirth, P. Dickerson, Welding Journal. 76 (6): 20. and A. Misra. Mechanism for shear banding in nano- layered composites. 2010. Applied Physics Letters. 97: 22. Galvao, I., C. Leitao, A. Loureiro, and D. Rodrigues. Fric- 021909. tion stir welding of very thin plates. 2012. In Soldagem & Inspecao. (Sao Paulo, Brazil, March 2012). , p. 2. Sao 12. Han, W. Z., A. Misra, N. A. Mara, T. C. Germann, J. K. Paulo: Soldagem & Inspecao. Baldwin, and T. Shimada. Role of interfaces in shock- induced plasticity in Cu/Nb nanolaminates . 2011. 23. Jata, K. V., and S. L. Semiatin. Continuous dynamic Philosophical Magazine. 91 (32): 4172. 363

Page 364

recrystallization during friction stir welding of high strength aluminum alloys. 2000. Scripta Materialia. 43 (8): 743. 24. Cobb, J., S. Vachhani, R. M. Dickerson, P. O. Dickerson, W. Z. Han, N. A. Mara, J. S. Carpenter, and J. A. Schnei- der. Layer stability and material properties of friction stir welded Cu-Nb nanolamellar composite plates. 2014. Materials Research Letters. 2: 227. 25. Ekiz, E. H., T. G. Lach, R. S. Averback, N. A. Mara, I. J. Beyerlein, M. Pouryazdan, H. Hahn, and P. Bellon. Microstructural evolution of nanolayered Cu-Nb com- posites subjected to high-pressure torsion. 2014. Acta Materialia. 72: 178. 26. Beyerlein, I. J., N. A. Mara, J. S. Carpenter, T. Nizolek, W. M. Mook, T. A. Wynn, R. J. McCabe, J. R. Mayeur, K. W. Kang, S. J. Zheng, J. Wang, and T. M. Pollock. Interface-driven microstructure development and ultra high strength of bulk nanostructured Cu-Nb multilay- ers fabricated by severe plastic deformation . 2013. Journal of Materials Research. 28 (13): 1799. Publications Cobb, J., S. Vachhani, R. Dickerson, P. Dickerson, W. Han, N. Mara, J. Carpenter, and J. Schneider. Layer stability and material properties of friction stir welded Cu-Nb nanol- amellar composite plates. 2014. Materials Research Letters. 2 (4): 227. 364

Page 365

Materials for the Future Early Career Research and Development Final Report Probing Interface Reactions of Calcite Nanocrystals at Elevated Temperatures and Pressures Rex P. Hjelm Jr 20130772ECR Abstract have hindered progress in experimentally probing the Calcite (the most stable form of calcium carbonate, atomic interface structures controlling precipitation and CaCO3) is an important mineral phase in both environ- dissolution of carbonates at conditions encountered in mental systems and energy applications. Calcium car- the Earth’s crust. bonate is considered the long-term sink for atmospheric This ECR project was designed to reduce the complexity carbon dioxide (CO2), and thus the process by which of calcium carbonate mineral growth behaviors (which CO2 is trapped and transformed to stable carbonate depend on substrate and solution chemistries, as well minerals is of great fundamental and practical interest. as macroscopic variables such as temperature, transport The main goal of this project was to isolate and char- and thermodynamic stability) into elementary processes. acterize the atomic interface structures that facilitate Specifically, we pursued advanced total scattering tech- precipitation and dissolution of carbonate species under niques based on neutron and X-ray total scattering mea- conditions found in underground reservoirs for CO2 se- surements to target the atomic structure at the interface questration. Our efforts produced three sets of results: between substrate minerals and water/CO2 fluids at (1) the design, building, and commissioning of a high elevated temperatures and pressures. The approach in- pressure and temperature sample environment com- volved the use of porous substrates with tunable surface patible with neutron local structure studies; (2) novel chemistry, a size dependent (and therefore surface area experimental observation of the atomic structure of dependent) set of carefully characterized nanocrystal- the first few layers of growth and dissolution in carbon- line carbonate samples, high pressure and temperature ate mineral growth in an aqueous environment; and (3) sample environment development for local structure methodology advancements for finite material structure measurements, and differential total scattering measure- refinement. ments with fluid-solid mixtures. A complementary effort targeting the integration of multi-length scale scattering Background and Research Objectives measurements was also supporting this work. There are many challenges to probing the relevant interface reactions experimentally: first, the solid/fluid Scientific Approach and Accomplishments interface represents a very small proportion of the total The original project scope had four tasks associated system, resulting in low signal strength for many experi- with it: (1) careful synthesis and characterization of size mental measurements. Second, standard diffraction (the dependent nanocrystalline materials; (2) development usual means of determining atomic structure) is intrinsi- of high pressure and temperature cells compatible with cally complex due to the amorphous nature of the fluid local structure measurements on the Neutron Powder phase and the non-periodic/highly disordered interface Diffractiometer (NPDF) instrument; (3) neutron total within the solid phase. Third, the sequestering of CO2 in scattering measurements of calcite-fluid solutions at el- underground reservoirs occurs under hot, pressurized evated pressure and temperature; and (4) a data simula- conditions. Sample environments used to achieve these tion and data modeling effort to interpret experimental conditions (e.g., pressure cell) further reduce signal from results. The progress in each of these areas is described the interface structure during experiments. As a further below. complication, CO2 trapping mechanisms are governed by emergent processes from pore (micro) down to molecu- Task 1: We have identified that very small and mono- lar (atomic) scales, and experimental approaches typical- disperse calcite nanocrystal samples are quite difficult ly focus on one length scale or another. All these factors 365

Page 366

to synthesize. Nonetheless, a large (100 gram) calcite tron total scattering measurements (which require a very nanoparticulate sample with an average particle diameter low and stable background scattering signal). The suite of of 40 nm has been prepared at the Lujan Center by a se- gas pressure cells includes both the vanadium and titani- nior advisor for this project, Dr. Luke Daemon. This sample um-zirconium materials, and two different cell thicknesses. has been characterized by laboratory X-ray diffraction, Finite element analysis was applied to the final designs to thermogravimetric analysis, transmission electron micros- validate safety factors for experimental measurements. copy (for morphology), and surface area measurements. The commissioning of the new NPDF fluid cell apparatus Plans for both mechanical milling and digestive ripening of for pressure and temperature ranges up to 100 MPa and this sample have been made in hopes to produce smaller 350 K was completed in the 2014 Lujan Center run cycle. and larger average particle size samples, respectively. Addi- High quality PDF data were collected in the apparatus for tionally, neutron and X-ray total scattering measurements calibration samples (Si powder) at room temperature, and and atomic pair distribution function (PDF) analyses of additionally for supercritical and subcritical CO2 fluids at calcium carbonate crystalline polymorphs, including calcite elevated temperatures and pressures. This result and high (micron-sized bulk grains and 40 nm particles), aragonite pressure/temperature fluid cell applications were adver- and monohydrocalcite have been completed during the tised at the 2014 ACS spring meeting (March 16-20) and 2014 run cycle at the Lujan Center, LANL and the Advanced 2014 American Crystallography Association Annual Meet- Photon Source (APS), ANL. Characterization of these car- ing (May 17-21) and American Conference on Neutron bonate polymorphs provides reference structural informa- Scattering (June 1-6). This work was published in the Re- tion and model-independent local structure comparisons view of Scientific Instruments, describing the new sample for the early stages of carbonate mineral growth. Figure environment capability at the Lujan Center. 1 shows the comparison of NPDF data measured for both Task 3: Investigation of the first series of nanoparticulate bulk and 40 nm calcite samples. calcite samples with the commissioned high temperature and high pressure cells on the NPDF instruments was originally scheduled in the 2014 Lujan Center run cycle for four days of neutron scattering beamtime. This part of the investigation was postponed to a later run cycle due to the difficulty in producing a size dependent set of calcite nanoparticles. As an alternative, the idea of follow- ing calcite and carbonate mineral growth within controlled pore glass (CPG) substrates was nucleated. Two days of synchrotron X-ray total scattering beamtime were awarded through the APS competitive proposal review process to investigate the interfacial reactions and growth of calcium carbonate minerals in CPG. The approach involved the use of well characterized CPG powders (with tunable pore wall chemistry) and a fixed chemistry of carbonate growth solution as a reasonable proxy for studying the carbon- ate mineral precipitation in reservoir rock for sequestra- tion. The results from this initial experiment prove the signal sensitivity of total scattering difference methods Figure 1. (a) Time-of-flight neutron powder diffraction pattern for monitoring CaCO3-H2O-CO2 fluid/solid interfaces and at a detector angle of 148° (bank 4) for both bulk and 40 nm growth processes. While the analysis is still preliminary calcite samples. Only one of four banks of neutron data is shown. (and on-going), the results provide a completely novel (b) Corresponding PDF G(r) obtained through Fourier transfor- experimental observation that the initial stages of interface mation of neutron total scattering data (after merging all four growth follow monohydrocalcite-like building blocks. De- banks). For 40 nm sample, note the presence of broad Bragg re- tailed modeling of the difference experiments will allow us flections in plot (a) and the finite size effects resulting in reduced correlation intensity in plot (b). to track the rate and size of particulate growth. Follow-on experiments are planned with different growth solutions Task 2: Our team, in collaboration with the Lujan Center’s and at different temperatures. sample environment, high pressure, and mechanical design teams, completed the engineering design of a series of Task 4: Neutron and x-ray pair distribution function data high pressure and temperature cells compatible with neu- modeling efforts have been incorporated in Tasks 1-3, 366

Page 367

Figure 2. Example of current approach. discussed above. Current approaches are focused on factor; dashed blue curve). The black curve in Figure 2e simulating nanocrystalline materials as well as porous sub- is what one would observe for the experimental PDF of strates from the atomic scale through the nanoscale. Two MCM-41 material (real experimental data is not shown examples are given in Figure 2. The first example (Figures here). PDF analysis based on diffraction only reflects the 2a-d) features a single donut-shaped nanoparticle. At low- amorphous nature of the silica framework and would not Q (Figure 2a), the scattering arises from the whole particle, show significant high-r oscillations beyond ~10 Å. Multiple allowing one to deduce information related to particle length scales are important for accessing features such shape. At high-Q (Figure 2b), the Bragg diffraction peaks as pore network connectivity and degree of pore filling. give information about the internal atomic structure of the While PDF measurements inform on length scales of large particle. The simulated PDF G(r) (Figure 2c; orange curve) Q-range of ~0.8-50 Å-1 (corresponding to ~1-100 Å length of such a configuration cannot be directly compared with scale), the low-Q diffractometer measurements can probe the experimental PDF data. This is due to the fact that the on the Q-range of ~0.03-0.5 Å-1 (corresponding to ~5-200 simulation is completed in vacuum while the experimental nm length scale). It will be important to experimentally PDF data contains the structural features of an “assembly” merge these length scales to approach applied scientific of many uncorrelated nanoparticles (all constituting an problems in CO2 sequestration. Our developed methods infinite but structure-less “background” of average non- for combining simulated information from multiple length zero density). To take the environment of a real system scales into a single atomic structure refinement have been into account during modeling, a correction function (Figure published in the Journal of Applied Crystallography. 2c; blue curve) must be applied to the model of a finite In summary, our efforts produced three sets of results: (1) nanoparticle to compare it to real data. With a correction, the design, building, and commissioning of a high pressure the computed model PDF oscillates around zero (Figure 2d; and temperature sample environment compatible with black curve) and can be directly compared with the experi- neutron local structure studies; (2) novel experimental mental PDF data. The correction function (Figures 2c and observation of the atomic structure of the first few layers d; blue curve) is actually the characteristic form factor of a of growth and dissolution in carbonate mineral growth in finite-shaped particle. In the second example, the uncor- an aqueous environment; and (3) methodology advance- rected model PDF of ordered mesoporous silica MCM-41 ments for finite material structure refinement. Two pub- (Figure 2e; orange curve) contains both the sharp peaks at lications have appeared and another is in progress. Three low-r (i.e., bonding pairs of the amorphous silica) and the publications resulted from this work and another manu- broad structure bump/oscillation at high-r. By analogy with script is in draft form. the first example, the broad structure feature at high-r in Figure 2e (solid blue curve) come from the nature of the Impact on National Missions channel packing (hexagonal close packing array; dashed gray curve) and the overall particle shape (spherical form Understanding the behavior of mineral interfaces in 367

Page 368

geothermal fluids is important for expanding many of the earth’s energy options, including geothermal production, hydrocarbon extraction, remediation of contaminants, and geologic carbon sequestration. Successful study of the interface reactions in the CaCO3-H2O/D2O-CO2 system directly supports the missions of reducing climate and energy impacts, and enabling environmental remediation and restoration. Sample environment capabilities and modeling methodol- ogy developed as part of this work will also provide new or improved research tools for probing complex materials, relevant to the MARIE and NNSA missions of LANL and the DOE Office of Science, as well as DOE Fossil Energy, and Nuclear Energy Programs. For instance, the temperature and pressure regimes and length scales of interest in this study are directly applicable to optimizing hydrocarbon extraction from oil and gas reservoirs and understanding corrosion and scaling in nuclear reactors. The developed sample environment remains at LANL’s Lu- jan Center, to be applied to NNSA and Geosciences related scientific studies on the NPDF beamline. A range of en- ergy applications are possible, such as catalysis, clathrate inhibition, carbon capture, ion exchange, and gas storage. The developments can also extend to combined methods. Isotherm sorption and neutron diffraction measurements, for instance, can be joined for elucidating the adsorption mechanism and the structure of molecules confined in po- rous solids. Another example is the derivation on equation- of-state for a system of interest from neutron diffraction data collected with a range of macroscopic thermodynamic parameters. The work on methods to combine data over multiple length scales will provide scientists with tools to under- stand the atomistic to mesoscale structure of materials. This work will support LANL’s effort in emergent phenom- ena and defects and interfaces to be responsive to Office Science initiatives in mesoscale science. This work will also benefit NNSA and potentially DOD funded projects in explosives and propellants. 368

Page 369

Materials for the Future Exploratory Research Continuing Project Spin-state Transitions as a Route to Multifunctionality Vivien Zapf 20140177ER Introduction Benefit to National Security Missions The ability to sense and manipulate magnetic spins with Our work primarily seeks to understand the chemistry tuning parameters such as electrical current, light, or and physics of metal-organic materials where magne- voltage, underlies a great number of technologies. Of tism such as spin-state transitions couples to the fer- the three, current and light are well-established or on roelectric properties. Secondary to that, the coupling to their way to technological use. However, coupling mag- light, pressure, and chemical absorbents can be studied. netic spins to voltage, e.g. the ferroelectric or dielectric The ultimate applications of these materials are new properties of a material, remains a fundamental scien- magnetic and electric sensors, tunable high-frequency tific challenge. Yet, coupling between magnetic spins devices, and computer memory have eventual applica- and voltage has significant potential to reduce energy tions to energy security and remote sensing. This work consumption and dissipation in the process of sensing closely relates to the DOE/SC. If we demonstrate this and manipulating spins, and the need for this is specified new approach to magneto-electric coupling, follow-on in a number of recent funding calls especially by military funding for applications of this work can be envisioned. agencies and metrology researchers seeking room-tem- Applied work on the very similar topic of multiferroics, perature low-powered magnetic sensors, in addition to which targets the same applications, is currently funded civilian applications in sensing, computation circuits and by DARPA, the ORNL, the Army, the Navy, and by the memory, liquid crystal displays, and electronics. We take electronics industry. a novel approach by investigating the coupling between spin-state transitions and ferroelectricity. Specifically, we Progress focus on metal-organic materials where these spin-state The research team is exploiting the flexible design and transitions are common under ambient conditions. unique magnetic characteristics of organic materials instead of the more traditionally-studied inorganic ma- When the spin-states in neighboring spins communicate terials to create and understand the coupling between via magnetic exchange or structural distortions, the re- magnetism and electric dipoles in insulating materi- sulting orderings and phase transitions can become com- als. The advantages of organic-based materials is their plex and fascinating, and their understanding requires a flexibility for design, the softness of the lattice that can careful collaboration between analytic, computational, easily deform to create electric dipoles, and the ability and experimental understanding. Spin-state transitions to encapsulate different types of polar molecules within in metal-organics have been found in the last decade pores of the structure. By ‘organic’ materials we refer to be sensitive to chemical absorption, light and small to several classes of materials. The first class consists pressures, and to be inherently coupled to the lattice of organic m, where the magnetism resides within a and in some cases drive structural phase transitions. We carbon-based molecule. The second consists of metal- explore using spin-state transitions to drive structural organic coordination compounds. Here the magnetism changes in materials that modify their ferroelectricity is created by a transition-metal ion like Fe or Co, that is and/or dielectric constant. We build on our recent and coordinated by organic ligands into a crystal structure. intriguing results in a metal-organic. The potential for These coordination compounds can form porous frame- multifunctional cross-coupling between the magnetic- works (“metal-organic frameworks”, or MOFs) where electric properties as well as the incident light, pressure, pores within the structure can host additional atoms or and chemical sensitivity opens the door to other potent molecules. Finally, molecular magnets are compounds multifunctional devices. where again a transition-metal ion is the source of the 369

Page 370

magnetism, but this magnetic ion resides in an organic long-range magnetic order, creating ferroelectricity. We molecular structure. These molecules pack together into synthesized the compound and performed magnetic and a crystalline structure. In all cases we are exploring single electric experiments to very high magnetic fields at LANL crystals of these materials. and at the DC facility of the National High Magnetic Field Facility in Florida. We are in the process of completing this Our search for new magnetoelectric coupling mechanisms work, extending the experiments to more voltages, mea- in organic-based materials is three-pronged, exploring suring the electron spin resonance. We have modeled this three different mechanisms for magnetoelectric coupling. compound and are refining the modeling parameters in a tight iterative loop with experiments. Building on this work In the first prong of this project, we search for trimer- we have identified further spin-state transition compounds ized magnetic structures in organic magnets. This is an that are promising for observing this phenomenology even example of theory driving experiment. Here we predicted at high temperatures up to room temperature. Synthesis of the existence of magnetoelectric coupling due to compet- these materials is in progress. ing interactions within trimer compounds and subsequent distortion of the electronic wave function or trimer struc- Future Work ture. We have identified this phenomenology at work in • Complete the phase diagram of TNNCH3CN and com- the organic magnet TNN#CH3CN and computed the phase pute the electric polarization in order to compare our diagram and electric polarization. Our theoretical results results with the future experiments by Yasu Takano agree well with the set of experiments by performed Yasu that are described in the progress report. Takano (these experiments are described in the previous • Propose a model and compute the corresponding progress report). We are currently writing a manuscript on phase diagram for the new trimer compound that was this topic. recently synthesized by the experimentalists in our The second prong of this project is to investigate metal- team. organic framework compounds where magnetism in the • Provide theoretical input for the new metal-organic porous framework can couple to ferroelectricity of a multiferroic that Nathan Smythe is producing in his molecule trapped within the pore. On this topic we have new lab. discovered two perovskite metal-organic framework com- • Continue exploration of new materials including the pounds with magnetoelectric coupling and investigated the La1-xSrxCoO3 family of spin-state transition materials, mechanisms. We actually find a magnetic order directly andmagnetic variants of the ferroelectric tris-sarcosine creating ferroelectricity at the magnetic transition in one calcium chloride of these materials, which is rare within metal-organics. • Flesh out our understanding of the most promising We use the uniquely sensitive measurement techniques at materials discovered in the first year; this includes the LANL in pulsed magnetic fields to measure the magneto- frustrated trimer materials where magnetoelectric electric coupling. Our results are in fact surprising because coupling will be explored. magnetoelectric coupling appears to be not the simplistic • Magnetoelectric coupling in tris[2-(pyrrol-2-ylmethyle- idea of a magnetic framework with electric dipoles encap- neamino)ethyl]amine-Manganese(II) will be mea- sulated, but rather a more complex scenario of ferroelec- sured, and we will grow family members with different tricity originating both within the framework and within magnetic ions and ligands. pores, influenced by the coupling between them. Two • Variants of triethylmethylammonium tetrabromofer- papers are in final draft form, and third paper is published. rate (III) with different ligand lengths will be explored to tune the magnetoelectric coupling above room Finally the third prong of this project is to investigate temperature. spin-state transitions as a route to multiferroic behavior. Spin-state transitions occur in certain transition-metal ions Conclusion with partially-filled d shells. When a local distortion of the Our deliverable will be an enhanced understanding of lattice changes the crystal electric field environment of the spin-state transitions and their coupling to the crystal- magnetic ion, the electrons in the partially-filled d-orbitals line lattice and it’s electrically polar properties, as well as can occupy a different pattern of orbitals that creates a dif- new metal-organic materials and materials approaches to ferent total magnetic moment for the ion. This is a mecha- magneto-electric coupling. The applications of magneto- nism for strong coupling between magnetism and crystal electric coupling are devices such as computer memory, structures, that we in turn exploit to create coupling be- filters, sensors, and antennas. Our approach achieves tween magnetism and electric dipoles.Here we discovered magneto-electric coupling with less power dissipation than the first example of a spin-state transition, rather than 370

Page 371

existing spintronics devices due to the insulating nature of our materials. Publications Gomez-Aguirre, L. C., B. Pato-Doldan, J. Mira, S. Castro- Garcia, M. A. Senaris-Rodriguez, M. Sanchez-Andujar, J. Singleton, and V. S. Zapf. Magnetic field-induced multiferroic behavior in [CH3NH3][Co(HCOO)3] Metal-Organic Framework. To appear in Journal of the American Chemical Society. Kamiya, , and C. D. Batista. Magnetic Vortex Crystals in Frustrated Mott Insulator. 2014. PHYSICAL REVIEW X. 4 (1). Koutroulakis, , Zhou, Kamiya, J. D. Thompson, H. D. Zhou, C. D. Batista, and S. E. Brown. Quantum phase diagram of the S=1/2 triangular-lattice antiferromagnet Ba3CoSb2O9. 2015. PHYSICAL REVIEW B. 91 (2). Lin, , Barros, Mun, Kim, Frontzek, Barilo, S. V. Shiryaev, V. S. Zapf, and C. D. Batista. Magnetic-field-induced phases in anisotropic triangular antiferromagnets: Application to CuCrO2. 2014. PHYSICAL REVIEW B. 89 (22). Ma, J., Y. Kamiya, T. Hong, H. B. Cao, G. Ehlers, W. Tian, C. D. Batista, Z. L. Dun, H. D. Zhou, and M. Matsuda. Static and Dynamical Properties of the Spin-1/2 Equilateral Triangular-Lattice Antiferromagnet Ba3CoSb2O9. Physical Review Letters. Mun, E. D., Chern, Pardo, Rivadulla, Sinclair, H. D. Zhou, V. S. Zapf, and C. D. Batista. Magnetic Field Induced Transition in Vanadium Spinels. 2014. PHYSICAL REVIEW LETTERS. 112 (1). Mun, E. D., J. Wilcox, J. L. Manson, B. Scott, P. Tobash, and V. S Zapf. The Origin and Coupling Mechanism of the Magnetoelectric Effect in TMCl2-4SC(NH2)2 (TM = Ni and Co). 2014. Advances in Condensed Matter Physics. 2014: 512621. Zapf, , Jaime, and C. D. Batista. Bose-Einstein condensation in quantum magnets. 2014. REVIEWS OF MODERN PHYSICS. 86 (2): 563. Zhu, S., Y. Q. Li, and C. D. Batista. Spin-orbit coupling and electronic charge effects in Mott insulators . 2014. Physical Review B. 90: 195107. 371

Page 372

Materials for the Future Exploratory Research Continuing Project Beyond the Chemical Reaction Zone: Detonation Product Gases in the Warm Dense Regime Dana M. Dattelbaum 20140261ER Introduction Progress The chemical reaction zone behind a detonation front Measured the Hugoniot of low-density ammonia based comprises a dense fluid mixture of components such as on gas gun experiments. Determined temperature of N2, CO2, CO, CH4, and NH3, at densities ranging from first- and second-shocked states based on pyrometry up 0.09 to 1.5 g/cm3. There is considerable overlap be- to 7000-8000 K. These were compared to values calcu- tween the constituents of detonation product mixtures lated in hydrodynamic simulation using a legacy SESAME and those of planetary ices and atmospheres. Condi- table and the PAGOSA software package. Calculated tions relevant to planetary physics are often those of temperatures were systematically low relative to mea- the warm dense matter (WDM) regime: too dense for surement. From the data we also determined that am- weakly-coupled plasma models but too hot for standard monia does not ionize on first shock. This work is being condensed matter techniques. While these states are reported at the Shock Compression Conference in June, hotter and denser than those of the chemical reaction and a journal manuscript is in preparation. zone, multiple shocks may readily carry the product Safety and procedures were completed to start firing mixture through the WDM and into the dense plasma methane at various initial densities this quarter. We region: standard SESAME tables extend to T10 keV and performed the first gas shock experiments on the LANL ρ/ρ010^6. Despite their vital importance in these con- 2-stage high performance gun. We also designed an texts, however, molecular dissociation and ionization at asymmetric cell to reach shock conditions three times elevated pressure and temperature are poorly character- greater in pressure for the same impact condition; this ized and constitute a significant source of potential error innovation is being published in HVIS Proceedings 2015. in our equations of state. To address these gaps, we will establish a synergistic experiment-theory-simulation We developed an interferometric fiber Bragg strain sen- effort focused on quantifying the P-V-T surfaces, and sor to provide early triggers for spectroscopic diagnos- associated dissociation and ionization processes for the tics. Again, there is a manuscript in preparation. principal detonation product species: N2, CO2, CH4 and NH3. The research team also completed theoretical investiga- tions of ammonia in a warm dense matter regime, using Benefit to National Security Missions newly developed simulation methodology. This enabled The successful execution of this experimental-theoretical us to more accurately assess the extent of molecular project provides critical and foundational data and decomposition. These results are being prepared for models for improved confidence in weapons simula- publication in the Journal of Chemical Physics. tions. This project will enable rapid development of the experimental and theoretical tools to demonstrate Future Work transformational improvements in the treatment of The objectives of the experimental effort are: 1) to quan- detonation product gases. The results of this project are titatively resolve the shocked states in N2, CO2, CH4 relevant to the National Boost initiative, Predictive Ca- and NH3 on the principal Hugoniot, and along a quasi- pability Framework, and objectives of some DOE/NNSA isentropic path to very high compression ratios and Campaigns, and Advanced Certification. In addition, temperatures, and 2) to determine at what conditions the fundamental data obtained is relevant to planetary dissociation and ionization (and possibly metallization) physics, and definitions of the conditions of the large gas occur, and attempt to measure spectroscopic evidence giant planets. 372

Page 373

of dissociation products. On- and off-Hugoniot states in projectiles. To appear in Shock Compression of molecular gases will be achieved via gas gun-driven plate Condensed Matter - 2015. (Tampa, FL, 14-19 June impact using a suite of high performance gas guns with 2015). impact velocities up to 8 km/s. The experimental program will consist of two directions. The first, and lower risk, are shock experiments on fluid (liquefied) species, with a focus on reaching higher compression-rate, off-Hugoniot states, and applying in situ spectroscopies for the first time. The second, parallel direction will be to apply our newly devel- oped experimental approach on elevated density gases. To determine the temperatures along the compression path, and to measure dissociation and ionization in the gases, we have a variety of time-resolved optical spectroscopic tools at our disposal. Ab initio molecular dynamics (AIMD) simulation will be performed at fixed (ρ,T) states matched to those of experi- ment. Average degrees of coordination (bonding) will be used to calibrate atomic fluid models that more effectively capture dissociation transitions. Ionization states, however, are very difficult to infer directly from AIMD, because the imposed periodic boundary conditions prevent electron density from actually leaving the simulation cell. Our initial approach to this problem will be to calculate electrical con- ductivity (ac) and interpret the degree of charge transport using the Drude model for a free electron gas. Simulations of the plate impact experiments will be performed using PAGOSA, a three-dimensional Eulerian hydrocode used to simulate compressive flow and high strain-rate deformation. Conclusion The expected result of this program is a significantly im- proved modeling capability for simple molecular gases that are relevant to detonation products and weapons simula- tions for the Laboratory. New EOS methodologies will pro- vide a physics-basis to bridge the gap between compressed fluid and Thomas-Fermi-Dirac regimes. Furthermore, we will lead the field by applying a comprehensive suite of ad- vanced spectroscopic diagnostics under shock conditions. Lastly, new models, based on experimental results, will be evaluated in the Eulerian code PAGOSA. Publications Gibson, L. L., B. D. Bartram, D. M. Dattelbaum, J. M. Lang, and J. S. Morris. A gas-loading system for LANL two- stage gas guns. To appear in Shock Compression of Condensed Mater - 2015. (Tampa, FL, 14-19 June 2015). Goodwin, P. M., B. R. Marshall, R. L. Gustavsen, J. M. Lang, A. H. Pacheco, E. N. Loomis, and D. M. Dattelbaum. Fiber-interferometric detection of gun-launched 373

Page 374

Materials for the Future Exploratory Research Continuing Project Topological Kondo Insulators Joe D. Thompson 20140271ER Introduction lating electrons for energy applications, such as thermo- Metals are materials in which there is a band of elec- electric cooling, as well as for discovering an exotic state tronic states that is only partially filled, allowing elec- in which elementary charge carriers act as their own trons to flow easily in response to a weak electric field. antiparticles, so-called Majorana fermions. This project In a pristine insulator, however, possible electronic states is primarily one of discovery science, focused on under- are fully occupied, and there is a large energy gap be- standing the fundamental consequences of theoretical tween filled and unfilled electronic states. As a result, no predictions that also have implications for new interpre- electrons are available to move in response to an electric tations of electronic behaviors in plutonium-based mate- field and, hence, the electrical resistivity is infinite. An rials. But, research will be pursued as well with potential entirely new state of electrons has been proposed that technological applications in mind, including the use of combines both electrically conducting and electrically those discoveries in the development of new spin-based insulating electrons in a single material. The concept logic devices that have broad relevance to DOE energy underlying this state of electrons is highly non-trivial and defense missions. and relies on the mathematical notion of topology that Progress ‘protects’ conducting electrons from acquiring the insu- lating behavior of the vast majority of other electrons. A metallic topological surface state is theoretically pre- As long as time-reversal symmetry is preserved (that is, dicted to be robust to perturbations that do not break an electron with spin ‘up’ moving with momentum +k time-reversal symmetry. To test the robustness of the is identical to an electron with spin ‘down’ moving with surface state of the theoretically proposed topological momentum –k), and there is strong coupling between Kondo insulator SmB6, we have continued experiments the electron spin and the electron angular momentum, in which we measure the response of SmB6 to surface this very unusual state of electrons should appear. The damage created by systematically irradiating the surface goal of this project is to establish experimentally the with heavy ions (krypton and argon) to controlled depths validity of this idea of topologically protected conduct- ranging from 10Å to over 400Å. In work published in ing electrons in materials that are electrically insulating Physical Review B we showed that, upon significant because of strong electron-electron correlations. Just as damage of the surface by these non-magnetic ions, the the theory is non-trivial, so is its experimental validation surface state is not destroyed. Instead, the surface state because of the inevitable complication of material’s im- reconstructs beneath the heavily damaged region that perfections which produce responses that mimic those becomes a poor conductor. This is consistent with expec- of topologically protected conducting electrons. tations for a topological surface state and an important confirmation of the theoretical prediction that SmB6 is Benefit to National Security Missions a topological Kondo insulator. In further work we have The theoretical proposal that certain materials should shown that this reconstruction only occurs once the have electrically insulating interiors but electrically con- concentration of defects in the damaged layer is suf- ducting surfaces has profound implications for our basic ficiently high, on the order of 1 displacement per atom. understanding of how electrons can organize them- At small levels of damage, the electrical conductivity of selves as a consequence of rules of quantum mechanics. the surface state appears to be moderately reduced, Experimentally validating this theoretical proposal in possibly indicating the presence of a ‘Kondo hole’ (the materials where strong electron-electron interactions absence of a Sm atom from its expected position in also are present would open opportunities for manipu- the lattice) that scatters the metallic surface electrons 374

Page 375

and reduces the surface conductivity. A further theoreti- in the bulk is systematically suppressed by applied pres- cal prediction is that the topologically protected metallic sure. With this range of techniques, we expect to prove or surface should be sensitive to the introduction of magnetic disprove several theoretical predictions. If theory is proven impurities that break time-reversal symmetry. We have to be correct, this work will lay the foundation for an en- tested this prediction by irradiating the surface of SmB6 tirely new field of research on strongly correlated electron with magnetic iron-ions. Surprisingly, we found that the materials and the substantial technological consequences surface state is not destroyed by this damage but recon- of topologically protected quantum states of matter. structs below the damaged surface, as with non-magnetic irradiation. This observation is not understood fully but Conclusion may be a consequence of the orientation of the magnetic This project will establish a new field of research on elec- moments of implanted iron ions. Nevertheless, this result tronic states that emerge in materials as a consequence has important implications for both the study of SmB6 with of strong electronic correlations and that hold promise for magnetic field and also for the application of topological enabling new technologies ranging from thermoelectric surface states, for example in magnetic heterostructures cooling and spintronic devices to quantum computing. form with SmB6. Publications The design of an experimental setup to measure the ther- Wakeham, N., Y. Wang, Z. Fisk, F. Ronning, and J. D. mal conductivity of single crystals of SmB6 under applied Thompson. Surface state reconstruction in ion- pressure has made good progress. Initial tests of several damaged SmB6. 2015. Physical Review B. 91: 0851007. experimental setups have led to the selection of a dynamic method, the so-called ‘3-Omega’ technique. In this method the thermal conductivity of a sample is measured by measuring the 3rd harmonic of the voltage that is induced across a crystal in the presence of a small oscillating ther- mal gradient generated by an oscillating current through a metallic strip on the crystal’s surface. Isolation of the small 3rd harmonic signal from the background is a challenge using conventional electronics but has been achieved through a special experimental arrangement. The applica- tion of a narrow metallic strip to the sample is also chal- lenging, but good progress has been made in that regard as well. The next step will be to show we get accurate and reproducible data for samples in a high pressure environ- ment. Future Work To achieve our goal of experimentally validating the theoretical prediction of topologically protected electronic states in strongly correlated electron materials, we will use heavy-ion irradiation to damage the surface systematically to controlled depths ranging from approximately ten to several hundred nanometers depth and measure and mod- el the effect of this damage on the temperature-depen- dent electrical resistivity of proposed topological Kondo insulators. If this damage does not break time-reversal symmetry, theoretically the metallic surface should survive damage, but if this symmetry is broken by irradiation dam- age, it is possible that the metallic surface state would be destroyed. This possibility has not been settled by theory but will be by proposed experiments. In parallel, we will pursue the use of pressure-dependent measurements of the thermal and electrical resistivities of SmB6 to probe the evolution of the metallic surface as the insulating gap 375

Page 376

Materials for the Future Exploratory Research Continuing Project Semiclassical Modeling of Non-adiabatic Processes in Molecular Materials Dima V. Mozyrsky 20140293ER Introduction marily addresses Energy and Earth Systems LANL Grand Present day computational quantum chemistry and ma- challenge by providing computational means for mo- terial science attempt to evaluate not only equilibrium lecular materials suitable for clean energy (solar energy properties of molecular materials, but also their dynam- capture and energy storage). Secondly, it strongly relates ics under different physical conditions. For example, to Materials: Discovery Science to Strategic Applications by absorbing a quantum of laser radiation a molecule challenge by discovering emergent phenomena in com- goes to an excited electronic state, which may lead to plex systems. In particular we expect that the project various possibilities in its subsequent evolution: It may will attract the attention of the world-wide community dissociate into to new molecules or atoms or, emitting of spectroscopists and computational material scientists a phonon, it may go to a long-living excited vibronic and will establish new collaborations with LANL and at- state without splitting into new molecules. Understand- tract new CINT users. ing and control over these processes (photoinduced Progress pathways) lies at the heart of all our efforts to design functional photoactive materials for many technological The research team tested the relative error on the Semi- applications. Such molecular dynamics is inevitably ac- Classical Monte-Carlo algorithm (SCMC) and found that companied by the transitions between neighboring elec- it scales as a square root of the number of trajectories. tronic states in a molecule that are not easy to describe We have also shown that the error is approximately within the existing approximation schemes. The majority independent on the number of nuclear degrees of free- of the existing schemes are either too computation- dom. The details can be found in Ref. (1). ally expensive or based on ad-hoc approximations that In order to apply the semiclassical Monte-Carlo ap- lead to inconsistencies between quantum and classical proach to the realistic models we needed to significantly mechanics. Therefore there is a clear need for a novel improve the convergence of the algorithm, i.e., by factor computational paradigm that, on one hand, would be ~ 50. This is because the electronic structure calculations sufficiently precise and based on well controlled physical for realistic molecules are typically very computationally approximations rather than ad hoc assumptions, and, expensive and constitute a major bottleneck for the use on the other hand, would be computationally feasible to of the semiclassical methods since these methods typi- allow for efficient numerical simulations of molecular dy- cally require a large number of trajectories. For example, namics in realistic multi-atomic molecules. In this project our initial approach (SCMC method) had required several we aim to accomplish this goal. thousands of trajectories to evaluate the transition prob- abilities for a system at a single level crossing, while the Benefit to National Security Missions contemporary computational resources at LANL would First and foremost our project upon completion will not allow for “such luxury”: Typical non-adiabatic surface provide novel computational capabilities critical for hopping simulations of realistic molecular systems would understanding light-induced dynamics in many techno- involve several hundred trajectories, at most a thou- logically relevant nanostructures. In fact, for the first sand of trajectories. The problem of convergence was time, experimental ultrafast spectroscopy will have its resolved by developing two new methods, the accelerat- theoretical counterpart able to treat the material on ed-SCMC and the accelerated-SCMC with re-Gaussianiza- the same footing. Consequently, we envision extremely tion, which reduce the cost of the SCMC algorithm up to broad applications of developed tools, relevant to the two orders of magnitude for certain systems. In the first current and future LANL/DOE missions. Our project pri- 376

Page 377

approach the cost efficiency is achieved by avoiding the deviation from the work plan stated in the original pro- repetitive trajectories, which increases the efficiency by posal. However, it may provide a more efficient solu- several times. Furthermore, the accelerated-SCMC method tion to the problem of non-adiabatic dynamics (which is “hybrid”, - it only uses Monte-Carlo method for evaluat- is the main goal of this project). ing the integrals only higher order terms in the expansion, while the first order integral is evaluated by finite differ- ences. This leads to a significant decrease in the number of Conclusion trajectories because Monte-Carlo method is efficient only The major technical goal of this project is to provide a for the integration of higher-dimensional functions. The computational algorithm to evaluate molecular dynam- re-Gaussianization procedure (i.e., the second method) ics in situations involving transitions between different turned out to be a very efficient approach in problems electronic states in a molecule. The algorithm will allow for involving several level crossings. It allows us to reconstruct a reliable and efficient evaluation of reaction pathways and the wavefunction after the system passes through a level product ratios in a variety of problems of photoinduced crossing, which leads to significant computational savings molecular dynamics and will complement, and, in some for two reasons: (1) It reduces the number of interfering cases, replace the existing approaches. We also intend to wavepackets at the next crossing, which leads to a very apply our method to study non-adiabatic dynamics in sev- strong reduction in the number of needed trajectories; eral molecular systems involving interaction of molecules (2) It removes the necessity to run multiple trajecto- with light relevant for photovoltaic applications. ries between level crossings, - instead one runs a single, re-Gaussianized wavepacket between the non-adiabatic Publications regions, which also significantly reduces the number of White, A. J., V. N. Gorshkov, S. Tretiak, and D. Mozyrsky. needed trajectories to maintain a desired accuracy. In most Non-adiabatic molecular dynamics by accelerated cases the new procedures are nearly as efficient as the semiclassical Monte Carlo. 2015. The Journal of commonly used surface hopping schemes, with little to no Chemical Physics . 143: 014115. loss of accuracy. This implies that these modified SCMC White, A., V. N. Gorshkov, R. Wang, S. Tretiak, and D. algorithms will be practical numerical solutions for simulat- Mozyrsky . Semiclassical Monte-Carlo: A First Principles ing non-adiabatic dynamics in realistic molecular systems. Approach to Non-adiabatic Molecular Dynamics . 2014. The details for the two approaches can be found in Ref. (3). The Journal of Chemical Physics . 141: 184101. We have studied analytically a model involving non- Zhou, N., L. Chen, Y. Zhao, D. Mozyrsky, V. Chernyak, adiabatic dynamics in a system with a few slow degrees and Y. Zhao. Ground state properties of sub-Ohmic of freedom. It was shown in such problems that one can spin-boson model with simultaneous diagonal and introduce an order parameter that properly describes off-diagonal coupling . 2014. Physical Review B. 90: the collective dynamics of slow degrees of freedom. Such 155135. study is relevant for better general understanding of non- adiabatic dynamics. The details can be found in Ref. (2). Future Work • Generalize the existing algorithm to multi-dimensional situations, i.e., for multi-atomic molecules. Apply the algorithm for several molecular systems of interest • Include description of noise and dissipation effects. In the majority of experimental situations the forces produced by molecule’s surrounding medium (typically a solution) are sufficiently strong and may significantly modify the system’s dynamics. Accounting for such forces is, therefore, imperative for the reliable numeri- cal simulation of realistic molecules in realistic physical conditions. • Test a deterministic approach developed in the course of the project. If successful, this approach may signifi- cantly improve the numerical cost of the non-adiabatic modeling algorithm. This direction represents a slight 377

Page 378

Materials for the Future Exploratory Research Continuing Project Making nano-Mg a reality Irene J. Beyerlein 20140348ER Introduction bio, and aerospace). Nano-Mg-X will no doubt exhibit With the need to reduce gas consumption, Mg alloys many other interesting properties apart from ultra high have risen to the top of the list as the next lightweight strength, such as corrosion resistance (aging), magne- structural materials to replace steel. However, even if tism, conductivity, thermal stability, biodegradation re- all the steel in your car (e.g., 30 mpg) were replaced sistance, and fracture toughness. Our project will funda- with today’s Mg, the weight reduction would boost fuel mentally transform Mg research and manufacturing and efficiency by only 54.6% (to 46 mpg), well below the make LANL the leader of this rapidly growing market. Corporate Average Fuel Economy standards for 2020 (61 This ER notably fulfills all three pillars of LANL’s Materials mpg). But if the strength of bulk Mg were to increase Strategy. 1) Defects and Interfaces: our innovation tunes ten-fold, the gas mileage would boost from 30 mpg to interfaces to suppress emission of twinning dislocations. 76 mpg, leading to a transformational impact on the 2) Emergent Phenomena: all matter is controlled by their automotive industry. For other structural metals (e.g., Al, atomic structure and nearly 30% of nano-Mg-X will be Cu, Ni, steel), up to ten-fold increases have been realized Mg-X interface with a differing atomic structure than by severe plastic deformation (SPD) techniques, which Mg or X. Hence new effects can be expected. 3) Extreme transform coarse-grained metals to nano-grained metals. Environments: we exploit one extreme (severe strains) Unfortunately, all attempts to make “nano- Mg” in bulk to produce nano-Mg-X, a material that will be tolerant in sizes have failed. Bulk nano-Mg cannot be made because extremes of high stress, high strains (formable), and we SPD causes deformation twinning in Mg. For this ER, we expect, high temperatures. propose a breakthrough in microstructural processing by Progress design, where the key is to create a composite of Mg and a smaller amount of second metal phase, metal X, such This past year, we made progress towards our goal of as Nb, Fe, or V, and tune the Mg-X interfaces to suppress developing a ground-breaking method for synthesizing a twinning, permitting Mg refinement to the nanoscale. nano layered composite metal that combines two phas- This innovative approach gives rise to new scientific is- es, one metal with a hexagonal close packed (hcp) crystal sues of Mg-X interface/defect processes. To tackle them, structure and another metal with a body centered cubic we pose two original hypotheses and use them to direct crystal (bcc) structure. Our main innovation behind this an integrated experimental and modeling s trategy. By plan is the recognition that deformation twinning is the the end of the ER, we will deliver nano-Mg-X in bulk chief obstacle that has prevented so many scientists sheet form, wherein the Mg crystals are <100 nm in size. before us from achieving this goal. Specifically last year, Being sought after for a decade, our bulk nano-Mg-X, as we achieved many milestones. First, we determined well as the science that enabled it and synthesis path- how to make nanolayered composite Zr/Nb, where Zr is ways that made it, will be in high demand by countless an hcp metal and Nb is a bcc metal. We determined that reputable research groups and industry (automotive, in order to refine from coarse-layered sizes to nano-lay- bio, and aerospace). ered sizes, it is necessary to eliminate shear banding, an instability that happens during the initial stages of SPD Benefit to National Security Missions processing. Once shear banding begins, the layers no Being sought after for a decade, our bulk nano-Mg-X, as longer co-deformed and layer refinement via mechanical well as the science that enabled it and synthesis path- processing stops. Second, we characterized the micro- ways that made it, will be in high demand by countless structure and we determined the interface crystallog- reputable research groups and industry (automotive, raphy. The texture of the Zr phase and Nb phase corre- 378

Page 379

sponded to the texture that would develop in each phase above three accomplishments in CPFE led to three papers, when rolled alone. However, when the layer thickness either being reviewed or accepted for publication. For the reduced below 100 nm, the texture strongly deviated from upcoming year, we are now ready to carry out fabrica- the conventional rolling texture. Instead it became sharp tion process simulations for Mg/Nb using our novel CPFE and correspondingly, a few predominant interfaces are sus- model. pected to emerge. From our analysis, the textures suggest that the two predominant interfaces are (a) \{110\}Nb||\{1 -1 Future Work 0 0\}Zr and \{111\}Nb||\{1 -2 1 10\}Zr and (b) \{110\}Nb||\{1 -1 0 The overarching goal of this project by the end of the three 0\}Zr and \{100\}Nb||\{1 -2 1 10\}Zr. Third, we then developed years is to develop a ground-breaking method for synthe- a multi-scale model to determine the microstructural evo- sizing nano layered composite metal that combines two lution during the processing of Zr/Nb, namely its disloca- phases, one metal with a hexagonal close packed (hcp) tion defect structure, its bimetal interface structure, and crystal structure and another metal with a body centered texture. Through this work, we also determined that it did cubic crystal (bcc) structure. In this past year, we made a not twin, why it did not twin, and why the predominant novel nanocomposite material in bulk form using severe interfaces emerged. We also determine the microstruc- plastic deformation (SPD). In these composites, the hcp tural features that caused shear banding. The fact that material was Zr and the bcc material was Nb. We were annealing reduced the dislocation density postponed shear able to model these Zr/Nb nano layered composites using banding. Thus, the changes annealing caused in grain size crystal plasticity finite element techniques. In the next and texture were not the main reasons it postponed shear year, we plan to work our way towards making Mg/Nb banding. composites. We plan to first apply our SPD technique to single phase Mg. Two types of Mg materials will be used. Fourth, we made progress towards our ultimate goal of One will be pure Mg, which is known to twin at room making Mg/Nb composites. To this end, we made Mg-Nb temperature. The second material will be an Mg alloy, nano layered composites via physical vapor deposition. called Mg-4Li. In planning for these tests, we have devel- With this fabrication method, we were able to modify the oped constitutive crystal plasticity based models to study interface structures in a controlled manner and to deter- the effect of starting microstructure (grain size, texture) mine the deformation behavior via micro pillar testing. We on twinning in both Mg materials. We find that we can found that when the layers are thin, 5 nm, the Mg takes on postpone twinning in Mg at elevated temperatures and a body center cubic (bcc) crystal structure. However, when we also found that twinning is postponed in Mg due to the they are 50 nm, while still being nano in size scale, the Mg lower activation barrier for pyramidal <c+a> slip. We will keeps its hcp crystal structure. As a result, we found that at the same time carry out crystal plasticity finite element the two pillars had very different constitutive behaviors. (CPFE) simulations of Mg and Mg/Nb for understanding The 5 nm Mg/Nb composite achieved both outstanding microstructure evolution during bulk deformation process- ductility and strength. At the same time we also carried ing. We have already started developing the CPFE models out atomic scale simulations to determine which interface for Mg and for modeling twin formation in Mg. In the next structures are energetically favorable and which crystal year, we will complete these efforts and begin extending it structure is preferred as a function of layer thickness. The to consider Mg/Nb. This model will help guide the process- simulations predicted that due to the coherency strains ing steps for achieving target interface configurations and in the interface, the Mg transforms from hcp to bcc. The character. coherency strains for bcc Mg are less than that for hcp Mg. Conclusion Using this knowledge, we performed crystal plasticity finite By the end of the ER, we will deliver 1) nano-Mg-X in bulk element (CPFE) simulations to understand microstructure sheet form, 2) characterization of its nanostructure and evolution during bulk deformation processing. In order mechanical properties, and 3) processing path designs for to extend these methods to Mg, we had to characterize making nano-Mg-X in the form of bulk sheets and tubes. In the hardening law for Mg, which is non-trivial since Mg addition, we will offer new understanding in the nascent deforms by several distinct slip and twinning modes and field of hcp-X interfaces, the associated hcp-X interface develop a way to treat the formation of twin lamellae. models, and patentable SPD techniques. The latter turned out to be the first modeling technique to treat deformation twinning in a microstructural materials Publications model. We also had to extend these models to Nb. For Ardeljan, , I. J. Beyerlein, and Knezevic. A dislocation this, we had to develop for the first time a CPFE model for density based crystal plasticity finite element model: bcc materials that accounted for non-Schmid effects. The Application to a two-phase polycrystalline HCP/BCC 379

Page 380

composites. 2014. JOURNAL OF THE MECHANICS AND MICROSTRUCTURE AND PROCESSING. 600: 200. PHYSICS OF SOLIDS. 66: 16. Jahedi, , Ardeljan, I. J. Beyerlein, M. H. Paydar, and Ardeljan, M., I. J. Beyerlein, and M. Knezevic. A dislocation Knezevic. Enhancement of orientation gradients during density based crystal plasticity finite element model: simple shear deformation by application of simple application to a two-phase polycrystalline HCP/BCC compression. 2015. JOURNAL OF APPLIED PHYSICS. composites. 2014. Journal of the Mechanics and 117 (21). Physics of Solids. 66: 16. Jahedi, , M. H. Paydar, Zheng, I. J. Beyerlein, and Knezevic. Ardeljan, M., M. Knezevic, T. Nizolek, I. J. Beyerlein, S. Texture evolution and enhanced grain refinement J. Zheng, J. S. Carpenter, R. J. McCabe, N. A. Mara, under high-pressure-double-torsion. 2014. MATERIALS and T. M. Pollock. A multi-scale model for texture SCIENCE AND ENGINEERING A-STRUCTURAL development in Zr/Nb nanolayered composites MATERIALS PROPERTIES MICROSTRUCTURE AND processed by accumulative roll bonding. 2014. Journal PROCESSING. 611: 29. of Materials Science and Engineering. 63: 012170. Knezevic, , Drach, Ardeljan, and I. J. Beyerlein. Three Beyerlein, I. J., J. S. Carpenter, Hunter, L. S. Toth, and dimensional predictions of grain scale plasticity and Skrotzki. Nano-enabled orientation alignment via grain boundaries using crystal plasticity finite element extreme shear strains. 2015. SCRIPTA MATERIALIA. 98: models. 2014. COMPUTER METHODS IN APPLIED 52. MECHANICS AND ENGINEERING. 277: 239. Cao, L. e. i., Hunter, I. J. Beyerlein, and Koslowski. The role Knezevic, , I. J. Beyerlein, M. L. Lovato, C. N. Tome, A. of partial mediated slip during quasi-static deformation W. Richards, and R. J. McCabe. A strain-rate and of 3D nanocrystalline metals. 2015. JOURNAL OF THE temperature dependent constitutive model for BCC MECHANICS AND PHYSICS OF SOLIDS. 78: 415. metals incorporating non-Schmid effects: Application to tantalum-tungsten alloys. 2014. INTERNATIONAL Carpenter, J. S., Nizolek, R. J. McCabe, Knezevic, S. J. Zheng, JOURNAL OF PLASTICITY. 62: 93. B. P. Eftink, J. E. Scott, S. C. Vogel, T. M. Pollock, N. A. Mara, and I. J. Beyerlein. Bulk texture evolution Knezevic, , Jahedi, Y. P. Korkolis, and I. J. Beyerlein. of nanolamellar Zr-Nb composites processed via Material-based design of the extrusion of bimetallic accumulative roll bonding. 2015. ACTA MATERIALIA. tubes. 2014. COMPUTATIONAL MATERIALS SCIENCE. 92: 97. 95: 63. Carpenter, J. S., R. J. McCabe, J. R. Mayeur, N. A. Mara, Knezevic, , Nizolek, Ardeljan, I. J. Beyerlein, N. A. Mara, and and I. J. Beyerlein. Interface-Driven Plasticity: The T. M. Pollock. Texture evolution in two-phase Zr/Nb Presence of an Interface Affected Zone in Metallic lamellar composites during accumulative roll bonding. Lamellar Composites. 2015. ADVANCED ENGINEERING 2014. INTERNATIONAL JOURNAL OF PLASTICITY. 57: 16. MATERIALS. 17 (1): 109. Knezevic, , Zecevic, I. J. Beyerlein, J. F. Bingert, and R. J. Hunter, , R. F. Zhang, and I. J. Beyerlein. The core structure McCabe. Strain rate and temperature effects on the of dislocations and their relationship to the material selection of primary and secondary slip and twinning gamma-surface. 2014. JOURNAL OF APPLIED PHYSICS. systems in HCP Zr. 2015. ACTA MATERIALIA. 88: 55. 115 (13). Knezevic, M., T. Nizolek, M. Ardeljan, I. J. Beyerlein, N. Hunter, , and I. J. Beyerlein. Relationship between A. Mara, and T. M. Pollock. Texture evolution in two- monolayer stacking faults and twins in nanocrystals. phase Zr/Nb lamellar composites during accumulative 2015. ACTA MATERIALIA. 88: 207. roll bonding. 2014. International Journal of Plasticity. 57: 16. Hunter, , and I. J. Beyerlein. Unprecedented grain size effect on stacking fault width. 2013. APL MATERIALS. 1 Kumar, A., I. J. Beyerlein, and J. Wang. First-principles study (3). of the structure of Mg/Nb multilayers. 2014. Applied Physics Letters. 105: 071602. Hunter, , and I. J. Beyerlein. Predictions of an alternative pathway for grain-boundary driven twinning. 2014. Lentz, , Klaus, I. J. Beyerlein, Zecevic, Reimers, and APPLIED PHYSICS LETTERS. 104 (23). Knezevic. In situ X-ray diffraction and crystal plasticity modeling of the deformation behavior of extruded Hunter, , and I. J. Beyerlein. Stacking fault emission Mg-Li-(Al) alloys: An uncommon tension-compression from grain boundaries: Material dependencies and asymmetry. 2015. ACTA MATERIALIA. 86: 254. grain size effects. 2014. MATERIALS SCIENCE AND ENGINEERING A-STRUCTURAL MATERIALS PROPERTIES Lentz, , Klaus, Wagner, Fahrenson, I. J. Beyerlein, 380

Page 381

Zecevic, Reimers, and Knezevic. Effect of age hardening on the deformation behavior of an Mg- Y-Nd alloy: In-situ X-ray diffraction and crystal plasticity modeling. 2015. MATERIALS SCIENCE AND ENGINEERING A-STRUCTURAL MATERIALS PROPERTIES MICROSTRUCTURE AND PROCESSING. 628: 396. Xu, X. F., Y. Jie, and I. J. Beyerlein. A note on statistical strength of carbon nanotubes. 2013. Computers, Materials & Continua. 38: 17. Xu, X. F., Y. Jie, and I. J. Beyerlein. A probability model for the strength of carbon nanotubes. 2014. American Institute of Physics Advances. 4: 077116. Yuan, R. u. i., I. J. Beyerlein, and Zhou. Emergence of grain- size effects in nanocrystalline metals from statistical activation of discrete dislocation sources. 2015. ACTA MATERIALIA. 90: 169. Zhou, , C. J. O. Reichhardt, Reichhardt, and Beyerlein. Random organization in periodically driven gliding dislocations. 2014. PHYSICS LETTERS A. 378 (22-23): 1675. Zhou, , Reichhardt, C. J. O. Reichhardt, and I. J. Beyerlein. Dynamic Phases, Pinning, and Pattern Formation for Driven Dislocation Assemblies. 2015. SCIENTIFIC REPORTS. 5. 381

Page 382

Materials for the Future Exploratory Research Continuing Project Toward Tunable Functionalties Using Epitaxial Nanoscaffolding Films Quanxi Jia 20140371ER Introduction understanding, and exploiting emergent phenomena in Lattice-strained epitaxial nanoscaffolding films (epi- materials. In particular, this research develops/expands NSFs), in which a parallel array of nanoscaled material A our key capabilities outlined in the EPM ER category, e.g. interfaces with another material B and forms a regular i) functional design and optimization, ii) novel control lateral arrangement of A:B on a substrate, provide a and metrology techniques exploiting emergent phe- new design paradigm to tune/manipulate functional- nomena, iii) understanding mechanisms for control of ities that cannot be obtained in individual constituents material properties, and iv) synthesis and processing (A or B). Tuning/controlling functionality of a broad techniques that lead to control of emergent phenom- range of materials is emerging as an exciting direction ena. Furthermore, this project enables the development in materials research community. Built on our expertise, of new experimental (advanced synthetic and diagnostic integrated capabilities (from CINT, NHMFL, and T-4), and techniques) and theoretical capabilities to probe the preliminary experimental results, this project aims to state of complex nanoscale materials. This effort can develop new capabilities toward tunable functionalities ensure LANL’s continued and expanded leadership in using epi-NSFs where the vertical interface can act as nanostructured materials, innovative probing methods, an “active device”. Specifically, we will, for the first time and theoretical predication and design of new/improved in the field, investigate the tunable optical and resistive functional materials. switching properties based on La0.7Sr0.3MnO3:ZnO The success of this project enables LANL to respond ef- (LSMO:ZnO) and SrTiO3:Sm2O3 (STO:SmO) epi-NSFs, re- fectively to future BES calls since our proposed work di- spectively. The impact of this project reaches far beyond rectly addresses the Grand Scientific Challenge identified these materials systems. in the BESAC report: how do remarkable properties of Leveraging CINT’s well controlled laser-MBE for the matter emerge from the complex correlations of atomic growth of complex metal-oxide films, we target to or electronic constituents and how can we control these synthesize both LSMO:ZnO and STO:SmO epi-NSFs with properties? Beyond BES, this work will have enormous controlled microstructures. In parallel with modeling ramifications in technological sectors such as memory effort, we also use computational results to optimize and electro-optical devices. Finally, this work enables the processing parameters (which can affect the surface an important capability for MaRIE, since the ability to morphology of the film) and to tune the optical proper- design, synthesize, measure, and model functionality at ties of ZnO and resistive switching property (or so called different length scales in complex materials is the central “memristive” behavior) of STO. Combining our unique competence underpinning MaRIE’s M4 facility. Addition- characterization tools (such as transport measurement ally, this work exploits and expands the capabilities at and photoluminescence (PL) spectroscopy at a wide two LANL National User Facilities: CINT and NHMFL. range of magnetic fields and temperatures) with ad- Progress vanced TEM and high-resolution x-ray diffraction capabil- ity, our objective is to establish the processing-structure- In the past year, we made tremendous progress for functionality relationship of epi-NSFs. this project. This can be testified by our many refereed journal articles published in 2014 and patent application Benefit to National Security Missions related to this project. In the following, we summarize This work supports and strengthens the Laboratory’s the key technical accomplishments. core scientific capabilities essential to discovering, 382

Page 383

We, using LSMO:ZnO epitaxial nanoscaffolding films (epi- Φ-scan results. Detailed analysis also shows that the micro- NSFs) as a model system, have studied the effect of film structure of epi-NSF LSMO:ZnO on LAO is independent of thickness on the physical properties of nanocomposites. the film thickness. We show that strain, microstructure, as well as magne- Magnetic properties of epi-NSF LSMO:ZnO toresistance strongly rely on film thickness. These results To reveal the property-structure correlation in these confirm the critical role of film thickness on the microstruc- epi-NSFs, the magnetotransport and magnetoresistance tures, the strain, and functionalities. It further shows that (MR) of these films have been investigated at different one can use film thickness as a key design parameter to temperatures and magnetic fields. We have found that the design nanocomposite with optimum functionalities. microstructure evaluation with the film thickness should Optimizing the processing parameters for high quality be the dominating factor in determining the transport LSMO:ZnO epi-NSFs properties of the films on sapphire substrates. In the case LSMO:ZnO films with different thicknesses (10 nm to 200 of LSMO:ZnO films on LAO substrates, TEM results show nm) were grown on r-plane sapphire, c-plane sapphire, and that the microstructure does not change with the film LAO (001) by using pulsed laser deposition (PLD) with a KrF thickness. Thus, the metal-insulator transition temperature excimer laser (Lambda Physik LPX 300, λ = 248 nm, 2Hz). and resistivity variation with the film thickness could be The rectangle laser beam with an area of 5.55 mm2 was correlated with the strain effect. focused onto the composite target with an energy density Parallel connection channel model of 2.0 J/cm2. A substrate temperature of 750 °C and an In order to fit the magnetotransport results in the whole oxygen pressure of 100 mTorr were found to be optimal for temperature range, including the abnormal high resistance high quality epi-NSF LSMO:ZnO. at low temperatures, a modified parallel connection chan- Structural characterization using x-ray diffraction nel model was used to analyze the experimental results. The structures of epi-NSFs with different film thicknesses We have found that our results are consistent with the mi- on various substrates have been analyzed by X-ray diffrac- crostructure variations of the nanostructures on sapphire tion (XRD). It is noted that both LSMO and ZnO phases substrates. The phenomenological model is not suitable for are preferentially oriented out-of-plane. The Φ-scans of the thin nanocomposite films on LAO substrates since the LSMO \{202\} and LAO \{202\} show a fourfold symmetry, large strain effect can significantly alter the ferromagnetic suggesting that the LSMO phase on the LAO substrate is behavior of the LSMO phase. epitaxially grown (cube-on-cube). Different from the films on sapphire substrates, the out-of-plane lattice constant Future Work of films on LAO (001) strongly depends on film thickness. • Probe both the transport and the structural prop- The LSMO (002) peak monotonically shifts to the larger erties of La0.7Sr0.3MnO3:ZnO (LSMO:ZnO) and angles with increasing film thickness, indicating that the SrTiO3:Sm2O3 epitaxial nanoscaffolding films. strain varies with the thickness. We believe that when the nanocomposite films are thin enough, the strain in the • Tune the optical properties of LSMO:ZnO using exter- LSMO phase is controlled by the LAO substrate. However, nal magnetic field. the strain of LSMO phase is dominated by the secondary • Understand the structure-property relationship of ZnO phase when the films become thicker (200 nm). these materials. Microstructural characterization using transmission Conclusion electron microscopy To investigate the thickness dependent microstructures Our integrated capabilities to design, synthesize, and char- of the ei-NSFs, transmission electron microscopy (TEM) acterize different epi-NSFs with a range of controllable and and scanning TEM (STEM) in high-angle annular dark-field tunable properties make it possible to explore new physics. (HAADF) mode were conducted. We have found that the This project has also great technological impact on areas of ZnO nanopillars tend to embed in the LSMO matrix. The functional materials for application in memories, photonic LSMO phase is connected into much larger domains. It is devices, and sensing components. The following outlines interesting to note that both LSMO and ZnO nanodomains the expected scientific and technical results: are vertically and alternatively aligned on LAO substrate • Development of a set of principles to design epi-NSFs with uniform feature size through the film thickness. The with targeted functionalities. plan-view STEM image further reveals that the ZnO rectan- gular nanopillars are embedded in the LSMO matrix. These • Demonstration of magnetic field tunable optical prop- perpendicular ZnO domains are consistent with the x-ray 383

Page 384

erties of ZnO in LSMO:ZnO epi-NSFs. 24: 5240. • Illustration of much enhanced tunable memristive switching in STO:SmO epi-NSFs. Publications Chen, A., M. Weigand, Z. Bi, W. Zhang, X. Lv, P. Dowden, H. Wang, J. L. MacManus-Driscoll, and Q. X. Jia. Role of microstructure and strain on the magnetoresistance of nanocomposite films. 2014. Scientific Reports. 4: 5426. Jia, Q. X.. Self-assembled epitaxial nanocomposite films: their strain control and functionalities. Invited presentation at New Mexico Tech. (Socorro, NM, 7 March, 2014). Jia, Q. X.. Effect of interfaces on competing interactions of functional complex metal-oxides. Invited presentation at Univ. of Connecticut. (Storrs, CT, 17 OCt. 2014). Lee, S. B., A. Sangle, P. Lu, A. Chen, W. Zhang, J. S. Lee, H. Wang, Q. X. Jia, and J. L. MacManus-Driscoll. Novel electroforming-free nanoscaffold memristor with very high uniformity, tunability and density. 2014. Adv. Mater.. 26: 6284. Lee, S. B., W. Zhang, F. Khatkhatay, H. Wang, Q. X. Jia, and J. L. MacManus-Driscoll. Ionic conductivity increased by two orders of magnitude in micrometer-thick vertical yttria-stabilized ZrO2 nanocomposite films. To appear in Nano Lett.. Lee, S. B., W. Zhang, Q. X. Jia, H. Wang, and J. L. MacManus-Dris. Strain tuning and strong enhancement of ionic conductivity in SrZrO3-RE2O3 (RE = Sm, Eu, Gd, Dy, and Er) nanocomposite films. 2015. Adv. Functional Mater.. 25: 4328. Lu, P., E. Roemro, S. B. Lee, J. L. MacManus-Driscoll, and Q. X. Jia. Chemical quantification of atomic-scale EDS maps under thin specimen conditions. 2014. Microscopy & Microanalysis . 20: 1782. MacManus-Driscoll, J. L., A. Suwardi, A. Kursumovic, Z. Bi, C. F. Tsai, H. Wang, Q. X. Jia, and Q. J. Lee. New strain states and radical property tuning of metal oxides using a nanocomposite thin film approach. 2015. APL Mater.. 3: 062507. Yang, S. M., S. B. Lee, J. Jian, W. Zhang, Q. X. Jia, H. Wang, T. W. Noh, S. V. Kalinin, and J. L. MacManus-Driscoll. Strongly enhanced oxygen ion transport through samarium-doped CeO2 nanopillars in nanocomposite films. To appear in Nat. Commun.. Zhao, R., W. Li, J. H. Lee, E. M. Choi, Y. Liang, W. Zhang, R. Tang, H. Wang, Q. X. Jia, and J. L. MacManus-Driscoll. Precise tuning of (YBa2Cu3O7)1-x:(BaZrO3)x thin film nanocomposite structures. 2014. Adv. Funct. Mater.. 384

Page 385

Materials for the Future Exploratory Research Continuing Project Direct-gap Group-IV Nanocrystals:Cheap, Versatile Materials for Solar Cells Sergei A. Ivanov 20140446ER Introduction Mission in the Basic Understanding of Materials, and Cheap, non-toxic and abundantly produced, Group IV in the process will involve exploration of Fundamental semiconductors (silicon (Si), germanium (Ge), tin (Sn)) Chemistry. These studies directly seek to understand and are versatile materials for a broad range of devices. exploit the power of Defects and Interfaces in a material However, Si and Ge are indirect semiconductors, which system of unique potential importance to applications make their interactions with light (absorption or emis- in Energy Security, particularly in the area of Renewable sion) much less efficient than those of direct-gap semi- Energy, and as such lie at the core of the LANL “Materials conductors. This places severe limitations on how these for the Future” Pillar. The development of direct band materials can be used in solar cells, driving prices up and gap Group IV NCs by low-cost methods is also extremely holding utilization down. If one could, instead, convert relevant to ongoing and future DOE Office of Science these materials into a more versatile form that could be programs. Development of low-cost silicon-based photo- handled and processed cheaply, and that offered better voltaic materials will result in long-term benefits to LANL absorption properties, the effect on solar cell develop- efforts and eventually to federal stakeholders in these ment and deployment would be transformative. Even areas. simply rendering these materials amenable to, e.g., Progress screen printing, spray deposition and other fabrica- tion techniques suitable for roll-to-roll processing or In the second year of our project, we continued the the coating of surfaces of complex morphology, would development of the versatile procedure for the synthesis enable a host of new civilian and military applications of SnGe alloy nanoparticles together with the investiga- that are incompatible with rigid, brittle crystalline-Si. tion of the mechanism leading to their formation. As we We propose to use the unique power and flexibility of achieved deeper understanding of the factors beneficial colloidal synthesis to accomplish exactly that. The use for the increased load of tin in germanium lattice, we of colloidal nanocrystal (NC) synthesis places the ability were able to synthesize SnGe alloy nanoparticles with to create size-/composition-/shape-controlled NCs in 95% of tin load. As such, we have achieved full compo- device-relevant quantities into the hands of nearly any sitional variation of SnGe alloy, where we can change its researcher. Importantly, colloidal NCs are amenable to composition from pure Ge to Sn0.95Ge0.05 nanocrys- low-cost solution-based processing techniques, making tals! them extremely versatile. We seek to exploit the power Synthesized nanocrystals were also studied with absorp- of colloidal synthesis to create a new class of Group IV tion spectrometry in the attempt to elucidate the nature alloy NCs with direct band-gap behavior at solar-relevant of changes in the electronic structure of those products energies (1.0-1.5 eV) for use in a wide-range of emerging as a function of the amount of tin. The results obtained solar technologies. to date indicate that previous theoretical models might have underestimated the amount of tin needed to con- Benefit to National Security Missions vert germanium into the direct-gap material. This work combines a number of marquee capabilities of the Laboratory, including world-recognized nanoscience In parallel with the synthesis of single-source Ge-Sn and efforts towards a goal of tremendous potential impact, Si-Sn precursors, where the same molecule contains especially with direct relevance to applications in solar both needed elements, we are working on the modifica- energy capture and energy-efficient solid-state lighting. tion of our original approach to the synthesis of SnGe In this way, our work is directly tied to the LANL Science 385

Page 386

nanocrystals so that we could introduce the third element underexplored interplay of bulk and small size effects in – silicon – into the composition of the alloy. Preliminary semiconductors, advancing the understanding of electronic results have already revealed that this approach appears to structure and behavior in essentially all nanomaterial and be more productive than the originally proposed synthesis atomically disordered systems. of SnSi followed by the introduction of germanium. Intro- duction of silicon during the SnGe nanoparticle synthesis Publications allowed us to obtain the first batch of ternary composition Ramasamy, , P. G. Kotula, A. F. Fidler, M. T. Brumbach, J. M. SiGeSn nanoparticles! Direct incorporation of Sn into Si Pietryga, and S. A. Ivanov. SnxGe1-x Alloy Nanocrystals: is yet an elusive goal, but the discovered modification of A First Step toward Solution-Processed Group IV Photovoltaics. 2015. CHEMISTRY OF MATERIALS. 27 this approach with the initial synthesis of the ternary alloy (13): 4640. followed by the minimization of germanium content is a vi- able alternative that might lead to the SnSi alloy formation, the ultimate goal of our project. Future Work As the project is continues to be short funded, the theoret- ical investigation of the systems of interest will again have to be de-emphasized (but not fully abandoned). In the year to date, the following tasks will be performed. We will model the alloy NCs with different compositions and type of protecting molecules on the NC surface seek- ing to understand the best formulations for stable GexSn(1- x) and SixSn(1-x) alloy NCs. Together with theoretical as- sessment of NC stability, we will continue the development of a synthetic procedure for heavily alloyed germanium-tin NCs (with tin more than 90%). The research team will fully switch to the silicon-germani- um-tin formulations utilizing the same synthetic approach to map the composition boundaries of alloyed nanocrys- tals. Additionally, we will develop the approach to prepare uniform films of synthesized nanocrystals to assess their application in photovoltaics as well as a cathode in lithium batteries. All synthesized nanocrystals will be continued to be spec- troscopically characterized in order to establish the under- standing of their electronic structure on the way from poor to efficient light absorber. Conclusion Utilizing two different synthetic approaches combined with two methods of reaction activation, the goal is to synthe- size nanocrystals with formulations SixSn1-x and GexSn1-x. The synthetic efforts will benefit greatly from the theo- retical component, predicting compositions of greatest stability and Sn-distribution effects. Technologically, the alloy NCs with enhanced absorption will have an immedi- ate, direction-changing impact on research within solar-cell community in general. Fundamental studies of these for- mulations will provide an insight into the fascinating and 386

Page 387

Materials for the Future Exploratory Research Continuing Project Metal and Semiconductor Nanocrystal Superlattices Under Pressure: Multiscale Tuning of Structure and Function Jennifer A. Hollingsworth 20140456ER Introduction Therefore, we aim to apply a novel approach—high- The overarching goal of this project is to explore and pressure methods as an analytical and synthesis tool—to exploit high pressure as a novel means to precisely tune establish new understanding of NCSL formation and colloidal nanocrystal superlattices (NCSLs) from the stability toward design of new structures and controlled atomic to the nano- and mesoscales to realize: (1) new functionality. understanding of forces and interactions affecting super- Benefit to National Security Missions crystal stability, (2) novel NCSLs structures by controlled compression of self-assembled NCSLs, and (3) collective The effort directly addresses Basic Science focus areas, behavior as a function of interparticle distance. including Materials for the Future Pillar Focus Area 2: Advancing the understanding of materials functional- On their own, nanoparticle colloids of carefully con- ity through co-design using extreme conditions, in situ trolled size/shape and surface properties will self- measurements and User Facilities, as we make use of assemble into a surprisingly wide array of ordered capabilities located at LANL CINT and LANSCE, as well two-dimensional (2D) and three-dimensional (3D) NCSLs as Cornell’s High-energy synchrotron source for x-ray that mimic atomic lattices. These can comprise single scattering analysis (the CHESS facility). The work is also or multiple components, e.g., binary and ternary su- clearly responsive to envisioned MaRIE mission areas, perlattices (SLs), where the latter afford a richer variety especially “high-pressure nanoscience.” We also sup- of accessible structures and opportunities for develop- port Focus Area 3: Developing multi-functional materials ing collective functional properties compared to their architectures to transform structure/function integra- single-component counterparts. Significant progress has tion, performance and emergent properties. A success- been made over the last decade in “nanoparticle-colloid ful program will lay the groundwork for a predictive self-assembly,” with ~20 binary NCSLs now known. These “closed-loop” capability for designing/implementing have been constructed from two metals, two semicon- novel/functional nanocrystal superlattice structures that ductors, two magnetic nanoparticles, or combinations take advantage of collective, “emergent” behavior to thereof. realize important optical performance, especially “super- radiance.” Superradiance can significantly enhance the Surprisingly, however, efforts to exploit new functionality efficiency of nanoscale light sources without the need resulting from ordered arrays of dissimilar but interact- for an external cavity, with the potential to transform ing nanoparticles has been limited to ~six examples, solid-state lighting and telecommunication technologies. including interactions in magnetic-nanoparticle pairs, Also, the ability to couple otherwise discrete quantum enhanced electronic conductivity in semiconductor emitters over distance is likely to play an important role mixtures, and fluorescence quenching of a metal/semi- in the realization of quantum computation. Beyond conductor pair. In addition, there are interesting, but superradiance, the effort will enable other “collective limited, examples of enhanced stability imparted to functionality,” such as the ability to manipulate funda- nanoparticle components as a result of superstructuring, mental statistical properties of light emission through, in particular, stability against effects of elevated tem- with implications for the development of single-photon perature. Beyond this, however, understanding of the and entangled-photon-pair sources that are critically comparative stability of different nanocolloidal-crystal needed for quantum-communication applications. lattices and their tendency to transform into other struc- Beyond “light”: the targeted ability to precisely orga- tures remains lacking. nize interacting-nanocrystal superstructures comprising 387

Page 388

components of selected composition/functionality is key that for NCSLs of reasonable size (~1 micron or more) we for energy harvesting (solar, thermoelectrics), catalysis, en- can physically manipulate them into the DAC. We have ergy storage, and sensing, where coupled interfacial (e.g., characterized NCSLs by optical spectroscopy and obtained charge-separation/transfer) and extended-transport issues parameters for realizing NCSL formation for both electro- dominate. static assembly and hard-sphere self assembly (latter in hy- drophobic media), e.g., concentration of “screening” ions Progress (as small charged NCs or salt additives), NC concentration, The research team completed high-pressure diamond- time, non-solvent concentration. We now have a good anvil-cell (DAC) synchrotron X-ray diffraction (XRD) studies understanding so that in year 3 we will be able to shift our of two series of core/shell QDs. Here, core/shell systems emphasis from NCSL nucleation/growth to in situ pres- were compared with different tendencies for alloy forma- sure studies. That said, thus far we have prepared NCSLs tion. Namely, one system was anticipated to be highly from “well-behaved” NCs of uniform size/shape. In year susceptible to alloy formation via cation-cation migration 3, we will transition to larger and less uniform thick-shell and alloying, while the other was presumed less likely to quantum dots (giant QDs) and the hybrid gQD-Au systems alloy as the mechanism of alloy formation would have to toward realizing truly “collective” optical properties from be mediated by anion-anion exchange. Results from this the mixed excitonic-plasmonic systems. analysis were published. Future Work In addition, we completed an analysis of observed varia- Overall, we aim to apply a novel approach—high-pressure tions of phase transformation pressures with shell thick- methods as an analytical and synthesis tool—to establish ness and core/shell composition. These results afford key new understanding of colloidal nanocrystal superlat- insight into fundamental stability parameters (bulk moduli/ tice (NCSL) formation and stability toward design of new compressibility) of the core/shell quantum dots, and are structures and controlled functionality. To this end, we are currently being written up for a late-summer submission. pursuing three Research Goals, with specific Year 3 Goals delineated as follows: Beyond these fundamental studies of the effects of pres- sure on nanocrystal stability (as input into our superstruc- NCSL stability studies ture analyses), we have made significant progress this FY in High-pressure diamond-anvil-cell (DAC) synchrotron x-ray other key aspects of the program. Namely, a DAC with opti- diffraction (XRD) has now been completed of two series cal windows for conducting in situ observations of NCSL of core/shell QDs. In year 3, we will follow-up with some photoluminescence under pressure (Research Goal 3) was iterations of one of these systems — variations of the core/ purchased and installed in the CINT “single-nanostructure” thick-shell CdSe/CdS system to assess unexplored effects spectroscopy laboratory. This instrument has a built-in of crystal structure (cubic vs. hexagonal) and particle shape capability for exciting/recording photoluminescence as (hexagonal/faceted, pyrimidal, elongated/asymmetric) as part of a ruby luminescence method for pressure calcula- key parameters potentially influencing optical properties tion. Our initial aim was to “piggy-back” on this capability (quantum yield and stability). That said, our focus in year 3 and attempt to simultaneously pull our emission spectra will transition to Goals 2 and 3. from our NCSLs. However, although not explicitly stated in the company’s technical specs, the wavelength win- New NCSL structures through pressure dow that can be accessed is very limited due to the optics NCSLs (QD-QD; QD-Au) prepared in microfluidic-droplet integrated in the system. That said, we have designed a reactors will be transferred to our PL-integrated DAC for “work-around” (physical modification to the setup) that in situ measurements of PL as a function of pressure. The will allow us to couple the DAC to our spectroscopy setup resulting novel fused structures will be further character- and, thereby, to still obtain direct/correlated assessment ized structurally/optically outside the DAC. Furthermore, of NCSL emission performance with a precise pressure we have also established a capability for fabricating 2-D measurement. NCSLs using Langmuir-Blodgett techniques. Thus, we now have access to both types of superstructure, representing We have also made significant progress in our efforts to two extremes of NC-NC interactions. We will perform mo- prepare phase-pure/discrete NCSLs in micronfluidic droplet lecular dynamics simulations using the LAMMPS software “microreactors.” We have identified this approach (only a to study the differing self-assembly dynamical processes single report in the literature on this novel application of that are active in the formation our 2-D/3-D nanoparticle microfluids to crystal growth) as the only route to obtain- arrays. ing large NCSLs phase separated from amorphous NC aggregates and even other NCSLs. We have demonstrated New binary NCSL functionality through pressure: In the 388

Page 389

second half of year 3, a few of the “most interesting” bina- ry systems explored in (2) will be investigated for collective optical properties, where we target “superradiant” modes that spread over the entire supercrystal domain. Realiza- tion of optimized, extended superradiance is anticipated for Year 3 of the effort. Conclusion Until now, the development of new nanocrystal superlat- tices (NCSLs) has depended upon empirical observations of effects of size-ratio, surface-ligand/charge, solvent dielec- tric constant and electrophoretic mobilities. By combining our expertise in synthesis and computational chemistry with in situ high-pressure x-ray scattering and optical spec- troscopy, we will establish correlated structure-stability relationships and principles for predictive development of new superstructures, aided by “pressure sintering.” We will also demonstrate unprecedented collective behavior (e.g., superradiance) by precise tuning of inter-nanocrystal distances toward designed “colloidal metamaterials” for applications from light-emission to catalysis, photovoltaics & thermoelectrics; whereas, current examples are limited to small-cluster behavior or uncontrolled/disordered as- semblies. Publications Acharya, K. P., H. M. Nguyen, Paulite, Piryatinski, J. u. n. Zhang, J. L. Casson, Xu, H. a. n. Htoon, and J. A. Hollingsworth. Elucidation of Two Giants: Challenges to Thick-Shell Synthesis in CdSe/ZnSe and ZnSe/CdS Core/Shell Quantum Dots. 2015. JOURNAL OF THE AMERICAN CHEMICAL SOCIETY. 137 (11): 3755. Li, , Bian, Wang, Xu, J. A. Hollingsworth, Hanrath, Fang, and Wang. An Obtuse Rhombohedral Superlattice Assembled by Pt Nanocubes. 2015. NANO LETTERS. 15 (9): 6254. Quan, Z., D. Wu, J. Zhu, W. H. Evers, J. M. Boncella, L. D. A. Siebbelese, Z. Wang, A. Navrotsky, and H. Xu. Energy landscape of self-assembled superlattices of PbSe nanocrystals. 2014. Proceedings of the National Academy of Sciences . : 9054–9057. 389

Page 390

Materials for the Future Exploratory Research Continuing Project Interactions of Electrons with Quantum-Confined Systems Probed by Scanning Tunneling Spectroscopy Victor I. Klimov 20140495ER Introduction the boundaries in advanced instrumentation for stud- The beneficial aspects of quantum confinement as a ies of interactions of charged particles with individual means for controlling light-matter interactions are well nanostructrues by combing scan-probe and optical documented. On the other hand, the influence of con- techniques. Further, this project exploits quantum con- finement on interactions of solids with charged particles finement effects (e.g., relaxation of momentum con- (electrons and holes) still remains largely unexplored. servation) for enhancing the detection sensitivity. Since The case of charge injection is particularly important as interactions of energetic radiation with matter are medi- most practical devices make use of this type of excita- ated by “hot” electrons, the effects under investigation tion. One expected effect of confinement is a diminished are directly relevant to radiation detection. Further, due role of momentum conservation, which might lead to an to their potential use in IR avalanche photodiodes, these intriguing situation where interactions of electrons with studies also enhance capabilities in remote sensing and matter are constrained only by energy conservation. This quantum communication making these studies relevant would have major implications for several detection/ to the “Energy Signatures” strategy. Further, impact ion- sensing technologies, where the use of nanostructures ization can be exploited to enhance solar photovoltaics. could greatly increase sensitivity, potentially up to the Additionally, our studies support the “Materials” Pillar fundamental limit defined only by energy conservation. under the theme emergent phenomena. Specifically, an Here, we address this problem via comprehensive stud- underlying theme of this project is materials by design, ies of interactions of single electrons/holes with individ- which is the central vision of the controlled functional- ual semiconductor nanocrystals (NCs). The specific focus ity strategy. The fundamental aspects of our work are of is on impact ionization (ImI), that is, the collision of an interest to the DOE Office of Science. injected charge with a valence-band electron, which is Progress subsequently promoted to the conduction band. Elec- trons at well-defined energies will be injected into the Experimental results NCs from the tip of a scanning tunneling microscope and Light-modulated scanning tunneling microscopy/spec- the ImI efficiency will be inferred from measurements troscopy of PbS quantum dots (QDs) of the resulting luminescence. In addition to single-NC One of the goals of this project is to investigate light- sensitivity in detecting ImI events, this study will provide matter interactions within an individual quantum- comprehensive information on how the ImI efficiency confined semiconductor nanocrystal using methods relates to NC morphology and the structure of electronic of scanning tunneling microscopy (STM). Toward this states. Based on these measurements, we will answer an goal, we have measured a tunneling current through important fundamental question: can the ImI efficiency the individual PbS QD as a function of voltage applied in NCs approach the energy-conservation-defined limit? to the STM tip with and without illumination. Using the We also aim to unequivocally demonstrate that ImI laser beam directed towards the QD sample through an performance in nanostructures is superior to that of bulk optical port of the STM, we were able to modulate the solids. tunneling current using excitation rates of 0.01 – 0.1 photons absorbed per QD per ms. Furthermore, the high Benefit to National Security Missions spatial resolution of the STM combined with the light- The proposed work directly addresses the Science Pillar modulation capability enabled spatial mapping of both “Science of Signatures” especially the second theme the photocurrent and the dark current in QD monolayer “Revolutionize Measurements”. This project pushes 390

Page 391

and multilayer films, providing information on correlations pleted and sent for manufacturing. between the film morphology and photoconductivity. System for tunneling gap monitoring: Precise control and While our initial room-temperature measurements yielded evaluation of tip/sample spatial separation is crucial for current-voltage characteristics with sufficient signal-to- STM measurements. Therefore, we have developed a tip/ noise ratio for determining a QD band gap, further noise sample spatial gap monitoring system using a telescope suppression via cryogenic cooling is required for resolving and a CCD camera. This system allows for the sample/tip individual quantized transitions within the QDs. To achieve spatial gap inside the ultra-high vacuum chamber to be this goal, we have modified the STM design to allow for monitored on the computer screen. cooling of the sample holder with liquid nitrogen. First test measurements of this cryogenic capability are underway. Future Work During the next year, we will focus on the following two Sample preparation tasks: Sample preparation is perhaps the most critical step in ensuring the success of STM measurements. To this end, The research team will continue studies of PbS and PbSe we have explored a large parameter space for prepar- nanoplatelets using scanning tunneling spectroscopy (STS) ing “STM-grade” QD samples including various ligand measurements in order to determine the structure of treatments, different regimes of annealing, different film electronic states in these novel nanostructures. Specifi- thicknesses, and several different substrate materials. We cally, we plan to investigate the difference (if any) between found that depending on the amount of short ligands such the edge and the interior states. We will also probe the as 1,2-ethanedithiol (EDT), the films can exhibit either effect of nanoplatelet stacking on the structure/energies of metallic or semiconducting behavior. As these samples are STS features and attempt to derive the strength of inter- highly air-sensitive with surfaces prone to oxidation, we platelet coupling. The results of the STS measurements will are developing an ultra-high vacuum heater/dock bundle be analyzed using model calculations conducted within a that allows for in situ annealing of the sample in the main multi-band effective mass approximation. STM chamber for convenient transfer to the STM head. Using the high-resolution spatial imaging capability of the We will continue our instrument development work. The STM, we ensure that annealing and ligand treatment do next stage is the incorporation of the optical system into not change the morphology of the close-packed QD films the STM to allow for efficient pick-up of light emitted by a (avoiding, for example, fusion of the QDs). nanostructure in the course of excitation with a tunneling current. The mechanical parts of the pick-up system have STM upgrades been already machined. The fiber optical elements have Integration with optics: We have integrated a laser system been acquired. After optical and mechanical pick-up parts into the STM setup. This system is equipped with a modu- are assembled and installed in the STM chamber, we will lator for high-sensitivity, low-noise lock-in measurements test the complete system by analyzing scanning tunneling of excited photocurrents. For all QD sizes used in the mea- emission from bulk substrates (e.g., CdSe and CdS). Eventu- surements excitation intensities are sufficient to maintain ally these measurements will be extended to CdSe-based average QD occupancies close to or exceeding unity. This nanocrystals. ensures a large modulation of the tunneling photocurrent excited by the incident laser beam. Conclusion • Development of a novel STL technique for quantitative Low-temperature capability: We have acquired a special characterization of ImI with single-NC sensitivity and custom-designed liquid nitrogen Dewar that enables safe complete morphological information. transfer of liquid nitrogen through a flexible transfer line • Development of a quantitative theoretical model for with a special “Z” shape. Our initial tests have demonstrat- treating ImI in semiconductor NCs of various composi- ed successful cooling of the STM head to about 165 Kelvin. tions, shapes and internal structures. • Detailed experimental information on ImI yields in NCs Heating capability We have designed an ultra-high vacuum as a function of composition, size, shape (dots vs. rods) heater/dock allowing for in situ sample annealing. It con- and internal structure (e.g., core-only vs. core/shell). sists of a one-inch diameter heater unit surrounded by a heat shield with a sample dock on the top of the shield. • Development of nanostructures with ImI yields ap- The sample dock is designed to allow for transfer of our proaching those for the energy-conservation-defined annealed samples from the heater dock to the STM sample limit. stage using the wobble stick. The design has been com- 391

Page 392

Materials for the Future Exploratory Research Continuing Project Unraveling Interfacial Charge and Energy Transfer Processes in Single Layer 2D Transition Metal Dichalcogenides Aditya Mohite 20140540ER Introduction pillar of Materials for the Future. The formation of Graphene research has initiated rejuvenated interest inter-face states between TMDs with an understanding and served as a catalyst for the birth of a new field, of their fate and transport will allow us to go beyond beyond graphene, in the form of novel two-dimensional intrinsic optical and electronic properties. This project, (2D) layered materials also referred to as Transition therefore, directly addresses the materials strategy of metal dichalcogenides (TMDs). TMDs exhibit remark- controlled functionality by tailoring materials to perform able electronic and optical properties arising from the beyond their basic properties. The successful completion reduced dimensionality of a single unit cell perpendicu- of this ER will have tremendous implications for agencies lar to the crystal plane that hold the promise for trans- like DoD, Space Research, NASA etc for the use of thin formational research and development for next genera- film optoelectronic devices for remote applications and tion thin film optoelectronic applications. Unlike the as power sources for on field personnel. semi-metal graphene (zero electronic band-gap), TMDs Progress such as MoS2, MoSe2, WS2, WSe2, GaS, GaSe, InAs etc. are intrinsic semiconductors with band gaps spanning In the past year we have made significant progress in the the entire solar spectrum. This provides a unique op- synthesis and optoelectronic characterization of hetero- portunity to create layered homo and heterostructures structures with similar and different layered transition by combining two or more layers of these 2D materials metal di-chalcogenides (TMD) and achieved most goals and tailoring their interfacial properties to pave the path established at the beginning of the year. for current applications in phototransistors, photovolta- The main highlight was our work on creating highly ics, photodetectors, photocatalysis (H2 evolution/water efficiency metal-semicondcutor contacts using a novel splitting) and light emitting devices (LEDs) and transfor- phase engineering approach, which can reversibly mational technologies in spin and valleyoptoelectronics. convert a TMD from a semiconductor to a metal. We de- However in order to take advantage of the rich optoelec- veloped this approach to make clean and perfect metal tronic properties, it is critical to understand the charge -semiconductor interfaces with TMDs and demonstrated generation, recombination and separation processes that excellent, loss free charge transfer can be achieved. at the homo (create by two layers with semiconducting Specifically, we showed that by we can reduced the con- phase (2H) and/or metallic phase (1T) or and hetero in- tact resistance in an optoelectronic device and achieve terfaces (created by dissimilar TMDs with semiconduct- highly efficiency photoresponse as compared to the ing or metallic phase). In this project, we will combine pristine semiconductor device. This work was published single atomic layers of MoS2 and MoSe2 to form the first in several high impact papers (Nature Materials, ACS of its kind, layered TMD heterostructures that will en- Nano and APL) and most recently in a review article able us to establish the interfacial design rules, vital for in Chem. Soc Reviews, which is a high impact journal the understanding, control and manipulation of the flow (impact factor of 33). The phase transformation strategy of charge and energy across the TMD heterostructure also allowed us to test the catalytic performance of the interfaces. TMDs and by comparing the performance of 2-dimen- sional TMDs to 1-dimensional nanowires, we were able Benefit to National Security Missions to show that the mechanism of catalytic activity in 1D The novel physical phenomena we aim to explore in and 2D TMDs is different. We also demonstrated that this project will pave the way to make transition metal the origin of catalytically activity in 2D sheets is from the dichalcogenides (TMDs) viable assets in LANL’s science 392

Page 393

edges and not from the basal plane. This work is expected Conclusion to resolve a long standing controversy in the use of layered We will combine single atomic layers of MoS2 and MoSe2 2D sheets as catalysts. to form the first of its kind, layered TMDC heterostructures that will enable us to establish the interfacial design rules, In addition to this work, we also demonstrated a unique vital for the understanding, control and manipulation of CVD growth of TMDs using a modified precursor, which the flow of charge and energy across the TMDC hetero- leads to the growth of high quality single-crystalline TMDs. structure interfaces. The optical and electrical characterization of the grown TMDs demonstrated properties that are consistent with The success of this proposed work promises to redefine high optical grade semiconductors. For e.g. we were able the field of thin film optoelectronic devices that have to observe new features in the photocurrent and Raman tremendous advantages in their overall performance, form spectrum that have not been observed before. We worked factor, flexibility and cost over existing thin film technolo- with theorists and collaborators at Northeatern university gies based on Silicon. and OIST (Japan) to validate our experimental observations using theory and structural characterization using diffrac- Publications tion techniques such as LEEM and PEEM. This work is now Bilgin, , Liu, Vargas, Winchester, M. K. L. Man, Upmanyu, accepted in ACS Nano and is expected to be a significant K. M. Dani, Gupta, Talapatra, A. D. Mohite, and Kar. step in achieving high quality TMD samples for research Chemical Vapor Deposition Synthesized Atomically groups across the world. Furthermore, extended our Thin Molybdenum Disulfide with Optoelectronic-Grade growth strategy to growth heterostructures of TMDs such Crystalline Quality. 2015. ACS NANO. 9 (9): 8822. as MoS2/MoSe2 or MoS2/WS2 etc. Preliminary character- Cummins, D. R., Martinez, Kappera, Voiry, Martinez-Garcia, ization of these heterostructures has shown that the inter- Jasinski, D. a. n. Kelly, Chhowalla, A. D. Mohite, M. face between the two TMDs is atomically sharp that should K. Sunkara, and Gupta. Catalytic Activity in Lithium- allow for extremely efficient charge and energy transfer Treated Core-Shell MoOx/MoS2 Nanowires. 2015. process between these layers. This also gives us a handle JOURNAL OF PHYSICAL CHEMISTRY C. 119 (40): 22908. on understanding the interface states that are known to be associated with loss mechanisms in optoelectronic devices. Kappera, , Voiry, S. E. Yalcin, Branch, Gupta, A. D. Mohite, and Chhowalla. Phase-engineered low-resistance In the coming fiscal year, we will continue our efforts on contacts for ultrathin MoS2 transistors (vol 13, pg the characterization of heteorstructures created using 1128, 2014). 2014. NATURE MATERIALS. 13 (12). our novel growth technique and demonstrate a highly Kappera, , Voiry, S. E. Yalcin, Jen, Acerce, S. o. l. Torrel, efficient TMD device with performance close to that Branch, Lei, Chen, Najmaei, J. u. n. Lou, P. M. Ajayan, predicted through theoretical calculations. We are on Gupta, A. D. Mohite, and Chhowalla. Metallic 1T phase track to achieve our goal and the work is expected to lead source/drain electrodes for field effect transistors to several high impact publications. Currently, we have from chemical vapor deposited MoS2. 2014. APL published, 9 peer reviewed papers and the work has led to MATERIALS. 2 (9). 5 invited talks at international conferences and 3 invited Lei, , A. l. i. Sobhani, Wen, George, Wang, Huang, P. e. i. seminars at Universities in the US. Dong, B. o. Li, Najmaei, Bellah, Gupta, A. D. Mohite, Ge, J. u. n. Lou, N. J. Halas, Vajtai, and Ajayan. Future Work Ternary CuIn7Se11 : Towards Ultra-Thin Layered The goals for FY16 are: Photodetectors and Photovoltaic Devices. 2014. ADVANCED MATERIALS. 26 (45): 7666. • CVD growth of vertical and lateral TMD hererostruc- tures made from MoS2/MoSe2 Lei, S., L. Ge, S. Najamaei, A. George, R. Kappera, J. Lou, • Correlated photo current and photoluminescence to M. Chhowalla, H. Yamaguchi, G. Gupta, R. Vajtai, P. understand charge and energy transfer processes at Ajayan, and A. Mohite. Evolution of the Electronic the interface of these hererostructures. Band Structure and Efficient Photo-Detection in Atomic Layers of InSe. 2014. ACS NANO. 8 (2): 1263. • Based on the measurements above, understand design principles for stacking 3 or more hererostructures with Voiry, , Mohite, and Chhowalla. Phase engineering of cascading energy levels to obtain high performance transition metal dichalcogenides. 2015. CHEMICAL optoelectronic devices. SOCIETY REVIEWS. 44 (9): 2702. 393

Page 394

Materials for the Future Exploratory Research Continuing Project Microstructure Based Continuum Process Modeling of Weapons Metals Rodney J. Mccabe 20140630ER Introduction reproducible, desired performance. However, despite its We will develop an engineering scale modeling capability importance, process control decisions do not involve the that describes the evolution of microstructure, prop- effect of the evolving microstructures and properties of erties, and performance that accompany production these complex metals. Lacking this effect prevents the processes of uranium weapons components. We will weapons complex from moving forward into process- accomplish this by experimentally measuring and model- aware manufacturing and product-based qualification. ing the change in microstructure and properties that For posterity, we will have an experimentally validated, accompany thermal processes such as annealing and computationally efficient microstructure based model- heat treating. This newly developed thermal processing ing capability. This capability is a long sought after tool evolution law will be coupled with our well established for processing-aware manufacturing and engineering deformation based evolution law for uranium that performance analysis. Within the project we intend to describes the evolution of microstructure and proper- focus on the physics of the microstructure evolution of ties accompanying mechanical processes such as rolling uranium during both deformation and thermal drives. and forming. A novel aspect of this approach is that the The methods can be adapted to other weapons and non- microstructure and property evolution accompanying a weapons metals. mechanical deformation process will propagate into the model of a subsequent thermal process and vice versa. The innovative processing modeling framework de- Thus, microstructure and property evolution will be veloped will not be limited to only the cases and ap- tracked from the initial process through the final part. proaches we employ here. The continuum modeling We will also develop a novel, computationally efficient platform will be adaptable to other regimes provided the representation of our evolution laws in order to over- microstructure based constitutive law can be developed. come the computational challenges of implementing Validation of the developed models will ultimately be microstructure based evolution laws within engineering accomplished using MaRIE like experiments. This type scale simulations. In addition, this new modeling capa- of experiment is important in the science it will uncover, bility will be validated using MaRIE-like experiments. Our but also because it will drive the needs for improved in deliverables will enable LANL to 1) gauge the properties situ capabilities; temporal, spatial, and angular resolu- of existing stockpile components, 2) analyze the sensi- tion; analysis and handling of large data sets; and for- tivity of manufacturing process variables, 3) examine ward modeling needs that will be desired with MaRIE. alternative manufacturing processes resulting in similar microstructures and properties, and 4) directly assess Progress the engineering performance of metal weapons parts in Our objective is to develop a computationally efficient service. engineering scale modeling capability that describes the evolution of microstructure, properties, and perfor- Benefit to National Security Missions mance that accompany production processes of weap- Our deliverables fill a large gap in the processing/micro- ons components. To accomplish this, the project consists structure/property knowledge of weapons metals, an of manufacturing operations, experimental characteriza- achievement that will have widespread impact on sev- tion, and model and computational development. In the eral National Security missions. Manufacturing of ura- past 12 months, the research time has accomplished the nium and other metal weapons components tradition- following tasks: ally relies on process control to ensure that parts have 394

Page 395

• Specimens have been machined from uranium plates have been laid out and modifications to the deforma- having undergone seven different rolling schedules tion model have begun. (rolled in FY14). Future Work • Recrystallization experiments have been performed on Our objective is to develop a computationally efficient specimens from all seven different rolling schedules. engineering scale modeling capability that describes the • The microstructures of specimens from all seven have evolution of microstructure, properties, and performance been electron backscatter diffraction (EBSD) character- that accompany production processes of weapons com- ized in the as-rolled and recyrstallized conditions as ponents. To accomplish this, the project consists of manu- well as characterization of the starting materials. facturing operations, experimental characterization, and • Correlations between the as-deformed microstructure model and computational development. Whereas, most of and recrystallized microstructure, particularly the tex- these efforts during the first 15 months of the project have ture evolution, have been experimentally determined been to develop the physically based models and compu- for clock-rolled uranium and the results have been tational tools, most of the efforts in the next 12 months published (McCabe et. al., J. Nucl. Mat., 265 (2105) concern model validation and improvement. 189-195) and presented at the TMS 2014 annual con- Tasks for the next 12 months will include: ference. • Uranium shapes have been formed from cross-rolled • Machine, test, and characterize specimens for in situ uranium (from FY14) with experimental measures of neutron and X-ray diffraction studies at LANSCE and local strain for model validation. APS. These tests will simulate a complete thermo- • Modeling of uranium cross-rolling (from FY14) is mechanical processing mapping the microstructure finished and results will soon be submitted for publica- evolution throughout a deformation and recrystalliza- tion (Zecevic et al., “Origin of texture formation in or- tion step on individual samples. thorhombic alpha-uranium under simple compression • Form parts with experimental strain mapping from and rolling to high strains”, J. Nucl. Mat. in preparation. straight and clock rolled uranium that have been • Computationally efficient, spectral crystal plasticity experimentally characterized and modeled through an for a surrogate material was successfully coupled with entire mechanical and thermal processing. Character- the implicit version of the commercial finite element ize the microstructures of the formed parts. package Abaqus. These results have been published • Finish analyzing the straight rolled electron backscatter (Zecevic et al., “A new implementation of the spectral data to quantitatively define the relationship between crystal plasticity framework in implicit finite elements”, initial grain size and rolling reduction to final recrystal- Mech. of Mat., 84 (2015) 114-126) lized grain size. • An efficient computational framework to model the • Finish integration of additional material models into mechanical processing operations has been imple- our computationally efficient framework. Develop mented for surrogate materials using previously spectral plasticity representation for uranium and test published data. Benchmarking of this framework is its validity against the visco-plastic self-consistent mod- nearly complete; preliminary results were presented el we have been using to successfully model uranium. at the 2015 Mach Conference in Annapolis, MD, and • Finish making modifications to our deformation model finished results will soon be submitted for publication to allow us to account for the microstructure charac- (Richards et al., “Acceleration of macroscale plasticity teristics important for the recrystallization model. simulations using GPUs”, in preparation). Integration • Run modeling simulations equivalent to our three pro- of additional material models into the framework is cessing step validation experiments (rolling, forming, underway. recrystallizing) and compare with experimental results • A toy model to visualize the effects of different recrys- Conclusion tallization laws has been implemented, and is being used to identify the necessary parameters to develop The overall technical goal is to develop a computation- physically realistic models of recrystallization. ally efficient, deformation and thermal processing physics based modeling capability allowing for accurate prediction • Based on experiments and the toy model, the founda- of microstructure, property, and performance evolution tion for using the output from our deformation model at the engineering level. This work will provide scientific as the input for the recrystallization model has been understanding of the evolution of the defects that are the established. The tasks necessary to modify the defor- mechanisms of deformation and the driving force for ther- mation model to account for the microstructure char- mally driven microstructure changes. The developed mod- acteristics important for the recrystallization model 395

Page 396

eling capability will allow for reliable modeling of metals processing and engineering performance in uranium and other complex metals. The computational tools developed will allow better physics to be added to larger simulations of components and engineering systems. Publications Ardeljan, , R. J. McCabe, I. J. Beyerlein, and Knezevic. Explicit incorporation of deformation twins into crystal plasticity finite element models. 2015. COMPUTER METHODS IN APPLIED MECHANICS AND ENGINEERING. 295: 396. Knezevic, M., J. Crapps, I. J. Beyerlein, D. R. Coughlin, K. D. Clark, and R. J. McCabe. Anisotropic modeling of structural components using embedded crystal plasticity constructive laws within finite elements. To appear in International Journal of Mechanical Sciences. Lebensohn, R. A., M. Zecevic, M. Knezevic, and R. J. McCabe. Average intragranular misorientation trends in polycrystalline materials predicted by a viscoplastic self-consistent approach. Acta Materialia. McCabe, R. J., A. W. Richards, D. R. Coughlin, K. D. Clarke, I. J. Beyerlein, and M. Knezevic. Microstructure effects on the recrystallization of low-symmetry alpha-uranium. 2015. Journal of Nuclear Materials. 465: 189. Zecevic, M., M. Knezevic, I. J. Beyerlein, and R. J. McCabe. Origin of texture formation in orthorhombic alpha- uranium under simple compression and rolling to high strains. Journal of Nuclear Materials. Zecevic, M., R. J. McCabe, and M. Knezevic. A new implementation of the spectral crystal plasticity framework in implicit finite elements. 2015. Mechanics of Materials. 84: 114. 396

Page 397

Materials for the Future Exploratory Research Continuing Project Solute and Microstructure Prediction during Processing (U) Amy J. Clarke 20140639ER Introduction the physics needed to capture non-continuum events All metallic alloys experience solidification, which occurs that dictate the microstructure, such as pattern forma- over multiple length and time scales. Thus, achieving tion at the solidification front or the geometric form of a intended solidification structures in metallic alloys is key dendrite. Our coupling of in-situ characterization during to achieving properties and performance. To predict so- processing with our materials and modeling expertise lidification structural development, we must understand will enable unprecedented processing and microstruc- how solidification structure develops from the micro- to tural linkages for high-density metallic alloys; it will also the macroscale. This can only be achieved through a create the necessary infrastructure for future applica- coupled experimental and modeling campaign. Our goal tion to actinides. The integration of multiscale physics is to make Los Alamos National Laboratory’s macro- into a continuum, engineering-scale code for structural scopic (or continuum engineering) scale casting simula- development is the critical link needed to perform tion code, Truchas, microstructure-aware. Our work will process-aware manufacturing. Our proposed work will combine direct observations of solidification in high-den- also highlight the need for advanced experimental and sity metallic alloys at the micro- and macroscale using computational tools to extend our predictive capabilities synchrotron x-ray and high-energy proton imaging, along to emerging manufacturing technologies, such as addi- with computational microstructural modeling, to permit tive or in-situ fabrication. the identification, development, and incorporation of Progress important analytical expressions for structural develop- ment into Truchas. Our coupled experimental and mod- We have performed in-situ casting mold filling and eling campaign will make unprecedented processing and solidification experiments at pRad (the 800 MeV proton microstructural linkages for high-density metallic alloys; radiography facility at Los Alamos National Laboratory, it will also create the necessary infrastructure for future LANL) to improve our understanding of alloy melt fluid application to actinides. flow and the role of thermal conditions on microstruc- tural development during the casting process. We Benefit to National Security Missions also performed initial in-situ x-ray imaging at Argonne We currently do not have experimentally informed National Laboratory’s Advanced Photon Source (APS) casting or solidification microstructural development to monitor dendritic microstructural development in an models at Los Alamos National Laboratory. While de- aluminum-silver alloy and subsequent chemical redis- tailed microscale structure predictions over engineering tribution (homogenization) during iso-thermal holding length scales (i.e., centimeters) are still a distant goal, at an elevated temperature. The casting experiments from this work first order approximations of structure are particularly valuable for assessing larger length-scale at the continuum level will be attainable with current solute segregation information and generating a range computational tools. Casting simulations have been of microstructural characteristics for comparison with performed at Los Alamos National Laboratory for over Truchas modeling. We have performed ex-situ micro- 20 years with casting codes based on continuum archi- structural (electron microscopy) and chemical (EDS, tecture that consists of coupled differential equations energy dispersive spectroscopy and MXRF (Macro X-Ray solved in a meshed environment. They perform well Fluorescence) examinations of the castings we produced for fluid flow and macroscopic heat flow and are able at pRad. The measurements are ongoing for tin-bismuth to predict what temperature exists at any given cell. and gold-zinc-aluminum alloy castings, and we are Nonetheless, continuum codes currently do not include beginning to explore mechanical properties for selected 397

Page 398

samples. We have also evaluated the deformation of local predictions made with microstructural modeling and aluminum-copper alloy samples made previously at APS to experiments, and the demonstration of MaRIE-like experi- start linking microstructural development to mechanical ments that follow solidification microstructural develop- performance, both experimentally and computationally. ment to mechanical properties (i.e., certification-type We also evaluated phase selection in gold-zinc-aluminum experiments). We will not develop a micro-scale homog- and bismuth-antimony alloys during solidification with enization model or link homogenized microstructures to HIPPO (High-Pressure-Preferred Orientation) at the Los properties, but we will have performed preliminary in-situ Alamos Neutron Science Center (LANSCE) at LANL to help homogenization experiments, identified suitable alloy(s) provide modeling inputs. for follow-on studies, and have the modeling framework within Truchas. A flexible mathematical framework has been developed within Truchas, which can accept analytical models for so- For FY16, our primary goal will be to inform and validate lidification that produce predictions of final microstructural solute and structure predictions from Truchas. We will features. The next step, which is underway, is to incorpo- also explore the link of solidification microstructural evolu- rate solute information into the final predictions. Our mod- tion to mechanical properties. To exploit opportunities eling focus this year has been to treat the flow conditions for strengthening the structure models, we will also use that assume rapid solute mixing in the liquid and minimal thermal parameters from Truchas as boundary conditions diffusion in the solid. Our methodology of incorporating for local micro/meso-scale modeling of microstructural analytical expression into Truchas permits the rapid assess- development with other techniques for comparisons. ment of model applicability and systematic improvements. Conclusion Exploration of various micro-scale modeling schemes (i.e., We will create a flexible computational framework for phase-field, front tracking, and dynamic needle network the multiscale prediction of high-density metallic alloys (DNN)) has led us to two important conclusions. Phase- during solidification from the micro- to the macroscale by field simulations, in conjunction with experimental micro- extending a continuum casting model (Truchas) to include structure development, have helped to identify trouble fast, analytical models for microstructure. We will use areas for both analytical models and other types of mod- proton imaging to experimentally measure macroscale eling schemes. The combination of computational speed phenomena and x-ray imaging to watch the evolution of and good experimental matching of DNN for features key microstructural features, similar to those in actinides, such as primary dendrite spacing has made it the primary in high-density alloys. Our experiments will allow us to candidate for conducting simulations based upon Truchas identify the critical parameters needed for the prediction outputs. Due to enhanced applicability to typical casting of microstructure in a casting with Truchas at a length scale conditions, the micro-modeling focus has naturally shifted never before accomplished. back to diffusion-limited, rather than kinetically-limited, growth conditions. Future Work The scope of work has changed due to continued un- certainty and reduction in budget allocation. While the end goal to make Truchas microstructure-aware has not changed, we have strengthened our focus on predicting dendritic microstructural evolution, especially in larger volumes relevant to casting, with comparisons to experi- mental microstructures and mechanical properties. We will focus less on homogenization, and more on dendritic microstructures produced by solidification. Essentially, we will perform dendritic microstructural predictions with Truchas and micro-scale models, evaluate solidification microstructures produced by in-situ and ex-situ micro- and macro-scale experiments for comparisons with modeling, and will explore the link to mechanical properties. Our end milestones will be the development of a process (cast- ing) model that can predict microstructure, tested against 398

Page 399

Materials for the Future Exploratory Research Continuing Project In situ X-ray Imaging and Diffraction to Understand the Mechanics of Initiation Mechanisms in Explosive Single Crystals Kyle J. Ramos 20140643ER Introduction Benefit to National Security Missions The project is a joint experimental/theoretical investiga- Success in this project will motivate and guide techno- tion of the deformation mechanisms in the energetic logical development and further experimental work at molecular crystal (cyclotrimethylene trinitramine) RDX. the NNSA’s Dynamic Compression Sector at the Ad- The localization of mechanical deformation in explosives vanced Photon Source and proposed facilities such as under shock compression has been linked to the on-set LANL’s MaRIE. of chemical reactions and the initiation of detonation. The development of a mechanistic understanding of New capabilities for real-time, in situ X-ray diffraction the response of energetic constituents to impact will and imaging using the IMPULSE capability at the Ad- provide the first framework for predicting the material vanced Photon Source will be employed to measure conditions for initiation. This capability will significantly the evolution of the average lattice strain and spatially affect high explosive science at LANL and beyond and is localized material failure in RDX single crystals during critical to our national security mission. NNSA Advanced dynamic compression. These data will be interpreted Simulation and Computing (ASC) and the Department and eventually predicted through the development of of Defense (DoD) strongly support in situ measurements a single crystal plasticity model. The development of a and emphasize the need for predictive materials mod- validated, predictive thermomechanical model for RDX els. In addition, new thermomechanical models and will enable us to model the response of more complex the ability to predict initiation are indispensable to the microstructures and defects at the completion of the development of reduced sensitivity explosives for DoD project. The single crystal plasticity models will take as and Department of Energy (DOE) insensitive munitions input existing data from the literature, in situ data from via particle and crystal engineering. IMPULSE and the results of first principles electronic structure calculations of the orientation-dependent Progress thermophysical properties of RDX. The use of first prin- ciples calculations are critical when we will model the The research team made outstanding progress on all as- high-pressure phase of RDX that cannot be recovered pects of the project over the last 12 months. Significant for characterization in the laboratory. The single crystal advances in theory and simulation have been achieved plasticity model will be validated via its implementation and in situ X-ray diffraction experiments at the Advanced into a finite element simulation code from which we will Photon Source (APS) have moved ahead. Details on spe- compute diagnostics, including interface velocimetry and cific technical advances are as follows. X-ray diffraction patterns that can be compared directly Major contributions were made to a recently published with in situ experiments. review article on the elasticity of energetic molecular The ability to measure the structure response of com- crystals [D. E. Hooks, K. J. Ramos, C. A. Bolme, and M. J. plex materials in the nanosecond time scales before Cawkwell, Propellants Explosives Pyrotechnics, 40, 333 they are destroyed during shock compression is a major (2015)]. This review article allowed us to dissect the innovation that this project will advance. Nevertheless, experimental protocols for measuring the elastic con- the integration of these measurements with theory and stants of brittle, low symmetry molecular crystals and simulation that can rationalize and predict responses how elastic tensors may be predicted by first principles based on the activity of the underlying deformation or molecular dynamics simulations. Since accurate ten- mechanisms breaks new ground in shock physics. sors of elastic constants are of critical importance to 399

Page 400

modeling the mechanical response of explosives, this work test our ‘anomalous hardening‘ model and subsequent satisfied major deliverables of the project and made a sig- discovering of a monoclinic phase as described in item six nificant contribution to the scientific literature. above. Four frames of data were captured per experiment with very high signal-to-noise and detailed the deforma- A new equation of state for alpha RDX was developed tion process. based on extremely accurate density functional theory calculations and quality literature data. We are currently Next steps include analyzing the X-ray diffraction patterns making good progress on the development of an equa- and comparing them to results from recently developed tion of state for the high pressure gamma polymorph of finite element model and X-ray diffraction simulations RDX that will allow us to model phase transformations in of the crystalline deformation during shock loading. The the material, and the tetragonal explosive PETN. A novel analysis will enable the extraction of previously unattain- anisotropic constitutive model has been proposed, param- able equation-of-state information and the development eterized to available literature and calculated data, and and validation of an anisotropic plasticity model crucial for implemented in the ABAQUS finite element simulation predicting thermomechanical localization of deformation package. Continuum-scale simulations of impact problems and associated heating. Further experiments have already in RDX single crystals, using known systems for plastic flow, been designed and built and will be completed in July. are currently underway. Future Work A state-of-the-art single crystal plasticity model was de- Our goals for the third year of the project include: veloped and implemented in a 1-D hydrocode. Small-scale simulations using these methods highlighted deficiencies • Continued application of shear stress to dislocations in in our understanding of dislocation-mediated plasticity in alpha-RDX in atomistic simulations to evaluate barri- RDX. These gaps in our knowledge will be addressed with ers to dislocation motion. These data will inform the new, carefully designed plate impact experiments. generation of single crystal plasticity models. • Continued development of an anisotropic equation of Code was developed for simulating X-ray diffraction pat- state, based on our new formalism, for gamma-RDX terns with the beam parameters used at the APS. We have and/or other potential phases that are discovered tested the code using datasets obtained from earlier large- in atomisitic simulations and XRD (X-ray diffraction) scale molecular dynamics simulations. experiments. • Perform plate impact experiments for phase transfor- We performed large-scale atomistic and molecular dynam- mation investigation using XRD diagnostic. ics simulations of the core structure of a dislocation in al- • Compare XRD experimental results with computed pha RDX. We hope these simulations will allow us to better diagnostic from simulations. understand the slip in molecular crystals at a fundamental level. • Apply the new models in the simulation of more com- plex microstructures Our ‘anomalous hardening’ model for the unusual shock response of specific orientations of RDX single crystals to shock was reevaluated. By reanalyzing data from large- Conclusion scale molecular dynamics simulations we discovered a We will deliver a validated, anistropic thermomechanical previously unknown monoclinic phase of RDX that can only model for the response of single crystals of the explosive be accessed by dynamic compression. Based on unit cell RDX to shock compression. We shall construct anisotropic, parameters obtained from density functional theory cal- temperature-dependent equation of state for RDX based culations and simulated X-ray diffraction patterns, we will on experiment and theory. The equation of state will be perform a series of in situ X-ray diffraction experiments at employed in a single crystal plasticity model that incorpo- the APS during the next run cycle to confirm the presence rates deformation mechanisms and phase transformations of the hypothetical new phase. that have been observed experimentally. Rates for the deformation mechanisms will be parameterized to experi- The first in situ diffraction patterns from explosives were mental data. The ability to model deformation processes recorded during impact at the APS. The hardware for the in explosives pertains directly to explosive initiation and X-ray diffraction was designed and successfully imple- safety, both of which are of considerable importance to on- mented on the IMPact System for Ultrafast Synchrotron going NNSA and DoD missions. Experiments (IMPULSE). Single crystal RDX samples were prepared for impact on \{210\} and \{021\}-orientations to 400

Page 401

Publications Hooks, D. E., K. J. Ramos, C. A. Bolme, and M. J. Cawkwell. Elasticity of Crystalline Molecular Explosives. 2015. PRO- PELLANTS EXPLOSIVES PYROTECHNICS. 40 (3): 333. 401

Page 402

Materials for the Future Exploratory Research Continuing Project Enabling Mesoscale Science: Nonlocal Dislocation-Flux Crystal Plasticity Under Shock Loading Conditions Darby J. Luscher 20140645ER Introduction Benefit to National Security Missions Our team will develop and implement a novel physics- The length and time scales relevant in simulations of based model for dislocation-mediated single crystal weapon performance motivate the development of mac- plasticity, applicable under shock loading regimes roscale models that capture the essence of dominant characterized by large deformation, deformation rate, physical processes at finer scales. Examples include plas- and high pressure and temperature. This endeavor is tic slip, void nucleation, growth, and coalescence, phase motivated by challenging mesoscale problems such as transformation, and twinning. These physical processes damage nucleation at material interfaces. The novelty of depend critically upon microstructural details, such as the proposed formulation stems mainly from its explicit grain morphology, orientation distributions, grain bound- representation of dislocation flow through the crystal lat- ary characteristics, and defects. The development of tice, coupled with a detailed accounting for elastic inter- macroscale models reflecting these processes demands actions between dislocations. These features establish a an understanding of the underlying physics, which is new paradigm for crystal plasticity, which is fundamen- often gained through experiment. Currently available tally different from existing models in that arena. diagnostic measurements for shock-regime experiments are often ambiguous in identifying specific mechanisms Specifically, our single crystal plasticity model will in- of these processes. For example, measurements of free clude (1) a proper treatment of the evolution of dislo- surface velocity (VISAR) require speculative inference to cation fields, built upon balance laws governing their draw conclusions about damage nucleation kinetics. transport through the lattice, (2) a physically consistent representation of long-range nonlocal interactions of Development of new in-situ diagnostic technologies, for dislocations, and their role in resisting (or enhancing) example, through MaRIE, promises to provide detailed slip, (3) kinetics models for (a) dislocation nucleation, data associated with these im-portant physical process- multiplication, and annihilation, consistent with large es. In many cases, such newly available data would not deformation rates and for (b) dislocation velocity, ac- be amenable to direct comparison with results obtained counting for phonon-drag and inertial effects acting on using existing models. The connection between micro- dislocations, and (4) boundary conditions that have a structural processes and simulations of weapon perfor- direct physical interpretation. Our work will deliver (1) mance demands insight from mesoscale modeling, to a coupled-physics, mixed-field implementation of the provide more detailed interpretation of currently avail- theory, amenable to numerical simulations of polycrystal able measurements and to be predictive on a commen- response, and (2) novel simulations of dislocation pile- surate level of resolution with anticipated diagnostics. up leading to damage nucleation in copper polycrystals Our research will deliver a mesoscale simulation capabil- under shock loading. ity implemented into an ASC LAP code. These simulation High-risk aspects are numerical issues related to the tools are expected to deliver on the promise of nonlocal computational implementation of mixed-field phys- modeling strategies, namely the ability to predict scale- ics calculations, development of models for computing dependent material response as part of the solution of stress fields in the vicinity of dislocations, and the ag- a multi-physics problem, without requiring any length- gressive schedule of code implementations. scale parameters to be specified as inputs. 402

Page 403

Progress problem will be solved within FLAG by leveraging the ad- In FY15, we continued to develop the prototype code vection schemes within FLAG to remap solution variables and apply this code to solving relevant boundary value for arbitrary Lagrange/Eulerian (ALE) calculations. Identify- problems. The extensions of the prototype code include a ing the solution strategy for the DDC problem within FLAG translation from the relatively slow running python source continues. We anticipate that these tasks will be prelimi- code into fortran code that typically runs two orders narily completed by the end of FY15 and will leave us in of magnitude faster. In several cases, numerical testing good position to successfully achieve FY16 stage gate. revealed deficiencies in the implementation scheme that Future Work were subsequently resolved. At this point, the 1D proto- type code is relatively mature and the overall implementa- In FY16 we will continue to apply the prototype 1D code tion strategy has stabilized. to simulations of various plate impact scenarios. These simulations are intended to challenge the numerical Theoretical development continued in FY15, focusing on implementation and reveal where improvements to the constitutive relations; specifically, we have compared the reliability of the coupling and integration schemes can be competing influence of various dislocation evolution terms. made. Also, we will develop an improved kinematic theory This work has led to an increased understanding of the role for solving the DDC problem. The original FY15 goal of of dislocation nucleation, multiplication, annihilation, and multi-dimensional implementation of each sub-problem motion within the context of shock loading. into an Advanced Simulation and Computing (ASC) La- grangian Applications Project (LAP) code will be completed We modeled the response of single-crystal plate impact during FY16. Full coupling of these sub-problems will begin problems with various combinations of simulation param- toward the latter part of FY16. eters in order to help identify where the computational algorithms should be improved. We have demonstrated Stage Gate 3: Tested implementation of each sub-problem application of the theory to a notional interface problem into ASC code. A demonstration of the full coupling of where dislocations are unable to slip across an explicit these sub-problems to a 3D shock loading problem. interface within the problem. We have also developed a test problem that explores the notion of mixed-domain Conclusion strategy where the DMB problem is solved everywhere We will deliver simulation tools built around the mesoscale throughout the problem domain, but the CDT and DDC physics-based nonlocal models of plasticity developed by problem is only solved near interfaces that would resist our research that will deliver on the promise of nonlocal dislocation motion. modeling strategies, namely the ability to predict scale- dependent material response as part of the solution of a The overall theoretical framework and application to three multiphysics problem, without requiring any length-scale plate impact scenarios has been included in a manuscript parameters to be specified as inputs. Our research will accepted for publication in the International Journal of have significant impact to LANL, the DOE’s Office of Sci- Plasticity. The reviews for this publication were favorable ence, and the U.S., by virtue of its critical role in enabling reinforcing the timeliness of this coordinated multi-physics mesoscale science needed for modeling nucleation and approach to dislocation mediated plasticity for shock load- evolution of defects at material interfaces. ing environments. The numerical implementation and an associated parameter sensitivity study were included in a Publications manuscript submitted for publication. Two invited lectures Luscher, D., J. Mayeur, H. Mourad, A. Hunter, and M. have been given based on this work in FY15. Kenamond. Coupling continuum dislocation transport with crystal plasticity for application to shock loading We initiated the multi-dimensional implementation of conditions. 2016. International Journal of Plasticity. 76: each sub-problem into an Advanced Simulation and Com- 111. puting (ASC) Lagrangian Applications Project (LAP) code. The DMB problem comprises a solution scheme to satisfy conservation of momentum and a constitutive integra- tion scheme to update crystal plasticity calculations. Our DMB problem will leverage the existing solver within FLAG. Implementation of the crystal plasticity update has been initiated via the creation of a “node” within the database hierarchy of FLAG. The subroutines to evaluate the consti- tutive model are being ported to this FLAG node. The CDT 403

Page 404

Materials for the Future Exploratory Research Continuing Project Embedded Fiber Sensor Approach for Dynamic Pressure and Temperature Measurements in Explosives George Rodriguez 20140650ER Introduction Benefit to National Security Missions From the standpoint of high explosive (HE) engineering The project relevance is intimately tied to new diagnos- and science, there is a critical need across the Depart- tic capability for the Laboratory’s core mission activi- ments of Energy and Defense for in-situ probes that can ties that include high explosive (HE) campaign science, continuously measure thermodynamic state variables weapons surety and engineering, explosives chemistry (pressure, temperature, and detonation velocity) under and equation-of-state, initiation systems, and others that conditions of extremes. To date, very few approaches will immediately benefit Department of Energy and De- exist that are capable of accurately measuring state partment of Defense programs in HE science, materials variables independently without referring to a model dynamics, and surety. The diagnostic probe developed or inferring from other parameters such as velocity or under this project will yield dynamic measurements density. This project will use an embedded fiber optic of pressure and temperature conditions in controlled approach that places sensors at the point where the thermal explosions with secondary explosives. The measurements need to be made, and makes it possible experimental information will advance the Laboratory’s to measure gradients from violent HE chemical reac- understanding of explosives by providing validation data tions if the sensors can be distributed in the HE. Fur- needed for reactive explosive models. The in-situ pres- thermore, the sensors need to be impervious to harsh sure and temperature probe being developed is clearly environments where HE reaction chemistry can produce aligned with Laboratory Mission relevance including extreme temperature and pressure in conditions from Matter in Extremes call Topical Area 3: Diagnostics to slow burn cookoff to detonation. Under this project, we Measure Temperature, Pressure (Stress), Strain, Veloc- will establish a new Laboratory diagnostic capability in ity, and their Gradients. Mission relevance to Nuclear high pressure science and thermometry. The diagnostic Weapons Programs in performance validation for is transformative, because it will provide the Laboratory Directed Stockpile Work, Surety, and Campaign Science the ability to measure conditions of HE burn, deflagra- is identified because the project develops a desirable tion, and detonation propagation. The diagnostic is sensor diagnostic capability currently not available. It is intimately tied to Laboratory models for fundamen- expected that other agencies involved in high explosive tal material inputs that initialize HE equation-of-state burn reactive flows ( i.e., NASA) would be interested in models (EOS), reaction rates, and by-product EOS for the technology as well. predictive HE science. The probe is indifferent to mate- rial type, and the diagnostic should be applicable to inert Progress materials (metals, plastics, liquids, etc.) where internal In FY15 we successfully completed our pressure-only conditions are often impossible do diagnose with surface based measurements in thermal explosion cook-off limited techniques. In this project, we focus on violent experiments comparing response between PBX9501 HE reactive conditions because of its complex thermo- and PBX9502. We related reaction violence to pres- dynamic behavior that challenges science models linking sure response in these secondary explosives. Reaction thermal explosion and detonation. The data resulting violence of a thermal explosion is determined by the from from this project will bound critical HE initiation energy release rate of the explosive and the coupling models relevant to Laboratory mission and put into of that energy to the case and surroundings. For the place a baseline diagnostic for pressure and temperature HMX (PBX9501) and TATB (PBX9502) based second- of materials in extremes. ary high explosives studied, we have observed that temperature controls the time to explosion and pres- 404

Page 405

sure controls the final energy release rate subsequent to this summer at the 2015 American Physical Society’s topi- ignition. Pressure measurements in the thermal explosion cal conference on Shock Compression of Condensed Mat- regime have been notoriously difficult to make due to the ter (“Relationship Between Pressure and Reaction Violence extreme rise in temperature which is also occurring during in Thermal Explosions,” Paper # K3.00001). We are follow- a thermal explosion. We utilized several different pressure ing up the presentation of this work with a peer reviewed measurement techniques for several different secondary journal publication. high explosives. These techniques include commercially available piezoelectric and piezoresistive sensors which Future Work we have utilized in the low pressure (sub 30 MPa) range of In FY16, we will complete the our studies of simultane- PBX9502 thermal explosions, and fiber Bragg grating sen- ous temperature and pressure extraction in the thermal sors for the higher pressure range (up to GPa) for PBX9501 response of the HMX based PBX9501 explosive and of and PBX9502 experiments. Simultaneous x-ray radiogra- the TATB based insensitive high explosive PBX9502. After phy measurements of burn velocity were also made, and these studies, the plan is to switch to elongated rate sticks correlations between pressure, burn velocity, and reaction of high explosives (PBX9501 and PBX 9502) where initial violence studied. burn is followed by the sub-sonic burn regime of deflagra- tion before crossover to full detonation. Recording of Design modifications to our pressure-only fiber Bragg deflagration pressure and temperature histories will be instrumentation were completed to accommodate the ex- attempted with pressures expected to eventually exceed traction of pressure and temperature independently with a the maximum range available with fused silica based fiber single fiber Bragg grating sensor. The modifications includ- optic Bragg sensors when reaching full detonation pres- ed the switch over of the instrument from a design that sures. Nonetheless, the build-up of pressure and tempera- was optically insensitive to fiber light polarization to one ture in the deflagration regime is of valuable interest in that is polarization dependent. This required all optical understanding the crossover from low velocity convective components in the system to be exchanged for polarization burn to pressures at high sub-sonic speeds. based encoded illumination and detection, including the fiber Bragg grating sensor. By using polarization encoded Conclusion pulses to interrogate the sensor, the temperature-pressure The project will yield a field tested rack-mounted portable extraction method relies on the polarization mode disper- diagnostic detection system for simultaneous dynamic sion properties of fiber (i.e, both polarizations propagate pressure and temperature sensing for explosive burn, thru the fiber system but because of the difference in blast waves, and detonation physics. This system will be the speed of light between each mode, a difference in a first-of-a-kind embedded probe for measuring dynamic response of the fiber Bragg grating sensor yields a differ- pressure and temperature for a variety of extreme en- ence in time-delay between modes). We use polarization vironments including shocked samples, deflagration-to- mode dispersion to simultaneously detect the pressure detonation transition and high explosive cookoff, and other and temperature while monitoring the polarization mode experiments. We expect a diagnostic that will be capable dispersion time-delay to yield a set of two measurements of measuring dynamic events (nanosecond time scale) of that can only be satisfied by a unique combination of pressure up to 100 kilobar and temperature up to 1000 pressure and temperature. Our first experimental tests of degrees Celsius. these temperature-pressure based HE thermal explosion measurements are to be made in late summer of 2015. Publications Rodriguez, G., M. Jaime, C. H. Mielke, F. F. Balakirev, A. Results from our fiber Bragg grating based sensor develop- Azad, R. L. Sandberg, B. Marshall, B. M. La Lone, B. F. ment and instrumentation for high-speed pressure diag- Henson, L. Smilowitz, M. Marr-Lyon, and T. Sandoval. nostics were published in the the peer-reviewed journal Insight into fiber Bragg sensor response at 100 MHz Optics Express (“Coherent Pulse Interrogation System for interrogation rates under various dynamic loading Fiber Bragg Grating Sensing of Strain and pressure in Dy- conditions. 2015. In 2015 SPIE International Society for namic Extremes of Materials, Opt. Express Vol. 23, 014219 Optics and Photonics Defense, Security, and Sensing (2015)), and a invited SPIE conference proceedings paper Conference: Fiber Optic Sensor and Application XII. was also published (“Insight into Fiber Bragg Sensor re- (Baltimore, 20-24 Apr. 2015). Vol. 9480, p. 948004. Bellingham: SPIE. sponse at 100 MHZ Interrogation Rates Under Various Dy- namic Loading Conditions,” SPIE Vol. 9480, 948004 (2015)). Rodriguez, G., M. Jaime, F. F. Balakirev, C. H. Mielke, A. The results showing details of the pressure response differ- Azad, B. Marshall, B. M. Lalone, B. H. Henson, and L. ences between PBX9501 and PBX9502 were also presented Smilowitz. Coherent pulse interrogation system for 405

Page 406

fiber Bragg grating sensing of strain and pressure in dynamic extremes of materials. 2015. Optics Express. 23 (11): 14219. Smilowitz, L., B. F. Henson, G. Rodriguez, D. Remelius, E. Baca, D. Oschwald, and N. Suvorova. Relationship between pressure and reaction violence in thermal explosions. To appear in Journal of Physics Conference Series. Smilowitz, L., B. Henson, G. Rodriguez, R. Sandberg, M. Holmes, A. Novak, E. Baca, and D. Oschwald. Following reaction progress from thermal decomposition to ignition and internal burning. To appear in Proceedings of the 15th International Detonation Symposium. (San Francisco, 13-18 Jul. 2014). 406

Page 407

Materials for the Future Exploratory Research Continuing Project Thin-Film Heat Switch for Active Thermal Management of CubeSat Payloads Alexander H. Mueller 20150375ER Introduction and communication missions in support of nuclear Satellites are exposed to a complex thermal environ- non-proliferation, and intelligence missions. Future ap- ment: external heat loads change rapidly between solar plication in waste heat capture and thermal regulation exposure and eclipse (Fig. 1a), and internal heat loads of buildings extends the relevance of this research to en- are dictated by the operation of electronic payloads1. ergy security and environmental stewardship missions. While heat is most efficiently dissipated from the satel- Progress lite during eclipsed (“cold”) periods, payload operation may follow different patterns as dictated by the mission. This LDRD project is developing an active thermal The resulting thermal management problems are mag- transfer technology to maintain optimal temperatures of nified in small form factor CubeSats that have limited cubesat electronic components. The initial component space to move, store, and release thermal energy. for demonstration of the integrated device was to be a power amplifier, however the research team decided The goal of this project is to fabricate the next genera- that maintaining a Li-ion battery pack at optimal charg- tion heat switch and demonstrate its performance in a ing and discharging temperatures would be a more satellite environment. Starting from our existing device general, less mission-specific application whose dem- and modeling capabilities, we will scale up the heat onstration would have higher impact in the community. switch area and improve the thermal conduction ratio This required an increase in the active area of the device between the “on” and “off” states. The new heat switch to 7.5 sq.in., and a more complex geometry involving a will then be integrated with a CubeSat thermal manage- phase change material (PCM) for thermal storage that ment system to maintain a nbatteryu pack at optimal would be sandwiched between two heat switches ca- operating temperature using the batteries own waste pable of controlling the flow of heat into and out of the heat. If successful, the heat switch will be tested to PCM. General Environmental Verification Standards (GEVS) in anticipation of a possible CubeSat launch scheduled for The team successfully demonstrated the scale up (by 50x late 2016. from previous devices) of the heat switch device and the fabrication of the PCM layer in time for the first mid-year Development of this active and compact thermal man- review of the project. The large area heat switch was agement system will allow for a denser packing of fabricated from photolithographically pattered Ti/Au payloads on cubesats, and after proven performance, electrode on sapphire wafer a planar copper electrode full size satellites. After development and production, and 3M Novec 7500 dielectric fluid in between. The this technology can be applied to terrestrial applications thermal transport properties of the heat switch were such as industrial waste heat management and concen- demonstrated and shown to be capable of a rapidly tration, thermal regulation of buildings, and electronic changing its thermal conduction coefficient in a tunable component cooling in consumer electronics. manner for differing heat loads. The thermal storage layer between the two heat switches was assembled by Benefit to National Security Missions melt casting a commercial PCM into a mold formed by Success in this project will advance the utility of cubesats the planar copper electrodes. Several thermistors were and have future applications in the thermal manage- embedded into the PCM for measuring and controlling ment of full size satellites. These platforms are important the heat flows during operation. The PCM was shown in our nation’s remote sensing, information collecting, to be able to store 2 kJ of heat in a 35 ml volume while 407

Page 408

maintaining an overall temperature near 25°C. The mate- ment system capable of moving 150W / 100cm^2 of waste rial accomplishes this by undergoing a solid-solid phase heat into a thermal storage medium. If successful, the sys- transition that absorbs thermal energy upon heating above tem will undergo NASA standard reliability testing and be 25 oC and will subsequently release this heat upon cooling deployed on a cubesat scheduled for future launch. Devel- below 12 oC due to the reverse phase transition. The de- opment of such a system will allow for increased cubesat vice will use the stored heat in anticipation of charging the payload densities and higher operational duty cycles, and batteries as soon as the satellite moves out of the eclipse thus increase the utility of the cubesat’s sensing and com- part of its orbit. To do so, the heat switch is activated to munication hardware. moving heat from the PCM to the battery pack so that the battery temperature is optimal for rapidly accepting charge from the solar panels as soon as they are illuminat- ed. This new mode of operation makes better use of the existing power and enables advanced payload capabilities. Currently a test frame for the integrated PCM / heat switch device with the battery pack is being fabricated. This first prototype will be tested in an environmental chamber that is capable of simulating the thermal environment that the assembly will be exposed to in space. The team is await- ing delivery of the battery charger in order to simulate the entire coordinated cycle of charging and discharging the batteries during orbit. The cycle consists of capturing the heat generated during discharge of the batteries in the illuminated (hot) part of the orbit in the PCM using the heat switch to mediate the heat transfer. This heat will be stored in the PCM during the eclipse (cold) cycle. Once the satellite moves towards the end its the eclipsed posi- tion, the stored heat will be used to increase the batteries temperature in anticipation of charging immediately upon leaving eclipse as the solar panels are illuminated. We anticipate testing the first prototype in the environmental chamber at end FY15 or early FY16. Future Work Project goals for FY 16 are to integrate the heat switch, phase change material (thermal storage) and battery pack into one monolithic device. This device will be tested in an environmental chamber and subjected to GEVS testing. Upon completion of testing, a revised device will be as- sembled into a cubesat housing. Project goals for FY15 will be to fabricate a large area (25 cm2) electrohydrodynamic (EHD) heat switch and evaluate its performance. Once the operational parameters and performance are determined we will refine the design in order to reach the 150 W/ 100cm2 heat transfer rate nec- essary for integration with a cubesat battery pack. This will involve iterative micro-fabrication of electrode geometries and investigation of stacked serial and parallel configura- tions. Conclusion We expect to develop an active, compact thermal manage- 408

Page 409

Materials for the Future Exploratory Research Continuing Project Sub-Grid Meso-Scale Model for Twinning and Slip Processes Curt A. Bronkhorst 20150431ER Introduction notional representation of complex physical processes Plastic deformation in metallic materials occurs by two leading to tenuous quantitative accuracy. In addition, physical processes - dislocation motion (slip) and me- this highly phenomenological material modeling strat- chanical twinning. Representing plasticity by slip pro- egy leads to numerical difficulties under certain loading cesses is computationally more straightforward since conditions. At present, these models completely ignore dislocations are very small (comparable to atomic length the physical process of twinning on the shock hardening scales) relative to most computational cells used in prob- of metallic materials and also the impact that twinning lems requiring models for plastic deformation. This is has upon the structural evolution of materials and its true even when modeling single crystals explicitly. Me- impact upon their damage and failure response. This chanical twinning is a mechanism which physically spans project will capitalize on LANL’s expertise in plasticity an entire single crystal and forms a plate-like morphol- modeling and begin to develop the meso-scale modeling ogy. The formation of these plate-like structures dur- tools to allow us to begin to quantitatively compute both ing deformation transforms the nature of the material dislocation slip and twinning processes in a way which is microstructure. These plate like structures have been structurally accurate and allow us to quantify the influ- impossible to represent numerically within single crystal ence of both of these processes on the way in which it models within traditional finite element codes since they changes under dynamic loading conditions. These devel- are many times larger than a single computational cell. opments can be used directly in weapons calculations Their representation has been limited to homogenized in the future (selected regions of interest) or used to theories which do not model microstructures or highly motivate the proper high length scale models to enable resolved phase-field theory which are extremely expen- representation of slip and twinning together in a physi- sive computationally. This project offers a new tech- cally accurate way across the entire weapons system. nique to explicitly represent the morphological transfor- Progress mation due to mechanical twinning within polycrystal plasticity models. This will be a transforming develop- The new numerical finite element which includes an ment which will allow us to directly study both slip and embedded region of additional deformation within a twinning processes computationally under general 3D planar region of finite thickness has been completed. loading conditions. If successful, this work will allow us Implementation of this new numerical tool has been to directly study all deformation involving both plastic performed in two-dimensions within a user-element mechanisms and more accurately quantify structural subroutine within ABAQUS. This numerical tool is evolution of materials under dynamic loading conditions. specifically being used in this project to represent the Under these conditions, mechanical twinning is generally explicit deformation mechanism of mechanical twinning more prominent and is believe to contribute substan- which exhibits planar, laminar topology within individual tially to shock hardening in strategic materials. single crystals of metallic and organic crystalline mate- rial. This numerical technique offers the opportunity Benefit to National Security Missions to finally represent this deformation mechanism explic- Much of our present weapon assessment and certifica- itly. The element has been successfully tested using tion strategy relies upon numerical simulation of the an elasto-viscoplastic material model under dynamic shock and dynamic deformation behavior of metallic ma- loading conditions using a macroscale continuum model. terials. At present, this deformation behavior is repre- These are the same loading conditions of interest to this sented by highly phenomenological equations with only project. The new technique has demonstrated adequate 409

Page 410

numerical robustness under these conditions. be performed on the deformed samples to characterize the statistics of twin formation in this material. These Integration of a finite deformation model for the viscoplas- experimentally results will define specific grain boundaries tic deformation of single crystal titanium will begin in the of interest which can be explore computationally with the fourth quarter of this year. Currently there is a difference above-mentioned model and embedded element. in the strain measures used within the single crystal model and the embedded region finite element under finite strain Future Work conditions. Changes in the measure of strain and the The anticipated tasks and accomplishment goals for this variables used to represent the kinematics will be changed project in fiscal year 2016 are as follows: within the embedded region finite element to be consis- tent with that used within the single crystal model to be • Develop the sub-grid twinning model within the frame- based upon the deformation gradient experienced locally work of the existing single crystal model for slip based within the element. This will facilitate the merging of the plastic processes. Consistency of strain and kinematic two currently separate subroutines into a single user ele- quantities between the sub-grid and crystal model will ment subroutine containing the model of the single crystal. be developed. This step will be completed by mid-sec- ond year with implementation to follow. Implementation of a model for twin nucleation and growth within the single crystal model code is underway. When • Develop and implement the analytical nucleation completed, the single crystal model will be capable of criteria for the initiation of twin lamellae within the representing thermally activated based dislocation slip ABAQUS code. We anticipate the development of the process as well as the nucleation and growth of mechani- nucleation model to be completed in the first year and cal twinning based upon the power law representation its implementation in the second year together with of the rate of twin growth. The twin nucleation model is twin growth. based upon the mechanism of partial dislocation accumu- lation at a grain boundary providing the driving force. Our • Develop the twin growth model for expansion of the first mechanical problem to exercise the above-mentioned laminate twin(s) with deformation. This model will be tools will be a simple two-dimensional periodic bi-crystal completed early in year two and will be coupled with system allowing for twin nucleation and growth within a the nucleation model and sub-grid code. Bi-crystal cal- single row of computational cells. Differing orientation culations will begin in year two to test numerical stabil- relationship between the two crystals will allow for prob- ity and begin to test the nucleation and growth criteria ing of the driving force for twin propagation. Simple shear for growth of twins under simple shear conditions. loading conditions will be applied to this simple boundary • Demonstrate the feasibility of the newly developed value problem to properly exercise the new numerical tool model by performing polycrystal simulations of dy- for robustness. In addition, we will also probe the physics namic experiments conducted on high purity Ti and of nucleation and growth of a single twin within this simple comparing results to the new experiments and quan- system. Bi-crystal combinations of interest as defined by titative metallographic information. A test plan has the experiments will also be easily examined. been developed and will be executed throughout the High purity polycrystalline titanium has been acquired for duration of this project. We will draw upon these this project and a procedure for combined cold-rolling experimental results in the simulations we choose to and heat-treatment to achieve a target equiaxed grain focus upon. size of 50 microns has been developed. This is a grain size Conclusion which is larger than traditionally achieved with high purity titanium to facilitate detailed examination of individual This project will develop the numerical element to repre- twin regions. A matrix of mechanical experiments has sent the morphological deformation of mechanical twin- been defined for the 50 micron grain size titanium. The ning in single crystal metallic materials. The kinetic and variables in the experimental component of this project kinematic theory to represent the mechanism of twin- include strain rate (0.1/s, 1.0/s, 100/s, and 2000/s), strain ning will be developed and taken from existing work. The (total strains of 0.01, 0.02, and 0.05), and the three princi- theory will be coupled with the developments of the new pal plate directions. All experiments will be conducted at numerical element. This development will be coupled with an initial temperature of 20 C. Execution of these experi- existing theory for dislocation slip processes. This work ments is underway. Once completed, in addition to the will be coupled with experimental data for HCP materials. mechanical behavior data which will be used by the single This is exciting work and is anticipated to produce sev- crystal model, metallographic EBSD characterization will eral publications. This will also allow us to institutionally 410

Page 411

broaden our computational material science capability. Publications Mourad, H. M., and C. A. Bronkhorst. Finite Element Simulations of Dynamic Shear Localization in Elasto- Viscroplastic Solids under Adiabatic Conditions. Presented at U. S. National Congress of Computational Mechanics. (San Diego, 26-30 July, 2015). 411

Page 412

Materials for the Future Exploratory Research Continuing Project Higher Order Spin Noise Spectroscopy: From Foundation of Quantum Mechanics to Applications Nikolai Sinitsyn 20150504ER Introduction not exist. The proposal will advance new Laboratory Complete information about an interacting system is capabilities that underpin a wide variety of Labora- contained in the full set of correlators of its variables. tory missions including Nanotechnology and Quantum However, one of the central results in statistical phys- Information Science. Demonstration of new effects ics, called the fluctuation-dissipation theorem tells us by measurements of higher order correlators in spin that information provided by standard experimental systems will generate worldwide attention and will form measurement tools (e.g. conductivity, susceptibilities, or a robust platform for a mesoscopic spintronics program pump-probe experiments) is intrinsically limited: Acces- at LANL. The present project will lay the groundwork for sible linear response characteristics are equivalent to future proposals for external funding to address national measurements of only 2nd order correlators taken at problems, including energy efficient applications (DOE). thermodynamic equilibrium. Our main idea is that Spin The team will work with A.Taylor to develop a BES Mate- Noise Spectroscopy, an optical technique pioneered at rials Science funded project. LANL, opens a unique opportunity to escape from the Progress constraints of the fluctuation-dissipation theorem and measure higher-order correlators, which we will use to: Project is going well and on track. The team has already published one combined experimental-theoretical paper • characterize decoherence mechanisms in solid state in Scientific Reports about the invention of the “Cross- qubits made of InGaAs quantum dots, Correlation Spectroscopy” - the technique that allows • reveal interactions and disorder characteristics of measurements of new types of spin correlators in con- conducting electrons in semiconductors, densed matter systems. • explore spins entanglement in atomic gases and test Several work are currently either submitted or at the quantum mechanics at new scales. final stage of preparation. This includes: • PI (Nikolai Sinitsyn) with a CNLS postdoc - Dr. Fux- Studies of higher-order spin correlations will constitute iang Li have collaborated with the group of experi- a significant advance in the field of quantum measure- mentalists in Germany, lead by Prof. Jonathan Finley ment science. We will achieve the most precise under- at Technische Universitat, Munchen. We already standing of the physics of a nuclear spin bath, determine prepared a LA-UR internal LANL publication that whether macroscopic systems of ~109 interacting spins reported the discovery of three-stage decoherence can still exhibit quantum properties that cannot be mechanisms of a solid state qubits. Article is cur- found in classical physics, and obtain qualitatively new rently submitted to Nature Physics. information about many-body quantum electron interac- • All members of the project have performed re- tions. search (both measurements and theory) on the Benefit to National Security Missions 4th order spin correlations for hot atomic vapors in external noisy magnetic fields. We are at the stage of This project will build capabilities in the novel measure- working on the final experiment that would generate ment methods that enable new scientific discovery at figures of good quality for publication. LANL. It is based on recent technological advancements that use the large data storage and manipulation capa- • PI and Fuxiang Li have been working on a very com- bilities. Until very recently this technology simply did plex fundamental theory of higher order spin corre- 412

Page 413

lators of conducting electrons and constraints imposed Publications by higher order fluctuation relations on such correla- Bechtold, A., D. Rauch, F. Li, T. Simmet, P. Ardelt, A. Regler, tions. We already obtain main results and are working K. Müller, N. A. Sinitsyn, and J. Finley. Three-stage currently on the final version of the article. decoherence dynamics of an electron spin qubit in an optically active quantum dot. 2015. Nature Physics. : x. • Dr. Scott Crooker (CI) has performed test measure- ments of the 4th order correlator of spins in hole Li, F., and N. A. Sinitsyn . Universality in higher order spin doped quantum dots in GaAs. Despite observing a noise spectroscopy. Physical Review Letters. complex signal, our analysis revealed that it is strongly affected by nonlinearities in the detector, so that alter- Rice, W., T. Baker, W. Liu, N. A. Sinitsyn, V. A. Klimov, and native approaches are needed. PI and Fuxiang Li have S. A. Crooker . Revealing Giant Internal Magnetic proposed a different type of correlator that is expected Fields due to Spin Fluctuations in Magnetically- Doped Colloidal Nanocrystals. To appear in Nature to be free of the problems observed by experimental- Nanotechnology. ists. However, this requires the change in the setup, which will take time (about 1/2 year) to make but it is Roy, D., L. Yang, S. A. Crooker , and N. A. Sinitsyn. Cross- totally within our capabilities. correlation spin noise spectroscopy of heterogeneous interacting spin systems. 2015. Scientific Reports. 5: 9573. Future Work Sinitsyn, N. A.. Exact transition probabilities in a 6-state Experimental studies will focus on measuring third-order Landau-Zener system with path interference. 2015. correlator of the solid state spin qubit realized in InGaAs Journal of Physics A: Mathematical and Theoretical. 48: quantum dots. These spins are entangled with a dense 195305. bath of ~105 nuclear spins. Measurements of 3rd order correlators should provide considerable new insight in Sinitsyn, N. A.. Solvable four-state Landau-Zener model of physics of this interaction. They will allow us to use our two interacting qubits with path interference. 2015. the central spin qubit as a nanoscale probe of local nuclear Physical Review B. 92: 205431. spin dynamics. Recently, we discovered that these correla- tors can be used to probe fundamental questions of quan- tum mechanics. Our team already obtained preliminary experimental results in this direction. Theoretically, we will focus on higher order correlators of conduction electrons. We will explore the universal- ity of their behavior due to thermodynamic constraints described by the higher order fluctuation dissipation relations. We will identify a small number of independent parameters that control the shape of higher order corre- lators and make predictions for their behavior in specific systems. Conclusion The goal of this project is to demonstrate the new mate- rial characterization method and use it to explore essen- tially new physical phenomena, previously unreachable by conventional means including some of the most funda- mental problems in science such as the emergence of the macroscopic classical realism from microscopic quantum mechanics. We will show that considerable stream of information pro- vided by the spin noise signal is sufficient to precisely and non-invasively determine higher-than-2nd order correla- tors of variables even in mesoscopic quantum systems at thermodynamic equilibrium. 413

Page 414

Materials for the Future Exploratory Research Continuing Project Three-Dimensional Porous Nanographene for Highly Efficient Energy Storage Hsing-Lin Wang 20150532ER Introduction storage applications. Graphene, due to its unique chemical and physical prop- Benefit to National Security Missions erties, has emerged as a new energy storage material for lithium ion battery anodes. However, over 50% of the The proposed research enables a better understanding graphene charge capacity is lost during charge-discharge of fundamental mechanisms for Li behavior in porous cycles due mainly to restacking of graphene sheets into graphene structures and provides a viable approach for graphite-like structures. A lack of ideal model graphene fabricating functional nanomaterials for advanced bat- systems that can be easily incorporated into Li battery tery technologies with high-power and high-rate perfor- anodes has hindered scientific investigation of Li reac- mance. The project addresses fundamental challenges tion mechanisms and prevented development of robust in materials, energy, and nanotechnology. Success will graphene-based anode materials. have wide-spread impact on the fabrication of high- performance rechargeable batteries for efficient energy Using a novel synthetic method developing at C-PCS/ storage. This research will strengthen our capability in MPA-11, we are able to synthesize a series of nitrogen- addressing key LANL missions of materials functionality doped nanographene structures that allow for finely- and energy security, positioning LANL in a new frontier tuned nitrogen doping and molecular structures and of energy storage research. With this proposed LDRD- sizes. In this proposal, using these ideal nanographene ER effort, understanding the mechanisms that underpin model systems, we aim to elucidate Li ion adsorption/ battery reactions and developing high-performance desorption/diffusion kinetics as functions of nitrogen electrode materials for energy storage will support new dopant, molecular size, and structures of the synthesized fundamental and applied follow-on projects in the Basic nanographene. Based on the acquired knowledge, the Energy Science (BES) and Energy Efficiency and Renew- 3D porous nanographene anodes with desired chemical able Energy (EERE) offices, respectively. and structural properties will be further prepared via a crosslinking synthetic strategy in order to maximize the Progress specific Li storage capacity, diffusivity, and long-term One of our main goals is to develop Such synthesis is charge/discharge stability. regarded as highly challenging as there is no guarantee success of such molecules. In the fiscal year 2015, we In order to achieve these goals, we propose an innova- have made significant progress in several fronts. First, we tive R&D approach by integrating theoretical predictions have demonstrated synthesis of two dimensional (2D) from density functional theory calculations and na- and three dimensional (3D) nanographene molecules noscale dynamic simulations with experimental char- through multi-step organic synthesis. Complete charac- acterization using well-defined nanographene model terization of such molecules through a suite of organic systems. In turn, nanographene with optimally designed spectroscopy probes and the results are in perfect agree- electronic and geometric structures will be realized ment with the hypothesized structure. though molecularly controlled synthetic methods. By combining the multidisciplinary expertise, we expect to Second, the electronic, optical and redox properties of understand fundamental reaction mechanisms of lithium these molecules have been fully characterized. These on the doped nanographene structures and ultimately, molecules reveal structure and redox potentials that are to propose a path forward to designing structurally functional group-dependent. Although such results are stable and high-capacity graphene anodes for energy somewhat expected, once fabricated into lithium ion 414

Page 415

battery, varying functional group have a huge impact in the performance in terms of the specific capacitance and stability. Third, Lithium battery from these nanographene anode have been prepared and the highest capacity is ~950-1000 mAh/g, the highest number ever reported for carbon based anode. We collaborate with Professor gang Wu 9A former scientist at MPA-11) at University Buffalo and Lim- ing Dai at Case Western Reserve who is the world expert in energy conversion and storage devices. Some of the results have been reproduced by more than one group. These results have been written and submitted for publica- tions. One of the manuscripts was sent to “Science” and currently evaluated by the Advisory board. Future Work In FY16, we will (1) realize atomic level control over the chemical synthesis of the nanographenes which include two dimensional (2D) and three dimensional (3D) nanog- raphene structures. We strive to show chemical synthesis of nanographenes and demonstrate how these nanog- raphenes self-assemble to form hierarchical structures, which is critical to the performance of lithium ion batteries with respect to their charge capacity and stability, and (2) optimize the electronic and geometric properties of nanographene for efficient Li ion adsorption and diffusion determined by electrochemical measurement. Batter- ies tests will provide the correlation between functional group, geometry and redox properties. Such correlation has never been realized and the demonstration of such correlation will represent a major advancement in rational design of anode materials for energy storage devices. Conclusion We expect to: (1) Develop new synthetic protocols to prepare a series of nanographenes with various functional groups (linker) and nitrogen doping to finely tune their physical and electronic properties. (2) Gain fundamental understanding of lithium adsorption/desorption and dif- fusion kinetics on doped graphene by using well-defined (e.g., molecular size, structure, and doping content and lo- cation) nanographene model systems. (3) Based on insight provided from experimental results and DFT predictions, design and synthesize novel 3D porous nanographene anodes with optimal chemical and structural properties to maximize Li capacity, diffusion rate, and cyclic stability for energy storage applications. 415

Page 416

Materials for the Future Exploratory Research Continuing Project Controlled Helium Release from Composite Plasma Facing Materials through Interface Design Yongqiang Wang 20150567ER Introduction Benefit to National Security Missions The overall project goal is to demonstrate a tungsten (W) This research is important to LANL’s core materials based plasma-facing material that continually outgasses capabilities: discovering, understanding,and exploiting helium (He) as it is being implanted, thereby preventing defects and interface in materials. To achieve the vision He assisted cavity growth and yielding a morphologically of prediction and control of materials functionality set stable plasma-facing surface. This demonstration will be by MaRIE vision, it is noted that “a key grand challenge accomplished by designing and testing a tungsten-metal is the ability to predictively manipulate microstructures (W-M) nano-composite containing interfaces that pro- to achieve desired macroscopic performance. Central vide stable pathways for controlled and continuous He to this challenge is the potency of defects, either to be outgassing. The interfaces will be designed such that the exploited intentionally for enhanced performance or pathways self-organize under He implantation. to suffer their deleterious effects.” Our proposed re- search directly addresses this vision and will strengthen This project is built on recent research by our team into LANL’s competency in the design of functional materials the interaction of He with solid-state interfaces in bcc- through exploiting defects and interfaces. fcc multilayer composites, in which we discovered that that certain metal-metal interfaces are capable of stably A recent report by the DOE Office of Fusion Energy storing up to several atomic % of He without forming He Sciences’ Advisory Committee emphasized that “the ma- bubbles: far in excess of the bulk He solubility limit. The jority of [plasma facing component] materials research reason for this surprising lack bubbles as compared to should be oriented towards tungsten” (p. xvi) with the the bulk counterparts is that implanted He aggregates specific goal of identifying and characterizing W-based into platelets that wet high-energy parts of the inter- materials suitable for fusion-relevant environments. The faces of interest and do not grow into voids even in the proposed project directly addresses the needs in this presence of radiation-induced vacancy supersaturations. report. We plan to use this insight to design W-M interfaces with We plan to design W-M interfaces with high-energy high-energy regions patterned into continuous path- regions patterned into continuous pathways (nanochan- ways (nanochannels) that will first trap (or store) He and nels) that will first trap (or store) He and then allow it to then allow it to outgas while maintaining morphological outgas while maintaining morphological flatness, me- flatness, mechanical cohesion and thermal conductivity chanical cohesion and thermal conductivity across the across the interface. Successful demonstration of this interface. Successful demonstration of this novel inter- novel interface design concept will enable breakthrough face design concept will enable breakthrough improve- improvements in the performance of W based plasma- ments in the performance of W based plasma-facing facing materials under fusion-relevant conditions, and materials under fusion-relevant conditions, and the the insights gained will become a broader “toolkit” of insights gained will become a broader “toolkit” of mate- materials science and engineering methods that may rials science and engineering methods that may later be later be used in other fusion and non-fusion related used in other fusion and non-fusion related applications applications where precipitation of impurities play an where precipitation of impurities play an important role. important role. 416

Page 417

Progress function of wetting angle and spacing of misfit dislocation Experimental efforts: the key accomplishment is to synthe- intersections. size nanostructured thin films with high quality Cu-V inter- We have presented our modeling work at the 2014 MRS faces and to capture the interface dislocation network by Fall Meeting (contributed talk) and at the 15th Internation- transmission electron microscope. To capture dislocation al Conference on Plasma Facing Materials and Components network at Cu-V interfaces is one of the grand challenges held at Aix en Provence, France in May 2015 (poster). The in this project. Here we hypothesize that the diffusivity of experimental results have been accepted to present at the helium along dislocation line will be much faster than that 17th International Conference on Fusion Reactor Materials in bulk and the dislocation network at interfaces has been to be held at Eurogress Aachen, Germany in October 2015 treated as the channel to effectively diffuse out the extra (poster). By invitation, we gave a presentation and submit- helium in materials. To form effective channels for helium ted a white paper on this subject at a recent Plasma-Mate- diffusion, the interface structure needs to be atomic sharp rials Interactions Workshop organized by Princeton Plasma between Cu and V and also needs large grain size in both Physics Laboratory for Office of Fusion Energy Sciences in Cu and V layers. In order to synthesize high quality thin May, 2015. films, we have performed many batches of depositions with different sputtering rate, temperature, film thickness, Future Work and substrate. The grain size has been improved from 20 The goal of the second year is to demonstrate through nm at very beginning of the project to 200 nm, an increase modeling, simulation and experiment that W-Cu nanocom- of one order of magnitude. The large grain size provides us posites possess nanochannel structures for desired helium the possibility to see the structure of dislocation network storage and controlled helium release. Experimentally, under transmission electron microscope (TEM). We think we will first synthesize theory-predicted W-Cu interfaces we have found the recipe to synthesize high quality Cu/V (bi-layer, tri-layer, and multi-layer) using physical vapor bilayer, tri-layer and multi-layer thin films. Helium implan- deposition methods; then perform He ion implantation tation has been arranged in this month along with the in these structures using various beam energies, ion flu- following up TEM characterization to observe the nano- ences, beam fluxes, and temperatures; and finally perform channel structure. We believe within the first 12 months detailed characterizations of the as-deposited and ion of this project, we will figure out how helium atoms diffuse implanted samples for nanochannel distribution, He- out of the Cu/V thin films with presence of the dislocation bubble formation, He retention and release, and surface network at Cu/V interface. morphology etc. On the modeling side, we will investigate Modeling effort: During the first year, the team has con- the large-scale morphology evolution during continuous structed a phase field model of He precipitate growth at vacancy loading both at isolated interfaces and in multilay- patterned interfaces. This approach uses the Cahn-Hilliard ers. These simulations will be performed both for Cu-V and equation to evolve an order parameter field that describes Cu-W interface structures. Preparation of two publications the He precipitates as well as the metal into which the He is anticipated: one that describes the phase field method precipitates grow (Cu, in the case of Cu-V precipitates). developed during the first year and another that explores The interface is described as a surface with a location the He large-scale precipitate morphology evolution. dependent energy, which gives rise to wetting at speci- Graduate student D. Y. Yuryev will spend the fall semester fied contact angles. The non-wettable part of the interface of 2015 at Los Alamos National Laboratory working with is modeled by a Dirichlet boundary condition that fixes Project PI, Y. Q. Wang, on integration of her simulations the order parameter at the value corresponding to the with experiments being performed at LANL. Her stay will metal. A custom, quasistatic helium loading algorithm was be supported by a DOE Office of Sciences Graduate Fel- developed whereby the net helium content is increased by lowship. We plan to present our work at the 2015 fall MRS changing the order parameter in the helium precipitates. meeting in Boston. After each helium loading step, the precipitate morphol- Conclusion ogy is relaxed by solving the time-dependent Cahn-Hilliard equation until a static state is reached. The process is then The overall technical goals are to demonstrate a tung- repeated as often as needed to simulate the desired extent sten (W) based plasma-facing material that continually of precipitate growth. We use the MOOSE phase field outgasses helium (He) as it is being implanted, thereby code developed at Idaho National Laboratory to run these preventing He assisted cavity growth and yielding a stable simulations. We are currently conducting a parametric plasma-facing surface. This demonstration will be accom- study that investigates the critical helium content needed plished by designing and testing a tungsten-metal (W-M) to cause precipitates to de-wet from the interface as a nano-composite containing interfaces that provide stable 417

Page 418

pathways for controlled and continuous He outgassing. Successful demonstration of this novel concept will en- able breakthrough improvements in the performance of W based plasma-facing materials under fusion-relevant conditions as well as enhancing other applications where precipitation of impurities play an important role. Publications Li, N., M. Demkowicz, N. Mara, Y. Q. Wang, and A. Misra. Hardening due to interfacial He bubbles in nanolayered composites . To appear in MATERIALS RESEARCH LETTERS. Zhang, H. X., F. Ren, Y. Q. Wang, M. Q. Hong, X. H. Xiao, W. J. Qin, and C. Z. Jiang. In situ TEM observation of helium bubble evolution in V/Ag multilayer during annealing. To appear in Journal of Nuclear Materials. 418

Page 419

Materials for the Future Exploratory Research Continuing Project Precision ‘Bottom-Up’ Fabrication of Non-classical Photon Sources Jennifer A. Hollingsworth 20150604ER Introduction room-temperature photon purity, and (3) fast emission. “Quantum” light sources deliver photons in a strictly reg- Specifically, we will embed our so-called ‘giant’ QDs ulated fashion that defies the statistical distribution of (g-QDs) within zinc oxide NWs, where the g-QDs will be photons that result from classical light sources, such as structured to afford both room-temperature spectral lasers, light-emitting diodes and thermal sources. A true purity and single-dot level optical stability (photons-on- single-photon source yields one photon when optically demand), while the NWs will provide efficient subwave- or electrically prompted, while non-quantum sources are length waveguiding. always subject to a Poisson distribution in their photon Benefit to National Security Missions statistics. Single-photon sources are needed as ‘building blocks’ toward next-generation quantum-information With respect to our Science Mission, especially in regard technologies – secure quantum communication (optical to Materials Chemistry, Materials-by-Design, and Materi- quantum bits/‘qubits’), quantum networking (optical als Properties, we will establish new methods for fabri- ‘messengers’), quantum cryptography, quantum compu- cating functional hybrid nanoscale structures that will tation, and quantum sensing (including aiding develop- exhibit designed emergent properties. We will establish ment/qualification of sensors for quantum-information new knowledge/new understanding regarding how to experiments and performance testing of low-light imag- enhance/tune fundamental semiconductor radiative ing systems). rates by field effects (proposed novel quantum-dot/ dielectric metamaterial couple) and/or charge enhance- Nanowire (NW) elements are nearly ideal structures for ment (quantum-dot/diode couple). More practically, assembling photonic circuits due to their extreme shape we will demonstrate the first waveguided optically asymmetry, which can afford optical confinement in and electrically pumped single-photon sources based two dimensions and opportunities for extensive trans- on solution-synthesized quantum emitters, character- port in the long, third dimension, followed by photon ized by an unprecedented combination of: ultra-pure out-coupling at the distal end. Integrating individual, photon statistics, high room-temperature efficiencies, otherwise isolated single quantum emitters within a and “on-demand” response that is not possible using NW cavity affords clear benefits for both systems – the conventional approaches. We also anticipate proving a wire is functionally “activated” by the emitter, while the path-forward to very low-threshold lasing resulting from emitter becomes accessible for a range of excitation our coupling of a novel emitter with a nanowire cavity routes, as well as to photon-information transport and (for strong cavity-emitter coupling). Thus, our Discovery collection. Although significant progress has been made Science aims will enable understanding/controlling col- in the past decade in emitter-in-wire single-photon lective properties of excitonic/dielectric hybrids using sources by epitaxial-growth methods (top-down etching advanced spectroscopies, in-situ measurements & CINT of planar structures containing self-assembled QDs, QDs (user facility). Toward applications relevant to nuclear self-assembled in a NW, and bottom-up nanodisc-in-wire nonproliferation, DoD, NIST, and IC, the work aims to structures), outstanding challenges remain. establish new single-photon source capabilities for next- generation quantum information technologies – secure The overarching goal is to address these challenges by quantum communication, networking, cryptography, combining novel fabrication principles with unique emit- computation and sensing, with low-threshold lasers also ters to create quantum-emitter/NW-waveguide hybrid useful for next-gen solid-state lighting and conventional structures exhibiting: (1) on-demand single photons, (2) 419

Page 420

communications. More specifically — Photon-on-demand impossible to create using traditional lithographic tech- are needed for DoD, DARPA secure communication, as niques. We have also established a quantitative ‘guide’ for well as non-proliferation programs (e.g., next-gen NCAM) - correlating pen-substrate contact time with the number performance testing of low-light imaging systems, develop- of nanoparticles deposited per writing step. Prior to our ment & qualification of sensors for quantum information work, this type of “predictive” capability was believed to (QI) experiments, while entangled photon source will serve be limited to simpler, molecular (rather than nanoparticle) broad needs in QI science & secure communication (QKD & inks. Finally, we demonstrated this level of control for a beyond). range of nanocrystal compositions (red-emitting CdSe/ CdS and IR-emitting InP/CdSe QDs), sizes (~10-50 nm) Progress and surface chemistries (bare g-QDs and silica-coated In the first nine months of this project, we made signifi- g-QDs), thereby addressing another project goal – to have cant progress toward research Objective 1, which is the the flexibility in the fabrication approach to use emitters foundational science on which Objectives 2 and 3 will ‘as-needed’ for proper emitter-waveguide coupling and to build. Specifically, a key aspect of the proposed program modify the emitter structure/chemistry to ensure proper- was to establish a method for coupling quantum emit- ties stability throughout the processing/integration steps. ters with optical waveguide structures. The proposed We have recently submitted this work for publication: method was to use the ‘direct-write’ capability afforded The rapid progress in integration aspects of the project by dip-pen nanolithography (DPN) to literally place from will be complemented in Year 2 by equally rapid progress few to a single giant quantum dot (g-QD) onto the tops in photo-physical characterization of the hybrid QD-ZnO of ZnO nanowire (NW) waveguides, where the ZnO NWs structures and in the synthesis of doped ZnO NWs. are sub-micron in diameter and grown in organized arrays across a solid substrate. Such post-synthesis “nanoscale Future Work manipulation” would be unprecedented for a mask-less In the second year of the project we will transition our technique in creating coupling between two nanoscale success in nanoscale precision integration of ‘Giant’ quan- materials systems over many such systems (as opposed tum dot (g-QD) quantum emitters with silica dielectric to “one-off” approaches). We successfully used DPN to nanodisks by dip-pen nanolithography (DPN) to coupling precisely position g-QDs exclusively on top of dielectric gQDs with in-house synthesized ZnO nanowires (NWs). nanowires. However, for this initial work, we employed sili- This will allow us to establish functional coupling via the con nanodisk antennas as opposed to ZnO nanowires. The proposed “post-synthesis nanomanipulation” strategy. We former were obtained from an external collaborator, while will characterize color and NW-diameter/length-dependent the ZnO structures had to be synthesized by our team. waveguiding and assess single-photon emission perfor- Several months of development were required to realize mance. Tasks to be accomplished include an initial assess- the desired ZnO diameters, aspect ratios and pitch (rod-to- ment of ZnO NW waveguide properties, which will entail rod spacing). We recently achieved this level of control and further g-QD synthesis and ZnO NW synthesis (400-600 are ready to move forward with the initially targeted ZnO nm diameter – sizes initially guided using finite-domain nanowires in year 2, as the DPN approaches developed in time-difference (FDTD) calculations); preparation of simple year 1 using Si nanodisks are directly applicable to the ZnO gQD/ZnO NW couples for study of basic wave-guiding be- structures. havior of ZnO NWs for red (CdSe/CdS gQD: ~650 nm) and Importantly, we established a new DPN strategy to obtain near-infrared (InP/CdS gQD: 1 μm) emission, and confocal accurate nanoscale integration of the nanocrystals with microscopy studies to determine the efficiency of ZnO NW their nano-sized substrates. Namely, we developed a novel waveguiding as a function of NW diameter/length and gQD three-step ‘reading-inking-writing’ approach, where atom- color. Based on this baseline assessment, we will attempt ic force microscopy (AFM) images of the pre-patterned to fabricate a waveguided single-photon source comprising substrate topography are used as ‘maps’ to accurately a gQD embedded in a ZnO-NW, which will entail: synthesis place nanocrystal ink. We also explored two methods for of an array of partially grown ZnO NWs; direct-writing of the ‘inking’ step — dip-coating and scan-coating of the AFM g-QD emitters using DPN onto the tops of the ZnO NWs, ‘writing’ tip. Only the latter was found to afford controlled and completion of NW growth. Emission characteristics nanocrystal deposition (number of particles deposited (stability, linewidth, emission rate, polarization, waveguid- per pillar) and repeatable deposition across the array of ing) of the new hybrid gQD/ZnO NW structures will be de- nanodisks. For the first time, we demonstrated that DPN termined by performing single-nanostructure time-tagged, is indeed a viable technique to fabricate multicomponent time-resolved photoluminescence spectroscopy, including (hybrid) nanostructures that would be challenging or Hanbury Brown and Twiss and Hong-Ou-Mandel experi- 420

Page 421

ments to assess the purity and indistinguishability of the single-photon emission events (employing a new super- conducting NW single-photon detector for experiments in the near-infrared). Toward Objectives 2 (Electrically driven single-photon emission) and 3 (‘Speeding up’ colloidally synthesized emitters), we will begin efforts to fabricate and test stable ZnO p-n junctions. Also, Objective 1 will lay the groundwork for Objectives 2 and 3. Conclusion We will demonstrate the first waveguided optically (Objec- tive 1) and electrically (Objective 2) pumped single-photon sources based on solution-synthesized quantum emitters, where our gQD emitters possess the unique advantages of ultra-pure photon statistics (complete antibunching) and high room-temperature efficiencies compared to epitaxi- ally grown emitters, combined with “on-demand” (non- blinking) response that is not possible using conventional colloidal QDs. We will also provide new understanding toward enhancing radiative rates by field effects (gQD/ dielectric couple) and/or charge enhancement (gQD/diode couple) (Objective 3). Each new property will “emerge” as a direct result of the proposed design for a novel hybrid- material constructs. Publications Dawood, F., P. A. Schulze, C. J. Sheehan, M. R. Buck, A. M. Dennis, N. Karan, I. Staude, J. Dominguez, G. S. Subramania, A. R. James, I. Brener, N. A. Amro, and J. A. Hollingsworth. Precision Placement of Nanocrystals Onto Sub-Micron Three-Dimensional Dielectric Antenna by Dip-Pen Nanolithography. ACS Nano. 421

Page 422

Materials for the Future Exploratory Research Continuing Project Perovskite Solar Cells: The Next Frontier in Energy Harvesting Aditya Mohite 20150612ER Introduction films consisting of large crystal grains. Our unique solu- Recent discovery of incorporating organic–inorganic tion growth process led to reproducible and efficient hybrid perovskites such as CH3NH3PbX3 (X = Cl, Br, I) as hysteresis-free solar cells with peak efficiency approach- donor materials in a simple planar bilayer photovoltaic ing ~18%. We are now among only 3 teams in the world devices has revealed very high power conversion ef- able to fabricate hysteresis-free devices with this ef- ficiency exceeding 15%1-4 without the need for nano- ficiency; the others include H. Snaith’s vapor deposition structuring or creating complex device architectures. In approach (Cambridge, UK) and M. Graetzel’s mesopo- less than eight months, this work published in Nature1, rous architectures (Switzerland). There are several dis- 3 and Science4 has been cited more than 200 times. In tinct advantages to our process over the other systems. spite of the celebrity status for this amazing new materi- Notably, our subsequent preliminary studies suggest that al, there is extremely limited scientific understanding on these high crystalline quality perovskite films are poten- what intrinsic properties are responsible for endowing it tially behaving as classical direct band-gap crystalline with near ideal optical and electrical properties. Building semiconductors like GaAs. Specifically, we observe that on our extensive expertise in the organic photovoltaic like GaAs optically excited electrons and holes occurs via (PV) devices 5, through a combined experimental and broadband absorption and predominately recombine modeling effort, here we will formulate critical design radiatively (so called bimolecular recombination. The PL strategies for high-performing devices based on hybrid quantum yield is near unity, and the PV device ideality perovskites by achieving synthetic control over light factor is close to 1 indicating defect-free behavior. Our absorption properties, conversion of optical-to-electric theoretical DFT calculations [Science] explained these energy, manipulation over charge separation and recom- observation by predicting the presence of blue-shifted bination. electronic states at the interfaces due to excess chlorine (Cl) at the grain-boundaries, providing a natural passiv- Benefit to National Security Missions ation layer that minimizes internal charge carrier losses This proposal focuses on developing perovskites and thus contributing to the high efficiencies. to gain unprecedented power conversion efficiency Further, our initial measurements of perovskite photo- in photovoltaic (solar cell) devices. Perovskites with stability under constant solar illumination reveal a major controlled crystallite domain size, defects, and morphol- breakthrough in obtaining stable perovskite solar cells ogy, coupled with an interfacial layer could redefine with high efficiency. The observed degradation depends photovoltaic research as next generation solar cells with on the experimental conditions. Theoretical model- unprecedented high efficiency. The proposed work will ing attributes photo degradation to the formation of strengthen our capability in addressing laboratory mis- macroscopic charged domains, which is confirmed by sions, particularly in the areas of materials development experiments. Such a joint theory-experimental effort and energy security, and support new program develop- can clarify the observed physical phenomena and sug- ment in Office of Science. gests synthetic strategies to control photo degradation (elaborated below). These examples also highlight on- Progress going integrated efforts between theory, synthesis and Work by our team at LANL that was recently published characterization. These results place our LANL team at a in Science has demonstrated a simple solution-based distinct advantage over other research groups across the approach to grow films of CH3NH3PbIxCl3-x perovskite world and as we stand today, we are perceived as among 422

Page 423

the top 3 research groups in the area of perovskite solar cells. Future Work In the coming fiscal year, we will focus on two aspects: • Ferroelectric effect in perovskite grains: We will un- derstand through experiments and theoretical efforts tif the large grain perovskites demonstrated exhibit ferroelectric effect. We will perform charge polariza- tion measurements in the dark and also study light induced polarization. We will use this information to understand the role of ferroelectricity in photovoltaic devices. We will answer questions like what is the ori- gin of the high open circuit voltage and high current. • We will perform detailed structure function relation- ship between the large grains and its optical proper- ties. We will perform temperature dependent XRD and optical measurements to understand how the struc- tural phases play a role in perovskites and how do they influence the overall macroscopic behavior. Conclusion We expect to achieve an understanding of the intrinsic source of high photocurrent and voltage (open circuit volt- age; Voc) in perovskite-based solar cell devices, as well as how the emergent ferroelectric properties of these unique materials can be tuned to surpass current device efficien- cies and develop high efficiency low cost PV devices. Publications Nie, , Tsai, Asadpour, Blancon, A. J. Neukirch, Gupta, J. J. Crochet, Chhowalla, Tretiak, M. A. Alam, Wang, and A. D. Mohite. High-efficiency solution-processed perovskite solar cells with millimeter-scale grains. 2015. SCIENCE. 347 (6221): 522. Tsai, , Nie, Cheruku, N. H. Mack, Xu, Gupta, A. D. Mohite, and Wang. Optimizing Composition and Morphology for Large-Grain Perovskite Solar Cells via Chemical Control. 2015. CHEMISTRY OF MATERIALS. 27 (16): 5570. 423

Page 424

Materials for the Future Exploratory Research Continuing Project Defect-Induced Emergent Magnetism in (Nonmagnetic) Complex Oxides and their Interfaces Scott A. Crooker 20150613ER Introduction Time-Resolved Faraday Rotation. Controlled introduction and removal of VO defects via annealing protocols is a Interest in oxide-based electronics has exploded in re- key component of this program. cent years, fueled by the ability to grow atomically-pre- cise interfaces and heterostructures of various complex Benefit to National Security Missions oxides. Strontium titanate (SrTiO3) — a nonmagnetic wide-bandgap insulator — is a foundational material in This research is vital to LANL’s core materials capabilities: this field. Owing to its widespread use in materials sci- Discovering, understanding, and exploiting defects and ence, its dielectric and lattice properties are well known. interfaces in materials. It not only strengthens LANL’s However, tantalizing evidence in recent years indicates existing competency, but also significantly expands that there may yet be much more to SrTiO3 than previ- capabilities for our future nanostructured materials ously realized: Several groups around the world have via “external” control of interfaces and defects. Fur- recently reported unexpected emergent phenomena thermore, the MaRIE Workshops report indicates that at these oxide interfaces, most strikingly magnetism “interfaces and defects” must be addressed to solve the and superconductivity. This is particularly remarkable decadal challenges for accelerating materials discovery. because the constituent oxides (e.g., SrTiO3 and LaAlO3) To achieve the vision of prediction and control of ma- are nominally nonmagnetic and insulating. Crucially, the terials functionality, it states “a key grand challenge is formation and distribution of oxygen vacancies defects the ability to predictively manipulate microstructures (VO) – which form readily in SrTiO3 and are mobile – are to achieve desired macroscopic performance. Central widely suspected to play an essential but as-yet-poorly to this challenge is the potency of defects, either to be understood role in these emergent interfacial phenom- exploited intentionally for enhanced performance or ena. to suffer their deleterious effects. The role of defects is a microcosm of the broader impact of rare events and Our team recently applied powerful magneto-optical fluctuations that are a key stumbling block in achieving techniques to bulk SrTiO3 crystals for the first time and prediction and control.” Our project directly addresses discovered that, surprisingly, VO defects can generate an this vision. The ability to control defects will yield new optically-induced and persistent magnetism in this nomi- and unique results since the materials we are going to nally nonmagnetic material. However, the underlying explore will provide us the opportunity to obtain new physical mechanisms remain unknown. Nor is it known and/or improved functionalities not obtainable through how deeply these findings inform on the interfacial mag- bulk materials or by simply changing material chemistry. netism discussed above. Therefore, motivated by these Moreover, this research represents an opportunity for results and coupled with our state-of-the-art capabili- LANL to take the global lead in the fundamental under- ties for pulsed-laser deposition (PLD) oxide growth at standing of oxide materials. LDRD’s investment in such LANL, we propose to reveal the origin of (and ultimately an effort will ensure continued leadership in nanotech- engineer) the defect-related emergent magnetic proper- nology and functional materials, which will enable LANL ties in these bulk and interfacial complex oxides, focus- to respond effectively to future calls by BES and other ing on the archetype SrTiO3. We will apply a powerful funding agencies. suite of magneto-optical techniques to SrTiO3 epilayers and SrTiO3/LaAlO3 interfaces grown by PLD and MBE Progress techniques, including Optically-Coupled SQUID Magne- Over the past year, postdocs Bill Rice and Luyi Yang have tometry and also ultrafast magnetization dynamics via made substantial progress on various aspects of this 424

Page 425

project. Bill has established a fully-functional capability Controlled introduction and removal of VO via annealing for Time Resolved Faraday Rotation, which is an ultrafast in UHV and oxygen environments. This essential capabil- optical technique for measuring the picosecond-resolved ity currently exists in U. Minnesota but we will emulate dynamics of spin and magnetization in semiconductor (and and improve it here at CINT to ensure quick turnaround of metallic) materials. This new LANL capability employs an samples and rapid convergence on ideal annealing recipes. ultrafast Ti:sapphire laser coupled to an optical parametric oscillator, which can deliver 100fs optical pulses at wave- Conclusion lengths spanning the optical spectrum. This new capability The overall technical goal of this project is to study and has been tested and benchmarked in two semiconductor ultimately reveal the origin of the recently-discovered systems: 1) magnetically-doped semiconductor nanocrys- magnetism and magnetic effects that emerge in many tals, and 2) new atomically-thin MoS2 semiconductors. (nominally nonmagnetic) oxide semiconductors. We will Both material systems have yielded new and exciting focus on strontium titanate, which is the archetypical experimental results. In the first case, we have a nearly- and foundational material in this new burgeoning field of completed draft of a paper written up, entitled “Revealing ‘complex oxide electronics’. Our plan is to directly compare Giant Internal Magnetic Fields due to Spin Fluctuations in magnetism and magneto-optical phenomena in SrTiO3 Magnetically-doped Colloidal Nanocrystals”, that we intend grown by bulk (commercial) methods, by pulsed laser to submit to Nature Nanotechnology. In the second case, deposition (PLD), and by molecular beam epitaxy (MBE). A we have written up a paper, “Long-lived Nanosecond Spin key component is the rapid iteration of samples following Relaxation and Spin Coherence of Electrons in Monolayer controlled introduction and removal of oxygen vacancies. MoS2 and WS2”, that is currently under review at Nature Physics (in fact, we just recently received very favorable Publications reviews in the first round of refereeing). In parallel with Yang, L., N. A. Sinitsyn, W. Chen, J. Yuan, J. Lou, and S. A. the above, we obtained new SrTiO3 crystals that we have Crooker. Long-lived nanosecond spin relaxation and characterized and are intending to measure in the SQUID spin coherence of electrons in monolayer MoS2 and WS2. 2015. Nature Physics. 11: 830. magnetometer within the next month. Future Work Growth and preparation of SrTiO3 epilayers and SrTiO3 / LaAlO3 interfaces will be performed at CINT-Los Alamos using our state-of-the-art PLD techniques. Direct compari- son with similar samples grown by MBE forms a key piece of this proposal; these will come from established con- nections with Chris Leighton & Bharat Jalan (Minnesota ) and Suzanne Stemmer (UCSB). Reduced concentrations of elemental impurities in MBE samples will allow us to as- certain whether, e.g., VO-Fe complexes play any role in the observed magnetic phenomena. In parallel, we will also continue to measure commercial “pure” SrTiO3 crystals as well as crystals that are intentionally and very lightly doped (0.01%) with Fe and Nb. Magneto-optical studies will primarily rely on our “work- horse” setup for magnetic circular dichroism (MCD;), which has proven extremely effective at revealing magnetism in SrTiO3. Our Oxford Spectramag split-coil magnet at the NHMFL has direct optical access and is ideally suited for this purpose (T=1.5-300 Kelvin, B= 0-8 Tesla). Tunable light sources include a spectrally filtered Xe lamp and tunable dye / Ti:sapphire /OPO lasers. Planned modifications in- clude more efficient frequency-doubling crystals to access deeper UV wavelengths (~350 nm), necessary for above- bandgap studies of SrTiO3 interfaces and thin epilayers. 425

Page 426

Materials for the Future Exploratory Research Continuing Project Energetic Materials Cocrystal Engineering: Toward Superior Munitions Philip Leonard 20150623ER Introduction An increased understanding of intermolecular bonding This research is developing energetic materials with new will be crucial to the development of explosives with and hopefully superior performance and safety proper- superior properties in terms of energy density, handling ties via cocrystallization. Cocrystallization will allow us safety, and specialized properties such as high melting to exploit synergistic properties in existing energetic and decomposition temperatures. Because explosives materials (EM) by alloying on a molecular level. Property are widely used in the harvesting of fossil fuels, this modification via creation of stoichiometric cocrystals research will improve our ability to extract energy re- has been demonstrated with great success in pharma- sources from deeper or more difficult terrain. Improved ceuticals and is beginning to bear fruit for explosives as explosive safety has direct application to DOD safety and well. For example, two EM that are powerful performers DOE safety and surety missions. We envision that the in terms of detonation velocity and shock pressure may capability to modify detonation velocity and pressure separately be highly sensitive to impact, by cocrystal- as a function of density will have important implications lizing those molecules one might be able to achieve a for the design of precision explosive applications such as layered structure where planar dislocations dissipate linear and conical shaped charges and explosive lenses, energy non-violently resulting in an acceptable safety applications which may be of interest to DARPA or NASA margin. In the same vane, two insensitive materials for controlled demolition such as rocket stage separa- which are non-performers due to inadequate density tion. could cocrystalize into a form with a higher than average Progress density as a result of superior packing. Given challenges in scaling and adopting new materials, creating bet- This project comprises three main pursuits, development ter EM through cocrystallization has great advantages. of computational ability, synthesis of cocrystalline mate- Cocrystallization of HE is also critical as an alternative rial, and characterization by X-ray and other methods. X- to formulation where the properties of molecules are ray characterization has been delayed due to instrument always averaged - often with trade-offs in sensitivity and problems which have now been resolved, the other two performance. focus areas are proceeding well. One great challenge in this effort is coupling modeling A systematic screening for cocrystal formation between and experiment. We are applying the latest in density sixteen target molecules was initiated and is ongoing. functional theory (DFT) functionals to predict packing Cocrystallization from solvent and during solvent grind- structures as well as modeling of single crystal X-ray data ing has been used for screening. The effects of the two to determine structures. This active discourse will ac- different crystallization methods on cocrystalization celerate experimental achievement while also providing formation and phase purity are being assessed. Charac- insight into the failings of our modeling assumptions. terization by x-ray diffraction and vibrational spectros- copy is ongoing. Reitveld refinement of powder x-ray Benefit to National Security Missions diffraction data is being developed to provide unit cell Cocrystals have great potential for technological innova- parameters and quantitative measures of phase purity. tion. This project will perform fundamental research into Assessing and attaining phase purity is crucial before the nature of intermolecular interactions in condensed transitioning to explosive property testing because of matter that will enlighten cocrytallization in defense- deleterious effects of remaining pure parent phases used related applications as well as life-science applications. to form the cocrystal. 426

Page 427

A new approach and code for the prediction of the atomic Future Work structure of co-crystals has been implemented in our Initial screening ‘APES’ code (Automated Prediction of Explosive Struc- Cocrystal formation will be performed using solvent grind- tures). APES employs efficient algorithms to scan and ing and wet chemistry techniques. Solvent grinding has rank potential crystal structures as a function of unit cell been used as an efficient method for pharmaceuticals but parameters (lattice parameters and angles), the position not in explosives because of the inherent hazards which and orientation the molecules, and stoichiometry. All of we have overcome. The LANL HE Crystal Lab has constant these degrees of freedom are searched using a stochastic, temperature and continuous circulation growth vessels, simulated annealing algorithm and the code is heavily par- specifically designed for growth of cocrystals. Our goal allelized (MPI) such that extremely high throughputs can is to create multiple highly crystalline materials for X-ray be achieved. The code makes no assumption of the space analysis. group, leaving that step instead to the post-processing of promising crystal structure with ab initio electronic struc- X-ray analysis ture codes. In order to understand intermolecular interactions in the synthesized co-crystals we must determine their full atom- While APES can quickly scan and rank millions of potential ic structure. We will apply atomistic simulation techniques structures, the accuracy of its predictions are controlled to aid solution. Powder XRD on the cocrystals will enable by the molecular geometries and descriptions of inter- us to identify their Bravais lattice and unit cell parameters. molecular interactions that we input. We found that the We will also measure vibrational spectroscopy in order to set of Buckingham potentials developed for C, H, N, O- determine the number of unique molecules per unit cell. containing molecules developed by D. E. Williams provides Our goal is to generate data sets for each new material for an adequate and transferable description of non-bonded modeling. interactions in organic molecular crystals. The set of partial charges assigned to every atom are extremely sensitive Modelling to the method used in their calculation. Hence, we dis- Armed with the Bravais lattice, unit cell parameters, covered that only relatively high-level approaches such at possible synthon structures, and the number of unique CHELPG give reliable sets of partials charges whereas sets molecules per unit cell, we will perform a series of Monte of simple Mulliken partial charges do not give the correct Carlo simulations with a generic force field to determine an ranking of crystal structures in APES simulations. We will optimal set of coordinates for the molecules of that lattice. soon use spectroscopic data from the experiments men- The stability and heats of formation of those structures tioned above to estimate the number of unique molecule will be computed by condensed phase DFT calculations in each unit cell in order to narrow the search space for in order to identify which is energetically most favorable. APES in terms of stoichiometry. Our goal is to be able to solve single crystal X-ray data and predict the correct packing structure (although possibly We are continuing to improve the purity of our most prom- not exclusively). ising candidate, TNX. We have discovered that solvent choice is critical for producing material with acceptable Sensitivity and performance characterization phase purity. TNX possesses special challenges for scale Sensitivity of new cocrystals will be assessed by the WX-7 up because it is a co-crystal of melamine and trinitrophlo- analytical team using standard drop-hammer, BAM friction, roglucinol which have very different solubilities. Never- and electrostatic discharge (ESD) testing. Performance theless, by transitioning from DMSO to ethylene glycol testing is not envisioned as occurring in the first year. Our followed by washing with methanol, we have been able to goal is to make safe energetic materials. produce high-purity material in excellent yield (~70- 75%). Combustion studies on pure parent HE (TNPG) and TNX Conclusion have been performed. TNPG was found to behave anoma- Our technical goals are to (1) develop and utilize rational lously: it burned at low pressures (< 800 psi), but would crystal engineering strategies to discover cocrystals and (2) not sustain combustion at high pressures. Due to smoke demonstrate superior explosive safety and performance obscuring the burn front at low pressures (≤ 1000 psi), TNX and improved material properties. The first goal has broad could only have data fit between 1200 and 1800 psi. Scale relevance to chemistry in general as modification of physi- up to 100 g scale of TNX for 1” rate stick will proceed next cal properties are significant to everything from pharme- quarter. ceuticals to non-linear optical (NLO) materials. Achieving a rational design process for cocrystallization based on more than intuition is critical to the future of the field. Develop- ing new explosives is no less critical to national security as 427

Page 428

existing mainstays are being threatened due to environ- mental concerns. Publications Bowden, P. R., P. W. Leonard, J. P. Lichtardht, B. C. Tappan, B. L. Scott, and K. J. Ramos. Energetic salt of trinitrophloro- glucinol and melamine. To appear in APS Shock Compres- sion of Condensed Matter. (Tampa, 14-19 Jun. 2015). 428

Page 429

Materials for the Future Exploratory Research Continuing Project Majorana Fermions for Quantum Information Filip Ronning 20150628ER Introduction the prospect of using them for quantum information Quantum information processing offers great potential technology. The most promising route to accomplish this in terms of National Security, fundamental science, and employs superconductivity in a topologically non-trivial commercial applications. However, quantum information material. Here topology refers to the “curvature” of the is notoriously sensitive to small amounts of noise. A so- electronic structure in analogy to the curvature of a lution to this problem lies in using so-called fault tolerant sphere versus a donut in geometry. The work we per- topologically protected states to store quantum informa- formed in the first year took three approaches. First, we tion. Majorana fermions, which were predicted over 70 aimed to identify that the semimetal, niobium arsenide years ago, but not yet definitely observed in nature, are (NbAs) is topologically non-trivial, and whether super- examples of a topologically protected state that could conductivity could be found in this material. Second, we be used for quantum information technology. Recent attempted to create clean interfaces of superconducting progress has suggested that appropriately engineered aluminum (Al) on the surface of samarium hexaboride structures in solid state systems could realize Majorana (SmB6), a proposed “topological insulator.” Third, we are fermions. Several promising first experiments have been exploring the robustness of using Majorana fermions for performed, but disorder has limited further progress on quantum memories in the presence of thermal decoher- existing materials. Instead, strongly correlated f-electron ence effects. materials, such as SmB6 (samarium hexaboride), can NbAs was recently predicted to be a so-called Weyl- also host Majorana fermions in appropriately engineered semimetal. This metallic state has many unique proper- structures, and are much cleaner. ties due to the electronic structure being topologically non-trivial. We were able to synthesize single crystals Benefit to National Security Missions of NbAs, and found that they have an amazing mobil- This work will address DOE Office of Science and na- ity at low temperatures (2 K), which is even larger than tional security mission needs. We propose to provide that found in graphene. Furthermore, by studying the evidence for an exotic particle known as a Majorana electronic structure using ultra high magnetic fields we fermion. A Majorana fermion has potentially a huge im- could determine that the electronic structure is indeed pact on national security as it can provide the platform topologically non-trivial as predicted by theory. Unfortu- for quantum information processing, which enables nately, our attempts to discover superconductivity in this computations not possible on a classical computing ar- material have been unsuccessful down to 0.050 K. chitecture. The advantage of Majorana fermions is that they are much less sensitive to noise effects which can As originally proposed, we are exploring devices made destroy other quantum information architectures. We out of the topologically non-trivial material SmB6, and will work on a material known as Samarium hexaboride. superconducting aluminum. We have been able to This material is interesting by itself as it may possess a deposit superconducting pads of Al on SmB6 that are unique surface state. Thus our studies will also shed light 100 nm thick. Importantly, these pads have been shown on a new type of matter found in Samarium hexaboride. to be consistently and robustly adhered to the sample surface, and fully superconducting from resistance Progress measurements. Initial problems with the Al not becom- The objective of this proposal is to find evidence for a ing superconducting were resolved through changing the Majorana fermion in a condensed matter system with metal source to reduce any contaminants. In early work 429

Page 430

the resistance of the patterned samples showed irrepro- We will explore the properties of devices which induce ducibility because of intermittent contact between the superconductivity on the surface state of SmB6. This will SmB6 and the Al. These issues were overcome by improv- be done by depositing aluminum (itself a superconducting ing the process by which the sample surface is prepared. material) on top of the SmB6 crystal. If the interface be- The samples are now aggressively cleaned and undergo a tween the SmB6 crystal and the aluminum is clean enough, thorough oxide removal step. In addition, a 3nm titanium then we will have the possibility of creating non-linear layer is deposited in between the SmB6 and Al layer to transport in such a device from which we could identify if improve adhesion. the SmB6 does indeed possess the exotic electronic state. In fact, the signatures in the transport measurement if Progress has also been made on the patterning of the successful may also reveal the first evidence of a Majorana narrow junction between two Al strips. Small samples are state. We will also investigate how thermal decoherence difficult to process due to surface non-uniformities and the can affect the utility of Majorana fermions for quantum small dimensions required during the lithographic process. memories. A procedure has been developed that has resulted in 70 nm gaps between contacts, but progress still needs to be Conclusion made to reduce the linewidth variation in the gap and to We propose to provide evidence for the existence of the improve the adhesion of the e-beam resist to the sample. elusive Majorana fermions in hybrid structures using SmB6, and to further demonstrate that the position of the In order for us to show evidence for Majorana fermions in Majorana fermion can be moved. The physical exchange the junctions it is necessary to perform low power mea- of Majorana states leads to exotic properties, which in the surements of the current-voltage characteristics of the future can form the basis for quantum computation. junctions. The experimental setup to perform these mea- surements is now operational. Initial measurements show Publications qualitatively that the contact between the Al and SmB6 is Ghimire, N. J., Y. Luo, M. Neupane, D. J. Williams, E. D. good. The next step will be to get quantitative information Bauer, and F. Ronning. Magnetotransport of single on the contact quality, and once patterned junctions are crystalline NbAs. 2015. JOURNAL OF PHYSICS- produced we can further optimize the electronic setup for CONDENSED MATTER. 27 (15): 152201. their characterization. A self-correcting quantum memory is a quantum many- body system that can protect its quantum state against er- rors for a period of time that grows rapidly with the system size. Most proposals for quantum memories in 2 dimen- sions, such as those using Majorana fermions, have issues that make them unstable with respect to small local static perturbations. We recently introduced a new approach to quantum memories where the interactions of the system are random, according to particular distribution. Small- scale numerical simulations show that this approach may also work at larger sizes. Our line of research for the next few months is to investigate the role of thermal fluctua- tions in these systems and whether practical methods for “storing” and “reading” a quantum state are possible. One paper was published in JPCM, and another submitted to PRX on the above work. Also, the PI gave an invited talk at an international conference on SmB6. Future Work Samarium hexaboride (SmB6) may or may not possess an exotic electronic state on the surface of the material. If it exists, and can be made superconducting one has the chance of creating the particle called a Majorana fermion, which could be used in quantum information processing. 430

Page 431

Materials for the Future Exploratory Research Final Report 3-Dimensional Characterization of Nuclear Fuels: Microstructural Evolution under Representative Temperature and Thermal Gradients Donald W. Brown 20130232ER Abstract Background and Research Objectives X-rays and neutrons have been used as probe particles Recently, interest in nuclear energy in the Unites States to monitor the microstructure of nuclear fuel under con- (and worldwide) has surged, with increased efforts to ditions simulating those perceived by the fuel in reactor. extend the lifetime of the current fleet of reactors and Ceramic (UO2) has been the focus of the experimental to develop advanced reactors that operate at higher work, but other metallic and composite fuels have been temperatures and use fuel to higher burn-up. The acci- studied when strategic. High energy x=ray diffraction dent at Fukushima has not changed the fact that nuclear microscopy (HEDM) has been used to monitor grain energy will remain part of the US energy mix for the morphology under extreme temperature (2000°C), foreseeable future, but has shifted research priorities to x-ray micro-tomography (μΤ) has been used to monitor accident tolerant designs. pore and crack structure under temperature and irradia- Nuclear fuel functions in possibly the most extreme tion (separately) conditions, and small angle scattering environment. The extreme thermal gradient drives radial (SAS) has been used to monitor void and pore evolution grain growth and void migration in ceramic (UO2) fuel under high temperature. In many cases, the analysis of which degrades thermal conductivity and facilitates the scattering data is still on-going, but clear effects are transport of fission products which is important in the available. case of cladding failure. Current models attempt to Diffusion of xenon gas (a fission product) from within the predict useful lifetimes of fuel assemblies. Safety mar- microstructure to the volume external to the fuel and gins and predictions of the engineering performance within cladding (the plenum) is critical to the useful life- of nuclear reactor fuel rely on modeling codes used to time of ceramic nuclear fuel. The xenon atoms initiate as predict dimensional change, stress state and fission gas interstitial impurities and diffuse to form clusters, voids, release as a function of burn-up and temperature his- and finally pores on their way to the plenum. The HEDM tory. Thermal conductivity is the single most important data shows clearly that increases in the oxygen stoichi- material parameter in these codes and is heavily depen- ometry (UO2+x, where x0.1) increases the grain size dent on microstructure (e.g. grain morphology, porosity, following sintering under the same conditions. This is im- etc.), oxygen stoichiometry, chemical segregation, etc. portant because grain boundaries act as short circuits for To date, empirical and phenomenological descriptions the fission produced xenon to reach the plenum. Thus, of the evolution of the microstructure of fuel materials large grains slow the transfer of fission gases (e.g. Xe are used to calculate the thermal conductivity, but this and Kr) from the microstructure to the plenum). Small leaves huge uncertainties and necessarily results in large angle x-ray scattering (SAXS) has been utilized to provide safety margins and inherently inefficient use of resourc- statistical information related to the development of es as well as increased production of waste. xenon bubbles at temperature. X-ray μΤ shows clearly Microstructural evolution occurs in nuclear fuels due the development of small pores at the interface of a to the high temperature, extreme thermal gradient and metallic fuel and its bonded cladding during irradiation. radiation damage. In ceramic fuels, grains grow in the This is possibly related to the formation of bubbles on direction of extreme thermal gradient, that is, radially, the surface of the metallic fuel which block coolant flow while voids elongate along the gradient. Cracks develop channels resulting in dangerous hot spots in research during the first thermal cycle due to the stresses associ- reactors. ated with the strong thermal gradient. Oxygen stoichi- 431

Page 432

ometry in pure UO2 and actinide content in mixed oxide ments during deformation (Cu) and annealing (Al and Ni) fuels (MOX), evolve under these conditions. All of these have been completed. changes in the microstructure affect both the thermal con- We have evolved the technique as necessary to study ductivity and transport of fission products in ways which high-Z materials. In particular, we have completed studies we can neither predict accurately nor fully understand, but of ceramic UO2 as a function of processing and after time these are the lifetime limiting behavior of ceramic nuclear at temperature simulating operation conditions. Figures fuels. 1a-c show grain maps and grain size distributions in UO2, The objective of the research was to characterize the UO2.11 and UO2.16 material sintered at 1450°C for 4 microstructural evolution of the fuel under conditions hours. Note the difference in scale of figure 1a. Clearly, the simulating those perceived during operation, that is, at higher oxygen content leads to larger grain sizes in the as high temperature and high thermal gradient. High energy sintered material. x-ray diffraction and imaging can probe the microstruc- ture of bulk nuclear fuel at relevant length (μm) and time (seconds) scales while in this demanding environment. The understanding gleaned from the experimental work has been used to develop microstructural models associated with the MARMOT code being developed at Idaho National Lab. Scientific Approach and Accomplishments We have used x-ray scattering experiments on nuclear fuels to monitor the grain morphology, orientation, void structure, and chemical segregation in their respective mi- crostructures as a function of external perturbation, such as time at temperature. The primary subject has been ce- ramic UO2 in an extreme thermal gradient. However, stra- tegic studies of other fuel types were completed as well to 1.) develop techniques for probable use in MaRIE and 2.) develop strategic partnerships in the arena of nuclear fuel with other institutions such as INL and UC Berkeley. High energy diffraction microscopy (HEDM) was used to probe the grain morphology before and after time at elevated temperature to inform microstructural models (e.g. MARMOT) of the growth of grains in specific neigh- borhoods under operation conditions. X-ray micro-tomog- raphy (μT) has been used to characterize the evolution of pores (a key component of the microstructure) both under high temperature and after irradiation. Finally, Small Angle Scattering (SAS) has been used to characterize heteroge- neities on nm length scales during manufacture of the fuel and as bubbles develop in fuel that was implanted with Xe Figure 1. Grain map and grain size distribution for a.) UO2, b.) atoms as interstitials. UO2.11, and c.) UO2.16. High-Energy Diffraction Microscopy The HEDM technique The HEDM technique allows us to pick out specific grains was developed in parallel by the Suter group at Carnegie that are, for instance, far in the tails of the grain size dis- Mellon University and by the Poulson group in Denmark to tribution to search for the cause of such anomalous grains study grain and sub-grain level crystal orientation, strain and evaluate if they have an inordinate control on the and defect density. The technique uses a near field area properties of the overall sample. Figure 2 shows an inverse detector to record diffraction spots and directly image pole figure with all of the 20018 identified grains located. grains in a polycrystal based on orientation. To date, the The colors are indicative of the size of the individual grains. technique has been used on annealed metals with relative- There is no obvious correlation between grain size and ly low z-number, e.g. copper, and recently in-situ measure- orientation. Six grains have been extracted from the micro- 432

Page 433

structure which have sizes of greater that 3x the average 250keV X-rays at the 10BM beamline at the Advanced Pho- grain size. Again, these six grains span the range of orienta- ton Source. This clearly demonstrates the ability to non- tions, indicating that grain orientation does not play a role destructively visualize the interior of a full fuel pellet using in the anomalous growth of these specific grains. A deeper high energy x-rays. search into the effect of the grain neighborhood effect on the grain growth is ongoing. Figure 3. Grain map and grain size distribution for UO2 a.) as- sintered and b.) after 3 hours at 1950C. Figure 4c shows a very high resolution tomographic slice Figure 2. Mapping of the orientations of identified grains on of a section from a metallic uranium fuel foil after irradia- the stereographic projection. The color of each spot (grain) is tion, demonstrating the ability to characterize very active indicative of the grain size, as per the color bar. Six grains with material. The sample is a monolithic uranium-molybdenum anomalously large size have been identified and located on the stereograph. alloy clad aluminum, with a 15 μm thick zirconium interac- tion barrier separating fuel from the cladding. Each pixel is Figure 3 shows grain maps and grain size distribution of 0.7μm. The three layers of the sample are clearly evident UO2 in the as-sintered state and after 3 hours at 1950°C. as are pores that have formed during the irradiation. The Grain growth during time at temperature is readily appar- failure mechanism of these monolithic fuel plates is de- ent. As above, we are currently searching the local grain bonding of the fuel from the cladding, resulting in bubbles morphologies in an attempt to understand the driver for forming on the surface of the fuel and block coolant flow selection of preferred grains for growth. This information channels. The mT not only probes the possible failure will prove critical for the development of models that can mechanism of the research reactor fuel, it also demon- predict grain growth, which controls the diffusion of fission strates the capability to non-destructively probe highly gases from the microstructure to the space surrounding activated spent fuels. the fuel. Small Angle Scattering SAS techniques are well established Micro-Tomography The pore and crack morphology has but application to UO2 at high temperature is novel. Thin been imaged with x-ray μT. Several groups have pushed film samples (1μm) of UO2 were grown using Ion Bom- the limits of μT to sub-micron resolution in some highly bardment Assisted Deposition (IBAD). A pure UO2 sample optimized cases, but in general, the current resolution limit was grown as well as samples with 2 and 4 weight per- is ~1-2μm. To study the growth and migration of pores and cent xenon. The xenon atoms are trapped in the lattice as cracks in UO2, we have pushed this resolution to ~0.7 μm interstitials in a similar fashion to the xenon produced as as well as adapt the technique to higher energy incident a fission product during use. Small Angle Xray Scattering x-rays. Limitations in scintillator technology prevent further (SAXS) was completed during heating of the samples to reduction of the resolution at high x-ray energy. Figures 4a monitor the motion of the xenon from individual intersti- and b show (a.) a relatively low resolution (6μm) tomo- tials to form clusters, voids, bubbles, etc. graphic slice through an engineering scale (5mm diameter) ceramic nuclear fuel pin as well as (b.) a micrograph of Figure 5 shows SAXS data from a.) pure UO2 and b.) the same area. The x-ray tomograph was collected with UO2 with 4wt% xenon with time at 225°C. The pure UO2 433

Page 434

sample shows absolutely no evolution in the scattering in-situ experiments designed to reach 2000°C were halted over a 4 hour period. In contrast, the 4wt% xenon sample at 1500°C when oxide was detected in the diffraction shows an increase of scattering at ~0.02 1/Å in 10’s of pattern, suggesting (correctly) that the tantalum sample minutes, which indicates the development of a heteroge- containment had oxidized. The experiments were stopped neity (presumably bubbles) with a length scale of ~30nm. with no breach of containment. Further development of a Transmission electron microscopy (TEM) analysis has just boron-nitride container for in-situ experiments occurred. begun. Figure 5c shows elemental maps on a line across BN was chosen because it is both more stable than re- the sample. The xenon content fluctuates with a length fractory metals, and is transparent to high energy x-rays. scale on the order of 10’s of nm, consistent with the het- However, we were unable to demonstrate the robustness erogeneity indicated by the SAXS data. Complete analysis of the BN containment to the satisfaction of either the ex- of this data will inform models of the motion of xenon in perimental team or the APS radiation safety committee in the microstructure as the release of the xenon from the time for the final experimental run of this project. Because microstructure is a primary driver in the overall burn-up of the ramifications of a containment breach on a radioactive the fuel. sample at a synchrotron light source are considerable, a very conservative approach was taken. Future Much of what is shown in this report is “prelimi- nary” results which have not been analyzed to the fullest extent. Following the completion of this project, Reeju Pokharel (postdoc) will continue with roughly 0.3FTE of support from the civilian nuclear energy project to finish processing this data as well as continue work on UN/U3Si5 composite fuels. The team is part of multiple proposals to the DOE Office of Nuclear Energy. In 2014, Brown headed an effort from which a full proposal was requested. The Figure 4. a.) Slice from a tomographic reconstruction of a proposal was not funded, but has been re-submitted in 5mm diameter UO2 fuel pin. b.) Optical micrograph of same. c.) 2015 with more results and suggested revisions request- Tomographic reconstruction of a section of an irradiated metallic ing $1M over 3 years. Proposals from Idaho National Lab (uranium molybdenum) fuel. (metallic uranium fuels) and Advanced Photon Source (cladding) including significant participation from the LANL team have also been submitted. Impact on National Missions The work described here-in has provided previously unob- tainable microstructural data useful for the CASL, NEAMS and Fuel Cycle R&D programs residing at LANL, as well as the MARMOT code developers at INL (and now Penn State University). Mike Tonks at PSU and Brad Fromm at INL are currently using the 3D microstructures generated with HEDM in the as-sintered state as starting points for predic- tive calculations. The models will evolve the microstructure under the same conditions achieved in our experiments, to be compared directly to the final microstructures we Figure 5. SAXS data collected on (a.) pure UO2 and (b.) Xe im- observed. This provides unprecedented validation oppor- planted UO2 during time at 225°C . tunities for the code developers. In-Situ Experiments Above is an example of a very suc- Publications cessful in-situ x-ray experiment completed as part of this Brown, D. W., Balogh, Byler, C. M. Hefferan, J. F. Hunter, project, i.e. monitoring the microstructure while at tem- Kenesei, S. F. Li, Lind, S. R. Niezgoda, and R. M. Suter. peratures approaching 700°C. While not shown here be- Demonstration of Near Field High Energy X-Ray Diffrac- cause of space limitations, in-situ SANS experiments during tion Microscopy on High-Z Ceramic Nuclear Fuel Mate- sintering of UO2 powder to 1200°C were also successfully rial. 2014. MECHANICAL STRESS EVALUATION BY NEU- completed. However, a goal of this project was to com- TRONS AND SYNCHROTRON RADIATION VII. 777: 112. plete in-situ HEDM to ~2000°C. At Diamond Light Source, 434

Page 435

Materials for the Future Exploratory Research Final Report Very Low Temperature Scanning Point Contact Spectroscopy Investigation of Inhomogeneous States on the Nano-scale Roman Movshovich 20130285ER Abstract unique apparatus developed within this proposal will This project aims to push the latest advances in scanned open new avenues of the correlated electron research, probes techniques to a new state-of-the-art level to ad- notably within unconventional superconductivity and dress a number of outstanding problems from the “top quantum critical phenomena, an area of LANL’s tradi- shelf” issues in condensed matter physics. It combines tional excellence and world leadership our expertise in several Scanning Probe Microscopy Background and Research Objectives (SPM) and Spectroscopy (SPS) techniques and a dilution refrigerator, instrumented with a 14 Tesla magnet, to Our proposal involves a broad spectrum of outstanding develop a novel Very Low Temperature Scanning Point topics of research. It is the result of the confluence of Contact Spectroscopy (VLTSPCS) apparatus. The VLTSPCS new capabilities afforded by the proposed novel VLTSPCS can be used to investigate several outstanding high apparatus and our interests, demonstrated by research profile problems in correlated electron physics, where conducted both at present and in recent past at LANL. our group at LANL has both demonstrated expertise and The High-Field Superconducting (HFSC) state of CeCoIn5 world leadership. Point Contact Spectroscopy (PCS) has [3] is one of the confluences, and will be one of the fo- been used extensively for investigating homogeneous cal points of this proposal. This state was proposed [3] as superconducting (SC) states mostly in zero magnetic realization of a spatially inhomogeneous superconduct- field. At the interface between the normal tip and super- ing state, predicted by Fulde and Ferrell and Larkin and conducting sample, Andreev Reflection (AR) of quasi- Ovchinnikov (FFLO) [1,2] in the early 1960’s. particles creates a bound state, which reveals itself as a zero-voltage resonance peak in the dI/dV spectra inside The FFLO state is a straightforward consequence of the the SC energy gap. Our unique VLTSPCS apparatus will go very successful BCS (Bardeen-Cooper-Schrieffer) theory well beyond state-of-the-art PCS capabilities: sub-nano- of superconductivity, and a direct outcome of the pair- meter resolution of the PCS combined with the scanning breaking effect of the magnetic field acting on the spins capability and in-situ control of the force between the of the superconducting electrons. In spite of the appar- PCS tip and the sample, high magnetic fields, and very ent theoretical simplicity (Fig. 1), an FFLO state was not low temperatures will allow us to explore the spatially observed so far, likely due to several stringent require- inhomogeneous states in an unprecedented part of the ments on materials and physics aspects needed to sup- phase space. Within this proposal we will address some port an FFLO state. of the hottest topics in correlated electron physics: (1) spatial variation of the energy gap in the postulated Experimental results stimulated great theoretical activity Fulde-Ferrell-Larkin-Ovchinnikov [1,2] superconducting in this field, with diverse scenarios based on attraction state in CeCoIn5 at high fields [3]; (2) direct visualiza- or competition between superconductivity and magne- tion of the exotic high-field vortex structures, up till now tism. A number of theories even suggest that the FFLO studied only by Small Angle Neutron Scattering (SANS) state underlies the physics of the HFSC state of CeCoIn5 techniques, e.g., in CeCoIn5 [5,6] and UPt3 [7]; (3) high- [14-16]. In particular, enhancement of the normal quasi- field capability will give access to field-tuned Quantum particle density of states at the nodes of the FFLO state, Critical Points (QCP), where some theories predict the where the superconducting energy gap goes to zero, was breakdown of the quasiparticle concept [8]. We will be shown to lead to the formation of the MLRO [15]. On the able to test this directly, as PCS spectra are sensitive other hand, there are also theories that do not require to the sample’s local electronic density of states. The 435

Page 436

FFLO for formation of the HFSC phase of CeCoIn5. Some of for or against an FFLO state. We will resolve this contro- these theories show that MLRO can be a result of coupling versy by directly visualizing the spatial variation of the of the superconducting and magnetic order, usually with superconducting energy gap on the sub-nanometer scale, a third fast-spatially-varying (on the order of a unit cell) using the VLTSPCS apparatus to be developed within this superconducting order parameter [5,17,18]. Furthermore, proposal (details below). pair-breaking along the nodes of the d-wave superconduc- The VLTSPCS apparatus will enable the visualization of tor by a magnetic field due to the Zeeman effect was also vortex lattices in high magnetic fields, a table-top capabil- shown to lead to MLRO [19]. ity complimentary to the Small Angle Neutron Scattering (SANS). The VLTSPCS will fill several niche applications, for example, in cases where no large samples are available (required for SANS), or for studies of thin films. The PCS couples directly to the quasiparticles in the SC state and thus to the superconducting order parameter. Therefore, the contrast at high fields will be much greater than in SANS experiments that couples to variations of the mag- netic field only. The significance of such apparatus itself cannot be overes- timated. The combination of a dilution refrigerator with a new 14 T superconducting magnet, a scanning platform for a variety of surface probes options, including conventional Magnetic Force Microscopy (MFM), Atomic Force Micros- copy (AFM), and Kelvin Force Microscopy (KFM), and the technical focus of this proposal - an innovative Scanning Point Contact Spectroscopy (SPCS) probe, capable of prob- ing the magnitude of the superconducting energy gap on the sub-nanometer scale – is unique. It will open a vast ex- perimental phase space and a new capability to interrogate correlated electrons at very low temperatures and high fields, and help assure the continued preeminent position of LANL in this research field. Scientific Approach and Accomplishments The phase space occupied by the HFSC phase of CeCoIn5 Figure 1. Zeeman splitting of the spin up and spin down bands presents formidable challenges of very low temperatures leads to a finite momentum of superconducting Cooper pairs of (below 300 mK) and high magnetic fields (up to 12 T). electrons with opposite spins, resulting in the spatially varying Probing the spatial dependence of the SC energy gap in energy gap D() with nodal ( D() =0) planes perpendicular. CeCoIn5, visualizing the vortex lattice, or the SC energy gap around an impurity, requires a scanned probe capability. Whether or not the FFLO physics plays an important role in The STM is a powerful tool that has been used to investi- the formation of the HFSC state, the MLRO state is unique gate a number of unconventional superconductors, such as in its own right: magnetism is confined to the supercon- BSCCO, a high-temperature copper-oxide superconductor. ducting state, and disappears when CeCoIn5 is driven into However, no STM system exists that satisfies the extreme the normal state with increasing magnetic field or tem- requirements listed above (the field range; the tunneling perature. These properties are contrary to the canonical current in existing STMs is always parallel to the direction paradigm of competition between superconductivity and of the magnetic field, contrary to some of our measure- magnetism. Therefore, the HFSC state in CeCoIn5 is a new ment requirements, see below). The STM is extremely exotic state of matter and understanding it is of paramount sensitive to the sample surface quality and cleanliness. importance. In addition, atomic resolution of STM imposes stringent Bulk measurements, whether thermodynamic, transport, requirements on the control of the distance between the or spectroscopic (NMR and neutron scattering), so far have STM tip and the substrate, since the tunneling current is failed to provide a definitive “smoking gun” result either exponentially sensitive to the probe-sample separation. 436

Page 437

Therefore the STM is very sensitive to the mechanical imaging of the ferromagnetic domains. This resulted in vibrations and electrical noise in the system. The superior three recent articles on the details of formation and evolu- atomic resolution of the STM is not required, however, to tion of the ferromagnetic domains in electron correlated achieve our goals, since the characteristic length scales of materials (two submitted), see the publication section of the FFLO state, vortex separation, and impurity-induced the report. suppression of the SC energy gap are all on the order of Dr. Neliza Leon Brito was hired to take over from Dr. Jee- the superconducting coherence length ξ ≈ 2-10 nm. The hoon Kim, who left the project to take a professor position sub-nanometer spatial resolution of the PCS is therefore in Postech, South Korea. Neliza came with experience in fully sufficient to achieve our experimental goal. The SPCS designing and building a Scanning Hall Microscopy, for Alex will employ the Contact-mode Atomic Force Microscopy de Lozanne group in University of Texas, Austin, the same (CAFM) technique. It presents us with a perfect compro- group where Dr. Jeehoon Kim obtained his Ph.D. mise: the spatial resolution will increase to sub-nm range (still satisfying the experimental requirements), with a ben- First, to get up to speed with MFM technique, Dr. Leon efit of drastically reduced sensitivity to surface cleanliness Brito used Meissner technique of measuring magnetic and preparation and to mechanical and electrical noise. penetration depth in several pnictide superconductors of The “lower tech” SPCS presents us with the best tool to ad- interest (and Nb, our standard for these measurements). dress the science objectives within this proposal. These experiments require fine field control, and ability to take cantilever response at a single point over the sample. Next step was to extend the operation of the MFM ap- paratus into scanning mode and for Neliza to acquire this expertise, a necessary part of the SPCS technique. Subsequent MFM investigations focused on several ferro- magnetic (FM) compounds, and in particular, an easy-axis ferromagnet Fe3GeTe2. For the easy-axis FM compound the expected domain structure observed on the surface perpendicular to the easy axis display a very characteristic domain branching structure (see Figure 2(d)), where the large domains in the bulk of the material break up into Figure 2. Magnetic domain structure of FeGeTe observed with smaller and smaller domains as they approach the surface, MFM for the (a) (101) surface (lift height = 300 nm). (b) (001) thereby reducing the leakage (or stray) field outside of surface for a lift height 1,500 nm. (c) (001) surface for a lift height the sample and lowering the total magnetic energy of the of 600 nm. All MFM images, a - c, were taken at the same scan- system. One expects just two dominant magnetizations of ning size (28 × 28 µm) and experimental conditions (zero field the domains: up and down. The resulting color image of cooled; T = 4.0 K; P the schematic) (d) schematic of the domain the surface would correspondingly be expected to have structure near the surface of uniaxial ferromagnets, where the two dominant colors (contrasts), consistent with the im- c-axis is the magnetic easy axis, showing how the branching age shown in Figure 2(b). Instead, we observed that the of bulk stripe domains of alternating magnetization near the sample edge results in a surface magnetic structure consisting smallest “dagger” domains (seen as dots within the stripe of wavy stripes with rows of dagger domains with the opposite domains) became bright color, indicating repulsion of the magnetization; (e) the unperturbed FeGeTe microstructure where cantilever from the sample, as shown in Figure 2(c). This the easy axis is tilted away from the c-axis and only two MFM effect appeared only when the lift height of the cantilever contrast states (indicating only two domain magnetization direc- (distance above the surface) was reduced below a critical tions) are observed, and (f) the microstructure observed with value. Two possible scenarios need to be considered: (1) MFM for lift heights ≤ 600 nm due to influence of the field H of MFM magnetic tip flips the smallest domains of the sam- cantilever on the sample, creating a third magnetization direc- ple, illustrated in Figures 2(e,f) or (2) the stray field from tion and an MFM contrast. the sample is large enough to flip the magnetization of the Within experimental part of the proposal the focus was tip of the MFM cantilever (so called soft cantilever regime). on developing MFM expertise of Dr. Neliza Leon Brito – a We initiated a collaboration on the modeling of the MFM postdoc who started to work on this project in December signal (in this case the resonant frequency of the magnetic 2013. Neliza needed to become an expert in operation of cantilever, which is proportional to the second gradient of MFM, crucial for development of the SPCS technique. She magnetic field) in a complicated geometry of the surface ran MFM in several modes, successfully performing Meiss- branching domains, with Dr. Denis Pelekhov of the Ohio ner force technique penetration depth measurements, and State University. One simulation for a cantilever lift height 437

Page 438

of 600nm is shown in Figure 3. Such simulations point We continue a strong collaboration with Pr. Jeehoon Kim towards the second scenario (cantilever direction is influ- from Postech, Korea, who is continuing the work of devel- enced by the sample’s stray field) as the most likely one, oping low temperature SPCS that he started as a postdoc but not completely, as the cantilever’s behavior is some- here at LANL. We have hosted several of his students, and where between the soft and hard regimes. These results exchanged samples and components. In particular, he sup- are being written up two papers, one focusing of physical plied us with several home-made platinum point contact properties of Fe3GeTe2, and another on general subtleties cantelevers. This collaboration resulted in several publica- of the MFM investigation of magnetic compounds, and a tions, one to be submitted to Nature Physics (see publica- great care that needs to be taken while interpreting MFM tion section of the progress report). data. This paper will be of significant value to the scanning Theoretically, previous studies of various aspects of un- probes community. conventional heavy-fermion superconductivity, such as the FFLO state, vortex state and tunneling into the HFSC have thoroughly prepared us for this project. We developed theoretical modeling capabilities to tackle the spatially nonuniform FFLO state [22] as well as vortex state [4] in unconventional superconductors. More recently, we de- veloped a multi-channel tunneling model for point-contact spectroscopy (PCS) in heavy-fermion systems and applied it to CeCoIn5 [23,24]. This unique problem presented several challenges, as the available PCS data on heavy fermion compounds cannot be fitted by the standard BTK (Blonder- Tinkham-Klapwijk) theory of PCS, which was extremely successful for conventional superconductors. It turned out, for example, that one needed to account for the localized states at the surface, as well as for both heavy electron (correlated) and light electron bands. The quantum inter- ference between localized and itinerant electrons sufficed to describe the asymmetric (Fano) line shape of the differ- ential conductance. Detailed modeling was required to ex- Figure 3. Micro-magnetic LLG simulations of a model sample tract the Fano and superconducting parameters of interest system, with stripes of alternating up-down magnetization per- from measured dI/dV spectra, such as the superconduct- pendicular to the sample’s surface, with small cylindrical islands ing energy gap and energy level of the localized f-electron in the middle of the all stripes and regularly spaced along them. state. That work demonstrated that a full understanding The top graph gives the total stray field of the sample, next is of the SPCS data and extraction of relevant physics (energy the orientation of the soft cantilever that follows the direction of gap and other parameters) requires state-of-the-art model- the resulting combined magnetic field. Second graph from the ing capabilities of surfaces within a multi-channel model bottom gives a second gradient of field components along the for electrons hopping from the tip to the sample. Based on z-direction, which determines the forces on the MFM tip. The the previous expertise we have accomplished the following bottom graph shows the expected MFM signal (proportional to the second derivative of field along z) for soft (magnetization fol- within the theoretical component of this project: lows the direction of the external field) and hard (magnetization is constant, parallel to z-axis). 1) We derived an analytic formulation of the tunneling conductance in the normal state based on our multi- On the theory side, the objective was to explore the ques- channel point-contact tunneling model for heavy-fermion tion of how the FFLO state will be revealed by SPCS. Our materials. This new formulation provides a great simplifi- combined expertise in both FFLO physics [18] and PCS of cation over previous formulations and allows us to better heavy-fermion superconductors [19] enables the modeling understand qualitatively the normal state properties of of the SPCS spectra and extraction of the superconducting electrons tunneling from a metallic tip into a heavy-fermi- energy gap in scanned mode of operation, when the point on metal beyond numerical results. We will continue with contact parameters may vary during the surface scan. The our efforts of the construction of a tunneling Hamiltonian model calculations include the effects of high magnetic that incorporates the magnetic field, correlations effects, fields and strong Pauli limiting of CeCoIn5 on the SPCS and other ordered states, such as a spin-density wave. spectra. 2) We investigated the evolution of the spectral function, 438

Page 439

tunneling density of states of quasiparticles, and the local superconducting order parameter and the superfluid den- magnetic penetration depth in doped 122 iron pnictide su- sity were dramatically suppressed toward the dirty limit, perconductors. Since our main interest is toward scanned indicating the violation of the Anderson theorem for con- probes, a special emphasis was given to the local effects in ventional s-wave superconductors and the breakdown of the presence of a single Zn impurity and the overall sup- the Abrikosov-Gorkov theory for impurity-averaged Green’s pression of superconductivity at arbitrary impurity concen- functions. Furthermore, we found that the superconduct- trations. ing phase was fully suppressed close to the critical impu- rity concentration of roughly n_(imp) ≈ 10%, in agreement Technical details with recent experiments. We introduced a two-channel tunneling model to general- ize the widely used BTK (Blonder-Tinkham-Klapwijk) theory of point-contact conductance between a normal metal contact and superconductor. Tunneling of electrons can occur via localized surface states or directly, resulting in a Fano resonance in the differential conductance G=dI/ dV. We presented an analysis of G within the two-channel model when applied to soft point-contacts between nor- mal metallic silver particles and prototypical heavy-fermion superconductors CeCoIn5 and CeRhIn5 at high pressures. In the normal state the Fano line shape of the measured G was well described by a model with two tunneling chan- nels and a large temperature-independent background conductance (see Figure 4). In the superconducting state a Figure 4. Fitting the Fano line shape of CeCoIn at ambient pres- strongly suppressed Andreev reflection signal was ex- sure in the normal state. The transparency D = 0.210 is in the plained by the presence of the background conductance. tunneling rather than high transparency limit. The solid (black) We reported Andreev signal in CeCoIn5 (shown in Figure lines are fits with constant background conductance G0 = 282.6 5) consistent with standard d_(x2-y2)-wave pairing, as- (kΩ), which is shown as red-shaded background. suming an equal mixture of tunneling into [100] and [110] crystallographic interfaces. Whereas in CeRhIn5 at 1.8 and 2.0 GPa the signal was described by a d_(x2-y2)-wave gap with reduced nodal region, i.e., increased slope of the gap opening on the Fermi surface. The is a possibility is that the shape of the high-pressure Andreev signal is affected by the proximity of a line of quantum critical points that extends from 1.75 to 2.3 GPa, which is not accounted for in our description of the heavy-fermion superconductor. Recent experiments on Zn-doped 122-type iron pnic- tides, Ba(Fe,Co,Zn )2As2, challenged our understanding of electron doping the 122s and the interplay between doping and impurity scattering. To resolve this enigma, we investigated the disorder effects of nonmagnetic Zn impurities in the strong (unitary) scattering limit on various properties of the system in the s+- wave superconducting Figure 5. Fitting the Andreev signal of CeCoIn in the super- pairing state. The lattice Bogoliubov–de Gennes equation conducting state at ambient pressure and at 1.31 K with = 23K. (BdG) was solved self-consistently based on a minimal We assumed isotropic -wave gap (left) and -wave gap (right). The wave is plotted for standard (black) and five times steeper two-orbital model with an extended range of impurity (red) slope of the gap at nodes with equal weight of tunneling concentrations. We found that Zn impurity is best modeled into [100] (0°) and [110] (45°) surface orientations. Inset: angle as a defect, where charge is mainly localized, but scattering dependence of the gap function. is extended over a few lattice sites. With increasing Zn con- centration, the density of states showed a gradual filling Impact on National Missions of the gap, revealing the impurity-induced pair-breaking This project supports several BES programs: Novel Materi- effect. Moreover, both the disorder configuration-averaged als (PI Joe D. Thompson) and Vortex Physics (PI- L. Civale). 439

Page 440

It will open new areas of research in the areas of the by nuclear magnetic resonance. 2005. PHYSICAL RE- Strongly Correlated Electron Physics, magnetism, and gen- VIEW LETTERS. 94 (4). eral superconductivity. This project supports both themes 11. Kumagai, , Shishido, Shibauchi, and Matsuda. Evolution of Emergent Phenomena and Defects and Interfaces, two of Paramagnetic Quasiparticle Excitations Emerged of the three areas of the Materials Strategy at LANL. in the High-Field Superconducting Phase of CeCoIn5. References 2011. PHYSICAL REVIEW LETTERS. 106 (13). 12. Tokiwa, , Movshovich, Ronning, E. D. Bauer, Papin, A.

1. FULDE, P., and R. A. FERRELL. SUPERCONDUCTIVITY D. Bianchi, J. F. Rauscher, S. M. Kauzlarich, and Fisk. IN STRONG SPIN-EXCHANGE FIELD. 1964. PHYSICAL Anisotropic effect of Cd and Hg doping on the Pauli REVIEW. 135 (3A): A550. limited superconductor CeCoIn(5). 2008. PHYSICAL
2. LARKIN, A. I., and Y. N. OVCHINNI. INHOMOGENEOUS REVIEW LETTERS. 101 (3). STATE OF SUPERCONDUCTORS. 1965. SOVIET PHYSICS
3. Ikeda, . Impurity-induced broadening of the transition JETP-USSR. 20 (3): 762. to a Fulde-Ferrell-Larkin-Ovchinnikov phase. 2010.
4. Bianchi, A., R. Movshovich, C. Capan, P. G. Pagliuso, PHYSICAL REVIEW B. 81 (6). and J. L. Sarrao. Possible Fulde-Ferrell-Larkin-Ovchin-
5. Miyake, . Theory for Coupled SDW and Superconduct- nikov superconducting state in CeCoIn5. 2003. PHYSI- ing Order in FFLO State of CeCoIn5. 2008. JOURNAL OF CAL REVIEW LETTERS. 91 (18). THE PHYSICAL SOCIETY OF JAPAN. 77 (12).
6. Young, B. -., R. R. Urbano, N. J. Curro, J. D. Thompson,
7. Yanase, , and Sigrist. Antiferromagnetic order in the J. L. Sarrao, A. B. Vorontsov, and M. J. Graf. Microscopic FFLO state. 2009. In 25TH INTERNATIONAL CONFER- evidence for field-induced magnetism in CeCoIn5. ENCE ON LOW TEMPERATURE PHYSICS (LT25), PT 5A.
8. PHYSICAL REVIEW LETTERS. 98 (3). Vol. 1505a Edition.
9. Kenzelmann, , Straessle, Niedermayer, Sigrist, Pad-
10. Ikeda, , Hatakeyama, and Aoyama. Antiferromagnetic manabhan, Zolliker, A. D. Bianchi, Movshovich, E. D. ordering induced by paramagnetic depairing in uncon- Bauer, J. L. Sarrao, and J. D. Thompson. Coupled su- ventional superconductors. 2010. PHYSICAL REVIEW B. perconducting and magnetic order in CeCoIn(5). 2008. 82 (6). SCIENCE. 321 (5896): 1652.
11. Agterberg, D. F., Sigrist, and Tsunetsugu. Order Param-
12. Kenzelmann, , Gerber, Egetenmeyer, J. L. Gavilano, T. eter and Vortices in the Superconducting Q Phase of h. Straessle, A. D. Bianchi, Ressouche, Movshovich, E. CeCoIn5. 2009. PHYSICAL REVIEW LETTERS. 102 (20). D. Bauer, J. L. Sarrao, and J. D. Thompson. Evidence for a Magnetically Driven Superconducting Q Phase of
13. Aperis, , Varelogiannis, P. B. Littlewood, and B. D. CeCoIn5. 2010. PHYSICAL REVIEW LETTERS. 104 (12). Simons. Magnetic Properties of an Antiferromagnetic d-Wave Singlet and pi-Triplet Superconductor. 2009.
14. Gannon, W. J., W. P. Halperin, Rastovski, M. R. Eskild- JOURNAL OF SUPERCONDUCTIVITY AND NOVEL MAG- sen, Dai, and Stunault. Magnetization in the supercon- NETISM. 22 (2): 115. ducting state of UPt3 from polarized neutron diffrac- tion. 2012. PHYSICAL REVIEW B. 86 (10).
15. Suzuki, K. M., Ichioka, and Machida. Theory of an in- herent spin-density-wave instability due to vortices in
16. Si, Q. M., S. Rabello, K. Ingersent, and J. L. Smith. Lo- superconductors with strong Pauli effects. 2011. PHYSI- cally critical quantum phase transitions in strongly cor- CAL REVIEW B. 83 (14). related metals. 2001. NATURE. 413 (6858): 804.
17. Bianchi, A. D., Kenzelmann, DeBeer-Schmitt, J. S.
18. Koutroulakis, , M. D. Stewart Jr., V. F. Mitrovic, Horvatic, White, E. M. Forgan, Mesot, Zolliker, Kohlbrecher, Berthier, Lapertot, and Flouquet. Field Evolution of Movshovich, E. D. Bauer, J. L. Sarrao, Fisk, Petrovic, and Coexisting Superconducting and Magnetic Orders in M. R. Eskildsen. Superconducting vortices in CeCoIn5: CeCoIn5. 2010. PHYSICAL REVIEW LETTERS. 104 (8). Toward the Pauli-limiting field. 2008. SCIENCE. 319
19. Kakuyanagi, K., M. Saitoh, K. Kumagai, S. Takashima, (5860): 177. M. Nohara, H. Takagi, and Y. Matsuda. Texture in the
20. Das, , J. S. White, A. T. Holmes, Gerber, E. M. Forgan, superconducting order parameter of CeCoIn5 revealed A. D. Bianchi, Kenzelmann, Zolliker, J. L. Gavilano, E. D. 440

Page 441

Bauer, J. L. Sarrao, Petrovic, and M. R. Eskildsen. Vortex Fermi surface topology in doped 122 iron pnictides. Lattice Studies in CeCoIn5 with H perpendicular to c. 2013. PHYSICAL REVIEW B. 88 (21). 2012. PHYSICAL REVIEW LETTERS. 108 (8). Yang, J., I. Yang, D. Wulferding, S. Chung, R. Movshovich, H. W. Yeom, K. S. Kim, and J. Kim. Confinement of projec- 22. Vorontsov, A. B., J. A. Sauls, and M. J. Graf. Phase dia- tive vortices via an extra dimension. Nature Physics. gram and spectroscopy of Fulde-Ferrell-Larkin-Ovchin- nikov states of two-dimensional d-wave superconduc- tors. 2005. PHYSICAL REVIEW B. 72 (18). 23. Fogelstrom, , W. K. Park, L. H. Greene, Goll, and M. J. Graf. Point-contact spectroscopy in heavy-fermion superconductors. 2010. PHYSICAL REVIEW B. 82 (1). 24. Park, W. K., and L. H. Greene. Andreev reflection and order parameter symmetry in heavy-fermion super- conductors: the case of CeCoIn(5). 2009. JOURNAL OF PHYSICS-CONDENSED MATTER. 21 (10). Publications Chen, H., Y. Y. Tai, C. S. Ting, M. J. Graf, D. Jianhui, and J. Zhu. Disorder effects in multiorbital s(+/-)-wave super- conductors: Implications for Zn-doped BaFe2As2 com- pounds. 2013. PHYSICAL REVIEW B. 88 (18): 184501. Fogelstrom, , M. J. Graf, V. A. Sidorov, X. i. n. Lu, E. D. Bauer, and J. D. Thompson. Two-channel point-contact tun- neling theory of superconductors. 2014. PHYSICAL RE- VIEW B. 90 (10). Jeong, , Yang, Yang, O. E. Ayala-Valenzuela, Wulferding, J. -. Zhou, J. B. Goodenough, de Lozanne, J. F. Mitchell, Leon, Movshovich, Y. H. Jeong, H. W. Yeom, and Kim. Magnetic domain tuning and the emergence of bubble domains in the bilayer manganite La2-2xSr1+2xMn2O7 (x=0.32). 2015. PHYSICAL REVIEW B. 92 (5). Kim, , Haberkorn, Kim, Civale, P. C. Dowden, and Movshovich. Ferromagnetic bubble clusters in Y0.67Ca0.33MnO3 thin films. 2013. APPLIED PHYSICS LETTERS. 102 (19). Kim, J., N. Haberkorn, K. Gofryk, M. J. Graf, L. Civale, F. Ron- ning, A. S. Sefat, and R. Movshovich. Intrinsic super- conducting properties and vortex dynamics in heavily overdoped Ba(Fe0.86Co0.14)2As2 single crystal. 2015. Solid State Communications. 201: 20. Leon-Brito, N., E. D. Bauer, F. Ronning, J. D. Thompson, and R. Movshovich. Study of magnetic microstructure of Fe3GeTe2 using magnetic force microscopy. Apllied Physics Letters. Leon-Brito, N., E. D. Bauer, F. Ronning, J. D. Thompson, and R. Movshovich. Surface magnetometry of Fe3GeTe2 via magnetic force microscopy. Journal of Physics: Con- densed Matter. Pan, , Li, Tai, M. J. Graf, Zhu, and C. S. Ting. Evolution of the 441

Page 442

Materials for the Future Exploratory Research Final Report Excited State Quantum Interactions in Carbon Nanotubes Stephen K. Doorn 20130309ER Abstract opens the way for new quantum physics to be explored This project was motivated by two unexpected discov- in these systems. In particular, Raman excitation pro- eries revealed in our work at LANL. The first was the files (REPs) of pure, single-nanotube chirality samples, discovery of strong asymmetries in the Raman excita- demonstrated for the first time two peaks in the REPs. tion profiles of carbon nanotubes (CNTs), when previous These corresponded to the long-expected incoming and expectations were that no such asymmetries should outgoing resonances. Surprisingly, however, the outgo- exist. The second was the discovery of strong quantum ing resonance was significantly weaker than the incom- interference effects occurring between closely spaced ing (Figure 1), as a consequence of non-Condon effects. electronic levels of CNTs. Our project goals were to While these fundamental behaviors are fascinating by study these behaviors in depth to develop an under- themselves, they have the potential to alter relaxation standing of the origins and CNT structural (or (n,m)) pathways and dynamics underlying anticipated appli- dependence of these behaviors. We also wished to cations in photonics, optoelectronics and energy har- explore the consequences these behaviors may have for vesting. It is therefore important to understand both other photophysical responses of CNTs. As a result of the origins and consequences of these phenomena in this project, we significantly advance the state-of-the-art CNTs. Our goal was to understand how the interplay of in CNT separations for isolation of specific (n,m) species. electronic and vibrational structure in CNTs gives rise to We established the metallicity, excited-state, and diam- non-Condon and interference behaviors and determine eter dependences of the REP asymmetry behavior. We what consequences they hold for observed photophysi- developed a robust model for the asymmetries showing cal responses. they originate in non-Condon behaviors as a result of We focused on three objectives: state-mixing of different exciton levels. We revealed the (n,m) dependence of several new vibrational modes of

1. Non-Condon Behaviors were probed for multiple CNTs. And finally, we probed and accurately modeled the nanotube structures, vibrational modes, and excited excited state dynamics of model ring compounds that states using Raman and modeled with quantum chemi- served as molecular mimics of CNTs. The implications of cal theory and condensed matter approaches to gain an these results are far reaching. They further demonstrate understanding of their origins. the importance of access to pure (n,m) samples for revealing new CNT behaviors. Our understanding of fun- 2) We advanced the state of the art in nanotube puri- damental optical processes in the nanotubes is greatly fication to access more tube structures and to enable strengthened. We have also significantly strengthened extensive probing of other phonon mode behaviors to LANL capabilities for theory and experimental probing contrast to those exhibiting the non-Condon response. of photonic behaviors at the nanoscale. Together, these results form a robust basis for future efforts targeting 3) Relaxation Pathways and Dynamics were probed in emerging applications of nanotube optics. model systems using time-resolved spectroscopies to learn the fate of optical excitations as determined by Background and Research Objectives phonon-mediated electronic state interactions. Our recent first-time demonstration of quantum inter- Each of these focal areas represents pioneering stud- ference [1] and violation of the Condon approximation ies that open significant new areas of CNT physics for [2] in carbon nanotubes (CNTs) overturns long-held further investigation. No other molecular or condensed- assumptions regarding their optical response and 442

Page 443

matter system offers the ability to study non-Condon and new unambiguous information on (n,m) dependent fre- interference behaviors in such a controlled and systematic quencies of additional modes not previously accessible. way and the information gained impacts our understand- Origins, Transition Level, and Structural Dependences of ing of photophysical behaviors of broad classes of optical Non-Condon Behavior materials and serves as a bridge between molecular and With our advances in separations we had the opportunity condensed matter systems. to significantly expand our Raman excitation profile (REP) data beyond the original work on semiconductors [2]. We produced REPs for four so-called “armchair” metallic struc- tures ((8,8), (7,7), (6,6), (5,5); diameter of 0.68 to 1.09 nm) [5]. In all cases significant REP asymmetry was demonstrat- ed (Figure 2), extending our observed trends to metallic tubes and further cementing our finding that the behavior is intrinsic and is a general behavior across all nanotube types (semiconducting and metallic species). The high quality of the new metallic REPs, paired with previous semiconducting results, allowed us to definitively demon- strate the asymmetry weakens as diameter decreases. Figure 1. Example Raman excitation profiles of the LO mode for several semiconducing carbon nanotube structures. REPs show two peaks arising from incoming and outgoing resonances. Clear asymmetry is seen, with the outgoing resonance peak always being weaker. Scientific Approach and Accomplishments Generation of targeted nanotube structures To meet our goals, advances in nanotube separations needed to be made. These provided specific new chirali- ties while efficiently generating large quantities (up to mL scale) of enriched material. We established a new tech- nique based on aqueous two-phase separation concepts that allow rapid separation and efficient scale-up of Figure 2. Raman excitation profiles of the TO mode for the (5,5), targeted chiralities [3]. We produced up to 10+ mL scale (6,6), (7,7), and (8,8) armchair metallic nanotube structures. All volumes of new armchair metallic species ((8,8), (7,7), REPs show an asymmetry similar to that found for semiconduct- (6,6), and (5,5)), for metallic studies, advances in large-di- ing nanotubes. ameter separations (a particularly difficult range intended to give (10,5), (10,2), (12,4) for pump probe and interfer- Through a collaboration with Eduardo Barros of MIT, we ence studies) [4], novel non-armchair metallics ((10,4), also generated a theoretical description of the REP be- (7,4)) for interference work and (7,5) and (7,6) for our work haviors that bridged the molecular picture of the result on isotopically substituted tubes. Isotopically enriched being a non-Condon effect with an approach based on (10,5) tubes produced via a selective polymer suspension condensed matter theory. We contrasted our results with technique were also available. Finally, we also pushed the a competing condensed-matter theory [6] that suggested ability to generate new small-d and intermediate diameter the asymmetries arise solely as a consequence of features chiralities (including (5,4), (10,0), and (11,0) structures) at of the nanotube density of states. While we showed this the extremes of the diameter and chiral angle distributions approach can describe our diameter-dependent trends, it of our samples. is inadequate to describe the magnitude of the asymme- tries. As an alternative we have developed a description Such a broad range of structures is an unprecedented that combines molecular concepts with condensed matter range of pure chiralities. These were essential for generat- that shows the asymmetries arise as the result of mixing of ing the high quality Raman excitation profiles in our stud- transition density between allowed and forbidden exciton ies. They allowed probing the limits of chirality-dependent states. Such a result is a type of non-Condon effect, thus quantum behaviors. They also were critical for establishing confirming the originally suggested basis for the observed 443

Page 444

asymmetries. The flexible approach is also general to semi- to the energy mismatch between the two mixing exciton conductors (see below). Significantly, the approach is also states, the stronger the asymmetry will be. This accurately testable and provides a route to understand the varying describes the trends we observe for phonon behaviors. effects observed for different transition levels and for dif- The results also suggest that isotopic substitution should ferent phonon modes. modulate the observed asymmetries. This prediction is be- ing tested in the last weeks of this project. We are finish- Transition Energy and Phonon Dependence of REP ing REP measurements on C13 isotopically substituted Asymmetry in Semiconductors nanotubes of the (7,5), (7,6), and (10,5) structures. Pre- We have established the degree of REP asymmetry in semi- liminary data on the (10,5) is already showing the expected conductors for a wide range of chiralities across the first effects, thus supporting the new theoretical framework we 4 exciton excited states and for the radial breathing mode have developed. (RBM), G- (or TO) and G+ (or LO) modes. For transition level dependence, we find that the asymmetry in the low- New Phonon Modes est transition is weakened for the G modes, and disappears Access to significant numbers of new pure nanotube chi- for the RBM (Figure 3). We also always find the TO mode ralities has allowed us to measure unambiguously for the displays a stronger asymmetry than is found for the LO first time the chirality dependence of metallic TO modes mode [2]. We have recently added data for our smallest and the so-called “M mode” of the semiconductors. These diameter nanotube (the (5,4)) to this set and find asymme- are providing new insights into coupling to the electronic try for all modes nearly disappears, which adds significant structure of the tubes and also inform back to the non- weight to our assertion that smaller diameter tubes give Condon effects. weaker effects. The armchair work has allowed us to obtain the first ever detailed frequency values for the TO phonon mode (Fig- ure 4) over 6 different structures ((5,5) through (10,10)). Frequencies show a sharp decrease to smaller diameter tubes and exhibit anomalous behavior of the phonon dispersion. The behavior can be described by introducing a new phonon scattering mechanism in nanotubes. We have confirmed this mechanistic description by implement- ing electrochemical doping measurements on pure (6,6) structures in our armchair series. Our results now resolve a controversy in the theoretical community over how to de- scribe electron-phonon coupling in metallic nanotubes [7]. Figure 3. A comparison of how REP asymmetry for the G-mode evolves on going from E11 to E22 to E33. E11 shows the weakest asymmetry. A theoretical description for understanding these trends had been lacking, but as an extension of the work on armchair structures we have demonstrated the generality of the condensed matter state-mixing approach to non- Condon modeling to include the semiconductor structures. Of significance, we demonstrate the excitonic nature of the optical transitions plays a significant role in the origins of the behavior. State-mixing surprisingly evolves from the strongest allowed excitonic transitions. The modeling results converge on a description that involves mixing of weakly allowed states associated with phonon sidebands. Figure 4. Raman spectra of the TO mode for the labeled The theory is beginning to have predictive power, sug- armchair metallic nanotube structures. Frequency downshifts gesting that the closer the phonon mode is in frequency significantly as nanotube diameter decreases. 444

Page 445

We also obtained M-mode frequencies for 12 separate lished clear pictures of phonon-mediated exciton behav- tube chiralities. Again a strong dependence on diameter iors in CPPs. The quantum chemical modeling provides an is found, but in a unique manner. Within a given (2n+m) exact description of the pump probe results. As a comple- family of tubes, the frequencies deviate from an “ideal” ment to the dynamics studies and the armchair nanotube armchair behavior in much the same way that nanotube studies, the effort has also pursued Raman studies of the transition energies do! This behavior has no other paral- CPPs and has revealed a clear size dependence in many lels for other nanotube phonon modes. Additionally, we of the vibrational modes, analogous to what is commonly find strong chirality dependences in the intensities. Most found for many of the nanotube phonon modes. significantly, for tubes like the (11,0), for which we are also able to obtain M-Mode REPs, we find the REP shape is completely unlike the excitonically-determined line shapes found for the RBMs and G-modes (see Figures 1-3 as representative examples). The M-mode REP maxima do not match up with those found for the G and RBM modes. Tied to this behavior is that near resonance, the M-mode peaks themselves show two clear components. This con- trasting behavior to the RBM and G modes suggests that the M-mode behavior is completely determined by the nanotube electron and phonon dispersions, as opposed Figure 5. Example pump-probe results for the 12-ring cyclo- to coupling to exciton levels. This provides an explanation paraphenylene structure 12 [CPP]. Low (left) and high (right) for lack of asymmetry in the M-mode REPs. As determined time resolutions of several spectral regions are shown. from our results on the G-mode REPs, such non-Condon effects are driven by coupling to excitons. In the absence In summary, this ER effort has successfully met most of of such coupling (as apparently for the M-mode) no asym- the goals set forth at its beginning. We now understand metry will be observed! the excitonic origins of the non-Condon effects observed in their Raman excitation profiles. We have established Excited-State Dynamics in Model Molecular Surrogates significant new capability for nanotube processing, prob- Unfortunately, our separations efforts were not able to ing their behaviors optically, and understanding their provide the levels of (10,5) and (10,2) structures we re- photophysics with new theory efforts. Finally, we continue quired to pursue studies of the effects of quantum inter- to highlight the importance of pure chirality nanotube ference on excited-state dynamics in nanotubes. As an samples for revealing new and unexpected nanotube phys- alternative, we pursued studies of dynamics for cyclopara- ics. Overall, the effort to date has been highly productive, phenylenes, the simplest structural unit of armchair car- with nine papers published, one more submitted, and five bon nanotubes, and which we find also display very strong others in preparation. non-Condon effects. In these studies, femtosecond pump probe measurements were paired with quantum chemical Impact on National Missions theory modeling of exciton behaviors. Our pump probe Results of this project will have a direct bearing on devel- measurements (see Figure 5 as an example) are the first oping optical and electronic properties of carbon nano- ever probes of the early time dynamics in these systems. tubes for photonic and energy harvesting applications and We are able to resolve the growth of optically excited may also contribute to sensing and spectral tagging appli- states and initial decay to intermediate states. Most im- cations. As a result, this work will have direct relevance to portantly, we are able to spectrally and temporally resolve the mission of the DOE-BES funded Center for Integrated coupling of the electronic degrees of freedom to phonon- Nanotechnologies and the potential applications will be driven localization of excitons. of interest to agencies including NIH, DOE, DHS, and DOD with potential impact on threat reduction and renewable Our time-dependent density functional theory study pro- energy missions. Our efforts have also driven development vides a simple and conceptually appealing physical picture and understanding of fundamental surface chemistry of explaining all experimentally observed trends in optical low-dimensional properties in this family of molecules [8]. We found that the commonly used Condon approximation is invalid in this materials. Additionally, the quantum behaviors we studied molecular family and breaks optical selection rules, making are found in broad classes of materials, but only carbon these materials superior fluorophores. Using LANL-unique nanotubes allow the detailed study that we pursued. As non-adiabatic excited state dynamics modeling we estab- a result, the nanotubes serve as ideal model systems for 445

Page 446

obtaining a basic understanding of several novel quantum Publications behaviors. As such, their study supports DOE goals for Adamska, , Nayyar, Chen, A. K. Swan, Oldani, Fernandez- expanding fundamental understanding of functional nano- Alberti, M. R. , et.al. Self-Trapping of Excitons, Violation materials. of Condon Approximation, and Efficient Fluorescence in Conjugated Cycloparaphenylenes. 2014. NANO LET- References TERS. 14 (11): 6539.

1. Duque, J. G., Chen, A. K. Swan, A. P. Shreve, Kilina, Berciaud, S., X. Li, H. Htoon, L. E. Brus, S. K. Doorn, and T. Tretiak, Tu, Zheng, and S. K. Doorn. Violation of the F. Heinz. Intrinsic line shape of the Raman 2D mode in Condon Approximation in Semiconducting Carbon freestanding graphene monolayers. 2013. Nano Let- Nanotubes. 2011. ACS NANO. 5 (6): 5233. ters. 13: 3517.
2. Duque, J. G., Telg, Chen, A. K. Swan, A. P. Shreve, Tu, Chen, H., M. R. Golder, F. Wang, S. K. Doorn, R. Jasti, Zheng, and S. K. Doorn. Quantum Interference be- S. Tretiak, and A. K. Swan. Raman-active modes of tween the Third and Fourth Exciton States in Semicon- even-numbered cycloparaphenylenes: comparisons ducting Carbon Nanotubes Using Resonance Raman between experiment and DFT calculations with group theory arguments. 2015. Journal of Physical Chemistry Spectroscopy. 2012. PHYSICAL REVIEW LETTERS. 108 C. 119: 2879. (11). Ciesielski, R., V. Giegold, N. F. Hartmann, E. H. et.al, Elec-
3. Fagan, J. A., C. Y. Khripin, C. A. S. Batista, J. R. Simpson, tronic and vibrational coherences in single semicon- E. H. Haroz, A. R. H. Walker, and Zheng. Isolation of ducting carbon nanotubes probed by femtosecond Specific Small-Diameter Single-Wall Carbon Nanotube pulse-shaping microscopy at room temperature. Na- Species via Aqueous Two-Phase Extraction. 2014. AD- ture Nanotechnology. VANCED MATERIALS. 26 (18): 2800. Fagan, J. A., C. Y. Khripin, C. A. S. et.al, Isolation of Specific
4. Fagan, J. A., E. H. Haroz, Ihly, H. u. i. Gui, J. L. Black- Small-Diameter Single-Wall Carbon Nanotube Species burn, J. R. Simpson, Lam, A. R. H. Walker, S. K. Doorn, via Aqueous Two-Phase Extraction. 2014. ADVANCED and Zheng. Isolation of > 1 nm Diameter Single-Wall MATERIALS. 26 (18): 2800. Carbon Nanotube Species Using Aqueous Two-Phase Fagan, J. A., E. H. Haroz, Ihly, H. u. i. Gui, J. L. Blackburn, Extraction. 2015. ACS NANO. 9 (5): 5377. J. R. Simpson, Lam, A. R. H. Walker, S. K. Doorn, and Zheng. Isolation of > 1 nm Diameter Single-Wall Car-
5. Haroz, E. H., J. G. Duque, E. B. Barros, Telg, J. R. Simp- bon Nanotube Species Using Aqueous Two-Phase Ex- son, A. R. H. Walker, C. Y. Khripin, J. A. Fagan, Tu, traction. 2015. ACS NANO. 9 (5): 5377. Zheng, Kono, and S. K. Doorn. Asymmetric excitation profiles in the resonance Raman response of armchair Haroz, E. H., J. G. Duque, J. Simpson, E. Barros, H. Telg, carbon nanotubes. 2015. PHYSICAL REVIEW B. 91 (20). A. Height-Walker, X. Tu, M. Zheng, J. Kono, and S. K. Doorn. Asymmetric excitation profiles in the resonance
6. Moura, L. G., M. V. O. Moutinho, Venezuela, Fantini, Raman response of armchair carbon nanotubes. 2015. Righi, M. S. Strano, and M. A. Pimenta. Raman excita- Physical Review B. 91: 205446. tion profile of the G band in single-chirality carbon Haroz, E. H., J. G. Duque, X. Tu, M. Zheng, A. R. Hight Walk- nanotubes. 2014. PHYSICAL REVIEW B. 89 (3). er, R. H. Hauge, S. K. Doorn, and J. Kono. Fundamental
7. Telg, , E. H. Haroz, J. G. Duque, Tu, C. Y. Khripin, J. A. optical processes in armchair carbon nanotubes. 2013. Nanoscale. 5: 1411. Fagan, Zheng, Kono, and S. K. Doorn. Diameter depen- dence of TO phonon frequencies and the Kohn anom- Lim, , A. R. T. Nugraha, Cho, Noh, Yoon, Liu, Kim, Telg, E. H. aly in armchair single-wall carbon nanotubes. 2014. Haroz, G. D. Sanders, Baik, Kataura, S. K. Doorn, C. J. PHYSICAL REVIEW B. 90 (24). Stanton, Saito, Kono, and Joo. Ultrafast Generation of Fundamental and Multiple-Order Phonon Excitations
8. Adamska, , Nayyar, Chen, A. K. Swan, Oldani, Fernan- in Highly Enriched (6,5) Single-Wall Carbon Nanotubes. dez-Alberti, M. R. Golder, Jasti, S. K. Doorn, and Tretiak.
9. NANO LETTERS. 14 (3): 1426. Self-Trapping of Excitons, Violation of Condon Ap- proximation, and Efficient Fluorescence in Conjugated Telg, H., E. H. Haroz, J. G. Duque, X. Tu, C. Khripin, J. Fagan, Cycloparaphenylenes. 2014. NANO LETTERS. 14 (11): M. Zheng, J. Kono, and S. K. Doorn. Diameter depen- dence of TO phonon frequencies and the Kohn anom-

aly in armchair single-wall carbon nanotubes. 2014. Physical Review B. 90: 245422. 446

Page 447

Materials for the Future Exploratory Research Final Report Enhancing Thermoelectric Properties of Topological Insulators through Nanostructuring Nikolai Sinitsyn 20130348ER Abstract The prediction of TIs in 2005 had quickly evolved to the This project was motivated by the discovery of new discovery that some of the most important thermoelec- materials, called Topological Insulators (TI), which dem- tric materials, Bi2Se3 and Bi2Te3, belong to this class. onstrate a new state of electrons that is topologically Due to specific topological properties of the electronic distinct from the conventional band insulators or metals. band structure, these insulators possess gapless (metal- One of the manifestations of the topological proper- lic) modes at the surfaces. Unlike usual 2D electron sys- ties is the appearance of conducting electron states, on tems, such surface states have a Dirac spectrum. When the surface of material, which propagate throughout geometry is confined to quasi-1D, the Dirac Hamiltonian the sample without elastic scattering. We believe that completely forbids the electron backward scatterings, nanostructuring of TIs can substantially increase the role thus opening up ballistic channels. of these states in charge and thermal transport to the By increasing the figure of merit of materials, a wide level of obtaining a device structure with record high range of new energy saving technologies is expected to thermoelectric characteristics. We develop a theory of become possible. For example, spinning components of thermoelectric effects in thin films and nanowires made refrigerators will be replaced by energy efficient, com- of TIs. This goal will require an understanding the role of pact cheap and silent solid-state devices. Thermoelectric disorder and quantum confinement on the topologically generators and refrigerators are mainly used in applica- protected surface states. Complimentary experimental tions requiring reliability, including the power generation efforts were focused on measurements of thermal trans- sources for deep space missions and cooling of comput- port in topological insulator nanowires using state of the er central processing units. The environmental concerns art CINT Discovery platform. This provided a selective and demands to explore alternative energy sources (e.g., probe for conflicting processes determining the figure generation of electric power from waste heat) have been of merit. Theoretical results were benchmarked against motivating the search for new high-ZT thermoelectric experimental data and used to provide a guideline for materials. We explored two general strategies for further subsequent experimental studies. Our studies reveal increasing ZT: several effects in TI-nanowires that will likely lead to new advances in thermoelectricity.

1. Make electrons as ballistic as possible, to increase σS2. Background and Research Objectives
2. Suppress the phonon heat conductivity, κ. The effectiveness of a thermoelectric material is de- termined by the ratio of electrical (σ) to thermal (κ) The above two mechanisms are usually conflicting. conductivity and its thermopower (Seebeck coefficient, However, in TI naonowires, this is not the case for the S), with a material’s figure of merit given by ZT=σS2T/κ. following reason. Localization of phonons in a confined, For high performance one wants large thermopower, quasi-1D, geometry will lead to the reduction of κ, as in high electrical conductivity and low thermal conductiv- other materials, but the product of electronic conductiv- ity, posing an optimization problem involving conflicting ity and thermopower, σS2, should actually, substantially materials properties. While there are no fundamental increase in Tis due to the protected ballistic character of theoretical or thermodynamic limits to increases, ZT~1 Dirac electrons in quasi-1D geometry. This distinguishes has remained the upper limit for materials in use for the TI nanowires from other nanostructurs because, in last 30 years. standard quasi-1D materials, transport is suppressed by 447

Page 448

elastic scattering both for phonons and electrons. that also strongly supports LANL’s long-term development of new functional materials capabilities. We built funda- Scientific Approach and Accomplishments mental understanding of topological effects in heat trans- Our team fabricated the nanowire from a topological in- port. sulator SnTe and performed measurements of the trans- port characteristics, including the thermoelectric figure of Publications merit (Xu et al, Nanoscale, 2015). We achieved particu- Lin, J., and N. A. Sinitsyn. Three level Landau-Zener-Cou- larly interesting progress with relatively thick nanowires lomb model: Exact transition probabilities in the three- state Landau-Zener-Coulomb model. 2014. Journal of (1000-100nm) at which specific topological physics is not Physics A: Mathematical Theoretical. 47: 015301. expected to play big role. However, experiments revealed unexpected phenomena – the growth of the Seebeck coef- Lin, J., and N. A. Sinitsyn. The model of a level crossing with ficient by a factor 4 (and the increase of figure of merit by a Coulomb band: exact probabilities of nonadiabatic an order of magnitude). It is a surprising and very useful transitions. 2014. Journal of Physics A: Mathematical effect, meaning that there are additional factors that can Theoretical. 47: 175301. improve thermoelectric properties of those materials. Ren, J., and N. A. Sinitsyn. Braid Group and Topological Following studies of SnTe we turned to the nanowires Phase Transitions in Nonequilibrium Stochastic Dynam- made of lead telluride (PbTe) and its alloy compounds, ics. 2013. Physical Review E. 87: 050101(R) . which are the most promising thermoelectric materials Roslyak, O., and A. Piryatisnki. Sonoluminescence of car- for electric power generation (Xu et al, Nanoscale, sub- bon nanotubes. 2013. Unpublished Document. mitted). We developed an approach to grow high quality single-crystalline PbTe and PbSnTe nanowires by a vapor Sinitsyn, N. A.. Landau-Zener Transitions in Chains. 2013. transport approach and performed the thermoelectric Physical Review A. 87: 032701. studies of these individual nanowires. We measured the Sinitsyn, N. A.. Nonadiabatic transitions in exactly solv- electrical conductivity, thermopower and thermal conduc- able quantum mechanical multi-channel model: role tivity of the same individual nanowires to determine their of level curvature and counterintuitive behavior. 2013. thermoelectric figure of merit ZTs. In comparison to bulk Physical Review Letters. 110: 150603. samples, the PbSnTe nanowires showed both improved thermopower and suppressed thermal conductivity, lead- Tse, W., A. Saxena, D. Smith, and N. A. Sinitsyn. Spin and ing to an enhancement of ZT by a factor of ~ 10. The ZTs of Valley Noise in Two-Dimensional Dirac Materials. 2014. the PbSnTe nanowires were also several times higher than Physical Review Letters. 113: 046602. that of PbTe nanowires reported in the literature. Our work Xu, E. Z., Z. Li, J. A. Martinez, N. A. Sinitsyn, N. Li, B. Swartz- clearly demonstrated that nanostructuring, in combination entruber, and J. A. Hollingsworth. Diameter dependent with alloying, is an effective approach to enhancing ther- thermoelectric properties of individual SnTe nanow- moelectric properties of topological insulators. ires. 2015. Nanoscale. 7: 2869. Our team has achieved considerable theoretical progress Xu, E. Z., Z. Li, J. Aviles, N. Li, B. Swartzentruber, N. Sinitsyn, in understanding transport properties in Dirac semiconduc- H. Htoon, J. Wang, and S. X. Zhang. Vapor transport tors, which class includes topological insulators. The PI has growth and enhanced thermoelectric properties of published a series of articles in Phys. Rev. Lett., Phys. Rev. single-crystalline PbSnTe nanowires. Nanoscale. B, J. Phys. A, and contributed to a Nature Physics article Yang, L., N. A. Sinitsyn, W. Chen, J. Yuan, J. Zhang, J. Lou, with the discovery of unusual behavior of spin relaxation and S. A. Crooker. Long-lived nanosecond spin relax- in transition metal dicalcogenides (L. Yang, et al, Nature ation and spin coherence of electrons in monolayer Physics, 2015). MoS2 and WS2. 2015. Nature Physics . : doi:10.1038/ nphys3419. Impact on National Missions The project supported the Threat Reduction and Energy Security Mission, and advanced Science and Technol- ogy Grand Challenges in Materials: Discovery Science to Strategic Applications (Emerging Phenomena) and Energy & Earth Systems (Concepts & Materials for Clean Energy). It offered a unique opportunity to exploit novel materials

• topological insulators - for energy applications, an area 448

Page 449

Materials for the Future Exploratory Research Final Report “Upscaling” Nanoscale Thermoelectrics: The Meso-macroscale Design Challenge for Real-World Energy Needs Jennifer A. Hollingsworth 20130350ER Abstract address each of the above problems and, therefore, to Next-generation thermoelectrics (TEs) have the potential have a crosscutting and significant impact on all aspects to dramatically impact a wide range of energy harvest- of Global Energy Security and Climate Change. In addi- ing and utilization schemes underpinning a sustainable tion, TEs can complement other technologies, e.g., solar and secure energy supply for the 21st century. To do cell/TE hybrid systems, where the TE component boosts so, conventional TE technologies must be transformed energy output by converting solar-generated heat into to achieve unprecedented performance efficiencies. “extra” electricity. By improving the economics of other Nanoscale size effects have been identified as a means nascent technologies, TEs can hasten their progression to realizing the needed advances, and theoretical and to widespread implementation. To date, though, low TE experimental demonstrations of important improve- efficiencies—well below the Carnot efficiency limit— ments in TE parameters have been shown at the level have relegated TEs to important but niche applications in of individual nanoscale structures. Critically, however, remote power generation (e.g., space/deep-sea), remote general strategies are lacking to translate success at the cooling (e.g., ionizing radiation detector), and specialty nanoscale into performance requirements at the meso- applications (e.g., DNA synthesizers), where benefits of macroscale. We address this gap with a stepwise ap- reliability, compact size, and no moving parts trump ef- proach to the integration of optimized TE nanomaterials ficiency & cost considerations. into structures of increasing complexity and scale, estab- Currently, the TE figure-of-merit, ZT, remains at ~1 for lishing a path from the single nanostructure to function- most deployed systems,1 where ZT = S2σ/κ and S = See- al bulk composites capable of macroscale manipulation beck coefficient, σ = electrical conductivity, κ = thermal and device-level performance. conductivity. A ZT of ~2-3 is considered necessary for TEs to significantly penetrate general energy use. Optimizing Background and Research Objectives ZT requires minimizing κ (to prevent short-circuiting the More than half of the energy generated worldwide is thermal gradient) while simultaneously maximizing σ lost as heat. Such ‘waste heat’ can originate from large and S. Referred to as a “combination of opposites,” this point sources (e.g., industrial processes) or smaller dis- challenging materials requirement demands separate tributed sources (e.g., automobiles). Even partial recov- tuning of bandgap, electronic density-of-states (DOS), ery of this lost energy would have a dramatic impact on and electron/ phonon mean free path, given that in the economic and environmental costs associated with conventional bulk materials these parameters are cor- the increasing global appetite for energy. Furthermore, related. “heat management” problems plague the electronics industry and the emerging solid-state lighting industry. Nanotechnology can address the ZT challenge, and dra- In both cases, performance efficiencies and product matic improvements in specific TE parameters and even lifetime are impacted by high operating temperatures. in ZT (from ~1.1 to 2.5) have been demonstrated experi- On-demand/site-specific cooling would eliminate costly, mentally, with ZTs as high as 10 predicted. Semiconduc- entire system cooling approaches and boost system ef- tor NWs are a promising nanomaterial building block for ficiencies. TE applications. Their anisotropic structure provides a path for charge transport, while their width can be ma- TE conversion of heat (temperature gradients) into elec- nipulated in the nanoscale regime to cause reductions in tricity (the ‘Seebeck Effect’) or the converse, TE cooling κ or enhancements in S. In addition, using various in situ (‘Peltier Effect’), is a technology that has the potential to 449

Page 450

techniques, NW surfaces can be “roughened” or internal 2. Understanding and controlling meso-macroscale TE composition can be modulated to form axial heterostruc- properties of novel branched/networked NW con- tures or NW “superlattices.” Each synthetic “knob” pro- structs. vides opportunities to tune κ and S, while retaining σ. 3. Assessing the viability of aerogels as a NW TEs matrix In addition to high ZTs, however, TE device efficiency de- and/or functional spacer layer. pends on maintaining large temperature differences (ΔTs) across the hot and cold junctions. If device thickness is Scientific Approach and Accomplishments small (<<1 mm), then extremely large heat fluxes (>100’s Our original approach is outlined in Figure 1. The strategy kW) are required to maintain large ΔTs. The practical of studying differently sized and shaped lead chalcogenide limitations on engineering a device for such massive heat nanowires (NWs) (of increasing network complexity) as input imply that TE devices need to be significantly thicker building blocks for new thermoelectric (TE) devices was than the longest dimension of typical nanoscale structures largely followed. However, during the execution of the (>>1μm). This requirement constitutes a potential “show project, we determined that TE properties could most stopper” for nano-enabled TEs, and it is this challenge of readily be obtained from thin-film devices and, further, scale that we address here. that even thermal conductivity was accessible if we placed an aerogel (non-thermally conducting) spacer in between In order to use NWs, for example, as the elementary the NW film and the solid support (Figure 2). units of a TE device, they need to be scaled to the above- described device-scale dimensions. It is unlikely that individual NWs will reach the lengths of centimeters with retention of optimized nanoscale features (diameters <10 nm, roughened surfaces, internal superlattice structure, etc.). Therefore, arrays or forests of shorter NWs (microns) will have to be “stacked” or individual NWs interconnected to form 3-D networks. For either scenario, a supporting matrix (array backfill material) or framework is presumably required for bulk-level manipulation and device process- Figure 1. Schematic showing increasing complexity in nanowire ing. This aspect of NW-enabled TEs remains little explored structures and integration with aerogel (thermally insulating, in the literature. For these reasons, we aimed to explore structurally supportive) matrix material. the effects on nanoscale TE properties of NW branch and network formation at progressively increasing levels of complexity, as well as assess the impact of matrix integra- tion on these properties. Thus, the overarching goal of the program was to address the challenge of scale that is uniquely relevant to nano- enabled TEs. In so doing, we discovered an alternative approach to addressing scale issues in TEs, namely, by using non-thermally conductive spacer layers as a means for avoiding “thermal shorting” in TE thin films. We have also addressed an underpinning “knowledge gap” in terms of understanding how the key transport properties—elec- trical and thermal—change across the extremity of length scales from nano/meso- to macroscale. Though the prob- lem of scale is paramount for TEs, the synthetic strategies and new principles of multi-scale transport can be applied Figure 2. (top) Schematic of the approaches developed in this to a wide range of nano-enabled technologies. The three program to either fabricate a nanowire thin film on top of an objectives that were pursued are: aerogel support layer or to fabricate an aerogel-nanowire bulk composite material (with sufficient nanowire density to support

1. Assessing and optimizing nanoscale TE functionality electrical conductivity through the construct). (bottom) Photo- by way of baseline studies in single-NW systems of graphs of nanowire film grown on top of an aerogel layer (left) controlled size and surface roughness. and of an aerogel-nanowire bulk composite (right). 450

Page 451

Without the aerogel spacer, the thick solid support would large electrodes; electron-beam lithography and metal cause rapid thermal shorting of the NW thin film, prevent- deposition to link NWs to electrodes), for which microamp ing meaningful data from being obtained. In this way, we currents were occasionally obtained, confirming reason- were able to determine ZT in direct measurements of all able conductivity for these thin wires. Challenges: the ZT components, without having to fabricate macroscale consistency of single-NW device performance and lifetime structures comprising aerogel-NW composites. Although in air would need to be improved to move this aspect of these thin-film devices became the focus of our research, the work forward. More fruitful, we fabricated thin-film we also developed the synthetic strategies to fabricate devices comprising NW mats, for which we developed a bulk aerogel-NW composites that can potentially be used unique procedure for improving electrical conductivity as large-scale building blocks for traditional TE device that involved multiple surface-ligand exchanges followed constructs. In addition to this experimental modifica- by annealing in forming gas, where anneal temperatures tion, we also expanded our NW materials to include core/ were sufficiently low to avoid particle sintering. With such shell NWs, rods and tetrapods. Therefore, in addition to modified films, we conducted thermoelectric parameter the Publications reported within this document, this ER measurements with Marcelo Jaime (MPA-CMMS). will result in four further publications with project post- doc, Nimai Mishra, as first author: (1) Lead chalcogenide nanowire building blocks for thin-film thermoelectrics: Im- pact of size, shape and interfaces; (2) Bulk thermoelectrics enabled by hybrid semiconductor-aerogel compositing, (3) Controlled branching in lead chalcogenide nanowires, and (4) Engineering Auger recombination in core/shell rods and tetrapods. Significant results were obtained in this program in both theory and experiment thrusts. Examples are provided below for each Objective Objective 1: We completed a systematic study of PbS, PbSSe, and PbSe nanowire (NW) synthesis by catalyzed solution-liquid-solid (SLS) growth, investigating numerous synthetic variables to obtain size, surface, and branching- controlled NW growth. The variables were: bismuth (Bi) nanoparticle catalyst size (5-38 nm diameters); tempera- ture (175-320 oC): purity of coordinating-ligand (e.g., 90% trioctylphosphine versus 97%; 90% trioctylphosphine oxide versus 99%); and precursor (dual-source versus single source). In support of these studies, single-source precur- sors for PbS and PbSe were synthesized and character- ized by NMR (working with Ryszard Michalczyk, B-11, and Figure 3. (a) Data demonstrating the thermoelectric effect in Koushik Ghosh, MPA-CINT). Overall, we currently favor thin films comprising PbSe nanowires. Voltage (top) changes as use of higher purity ligands and single-source precursors, a clear function of changing film delta-temperature (bottom). and we discovered that temperature is the key variable for (b) First Seebeck coefficient values obtained in this program for tuning surface morphology from rough to smooth and NW differently sized PbSe nanowires, demonstrating the expected impact of quantum confinement for the first time in ensembles of branching from straight to multiply branched, both trend- ultra-thin nanowires. ing with increasing temperature, where our temperature range spans below and above the melting point of bulk Objective 2: Manuscript titled, Enhanced Thermoelectric Bi (271 oC). In fact, we obtained uniquely high yields of Properties of Semiconductor Nanowire Networks, was pre- branched NWs, e.g. 90% for PbSe. pared that describes a novel theoretical approach to map- ping TE “networks” on two-port networks. Here, Piryatinski NW building blocks in hand, we fabricated and tested argues that in contrast to a conventional single-port (i.e., single-NW and thin-film devices, capitalizing on new col- resistor) network model, our model allows for large-scale laborations within and outside LANL. With external col- calculations and also predicts convergent TE figure-of-mer- laborator, Enzhi Xu (Indiana Un.), we fabricated single-NW it, ZT, behavior with increasing number of junctions. Using devices from PbS NWs (dropcast NWs onto substrate with 451

Page 452

this model, the numerical simulations of ZT are performed extremes (cold and hot) and entailed conversion of en- for branched and Cayley-tree NW networks. The simula- ergy types (e.g., heat to electricity) and energy translation tions show that the phonon impedance of the network across interfaces, with direct relevance to complex materi- junctions plays the dominant role in the enhancement of als design and characterization for MaRIE. In addition, the the network performance. Importantly, depending on the technology developed can impact longer term develop- phonon impedance mean value, the ZT enhancement of ment of small-size, rapid-response temperature probes for random networks can vary in the range of 20-150%. To pro- extreme environments, e.g., explosives. vide connection with experiment, the manuscript further demonstrates how the model parameters can be related to References the observables available in TE measurements. 1. Bell, L. E.. Cooling, heating, generating power, and recovering waste heat with thermoelectric systems. Objective 3: We successfully incorporated PbSe NWs 2008. SCIENCE. 321 (5895): 1457. into silica-aerogel monoliths. In addition, we determined optimal NW solution chemistry to realize placement of Publications high quality NW films on top of silica-aerogel supports. Acharya, K. P., T. G. Holesinger, J. A. Crisp, S. A. Ivanov, D. Both types of device components have been demonstrated J. Williams, J. L. Casson, M. Sykora, and J. A. Holling- experimentally (conductivity measurement in former case, sworth. Layer-by-layer fabrication of nanowire sensi- and full TE characterizations in case of latter construct) tized solar cells: geometry-independent integration. to be potentially important approaches to moving nano- 2014. Advanced Functional Materials. 24: 6843–6852. enabled TE forward. Mishra, N., W. Y. Wu, M. S. Bharathi, H. Ramanarayan, J. A. Hollingsworth, Y. W. Zhang, and Y. Chan. Continuous Impact on National Missions shape-tuning of nanotetrapods and their shape-medi- The program has impacted multiple national missions. ated self-assembly behavior. ACS Nano. First, with respect to emerging materials technologies underpinned by advanced materials R&D, the program has Roslyak, O.. Enhanced thermoelectric effect of branched provided important insights in the areas of nanomateri- nano-wire trees. Invited presentation at Fordham Uni- als synthesis and integration, characterization and theory. versity Seminar. (New York, 12 May 2013). These lay the foundation for successful translation of na- Roslyak, O., and A. Piryatinski. Thermoelectric effects in noscale thermoelectric phenomena to the macroscale and disordered branched nano-wires. 2013. In American afford new understanding with respect to how properties Physical Society Annual Meeting 2013. (Baltimore, scale with changes in length scale and complexity (Meso- MD, 18-22 March 2013). , p. BAPS.2013.MAR.W12.14. scale science impact). This aspect of the work contributes College Park, MD: Bulletin of the American Physical directly to DOE Basic Science initiatives and is relevant to Society. the Scientific Discovery and Innovation Mission in the Basic Roslyak, O., and A. Piryatinski. Enhanced thermoelectric Understanding of Materials. We have further demonstrat- properties of semiconductor nanowire networks. 2015. ed novel functional composites comprising nanowire (NW) arXiv.org. : arXiv:1501.03229 . networks in aerogels as well as novel multi-layer structures comprising NW thin films on aerogel spacer layers for ap- Roslyak, O., and A. Piryatinski. Enhanced thermoelectric plications in thermoelectrics. In this way, the work also ad- properties of semiconductor nanowire networks. To dresses challenges to the implementation of nano-enabled appear in Nano Letters. thermoelectrics in real-world device architectures, where Xu, E. Z., Z. Li, J. A. Martinez, N. Sinitsyn, H. Htoon, N. Li, the new “paths forward” provided by the project can have B. Swartzentruber, J. A. Hollingsworth, J. Wang, and S. a transformative impact on Energy Security (Renewable En- X. Zhang. Diameter dependent thermoelectric proper- ergy) and the Environment (Climate and Energy Impact and ties of individual SnTe nanowires. 2015. Nanoscale. 7: even “waste” management), as thermoelectrics tackles 2869. both waste heat utilization and heat management issues. Furthermore, we have established new modeling tools for addressing transport processes from simple to branched NWs and their composites over multiple length-scales, bringing new capabilities for theory-guided materials dis- covery and development and, again, impacting mesoscale science through truly multi-scale theory advances. Lastly, the new composite materials were tested at temperature 452

Page 453

Materials for the Future Exploratory Research Final Report Giving Cold Atoms Weight: Creating Heavy Fermions in Optical Lattices Cristian D. Batista 20130385ER Abstract nature. For this reason, it is very challenging to under- Lanthanide and actinide based compounds are strongly stand and predict its electronic and structural properties. correlated materials of strategic interest for LANL. The Cold atoms offer a unique opportunity of isolating the coexistence of localized f-moments with itinerant elec- essential ingredients that control the properties of these trons in broad bands leads to unusual states of matter, materials and tuning the most relevant parameters in such as unconventional superconductivity and heavy a way that would not be possible by directly studying fermion behavior, which are far from being understood these compounds. in spite of more than 30 years of intense research. These We argue that the closed channel amplitude physics novel states typically appear near quantum critical of the Feshbach resonance gives access to the “Kondo points induced by variations of different external pa- physics” responsible for heavy fermion behavior in inter- rameters such as pressure, magnetic field or chemical metallic compounds. The first goal of this project is to doping. The main bottleneck for characterizing these describe the narrow resonance properties and derive states is that the experimental characterization is limited a model for the effective atom-atom interactions from by many factors. Consequently, there is a need for find- first principles. The second goal is to connect this model ing controllable realizations of this physics that can shed with the one that is traditionally used to describe the light on the basic and universal properties of the novel electronic properties of inter-metallic compounds. By states that emerge near quantum critical points. We pro- guiding cold atoms towards a first realization of heavy pose to address this need by modeling a class of atomic fermion physics, this project will establish the first gases known as Bose-Fermi resonant systems in an one-to-one connection between cold atoms and inter- optical lattice, and identifying regimes of parameters for metallic compounds. This research is relevant for under- which the atomic system can be mapped into a lattice standing the origin of the complex electronic properties of local moments (localized f-electrons) that interact via of actinide-based materials, which are relevant for the exchange with itinerant electrons. Describing and test- nuclear energy security mission. ing this non-trivial mapping requires a description that bridges the Angstrom-sized atomic physics scale with the Scientific Approach and Accomplishments micron-sized optical lattice physics. The aim of this inter- In a magnetically controlled cold atom Feshbach reso- disciplinary approach is to solve a long-standing problem nance, cold alkali atoms are placed in a homogeneous of strongly correlated materials that are relevant for our magnetic field of strength B. When interacting, the LANL’s mission. atoms temporarily rearrange their spins (nuclear spins and valence electron spins) thereby bringing their rela- Background and Research Objectives tive coordinate wave function into a quasi-bound state The possibility of trapping cold atoms in periodic poten- that we will call “molecule”. This process is described by tials opens new horizons for the experimental and theo- a mixed fermion-boson Hamiltonian, in which the boson retical study of correlated matter. We are slowly discov- represents the amplitude of the quasi-bound state. ering that the physical laws that govern the interactions between these atoms are very similar to the ones that We completed the first principles derivation of a model describe the interaction between electrons in actinide for describing this system of cold alkali atoms. We also based inter- metallic compounds. Besides being the ba- mapped the model into a Kondo Lattice Hamiltonian and sic materials for nuclear applications, these compounds studied ways of tuning the parameters of the equivalent are among the most complex correlated materials in 453

Page 454

Kondo Lattice model (KLM) by controlling the parameters bach resonances depends sensitively on the scattering en- of the original system of alkali atoms. The KLM is relevant ergy. As demonstrated by the eight-decade old challenge for lanthanide and actinide based materials because it de- of constructing good effective interactions to describe the scribes the interaction between local moments (localized nucleon-nucleon interactions in low energy nuclear phys- f-electrons) and conduction electrons that form a broad ics, the effective description of energy-dependent interac- band (itinerant electrons moving in s, p or d-bands). tions can be daunting. However, we found that in the limit of sufficiently narrow resonances, the two-body resonanc- We used the Density Matrix Renormalization (DMRG) es fall into two universal categories: potential resonances code, developed during the first year of the project, to and quantum tunneling resonances. We show that the nar- study the phase diagram of the one-dimensional version of row Feshbach resonances fall into the latter category and the model derived from first principles. In agreement with we show that similarity scaling can map results from vastly our expectations, the phase diagram includes quantum different cold atom experiments onto each other. phase transition between a molecular superfluid and a dis- ordered state. The disordered state is the one-dimensional The many-body techniques developed under this project, version of the heavy fermion phase that appears in higher such as the development of a DMRG code, were applied dimensions (the Fermi liquid is unstable in one-dimension). to solve different many-body problems involving compet- In terms of the KLM, the counterpart of the molecular ing interactions and quantum phase transitions [2-5]. The superfluid is a magnetically ordered state in which the local multiple results are described in the manuscripts listed magnetic moments order antiferromagnetically in a plane below. In particular, we found new plateux phases in- perpendicular to an effective applied magnetic field. Along duced by pressure in the so-called “Shastry-Sutherland” this process, we realized that a precondition for obtain- compound SrCu2(BO3)2 [2] and a novel anyonic liquid at ing the Kondo Lattice model physics with cold atoms is to a field-induced quantum critical point [3]. We have also have a strong (hard-core) repulsion between the molecules reproduced the inelastic neutron scattering data of the (closed channel). Unfortunately, there is no obvious way heavy fermion antiferromagnetic compound CeRhIn5 [5]. of producing this interaction in current implementations This work was done in collaboration with experimentalists of atomic gases. However, this obstacle could be removed from MPA Division. We are now applying our many-body in the future by the use of other atomic species or actual (DMRG) techniques to understanding the magnetic field molecules. and temperature phase diagram of this compound. Our preliminary results indicate that the competition between We have recently submitted a manuscript entitled “The Kondo screening and the RKKY interaction leads to non- narrow Feshbach Resonance as a long-delay interac- trivial effects induced by magnetic field, such as a non- tion” [1]. The scattering amplitude of cold atom narrow monotonic behavior of the ordering temperature as a func- Feshbach resonances exhibits a steep dependence upon tion of the magnetic field. scattering energy. We derived the energy dependence of the effective energy-dependent scattering length from the Impact on National Missions scattering description of two coupled channels and find Lanthanide and actinide based compounds are strongly a universal three-parameter description in the narrow correlated materials of strategic interest for LANL and the resonance limit. We test the energy dependence for two DOE complex. We have developed methods for studying specific resonances in the 40K-87Rb interaction complex. and emulating the competing phases of these materials The energy dependence implies a negative effective range under control, i.e., without relying on uncontrolled approx- close to resonance with an absolute value that can exceed imations. As we demonstrated in this project, these unbi- cold atom inter-particle distances. Rather than attributing ased methods can be applied to any correlated electron the large absolute value of the effective range to an effec- material and they are promising for achieving predictive tive interaction-at-a-distance, we point out that the sign power for these classes of complex materials. and magnitude of the effective range are consequences of a long-lasting interaction. We determined the delay-time References and recover the time that the interacting atoms spend in

1. Hazra, J., B. Naduvalath, and E. Timmermans. The nar- the closed channel during the collision. We pointed out row Feshbach Resonance as a long-delay interaction. that the long time spent in the vibrationally excited closed Physical Review B. channel state can give rise to an effective three-body loss- process that involves the vibrational de-excitation by a 2. Haravifard, S., D. Graf, A. Feiguin, C. D. Batista, J. C. third atom. Lang, D. M. Silevitch, G. Srajer, B. D. Gaulin, H. A. Dabkowska, and T. F. Rosenbaum. Crystallization of The scattering amplitude of the cold atom narrow Fesh- 454

Page 455

Spin Superlattices with Pressure and Field in a Layered Magnet. Nature Communications. 3. Rahmani, , A. E. Feiguin, and C. D. Batista. Anyonic Liq- uids in Nearly Saturated Spin Chains. 2014. PHYSICAL REVIEW LETTERS. 113 (26). 4. Wang, , Kamiya, A. H. Nevidomskyy, and C. D. Batista. Three-Dimensional Crystallization of Vortex Strings in Frustrated Quantum Magnets. 2015. PHYSICAL REVIEW LETTERS. 115 (10). 5. Das, , S. -. Lin, N. J. Ghimire, Huang, Ronning, E. D. Bauer, J. D. Thompson, C. D. Batista, Ehlers, and Janos- chek. Magnitude of the Magnetic Exchange Interaction in the Heavy-Fermion Antiferromagnet CeRhIn5. 2014. PHYSICAL REVIEW LETTERS. 113 (24). Publications Das, , S. -. Lin, N. J. Ghimire, Huang, Ronning, E. D. Bauer, J. D. Thompson, C. D. Batista, Ehlers, and Janoschek. Magnitude of the Magnetic Exchange Interaction in the Heavy-Fermion Antiferromagnet CeRhIn5. 2014. PHYSI- CAL REVIEW LETTERS. 113 (24). Haravifard, S., D. Graf, A. Feiguin, C. D. Batista, J. C. Lang, D. M. Silevitch, G. Srajer, B. D. Gaulin, H. A. Dabkowska, and T. F. Rosenbaum. Crystallization of Spin Superlat- tices with Pressure and Field in a Layered Magnet. Nature Communications. Hazra, J., B. Naduvalath, and E. Timmermans. The narrow Feshbach Resonance as a long-delay interaction. Physi- cal Review B. Rahmani, . QUANTUM DYNAMICS WITH AN ENSEMBLE OF HAMILTONIANS. 2013. MODERN PHYSICS LETTERS B. 27 (26). Rahmani, , A. E. Feiguin, and C. D. Batista. Anyonic Liquids in Nearly Saturated Spin Chains. 2014. PHYSICAL RE- VIEW LETTERS. 113 (26). Rahmani, A., T. Kitagawa, E. Demler, and C. Chamon. Cool- ing through optimal control of quantum evolution. 2013. PHYSICAL REVIEW A. 87 (4): -. Wang, , Kamiya, A. H. Nevidomskyy, and C. D. Batista. Three-Dimensional Crystallization of Vortex Strings in Frustrated Quantum Magnets. 2015. PHYSICAL REVIEW LETTERS. 115 (10). 455

Page 456

Materials for the Future Exploratory Research Final Report Topology in Superposition: Quantum Decoherence in Many-body Systems Wojciech H. Zurek 20130409ER Abstract Background and Research Objectives Decoherence is responsible for eliminating symptoms Decoherence is by now widely regarded as a “quan- of quantumness from the familiar, macroscopic world. tum fact of life.” It is often cited as the bridge over the Thus, even though our Universe is undeniably quantum, quantum - classical divide. Above all, it is investigated our everyday reality appears to be classical. The origin as both an obstacle to practical applications of quantum of decoherence responsible for the emergence of the theory (in quantum computing or quantum metrology) classical is the immersion of the macroscopic objects in and – more recently – as an important resource (e.g., our world in the environments (such as air or photons) in creating entanglement or in speeding up controlled that, via incessant interactions, constantly monitor their transitions between quantum states). states. Quantum states – in contrast to classical states Yet, models of the environment used in its study – the – are perturbed by such monitoring. Therefore, only cer- central spin model and quantum Brownian motion – tain states of the systems can survive such “environmen- were until recently embarrassingly primitive. Both were tal monitoring” intact. They are the candidate classical introduced in the early 80’s and suffer from the same states. The usual objective of the studies of decoherence oversimplification: An environment consisting of many was to assess the effect of the environment on the state non-interacting subsystems (spins and harmonic oscil- of the system. This was done using fairly primitive mod- lators, respectively) that have only limited relation with els of the environment that acted as a “quantum garbage realistic environments encountered in condensed matter dump” for the excess (quantum) information. The aim of or everyday life. this proposal was to shift the focus of attention from the state of the system to the state of the environment. We We have studied decoherence caused by quantum many characterized how many copies of what sort of informa- – body environments. We have, in particular, investi- tion about the system becomes deposited in the envi- gated the flow of information and its amplification in ronment, and assessed its role in the emergence of the settings that are more realistic, and that recognize the classical. We also wanted to study more realistic environ- “facts of life” ignored so far in discussions of the tran- ments. We have succeeded (to some degree, better than sition form quantum to classical: The information we we expected) in meeting the goals of the proposal. Thus, gain about our quantum Universe is usually obtained we have arrived at a very satisfying quantitative theory indirectly, by measuring fragments of the environ- of the amplification that occurs when a photon environ- ment (e.g., photons) that has interacted (e.g., scattered ment monitors objects (e.g., when sunlight scatters from from) the “systems of interest.” This is possible because them). We have also obtained additional insights into interactions capable of decohering are also imprint- excitations arising in many-body environments, and, in ing – in many copies the information about the state of particular, into excitations left behind by quantum phase such systems in the environment. Thus, in our quantum transitions. Understanding quantum decoherence and Universe, at least some of the environments are not just the quantum-classical border is essential for quantum “garbage dumps” for spurious information, but are de information processing, and for the quantum aspects facto communication channels that allow observers to materials science, of interest to LANL and DoE because gain information indirectly – i.e., without direct interac- of their impact on national missions. tion – about objects of their interest. 456

Page 457

This point of view, championed by our group (and known the Kosterlitz-Thouless universality class) add urgency to as “Quantum Darwinism”) shifts the focus from the study this project. of the effect of the environment on the “system of inter- The unexpected answer is that KZM predictions apply, but est” to the investigation of the effect of the system of not in the naïve form (that depends only on the asymptotic interest on the environment. The basic observation is that values of critical exponents). Rather, asymptotic scalings the preferred states of the system (so – called “pointer are expected only in the limit when the quench is very states”) that are singled by their stability in spite of de- slow (which also means that the size of the system has to coherence are becoming imprinted, in many copies, on be very large). These advances have been reported in a pa- the state of the environment. Observers can than obtain per by Dziarmaga and Zurek (Quench in 1D Bose- Hubbard information about such systems by intercepting small frag- Model: Topological Defects and Excitations from Kosterlitz- ments of the environment. (For example, you, the reader, Thouless Phase Transition Dynamics). are now intercepting fragments of the photon environ- ment to find out the content of this report). With Julian Sonner of MIT and del Campo, Zurek has initi- ated investigation of KZM in the context of holographic The study of quantum Darwinism requires new tools. duality. Holographic duality is a powerful approach that al- These are supplied by the modern developments of quan- lows one to treat strongly coupled (i.e., intractable via any tum information theory. For instance, one of the key ques- “direct” approach) field theoretic and many – body models tions is the number of copies of the state of the system using a “dual” system – anti de Sitter Universe – that is that are deposited in the environment in the course of de- weakly coupled, and, therefore, more tractable. coherence. This is de facto a process of amplification that happens because of decoherence. It can be analyzed using Our preprint with these results has just appeared (“Univer- tools such as quantum Chernoff information that allow one sal far-from-equilibrium Dynamics of a Holographic Super- to estimate redundancy of the information deposited in conductor”). It confirms KZM predictions for the winding the environmnet. numbers, but it also opens several interesting questions about KZM and holographic duality we intend to pursue in Scientific Approach and Accomplishments the future. Our study of quantum coherence and decoherence in situ- ations involving many body systems has progressed along Zurek has continued his work on quantum foundations, in- several parallel lines, some of them stimulated by recent cluding an update prepared for the Solvay conference “The experimental developments. Theory of the Quantum World” (edited by D. Gross et al., World Scientific, 2013). These advances have also led to a Our paper connecting decoherence with amplification mini-review in the 2014 Physics Today, and a followup dis- (Zwolak, Riedel and Zurek, Amplification, Redundancy, cussion, in May 2015 issue of Physics Today. These papers and the Quantum Chernoff Information, Phys. Rev. Lett. provide an accessible discussion of the propagation of in- 112, 140406 (2014)) has been just published. It provides formation through many body environments. The interplay modern, quantum information–based understanding of between the process of decoherence and the amplification the nature and of the role of amplification in the transition that is naturally caused by decoherence is especially high- from quantum to classical. This is an important step in the lighted and placed in the context of quantum foundations. development of the “quantum Darwinism” program we This (as well as other advances summed up in that article) have championed. is a major step forward in the understanding of the transi- tion from quantum to classical. The tools developed in this With Dziarmaga, Zurek has investigated dynamics of phase study (such as quantum Chernoff information) will be also transitions in the Kosterlitz- Thouless universality class. suitable in other contexts (such a detection of weak forces, This is important, as there are several interesting systems etc.). They are also essential for the “Quantum Darwinism” that fall into this category, including paradigmatic quantum program our group has championed. Bose-Hubbard model in one dimension. In the Kosterlitz- Thouless universality class the near-criti- Impact on National Missions cal behavior of quantities such as the relaxation time is not This research contributed to the condensed matter science polynomial (as is the norm for most other second order at Los Alamos, providing critical insights into quantum phase transitions) but, rather, exponential. The question is technological applications. These are diverse, including then how to adapt the Kibble-Zurek mechanism (KZM) to superconductors, metrology, etc., and have clear impact on this non-polynomial setting. Ongoing experiments on the national missions. one–dimensional Bose–Hubbard model (that belongs to 457

Page 458

Quantum computing may be of the most significant and Riedel, C. J., W. H. Zurek, and M. Zwolak. The rise and fall urgent potential impact. There is the dream of quantum of redundancy in decoherence and quantum Darwin- information processing, but even if it turns out to be more ism (vol 14, 083010, 2012). 2013. NEW JOURNAL OF PHYSICS. 15: Article Number: 039503. distant than some hope, the fundamental elements of computers will reach single atom size this decade (this is Streltsov, A., and W. H. Zurek. Quantum Discord Cannot Be based on extrapolation of the famous “Moore’s law”). Un- Shared. 2013. PHYSICAL REVIEW LETTERS. 111 (4): Ar- derstanding what happens to superpositions of collective ticle Number: 040401. degrees of freedom and of relevant sources of decoher- ence is then indispensable in these (and similar) applica- Tylutki, M., J. Dziarmaga, and W. H. Zurek. Dynamics of the tions of interest to LANL, the NSSA and DOE. Mott Insulator to Superfluid quantum phase transition in the truncated Wigner approximation. 2013. 21ST IN- Moreover, quantum superpositions in cold atom systems TERNATIONAL LASER PHYSICS WORKSHOP. 414: Article can possibly be used to enhance sensitivity of measuring Number: 012029. devices. Capabilities foreseen for MaRIE seem well suited Zurek, W. H.. Wave-packet collapse and the core quantum for condensed matter studies related to this project (dy- postulates: Discreteness of quantum jumps from uni- namics of phase transitions). tarity, repeatability, and actionable information. 2013. Publications PHYSICAL REVIEW A. 87 (5): Article Number: 052111. Sonner, Julian; del Campo, Adolfo; Zurek, Wojciech H., Uni- versal Far-from-equilibrium Zwolak, M., and W. H. Zurek. Complementarity of quantum discord and classically accessible information. 2013. Dynamics of a Holographic Superconductor, NATURE COM- SCIENTIFIC REPORTS. 3: Article Number: 1729. MUNICATIONS, Volume: 6 Article Number: 7406 Published: JUN 2015 Partner, Heather L.; Nigmatullin, Ramil; Burgermeister, Tobias; et al. Structural phase transitions and topo- logical defects in ion Coulomb crystals, PHYSICA B- CONDENSED MATTER Volume: 460 Special Issue: SI Pages: 114-118 Published: MAR 1 2015 Zurek, Wojciech H. QUANTUM DARWINISM, CLASSICAL REALITY, and the randomness of quantum jumps PHYSICS TODAY Volume: 67 Issue: 10 Pages: 44-50 Published: OCT 2014 Zwolak, , C. J. Riedel, and W. H. Zurek. Amplification, Re- dundancy, and Quantum Chernoff Information. 2014. PHYSICAL REVIEW LETTERS. 112 (14) Article Number: 140406. Campo, del, and W. H. Zurek. Universality of phase transi- tion dynamics: Topological defects from symmetry breaking. 2014. INTERNATIONAL JOURNAL OF MOD- ERN PHYSICS A. 29 (8) Article Number: 1430018. Dziarmaga, J., and W. H. Zurek. Quench in 1D Bose-Hub- bard model: topological defects and excitations from Kosterlitz-Thouless phase transition dynamics. 2014. Scientific Reports. 4: 5950. Pyka, K., J. Keller, H. L. Partner, R. Nigmatullin, T. Burger- meister, D. M. Meier, K. Kuhlmann, A. Retzker, M. B. Plenio, W. H. Zurek, A. del Campo, and T. E. Mehl- staubler. Topological defect formation and spontane- ous symmetry breaking in ion Coulomb crystals. 2013. NATURE COMMUNICATIONS. 4, Article Number: 2291. 458

Page 459

Materials for the Future Exploratory Research Final Report Accurate Interfacial Structures for Atomistic Simulations: Minimizing the Grand- Canonical Free Energy Danny Perez 20130517ER Abstract relevance. However, due to the astronomical size of the To carry out effective simulations at the atomic scale, it accessible configuration space, unconstrained searches is crucial to know the details of the atomistic structure of are scarcely attempted. While sampling algorithms are materials. However, obtaining such information experi- available, materials are notoriously difficult to sample mentally is often impractical due to technical limitations thoroughly, because their energy landscape is very or high costs. It is therefore crucial to be able to directly rough, so that exploring it in search of stable configura- discover these relevant atomic scale configurations using tions is extremely time-consuming. This problem is even computer simulations. In this project, we take on this more severe when the number of atoms in the struc- challenge by developing a suite of novel simulation tech- tures of interest is unknown a priori. This is often the niques that allow the configuration space of complex case for defects such as grain boundaries (the interface systems to be efficiently sampled. We demonstrate their between regions of the materials with different crystal- power by investigating the behavior of defects in materi- line orientations). Indeed, these defects are by definition als, and more specifically of grain boundaries in metals. coupled to the bulk material. They can therefore ex- Our method provides novel insights into the behavior of change atoms with their environment, which can affect these complex systems, insights that were previously out the density, and hence the structure, of the boundaries. of reach. In this case, simulations have to tackle the difficulty of accurately accounting for these changes in the number Background and Research Objectives of atoms, which in an ongoing challenge. Overall, this To carry out effective simulations at the atomic scale, it inability to thoroughly search the space of possible is crucial to know the details of the atomistic structures configurations severely limits the accuracy of atomistic of materials. Indeed, many of the crucial properties of simulations: again, even the most accurate simulation is materials (e.g., strength) are controlled by the details only as good as its starting point. Further, being unable of their atomic structure, for example, the nature of to sample the landscape of these materials implies that the different defects in the material, the ease at which accurate prediction of their properties in various condi- defects move, and how they interact. In order to be able tions of interest to applications cannot be expected. In to make reliable predictions from simulations, one needs this project, we developed efficient sampling algorithms to be able to create representative configurations of the to systematically and autonomously search for the most atoms in the material, so the properties and evolution relevant structures. Our work is based on a grand-canon- of such configurations can be predicted. Even the most ical approach where we are able to efficiently stitch to- accurate methods are useless if such a starting point gether simulations of different number of atoms in order cannot be obtained. A pressing challenge is that, often, to make predictions on open systems where the number obtaining that information experimentally is impractical of atoms is able to naturally fluctuate. This framework is due to technical limitations or to the high cost of such coupled to state-of-the-art sampling algorithms that will a procedure. When lacking better approaches, simple allow an efficient search of this huge space. Our method rules of thumb or geometric constructions are often is demonstrated on problems related to defects and used to fill in the missing details. Sometimes, searches grain boundary in materials. Our approach could have in limited spaces are used to explore the neighborhood a profound impact in the field by leveraging modern of these guesses to assess their thermodynamic stability simulation techniques to turn the “art” of predicting the (i.e., how comparatively stable they are in conditions of atomic structure of materials into a systematic search interest), which is often a good metric of their practical process. This could dramatically improve the accuracy 459

Page 460

and predictive power of atomistic simulations. set of simulations (the thermodynamic integration part) is then carried out at high temperature to “stitch” the Scientific Approach and Accomplishments different simulations together. This is done by introducing An important part of the project has been dedicated to the “fractional particles”, i.e., particles that can come in and development, implementation, and comparison of differ- out of the simulation cell in a continuous fashion by scaling ent sampling approaches. We first investigated the formal their interaction with the rest of the system. This enables underpinnings of different potential simulation methodolo- us to then predict the probability that the system would gies in order to identify the most promising. In doing so, contain a given number of particles at a given temperature we uncovered an unexpected equivalence between three and to compute any thermodynamic quantity of interest at simulations methods that are widely used in the field, any temperature. Our approach combines several state-of- namely Wang-Landau sampling, statistical temperature the-art approaches into a unique new capability that has molecular dynamics, and metadynamics. The first and last not been applied to materials before. methods are workhorses of computational thermodynam- Despite initial encouraging results, we rapidly realized ics. While they superficially share some resemblance, their that, in order to be practical, the efficiency of the parallel formal relationships were never investigated. This has tempering simulations would have to be improved. This lead to many “rediscoveries” where very similar enhance- is because the thermodynamic integration part can only ment or analysis were proposed for each of the methods, be done in a liquid, but we are really interested in the without putting these in the wider context of an “ecosys- properties of the solid. Therefore, our simulations have tem” of related techniques. This duplication of effort is to extend across the melting transition in order to make costly and slows down the pace of innovation. In our work, grand-canonical predictions. Crossing phase transitions is we demonstrated that the three methods can be made a notorious problem in sampling because the configura- absolutely identical through a consistent choice of initial tions on either side of the transitions are extremely dif- conditions and update rules. This unexpected equivalence ferent and hard to reach from the other side. Because the opens the door to technology transfer between different simulation cannot easily cross the transition, convergence communities that were previously working mostly inde- is extremely slow, which introduces errors in our low- pendently. We published our findings in the Journal of temperature predictions. This is an extremely challenging, Chemical Theory and Computation [1]. This work is already and well known, problem that we had to tackle in order well cited, demonstrating the usefulness of this insight. It to make further progress. Our approach to the problem has also been presented in leading national conferences was two-pronged: first, optimally adjust the set of tem- and was very well received. peratures during the parallel tempering stage, and second, Building on this improved understanding of the different introduce a novel database resampling strategy that allows methods, we then focused on assessing different tech- replication of important configurations across the different niques that can be applied to the problem of so-called replicas of the system, instead of just shuffling the same grand-canonical thermodynamics in materials. These configurations around as is usually done in parallel tem- methods allow one to simulate systems where the num- pering. The first problem is a long-standing one and, since ber of atoms is not constant because of a coupling with the introduction of parallel tempering, a huge number of a reservoir of atoms (usually the extended environment solutions have been proposed. These solutions are how- in which a particular configuration is embedded). Upon ever very cumbersome to apply in practice. We realized testing many different approaches, we settled on a com- that, because we were using parallel tempering in conjunc- bination of “adaptive force biasing” sampling combined tion with other state-of-the-art methods to compute the with a parallel tempering approach and a thermodynamic density of states, we could directly predict the optimal integration at different particle number. In simple terms, temperature distribution, instead of slowly pushing the we first run different simulations with a varying number temperatures around in the right direction. This is a major of atoms. In these simulations, each running on multiple advance, because our method does not require running processors, we characterize the thermodynamics on a wide preliminary simulations: instead we continuously adjust range of temperature. This is done by running simulations the optimal temperature set based on the latest thermo- at different temperatures side-by-side and allowing them dynamic data gathered during the simulation. This last to interact by exchanging configurations between them; point is crucial in the case of phase transitions, because this is parallel tempering. This insures the “mixing” of the the transition point tends to move as it converges. There- different simulations and greatly accelerates convergence. fore, a priori identifying the optimal temperature set is just We further accumulate crucial information on the density as difficult as solving the original problem. The method is of states of the system during the simulation. A second first described in Ref. [2], and will be covered in details in 460

Page 461

Ref. [3]. The second improvement is the introduction of a database of configurations that can be used to resample during parallel tempering. The idea is simple but extremely far reaching: by measuring the density of states as we perform parallel tempering, we can automatically discover when some replicas are trapped in “metastable” states, i.e., long-lived states which are nonetheless irrelevant in the conditions of interest. Even better, we can resample a more appropriate state for these replicas from the set of all configurations visited by any other replica at any point in the past [3]. Taken together, these two improvements have delivered dramatic speedups for simulations across the melting transition. These improvements are critical for our needs, but are also extremely widely applicable to other fields, as parallel tempering is a workhorse in atomis- tic simulations. We have implemented our approach into our in-house sim- Figure 1. Illustration of a Sigma 45 grain boundary in copper. ulations code AMDF and demonstrated its efficiency. We The atoms in red a constrained to move as rigid blocks in order to thoroughly validated the codes by comparing the results impose a specific grain boundary orientation. The configurations of different sampling techniques. As a first application, we of the atoms in blue are fully sampled with our method. This is investigated, in collaboration with an ongoing DOE SciDAC repeated by changing the number of atoms in the boundary until project, the behavior of small helium clusters in bulk tung- a whole layer has been added or removed. sten, a system of relevance to fusion energy production. Understanding the behavior of these clusters is critical, as they are known to nucleate helium bubbles, which can lead to dramatic degradation of the performance of the material. We predicted the binding free energies of these clusters over a large range of temperature, which can be used to predict the relative abundance of clusters of dif- ferent sizes. This result nicely complemented the other techniques used in that project. An article describing our results was published in Physical Review B [5] while the details of the method are discussed in another paper pub- lished in Physics Procedia [4]. Using our technique, we have begun exploring the behav- ior of realistic grain boundaries in copper. Copper is an ideal test material because it can be very well described using empirical potentials. We have selected three bound- Figure 2. Evolution of the probability of observing a Sigma 45 aries of different character in order to sample from the grain boundary containing a given number of atoms as a func- very wide range of possible boundaries. For each bound- tion of temperature. The height of each vertical bar is propor- ary, we have performed a complete grand-canonical scan tional to the probability that the corresponding number of atoms is found in the boundary. that enables us to predict their properties at any tempera- ture and chemical potential (the “strength” of the source/ likely. This is expected, as, in the limit as the temperature sink of atoms the system is coupled with). To our knowl- goes to zero, only the state with the very lowest energy edge, this is the first time this was achieved. An example will be observed. However, as the temperature increases, of such a boundary is shown in Figure 1. A typical example the contribution of that state falls sharply. By the time of a result of our approach is shown in Figure 2. There, we the temperature reaches 500K, a wide range of different show the evolution of the probability of observing a given number of atoms, and hence of structures of the bound- number of atoms in the grain boundary as a function of ary can be expected to coexist. Even more interesting, as temperature. The observed behavior is extremely complex. the temperature keeps increasing, the configurations that At low temperatures (200K), only one number of atoms is dominated at low temperatures become increasingly irrel- 461

Page 462

evant. This is a clear demonstration of the need to properly 3. Vogel, T., and D. Perez. Towards an optimal flow: identify structures that are relevant in different condi- Density-of-states-informed replica-exchange simula- tions: predictions of the properties of the boundary will tions. Physical Review Letters. obviously be adversely affected by the postulate that the 4. Vogel, T., and D. Perez. Sampling in the multicanonical relevant structures remain the same at any temperature. ensemble: small He clusters in W. 2014. Physics Proce- Using our approach, we can do away with such assump- dia. 57: 104. tions and directly explore the possible configurations of such complex systems in their full complexity, leading to a 5. Perez, , Vogel, and B. P. Uberuaga. Diffusion and dramatic improvement in our ability to model, understand, transformation kinetics of small helium clusters in bulk and eventually control, such systems. tungsten. 2014. PHYSICAL REVIEW B. 90 (1). Impact on National Missions Publications Atomistic computer simulations play an increasingly impor- Junghans, , Perez, and Vogel. Molecular Dynamics in the tant role in the prediction of properties of materials, or in Multicanonical Ensemble: Equivalence of Wang-Landau the interpretation of experimental results. However, to be Sampling, Statistical Temperature Molecular Dynam- effective, these simulations need adequate starting points ics, and Metadynamics. 2014. JOURNAL OF CHEMICAL that properly represent the most probable atomic-scale THEORY AND COMPUTATION. 10 (5): 1843. structure of the material. This is crucial for a number of the Perez, , Vogel, and B. P. Uberuaga. Diffusion and transfor- Laboratory’s missions, especially those dealing with mate- mation kinetics of small helium clusters in bulk tung- rials in extreme conditions. For example, we recently found sten. 2014. PHYSICAL REVIEW B. 90 (1). that adding a small proportion of interstitials to a grain boundary in tungsten (a leading candidate for the first wall Vogel, T., and D. Perez. Sampling in the multicanonical en- of fusion reactors) can dramatically lower the stress at semble: small He clusters in W. 2014. Physics Procedia. which the boundaries slide, which could lead to creep, but 57: 104. also enhance the ability of the materials to heal radiation- induced defects. To quantify these effects, it is paramount Vogel, T., and D. Perez. Accelerating the convergence of replica exchange simulations using Gibbs sampling and to first identify the structures that are thermodynamically adaptive temperature sets. 2015. Physics Procedia. 68: relevant. These same challenges exist everywhere atomis- 125. tic simulations of materials are used. Our proposed meth- odology will assist in the sampling of the possible struc- Vogel, T., and D. Perez. Towards an optimal flow: Density- tures and greatly help in the identification of the relevant of-states-informed replica-exchange simulations. Physi- ones. This has the potentials to significantly increase the cal Review Letters. power of atomistic simulations across a very wide range of problems of direct relevance to the lab’s missions, for example to investigate the properties of cladding materials in nuclear reactors, or to be able to predict the response of polycrystalline materials to shock, to name only a few examples. This capability directly supports mission needs for DOE Office of Science, Nuclear Energy, and MaRIE while enhancing our basic understanding of materials. References

1. Junghans, , Perez, and Vogel. Molecular Dynamics in the Multicanonical Ensemble: Equivalence of Wang- Landau Sampling, Statistical Temperature Molecular Dynamics, and Metadynamics. 2014. JOURNAL OF CHEMICAL THEORY AND COMPUTATION. 10 (5): 1843.
2. Vogel, T., and D. Perez. Accelerating the convergence of replica exchange simulations using Gibbs sampling and adaptive temperature sets. 2015. Physics Procedia. 68:

462

Page 463

Materials for the Future Exploratory Research Final Report Understanding and Controlling Magneto-Electric Coupling in Multiferroic Materials Dmitry A. Yarotski 20130525ER Abstract matter physics and will open new directions in complex The last few decades have seen the discovery of novel materials research. classes of correlated electron materials with exotic prop- Background and Research Objectives erties that few would have imagined. Multiferroic mate- rials stand out among them, due to the ability to change Multiferroics are technologically important materials their magnetization and/or polarization state with either with co-existing ferroelectric (FE) and (anti) ferromag- electric or magnetic fields. This has the potential to revo- netic (FM/AFM) orders. The allure of these materials lies lutionize future energy, sensing and information technol- in the possibility of controlling the electric and magnetic ogies, as multiferroic circuits will combine the low power responses with either electric or magnetic fields. This consumption and speed of field-effect devices with the has the potential to revolutionize future technology by permanence of magnetic elements. The remarkable combining the low power consumption and speed of properties of multiferroics emerge from strong coupling field-effect devices with the permanence of magnetic between coexisting electric and magnetic orders. De- elements. In addition to their technological appeal, spite recent progress in the synthesis and characteriza- multiferroic materials exhibit a variety of novel physical tion of single-phase and heterostructured multiferroic phenomena, e.g. new elementary excitations, like “elec- materials, the existing understanding of magnetoelectric tromagnons” (magnetic excitations driven by the electric (ME) coupling mechanisms is still controversial, and the field of light), and offer fertile ground for exploring the dynamical properties of multiferroics remain practically fundamental mechanisms governing ME coupling (which unexplored. In particular, the role of low-energy excita- differ between various materials). Despite substantial tions (phonons, magnons, etc.) in the emergence of effort over the past decade, single-phase multiferroic multiferroic functionality has not been clarified. materials with strong ME coupling have only been real- ized at cryogenic temperatures. Our ultimate goal is to leverage the unmatched LANL integration of material synthesis, advanced ultrafast Currently, there is a consensus that the major obstacle optical techniques and forefront solid-state theory to on the way to implementation of room-temperature test the dynamic limits of ME phenomena and reveal the multiferroic functionality is the lack of understanding mechanisms governing ME coupling in representative of microscopic mechanisms governing the coupling multiferroic materials. In pursuit of this goal, we used between magnetization and ferroelectric polarization. sub-picosecond optical and terahertz (THz) pulses to In particular, the role of low-energy excitations in emer- directly excite electronic degrees of freedom and then gence of ME coupling and their influence on dynamics probe the ensuing energy transfer to the low-energy of multiferroic response remain practically unexplored. modes in order to investigate their effect on magnetic Although great strides have been made in characteriza- and electric orders in each material. Our unique com- tion of the static ME response by standard methods bination of broadband time-resolved probes allowed consisting of measuring the magnetization as a function unambiguous separation of spin, charge and lattice of applied electric field or vice versa, they provided little dynamics, and in conjunction with theoretical modeling, information on the underlying microscopic physics. Argu- revealed the mechanisms of coupling between low-en- ably, more insight has come from optical spectroscopy, ergy modes responsible for the emergence of magneto- which hinted on an intimate connection between the electric interactions in multiferroics. The results of this spectrum of low-energy excitations (phonons, magnons, work are poised to make a broad impact on condensed and electromagnons) and ME coupling. However, there 463

Page 464

is still strong disagreement as to whether these low-energy function of temperature. Then, by varying the temperature excitations are directly linked to the ME coupling. above and below the AFM ordering temperature, TN, while tracking the maximum photoinduced reflectivity change, In this project, we have leveraged recent developments in we revealed the role of these fluctuations in governing nonlinear optical techniques and have taken an innovative the coupling between spin and charge orders. The experi- approach to resolving this controversy by direct excita- mental results were interpreted by developing a model for tion of specific low-energy modes and subsequent obser- polaronic hopping between two atomic sites, governed vation of their effect on magnetic/electric orders, and, by the double exchange mechanism. Our studies revealed importantly, coupling between them. More specifically, we that charge fluctuations are coupled to the electric po- have targeted systems where either lattice distortions or larization, and furthermore that magnetic order strongly charge ordering drive ferroelectric response of the materi- influences this coupling. This explicitly demonstrates that als, and investigated how an introduction of elementary quantum charge fluctuations may be the mechanism structural (phonons) and magnetic (magnons) excitations coupling electric polarization and magnetic order in this affects the spin, lattice and charge behavior and multifer- unique multiferroic material. In addition, our time-resolved roic properties of the material. Our ultimate objective for data, particularly the temperature dependence of coherent this project was to address three longstanding problems in phonon oscillations, indicates that magnetic order influ- the basic and applied science of ME systems: 1) How fast ences polaron hopping and also suggests that polarized can the magnetization/polarization be switched, and how bilayers in LuFe2O4 are stacked anti-ferroelectrically along can this be improved? 2) What low-energy excitations are the crystal axis, in agreement with recent predictions. responsible for the strong interactions between electric and magnetic orders? 3) How can we manipulate these Magnon-assisted relaxation of photo-carriers in excitations to enhance the ME coupling and design better multiferroic Eu0.75Y0.25MnO3 multiferroics? The results of our work provide the answers One of the most dramatic effects of strong ME interactions to these questions for several material systems, which are is an existence of a new elementary excitations, the elec- by no means complete but pave the way to better under- tromagnons, that consist of coupled magnons and polar standing of multiferroic functionality and development of phonons and have resonant frequencies in THz range. the basic principles of ME material design. The impact of However, it is still unclear whether the electromagnons are this project is expected to reach well beyond these particu- also coupled to electronic degrees of freedom or not. Here, lar materials as this study has enabled the transition from we address this question by applying optical pump-THz “passive” observation of quasiparticle and order dynamics probe spectroscopy to study the effects of photo-induced to an “active” excitation and control of materials response electronic perturbations on THz electromagnon absorption with coherent photon sources. in a multiferroic Eu0.75Y0.25MnO3. This material belongs to distorted perovskite manganese oxides RMnO3, whose Scientific Approach and Accomplishments magnetic structure evolves with rare-earth ion R from an Our scientific accomplishments are exceptional considering A-type antiferromagnet through incommensurate spiral the scale of the project. This can be testified by our near spin order to a collinear E-type antiferromagnet. Onset 11 refereed journal articles published and in preparation. of the spiral spin order below TN=47 K breaks the spatial In the following, we highlight most important technical ac- inversion symmetry and induces a ferroelectric polarization complishments. in the a-c plane at TFE=30 K. In the temperature range of coexisting spin spiral and ferroelectric orders the electro- Strong coupling of charge fluctuations to ferroelectric and magnon modes at 0.7 and 2.5 THz have been observed, magnetic orders in LuFe2O4 hinting at a strong magneto-electric coupling in this mate- Multiferroic LuFe2O4 has attracted much recent attention rial. We have observed both electromagnon modes in THz due to its strong ME coupling near room temperature. absorption spectra of Eu0.75Y0.25MnO3 crystal below TFE. Theoretical studies have linked the giant ME coupling in The temporal evolution of THz absorption following an LuFe2O4 bilayers to quantum charge fluctuations, but excitation with 1.5 eV pump (above band gap) photons at experimental evidence supporting this hypothesis has temperatures below and above magnetic phase transition been scarce. We used femtosecond optical pump-probe showed no pronounced changes of electromagnon ab- spectroscopy, closely coupled to theory, to examine the sorption. This behavior points at relatively weak coupling effect of photoexciting either intralayer or interlayer Fe2+ of electronic degrees of freedom to the electromagnon -> Fe3+ charge transfer transitions at 1.5 eV and 1.1 eV, modes. Alternatively, the changes might be obscured by respectively, (in effect externally driving polaronic charge a strong frequency dependence of photoinduced THz fluctuations) on the interlayer charge transfer energy as a transmission that resembles free carrier response and can 464

Page 465

be described by the Drude model with scattering rates of an incommensurate antiferromagnetic structure (TN = 650 <1 THz for all measured temperatures. Another notable K) and ferroelectricity (TC = 1103 K). With the aim of en- feature of THz transmission dynamics is very slow (>>ns) hancing the magnetic moment in BiMnO3 while maintain- relaxation at temperatures of 50 K and above, which accel- ing ferroelectricity, various approaches have been followed erates significantly (100’s fs) at T < 35K. The photo-excited including attempting to induce a local spin configuration in carriers relax by either recombining across the bandgap or a B-site ordered structure, or creating a complex chiral or depositing the energy to other quasiparticles. Apparently, canted spin structure via substitution of different 3d transi- a new scattering channel opens at lower temperatures tion metal cations at the B-sites. Using epitaxial scheme, and promotes carrier relaxation back to equilibrium. The we have achieved co-existence of magnetic and ferroelec- proximity of crossover temperature to the magnetic transi- tric orders in BiFe0.5Mn0.5O3 (BFMO) at room tempera- tion near TN suggests that this scattering process involves ture by strain engineering. We have found that the BFMO magnons, similar to another multiferroic, TbMnO3, where perovskite phase only forms if there is a close lattice match magnon-assisted hopping is known to contribute to the with the substrate (i.e., (001) SrTiO3), and if the films are relaxation dynamics. This behavior clearly reveals strong very thin and highly strained. Meticulous growth control spin-charge coupling in the dynamic response of this class has achieved clean, highly epitaxial, strained BFMO layers. of multiferroic materials and might have implications for Using first-principles density-functional theory calculations their future application in optoelectronic devices. we predicted anti-parallel alignment of magnetic moments on Fe and Mn ions and revealed the strong dependence We have also applied optical pump-optical probe spectros- of the total magnetic moment on the strain. Subsequent copy to the same material to extract more details about measurements using X-ray magnetic absorption dichroism electron-magnon coupling. While the THz probe is more spectroscopy at Advanced Photon Source confirmed the sensitive to the collective motion of delocalized photo-ex- predicted complex spin structure in this material, while cited carriers, optical probes are nearly resonant with d-d strain effects are yet to be verified by growing Bi2FeMnO6 transitions of the magnetic ions and provide indirect infor- films on substrates with variable lattice mismatch. These mation about magnetic re-ordering that follows photon results will provide a solid foundation for modeling and absorption. As expected, our results demonstrated that creating more complex heterostructured materials where the optical response of Eu0.75Y0.25MnO3 to photons with the interfaces are used to control strain state and order energies of 1.55 and 3.1 eV is dominated by the d-d and coupling. p-d transitions of magnetic Mn ions. The photo-induced charges increase the occupancy of higher d-orbitals and Other accomplishments create a localized spin excitation that results in ultrafast Experimentally, we have explored several novel materials switching of super-exchange interactions. The decay of this that hold the promise for magnetoelectric functionality. localized spin state appears as the tremendous increase We used optical pump–THz probe spectroscopy at low in the amplitude of photoinduced reflectance due to the temperatures to study the charge dynamics in topologi- strong coupling of optical transitions to the spin-spin cal insulator Bi2Se3. Film thickness variation allowed us correlations. The decay involves emission of spin waves to separate the bulk from the surface transient response. (magnons) and occurs within hundreds of picoseconds of We demonstrated that for thinner films the photoexcita- the pump pulse, in agreement with our optical pump-THz tion changes the transport scattering rate and reduces the probe results. In addition, the relaxation of photoexcited THz conductivity, whereas in thicker films the conductivity electrons to equilibrium included the formation and trap- increases upon photoexcitation and scales with increas- ping of the Jahn-Teller polarons on the Mn sites, which ing both the film thickness and the optical fluence. These affects the magneto-electric coupling strength in this mate- different dynamics are attributed to the surface and bulk rial. electrons, respectively, and demonstrate that long-lived mobile surface photo-carriers can be accessed indepen- Complex spin structure in magetoelectric BiFeMnO3 dently below certain film thicknesses for possible opto- Among single-phase multiferroic materials, BiMnO3 and electronic applications. In the large magnetoresistance BiFeO3 have been extensively investigated due to the material WTe2, ultrafast optical pump-probe spectroscopy promise of strong magnetoelectric coupling close to room- revealed several characteristic timescales in quasiparticle temperature. BiMnO3 is a rare multiferroic material which dynamics. The fast relaxation process occurring on a sub- exhibits ferromagnetism. While the ferroelectric transition picosecond time scale was caused by electron-phonon temperature is relatively high at 400 K, the ferromagnetic thermalization, allowing us to extract the electron-phonon transition temperature is far below room temperature (TC coupling constant. An additional slower relaxation process, = 105 K), and the magnetization is rather small. BiFeO3 has occurring on a time scale of 5-15 picoseconds, could be 465

Page 466

attributed to phonon-assisted electron-hole recombina- Publications tion. As the temperature decreased below ambient, the Aguilar, R. V., Qi, Brahlek, Bansal, Azad, Bowlan, Oh, A. J. timescale governing this process increased due to the Taylor, R. P. Prasankumar, and D. A. Yarotski. Time-resolved reduction of the phonon population. However, below 50 K, terahertz dynamics in thin films of the topological insulator an unusual decrease of the recombination time occurred, Bi2Se3. 2015. APPLIED PHYSICS LETTERS. 106 (1). most likely due to a change in the electronic structure that has been linked to the large magnetoresistance observed Aguilar, R. Valdes, A. Azad, S. W. Cheong, A. J. Taylor, R. P. in this material. Prasankumar, and D. A. Yarotski. Slow relaxation of photo- carriers and their coupling to magnons in multiferroic Experimental advances were matched by theory devel- Eu0.75Y0.25MnO3. Applied Physics Letters. opment aimed at explaining the data and directing the experiment to discovering the mechanisms underpinning Aguilar, R. Valdes, Y. M. Sheu, A. J. Taylor, R. P. Prasan- multiferroic response. Beside modeling the static and kumar, and D. A . Yarotski. Time resolved terahertz and dynamic signatures of magnetoelectric coupling in optical second harmonic investigations in multiferroic RMnO3 and and THz ranges described above, we have focused on de- RMn2O5. Presented at 2013 March Meeting of The Ameri- veloping the theory of ultrafast time-resolved X-ray mag- can Physical Society. (Baltimore, MD, 18-22 March, 2013). netic linear and circular dichroism (XMLD/XMCD) spectros- Chen, A. P., H. H. Zhou, Z. X. Bi, Y. Y. Zhu, Z. P. Luo, A. copy of magnetic materials. A model Hamiltonian including Bayraktaroglu, J. Phillips, E. M. Choi, J. L. MacManus- the core-levels has been considered. The time evolution Driscoll, S. J. Pennycook, J. Narayan, Q. X. Jia, X. H. Zhang, for the spin and charge density has been formulated to de- and H. Y. Wang. A New Class of Room-Temperature Multi- rive the dielectric function. A numerical algorithm is being ferroic Thin Films with Bismuth-Based Supercell Structure. developed and will be tested against the data from ultra- 2013. ADVANCED MATERIALS. 25 (7): 1028. fast XMLD/XMCD capabilities recently developed at LANL. Chen, A. P., W. R. Zhang, F. Khatkatay, Q. Su, C. F. Tsai, L. Impact on National Missions Chen, Q. X. Jia, J. L. MacManus-Driscoll, and H. Wang. Our work directly addresses the Grand Scientific Challeng- Magnetotransport properties of quasi-one-dimensionally es identified in the Basic Energy Sciences Advisory Commit- channeled vertically aligned heteroepitaxial nanomazes. tee (BESAC) report, which are central to DOE’s missions in 2013. APPLIED PHYSICS LETTERS. 102 (9): -. energy, science, and security in general, and to the LANL Materials Grand Challenge in particular. The demonstrated Chen, A. P., Z. X. Bi, Q. X. Jia, J. L. MacManus-Driscoll, and integration of material synthesis, novel ultrafast spectro- H. Y. Wang. Microstructure, vertical strain control and scopic capabilities (intense THz sources and time-resolved tunable functionalities in self-assembled, vertically aligned X-ray probes) and forefront condensed matter theory nanocomposite thin films. 2013. ACTA MATERIALIA. 61 (8): represents the LANL Materials Strategy and provides LANL 2783. with the capability to investigate emergent properties of complex materials through observation of the dynami- Choi, , Fix, Kursumovic, C. J. Kinane, Arena, Sahonta, Bi, J. i. cal behavior of relevant order parameters and through e. Xiong, L. i. Yan, Lee, Wang, Langridge, Kim, A. Y. Borise- selective excitation of the low-energy modes responsible vich, I. a. n. MacLaren, Q. M. Ramasse, M. G. Blamire, Jia, for material functionality. Our thrust to interface materials and J. L. MacManus-Driscoll. Room Temperature Ferri- science with ultrafast THz, optical and X-Ray coherent pho- magnetism and Ferroelectricity in Strained, Thin Films of ton probes represents an essential element in the MaRIE BiFe0.5Mn0.5O3. 2014. ADVANCED FUNCTIONAL MATERI- strategy that connects the M4 facility to the Multi-Probe ALS. 24 (47): 7478. Diagnostic Hall. Proposed experiments will provide criti- Dai, Y. M., J. Bowlan, H. Li, H. Miao, Y. G. Shi, S. A. Trug- cal understanding of the mechanisms of magnetoelectric man, J. X. Zhu, H. Ding, A. J. Taylor, D. A. Yarotski, and R. coupling and thus enable the design and synthesis of new P. Prasankumar. Ultrafast Carrier Dynamics in the Large multiferroic materials with controlled functionalities. Ma- Magnetoresistance Material WTe2. To appear in Physical terials with tunable and novel functionality are an enabling Review B Rapid Communications. component in the development of next-generation devices for sensing, information storage, and spintronics applica- Lee, , S. A. Trugman, C. D. Batista, C. L. Zhang, Talbayev, tions. We believe that our integrated capabilities in com- X. S. Xu, S. -. Cheong, D. A. Yarotski, A. J. Taylor, and R. P. plex material design, synthesis and characterization will be Prasankumar. Probing the Interplay between Quantum of great interest to multiple sponsors, including DOE-BES, Charge Fluctuations and Magnetic Ordering in LuFe2O4. DOD, IC, and industry. 2013. SCIENTIFIC REPORTS. 3. 466

Page 467

Lee, , and R. P. Prasankumar. Correlation between quantum charge fluctuations and magnetic ordering in multiferroic LuFe2O4. 2014. EUROPEAN PHYSICAL JOURNAL B. 87 (11). Lee, J., S. A. Trugman, C. L. Zhang, D. Talbayev, X. S. Xu, S. W. Cheong, D. A. Yarotski, A. J. Taylor, and R. P. Prasan- kumar. The influence of charge and magnetic order on polaron and acoustic phonon dynamics in LuFe2O4. 2015. Applied Physics Letters. 107 (4): 042906. Talbayev, , Lee, S. A. Trugman, C. L. Zhang, S. -. Cheong, R. D. Averitt, A. J. Taylor, and R. P. Prasankumar. Spin-depen- dent polaron formation dynamics in Eu0.75Y0.25MnO3 probed by femtosecond pump-probe spectroscopy. 2015. PHYSICAL REVIEW B. 91 (6). Xu, B., Y. M. Dai, L. X. Zhao, K. Wang, R. Yang, W. Zhang, J. Y. Liu, H. Xiao, G. F. Chen, A. J. Taylor, D. A. Yarotski, R. P. Prasankumar, and X. G. Qiu. Optical Signatures of Weyl Points in TaAs. Physical Review Letters. 467

Page 468

Materials for the Future Exploratory Research Final Report Understanding of Nanoscale Fracture and Its Application in Developing High Fracture Toughness Nanoscale Composites Nan Li 20140450ER Abstract the science of nanoscale fracture and the engineering In this project, we for the first time in the world aim at issue of designing nanoscale materials with high frac- studying fracture toughness of nano/micro-dimensional ture toughness. The development of both modeling and samples and nanostructured materials. High strength experimental capability will also benefit other programs materials generally have low fracture toughness be- at LANL. cause of reduced plastic deformation at the crack tip. Background and Research Objectives We aim to design materials with high fracture toughness and good damage tolerance as well as high strength. Future energy technologies demand novel materials Nanocomposites are promising candidates because of with high fracture toughness and high strength that far the role of interfaces in plastic deformation, including exceed the limits of even the most advanced materials impeding dislocation motion as a barrier, facilitating the to date. Materials with the high strength can have low recovery and emission of dislocation at/from boundar- fracture toughness provided a crack nucleates. We hy- ies, providing additional plastic deformation pathways, pothesize that nanocomposites could hold high strength and assisting in accommodation of plastic deformation. and high toughness by virtue of their high density of These aforementioned roles are strongly determined by interphase boundaries, largely because irreversible the microstructures of nanocomposites and structures energy is associated with plastic deformation. However, and properties of boundaries. Thus, we hypothesize “By fracture mechanics of nanoscale materials (referred virtue of microstructures and structures of nanocompos- to as Nanoscale Fracture), including the measurement ites, dislocations can be dynamically recovered within of fracture toughness and characterization of fracture interface, preventing the formation of dislocation forest, behaviors of nanoscale matertials, is less understood in assisting the further plastic deformation and suppressing contrast to fracture mechanics of conventional coarse- the damage nucleation and propagation. The combina- grained materials. Understanding “Nanoscale Fracture” tion of these advantages will enable the development of involves both (1) defining/evaluating facture toughness high strength and high toughness composites.” However, and (2) determining the influence of microstructures and the current fracture theory and models are unable to boundary structures on fracture behaviors. However, quantitatively and qualitatively test this hypothesis be- current theoretical and modeling tools are unable to ad- cause of the lack of understanding of interface physics. dress the two major issues because they do not have the Moreover, the current experimental techniques cannot capability of dealing with the influence of boundaries on measure and characterize fracture toughness and frac- plastic deformation. Therefore, we propose to develop ture behavior of nanocomposites. Therefore, we have a first of its kind of Nanoscale Fracture (NF) modeling (1) for the first time developed a Nanoscale Fracture tool. The most challenging technical part in terms of this modeling tool and (2) developed the capability of mea- theoretical approach is to include the influence of micro- suring and characterizing fracture behaviors at nanoscale structures and boundary structures on the dislocation (including nano-sized specimen and nano-grained com- networks ahead of the crack tip and the sliding of the posites). Through this ER, we are able to address “what boundaries. Both determine the plastic work that can be is Nanoscale Fracture?” and measure fracture tough- done in the crack propagation process. To do so, we have ness at nanoscale. With this technique, we are able to developed our interface-dislocation dynamics model to design the high strength and high toughness nanoscale deal with interfaces and grain boundaries in polycrystal- materials by virtue of microstructures and boundaries. line composites and integrate it into meso-scale/continu- This significant contribution enables one to address both um-scale models to simulate the evolution of dislocation 468

Page 469

networks in grains and boundaries and the contribution path and (b) a propagating crack also facilitates twinning. of boundary sliding to the plastic deformation. These have This work has published in Philosophical Magazine Letters not been included in any of the existing materials modeling 94 (4), 225-232 (2014). tools because of the lack of systematic fundamental under- Continuum-scale: elastic-plastic deformation in the far- standing of interface physics at atomic scales. field of crack Scientific Approach and Accomplishments To interpret in situ SEM bending testing, we have de- veloped a finite element method to characterize stress Atomic-scale: modeling and experimental intensity factor of crack at a given geometrical dimension. characterization of cracking zone Using this method, we have formulated the dependence By combining in situ high resolution transmission electron of stress intensity factor on sample’s dimensions, loading, microscopic observation and molecular dynamics simula- and materials properties. These results have been used to tions, we for the first time performed atomistic observa- link with experimental data. tion of a crack tip approaching coherent twin boundaries in face centered cubic structures. Coherent twin boundar- Evaluation of fracture toughness in small sample ies (CTBs) in nano-twinned materials could improve crack Evaluation of fracture toughness of materials is dependent resistance. However, the role of the CTBs during crack of the geometry and dimensions of the tested samples. For penetration has never been explored at atomic scale. coarse-grained materials, the sample can be made as big Our in situ observation on nano-twinned Ag under a high as the ASTM standards. Correspondingly, many theoretical resolution transmission electron microscope reveals the models have been developed to evaluate fracture tough- dynamic processes of a crack penetration across the CTBs, ness by combining experimental measurements. However, which involve alternated crack tip blunting, crack deflec- we found such models could not describe fracture tough- tion, twinning/detwinning and slip transmission across the ness for small samples. Figure 1(a) shows the three-point CTBs. The alternated blunting processes are related to the bending sample and loading condition. The sample size is emission of different types of dislocations at the crack tip about 30 um in wide and 6~10 um in height. Figure 1(b) and vary with the distance of the crack tip from the CTBs. shows the simulation model used in finite element model- The results suggest that there is a critical twin thickness ing. Figure 1(c) shows variation of fracture toughness with corresponding to both high strength and good toughness the crack length for two different height of samples (6.2 in nanotwinned materials. This work was published in Na- and 9.2 um). The blue dashed lines show the prediction ture Group Journals, Scientific Reports 4:4397 (2014). from the existing models. We have developed an empiri- cal model, which can predict fracture toughness for small Meso-scale: modeling and experimental characterization sample. Currently, we are working on a paper to report our Combining in situ transmission electron microscopic obser- contribution to the field of nanoscale fracture and na- vation and crystallographic analysis, we characterized twin- noscale mechanics. ning-dominated nucleation, propagation and deflection of crack in molybdenum. Corresponding to previous studies, the links between twins and cracks in bcc structures were proposed and demonstrated, including that (1) cracks and twins are independently produced, (2) cracks are induced by twins, and (3) twins are induced by cracks. According to these understandings, a strong correlation between twins and cracks in bcc structure could exist. However, the ex situ experimental setup could not directly demonstrate this correlation. In our work, we demonstrate a strong twin- crack relation in Mo. \{112\} twinning is characterized to be responsible for the dynamic process of crack zigzag propa- gation. The activation of the \{112\} twinning systems at the crack tips is ascribed to the high local stress at crack tips. Figure 1. Prediction of fracture toughness for small samples. The deflection is ascribed to the alternative of twinning (a) in situ indentation testing of three-point bending micrometer samples, (b) Finite element simulation model, (c) variation of variants that nucleate at the crack tip and are facilitated fracture toughness with crack length. by twinning reflection at GBs. Moreover, in situ HRTEM images suggested de-twinning on crack surfaces. These observations demonstrate a twin-crack relation that (a) twins nucleate cracks and dominate a preferred cracking 469

Page 470

Phase transformation enhanced fracture toughness the interface between the z-AlN and w-AlN crystals is sharp The enhancement of plastic deformability in ceramics will and the layer thickness remains unchanged after the phase enable the design of novel composites with high hard- transformation. These two features can be well addressed ness and measurable ductility that are used in energy and based on the collective gliding of three Shockley partials defense related technologies and aerospace engineering. on every two (111) planes. It is worth mentioning that the Metal/ceramic multilayers have come into greater focus three Shockley partials have the net zero Burgers vector, due to their promising mechanical, physical and chemi- attract each other and form the sharp interface between cal properties, making them practically useful for harsh the z-AlN and w-AlN, as demonstrated by our atomistic environments and extreme loading. Compared to metals, simulations. ceramics are lacking in room temperature plasticity and fracture toughness, because of a high lattice friction stress on most slip systems at low homologous temperatures. Metal/ceramic layered composites have been fabricated and extensively tested under various types of loading. As the bilayer period is reduced to a few nanometers, plastic co-deformability was inferred in Al-TiN multilayers from compressive stress-strain curves. Besides plastic code- formation, we found that thin ceramics layers can easily undergo phase transformation that releases elastic energy and offer additional plastic deformation. This can enhance fracture toughness in nanolayered metal/ceramics com- posites. High-resolution TEM (HRTEM) images in Figures 2a and 2b clearly show the initial Al-AlN-TiN trilayers. Cracks initiate in the TiN layer adjacent to the AlN layer. When the crack opens, the AlN layer does not crack, instead a phase transformation occurs in the AlN component from the zinc blende to the wurtzite phase. The crack first initiates in the TiN layer from the AlN-TiN interface (Figure 2b). This is a Figure 2. (a) and (b) HRTEM images of the Al-TiN interface mode I crack due to the tensile stress resulting from the before and after indentation. The red lines indicate the (111) bending stress state beneath the indenter tip. The second plane. (c) and (d) Initiation and propagation of the crack II. The crack initiates in the TiN layer from the top surface (Figure Al layer near the crack II does not fracture and the layer thick- 2c). According to the crystallography of the TiN layer, the ness reduces from 6.1 nm to 3.8 nm. second crack surface is close to a \{111\} plane (Figure 2c). It is also noticed that the AlN and Al layers near the second crack do not fracture associated with the opening of the crack (Figure 2d). We studied the stress state associated with the crack initiation using finite element method and found that both the cracks are initiated by the tensile stress. To characterize the two phases of AlN and the transition path, we performed in situ TEM indentation tests. Phase transformation processes are revealed with several high- resolution transmission electron microscope (HRTEM) images of the Al-AlN-TiN trilayer during in situ indentation testing. Figures 3b-3e show several TEM images of the AlN layer with respect to the indentation time. Six-atomic lay- ers of w-AlN forms and propagates towards the left direc- tion in Figures 3b-3d. A sharp interface between the z-AlN and w-AlN forms in the AlN layer, as shown in Figure 3d and magnified in Figure 3d’. Finally, the z-AlN layer trans- forms entirely into the w-AlN layer (Figure 3e). Two micro- Figure 3. Phase transformation in the AlN layer. (a) one HRTEM structural features are worthy of further discussion, i.e., image from in situ indentation test. WZ and ZB indicate the 470

Page 471

w-AlN and z-AlN phase. Along the layer, phase transformation occurs heterogeneously. (b)-(e) HRTEM images of four snapshots of the region outlined in the blue rectangle in (a), showing the w-AlN nucleated and propagate towards the right direction. A sharp interface between the w-AlN and z-AlN was observed in all images and magnified in (d’). Interface sliding enhanced fracture toughness Interface sliding for layered composites can be a major plastic deformation mode, especially for weak shear in- terfaces. We investigated the shape and size of the plastic zone at crack tip in layered composites with respect to in- terface shear resistance. Figure 4(a) shows the simulation Figure 5. Crack tip blunts associated with interface sliding in model where interface sliding is described with additional Mg/Nb multilayered composites. slip system with the shear resistance interface. Crystal plas- ticity is described with slips that have a shear resistance Impact on National Missions dislocation. By varying the interface shear resistance, we The proposed work aligns with the call of “Materials: can study the shape and size of the plastic zone at the Discovery Science to Strategic Applications” to address the crack tip. As shown in Figures 4(b)-4(d), with decreasing priority in materials at LANL: “the mechanistic multi-scale interface shear resistance, the plastic zone increases along understanding and control of inhomogeneities, intrinsic the interface plane. As a consequence, the loading at same and engineered, across all appropriate length and time amount displacement decreases as well (Figure 4e). This scales that govern materials functionality”. This ER project means more plastic work associated with interface sliding, addresses two themes of our Materials Grand Challenge: enhancing fracture toughness. Defects & Interfaces and Extreme environments. This work has developed a new materials modeling tool and provided We recently performed in situ indentation testing for Mg/ insights into the roles of interfaces/grain boundaries in Nb multilayers in a SEM. We observed formation of a sharp fracture resistance, which is of mission relevance to work crack from the surface due to the tension stress associated at LANL. elastic bending (Figure 5a). The crack does not propagate but the crack tip blunts associated with interface sliding Publications (Figure 5b to 5e), as evident from the observation of crack Bufford, D.. In situ nanoindentation study on plasticity and tip and a white strip (marked in a red circle). Finally, the work hardening in aluminium with incoherent twin crack propagates and forms a sharp tip again. boundaries. 2014. Nature Communications. 5: 4864. Chen, Y., K. Y. Yu, Y. Liu, S. Shao, H. Wang, M. A. Kirk, J. Wang, and X. Zhang. Damage-tolerant nanotwinned metals with nanovoids under radiation environments. 2015. Nature Communications. 6: 7036. Huang, , Wang, and Zhou. Effect of plastic incompatibility on the strain hardening behavior of Al-TiN nanolayered composites. 2015. MATERIALS SCIENCE AND ENGI- NEERING A-STRUCTURAL MATERIALS PROPERTIES MI- CROSTRUCTURE AND PROCESSING. 636: 430. Li, , Shao, Li, McCall, Wang, and S. X. Zhang. Single Crystal- line Nanostructures of Topological Crystalline Insulator SnTe with Distinct Facets and Morphologies. 2013. NANO LETTERS. 13 (11): 5443. Figure 4. Interface sliding at crack tip. (a) FEM simulation Li, N., and J. Wang. In-situ TEM Study of Dislocation-Inter- model and method, (b), (c), and (d) Variation of shape and size of face Interactions. 2015. In The Transmission Electron plastic zone at crack tip with respect to interface shear resis- Microscope – Theory and Applications. Edited by tance, (e) the loading-displacement curves. Maaz, K.. , p. 69. Croatia: InTech. Liu, L.. Twinning-dominated nucleation, propagation and deflection of crack in molybdenum characterized with 471

Page 472

in situ transmission electron microscopy. 2014. Philo- wang, J.. Reactions of lattice dislocations with grain bound- sophical Magazine Letters. 94 (4): 225. aries in Mg: Implications on the micro scale from atomic-scale calculations. 2014. International Journal Liu, L.. Atomistic observation of a crack tip approaching of Plasticity. 56: 156. coherent twin boundaries. 2014. Scientific Reports. 4: 4397. Salehinia, , Shao, Wang, and H. M. Zbib. Interface structure and the inception of plasticity in Nb/NbC nanolayered composites. 2015. ACTA MATERIALIA. 86: 331. Salehinia, I.. Molecular dynamics simulations of plastic de- formation in Nb/NbC multilayers. 2014. International Journal of Plasticity. 59: 119. Salehinia, I.. Plastic deformation of metal/ceramic nanolay- ered composites. 2014. The Journal of The Minerals, Metals & Materials Society. 66 (10): 1. Shao, , Wang, Misra, and R. G. Hoagland. Spiral Patterns of Dislocations at Nodes in (111) Semi-coherent FCC In- terfaces. 2013. SCIENTIFIC REPORTS. 3. Shao, S.. Energy minimization mechanisms of semi-coher- ent interfaces. 2014. Journal of Applied Physics. 116: 023508. Shao, S., J. Wang, A. Misra, and R. G. Hoagland. Relaxation of Misfit Dislocations at Nodes. 2014. Materials Sci- ence Forum. 783: 515. Wang, . Atomistic Simulations of Dislocation Pileup: Grain Boundaries Interaction. 2015. JOM. 67 (7): 1515. Wang, , R. F. Zhang, C. Z. Zhou, I. J. Beyerlein, and Misra. Interface dislocation patterns and dislocation nucle- ation in face-centered-cubic and body-centered-cubic bicrystal interfaces. 2014. INTERNATIONAL JOURNAL OF PLASTICITY. 53: 40. Wang, , Zhou, I. J. Beyerlein, and Shao. Modeling Interface- Dominated Mechanical Behavior of Nanolayered Crys- talline Composites. 2014. JOM. 66 (1): 102. Xu, E. Z., Li, J. A. Martinez, Sinitsyn, Htoon, N. a. n. Li, Swartzentruber, J. A. Hollingsworth, Wang, and S. X. Zhang. Diameter dependent thermoelectric properties of individual SnTe nanowires. 2015. NANOSCALE. 7 (7): 2869. Yu, Q. i. n., Jiang, and Wang. Tension-compression-tension tertiary twins in coarse-grained polycrystalline pure magnesium at room temperature. 2015. PHILOSOPHI- CAL MAGAZINE LETTERS. 95 (4): 194. Yu, Q. i. n., Jiang, and Wang. Cyclic deformation and fa- tigue damage in single-crystal magnesium under fully reversed strain-controlled tension-compression in the [10(1)over-bar0] direction. 2015. SCRIPTA MATERIALIA. 96: 41. 472

Page 473

Materials for the Future Exploratory Research Final Report Additive Manufacturing of Mesoscale Energetic Materials: Tailoring Explosive Response through Controlled 3D Microstructure Alexander H. Mueller 20150753ER Abstract through manipulation of defects, wave interactions Additive Manufacturing (AM) provides control over the leading to the formation of “hot spots,” - regions of high mesoscale (micron-to-mm) internal structures of com- temperature and pressure from which chemical reac- ponents at a level difficult to achieve through traditional tions initiate, and ultimately control the reactive flow machining techniques. These new structural design and explosive performance properties. parameters have a profound effect on the mechanical AM can elucidate the mechanisms that allow control properties of the AM parts, and provide a previously un- of performance through structure by developing the available avenue to design damage propagation through capabilities and materials to produce reproducible, the part. Control of the components’ microstructure has micro-structured samples for high fidelity dynamic particularly exciting applications to high explosives (HE) experiments. This includes the foundation for tailored whose sensitivity and performance are dominated by sensitivity, directional performance, vibrational toler- structural features at the mesoscale. Modeling the ef- ance, and other properties not currently accessible to fects of mesoscale structure on the mechanical proper- DoD, DOE and industrial reactive material applications. ties, and damage and shock propagation within the part To achieve this goal, mesoscale HE modeling will play an will allow for development a predictive capability to essential roll. fully exploit these new manufacturing capabilities. The methodologies developed within this project will allow The first objective of the modeling effort is to be able for the in-silico building of structures with mesoscale to determine and predict the mechanical response of inclusions, three-dimensional mesh generation of these a design meso-structured AM-HE. This can be accom- structures, and Finite Element Method (FEM) modeling plished by using Finite Element Methods (FEM) to per- of the mechanical properties and reaction under shock. form numerical simulations of AM-HE meso-structures. This generates a modeling foundation for future handoff The meso-structure is usually designed using a CAD to simulations of reactive burn and will allow for the (Computer Aid Design) system. For the FEM analysis it predictive design of the detonation characteristics of ad- is necessary first to mesh the CAD geometry and then ditively manufactured mesostructured HE components. to transfer the mesh to the FEM code. This requires an interface between the CAD system and the Mesh Gen- Background and Research Objectives eration program (MG) and an interface between the MG Additive Manufacturing (AM) provides control over the program and the FEM code. mesoscale (micron-to-mm) internal structures of com- ponents at a level difficult to achieve through traditional In this report we outline the steps developed to bring machining techniques. Application of AM to high explo- meso-structure designs from a CAD modeling to the FEM sives (HE) allows control of the HE microstructure and analysis. properties so that the interaction between the structure, reactive chemistry, and detonation properties may be Scientific Approach and Accomplishments better understood and used. The approach presented in this report for AM-HE meso- structure modeling follows the above mentioned steps: Modern explosives have stochastic, un-optimized develop CAD geometry, generate the mesh, and perform microstructures. AM seeks to change these material FEM analysis. In this study CAD design was accom- limitations by directly assembling hierarchical structures plished using the BRL-CAD program, the TetGen v1.4.3 to tailor and control the functionality of high explosives 473

Page 474

(tetrahedron generation) program was used for mesh generation, and the FEM analysis was performed using the ABAQUS finite element code. The meso-structure was designed in BRL-CAD. Then the geometry of the microstructure was extracted as a STL (STereoLithography) file. The STL file represents the trian- gulated surface of the meso-structure. In the next step the STL file was used as input for the mesh Figure 1. BRL-CAD geometry for Ply90 (a,b) and Ply45 (c,d) meso- generation in TetGen and the geometry was discretized in structures. tetrahedrons. Using different options, the size, the aspect ratio of the element, etc. of the mesh can be controlled. FEM analysis TetGen allows for multiple material regions and holes i.e. Once the FEM mesh for the meso-structure is obtained, regions with no material. the different analyses can be performed. In the present study different examples involving dynamic mechanical In the FEM analysis step the tetrahedron mesh gener- response are shown. The material in all cases was PBX9501 ated with TetGen was used as input for the ABAQUS-FEM modeled as a viscoelastic material with damage. The dam- analyses. In this step, material models, boundary and age in these results models the shear modulus degradation initial conditions were added to the problem. For a given due to microcracking. The analysis was performed using mesh, different analyses can be performed: static and ABAQUS with a user material subroutine implementing the dynamic response, frequency and modal analyses, buckling PBX9501 material mechanical model. stability, heat transfer, burning models, etc. The analyses can be combined; however, in the present report only the Example 1 mechanical response was considered. Several examples In this example the meso-structure Ply90 was subjected involving different structures and different analyses are dynamically to a uniform compression strain rate in z and x presented. direction. Fig. 2(c) shows the damage distribution for load- Geometry and mesh generation ing in z-direction and Fig. 2(d) shows damage distribution As an example of the aforementioned approach, two types for loading in x-direction. The resulting damage is found of meso-structures were considered in this report. The first to be localized at the cylinders intersections between the type follows the idea of laminate composites. The meso- layers. structure consists of a stack of layers each layer being composed of a number of cylinders with stacking sequence at 900 and at 450 as shown in Fig 1. One refers to these structures as Ply90 and Ply45. The FEM mesh for these structures is shown in Fig. 2(a,b). For Ply90 the size of the structure was 2 cm in x and y direction and 0.9 cm in z direction. The cylinder diameter was 1 mm. For Ply45 the size of the structure was 1 cm in x and y directions and 0.3 cm in the z direction. The cylinder diameter was 0.35 mm. The second type of meso-structure follows the idea of par- ticles distributed within a matrix. For this report a distri- Figure 2. TetGen FEM mesh generation for Ply90 (a) and Ply45 bution of spheres inside of a cube was considered. In this (b). (c,d) Damage plot from ABAQUS FEM simulations for Ply90 case, an initial sphere was designed in BRL-CAD and then under differing dynamic loading conditions. discretized using the extracted STL file from CAD geometry. Example 2 Then, using this sphere as a pattern, a number of spheres were generated inside of a cube following a given spatial In this case natural frequencies were determined for Ply90 distribution: uniform (refers as Sph01) and a non-uniform and Ply45 meso-structures simple supported in x, y and z gradient-like distribution (refers as Sph02) in x,y, and z direction. Fig 4 shows different vibration modes for Ply90 direction. Fig. 3 shows the FEM mesh in these cases. In this and Ply45. This analysis used the elastic bulk and shear case the size of the cube was 1 cm in all directions and the modulus extracted as instantaneous values forming the spheres size were scaled accordingly. viscoelastic PBX9501 material model. 474

Page 475

One important aspect of the approach used in this work is that the MG and FEM analysis can be incorporated in more complex code. For example in a shape optimization prob- lem one starts with an initial CAD design. During the opti- mization the shape changes and the geometry needs to be remeshed and FEM analysis performed on new geometry. In the presented approach the mesh procedure and FEM analysis can be included as internal subroutines in the optimization code. In this case the shape of the object is Figure 4. Natural frequency modes of the meso-structures Ply90 represented a triangulated surface which can be available (a,b) and Ply45 (c,d). from the MG program (TetGen in our case). Example 3 In addition, analyses of different array configurations of HE designs can provide a reference database, which can be In this example, a meso-structure with non-uniform distri- used later in design and manufacturing of HE. bution of spheres Sph02 was subjected to a compressive uniform strain rate in the z-direction. This problem used the mesh shown in Fig. 3(c). The loading was applied on one face of the cube on the z-direction while the opposite face was simply supported in z-direction. Material model was PBX9501 as described above. The analysis was per- formed for the time duration necessary for the waves to travel along the length of the cube in z-direction. Figure 5. Dynamic loading of meso-structure Sph02 at uniform compression strain rate in the z-direction, using PBX9501 mate- rial model. Isosurfaces of the pressure wave: (a) initial, (b,c,d) intermediate sequences of time, (e,f) wave reaches the end face of the cube. Impact on National Mission Achieving control of explosive sensitivity and perfor- mance by the hierarchical assembly of energetic materials lies squarely at the intersection of explosives chemis- try, materials science, and detonation physics; core and Figure 3. TetGen FEM mesh generation for uniform (a), (Sph01), unique competencies of LANL. The materials capabilities and non-uniform (b,c), (Sph02), distribution of spherical defects developed by this work align with the lab’s core mission within a cube matrix. as detailed by the NNSA Additive Manufacturing Initiative Roadmap, have utility in support of MaRIE, the Materials Figure 5 shows the pressure wave contours at different Strategy in Defects and Interfaces, Extreme Conditions and moments in time as the wave advances in the material. It Emergent Phenomena areas, and the DOE-BES Mesoscale can be observed that the introduction of particles affects Report; while the modeling and simulation tools developed the wave profile compared with the wave profile in a ho- will position LANL at the lead for HE materials by design. mogeneous material. In the present work an approach for integrating AM-HE de- It should be noted that for designed AM-HE meso-struc- sign with FEM analysis was presented. Additive manufac- tures more complex analyses can be performed such turing (AM) of energetic materials has been designated as as shock problems, modal analyses, heat transfer, burn a priority to the NNSA and their stockpile stewardship mis- models, etc. Bulk properties of microstructure can be de- sion. The development of this technology with relevant ap- termined also from global response of the meso-structure. plications in lifetime extension programs (LEPs) and for the For a meso-structure like Sph01 or Sph02, by adjusting the development of next generation materials (components distribution of spheres or other inclusions in the HE matrix, of the future) has garnered much support throughout the the wave profile can be modified. agency. Furthermore, the development of AM processes 475

Page 476

for HE has received considerable interest in DoD programs and Intelligence sponsors to address their particular needs. The 2016 NNSA stockpile stewardship plan specifically ac- knowledges the importance of AM technologies and their tie-in to MARIE, with the understanding of the interactions of defects in HE during the buildup to detonation being an obvious material interaction in an extreme environment. The understanding that we will gain over the behavior of explosives gives this basic research relevance to the above program areas. 476

Page 477

Materials for the Future Postdoctoral Research and Development Continuing Project Efficient Carbon Nanotube Growth on Graphene-Metal Surfaces Enkeleda Dervishi 20130785PRD2 Introduction Benefit to National Security Missions The project goal is to optimize the synthesis conditions The work is directly relevant to DOE Office of Science in- that promote the catalytic conversion of graphene into terests in the fundamental science of and development high yield few-walled (1-2 layers) carbon nanotubes of multifunctional nanoscale materials. The work will (CNTs). The approach will overcome typical drawbacks of also probe the fundamentals of carbon nanomaterial in- nanotube synthesis that include use of harsh chemical teractions with other materials, also of interest to DOE/ treatments of growth substrates and include the need SC. The nanotube/graphene materials are of significant for multi-step purification processes. These challenges interest for energy storage and energy harvesting, giving will be overcome by using graphene sheets as an in- the work mission relevance for both energy security novative new substrate for growth that allows one-step and the environment (climate and energy impacts). The synthesis of composite materials. Growth can also be at- effort is also directly related to development and basic tained at low-temperatures (~200 degrees C vs the typi- understanding of nanomaterials with ultimate impact on cal 800 degrees or higher) which is essential for enabling electronics, sensing, and imaging needs as well. development of composites based on multiple materials with very different ranges of temperature compatibility. Progress Another innovative aspect is to use dip-pen lithography Throughout FY15, we made the following progress in capability for fine-control over placement and morphol- the areas of large area and single-domain graphene and ogies of metallic nanoparticle growth catalysts. Optimi- grapheme film synthesis, and envelopment of graphene zation of our growth processes will require variation and heterostructures with oxides, nano wires, and quantum study of the influence of multiple growth parameters. dots, as well as spectroscopic characterization of gra- These will include modulation of catalyst particle size, phene molecules. composition, and placement. We will also study the ef- fects of variable temperatures and reactant gas composi- Large area graphenes for QD heterostructure plasmon- tion and flow rates. The resultant synthesis will directly ics: We studied interaction of semiconducting quantum provide nanotube/graphene composites that will be dots with large area graphenes. Interactions in these studied for their energy harvesting and storage poten- heterostructures create new optical responses from tial. Low temperature growth processes are anticipated the quantum dots. Specifically, optical excitation of the to also allow the use of gold nanoparticles for growth as quantum dots creates localized charging of the gra- well, which will be the key to enabling development of phene resulting in optical plasmons that in turn interact nanotube-nanowire hybrids (not yet attained) of interest with the quantum dots to modify their optical emission for optical and electronics needs. The knowledge gained properties. The result is an enhanced rate of biexciton in these first synthesis directions will also allow us to emission. tailor synthesis conditions to pursue growth of other Single domain graphene synthesis: Efforts are proceed- novel composite materials including carbon nanotube ing at generating large area, single domain graphenes hybrids with complex metal oxides. Such oxide thin films and evaluating the importance of different growth form the basis for superconducting and multi-ferroic parameters for optimization. The single domain materi- materials of technological interest. Interactions with na- als will be incorporated into a number of our hetero- noelectronic materials such as nanotubes can introduce structure studies. In particular, we will target generation emergent electronic behaviors that may enhance the of twisted bilayer graphenes that produce new optical properties of interest. 477

Page 478

resonances. Pairing of these systems with quantum dots tures from more molecular-like to large-scale extended should further enhance the plasmonic behaviors discussed behaviors. These results also provide a new characteriza- in the above paragraph. Towards that end we are experi- tion tool for molecular graphenes complementary to other menting with iron films as growth substrates for prefer- approaches but providing more structural information. ential large-area bilayer growth and will continue that approach into FY16. Future Work Large area, single domain graphene synthesis: In FY16 we Graphene films: We have generated graphene films from will continue with optimization of single domain large area graphene disk syntheses for use by collaborators in bio- growth and push Fe-film catalyzed systems for “twisted” sensing applications and bioviability testing. The bio viabil- bilayer growth. Large area materials will be fed into our ity tests are noteworthy in demonstrating that graphene heterostructure studies. The twisted bilayer work in par- films are capable of killing bacterial populations and may ticular will be paired with quantum dot efforts to further prove to be useful for medical coatings technology. probe optical coupling in these systems as follow-on to the FY15 results (see Paste-Year’s Accomplishments). Nanowire heterostructures: Our goal is to access the potential high performance of Si nanowires for energy In FY16, graphene films will continue to be produced to storage in batteries, but overcome current limitations of supply collaborators in the areas of biosensing and biovi- physical degradation of the nanowires on electrochemical ability. cycling by pairing them with a highly conductive overcoat of graphene to provide physical strength. We have pur- In FY16 we will also finish up evaluation of the Ge nanow- sued Ge nanowires as a test system, but have encountered ire melting mechanism and proceed with generation of difficulties in melting of the nanowires during the gra- test heterostructures for use in Raman spectroelectro- phene overcoating process. We have recently discovered chemical probing of the nanowire expansion process that the melting is a function of ambient pressure in the during electrochemical cycling. Si heterostructure growth growth chamber and are closing in on functional growth strategies will be continued. parameters. In parallel we have also developed a number of synthesis strategies for overcoating of Si nanowires by Oxide heterostructures: We now have all components using an etching and catalyst nucleation process. Pursuit of required for generation of the oxide heterostructures and these new strategies will be pursued into FY16. will proceed through FY16 with their optimized generation and related Raman studies. Oxide heterostructures: Our first target is to integrate large-area graphenes with ZnO for optimal electro-optic Graphen Molecule Raman: In FY16 final spectra will be ob- behavior. We have worked out the necessary transfer tech- tained and analyzed. We are at the intial stages of writing niques and are currently generating improved ZnO for inte- up the work for publication. gration to pursue Raman and electrical transport measure- Conclusion ments in these systems. We are pushing to also integrate single domain graphene for performance optimization. We Our primary technical goal is now focused on generation have also begun integrating large area graphene to ferro- of large-area graphene heterostructures. The methods magnetic oxides with the goal of probing strain transfer to we anticipate using will result in generation of functional graphene using Raman spectroscopy. hybrid carbon nanomaterials of interest for energy storage, electronic, and multi-ferroic applications. These may in- Graphene molecule Raman spectroscopy: In FY15 we clude hybrid quantum dot/graphene materials, graphene/ have evaluated size dependent trends in the fundamental nanowire heterojunctions, and graphene composites with Raman response of the D, G, and 2D vibrational modesvof complex oxide thin films. synthesized graphene molecules. These systems have well- defined structures produced by direct chemical synthesis, Publications allowing for the first time exact correlation of Raman be- Gao, Y., O. Roslyak, E. Dervishi, N. S. Karan, Y. Ghosh, C. J. haviors to the graphene structure. We observe clear rise Sheehan, F. Wang, G. Gupta, A. Mohite, A, M. Dattel- and fall of D to G intensity ratios as structures increase in baum, S. K. Doorn, J. A. Hollingsworth, A. Piryatinski, and H. Htoon. Hybrid Graphene-Giant Nanocrystal size to the large area graphene sheet limit. These results Quantum Dot Assemblies with Highly Efficient Biex- can be understood as due to explicit structural changes citon Emission. . 2015. Advanced Optical Materials that vary the number of graphene edge/defect states. We Highlighted on the Inside Front Cover. . 3 (1): 39. also find that the size-dependent vibrational frequency changes can be used to observe the transition of our struc- 478

Page 479

Materials for the Future Postdoctoral Research and Development Continuing Project Understanding and Controlling Magnetism in Multiferroics with THz Pulses Rohit P. Prasankumar 20130812PRD3 Introduction exciting the low-energy modes responsible for material In ferromagnets, the magnetization can be controlled functionality and probing their effects on this coupling. with an external magnetic field, an idea widely used, for More generally, our effort to interface materials science example, in magnetic data storage. However, magnetic with ultrafast terahertz (THz) and optical probes repre- fields have some drawbacks: there are limitations to how sents an essential element in the MaRIE strategy that fast they can be switched, they are inconvenient to use, connects the M4 facility to the Multi-Probe Diagnostic and they consume extra energy. Much effort has thus Hall. This work directly addresses the LDRD Grand Chal- been devoted to multiferroic materials, in which electric lenge in Materials, which underpins all three Laboratory and magnetic polarizations exist simultaneously and mission areas. It also addresses several of the Grand can be coupled, enabling control of magnetism with an Challenges for Basic Energy Sciences identified by the electric field. DOE Office of Science. We will work with the Program Director for Basic Energy Sciences to explore future Recently, there have been significant advances in the funding opportunities within BES in the growing areas of ability to synthesize these materials, which has led to ultrafast materials science. much effort focused on increasing the magnetoelectric (ME) coupling between magnetic and ferroelectric (FE) Progress order, especially near room temperature. However, the In FY15, one of the main things that we worked on is microscopic origin of this coupling can vary across differ- optical-pump, THz probe (OPTP) measurements in the ent classes of materials, and is not well understood. Low multiferroic HoMnO3. Unlike typical OPTP measure- energy excitations (i.e. soft modes, magnons, etc.) are ments, THz probing in HoMnO3 shows changes only at thought to play a key role in ME coupling, and therefore the magnon resonance, allowing us to directly track spin also offer a promising route to controlling it. dynamics after optically exciting electrons. We learned from additional measurements at different pumping Here, we will use ultrashort optical and terahertz (THz) fluences and temperatures that the excited electrons pulses to shed light on the microscopic origin of mag- couple to the spins via phonons, since the optically netoelectric coupling in different multiferroic materials induced changes in the magnon resonance are consis- (building on our previous work in this area). In these tent with steady state temperature increases. To test experiments, we will use these pulses to separately ma- this physical picture, we did optical pump, optical probe nipulate and probe the magnetic and electric orders in a measurements. A slow rise time is seen in these signals, given material, enabling us to, e.g., modify the magnetic which is nearly identical to that seen in the THz data. The order and probe the resulting changes in ferroelectric slow rise time in the optical data can be attributed to order, or vice versa. This will provide much insight into spin-phonon relaxation through its temperature depen- magnetoelectric coupling in multiferroics, which should dence. This confirms that the magnon dynamics seen in extend our knowledge of their basic physics and also en- our THz signals result from a phonon-mediated transfer able researchers to optimize them for applications. of energy from electrons to spins, since they correspond to the spin-phonon relaxation time. We have written a Benefit to National Security Missions manuscript about these results, which should be submit- The proposed experiments will provide LANL, as well as ted to Physical Review Letters in in FY15. CINT, with the capability to investigate magnetoelectric coupling in multiferroic materials through selectively The research team gave talks about this work at the Ma- 479

Page 480

terials Research Society spring meeting and the Quantum them; we have extensively used optical pump-probe spec- Electronics and Laser Science conference. troscopy to study these heterostructures, and THz pump- ing should give more insight into ME coupling in these We spent some time looking for nonlinear THz pulse driven systems. transport in LCMO and LSMO thin films (which are simulta- neously magnetic and conducting). The current conclusion Finally, we will extend our initial experiments on THz trans- is that our THz fields are not strong enough to drive non- port in thin manganite films to explore the role of polarons linear transport, since all measurements seem to reveal a vs. free carriers in transport, particularly under high driving linear response. Considering the measured values for the fields. This will likely include the fabrication of a metama- binding energies of the polarons in these materials, we do terial structure on top of these films to enhance the field expect that 10-100 times larger THz electric fields would strengths in the manganite layer, since our THz electric be needed to be able to “untrap” or pull electrons out of fields are not yet strong enough to liberate polarons from bound polaronic states. The next step in this project is to their bound states. This is a new regime of transport in work with other colleagues in the LUMOS group to pattern these extensively studied systems, which should reveal a metamaterial on top of the film, which should enhance novel physical phenomena. Overall, the proposed experi- the incident THz pulse at the film by a factor of as much as ments should thus provide much insight into ME coupling 50. in multiferroics and transport in colossal magnetoresis- tance manganites, with wide ranging impact in physics and We recently obtained some TbMnO3 crystals. These are materials science. multiferroic, similar to HoMnO3 discussed above, but rath- er than having a magnetic dipole active magnon mode, it Conclusion is electric dipole active, or an electromagnon, and also lies Multiferroic materials, in which magnetic and ferroelec- within the spectrum of our THz pulse. In initial tempera- tric order can be closely coupled, offer much promise for ture dependent transmission measurements it was difficult a variety of applications in data storage, novel logic ele- to identify this mode, but we are working to better polish ments, and sensing. However, the mechanisms underlying the crystal and solving some other experimental problems this coupling are not well understood, limiting the poten- so that it should be observable. Then we will optically tial of these materials. Here, we will use ultrashort optical pump and probe changes in the electromagnon as done in and terahertz (THz) pulses to shed light on the microscopic HoMnO3. Since this magnon can be excited with our THz origin of magnetoelectric coupling in different multiferroic E-field which is significantly stronger than our THz B-field, materials. This will provide much insight into magneto- we will also try to detect changes to the ferroelectricity electric coupling in multiferroics, which should extend our by probing second harmonic generation after exciting this knowledge of their basic physics and also enable research- mode. ers to optimize them for applications. Future Work Publications In FY16, we will initially focus on submitting our optical- Bowlan, P., D. A. Yarotski, N. J. Hur, A. J. Taylor, and R. P. pump, THz-probe results on HoMnO3 for publication. We Prasankumar. Direct observation of magnon dynamics will then optimize our system for performing mid-infrared in multiferroic HoMnO3. Presented at Quantum Elec- (IR) pump, THz-probe experiments on HoMnO3, which tronics and Laser Science Conference. (San Jose, CA, will shed light on the coupling between lattice and mag- 10-15 May 2015). netic degrees of freedom. We will then perform similar Bowlan, P., S. W. Cheong, D. A. Yarotski, A. J. Taylor, and R. experiments on the manganite TbMnO3, which is well P. Prasankumar. Directly probing antiferromagnetic or- known for its strong magnetoelectric coupling, albeit at der in HoMnO3 on an ultrafast time scale using optical- low temperatures. This material possesses an electroma- pump, THz-probe spectroscopy. Presented at Materials gnon resonance that is directly linked to ME coupling, and Research Society Spring Meeting. (San Francisco, CA, the proposed experiments should thus shed much light 6-10 April 2015). on these phenomena. We should be able to photoexcite the electromagnon resonance while probing ferroelectric order with second harmonic generation, shedding light on the coupling between these degrees of freedom. We will also perform similar experiments on multiferroic hetero- structures, consisting of thin FE and magnetic films grown on top of one another to maximize the coupling between 480

Page 481

Materials for the Future Postdoctoral Research and Development Continuing Project Design Principles for High Performance Organic Photovoltaics Aditya Mohite 20140658PRD1 Introduction interfaces and specifically provide the following in an In this project we will apply innovative interface modi- attempt to develop high efficiency solar cells: fication techniques to alter the charge recombination • Reveal the design principles for the efficient conver- rates across an organic solar cell interface in a bilayer or- sion of light to energy ganic photovoltaic device by incorporating a spacer layer • Develop mesoscale devices starting from model with designed functionality. The photo-generated exci- systems ton after migration to the donor-acceptor (D-A) interface dissociates due to a strong built-in field setup by the • Apply the fundamental understanding to practical energy level alignment of the donor and acceptor form- architectures such as bulk heterojunctions (BHJ) for ing a charge transfer or exciplex state (electron is on the organic solar cells and also for other types of light to acceptor molecule, C60, and hole on donor P3HT). There energy conversion applications is a strong tendency for the exciplex state to recombine, Progress which leads to poor photocurrent and thus low power conversion of sunlight to electricity. A few monolayers of There has been significant progress towards the goals the spacer layers are expected to suppress the electron described in FY15. We have achieved two key results to recombine with the hole and promote the separated that are fundamentally transformational. electron away from the hole to generate photocurrent. First, by performing carefully designed experiments We will use three different types of spacer layers: (a) In- on organic solar cell devices, we have understood the sulator (LiF) (b) Oligomer molecules and (c) Heavy Atom. role of microscopic ordering at the interface (donor/ Each of these spacer layers will result in the suppression acceptor), where the exchange of charge and energy of the exciplex recombination using different mecha- takes place. Given the spaghetti like morphology of the nisms. For example, LiF will act like a tunnel barrier, polymer systems, this was a challenging problem as where an optimum thickness of the barrier is expected different morphologies lead to different optoelectronic to completely suppress back recombination. Similarly, in response and manifests as a different interface. There the case of Oligomers, ordering at the D-A interface will was no control over the morphology of the polymers control the recombination rate and for the Heavy Atom at the interface which led to low efficiency devices. We case, the exciton lifetime is expected to be long due to have demonstrated the role of ordering by using small enhanced spin-orbit coupling, which will impart some molecules (Oligothiphenes) with specific functional triplet character to the singlet state. These strategies are groups that microscopic ordering can be controlled also expected to bring about a deep scientific under- and this leads to a 500% increase in the overall power standing of interface photo-physical processes that will conversion efficiency in a bilayer device. These small be universal for heterostructures interfaces created with molecules can be deposited by thermal evaporation or nanostructures. via Langmuir-Blodgett trough process. We determined Benefit to National Security Missions that when the oligomers are arranged out of plane to the donor (polymer that absorbs the light), extremely This project aligns with LANL’s Energy Security efficient charge and energy flow is observed, in contrast and Environment Mission areas, the DOE-BES Sunshot to when they are lying in plane or are pi-stacked. The initiative, and Grand challenges addressed by the BES. microscopic structure was verified using one and two Through studying this area of research, we will gain con- dimensional XRD at LANL and at the Advanced Photon trol over the transfer of charge and energy flow across 481

Page 482

Source at Argonne National Lab. Furthermore, using these Publications design principles established on proof of principle bilayer Dun, , Huang, Huang, Xu, Zhou, Y. e. Zheng, Tsai, Nie, D. R. solar cells, we translated them to practical polymer solar Onken, Li, and D. L. Carrollt. Hydrazine-Free Surface cell devices also known as bulk heterojunction solar cells. Modification of CZTSe Nanocrystals with All-Inorganic We observed an increase in the efficiency from 4% to 8% Ligand. 2014. JOURNAL OF PHYSICAL CHEMISTRY C. 118 (51): 30302. with these strategies in a P3HT/PCBM system. This is the highest efficiency reported for a P3HT/PCBM solar cell to Kuo, , Nie, Tsai, Yen, A. D. Mohite, Gupta, A. M. Dattel- date. This work was in review in Science for 9 months and baum, D. J. William, K. C. Cha, Yang, Wang, and Wang. unfortunately was rejected. It was the editors call. We are Structural Design of Benzo[1,2-b:4,5-b ‘]dithiophene- writing a manuscript for Nature Materials on the ordering Based 2D Conjugated Polymers with Bithienyl and Ter- work. thienyl Substituents toward Photovoltaic Applications. 2014. MACROMOLECULES. 47 (3): 1008. In addition to the ordering work, we have also embarked upon a new research direction of controlling morphology Liu, , M. R. Kelley, S. A. Crooker, Nie, A. D. Mohite, P. P. in the recently discovered perovskite solar cells. One of Ruden, and D. L. Smith. Magnetoelectroluminescence of organic heterostructures: Analytical theory and the key challenges in perovskite solar cell is controlling the spectrally resolved measurements. 2014. PHYSICAL RE- crystal structure and morphology, which determines its VIEW B. 90 (23). optical and electrical properties. We have developed three new strategies to develop highly crystalline films with large Mok, J. W., Lin, K. G. Yager, A. D. Mohite, Nie, S. B. Darling, single crystalline grains that have quality similar to that of Lee, Gomez, Gosztola, R. D. Schaller, and Verduzco. conventional direct band-gap semiconductors like GaAs Linking Group Influences Charge Separation and Re- and defect purity like crystalline Si. These techniques are combination in All-Conjugated Block Copolymer Pho- (1) Hot casting using an IR heat gun (2) Doctor blading for tovoltaics. 2015. ADVANCED FUNCTIONAL MATERIALS. making large area films and (3) Hot-casting using spin coat- 25 (35): 5578. ing. Nie, , Chen, Smith, Xia, Hewitt, and Carroll. Nano graphite platelets enhanced blue emission in alternating cur- We get power conversion efficiencies of >18% with thee rent field induced polymer based electroluminescence films and writing several manuscripts for high impact jour- devices using Poly (9,9-dioctylfluorenyl-2,7-diyl) as the nals like Nature Materials, Nature Photonics and Science. emitter. 2014. ORGANIC ELECTRONICS. 15 (1): 99. This work is expected to have transformation impact in the field of thin-film solar cells. Nie, , Tsai, Asadpour, Blancon, A. J. Neukirch, Gupta, J. J. Crochet, Chhowalla, Tretiak, M. A. Alam, Wang, Future Work and A. D. Mohite. High-efficiency solution-processed In FY16, we will focus on the aspect of photostability perovskite solar cells with millimeter-scale grains. 2015. SCIENCE. 347 (6221): 522. and reliability in the large grain perovskites that we have developed using several techniques. We will measure Sun, , Asadpour, Nie, A. D. Mohite, and M. A. Alam. A Phys- the efficiency of the perovskite solar cells as a function of ics-Based Analytical Model for Perovskite Solar Cells. constant solar light using a solar simulator. We will test the 2015. IEEE JOURNAL OF PHOTOVOLTAICS. 5 (5): 1389. device for chemical and structural degradation. We will also explore the effect of UV light on the photostability Tsai, , Nie, Cheruku, N. H. Mack, Xu, Gupta, A. D. Mohite, of perovskite solar cells. In addition, we will also investi- and Wang. Optimizing Composition and Morphology for Large-Grain Perovskite Solar Cells via Chemical Con- gate the thermal and mechanical stability in these devices trol. 2015. CHEMISTRY OF MATERIALS. 27 (16): 5570. through a collaboration with Prof. Reinhold at Stanford University. Xia, , Chen, G. M. Smith, Sun, Yang, Nie, Li, Huang, Ma, and D. L. Carroll. High-performance alternating current Conclusion field-induced chromatic-stable white polymer elec- The success of this proposal will result in obtaining uni- troluminescent devices employing a down-conversion versal design rules that will guide the fabrication of high layer. 2015. JOURNAL OF LUMINESCENCE. 161: 82. efficiency photovoltaics using nanostructure materials. Yamaguchi, , Granstrom, Nie, Sojoudi, Fujita, Voiry, Chen, These design principles will be applicable to several other Gupta, A. D. Mohite, Graham, and Chhowalla. Reduced applications such as organic light-emitting diodes (OLEDs), Graphene Oxide Thin Films as Ultrabarriers for Organic sensors, photo-catalysis, etc. Electronics. 2014. ADVANCED ENERGY MATERIALS. 4 (4). 482

Page 483

Yen, , Tsai, Kuo, Nie, A. D. Mohite, Gupta, Wang, Wu, Liou, and Wang. Flexible memory devices with tunable electrical bistability via controlled energetics in donor- donor and donor-acceptor conjugated polymers. 2014. JOURNAL OF MATERIALS CHEMISTRY C. 2 (22): 4374. Yin, S. u. n., Nie, A. D. Mohite, Saxena, D. L. Smith, and P. P. Ruden. Current-voltage characteristics of organic heterostructure devices with insulating spacer layers. 2015. ORGANIC ELECTRONICS. 24: 26. 483

Page 484

Materials for the Future Postdoctoral Research and Development Continuing Project Synthesis of Novel Energetic Materials David E. Chavez 20140659PRD1 Introduction zines led to compounds with exceptional absorption of This project will focus on the development of novel visible light. energetic materials that are thermally stable but can be Twelve compounds of non-explosive tetrazines and initiated through the use of unprecedented mechanisms explosive tetrazines with zinc, copper, nickel, cobalt such as photochemical or electrochemical stimulation and iron were synthesized and their optical, explosive to novel excited states. The research will focus on the and chemical properties characterized in FY15. The iron synthesis of molecules with kinetic barriers large enough based compounds had the largest absorption of visible to prevent thermally induced detonation and excited light, the copper based compounds had significant ab- states capable of bypassing those barriers in response to sorption of near-infrared light, while the zinc, nickel and non-thermal stimulus (optical or electrochemical stimu- cobalt based compounds did not show any promising lation). absorption of light. In addition it was found that the ab- sorption of light was dependent on the metal used, and Benefit to National Security Missions independent of whether the compounds were explosive Los Alamos National Laboratory is a world leader in the or not. This was significant because it demonstrated that synthesis and development of energetic materials. The it is possible to control the optical and explosive proper- design of a new class of explosives with controlled initia- ties of compounds independently of each other. A manu- tion properties has the potential to enhance safety and script of these efforts has been submitted for publication could have applications in detonator development and in Inorganic Chemistry in FY15. weapons programs. The proposed work will contribute to the Laboratory’s core missions of new energetic ma- The strong absorption of visible light in iron based terials development and threat reduction science. The compounds prompted further investigation into this fundamental nature of the work could have an impact in class of materials. It was found that chemically altering the basic understanding of materials and chemistry. the tetrazine portion of the compounds could alter the wavelength of light absorbed without significantly af- Progress fecting the explosive properties of the compounds. This In FY15 significant progress was made towards the de- was significant because it allowed for the compounds velopment of materials explosive for direct optical initia- to be tuned to the wavelengths of light produced by tion. The overall goal of the project is the development commercially available lasers. This match between the of new explosive materials, which are stable to external wavelength of light emitted by laser sources and the stimuli (impact, friction, heat, and spark) that can be wavelength of light absorbed by the compounds will detonated when exposed to laser irradiation. Conven- lead to move efficient conversion of the laser energy into tional explosives are unable to undergo this detonation the initiation of the explosive materials. A collaboration process since they do not strongly absorb laser light. between M division and T division was established to Compounds consisting of explosive tetrazines (tetrazines better understand the properties of these new materials. are nitrogen rich small dye molecules) with metal ions Overall these six new compounds have been synthesized were targeted due to their potential to absorb large and their optical, explosive, and chemical properties amounts of light and their potentially tunable explosive have been characterized. A publication summarizing properties. Initial results from previous work had shown these efforts is currently (FY15) in preparation. that the combination of iron with non-explosive tetra- 484

Page 485

The absorption of near-infrared light by the copper based Conclusion compounds was significant since the most common com- The technical goals of the project are to produce new mercially available laser emits near-infrared light. Addition- optically active energetic materials that respond to non- al investigation into these materials led to the discovery thermal stimuli. The anticipated impact will be in the area that the copper based compounds could adopt two dif- of controlled initiation of energetic materials for applica- ferent forms. The first form has a higher ratio of tetrazine tions. Access to such materials will significantly enhance to copper, while the second form has higher density and the safety and surety of many systems employing energetic more favorable explosive properties. Tetrazines with differ- materials. ing explosive properties were used to synthesize additional copper based compounds of both forms. These resulting Publications compounds have a wide range of explosive properties, Breiner, M. M., D. E. Chavez, T. W. Myers, and R. D. Gilardi. but nearly identical optical properties. This is significant 1,2,4,5-Tetrazinyl-Substituted Amino-1,2,4,5-Tetra- because it will allow for the determination of what explo- zines. 2015. SYNLETT. 26 (4): 557. sive properties are important for direct optical initiation. Chavez, D. E., T. W. Myers, J. M. Veauthier, M. T. Greenfield, A collaboration with W division has been established to R. J. Scharff, and D. A. Parrish. Pentaerythritol Trinitrate test the response of these materials to laser irradiation. A Substituted s-Tetrazine and s-Triazine. 2015. SYNLETT. manuscript is being prepared on this work in FY15 with the 26 (14): 2029. expectation that publication will occur in FY16. Myers, T. W., D. E. Chavez, S. K. Hanson, R. J. Scharff, B. L. In FY15 this project allowed additional capabilities to be Scott, J. M. Veauthier, and Wu. Independent Control established. Single crystal X-ray diffraction capability was of Optical and Explosive Properties: Pyrazole-Tetrazine developed at TA-9 which allows for the determination of Complexes of First Row Transition Metals. 2015. INOR- the molecular structure of new materials. Although this GANIC CHEMISTRY. 54 (16): 8077. capability existed off site, the ability to run these experi- Myers, T. W., S. K. Hanson, Jacqueline M. Veauthier, and ments at TA-9 avoids the cost and safety concerns of trans- David E. Chavez. First row transition metal complexes porting explosive materials off-site. In addition protocols of tetrazine based ligands as a new class of energetic for the elemental analysis of compounds with both organic materials . Presented at 249th ACS National Meeting & and metallic components were developed. This type of Exposition. (Denver, CO, 22-26, Mar. 2015). analysis allows for the determination of the mass composi- tion of new materials. Overall, a large number of new compounds have been synthesized as candidates for direct optical initiation. In FY16 the existing materials will be further investigated to determine the most promising materials. In addition the results obtained and the collaborations established will be used to guide the design of the next generation of these materials. Future Work Thomas was very productive in FY15, developing and char- acterizing new photoactive energetic molecules. In FY16 he will continue to make new molecules and examine their properties for optical initiation. Specific FY16 tasks to be accomplished: • Characterization of tetrazine ligands • Reactions of tetrazine ligands with metals, character- ization of products • Scale up new selected compounds and collaborate with W division for laser initiation testing. 485

Page 486

Materials for the Future Postdoctoral Research and Development Continuing Project Investigating Structure-Directing Agents in Nonconventional Nanowire Synthesis Using a Transmission-Electron-Microscope Flow-Cell Holder Jennifer A. Hollingsworth 20140661PRD1 Introduction ated worldwide is lost as heat), heat management in the Understanding the growth mechanisms for one-dimen- electronics industry, remote power generation, etc. If sional (1D) semiconductor nanostructures, especially successful, to our knowledge, this will also be the first more complex 1D heterostructures including superlat- real-time in situ imaging of nanowire growth processes tice, core/shell, and dumbbell-like structures, is essential in solution. for the controlled synthesis of these novel architectures. Benefit to National Security Missions Importantly, structural control is anticipated to enable properties control, where the ability to tune transport The work will advance fundamental understanding of properties, charge separation efficiencies, and charge or nanomaterials growth mechanisms, which will enable energy transfer processes underpins a range of appli- new materials-by-design strategies. It will also advance cations from energy conversion and storage to cataly- our understanding of chemical processes that take place sis and photodetection. To date, however, controlled at nano/meso-interfaces and how these influence struc- solution-phase growth of high quality and intentionally ture and ultimately function. Together, these contribute designed 1D heterostructures is still in the very early directly to DOE/SC, Basic Understanding of Materials and stage. Improved methods for establishing the structure- Fundamental Chemistry missions. It will also advance a controlling mechanisms and, thereby, for increasing new capability in liquid-phase transmission electron mi- the pace of materials discovery and development are croscopy, propelling LANL to the forefront of this emerg- needed. ing and important research area. This capability will enable direct assessments (with enhanced efficiency and The work will take advantage of the postdoc’s exper- accuracy compared to ‘post-mortem’ assessments) of tise in glycol-based solution-assisted low-temperature the impact of growth parameters on morphology, com- refluxing processes for synthesizing tellurium-based position and interface evolution. This same technique semiconductor nano-heterostructures for thermoelectric can also be applied to studying corrosion processes and applications, along with LANL’s new capability in in situ the impact of chemical or high-energy electron dosing liquid-phase transmission electron microscopy (TEM). on both growth and degradation of surfaces or three- The latter will enable an unprecedented systematic as- dimensional structures, with relevance to MaRIE and sessment of nanowire and nanowire heterostructure stockpile stewardship. Finally, the nanowire materials formation real-time by direct imaging under the condi- produced will have direct relevance to energy harvesting tions of growth. Nanowires will be synthesized in the (energy security, remote power generation) and cooling TEM flow cell, permitting direct assessment of the im- (e.g., spot cooling in electronics) applications, with fun- pact of reaction parameters on nanowire formation and damental understanding translatable to other nanowire growth progression. Glycol-based syntheses are scalable materials systems that have potential applications in and afford a lower temperature alternative compared to chem/bio sensing. higher temperature and more costly solution-phase and vapor-phase approaches. And yet, they are also capable Progress of yielding high-quality, single-crystalline materials. The In FY15, the postdoc focused on flow-enabled synthesis specific 1D nanostructures studied here will support ad- of hetero-structured nanowires (toward addressing a vances in thermoelectrics and topological insulator ap- need in the fields of thermoelectrics and topological plications, with important consequences for waste-heat insulators for ultrathin bismuth or lead chalcogenide energy harvesting (more than half of the energy gener- nanowires (e.g., Bi2Te3 or PbTe, respectively) and their 486

Page 487

heterostructures (e.g., where PbSe/PbTe superlattices Bi2Te3 or PbTe, respectively) and their heterostructures were predicted by Dresselhaus (PRB 2003) to afford a (e.g., where PbSe/PbTe superlattices were predicted by game-changing thermoelectric figure-of-merit: ZT >6, Dresselhaus (PRB 2003) to afford a game-changing ther- if nanowire diameter could be controlled to <5 nm and moelectric figure-of-merit: ZT >6, if nanowire diameter segment lengths to only 2-3 nm). He has successfully could be controlled to <5 nm and segment lengths to only re-established our laboratory’s novel flow-solution-liquid- 2-3 nm). Notably, the catalyzed flow-based approach may solid method for hetero-nanowire synthesis. He deter- prove even more amenable to addressing this need for mined that lack of reproducibility in nanowire growth ultra-thin / controlled-heterostructure nanowire growth, (even whether or not nanowire growth took place, e.g., in as we have shown that by simply tuning the flow-rate (ad- competition with quantum dot growth) could largely be dition rate) of the nanowire precursor compounds during attributed to the treatment of the growth substrate prior growth affords dramatic control over nanowire diameter to deposition of the low-melting metal catalyst (bismuth). (faster yields narrower by enhancing supersaturation). We So far, only HF treatment ensures that the metal film is also aim to shift our solution methods to study in situ in sufficiently “active” and able to support/catalyze nanowire LANL liquid-phase TEM holder. growth (without HF treatment, the metal appears not to break apart into the nanoscale catalytic “islands”). More Conclusion recently, he has successfully synthesized heterostructure The project will advance the state-of-the-art in liquid- nanowires comprising controllably abrupt or alloyed phase transmission electron microscopy (TEM). By pushing interfaces between segments of defined length. He has boundaries in terms of how the parameters and processes controlled interface properties by exploring methods for that influence nanomaterial’s growth are assessed, the depleting the catalyst droplets of precursor prior to initia- project will also afford new mechanistic understanding that tion of growth of a second or third segment (eliminated will enable ‘materials-by-design’ of complex, functional the “reservoir” effect so common in gas-phase nanowire one-dimensional semiconductor nanostructures for ap- growth). His nanowires of alternating segments will be plications from energy conversion and storage to catalysis explored for thermoelectric properties, while extremely and photodetection. short segments will be analyzed for “quantum dot-in-wire” properties, e.g, enhanced photoemission from an embed- Publications ded emitter in a wire. Lastly, we have established a capabil- Hollingsworth, J. A., G. Zhang, and R. Laocharoensuk. Con- ity for templating the growth of catalyst-metal arrays for trolling composition and interfaces in solution-grown ordered nanowire growth (polystyrene sphere template semiconductor nanowires. Langmuir (Invited Feature Article). for directing metal deposition prepared using Langmuir- Blodgett sphere assembly). This will be the first demonstra- tion of ordered arrays of nanowires grown by a solution- phase method. Future Work In FY16, the postdoc will continue to target specific, novel nanowire heterostructures (nano-segmented for thermo- electrics and dot-in-wire for single-photon source applica- tion) by solution-phase growth methods. This will entail continuing to pursue significant progress being made in flow-based synthesis as described in the Progress sec- tion. He will also pursue a second option: polyol synthe- sis, where the aim will be to understand the controlling variables in polyol syntheses that determine nanowire diameter and use this understanding to synthesize ultra- thin sub-5 nm wires (currently a key synthetic challenge to access nanowires below 10 nm) and to monitor the impact of molecular “capping agents” or reactive cations on the formation of core/shell or superlattice heterostructures, respectively. Together, these aims address a need in the fields of thermoelectrics and topological insulators for ultrathin bismuth or lead chalcogenide nanowires (e.g., 487

Page 488

Materials for the Future Postdoctoral Research and Development Continuing Project Quantum Control of Tailor-designed Photoactive Energetic Materials Tammie R. Nelson 20140668PRD2 Introduction to compute gradients along the excited state potential A critical need for the prediction and control of explo- energy surfaces. Several numerical tests on represen- sives behavior has been identified as a “grand challenge” tative molecules in solvent environment have been in the area of dynamic and energetic materials science. conducted and the results have been summarized in an Development of photoactive energetic materials that article “Simulations of fluorescence solvatochromisms in optically initiate by a laser light, can address future mile- substituted PPV oligomers from excited state molecular stones for enhanced safety and security of the nuclear dynamics with implicit solvent” recently published in stockpile. Here, photoinitiation requires non-adiabatic Chemical Physics Letters. In terms of applications, using (non-radiative) conversion of excess electronic energy developed theoretical methodology, we have conducted into specific vibrational degrees of freedom mediated nonadiabatic dynamics simulations that show that the by the electron-phonon coupling. Excited states have deactivation mechanism of tetrazine photochemistry in varying electron-phonon couplings, and depending on PetrinTzCl is due to vibrational excitation of the Petrin their localization and energy, lead to different relaxation moiety during internal conversion. Simulation results pathways. This structure provides a “landscape” for opti- along with respective experimental data offer potential cal control (steering the outcome of a chemical reaction for this material to be stable in ambient lighting, yet with light) based on which excitations are selected. Non- possible to initiate with short pulsed lasers. An article adiabatic excited-state molecular dynamics (NA-ESMD) on this subject “Photoactive high explosives: linear and simulations provide a detailed understanding of key nonlinear photochemistry of petrin tetrazine chloride” photoinduced phenomena controlling competing inter- has been recently published in the Journal of Physical actions and relaxation pathways in complex materials. Chemistry C. Benefit to National Security Missions Future Work The proposed work will increase the controllability of We have already made a prototype code incorporating chemical dynamics by predicting and controlling the solvent at the Polarizable Continuum Model (PCM) level. functionality of materials. The proposed work will fur- During the second year of the project, our goal is for- ther LANL’s reputation and prestige in quantum control mulate a combined quantum mechanical and molecular and molecular design. Control of explosive initiation mechanical (QM/MM) approach to treat large systems would be transformational for LANL’s core missions of and include solvent and thermal bath effects will be stockpile safety and energetic materials. developed. QM/MM is a much more accurate approach in which a system is divided into a quantum mechani- Progress cal region and a molecular mechanical region allowing Over the past year, significant changes have been done realistic modeling of dielectric media. Depending on the to the non-adiabatic excited state molecular dynamics molecular system, different force fields will be interfaced (NA-ESMD) codes toward implementation of solvent with our existing non-adiabatic excited-state molecular contributions. A Polarizable Continuum Model (PCM) dynamics (NA-ESMD) code. In terms of applications, has been implemented and numerically tested. This photoactive energetic materials with controllable optical allows us to conduct simulations of excited state dynam- functionality will be studied. Specific vibrational de- ics in the presence of the simple solvent environment. grees of freedom responsible for bond breakage, rapid We have also developed theoretical formalism on how decomposition, and the onset of the exothermic chain reactions relevant to explosive initiation processes in the 488

Page 489

surrounding material will be identified. In the future, we will use the developed methodology to design and propose controllable materials. Detailed numerical NA-ESMD simu- lations to investigate photoinduced pathways, timescales, branching effects, and multiple product formation, which can be experimentally validated by ultrafast spectroscopy capabilities, will be performed. Possible photoinduced reaction coordinates and electron-vibrational relaxation pathways to elucidate mechanisms for optical initiation will be assessed. Moreover, first principle electronic structure calculations will be used to quantitatively evaluate the nonlinear optical responses to understand excited-state properties and to provide design strategy for new photoac- tive energetic materials. Conclusion The proposed implementation of the quantum mechani- cal and molecular mechanical (QM/MM) approach will provide novel computational capabilities critical for under- standing light-induced dynamics in many technologically relevant materials. Specifically, this will allow us to simu- late large molecular systems where full ab-initio calcula- tions are prohibitively expensive, and to describe systems that interact strongly with solvent environments. Such simulations have been done previously for the ground state but were never attempted for excited states due to computational complexity. Publications Bjorgaard, J. A., T. Nelson, K. Kalinin, V. Kuzmenko, K. A. Velizhanin, and S. Tretiak. Simulations of fluorescence solvatochromisms in substituted PPV oligomers from excited state molecular dynamics with implicit solvent. 2015. Chemical Physics Letters. 631: 66. Greenfield, M. T., S. D. McGrane, C. A. Bolme, J. A. Bjor- gaard, T. R. Nelson, S. Tretiak, and R. J. Scharff. Pho- toactive high explosives: linear and nonlinear photo- chemistry of petrin tetrazine chloride. 2015. Journal of Physical Chemistry A. 119: 4846. Nelson, , Fernandez-Alberti, A. E. Roitberg, and Tretiak. Nonadiabatic Excited-State Molecular Dynamics: Mod- eling Photophysics in Organic Conjugated Materials. 2014. ACCOUNTS OF CHEMICAL RESEARCH. 47 (4): 1155. 489

Page 490

Materials for the Future Postdoctoral Research and Development Continuing Project Ultrafast Carrier Dynamics in Novel Two-Dimensional Nanomaterials Victor I. Klimov 20140675PRD3 Introduction project could potentially benefit MaRIE. Semiconductor nanomaterials are a versatile materials Scientific Discovery and Innovation/Basic Understand- platform for exploring novel physical phenomena not ing of Materials. Through ultrafast laser spectroscopy of found in their bulk counterparts. A detailed understand- these semiconductor nanosheets, we will establish the ing of the physical factors controlling the electronic prop- role of two-dimensional quantum confinement on in- erties will allow for the development of materials with traband carrier cooling, Auger recombination and other tailored electronic properties. Recent advances in syn- multi-carrier interactions, significantly contributing to thetic methods have allowed for the synthesis of nearly the basic understanding of these novel nanomaterials. atomically thin two-dimensional nanoplatelets. These materials have already shown promise for use in next The materials developed in this project can be used as generation transistors as well as other electronic applica- spectrally tunable light harvesters in photoconversion in- tions. However, the underlying electronic properties of cluding both photovoltaics and generation of solar fuels. these materials remain largely unexplored. Specifically, it This makes this project directly relevant to the LANL’s still remains largely unknown how the effect of confine- Energy Security and Environment missions. ment in only one dimension is different from the effects of the two-dimensional and three-dimensional confine- Progress ment realized, respectively, in nanorods and quantum Our initial target materials in this project were transition dots. To this end, we will conduct a side-by-side study of metal chalcogenide layered semiconductors that were materials of different dimensionalities with femtosecond recently isolated in the form of single-layered plates. We spectroscopy. Experiments will elucidate the time scales anticipated that following the methodology outlined in of carrier relaxation, exciton binding energies, as well the published literature would enable us to prepare col- as the strength of multicarrier interactions. Due to the loidal suspensions of single-layered MoS2 and character- quasi-two-dimensional confinement of the electrons in ize them with femtosecond optical spectroscopy. How- the material, the Coulomb interactions between carriers ever, the sample preparation did not produce materials are expected to be significantly enhanced relative to ex- of sufficient quality for spectroscopic studies, with the isting semiconductor nanocrystals. The strong Coulomb suspensions containing a mixture of plates of differing interaction can give rise to room temperature excitonic thicknesses. Therefore, in FY15, we instead focused on a phenomena and enhance multicarrier processes that are different class of nanomaterials, namely PbSe. We syn- typically only observed at extremely low temperatures. thesized two-dimensional platelets of varying thickness of sufficiently high quality for optical studies. Femtosec- Benefit to National Security Missions ond transient absorption and time resolved photolumi- Agency Relevance/DOE/SC. The primary goal of this nescence experiments were utilized to characterize the project is to explore the electronic properties of a new interaction strength between carriers as well as the rate class of two-dimensional semiconductor materials. As of carrier cooling. Additionally, using a hyperbolic effec- such, the results will be valuable to the DOE’s Office of tive mass model we were able to explain the size de- Science. pendence of the band gap as a function of the thickness of the material. Furthermore, we were able to quantify Mission Relevance/MaRIE. Since we are studying charge the magnitude of the carrier-carrier interactions in bulk, carrier interactions at extremely high densities in these two-dimensional plates, one-dimensional nanorods, and two-dimensional semiconductors, the insights from this 490

Page 491

zero-dimensional quantum dots. Interestingly, we found that one-dimensional materials exhibited the strongest interaction between carriers through a combination of di- electric and geometrical effects. These results are currently being prepared for publication. Future Work This project focuses on better understanding the role that dimensionality plays in controlling the optical and electron- ic properties of nanometer-sized objects. In particular, we are interested in how the interactions between electrons are modified when the size of a material was confined to the nanoscale in one (plates), two (rods), and three (dots) dimensions. The extremely thin two-dimensional materi- als facilitate increased interactions between charge car- riers, which should result in novel physical phenomena. To gain insight into the new physics that arise in strongly confined two-dimensional materials we will utilize femto- second transient absorption spectroscopy in conjunction with time-resolved photoluminescence experiments to probe the carrier dynamics. Through tuning of the power of the excitation beam, the dynamics of multicarrier processes will be measured. Due to the increased interac- tions between carriers we expect to find a corresponding increase in the multicarrier recombination rate, as well as potentially bound multiexciton states that are typically only observed at temperatures approaching absolute zero. Tuning of the probe wavelength across the spectra will al- low for the relaxation process to be directly monitored as well as resolving the energies of higher-lying excited states. Saturation experiments will also reveal the degeneracy of the transitions, providing a detailed understanding of the energy level structures that lead to the unique properties found in ultra thin semiconductor layers. Conclusion These studies of two-dimensional nanostructures will help elucidate how the precise shape and size of the nanoma- terial control their optical and electronic properties. We expect that the two-dimensional confinement of charge carriers will enable increased interactions between them leading to novel physics. The insights gained from this project will advance our understanding of the fundamental electronic properties of two-dimensional materials and facilitate their applications in real-life technologies. Publications Gao, J., A. F. Fidler, and V. I. Klimov. Carrier multiplication detected through transient photocurrent in device- grade films of lead selenide quantum dots. 2015. Na- ture Comm.. 6: 8185. 491

Page 492

Materials for the Future Postdoctoral Research and Development Continuing Project New Room Temperature Multiferroic Thin Films Enabled by Strain Engineering Quanxi Jia 20140676PRD3 Introduction Progress Multiferroic materials have attracted extensive inter- Owing to the potential applications in the development ests due to their potential technological applications of high performance solid-state devices, multiferroics as multifunctional devices. However, the single-phase materials have recently attracted a great deal of atten- multiferroic materials are rare because of the distinct tion. Single-phase multiferroic materials are very scarce nature of magnetism and ferroelectricity. Seeking nano- due to their exclusive nature origins. Another way to composite multiferroic materials, therefore, has become achieve multiferroicity is by combining individual fer- our recent research focus. Strain engineering provides roelectric and ferromagnetic materials into composites. an important approach to yield new functional materi- This allows to choose components with large ferroelec- als with novel properties. We have demonstrated a new tric properties (d33 and g33) and magnetostriction, so class of multiferroics (BiFe0.5Mn0.5MnO3 or BFMO) the resulting composites may have superior multiferroic based on two partially miscible phases of BiFeO3 and properties and magnetoelectric couplings. BiMnO3 using epitaxial strain. Fully strained BFMO/LaA- The following highlights our accomplishments from Aug. lO3 (LAO) superlattices with controlled layer thickness 2014 to June 2015. and periodicity could be a very promising way to realize new room temperature multiferroic materials. We will We have fabricated CoFe2O4 films with different thick- take advantage of the well-equipped capabilities at Los nesses, grown on MgO substrates. The strain effect on Alamos National Laboratory (such as laser molecular magnetism and magnetic anisotropy were studied. We beam epitaxy, high resolution x-ray diffraction, transmis- found that strain plays a critical role in determining the sion electron microscopy, and transport characterization ferromagnetism of CoFe2O4 films. The magnetostric- tools) to systematically study the highly strained BFMO/ tion coefficient of CFO films is estimated to be λ[001] = LAO superlattice structures. −188×10−6. This study is important for the design and growth of CoFe2O4 based multiferroic superlattices such Benefit to National Security Missions as BaTiO3:CoFe2O4. This project focuses on fabrication, properties, and microstructural characterization of advanced nanostruc- We grew Bi2FeMnO6 (BFMO) thin films with both tured superlattice with an emphasis on room tempera- conventional pseudocubic structure and novel super- ture multiferroic materials. The proposed effort supports cell structure on SrTiO3 (001) substrates with different and strengthens the Laboratory’s core scientific capabili- thicknesses of CeO2 buffer layers (ranging from 6.7 to ties essential to discovering, understanding, and exploit- 50.0 nm) using pulsed laser deposition. The correlation ing emergent phenomena in materials. This project between the thickness of the CeO2 buffer layer and the also directly addresses the LANL Energy Security and structure of the BFMO films shows that the CeO2 buffer Environment mission areas as well as the Grand Scien- layer, as thin as 6.7 nm, is sufficient in triggering the nov- tific Challenge identified in the DOE BESAC report: how el BFMO supercell structure. The buffer layer thickness do remarkable properties of matter emerge from the is found to be critical in controlling the microstructure complex correlations of atomic or electronic constituents and magnetism of the BFMO supercell structures. Thin and how can we control these properties? seed layers can produce a smoother interface between the BFMO film and the CeO2 buffer layer, and there- fore better ferrimagnetic properties. Our results have 492

Page 493

demonstrated that strain and interface could be utilized to Choi, E. M., J. E. Kleibeuker, T. Fix, J. Xiong, C. J. Kinane, generate novel thin film structures and to tune the func- D. Arena, S. Langridge, A. P. Chen, Z. Bi, J. H. Lee, H. tionalities of thin films. Wang, Q, X. Jia, M. G. Blamire, and J. L. MacManus- Driscoll. Interface coupled BiFeO3/BiMnO3 superlat- We have optimized the growth and characterized the tices with magnetic transition temperature up to 410 growth rate of perovskite thin films in laser MBE system by K. Adv. Mater.. the advanced RHEED system. Zhang, W., A. P. Chen, M. Li, L. Li, Z. Xia, P. Boullay, L. Wu, Y. Zhu, J. L. MacManus-Driscoll, Q. X. Jia, and H. Wang. Future Work Two-dimensional layered oxide structures tailored by Fully strained superlattices with controlled layer thickness self-assembled layer stacking via interfacial strain. Adv. and periodicity provide very promising ways to realize Funct. mater.. new room temperature multiferroic materials. The well- equipped capabilities at CINT (such as laser molecular beam epitaxy, high resolution x-ray diffraction, transmis- sion electron microscopy, and transport characterization tools) will enable us to carry out the following proposed FY16 tasks with success. We will grow high quality multiferroic superlattices such as BaTiO3:CoFe2O4 and La0.7Sr0.3MnO3:BaTiO3 by using the Laser MBE system. With the RHEED system, we are able to monitor the layer by layer growth and we can design the interface with atomic sharpness. We will perform magnetoelectric coupling (ME) measure- ments of these superlattices using VSM and SANS. As this project will be complete at the end of July 2016, we will wrap up this project and explore new funding opportu- nities. Conclusion This project focuses on fabrication, properties, and micro- structural characterization of advanced nanostructured su- perlattice with an emphasis on room temperature multifer- roic materials. We expect that the structure-functionality relationship of this new class of multifunctional materials can be established at the end of the project. Publications Ahmed, T., A. P. Chen, D. A. Yarotski, S. Rudin, S. A. Trug- man, Q. X. Jia, and J. X. Zhu. First-principles study of magnetic, electronic and optical properties of double perovskite Bi2FeMnO6. Phys. Rev. B. Ahmed, T., A. P. Chen, Q. X. Jia, D. A. Yarotski, and J. X. Zhu. Site dislocation between Fe and Mn sites in double perovskite Bi2FeMnO6:A theoretical explanation for XMCD anomaly. To appear in Scientific Reports. Chen, A. P., W. Zhang, J. L. MacManus-Driscoll, H. Wang, M. R. Fitzsimmons, and Q. X. Jia. Role of interfaces on competing interactions of ferroic films. Invited pre- sentation at 2015 MRS Fall meeting. (Boston, Nov. 29

• Dec. 4). 493

Page 494

Materials for the Future Postdoctoral Research and Development Continuing Project Search for the Topological States in F-electron Systems Tomasz Durakiewicz 20140678PRD3 Introduction tor, Sr-doped BiSe2. Data reduction and analysis on The broad category of correlated electron systems, with those systems is ongoing. His work in this period of time eminent examples like high-temperature superconduc- resulted in several papers published or at various stages tors, also includes a family of f-electron compounds of preparation; see a full list of papers published and in based on lanthanides (4f) or actinides (5f). This area of submission for 2015. research is a LANL specialty. This proposal bridges the Madhab also organized and run several synchrotron LANL expertise in basic physical properties of f-electron based experiments at various beamlines: 4 runs at materials, and the cutting edge novel science of topolog- Advanced Light SOurce in Berkeley (Feb, Mar and twice ical surface states so far known only for non-f electrons. in Apr) as well as Univ of Tokyo in May and Swiss Light An extensive Angle-Resolved Photoelectron Spectros- Source in June. copy (ARPES) study of several candidate systems will be performed, aimed at identification and characterization Future Work of such states. A spin-resolved variant and temperature- During the first year of the project a series of Angle- dependent studies will provide evidence for spin texture, Resolved Photoelectron Spectroscopy (ARPES) studies an important and now still missing indicator of topologi- will be performed in specialized synchrotron facilities. cal protection. Results will be directly applicable to a The goal is twofold: to perform a detailed temperature wider set of problems found in correlated systems and

• dependent study with high energy resolution in order their surface properties. to resolve the band dispersion of the surface states and and to work on the spin-resolved variant in search of Benefit to National Security Missions signatures of spin texture. These measurement will ulti- This project has direct ties to basic research of novel mately provide evidence by identifying (i) the massless materials, and advancing our ability to provide design (linear) band dispersion and (ii) spin texture characteris- principles for future materials. It is linked to e.g. NSF tic for topological surface states. funding efforts in the area of topological insulators. By focusing on f-electrons it is relevant to enhancing and During the second year, the results of the first year’s control of the transport properties of lanthanides and work will be prepared for publications. Additional data actinides, including nuclear fuels. will be collected if missing information about the elec- tronic structure in specific parts of energy-momentum Progress space is found. During the first six months of his appointment, between Jan 2015 and June 2105, Madhab Neupane was focusing Conclusion on (1) taking photoemission data at synchrotron facili- It is expected that this project will lead to (i) proving or ties, (2) analyzing data and (3) writing manuscripts. disproving the existence of topological surface states in f-electron systems, (ii) measuring the temperature evolu- The search for novel topological surface states and tion of the hybridization gap and analyzing its influence novel properties of materials being hosts to such states and role in forming the surface states, (iii) performing an is ongoing. Madhab measured, for the first time, the detailed spectral analysis in order to obtain information electronic structure of a novel system BiPd, contin- about the renormalization energy scales and (iv) extend- ued measurements of SmB6 and CeB6 and performed ing the search for surface states towards a broader class initial measurements of the topological superconduc- 494

Page 495

of f-electron systems. Publications Chang, , Das, Chen, Neupane, Xu, M. Z. Hasan, Lin, Jeng, and Bansil. Two distinct topological phases in the mixed-valence compound YbB6 and its differences from SmB6. 2015. PHYSICAL REVIEW B. 91 (15). Ghimire, N. J., Luo, Neupane, D. J. Williams, E. D. Bauer, and Ronning. Magnetotransport of single crystalline NbAs. 2015. JOURNAL OF PHYSICS-CONDENSED MAT- TER. 27 (15). Huang, , Xu, Belopolski, Lee, Chang, Wang, Alidoust, Bian, Neupane, Zhang, Jia, Bansil, Lin, and M. Z. Hasan. A Weyl Fermion semimetal with surface Fermi arcs in the transition metal monopnictide TaAs class. 2015. NA- TURE COMMUNICATIONS. 6. Neupane, , Xu, Alidoust, Bian, D. J. Kim, Liu, Belopolski, T. -. Chang, H. -. Jeng, Durakiewicz, Lin, Bansil, Fisk, and M. Z. Hasan. Non-Kondo-like Electronic Structure in the Correlated Rare-Earth Hexaboride YbB6. 2015. PHYSI- CAL REVIEW LETTERS. 114 (1). Neupane, , Xu, Alidoust, Sankar, Belopolski, D. S. Sanchez, Bian, Liu, Chang, Jeng, Wang, Chang, Lin, Bansil, Chou, and M. Z. Hasan. Surface versus bulk Dirac state tuning in a three-dimensional topological Dirac semimetal. 2015. PHYSICAL REVIEW B. 91 (24). Xu, , Liu, S. K. Kushwaha, Sankar, J. W. Krizan, Belopolski, Neupane, Bian, Alidoust, Chang, Jeng, Huang, Tsai, Lin, P. P. Shibayev, Chou, R. J. Cava, and M. Z. Hasan. Observation of Fermi arc surface states in a topological metal. 2015. SCIENCE. 347 (6219): 294. Xu, , Neupane, Belopolski, Liu, Alidoust, Bian, Jia, Landolt, Slomski, J. H. Dil, P. P. Shibayev, Basak, Chang, Jeng, R. J. Cava, Lin, Bansil, and M. Z. Hasan. Unconventional transformation of spin Dirac phase across a topological quantum phase transition. 2015. NATURE COMMUNI- CATIONS. 6. Xu, S. Y., I. Belopolski, N. Alidoust, M. Neupane, G. Bian, C. Zhang, R. Sankar, G. Chang, Z. Yuan, C. C. Lee, S. M. Huang, H. Zheng, J. Ma, D. S. Sanchez, B. K. Wang, A. Bansil, F. Chou, P. P. Shibaye, H. Lin, S. Jia, and M. Z. Hasan. Discovery of a Weyl Fermion semimetal and to- pological Fermi arcs. 2015. Science. : 1. 495

Page 496

Materials for the Future Postdoctoral Research and Development Continuing Project Rational Design of Multiferroics and Influence of Cationic Disorder on Multiferroicity in Perovskites Blas P. Uberuaga 20140679PRD3 Introduction not well understood at present and is specifically tar- Ferroelectricity and magnetism in solids have seemingly geted in this work. different origins. While magnetism is related to ordering Benefit to National Security Missions of spins of electrons in incomplete ionic shells, ferroelec- tricity results from a delicate balance between long- This research is of direct relevance to the missions above range coulombic dipole-dipole interactions and short- as the ability to rationally design multiferroics will en- range repulsive forces. Materials in which ferroelectricity able novel and revolutionary future energy, sensing, and and ferromagnetism exist simultaneously are known as information technologies in which the low power and multiferroic materials. Multiferroics can, in principle, en- high speed of field-effect electronics are combined with able novel and revolutionary future energy, sensing, and the permanence and routability of voltage-controlled information technologies. However, ferroelectric ferro- ferromagnetism. This paradigm has the potential to magnets are exceedingly rare. This research effort seeks mitigate the costs, risks and time involved in preparing to enable rational design of novel thin-film multiferroic and testing of multi-functional materials. Further, this materials. research should yield insights into the fundamental fac- tors underlying materials behavior, providing the basic The research team is developing a design strategy to understanding needed to further develop novel materi- systematically navigate through the chemical space of als and material architectures for advanced energy ap- magnetic perovskite oxides (ABO3 systems) using quan- plications. Finally, an atomic-level understanding of the tum mechanical computations in combination with ther- influence of crystalline disorder on multiferroic behavior modynamics based models and materials informatics. would directly translate to better device performance. State-of-the-art computations are being used to explore and understand coupling between ferroelectricity and These goals are of direct relevance to the missions of magnetism for a range of ABO3 compounds as a func- DOE/BES and LANL, particularly the laboratory’s science tion of composition, strain and nominal charge states pillar of Materials for the Future. of A and B site cations. Advanced density functional Progress theory based computations such as those with hybrid functionals (to appropriately account for the self-energy Over the past year, we have built machine learning of a many-body system of electrons) are being used for a models to examine various aspects of material discovery. further careful investigation and validation of the identi- In particular, we have examined important properties of fied promising material candidates with multiferroic perovskite materials as these hold particular promise for behavior. a number of applications involving multiferroic proper- ties and fast ion conduction. We highlight three papers A key objective of this effort is to assess the effect of that emphasize this work. crystalline cationic disorder on the magnetism, ferro- First, in a Journal of Applied Physics paper, we showed electricity and their coupling in perovskites. Crystalline how one can combine two perovskite materials (with disorder in magnetic perovskites is expected to strongly chemical formula ABX3, where A and B are cations and influence both the local magnetic structure as well as X is an anion, in this case Cl) to form a double perovskite the polar phonon modes. However, the atomic scale that emphasized the strengths of each and minimized mechanisms through which crystalline cationic disorder the weaknesses. In particular, we showed how combin- and multiferroicity couple in magnetic perovskites are 496

Page 497

ing CsCaCl3 and RbZnCl3 lead to a stable cubic double it lends itself as an ideal test-bed for materials informatics perovskite that had a strong ferroelectric response to based approaches involving big-data and machine learn- strain. By them selves, CsCaCl3 is a very stable cubic ing methods. We plan to use the data generated in the perovskite that shows little sensitivity to strain while quantum mechanical computations to develop efficient, RbZnCl3 is unstable and easily distorts. By combining the robust and validated machine learning models for property two, we showed how a perovskite that is stable in the predictions for perovskites and double perovskites, which cubic structure but undergoes ferroelectric distortions can then be used in a high throughput screening studied under small strains leads to a new material with potentially to identify potential material candidates with a specific ap- improved functionality. plication in mind. In a second paper, we used accelerated molecular dy- In parallel to these bulk studies, we will also focus on namics methods to examine the kinetic pathways for interface structures in complex oxides. Charge transfer oxygen vacancies in double perovskites as a function of effects at hetero-interfaces between two appropriately the cation ordering. Because of the structure of these chosen ferroelectric materials can also lead to an inter- materials, there are 9 high symmetry orderings of the A face magnetization. If the interface magnetization couples and B cations. We examined how oxygen vacancies, the strongly with either strain or polarization in either side of important carrier of ionic conductivity in these materials, the materials, an efficient multiferroic device can be real- move through the material for the different cation order- ized. This mechanism has not been explored in detail to ings. We find that both the absolute mobility as well as the date. Furthermore, presence or absence of misfit disloca- nature (one-dimensional or three-dimensional) depends tions or controllable point defects (such as O vacancies) on significantly on the cation ordering. In particular, there are ferroelectricity, ferromagnetism and their coupling will be orderings in which the mobility is higher than in either of studied. the component single perovskites (in this case, SrTiO3 and LaAlO3). This work is under review at Chemistry of Materi- Conclusion als, where the first round of reviews were very favorable. We are systematically screening the chemical space of magnetic perovskites to identify promising multiferro- These two papers showed the promise of double ics. This research effort will deliver new insights into the perovskites for advance functionality, but did not demon- factors that govern coupling between ferroelectricity and strate their discovery. In a third paper accepted to Phys. ferromagnetism in perovskite materials. By systematically Rev. B, we build and critically access state-of-the-art ma- exploring strain-(polar) phonon coupling and spin-phonon chine machine approaches for materials classification and coupling in a range of material systems, we will identify property predictions by taking a specific example of classi- and understand mechanisms that control spin-phonon- cal AB crystalline solids. Once validated, these approaches strain behavior in these materials. An atomic-level under- are now being applied to perovskite systems. In a fourth standing of the influence of crystalline disorder on multi- paper submitted to Acta Crystallographica, we have built a ferroic behavior would directly translate to a better device machine learning approach that examines those “features” performance and an improved control over the functional- of perovskites that predict their formability. Importantly, ity. predicting which perovskite chemistries lead to stable perovskites with simple heuristics significantly accelerates Publications the discovery process. We found that new features that Kim, C., G. Pilania, and R. Ramprasad. From organized high- were not previously considered do a better job of predict- throughput data to phenomenological theory: the ex- ing formability than those previously considered. This ample of dielectric breakdown. Chemistry of Materials. demonstrates the power of the approach. Kumar, A., G. Pilania, T. D. Huan, T. Lookman, and R. Ram- prasad. Informatics-driven strategy for the accelerated Future Work design of dielectrics. Scientific Reports. In the coming fiscal year, we plan to build on our on- going work by considering double perovskites–ordered Lookman, T., P. V. Balachandran, D. Xue, G. Pilania, T. Shear- perovskites with multiple A and/or B site cations. These man, J. Theiler, J. E. Gubernatis, J. Hogden, K. Barros, materials can be synthesized with a variety of cation order- E. BenNaim, and F. J. Alexander. A perspective on ings, each of which shows a different tendency towards materials informatics: state-of-the-art and challenges. exhibiting ferroelectric polarization and it’s coupling with Information science for materials discovery and design. strain. Furthermore, due to enormous combinatorial pos- Edited by Lookman, T.. sibilities available in the double perovskite chemical space, Pilania, , J. E. Gubernatis, and Lookman. Structure classifi- 497

Page 498

cation and melting temperature prediction in octet AB solids via machine learning. 2015. PHYSICAL REVIEW B. 91 (21). Pilania, , and B. P. Uberuaga. Cation ordering and effect of biaxial strain in double perovskite CsRbCaZnCl6. 2015. JOURNAL OF APPLIED PHYSICS. 117 (11). Pilania, , and Lookman. Electronic structure and biaxial strain in RbHgF3 perovskite and hybrid improper ferro- electricity in (Na,Rb)Hg2F6 and (K,Rb)Hg2F6 superlat- tices. 2014. PHYSICAL REVIEW B. 90 (11). Pilania, G., J. E. Gubernatis, and T. Lookman. Classification of octet AB-type binary compounds using dynamical charges: a materials informatics perspective. Scientific Reports. Pilania, G., P. V. Balachandran, J. E. Gubernatis, and T. Look- man. Predicting the formability of ABO3 perovskite solids: An initial machine learning study. 2015. Acta Crystallographica Section B. 71: 507. Uberuaga, B. P., and Pilania. Effect of Cation Ordering on Oxygen Vacancy Diffusion Pathways in Double Perovskites. 2015. CHEMISTRY OF MATERIALS. 27 (14): 5020. 498

Page 499

Materials for the Future Postdoctoral Research and Development Continuing Project Studies on Functional Materials: Design and Optimization Turab Lookman 20140682PRD4 Introduction Thus, it strengthens existing connections and will forge Accelerated materials discovery and design is the cor- other connections, such as the use of coherent x-ray nerstone of the US Materials Genome Initiative (MGI). diffraction at LCLS/SLAC on functional materials, which Costly time-consuming trial and error methods have connects to LANL’s signature facility concept, MaRIE. This been the dominant approaches for searching for new project goes beyond concurrent DR work in that its focus materials with desired, target properties. However, is on fundamental science governing optimal properties computational material design based on data mining and the use of large scale facilities (such as CINT, LCLS) to techniques and informatics tools are increasingly play- probe mechanisms. In a more general sense, accelerated ing an important role in the discovery and design efforts. materials discovery, if we can demonstrate its effective- Very few examples exist and this work will provide spe- ness, is an approach to develop new materials for a cific examples for given materials classes that combine broad range of missions including sensor and structural experiments, data and inference modeling. The implica- materials for nuclear nonproliferation, and replacement tions for applications of finding materials with targeted materials for stockpile stewardship. properties are enormous. The focus will be on ferroic Progress materials, which form an essential family of functional (or smart) materials, which combine polarization, mag- Training and using state-of-art pulse laser deposition netization and/or electronic properties to yield a non- (PLD) equipment in the DOE facility, CINT, took up a linear response to external stimuli such as temperature, substantial fraction of the year. Towards the latter part, pressure, electric, and magnetic fields. The huge de- a number of thin film systems have been fabricated and mand to replace PZT (PbZrxTi1-xO3) containing Pb with their properties measured. These included manganites Pb-free piezoelectrics, the attempts to minimize energy on a number of substrates. The magnetization response dissipation during thermo-mechanical energy conver- was also characterized. Work on using novel substrates, sion in shape memory alloys (SMAs), and the efforts to such as the shape memory alloy, NiTi, has now begun. generate greater temperature differences in electrocalo- Having ensured that we are fabricating well controlled ric materials, are examples of intensive efforts involving thin films, the set is stage for synthesizing functional functional materials design. materials and discovering systems with targeted proper- ties, such as films with optimal dielectric tenability and Benefit to National Security Missions new electrocalorics. The work being proposed is at the interface of two of Another major effort that was completed over the LANL’s pillars, namely Materials for future and IS&T course of the year was synthesizing and characterizing (Information Science and Technology) and is relevant to new NiTi based alloys with very low thermal dissipation. BES/DOE and the Office of Science initiatives under the This effort was was synergistic with the LDRD-DR in ma- “Materials Genome Initiative (MGI)”. Ongoing LDRD- terials informatics. The inference and machine learning DR funded work at LANL is addressing how accelerated models studied there predicted a number of alloys with materials discovery might be realized. Several materi- anticipated targeted properties. Compounds were then als classes are under investigation, principally based on synthesized based on these predictions and their proper- perovskites. The proposal complements these efforts in ties measured. A number of new alloys (e.g. NiTiCuFePd) using the codesign strategy on materials such as electro- with better thermal dissipation properties than known caloric materials, important in solid-state refrigeration, previously in the literature have been synthesized and and energy harvesting materials such as ferroelectrics. 499

Page 500

the new discovery is now being prepared for publication. celerated search for materials with targeted properties Such alloys are important as they minimize fatigue under by adaptive design. Nature Communications. thermal and mechanical cycling and have widespread use Xue, D., R. Yuan, Y. Zhou, D. Xue, T. Lookman, X. Ding, and in industry. J. Sun. Design of High Temperature Ti-Pd-Cr Shape Memory Alloys with Small Thermal Hysteresis and High Future Work Reversibility. Scientific Reports (2015). The focus in the second year will be on layered systems. Experiments will be developed using Pulsed Laser Deposi- tion (PLD), which with existing data and statistical infer- ence methods, will guide materials design. This work will complement other work using ab initio methods. I will first focus on dielectric systems to search for thin films (Barium Strontium based) deposited on specific substrates to find dielectrics with better tunable characteristics. I will also focus on developing a strategy for studying the electrocalo- ric effect in layered systems. The electrocaloric effect, the adiabatic temperature change that occurs in polar mate- rials upon application of an electric field, is a promising route to solid-state refrigeration. My approach will be to optimize the large response of the polarization to electric field around the phase transition, as up to now the tem- perature window is too limited. I will therefore investigate bilayers of ferroelectrics with different transition tem- peratures (e.g. by doping with different amounts of Zr4+ into BaTiO3) in order to obtain a wider temperature range without sacrificing a large electrocaloric response (change in temperature induced by the electric field). The data will comprise thickness of layers, dopant concentrations and types and aspects of structure and chemistry, and will be assembled via a suite of experiments guided by statistical methods. Conclusion The research will focus on utilizing statistical methods to accelerate the discovery process with an emphasis on dielelectrics, ferroelectrics and shape memory alloys (SMAs). Augmented by well-established phenomenological approaches in the field, such as Landau theory, together with statistical and machine learning methods, as well as experiments in bulk and on layered systems, the work will demonstrate the essential components needed for finding materials with targeted properties. Publications Balachandran, P. V., D. Xue, J. E. Hogden, J. Theiler, and T. Lookman. Adaptive design strategies for materials de- sign. Scientific Reports (2015). Xue, D., O. V. Balachandran, H. Wu, R. Yuan, Y. Zhou, X. Ding, J. Sun, and T. Lookman. Designing vertical mor- photropic phase boundaries in BaT iO3-based lead-free piezoelectrics. Science Reports (2015). Xue, D., P. V. Balachandran, J. E. Hogden, and J. Theiler. Ac- 500

Page 501

Materials for the Future Postdoctoral Research and Development Continuing Project Probing and Controlling the Surface States of Topological Insulators Scott A. Crooker 20140683PRD4 Introduction is based on the very recent discovery of a new class of Topological Insulators (TIs) are a recently discovered so-called “topologically protected” materials, in which class of materials that have garnered significant atten- scattering and dissipative processes are prohibited from tion in the last 5 years within the condensed matter fundamental physics considerations. The experimental physics community. They are bulk semiconductors such techniques that we will develop as part of this project as Bi2Se3 and Bi2Te3 in which spin-orbit coupling gives will advance new Laboratory capabilities that underpin rise to a topologically protected two-dimensional (2D) a wide variety of Laboratory missions related to Materi- metallic surface, with a surface electronic structure simi- als Science and Controlled Functionality in Materials. lar to the Dirac cone spectrum of graphene. The surface Beyond direct scientific impact, these studies have the states are protected from backscattering by time reversal potential to pave the pathways for new “topologically symmetry, and therefore TIs have potential applica- protected electronics”; that is, transport of charge and tions for spin logic devices and low-power electronics. spin in new generations of lower-power and more ef- Although there is tremendous excitement and interest ficient electronic devices, which has direct impacts on in these novel materials within the physics and materials securing our energy future. science community, the exploration of exotic spintronic Progress properties in TIs such as spin photocurrent and current- induced spin-density properties are still at an early In FY16, Director’s-funded postdoc Luyi Yang built an stage. The ability to control the topological surface states experimental setup for measuring the dynamics of spin electrically or optically is the crucial key to the develop- relaxation in the new atomically-thin semiconductors ment of spin-based devices. such as MoS2 and WSe2 (molybdenum disulphide and tungsten diselenide). This experimental capability is now Luyi will probe, and ultimately manipulate, the spin fully functional, and involves the use of a tunable dye degrees of freedom in TIs. A major part of this research laser, and also an ultrafast optical parametric oscillator will combine the unique optical and pulsed magnetic laser for time-resolved studies. She has successfully transport techniques at the National High Magnetic Field applied this technique to measure for the first time the Laboratory at LANL to characterize spin structure and very long spin relaxation in these new 2D materials. Her transport. In TIs it is possible to photoexcite currents paper on this subject has been accepted at Nature Phys- whose charge, spin, and direction can be controlled by ics (July 2015); the first round of referee reports being varying the helicity of the light. The subsequent spin cur- quite favorable. In addition, Luyi published a paper in rents and spin dynamics can be monitored and imaged Scientific Reports in April 2015 (“Cross-correlation spin via the magneto-optical Kerr effect. The unique nature of noise spectroscopy of heterogeneous interacting spin spin-momentum locking in the SSs suggests that the spin systems”). current in a TI can be generated with the same efficiency as charge photocurrent in a conventional insulator. Future Work In FY16 we will focus primarily on constructing the Benefit to National Security Missions infrared Magneto-Optical Kerr imaging microscope that This project will build Laboratory capabilities in the will allow us to directly image the spin currents in these areas of Materials Science, Emergent Phenomena, and topological insulator materials. This will require the Materials-by-Design, and also in novel measurement purchase and benchmarking of a variety of mid-infrared methods that enable new scientific discovery at LANL. It quantum cascade lasers. The first step will be to per- 501

Page 502

form magneto-reflectivity and absorption studies, using these lasers, of a variety of TI materials, which will allow us to screen for the best ones. Conclusion The principal goal of this project is to image, via magneto- optical techniques pioneered at the NHMFL-Los Alamos, the spin currents that are predicted to form at the surface of TI materials. The results will allow direct visualiza- tion and spatial mapping of the topologically protected electronic states for the first time. Beyond direct scien- tific impact, these studies have the potential to pave the pathways for new “topologically protected electronics”; that is, transport of charge and spin in new generations of lower-power and more efficient electronic devices, which has direct impacts on securing our energy future. Publications Roy, D., L. Yang, N. A. Sinitsyn, and S. A. Crooker. Cross- correlation spin noise spectroscopy of heterogeneous interacting spin systems. 2015. Scientific Reports. 5: 9573. Yang, L., N. A. Sinitsyn, J. Yuan, J. Zhang, J. Lou, and S. A. Crooker. Long-lived nanosecond spin relaxation and spin coherence of electrons in monolayer MoS2 and WS2. 2015. Nature Physics. 11: 830. 502

Page 503

Materials for the Future Postdoctoral Research and Development Continuing Project Three-Dimensional Nitrogen-Doped Porous Nanographene for High- Performance Supercapacitor Hsing-Lin Wang 20140684PRD4 Introduction ratory and technologies well established at LANL. For the past two decades, graphene has emerged as one Benefit to National Security Missions of the most exciting carbon based materials due to its superior electrical conductivity, exceptional high surface Controlled synthesis of 3D porous nitrogen-doped area, and excellent structural stability. In recent years, nanographene with defined flake sizes and doping (type, graphene has revealed tremendous potential as an ideal content, and location) will enhance graphene electron electrode material in energy storage devices such as lith- mobility and the as-prepared 3D porous structure, which ium ion battery and supercapacitor. However, graphene is desired to overcome the restacking and defects ob- suffers from low processibility, difficult to scale up, and served in traditionally-reduced graphene oxides sheets. lack of tunability in structure and properties. In addi- We expect to design and synthesize novel and high- tion, as part of the anode materials in lithium ion battery performance 3D porous nitrogen-doped nanographene and supercapacitors, graphene sheets often aggregate supercapacitors with optimal chemical and structural and restack with increasing operation time and charg- properties to maximize specific capacitance, power ing/recharging cycles; this structure instability leads to a density, energy density, and cycle life, relevant to LANL’s decrease in surface area and deter the electrolyte from Energy Security and Environment missions. If successful, accessing to graphene sheets surface. To mitigate the this project will result in a deep understanding of the processibility, graphene are oxidized and then reduced electrochemistry of 3D porous nanographene as well to form the so called “reduced graphene oxide (RGO), as the invention of the first 3D porous nitrogen-doped which has limited solubility in certain organic solvents. nanographene-based supercapacitor. The results will However, this process also induced many structural de- have wide-spread impact on the fabrication of high-per- fects that hamper the performance of in energy storage formance supercapacitors. devices. Progress Our approach is to use three dimensional nanogra- phenes synthesized in our laboratory and crosslinking In FY15, we developed a novel approach to synthesize a these small nanographene moieties to form mesoporous series of three-dimensional (3D) nanographenes (NGs) nanographene which will have a well-defined molecular with well-defined molecular structures; in our synthe- structure, high surface area and most importantly struc- sis, a variety of functional groups have been covalently tural homogeneity. The design of this novel mesoproous attached to the NG flakes that enable fine-tuning of nanographene is expected to exhibit very high capaci- electron density on flakes and d-spacing between flakes. tance and operational stability over a very long period of The multistep organic synthesis of 3D NG is designed time. Our system is the first ever demonstrated to use to overcome the restacking observed in traditional rGO mesoporous nanographene in energy storage devices. sheets and achieve graphene’s intrinsically high Li stor- In addition, we will develop ways to dope these porous age capacity and diffusivity while simultaneously ad- ographenes to render high conductivity to further im- dressing the long-term stability issues. prove their performances. Success of this project could We have carried out comprehensive structural charac- have huge impact in the development of next generation terization of the intermediate and NGs using MALDI-TOF, energy storage devices . 1H NMR, and 13C NMR spectra. All the spectroscopic This project is based on materials developed in our labo- results are in perfect agreement with the proposed mo- 503

Page 504

lecular structures, confirming the successful preparation of NGs. The thermal stability of NGs was examined by TGA . These NGs with various functional groups exhibit excellent thermal stability without significant mass loss up to 300oC under N2 atmosphere. The graphitized residue (char yield) of these NGs was more than 60% at 1000oC, which is at- tributed to their high aromatic (graphene-like) content. Our 3D NGs reveal great promises for the fabrication of energy conversion and storage devices. We use NG as as anode for the Li ion battery, which was prepared for the first time to exhibit fantastic specific capacitance (> 900 mAh/g) - one of the highest ever reported for carbon based Lithium ion battery, and stability through hundreds of charging discharging cycles. Such performances are ranked the very top among all graphene based materials. Future Work There are two major tasks to be accomplished in FY16: Electrochemistry and device tests The electrodes prepared from the three-dimensional porous nitrogen-doped nanographene will be studied in supercapacitor and lithium-ion battery: In particular, the nanographene composite anode materials will be tested in thin-film and standard CR-2325-type coin cells for elec- trochemical and battery characterization, respectively, to study their lithium-ion diffusion coefficient, interfacial characteristics, reversible capacity, electrochemical resis- tance, and charge/discharge characteristics. We will study the charge transfer limitations (electronic and ionic) in various porous nanographene structures using experimen- tal techniques such as chemical delithiation, galvanostatic charge and discharge, galvanostatic and potentiostatic intermittent titration techniques, cyclic voltammetry, and impedance measurements, to gain insights into the structure-property-synthesis correlations. Physical characterization Extensive physical characterization (high-resolution trans- mission electron microscope, scanning electron micro- scope, X-ray diffraction, X-ray photoelectron spectroscopy, Raman spectroscopy, Fourier transform infrared spectros- copy, Brunauer-Emmett-Teller surface area analysis, etc,) will be employed to determine the morphology, phase composition, heteroatom doping, surface area and poros- ity of the as-prepared porous nanographene materials. Conclusion We will synthesize the first porous nanographenes with a well defined molecular structure and surface area. We expect to demonstrate, for the first time, fabrication of supercapacitors based on a series of three dimensional nanographenes. 504

Page 505

Materials for the Future Postdoctoral Research and Development Continuing Project Photophysical Properties of Self-Assembled Nanoclusters Jennifer Martinez 20150704PRD1 Introduction as molecules in one instance and particles in another Artists and scientists alike have made and studied metal instance. This basic research project has applications to nanoparticles for centuries. These early particles were many important needs including biosensors, catalysis, the dichroic glass of the Lycurgus Cup (4th century), the batteries and solar harvesting and thus supports labora- lustrous ceramic glazes (> 9th century) of the Islamic tory missions in emerging threats (science of signatures), world, and even the rich colors of stained glass (e.g. Rose materials for the future and energy security. Window of Notre Dame Catherdral, ca 1250). Those The control of light energy and control of structure in particles (hundreds of atoms) have been well studied, as self-assembled and biologically-assembled material has been the electronic phenomena, which allows their architectures, as proposed here, is important for nu- technological and artistic use. When we shrink the num- merous applications of relevance to the Department of ber of atoms from the hundreds or thousands of atoms Energy. “From Quanta to the Continuum: Opportuni- to just a few atoms of gold or silver, we change the prop- ties for Mesoscale Science”, a BESAC report, identified erties of the collection (i.e. they are now fluorescent). a “grand challenge in materials science is to create These new, very small collections are called “nanoclus- multiscale functional structures that encode content; ters”. Little is understood about their optical properties adaptively respond to their environment; and capture- (e.g. origins of their photophysical properties), how to transport, and utilize energy”. Basic Research Needs for assemble those clusters into useful organizations (i.e. for Solar Energy Utilization was the focus of a BES workshop. solar harvesting or catalysis), nor how the clusters might Here, one of five identified cross-cutting issues relevant electronically “talk” to one another. In this project we to priority research directions was “… novel methods for aim to produce these clusters within DNA and to assem- self-assembly of molecular components into function- ble these clusters using the exacting nature of DNA-DNA ally integrated systems.” Each of these areas is being interactions. It is only through the use of biological tem- addressed within this PD proposal. plation of materials that we can ultimately control their structure and thus their properties. While we have start- Progress ed to make progress in the controlled synthesis of these Nanoclusters are collections of a few atoms of silver or materials, we have little utilized the biological control of gold that exhibit size dependent optical properties, such synthesis and assembly to understand and manipulate as strong fluorescence. Little is understood about their these new materials. If we are able to not only create, optical properties (e.g. origins of their photophysical but understand these new materials we can create the properties), how to assemble clusters into useful orga- next generation of optical devices (brighter and more nizations (i.e. for solar harvesting or catalysis), nor how touch sensitive phones), catalysts (energy for cars), and the clusters might electronically “talk” to one another. In energy collecting devices (solar harvesting). order to understand these questions we have developed strategies to self-assemble the clusters and study their Benefit to National Security Missions photophysics (using a genetically engineered polymer This project will answer long standing questions within and DNA) and to co-assemble the clusters with carbon the chemistry, materials and physics communities, by nanotubes and enzymes to create a biocatalyst for alter- bringing a more thorough structure-function under- native fuels. standing of fluorescent silver and gold nanoclusters. Fur- ther, we will bring fundamental understanding of how to To study how clusters talk to each other and how they modulate nanocluster physics so that they can behave 505

Page 506

may transition to have more advanced properties we ficient electrode that can be used as a catalyst for alterna- have created a cluster templated within a genetically tive fuels. With sources of fossil fuels dwindling, there is engineered polymer. Genetically engineered polymers an urgent need to find cheap, renewable, and alternate consist of repeating units of amino acid sequence yielding forms of energy using naturally abundant resources such a polymer with defined stimuli response, biocompatibility as sunlight, air, and water. Nanostructured materials and and programed assembly (as a function of changes in tem- enzymatic fuel cells are showing great promise in this perature). For the first time we report the direct synthesis respect, if only we can increase their efficiency. We have of pure, single sized, stimuli responsive, water-soluble created a new DNA-templated gold nanocluster (AuNC) of and biocompatible elastin like polymer (ELP) templated ~1 nm in diameter. When integrated with bilirubin oxidase Au nanoclusters. Optical absorption spectra and electron (BOD) and carbon nanotubes (CNTs), the AuNC acts as an microscopy indicates that the particles synthesized were enhancer of electron transfer (ET) and lowers the over- nanoclusters and their size was close to the boundary of potential of electrocatalytic oxygen reduction reaction plasmonic nanoparticles. From dynamic light scattering (ORR) by ~15 mV as compared to the enzyme alone (e.g. and electron microscopy we can observe that the clusters makes the electrode/alternative fuel more efficient). This assemble with increase in temperature. We are now in the unique role of the AuNC as enhancer of ET at the enzyme- process of understanding how these clusters talk or change electrode interface makes it a potential candidate for the their electronic properties within these assemblies. development of cathodes in enzymatic fuel cells, which often suffer from poor electronic communication between To more specifically assemble fluorescent clusters and the electrode surface and the enzyme. study their photophysics we have used DNA. DNA is much like Velcro in that it can be designed to specifically bind its Future Work compliment and can serve as a molecular ruler by precisely Goals and Tasks for FY16 include: spacing two materials away from each other to better understand at what distances the materials can commu- • Synthesize different cluster sizes, protected with either nicate. We have created a highly fluorescent silver nano- DNA or proteins. cluster templated within a piece of DNA. We observed • Design DNA clusters to produce self-assembled struc- that as the clusters were concentrated (effectively causing tures, utilizing the length of DNA as a molecular ruler the clusters to assemble and touch each other) that we to control the distance between clusters in the assem- substantially red shifted the resultant fluorescence and bly. absorbance of the material. This new “emergent” fluores- • Use these created assemblies to understand distance cence was reversible with simple dilution of the clusters to dependent cluster-cluster interactions (energy trans- reproduce the monomeric fluorescence. These initial re- fer) and to address the nature of clusters interaction sults suggested that the clusters could electronically com- (plasmonic or nonplasmonic). municate. To better control this interaction we designed • Probe DNA protected nanoclusters with EXAFS (ex- the cluster templating DNA to have a piece of overhanging tended X-ray absorpotion spectroscopy) to understand DNA that could zipper together with its compliment (like the nanocluster size and bonding. Velcro). That compliment contained the identical type of • Ustilize steady state and time resolved fluorescence fluorescent cluster with a complimentary overhanging spectroscopy and fluorescence microscopy to probe DNA. We can control the separation of the two clusters photophysical properties of clusters and their self as- with distances of 2, 4, and 6 nm. We find that the clus- sembly the combination of UV-Vis absorption spectros- ters communication (as seen by a substantial red shift in copy the fluorescence of 100 nm) decreases substantially with distance and drops off completely by 6 nm. Working with Conclusion our theory colleagues we have indication that the clusters Overall technical goals: We will synthesize different cluster are coupling through dipole interactions that are as strong sizes, protected with either DNA or proteins; those clus- as strongly coupled (and touching) conjugated polymers, ters will be self-assembled utilizing the length of DNA as a even though the clusters are far more separated than are molecular ruler to control the distance between clusters; polymers (suggesting that the clusters couple much more these created assemblies will then be used to understand strongly than current materials in the literature). distance dependent cluster-cluster interactions (energy Beyond organizing cluster-cluster interactions, we have transfer) and to address the nature of clusters interaction organized clusters to assemble with nanomaterials (carbon (plasmonic or nonplasmonic). nanotubes) and enzymes in the hope of creating an ef- Anticipated Impact: Residing directly between particles 506

Page 507

and molecules, nanoclusters are the “missing link” in mate- rials chemistry. Understanding how to create and control their properties not only solves an important mystery in science, but creates materials for new commercial technol- ogy. Publications Chakraborty, S., S. Babanova, R. C. Rocha, A. Desireddy, K. Artyushkova, P. Atanassov, and J. S. Martinez. A DNA- Hosted Gold Nanocluster Enhances Enzymatic Reduc- tion of Oxygen by Facilitating Efficient Electron Trans- fer. 2015. Journal of the American Chemical Society. 137 (36): 11678. 507

Page 508

Materials for the Future Postdoctoral Research and Development Continuing Project Dynamic Strength and Phase Transition Kinetics in Geophysical Materials Cynthia A. Bolme 20150707PRD2 Introduction electron laser and the details of the experimental sta- This project is performing cutting-edge science by using tions for MaRIE. Understanding the behavior of matter the most brilliant hard x-ray source available to examine during extreme shocks is directly relevant to the nuclear the response of geophysical materials to strong shock weapons program. waves. Only in the last few years has the ability to gen- Progress erate conditions of geophysical impacts been available at the same facility with a high brilliance x-ray source. This In FY15, we have gone through the journal review pro- x-ray source is necessary to observe how materials like cess for a journal article on the phase transition of amor- SiO2 and H2O change their structures from crystalline phous fused silica to the crystalline stishovite phase dur- solids to disordered materials, like liquids or glasses. Un- ing shock compression. The article is in the final stages derstanding the changes in structure of these materials of review, and we anticipate that it will be published in will provide new data about how these materials behave Nature Communication in the next few months. We also during geophysical impact events and about the proper- examined the structure of fused silica that was recov- ties of these materials when they exist at high pressures ered after shock compression to learn about the phase and high temperatures, such as inside planet Earth. reversion process to a high density amorphous phase. An experiment will be performed at the Linac Coherent We received x-ray beam time to perform an experiment Light Source (LCLS) to specifically examine the strength at the LCLS to study the strength of shock compressed of iron at high temperatures and pressures. These pres- iron. We designed the iron experiment and prepared sure and temperature conditions will be similar to those some of the samples that will be consumed during the found inside the earth. shock experiment in December of FY16. Benefit to National Security Missions Future Work This work contributes strongly to the basic understand- In FY16, we will perform an experiment at the Linac ing of materials by producing a new class of data that Coherent Light Source (LCLS) to study the strength of can be used to develop physically based models of iron under shock compression. The experiment will dynamic materials response. This new class of data include polycrystalline and single crystal iron samples. provides detailed information about the interatomic The samples will be shock compressed to simulate the distances and the sizes and orientations of crystals in pressure and temperature conditions of iron inside the materials. Due to the specific materials being investi- earth - from the crust to the molten iron core. gated (i.e. materials that are very common in geophysi- We will also complete the analysis of a previously cal bodies, such as asteroids, comets, and rocky planets), acquired data set showing the fast dynamics of water this work will also directly contribute to astrophysical freezing to the ice VII phase, and we will submit a jour- areas of research. nal article on these results. The Laboratory’s proposed Matter-Radiation Interactions Conclusion in Extremes (MaRIE) Signature Facility includes a hard x- ray free electron laser, similar to the facility in which this This project will provide unique new data about how the work performs experiments, and the work in this project atoms in geophysical materials (for example, water, SiO2, will help to define the characteristics of the MaRIE free and iron) respond under conditions of high temperature 508

Page 509

and pressure. This information will progress the current level of understanding about how these materials behave in geophysical events, such as asteroid impacts and the dynamics of the earth’s molten iron core. Publications Gleason, A. E., C. A. Bolme, H. J. Lee, B. Nagler, E. Galtier, D. Milathianaki, J. Hawreliak, R. G. Kraus, J. H. Eggert, D. E. Fratanduono, G. W. Collins, R. Sandberg, W. Yang, and W. L. Mao. Ultrafast visualization of crystallization and grain growth in shock-compressed SiO2. 2015. Na- ture Communications. 6: 8191. 509

Page 510

Materials for the Future Postdoctoral Research and Development Continuing Project In-situ, 3D Characterization of Solidification in Metals Amy J. Clarke 20150709PRD2 Introduction solidification experiments in which the thermal gradient, The pRad facility at LANL provides the optimal size cooling rate, and alloy composition are independently scale for doing solidification experiments with spatial and systematically varied. These datasets will give us a resolution of ~20 µm and a field-of-view of several better understanding of solidification and will be made centimeters. This is superior to the more common x-ray available to validate simulation methods, including a experiments, because it allows for a sample size that is microstructure prediction model that is currently being more representative of an actual casting, but the spa- implemented in the advanced simulation and computing tial resolution is still sufficient to capture the details of (ASC) code Truchas. microstructural development. However, proton experi- MaRIE: In this project, we will be using multiple beams ments have been limited by the relatively low image (e.g. proton and x-ray) to interrogate samples during acquisition rate of 20 Hz, which results in an acquisition solidification. This data can be used to access multiple time of almost 30 seconds for a full 3D dataset. size scales, since proton imaging has a resolution of ~20 Synchrotron x-rays, on the other hand, provide the µm and x-ray imaging typically has a resolution of ~1 temporal resolution needed to collect a 3D dataset in µm. As both methods have a field of view about 1000 just a few seconds; however, x-rays are limited to about times their resolution, protons can be used to see large a millimeter of penetrating power. Furthermore, none volumes of material, while x-rays can give higher resolu- of the in-situ x-ray experiments of metal solidification tion views of smaller volumes. The multiple characteriza- that have been done to date have used a well controlled tion methods also provide opportunities to examine the furnace that can provide a given thermal gradient and chemistry of samples in higher detail, due to the differ- cooling rate. ing contrast mechanisms between the two beam types. This work is in alignment with MaRIE, as multiple beams We propose to overcome the pRad temporal resolution are envisioned to simultaneously characterize samples in limits by using the recently developed time-interlaced the future. This research will also help to define what is model-based iterative reconstruction (TIMBIR) method. currently possible with existing characterization meth- This will provide temporal resolution that is on par with ods and outline what capabilities should be included in synchrotron x-ray experiments, but with the superior future instruments. field-of-view provided by protons. Improvements on the scientific front will come from building and using a Progress furnace that can provide controlled temperature profiles Since the fellowship started in April 2015, the main ac- in space and time while also rotating, which is neces- complishment has been the design of a furnace that will sary for collecting a computed tomography dataset. This permit computed tomography (CT) during metal alloy furnace will allow for a much better temperature control solidification. The rotation and translation stages and the than has been achieved before, making it possible to do necessary controllers to perform CT experiments have systematic studies of cooling rate, temperature gradient, been identified and quoted, and will be ordered and and alloy composition on solidification behavior. received by the end of the FY. The delivery timeframe for these parts is 8 to 12 weeks. Benefit to National Security Missions DOE/SC: This project is tied to science campaigns by The furnace portion of the experimental setup has un- generating datasets of microstructure evolution during dergone a significant design change from our radiogra- 510

Page 511

phy experiments, making it simpler and more capable. This the new x7 magnifier at pRad and the recently developed design change permits the use of simpler resistive heating time-interlaced, model-based iterative reconstruction elements, rather than induction heating coils. The resistive (TIMBIR) methodology, we will be able to improve the heating elements have been specced out, and we are in spatial and temporal resolution from 50 µm to 20 µm and the process of getting quotes. from 20 seconds to 1 second per 3D dataset. On the scientific front, two papers have been published on Publications the time-interlaced model-based iterative reconstruction K.A. Mohan, S.V. Venkatakrishnan, J.W. Gibbs, E.B. Gulsoy, (TIMBIR) methodology, which is one of the key enabling X. Xiao, M. De Graef, P.W. Voorhees, C.A. Bouman. technologies of this project for achieving improved tem- “TIMBIR: A Method for Time-Space Reconstruction poral resolution for the proposed dynamic solidification from Interlaced Views.” IEEE Transactions on Com- experiments. The first paper focuses on the mathematical putational Imaging. DOI: 10.1109/TCI.2015.2431913 formulation for TIMBIR, and the second paper showcases (2015). an application of TIMBIR to studying the morphology of J.W. Gibbs, K.A. Mohan, E.B. Gulsoy, A.J. Shahani, X. Xiao, free-growing metallic dendrites. C.A. Bouman, M. De Graef, P.W. Voorhees. The Three- Dimensional Morphology of Growing Dendrites. Scien- Future Work tific Reports. Accepted (2015). During the remainder of FY15, the motion stages and fur- nace components will be specified, purchased, and deliv- ered. Since there is a lead-time of several months between ordering and delivery of the components, the end of FY15 will also provide time to test the heavily revised TIMBIR code [1]. These tests will be used to determine if our 40- core computer cluster is sufficient for data processing, or if it will be necessary to request computing time on high performance computing machines. With delivery of the experimental furnace and stage components by the end of FY15, assembly and testing of the system will start in early FY16. The testing will include programming the rotation stage for the particular motion that we want, and coordinating the rotation and heating control systems with the beamline/camera data acquisition system. The first “in-beam” tests of the system should occur in the early part of FY16 with experiments at pRad, provided beam-time is received from a competitive, institutional proposal process. To our knowledge, this will be the first time that well-controlled, directional solidification experi- ments and 3D, in-situ characterization are combined. Initial experiments will look at void formation in castings, inter- facial morphology and dynamics during solidification, and the evolution of solute distribution during solidification. The processing and analysis of these datasets will likely require most of FY16. Conclusion We will develop the capability to do controlled solidifica- tion experiments in any type of beamline, including the proton, neutron, and x-ray beamlines across the U.S. DOE complex. We will use this capability to probe solidifying metals with a variety of different diagnostic means to ac- cess a wide variety of spatial and chemical data. By using 511

Page 512

Materials for the Future Postdoctoral Research and Development Continuing Project Dendritic Microstructure Selection in Cast Metallic Alloys Amy J. Clarke 20150713PRD2 Introduction that predict realistic solidification scenarios will provide Complex – so-called dendritic – microstructures are most LANL with unique capabilities for the prediction and commonly observed in metallic alloys. Their morpho- control of solidification microstructures. logical and compositional characteristics determine the Benefit to National Security Missions mechanical properties and performance of cast parts. DOE/SC: The development of a simulation capability From a theoretical standpoint, dendritic growth is a long- for microstructure solidification at realistic time/length standing example of complex pattern formation, involv- scales and gravity conditions, which is currently miss- ing structural and chemical changes over multiple length ing at LANL, will be a critical step toward predicting and and time scales. Most theoretical studies have focused controlling solidification structure evolution, which is key on diffusive dendritic growth, neglecting convective for LANL to support advanced manufacturing initiatives solute transport in the liquid. Yet, gravity yields signifi- and the design of innovative processing studies. cant convection, resulting in detrimental microstructure inhomogeneities. MaRIE: This project is strongly motivated by the capabil- ity to quantitatively simulate solidification microstruc- The first implementation of a novel, multi-scale dendritic tures in realistic gravity conditions to support the design needle network (DNN) model is orders-of-magnitude of solidification experiments with in situ imaging at LANL faster than simulation methods like phase-field and with high energy X-ray and proton beams. The multiscale hence permits simulations at the mesoscale. In this modeling approach will not only guide multi-scale char- project, liquid convection will be added into this model, acterization experiments with MaRIE, but should also enabling dendritic solidification simulations under re- bring fundamental insight into the relevant physics that alistic gravity conditions at length/time scales relevant need to be included in our models. to experiments and casting processes for the first time. Computationally intensive implementations of phase- Basic Understanding of Materials: Because dendritic field models will also be used as a validation tool, in solidification involves a broad range of length and time support of the DNN model development. scales, it has only been approached theoretically at the scale of a few dendrites in idealized conditions, such as a Convection effects at the microscopic scale will be stud- purely diffusive regime. The current project will provide ied with phase-field and at the scale of experiments with the first tool to quantitatively explore the crucial effect multi-scale DNN simulations. The model developments of gravity-induced convection on spatially extended will be validated by quantitative comparisons to experi- dendritic arrays. ments performed by Clarke’s team at LANL. The simula- tions will guide the design of new experiments that will Climate and Energy Impact: The reduction of energy be executed at U.S. DOE User Facilities. By combining consumption, for instance in transportation, requires controlled experiments with intensive simulations, we the development of innovative materials with tailored will reveal the fundamental mechanisms of microstruc- microstructures and properties. The ability to explore ture selection in the presence of unavoidable convec- microstructure design through simulations will also ac- tion, which is essential to the properties and perfor- celerate the development and deployment of energy mance of cast materials. efficient processes. The development of cutting-edge simulation techniques 512

Page 513

Progress The final goal at the end of the project in FY17 is an imple- Since the beginning of the project in May 2015, we are mentation in three dimensions (3D). Hence, each step will investigating which CFD numerical framework would be be toward this aim. Moreover, in the choice of the appro- most suitable to a direct coupling with the multiscale priate CFD approach, simulation tools already available at Dendritic Needle Network (DNN) crystal growth model. LANL (e.g. Truchas) will be strongly favored to ideally build The potential candidates for CFD approaches include Finite a long-lasting, in-house capability for simulating dendritic Volume, Finite Element, Finite Difference, and potential solidification under realistic gravity and processing condi- alternative approaches like Lattice-Boltzmann methods. tions. The selection is based upon the potential of the approach While the anticipated lower budget allocation may affect to deal with fluid-structure and fluid-particle interactions the extent of the model deployment and coupling with in a quantitative way, as well as the potential to achieve other codes at LANL, the final goal remains a 3D coupled highly scalable simulations. At this point, the two most DNN-CFD implementation. promising approaches are the Finite Volume and Lattice- Boltzmann methods. The former benefits from broad Conclusion expertise and various optimized tools already in place at Cutting-edge, multi-scale simulations validated by in situ LANL. The latter has recently shown great promise in the experiments will shed light on the poorly understood, yet field of multiphase flows, and would provide an alternative crucial effects of gravity on microstructure selection in me- approach that would be relatively simple to implement, be tallic alloys during metallurgical processing. Understanding effective in handling microscopic fluid-particle interactions, crystal growth under gravity-induced liquid flow across and would offer a highly parallelizable framework. While length scales will enable the control of microstructure no particular approach has been ruled out of the selection and defects in cast parts. This project will mark a trans- process yet, we are currently considering the feasibility of formational leap toward reaching predictive capability for pursuing both Finite Volume and Lattice-Boltzmann meth- advanced manufacturing at LANL, aimed at tailoring ma- ods and will benchmark them for DNN calculations. terials microstructures and properties relevant to national security and energy challenges. It will also provide new, Future Work predictive computational tools needed for future MaRIE The main goal for FY16 is to design a two-dimensional studies. model coupling the multi-scale Dendritic Needle Network (DNN) model for crystal growth to fluid dynamics in the liq- uid. Since Computational Fluid Dynamics (CFD) equations are harder to solve than the intentionally simplified DNN equations, it is more appropriate to incorporate the DNN model into a CFD code than vice-versa. Toward that aim, in FY15 we will identify the CFD numeri- cal framework most suitable to direct coupling with DNN crystal growth modeling. The potential candidates for CFD approaches include Finite Volume, Finite Element, Finite Difference, and potential alternative approaches like Lattice-Boltzmann methods. The choice will be based upon the potential of the approach to deal with fluid-structure and fluid-particle interactions in a quantitative way. In FY16, we will develop and validate a new mathematical formulation of the DNN equations in the framework of the chosen CFD approach. This core step may require sub- stantial DNN redesign, and a subsequent set of thorough validation tests to ensure that the crystal growth modeling provides reliable, quantitative predictions. Finally, we will develop a fully coupled DNN-CFD imple- mentation, which will be made computationally efficient by the use of advanced numerical techniques, such as in- tensive parallelization (e.g. using Graphic Processing Units). 513

Page 514

Materials for the Future Postdoctoral Research and Development Final Report Frustrated Materials Cristian D. Batista 20120751PRD2 Abstract leading to an effective non-local classical action for Competing or conflicting interactions are ubiquitous in the local moments that will be solved by applying the physical and biological systems that range from water ice Metropolis algorithm. All the thermodynamic properties and glass formation to neural networks and protein fold- could be computed under control because the statisti- ing. Frustration in these many-body systems refers to the cal error can be made arbitrarily small by increasing the fact that local pairwise interactions cannot be satisfied sample size. The temperature dependence of the ther- simultaneously. Condensed matter systems with strong modynamic quantities could thus be directly compared frustration exhibit a variety of fascinating features such against the experiments. as emergent gauge field, topological order, and fraction- Scientific Approach and Accomplishments alized excitations. They are also hosts to various complex orders and exotic phases such as spin ice and topological The research of Dr. Gia-Wei Chern has been focused on insulators. The main goal of this project is to investigate frustrated systems in general, and particularly in highly various intriguing phenomena resulting from strongly frustrated magnetism. The main goal here is to have a frustrated interactions. better understanding of these intriguing complex sys- tems and to design and control their novel properties for Background and Research Objectives applications to, e.g. magnetic storage, computing, and One of the most spectacular manifestations of geo- sensing. metrical frustration is the presence of the novel states Working with his collaborators at LANL, Dr. Chern has of matter and associated phase transitions. These novel studied a novel magnetic field induced phase transi- phases arise as a compromise between the opposite tion in vanadium spinels. The magnetic ions in these tendencies that are inherent to frustrated geometries. compounds form a highly frustrated pyrochlore lattice, A triangular plaquette is an archetypal building block which is a three-dimensional network of corner-sharing for lattices with geometrical frustration. Many unusual tetrahedron. The vanadium spinels have been a canoni- magnetic and orbital orderings in frustrated magnets are cal system for studying frustrated spin-orbital interac- best understood from the viewpoint that the system is in tions. Taking advantage of the high magnetic field facility the vicinity at LANL, experimentalists have uncovered an interesting of a metal-insulator transition. It is a challenging theo- phase transition induced by magnetic field in these com- retical task to model these systems, which lie in the pounds. Gia-Wei and his colleagues have developed a intermediate-coupling regime between deep Mott insu- theoretical model that successively explains the experi- lator and the metallic phase. To carry out such studies, mental results. we employed a combination of analytical and numeri- Another main area of his research is the study of ar- cal techniques. We used mean field and semi-classical tificially created frustrated systems, in particular the techniques for computing thermodynamic properties. so-called artificial spin ice. Artificial spin ice is a boom- For the numerical approach we implemented ing field in condensed matter physics. Among the most a Mote Carlo algorithm for solving the problem of itiner- exciting recent developments is the emergence of mag- ant electrons interacting with localized classical spins. netic monopoles in these materials. Emergent phenom- Since the action is quadratic in the fermion operators, ena, which until now have only been accessible at low-T, the itinerant electrons can be integrated out exactly, can thus be tailor-designed to manifest at room tem- 514

Page 515

perature. In particular, the existence of mobile magnetic and are also important to the strategic mission of LANL. monopoles coupled to magnetic fields provides a promis- ing approach to realize magnetricity, the control and ma- Impact on National Missions nipulation of magnetic currents. With the rapid progress in Given the complexity of correlated materials and the un- nano-fabrications, it is now possible to realize artificial spin limited number of compounds, the discovery of new states ice in various lattice geometries. One of the key advantages of matter and functionalities must be organized around of these systems is that they allow direct experimental ac- simple guiding principles. Our basic principle for discover- cess to the microscopic degrees of freedom. ing complex and collective forms of matter that exhibit novel properties and respond in new ways to environmen- With collaborators in LANL, Dr. Chern has proposed the tal conditions is to make, characterize and model classes first feasible design of a three- dimensional artificial spin of materials with controllable degree of electronic localiza- ice that has exactly the same lattice topology as the natural tion. This is a necessary step towards the dream of “mate- spin-ice compounds. Experimental realization of such rials by design.” systems will allow us to study the dynamics of magnetic monopoles in these novel materials taking advantage of With such tools, we can better understand the complexity the powerful tools of modern nanotechnology. In addition, of actinides and other correlated materials. Actinide based the earlier proposal of composite monopole quasiparticles, materials are of primary interest for energy security appli- namely an emergent monopole surrounded by background cations relevant for DOE/NNSA and DOE/SC. Moreover, our magnetic charges, has been experimentally confirmed; Dr. ability of predicting the behavior of correlated materials is Chern was also involved in the theoretical analysis of this crucial for optimizing response functions, such as magne- work. to-electric susceptibility, which are necessary for sensing applications relevant for DoD basic defense applications Motivated by recent advances in thin-film growth tech- (including detection), DHS basic homeland security (includ- nology, Dr. Chern also studied the triangular Ising antifer- ing detection) and Intelligence agencies. We are respond- romagnetic thin films. He and his colleagues discovered ing to the national need of predicting actinide, and more in unusual re-entrant phase transitions, which depend on the general, correlated materials properties for energy security number of layers in the system. This study added a surpris- applications. Our work is a first step towards reaching the ing chapter to this well studied canonical system of frus- level where studies of actinide properties can be predicted trated magnetism. from first principles computer-based simulations, thus avoiding costly experiments and contributing significantly Collaborating with colleagues in the Theoretical Division to the reliability of the nuclear stockpile. The outcome of (T-1), Dr. Chern has pioneered in applying the state-of-the- this project (new code for performing molecular dynamics art large-scale Langevin dynamics simulations to investigat- of correlated materials) can also be used for studying the ing novel magnetic orders in metallic frustrated magnets. effects of matter and radiation in extremes, the mission of The Langevin dynamics method was developed at LANL for MARIE. efficient simulations of electrons interacting with classical fields. One particular interest is the complex spin textures Publications stabilized by electrons in the frustrated kagome lattice. In Chern, , Chien, and Di Ventra. Dynamically generated flat- this study, Gia-Wei and his collaborators have uncovered band phases in optical kagome lattices. 2014. PHYSI- several novel non-coplanar magnetic orders, some of CAL REVIEW A. 90 (1). which also support topological electron states. Chern, , Maiti, R. M. Fernandes, and Woelfle. Electronic The Langevin method can also be applied to improve the Transport in the Coulomb Phase of the Pyrochlore Spin efficiency of quantum molecular dynamics (QMD) simula- Ice. 2013. PHYSICAL REVIEW LETTERS. 110 (14). tions, especially for metallic systems. With his collabora- Chern, , Reichhardt, and C. J. O. Reichhardt. Avalanches tors, Dr. Chern has developed numerical schemes that inte- and disorder-induced criticality in artificial spin ices. grate techniques for studying strongly correlated electrons 2014. NEW JOURNAL OF PHYSICS. 16. with the efficient QMD simulations. More specifically, Dr. Chern worked on developing a quantum molecular dynam- Chern, , Reichhardt, and Nisoli. Realizing three-dimensional ics simulations based on an efficient Gutzwiller method. He artificial spin ice by stacking planar nano-arrays. 2014. successfully developed a prototype Gutzwiller QMD code APPLIED PHYSICS LETTERS. 104 (1). and discovered interesting new results based on this code. Chern, , and Wu. Four-Coloring Model and Frustrated Su- These works will contribute to our fundamental under- perfluidity in the Diamond Lattice. 2014. PHYSICAL RE- standing of the dynamics of strongly correlated materials 515

Page 516

VIEW LETTERS. 112 (2). anisotropic triangular lattice. 2013. PHYSICAL REVIEW B. 87 (17): -. Chern, G. W., C. Reichhardt, and C. J. O. Reichhardt. Frus- trated colloidal ordering and fully packed loops in ar- Rahmani, , and Chern. Universal Renyi mutual information rays of optical traps. 2013. PHYSICAL REVIEW E. 87 (6): in classical systems: The case of kagome ice. 2013. -. PHYSICAL REVIEW B. 88 (5). Chern, G. W., M. Morrison, and C. Nisoli. Degeneracy and Rahmani, A., and G. W. Chern. Universal Renyi mutual in- criticality from emergent frustration in artificial spin formation in classical systems: The case of kagome ice. ice. 2013. Physical Review Letters. 111: 177201. 2013. PHYSICAL REVIEW B. 88 (5): -. Chern, G. W., R. M. Fernandes, R. Nandkishore, and A. V. Zhang, S., I. Gilbert, C. Nisoli, G. W. Chern, M. J. Erickson, Chubukov. Broken translational symmetry in an emer- L. O’Brien, C. Leighton, P. E. Lammert, V. H. Crespi, and gent paramagnetic phase of graphene. 2012. PHYSICAL P. Schiffer. Crystallites of magnetic charges in artificial REVIEW B. 86 (11): -. spin ice. 2013. NATURE. 500 (7464): 553. Chern, G. W., and C. D. Batista. Spontaneous quantum Hall effect via a thermally induced quadratic Fermi point . 2012. Physical Review Letters. 109: 156801. Chern, G. W., and O. Tchernyshyov. Magnetic charge and ordering in kagome spin ice. 2012. PHILOSOPHICAL TRANSACTIONS OF THE ROYAL SOCIETY A-MATHE- MATICAL PHYSICAL AND ENGINEERING SCIENCES. 370 (1981): 5718. Chern, G. W., and R. Moessner. Dipolar Order by Disorder in the Classical Heisenberg Antiferromagnet on the Kagome Lattice. 2013. PHYSICAL REVIEW LETTERS. 110 (7): -. Choi, E. S., G. W. Chern, and N. B. Perkins. Chiral magne- tism and helimagnons in a pyrochlore antiferromagnet. 2013. PHYSICAL REVIEW B. 87 (5): -. Choi, E., G. W. Chern, and N. B. Perkins. Helimagnons in a chiral ground state of the pyrochlore antiferromagnets. 2013. EPL. 101 (3): -. Gilbert, I. a. n., Chern, Zhang, O’Brien, Fore, Nisoli, and Schiffer. Emergent ice rule and magnetic charge screening from vertex frustration in artificial spin ice. 2014. NATURE PHYSICS. 10 (9): 671. Knolle, J., G. -W. Chern, D. D. L. Kovrizhin, R. Moessner, and N. B. Perkins. Raman Scattering Signatures of Kitaev Spin Liquids in A2IrO3 Iridates. 2014. Physical Review Letters. 113: 187201. Lin, , Kamiya, Chern, and C. D. Batista. Stiffness from Disor- der in Triangular-Lattice Ising Thin Films. 2014. PHYSI- CAL REVIEW LETTERS. 112 (15). Mun, E. D., Chern, Pardo, Rivadulla, Sinclair, H. D. Zhou, V. S. Zapf, and C. D. Batista. Magnetic Field Induced Tran- sition in Vanadium Spinels. 2014. PHYSICAL REVIEW LETTERS. 112 (1). Perkins, N. B., G. W. Chern, and W. Brenig. Raman scat- tering in a Heisenberg S=1/2 antiferromagnet on the 516

Page 517

Materials for the Future Postdoctoral Research and Development Final Report Designing and Probing Novel Materials by Pressure Tuning of Nanocrystals Hongwu Xu 20120753PRD2 Abstract and are structurally defect-free, possess unique prop- Self-assembly of nanocrystals (NCs) into superlattices erties compared to their bulk counterparts. Pressure is an intriguing phenomenon that may lead to materi- tuning of nanocrystals is a unique method for design- als with novel collective properties, in addition to the ing and making novel nanostructured materials. Typi- unique properties of individual NCs compared with their cally, NCs undergo structural transformations via single bulk counterparts. Pressure tuning of nanocrystals (NCs) nucleation events at high pressures [1]. As the surface and their assembled superlattices is a novel approach to energy of NCs overwhelms their bulk energy, these design, develop and probe new nanostructured materi- phase transitions may be arrested at ambient conditions als. In this research, we have determined the structural through surface stresses that are equivalent to external evolution and stability relations of several NC superlat- compressive pressures. Moreover, recent studies indi- tices (e.g. PbSe, PbTe and Pt) at high pressure (P) condi- cate that high pressure and deviatoric stress can be a tions. We started with controlled synthesis of nanopar- more efficient way to tune NC superlattices and 1D/2D ticles with desired characteristics and then assembled mesostructures [2]. These assembled NC architectures them into ordered superlattices or supercrystals. We not only exhibit a single-crystal-like structure but also systematically investigated their high-pressure behav- possess strong size-quantization effects and enhanced ior using small-angle and wide-angle synchrotron X-ray properties. Hence, studying nanoparticles and their scattering (SAXS and WAXS) coupled with diamond-anvil assemblies under high pressures enables discovery and cell (DAC) techniques, and we determined the relative continuous synthesis of novel materials with potential stabilities of different NC superlattice polymorphs us- applications in catalysis, solar cells and other areas [3]. ing calorimetric techniques. Our results have provided Scientific Approach and Accomplishments important insights into the fundamental mechanisms underlying the formation of high-pressure nanocrystal The general approaches of our research included: 1) phases and the coalesced superlattices potentially with synthesis of colloidal nanoparticles with desired char- enhanced or newly emerging properties. acteristics (structure, size, shape etc.), 2) self-assembly of these nanoparticles into superlattices/supercrystals, Background and Research Objectives 3) structural characterization of the nanoparticles and Novel materials are often found by modifying the struc- their assemblies at high pressures, and 4) determination tures of existing materials without altering their compo- of the stability relations among different NC superlattice sitions. As an important but underexplored thermody- polymorphs. The studied systems include PbTe, PbSe, Pt namic variable, pressure can induce extreme changes in and NaYF4. We describe the specific results below. material structures, potentially leading to new physical Pressure-induced switching between amorphization properties. For example, soft, opaque graphite is the and crystallization in PbTe nanoparticles form of carbon stable at ambient conditions, but it can Combining in-situ high-P synchrotron X-ray scattering be turned into ultra-hard, transparent diamond at high with transmission electron microscopy (TEM), we inves- pressures. To expand the range of available solid-state tigated P-induced structural switches between the rock materials, it is important to find pathways leading to salt and amorphous phases as well as the associated metastable, high-pressure phases that can be brought to mechanisms of their formation in 6 nm PbTe nanocrys- ambient conditions for real-world applications. tals [4]. To harvest the metastable phases of PbTe, we Nanocrystals, which bridge single atoms to bulk solids applied different maximum pressures on PbTe NCs and 517

Page 518

selectively obtained the stable crystalline or metastable amorphous PbTe phases at ambient conditions. In addi- tion, the conversion from the amorphous to crystalline PbTe phase was observed under electron beam irradiation with a TEM. Synthesis of PbTe NCs and subsequent loading into a DAC. PbTe colloidal nanoparticles with an average size of 6 nm were synthesized as follows. 0.45 g of lead oxide (2 mmol), 4.2 mL of oleic acid and 12.0 mL of octadecene were loaded in a three-neck round-bottom flask under airless condition, heated to 170 °C for 30 min, and subsequently stabilized at 150 °C for 5 min under agitation and argon Figure 1. Several representative 2D WAXS images showing the protection. Once the solution temperature was about 110 two types of processes and their correlations. °C, 2 mL of Te-trioctyl phosphine (0.5 mol/L) was rapidly in- jected to initiate the formation of PbTe NCs, and was then Upon gradual increase of irradiation dose, there were ceased after 7 min with a cold water bath. Lastly, the PbTe more and more electron diffraction rings, and further, their NCs were dispersed in hexane, forming black transparent intensities increased with irradiation time. These changes colloidal solutions. indicate that amorphous PbTe nanoparticles became crystallized upon electron irradiation, and the resulted NCs The sample was loaded into a DAC by drop-casting a were confirmed to have the rock salt structure. Thus the suspension of PbTe NCs into the gasket hole seated on the recovered high-P amorphous PbTe nanoparticles can re- tip of one diamond anvil. This procedure was repeated verse back to the stable rock salt phase due to the thermal 2-3 times to ensure that there were sufficient PbTe NCs in energy from electron beams. the DAC sample chamber. To determine pressure depen- dence of the phase transformation behavior of PbTe NCs, two separate runs were conducted to reach the maxi- mum pressures of 15.6 and 19.7 GPa. Upon release of the pressure to ambient conditions, the sample was carefully transferred from the gasket hole to a Cu grid for later TEM observation. High-P behavior of PbTe nanoparticles. Starting with the same rock salt PbTe NCs, depending on the maximum pressure, we harvested either rock salt or amorphous phase upon decompression (Fig. 1). When PbTe NCs were pressurized above 10 GPa, amorphization, rather than the rock salt to orthohombic transition observed in bulk PbTe, started to occur. When the maximum pressure reached 15.6 GPa, the amorphous phase transferred back to the starting rock salt phase after releasing the pressure. How- ever, when the maximum pressure reached 19.7 GPa, the metastable amorphous phase could be kept at ambient conditions. The difference between these two recovered samples - crystalline phase from the 15.6 GPa run and amorphous phase from the 19.7 GPa run - can be readily Figure 2. Schematic diagram showing the Gibbs free energy re- discerned from their diffraction patterns. Hence, the meta- lation of the phase transitions observed in PbTe NPs and HRTEM stable amorphous PbTe nanoparticles can be harvested by images of the samples recovered from 15.6 GPa (left inset) and applying higher pressures on PbSe NCs. from 19.7 GPa at the initial stage of electron beam irradiation (right inset). The scale bars in both images represent 5 nm. Transition of PbTe nanoparticles from metastable amor- phous to stable rock salt structure. We used TEM to In summary, high pressure and thermal energy can be investigate the effects of electron beam irradiation on the used to achieve multiple switching between stable rock- stability of amorphous PbTe nanoparticles (Fig. 1). salt structure and metastable amorphous structure in 518

Page 519

PbTe nanoparticles. Figure 2 shows the Gibbs free energy of nanocubes in different layers, it is clear that nanocubes relations between the two structures, illustrating their in the top layer directly sit on the top surface of nanocubes switching mechanisms. The amorphous nanophase formed in the bottom layer (Figs. 3d-g). Therefore, the formed Pt at 15.6 GPa has a lower density (LDA – low-density amor- NC assembly has a simple cubic superstructure (Fig. 3h). phous), is energetically unfavorable, and would revert back to the rock salt structure (memory effect). In contrast, the amorphous nanophase formed at 19.6 GPa possesses a higher density (HDA – high-density amorphous) and has a larger energy barrier, preventing it from reversion back to the rock salt structure. Nevertheless, the energy barrier between the recovered amorphous and rock salt struc- tures of PbTe nanoparticles can be overcome by additional thermal energy under electron irradiation. Controlled self-assembly of non-spherical Pt nanoparticles Self-assembly of small building blocks such as atoms, mol- ecules and nanocrystals into mesoscopic and macroscopic structures is an intriguing phenomenon and is an active research area in chemistry, biology and materials science [5]. Most of the earlier self-assembly preparations used spherical nanoparticles as their building blocks. Recently, progresses in colloid chemistry have made it feasible to prepare well-defined non-spherical nanocrystals, allowing design and fabrication of non-spherical NC-based super- Figure 3. Self-assembly patterns of Pt nanocubes generated from a toluene solvent. (a) TEM image. (b) Magnified image from crystals [6]. Through the delicate control over solvent- (a) and (c) SAED pattern from the same area. (d) 3D reconstruc- ligand interactions, we have shown that Pt nanocubes tion (left) with two orthogonal views AB and CD (right). (e, f) Slice can be selectively assembled into two distinct types of views of the top and bottom layers at the heights of z1 and z2, superlattices in the form of 2D thin films: simple cubic and respectively, as indicated in (d). (g) Calculated image obtained by body-centered tetragonal [7], as described below. averaging (e) and (f). (h) Structure model of the sc superstructure assembly. Self-assembly of Pt nanocubes using a “home-built” setup. Uniform Pt nanocubes with an edge length of 10 nm and When hexane was used as the solvent (while other condi- coated with oleic acid and oleylamine ligands were used as tions were kept the same), a different assembly pattern the building blocks. These nanocubes were separately dis- resulted (Figs. 4a and 4b). Based on the TEM tomography persed in two solvents: hexane and toluene, forming col- results, Pt nanocubes in the top and bottom layers of a loidal solutions, and then these colloidal nanocubes were three-layer assembly occupy the same horizontal positions, assembled in a controlled manner as follows. A two-layer while those in the middle layer situate in the interstices, substrate was used: the bottom layer is a square Si wafer, each of which is produced by four nanocubes from the top and the top layer is a TEM Cu grid placed on the Si wafer. and bottom layers (Figs. 4d-i). As a result, a body-centered- Several drops of Pt nanocube solution were put onto the tetragonal superstructure is formed (Fig. 4h). substrate, and, due to the surface tension of solvent on the Si wafer, the colloidal solution was constrained within In summary, due to the presence of organic ligands out- the wafer. To slow down solvent evaporation, the substrate side Pt nanocubes, one can use different solvent types to was immediately covered with a weighing dish and then modify the weak ligand-solvent and solvent-solvent inter- sealed. After complete solvent evaporation, Pt nanocube actions during the self-assembly process, thereby changing assemblies formed on the substrates, which were later the relative positions of Pt nanocubes in different layers to studied using TEM. produce different packing structures. Thus our developed method is an effective approach to synthesize different Controlling self-assembly of Pt nanocubes with differ- types of assemblies from the same non-spherical nanopar- ent dispersion solvents. When toluene was used as the ticles. solvent, Pt nanocubes in different layers were stacked together with the face-to-face contact (Figs. 3a and 3b). From the TEM tomography results of the relative positions 519

Page 520

NCs, ligand interactions were expected to decrease in this order. Measurements of the disassembly enthalpies of superlat- tice polymorphs. The enthalpies of dissolution of the three supercrystals into well dispersed and separate NCs were measured using solution calorimetry with hexane as the solvent. Our results show that the dissolution enthalpies of all these polymorphs are on the order of 2 kJ per mol of PbSe, which is comparable to the energy of van der Waals interactions. Further calorimetric data analyses reveal that the bcc superlattice is the energetically most stable polymorph, with the bct being 0.32 and the fcc 0.55 kJ/ mol higher in enthalpy. This stability sequence is consistent with the decreased packing efficiency of PbSe NCs from bcc (17.2%) to bct (16.0%) and to fcc (15.2%) (Fig. 5b). On the other hand, the enthalpy differences among the three polymorphs are small, confirming a closely spaced energy landscape and explain the ease of formation of different NC superlattices at slightly different synthesis conditions. Figure 4. TEM and tomographic results of Pt nanocube self-as- sembly generated from hexane as the solvent. (a) TEM image. (b) Magnified image and (c) SAED pattern of the framed area in (a). (d) 3D reconstruction (left) with two orthogonal views AB and CD (right). (e−g) Slice views of the top, middle, and bottom layers at the heights of z1, z2, and z3, respectively, as indicated in (d). (h, i) Images obtained as the average of (e) and (g) and average of (e), (f) and (g), respectively. Unit cells are outlined. (j) Structure Figure 5. Three kinds of PbSe nanoparticle superstructures and model of the bct assembly. their thermodynamic stability relations. Energy Landscape of PbSe nanoparticle superlattice Other synchrotron X-ray scattering and calorimetry polymorphs experiments Different nanoparticle superlattice polymorphs with similar High-P SAXS/WAXS experiments of PbSe NC superlattices. energetics can be made by slight changes in growth condi- We conducted in-situ high-P small-angle and wide-angle X- tions. Using solution calorimetry with hexane as the sol- ray scattering experiments of PbSe nanoparticles with dif- vent, we measured the enthalpies of disassembly of three ferent sizes ranging from 3 to 10 nm and their assembled PbSe NC superlattices and thus determined their stability superlattices using DAC techniques. Depending on the size relations [8]. and/or shape of PbSe NCs, either a body-centered-cubic or face-centered-cubic superstructure is adopted. Two major Self-assembly of PbSe NCs into three superlattice poly- results: First, we successfully prepared a series of nano- morphs. Uniform PbSe nanoparticles with the shape of structures such as nanocube dimers and 1D nanorods/ truncated ocathedron were used as the building blocks. nanowires. Thus pressure tuning is an effective approach Through controlling solvent type and substrate curvature, to produce novel types of nanoparticle mesostructures we assembled the PbSe NCs into three different supercrys- that are difficult to synthesize with conventional chemical tals (Fig. 5a) - body-centered cubic (bcc), body-centered methods. Second, the combined SXAS and WAXS measure- tetragonal (bct) and face-centered cubic (fcc). The nearest ments allowed examination of structural evolution, phase interparticle surface distance increases from bcc to bct and transformations and mechanical behavior of both PbSe then to fcc supercrystals, and since they have the same nanocrystals (atomic-level) and their assembled architec- 520

Page 521

tures (meso-scale) as a function of pressure. (45): 8431. SAXS/WAXS measurements of NaYF4 nanorod supercrys- 3. Badding, J. V., J. F. Meng, and D. A. Polvani. Pressure tals. We successfully assembled NaYF4 nanorods into tuning in the search for new and improved solid state highly ordered supercrystals, in which both the atoms materials. 1998. CHEMISTRY OF MATERIALS. 10 (10): and nanoparticles have the same orientations. We have 2889. determined the 3D packing orientations and patterns of 4. Quan, , Luo, Wang, Xu, Wang, Wang, and Fang. Pres- nanorods in the “single supercrystal” using synchrotron sure-Induced Switching between Amorphization and SAXS and WAXS with the goal of understanding the correla- Crystallization in PbTe Nanoparticles. 2013. NANO LET- tions between atoms and nanoparticles as well as resulting TERS. 13 (8): 3729. physical properties. 5. Whitesides, G. M., and B. Grzybowski. Self-assembly at High-T calorimetric and SAXS/WAXS measurements of all scales. 2002. SCIENCE. 295 (5564): 2418. PbSe NC superlattices. Recently, we extended our ca- lorimetric investigations of PbSe NC superlattices from 6. Quan, , and Fang. Superlattices with non-spherical room to elevated temperatures. Combined with high-T building blocks. 2010. NANO TODAY. 5 (5): 390. synchrotron X-ray scattering, the measurements revealed a thermally-induced phase transitions of PbSe NC superlat- 7. Quan, , Xu, Wang, Wen, Wang, Zhu, Li, C. J. Sheehan, tices from bcc to fcc. The combination of calorimetric with Wang, Smilgies, Luo, and Fang. Solvent-Mediated Self- X-ray scattering measurements allowed correlation of the Assembly of Nanocube Superlattices. 2014. JOURNAL structural transition with its associated energetic change. OF THE AMERICAN CHEMICAL SOCIETY. 136 (4): 1352. Impact on National Missions 8. Quan, , D. i. Wu, Zhu, W. H. Evers, J. M. Boncella, L. D. The discovery of new materials that exhibit novel struc- A. Siebbeles, Wang, Navrotsky, and Xu. Energy land- tures and emergent properties at high pressures provides scape of self-assembled superlattices of PbSe nano- a strong impetus for harvesting such materials at ambient crystals. 2014. In PROCEEDINGS OF THE NATIONAL conditions. Through the state-of-the-art synchrotron X-ray ACADEMY OF SCIENCES OF THE UNITED STATES OF scattering and calorimetric measurements of series of AMERICA. Vol. 111, 25 Edition, p. 9054. lead chalcogenide, metal and other NC systems, we have determined the structural evolutions and transitions as a Publications function of pressure or solvent type, and thermodynamic Lu, , Yang, Quan, Lin, Bai, L. i. n. Wang, Huang, and Zhao. stability relations among various polymorphs for these Enhanced Electron Transport in Nb-Doped TiO2 nanoparticles and/or their assembled superlattices. More Nanoparticles via Pressure-Induced Phase Transitions. 2014. JOURNAL OF THE AMERICAN CHEMICAL SOCI- broadly, the obtained results have applications/implica- ETY. 136 (1): 419. tions for controlled nanoparticle synthesis and for design and fabrication of novel structured nanomaterials po- Quan, Z. W., Z. P. Luo, Y. X. Wang, H. W. Xu, C. Y. Wang, Z. tentially with enhanced or newly emerging properties. In W. Wang, and J. Y. Fang. Pressure-Induced Switching particular, the use of pressure enables systematic manipu- between Amorphization and Crystallization in PbTe lation of the nanophases and their assembled architec- Nanoparticles. 2013. NANO LETTERS. 13 (8): 3729. tures (e.g., interparticle spacing), thereby enabling tuning Quan, Z., D. Wu, J. Zhu, W. H. Evers, J. M. Boncella, L. D. A. of their properties for specific applications. This research Siebbeles, Z. Wang, A. Navrotsky, and H. Xu. Energy represents a new direction in high-pressure nanomaterials landscape of self-assembled superlattices of PbSe science with great potential for technological applications. nanocrystals. 2014. Proceedings of the National Acad- emy of Sciences . 111 (25): 9054–9057. References

1. TOLBERT, S. H., and A. P. ALIVISATOS. HIGH-PRESSURE Quan, Z., H. Xu, C. Wang, X. Wen, Y. Wang, J. Zhu, R. Li, C. STRUCTURAL TRANSFORMATIONS IN SEMICONDUC- Sheehan, Z. Wang, D. M. Smilgies, Z. Luo, and J. Fang. TOR NANOCRYSTALS. 1995. ANNUAL REVIEW OF Solvent-mediated self-assembly of nanocube superlat- tices. 2014. J. Am. Chem. Soc.. 136 (4): 1352. PHYSICAL CHEMISTRY. 46: 595. Wu, , T. M. McDonald, Quan, S. V. Ushakov, Zhang, J. R.
2. Wu, , Bai, Sun, R. E. Haddad, D. M. Boye, Wang, and Long, and Navrotsky. Thermodynamic complexity of Fan. Pressure-Driven Assembly of Spherical Nanoparti- carbon capture in alkylamine-functionalized metal- cles and Formation of 1D-Nanostructure Arrays. 2010. organic frameworks. 2015. JOURNAL OF MATERIALS ANGEWANDTE CHEMIE-INTERNATIONAL EDITION. 49 521

Page 522

CHEMISTRY A. 3 (8): 4248. Zhu, , Quan, Lin, Jiang, Wang, Zhang, Jin, Zhao, Liu, C. J. Brinker, and Xu. Porous Ice Phases with VI and Dis- torted VII Structures Constrained in Nanoporous Silica. 2014. NANO LETTERS. 14 (11): 6554. 522

Page 523

Materials for the Future Postdoctoral Research and Development Final Report NMR Study of Quantum States of Matter Joe D. Thompson 20130780PRD1 Abstract with existing theory. Further, experiments during the Quantum fluctuations that arise out of a zero-temper- last decade have found that an unconventional form of ature magnetic transition, a magnetic quantum-critical superconductivity tends to appear in the vicinity of a point, are a promising candidate for the mechanism quantum-magnetic transition, but the possible relation- that creates unconventional superconductivity, one ship between quantum fluctuations at this zero-temper- of the most intriguing of all quantum states of matter. ature transition and unconventional superconductivity Heavy-fermion materials, whose name comes from the remains an outstanding question. Nuclear magnetic heavy mass of their itinerant electrons, have emerged resonance (NMR) is a particularly powerful microscopic as prototypes for discovering new types of quantum- probe of magnetism as well as of the nature of quantum critical points and for exploring the relationship between fluctuations and their possible role in producing uncon- quantum fluctuations and unconventional superconduc- ventional superconductivity. tivity that also is known to be hosted in materials based Scientific Approach and Accomplishments on plutonium. In this project, we have used nuclear magnetic resonance (NMR) to probe and understand Nuclear magnetic resonance is most familiar in the con- quantum criticality in the heavy-fermion compound text of magnetic resonance imaging (MRI) that is used to CeRhIn5, the nature of unconventional superconductiv- probe soft body tissue, but because the nuclei of atoms ity in U2PtC2 and its possible relationship to quantum in a metal often have a spin like the nuclei of soft tissue, criticality, and the new heavy-fermion superconductor NMR also can be used to ‘see what’s inside’ a metal and PuRhIn5, which is an isostructural Pu-analog of CeRhIn5. more importantly to determine at a microscopic level In addition, NMR studies have explored the effect of sub- the nature of spins on electrons through their coupling stituting Nd atoms for Ce in the quantum-critical heavy- to nuclear spins. Electrons and the spins that they carry fermion superconductor CeCoIn5 and to determine the are responsible for magnetic order that can be tuned magnetic ground state of (Me4N)2PuCl6. Progress in this by a non-thermal control parameter, such as an applied project has advanced our understanding of the conse- magnetic field, chemical composition or pressure, to quences of quantum fluctuations that emerge from a absolute zero temperature. By analyzing NMR measure- zero-temperature magnetic transition. ments it is possible to deduce microscopic properties of magnetic order, information about quantum-fluctuations Background and Research Objectives of that order as its transition temperature is tuned to There are thousands of examples of a transition at finite zero temperature and how electrons form superconduc- temperature from a magnetically ordered state to a tivity that appears in the vicinity of the zero-temperature magnetically disordered state. At such a classical phase magnetic boundary. Heavy-fermion materials, in which transition, thermal fluctuations due to the finite tem- electron-electron interactions effectively increase the perature environment drive the transition. But, it also is mass of electrons by a factor of 100 to 1000 times the possible theoretically to have a transition from magneti- mass of non-interacting electrons, are prototypes for cally ordered to disordered states at absolute zero tem- discovering and understanding quantum states of mat- perature where there are no thermal fluctuations. In this ter, such as unconventional forms of superconductivity, case the transition is driven solely by rules of quantum and have been the focus of this research. In pursuing mechanics. There is a growing number of experiments this research, this project has used the power of NMR suggesting that such a quantum-driven transition exists, to make new discoveries as well as to understand better but none of these examples can be explained completely the complex quantum states that can appear in heavy- 523

Page 524

fermion materials, specifically U2PtC2, CeRhIn5, and Pu- field) to search for field-induced magnetic order in the RhIn5. In the following we highlight accomplishments from superconducting state of CeRhIn5. We found no evidence the study of these materials and more. for the expected magnetic order, possibly because the coupling between nuclear and electronic spins changed U2PtC2 has been known for nearly 30 years to be a su- from its value at atmospheric pressure and made NMR- perconductor below about 1.5K but the nature of the detection of magnetic order impossible. superconductivity has never been studied. NMR and NMR- Knight shift, a measure of the electron-spin response to a PuRhIn5 has exactly the same crystal structure as CeR- magnetic field, experiments were carried out on U2PtC2 hIn5 and also is a heavy-fermion material, but instead of at temperatures to 50 mK. The surprising conclusion from ordering antiferromagnetically like CeRhIn5 at atmospheric these experiments was that the superconductivity is highly pressure, it is superconducting near 1.5 K. Our NMR ex- unusual: superconducting electrons form with parallel spin periments on PuRhIn5 discovered that its superconductiv- alignment in contrast to the more typical antiparallel align- ity also is unconventional, so-called d-wave with zero net ment. This discovery is one of the rare examples of equal spin but finite orbital moment of superconducting elec- spin-pairing superconductivity.[1] Additionally, analysis of trons.[2] NMR measurements at temperatures above the the Knight shift showed that this unconventional supercon- superconducting transition were consistent with proximity ductivity developed out of a state in which ferromagnetic of PuRhIn5 to an antiferromagnetic quantum-critical point. correlations among electrons were particularly strong, rais- Indeed, similar measurements made on a sample in which ing the possibility that U2PtC2 is close to a ferromagnetic a small number of In (indium) atoms were replaced by Cd quantum-critical point. This possibility was supported by (cadmium) found that superconductivity was replaced with finding ferromagnetic order in samples of U2PtC2 where antiferromagnetic order. It, therefore, seems likely that a small number of Pt (platinum) atoms were replaced by quantum fluctuations of antiferromagnetic order play an nickel (Ni). essential role in producing the unconventional supercon- ductivity. During the course of this project, collaborators found evi- dence for two magnetic quantum-critical points in CeRhIn5 CeCoIn5 is an unconventional (d-wave) heavy-fermion induced by magnetic fields of 35 and 50 T at atmospheric superconductor at atmospheric pressure and its supercon- pressure. An interpretation of those discoveries requires ductivity is ‘born’ out of a quantum-critical state. Replac- assumptions about the nature of the magnetic order at ing 5% of the Ce (cerium) atoms by Nd (neodymium) to these fields. Because NMR can detect the nature of the make Ce0.95Nd0.05CoIn5 induces magnetic order in the magnetic order, proposals were submitted to the National superconducting state; however, microscopic coexistence High Magnetic Field Laboratory in Tallahassee for NMR of the superconductivity and magnetism has yet to be measurements to 30 T and at temperatures to as low as determined. Alternating current (ac)-susceptibility mea- 300 mK. Those proposals were accepted and we carried surements determined that the superconducting transition out the proposed experiments under these very demand- temperature is homogeneous throughout the material, ing conditions. The important conclusion is that a field indicating homogenously distributed Nd atoms. NMR mea- of 30 T does not change the magnetic structure, but the surements were performed on crystals of this compound ordered magnetic moment appears to have decreased by at temperatures 1.5K to 295K and demonstrated that the about 25 %. This result is consistent with the proposed in- normal state electrons behaved like local magnetic mo- terpretations that the field-induced quantum-critical states ments at sufficiently high temperatures. This is contrast to in CeRhIn5 involve both critical magnetic fluctuations and CeCoIn5 in which NMR indicated a quantum-critical state critical electronic degrees of freedom, a particularly inter- above the superconducting transition. This difference sug- esting and unusual type of quantum criticality. gests that the Nd ions are removing some of the quantum fluctuations and in the process creating the unusual situa- This unusual criticality also might be responsible for tion where magnetic order develops inside the supercon- superconductivity that appears in CeRhIn5 as its antifer- ducting state. romagnetic transition temperature is tuned toward zero- temperature by applied pressure. Earlier experiments had With interest in looking for an NMR signal from Pu (plu- suggested that magnetic order could be induced inside the tonium), we also pursued a search for this signal in the superconducting state by an applied magnetic field. To test organic complex (Me4N)2PuCl6. Initial experiments in for this possibility, we used NMR measurements at pres- a sample with high abundance of the 239-isotope of Pu sures to 2 GPa (approximately 20,000 times atmospheric 239Pu, which is most desirable for NMR, showed several pressure), temperatures to 50 mK and magnetic fields to signals. To identify the nature of these signals, NMR was 12 T (approximately 200,000 times the earth’s magnetic performed on a second sample with high 242Pu isotopic 524

Page 525

abundance, an isotope that does not give an NMR signal. nance study of the unconventional superconductor These experiments and simulation of NMR measurements PuRhIn5. Physical Review B. identified the measured signals as coming from NMR of Mounce, A. M., G. Koutroulakis, H. Yasuoka, J. Autschbach, the Cl (chlorine) nuclei and not Pu. Once this was clarified, K. A. Stosh, E. D. Bauer, J. D. Thompson, and D. L. Clark. we could fully replicate the NMR data by theoretical cal- Search for 239-Pu NMR in [(CH3)4N]2PuIVCl6. Physical culations. These results, the first NMR experiments on this Review B. Pu-based material, will be published in the near future. [3] Mounce, A. M., H. Yasuoka, G. Koutroulakis, N. Ni, E. D. Impact on National Missions Bauer, F. Ronning, and J. D. Thompson. Evidence for Unconventional superconductivity holds promise for excit- spin-triplet superconductivity in U2PtC2 from 195Pt NMR. 2015. Physical Review Letters. 144: 127001. ing new energy and defense technologies, but how and where that superconductivity might appear cannot be predicted. Empirically, experiments have shown that this superconductivity frequently appears as a magnetic transi- tion is tuned to zero temperature, but the nature of the quantum excitations and how they might induce supercon- ductivity are major unanswered questions. Fundamentally new understanding of the mechanism of unconventional superconductivity and its possible relationship to quantum fluctuations is needed to be able to predict how and when superconductivity appears. With unconventional supercon- ductivity also found recently in plutonium-based materi- als, superconductivity itself can be used to inform a much clearer perspective of states of matter than can arise from the electronic complexity of Pu. Using NMR to probe and understand novel quantum states of matter, this project has addressed DOE, Office of Science priorities called out in “Directing Matter and Energy: Five Challenges for Sci- ence and the Imagination,” in particular addressed the the need for science that is required to discover, understand and control complex and collective forms of matter that exhibit novel properties and respond to environmental conditions, enabling the creation of materials with novel innate functionality. References

1. Mounce, A. M., H. Yasuoka, G. Koutroulakis, N. Ni, E. D. Bauer, F. Ronning, and J. D. Thompson. Evidence for spin-triplet superconductivity in U2PtC2 from 195Pt NMR. 2015. Physical Review Letters. 144: 127001.
2. Koutroulakis, G., A. M. Mounce, H. Yasuoka, J. Mitch- ell, E. D. Bauer, and J. D. Thompson. Nuclear magnetic resonance study of unconventional superconductivity in PuRhIn5. Physical Review B.
3. Mounce, A. M., G. Koutroulakis, H. Yasuoka, J. Autsch- bach, K. A. Stosh, E. D. Bauer, J. D. Thompson, and D. L. Clark. Search for 239-Pu NMR in [(CH3)4N]2PuIVCl6. Physical Review B. Publications Koutroulakis, G., A. M. Mounce, H. Yasuoka, J. Mitchell, E. D. Bauer, and J. D. Thompson. A nuclear magnetic reso- 525

Page 526

Materials for the Future Postdoctoral Research and Development Final Report Electronic and Photonic Transport in Chiral Materials and Nanostructures Diego A. Dalvit 20130781PRD1 Abstract degrees of freedom in 2D Dirac materials. The work was We have performed extensive studies of the spin physics published in the Physical Review Letters [1]. In connec- and optical properties of 2D Dirac materials, which are tion with this work, they also collaborated with the ex- atomically thin materials with elementary excitations perimental group of Scott Crooker at the High Magnetic governed by the Dirac equation. These include graphene, Field Laboratory and expect continuing collaboration transition metal dichalcogenides, and topological insula- with theory and experiment. Wang-Kong has also initi- tors. ated a study of excitonic effects on the optical properties of 2D transition metal dichalcogenides and work is in Background and Research Objectives progress. Chirality is prevalent in Nature. Certain molecules exhibit Wang-Kong studied the heat transfer in topological intrinsic optical chirality in which left-handed and right- insulator systems in collaboration with T-4 staff member handed circularly polarized light propagate differently. Diego Dalvit and external collaborator Pablo Rodriguez With the advent of metamaterials, optical chirality has (U. Paris Sud, France). They developed a theory for the been engineered in artificial systems allowing control far field and near field radiative heat transfer between of the photon polarization. Spin-orbit (SO) coupling in topological insulator surfaces, and discovered a simple solids provides yet another important kind of chirality. power law for the near field heat transfer flux. The work In the last 20 years, research in SO coupling has been ac- was published as an invited article in the Special Issue tively pursued in spintronics, and has recently led to the on Casimir Physics of the Journal of Physics: Condensed discovery of new classes of materials known as topologi- Matter [2]. cal insulators and superconductors. Graphene and other newly identified 2D materials, on the other hand, also He has also collaborated with the experimental optics demonstrate a chiral property closely akin to SO cou- group of Prof. Peter Armitage at the Johns Hopkins Uni- pling. Study of electronic and photonic chiral materials versity. He performed the theoretical analysis in conjunc- offers exciting opportunities for discovering interesting tion with Armitage group’s experimental Faraday effect new physics and technologically important applications, measurements of topological insulators. They found an including spintronics-based information processing, new interesting magnetic field dependence on the transport material probe techniques, Casimir force repulsion, and scattering lifetime for electrons in topological insulators. tunable near-field radiative heat transfer. The work was submitted to the Physical Review Let- ters [3] and review is in progress. Wang-Kong has also Scientific Approach and Accomplishments maintained an ongoing collaboration with Prof. Allan The Director funded postdoc associate Wang-Kong Tse MacDonald at the University of Texas at Austin to study has studied the spin dynamics and spin noise of 2D the D.C. Coulomb drag between graphene sheets in the transition metal dichalcogenides in collaboration with quantum Hall regime. Preliminary results have been T-4 staff members Nikolai Sinitsyn, Darryl Smith and obtained and a draft of the manuscript for this work has Avadh Saxena. They derived the equations of motion for been developed. This is a difficult problem because the spin dynamics and spin noise, and developed a theory interaction effects in quantum Hall systems are not easily for optical observation of spin noise in these materials. amenable to precise quantitative description. Despite A unique property that was discovered from our work is achieving considerable progress in this problem, there the emergence of valley noise due to the unique valley remain subtle issues that need further investigation and 526

Page 527

understanding and work is ongoing in this problem. In tial progress in understanding the electronic and optical addition, Wang-Kong has recently initiated a problem to properties of chiral materials, and provided a strong theory study quantum friction and Coulomb drag with Diego Dal- support for the ongoing LANL programs at CINT on nano- vit. He expects to maintain a collaboration with Diego to science and technology. continue studying this problem after leaving from the lab. References Wang-Kong has mentored a Ph.D. student from University

1. Tse, W. K., A. Saxena, D. L. Smith, and N. A. Sinitsyn. of Texas at Austin to study the renormalization of optical Spin and Valley Noise in Two-Dimensional Dirac Mate- properties and plasmon properties in multilayer graphene rials. 2014. Physical Review Letters. 113: 046602. systems due to electron-electron Coulomb interaction. They formulated a theory based on the quantum kinetic 2. Rodriguez-Lopez, P., W. K. Tse, and D. A. R. Dalvit. Radi- equation to take into account both the band-renormaliza- ate Heat Transfer in 2D Dirac Materials. 2015. J. Phys.: tion and excitonic effects on an equal footing. Their results Condens. Matter. 27: 214019. show that there can be appreciable renormalization effects on the optical Drude weight and the plasmon frequencies 3. Wu, L., W. K. Tse, M. Brahlek, C. M. Morris, R. Valdez in graphene systems, which was reported in the T-4 BLABS Aguilar, N. Koirala, S. Oh, and N. P. Armitage. Obser- seminar. vation of cyclotron resonance and electron-phonon coupling in surface states of bulk-insulating topological He participated in grant proposal writing for one ER project insulator Bi2Se3. Physical Review Letters. and one DR project. The ER project proposal was writ- ten with Nikolai Sinitsyn and Scott Crooker, in which they Publications proposed to expand the studies on spin and valley noise, Lopez, P. Rodriguez, W. K. Tse, and D. A. R. Dalvit. Radiate the spin and valley dynamics of 2D transition metal dichal- Heat Transfer in 2D Dirac Materials. 2015. Journal of cogenides. The DR project was written with Houtong Chen, Physics: Condensed Matter. 27: 214019. Diego Dalvit and Nikolai Sinitsyn in which they proposed Tse, W. K., A. Saxena, D. L. Smith, and N. A. Sinitsyn. Spin to work on the optical properties of metamaterial hetero- and Valley Noise in Two-Dimensional Dirac Materials. structures comprising metamolecules and Dirac materials.
2. Physical Review Letters. 113: 046602. The DR project was finally funded. Wu, L., W. K. Tse, M. Brahlek, C. M. Morris, R. Valdes Agui- Wang-Kong attended the 2014 and 2015 APS March meet- lar, N. Koirala, S. Oh, and N. P. Armitage. Observation ings in which he delivered talks on “Spin and Valley Noise of cyclotron resonance and electron-phonon coupling in Dirac materials” and “Nonlinear Optical and Excitonic in surface states of bulk-insulating topological insulator Effects in Two-Dimensional Transition Metal Dichalco- Bi2Se3 . Physical Review Letters. genides”. In 2014, he visited the New York City College of Technology and delivered a talk on “Spintronics with Dirac electrons”. He delivered several internal talks at LANL: a CINT seminar on “Magneto-optical and Magneto-electric Effects in Topological Insulators”, a T-4 BLABS seminar “Many-body renormalization in Dirac Electronic Systems: Optical Properties”, and a T-4 BLABS seminar “Magneto- Coulomb drag in double-layer graphene”. He attended the LANL conferences “Information Science for Materials Discovery and Design” and “Mesoscale Science Frontiers”. In 2015, he visited and interviewed for faculty positions at the University of Delaware, Purdue University, University of Alabama, University of New Hampshire, and University of Massachusetts, where he delivered interview talks. He received an offer from the University of Alabama and gladly accepted the position. He will begin his new position in mid-August 2015. Impact on National Missions The research performed in this project made substan- 527

Page 528

Materials for the Future Postdoctoral Research and Development Final Report Alternating Positive-Negative Charge Systems: New Compounds and Synthetic Routes David E. Chavez 20130788PRD2 Abstract are two broad categories of strategies to incorporate This report describes the preparation of 1-amino- energetic properties in a molecule: conventional fuel/ 1,2,3-triazole-3-oxide (DPX2) and its transformation to oxidizer strategies seen in explosives such as TNT 1,2,3,4-Tetrazine-1-oxide. DPX-2 provides insight into (2,4,6-trinitrotoluene) or PETN (pentaerythritol tetra- a novel N-Oxide/N-Amino compounds, being the first en- nitrate) and high heat of formation compounds (ring ergetic material in this class. The ability of this material strain, high-nitrogen content). While high heat of forma- to undergo a nitrene insertion forming 1,2,3,4-tetrazine- tion compounds display unique properties, increasing 1-oxide was also studied, and evidence for this material, energy content by increasing nitrogen content can lead the first non-benzoannulated 1,2,3,4-tetrazine-1-oxide, to explosive sensitivity and decrease in thermal stabil- is presented. The existence of both of these materi- ity. This can be seen when 1,1’-azobis(1,2,3-triazole) [4] als opens new strategies in energetic materials design. and 1,1’-azobis(tetrazole) [3] are compared; the former DPX2 was characterized chemically (Infrared, Raman, can easily be handled, and the latter, despite only two NMR, X-ray) and as a high explosive in terms of energetic CH groups being replaced by N, often spontaneously performances (detonation velocity, pressure, etc.) and detonates during handling, and decomposes at a tem- sensitivities (impact, friction, electrostatic). DPX-2 was perature over 100 oC lower. Additionally, highly strained found to possess good thermal stability and moderate compounds such as 1,3,3-trinitroazetidine or octanitro- sensitivities, indicating the viability of N-amino N-oxides cubane have exceedingly long synthese times [5]. as a strategy for the preparation of new energetic mate- Strategies have been developed to stabilize high heats rials. of formation compounds [6,7]. This can be achieved by tailoring both molecular shape as well as the addition Background and Research Objectives of electron withdrawing substituents [8]. Perhaps the In the field of energetic materials chemistry a major most striking effect of removal of electron density from challenge is to prepare extremely high performance a nitrogen system, thereby increasing stability, is through materials while maintaining stability towards destruc- N-oxidation. 1,2,3,4-tetrazine-1,3-dioxide heterocycles tive stimuli such as heat or shock. These metastable are stable class of energetic compounds, often stable at materials offer unique insights into the factors affecting temperatures above 200 oC [9], while only one, ther- molecular stability [1,2] as a result of their lying on the mally labile, unoxidized 1,2,3,4-tetrazine ring system borderline of existence and non-existence. Efforts such is known [10]. Beyond the stabilization of a nitrogen as these require a unique mix of theory and experimen- system by a N-oxide, the zwitterionic nature of this func- tal work in order to fulfill practical requirements; how tional group leads to strong dipoles and intermolecular much energy can be packed into a molecule before it interactions which are known to stabilize energetic becomes too unstable for practical use [3]. Beyond this materials. 2,6-diamino-3,5-dinitro-1,4,-pyrazine-1-oxide area, energetic materials research is also actively fo- is an example of such an explosive [11], and our recent cused on novel energetic compounds and strategies for work with nitrotetrazole, azidotetrazole, and bistetrazole creating materials that possess reduced toxicological or oxides also confirm the ability of N-oxides to stabilize environmental footprints. energetic nitrogen heterocycles [12-14]. At the forefront of energetics research are new syn- Beyond N-oxides, N-amines have also demonstrated thetic strategies to impart unique energetic qualities to utility in energetic materials; the additional catenated a molecule. Within energetic materials design, there 528

Page 529

nitrogen atom increases the heat of formation and the mental sensitivities are 8.5 Joules towards impact and 192 NH2 unit is available for intermolecular interactions [15- N towards friction,(slightly less sensitive than RDX) and 16]. Molecules containing both N-oxides and N-amines insensitive toward electrostatic discharge. These results on the same ring could be very high performing explosives indicate that the novel energetic N-amino/N-oxide combi- [17-20]. Unfortunately the synthetic paths to N-amines and nation is capable of forming energetic materials that are N-oxides often preclude them being on the same molecule, not overly sensitive to mechanical or thermal stimuli, both let alone the same heterocycle. important for the design of new explosives and thus con- firming that this new energetic material strategy is viable Scientific Approach and Accomplishments and highly promising. In this work, we prepare the first energetic molecule containing both an N-oxide and N-amine on the same heterocyclic ring, by two distinct routes. The first route to the preparation of the new energetic material 1-amino- 1,2,3-triazole-3-oxide (DPX2, 1 ) was via hypofluorous acid oxidation of 1-amino-1,2,3-triazole. An unoptimized pro- cedure found that one equivalent of hypofluorous acid at -23 oC produced DPX2 (Figure 1). At higher temperatures the reaction of the hypofluorous acid appears to occur increasingly at the amine as opposed to the ring, result- ing in loss of the amino group and formation of 1,2,3-tri- azole as identified by NMR and mass spectrometry. When Figure 2. In order to confirm the molecular structure of , crystals increased equivalents of hypofluorous acid were used, suitable for x-ray crystallographic analysis were grown from ni- oxidized and deaminated products were formed including tromethane solvent. This figure confirms our proposed structure triazole-1-oxide, as well as higher unknown oxidation prod- and displays the molecule as it appears in the crystal structure. ucts. Recrystallization of the desired material from nitro- Non-H atomic displacement ellipsoids are 50% probability. methane gave crystals suitable for diffraction. X-ray quality With DPX-2 in hand we sought new routes to produce the crystals of DPX2 were grown from nitromethane as the material. In an alternate route, 1-(benzyloxy)-1,2,3-triazole solvent. Single crystal X-ray analysis of DPX confirmed the was aminated with tosylhydroxylamine (THA) producing molecular structure of the material and also showed that 1-amino-3-(benzyloxy)-1,2,3-triazolium tosylate. This mol- the compound has a density of 1.571 g cm-3 (Figure 2). ecule was then reduced without purification with hydrogen (Pd/C catalyst) giving the target compound (Figure 3). Figure 1. DPX-2 (1) was first synthesized through the use of the very powerful oxygen transfer reagent hypofluorous acid (HOF). The reaction proceeds to provide numerous oxidation products of 1,2,3-triazole, including (1). This product represents the first Figure 3. An alternate synthesis route to DPX-2 (1) was pursued example of an energetic materials containing both N-oxide and as the original process was low yielding. In this alternate route, N-amino moities in the same molecule. This combination of 1-(benzyloxy)-1,2,3-triazole was aminated with tosylhydroxyl- energetic functional groups may lead to novel explosive perfor- amine (THA), to provide the N-amino intermediate. This inter- mance and safety properties and will serve as the basis for the mediate compound was converted to DPX-2 (1) using hydrogen gas in the presence of a palladium catalyst. Although this route study of future materials in this class of energetic materials requires two steps, it is more efficient than the original process to The thermal stability of DPX-2 was also determined using prepare DPX-2 (1). differential scanning calorimetry, which showed that the In attempting to purify the 1-amino-3-(benzyloxy)-1,2,3-tri- material was stable up to 210 oC, which is comparable to azolium tosylate to obtain a crystal structure we observed many existing conventional energetic materials. The heat loss of the NH¬2 protons by nuclear magnetic resonance, of formation of DPX-2 was calculated to be 286.1 kJ/mol. a shift in the remaining two protons and additional shifts With this data in hand, we next sought to determine how in the carbon spectrum to 129.2 and 119.4 ppm. Mass sensitive DPX-2 was towards destructive stimuli such as spectrometry studies indicated that the produced material impact, electrostatic discharge, and friction. The experi- was two mass units lighter than the DPX-2. We suspected 529

Page 530

spontaneous nitrene formation and ring expansion, form- challenging, new routes to alternating positive and nega- ing the previously unknown 1,2,3,4-tetrazine-1-oxide (2). tive charge (ANPC) materials are possible, opening up new This result was verified by pursuing a deliberate route to potential strategies to access these novel materials for use produce 2 (Figure 4). in numerous applications within the weapons program. References

1. Klapoetke, T. M., and J. Stierstorfer. The CN7 Anion.
2. Journal of the American Chemical Society. :
4. Huynh, M. H. V., M. A. Hiskey, D. E. Chavez, and R. D. Gilardi. Synthesis, Characterization and Energetic Figure 4. Our work proceeded toward the synthesis of the 1,2,3,4-tetrazine-1-oxide (2) using the route displayed in Figure Properties of Diazido heteroaromatic High-Nitrogen
5. Synthesis route to 1,2,3,4-Tetrazine-1-Oxide (2). The first step C-N compound. 2005. Journal of the American Chemi- involved a nitrogen insertion reaction using manganese dioxide cal Society. : 12537. as the oxidizing reagent to provide the desired 1,2,3,4-tetrazine oxide ring system. This intermediate has a protecting group on 3. Klapoetke, T. M., and D. G. Piercey. the N-oxide however and required removal to attain the desired 1,1’-Azobis(tetrazole): A Highly Energetic Nitrogen-Rich compound (2). Removal of the protecting group using hydrogen COmpound with a N10 Chain. 2011. Inorganic Chemis- gas in the presence of a palladium catalyst provided the desired try. : 2732. compound, the first example of a non-fused 1,2,3,4-tetrazine- oxide compound. 4. Li, Y. -C., C. Qi, S. -H. Li, H. -J. Zhang, C. -H. Sun, Y. -Z. Yu, and S. -P. Pang. 1,1’-Azobis-1,2,3-triazole: A High- This work is both the first non-annulated 1,2,3,4-tetrazine- Nitrogen Compound with Stable N8 STructure and Pho- 1-oxide reported, as well as a new route to 1,2,3,4-tet- tochromism. 2010. Journal of the American Chemical razine-1-oxides via the nitrene insertion of a 1-amino- Society. : 12172. 1,2,3-triazole-3-oxide. Attempts to oxidize it to the 1,2,3,4-tetrazine-1,3-dioxide, providing a new route to this 5. Zhang, M. -X., P. E. Eaton, and R. D. Gilardi. Hepta- and energetically-useful system have so far not been fruitful. Octanitrocubanes. 2000. Angewandte Chemie Interna- We are currently attempting to obtain a crystal structure of tional Edition. : 401. the 1,2,3,4-tetrazine-1-oxide. Further attempts to convert 1 to 2 directly are currently ongoing. 6. Inagake, S., and N. Goto. Nature of Conjugation in Hydronitrogens and Fluoronitrogens. Excessive Flow of In conclusion, we have prepared and characterized the first Unshared Electron Pairs Into Single Bonds. 1987. Jour- energetic N-amino-N-oxide-substituted heterocycle, using nal of the American Chemical Society. : 3234. two routes to synthesize 1-amino-1,2,3-triazole-3-oxide from 1-amino-1,2,3-triazoles by hypofluorous acid oxida- 7. Kalinin, A. V., E. T. Apasov, S. L. Ioffe, and V. A. Tarta- tion, or from the amination of 1-(benzyloxy)-1,2,3-triazole kovsky. N-nitrohydrazines and their Salts. 1991. Bulle- followed by debenzylation. This unique energetic material tin of the Academy of Science USSR Division of Chemi- possesses no undo sensitivities towards impact, friction, cal Science (English Translation). : 988. or thermal stimuli, and confirms the potential utility of
6. Klapoetke, T. M., C. Petermayer, D. G. Piercey, and J. our strategy in energetics design. Additionally, we have Stierstorfer. 1,3-Bis(nitroimido)-1,2,3-triazolate Anion, prepared 1,2,3,4-tetrazine-1-oxide, the first non-benzoan- the N-Nitroimide Moiety, and the Strategy of Alter- nulated member of this class of compounds, using an ap- nating Positive and Negative Charges in the Design of proach that has not been utilized previously to reach this Energetic Materials. 2012. Journal of the American class of compounds. Chemical Society. : 20827. Impact on National Missions
7. Churakov, A. M., and V. A. Tartakovsky. Progess in This work will have a direct impact on energetic materials 1,2,3,4-Tetrazine Chemistry. 2004. Chemical Reviews. : for the future within the weapons program. The work has

shown that the strategy of N-Amino/N-oxide functionalized molecules can lead to interesting new molecular structures 10. Kaihoh, T., T. Itoh, T. K. Yamaguchi, and A. Ohsawa. First and energetic materials that are stable and safe to work Synthesis of a 1,2,3,4-Tetrazine. 1988. Chemical Com- with. Additionally, our work has shown that, while still munications. : 1608. 530

Page 531

11. Gokcinar, E., T. M. Klapoetke, and A. J. Bellamy. Com- putational Study on 2,6-Diamino-3,5-dinitropyrazine and its 1-Oxide and 1,4-Dioxide Derivatives. 2010. Journal of Molecular Structure: THEOCHEM. : 18.
12. Fischer, N., D. Fischer, T. M. Klapoetke, D. G. Piercey, and J. Stierstofer. Pushing the Limits of Energetic Mate- rials - the Sythesis and Characterizations of Dihydroxyl- ammonium 5,5’Bistetrazole-1,1’-diolate. 2012. Journal of Materials Chemistry. : 20418.
13. Gobel, M., K. Karaghiosoff, T. M. Klapoetke, D. G. Pierc- ey, and J. Stierstorfer. Nitrotetrazolate-2N-oxides and the Strategy of N-Oxide Introduction. 2010. Journal of the American Chemical Society. : 17216.
14. Klapoetke, T. M., D. G. Piercey, and J. Stierstorfer. The Taming of CN7: The azidotetrazolate-2-oxide Anion.
15. Chemistry a European Journal. : 13068.
16. Klapoetke, T. M., D. G. Piercey, and J. Stierstorfer. Ami- nation of Energetic Anions: High-Performing Energetic Materials. 2012. Dalton Transactions. : 9451.
17. Klapoetke, T. M., D. G. Piercey, and J. Stierstorfer. The 1,3-Diamino-1,2,3-triazolium Cation: A Highly Energetic Moiety. 2013. European Journal of Inorganic Chemis- try. : 1509. Publications Chavez, D. E., D. G. Piercey, S. Hemisch, C. Kirst, T. M. Klapoetke, and J. Stierstorfer. An energetic N-oxide and N-amino heterocycle and its transformation to 1,2,3,4-tetrazine-1-oxide. 2015. Propellants, Explo- sives, Pyrotechnics. : 491. Chavez, D. E., T. M. Klapoetke, D. A. Parrish, D. G. Piercey, and J. Stierstorfer. The synthesis and energetic proper- ties of 3,4-bis(2,2,2-trinitroethylamino)furazan. 2014. Propellants, Explosives, Pyrotechnics. : 641. 531

Page 532

Materials for the Future Postdoctoral Research and Development Final Report Microstructured Biohybrid Synthesis of Photosynthetic Assemblies Gabriel A. Montano 20130796PRD2 Abstract cal systems that obfuscates the key design principles This project was focused on developing novel photonic for optimizing solar light harvesting in artificial systems. materials that mimic the complex processes of light-har- Bio-inspired LH-materials should possess the desir- vesting and energy transfer in biological photosynthesis able properties of biological systems such as efficiency through materials self-assembly. Our approach was without necessarily mimicking the atomic level details of inspired by the most archaic of photosynthetic species, most natural antenna systems. In this work, we pro- photosynthetic bacteria. First, we use solvent exchange posed to replicate the function and efficiency of bio- strategies to generate amphiphilic polymer nanocom- logical light-harvesting complexes using self-assembly posites that demonstrate simple, two-component, approaches that are dictated by the natural propensities light-harvesting assemblies that exhibit morphology- of amphiphilic copolymers to self-assemble in solution dependent energy transfer efficiency. We extended our into membrane-like materials (micelles, bilayers, and solvent exchange process to generate polymer nano- monolyaers). composites that structurally and functional mimic one of Our objectives were to 1)demonstrate efficient, modu- the most efficient biological light-harvesting complexes lar light-harvesting and energy transfer in amphiphilic known, the chlorosome of green bacteria. Finally, we polymer nanocomposites through self-assembly, and create a supramolecular energy landscape consisting 2) develop supramolecular photonic nanocomposites of artificial chlorosome donors and planar, polymer that rival biological light-harvesting structure/function membrane embedded acceptors. Our described results relationships. indicate the potential of amphiphilic polymers in gener- ating modular, dynamic photonic materials using simple The goal of our work was to develop methods that allow materials approaches. for predicting and designing photonic nanocomposites that rival biological light-harvesting in function and ef- Background and Research Objectives ficiency, but completely through self-assembly of materi- Photosynthesis is the most prolific biological process als. on the planet with net primary production of biomass exceeding 100 Gt C yr-1. Photosynthetic organisms use Scientific Approach and Accomplishments complex and regulated multichromophore assemblies, Our approach in this effort was to build upon previous called light-harvesting (LH) antennas, to capture, con- research in our laboratory that found short-chain am- centrate and direct solar radiation to reaction centers phiphilic block copolymers (Poly(ethylene oxide)-block- that then carry out concomitant chemistry. Nature’s LH poly(butadiene) (PEO-b-PBD) are capable of forming antennas are remarkable, operating with high efficiency micelles in solution and monolayer or bilayers on solid in fluctuating environmental and photic conditions supports of thickness similar to biological membranes through nanoscale precision thus, serving as inspiration (~4-5 nm for bilayers) [1]. In this work, we exploit the in material design. LH antennas are diverse in structure thickness of these self-assembled polymer structures to and chromophore composition, but have ubiquitous induce the organization of chromophores in a manner function: to provide a network of light-absorbing mol- that allows for efficient energy transfer between donor ecules that move electronic excitation energy to a “trap” and acceptor molecules. that catalyzes endergonic electron transfer. Biological systems have accumulated structural complexity through Polymer Nanocomposites for efficient light-harvesting evolution and natural selection necessary for biologi- 532

Page 533

and energy transfer: Our initial strategy for develop- In addition to miceller nanocomposites, we tested the pro- ing photonic nanocomposites was based upon the self- pensity of PEO-b-PBD to spontaneously form monolayer/ assembly of PEO-b-PBD in solution to form micelles and bilayer films on solid supports. Indeed, polymers contain- subsequent bilayers. We introduced two chromophores, ing our donor-acceptor chromophores formed monolayer/ a donor (Bodipy-labeled lipid) and acceptor bacterio- bilayer supports that exhibited ETE greater than 95% chlorin (BC-116), that have energy overlap favorable for (Figure 1). These structures rival the ETE of natural photo- Förster resonance energy transfer (FRET) and naturally synthetic light-harvesting complexes and do so completely self-assembled with the polymer micelles via solvent through self-assembly of materials. Further, understanding exchange. Since the bodipy dye is anchored to the lipid, it the mechanisms that control ETE allows for design of pho- is expected to anchor close to the interface of the am- tonic response and potential of modular behavior [2]. phiphilic polymers. Similarly, the BC-116 is amphiphilic Polymer Chlorosome Nanocomposites (PCNs): Our work in nature, thus is also thought to anchor to the interface in this area was predicated upon the propensity for pEO-b- of the two polymer blocks. Thus, by intentional design pBD to form micelles in solution. One of the most archaic using amphiphilic chromophores, we allow self-assembly yet efficient light-harvesting complexes known is the chlo- to localize our molecules in proximity to one another for rosome of green photosynthetic bacteria. Chlorosomes energy transfer (Figure 1). Thus, our resultant structures are supramolecular structures containing 10-100’s of should exhibit the two necessary characteristics for FRET, thousands of bacteriochlorophyll (BChl) c (d,e, or f), which being energetically matched to transfer energy and in close are self-assembled through H-bonding and π- π stacking in- enough proximity to one another to transfer energy. We teractions, and are bounded by a single membrane leaflet found that energy transfer efficiency was dependent upon containing lipid and proteins (Figure 2). the amount of quencher molecule (BC116) added to the sample and independent of donor, which is consistent with previous theories of energy transfer in photosynthesis. Micelles exhibiting a maximum energy transfer efficiency (ETE) of >65% were observed (Figure 1). In this result, the polymer plays two important roles: 1) it provides the nanoscale framework necessary to keep the chromophores in close enough proximity for energy transfer and, 2) the Figure 2. Model organization in natural and nanocomposite amphiphilic nature provides the chemical structure neces- systems. In natural chlorosomes (A), a single membrane leaflet sary to organize the amphiphilic chromophores without containing lipids and protein provides a hydrophobic environ- ment for BChl c aggregation in the interior. Chlorosomes are and secondary type of coordination. attached to the cytoplasmic membrane through the baseplate protein that contains BChl a and serves as a mediator of energy transfer from BChl c in the chlorosome interior and BChl a in the membrane associated photosystem. The interior of the chloro- some also contains carotenoid molecules that aid light harvest- ing as well as photoprotection (not shown). (B) In the nanocom- posite, the lipid and protein envelop of the natural system is replaced by an amphipathic diblock copolymer. This organization scheme is starkly dissimilar to all other LH antennas where intricate protein coordination of chro- mophores is a requisite. Chlorosomes have considerable BChl concentration that results in effective harvesting of photonic energy while also being efficient at transferring energy to BChl a existing as a baseplate light-harvesting complex in the monolayer membrane of the chlorosome Figure 1. A. Components and structural assembly of photonic and to subsequent membrane-associated acceptors. From polymer micelle nanocomposites. B. Absorption spectrum show- a biomaterials perspective, the spectroscopic and physi- ing micelles of the bodipy donor (green) and BC-116 acceptor cal attributes of chlorosomes and the green bacteria LH (red) chromophores and micelles containing both species (blue). system are very attractive, in particular the predominantly C. Plots of Energy transfer efficiencey in micelles (solid circles) self-assembled nature of the system. Because pEO-b-pBD and bilayer films (open circles) as a function of acceptor concen- has a propensity to form micelles in solution, we attempt- tration. Dotted line is a theoretical fit of micelle data indicating ed to use the micelle interior as a potential reservoir to an ability to predict and design photonic response. 533

Page 534

allow for organization of Bchl c to form functional mimics assemblies capable of performing supramolecular energy of chlorosome. Through our work, we found that we were transfer in a model that is structurally similar to biological capable of generating structural/functional mimics of chlo- assemblies, however again, is completely through self- rosomes of long-range order (Figure 3) [3]. assembly. A schematic representation of our assembly can be observed in Figure 4. We then utilized a technique called HyperSpectral Imaging (HSI) to determine ETE of our nanocomposite assemblies. HSI allows one to deconvolve the components contributing to fluorescence observed at any given pixel thus, in the case of our system, allowing us to determine if contribution from an energy donor (PCN) contributes to the fluorescence of an energy acceptor Figure 3. Spectral properties of PCNs containing additional (membrane-DiR) (Figure 4). Our resultant nanocomposites chromophores. (A) Absorbance spectra of PCNs containing ad- exhibit ~55% energy transfer, [3] which is exceptional con- ditional amounts of Bchl . Additions of Bchl result in a shoulder sidering it is a proof of principle concept that has not been centered at ~790 nm. (B) Fluorescence emission spectra of optimized for efficiency. samples containing Bchl . (C)- AFM topograph of PCNs. Data are displayed in 3D and demonstrate chlorosome-like morphology of PCNs. Again, these nanocomposite structures were formed through solvent exchanged self-assembly allowing for modular control of composition in a scalable process. In addition, we were able to add additional chromophores such as Bchl a and carotenoid while retaining chlorosome function that resulted in supramolecular energy transfer. The generation of PCN’s is significant for several reasons. Firstly, prior to these results, it had been speculated that some level of coordination via protein-interactions in na- tive chlorosomes was necessary to generate the type of long-range order observed. Our results demonstrate that by providing a suitable environment, long-range order Figure 4. Three dimensional energy transfer in a spatial-en- of the Bchl c molecules will happen spontaneously. Sec- ergetic landscape. (A) Representation of the assembly process. ondly, the Bchl a present in chlorosomes that serves as an PCNs are drop cast on hydrophilic glass and become immobi- energy acceptor from Bchl c is thought to do so in the form lized on the substrate. The sample is then washed to remove of a protein-pigment highly ordered structure called the loosely-bound PCNs and then backfilled with PEO-b-PBD mi- baseplate. Our results suggest that protein is 1) unneces- celles containing DiR acceptor chromophores. PEO-b-PBD forms sary to induce the spectral properties observed by Bchl a continuous bilayer film between and around the deposited a in the baseplate and 2) self-assembly of Bchl a in PCNs PCNs. With selective excitation in to BChl c, emission from DiR is is sufficient to obtain Bchl c-a ETEs observed in biological observed demonstrating energy transfer. (B) Spectral overlap of PCN emission (green) and DiR absorbance (red). The data have assemblies. This does not necessarily mean that the Bchl been normalized to the peak of each trace and the overlap region a observed in native chlorosomes is not in the form of a is shaded. (C-E) fluorescence images showing the distribution and baseplate structure, but simply that such a structure is un- location fluorescence intensity from PCNs (C) and DiR (D) and the necessary to replicate similar function in PCNs. merged image (E). The images are scaled from zero to maximum intensity within the image. (F) Zoomed in view of the boxed area Multimembrane Supramolecular Energy Transfer: In an in (E) along with the intensity profile of the PCN and DiR compo- attempt to mimic the type of complex to complex energy nents along the indicated line. The scale bar in (C) represents 5 transfer observed in biological photosynthesis, we gener- μm. ated three-dimensional photonic-assemblies based upon the above described systems [3]. First, we adhered PCNs Ultimately, the results of this project demonstrate the to a hydrophilic substrate. Next, we added a solution of potential of generating modular photonic assemblies pBD-b-pEO polymer micelles containing an energy accep- based upon amphiphilic block copolymers. There are tor (fluorescent dye 1,1’-Dioctadecyl-3,3,3’,3’-Tetrame- significant advantages to the systems described herein to thylindotricarbocyanine Iodide (DiR)) molecule to create natural systems: 1) all of our approaches are based upon polymer bilayers. This process results in multi-membrane self-assembly and thus overcome many of the limitations 534

Page 535

of more synthetically arduous approaches: namely, our mophores: A Modular Platform for Efficient Energy approach is quick and modular, scalable and inexpensive; Transfer. 2015. NANO LETTERS. 15 (4): 2422. 2) the use of amphiphilic polymers allows for functional/ Collins, A. M., G. A. Montano, Y. M. Tian, J. Duque, and dynamic response as it is possible to include functional T. Sahin. Block Copolymers for Responsive, Energetic groups in polymer design that can be exploited by changes Nanocomposite Membrane Assemblies. 2014. In 58th in the polymer environment. In addition, our designs Annual Meeting of the Biophysical Society. (San Fran- have significantly added to the understanding of organiza- cisco, Feb. 15-19, 2014). , p. 420A. Cambridge, MA: Cell tion in chlorosomes, indicating that long-range order and Press. morphology and coupling of energy donor/acceptors is not dependent on proteins of the chlorosome, but can instead be induced by providing the proper environment such as the pEO-pBD micelle. Together, the results of our project significantly advance the criteria for designing polymer membrane nanocomposites for photonic applications. Impact on National Missions The work performed in our project directly addresses labo- ratory missions in energy security and controlled function- ality of materials. We have simplified the process of pho- tosynthetic light-harvesting and energy transfer to its basic principles, removing the intricate coordination observed in most previous attempts and in nature itself. Our materials are simple, modular and efficient, all of which are goals of next-generation photonic materials. In addition, the in- formation gained on polymer self-assembly and polymeric assembly of components can be utilized for other materials applications such as sensor design or responsive mem- brane systems allowing for design and control of emergent behavior of polymer nanocomposites. References

1. Goertz, M. P., L. E. Marks, and G. A. Montano. Biomi- metic Mono layer and Bilayer Membranes Made From Amphiphilic Block Copolymer Micelles. 2012. ACS NANO. 6 (2): 1532.
2. Adams, P. G., A. M. Collins, Sahin, Subramanian, V. S. Urban, Vairaprakash, Tian, D. G. Evans, A. P. Shreve, and G. A. Montano. Diblock Copolymer Micelles and Supported Films with Noncovalently Incorporated Chromophores: A Modular Platform for Efficient En- ergy Transfer. 2015. NANO LETTERS. 15 (4): 2422.
3. Collins, Aaron M., Jerilyn A. Timlin, Stephen M. An- thony, and Gabriel A. Montano. Polymer-dependent, self-assembled nanocomposites that structurally and functionally mimic green bacterial antenna complexes . Nature Communications. Publications Adams, P. G., A. M. Collins, Sahin, Subramanian, V. S. Ur- ban, Vairaprakash, Tian, D. G. Evans, A. P. Shreve, and G. A. Montano. Diblock Copolymer Micelles and Sup- ported Films with Noncovalently Incorporated Chro- 535

Page 536

Materials for the Future Postdoctoral Research and Development Final Report Broken Symmetries in Superconductors Albert Migliori 20130813PRD4 Abstract Understanding high-Tc is closely linked to understanding Broken symmetries lie at the heart of almost all mod- how a strongly correlated Mott insulator evolves into a ern physics problems — from the broken rotational and Fermi liquid. translational symmetries in magnets to the rather subtle In this research we employed resonant ultrasound spec- gauge symmetry breaking in electro-weak theory (the troscopy (RUS), which measures the resonant vibrational Higgs mechanism) and superconductivity. Sometimes frequencies in a crystal to better than part-per-million the symmetry breaking is known to exist, yet its form accuracy. Since the elastic moduli connect directly to remains a mystery. Two good examples of this are the the thermodynamics of the system, small changes in the hidden-order in URu2Si2, and the pseudogap in high- state of the system can be detected, even at high tem- temperature cuprate superconductors. Both problems peratures4, allowing T* to be measured across the phase are grand challenges in condensed-matter physics. We diagram. Further, the analysis of which elastic modes are developed new high-field transport techniques, and coupled to the phase transition gives information about improved the resonant ultrasound spectroscopy capa- the symmetry breaking at T*. bilities at LANL, to tackle these challanges. A second research goal was to employ ultra-high mag- Background and Research Objectives netic field transport to map out the Fermi surface of The high temperature superconductor YBa2Cu3O6+x the high-Tc superconductor YBa2Cu3O6+x. Using the (YBCO) has been at the centre of many important results unique capabilities at LANL, specifically the 100 T non- in condensed matter physics research, including mea- destructive magnet, we proposed to map out the Fermi surement of the London penetration depth and deter- surface in the underdoped pseudogap regime, providing mination of the Fermi surface by quantum oscillation complementary information to what was learned in the measurements. Despite these and other achievements RUS experiments. over the past 22 years, the fundamental nature of the physics in the pseudogap region of the cuprate (materi- Scientific Approach and Accomplishments als with copper-oxygen planes) phase diagram remains Dr. Albert Migliori, a pioneer in the field of RUS, has a mystery, even though it is now increasingly evident a well-established group at Los Alamos National Labs that it is paramount to understanding high temperature (LANL). Prior to Dr. Ramshaw’s arrival, preliminary mea- superconductivity (high-Tc) in these materials. While the surements by his group on YBCO single crystals showed breaking of both rotational and time reversal symmetry discontinuities in the sound propagation coefficients as a has been observed at the pseudogap onset temperature function of temperature, with features showing up at T*. T* in neutron scattering experiments, thermodynamic One of the goals of this project was to complete these evidence for a phase transition has yet to be measured. measurements and publish them. The experimental difficulty is that T* ranges from near 150 K to more than 300 K in underdoped YBCO: at these Concurrent with the RUS experiments, Dr. Ramshaw temperatures, signals from typical thermodynamic worked with Dr. Neil Harrison, performing high field probes such as specific heat are dominated by phononic quantum oscillation measurements on YBCO. The recent contributions. Another important question in the cu- development of the 100 Tesla pulsed field at LANL al- prates is how this pseudogap state and the small carrier lowed them to measure quantum oscillations in YBCO pockets evolve with hole doping into the large Fermi sur- at dopings higher than were previously possible, bridg- face seen in overdoped cuprates such as Tl2Ba2CuO6+x. 536

Page 537

ing the gap between the overdoped and the underdoped prate superconductors. 2014. NATURE COMMUNICA- Fermi surface measurements. Quantum oscillation mea- TIONS. 5. surements of the electron effective mass and Fermi surface Harrison, , B. J. Ramshaw, and Shekhter. Nodal bilayer- size nicely complemented the RUS projec. splitting controlled by spin-orbit interactions in under- doped high-T-c cuprates. 2015. SCIENTIFIC REPORTS. 5. We succeeded in publishing results on both of these main goals. Shekhter et al. Nature (2013) (see attached publica- Ramshaw, B. J., Day, Vignolle, LeBoeuf, Dosanjh, Proust, tion list) reports on a thermodynamic transition into the Taillefer, Liang, W. N. Hardy, and D. A. Bonn. Vortex pseudogap state in both the underdoped and near-optimal lattice melting and Hc2 in underdoped YBa(2)Cu(3)Oy. doped regions of the phase diagram. Ramshaw et al. Sci- 2012. PHYSICAL REVIEW B. 86 (17). ence (2015) shows the evolution of the cyclotron effective Ramshaw, B. J., S. E. Sebastian, R. D. McDonald, Day, B. S. mass and Fermi surface area in underdoped YBa2Cu3O6+x, Tan, Zhu, J. B. Betts, Liang, D. A. Bonn, W. N. Hardy, and and found a strongly enhanced electron effective mass— Harrison. Quasiparticle mass enhancement approach- indicative of a quantum-critical point near optimal doping. ing optimal doping in a high-T-c superconductor. 2015. In collaboration with other LANL scientists, we used our SCIENCE. 348 (6232): 317. newly-enhanced RUS capabilities to measure the heavy- Fermion superconductor PuCoGa5 (see Ramshaw et al. Ramshaw, B. J., Shekhter, R. D. McDonald, J. B. Betts, J. N. Proceedings of the National Academy of Science (2015)). Mitchell, P. H. Tobash, C. H. Mielke, E. D. Bauer, and Mi- We discovered valence fluctuations in this material, and gliori. Avoided valence transition in a plutonium super- found that they are likely the driver of the unusually high conductor. 2015. In PROCEEDINGS OF THE NATIONAL ACADEMY OF SCIENCES OF THE UNITED STATES OF superconducting transition in this material. AMERICA. Vol. 112, 11 Edition, p. 3285. A number of other publications were completed as well, Shekhter, , B. J. Ramshaw, Liang, W. N. Hardy, D. A. Bonn, F. particularly relating to pulsed-field measurements on high- F. Balakirev, R. D. McDonald, J. B. Betts, S. C. Riggs, and Tc superconductors (See publication list). Migliori. Bounding the pseudogap with a line of phase transitions in YBa2Cu3O6+delta. 2013. NATURE. 498 Impact on National Missions (7452): 75. The study of the cuprates and heavy Fermions has always pushed the envelope of experimental techniques–scanning tunnelling microscopy and angle resolved photoemission are but two examples with broad applications–Dr. Ram- shaw’s advancements during this project further our ability to perform RUS on small samples, and to perform high- precision transport in magnetic fields approaching 100 T. Mapping the effective mass of YBCO across phase diagram furthers our understanding of how these complicated materials become high-Tc superconductors. The study of PuCoGa5¬ provides insight into the unusual valence of plutonium—a material that is directly related to the core missions of LANL. Publications Doiron-Leyraud, , Badoux, S. R. de Cotret, Lepault, LeBoeuf, Laliberte, Hassinger, B. J. Ramshaw, D. A. Bonn, W. N. Hardy, Liang, J. -. Park, Vignolles, Vignolle, Taillefer, and Proust. Evidence for a small hole pocket in the Fermi surface of underdoped YBa2Cu3Oy. 2015. NATURE COMMUNICATIONS. 6. Grissonnanche, , Cyr-Choiniere, Laliberte, S. R. de Cotret, Juneau-Fecteau, Dufour-Beausejour, M. -. Delage, LeB- oeuf, Chang, B. J. Ramshaw, D. A. Bonn, W. N. Hardy, Li- ang, Adachi, N. E. Hussey, Vignolle, Proust, Sutherland, Kraemer, J. -. Park, Graf, Doiron-Leyraud, and Taillefer. Direct measurement of the upper critical field in cu- 537

Page 538

Materials for the Future Postdoctoral Research and Development Final Report Hybrid Metal-Semiconductor Nanostructures for Optimized Photosynthetic Algal Growth Jennifer A. Hollingsworth 20130816PRD4 Abstract ment and connectivity of these functional nano-building Our aim has been to develop approaches for increas- blocks can give rise to synergistic “emergent” inter-par- ing algal growth by way of a novel materials strategy. ticle interactions and properties. Thus, beyond address- Namely, we sought to develop quantum dots (QDs) ing technical challenges facing algal fuels development, and hybrid nanoscale constructs that could efficiently the proposed effort was also conceived to advance the color-shift solar light to the wavelengths most effec- development of functional materials-by-design, taking tively utilized by algae for growth and algal-oil produc- advantage of the postdoc’s previous work in establish- tion. Moreover, we pursued QD materials that would ing a synthetic framework for constructing multi-domain be stable and would easily incorporate into a hydrogel hybrid nanoparticles. His, as well as other, synthetic framework, where the hydrogel beads would serve as strategies for nanoscale architectures-by-design were the “vessels” in which algae would be grown (concept applied to the synthesis of novel, functional core/shell captured in patent application: 14/170,400). We further optical nanomaterials. Furthermore, in the context aimed to establish the nanomaterial-design principles of creating truly hybrid, multi-component structures, for realizing high-performance QD and hybrid QD-metal strategies other than solution chemical synthesis were or QD-antenna structures most useful for effective light- ultimately explored as more direct routes to “mixed-sys- conversion processes. tem” nanostructures. Specifically, we adapted scanning probe “direct-write” techniques to place QDs onto three- Background and Research Objectives dimensional nanoscale antenna/plasmonic structures, Fuels derived from biomass are a promising source of where the resulting hybrid systems are anticipated to energy. Cultivating aquatic microalgae in open ponds can afford similar enhancement of QD photoluminescence generate large quantities of biomass, and recently, Sayre properties as realized in QD-MNP couples. et al. showed a remarkable enhancement in algal growth Scientific Approach and Accomplishments inside hydrogel polymer beads. Further improvements in hydrogel-confined algae growth rely on efficient light Early in the project, the postdoc became most interested utilization. Photosynthesis is facilitated by absorption in the “multi-domain” aspect of QD hetero-structures, of blue and red solar photons by chlorophyll, but green i.e., how within nominally a single nanostructure, it is photons are transmitted. Our aim has been to develop possible to incorporate complexity to evoke emergent approaches for increasing algal growth by way of “opti- behavior through effects brought about by internal cally activating” hydrogel beads using fluorescent inor- nanoscale interfaces. He pursued novel chemistry and ganic quantum dots (QDs) and, possibly, metal nanopar- photophysical properties of multi-domain structures in ticles (MNPs) that exhibit surface plasmon resonance the form of core/multi-shell QD systems. The emphasis (SPR) peaks overlapped with QD absorption or emission. was on new non-toxic compositions for best compat- These nanoparticle inclusions would color-shift incident ibility as color-converting phosphors in algae produc- sunlight to wavelengths useful for photosynthesis (UV to tion. Specifically, he investigated indium phosphide (InP) blue, green to red), and, further, screen algal cells from QDs as core materials. Although not ideal for the algal- damaging UV radiation. sensitization application, he determined that CdS-coated InP nanoparticles produced the unprecedented prop- The intrinsic optical properties of single-domain QD and erty of dual suppressed-blinking emission. As a result MNP nanoparticles can be finely tuned by controlling of its remarkable behavior, this system was explored in particle size and shape. In addition, the controlled place- collaboration with researchers in MPA-CINT, C-IIAC and 538

Page 539

Vanderbilt University. The approach aimed to understand onto atomic force microscopy tips and subsequently writ- the underlying mechanism for both dual emission and ten onto pre-fabricated substrates, including 3D nanoscale dual-color blinking suppression by combining results from dielectric antenna, for which enhancements in radiative time-resolved photoluminescence (ensemble and single- rates and control of emission direction are expected. nanostructure), transient absorption, imaging and elemen- In summary, the postdoc will be a co-author on two journal tal mapping, and correlating these results with known publications (down-conversion QDs for solid-state lighting synthetic experimental conditions. With respect to syn- and near-infrared non-blinking QDs for direct-placement thesis, numerous surface treatments were applied prior to on optical antenna) and first-author on one journal pub- addition of the outer layers of CdS that caused dual emis- lication (novel dual-suppressed blinking QDs). Additional sion to either “turn on” or “turn off.” From this correlated collaborative publications are also expected to result from study, we were able to ascertain the design parameters for his work. Also significantly, he made significant contribu- achieving dual emission and for obtaining the remarkable tions to capabilities that will impact ongoing research in stability associated with non-blinking behavior for both MPA-CINT, including in the area of advanced nanomaterial emitted colors (red and infrared). In this way, the work manufacturing, while the materials he explored are now as implications for the realization of other multi-domain well suited for application as down-conversion (color-con- nanostructures possessing designed properties. verting) materials in optically enhanced algal production. The postdoc also explored a new approach to synthesis. As a first user of our custom automated-reactor system, Impact on National Missions he investigated our initial choice for the optically active DOE (basic and applied, EERE), DoD, NASA and NSF have all nanoparticle for incorporation into hydrogel polymer funded open or internal programs in algal biofuels, where beads — the ultrastable CdSe/CdS core/thick-shell QD. His NASA also aims to use the technology to treat wastewater. approach was to establish a better understanding of the DHS has reported on the potential impact that a successful colloidal growth conditions that support fabrication of biofuels program could have on addressing both national high-performance hierarchical nanostructures. In so doing, economic and environmental/energy security needs. Thus, he discovered that reproducibility in QD structure and it is now broadly recognized that algal biofuels develop- performance could be significantly enhanced by controlling ment will be a key component in our future energy pro- the quality of the core/shell interface, e.g., by careful han- file. To this end, the project aimed to work with leading dling of the cores prior to shell growth and by controlling algal biofuels experts in further enhancing the efficacy precursor addition at the shell-solvent interface. The latter of algal production. In doing so, it helped to establish a ultimately entailed manipulating QD-QD dipole-dipole novel materials strategy for optimizing light utilization by interactions to minimize the impact of particle assembly algae, namely, an approach based on optically functional- on the shell growth process, controlling precursor rate of izing a hydrogel algal growth matrix with QDs and MNPs. addition, optimizing the growth temperature for shell addi- At the same time, the postdoc developed novel optical tion, and choosing a core QD that affords minimal interfa- nanomaterials (see accomplishments) and worked with cial defects upon epitaxial shell growth. These insights are others on new approaches to creating hybrid semiconduc- proving invaluable as we move forward with applications tor-nanoparticle/dielectric-antenna 3D constructs using of this material system in other light-conversion applica- direct-write techniques. In this way, the effort contributes tions, including solid-state lighting. directly to DOE Basic Science initiatives, relevant to Me- soscale Science, Additive (Advanced) Manufacturing, and Two approaches to creating QD-plasmonic/antenna assem- the Scientific Discovery and Innovation Mission in the Basic blies were attempted — self-assembly from solution and di- Understanding of Materials. Importantly, the effort forges rected assembly using a scanning probe direct-write nano- new expertise at the interface between materials chemis- lithography system. The principal focus using the postdoc’s try/physics and bioscience. novel non-blinking infrared QD building blocks was the latter, as the antenna structure was theoretically predicted Publications to provide enhancement in the infrared, while metal-shell Buck, M. R., A. M. Dennis, H. M. Nguyen, H. Htoon, and J. synthetic hybrids developed in our lab (to date) are better A. Hollingsworth. Simultaneous near-infrared and vis- overlapped with visible-emitting QDs. Namely, the postdoc ible photoluminescence with suppressed blinking from synthesized stable, non-blinking near-infrared emitting QDs type-II InP core/shell nanocrystal quantum dots. Pre- with and without a silica spacer/protective layer. Through a sented at 248th American Chemical Society Meeting, collaboration and with the postdoc’s guidance, these QDs Fall, 2014. (San Francisco, August 10-14, 2014). were developed into liquid “inks.” The inks were loaded Buck, M., A. M. Dennis, F. Wang, H. M. Nguyenm, J. L. 539

Page 540

Casson, M. Sykora, H. Htoon, and J. A. Hollingsworth. Dual-color suppressed blinking in InP/CdSe/CdS core/ shell/shell giant quantum dots. Nano Letters (in prepa- ration). Dawood, F., C. J. Sheehan, M. R. Buck, A. M. Dennis, N. Karan, I. Staude, J. Dominguez, G. S. Subramania, A. R. James, I. Brener, N. A. Amro, and J. A. Hollingsworth. Direct Placement of Colloidal Nanocrystal Quantum Dots on Three-Dimensional Nanostructured Dielectric Antenna by Dip-Pen Nanolithography. ACS Nano. Hanson, C. J., M. R. Buck, K. Acharya, J. A. Torres, J. Kundu, X. Ma, S. Bouquin, C. E. Hamilton, H. Htoon, and J. A. Hollingsworth. Giant Quantum Dots: Matching Solid- State to Solution-Phase Photoluminescence Perfor- mance for Near-Unity Down-Conversion Efficiency. 2015. ACS Applied Materials and Interfaces . 7: 13125. 540

Page 541

Materials for the Future Postdoctoral Research and Development Final Report Ultrafast Measurements of Emergent Magnetism in New Complex Oxide Materials Scott A. Crooker 20140657PRD1 Abstract science efforts at Los Alamos. We expected to develop Strontium titanate (SrTiO3) is a foundational material in experimental tools (based on ultrafast pulsed lasers) the emerging field of complex oxide electronics. While that allowed us to “watch” — in real time — how magne- its electronic, optical, and lattice properties have been tization develops and decays in so-called new “complex studied for decades, this non-magnetic semiconductor oxide materials” such as strontium titanate (SrTiO3) and has recently become a renewed materials research focus related compounds. Specifically, we planned to build catalyzed by the surprising discovery of magnetism and and establish the technique of Time-Resolved Faraday superconductivity at interfaces between SrTiO3 and Rotation (TRFR), which is an all-optical method using other non-magnetic semiconducting oxides. The forma- polarized pulsed of light that allows one to measure how tion and distribution of oxygen vacancies, which donate magnetization develops on the timescale of atomic inter- two electrons for every one vacancy, is widely thought actions: millionths of millionths of seconds. to play an essential but as-yet-incompletely understood Scientific Approach and Accomplishments role in these observations. Recent signatures of mag- netization in gated bulk SrTiO3 have further galvanized Director’s-funded postdoc William Rice had a very suc- interest in the emergent properties of this material. cessful postdoctoral appointment at the NHMFL in Los Alamos, with important papers published in high-impact Background and Research Objectives scientific journals (see reference list). As originally To better understand how magnetization is created in proposed, he successfully established a fully-operational these nominally non-magnetic complex oxide materi- capability for measuring magnetization dynamics on als, we originally proposed to build and use an ultrafast picosecond timescales using the optical technique of pulsed pump-probe system in the National High Mag- Time-Resolved Faraday Rotation (TRFR). Bill used this netic Field Laboratory (NHMFL) optics lab. In this experi- new capability to measure the optically-induced magne- ment, wavelength-tunable light pumps the sample with tization in magnetically-doped semiconductor nanocrys- circularly polarized pulses and probes it with linearly po- tals for the first time. He wrote a paper on his data that larized light. In a Time-Resolved Faraday Rotation (TRFR) was accepted at the journal Nature Nanotechnology. scheme, a magnetic field is applied to a semiconductor Prior to this, Bill’s work on the static magnetic proper- perpendicular to the direction of the light. The pump ties of strontium titanate was successfully published in photo generates short-lived spins that are initially polar- Nature Materials. ized along the direction of the pump beam, which then precess about the applied magnetic field. Because the Additionally, as part of these magneto-optical studies, spin precession, as well as the excited spin relaxation, Bill also published a paper in the Journal of Physical alters the polarization of the linear probe light as a func- Chemistry Letters. tion of time between the pump and probe pulses, the spin relaxation dynamics can be measured by monitoring The TRFR capability established by Bill is now in constant the probe polarization. use to study other material systems such as the new atomically-thin semiconductors as represented by MoS2, A consistent theme in this work was to establish a new, and is proving to be a valuable tool for materials science powerful, general-purpose ultrafast capability for eluci- here at LANL. dating the picosecond-timescale dynamics of magnetiza- tion in new magnetic materials of interest to materials 541

Page 542

Impact on National Missions This project has established Laboratory capabilities in the areas of Materials Science, Emergent Phenomena, and Ma- terials-by-Design, and also in novel measurement methods that enable new scientific discovery at LANL. It is based on the very recent discovery of a new Emergent Phenomena: optically induced magnetization in complex oxide materi- als. Our approach for developing ultrafast optical methods to time-resolve magnetization dynamics with sub-pico- second resolution goes well beyond the specific example of strontium titanate that were investigated here and is applicable to a wide range of new materials in general, such as semiconductor nanocrystals for optoelectronic ap- plications (see results above). The experimental techniques (ultrafast optical studies based on Time-Resolved Faraday Rotation) that we developed as part of this project will advance new Laboratory capabilities that underpin a wide variety of Laboratory missions related to Materials Science and Controlled Functionality in Materials. Publications Rice, W. D., J. D. Thompson, P. Ambwani, C. Leighton, and S. A. Crooker. Revealing optically induced magnetiza- tion in SrTiO_3 using optically-coupled SQUID magne- tometry and magnetic circular dichroism . 2014. Jour- nal of Vacuum Science and Technology B. 32: 04E102. Rice, W. D., McDaniel, V. I. Klimov, and S. A. Crooker. Mag- neto-Optical Properties of CuInS2 Nanocrystals. 2014. JOURNAL OF PHYSICAL CHEMISTRY LETTERS. 5 (23): 4105. Rice, W. D., S. A. Crooker, P. Ambwani, C. Leighton, J. D. Thompson, G. Haugstad, and M. Bombeck. Persistent optically-induced magnetization in oxygen-deficient strontium titanate . 2014. Nature Materials. 13: 481. Rice, W. D., W. Liu, T. A. Baker, N. Sinitsyn, V. Kilmov, and S. A. Crooker. Revealing giant internal magnetic fields due to spin fluctuations in magnetically-doped colloidal nanocrystals. To appear in Nature Nanotechnology. 542

Page 543

Materials for the Future Postdoctoral Research and Development Final Report Shock-Driven Material Dynamics Investigated by Ultrafast X-ray Diffraction Cynthia A. Bolme 20140680PRD3 Abstract to generate high pressure-temperature conditions and The basic physical properties that dominate a material’s the highest available x-ray flux, which can be used to behavior under dynamic compression are frequently probe the structure. We studied the shock driven pro- governed by the material’s response on the scale of the cess of crystallization from an amorphous phase in fused atomic lattice. Traditionally, shock compression research silica to gain new insight into crystallization processes on has relied upon measurements of the material’s density the scale of the atomic lattice. and sound speed to deduce the lattice level response of Scientific Approach and Accomplishments the material. The recent advent of high brilliance x-ray sources (third generation synchrotrons and x-ray free We analyzed laser shock driven experimental data that electron lasers) has enabled the ability to directly probe was collected in situ during the shock event on fused the lattice level response of materials during shock silica. The experimental data included laser velocimetry compression. The Linac Coherent Light Source at SLAC and x-ray diffraction using the x-ray free electron laser National Laboratory was used to investigate the kinetics at the LCLS. By varying the laser pulse and peak power, and mechanisms of deformation of shock compressed the pressure and temperature conditions in the material fused silica using x-ray diffraction. The fused silica un- were tuned to examine the change in the nucleation and derwent a transition from amorphous to the crystalline growth kinetics of the stishovite grains. We completed stishovite phase. This x-ray diffraction data was analyzed data analysis of these crystallization kinetics, and we and a journal manuscript was submitted for publication. submitted a journal manuscript for review. We also ap- plied for future LCLS beam time to examine strength in Background and Research Objectives materials at high pressure. Understanding the fundamental processes which dictate Impact on National Missions basic physical properties such as strength, elasticity, plasticity and the kinetics of phase transformations/ This project furthered our understanding about how crystallization or recrystallization requires studies at the to perform shock compression experiments with the relevant length-scales (e.g., interatomic spacing and unique x-ray characteristics provided by an x-ray free grain size) and time-scales (e.g., phonon period). Infor- electron laser. Specifically, we improved our knowledge mation about these processes is important for physics about x-ray diffraction experiments in the complicated and chemistry, as well as applied sciences – geophysics regime involving non-crystalline (amorphous or liquid) and materials science as well as industry. Shock com- states. Work like ours builds capability for the proposed pression is an excellent platform for studying phase MaRIE signature facility. transformation kinetics and changes in microstructure This project also yielded knowledge about the rate of (to see deformation mechanism) because the nearly crystalline grain growth in the transformation of amor- instantaneous stress loading provides a fiducial from phous fused silica into crystalline stishovite. This infor- which to measure the material response. Until now, due mation is the subject of the journal manuscript resulting to the challenges of performing destructive shock wave from this project and adds to our basic understanding of experiments, direct observation of the material structure materials. during transformation has been difficult or impossible to resolve. The Linac Coherent Light Source (LCLS) at SLAC National Laboratory is unique in the world in that it has a dynamic compression driver (nanosecond optical laser) 543

Page 544

Publications Gleason, A. E., C. A. Bolme, H. J. Lee, B. Nagler, E. Galtier, D. Milathianaki, J. Eggert, J. Hawreliak, D. Fratanduono, R. G. Kraus, G. W. Collins, W. Yang, and W. L. Mao. Ultrafast visualization of crystallization and grain growth in shock- compressed SiO2. Nature Communications. 544

Page 545

Nuclear and Particle Futures

Page 546

Nuclear and Particle Futures Directed Research Continuing Project Illuminating the Origin of the Nucleon Spin Ivan M. Vitev 20130019DR Introduction Investment Plan of Nuclear and Particle Futures. Build- Nucleons (protons and neutrons) are not fundamen- ing upon our existing strategic partnership with Fer- tal constituents of matter, but are instead made up milab (E906, MiniBoone and LBNE), this project will of quarks and gluons, the elementary particles of the strengthen our fundamental science capabilities, bring strong interaction. There is compelling experimental new high-luminosity polarized target technology to evidence that the sum of the quark and gluon intrinsic LANL and provide a “major physics thrust to follow cur- angular momenta only contributes ~1/3 of the total pro- rent commitments to RHIC”. This project will maintain ton spin. Thus, the majority of the proton spin is unac- LANL’s leadership position in the field of spin physics and counted for, which has been referred to as the “proton produce the world’s most accurate polarized Drell-Yan spin crisis”. The missing fraction of the spin is likely car- measure-ment in proton-proton reactions. Our project ried by the orbital angular momentum of the quarks and is a timely and direct response to the DOE Mile-stone gluons. The long-term goal of this project is to develop (HP13) to “test unique QCD predictions for relations the experimental capability to measure the spin of the between single-spin phenomena in p-p scattering and proton in terms of contributions from the spins of the those observed in deep-inelastic scattering”. We antici- quarks, gluons and their orbital angular momentum. pate that this LDRD project will result in DOE Office of An equally important goal is to understand the relative Science funding for a LANL-lead spin physics program significance of these spin contributions in the theory of at Fermilab. Our integrated experimental and theoreti- strong interactions, Quantum Chromodynamics (QCD), cal program will allow us to lead in a major advance in and how they manifest them-selves in reactions with po- understanding the polarized nucleon structure through larized proton beams and/or targets. To this end, we will the only U.S.-based dedicated polarized DY experiment. determine the momentum distribution of quarks inside Providing a polarized target to FNAL will greatly enhance the proton, transverse to the proton momentum, from Fermilab’s capabilities and provide a much needed user which one can deduce whether quark orbital motion facility for spin physics. Our target will also be able to contributes significantly to the proton spin. polarize ND3, thus enabling one to extend the spin phys- ics to polarized neutrons. Furthermore, developing and To determine the distribution of quarks and gluons testing particle detector technology at high luminosity within a nucleon, we will construct a polarized target will directly benefit MaRIE, a LANL institutional prior- to install in Fermiab E906 experiment and carry out the ity. This detector development is also of fundamental world’s first measurement of the production of two importance to nuclear detection for applied missions like simultaneous leptons (electrons or muons) from a polar- nuclear nonproliferation. Such detectors often operate in ized proton target bombarded by a high-energy proton a high background environment. beam. From a detailed analysis of the azimuthal distribu- tion of such di-leptons, one can deduce properties of the Progress polarized nucleon structure. In particular, we will mea- Experimental: We presented an update of the experi- sure both the sign and magnitude of the quark Sivers ment to the Fermilab Program Advisory Committee in distribution, which is expected to be zero if the quarks January, with a new budget. We have started to collabo- have no orbital angular momentum. rate with FNAL on the necessary changes to the experi- ment. We met with the FNAL leadership to discuss the Benefit to National Security Missions support we will need from FNAL. Developed new beam This work is central to the FY13, FY14 LDRD Strategic line design with new optics. performed the acceptance 546

Page 547

tests of the magnet at Oxford after reorientation of coils guidance for what values of Pz the quasi-PDFs are good and refurbishing of the system. Magnet performed flaw- approximations of standard PDFs. Within the framework of lessly, we took new field homogeneity data, and shipped the spectator diquark model, we evaluate quark’s quasi- the magnet to UVa. The refrigerator has been refurbished PDFs and standard PDFs. We find that, for intermediate and leaks have been fixed. The refrigerator is ready for parton momentum fractions x, quasi-PDFs are good ap- system tests in July. We had a high level meeting with the proximations to standard PDFs (within 20-30%) when Pz nuclear physics (NP) program manager from DOE office of ≥1.5-2 GeV. On the other hand, for large x~1 much larger science to discuss future funding needs, and collaboration Pz > 4 GeV is necessary to obtain a satisfactory agreement between DOE NP and Fermilab. between the two sets. This study provides useful guidance for our lattice team in simulating PDFs on the lattice. We designed and manufactured first prototype of the target ladder and target cells. The nuclear magnetic reso- The calculations of Sivers and Boer-Mulders functions have nance (NMR) VME-system first prototype is currently in been completed on two ensembles with different lattice testing mode and will go in production after it has been de- actions to study discretization errors. We find that both bugged. The large ROOTS pumping system has been setup the domain wall and clover discretization schemes, which and is in the process of being tested at the Lujan center have very different chiral properties, give consistent results before being shipped to Fermilab. The slow control system at pion masses of about 300 MeV. These results are being for the polarized target has been designed. written up and will be presented at the Lattice 2015 Inter- national Conference. We are now developing and testing Theoretical: We investigated for the first time the role of the all-mode-averaging technique to improve the statistical gluons in generating the single transverse spin asymmetry. signal which is necessary to extend these calculations to Through such a study, we understand better how gluon’s lighter pion masses and higher momenta. transverse momentum is correlated with the proton’s transverse spin. We derived the next-to-leading order Future Work transverse momentum-weighted Sivers asymmetry in Experiment: Following a successful test of the supercon- semi-inclusive deep inelastic scattering (SIDIS) – those con- ducting magnet at University of Virginia (UVa) in FY15, tributions from the so-called three-gluon correlation func- the magnet and power supply will be shipped to Fermilab tions. Such a study provides a way to connect the quark (FNAL) and installed in the FNAL test area. The complete and gluon’s transverse motion, i.e., through QCD evolution nuclear magnetic resonance (NMR) and microwave sys- of the relevant quark-gluon as well as three-gluon correla- tems will be tested at UVa, then shipped to FNAL and in- tion functions. We will further generalize such a study to stalled. We will irradiate the frozen ammonium (NH3) tar- Drell-Yan production, which could help make more reliable get material, test a sample at UVa and transfer it to FNAL. prediction for spin asymmetry for our experiment. The ROOTS He pumping system will be moved to FNAL, assembled, tested and installed. UVa will complete any Lee and collaborators D. Kang (LANL) and O.Z. Labun (Ari- remaining work on the two target ladders and test them in zona) completed a paper on the first proof that the soft house. After testing, the liquid He refrigerator and target function for thrust distributions in Drell-Yan (DY), Deep ladders will be sent to FNAL. A complete set of software for Inelastic Scattering (DIS), and electron-positron (e+e-) col- slow controls and data acquisition for the polarized target lisions are equal up to two-loop order (second order in the system will be developed and tested. In collaboration with strong coupling alpha_s). With existing known results on FNAL we will design and build a cryogenic system, which the two-loop e+e- soft function, this is the first two-loop will be installed in the Seaquest area. result for the DIS and DY soft functions, and makes possi- ble up to NNNLL accuracy in predictions for all three types Theory: We will perform a global analysis of the spin-aver- of cross sections. This paper has been submitted to Physics aged cross sections for both semi-inclusive deep inelastic Letters B and is under review. Labun will visit T-2 this sum- scattering (SIDIS) and Drell-Yan (DY) production, as well as mer to collaborate with Lee on the definition, computa- the spin-dependent cross sections (Sivers asymmetry) for tion, and evolution of TMDPDFs in SCET using the rapidity SIDIS. We will promote our previous study to next-to-next- renormalization group. to-leading-log (NNLL) accuracy and make prediction of the DY Sivers asymmetry in the kinematics of our experimental Quasi-parton distribution functions (PDFs) are new theo- set-up. In lattice QCD, we will develop and test the all- retical objects, which could be computed directly on the mode-averaging technique to improve the statistical signal lattice, and will approach the standard PDFs in the large which is necessary to extend calculations to lighter pion proton momentum Pz. Since taking the Pz→∞ limit is not masses and higher momenta and carry out simulations feasible in lattice simulations, it is essential to provide 547

Page 548

on a new set of clover configurations being generated at 1560067. SIngapore: World Scientific. lighter pion mass and closer to the continuum limit. In Kang, D., O. Labun, and C. Lee. Equality of hemisphere on the one-loop computation of Transverse Momentum soft functions for e+e-, DIS and pp collisions at Distributions in Soft Collinear Effective Theory using the O(alphas**2). 2015. Physics Letters B. 748: 45. rapidity regulator and the rapidity renormalization group to compute their evolution to NLL accuracy and beyond. Kang, Z.. Single transverse spin asymmetries in polarized SIDIS and pp scattering. Invited presentation at The Conclusion 5th Workshop of the APS Topical Group on Hadronic Through a synergy between theory and experiment, we Physics. (Denver, 10-12 April 2013). will make a fundamental advance in our understanding of Kang, Z.. Nucleon spin: longitudinal, transverse, and the origin of the nucleon spin. In a strategic partnership evolution. Invited presentation at 2014 RHIC and AGS with Fermilab, we will construct a polarized target to carry Annual Users Meeting. (Upton, NY, 17-20 Jun. 2014). out the world’s most accurate measurement of di-lepton production from a polarized proton target bombarded by a Kang, Z.. Advances in the determination of TMDs from high-energy proton beam. We will develop state-of-the-art global analysis. Invited presentation at E1039/E906 theoretical and computational tools necessary to interpret Collaboration Meeting. (Santa Fe, NM, February 11-13, the experimental results that will directly lead to a major 2015). breakthrough in our understanding of the structure of mat- Kang, Z.. TMD evolution and global analysis. Invited ter and the theory of strong interactions. presentation at The 6th Workshop of the APS Topical Group on Hadronic Physics. (Baltimore, MD, April 8-10, Publications 2015). Almeida, L., C. Lee, S. Ellis, I. Sung, G. Sterman, and J. Walsh. Comparing and counting logs in direct and Kang, Z.. Transverse single spin asymmetry of the W effective methods of QCD resummation. 2014. Journal production at RHIC. Invited presentation at 2015 RHIC of High Energy Physics. 1404: 174. and AGS Annual Users Meeting. (Upton, NY, June 9-12, 2015). Dai, L., Z. Kang, A. Prokudin, and I. Vitev. Next-to-leading order transverse momentum-weighted Sivers Kang, Z. B., I. Vitev, and H. X. Xing. Transverse momentum- asymmetry in semi-inclusive deep inelastic scattering: weighted Sivers asymmetry in semi-inclusive deep the role of the three-gluon correlator . To appear in inelastic scattering at next-to-leading order. 2013. Physical Review D. PHYSICAL REVIEW D. 87 (3): -. Echevarria, M.. QCD Evolution of the Sivers Asymmetry. Klein, A.. An Experiment to Measure the Sivers Asymmetry 2015. In QCD Evolution Workshop 2014. (Santa Fe, of the Sea Quarks. 2015. In QCD Evolution Workshop. NM, 12-16 May, 2014). , p. 1560025. Singapore: World (Santa Fe, NM, 12-16 May, 2014). , p. 1560064 . Scientific. Singapore: World Scientific. Echevarria, M., A. Idilbi, Z. Kang, and I. Vitev. QCD Evolution Klein, A.. Drell Yan at FNAL with a Polarized Target. Invited of the Sivers Asymmetry . 2014. Physical Review D. 89: presentation at DNP Fall Meeting. (Waikoloa, HI, 6-11 074013. Oct. 2014). Engelhardt, M., B. Musch, T. Bhattacharya, R. Gupta, P. Klein, A.. The status of the polarized target and E1039. Hagler, A. Schaefer, S. Syritsin, and B. Yoon. Nucleon Invited presentation at Fermilab PAC Meeting. (Batavia, transverse momentum-dependent parton distributions IL, January 2015). from domain wall fermion calculations at 297 MeV pion mass . 2015. In 32nd International Symposium on Klein, A., P. McGaughey, and I. Vitev. Letter of Intent for a Lattice Field Theory (Lattice 2014). (Brookhaven, NY, Drell-Yan experiment with a polarized proton target. June 23-28, 2014). , p. 23. Trieste23: SISSA. 2013. Fermilab PAC . Gamberg, L., Z. Kang, I. Vitev, and H. Xing. Quasi-parton Kleinjan, D.. A polarized Drell-Yan experiment at Fermilab. distribution functions: a study in the diquark spectator Invited presentation at CIPANP2015. (Vail, CO, 19-24 model . 2015. Physics Letters B. : 112. May 2015). Jian, X.. Single Spin Asymmetries of Inclusive Hadrons Kleinjan, D., and A. Klein. A Polarized Drell-Yan experiment Produced in Electron Scattering from a Transversely to probe the dynamics of the nucleon sea. Invited Polarized 3He Target. 2015. In QCD Evolution presentation at Diffraction 2014. (Primošten (Croatia), Workshop. (Santa Fe, NM, 12-16 May 2014). , p. 10-15 Sept. 2014). 548

Page 549

Lee, C.. QCD Resummation: Comparing Direct and Effective parton distributions: Comparing Clover and Domain Methods. Invited presentation at Boston Jet Workshop. wall fermion results at ~300 MeV pion mass. Invited (Boston, MA, 21-23 Jan. 2014). presentation at The 33rd International Symposium on Lattice Field Theory. (Kobe, 14-18 July, 2015). Lee, C.. Comparing and Counting Logs in Direct and Effective Methods of Resummation. Invited presentation at SCET 2014. (Munich, Germany, 24-26 March, 2014). Lee, C.. The Evolution of Soft-Collinear Effective Theory. 2015. In QCD Evolution Workshop. (Santa Fe, NM, 12-16 May, 2014). , p. 1560045. Singapore: World Scientific. Lee, C.. Towards High Precision Resummation and Evolution Using SCET. Invited presentation at E1039/ E906 Collaboration Meeting. (Santa Fe, Feb. 2015). Lee, C.. Direct and Effective Methods of QCD Resummation. Invited presentation at Loopfest XIII. (New York, 18-20 June, 2014). Liu, M.. A new polarized solid proton target for Fermilab E1039 Drell-Yan experiment . Invited presentation at Spin2014. (Beijing, China, 20-24 Oct. 2014). Liu, M.. A new NMR system for E1039 polarized target. Invited presentation at Spin2014 . (Beijing, October 2014). Prokudin, A.. Transverse Spin Asymmetries and TMD Evolution. Invited presentation at POETIC V. (New Naven, CT, 22-26 Sep. 2014). Vitev, I.. Transverse momentum-wighted Sivers asymmetry in semi-inclusive deep inelastic scattering at next-to- leading order . 2014. In QCD Evolution Workshop 2013. (Newport News, 14-17 May 2013). , p. 1460019. New York: International Journal of Modern Physics. Vitev, I.. Theoretical Spin Effort at LANL. Invited presentation at E1039 Collaboration Workshop. (Batavia, IL, 18 March, 2014). Vitev, I.. A study of quasi-parton distribution functions in the diquark spectator model . Invited presentation at CD Evolution 2015. (Newport News, VA, May 26-30, 2015). Vitev, I., Z. Kang, and H. Xing. Transverse Momentum- weighted Sivers Asymmetry in Semi-inclusive Deep Inelastic Scattering at Next-to-leading Order . 2014. International Journal of Modern Physics Conference Series. 25: 1460019. Yoon, B.. Disconnected Contributions to Nucleon Structure. Invited presentation at Lattice QCD 2014. (Upton, NY, 23-28 Jun. 2014). Yoon, B.. Nucleon transverse momentum-dependent 549

Page 550

Nuclear and Particle Futures Directed Research Continuing Project Probing New Sources of Time-Reversal Violation with Neutron EDM Takeyasu Ito 20140015DR Introduction fundamental symmetries in nuclear science. The pro- An electric dipole moment (EDM) measures the separa- posed upgrade of the UCN (Ultracold Neutron) source tion of positive and negative charges within a system and is in line with the 2010 LANL NPAC (Nuclear, Particle, is an extremely sensitive probe of physics beyond the Cosmology, and Astrophysics) strategic plan that recom- standard model, the accepted theory of elementary par- mended development of a medium scale facility, and ticles. The neutron EDM (nEDM) is said to have “killed will benefit other experiments such as a neutron lifetime more theories than any other single measurement.” experiment, which received a strong community en- dorsement in a recent international workshop. Further- This LDRD project is a joint experimental and theoretical more, completion of the proposed experimental tasks project to probe new sources of time reversal violation will provide a viable alternate approach to the high-risk (T violation) with the nEDM. Using the LANSCE (Los Ala- SNS (Spallation Neutron Source) nEDM (neutron electric mos Neutron Science Center) Ultracold Neutron (UCN) dipole moment) experiment, which has a large P division source, we will develop a new nEDM experiment with a involvement. This will also lead to project funds coming sensitivity goal of 3x10-27 e-cm, a 10-fold improvement to LANL if a decision is made to go forward with this new over the current limit. More specifically, we will upgrade nEDM experiment. the existing UCN source which is expected to result in a 10-fold performance increase. In addition, we will build In addition, this project will strengthen an already a prototype nEDM experimental apparatus. With the up- existing successful synergy of theory and experiment in graded UCN source and prototype nEDM apparatus, we fundamental neutron physics. This is a unique strength will demonstrate that it is possible to perform an nEDM of LANL across US laboratories. This LDRD (Laboratory experiment with a sensitivity of 3x10-27 e-cm. Note that Directed Research and Development) will establish LANL the UCN density is the key parameter that is necessary T-division as one of the main US centers for nucleon for reaching the sensitivity goal. matrix elements computations with lattice QCD (Quan- tum Chromodynamics). This will put a solid basis to seek At the same time, to match the anticipated experimental additional external funding, through the channels of improvement, we will perform a comprehensive model- DOE (Department of Energy) Nuclear Theory topical col- independent analysis of T violation beyond the standard laboration centers as well as SciDAC grants in both High model (BSM) and we will pioneer first-principles calcula- Energy and Nuclear Physics. tions of the matrix elements, essential in extracting and bounding the underlying BSM sources of T violation from Additionally, the UCN source improvement will benefit EDM measurements. fundamental neutron physics experiments, actinide sci- ences, and detection technology. The goal of this LDRD project is to reduce the techni- cal risk of a full-fledged nEDM experiment at LANL with Progress sensitivity of 3x10-27 e-cm. In FY15, the experimental effort has made a lot of prog- ress in FY15, including: Benefit to National Security Missions The proposed research is central to both the LANL (Los (i) We tested the new guide coating material as well as Alamos National Laboratory) nuclear physics program the new guide coupling technology during the 2014- and the future of the US (United States) program on 2015 LANSCE beam time. The results obtained for the 550

Page 551

new UCN guide coating was very promising. For the new new physics scenario in the near future. We have written guide technology, while the results were encouraging we up our results in two manuscript, a long one with all the decided that further tests were necessary. We are planning details to be submitted to Physical Review D, and a shorter further tests at Institut Laue Langevin in France in July- one highlighting the results and impact, to be submitted to August 2015. Physical Review Letter. (ii) We have completed the UCN source design. The results Future Work were summarized in two documents, one describing the In FY16, the main focus of the experimental part will be physics design the other describing the engineering design. to 1) complete the construction of the ultracold neutron We held reviews to review both the physics and engineer- (UCN) source and the new UCN guides, 2) test them using ing design and out design was supported by the review UCN 3) install them, and 4) test their performance using committees. UCN. During the LANSCE beam time in late 2015-early 2016, the new UCN source and the UCN guides will be The theory effort has made progress on several fronts in tested using UCN from the existing UCN source. As soon as FY15: the beam time is over, the shielding blocks will be removed and the new UCN source and the UCN guides will be (i) We have performed a comprehensive analysis of the installed. Then when the LANSCE beam comes back in the ultraviolet divergence structure of the dominant new fall of 2016, the performance of the new UCN source and operators that violate time-reversal (T) and parity (P ) sym- the guide will be tested. At the same time we will make metries, thus inducing a neutron EDM. We have studied progress on the prototyping of the neutron electric dipole the off-shell mixing and renormalization of flavor-diagonal moment (nEDM) apparatus. This includes: a) continued dimension-5 T- and P-odd operators These operators testing of the high voltage, b) construction of the magnetic include quark electric dipole and chromo-electric dipole shielding and coils, c) construction of UCN detector and operators. In the process, we also provided a definition of spin analyzer, and d) construction of the 199Hg comagne- the quark chromo-electric dipole operator in a regulariza- tometer system. tion-independent momentum-subtraction scheme suitable for non-perturbative lattice calculations. We have also The theory effort will focus on two topics in FY16. (i) We presented results on how to combine the non-perturbative will perform the first numerical implementation to deter- numerical lattice QCD results and the perturbative analytic mine the size of the neutron EDM induced by the so-called results to make consistent physical predictions. These quark chromo-EDM operator, i.e. a CP-violating interaction results have appeared in a preprint (“Dimension-5 CP-odd of the quarks (that make up the neutron) with the electric operators: QCD mixing and renormalization”, 1502.07325), field associated to the strong force. (ii) We will carry out submitted for publication in Physical Review D. a comprehensive phenomenological analysis of the CP- violating couplings of the newly discovered Higgs boson (ii) We have obtained Lattice QCD (LQCD) results on to quarks and gluons, the carriers of the strong force. The the nucleon tensor charges including for the first time an analysis will include a comparative study of the sensitivity estimate of all systematic errors and a simultaneous ex- to these new interactions from both EDM searches and trapolation in the lattice spacing, volume, and light quark high-energy collider processes. masses. We have included in the analysis both connected and disconnected diagrams, and find that the discon- Conclusion nected contribution is smaller than the statistical error in the connected contribution. We found an overall error of By the end of the 3 year period, we expect to achieve the 10% on these matrix elements, a significant improvement following: over the previously estimated 50% uncertainty. Nucleon

1. Improved UCN source that increases flux by 20-fold tensor charges are highly relevant to our project because they determine the size of the nucleon electric dipole
2. Demonstration of the feasibility of an nEDM experiment moment (EDM) induced by quark EDMs, generated in with a sensitivity of 3x10-27 e-cm, through demonstra- many new scenarios of CP-violation beyond the Standard tion of the necessary UCN density stored in the prototype Model (BSM). We have used our results to derive model- nEDM chamber independent bounds on the EDMs of light quarks and to update the EDM phenomenology in split Supersymmetry, 3) First-principles calculation based determination of a popular supersymmetric scenario that survives all the the relevant matrix elements, essential in extracting and current constraints from collider searches. We are able to bounding the underlying BSM sources of T violation from propose a stringent test that might be able to falsify this EDM measurements 551

Page 552

Publications Bhattacharya, T., V. Cirigliano, R. Gupta, E. Mereghetti, and B. Yoon. Dimension-5 CP-odd operators: QCD mixing and renormalization. To appear in arXiv:1502.07325 [hep-ph], submitted to Physical Review D. Bhattacharya, T., V. Cirigliano, R. Gupta, H. W. Lin, and B. Yoon. Neutron Electric Dipole Moment and Tensor Charges from Lattice QCD. To appear in Physical Review Letters. Bhattacharya, T., V. Cirigliano, S. Cohen, R. Gupta, A. Joseph, H. W. Lin, and B. Yoon. Iso-vector and Iso-scalar Tensor Charges of the Nucleon from Lattice QCD. To appear in Physical Review D. Bhattacharya, T., V. Cirigliano, and R. Gupta. Neutron Electric Dipole Moments from Beyond the Standard Model Physics . 2014. In 31st International Symposium on Lattice Field Theory (Lattice 2013). (Mainz, Germany, July 29- Aug 3 2013). , p. 299. Trieste, Italy: PoS LATTICE2013 . Ito, T.. Science program at lanl ucn source. Invited presentation at The 2nd International Symposium on Science at J-PARC (J-PARC 2014). (Tsukuba, 12-15 Jul. 2014). Tang, Z., E. R. Adamek, A. Brandt, N. B. Callahan, S. M. Clayton, S. A. Currie, T. M. Ito, M. Makela, Y. Masuda, C. L. Morris, R. Pattie, Jr. , J. C. Ramsey, D. J. Salvat, A. Saunders, and A. R. Young. Measurement of spin- flip probabilities for ultracold neutrons interacting with nickel phosphorus coated surfaces . Nuclear Instruments and Methods in Physics Research Section A. 552

Page 553

Nuclear and Particle Futures Directed Research Continuing Project The Role of Short-lived Actinide Isomers in High Fluence Environments (U) Marian Jandel 20140046DR Introduction Benefit to National Security Missions In this project we will determine the nuclear physics Recent and urgent interest of various governmental properties of the actinide isomers, namely 236U and agencies, such as the Office of Non-Proliferation Re- 239U, and quantify their role in explosive environments. search and Development (NNSA NA-22), in more precise Apart from the knowledge of their existence and half- and correlated data on neutron-induced reactions will lives, there are no available data on the quantitative put the capabilities developed during this project into a population of these short-lived actinides isomeric states great position to support mission needs for systematic following neutron capture. Our preliminary experimen- measurements of isomeric state production in neutron tal studies at LANSCE suggest there may be substantial capture, correlated induced-fission data, predictive mod- population of high energy isomeric states in the neutron eling of capture and fission and refinement of transport capture process on actinides. These new, unanticipated codes such as MCNP to include important details on cor- results may have significant implications on our under- related radiation from neutron-induced reactions. New standing of nuclear reaction networks in high neutron signatures for special material detection, nuclear prolif- fluence environments such as those occurring in nuclear eration and forensics will be the objectives of the follow- explosions. New experiments will be staged and per- up research, which is a core mission of this laboratory. formed at LANSCE at the Detector for Advanced Neutron Knowledge of the production and de-excitation proper- Capture Experiments facility (DANCE), that will provide ties of actinide isomers will provide constraints on our new information about neutron-induced reactions on understanding of the performance of our well character- fissile actinides. With improved experimental capa- ized US nuclear tests. Of equal importance it will pro- bilities, we will obtain new precision data on yields of vide insights to global security missions associated with isomers after neutron capture on 235U and information nuclear forensics efforts where the ingoing properties of on prompt fission neutrons and gamma rays. Experi- devices will be unknown and actinide observables such mental and theoretical work on reaction rates leading to as 237U, 240U, are what will be available to diagnose (and on) isomeric states is very complex and represents the device characteristics. It is important that we under- a high risk part of this proposal. Low risk objectives stand how actinide isomeric population can influence include neutron-induced fission studies, which must be the production of the actinide observables. performed very accurately in order to address the pro- cess of neutron capture in detail.Experimental work will Progress include developments of a compact 4pi neutron detec- A prototype NEUANCE detector - NEUtron detector Ar- tor array that will be installed in the central cavity of the ray at DANCE (Detector for Advanced Neutron Capture DANCE detector array to detect prompt fission neutrons Experiments) consisting of six, each 5 cm square by 10 – NEUANCE (the NEUtron detector at dANCE). On the cm long, EJ-309 liquid scintillators was tested in beam theoretical front we will develop models for calculation with the new data acquisition system at flight path 14 of of neutron-induced reactions on excited states, neutron the Lujan Center at Los Alamos Neutron Science Center transport codes to populate excited isomeric states in (LANSCE). Data was taken using a 28 mg/cm2 thick 235U actinides, and nuclear de-excitation models to inter- target as well as non-fissioning capture samples. Both pret isomeric yields, and fission fragment de-excitation photomultiplier and “silicon photomultiplier” sensors including prompt fission n-gamma-fission fragment cor- were tested. Properties of the detectors, such as neu- relations. tron/gamma discrimination and pulse-height thresholds, were investigated. Although the details of the measure- 553

Page 554

ment are still being investigated, some preliminary results Finally, the reaction cross section calculation includes two are already driving design considerations for the full neu- parts, (1) the isomeric state production cross section by tron detector array. a neutron capture, and (2) the destruction of the isomers by neutron-induced fission. For the isomeric state produc- We were actively working on designing and testing a tion, we implemented a so-called discrete states embed- fission-fragment detector for DANCE. We built a prototype ded in the continuum in our Hauser-Feshbach code, which fission-fragment detector (FFD) based on thin scintillating allows us to produced more discrete gamma-ray lines. films. The detector consisted of 10 films attached to light- This is particularly important to estimate the production guide rings and all rings glued together to a solid cylindrical of the isomeric states whose excitation energies are high. body. We deposited a negligible amount (150 Bq) of Cf-252 We performed cross section calculations for three differ- at the last film. This prototype demonstrated that fission ent isomeric states. For the destruction of the isomers, fragments can be detected with the thin scintillating films the fission cross sections are calculated for the target in its and the scintillation can be transported through the whole ground and excited states, and an enhancement of the fis- system of light-guide rings. The next step was to design a sion cross section for the isomers is confirmed. Calculation real FFD for DANCE that to be tested with a neutron beam. of the fission transmission coefficient including the Class-I A FFD consisting five scintillating films was produced with and Class-II coupling is on-going, and preliminary imple- three of the foils loaded with small amount of U-235, 0.9 mentation was done. mg in total. The detector was used in a test run in the Lujan Center in the middle of January 2015. A good time- Future Work of-flight spectrum of the U-235(n,f) reaction was obtained • Carry out the measurement of neutron capture of and efficiency of the FFD of 14(2)% was estimated. Cur- U-235 with the thin film scintilator detector for fission rently we are working on production of a 10-film FFD fragments, at Lujan Center, LANSCE (October 2015) loaded with 20 mg of U-235 for a production data mea- and analyze data. surement at DANCE planned early in October, 2015. • Carry out the measurement of neutron capture of U-235 using the NEUANCE detector at Lujan Center, Measurements were carried out using a new data acquisi- LANSCE (January 2016) and analyze data. tion (DAQ) system. We received the DAQ hardware at the very end of FY14. In FY15 we installed, coded, and vali- • Implement the new DAQ using VME CAEN digitizers for dated the operation of the new DAQ such that it meets experiment 1) and 2) or exceeds all the performance specifications imposed by • Development of CGM: gamma-ray de-excitation model the project. We demonstrated the integrated operation of development using preliminary U236* data. Imple- different firmware types to handle the disparate detector ment data from Task 1) and 2) types required for the project to succeed. Current capa- • Development of CGMF - fission event generator for bilities include processing of signals from nearly any type NEUANCE design and optimization. Implement data of radiation detectors, including Germanium and Silicon from Task 1) and 2) detectors. In short, we demonstrated the DAQ’s ability to • Perform cross section calculations on the excited states perform the necessary experiments for the project to suc- calculations of U236*. Implement data from Task 1) ceed. and 2) • Sensitivity studies in transport simulations including On a theoretical side of the project, the CGMF (cascade delayed gamma-ray emission. Implement data from gamma multiplicity for fission) code was developed to Task 1) and 2) model the de-excitation of the fission fragments on an event-by-event basis, following the successive emissions • Develop a project continuation strategy and plan new of neutrons and photons. The emissions are treated in measurements and attract sponsors. the Hauser-Feshbach statistical model. Part of the current proposal, we have improved the treatment of emission Conclusion of electromagnetic radiation by taking into account the New data on isomer production after neutron capture will lifetimes of longer lives nuclear states (isomers), by imple- help benchmark the theory of compound nucleus de- menting in CGMF a sampling procedure that correlate the excitation and theory of the cross section on excited states. measurement coincidence window with the lifetimes of New capability will be created after completion of this the decaying nuclear states involved. We have made simu- project that will enable measurements of correlated data lations of neutrons and gamma rays for several coincidence on fission relevant to applications of National Security, windows and provided the results to the experimental col- Nuclear Forensics and Energy. leagues for detector analysis and design. 554

Page 555

Publications Baramsai, B., M. Jandel, T. A. Bredeweg, A. Couture, S. Mosby, G. Rusev, J. L. Ullmann, and C. L. Walker. Characterization and testing of EJ-309 and Stilbene scintillation detectors . 2015. In Proc. SPIE 9593, Hard X-Ray, Gamma-Ray, and Neutron Detector Physics. (San Diego, Aug 9). , p. 1. San Diego: SPIE. Jandel, , Baramsai, T. A. Bredeweg, Couture, Hayes, Kawano, Mosby, Rusev, Stetcu, T. N. Taddeucci, Talou, J. L. Ullmann, C. L. Walker, and J. B. Wilhelmy. Current and Future Research at DANCE. 2015. CGS15 - CAPTURE GAMMA-RAY SPECTROSCOPY AND RELATED TOPICS. 93. Rusev, G., B. Baramsai, E. M. Bond, and M. Jandel. Fission- neutrons source with fast neutron-emission timing. Nuclear Instruments and Methods A. Rusev, G., M. Jandel, B. Baramsai, A. R. Roman, J. K. Daum, R. K. Springs, E. M. Bond, T. A. Bredeweg, A. Couture, A. Favalli, K. D. Ianakiev, M. L. Iliev, S. Mosby, J. L. Ullmann, and C. L. Walker. Fission-fragment detector for DANCE based on thin scintillating films. To appear in Nuclear Instruments and Methods A, in print doi:10.1016/j.nima.2015.09.078. Rusev, G., M. Jandel, B. Baramsai, E. M. Bond, T. A. Bredeweg, A. Couture, J. K. Daum, A. Favalli, K. D. Ianakiev, M. L. Iliev, S. Mosby, A. R. Roman, R. K. Springs, J. L. Ullmann, and C. L. Walker. Development of a thin scintillation films fission-fragment detector and a novel neutron source. 2015. In Proc. SPIE 9593, Hard X-Ray, Gamma-Ray, and Neutron Detector Physics XVII, 959314 (August 26, 2015); doi:10.1117/12.2192440. (San Diego, Aug 9). , p. 1. San Diego: SPIE. al, A. Couture et. Enhancing the Detector for Advanced neutron Capture Experiments. 2015. EPJ Web of Conferences. 93: 07003. al, I. Stetcu et. Properties of prompt-fission gamma rays. 2014. Phys Rev C. 90: 024617. al, M. Jandel et. Current and Future Research at DANCE. 2015. In CGS15 – Capture Gamma-Ray Spectroscopy and Related Topics . (Dresden, August 25-29). , p. 1. Online: EPJ Web of Conferences. al, M. Jandel et. Capture and fission with DANCE and NEUANCE. To appear in European Journal of Physics A. al, S. Mosby et. A fission fragment detector for correlated fission output studies. 2014. Nuclear Instruments and Methods A. 757: 75. 555

Page 556

Nuclear and Particle Futures Directed Research Continuing Project Research Enabling a Next Generation Neutron Lifetime Measurement Steven Clayton 20140568DR Introduction Nuclear physics experimental capability, specifically the The lifetime of the free neutron is a fundamental input ability to understand and make measurements with to the Standard Model of particle physics. At present and about neutrons is a core capability of the nuclear the experimental situation with lifetime measurements weapons and nuclear nonproliferation programs which is shows a distribution of values for the lifetime that is sustained and advanced by this project. outside the claimed accuracy of these measurements. Progress Trapped ultra-cold neutrons (UCN) show great promise to both resolve existing discrepancies and to push the In the past 12 months, we met all of our second-year precision of lifetime measurements to the level of 0.1 goals, and furthermore developed a new detector for sec . A major focus of this project will be the implemen- ultracold neutrons that is highly promising for the future tation of a new technique for in situ detection of UCN of this experiment. The second year goals were to test inside the trap. In the first year, we will test a proto-type and characterize improvements to the experiment: system to do this. Subsequent years will make improve- • A new trap door system for improved loading. This ments to this system based on the experience in the was constructed and installed, and shows good load- first year. In the third year we expect to make extended ing efficiency consistent with the mock-up tested in running to make an initial lifetime measurement with the previous year. well-characterized uncertainties controlled at the level • Improved system for detecting the decay products of 1 second. from the in situ vanadium detector, and a new This project will develop and apply two major innova- vacuum jacket for the 52V decay products detector tions to the study of neutron decay with UCN. We will package. The new vacuum jacket was designed and make the first measurement of the UCN lifetime using an installed, allowing the gamma detectors to be placed asymmetric trap that will serve to minimize systematic closer to the vanadium sheet for improved solid uncertainties caused by marginally trapped neutrons. angle coverage, and improved shielding comprising Our measurement will also be the first to make an in lead bricks and borated plastic was installed that situ measurement of the number of trapped UCN, which greatly reduced the background event rate. will eliminate uncertainties introduced in the alternate • Measurements of the lifetime of ultracold neutrons technique of measuring the surviving neutrons by drain- in various gases were performed, setting detailed ing them from the trap. Measurements at comparable limits on the amount of residual gas permitted in accuracy to the present suite of experiments – but with the neutron lifetime apparatus to avoid a significant fundamentally different and understood systematic correction. uncertainties – will be key in understanding why earlier measurements underestimated their systematic uncer- With the improvements to the experimental apparatus, tainties. we collected data for a ~3 second (statistical) precision measurement of the neutron lifetime, and performed Benefit to National Security Missions several systematic studies, including tests of the normal- Better understanding the physical world - such as un- ization of the initial ultracold neutron population loaded derstanding the lifetime of the fundamental building into the trap, and tests of the time required to remove blocks of matter, is central to the mission of the Office of marginally-trapped ultracold neutrons from the trap (the science. neutrons that could slowly leak out from the trap dur- 556

Page 557

ing the holding period, thus leading to a loss mechanism a long time to be cleaned; and 3) studies of any vibration other than beta decay). Analysis of these data is ongoing induced heating of the trapped neutrons. The new, ac- and will result in either a physics result (if we have enough tive detector to replace the vanadium dagger offers three information to understand systematic errors at this level), distinct advantages. The first is improved statistics, due to or at least a measurement of the apparent holding time of higher detection efficiency of individual ultracold neutrons, the trap; we note that a physics result with ~3 second pre- and better beam utilization by eliminating the long period cision is already interesting in the context of the 6 second after every fill required to count decays from the vanadium discrepancy between the world average neutron lifetime detector. The second advantage of the active in situ detec- measured with the beam versus material bottle methods tor is that, because we will count each ultracold neutron as of previous workers. it is absorbed in the detector, a possible systematic uncer- tainty due to changes in characteristic absorption time into We have also developed a new ultracold neutron detector the detector is eliminated. Finally, the active detector is technology based on a thin film of 10B on a ZnS scintillator expected to be much less sensitive to background events. medium. Ultracold neutrons capture on the 10B layer, and Another goal for the third year of this project is detailed energetic reaction products create scintillation light in the mapping of the magnetic field of the trap, following up underlying ZnS layer, which is viewed by a light sensor such on a rough mapping this year to search for any low-field as a photomultiplier tube. Unlike gas-phase detectors such regions on the trap surface, where neutrons could be lost, as 3He-based technologies, the ultracold neutrons do not or zero-field regions in the trap volume, where neutrons have to pass through a window, avoiding the potential bar- could be depolarized and thus subsequently lost. rier and losses in the window material. We demonstrated these detectors with ultracold neutrons from the LANSCE Conclusion UCN facility, showing good efficiency and without the drop The overall goal of the project will be to characterize all tube necessary for windowed detectors; a manuscript de- systematic uncertainities involved in using this system to scribing this work is near publication in the journal Nuclear make a measurement of the neutron lifetime at the level Instruments and Methods A. We have begun develop- of 1 second. A new measurement at the 1-second level will ment of upgrades to the UCNtau experiment based on this lower the uncertainty of the PDG evaluation of the life- windowless detector technology, with plans to install them time. The neutron lifetime is one of its basic properties. It next year. The “cleaner” to remove superbarrier ultracold is an important input to the standard model of cosmology neutrons (those with energy greater than the trap poten- and understanding primordial nucleosynthesis. Combin- tial), now a passive upscattering material, will be replaced ing neutron lifetime with the other parameters of neutron with an active sheet that will enable a series of unprec- decay can be used to construct sensitive tests of the stan- edented systematics studies. Replacing the current in situ dard model of particle physics. detector, based on absorption onto a vanadium sheet and subsequent counting of the decay of 52V, with a new ac- Publications tive in situ detector that efficiently registers each neutron Salvat, D. J., E. R. Adamek, D. Barlow, J. D. Bowman, L. as it is absorbed, effectively improves the statistical reach J. Broussard, N. B. Callahan, S. M. Clayton, C. Cude- of the experiment as well as enables systematic studies not Woods, S. Currie, E. B. Dees, W. Fox, P. Geltenbort, possible with the current detector. K. P. Hickerson, A. T. Holley, C. Y. Liu, M. Makela, J. Medina, D. J. Morley, C. L. Morris, S. I. Penttilä, J. Future Work Ramsey, A. Saunders, S. J. Seestrom, E. I. Sharapov, S. K. Sjue, B. A. Slaughter, J. Vanderwerp, B. VornDick, In the third year of this project, we will take production P. L. Walstrom, Z. Wang, T. L. Womack, and A. R. data for a 1-second precision neutron lifetime measure- Young. Storage of ultracold neutrons in the magneto- ment (i.e. comparable to the current world-average preci- gravitational trap of the <span class=”aps-inline- sion), study systematic effects, and implement a new idea formula”><math xmlns=”http://www.w3.org/1998/ for an active “cleaner” for marginally-trapped ultracold Math/MathML”><mi>UCN</mi><mi>τ</mi></math></ neutrons, as well as an active, in situ, ultracold neutron span> experiment. 2014. Physical Review C. 89 (5): detector to replace the vanadium dagger. By counting 052501. superbarrier (above the trap potential) ultracold neutrons with good efficiency as they hit the cleaner, the active Wang, , M. A. Hoffbauer, C. L. Morris, N. B. Callahan, E. R. Adamek, J. D. Bacon, Blatnik, A. E. Brandt, L. J. cleaner will allow us to perform unprecedented systematic Broussard, S. M. Clayton, Cude-Woods, Currie, E. studies, such as 1) cross-checks of the normalization of the B. Dees, Ding, Gao, F. E. Gray, K. P. Hickerson, A. T. initial ultracold neutron population in the trap; 2) observa- Holley, T. M. Ito, C. -. Liu, Makela, J. C. Ramsey, R. tion of any marginally-trapped ultracold neutrons that take W. Pattie Jr., D. J. Salvat, Saunders, D. W. Schmidt, 557

Page 558

R. K. Schulze, S. J. Seestrom, E. I. Sharapov, Sprow, Tang, Wei, Wexler, T. L. Womack, A. R. Young, and B. A. Zeck. A multilayer surface detector for ultracold neutrons. 2015. NUCLEAR INSTRUMENTS & METHODS IN PHYSICS RESEARCH SECTION A-ACCELERATORS SPECTROMETERS DETECTORS AND ASSOCIATED EQUIPMENT. 798: 30. 558

Page 559

Nuclear and Particle Futures Directed Research Continuing Project k_effective: First Measurement of a Nanosecond-Pulsed Neutron Diagnosed Subcritical Assembly Anemarie Deyoung 20150044DR Introduction advanced the NDSE concept to a level that has a high Neutron generation is extremely sensitive to plutonium chance of succeeding at a first ns-scale pulsed neutron compressibility, and understanding the compressibility generation measurement. under the conditions encountered in a nuclear primary Benefit to National Security Missions is key to Life-Extension Program (LEP) options (includ- ing pit reuse), guarding against untoward aging effects, Success of this LDRD project would establish the tech- and establishing the safety/surety characteristics for the nical feasibility of proceeding toward a large future future stockpile. These and other key drivers have been weapons program in dynamic NDSE at the NNSS. Such studied in a classified venue by our XTD Division advisors a program would provide the first modern quantitative to the LDRD. In this project we will develop a precision understanding of the neutronic properties of Pu at high measurement technique to determine the nuclear gen- pressures and thus provide experimental validation of eration in a subcritical system, with an accuracy compa- weapons models which is key to Life-Extension Program rable to that obtained in nuclear testing. The Weapons (LEP) options (including pit reuse), guarding against Program requirement for an innovative alternative to untoward aging effects, and establishing the safety/ nuclear testing is a dynamic pulsed measurement of surety characteristics for the future stockpile. This also neutron generation. The focus of this proposal is the fea- has possible applications in the Global Security arena. sibility of a subcritical static neutron generation experi- The Nuclear Engineering and Non-proliferation (NEN) ment. Division is expert in the use of passive and active diag- nostic techniques for detection and characterization of Previous subcritical experiments have not had the configurations of special nuclear materials. The neutron diagnostic capability to infer nuclear generation. Our generation measurement offers an opportunity to ex- project involving Neutron Diagnosed Subcritical Experi- plore different techniques for simultaneously collecting, ments (NDSE) provides neutron generation, which is utilizing, and analyzing all the available time-dependent an extremely sensitive integral constraint on both the induced signatures. distribution and nuclear properties of materials. Our proposed NDSE capability would enable inference of Additionally, success of this project would advance the neutron generation or, more precisely, “alpha” (α), with scientific state-of-the-art in gamma-ray detectors and the accuracy needed for weapons analysis. We will per- DPF neutron sources. As well it would enable advanced form a high precision measurement of α using a static criticality studies, e.g., study of fast pulse excitation of subcritical assembly. The challenge is to make the mea- complex static critical assemblies. Measuring the alpha, surement under laboratory conditions with ~10 orders using a fast neutron pulse, of a static but complex critical of magnitude fewer gammas than were available during assembly, creates a new regime which permits advanced nuclear testing. testing of simulation codes. Observation of alpha as it changes from an initial value (14 MeV excitation) to its We will develop a capability that includes (1) a pulsed equilibrium value, would allow separation of various neutron source and (2) a detector and data acquisition contributions such as delayed neutrons, effects of cross system with proper shielding and collimation to measure section libraries for the unresolved resonance regions, the gammas emitted from the package. The recent ad- and the time-dependence of complex moderation and vances in pulsed neutron sources, detector technologies, reflection processes. fast electronics, and simulation-based analyses, have 559

Page 560

Progress Special Nuclear Material (SNM) target; and it enables The safety and performance of the stockpile could benefit validation - gamma falloff will be measured in the lab from a neutron-diagnosed subcritical experiment (NDSE) (future Pu sphere gamma measurement would have capability to challenge and inform our judgment in pre- independent Rossi alpha measurement) dicting the nuclear trigger’s capability to generate enough • DPF experiment – Scott Hsu, Robert Rundberg /LANL neutrons to go supercritical. NDSEs quantify the genera- tion of neutrons in a subcritical pit that is the fundamen- • Use of optical spectrometer allowed first measurement tal mechanism generating energy in nuclear weapons. of contaminants in the DD DPF pinch as well as elec- Neutron generation is extremely sensitive to plutonium tron density; use of shadowbar designed by George compressibility, and understanding the compressibility of Morgan allowed for precision measurement of DT DPF plutonium under the conditions encountered in a nuclear neutron profile weapon primary will be a key factor in their developing life-extension program (LEP) options (including pit reuse), • Boron target has been procured and testing has begun understanding aging, and establishing the safety/surety at LANL. The Laboratory is designing and procuring the characteristics for the future stockpile. iron and borated poly collimators to use as shielding for the static test. LANL is ready for the gamma mea- To these ends, we are developing an NDSE capability com- surement in September or October. prising (1) a pulsed neutron source producing a neutron flash and (2) a detector and data acquisition system on a Future Work dedicated line of sight that will measure gamma-rays emit- Our goal is to perform an accelerated gamma measure- ted from the dynamic package under interrogation. This ment on a 10B target using concrete line of sight at NNSS capability would provide a unique ability to infer neutron by Oct 2015, and to perform an NDSE measurement on a generation and answer critical questions related to integral static bare SNM assembly at the NNSS by September 2016. plutonium behavior. Here we describe our recent R&D ad- vances in pulsed neutron sources, detector technologies, Conclusion fast electronics, and simulation-based analyses. We must develop novel gamma-ray detector technology to meet the requirements of the neutron generation mea- Accomplishments to date are summarized as follows. surement; this includes new materials to achieve a signal • DPF modeling – Hui Li, Anna Hayes /LANL with large dynamic range and ultrafast few-ns scale time response. We must also develop a robust pulsed neutron • Provided initial confirmation of physics behind reen- source based on Dense Plasma Focus (DPF) technology; trant tube capability to make single pulse; intent is to this lies at the forefront of plasma physics. The reduced feed neutron output into MCNP modeling funding of this project delays the first measurement on a static subcritical assembly to 20 months; reduced funding • Gamma ray detector development – Vincent Yuan / for criticality expertise increases our reliance on infrastruc- LANL ture support and generally increases the risk of a further delay in the experiment. • Detector lab testing of fast scintillators and Cerenkov media led to a downselect where we chose Liquid VI Publications scintillator and manufactured a 2x2 array with each DeYoung, A., T. Goorley, A. Hayes, G. L. Morgan, V. W. C. cell 15”x15”x8” for use on the gamma falloff experi- Yuan, R. S. Rundberg, H. Li, G. Jungman, W. Myers, ment S. C. Hsu, T. J. Haines, E. Guardincerri, C. Hagen, D. Lowe, S. Molnar, P. O’Gara, I. Garza, E. Hunt, J. • MCNP modeling of experiment – Tim Goorley /LANL Tinsley, M. Kelley, J. Gatling, T. Waltman, R. Buckles, A. Luttman, N. Sipe, J. Friedman, N. Bennett, and E. • Modeling of experiment design reduces background Christensen. (U) Progress towards the NDSE LDRD so that we can design the line of sight and make a gamma measurement on a static assembly. To appear good measurement; provides for prediction of gamma in Nuclear Physics Design Physics Conference. (Los measurement Alamos, NM, 19-23 Oct 2015). • Design of novel Boron 10 target – George Morgan / DeYoung, A., V. Yuan, G. Morgan, C. Hagan, A. Obst, R. LANL Rundberg, A. Hayes, T. Goorley, H. Li, G. Jungman, W. Myers, S. Hsu, T. Haines, J. Lestone, E. Guardincerri, • Boron target avoids the large costs required for a M. Fowler, P. Koehler, J. Gatling, E. Hunt, M. Kelley, D. 560

Page 561

Lowe, B. Davis, and J. Tinsley. New diagnostic: neutron diagnosed subcritical experiments (LA-CP-14-20196). Presented at Nuclear Explosive Code Development Conference 2014. (Los Alamos, NM, 20-24 October 2014). Hsu, S.. DPF benchmarking diagnostics: present and future. Presented at DPF Workshop. (LLNL, Livermore, CA, 29 June 2015). McKenzie, G. E., T. J. Grove, W. L. Myers, and R. G. Sanchez. Reactivity worth studies associated with a Rossi-Alpha measurement system (LA-UR-15-25941) . Presented at American Nuclear Society Winter Meeting. (Washington DC, November 2015). Myers, W., J. Bounds, T. Cutler, J. Goda, J. Goettee, T. Goorley, S. Harbour, D. Hayes, R. Kimpland, S. Klein, D. Mayo, C. Moss, and S. Watson. Applications of fast burst reactors in support of data analysis tool development for the Neutron Diagnosed Subcritical Experiment ( LA-UR-15-27621). To appear in Nuclear Explosives Design Physics Conference. (Los Alamos, NM, 19-23 Oct 2015). Yuan, V., A. DeYoung, A. Obst, and G. Morgan. NDSE detectors development. Presented at 2015 Nevada working group meeting, LA-UR-15-22230 . (Las Vegas, Nevada, 31 March 2015). 561

Page 562

Nuclear and Particle Futures Directed Research Continuing Project Multi-Scale Kinetics of Self-Regulating Nuclear Reactors Venkateswara R. Dasari 20150058DR Introduction of the full-scale reactor to confirm self-regulation and Los Alamos is pioneering the development of self-reg- quantify uncertainty. ulating small compact nuclear reactors (SCRs). These Benefit to National Security Missions reactors are of immense interest to the community of sponsors who have power needs in space, underwater Los Alamos is pioneering the design of self-regulating and in remote areas. In this research, we propose to low-power small compact reactors. Such “fission bat- demonstrate the feasibility of self-regulation by develop- teries” are of immense interest to the community of ing and validating models for reactor behavior in tunable sponsors who have power needs in space (NASA/DoD/ ‘micro-engineered’ materials specifically designed to IC), underwater (DoD/IC/DHS) and in remote areas achieve self- regulation. Reactors that self-regulate fis- coupled with renewable sources (DOE). Despite these sion power to match varying operational needs and that transformative aspects, there is reluctance to adapt protect the reactor core from exceeding thermal limits reactor technologies because they are perceived to be during accidents are ideal for unattended operations. too financially risky and complex to operate. In this Di- rected Research, we propose to demonstrate feasibility SCRs are significantly different in scale and design from of self-regulation and develop and validate constitutive conventional reactors. Making use of precisely engi- models of neutron kinetic behavior in tunable ‘micro- neered core and reflectors, SCRs can be designed to engineered’ materials specifically designed to achieve have relatively longer neutron life time, small excess re- self-regulation. activity in the core and low power density making them an ideal candidate for achieving self- regulation. The Los Alamos National Laboratory is singularly qualified to transformational goal of this DR is to precisely under- undertake this research because of our unique ability to stand how novel SCRs can be designed to self-regulate bring together computational nuclear engineering, com- with minimal control system or human intervention. In putational and experimental nuclear materials science, this research we are seeking a disruptive alternate based and criticality testing facilities at NCERC. Of particular on the hypothesis that certain mesoscale properties of importance are our capabilities (1) to precision design, micro-engineered fuels can be ‘tuned’ so that emergent engineer and fabricate nuclear components, (2) to build neutronic functionality maximizes SCR self-regulation. in proliferation resistance and (3) to test the system and Building on this hypothesis, we have designed micro- its components relying upon methods developed as part engineered dispersed fuel composites that trigger and of science-based stockpile stewardship. In addition to amplify negative reactivity feedback at different times- providing fundamental knowledge on nuclear materi- cales and lengthen neutron lifetime by ‘geometrically als, it also provides nuclear criticality data necessary for delaying’ a significant portion of fissioning neutrons. characterizing safety and non-proliferation implications of dispersed fuels and are likely to lead to fabrication The focus of this research is to demonstrate the feasibili- of accident tolerant nuclear fuels. In summary we are ty of our SCR concept by accomplishing three inter-relat- leveraging specific capabilities being maintained at LANL ed R&D objectives: to research and develop techniques for the purpose of carry out our primary nuclear security to make limited quantities of tunable fuel composites mission, to meet needs of other national security spon- that maximize self-regulation; to measure their neutron- sors, and refining those actinide science skills for future ic and thermo-mechanical properties at different length- weapons and nonproliferation missions. scales; and using this data to model the performance 562

Page 563

Progress simultaneously evaluate neutronic, thermal and structural The project has made progress in all key areas necessary to performance of real-scale reactor (a national first). This achieve the goal of improving the ability of a small reactor will enable us to quantify how a microscopic change, such to self regulate using the physics of neutron reactions, in as coating on a fuel particle, could impact the macroscopic specially designed nuclear materials. Key areas include: 1) self-regulation performance of a reactor. NEN and AET the ability to make a dispersed reactor fuel with proper- post-doctoral students are pushing this capability forward. ties that enhance the ability to provide negative reactor We are already benchmarking these codes using past criti- feedback; 2) Modeling the neutron and thermal physics of cality experiments conducted by NEN division. the reactor at a level that will allow for the physical effects Another milestone relates to developing and refining to be predicted; and 3) determining the experiments that experimental plans for neutronic experiments to be per- can be performed that will allow direct measurement of formed over the next two years; it was recognized that self-regulation. developing such a plan early on in the project and get- A dispersed nuclear fuel is one in which uranium particle ting nuclear authorization will reduce project risk. Due to (kernels) are dispersed and compacted in robust materi- reduced funding received for the project and due to the als, such as graphite and tungsten. The purpose of our fact that some of the nuclear specialists were pre-occupied research is to specially coat these fuel particles to enhance with DAF restart, progress on this subtask is slightly behind its neutronic performance. It has been shown theoreti- the schedule, but will be completed as planned by the cally that self-regulation can be significantly enhanced if end of the fiscal year. We have already developed a list of the coatings are able to temporarily “trap” heat inside the experiments that will allow for the neutronic performance fuel particle. This rapid heat up of the particle in relation of the fuel described above to be verified and finalizing to its surroundings produces the desired negative reac- strategy for instrumenting them. The experiments will tivity feedback necessary for self-regulation. Therefore use COMET machine at DAF in the epithermal neutron flux thickness, conductivity and quality of the coatings is very configuration named “Zeus”. First series of experiments, important to this project. A key milestone for the first year named “Helios”, will be a variation of the “Zeus” experi- is to design, make and characterize a limited amount of a ment to accurately measure neutron flux and fuel temper- dispersed fuel to demonstrate that such coatings can be atures induced by traditional fuel materials. The materials fabricated and that they are robust enough to be extruded will be changed in a step-wise fashion to eventually arrive into a fuel compact. MST Division has developed the nec- at a configuration that will use the fuel being developed essary hardware to coat fuel particles. Team has overcome by MST. By performing the experiments in a step-wise numerous logistical and technical challenges to produce fashion, the uncertainty of each experiment can be better several batches of coated fuel-surrogates (Zr and Tung- estimated which is essential for validating the aforemen- sten were instead of Uranium). The coatings produced to tioned simulations. date have been very satisfactory in that they are uniform, The project to date is on track and meeting all proposed robust and are of the required thickness. Coated particles milestones. The early results have demonstrated the capa- have already been dispersed into graphite and extruded bilities necessary to make this project successful. as fuel blocks to demonstrate that coatings are robust enough. A multitude of tests are being run at LANL and INL Future Work to exquisitely characterize their performance. We expect The project team will design, synthesize and characterize to repeat these activities using natural and low enriched a limited quantity of fuel composite. Measurements of its uranium over the next two months. A contract has been materials and neutronic properties at various time and signed with General Atomics to make UO2 and UC fuel length-scales will provide data necessary to refine and vali- particles from the feedstock obtained courtesy of the Y-12 date models use in the design of small compact reactors. plant. This positions us well to fabricate larger quantity of fuel next fiscal year which will ultimately be used in the As a first step, novel fuel materials will be designed, devel- neutronic experiments. oped, fabricated and characterized. Candidate fuels will be manufactured at a specific micro-scale and then coated in Simultaneously modeling tools are being developed or a low thermal conductivity outer layer. The coated particles have been already developed to examine the reactivity will then be added to a material material like graphite. The feedback mechanisms for self-regulation. We have linked choice of fuel size, coating thickness, materials thermal important LANL computer codes, e.g. MCNP, with the conductivity will produce a thermal behavior that in turn heat transfer and solid modeling codes that specialize in will change the neutronic behavior of the material. reactor scale modeling. We now have the capability to 563

Page 564

The second step is to perform nuclear reactor physics experiments on the fuel samples. In reactor physics, an integral property ‘reactivity worth’ is often used to rep- resent aggregated effect of various nuclear reactions. Los Alamos maintains nuclear criticality benchmark machines for the explicit purpose of measuring reactivity worth of new materials. Samples will be placed in the benchmark machines to provide measurements of reactivity worth at various temperatures. Finally, we will build refined models that will be used to perform an integrated assessment of self-regulation. We plan to model the experiments and parts of the core in ex- plicit detail to capture the effects of the micro engineered materials. Conclusion We are seeking a disruptive solution based on the prem- ise that dispersed fuel composites can be ‘tuned’ at the micro-scale such that emergent neutron behavior can be designed to self-regulate a reactor with minimal control system or human intervention. Operationally, the De- fense Science Board identified small compact reactors to be “game-changers” whose demand cannot be underesti- mated for space exploration, for underwater vehicles, for assured arctic awareness and satellites that can operate on the dark side of the globe. This research will demonstrate feasibility and readiness of such a technology for near- term use by a community of sponsors. Publications Cummins, D. R., E. Luther, A. Telles, and P. Pappin. Low Temperature Sintering of Silicon Carbide for New Self Regulating Reactor Design. Presented at Rio Grande Fuels Conference. (Los Alamos/ Albuquerque, September 2015). Luther, E., D. V. Rao, I. Usov, A. Telles, A. Nelson, and D. Hurley. Fabrication of Graphite Composite Fuel with Controlled Thermal Transport Properties . To appear in TMS 2016: the Materials and Fuels for the Current and Advanced Nuclear Reactors V symposium . (Nashville, TN, Feb 14-18). Rao, D., A. Fallgren, and P. McClure. The Use of Moderation in Compact Fast Reactors to Improve Safety Performance and Reduce Proliferation Risks . To appear in Nuclear Technology. Telles, A., E. Luther, P. Papin, and D. V. Rao. The Effect of Compositional Variations on the Fabrication of Extruded Graphite Composite Fuel. To appear in Rio Grande Fuels Conference. (Los Alamos/ Albuquerque, September 2015). 564

Page 565

Nuclear and Particle Futures Directed Research Continuing Project Next-Generation Double Beta Decay Experiment Steven R. Elliott 20150088DR Introduction (Global Security) relevant missions. The data acquisition If neutrinos are their own antiparticles, technically and multiplexing will provide an excellent basis for bet- referred to as Majorana particles, theories that explain ter throughput in data collection and storage. The pulse the matter-antimatter asymmetry of the Universe and shape discrimination (PSD) analysis work also may serve hence our presence become plausible. Neutrinoless to improve detection sensitivities through recognition double beta decay experiments are the only practical of background versus sample events. The PSD may also technique to establish that neutrinos are Majorana par- allow us to do better time coincidence measurements by ticles. A next generation experiment using approximately looking for only coincidences with certain pulse shapes. a tonne of source will have an exciting opportunity to Low background gamma-ray detection has application make this determination. This is especially true given the to non-proliferation by lowering the limit of detection of knowledge we have regarding neutrino mass from other radioactive effluent from reprocessing and other weap- experiments. It is clear that this is an ideal time to begin ons production activity. This supports the Laboratories planning such an experiment. Science of Signatures Pillar. Some remaining R&D is required to eliminate risks asso- Progress ciated with a next-generation experiment. This includes a number of technical and theoretical topics. This project This past year we made progress on all of the activities intends to perform that R&D. associated with our project. Our efforts are R&D aimed at improving the performance of a tonne-scale double The Department of Energy is planning for a future large beta decay experiment based on Ge detector technology double beta decay experiment, but they haven’t chosen (TSGe). Some of our efforts are dedicated experiments the technology yet. With this R&D, we hope to improve at LANL, whereas other efforts exploit data from the Ma- the case for the use of Ge detectors. jorana Demonstrator (MJD) presently operating at the Sanford Underground Research Facility (SURF) and criti- Benefit to National Security Missions cal theory for understanding the expected decay rate. This is a critical time for this R&D. The DOE-NP office is planning to fund a large-scale double beta decay experi- A stainless steel test cryostat has been fabricated and a ment and is making plans for that eventuality. Presently, gas handling system has been constructed. This system NSAC is doing an assessment of the possible techniques will be used to evaluate several aspects of the hybrid that can accomplish the goal and a sub-committee has cryostat design. This system will evaluate the cooling posted its report. In April of this year, the committee efficiency of a gas filled cryostat, determine the proper submitted its initial report to NSAC indicating the impor- operating pressure and purity for the gas, and a Ge de- tance of the science and the critical nature of the next tector will be cooled and operated in the cryostat. Other 2-3 years in determining the final choice of technology. development work is ongoing for a second radio-pure Therefore, the R&D accomplished in the next few years test cryostat made of copper. will greatly influence the final decision. With the continued reliability issues for high voltage and This research will also provide a significant improvement signal connectors being uncovered by the operation of in HPGe detection limits, and the knowledge gained the MJD, we have begun pursuing more reliable connec- can be leveraged to lower backgrounds for operational tors and working with several vendors on new radio- 565

Page 566

pure options. paper has been published on this topic. We are presently studying operators relevant to beta decay and double We have repaired our robot and are beginning studies of beta decay. Some initial application of this work has been how to use robotics to reduce labor requirements for a published this year. During the next year we plan to deter- large-scale double beta decay experiment. mine if these models adequately reproduce the decay-rate quenching found in essentially all nuclear beta decays, and A white-paper design was completed to form the basis of a start initial applications to double beta decay. design specification document for two signal-management concepts: radio-frequency domain multiplexing (FDM) In BSM physics, we have investigated alternative new phys- and time-domain multiplexing (TDM). Both concepts are ics that could allow neutrinoless double beta decay. Some intended to reduce the required number of cables, and of these models are subject to tests in high-energy phys- hence background, for TSGe. We are now working to ics experiments at the LHC, where new particles could be down-select between these technologies. produced directly. A depth requirement concern is that of cosmic-ray mu- Future Work ons. Double beta decay experiments are conducted deep In the next fiscal year, we plan to make progress on several underground to avoid these muons and any associated separate experimental R&D tasks. We plan to begin use particles they may produce, such as neutrons, that may of a new cryostat concept. We plan to make down-select produce background. Hence, an important question for choices within the various possible multiplexing schemes TSGe is how deep underground should the experiment be and then test a configuration. We plan to begin using a located to avoid this background. This is a key question robot to perform some of the repetitive tasks in detec- for siting TSGe. We have begun preparing a simulation of tor assessment and plan for future uses. We plan to study muon propagation through the Earth and the subsequent improvements in low-radioactivity connectors. We plan to particle production. This simulation framework combines begin analysis of data from a present underground experi- particle reaction information and the experiment and the ment to assess the depth requirement for such experi- Earth-surface geometry. It is then possible to simulate ments. the interaction of the particles with the experiment and the surrounding rock material. The current status of this On the theory front, we will continue our studies on the in- analysis is that the experimental geometry has been imple- fluence of the axial vector coupling constants on the matrix mented and the muon distributions at various depths have element for double beta decay. We will also continue the been simulated. These first results are promising, agreeing study of other potential influences from particle physics on with other simulations and are being compared to data the double beta decay rate. from the MJD. Conclusion Environmental neutrons can interact with materials and Our expected results include assessments of new cryo- generate radioactive isotopes, which can then decay and stat designs for Ge detectors, use of robotics in assessing produce background for TSGe. Many of these neutrons detectors, new data acquisition techniques, understand- are produced by cosmic rays and hence the background is ing the depth requirement for such experiments and key depth dependent. Some of these isotopes have a delayed- theoretical issues in nuclear and particles physics required coincidence signature, allowing us to study and estimate to fully understand a measurement of double beta decay. the flux of neutrons in the deep underground environ- ment. We have identified a number of such candidate isotopes and have begun searching for them within the MJD data. Placing constraints on the rate of such isotopes within MJD will permit us to define a depth requirement for TSGe. The theory effort consists of two components, the cal- culations of nuclear physics matrix elements relevant to beta decay and double beta decay and the investigation of alternative Beyond-the-Standard-Model (BSM) mecha- nisms for neutrinoless double beta decay. For the matrix elements we have developed new algorithms to solve for the ground states in medium-mass and heavy nuclei. A 566

Page 567

Nuclear and Particle Futures Directed Research Continuing Project Cold Cathodes for Next Generation Electron Accelerators: Methodologies for Radically Improving Performance and Robustness Nathan A. Moody 20150394DR Introduction Benefit to National Security Missions Many grand challenges, including the quest for sustain- DOE/DOD/SC/MaRIE able energy, continued scaling of computational power, This project represents an enabling technology for exist- detection and mitigation of pathogens, and study of ing and future light sources and particle accelerators, the structure and dynamics of the building blocks of life including MaRIE, which is intended as a tool to study require the ability to access, observe, and control mat- weapons-relevant materials among others. Light sources ter on the frontier timescale of electronic motion and and particle accelerators are key to the DOE/SC mission the spatial scale of atomic bonds. The only instruments requiring state of the art electron beams for generation with such capability are future coherent x-ray sources of coherent x-ray light as well as all electron-beam-based and advanced colliders, which demand increasingly high DOD missions, such a high power free electron lasers performance electron beams well beyond the present and radio frequency source technologies. Improved cold state-of-the-art. The purpose of this project is to address cathodes with controllable parameters allows for higher this technology gap and make parallel, transformational performance, reduced complexity, reduced cost, and advances in the two critical performance areas of an reduced system maintenance in almost every relevant electron source: lifetime and efficiency. Traditional ap- application area. It is directly enabling for all accelerator- proaches have failed because advances in one of these based approaches to remote detection of chemicals, parameters have come at the expense of the other. pathogens, and nuclear materials . It would benefit accelerator-based solutions for management of nuclear Transformational advance in electron source (cold cath- waste as well as the production of critical medical radio- ode) performance requires that radical improvements in isotopes. lifetime and efficiency be achieved simultaneously. Previ- ous work has succeeded in delivering increases in one, The monolayer coatings we employ can potentially be but at considerable expense of the other because of functionalized to act as detectors, of interest to DHS/ the competing physical processes underlying traditional DHHS, or as a key element in energetic neutral atom approaches to cathode design and optimization. The imaging for remote space sensing (NASA). unique approach of this project is to decouple the com- Basic Understanding of Materials peting mechanisms so that both lifetime and efficiency X-ray free electron lasers (X-FELs) are the ideal tool to can be independently and substantially improved. The interrogate, understand, and even control matter in generalized hypothesis is that lifetime and efficiency of extremes. To date, nineteen Nobel Prizes have been electron sources can be controlled by manipulating na- awarded for x-ray science using beam-based x-ray light no-scale structure of cathode materials, yielding emer- sources, and we can expect more in the future as these gent behavior which radically enhances performance. tools open vast science frontiers to probe matter-in- Our innovative approaches span three thrust areas and extremes at unprecedented temporal, spatial, and share a common theoretical framework: a.) Enhanced energetic scales. The work represented in this project performance through nanostructure; b.) Enhanced closes critical technology gaps in the fielding of X-FELs as lifetime through monolayer protective coatings, and c.) instruments of discovery science. Integration of new physics in models, simulations, and technology demonstrations. Our progress to date has Progress validated our research approach, which includes key risk mitigation and supporting engineering efforts. This project is organized around 3 thrust areas, each 567

Page 568

described below, and work has progressed on schedule Thrust Area #3: “Integration of new physics in models, relative to the goals and major milestones associated with simulations, and technology demonstrations.” each thrust area. A strong theoretical framework enables understanding of Thrust Area #1: “Controlling and enhancing cold cathode photoexcitation and photoemission from nanostructured performance through nanostructure.” materials. Firstly, our density functional models generated exciting theoretical predictions of reduced work functions The unique control over electronic structure and optical for specific copper crystal faces coated with monolayer(s) dynamics in thin film nanomaterials suggests new methods of graphene. These unexpected predictions may have sig- for improving photoemission performance. We estab- nificant impact on accelerators utilizing copper photocath- lished a reference for comparing our methods and allow- odes. Secondly, we developed a preliminary photoemission ing systematic investigation of film composition, guided simulation tool that assumes single photon excitation and by our understanding of how composition relates to work predicts quantum efficiency (QE) for metal nanostructures function (or Fermi level), absorption cross section, carrier including their power dependence and Schottky barriers. density, and other properties. We commissioned a dedi- We used the tool to investigate the influence of plasmonic cated electron gun chamber to conduct these studies and subwavelength gratings on QE in the low power limit. We started investigating semiconductor nanocrystal quantum found substantially improved QE, attributable primarily to dot (QD) films by varying thickness (from 90 – 300nm) and increased photon absorption, but also impacted by surface using bare and gold-coated glass substrates. The QD diam- structure. We began investigating plasmonic field enhance- eter was ~7 nanometers (nm) and the films were made via ment in various laser and electric field regimes and extend sequential spin coating, followed by treatments rendering these models to a variety of semiconductor photocath- higher conductivity. Initial results using ~50 femtosecond odes. pulsed laser excitation of 266 nm show a photoemission response with a power dependence suggesting that film Regarding our technology demonstration, we have com- thickness and charge mobility are important parameters. missioned three separate experimental stations, with We began studying the role of ligand-removal, size, and targeted experiments in progress on QDs, graphene, gra- composition variations of QDs, especially as they impact phene oxide, reduced graphene oxide, and indium-gallium- changes in mobility and Fermi levels. nitride photocathode films. The project has identified key performance parameters specific to the film-on-barrier Thrust Area #2: “Enhancing robustness and lifetime using cathodes (e.g. barrier smoothness and porosity) and dem- monolayer barrier-shield coatings.” onstrated the ability to simultaneously probe both sides of film-barrier cathode structures. We demonstrated repro- We established synthesis conditions for obtaining both ducible growth of alkali antimonide photocathodes (QE defect-free graphene films (0.5 square cm) using single ~2%), providing an ideal system to explore surface layer layer chemical vapor deposition (CVD) growth on Cu protection by nanometer-thick graphene shields (since substrates and reduced graphene oxide (RGO) transferred these films respond to controlled gas contamination). We and suspended on wire mesh. CVD graphene was grown were able to quantitatively characterize the degradation directly on Cu with different crystal orientations: 100, 110 process of an unprotected photocathode with data agree- and 111. Measurements are underway to understand the ing with an emerging surface degradation model, which influence of monoatomic graphene layers on the Cu photo- predicts an exponential decay at a constant pressure. Our emission processes and comparing to theoretical predic- rapid design, fabrication and operational testing of the tion. We demonstrated key results for multi-layer graphene low-voltage DC cathode gun, and its accompanying femto- growth (2-40 atomic layers) by utilizing nickel substrates. second tunable laser source, has led to world-first mea- The resulting films were transferred to Cu and various wire surements of electron beams launched from a quantum meshes and the photoemission properties were corre- dot photocathode. lated to thickness of the barrier film. Based on our initial evaluation, both routes (graphene and RGO) yield viable Future Work barrier membranes and need to be further optimized. We Theoretical Framework: Establishing a validated under- have focused on mesostructure control of graphene oxide standing of photoexcitation and photoemission from nano- precursors using different solvents, successfully identifying structured materials using first principles and experimental those which yield defect free films. Thin-films have begun comparisons. gas barrier testing, and we have successfully incorporated monovalent ions (e.g. Na) into the graphitic sheets that impact their structural and photo-emission properties. 568

Page 569

Goal 1: Understand applicability of existing models address any compatibility issues Task: Comparison between families of density functionals Goal 2: Prepare existing rudimentary test hardware to scale to project needs Goal 2: Down-select from among existing models and adapt to include nanostructure Task: Update/upgrade vacuum system to provide immedi- ate photoemission data from first FY15 samples to acceler- Task: Determine and incrementally implement methods ate model validation and overall interface between theory to account for effect of quantum dots, surface plasmonic and experiment structures, and monolayer protective coatings Conclusion Goal 3: Identify highest priority experimental parameters We will develop and demonstrate ‘designer’ cold cath- and measurements ode electron sources with tunable parameters (bandgap, efficiency, optical absorption) that outperform present Task: Direct validation of theoretical models using test-case technologies in terms of efficiency and lifetime, where scenario success in either of these is considered transformational. Thrust Area #1: Controlling and enhancing cold cathode We introduce fundamentally new approaches to address performance through nanostructure decadal weaknesses in performance and enable cathode properties to be tuned or engineered for specific Depart- Goal 1: Validation of methods for controlling surface cath- ment of Energy missions and related applications. This ode properties project will specifically yield new understanding of nano- scale structure on cathode performance, new fabrication Task: Fabrication and measurement of straightforward, methodologies, validation of simulation and modeling first-generation plasmonic structures on photocathodes, tools, and prototype demonstrations of novel cold cath- allowing immediate comparison to theory. odes incorporating these new features. Goal 2: Validation of methods for controlling bulk cathode Publications properties Alexander, A., N. A. Moody, and P. Bandaru. Enhanced Task: Fabrication and measurement of straightforward, quantum efficiency of photoelectron emission, through surface textured metal electrodes. Journal of Vacuum first-generation quantum dot (QD) photocathode samples, Science and Technology A (JVSTA-A-15-292). allowing immediate comparison to theory. Carlsten, B. E, S. J. Russell, J. W. Lewellen, D. C. Nguyen, P. Thrust Area #2: Enhancing robustness and lifetime using M. Anisimov, C. E. Buechler, K. A. Bishofberger, L. D. monolayer barrier-shield coatings Duffy, F. L. Krawczyk, Q. R. Marksteiner, N. A. Moody, N. Yampolsky, and R. L. Sheffield. MaRIE XFEL physics Goal 1: Comparison between three primary methods for design risks and risk mitigation plans. 2015. Los growing graphene Alamos National Lab. (LANL), Los Alamos, NM (United States); DOE Contract Number: AC52-06NA25396; LA- Task: Determine the advantages of each and degree of UR—15-22069. control over relevant parameters Jensen, K. L., N. A. Moody, A. Shabaev, S. G. Lambrakos, Goal 2: Demonstrate transfer of large-area films D. Finkenstadt, A. Mohite, G. Gupta, and F. Liu. Photoemission from graphene on copper: theory and Task: Adapt existing glove-box transfer methods to vacuum experiment. To appear in Journal of Applied Physics. environment and provide demonstration of prototype coatings Thrust Area #3: Integration of new physics in technology demonstrations Goal 1: Specify requirements, functionality, and design of multi-use cathode test environments Task: Complete of engineering designs for all cathode test hardware, begin procurement/fabrication/assembly, and 569

Page 570

Nuclear and Particle Futures Directed Research Continuing Project Nuclear Science for Signatures, Energy, Security, Environment Albert Migliori 20150646DR Introduction Benefit to National Security Missions This project, “Nuclear Science Fellowships for Signatures, Actinide and nuclear science continues to be essential Energy, Security, Environment” will support research in to the U.S. and central to the missions of the DOE and nuclear science relevant to Laboratory mission areas by its NNSA laboratories, including nuclear weapons, global attracting and funding projects of a future generation of security, energy security, nuclear safeguards, nonpro- scientists and engineers with a high expectation, based liferation, environmental restoration, and radioactive on results from previous actinide research projects, of waste management. With nuclear weapons technology exceptional science and of retaining them as permanent continuing to play an enduring role in defense policy for employees. Nuclear science areas include materials, the foreseeable future, knowledge and expertise in the material properties, signatures, modeling, predictions, production, processing, purification, characterization, fabrication, detection, disposal, global security implica- analysis, and disposal of actinide series elements specifi- tions, forensics, and the specialized science surrounding cally and nuclear materials in general is essential to U.S. actinides and especially plutonium, uranium, and their national security. Nuclear science and detection technol- surrogates as fuels for energy and nuclear weapons. The ogy development is extremely important to nonprolif- aim is to advance nuclear science in a comprehensive eration and global security. Moreover, the risks of global project that ties targeted research with Los Alamos mis- warming, and the environmentally destructive effects of sion imperatives. The mechanisms for this are support of burning coal are such that nuclear energy is expected to the research arising from a multi-year broad-based post- assume a greater role in the nation’s electrical energy doctoral fellows project, and a summer student fellows production in the future. research project. The project will be administered by the Of the actinide elements, plutonium, uranium and Los Alamos G. T. Seaborg Institute, which is a center for neptunium are especially important to Los Alamos mis- actinide and transactinide science. sions. These elements are of technological and scientific An internal cross-institutional LANL advisory committee interest largely due to radioactivity and it is this property meets to select Nuclear Science Fellows. Fellow selection which makes their study particularly challenging. Special is based on the excellence of the science in the proposed facilities, instrumentation, and training, existing in only projects, the quality of the candidate, and the strategic a small number of locations worldwide, are required value of the proposed science. The quality of a candidate for safe and secure handling of these elements, distin- is measured by the candidate’s transcripts, publication guishing actinide science from most other research. and presentation record, educational background and Fundamental actinide and nuclear science provides the relevance, and letters of reference. Successful proposals technical basis for process and separations chemistry, have a clearly defined research proposal that supports metallurgy, characterization, and detection related to nuclear science at the single investigator or small-team the national security mission of the Laboratory and the level, described in a one-page abstract, and is written by national Integrated Plutonium Science and Research the student and mentor. These research plans address Strategy. nuclear science topics connected to the objectives of this project and span the strategic breadth of nuclear Progress science at LANL. Postdoctoral fellows work on a project The scope of this project has been broadened in FY15 that connects directly to nuclear science and that is not from actinide science and research to include nuclear otherwise funded. science, research and technology. 570

Page 571

There were several notable R&D accomplishments of post- energy scale associated with the presence of hybrid- doctoral fellows projects. ization between the 5f electron of uranium and the conduction electrons. In the Ce1-xUxMAl4Si2 (M=Rh, • Completed major-element and trace-element analyses Ir), we substituted 5% U for Ce to characterize energy of historic Trinity test samples. Developed method for scale associated with removing Ce ions from the lat- improved measurement of U isotopes in fallout debris tice, or “Kondo Holes”. Efforts are underway to per- using the Los Alamos Cameca ins-1280 secondary ion form similar measurements on U2-xThxRhIn8 to see mass spectrometer and measured U and Pu isotopes how the magnetism changes with Th substitution into in Trinity test samples. Submitted an article for pub- U2RhIn8. lication on the main formation mechanism of Trinity fallout debris. • Development of a containment vessel for the study of radioactive materials under pressure in a safe man- • Developed new protocol and containment in order to ner has been developed and constructed and safety analyze actinium at Stanford Synchrotron Radiation approval from the Health Physics group at the Argonne Lightsource (SSRL) for X-ray Absorption Spectroscopy National Laboratory Advanced Photon Source has been (XAS) experiment. For the first time, actinium was completed. Demonstration that this system works on a measured using XAS technique and the first XANES and non-hazardous material is complete and analysis of the XAFS were collected for actinium acetate. In conjunc- data is proceeding. tion with DFT and molecular dynamics calculations (performed by Justin Wilson and Enrique Batista), • Designed and arranged the working environment for EXAFS data analysis is underway. A manuscript is in calorimetric experiment; Performed power-up proce- preparation to communicate these new results. dures on the core equipment - calorimeter; set up and calibrated the calorimeter to its best status; worked • Commissioned a new beamline for parasitic data-tak- with mentors on revising IWD for the calorimeter; col- ing without disturbing other experimental operations. laborated on designing necessary parts for accomplish- We have explored methods for producing and charac- ing Pu calorimetric experiments; Contacted potential terizing actinide samples for optimal use in ultracold collaborators in Advanced Photon Source for future neutron experiments. We have built a new sample material characterization (X-ray adsorption and diffrac- handling system which reduced contamination. We tion). have placed first limits on the fission cross section at ultracold neutron energies. • Developed a technique to exchange oxygen isotopes with uranium oxide ions by irradiating material dis- • Pure iron and ferritic Fe-9Cr alloys with specific solved in desired concentrations of isotope water. amount of carbon were received and metallurgically Multiple steps were used to ensure full exchange, each prepared. Microstructures were characterized by elec- step replenishing the isotope water with the desired tron microscopy and electron backscattering diffrac- isotope concentration. NMR was used to track ex- tion techniques. Actinide-based nuclear fuel materials change of 17O while mass spectroscopy tracks 18O. fabrication initiated. Upon completion, the water is removed and the resul- tant powder is heated under nitrogen to produce UO2 • Preparation and measurement of 3 dimensional stan- powders for analysis. dards for evaluation of analytical capabilities, i.e., limit of detection for the XRF techniques. Publication of Future Work manuscript on determination of Pu in spent fuel with Future R&D tasks for this project include: high resolution X-ray (McIntosh, et al., Spectrochimica Acta Part B, 2015, doi:10.1016/j.sab.2015.05.014). • Develop a portable, plasma-based, system for separat- Publication of manuscript on characterization of Pu in ing and detecting of actinides. soil samples with multiple XRF techniques (McIntosh, et al., Journal of Analytical Atomic Spectrometry, 2015, • Chemical and isotopic measurement to understand doi: 10.1039/C5JA00068H). nuclear fallout formation for nuclear forensics tools • Synthesized single crystals of the following systems: • Understand trivalent actinide chemistry from electron- Ce1-xUxMAl4Si2 (M=Rh, Ir), La1-xUxMAl4Si2 (M=Rh, ic structure and bonding. Ir), and U2RhIn8. In La1-xUxMAl4Si2 (M=Rh, Ir) materi- als, substituting 1, 3, and 5% U for La to characterize • Utilize extremely low energy “ultracold” neutrons to induce fission in actinides near the material surface 571

Page 572

and characterize sputtered materials. Conclusion This research aims to accomplish the following: • Explore the ternary U-Si-X, (X=Al and Ti) systems to as- sess oxidation resistance for accident tolerant fuels in • Improve our understanding of the electronic structure, nuclear reactor concepts. phase stability, thermodynamics and thermal proper- ties of nuclear materials and the dynamic behavior of • Investigate the effects of carbon on defects involved plutonium, uranium, and some of their compounds in hardening and void swelling in ferritic Fe-Cr alloys across pressure, temperature, time, phase space, sur- under irradiation. faces and interfaces for nuclear materials. • Enhance detection and characterization of actinides • Develop advanced chemical separations and synthesis in complex matrices using multiple, complementary, processes and determine their signatures. nondestructive analytical methods based on X-ray fluo- • Expand capabilities in detection, measurement, and rescence spectrometry. analysis of signatures of nuclear and radiological ma- terials. Improve understanding of the environmental • Enhance understanding of correlated electron behav- behavior and signatures of nuclear materials. ior of U-based compounds, to elucidate their exotic • Enhance our understanding of plutonium aging, elec- ordered states, and to discover new U-based supercon- tronic structure and chemistry with minimal impact ductors. from nuclear decay processes. • Validate sole observation of magnetic resonance signa- Publications ture of the 239Pu nucleus. (McDonald), J. L. Brown, B. L. Davis, B. L. Scott, and A. J. Gaunt. Comparative structural studies of early • Demonstrate potential of magnetic resonance for lanthanide acetonitrile solvates. Inorganic Chemistry. nondestructive measurements of enrichment levels for nuclear forensics tools. (McDonald), J. L. Brown, M. B. Jones, A. J. Gaunt, B. L. Scott, C. E. Macbeth, and J. C. Gordon. Lanthanide(III) • Develop methods to understand how actinides form di- and tetra-nuclear complexes supported by a bonds with other elements. chelating tripodal tris(amidate) ligand. 2015. Inorganic Chemistry. 54 (8): 4064. • Observe elastic mechanical properties of actinide and lanthanide compounds at extremes of pressure and (McDonald), J. L. Brown, and A. J. Gaunt. Non-aqueous and organometallic chemistry of plutonium. Plutonium temperature. handbook: a guide to technology. Edited by Clark, D. L.. • Measure thermodynamic parameters for Pu-bearing Ardeljan, M., R. J. McCabe, I. J. Beyerlein, and M. Knezevic. oxides and alloys, U nitride/carbide, and other related Explicit incorporation of deformation twins into crystal materials to better understand structure-stability rela- plasticity finite element models. 2015. Computer tions and complex behavior of 5f electrons in transura- Methods in Applied Mechanics and Engineering. 295 nic compounds. (1 Oct): 396. • Develop characterization methods to measure O iso- Barker, B. J., J. Berg, and M. Wilkerson. Visible emission tope concentration in UO2, which will then be used to and low-lying LMCT states in Cs2NpO2Cl4. Journal of test the feasibility of O isotope forensics. Chemical Physics. • Fabricate U and Th nitride nuclear fuels via a novel Bonamici, C. E.. Formation of glassy nuclear fallout: An chemical process, in which the precursors contain a updated model. Geoghimica et Cosmochimica Acta. high percentage of nitrogen and undergo combustion Bonamici, C. E., W. S. Kinman, J. H. Fournelle, M. M. to make the nitride material. Zimmer, A. D. Pollington, and K. D. Rector. Formation of glassy nuclear fallout: An updated model. To appear in • Solution chemistry approaches to explore solid state Geochimica et Cosmochimica Acta. chemistry of nuclear fuels. Bonamici, C. E., and W. S. Kinman. Tracking volatility and • Observe the effect that Ga has on the oxidation state fractionation in the Trinity fireball through classgy of δ-Pu with surface electronic studies. fallout debris. Science. Bonamici, C. E., and W. S. Kinman. Volatility in the Trinity 572

Page 573

fireball as recorded in glass fallout debris. Actinide and J. L. Kiplinger. Expanding the chemistry of actinide Research Quarterly. metallocene bromides. Synthesis, properties and molecular structures of the tetravalent and trivalent Ellis, J. K., J. A. Leiding, W. M. Kerling, and A. P. Sattelberger. uranium bromide complexes: (C5Me4R)2UBr2, Novel metal-metal bonding in K[Tc8(μ-I)8I4]. Inorganic (C5Me4R)2U(O-2,6-iPr2C6H3)(Br), and [K(THF)] Chemistry. [(C5Me4R)2UBr2] (R = Me, Et). Inorganics. Guo, X., A. Navrotsky, R. K. Kukkadapu, M. Engelhard, A. Lichtscheidl, A. G., K. P. Browne, M. J. Monreal, D. E. Lanzirotti, M. Newville, and S. R. Sutton. Structure Morris, B. L. Scott, A. Nelson, and J. L. Kiplinger. and thermodynamics of uranium containing garnets. Synthesis, reactivity, spectroscopic and electrochemical Energy and Environmental Science. characterization of uranium(III) organometallic complexes. Organometallics. Guo, X., E. Tiferet, Q. Liang, A. Lanzirotti, M. Newville, M. Engelhard, R. K. Kukkadapu, D. Wu, S. R. Sutton, Lichtscheidl, A. G., M. T. Janicke, B. L. Scott, A. T. Nelson, M. Asta, and A. Navrotsky. Structure characteristics and J. L. Kiplinger. Syntheses, structures, and 1H, and thermodynamic stability of metal monouranates 13C\{1H\} and 119Sn\{1H\} NMR chemical shifts of MgUO4, CrUO4, and FeUO4: A combined experimental a family of trimethyltin alkoxide, amide, halide and theoretical study. Journal of the American and cyclopentadienyl compounds . 2015. Dalton Chemical Society. Transactions. 44: 16156. Guo, X., R. K. Kukkadapu, A. Lanzirotti, M. Newville, S. R. Loble, M., J. M. Keith, A. B. Altman, S. C. E. Stieber, E. R. Sutton, and A. Navrotsky. Charge-coupled substituted Batista, K. S. Boland, S. Conradson, D. L. Clark, J. L. garnets (Y3-xCa0.5xM0.5x)Fe5O12 (M = Ce, Th): Pacheco, S. A. Kozimor, R. L. Martin, S. G. Minasian, structure and stability as crystalline waste forms. 2015. A. C. Olson, B. L. Scott, D. K. Shuh, T. Tyliszczak, Inorganic Chemistry. 54 (8): 4156. M. P. Wilkerson, and R. A. Zehnder. Covalency in lanthanides. An x-ray absorption spectroscopy and Guo, X., S. Szenknect, A. Mesbah, S. Labs, N. Clavier, C. density functional theory study of LnCl6x-(x = 3,2). Poinssot, S. V. Ushakov, H. Curtius, D. Bosbach, R. 2015. Journal of the American Chemical Society. 137: C. Rodney, P. Burns, N. Dacheux, and A. Navrotsky. 2506. Thermodynamics of formation of coffinite, USiO4. 2015. Proceedings of the National Academy of Science, Macor, J. A., J. L. Brown, J. N. Cross, S. R. Daly, A. J. Gaunt, USA. 112 (21): 6551. G. S. Girolami, M. T. Janicke, S. A. Kozimor, M. P. Neu, A. C. Olson, S. D. Reilly, and B. L. Scott. Coordination Hartig, K. C., J. Colgan, D. P. Kilcrease, J. E. Barefield II, and chemistry of 2,2′-biphenylenedithiophosphinate and I. Jovanovic. Laser-induced breakdown spectroscopy diphenyldithiophosphinate with U, Np, and Pu. To using mid-infrared femtosecond pulse. 2015. Journal of appear in Dalton Transactions. Applied Physics. 118 (4): 043107. McIntosh, K. G., N. L. Cordes, B. M. Patterson, and G. Jacobsen, J. K., N. Velisavljevic, D. M. Dattelbaum, and R. S. J. Havrilla. Laboratory-based characterization of Chellappa. High pressure and temperature equation of plutonium in soil particles using micro-XRF and 3D state and spectroscopic study of CeO2. Physical Review confocal XRF. 2015. Journal of Analytical Atomic B. Spectrometry. 30: 1511. Jacobsen, M. K., and N. Velisavljevic. Development McIntosh, K. G., N. L. Cordes, B. M. Patterson, and G. of a containment setup for radioactive materials J. Havrilla. Laboratory-based characterization of in the Paris-Edinburgh press. Review of Scientific plutonium in soil particles using micro-XRF and 3D Instruments. confocal XRF. 2015. Journal of Analytical Atomic Spectroscopy. 30: 1511. Koutroulakis, G., T. Zhou, Y. Kamiya, J. D. Thompson, H. D. Zhou, C. D. Batista, and S. E. Brown. Quantum McIntosh, K. G., S. D. Reilly, and G. J. Havrilla. phase diagram of the S = 12 triangular-lattice Determination of plutonium in spent nuclear fuel using antiferromagnet Ba3CoSb2O9. 2015. Physical Review high resolution X-ray. 2015. Spectrochimica Acta Part B. 91: 024410. B: Atomic Spectroscopy. 110: 91. Lichtscheidl, A. G., B. L. Scott, A. T. Nelson, and J. L. McIntosh, K. G., S. D. Reilly, and G. J. Havrilla. Kiplinger. Convenient and green synthetic methods to Determination of plutonium in spent nuclear fuel using uranium(III) and uranium(IV) halide pyridyl starting high resolution X-ray. 2015. Spectrochimica Acta Part materials. Green Chemistry. B: Atomic Spectroscopy. 110: 91. Lichtscheidl, A. G., J. K. Pagano, B. L. Scott, A. T. Nelson, 573

Page 574

Radchenko, V., J. W. Engle, J. J. Wilson, J. R. Maasen, F. M. MoSi2 oxidation in 670-1498 K water vapor. To appear Nortier, W. A. Taylor, E. R. Birnbaum, L. A. Hudston, in Journal of the American Ceramic SocietyJ. K. D. John, and M. E. Fassbender. Application of ion exchange and extraction chromatography to the separation of actinium from protron-irradiated thorium metal for analytical purposes. 2015. Journal of Chromatography A. 2015 (1380): 55. Radchenko, V., J. W. Engle, J. J. Wilson, J. R. Maassen, F. M. Nortier, E. R. Birnbaum, K. D. John, and M. E. Fassbender. Production yields and chromatographic separation of protactinium isotopes formed in protron- irradiated thorium metal. Journal of Chromatography A. Rusev, G., A. R. Roman, J. K. Daum, R. K. Springs, E. M. Bond, M. Jandel, B. Baramsai, T. A. Bredeweg, A. Couture, A. Favalli, K. D. Ianakiev, M. L. Illiev, S. Mosby, J. L. Ullmann, and C. L. Walker. Fission-fragment detector for DANCE based on thin scintillating films. To appear in Nuclear Instruments & Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment. Rusev, G., M. Jandel, B. Baramsai, E. M. Bond, T. A. Bredeweg, A. Couture, J. K. Daum, A. Favalli, K. D. Ianakiev, M. L. Iliev, S. Mosby, A. R. Roman, R. K. Springs, J. L. Ullmann, and C. L. Walker. Development of a thin scintillation films fission-fragment detector and a novel neutron source. 2015. SPIE ProceedingsSPIE Proceedings. 9593. Sooby, E. S., A. T. Nelson, J. T. White, and P. M. McIntyre. Measurements of the liquidus surface and solidus transitions of the NaCl-UCl3 and NaCl-UCl3-CeCl3 phase diagrams. 2015. Journal of Nuclear Materials. 466 (November): 280. Tamasi, A. L., K. S. Boland, K. Czerwinski, J. K. Ellis, S. A. Kozimor, R. L. Martin, A. L. Pugmire, D. Reilly, B. L. Scott, A. D. Sutton, G. L. Wagner, J. R. Walensky, and M. P. Wilkerson. Oxidation and hydration of U3O8 materials following controlled exposure to temperature and humidity. 2015. Analytical Chemistry. 87 (8): 4210. Wilson, J. J., B. J. Barker, M. Ferrier, E. R. Batista, M. P. Wilkerson, J. M. Berg, S. A. Kozimor, K. D. John, J. W. Engle, and E. R. Birnbaum. Interactions of M3+ ions with Ds-DOTAM, a flouorescet macrocyclic ligand. Inorganica Chimica Acta. Wilson, J. J., E. R. Birnbaum, E. R. Batista, R. L. Martin, and K. D. John. Synthesis and characterization of nitrogen- rich macrocyclic ligands and an investigation of their coordination chemistry with lanthanum(III). 2015. Inorganic ChemistryInorganic Chemistry. 54 (1): 97. Wood, E. S., S. S. Parker, A. T. Nelson, and S. A. Maloy. 574

Page 575

Nuclear and Particle Futures Directed Research Final Report Peta-scale Studies of Cosmic Explosions and Supernova Shock Breakout with Palomar Transient Factory Przemyslaw R. Wozniak 20130030DR Abstract Background and Research Objectives Time-domain astronomy and the science of cosmic ex- The next generation Palomar Transient Factory (iPTF) plosions are currently at the forefront of modern astro- and its successor, the Zwicky Transient Facility (ZTF) are physics. The next generation Palomar Transient Factory poised to revolutionize the field of time-domain astro- (iPTF) and its successor project, the Zwicky Transient physics. PTF and ZTF have emerged as world leaders in Facility (ZTF), are poised to revolutionize the field, but the exploration of cosmic explosions such as supernovae they require transformational advances in peta-scale (SN) and the most successful precursors to Large Syn- data processing, as well as a comprehensive theory pro- optic Survey Telescope (LSST)---the first multi-Petabyte gram. The main science goals of this multi-disciplinary time-resolved sky survey scheduled to begin operations LDRD project aimed to address the most pressing issues in 2022. iPTF is hosted on the 48-inch Schmidt telescope in modern data intensive sky surveys. at Palomar (P48) and continuously surveys 8000 square degrees of sky on a 5-day cadence. In order to succeed, The most significant accomplishment of this project was these data intensive surveys must promptly identify the development and operation of the online transient actionable information from the torrent of imaging data, selection system based on state of the art machine classify emerging events, and obtain detailed follow-up learning approach that relies on random forest and observations using limited resources around the world sparse representation algorithms. Our research and and in space. They also require transformational ad- development program coupled observation, large-scale vances in big data processing as well as a comprehensive data processing, and computing technology with theory theory program. In August 2013, under the leadership of and simulation to uncover thousands of cosmic explo- our project team, LANL became a full institutional mem- sions. These discoveries and observations were used to ber of the iPTF Consortium of about half a dozen interna- explain the underlying physical mechanisms and contrib- tional partners led by California Institute of Technology. uted to more than three dozen refereed publications. The two most notable discoveries are: 1) observation of Our multi-disciplinary LDRD project was conceived to an early ultraviolet excess emission in supernova iPT- address the needs of modern data intensive sky surveys F14atg that supports single degenerate scenario (with and had three tightly interconnected major science only one white dwarf object) in a least some type Ia goals: 1) conduct the first deep photometric search for supernovae; and 2) discovery of ASAS-SN-15lh, the most supernova shock breakout events and other explosive luminous supernova recorded in history. phenomena on time-scales of less than a day; 2) perform extensive radiation hydrodynamics simulations and theo- We also carried out extensive radiation hydrodynamics retical modeling to understand the physics of transients calculations of transient emission from hot shocks over observed by time-domain surveys; 3) develop novel algo- a wide range of astrophysical scenarios to help interpret rithms and computing capabilities that support real-time observations, inform transient classifiers, and test phys- selection of astrophysical transients and optimization ics code performance in a critically important range of of follow-up observations based on machine learning physical conditions. The results of this project benefit algorithms. two major areas of the national security mission through development and testing of advanced technologies for The current-generation data intensive sky surveys Global Security and simulation codes for the Weapons searching for low probability explosive events produce Program. data streams in which false positive detections due to 575

Page 576

various artifacts and poorly controlled experimental factors for iPTF, one of the key documents compiled for any major outnumber real detections by several orders of magnitude. sky survey that codifies the most important credits. The To address this challenge, we set out to develop an auton- remaining paragraphs of this section address the most omous event broker that integrates cutting edge machine important elements of our scientific approach and the cor- learning algorithms with high performance computing responding results and accomplishments. infrastructure. Given new iPTF capabilities we placed a The most significant accomplishment of this project was major emphasis on studies of supernova shock breakout. the development and operation of the online transient After the initial collapse of the progenitor star, the shock selection system. We adopted a state of the art approach wave that eventually disrupts the entire star and produces based on direct image differencing (Figure 1). a core collapse supernova bounces off the dense core and propagates through outer layers of the star. As the shock breaks out of the stellar surface, it produces a bright flash prior to the main supernova event that can be detected as excess blue emission in visual band pass. Our plan was to observe up to half dozen such events in three years. We also anticipated discoveries of several thousand new tran- sients from the comprehensive search. We also planned to carry out extensive radiation hydrody- namics calculations of transient emission from hot shocks over a wide range of astrophysical scenarios to interpret observations from both iPTF and future experiments, Figure 1. Selection of transient variability of astrophysical origin inform transient classifiers, and test the performance of in optical images of the sky using direct pixel-by-pixel subtration. radiative transport simulation codes in a critically impor- After a sophisticated preprocessing to match the point spread tant range of physical conditions. Our goal was to build and function of the new science image (left), the reference image publish a suite of approximately 100 explosion models with (center) is subtracted to obtain the uncluttered difference time-resolved spectra and synthetic light curves at about image (right). A real transient (upper row) has the appearance 100 epochs. A detailed comparison of these models with consistent with the point spread function. However, candidate observed light curves provides valuable constraints on the detections from data intensive sky surveys are dominated by a structure of the supernova progenitor stars and lead to a variety of false positives including saturation trails (bottom row). better understanding stellar deaths and the physics of mat- The key innovation here was an automated classification ter under extreme conditions. engine for separating real transients from bogus detections based on the random forest algorithm in which a collection Scientific Approach and Accomplishments of randomized (and therefore de-correlated) decision trees Drawing on unique capabilities across LANL, we conducted is used in a voting scheme to label incoming candidate a comprehensive research and development program that objects. Using sparse representations with learned diction- coupled observation with theory, simulation, and comput- aries to provide additional image based features combined ing technology to discover cosmic explosions and explain with the decision forest framework, we improved the their mechanisms. Our project team contributed to a total figure of merit (missed detection rate at 1% false positives of 41 refereed publications (32 published and 9 submit- rate) by a factor of 2 compared to the system previously ted) with about equal balance between theory/simulation employed by PTF (Figure 2). This system also incorporates and observations/data mining. This includes 3 papers in online learning, i.e. it autonomously adapts as more data high impact journals Nature and Science (1 Nature paper becomes available. To our knowledge it is the first such published, 1 submitted to Nature and 1 to Science). We system deployed on a live data stream in astronomy. LANL also co-authored nearly 200 short communications (As- software has been deployed at NERSC supercomputing tronomers Telegrams a.k.a ATel). Senior investigators on center since July 2013. As of September 2015 our tran- the project delivered 16 presentations, about half of them sient selection software has been licensed to two differ- were invited. In November 2013, in Santa Fe, NM we host- ent collaborations: iPTF and All Sky Automated Survey for ed a hugely successful 3rd edition of the “Hotwiring Tran- Supernovae (ASAS-SN). We also developed automated sient Universe” workshop attended by over 80 participants classification algorithms for SN spectra with signal-to-noise representing leading forces in time-domain astrophysics. ratio (SNR), sampling, and wavelength coverage corre- In recognition of our contributions to iPTF, our machine sponding to the specification of SEDMachine, a new iPTF learning team was included in a prestigious “Builder’s List” instrument currently in commissioning. Tests on simulated 576

Page 577

data achieved unprecedented 95% classification accuracy with a single degenerate object (white dwarf), and against at SNR = 15 per pixel. The classifier can also be used to the competing double degenerate scenario [1]. The second identify very young supernovae. is the discovery of ASAS-SN-15lh, the most luminous super- nova recorded in history [2]. A distributed event broker architecture that supports automated inference of transient classifications has been developed and implemented (with the exception of the ac- tual follow-up algorithm). The top level view of the system is shown in Figure 3. The system includes a selection of classification algorithms: Bayesian belief networks, sup- port vector machines, random forest (decision trees), and sparse representations with learned dictionaries. The sys- tem can now be connected to data streams from Thinking Telescopes at LANL and high-energy transient localizations from Gamma-ray Burst Coordinates Network. Monte-Carlo simulations of real-time transient detections have demon- strated that the distributed event broker can accept input and process spatial and temporal cross correlations at a rate of 1 transient per second. This is more than sufficient for ZTF and we expect the system to be scalable to the Figure 2. Performance of the LANL transient selection system LSST regime (10 times higher event rate). in various configurations. The figure of merit (FOM) is defined as the rate of missed detections for a fixed false positive rate of 1%. This corresponds to the intersection between the receiver operating characteristic (ROC) curve for a given algorithm and a threshold at 0.01. The best performing LANL solution achieves FOM = 4.5%, more than a factor of 2 better than a competing algorithm that does not include machine learned features. These decision support tools have enabled discovery of more than 10,000 named iPTF transients, including 851 confirmed supernovae (others awaiting confirmation), mostly type-Ia, but also 215 type-II and 80 type-Ibc. Many exotic transients have been identified, including 62 active galactic nuclei, 20 super-luminous supernovae, 3 tidal dis- ruption flares, 15 cataclysmic variable outbursts, 8 GRB af- terglows, and 1 calcium-rich transient that likely represents Figure 3. Top level architecture of the real time event broker. a new class. It is anticipated that ~1300 confirmed super- The role of the event classification engine (left) is to correlate novae will be identified by the end of the project. Follow- new incoming observations with historical context, classify up observations with the instruments onboard the Swift events, and alert subscribers about interesting changes. The satellite were triggered for 35 transients, approximately role of the response optimization engine (right) is to coordinate 12 of which were detected in UV and resolved in time follow-up observations given limited resources. by the UVOT instrument. Six events have demonstrated some aspect of shock breakout and propagation with the The project performed new opacity calculations and in- corresponding analysis is presented in dedicated publica- corporated new physics into the simulation codes ([3] and tions. Nearly 200 additional supernovae have been found Figure 4). The process of running grids of radiation hydro- by ASAS-SN collaboration using our software. Two of the dynamics models and post-processing in the SPECTRUM above discoveries are having a particularly high scientific framework has been automated. A new public user inter- impact. The first is the observation of an early ultraviolet face (supernova.lanl.gov) allows plotting multiple models excess emission in supernova iPTF14atg that is best inter- and user data for a chosen scenario as well as tracking preted as the interaction of the material ejected by the ex- supernova emissions over time. Users can obtain publi- plosion with a companion star. This provides evidence that cation quality plots and metadata of both time resolved at least some type Ia supernovae occur in binary systems spectra and light curves. 60 models are currently available. 577

Page 578

Each model provides broad band spectra extending from time histories that does not require issuing unique object radio frequency to gamma ray energies sampled at approx- identifiers or frequent updates. This eliminates a very imately 200 epochs. An important milestone was the study constraining performance bottleneck and the need for of the effects of the thickness of the hydrogen envelope cumbersome book keeping. These results were presented on supernova shock breakout and observed light curves at the Databases in Networked Information Systems 2014 [4]. The timing and strength of the shock breakout flash workshop in Aizu, Japan. relative to the main peak of the supernova emission turns out to be very sensitive to the mass of the envelope (Fig- ure 5). As more and more mass is removed (plotted color moving from purple to red), the star becomes smaller, the breakout flash comes sooner, and the maximum brightness drops dramatically. Figure 5. Time dependence of optical emission at shock breakout during a simulated supernova explosion. The progenitor is a massive star with the mass 23 times that of the Sun. Color coded lines correspond to different masses of the hydrogen- rich envelope material that was lost prior to the explosion. Figure 4. Monochromatic opacity of samarium (Sm, Z=62) over LANL simulations such as these allow observers to interpret a broad range of photon energies. Our new local thermodynamic the relative timing and strength of the shock breakout peak in equilibrium (LTE) calculations show that line-rich spectra of relation to the main supernova light curve for transients detected heavier elements such as lanthanides increase the opacity in by astrophysical time-domain surveys. the visual and infra-red range by several orders of magnitude. The effect strongly impacts theoretical predictions of the We also developed a new hierarchical pixel-level indexing electromagnetic emission from neutron star mergers widely scheme for sky monitoring images called Spherical-cube anticipated to become the first directly detected gravitational Quad-tree Unique ID (SQUID). SQUID maps patches of wave (GW) sources. Early detection of optical counterparts to sky to integers in a way consistent with low level storage GW sources is currently a topic of intense interest both at LANL access patterns. Our reference implementation LibSQUID and across the astrophysics community. also supports very fast quad-cube map projections for resampling pixelated images to an efficient format and The work on data intensive computing led to the develop- was released in public domain on GitHub (libsquid.github. ment of a new distributed database architecture and work- io). Another technology product we expect to provide a ing prototype for time-resolved photometric measure- lasting value is the distributed data storage and manipula- ments from massive sky monitoring surveys. The data is tion system (for both large-volume data and simulation partitioned according to the Quad Tree Cube (Q3C) index- outputs) based on web access protocols. The system cur- ing scheme that allows rapid storage and retrieval of mea- rently supports online data access to Dark Sky Simulations surements covering a random patch of sky. Physical storage (55 Terabytes of data), the first publicly released trillion is then mapped to a set of PostgreSQL database servers, particle cosmological N-body simulation. each holding a fixed set of Q3C cells. The cells have fuzzy edges, i.e. measurements near their edges are replicated Impact on National Missions in neighboring cells, so that the data can be clustered and In the context of national security missions, the main queried locally, without crossing server boundaries. This focus of this project was to build new astro-informatics allows a linear scaling for the most time-consuming data competency to address growing needs in Space Situational manipulation tasks. We also invented a progressive hier- Awareness (SSA), develop new capabilities for the Weap- archical clustering algorithm for building long photometric 578

Page 579

ons Program, and test them using actual observations. Moody, U. D. Rebbapragada, and B. D. Bue. Ultraviolet In the process we also helped to facilitate hiring of three Radiation from Supernova-Companion Collision in a postdocs and five students, with one postdoc and one Type Ia Supernova . 2015. Nature. 521 (7552): 328. student converted to LANL staff members. 2. Dong, S., B. J. Shappee, J. L. Prieto, S. W. Jha, K. Z. Stanek, T. W. Holoien, C. S. Kochanek, T. A. Thompson, The image-based transient selection and spectral classifica- N. Morrell, I. B. Thompson, U. Basu, J. F. Beacom, D. tion tools developed by our team will become a perma- Bersier, J. Brimacombe, J. S. Brown, P. Chen, E. Conseil, nent addition to LANL SSA capabilities. To address the for- A. B. Danilet, E. Falco, D. Grupe, S. Kiyota, G. Masi, B. midable complexity of optimizing the custody of resident Nicholls, F. Olivares, G. Pignata, G. Pojmanski, G. V. space objects, we also developed a simulation toolbox for Simonian, D. M. Szczygiel, and P. R. Wozniak. ASASSN- evaluating the relative merits of various observing strate- 15lh: The Most Luminous Supernova Ever Discovered. gies in detecting anomalous behaviors of Earth satellites. 2015. ArXiv e-prints. The software maintains a set of simulated orbits and predicts viewing conditions for satellites given the position 3. Fontes, C. J., C. L. Fryer, A. L. Hungerford, P. Hakel, J. of the Sun and a network of observatories placed around Colgan, D. P. Kilcrease, and M. E. Sherrill. Relativistic the globe. We used this framework to simulate interest- opacities for astrophysical applications. 2015. High En- ing orbital anomalies and simple yet realistic collaborative ergy Density Physics. 16: 53. and non-collaborative follow-up response scenarios. Our 4. Bayless, A. J., W. Even, L. H. Frey, C. L. Fyer, P. W. A. preliminary results indicate that our proposed coordinated Roming, and P. A. Young. The Effects on Supernova observing approach (Dynamic Coalition Architecture) Shock Breakout and Swift Light Curves Due to the Mass provides a substantial advantage over the un-coordinated of the Hydrogen-Rich Envelope. 2015. The Astrophysi- approach with each telescope separately optimizing its cal Journal. 805 (2): 98. own schedule. In the future, our statistical SSA simulation toolbox can be used to study a wide range of scenarios and telescope networking strategies. Publications Barth, A. J., A. Voevodkin, D. J. Carson, and P. Wozniak. A Our detailed models of supernova shock breakout and Search for Optical Variability of Type 2 Quasars in SDSS other explosive transients were used to test Advanced Sim- Stripe 82. 2014. The Astronomical Journal. 147 (1): 12. ulation and Computing (ASC) capabilities in the critically important regime at the transition between diffusion and Bayless, A. J., T. A. Pritchard, P. W. A. Roming, P. Kuin, P. J. free streaming using both flux diffusion and implicit Monte Brown, M. T. Botticella, M. Dall’Ora, L. H. Frey, W. Even, Carlo codes. New state of the art plasma opacity calcula- C. L. Fryer, J. R. Maund, and M. Fraser. The Long-lived tions for heavier elements such as lanthanides (Figure 4) UV “Plateau” of SN 2012aw. 2013. The Astrophysical completed over the course of this project will permanently Journal. 764: L13. enhance the physics simulation tools at the disposal of the Bayless, A. J., W. Even, L. H. Frey, C. L. Fyer, P. W. A. Rom- Weapons Program at LANL. ing, and P. A. Young. The Effects on Supernova Shock Breakout and Swift Light Curves Due to the Mass of the Our novel SSA approach has been presented to numer- Hydrogen-Rich Envelope. 2015. The Astrophysical Jour- ous sponsors within the Department of Defense and other nal. 805 (2): 98. WFO organizations attracting significant attention each time. These program development efforts are already Bellm, E. C., N. M. Barriere, Bhalerao, S. E. Boggs, S. B. beginning to pay off with $250,000 of funding from a WFO Cenko, F. E. Christensen, W. W. Craig, Forster, C. L. sponsor in FY2015. This project also positioned us for a fu- Fryer, C. J. Hailey, F. A. Harrison, Horesh, Kouveliotou, ture role in the Zwicky Transient Facility (ZTF) and the Large K. K. Madsen, J. M. Miller, E. O. Ofek, D. A. Perley, V. Synoptic Survey Telescope (LSST) surveys as well as new R. Rana, S. P. Reynolds, Stern, J. A. Tomsick, and W. W. opportunities for a major involvement in NASA missions Zhang. X-RAY SPECTRAL COMPONENTS OBSERVED IN THE AFTERGLOW OF GRB 130925A. 2014. ASTROPHYS- and dark energy investigations identified in the strategic ICAL JOURNAL LETTERS. 784 (2). plan of the DOE Office of Science. Cao, Y., S. R. Kulkarni, S. Valenti, D. A. Howell, I. Arcavi, J. References Johansson, R. Amanullah, A. Goobar, J. Sollerman, F.

1. Cao, Y., S. R. Kulkarni, S. Valenti, D. A. Howell, I. Arcavi, Taddia, A. Gal-Yam, A. Horesh, I. Sagiv, S. B. Cenko, M. J. Johansson, R. Amanullah, A. Goobar, J. Sollerman, F. M. Kasliwal, P. E. Nugent, J. Surace, P. R. Wozniak, D. I. Taddia, A. Gal-Yam, A. Horesh, I. Sagiv, S. B. Cenko, M. Moody, U. D. Rebbapragada, and B. D. Bue. Ultraviolet M. Kasliwal, P. E. Nugent, J. Surace, P. R. Wozniak, D. I. Radiation from Supernova-Companion Collision in a 579

Page 580

Type Ia Supernova . 2015. Nature. 521 (7552): 328. Vestrand, E. S. Walker, and P. Wozniak. GRB 081007 and GRB 090424: The Surrounding Medium, Outflows, Dong, S., B. J. Shappee, J. L. Prieto, S. W. Jha, K. Z. Stanek, and Supernovae. 2013. The Astrophysical Journal. 774: T. W. Holoien, C. S. Kochanek, T. A. Thompson, N. Mor- 114. rell, I. B. Thompson, U. Basu, J. F. Beacom, D. Bersier, J. Brimacombe, J. S. Brown, P. Chen, E. Conseil, A. B. Johnson, J., D. Whalen, W. Even, C. Fryer, A. Heger, J. Smidt, Danilet, E. Falco, D. Grupe, S. Kiyota, G. Masi, B. Nich- and K. J. Chen. The Biggest Explosions in the Universe. olls, F. Olivares, G. Pignata, G. Pojmanski, G. V. Simo- 2013. The Astrophysical Journal. 775: 107. nian, D. M. Szczygiel, and P. R. Wozniak. ASASSN-15lh: The Most Luminous Supernova Ever Discovered. 2015. Khazov, D., O. Yaron, A. Gal-Yam, I. Manulis, A. Rubin, S. R. ArXiv e-prints. Kulkarni, I. Arcavi, M. M. Kasliwal, E. Ofek, Y. Cao, D. Perley, J. Sollerman, A. Horesh, M. Sullivan, A. V. Filip- Fontes, C. J., C. L. Fryer, A. L. Hungerford, P. Hakel, J. Col- penko, P. E. Nugent, D. A. Howell, S. B. Cenko, J. M. Sil- gan, D. P. Kilcrease, and M. E. Sherrill. Relativistic opac- verman, H. Ebeling, F. Taddia, J. Johansson, R. R. Laher, ities for astrophysical applications. 2015. High Energy J. Surace, U. D. Rebbapragada, and P. R. Wozniak. Flash Density Physics. 16: 53. Spectroscopy: Emission Lines from the Ionized Circum- stellar Material around < 10-Day-Old Type II Superno- Frey, L. H., W. Even, D. J. Whalen, C. L. Fryer, A. L. Hunger- vae. The Astrophysical Journal. ford, and C. J. Fontes. The Los Alamos Supernova Light- curve Project: Computational Methods. 2014. The Ofek, E., A. Zoglauer, S. Boggs, N. Barriere, S. Reynolds, C. Astrophysical Journal Supplement. 204: 16. Fryer, F. Harrison, S. Cenko, S. Kulkarni, A. Gal-Yam, I. Arcavi, E. Bellm, J. Bloom, F. Christensen, W. Craig, W. Frey, L., C. Fryer, and P. Young. Can Stellar Mixing Explain Even, A. Filippenko, B. Grefenstette, C. Hailey, R. Laher, the Lack of Type Ib Supernovae in Long-duration K. Madsen, E. Nakar, P. Nugent, D. Stern, M. Sullivan, J. Gamma-Ray Bursts?. 2013. The Astrophysical Journal Surace, and W. Zhang. SN 2010jl: Optical to hard X-ray Letters. 773: L7. observations reveal an explosion embedded in a ten solar mass cocoon. 2014. The Astrophysical Journal. Fryer, C.. Compact Object Formation and the Supernova 781 (1): 420. Explosion Engine. 2013. Classical and Quantum Gravity. 30 (24): 244002. Palaversa, L., Z. Ivezic, L. Eyer, D. Ruzdjak, D. Sudar, M. Galin, A. Kroflin, M. Mesaric, P. Munk, D. Vrbanec, H. Fryer, C. L., W. Even, B. W. Grefenstette, and T. W. Wong. Bozic, S. Loebman, B. Sesar, L. Rimoldini, N. Hunt-Walk- Observational constraints of stellar collapse: Diagnostic er, J. VanderPlas, D. Westman, S. J. Stuart, A. Becker, probes of nature’s extreme matter experiment. 2014. G. Srdoc, P. Wozniak, and H. Olusei. Exploring the Vari- AIP Advances. 4 (4): 041014. able Sky with LINEAR. III. Classification of Periodic Light Curves. 2013. The Astronomical Journal. 146: 101. Holoien, T. W., C. S. Kochanek, J. L. Prieto, K. Z. Stanek, S. Dong, B. J. Shappee, D. Grupe, J. S. Brown, U. Basu, J. Panaitescu, A.. Optical Flashes from Internal Pairs Formed F. Beacom, D. Bersier, J. Brimacombe, A. B. Danilet, E. in Gamma-Ray Burst Afterglows. 2015. apj. 806: 64. Falco, Z. Guo, J. Jose, G. J. Herczeg, F. Long, G. Pojman- ski, G. V. Simonian, D. M. Szczygiel, T. A. Thompson, Panaitescu, A., W. T. Vestrand, and P. Wo’zniak. Peaks of J. R. Thorstensen, and P. R. Wozniak. Six Months of optical and X-ray afterglow light curves. 2013. mnras. Multi-Wavelength Follow-up of the Tidal Disruption 433: 759. Candidate ASASSN-14li and Implied TDE Rates from ASAS-SN. 2015. ArXiv e-prints. Panaitescu, A., W. T. Vestrand, and P. Wo’zniak. ‘’Self-ab- sorbed’’ GeV Light Curves of Gamma-Ray Burst After- Hsiao, E. Y., M. A. Nyothers, P. W. Wozniak, and O. Yaron. glows. 2014. apj. 788: 70. Strong near-infrared carbon in the Type Ia supernova iPTF13ebh . 2015. Astronomy and Astrophysics. 578: Panaitescu, A., and T. Vestrand. Synchrotron and Inverse- A9. Compton Emissions from Pairs Formed in GRB After- glows (Analytical Treatment). 2014. The Astrophysical Jin, Z. P., S. Covino, M. Della Valle, P. Ferrero, D. Fugazza, D. Journal. 793: 104. Malesani, A. Melandri, E. Pian, R. Salvaterra, D. Bersier, S. Campana, Z. Cano, A. J. Castro-Tirado, P. D’Avanzo, J. Pastorello, A., J. L. Prieto, N. Elias-Rosa, D. Bersier, G. Hos- P. U. Fynbo, A. Gomboc, J. Gorosabel, C. Guidorzi, J. B. seinzadeh, A. Morales-Garoffolo, U. M. Noebauer, S. Haislip, J. Hjorth, S. Kobayashi, A. P. LaCluyze, G. Mar- Taubenberger, L. Tomasella, C. S. Kochanek, E. Falco, coni, P. A. Mazzali, C. G. Mundell, S. Piranomonte, D. E. U. Basu, J. F. Beacom, S. Benetti, J. Brimacombe, E. Reichart, R. Sanchez-Ramirez, R. J. Smith, I. A. Steele, Cappellaro, A. B. Danilet, S. Dong, J. M. Fernandez, N. G. Tagliaferri, N. R. Tanvir, S. Valenti, S. D. Vergani, T. Goss, V. Granata, A. Harutyunyan, T. W. Holoien, E. E. 580

Page 581

Ishida, S. Kiyota, G. Krannich, B. Nicholls, P. Ochner, G. variability due to GRB 080310. 2013. Astronomy and Pojma’nski, B. J. Shappee, G. V. Simonian, K. Z. Stanek, Astrophysics. 549: A22. S. Starrfield, D. Szczygiel, L. Tartaglia, G. Terreran, T. A. Thompson, M. Turatto, R. M. Wagner, W. S. Wiethoff, Whalen, D. J., J. Smidt, W. Even, S. E. Woosley, A. Heger, M. A. Wilber, and P. R. Wo’zniak. Massive stars exploding Stiavelli, and C. L. Fryer. Finding the First Cosmic Explo- in a He-rich circumstellar medium - VII. The metamor- sions. III. Pulsational Pair-instability Supernovae. 2014. phosis of ASASSN-15ed from a narrow line Type Ibn to apj. 781: 106. a normal Type Ib Supernova. 2015. mnras. 453: 3649. Whalen, D. J., W. Even, J. Smidt, A. Heger, K. J. Chen, C. L. Rubin, A., A. Gal-Yam, A. De Cia, A. Horesh, D. Khazov, E. Fryer, M. Stiavelli, H. Xu, and C. C. Joggerst. Supermas- O. Ofek, I. Manulis, O. Yaron, P. Vreeswijk, S. Ben-Ami, sive Population III Supernovae and the Birth of the First S. R. Kulkarni, D. A. Perley, Y. Cao, M. M. Kasliwal, S. B. Quasars. 2013. apj. 778: 17. Cenko, D. Xu, A. V. Filippenko, J. S. Bloom, K. I. Clubb, P. Whalen, D. J., W. Even, J. Smidt, A. Heger, R. Hirschi, N. Nugent, Y. C. Pan, J. M. Silverman, C. Badenes, I. Arcavi, Yusof, M. Stiavelli, C. L. Fryer, K. J. Chen, and C. C. Jog- D. A. Howell, S. Valenti, D. Sand, J. Sollerman, J. Johans- gerst. Pair-Instability Supernovae in the Local Universe. son, D. C. Leonard, J. C. Horst, J. M. Fedrow, S. F. Ar- 2014. The Astrophysical Journal. 797: 9. men, R. M. Quimby, A. Sternberg, T. Matheson, M. Sul- livan, K. Maguire, J. Surace, U. D. Rebbapragada, and P. Whalen, D. J., W. Even, L. H. Frey, J. Smidt, J. L. Johnson, C. R. Wozniak. Early Type II Supernova Light Curves. The C. Lovekin, C. L. Fryer, M. Stiavelli, D. E. Holz, A. Heger, Astrophysical Journal. S. E. Woosley, and A. L. Hungerford. Finding the First Cosmic Explosions. I. Pair-instability Supernovae. 2013. Sesar, B., Z. Ivezic, S. J. Stuart, D. M. Morgan, A. Becker, apj. 777: 110. S. Sharma, L. Palaversa, M. Juric, P. Wozniak, and H. Olusei. Exploring the Variable Sky with LINEAR. II. Halo Whalen, D., J. Johnson, J. Smidt, A. Heger, W. Even, and C. Structure and Substructure Traced by RR Lyrae Stars to Fryer. The Biggest Explosions in the Universe. II. 2013. 30 kpc. 2013. The Astronomical Journal. 146: 21. The Astrophysical Journal. 777 (2): 99. Shappee, B. J., A. L. Piro, T. W. Holoien, J. L. Prieto, C. Whalen, D., J. Johnson, J. Smidt, A. Meiksin, A. Heger, W. Contreras, K. Itagaki, C. R. Burns, C. S. Kochanek, K. Z. Even, and C. Fryer. The Supernova that Destroyed a Stanek, E. Alper, U. Basu, J. F. Beacom, D. Bersier, J. Bri- Protogalaxy: Prompt Chemical Enrichment and Super- macombe, E. Conseil, A. B. Danilet, S. Dong, E. Falco, D. massive Black Hole Growth. 2013. The Astrophysical Grupe, E. Y. Hsiao, S. Kiyota, N. Morrell, J. Nicolas, M. Journal. 774: 64. M. Phillips, G. Pojmanski, G. Simonian, M. Stritzinger, D. M. Szczygiel, T. A. Thompson, J. Thorstensen, M. Whalen, D., W. Even, C. Lovekin, C. Fryer, M. Stiavelli, P. Wagner, and P. R. Wo’zniak. The Young and Bright Type Roming, J. Cooke, T. Pritchard, D. Holz, and C. Knight. Ia Supernova ASASSN-14lp: Discovery, Early-Time Ob- Illuminating the Primeval Universe with Type IIn Super- servations, First-Light Time, Distance to NGC 4666, and novae. 2013. The Astrophysical Journal. 768: 195. Progenitor Constraints. 2015. ArXiv e-prints. Wozniak, P. R., D. I. Moody, Z. Ji, S. P. Brumby, H. Brink, J. Skillman, S. W., M. S. Warren, M. J. Turk, R. H. Wechsler, D. Richards, and J. S. Bloom. Automated Variability Selec- E. Holz, and P. M. Sutter. Dark Sky Simulations: Early tion in Time-domain Imaging Surveys Using Sparse Data Release. The Astrophysical Journal. Representations with Learned Dictionaries. 2013. In American Astronomical Society, AAS Meeting #221. Smidt, J., D. J. Whalen, E. Chatzopoulos, B. K. Wiggins, K. J. (Long Beach, CA, 6-10 January 2013). , p. #431.05. Chen, A. Kozyreva, and W. Even. Finding the First Cos- Washington, DC: American Astronomical Society. mic Explosions. IV. 90 - 140 M⊙ Pair-Instability Super- novae. 2015. The Astrophysical Journal. 805 (1): 44. Yan, L., R. Quimby, E. Ofek, A. Gal-Yam, P. Mazzali, D. Per- ley, P. Vreeswijk, G. Leloudas, A. de Cia, F. Masci, S. B. Smidt, J., D. J. Whalen, W. Even, B. Wiggins, J. L. Johnson, Cenko, Y. Cao, S. R. Kulkarni, P. E. Nugent, U. D. Reb- and C. L. Fryer. Population III Hypernovae. 2014. The bapragada, P. R. Wo’zniak, and O. Yaron. Detection Astrophysical Journal. 797: 97. of Broad H$$backslash$alpha$ Emission Lines in the Late-time Spectra of a Hydrogen-poor Superluminous Ukwatta, T. N., and P. R. Wozniak. Investigation of Redshift- Supernova. 2015. ArXiv e-prints. and Duration-Dependent Clustering of Gamma-ray Bursts. 2015. ArXiv e-prints. Yaron, O., D. A. Perley, A. Gal-Yam, J. H. Groh, A. Horesh, E. O. Ofek, S. R. Kulkarni, J. Sollerman, C. Fransson, A. Ru- Vreeswijk, P., A. Raassen, A. Smette, A. De Cia, P. Wozniak, bin, N. Sapir, P. Szabo, F. Taddia, S. B. Cenko, I. Arcavi, A. Fox, W. Vestrand, and P. Jakobsson. Time-dependent S. Valenti, M. M. Kasliwal, P. Vreeswijk, D. Khazov, O. D. excitation and ionization modelling of absorption-line 581

Page 582

Fox, Y. Cao, O. Gnat, P. L. Kelley, P. E. Nugent, A. V. Filip- penko, R. R. Laher, P. W. Wozniak, U. D. Rebbapragada, K. Maguire, and M. Sullivan. Confined Dense Circum- stellar Material Surrounding a Regular Type II Super- nova. Nature. 582

Page 583

Nuclear and Particle Futures Early Career Research Continuing Project Microscopic Fission Model for Data Needs Ionel Stetcu 20140581ECR Introduction Benefit to National Security Missions The current project is motivated by both basic-science The successful completion of this project will contribute needs and applications to new signatures for material to our ability to provide reliable nuclear reaction data for detection, proliferation and forensics. The development both basic and applied physics need in direct support of of new detection capabilities such as the time-projection the nuclear security mission. In particular, the principal chamber or SPIDER (Spectrometer for ion detection in investigator is working on fission simulations using the fission research) is complemented by an equally impor- code CGMF (cascade gamma multiplicity for fission), tant theoretical modeling effort, leading to updates in which provides a description of the prompt neutron and the state-of-the-art U.S. ENDF/B-VII.1 evaluated library gamma properties. This understanding is essential for that is used in applications. This project aims at comple- developing new signatures for nuclear material detec- menting planned experiments at LANSCE (Los Alamos tion, nuclear proliferation, forensics and energy genera- Neutron Science Center) by providing reliable modeling tion, which are at the core mission of the laboratory for relevant fission quantities (fission fragment yields, and NNSA. More reliable input provided by this work kinetic energy distribution). We will also obtain informa- will lead to more reliable predictions. Finally, given the tion about other quantities about which only indirect planned experimental studies at the Los Alamos Neu- experimental information can be inferred (angular tron Science Center (LANSCE), a successful project will momentum of the fission fragments, excitation energy allow the development of an important capability at Los sharing between fragments). Hence, once implemented Alamos National Laboratory in both fundamental and ap- and tested, it will constitute a powerful tool in fission plied nuclear physics, and it will complement the current research. experimental program. For the execution of the project, we will make use of Progress the latest-generation supercomputers to implement and In the last fiscal year, I have finished all the major com- study nuclear fission of actinides in the density function- putational implementation proposed. The code has been al theory (DFT) framework. DFT is the only microscopic tested on titan at ORNL and moonlight at LANL, scales approach currently feasible for heavy nuclei, and has the very well and makes efficient use of the GPUs (graphical additional advantage of being able to treat on the same processing units). Depending on the size of the problem, footing the isolated nuclear system and the interaction the GPU code is faster by factors of 15 to 25 with respect and response in time of a nucleus with external probes to the MPI (message passing interface) code from which (e.g., incoming neutrons, photons, etc). In the current it originated. project, we will stretch the nucleus close to the breaking point, release it, and follow the dynamics of fission. Such For a particular nucleus, the code starts with a set of calculations require extensive computational resources, self-consistent densities (obtained via another code). Dif- as one has to solve tens to hundreds of thousands of ferent external perturbations can be applied in order to time-dependent non-linear coupled three-dimensional drive the system toward the point of breaking into two partial differential equations and the developed software (or more fragments), process called fission. For applica- represents a true quantitative jump (it is at least a factor tions of interest, this process happens at low excitation of a 1000 to 2000 more complex than that of the “near- energies, and the goal of this project is to model this est” competition). Because of its complexity, this soft- process and extract quantities of interest. However, ap- ware represents a first step toward exascale computing. plying external fields (which model real external probes 583

Page 584

like photons and neutrons), has to be made with care be- of the processes that can be investigated with the current cause one needs to keep the system cool, that is as close as capabilities. This work was published in Phys. Rev. Lett. 114 possible to the lowest energy surface. In order to achieve (2015) 012701 and Acta Phys. Polonica B 46 (2015) 391 . this, I have implemented the quantum friction (QF), which acts like an external field, designed to keep the energy low, Future Work while the system is deformed toward fission. QF has two In the period remaining until the end of the funding pe- components, one of them in the so called normal channel, riod, I will concentrate on obtaining, in a time-dependent and another one in the pairing channel. I have found that approach, asymmetric fission at low energies. I will in- the cooling in the normal channel works very well, and vestigate several scenarios in which the asymmetry could reduces the final energy (when the external field is turned be introduced via a perturbation in time. If this approach off) by about 10 MeV. However, in the pairing channel, the pumps too much energy into the system, I will have to QF did not perform as expected, and the energy cannot revert to introducing the asymmetry directly into the be reduced by including QF. As a result, the evolution is initial conditions, that is the starting densities and wave not adiabatic, and the nucleus fission occurs only at high functions which are evolved in time. This is possible, but I energies, of the order of a few tens to hundreds of MeV. would like to avoid it because of the high cost in computing This is not satisfactory, as the scope of the current project power. I will continue the investigations of 240Pu in smaller is obtaining fission at low energies. Hence, I was forced boxes, and once fission achieved, I would like to extract to look for other solutions that would allow a description the total kinetic energy of the two fragments as a function of fission at low energies. For this, I have implemented of mass asymmetry, at least for a few light to heavy mass modifications in the code that provides the initial condi- fragments. I also plan to investigate whether neutrons are tions constraints that allow me to obtain the initial system emitted during the breakup in this approach. already deformed at low energies. The initial conditions are used as input in the time-dependent code, where I Conclusion release the constraint in a reasonably small amount of The main goal of the current project is to provide a first time, and follow the evolution in time. In this way, keeping fully quantum-mechanical description of nuclear fission, the excitation energy very low, I was able to increase the the physics process in which a heavy nucleus breaks apart. deformation of the system by 30%, as a consequence of Our approach can provide a reliable method to extract Coulomb interaction, which elongates the system, driving it fission properties that can be later used as input in simula- toward fission. While the system did not break as of today, tions with applications to energy generation, global secu- this result can be understood: in the solutions that I have, rity, weapons, astrophysics. We plan to obtain fission yields one needs to add another perturbation, that would make in charge, mass, kinetic energy and angular momentum, the system break asymmetrically, which is indeed observed and to study the properties of neutrons emitted during the (it is almost impossible to break the system at low energies breakup. Selected results will be tested against experimen- in two identical fragments). I am exploring at the moment tal data and predictions of other theoretical developments how one can produce two asymmetric fragments, without in our group. adding additional constraints in the initial conditions (the code that produces initial conditions is very demanding Publications and can converge to the solutions slowly). If this is not Stetcu, . NUCLEAR STRUCTURE AND DYNAMICS WITH DEN- possible with external probes, one can still implement ad- SITY FUNCTIONAL THEORY. 2015. ACTA PHYSICA PO- ditional conditions in the initial solutions. LONICA B. 46 (3): 391. Stetcu, , C. A. Bertulani, Bulgac, Magierski, and K. J. Roche. Together with collaborators in Warsaw and Seattle, I have Relativistic Coulomb Excitation within the Time Depen- also implemented the required tools to calculate the total dent Superfluid Local Density Approximation. 2015. kinetic energy, once the two fragments separate. The code PHYSICAL REVIEW LETTERS. 114 (1). has been tested using fission at high energies, but it does not require any changes for low-energy processes. Thus, I am well positioned to calculate this very important observ- able, which is part of the goals of the current investigation, once the two fragments separate at low energies. I have also worked to extend a calculation involving a photon interacting with a 238U nucleus. This is important and relevant for the current project as photo-fission is one 584

Page 585

Nuclear and Particle Futures Early Career Research Continuing Project Relativistic Electrons in Magnetized Plasmas Zehua Guo 20140605ECR Introduction mitigation of disruption-induced runaway electrons is of This research project will develop a comprehensive phys- great importance to overcome the plasma-wall interac- ics model and the predictive capability for the genera- tion challenge on the ITER fusion reactor experiment, tion and transport of relativistic electrons in magnetized which is of high priority to the DOE Office of Science. plasmas with applications to tokamak fusion reactor, The comprehensive Fokker-Planck solver developed earth’s radiation belt and solar corona. Runaway elec- in this project is also of interest to earth radiation belt trons (RE) during tokamak disruptions have been iden- modeling, a priority for our national security mission via tified to be a potential showstopper for ITER tokamak LANL’s Global Security program in space weather model- reactor (ITER being built in France), because the pro- ing and mitigation of intense relativistic electrons, which jected tens of mega-ampere REs will cause intolerable pose a threat to the satellites, spacecraft and astronauts. materials damage when they intercept the reactor wall. Other applications include high energy astrophysics and To develop control and mitigation strategies, it is impera- solar coronal heating. tive to understand the physics underlying the generation Progress and transport of these REs. We will upgrade our existing Fokker-Planck solver in the tokamak geometry to study We have derived the drift-kinetic equations for relativ- the distribution functions of REs. Essential effects due istic electrons in tokamaks and extended our existing to electron radial drift, radiation damping, Coulomb Fokker-Planck code to be suitable for runaway electrons collisions and wave-particle interactions will be treated (RE). The relativistic collision operator for small-angle on the same footings for the first time. Stability analysis collisions has been implemented and the radial drift due will be performed to identify unstable waves, and the to tokamak magnetic geometry is included. We have subsequent wave-particle interaction will be incorpo- also studied the instabilities using analytical distribution rated into the Fokker-Planck model with quasi-linear functions of primary and secondary REs. The electrostat- theories. Since the collisional slowing down is ineffec- ic Trivelpiece-Gould mode dominates for the primary RE tive for REs, we will explore the intriguing possibility of distribution and the sonic-whistler type modes dominate using waves, either self-excited or externally injected, for the secondary REs. A rather general tool for analyz- to control the REs. We will focus on a novel mechanism ing linear instabilities of arbitrary distribution functions that enhances the pitch angle scattering of REs and cre- has been developed and is ready to be verified against ates an energy exchange channel between thermal and the analytical results. For the remaining four months REs through nonlinear wave-wave coupling between the of FY15, we will continue to benchmark our result with whistler wave and the much slower electrostatic modes. previous analytical and numerical solutions of the RE This study will help establish the physics basis for reduc- distributions and study the effects of radial transport on ing/removing relativistic electrons via wave injection or RE distribution. selective excitation. Future Work Benefit to National Security Missions This project started in Oct 2014. For FY15, we focused This project aims to develop a one of a kind predic- on verifications against previous analytical and numeri- tive modeling capability for runaway electrons in a cal results of the RE distributions, extension to include magnetized plasma, commonly observed in space and synchrotron radiation losses and analysis of their im- astrophysical plasmas, and in the laboratory tokamak portance in determining the RE distribution. In the next experiments. The immediate impact on the control and fiscal year, our major tasks include: (1) understand the 585

Page 586

radial losses of REs due to natural drift motion in tokamak geometry; (2) analyze the kinetic and fluid instabilities; (3) study the effects of wave-particle interactions on RE distri- butions. Using the steady-state (or transient) RE distribu- tion functions, detailed stability analysis of the RE distribu- tion function will be performed to understand the unstable waves. With the new distribution function, we will analyze the wave characteristics and compare to previous work where the radiation damping and radial transport physics were not accounted for. Furthermore, we will include the anomalous transport due to wave-particle interactions into the Fokker-Planck equation and study how RE distributions are modified. We will examine the idea that wave/turbu- lence induced transport can help to limit the growth rate of runaway electrons and quantity the effect together with the radial losses. Conclusion The project will produce a reliable model and code to simulate the relativistic electrons in magnetic fields for studying the runaway electron formation and control in a tokamak. The predictive capability will also be applicable to radiation belt dynamics and coronal flare physics. This work will help elucidate the role of nonlinear wave-wave coupling processes in the presence of relativistic electrons, and establish the physics bases for using wave-particle interaction to control runaway electron generation and radial transport in ITER. 586

Page 587

Nuclear and Particle Futures Early Career Research Continuing Project Electron Transport in Warm and Hot Dense Matter Charles E. Starrett 20150656ECR Introduction Green’s function code. This has been tested against the In this project we will develop a completely novel PI’s existing average atom code that uses the standard simulation capability for dense plasmas. Such plasmas orbital based Kohn-Sham density functional theory. The exist in inertial confinement fusion plasmas (eg. at the Green’s function code gives the same results to within national ignition facility (NIF)) as well as in astrophys- numerical tolerance, as expected, but is a factor of 5 ics, for example, in the interiors of giant planets and in faster than the orbital based code. white dwarfs. Building on a existing computational tool The Green’s function implementation also avoids two developed by the PI, we will extend a method that has key numerical difficulties that are present in orbital been very successful in condensed matter physics, into based implementations - namely the search for weakly the dense plasma regime. This method, known as KKR- bound states and the tracking of resonance states. This Green’s function, will provide an description of these numerical speed-up and fundamental new approach to plasmas with unparalleled realism across the tempera- the single center model have been accepted for publica- ture and density regime of interest. We will apply the tion in High Energy Density Physics. method to the calculation of electrical conductivity, which is of importance to the modeling of inertial con- Future Work finement fusion experiments. In the lower temperature In this fiscal year we will develop the implementation regime the new capability will complement current high of the single center solution. This will involve solving accuracy tools, while for hot dense plasmas the method the Schrodinger equation for complex energies. The will provide a gold standard where none currently exists. implementation will be tested against the PI’s existing real-space implementation. Once this is completed we Benefit to National Security Missions will begin the implementation of the so-called structure The project aims to greatly improve our understanding constants using the Ewald technique. of electron transport in dense plasmas. As such, it will ultimately provide a tool that could be used to provide In FY16 we will complete the implementation of the crucial input information into hydrodynamical simula- so-called structure constants and combine this with the tions of implosion experiments of inertial confinement single center code that was completed in FY15. We will fusion. Such experiments, like those conducted at the test this and implement the calculation of the conductiv- national ignition facility and elsewhere, are essential for ity. the stockpile stewardship mission of the DOE NNSA. Conclusion The new method that we aim to develop will provide The ultimate result of this work will be a completely simulations of of unparalleled physical realism across new computational framework for calculating electri- the temperature and density regime of interest to the cal conductivity in warm and hot dense plasmas. The inertial confinement fusion community. It will put LANL method will compare favorably to the existing gold stan- at the forefront of the rapidly evolving field of electron dard methods at low temperature, and provide the gold transport in warm and hot dense matter. standard at higher temperatures, where none currently exists. The method will lead to a new understanding of Progress electron transport in dense plasmas and the resulting Since the beginning of this project at the end of Janu- calculations will of high relevance to the modeling of in- ary 2015, we have implemented the single center 587

Page 588

ertial confinement fusion experiments (eg. at the National Ignition Facility). Publications Starrett, C. E.. A Green’s function quantum average atom model. 2015. HIGH ENERGY DENSITY PHYSICS. 16: 18. 588

Page 589

Nuclear and Particle Futures Early Career Research Continuing Project A Step toward Nuclear Reaction Studies for Applications at FRIB Shea M. Mosby 20150683ECR Introduction lived nuclei. The project will do this by performing two The neutron capture reaction - in particular, the cross independent experiments on the same isotope - one section - is of great importance for a variety of fields using an established method for studying neutron ranging from nuclear astrophysics to defense programs. capture, and one using the new method. By the end of This process must be understood across a whole host the project, the two methods should provide consistent of elements and isotopes, with half-lives ranging from results for neutron capture, thus validating the new thousands of years to a fraction of a second. Current method and opening up hundreds of important isotopes experimental access to this process is limited to stable for study. and near-stable isotopes (half-lives ranging from months Progress to years), and theoretical calculations are comparatively unreliable due to uncertainty in the nuclear structure The project start and end date was delayed until June inputs they require. to better match with facility run schedules. Neverthe- less, there is significant progress in key areas needed to This project seeks to validate a potentially game-chang- execute the experiments early next FY. ing technique for measuring the necessary nuclear structure to accurately calculate the neutron capture • The Argonne proposal was submitted and accepted cross section. Direct measurements using the Detector such that all of our requested beam time is ap- for Advanced Neutron Capture Experiments (DANCE) at proved. We are communicating with them now to Los Alamos National Laboratory will be made in parallel schedule the experiment. with indirect measurements using the Apollo instrument • There has been no call for proposals for the Los at Argonne National Laboratory. If successful, Apollo can Alamos aspect of the project. Nevertheless, we’ve be used to vastly improve our understanding of neutron performed rate calculations and analyzed some past capture on short-lived isotopes relevant for both nuclear data under similar conditions such that we under- astrophysics and defense programs. stand what our beam time request will be. Benefit to National Security Missions • We’ve begun the analysis and simulation tool chain Nuclear reaction rates (particularly neutron capture) on work such that we understand the scope of work short-lived isotopes are known to have impact for basic needed. We believe that these tools will be ready science questions such as the origin of the elements prior to each experiment running. heavier than iron, as well as radiochemical diagnostics used as part of our nuclear security mission. The cur- Future Work rent state of the art experimental methods for studying • Execute experiments at both Argonne National these reaction rates are limited to stable or long-lived Laboratory (ANL) and the Los Alamos Neutron isotopes. While several methods have been explored, no Science Center. Based on current information, we truly robust method of constraining neutron capture on expect to run both experiments between November short-lived isotopes has been demonstrated. 2015 and January 2016. Both experiments will run at the behest of their respective facilities’ operations This project seeks to remedy that situation by perform- schedule. ing the experimental work needed to validate a new method to constrain neutron capture rates on short- • Complete analysis software framework to move 589

Page 590

from raw data to the relevant gamma-ray cascade spectra for comparison with theory. • Complete the theory code / detector simulation inte- gration so the experimental data can be connected to the relevant physics codes. The detector simulation output should be in the same data format as the ex- perimental data so the same analysis toolchain can be used for consistency. • Preliminary analysis of data performed such that rough results are in hand to guide the final stages on analysis. Detector calibrations (e.g. gain matching, time align- ment) will be complete, and detector coincidence (e.g. particle / gamma-ray coincidences for the ANL experi- ment) spectra constructed. Conclusion The project will directly measure the 96Zr neutron capture cross section using DANCE, and constrain a theoretical calculation of the cross section using indirect techniques with Apollo. Furthermore, DANCE can make an indepen- dent measurement of the nuclear structure properties of 96Zr and independently constrain theoretical calculations of the capture cross section. It is expected that these three independent measurements will result in a consistent cross section prediction. 590

Page 591

Nuclear and Particle Futures Early Career Research Continuing Project Optimization of Compton Source Performance through Electron Beam Shaping Nikolai Yampolsky 20150690ECR Introduction has been found. This work resulted in the finding of the The project aims to investigate a novel way for increas- photon Wigner distribution function of a single-electron ing brightness of light sources based on inverse Comp- emission. The Wigner distribution function for the ton scattering (ICS). The key idea is to manipulate the ensemble of electrons has been found and was reduced electron beam phase space so that all the electrons emit to the conventional form of a convolution between the photons of the same energy in some direction. During electron distribution function in the bunch and a single- the course of the project we will study the feasibility of particle Wigner distribution function. The convolution the scheme and determine how large an improvement has been calculated analytically for quasi-mono-energet- can be achieved. Upon success of the project we will ic and low-divergence electron bunch and the result has describe practical modifications to currently existing been found in a compact form suitable for further analy- and future ICS light sources which should significantly sis. The resulting form of the photon Wigner distribution improve their quality. function clearly shows the optimal beam conditioning resulting in the brightest inverse Compton scattering Benefit to National Security Missions source. The proposed scheme may have a strong impact on Ma- Future Work RIE in its early stage. There is a possibility of reusing the planned MaRIE Injector Test Stand (MITS) as an ICS light • Determine the required conditioning of the electron source to enable early experiments. This project will also beam phase space be beneficial for testing X-ray detectors planned for the • Determine the beamline which is required to obtain MaRIE FEL. the optimal beam distribution. • Design and simulate beamline using standard accel- Two ICS sources have been built and two future facilities erator code ELEGANT. are proposed to generate γ-rays through ICS for material • Optimize beamline to minimize deleterious higher- science applications. Increasing the brightness of those order transport effects. sources will be beneficial for their host DOE facilities. Conclusion Recently, high-flux gamma rays have been proposed as approaches for detecting special nuclear materials and We expect to demonstrate that ICS brightness can be to address international nuclear proliferation concerns. increased through appropriate conditioning of the Compact ICS source is an attractive option for generating electron beam phase space. We expect to eliminate these photons. The proposed method for conditioning the largest contribution to the brightness degradation, the beam phase space will reduce the ICS bandwidth i.e. either due to the angular divergence or the energy which will allow meeting the requirements. Moreover, spread of the electron beam. At the moment, it is an increasing the brightness of such a source will allow open question whether both effects can be suppressed reducing the power required to operate such a machine. simultaneously. We conservatively anticipate that only one of them can be compensated and the final source Progress brightness will be defined by the smallest rather than the largest effect. The equation of motion of an arbitrary relativistic elec- tron in the fields of the plane electro-magnetic wave has been re-derived and the resulting radiation field 591

Page 592

Nuclear and Particle Futures Early Career Research Final Report First Principle Study of Relativistic Beam and Plasma Physics Enabled by Enhanced Particle-In-Cell Capability Chengkun Huang 20130744ECR Abstract investigating future light source development, astrophys- Relativistic charged particles are commonly generated ical processes, kinetic high energy density laboratory in nature and man-made experiments. They are ac- plasma (HEDLP) and alterative ICF concepts. celerated/decelerated either naturally or deliberately Background and Research Objectives for various purposes, e.g. 10s of GeV electron beams can be manipulated to produce intense coherent short- The Particle-In-Cell (PIC) [1] method is a computation- wavelength radiation in an X-ray Free Electron Laser ally efficient, fully kinetic ab-initio method widely used (FEL) — an indispensible tool for material research at in plasma and particle beam modeling. Specifically, PIC many scientific frontiers. Relativistic particles also play models have been applied to a number of challenging important roles in Inertial Confinement Fusion (ICF) en- problems involving relativistic particles: (1) understand- ergy generation employing laser-driven particle beams ing how the microbunching instability can arise from the as ignitor and in various astrophysical phenomena such interactions between a highly relativistic beam and its as jets and cosmic rays. The Particle-In-Cell (PIC) method Coherent Synchrotron Radiation (CSR) in a FEL [2] bunch has proven successful for kinetic simulations of weakly compressor and how it affects the beam quality and to mildly relativistic beams and plasmas, but its accuracy FEL performance; (2) investigation of the collisionless will decrease for highly relativistic particles due to the particle acceleration process in astrophysical shock and imprecise representation of the particles and their fields. the deceleration process related to the collective wave- In many cases, such inaccuracy can lead to numerical particle interaction in high-energy-density plasmas. The instabilities detrimental to the simulations. In this work, former application is directly relevant to the success of we systematically developed, validated and integrated the MaRIE X-ray FEL signature facility at LANL, while the recent improvements on the major components of the latter addresses fundamental science questions in both PIC method, enhancing its capability for the model- astrophysics and a high-risk yet high-potential Fast-Igni- ing of relativistic phenomena. A new high-order field tion (FI) [3] technology for inertial fusion energy. Al- solver with careful control of the dispersion properties though the standard PIC method with the second order is developed and benchmarked for the modeling of syn- accurate numerical algorithm is well suited for the study chrotron radiation of highly relativistic charged particles. of beam and plasma dynamics at low speed, its applica- The benefits of high order particle shapes, digital filters tion to relativistic phenomena is severely limited. The and maintaining conservative properties are elucidated, major limitations come from the dispersive and anisotro- especially with regard to the suppression of numerical pic propagation properties of the electromagnetic wave instabilities. An unexplored attribute, the spectral fidel- on the computation grid, the electrostatic and electro- ity, is found to have a critical role in the stability of the magnetic numerical instabilities associated with the use simulation model and used in the further development of a computation grid, and the undesirable noise due to of a novel PIC method that is intrinsically stable. We the unphysical Cherenkov radiation of relativistic parti- already applied the enhanced capability for high fidelity cles and under-resolved dynamics. These limitations ne- self-consistent study of a class of accelerator and plasma cessitate a new computational framework based on the physics problems involving such particles and the radia- PIC method to compute and understand cutting-edge tion they produced, including micro-bunching instability problems involving highly relativistic charged particles in FEL bunch compressor, particle acceleration in astro- for applications such as FEL and FI-ICF. Our approach physics and isochoric deposition of energetic particles in is to address the issues of the standard PIC method by ICF capsule, thus establishing a numerical framework for identifying the key deficiencies in PIC components and 592

Page 593

by carefully redesigning them. These include the lack of obtained accurate numerical solutions and corresponding the spectral fidelity in the charge/current deposition and approximated growth rates. It has been shown that in the the accumulation of the phase error in the electromagnetic infinitesimal time step limit, and for sufficiently large grid field solver. We will devise new numerical schemes to over- size in the finite time step case, the dominant FGI is caused come these deficiencies. With this enhanced PIC capability, by the intersection of the finite grid plasma mode and the our ultimate goal is to provide accurate and self-consistent two lowest order alias modes. On the other hand, for rela- modeling and understanding to the relativistic beam/ tivistic cold plasma flow under the most relevant simula- plasma problems of importance to LANL’s mission and to tion parameters, an intersection of a stationary mode and the broader scientific community, particularly related to alias modes, due to inexact charge conservation, can lead the coherent synchrotron radiation and collective particle to the most dominant instability. The growth rate for this acceleration/deceleration processes. instability is obtained analytically and there is no threshold grid size or time step size below which this instability will Scientific Approach and Accomplishments not occur. This instability and associated growth rate have In the PIC method, a fundamental incompatibility between been confirmed in our simulations with excellent agree- the continuous particle model and the discrete field repre- ment with our predictions [6]. An investigation of charge sentation causes an aliasing effect. Under certain numeri- conserving schemes commonly used in 1D electromagnetic cal conditions, the alias modes can interact with the other codes indicates that each alias mode is unstable by itself, modes admitted in the electromagnetic PIC model, causing similar to the electrostatic case. numerical instabilities that can render unphysical simula- Next, the origin of the FGI is studied by resolving the dy- tion results or even destroy the simulation. We found that namics in spectral domain at the single particle level and alias modes can also be unstable by themselves, depen- at the collective motion level [7]. The concept of spectral dent on the deposition scheme employed. Such numerical fidelity, i.e., how closely the PIC model resembles the un- instabilities will substantially affect the modeling of kinetic derlying physical system in the spectral domain, is devel- of the relativistic particles. oped and used to contrast the PIC model with the grid- To thoroughly understand the characteristics and the origin less model. The systematic spectral phase and amplitude of the grid type instabilities, we first reviewed the numeri- errors from the charge deposition and field interpolation cal dispersion relation of the standard electromagnetic PIC are quantified for common particle shapes used in the PIC model based on the Yee Finite Difference Time Domain models for the first time. It is shown through such analysis (FDTD) solver [4]. We abided by the following two guide- and in simulations that the lack of spectral fidelity relative lines that have been overlooked before: to the physical system due to the existence of aliased spa- tial modes is the major cause of the FGI in the PIC model. (1) The eigenmodes in a PIC model consist of spatially In this regard, the difference between the PIC model and continuous, temporally discrete particle eigen-distribution the physical system (or gridless model) is technical, not functions, and spatially and temporally discrete eigenfields. fundamental — with a properly chosen particle shape such The sole cause of aliasing in PIC is the sampling of the con- that its spectral content is band-limited to the maximum tinuous spatial variable onto the discrete one. As time and allowable wave-number in the simulation, the PIC mod- space are independent variables, this spatial aliasing effect els becomes effectively a gridless model. This opens the will not introduce any temporal aliasing. question of the optimal compact particle shapes for the PIC model. In addition, as with many physical instabilities (2) Unnecessary algebraic approximations that may result which result from the constructive feedback of modes, it is in an approximation to the original model should be understood that the systematic phase error plays a critical avoided. role in the development of FGI. This finding is verified in our simulation comparison with the gridless model. With Based on these guidelines, we rigorously derived the this new understanding of how and where FGI arises in the faithful numerical dispersion relation for the standard PIC model, the optimal strategy/method to mitigate FGI electromagnetic PIC model with a simple, direct current has been devised and implemented [8]. As this improve- deposition scheme, which does not exactly conserve elec- ment is readily extended to other particle mesh method, tric charge. We analyzed the numerical dispersion of the our work not only clarify and pave the way for the de- electrostatic-like mode in a 1-D electromagnetic PIC model velopment of better PIC models which is used in plasma with a drifting cold plasma and the Finite Grid Instability and beam physics, it is also applicable in a wide range of (FGI) [5] is predicted to be present. This instability in the N-body type computation physics problem. electromagnetic PIC model has not been previously ana- lyzed, nor has it been confirmed from simulation. We have 593

Page 594

A third area of our improvement focuses on the need to Concept workshop in 2014 and the International Confer- improve the dispersive properties of the field solver, which ence on Numerical Simulation of Plasmas (ICNSP) in 2015. is essential for modeling CSR from a relativistic beam. We have given one invited talk (ICNSP 2015) and received Although analytic solution of the CSR field exists in 1D, this another invitation at the 8th International West Lake problem is substantially complicated in realistic multiple Symposium (not given). M. D. Meyers, a graduate student dimension due to the lack of such understanding of the supported in the project, received a travel grant to the APS CSR field in multi-dimension thus there is no good known DPP annual meeting in 2014 for his work on the project. solution to benchmark the simulation. We approached Two publications, one on the 3D synchrotron radiation this problem from both the semi-analytic and computa- near field [11] and another one on the numerical disper- tion modeling aspects. For the synchrotron radiation near sion and finite grid instability [6] have resulted from the field, we extended the well-known 1D line charge model work in this project. One manuscript on the spectral fidel- [9] to two spatial dimensions [10] in the bending plane and ity of the PIC model is under peer-review [7] and another recently to full three dimensions [11]. These models over- one on the mitigation method is in preparation [8]. come the diverging space charge potential and fields issues in the 1D line charge model for a coasting beam, therefore Impact on National Missions they are amenable to numerical integration even in the Our work has established a high-fidelity first-principle presence of the singularity from space charge field. Both predictive modeling capability which resulted in improved our 2D and 3D models exhibit self-similar solution of the understanding of how relativistic charge particles interact single particle synchrotron radiation field kernel through a with the environment, including their own coherent radia- rescaling of the spatial variables. This important property tion. Modeling of other plasma physics and beam physics allows a convenient convolution procedure to replace the problems, especially of kinetic plasma behavior under summation process for the CSR calculation of a smooth extreme conditions, will further demonstrate its potential. beam profile, which can be justified when the beam size is The enhanced modeling capability provides important much smaller than the bending radius in the bunch com- impetus in many areas related to LANL’s missions for the pressor. We have built a 2D CSR numerical model based on national security, e.g., high energy density plasma, acceler- this procedure. Our team has also successfully developed ator beam dynamics, astrophysics and space applications, and tested a parallel high order finite volume field solver thus yielding many scientific and programmatic possi- and validated its accuracy in the phase velocity. The solver bilities. It provided training to student/postdoc in these has been implemented into a production parallel PIC code. important fields and the experience gained can be used We investigated the tunable dispersion property of this to enhance other state-of-the-art PIC codes used at many field solver and developed a tuning strategy optimized for frontiers, including LANL’s flagship VPIC code. The im- the low (or high) frequencies for CSR (or numerical Cher- proved modeling capability and the new insights concern- enkov noise suppression). Another key component of the ing coherent synchrotron radiation by a relativistic beam, accurate electromagnetic solver for our relativistic beam the microbunching instability in XFEL as well as the particle and plasma modeling is the absorbing boundary condition acceleration/deceleration mechanisms in high energy that can correctly matched to the order of the filed solver. density plasmas link directly into applications enhancing We have extended the state-of-the-art absorbing bound- national security and energy security. The new simulation ary condition — Perfectly Matched Layer (PML), into a high capability also supports future signature facilities at LANL, order PML boundary condition, which proves to be critical such as the Matter-Radiation Interactions in Extremes facil- for stable high order field solver. With this new high order ity. field solver and absorbing boundary condition, we have benchmarked CSR of a coasting beam with our 2D model. References In particular, our result indicated that (1) the velocity field 1. Birdsall, C. K., and B. A. Langdon. Plasma Physics via from the Liénard–Wiechert formulation dominates in the Computer Simulation. 1991. CSR at beam energy below ~ 50MeV, and (2) the transient effect at the magnetic dipole entrance leads to smaller 2. Emma, P., R. Akre, J. Arthur, R. Bionta, C. Bostedt, J. field at the front, opposite to the prediction from the 1D Bozek, A. Brachmann, P. Bucksbaum, R. Coffee, F. J. model. Decker, Y. Ding, D. Dowell, S. Edstrom, A. Fisher, J. Frisch, S. Gilevich, J. Hastings, G. Hays, P. Hering, Z. Over the two-year project period, we have made technical Huang, R. Iverson, H. Loos, M. Messerschmidt, A. presentations at the North America Particle Accelerator Miahnahri, S. Moeller, H. D. Nuhn, G. Pile, D. Ratner, J. Conference in 2013, the APS Division of Plasma Physics an- Rzepiela, D. Schultz, T. Smith, P. Stefan, H. Tompkins, nual meetings in 2013 & 2014, the Advanced Accelerator J. Turner, J. Welch, W. White, J. Wu, G. Yocky, and J. 594

Page 595

Galayda. First lasing and operation of an aa ngstrom- Huang, C., T. J. T. Kwan, and B. E. Carlsten. Synchrotron wavelength free-electron laser. 2010. Nature Photon- Radiation Near Field In 3D. 2014. In 25th Particle Accel- ics. 4 (9): 641. erator Conference. (Pasadena, 30 Sept. - 4 Oct. 2013). , p. 487. Pasadena: JACoW.org. 3. Fern’andez, J. C., J. J. Honrubia, B. J. Albright, K. A. Meyers, M. D., C. K. Huang, Y. Zeng, S. A. Yi, and B. J. Al- Flippo, D. C. Gautier, B. M. Hegelich, M. J. Schmitt, bright. On the numerical dispersion of electromagnetic M. Temporal, and L. Yin. Progress and prospects of particle-in-cell code: Finite grid instability. 2015. JOUR- ion-driven fast ignition. 2009. Nuclear Fusion. 49 (6): NAL OF COMPUTATIONAL PHYSICS. 297: 565. 065004. 4. Yee, K. S.. Numerical solution of initial boundary value problems involving maxwell’s equations in isotropic media. 1966. IEEE Transactions on Antennas and Propagation. 14 (3): 302. 5. Langdon, B. A.. Effects of the spatial grid in simulation plasmas. 1970. Journal of Computational Physics. 6 (2): 247. 6. Meyers, M. D., C. K. Huang, Y. Zeng, S. A. Yi, and B. J. Albright. On the numerical dispersion of electromag- netic particle-in-cell code: Finite grid instability. 2015. Journal of Computational Physics. 297: 565. 7. Huang, C. K., Y. Zeng, Wang Y., M. D. Meyers, S. A. Yi, and B. J. Albright. Finite grid instability and spectral fidelity of the electrostatic Particle-In-Cell algorithm. 2015. submitted, arXiv:1508.03360. 8. Huang, C. K., Y. Zeng, Wang Y., M. D. Meyers, S. A. Yi, and B. J. Albright. A simple method to mitigate finite grid instability in the Particle-In-Cell algorithm. 2015. in preparation. 9. Saldin, E. L., E. A. Schneidmiller, and M. V. Yurkov. On the coherent radiation of an electron bunch moving in an arc of a circle. 1997. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment. 398 (2-3): 373. 10. Huang, C., T. Kwan, and B. Carlsten. Two dimensional model for coherent synchrotron radiation. 2013. Physi- cal Review Special Topics - Accelerators and Beams. 16 (1): 010701. 11. Huang, C., T. J. T. Kwan, and B. E. Carlsten. Synchrotron Radiation Near Field In 3D. 2014. (Pasadena, 30 Sept. - 4 Oct. 2013). p. 487. Pasadena: JACoW.org. Publications Huang, C. K., Y. Zeng, Y. Wang, M. D. Meyers, S. Yi, and B. J. Albright. Finite grid instability and spectral fidelity of the electrostatic Particle-In-Cell algorithm. Computer Physics Communications. 595

Page 596

Nuclear and Particle Futures Early Career Research Final Report Answer to Heavy Element Production Puzzle by Measuring Neutron-induced Charged Particles at LANSCE Hye Young Lee 20130758ECR Abstract unmeasured nuclides will certainly exceed present data Despites decades of work on nuclear astrophysics, quality. how heavy elements are produced remains an actively Scientific Approach and Accomplishments studied field. Above iron, heavy nuclei are produced via neutron capture, via either a s(low) process or r(apid) We have developed the instrumentation enabling the process. The 22Ne(alpha,n)25Mg reaction is thought measurement of neutron-induced charged particle reac- to be the neutron source driving the synthesis of nu- tions. The LENZ (Low Energy n,z [3]) instrument is de- clides in the A = 60–90 mass range during the s process. signed to precisely measure (n,p) and (n,alpha) reaction Although many direct and indirect measurements were cross sections, using a time-of-flight method on the wide performed, the reaction rate still needs to be better incident neutron energy range from thermal to several determined due to poor resolution and limited sensitiv- tens of MeV at LANSCE. ity. With the WNR/LANSCE neutron beam upgrade and For better performance of detecting (n,alpha) cross the advanced charged particle detection techniques, the sections for the energy range of astrophysical interest goal of this research was to measure the time-inverse (below 3 MeV in neutron energy), we have enhanced reaction as 25Mg(n,alpha)22Ne to answer the heavy ele- the detection coverage and lowered the alpha detection ment production puzzle [1]. As a commissioning of the threshold. LANSCE provides the continuous-in-energy detection system, the detectors were calibrated using neutrons from thermal to several hundreds MeVs using an alpha-emitting radioactive source- 229Th and the in- a spallation tungsten target; therefore it is necessary to beam measurement has been performed for 59Co(n,p) identify different charged particles emitted from in-beam and 59Co(n,alpha) reactions. reactions. The collimators of the 0.71” opening size were inserted in the beam line in order to provide the beam Background and Research Objectives spot of interest for the LENZ chamber. The primary goal of this project is to extend our capabil- ity to measure neutron-induced reactions that produce The LENZ chamber (shown in Figure 1, 2, and 3) consist- and to further knowledge of these reaction that are a ed of a Twin Frisch-grid Ionization Chamber as a “delta- key to understanding of s-process nucleosynthesis. Due E” detector (partial energy deposit) and a double-sided to the small reaction cross sections associated with silicon strip detector (MICRON semiconductor prod- neutron-induced charged particle reactions because of uct [4]) as an “E” detector (full energy deposit). Since large Coulomb barriers and large signal-to-background charged-particle reactions often are suppressed from ratios in detecting low energy alphas [2], building a very a large Coulomb barrier, this instrument is designed to efficient detection system with the advanced technolo- enhance detection coverage. Initially, four different exist- gies to achieve a large solid angle and good angular and ing design concepts were reviewed for consideration; timing resolutions was proposed. Once reliable experi- however, the twin Frisch-grid ionization chamber was mental data are acquired, it was planned to compare the best suited for our purpose. them to verify theoretical predictions. Since heavy ele- ment production involves several thousands of nuclear The dimensions of the ionization volume are set to be reaction rates, the majority of rates heavily depend on 63.5 mm of the anode radius and 37 mm of the distance model predictions that still hold rather large uncertain- between the cathode and anode, in order to cover about ties. Detailed studies on these reactions over a much 120 degrees out of 180 degrees (66% coverage). Out- wider range of neutron energies and for previously 596

Page 597

side of the anode at a backward angle, there is an electric mum energy loss. Target thickness is also carefully deter- shield sheet for vetoing any beam-induced background mined by simulating Monte Carlo energy loss calculations charged particles outside of the ionization volume, which TRIM [9] for minimizing any self-absorption of reaction increases the signal-to-ratio and thereby lowers the detec- alpha particles in a target medium. tion threshold. Figure 1. Diagram of the Frisch-grid twin ionization chamber and the double-sided silicon strip detectors Figure 3. Picture of the interior of the LENZ instrument. The target wheel has 8 target locations, with one-inch openings. It is designed to readily accommodate multiple targets including a calibration target, such as 6LiF, an alpha source, reaction targets, and backing materials for back- ground subtraction, without breaking vacuum system. The ionization gas mixture was chosen to be a P10 gas – 90 % Argon gas and 10 % of CH4 gas, in order to reduce any beam-induced background from the counting gas. Often, neutron beams could create reactions with carbon, hydro- gen, oxygen, and argon gases, so the type of gas is chosen for each reaction to minimize interference with peaks of interest. We used the P10 gas and the pressure of 550 Torr for optimizing the exit energy of low energy alphas and the ionization detector’s performance. The pressure readout is well calibrated according to the manufacturer’s manual. Figure 2. Pictures of the LENZ chamber located at the flight For future experiments, this type and pressure of gas mix- path during the in-beam commissioning runs. ture can be changed to different reactions. For avoiding any alpha particle’s energy degradation, the Since LANSCE provides a white neutron energy spectrum, cathodes, grids, and anodes are made of thin tungsten the beam energy information is deduced from a time of wired grids instead of sheets or foils, so there is very mini- flight at a target, in this case cathode timing. Therefore, 597

Page 598

the timing resolution has to be preserved as well as pos- The in-beam commissioning measurements were com- sible. We used waveform digitizers, which allowed us to pleted on a 59Co target in Nov. 2014 – Jan. 2015. Cobalt store partial wavelet information to obtain the best timing is one of the materials that are used widely in nuclear and energy resolution, via post-processing using various applications [12]. Since the radiation damage of materi- digital filters. For the ionization chamber, the signals were als is strongly related to the accumulation of helium gas fed into the low-noise charge-sensitive preamplifiers with produced by the (n,x-alpha) reaction, it is important to fast rise time (CAEN product [10]), for obtaining the fast evaluate alpha-particle production cross sections. In this timing resolution. For the silicon strip detector, the signals commissioning measurement, we intended to measure were fed into the specially designed charge-sensitive (n,p) and (n,alpha) cross sections in comparison to Hauser- preamplifiers (MESYTEC product [11]) for double-sided Feshbach calculations [13]. The 59Co foil was installed in silicon detectors (DSSDs) whose output was captured by the target wheel; it ran for about three weeks of beam the digitizers directly. We have used x724 digitizers from time at the flight path 15R in the WNR target 4. Figure CAEN with a 14-bit resolution and a 100 MHz sampling 5. shows the particle identification measured from the frequency, for achieving the best pulse height resolution. “delta-E” signals from the gas detector and the “E” signals As shown in Figure 4, the energy resolution of DSSD using from DSSD detector. The proton and alpha groups are well a digitizer was obtained as good as 55 keV for the alpha separated without any beam energy gating. When the energy of 8.17 MeV. This LANL measured resolution is bet- beam energy is gated into the data, the (n,p) and (n,alpha) ter than the specification from the Micron semiconductor reaction thresholds will be confirmed. manufacturer, which was 75 keV. The data analysis is on going and the final results will be presented in the reaction cross sections of (n,p) and (n,alpha) on 59Co. The discussion with the nuclear reac- tion modeler (Kawano at T-2) on how to compare the measured cross sections to his Hauser-Feshbach calcula- tions has been initiated and we expect we can improve the theoretical prediction through these fruitful data on LENZ measurements at LANL. Impact on National Missions The neutron-induced charged particle detection system enhances our capability to provide not only crucial nuclear Figure 4. Double-sided silicon strip detector was calibrated for information in astrophysics, but also high-quality nuclear channel-to-energy and an energy resolution. Pulse height histo- data for nuclear applications. With this new capability, gram was measured with a Th-229 radioactive source. LANL is positioned to lead experimental efforts for detect- ing neutron-induced charged particles using the wide range of neutron energy. Completion of this additional capability of measuring neutron–induced charged particle reactions allowed us to attract multiple funding agencies inside LANL. One of them is the precision measurement on the 16O(n,alpha) cross sections, which has critical impact on nuclear applications. Since water is used as moderator or cooling agent, this cross section has been of great interest for reactor designs and applications. However, the current status of evalua- tion still needs a new independent set of data to validate the normalization among different measurements [14]. LENZ measurement at LANSCE will be a perfect project to provide this invaluable data set to the community. Figure 5. 2-dimensional plot of “delta-E” and “E” for demon- The other impact would be the continuous effort of strating different charged particle identifications improving the quality of nuclear physics input for better design of fusion reactors, nuclear fission reactors, accelera- tor-based advanced reactors, etc [15]. Estimating radiation 598

Page 599

damage of structural materials by measuring neutron- Physical Society of Japan. (Waikoloa, Hawaii, 7-11 Oct. induced hydrogen and helium gas production rates is a 2014). crucial part of new designs, in particular when associated with testing new materials. Therefore, new measurements on these test materials using LENZ at the relevant neutron energies available at LANSCE would be valuable. References

1. M. Jaeger, R. Kunz, A. Mayer, J. Hammer, G. Staudt, K. Kratz, and B. Pfeiffer, Phys. Rev. Lett. 87, 202501 (2001)
2. H.Y. Lee, PhD Thesis, University of Notre Dame (2006), H.Y. Lee, et al., Proceedings of Science, PoS(NIC-IX)082, (2006)
3. R.C. Haight, Proceeding of Int. Conf. on Nuclear Data for Science and Technology, DOI: 10.1051/ nda- ta:07518 (2007)
4. http://www.micronsemiconductor.co.uk
5. N. Patel, The Colorado School of Mines, PhD Thesis (2010)
6. A. Goepfert et al., Nucl. Instr. Meth. A 441, 438 (2000)
7. G. Zhang et al., Nucl. Sci. and Eng. 156, 115 (2007)
8. N. Ito et al., Nucl. Instr. Meth. A 337, 474 (1994)
9. http://www.srim.org
10. http://www.caen.it
11. http://www.mesytec.com
12. S. Kunieda et al., Phys. Rev. C85, 054602 (2012)
13. T. Kawano, P. Talou, M.B. Chadwick and T. Watanabe, J. Nucl. Sci. Tech (2009)
14. G. Giorginis et al., Int. Conf. on Nuclear Data for Sci- ence and Technology, DOI:10.1051/ndata:07481 (2007), OECD discussion, CIELO project
15. INDC(NDS)-358.Status report of the final CRP Meeting, Japan, Ed. A.B. Pashchenko (1995) Publications Lee, H. Y.. Neuctron-induced Reactions Nuclear Astrophys- ics. Invited presentation at Joint DNP Town Meetings on Nuclear Structure and Nuclear Astrophysics. (Texas A&M University, 21-23 Aug. 2014). Lee, H. Y., and R. C. Haight. Neutron-induced charged particle studies at LANSCE. Presented at The 4th Joint Meeting of the APS Division of Nuclear Physics and the 599

Page 600

Nuclear and Particle Futures Early Career Research Final Report Effects and Mitigation of Hot Electrons in Direct Drive Implosions Natalia S. Vinyard 20140575ECR Abstract from LANL, LLNL, and LLE will provide the required ex- The National Igniftion Facilty (NIF) provides a unique pertise. experimental platform to support stockpile stewardship, Background and Research Objectives study HEDP (high energy density physics), and advance our ultimate goal of inertial fusion energy (IFE). With Conducting high-profile cutting-edge experiments on NIF’s current configuration, efficiently utilizing available the NIF dictates that we accurately predict each experi- power is of crucial importance for investigating strong- ment’s performance to quantitatively assess its scientific shock regimes. The laser direct drive (DD) concept, ex- impact. Meanwhile, understanding the nuances of PDD tensively employed by LANL at both the Omega Facility allows us to develop experiments that are better aligned and NIF, can provide a superior experimental platform with the core mission of the Laboratory. For PDD, this by allowing ~7 times more energy to be coupled directly requires detailed knowledge of energy transfer from the to the capsule compared to ID. However, for DD experi- laser to the target as well as heat transport within the ments to be valuable to the Laboratory’s primary mis- target which is comprised of thermal diffusion, dominat- sion, it is crucial to understand the subtleties of DD, such ed by the background material properties, and energy as the impact of kinetic effects, particularly if the model- carried by the hot electrons. ing is done with rad-hydro codes. For example, recent Figure 1. Schematic diagram of targets for N130320(1) defect induced mix experiments (DIME) on NIF show a bright band of self-emission around the capsule equator It is well known that hot electrons play a significant role which was neither replicated in simulations nor seen at in laser-driven inertial fusion [1], both for ID [2] and lower laser intensities (I0<1015 W/cm2). At the same DD [3],[4] configurations. For example, hot electrons time, hard x-rays emission (hot electrons’ signature) oc- produced by Two-Plasmon Decay (TPD) [5], are hypoth- curred predominantly around the equator. Understand- esized to be the cause of a bright band of self-emission ing these laser-driven kinetic effects have the potential observed around the equator in recent DIME implosions to impact heat transport and burn models in Complex on NIF [6]. Moreover, the growth rate of the TPD insta- codes and have consequences for the DD intensity limit. bility, which scales like which scales like LnI/Te, (where I Managing these issues requires a detailed knowledge is the overlapped laser intensity, Ln is the density scale- of hot electron evolution under realistic DD conditions. length and Te is the electron temperature) was in quali- While recently there have been great strides taken in tative agreement with the observed data. However, to understanding supra-thermal electron dynamics, to-date explain these observations it is imperative to understand there has not been a detailed assessment of its impact the physical processes that drive hot electron evolution, nor means of mitigation for polar direct drive (PDD). from generation through transport to deposition under the direct-drive conditions. Therefore, the research objective is to investigate the generation, dynamics and impact of hot electrons on Modeling this chain requires a detailed physical under- the heat transport under PDD with intent to develop a standing of the dominate generation mechanisms, such better heat conduction model and use it to explore pos- as TPD, the resulting hot electron distribution function, sible mitigation schemes. We believe this research will and the details of energy transport that will occur in advance the predictive capabilities of our hydro codes, the target material. The implosion itself happens on a thereby improving the effectiveness of future implosion relatively slow, hydrodynamic time-scale and is modeled physics experiments. Our collaborations with scientists with rad-hydro codes. On the other hand, the part of the 600

Page 601

heat transport, that is associated with the hot electrons The first part of the project, which consisted of two dynamics, requires a much faster kinetic treatment. Since phases, examined the processes that result in hot electron it would be computationally prohibitive to simulate the creation under the conditions relevant to PDD implosions, whole implosion with kinetic codes, certain approxima- similar to the works of Myatt [3], Vu [12] and Yan [13], and tions are usually employed to provide some improvement assessed how global implosion parameters are affecting to the thermal transport accuracy. For example, when the their evolution. temperature gradients become so steep that their charac- The first phase was to identify the relevant global laser teristic scale length, , is surpassed by the electron mean and plasma parameters, such as laser intensity, density, free path, [7], the classical diffusion model for heat trans- pressure and temperature from 2D HYDRA simulations port (where the heat flow is given by with Spitzer–Harm PDD design. Two shots from Defect Induced Mix Experi- and Peltier thermal conductivities [8]) is no longer valid. ment (DIME) [6], N130320 and N130321 were taken as In this case, a constant thermal flux limiter, f, is typically base-line scenarios to obtain the drive and plasma pa- introduced to ensure that the heat flux never exceeds the rameters for the follow up theoretical and PIC code initial free-streaming limit and whose value is set to provide a conditions. These two shots were chosen in part due to reasonable match to the experimental observations (e.g. the plethora of available experimental data, particularly f=0.03 [9]). However, the effects of additional heating, pertaining to observation of the hot electrons. The targets say from a kinetic hot electron population are not taken imploded during these shots had plastic shells of 1100 into account in this simple model. While some rad-hydro μm nominal radius and 42 μm thickness. The inner 2 μm codes (e.g. HYDRA [10]) have an external energy source of the shell was doped with 2.1% of Ge, and a 2-μm layer, that can simulate hot electron behavior, it is up to the buried by 5 μm, was doped with 1.9% of Ga for spectro- user to ensure the appropriateness of its spectral, spatial scopic measurements. The capsules were mounted on a and temporal levels. Consequently, simulations have to be glass fill tube and filled with 5 atm of deuterium at shot benchmarked against experimental data to ensure the cor- time. The capsules were driven by all 192 NIF beams in a rect flux limiter and hot electron behavior are used before PDD configuration and delivered 319kJ and 462kJ in about one can have confidence in the modeling results. Our plan 2ns for N130320 and N130321, respectively. From prior is to use TPD theoretical models, combined with kinetic experiments on NIF and Omega, it was known that uniform VPIC [11] and rad-hydro HYDRA simulations, to enhance laser illumination, even with highly offset equatorial beams the understanding of hot electron dynamics and their resulted in oblate implosion. To compensate for this, the PI effects on thermal transport during DD implosions. This developed a Cone Power Tuning illumination scheme that work will have impact on burn and transport models in our resulted in a very symmetric image for N130320. This was radiation hydrodynamics codes as well as have possible achieved by increasing the energy in the two equatorial implication on DD intensity limit. Our scientific collabora- beams, while simultaneously decreasing the energy in the tors have been chosen based on their expertise in PDD two polar beams relative to the nominal setting. experiments, kinetic & TPD modeling and PDD rad-hydro modeling to maximize the probability of success for this Performing simulations of these two shots we were able to proposed work. obtain the drive and plasma parameters. These were ob- tained from simulations of N130320 DIME’s NIF shot. The Scientific Approach and Accomplishments experimental neutron bang time for this shot was 3.8±0.2 Our approach was to improve on a simulation technique ns while the neutron yield was 7.2±0.2e+11, compared to by J. F. Myatt, et.al of LLE [3] by investigating the impact simulated values of 4.19 ns and 6.14e+11, respectively. We of the TPD-generated hot electron heating on the dynam- plot simulated plasma and laser parameters at three dif- ics of the capsule implosion and incorporating 2D effects ferent times: at the beginning of the pulse when the laser which are crucial under PDD conditions. This project had has reached its maximum power (t=0.15ns), in the middle two main components. The objective of the first part was of the pulse (t=1.20ns) and at the end of the pulse right to study the creation and evolution of the hot electron before the laser is shut off (t=2.15ns). population, particularly for realistic 2D geometry that has not been deeply explored. The second part of this work N130320 simulated spatial profiles of the ions density, explored new scientific ground by determining how the electron temperature, ion temperature, electron number presence of hot electrons influences the global parameters density, and the laser energy absorbed per unit volume at of the implosion. three different times: at the beginning of the pulse when the laser has reached its maximum power (t=0.15ns), in Part 1: Production, evolution, and possible mitigation the middle of the pulse (t=1.20ns) and at the end of the mechanisms of hot electrons pulse right before the laser is shut off (t=2.15ns). 601

Page 602

In the second phase, we examined the implosion condi- transport is important for our ICF codes. This may clarify tions at a few polar points around the capsule, such as the some aspects of the current indirect-drive target design, as equator and pole and extracted 1D radial profiles to be well as improve the PDD ignition design. It is well known used as the initial value input parameters to VPIC simula- that the most serious impediment to PDD ignition is the tions. production and transport of hot electrons arising from the TPD instability. The PDD platform is important for LANL and N130320 simulated equatorial and polar spatial profiles of the National ICF Program because it may provide a high- the ions density, electron temperature, ion temperature, yield, more robust platform than the current ID (indirect electron number density, and the laser energy absorbed drive) platform.” per unit volume at three different times: at the beginning of the pulse when the laser has reached its maximum References power (t=0.15ns), in the middle of the pulse (t=1.20ns)

1. J. D. Lindl, Phys. Plasmas 2, 3933 (1995) and at the end of the pulse right before the laser is shut off (t=2.15ns). 2. L. Yin, et al. Phys. Plasmas 20, 012702 (2013) We can see that while some of the plasma parameters are 3. J. F. Myatt et al. Phys. Plasmas 19, 022707 (2012) very similar at the pole of the capsule as at the equator, others can differ by almost a factor of two. 4. D. Shvarts, et al. Journal of Physics: Conference Series 112 (2008) 022005 Part 2: Effects of hot electrons on direct-drive implosions The second part of the project investigated the effects of 5. B. Afeyan et al. Phys Rev. Lett. 75 (1995) hot electrons on DD implosions. Their main impact comes
2. M. Schmitt et al. Phys. Plasmas 20, 056310 (2013) in the form of preheat or deposited thermal energy into the shell and fuel, which, even in small quantities (<10%),
3. D. R. Gray and J. D. Kilkenny Plasma Phys. 22 81 (1980) can severely restrain the compression and degrade implo- sion performance. While we did not have time to input the 8. L. Spitzer and R. Harm Phys. Rev. 89 977 (1953) obtained profiles into VPIC simulations, we did perform preliminary assessment of the hot electrons effects [14]. 9. M. D. Rosen Comments Plasma Phys. Controlled Fusion This was done with the non-local heat transport (NLET) 8 165 (1984) model [15] in HYDRA, which takes into the account longer
4. M. Marinak, R. Tipton, O. Landen, et. al, Phys. Plasmas mean free paths of hot electrons. 3, 2070 (1996) Due to hotter temperatures and larger temperature gradi-
5. K. Bowers, B. Albright, L. Yin, B. Bergen, and T. Kwan, ents (by a factor of 1.8 at the beginning of the laser pulse) Phys. Plasmas 15, 055703 (2008) the hot electrons penetrate the capsule and deposit their energy deeper in the equatorial region than in the polar. As
6. H. X. Vu et al. Phys. Plasmas 17, 072701 (2010); Phys. a result, the implosion exhibits a highly prolate symmetry, Plasmas 19, 102703 (2012); Phys. Plasmas 19, 102708 relative to the flux-limited case. This is a stark manifesta- (2012) tion of hot electron effects on the capsule implosion.
7. R. Yan et al. PRL 103, 175002 (2009) Impact on National Missions This work used theoretical and computational approaches 14. T. J. Murphy, N. S. Krasheninnikova,G. A. Kyrala,et al. to advance our understanding of heat transport during po- Phys. Plasmas, submitted 2014) lar direct drive implosions. The accuracy of complex codes,
8. G.P. Schurtz, P.D. Nicolai and M. Busquet, Phys. Plas- i.e. HYDRA and RAGE, was improved, thereby enhancing its mas 7 4238 (2000) design capability for future laser-driven validation experi- ments. Moreover, this work can also impact stockpile stewardship. With ongoing science campaign experiments on OMEGA at LLE, in addition to planned direct drive igni- tion experiments on NIF within the next five years, it is crucial to address this issue now so our research results can be incorporated into these design efforts. Dr. Steven Batha, LANL ICF Program Manager, supports this work. He says, “Better understanding and modeling of hot electron 602

Page 603

Nuclear and Particle Futures Exploratory Research Continuing Project Hybrid Shock Ignition as an Alternate Concept for Fusion Energy Eric N. Loomis 20140180ER Introduction Benefit to National Security Missions Achieving energy gain from thermonuclear reacting Alternate Inertial Confinement Fusion (ICF) concepts, plasmas in a laboratory setting is a fundamental aim of such as Shock Ignition (SI), will stress different aspects of researchers around the world since, in time, it would the ignition problem compared to conventional indirect- lead to a new and controllable energy source, as well drive and will help elucidate shortcomings in our predic- as a much needed platform for supporting aspects of tive capabilities. This platform is also a potential low con- Stockpile Stewardship. Inertial Confinement Fusion (ICF) vergence design that at ignition-scale could out-perform has remained at the forefront of this research for many current NIF implosions or at least be more predictable. years and its flagship, the National Ignition Facility (NIF), This work addresses a major gap: LANL is currently the is currently performing the experiments towards ignition only major ICF laboratory without an investment in SI anticipated for well over a decade. In the past year, how- and, furthermore, LANL is the only ICF laboratory with- ever, thermonuclear ignition on NIF using conventional out an investment in any alternative ignition platform. indirect-drive ICF has proven more elusive than previ- The intent of this project is to begin to build an innova- ously expected. Even if ignition is achieved, new high tive Hybrid SI program with potential experiments on yield concepts are needed for useful energy production NIF within four years. as well as future “Applications of Ignition” experiments. In this project we present a novel platform with the po- Progress tential for high energy gain implosions, which integrates In FY15 we completed 2-D design studies of our OMEGA the existing advances made in Indirect-Drive Inertial scale targets, target fabrication of complex geometry Fusion with the promising new scheme of Shock Ignition gold hohlraums for experiments originally planned in (SI). If successful, our Hybrid SI concept will provide a FY16, began investigations into optimized laser pulse new platform by which to study hohlraum-based implo- shapes for hybrid-driven implosions, and discovered the sion physics, but which accesses different conditions potential detrimental effects of perturbations on con- than current NIF ignition attempts and stresses different verging shocks for shock ignition. Since efficient transfer aspects of our predictive capabilities, which we expect of energy during the collision of the ignitor shock with will lead to improved physics models in numerical radia- the rebounding initial compression shock is essential for tion hydrodynamics codes. reaching optimal conditions for thermonuclear ignition, this last accomplishment will likely prove to be of utmost Hybrid SI will use spherical hohlraums with a novel ar- importance. For this reason we plan to have it as our rangement of symmetric laser entrance holes that allow central focus for the remainder of FY15 and into FY16. both indirect-drive beam access to the hohlraum walls followed by shock ignitor drive beams that directly illu- This issue revealed itself as we were studying the results minate and provide requisite heating to the compressed of 2-D target simulations. These simulations studied the fuel. Our exploratory research into this concept will in- implosion physics of x-ray driven capsules with direct- volve multi-dimensional radiation hydrodynamics design drive ignitor shocks. Capsule diameters were main- calculations that demonstrate adequate implosion and tained at 550 microns and contained deuterium-tritium ignitor symmetry. These calculations will then be used to gas. Capsules had either 50 or 25 micron thickness study new physics associated with instabilities generated and 50 or 25 atmospheres of DT, respectively. Thinner by collisions of converging rippled shocks. capsules were expected to give greater neutron yield at the expense of diminished hydrodynamic stability. 603

Page 604

Low-mode drive asymmetry on the capsule was imposed decided to start exploring the effect of gas modulations on using spherical harmonic expansion of the radiation flux shocks in spherical geometry. Our first method for study- from the hohlraum wall plasma. Radiation from the ing this effect was to use very thin (i.e., 5 micron) capsules hohlraum wall emission to the capsule was then trans- known as Exploding Pushers because these types of targets ported using multi-radiation group implicit Monte Carlo due not suffer from deceleration Rayleigh-Taylor growth photonics. Finite statistics of this radiation field imposed and may be a good platform for studying them experi- high spatial frequency roughness at the ablation surface mentally. In Hydra we used radiation flux asymmetries to of the capsule; seeding ablative Rayleigh-Taylor perturba- change the gas shock conditions, but can now also place tions similar to what is imposed on real targets with finite single mode modulations at the inner surface. surface roughness. The combination of this high frequency roughness and the low mode drive asymmetry allowed Future Work us to study near experimental conditions in two dimen- We were not awarded facility time under the OMEGA laser sions. The hohlraum and laser geometry suggested that Laboratory Basic Sciences (LBS) program so experiments our targets are most likely to have mode 4 drive asym- are not likely to occur in this project as originally planned. metries so these were imposed on both the thick and thin For this reason we will continue to focus our efforts on capsule simulations. Modulations at the inner surface of detailed 2- and 3-D simulations of new physics issues asso- the capsule grew rapidly into the gas at the time of cap- ciated with Shock Ignition that we have discovered in this sule deceleration after feeding through the shell from the LDRD. One issue, specifically, is determining the loss in en- ablation surface. Thinner shells were observed to reach ergy transfer from the ignitor shock to the compressed fuel higher DT ion temperatures and neutron yield, but suffered when the fuel has a modulated density field or when the greater instability growth and mixing during deceleration. ignitor collides with a modulated rebounding shock. The Prior to the formation of the instability growth at capsule modulations in these cases grow much more rapidly due to stagnation, modulations in pressure were observed in the convergence of the ignitor shock reducing the amount of gas due to asymmetries carried by the initial shock conver- shock heating to the fuel. In FY16 we will study this behav- gence. These gas modulations reached their maximum at ior using the LANL Rage code and the LLNL Hydra code by the time of shock bounce at the center of the capsule, but applying inner surface perturbations to our OMEGA scale then reduced somewhat during the bounce and divergence capsules. These perturbations will imprint themselves stage of the shock. There is some theoretical evidence into the fuel, which will produce growing modulations in for this unstable/stable asymmetry transition in spherical the converging ignitor shock. By varying the initial wave- shock problems, but it has not yet been studied in regards length of the inner surface perturbation we can change the to shock ignition, which relies heavily on the strength and spatial and temporal distribution of fuel density modula- stability of converging shocks. tions and thus the amount of shock heating to the fuel as well. This effect should be present in both the direct-drive From 1-D simulations in the first year, we discovered a and hybrid concepts and may set lower limits on required need for tailored pulse shapes that had features well- ignitor shock pressure to overcome the loss in energy suited for the separate stages of fuel compression and transfer efficiency. Part of our FY16 efforts will go into ignitor shock heating. These features need to be contained designing actual experimental platforms to test this effect, within a single, shaped laser pulse due to the OMEGA facil- which we will then propose to the Inertial Confinement Fu- ity restrictions. To accomplish this a pulse was used that sion Science Campaign to support in out years. had a high-intensity short prepulse followed by a roughly 2 ns duration square pulse of moderate intensity. In a real In FY16 a new postdoc will be joining XCP to work on experiment, the beams creating the ignitor shock would be designs at NIF-scale. These designs will be used primarily time such that the high-intensity prepulse would irradiate to study instabilities, such as converging modulated shock the capsule just prior to capsule stagnation, thus maximiz- collisions, at ignition conditions. ing the strength of the ignitor shock. This pulse shape was simulated in a laser-hohlraum calculation using the code Conclusion Lasnex to generate the radiation flux source used in the Success of this project is based on demonstrating the novel capsule simulations discussed above. Under OMEGA laser integration and nuclear performance benefit of a hohl- intensities the hohlraums are predicted to reach about 200 raum-driven implosion with direct-drive Shock Ignition of eV radiation temperatures during the peak of the main 2 the fuel. Specifically, the outstanding questions that must ns pulse stage. be addressed to declare success regards the achievable symmetry of the hohlraum-driven capsule implosion and Following this set of 2-D OMEGA capsule simulations we the symmetry that must be attained by the ignitor beams 604

Page 605

to result in enhanced nuclear performance (neutron yield, ion temperature, and areal density) beyond standard implosion techniques. We will also produce ignition-scale design calculations in order to predict feasibility for NIF. 605

Page 606

Nuclear and Particle Futures Exploratory Research Continuing Project Quantum Kinetics of Neutrinos in the Early Universe and Supernovae Vincenzo Cirigliano 20140252ER Introduction the Nuclear Physics and High Energy Physics Office of Neutrinos are perhaps the most mysterious and elusive Science. of the known particles, and yet play a crucial role in the Moreover, our project will develop cutting edge capa- early universe and the life of stars. Neutrinos interact bility in transport theory. While this will be applied to only very weakly and have tiny masses, the heaviest neutrinos in supernovae and early universe, in the future neutrino being at least a million times lighter than the the very same tools could be applied to programmatic electron, the lightest charged particle. Moreover, obser- work supporting our national security mission. vations of solar, atmospheric, reactor, and accelerator neutrinos indicate that the neutrinos come in three dif- Progress ferent flavors, that can morph into one another as they Theory evolve, through a genuine quantum mechanical interfer- The research team demonstrated that n anisotropic en- ence effect. Despite their elusive nature, neutrinos play vironments a coherent spin-flip term arises in the Quan- a special role in the dynamics of the early universe (EU) tum Kinetic Equations (QKEs) which govern the evolution and supernovae (SN), because they come in huge num- of neutrino flavor and spin in hot and dense media. This bers and because through their flavor-dependent weak term can mediate neutrino-antineutrino transformation interactions they can set the ratio of neutrons to proton for Majorana neutrinos and active-sterile transformation in both the EU and in the heated ejecta of a supernova. for Dirac neutrinos. In an article published in Physics Such ratio is a key ingredient in understanding quantita- Letters B (V. Cirigliano, G. Fuller, A. Vlasenko, “A new spin tively what atomic nuclei are synthesized in these two on neutrino quantum kinetics”, Phys. Lett. B747 (2015) environment. 27-35) we discuss the physical origin of the coherent The overarching goal of this project is to set up the ana- spin-flip term: the spin of a massive neutrino can precess lytic and computational tools needed to describe neu- due to the interaction with an axial-vector field generat- trino kinetics in the EU and SN environments, simultane- ed by other neutrinos or other matter particles, such as ously keeping track of the effect of quantum mechanical electrons and protons, carrying weak charge. This spin morphing and the role of inelastic collisions with the precession does not require the presence of a magnetic medium. The appropriate tool to describe neutrino field nor a neutrino magnetic moments. In the same evolution in a hot and dense medium are the so-called article we have provided explicit expressions for the Quantum Kinetic Equations (QKEs). To date, no self-con- QKEs in a two-flavor model with spherical geometry, the sistent derivation of the QKEs exists, let alone numerical first step towards a computational implementation of solution. Our proposed research will improve on the the new effect. In the context of this two-flavor model, current state-of-the-art in two important aspects: first, we have also demonstrated that coherent neutrino spin it will provide a first-principle derivation of the QKEs, transformation depends on the absolute neutrino mass based on non-equilibrium field theory and a controlled and the so-called Majorana phases, parameters that are expansion in ratios of widely separated length scales. very difficult to measure in laboratory experiments. We obtained the analytic expression for the collision Benefit to National Security Missions term of the QKEs, accounting for neutrino scattering on Understanding how neutrinos have shaped the evolution other neutrinios, electrons, protons, and nucleons. We of the cosmos and of stars, including the implications for have also included neutrino pair production and an- the synthesis of atomic nuclei, is a major goal of both 606

Page 607

nihilation, in summary all the processes that are relevant 2. How do inelastic collisions affect collective flavor in the early universe and supernova environments. We transformation in a supernova envelope? What are are currently putting these results in a form that will be the consequences for the synthesis of heavy elements suitable for computational implementation (e.g. we are in the neutrino-heated supernova ejecta? What are reducing some of the integrals from five-dimensional to the consequences for the neutrino signal that can be two-dimensional, more amenable for an efficient numeri- potentially detected from galactic supernovae? cal calculation). Within FY15 we expect to summarize our results in a publication. Publications Blaschke, D., and V. Cirigliano. Neutrino quantum kinetics: Computation the collision term. 2015. LA-UR-15-25029. On the computational/astrophysical front, we examined whether the newly derived neutrino spin coherence could Cirigliano, V., G. M. Fuller, and A. Vlasenko. A new spin on neutrino quantum kinetics. 2015. Physics Letters B. lead to large-scale coherent neutrino-antineutrino conver- 747: 27. sion. We have done so in a simplified model of a supernova envelope, using the so-called single-angle approximation Vlasenko, A., G. M. Fuller, and V. Cirigliano. Prospects for for the neutrinos emitted by the surface of the compact neutrino-antineutrino transformation in astrophysical object. In a linear analysis we found that neutrino-an- environments . . To appear in arXiv:1406.6724. tineutrino transformation is largely suppressed, but we have demonstrate that nonlinear feedback can enhance Vlasenko, A., V. Cirigliano, and G. M. Fuller. Neutrino quan- tum kinetics. 2014. Physical Review D. 89 (105): 004. it. We have pointed out that conditions which favor this feedback may exist in core collapse supernovae and in binary neutron star mergers. Our results have appeared as a preprint (V. Cirigliano, G. Fuller, A. Vlasenko, “Prospects for neutrino-antineutrino transformations in astrophysi- cal environments”, 1406.6724) under review in Physical Review Letter. Future Work In the upcoming fiscal year, we will finish the analytic calculation of the collision terms, valid both for the early universe isotropic conditions and for anisotropic conditions relevant to compact astrophysical environments. Submit the results for publication in high impact journals Our goals are summarized as follows: • Perform numerical simulations of the full quantum kinetic equations (QKEs) describing the evolution and decoupling of neutrinos in the early universe. We plan to do this both for the case of three active neutrinos and for the case of active-sterile mixing. • Perform exploratory numerical simulations of the QKEs in a spherically symmetric supernova geometry. Conclusion The proposed research will enable us to address two sets of outstanding questions:

1. What is the energy and flavor composition of neutrinos about 1 sec after the Big Bang, when the light nuclei are first synthesized? How does this knowledge con- strain the existence of possible new neutrino states, so called “sterile” neutrinos? 607

Page 608

Nuclear and Particle Futures Exploratory Research Continuing Project Designing the Next Generation Compton Light Source Nikolai Yampolsky 20140269ER Introduction significantly better beam quality over existing ones. This The ultimate goal is to prove the feasibility of the new source will be compact (~100m) and relatively cheap technology for light sources based on inverse Compton (less than $100M), which makes it suitable to be widely scattering using microbunched electron beams. In this distributed to multiple national laboratories and major setup the scattered light will be coherently amplified research universities. The high brightness, tunability, since the radiation generated by each microbunch adds and short duration of these sources will allow for many up in phase. The idea has never been explored since experiments which currently have to be conducted at there was no recognized mechanism for creating very the 3rd and 4th generation light source facilities which short wavelength density modulations in relativistic are severely overbooked. Development of a cheap, com- electrons beams with relatively low energy (~100 MeV). petitive, high-brightness soft X-ray light source will cover We propose to create the beam distribution consisting several national light source needs and will boost many of several well separated energy bands through a series research areas such as Biology, Chemistry, Physics, and of manipulations with the bunch phase space using laser Material Science. modulation and conventional electron beam optics. Such AOT division is currently designing the Injector Test a distribution will drive the two-stream instability result- Stand (ITS) needed for testing novel ideas in Beam ing in the plasma wave, i.e. density modulation. The Physics required for successful commissioning of the proposed scheme can be split into three well separated MaRIE hard X-ray free electron laser (FEL). The ITS can stages, namely (1) creating required beam distribution, potentially be used as a Compton source for enabling (2) development of the two-stream instability caus- early MaRIE related experiments prior to construction ing beam microbunching, and (3) scattering laser pulse of the FEL both for refining MaRIE FEL parameters and off the microbunched beam to produce radiation with for providing additional FEL justification and motivation. increased coherency. We will study each of these stages The capability developed through this proposed work and analyze start-to-end performance of this novel will allow a larger number and breadth of preliminary scheme. We anticipate finding the parameters region for experiments to be done. This source can be used even which the novel scheme results in 5-6 orders of magni- after the commissioning of MaRIE FEL, adding a comple- tude increase in brightness over existing Compton light mentary soft X-ray capability to the facility. sources in 1-10nm wavelength range. Progress The largest challenge of this project is in mitigating vari- ous deleterious effects which may suppress the instabil- Start-to-end simulations have started. The simulations ity leading to beam microbunching. Preliminary analysis include simulation of the electron optics beamline shows that they are negligible at large wavelengths (on required to generate the unstable electron distribution the order of 1 micron) and dominant at small wave- (ELEGANT simulations) and further development of the lengths (on the order of 1 nm). During the course of the multi-stream instability in the following focusing channel project we will identify the precise limit of the novel (CPIC simulations). The analysis of the results is under- technology. way. Benefit to National Security Missions Future Work If successful, the project will lead to development of a • 3D simulations of the instability. new generation soft X-ray Compton light source having 608

Page 609

• Detailed start-to-end simulations using output of EL- EGANT as an input for CPIC. • Include intra-beam scattering into CPIC. • Optimize performance of the scheme resulting in the largest growth of the modulation. • Estimate the limit of the technology in terms of the shortest wavelength modulation which can be pro- duced. Conclusion Perform detailed study of the two-stream instability in relativistic beams driven by the multi-stream distribution. Support conclusions with detailed analytical estimates and numerical simulations including all relevant effects. Determine whether this mechanism for the microbunching instability may be used for improving quality of the Comp- ton light source. Provide reliable quantitative estimates of the resulting source parameters. Determine whether this mechanism for microbunch- ing may be significant in other applications in Accelera- tor Physics, e.g. seeding schemes for free electron lasers (FELs). Publications Yampolsky, N. A., G. L. Delzanno, C. Huang, and D. Shchegolkov. Two-stream Instability at Soft X-ray Wave- lengths for Increasing Brightness of Compton Sources. Presented at 35th International Free-Electron Laser Conference. (New York, NY, 25-30 August, 2013). Yampolsky, N. A., G. L. Delzanno, C. Huang, and D. Shchegolkov. Two-stream Instability at Soft X-ray Wave- lengths for Increasing Brightness of Compton Sources. Presented at North American Particle Accelerator Con- ference. (Pasadena, CA, 29/09/ - 4/10, 2013). Yampolsky, N. A., G. L. Delzanno, C. Huang, and D. Shchegolkov. Two-stream Instability at Soft X-ray Wave- lengths for Increasing Brightness of Compton Sources. Presented at 55th Annual Meeting of the APS Division of Plasma Physics. 609

Page 610

Nuclear and Particle Futures Exploratory Research Continuing Project Combined Klystron and Linac (Klynac) Bruce E. Carlsten 20140351ER Introduction our 50-kV electron gun and the final drawings for our We propose an engineering prototype demonstration Klynac structure. The first half of the klynac structure (Technology Readiness Level (TRL) of 5) of a novel ac- is currently being fabricated in a local machine shop, celerator system architecture, where the accelerator’s RF soon to be followed by the second half. After fabrication, power source is integrated with the accelerator struc- there will be a fine-tuning process where we ensure all ture itself. Using a klystron as the RF power source, we Klynac cavities have the correct resonant frequency. We have named this architecture “Klynac” to represent the anticipate the entire Klynac structure will be delivered functionalities of both the klystron and linear accelerator before the end of the fiscal year. By the end of the fiscal (linac) parts. To quantify its potential impact, the Klynac year, we will also have completed the necessary modifi- technology may lead to a reduction of a factor of 5 to 10 cation to our test beamline where we will experimentally in over-all weight of portable radiography systems when demonstrate our Klynac structure in FY16. These modi- used with an existing LANL TRL 6 capability, the resonant fications include support structures, vacuum pumping, air-core transformer (compared to the 1900-lb, 3-MeV new solenoids with power supplies, and modification of Varian Linatron M3, intended for fixed or truck-mounted the 50-kV CLIA pulse power supply. applications). The reduced weight and size of the Klynac Future Work will allow man-portable radiographic missions, including emergency response, and provide a new technology to All hardware components were received by the end of reduce the cost of medical radiography systems. FY15. Additionally, all modifications to our test beamline were completed at the end of the fiscal year. In FY16, we Benefit to National Security Missions will test operation of this device. The key test, the proof- This project aims to develop technology that can be used of-principle demonstration, will be to measure higher to reduce the size of portable MeV-class radiography sys- final electron beam energy than injected by the electron tems (including those use for medical cancer treatments) gun (50 kV). We will measure the final electron beam from ~ ton weights to a few hundred pounds. The NA-22 energy with an X-ray spectrometer. roadmap “Special Nuclear Materials Movement Detec- Conclusion tion Program Radiation Sensors and Sources Roadmap” calls for high-repetition-rate linacs (1–10 kHz) to increase The main technical goal of this project is to demon- photon flux compared to traditional linacs, allowing strate that a portion of the bunched klystron beam can detection of both prompt and delayed signatures. The be accelerated in a linac. For this demonstration, a low roadmap specifically calls for development of next-gen- voltage resonant klystron and linac will be built, driven eration accelerator concepts and development of com- at 50 kV and accelerating the electron beam to 1 MeV. pact, mobile photon sources, which can be addressed The S-band linac design is standard, and will be identical with the technologies developed by the proposed work. to that used in the previous linacs. Likewise, the electron gun design and the 4-cavity klystron design will follow Progress standard design practices. In FY15 we finished the stability analysis of the Klynac design, showing good operational stability (i.e., it will turn on and reach a stable operating point that is not highly sensitive to beam parameters. We also received 610

Page 611

Nuclear and Particle Futures Exploratory Research Continuing Project Multi-GeV Electron Radiography Frank E. Merrill 20140591ER Introduction Progress This project will investigate the potential for multi-GeV We have identified a location at SLAC to test eRad con- electron radiography to provide high spatial and tem- cepts. We developed the suite of simulation tools to poral resolution measurements of dynamic materials simulate eRad performance and we have identified exist- at the multi-probe diagnostic hall (MPDH) proposed at ing SLAC equipment for these measurements and have the Matter and Radiation in Extremes (MaRIE) facility. designed a system to utilize this equipment. The system This technique has been proposed as one of the major is being installed and first measurements will be collect- diagnostics on the MPDH, but has not yet been experi- ed in late July, fully meeting our second year goals. mentally demonstrated. We plan to design an imaging system to utilize the multi-GeV beams at Stanford Linear Future Work Accelerator Center to test the concepts and perfor- In the third year of this project: mance of high energy electron radiography. Although this type of imaging has been demonstrated with MeV- • The first data will be collected at the end of the sec- GeV protons, it has never been attempted with high ond fiscal year. energy electrons. The demonstration of this measure- • Analyzing the second year’s data, we will assess the ment capability would provide a new window into the performance of the fielded imaging system. dynamic process of opaque materials at unprecedented • From the analysis of this data we will determine if time and length scales. This new window would allow a further refinement of the system is needed. If so, more fundamental understanding of material response we will design and implement these modifications. to extreme environments such as is experienced within a • We will perform a second iteration of measurements functioning nuclear weapon system. at SLAC to determine the performance capability of multi-GeV electron radiography. Benefit to National Security Missions This work will test the concept of using multi-GeV Conclusion electrons for radiography of thin dynamic systems. This This project plans to design, assemble and test the first concept is a major pillar of the proposed Multi-Probe ever Multi-GeV electron radiography system. These Diagnostic Hall at the Matter and Radiation in Extremes measurements will focus on characterizing the perfor- (MaRIE) facility. It is envisioned that these high energy mance of this type of radiography system. This effort electrons will be used to study materials relevant to the will provide important information for the MaRIE project future weapons program, measuring fundamental mate- in determining the electron radiography capabilities as rials properties with very high temporal and spatial reso- well as specifying the requirements for a future facility. lution. The successful development of this technology will support the fundamental material research efforts Publications within the DOE Office of Science and the fundamental Merrill, F.. Imaging with penetrating radiation for the and applied research in the nuclear weapons program study of small dynamic physical processes. 2014. La- for both inert as well as reactive materials such as high ser and Particle Beams. explosives. This capability to measure dynamic material response will also be applicable to DOD programs for military applications. 611

Page 612

Nuclear and Particle Futures Exploratory Research Continuing Project Photocathodes in Extremes: Understanding and Mitigating High Gradient Effects on Semiconductor Cathodes in X-FELs Nathan A. Moody 20140616ER Introduction of high field gradient on semiconductor photocathodes. At the frontier of many scientific disciplines is the goal to Because these effects irreversibly damage the emitted understand, and even control matter, especially when it electron beam, they impact nearly every other aspect is subjected to extreme and dynamic conditions, such as of an X-FEL design and must therefore be rigorously and those encountered in combustion, explosions, or other experimentally understood, quantified, and mitigated excursions in temperature and pressure. The ideal tool before such designs mature beyond the conceptual to probe matter in such extreme conditions is what has stages. This project accomplishes all these goals. become known as an x-ray laser, or x-ray free electron DOE/DOD/SC: all next-generation light sources require a laser (X-FEL). An X-FEL produces pulses of light (coher- high brightness electron beam source. This research will ent hard x-rays) with a wavelength comparable to the answer fundamental questions concerning the behavior distance separating atoms in solid materials and a pulse of cathodes under extreme conditions. The results will duration nearly as short as the changes which occur in include fundamental research validating the basic ap- materials, such as the formation and breakage of chemi- proach to electron beam source design, and demonstra- cal bonds. Consecutive pulses of these hard x-rays would tions of the enabling technology required to successfully allow movies to be made of atoms or molecules, pro- utilize those cathodes in next generation light sources, viding first-ever crucial information about how matter user facilities, or weapons systems. We identify key tech- interacts under extraordinary environmental conditions. nical challenges and solutions which simplify FEL designs Any FEL relies upon a very strictly defined electron beam and/or reduce the commissioning and operatation costs and what stands in the way of building an X-FEL light significantly. source like the one described above is the lack of suf- ficiently bright electron beam. Designs for such a beam Basic Understanding of Materials: x-ray free electron exist but call for electron beam sources (cathodes) with lasers (X-FELs) are the ideal tool to interrogate, under- performance that has never been proven or verified. stand, and even control matter in extremes. To date, These large scale X-FEL instruments are very costly and nineteen Nobel Prizes have been awarded for x-ray sci- time consuming to construct, thus their design should ence using beam-based x-ray light sources, and we can be built on known facts regarding electron beam source expect more in the future as these tools open vast sci- performance and capability. The dominant question ence frontiers to probe matter-in-extremes at unprece- regarding most X-FEL designs relates to the maximum dented temporal, spatial, and energetic scales. The work electric field strength it can be exposed to (and for how represented in this project closes critical technology long) before the material breaks down due to the stress gaps in the fielding of X-FELs as instruments of discovery of the extreme electric field (> 100 MV/m). This project science. answers the question of whether present-day cathodes can survive high electric field gradient and it also pres- Progress ents methodologies for addressing and mitigating this RF cavity fabrication risk. The cavity proper has undergone its last fabrication step, high temperature brazing, and it passed leak rate, Benefit to National Security Missions flatness, surface finish, and dimensional tolerance MaRIE: This project addresses the highest risk element in inspections at the manufacturer on 6/18. The next step Matter-Radiation Interacting in Extremes (MaRIE) x-ray is similar inspections at LANL upon its arrival toward free electron laser (X-FEL) design: the unknown effect 612

Page 613

the end of June. Design considerations, preparation and InGaN semiconducting thin films with bandgaps in the machining of these cavity components, to accommodate 2.0 to 2.5 eV range are being screened with and without the high temperature braze process at a temperature of surface Cs treatment for their photocathode response 1700 Fahrenheit in hydrogen furnace, are similar to those following illumination with 405nm light. Other InGaN films which will be used during eventual fabrication of the will be grown with optimized photocathode properties and MaRIE photoinjector structure. Thus, this LDRD project characterized immediately following growth under ultra- has significantly advanced this critical LANL capability by high vacuum conditions using 405 nm light and a Faraday documenting and demonstrating best practices relating charge collection device recently installed on the LANL EN- to: braze joint design (horizontal versus vertical), compo- ABLE growth system. None of these semiconductor cath- nent modularity, braze alloy composition and temperature ode classes have been investigated at the high electric field dependence, braze form factor (shim versus paste), braze gradients targeted in this project. temperature dependencies (melting, diffusion, and ero- sion), component cooling rates, pre- and post-braze stress Future Work release anneals, braze dependencies on surface finish, and Complete low power RF Cavity Testing mitigation of virtual leaks. Having completed the fabrication of the RF cavity structure in FY15, we will repeat several of last year’s measurements The power coupler assembly was updated and modified but this time as a fully-assembled structure: RF coupling, to reduce fabrication cost and time, to make the coupler cavity resonance, coupling co-efficient, RF joint compatibil- more modular (for ease of later upgrade or, if necessary, ity, initial cathode integration using instrumented metallic repair). This further reduces project risk, enhances por- surrogate, and low power characterization of RF cavity. tability, and improves compatibility with our high-power Milestone: report data on each of the metrics listed above. testing host institution, the Argonne Wakefield Accelerator. High power RF testing We are preparing for fabrication and brazing of the RF Transport assembled RF structure (on strong-back chassis) power coupler assembly, and expect to have all parts of to RF testing station and integrate with the local control the test cell in-house by mid- to late-August. system, diagnostics, and drive laser. Dark current measure- ments using surrogate cathode first. Milestone: report data RF cavity testing on each of the metrics listed above. We are on-track to have essentially all of the test cell parts in-house and assembled and initial low-power RF tests Design and fabrication of movable cathode seal performed by early September 2015. The above tasks use a non-removable cathode for simplic- ity but first introduction of semiconductor cathode and/or Cathode fabrication other metal samples into cavity test cell at high power RF We have demonstrated in FY15 a LANL routine which require a mobile RF joint to allow cathodes to be removed allows us to reproducibly manufacture cesium antimo- and inserted. We will design, fabricate and test this seal, al- nide protocathodes with quantum efficiency (the ratio of lowing us to gain first indications of cathode performance vacuum emitted electrons to surface incident photons) as a function of electric field gradient for both removable of about 2% at 405 nm wavelength on thick optically flat and non-removable cathodes. Milestone: first publication substrates like silicon, glass, and single crystal metals. of results. Having demonstrated the requisite repeatability, we have also completed the required modifications to optimize Begin full cathode characterization cathode stoichiometry which will bring quantum efficiency Begin parametric testing at a variety of field gradients (QE) up to greater than 10%. This complements our earlier (starting with the lowest first), mitigation efforts of field demonstrations of utilizing encapsulated alkali-antimonide breakdown, quantify electric field damage thresholds cathodes in long-term storage. The biggest advantage of as a function of RF pulse duration, test multiple cathode the alkali antimonide photocathode family is very high, up samples involving more than one type of semiconductor to approximately 30%, quantum efficiency at the optimum film, scan surface area of cathode to provide a “map” of wavelength which lies in the visible part of the spectrum. quantum efficiency across the surface. Milestone: second The photocathodes of this class can only be manufactured publication. and operated under UHV conditions, which is why we have developed this LANL capability for the later stages of this Conclusion project. Wide tunable bandgap semiconductors, comprised This research will answer fundamental questions concern- of InGaN are also being investigated for their potential ing the behavior of electron beam sources (namely, upper as efficient and robust photocathode candidates. Several limit of electric field) when subjected to the conditions 613

Page 614

associated with an electron beam based x-ray free electron laser (X-FEL). The results will include both fundamental research validating the basic approach to designing elec- tron beam sources, and/or demonstrating the enabling technology required to successfully utilize those cathodes in a specific X-FEL design. Key technical challenges and solutions, such as cathode seal geometry and high electric field surface treatment, will emerge and the data will allow future X-FEL designs to be based on validated test results. Publications Carlsten, B. E., S. J. Russell, J. W. Lewellen, D. C. Nguyen, P. M. Anisimov, C. E. Buechler, K. A. Bishofberger, L. D. Duffy, F. L. Krawczyk, Q. R. Marksteiner, N. A. Moody, N. Yampolsky, and R. L. Sheffield. MaRIE XFEL phys- ics design risks and risk mitigation plans. 2015. Los Alamos National Lab. (LANL), Los Alamos, NM (United States); DOE Contract Number: AC52-06NA25396; LA- UR—15-22069. Lewellen, J. W., and N. A. Moody. High gradient cathode testing for MaRIE. 2014. In FEL2014. (Basel, Switzer- land). , p. THP024. Basel, Switzerland: JaCOW. Moody, N., H. Yamaguchi, G. Gupta, and A. Mohite. Gra- phene shield-enhancement of photosensitive surfaces and devices. 2014. Micro- and Nanotechnology Sen- sors, Systems, and Applications VI. 9083 (6): 9083331. 614

Page 615

Nuclear and Particle Futures Exploratory Research Continuing Project Superconducting Nuclear Recoil Sensor for Directional Dark Matter Detection Markus P. Hehlen 20150437ER Introduction Benefit to National Security Missions The universe consists of 72% dark energy, 23% dark This work builds new capabilities for the LANL Nuclear matter and only 5% of ordinary matter. One of the great- and Particle Futures Science Pillar, specifically the est challenges facing the scientific community today is Nuclear, Particle, Astrophysics, and Cosmology (NPAC) to understand the nature of dark matter. Dark matter thrust area. Developing advanced detectors that may detection is shaping experimental work in astrophys- enhance our understanding of the physics beyond the ics and particle physics, and it is widely recognized as current standard model is of particular importance and the most important problem in 21st century cosmol- directly ties into national funding agencies such as DOE ogy. Current astrophysics and particle physics models and NSF. In particular, the fundamental nature and the suggest that dark matter is made up of slowly moving, direct detection of dark matter is one of the key focus weakly interacting massive particles (WIMPs). They pass areas of the “Cosmic Frontier” research in DOE’s High through the solar system and should leave signals in de- Energy Physics (HEP) program in the Office of Science as tectors. However, their interaction with ordinary matter well as NSF’s Division of Physics (PHY). The DOE Cosmic is exceedingly weak, making their direct detection in the Frontier in particular focuses on new experimental con- presence of ubiquitous backgrounds a monumental task. cepts and dark matter detectors. Success in this project Detectors that can sense the predicted sidereal variation will allow us to develop new DOE-sponsored programs of the WIMP flux direction hold the greatest promise and will position LANL to assume the leading role in a to unambiguously prove the Galactic origin of WIMPs. next-generation directional dark matter experiment. The The goal of this project is to explore a novel concept of present work is also relevant to the Science of Signatures a solid-state WIMP detector with direction sensitivity. Science Pillar. It builds underlying science and technol- The high density of a solid detector may overcome the ogy in areas of interest to nuclear nonproliferation, background noise limitations of current large-volume treaty verification, and global security where directional gaseous detectors. The proposed detector consists of a detectors can play a critical role. layered structure comprising superconducting traces ap- plied to ultra-thin glass sheets. The anisotropic structure Progress is key to enabling the detector’s direction sensitivity. During an initial phase, we carefully studied various op- WIMP interactions within the glass create nuclear recoils tions of producing a multi-layer device consisting of al- that migrate forward through the glass and deposit their ternating WIMP interaction layers and superconducting energy in the adjacent superconductor. The respective traces. Several device architectures were assessed based heat can cause the superconductor to transition to an on material availability, ease of processing, scalability, electrically resistive state and produce a measurable and expected detector sensitivity. Practical consider- voltage spike. The results of this study will allow us to ations and model calculations informed each other. We formulate a roadmap for the development of the first concluded that the initial approach of stacking a large solid-state WIMP sensor that offers direction sensitivity number of sub-micron thick polymer layers faces signifi- and high density. Such a detector could usher in the next cant layer adhesion and surface roughness challenges. generation of dark-matter experiments looking to sense A new device architecture was developed that consists the unambiguous direction signature of WIMPs for the of superconducting Nb traces patterned onto commer- first time. cial ultra-thin glass sheets. This design (1) is significantly more scalable, (2) maintains the directionality of the de- tector, (3) can be tuned via geometry to a desired recoil 615

Page 616

energy range (enabling means for active background rejec- A new cryogenic test station was designed and fabricated tion), and (4) offers an attractive low-mass target (silica to allow for electrical testing of superconducting devices glass) that is suited for interaction with low-mass WIMPs. between room temperature and 3.8 K using an existing While previous dark matter searches have focused on the closed-cycle helium refrigerator. A sensitive 4-wire resis- ~100 GeV/c^2 WIMP mass scale, the lower WIMP mass tance measurement setup was implemented to allow for region (few GeV/c^2) has gained interest due to recent the automated acquisition of resistance vs. temperature observational and experimental hints as well as motiva- curves. tions from theory. The low mass regime is challenging as Superconductivity was observed at a critical temperature it requires very low (few keV) energy thresholds, a region of 5.9 K in a device containing a meandering trace of 5 µm dominated by large backgrounds that are very difficult to width, 1600 µm length, and 200 nm thickness deposited by discriminate. Due to these challenges, the current limits RF sputtering at 800 oC. Characterization of the other films on dark matter in the <=10 GeV/c^2 region are quite poor. is currently in progress, and the results will allow us to Therefore, the technology pursued by our experiment, downselect both the deposition chamber and the deposi- with its potential low-keV energy sensitivity, could make a tion process. strong impact in this regime even with quite small target masses, particularly if background discrimination/mitiga- Work has started to measure the critical current density tion and directional sensitivity can be achieved. of the Nb traces and to record the fast superconducting- to-normal transients expected upon an interaction. This Work was focused on modeling (Task 1) as well as fabricat- will initially be tested with an Americium-241 alpha source ing and characterizing (Task 2) the first single-layer proto- placed inside the cryostat. Once successful, the setup will type of the new device architecture. The key accomplish- be relocated to the LANL Ion Beam Materials Science Labo- ments are: ratory (IBML) for a systematic exploration of the detector’s We identified two types of ultra-thin (25-30 µm) glass energy and flux dependence. sheets that are commercially available as AF32 and D263 products in high quality and square-meter-scale areas from Future Work Schott Glass. Test samples were obtained from Schott, The work in the second project year (FY16) will focus on and their composition was measured using Secondary Ion assessing the performance of the superconducting nuclear Mass Spectrometry (SIMS). The compositional information recoil sensor (SNRS) (Task 3) and fabricating a multi-layer served as input for the nuclear recoil modeling task. device prototype (Task 4). The device modeling (Task 1) will be refined and continue to guide the design and data Our University of New Mexico collaborators calculated analysis. Specifically, Task 3 will involve the fabrication of the nuclear recoil energy distribution in the thin glass for a range of single-layer devices with varying geometry in interactions with WIMPs of different masses. This data was order to experimentally determine the energy response then used to perform SRIM (Stopping and Range of Ions in of the detector. Initial work will be performed with sealed Matter) calculations to determine the recoil range and thus sources on the existing instrumentation, followed by ex- predict the required superconductor thickness. From the periments on the beamline at IBML. If successful, we will calculated 50-300 nm thickness range we then predicted discover a lower energy threshold where the SNRS begins the width of the Nb trace to be on the order of 2-30 µm in to detect, and upper energy threshold where the SNRS order to offer sensitivity to 5-50 keV nuclear recoils. These latches (no self-recovery from normal to superconducting dimensions are routinely achieved with standard deposi- state), and possibly energy information in the pulse shape tion and photolithographic processes. within the active energy region. This information will be critical (1) to designing the Nb trace geometry to match Nb films were deposited on sapphire test substrates using the desired nuclear recoil energy range and (2) to devel- both RF sputtering and e-beam evaporation at both room oping active and passive background rejection strategies. temperature and 800 oC and using two different LANL These experiments will also demonstrate the directionality sputtering chambers. Superconductive film quality can be of the SNRS. Work will then begin to fabricate a multilayer strongly influenced by deposition parameters. Evaluating (e.g. 10 layers) device prototype. This will involve devel- different deposition techniques allows us to determine oping methods for mechanically stacking the individual the optimal process. The films were analyzed by X-ray dif- patterned ultra-thin glass sheets, establishing electrical fractometry to assess crystallinity and crystal phases. Test connections to each layer, providing means to cooling the structures with Nb traces of various lengths and widths structure below the critical temperature, and reading out were fabricated from these films using standard photo- the signals form all layers in parallel. lithographic processes. 616

Page 617

Conclusion The overall technical goal of the project is to demonstrate the novel concept of a layered glass/superconducting sensor for directional dark matter detection. We will (1) comprehensively model the nuclear recoil properties in the detector materials, (2) fabricate a single layered detector structure, (3) assess the sensitivity and directionality of the device using ion beams, (4) fabricate a multilayer proto- type to assess scalability, and (5) measure the detector background performance. We expect this work to produce numerous high-impact publications and to deliver the quantitative information needed to formulate a roadmap for the future development of a large-scale solid-state detector. 617

Page 618

Nuclear and Particle Futures Exploratory Research Continuing Project Neutrinos and Fundamental Symmetries in Nuclei Stefano Gandolfi 20150476ER Introduction Progress Nuclei are the testing ground for many measurements During the first year of the project we have finalized of neutrino physics and for tests of Physics Beyond the the calculation of electroweak imaginary-time response Standard Model. Experimental probes of neutrino phys- functions of 4He and 12C. In the electromagnetic case, ics and tests of fundamental symmetries have reached we found that the contribution of two-body operators astounding precision with recent measurements of neu- is very large as expected. The results are in excellent trino oscillations and limits on non-standard interactions agreement with the available experimental data, for in beta decays, for example. Commensurate theoretical both the longitudinal and the transverse response. We advances in weak interactions in nuclei are within reach. have also calculated the weak imaginary-time response In this project we will implement for the first time realis- functions, and we are working to extract information on tic treatments of two-nucleon correlations and currents the neutrino-nucleus cross sections. to enable higher-precision studies of neutrino physics at The research team developed the code to calculate the high energy and momenta, and use these same methods charge-changing weak currents. The matrix elements to improve the calculation of nuclear beta decays and for those operators are related to the beta decay. The related probes of physics beyond the Standard Model. subroutines have been incorporated in the variational It is already clear that improved models of the weak Monte Carlo code that we will use for light nuclei We currents are required. Neutrino cross-sections on nuclei are currently testing the code and then we will work to are 20-30% higher than predicted based upon simple optimize the subroutines and then proceed to the calcu- models, it is likely that the energy dependence produced lation in light nuclei. by these models is also incorrect as they only include We have made very important progresses through the simple single-nucleon kinematics. Similar discrepancies extension of Auxiliary Field Diffusion Monte Carlo code between theory and experiment in electron scattering to calculate properties of medium nuclei. The ground have been understood in terms of realistic nuclear in- state of closed shell nuclei, including oxygen and calci- teractions and currents. We will implement these same um, can now be solved for realistic nuclear Hamiltonians models for weak interactions and neutrino scattering to that include two-body forces. We are now working to enable much higher precision nuclear experiments. also include three-body forces in order to reproduce the Benefit to National Security Missions binding energies of nuclei. These are essential to get a reasonable description of the nuclear states involved in This research is important for the DOE Office of Science, the weak transition. nuclear and high energy physics offices. Understanding quantitatively the neutrino interac- Future Work tion with matter is relevant to experiments measuring Our goals for the next fiscal year can be summarize as neutrino oscillations, double beta decay, and beta decay follows: studies of physics beyond the standard model. These • Calculate the imaginary-time responses that are experiments need high precision weak matrix elements relevant for electron and neutrino scattering for a in the nuclei. variety of nuclei up to A=12. 618

Page 619

• Develop a theory to extract the high energy and mo- mentum part of response functions from two-nucleon propagators. • Develop the relevant two-body currents needed to calculate beta decay rates in light nuclei. • Extend existing Auxiliary Field Diffusion Monte Carlo codes to simulate the ground state of medium nuclei. Conclusion At the completion of this project we will have a vastly improved capability to predict weak interaction rates for nuclei, both at the low energy and momentum scales rel- evant for beta decays, the moderate momentum transfer relevant for neutrinoless double beta decay matrix ele- ments, and an accurate two-nucleon model for quasielastic neutrino scattering. Publications Carlson, J., S. Gandolfi, F. Pederiva, S. C. Pieper, R. Schia- villa, K. E. Schmidt, and R. B. Wiringa. Quantum Monte Carlo methods for nuclear physics. 2015. Rev. Mod. Phys.. 87: 1067. Gandolfi, S., A. Gezerlis, and J. Carlson. Neutron Matter from Low to High Density. 2015. Ann. Rev. Nucl. Part. Sci., in press. Lovato, A., S. Gandolfi, J. Carlson, S. C. Pieper, and R. Schia- villa. Electromagnetic and neutral-weak response func- tions of $4mathrmHe$ and $12mathrmC$. 2015. Phys. Rev. C. 91: 062501. Lynn, J. E., I. Tews, J. Carlson, S. Gandolfi, A. Gezerlis, K. E. Schmidt, and A. Schwenk. Chiral Three-Nucleon Inter- actions in Light Nuclei, Neutron-alpha Scattering, and Neutron Matter. 2015. Phys. Rev. Lett., submitted. Tews, I., S. Gandolfi, A. Gezerlis, and A. Schwenk. Quantum Monte Carlo calculations of neutron matter with chiral three-body forces. 2015. Phys. Rev. C, submitted. 619

Page 620

Nuclear and Particle Futures Exploratory Research Continuing Project Assessing the Quantum Physics Impacts on Future X-ray Free-electron lasers Mark J. Schmitt 20150508ER Introduction by the electron’s QM phase uncertainty. Applying this This project consists of theoretical and modeling ef- theory to existing and planned x-ray FELs, the current forts to develop a self-consistent model for the wave- model indicates that QM effects will cause electrons to electron interactions in x-ray free-electron lasers that spread by about 10% of the x-ray wavelength, resulting includes quantum mechanical effects of the electrons. in a reduction in 3rd harmonic gain by a factor of about We conjecture that quantum mechanical spreading of 2 for MaRIE-like parameters. A model for this reduced the electron wave function down the wiggler, left out QM bunching has been added to the GENESIS code (the of current free-electron laser models, will degrade the FEL code being used for the design of the MaRIE XFEL). extraction efficiency of these lasers. By including these Simulation results from the code show that degrada- effects in a new ab initio model, we hope to assess their tion in gain increases with increasing x-ray harmonic impact for various laser configurations (e.g. fundamental wavelength, making operation at harmonic wavelengths and harmonic lasing) and determine the best way to op- more difficult that thought from a classical point of view. timize laser performance in their presence. The results A presentation on the progress of this work is being of this new model will directly affect the design of future prepared for presentation at the 2015 International XFEL facilities and impact the capability for interrogating Free Electron Laser Conference, organized by the Korea extreme states of matter. Atomic Energy Research Institute (KAERI), that will take place in August 2015 in Daejeon, Korea. Benefit to National Security Missions Future Work This work directly impacts the design of the MaRIE 1.0 signature facility by providing an assessment of quantum Work for FY16 can be summarized as follows: effects potentially affecting the extraction of x-rays from • Continue work on 3D quantum theory of XFEL op- energetic electron beams. This work will extend our cur- eration incorporating feedback obtained from the rent understanding of the generation of x-rays from high scientific community. Milestone: Completion of 3D energy electrons relevant to astrophysics. The simulation quantum theory. models developed under this work will enhance the util- • Implement the improved QM theory in a multi-elec- ity of MaRIE 1.0 and its ability to be used to interrogate tron Genesis code in conjunction with an existing materials in extreme states, directly impacting the vali- classical radiation propagation model. Milestone: A dation of physics models used in weapons physics codes numerical simulation code that describes quantum and for defense applications. electrons and classical light in an XFEL. • Conduct theoretical studies of the quantum startup Progress regime where single photon light dynamics is impor- To accelerate the theoretical work of our research, Petr tant and compare results with the case where light Anisimov spent the first three months of the ER in Wako- is treated classically. Construct a fully QM startup shi, Japan working with Dr. Franco Nori at RIKEN on the module to be used in the main simulation code. quantization of relativistic electrons in the magnetic field Completed studies of the quantum startup regime needed for a full quantum (QM) mechanical treatment and QM model. of the FEL interaction. During this time Petr developed a • Submit publications/presentations of the work to QM mathematical formalism to describe the reduction refereed journals and scientific conferences and/or in electron bunching and concomitant x-ray gain caused workshops. 620

Page 621

Conclusion As we construct our model, we will verify its predictions against conventional free-electron laser models in classical regimes, and validate its results against data generated at the LCLS facility. This new model will be used to predict the performance of future free-electron lasers including MaRIE 1.0.The successful execution of this high-leverage research would make a major advance in the XFEL field, placing LANL at the forefront of the design community for future XFELs Publications Anisimov, P. M.. Quantum Nature of Electrons in Classical X-ray FELs. To appear in International Free Electron La- ser Conference - FEL 2015. (Daejeon, Korea, 23-28 Aug, 2015). 621

Page 622

Nuclear and Particle Futures Exploratory Research Continuing Project Transport Properties of Magnetized High-Energy Density Plasmas Jerome O. Daligault 20150520ER Introduction exploration of non-equilibrium conditions in dense plas- There is a rapidly growing interest in exploring the mas and materials under extreme conditions (e,g, radia- application of magnetic fields to inertial confinement tion damage, ion slowing-down, non-equilibrium phase fusion systems in both laser and pulsed-power driven transitions). By enhancing our basic science capabilities, scenarios. The presence of magnetic fields change the this project will help LANL meet its current and future fundamental properties of high energy density plasmas applied missions and challenges, including the science in ways that may bring significant rewards, but also pres- campaigns, Advanced Scientific Computing (ASC) and ent significant challenges for theory. Although decades national High Energy Density Laboratory Plasma (HEDLP) of innovative research has led to a basic understanding programs. This exploratory research has strong ties to of traditional inertial confinement fusion scenarios, the the Nuclear and Particles Futures pillar. combination of high-energy density and strong magnetic Progress fields requires extension of current physics models. In particular, understanding how strong magnetic fields, The research team developed simulation capability to either induced or spontaneous, affect transport proper- account for both Coulomb interactions and magnetic ties is a critical next step toward developing a theoretical fields over a wide range of Coulomb coupling and understanding and modeling capability of these next- magnetic field strengths. To this end, we added Lorentz generation systems. forces in the Newton’s equations for charged particles interacting through the Coulomb interaction in the Combining numerical simulations and analytical model- presence of an external, homogeneous magnetic field. ing, we will enable a first-principles exploration of the In the unmagnetized code developed by the PI in the transport properties of magnetized high energy density past, the particle dynamics is integrated using the Verlet plasmas. We will develop a molecular dynamics simula- algorithm. The challenge was to develop an extension of tion capability to account for both Coulomb interactions the Verlet algorithm than can handle arbitrarily strong and magnetic fields over a wide range of Coulomb cou- magnetic field. We first implemented existing velocity- pling and magnetic field strengths. This unique tool will dependent algorithms in the literature, including the be used to develop a fundamental physics understanding Störmer-Verlet and the Boris algorithms, which are as to how the combined effects of many-body interac- symplectic, stable, and arbitrarily accurate. However tions and magnetic fields affect transport properties, and our numerical tests revealed that these algorithms were to validate new models of transport properties. We will constrained to very small time steps (because the time augment our recent analytic transport theory for unmag- step must be small compared to the Larmor cyclotron netized plasmas to include magnetic fields. Expressions time, and therefore to short simulation time scales. This for the transport coefficients will be cast in a form that is limitation prevents systematic studies over wide range convenient for others to implement into the integrated of magnetic field strengths. We decided to design a new simulation codes used to model ICF plasmas. algorithm, where, unlike Störmer-Verlet and Boris, the choice of the time step is entirely independent of the Benefit to National Security Missions strength of the magnetic field. The new scheme has The results of this effort will introduce a new and unique been parallelized and incorporated successfully into our simulation capability in modeling the influence of code. magnetic fields on high-energy density fusion plasmas at LANL. The capability will allow significant scope for We have begun a detailed study of the effect of magnet- 622

Page 623

ic fields on the transport properties of high-energy density account for magnetic fields, and apply it to explore a wide plasmas. After a long period of trial and error to deter- range of Coulomb coupling and magnetic field strengths. mine the optimal simulation parameters, we have calculat- Two kinds of simulations will be done to explore the ed for the first time the self-diffusion coefficients as well as impact of magnetic fields: 1) We will develop our code to the parallel-to-perpendicular temperature relaxation rates incorporate arbitrarily strong external magnetic fields. This in one-component plasmas and in electron-ion plasmas, requires adding velocity-dependent Lorentz forces in addi- over a wide range of Coulomb couplings and magnetic field tion to the Coulomb forces in Newton’s equations describ- strengths. The analysis of the collected data and compari- ing the particle evolution. 2) We will develop the effective son with existing theoretical predictions are underway. potential to account for screened ion-ion interactions and Preliminary investigations reveal the limit of validity of the then implement those effective potentials in our MD code existing models and motivate the need for new theories. A and calculate ionic transport properties. paper on these results is in preparation. Conclusion We have derived and implemented the basic equations to The primary result will be a microscopic-level understand- include a magnetic field in our effective potential theory ing of the transport properties of magnetized plasmas described in the proposal. Results from the effective across a broad range of coupling and magnetization potential theory will soon be compared to the molecular strengths. Tangible products will be a numerical code and dynamics simulation results. theory capable of describing these plasmas, and practical formulas that can be implemented into integrated simula- Our efforts this past year have been focused on the devel- tion codes used to model ICF systems. opment of the new simulation capability, its testing, and the extension of the effective potential theory. We plan on Publications publishing next year our numerical, analytical and physics Baalrud, S., and J. Daligault. Modified Enskog kinetic theory results. for strongly coupled plasma. 2015. Phys. Rev. E. 91: 063107. Future Work The objective of this research is to develop a quantita- Daligault, J., and S. B. Baalrud. Plasma transport theory tive understanding of the dominant transport processes spanning weak to strong coupling. 2015. AIP Conf. Proc. . 1168: 040002. in magnetized high-energy density plasmas; including temperature relaxation, diffusion, electrical and thermal Sjostrom, T., and J. Daligault. Ionic and electronic transport conductivity and viscosity. Models of transport proper- properties in dense plasmas by orbital-free density ties will be developed and validated using molecular functional theory. Physical Review E Accessible at dynamics (MD) simulations. Our previous MD simulations http://arxiv.org/abs/1510.00647. and theoretical models have proven successful at captur- ing each of these transport properties in unmagnetized plasmas. The proposed work is a non-trivial extension to include an external magnetic field, which will contribute to individual particle equations of motion. Interaction forces will remain electrostatic. Extensive MD simulations will be carried out to unravel the microscopic particle dynamics in magnetized plasmas, and to develop a library of transport processes (diffusion, viscosity, resistivity, energy relaxation rate, dynamic structure factor, etc.) that will be used to validate the analytical models. Currently, our MD code allows for the simulation of classical systems of charges interacting through spherically symmetric potentials in a three-dimensional periodic domain containing up to sev- eral hundreds of thousands of particles and over long time scales. Currently, magnetic fields are not included. The inter-particle forces are obtained using the state-of-the-art particle-particle particle-mesh method, which combines high-resolution of close encounters and rapid, long-range force calculations. We will extend the present capability to 623

Page 624

Nuclear and Particle Futures Exploratory Research Continuing Project Magnetic Rayleigh-Taylor Instability Daniel Livescu 20150568ER Introduction of the mix in ICF. In addition, the study will also address While the goal of achieving economically viable con- several astrophysical phenomena where the combined trolled fusion remains remote in either the inertially MHD-Hall effect and RTI occur. (ICF) or magnetically confined configurations, it is Benefit to National Security Missions generally believed that one of the main obstacles is the development of hydrodynamic instabilities. However, the Since one of the major sources of mix in Inertial Con- complex physics associated with ICF makes such insta- finement Fusion (ICF) is Rayleigh-Taylor instability (RTI) bilities significantly more complicated than the classical development, controlling or suppressing this instability Rayleigh-Taylor instability (RTI). The influence of mag- could have major implications in achieving ignition. In netic fields, either self-generated or externally applied, order to investigate this possibility, we will provide the has recently been an active research topic in the hope first comprehensive studies using accurate numerical it may offer a way of controlling or inhibiting RTI and simulations of magneto-hydrodynamics (MHD)-Hall RTI mixing in ICF. However, so far numerical investigations of with realistic plasma transport properties. The work will this problem have been restricted to 2D, without realistic consider the role of external magnetic fields as well as plasma transport properties, and the few existent stud- various other effects (e.g. Biermann battery effect) on ies investigating the coupling between magnetic fields the development of RTI and electron thermal conductivi- and RTI seem to arrive at opposite conclusions. ty, at the conditions during the deceleration phase in ICF. In order to fill the gap in our knowledge of the coupling The results of the work could easily translate into future between RTI and magnetic fields, we will conduct a programmatic efforts. Thus, besides the importance of thorough study of the RTI in the variable-density (VD) answering the central questions of the proposal, the magneto-hydrodynamics (MHD)-Hall limit using the first mixing and turbulence characteristics we will identify high-resolution 3D numerical experiments of plasma may help address the physics of mix in ICF. For example, mixing with realistic plasma transport coefficients. In a recent DOE review regarding ICF concluded that the particular, we are aiming to address the following ques- modeling of mix is likely inadequate and recommended tions: “a program of scientific experiments and modeling focused on understanding the various physical effects, • Could a magnetic field suppress the late time growth in isolation, that impact the integrated implosion experi- of the RTI and, if yes, under what conditions and ments provides the best approach to eventually either configurations? achieve ignition, or to understand definitively why it may • Could a robust lower limit on mixing exist in the Hall- not be achievable.” In addition, the extensive database magnetized plasma, similar to the bounds obtained we will generate of accurate simulations could be used earlier in the hydrodynamic limit? The existence of to test and develop turbulence models and to test coarse such a lower bound could impact the feasibility of mesh simulation codes of programmatic relevance. To obtaining ignition. this end, we are in contact with project leaders within several campaigns and the PEM program. By examining a variety of configurations, we will not only answer the question regarding RTI growth in the pres- Progress ence of the MHD-Hall effect with realistic plasma trans- The project goals are to examine a variety of configura- port properties, but also contribute to the understanding tions, under late time ICF conditions, to understand the 624

Page 625

role of the combined buoyancy, magnetic, and Hall effects the second year we will focus on the following goals: on the Rayleigh-Taylor instability growth, turbulence, and Theory mixing. • Continue the analysis of the spectral range locality of Work started by writing the full set of compressible equa- the MHD-Hall equations, for the triply periodic case, tions describing the flow in a magnetized plasma: conti- following previous analysis of the pure MHD case. nuity, ion and electron momentum and energy, transport • Examine the linear behavior of the MHD-Hall equa- equations and Faraday’s law. The heat and mass fluxes tions derived. were derived, for the first time, using Onsager’s relations • Examine the structure of the MHD-Hall RT layer, in for irreversible processes and Onsager’s symmetry prin- particular the verify the existence of an inner turbulent ciple. The variable density density incompressible limit region and derive the scaling for the turbulent non- of these equations was derived assuming speed of sound turbulent interface at the edges of the layer. much larger than any characteristic velocity, which is rel- evant to late time ICF conditions. The resulting equations Computation have been implemented in the CFDNS code in both the • Continue with the low resolution simulations to scope Rayleigh-Taylor and homogeneous Rayleigh-Taylor configu- the parameter space, for the pure MHD and MHD-Hall rations. Preliminary simulations have been started in both cases in two configurations: triply periodic and classical configurations. While the Rayleigh-Taylor set-up is directly RT. relevant to ICF, the homogeneous set-up allows us to inves- • Start medium to high resolution simulations for the tigate the mixing problem under the ICF conditions, with- nominal cases. out the complications due to the presence of the edges. • Start analyzing the data. In particular, examine the role of the increase of plasma viscosity with tempera- In a related effort, we are finalizing a linear analysis of the ture on the turbulent evolution and/or suppression. compressible viscous Rayleigh-Taylor instability, with a background temperature gradient. The presence of a tem- perature gradient was never considered in such an analy- Conclusion sis, yet in many problems (e.g. ICF or solar corona), there The central questions of the proposal, which we expect to are strong temperature gradients. The results also address answer, are related to the late time Rayleigh-Taylor (RTI) a recent controversy on the role of large plasma viscosity instability growth suppression using magnetic fields and on the development of turbulence under ICF conditions. the existence of a lower bound in the mix development. Thus, we show, for the first time, using realistic plasma In seeking to address these questions, we will provide viscosities, what are the likely perturbation wavelengths the first comprehensive studies using accurate numerical damped, in ICF, by the large increase in plasma viscosity simulations of magneto-hydrodynamic (MHD)-Hall RTI with and what is the range of wavelengths least affected. realistic plasma transport properties. We expect, along Future Work the way, to be able to address questions related to mix properties and the possibility of controlling mix in Inertial This is the first detailed study of the combined buoyancy, Confinement Fusion as well as many questions relevant to magnetic, and Hall effects on the Rayleigh-Taylor insta- a multitude of astrophysical configurations. bility (RTI) growth, turbulence, and mixing. No previous study has been performed using realistic plasma transport Publications properties and in 3D. The nominal case for the study ad- Aslangil, D., D. Livescu, and A. Banerjee. Variable density dresses the conditions during the deceleration phase in mixing under variable mean pressure gradient. 2015. Inertial Confinement Fusion (ICF). During this phase, the In 15th European Turbulence Conference. (Delft, The RTI development mixes hot and warm plasmas, which en- Netherlands, 25-28 Aug. 2015). , p. Paper number 122. hances thermal energy loss and produces a larger “warm- Delft: TU Delft. spot” instead of a hot spot. This undesirable effect could be mitigated by reducing the electron thermal conduction Aslangil, D., D. Livescu, and A. Banerjee. Reynolds and coefficient and/or the growth of RTI. As such, we aim to in- Atwood numbers effects on homogeneous Rayleigh- vestigate the possibility of controlling the RTI development Taylor instability. To appear in 68th Annual Meeting of the American Physical Society Division of Fluid Dynam- and electron thermal conductivity at the nominal condi- ics . (Boston, 22-24 Nov. 2015). tions using external magnetic fields. We will also examine the turbulence and mix properties produced under these Pulido, J., D. Livescu, J. Woodring, J. Ahrens, and B. Ha- conditions. In order to be able to perform the study, during mann. Survey and analysis of multi-resolution repre- 625

Page 626

sentation methods using turbulence data. Computers and Fluids. Reckinger, S. J., D. Livescu, and O. V. Vasilyev. Comprehen- sive numerical methodology for Direct Numerical Sim- ulations of the compressible Rayleigh-Taylor instability. To appear in Journal of Computational Physics. Wieland, S. A., D. Livescu, O. V. Vasilyev, and S. J. Reckinger. The compressible Rayleigh-Taylor instability and vortex dynamics in stratified media. To appear in 68th Annual Meeting of the American Physical Society Division of Fluid Dynamics . (Boston, 22-24 Nov. 2015). 626

Page 627

Nuclear and Particle Futures Exploratory Research Continuing Project Enhancing the Long-Baseline Neutrino Experiment Oscillation Sensitivities with Neutron Measurements Keith R. Rielage 20150577ER Introduction several Giga-electron volt energies. At these energies, The matter-anti-matter asymmetry of the universe is neutrons are often emitted in the neutrino interaction one of the key scientific questions of our time. In order carrying away energy. Since the neutrino energy must to explain the asymmetry, a large violation of CP sym- be reconstructed using the outgoing particles, correlat- metry - particle vs. antiparticle behavior - is required. ing neutron interactions in the detector with neutron Of the two known types of matter particles - quarks and kinetic energies is paramount. LANL has a unique leptons - extensive studies of CP violation with quarks high-energy neutron facility that enables these measure- have been made and found to be too small to explain ments. the asymmetry of matter and anti-matter in the uni- The work will lead to a more robust approach to neutri- verse. The discovery of neutrino oscillations gives us a no oscillation analysis that will be employed for searches mechanism to search for CP violation with leptons. The of violation of CP symmetry (particle-anti-particle sym- Long-Baseline Neutrino Experiment (LBNE), endeavors to metry) as well as many other neutrino oscillation scenar- make a comprehensive measurement of neutrino oscil- ios. CP violation in particular may be tied to the origin of lation physics with a focus on measuring CP violation the matter-anti-matter asymmetry of the universe. in the lepton sector for the first time. In addition, LBNE endeavors to over-constrain the leptonic mixing matrix This work contributes to the DOE Office of Science giving unprecedented sensitivity to phenomena such as missions and directly supports the nuclear and particle Non-Standard Interactions. LBNE will measure CP viola- futures pillar at LANL. tion by making detailed studies of oscillation phenom- ena over a broad range of neutrino and anti-neutrino Progress energies. They require an event-by-event reconstruction This three-year project aims to measure neutron interac- of the neutrino energy from the outgoing particles of the tions with argon and interpret the data in the context interaction - including neutrons. In this project, we will of neutrino oscillation analyses relevant to the inter- measure neutron interactions in a liquid argon time- national long-baseline neutrino effort currently being projection chamber (TPC) in a well-understood neutron developed. The international long-baseline effort is beam for the very first time. being planned with the United States as the host coun- try. The effort is currently called the Deep Underground Los Alamos has a unique neutron facility with a well- Neutrino Experiment (DUNE). DUNE will employ an characterized, high-energy neutron beam. We will enormous liquid argon time-projection chamber (TPC) at deploy a large liquid argon TPC in this beam in order to a deep underground location in Western South Dakota. carry out the measurements. The theoretical part of the Our Laboratory Directed Research and Development effort will be focused on determining the sensitivity of a (LDRD) project is designed to measure the response of a variety of neutrino oscillation phenomena to the neu- liquid argon TPC to neutron interactions at well-defined trino and anti-neutrino energy resolutions. neutron kinetic energies up to 800 Mega-electron volts (MeV) using the Los Alamos Neutron Science Center’s Benefit to National Security Missions (LANSCE) Weapons Neutron Research (WNR) facility. For neutrino oscillation experiments using liquid argon time-projection chambers, neutrons are a major chal- Since the beginning of the project in October, 2014, the lenge. Neutrino oscillation experiments employ neu- team has carried out two liquid argon engineering runs trinos in the several hundred Mega-electron volts to 627

Page 628

with the prototype detector for the Cryogenic Apparatus will begin late this FY. The starting point is a detailed anal- for Precision Tests of Argon Interactions with Neutrinos ysis of the energy resolutions of various charged particles (CAPTAIN) program. The prototype, which we call Mini- and neutrons in liquid argon TPCs. The next step is to lay CAPTAIN, consists of 1000 pounds of instrumented liquid out what measurements will be made in the LBNE near argon. In our first test, we demonstrated our ability to neutrino detector. The final step is unifying the informa- achieve the very low electronics noise levels required for tion from the near and far detectors in such a way that we optimal operation. We then spent several months working can vary the resolutions of each piece of information in the on the purification system, conducting a variety of tests. oscillation analysis. We subsequently carried out the second engineering run and demonstrated good argon purity with respect to wa- Conclusion ter. Our challenge now is to demonstrate good purity with We will study neutron interactions on liquid argon in terms respect to oxygen. of their event signatures. We will use these results to test existing simulations of the interactions. We will then Since the second run, we have finished developing a gas- modify those simulations with our high statistics data-set. phase purification and condensing loop that will allow us Ultimately, we will produce an oscillation analysis approach to continuously purify the liquid argon and re-condense for LBNE that includes the impact of neutrons as well as a it. This system will allow us to obtain the optimal purity - requirement to be sensitive to new physics. This analysis especially with respect to oxygen. At the time of this writ- will create the near neutrino detector requirements essen- ing, we have begun our third engineering run and begun tial for making the optimal design choices. operating the recirculation system. We will understand the performance of the recirculation system and determine if further modifications are necessary after this run. A neu- tron run is anticipated in fiscal year, 2016. Several calculations and test effort has gone into the devel- opment of a neutron beam-flux monitoring system. Test data were taken in November and the system is now com- plete. The system will allow us to make flux and spectral measurements even at the low neutron rates we require to run the liquid argon TPC. Future Work During the first year, we have been working to commission the Mini-CAPTAIN detector. We are completing a second engineering run to demonstrate appropriate purity for the experiment. In tandem, we have put in a proposal to the LANSCE PAC to run at the WNR beamline 15R in the next run cycle. We expect to move the detector to the beam area in the beginning of FY16. The deployment and recommissioning of the detector requires significant set-up of cryogenic and electronics equipment. In addition, the detector will be filled with purified liquid argon that will be re-circulated to maintain optimum purity. The purity will be checked to assure high-quality data before taking beam. The detector will also be calibrated before taking beam. Subsequently, the detector must be drained and removed from the beamline. The next experimental task will be to carry out a detailed data analysis as well as continued simulations. The theoretical effort has stalled due to the departure of the T-division co-I. We are redirecting these resources into the commissioning of the detector. Detailed simulations 628

Page 629

Nuclear and Particle Futures Exploratory Research Continuing Project Direct Numerical Simulations of Magnetic Rayleigh-Taylor Instability Daniel Livescu 20150732ER Introduction Benefit to National Security Missions The development of hydrodynamic instabilities in the While the goal of achieving economically viable con- presence of complex plasma physics is one of the main trolled fusion remains remote in either the inertially obstacles in achieving ignition in both magnetic and iner- or magnetically confined configurations, it is generally tial confinement fusion (ICF). Besides the important role believed that one of the main obstacles is the devel- of pure Rayleigh-Taylor instability (RTI) in many practical opment of hydrodynamic instabilities. However, the problems, magnetic RTI is very important for the com- complex physics associated with confined fusion makes pression (or lack thereof) efficiency in magnetic or iner- such instabilities significantly more complicated than the tial confinement fusion. Since the instability can limit the classical Rayleigh-Taylor Instability (RTI). An example is maximum compression achieved in the pinch or other the interaction between the confined plasma and mag- inertial confinement fusion (ICF) implosions, controlling netic fields. The influence of magnetic fields has recently or, at least, understanding it is of critical importance. been an active (experimental + numerical) research topic in the hope it may offer a way of controlling or inhibit- At present, the role of a magnetic field on Rayleigh- ing RTI and mixing or, more generally, understanding Taylor Instability (RTI) growth is not well understood the role on the RTI development. For example, recent and the few existent studies seem to arrive at opposite experiments on the Omega Laser facility resulted in the conclusions. For example, recent numerical studies in compression of an externally imposed magnetic field to 2D conclude that magnetic fields suppress instabilities around 20MG, which is dynamically important compared and mixing while the only published study to explore the to characteristic RTI velocities. problem in 3D concludes that magnetic fields enhance the nonlinear growth of instabilities and mixing. Our re- Numerical investigations of this problem in the context cent simulations show that pure RTI growth is bounded of ICF have been limited to 2-dimensions (2D), and from below by the growth of the inner turbulent region, therefore are unable to capture the full 3-dimensional which is independent of initial conditions or fluid proper- (3D) nature of magnetic RTI. It is well known that mixing ties at high enough Reynolds numbers. There is no sys- and turbulence properties in 2D are completely differ- tematic study examining the combined (magneto-hydro- ent from 3D. We believe that conducting systematic dynamics) MHD-Hall RTI turbulence and it is not known investigations of magnetic field interaction with RTI in if such a lower growth bound exists for MHD-Hall RTI. 3D would allow us to answer definitively whether or not It is also not known what effects the plasma transport magnetic fields can be used to control RTI. properties could have on the growth of RTI. Moreover, This work will also benefit the stabilization of Trinity there are many astrophysical systems which are not well Phase 1 and further develop the CFDNS for efficient use understood, but where these effects likely play a role. on new and emerging architectures in support of our We propose to use Trinity Phase 1 to perform the first national security mission. fully resolved 3D simulations, with realistic plasma transport properties, of the combined buoyancy, mag- Progress netic, and Hall effects on the RTI growth, turbulence, and Our cycle on Trinity Phase 1 has not yet started. In mixing as well as electron thermal conduction, under preparation for the start date in mid July, the code has realistic ICF conditions. been ported to Haswell and small size simulations to test resolutions and parameter re=ranges are under way. 629

Page 630

Future Work This is a project for cycles on Trinity Phase 1 only. The sim- ulations will be performed during the two month window for the stabilization process during the present fiscal year. In order to study the instability growth and mixing charac- teristics due to the combined MHD-Hall RTI, we will con- duct a series of numerical experiments, in the Hall-MHD limit that will cover typical conditions under the decelera- tion phase in ICF. Furthermore, we will compare the results with the MHD and hydrodynamic (B→0) limits to isolate the effects associated with magnetic fields and finite plasma skin depth. We are planning to use realistic plasma transport coefficients and fully resolve dissipation scales. The project will use the OpenMP-MPI hybrid version of the CFDNS code. OpenMP is a library and API for shared mem- ory multiprocessing, and is an industry standard for using threads within a code. It is expected that using MPI every- where with upcoming many-core architectures, such as Trinity, will be too expensive. We plan on testing OpenMP by varying the number of threads used, and seeing if there are any performance increases. We have also made some changes to the code to increase the amount of vectoriza- tion to better utilize the more capable vector units found on Trinity. We will further test the implementation and use profilers such as Intel VTune and VampirTrace to identify potential bottlenecks. Conclusion In order to fill the gap in our knowledge of the coupling between instability development and magnetic fields in Inertial Confinement Fusion, we have proposed to conduct a thorough study of the Rayleigh-Taylor Instability (RTI) in the variable-density (magneto-hydrodynamics) MHD- Hall limit using high-resolution numerical experiments of plasma mixing with realistic plasma transport coefficients. In particular, we are aiming to address the following ques- tions: • Could a magnetic field suppress the late time growth of the RTI? • Could a robust lower limit on mixing exist in the Hall- magnetized plasma, similar to the bounds obtained earlier in the hydrodynamic limit? Publications Srinath, A., R. Miller, and D. Livescu. OpenCL-based tridi- agonal solvers for evaluation of compact finite differ- ences on parallel architectures. 2015. Under prepara- tion for a parallel computing conference. 630

Page 631

Nuclear and Particle Futures Exploratory Research Continuing Project Extreme-Scale Kinetic Plasma Modeling of Turbulence and Mix Using VPIC Brian J. Albright 20150751ER Introduction magnetic reconnection and turbulence processes. In- VPIC is general-purpose, kinetic plasma modeling code deed, multiple active missions have been fielded specifi- that has been applied to and validated on a variety of cally to advance our understanding of space science in problems, including laser-plasma instabilities in inertial these areas, which this project directly impacts. fusion, dense plasma transport, thermonuclear burn, Inertial fusion energy interests are also advanced by laser-driven particle acceleration, magnetic fusion, and this project; not only is the VPIC computer code used to applications in space and astrophysics. We propose to advance laser-plasma interaction science for direct- and apply VPIC to three basic science problems central to indirect-drive inertial fusion energetics, but also plasma- our understanding of turbulence and mix in space and phase mix of shell-fuel interfaces will be directly impact- laboratory plasmas and particle acceleration arising in ed by the atomic mixing studies we will conduct. magnetized plasmas. To conduct this work, we will make several changes to the VPIC code base in order to enable Finally, this work advances meaningfully the scientific its efficient use of the unique features of the Trinity su- community’s ability to harness the power of modern and percomputer. Specifically, we will adapt the VPIC archi- future supercomputers, making it of profound interest tecture to a four-tier structure that will enable exploita- to the information science and technology communities. tion of distributed-memory parallelism, task parallelism, Also, the techniques developed to process large data vol- data parallelism, and Trinity’s specific single-instruction- umes using the Trinity burst buffers may be of interest to multiple-data (SIMD) instructions. We will also examine information science applications involving similarly large the development of novel diagnostics taking advantage data sets. of the Burst Buffer capability Trinity has, a feature of the supercomputer enabling more extensive analysis of high Progress data volumes than ever before possible, a potentially We have begun to assemble our team and issue work groundbreaking achievement. This modification to our assignments for this project. No other progress yet to flagship VPIC kinetic plasma modeling code will position report. the scientific community for effective use of Exaflop/s supercomputers in future years. Future Work We will adapt VPIC to the Trinity architecture in a hier- Benefit to National Security Missions archical way that reflects the physical characteristics of This work will vastly improve the state of the art in plas- the system. As an overview, our approach will rely on ma kinetic modeling, a subject area with direct ties to asynchronous communication between compute nodes, many aspects of basic and applied science and national task-parallel division across the intra-node single ad- security. dress space, data-parallel division within a task, and full vectorization within a thread of execution using AVX-512 The ability to resolve outstanding problems associated intrinsics. with plasma-phase mix ties directly to outstanding prob- lems in high energy density science of direct importance • Test performance of finer-grained MPI parallelism to nuclear weapons programs. than employed previously. Assess performance trad- eoffs when replacing MPI ranks with threads. Moreover, as noted in the research proposal, NASA mis- • Experiment with restricting task parallel threads to sions rely heavily on large-scale simulations modeling 631

Page 632

optimize efficiency of the caches. Organize threads into user configurable groups such that each shares a replication of the field domain. • Develop a hierarchical threading model combining atomic memory operations with stream reductions. • Add support for AVX-512 intrinsics and find a suitable block size for processing particle data on Trinity. Ex- plore new intrinsics available in AVX-512 and evaluate their effects on performance. • Explore strategies for using Trinity high-bandwidth memory. • Explore ways to use the new burst buffer capabilities of Trinity for checkpoint/restart and novel methods of high-data-volume postprocessing of simulation data • Implement higher order particle weighting algorithm to VPIC to improve compute/data ratio • Define key physics problems to be run “at scale” using Trinity during Phases 1 and 2 related to turbulence and mix of plasmas and particle acceleration • Document and present results to the high performance computing and scientific communities. Conclusion At the culmination of this project, we will have made substantial progress toward resolving outstanding physics problems in plasma and high energy density science. Spe- cifically, we will have revealed through large-scale calcula- tions how dense plasmas mix with one another. We also will have advanced our understanding of the physics of magnetic reconnection and the acceleration of cosmic ray particles to very high energy. Another legacy of this project will be the development of our VPIC particle-in-cell kinetic plasma code for modern supercomputers such as Trinity and future multi-core platforms. Publications Huang, C. K., M. D. Meyers, Y. Zeng, S. Yi, and B. J. Albright. On the numerical dispersion and the spectral fidelity of the particle-in-cell method. To appear in 24th International Conference ont he Numerical Simulation of Plasmas . (Golden, Colorado, 12-14 Aug, 2015). Huang, C. K., M. D. Meyers, Y. Zeng, S. Yi, and B. J. Albright. On the numerical dispersion and the spectral fidelity of the particle-in-cell method. To appear in Computer Physics Communications. Nystrom, D.. Progress on optimizing VPIC for LLNL’s Se- quoia Platform. To appear in 24th International Conference on the Numerical Simulation of Plasmas . (Golden, Colo- rado, 12-14 Aug, 2015). 632

Page 633

Nuclear and Particle Futures Exploratory Research Final Report Ultra-Bright Electron Beam Acceleration in Dielectric Wake Accelerators Evgenya I. Simakov 20130463ER Abstract timescales from years to decades, and indeed centuries, Debris from historic nuclear weapons test rests on the after the detonation. We borrow a key insight in how to ground surface and within tunnels and underground accomplish this task from similar approaches that have cavities at all nuclear testing sites across the globe. This been successfully used in the field of cosmochemistry to material could be used to determine device design and measure extinct radionuclides in extraterrestrial samples performance but is currently of limited diagnostic value like meteorites. Chemical fractionation between Pu and because the short-lived fission products and key nuclear U that is known to occur during a nuclear event will activation species have long decayed away. The goal of cause systematic isotopic perturbation in 237Np and this project is to develop new measurement and assess- 239, 240Pu concentrations resulting from decay of short- ment techniques necessary to reconstruct the diagnostic lived U isotopes 237, 239, 240U. The proposed sampling information still present in solid debris as the stable and measurement techniques will be used to deter- decay end-members. The original presence of short-lived mine diagnostically sensitive and now extinct 237, 239, radionuclides alters the ultimate isotopic composition 240U isotopes in historic debris samples. A significant measured in historic nuclear debris in a pattern that challenge of the proposed research will be to develop can be interpreted for detailed forensic and diagnostic indirect methods to measure peak-yield fission products evaluation. The success of the proposed effort will be a to determine total fissions in a given sample. In order critical step towards a new capability with direct impact to reconstruct this information we propose to measure on nuclear forensics and on science based stockpile perturbations in stable element isotopic composition to stewardship. The duration of this Exploratory Research infer the original presence of short-lived fission prod- project was one year. Based on the success of this initial ucts. This idea depends on the presence of stable multi- “measurements focused” effort, a larger scale Directed isotope elements for which one isotope is the end-point Research project was begun under the leadership of of a fission decay chain and another is blocked from Hugh D. Selby in 2016. The title of the follow-on project fission decay by a different stable element. (20160011DR) is “Using extinct radionuclides for radio- Scientific Approach and Accomplishments chemical diagnostics.” This research aims to develop analytical techniques to Background and Research Objectives characterize the details of a nuclear detonation based on A nuclear explosion produces an intense burst of energy chemical analysis of aged nuclear debris. The analytes of and associated radiation driven by the violent assembly interest are short-lived radionuclides that have decayed of a critical mass of fissioning nuclear material. Details of beyond detection, but leave signatures of their original the nuclear event have traditionally been inferred from presence in the isotopic composition in archived debris measurements made during and immediately after the samples. We term these signatures, extinct radionu- detonation. In the case of radiochemical measurements, clides. The goals of the project are to recover two types many of the nuclear products that have been shown to of information, (1) short-lived uranium isotopes (237, be of most diagnostic value exist for but a moment in 239, 240U) and (2) fissions. For each of these goals good time, typically on timescales of hours to weeks. The key progress has been accomplished. In order to determine innovation of this proposal is the realization that short- the short-lived uranium isotopes, a collection of samples lived fission products and key nuclear activation prod- must span a suitable range of volatility as measured by ucts that form the basis of modern technical evaluation sample specific 236U/238Pu ratios. Although only one can in fact be indirectly measured over vastly longer unfractionated 236U/238Pu value is actually “true” of 633

Page 634

the device, chemical distortion that occurs during debris determination, a novel technique using high precision iso- condensation serves to generate a range of samples that topic measurements of stable molybdenum is being pur- can be relatively enriched (volatile) or depleted (refractory) sued. A sequential two-column ion-exchange protocol was in uranium compared to a reference 238Pu isotope. The developed and optimized. Purified Mo concentrates have proposed regression technique requires that 236U/238Pu been isolated with good decontamination against potential span a sufficient range within the sample set to allow a ro- interferences from Zr, Ru, Cr, Mn, and Fe. The purification bust linear fit. In the case of analyses of Trinitite, eight in- protocol performed well in an initial Mo analysis of a dis- dividual samples from archived LANL collections have been solved debris sample, which revealed significant deviation completed. The expected trends are observed, however from the natural isotopic composition. An optimized mea- the goodness of fit over a limited range in 236U/238Pu is surement routine using the Neptune Plus multicollector not sufficient to obtain a high-quality measure of 237, 239, ICP-MS is currently being developed. These high precision 240U. A new set of Trinitite samples obtained from a com- measurements will be used to quantitatively determine mercial rock and gem supplier has been procured. These perturbations in 95Mo/96Mo and 97Mo/96Mo isotopic samples appear as light green vesicular samples, distinctly ratios. Future analyses using an isotopically enriched 96Mo different from the black glass found in the LANL collection. spike will be interpreted for absolute fission concentration. It is hypothesized that the light green vesicular samples will be more volatile compared to the black glass samples Impact on National Missions and will provide the range in 236U/238Pu that is needed. A major limitation of current radiochemical diagnostics is Initial analyses of archived core samples from an under- the need to determine the concentration of key signature ground nuclear test have also been completed. For this radionuclides with short half-lives. Within a month after a test, archived samples that transit seven locations through nuclear detonation, these radionuclides are largely ex- the underground environment are available for analytical tinct. If successful, the proposed work will be a critical step study. Duplicate analyses of the first two locations have towards a new capability to extract this crucial information been completed. Five more duplicate analyses will provide from nuclear debris at later times. Such a capability would data to assess the 236U/238Pu volatility index and the fea- have direct and lasting impact on two key DOE/NNSA mis- sibility of the technique to measure 237, 239, 240U in un- sions: stockpile stewardship and post detonation nuclear derground debris samples. Two independent approaches forensics. A demonstrated capability to reanalyze archived are being pursued to determine fissions in historical debris debris samples from historic nuclear tests will improve the samples. These include direct measurements of 99Tc and technical foundation that supports Stockpile Stewardship indirect measurements of 95Zr and 97Zr through isotopic and will strengthen our Nation’s commitment to deter- perturbation of stable molybdenum isotopic ratios (e.g. rence without nuclear testing for decades to come. The 95Mo/96Mo and 97Mo/96Mo). Each of these analytes is limitations of historical measurements impacts nuclear fo- a peak yield fission product that can, in principle, be used rensics as well. Many of our earliest tests used simple de- to determine total fissions in the sample. The first phase signs that could be credibly replicated in a potential terror- of this project has focused on development of chemi- ist nuclear attack. Because many of these test were fielded cal methods used to isolate pure 99Tc and molybdenum in the late 1940’s and early 1950’s, the radiochemical data concentrates that can be assayed using both single and for these events is very sparse and/or of low quality. This multi-collector ICP-MS instrumentation. In the case of 99Tc leads to large uncertainties in evaluation of performance determination, a method has been developed during the and yield, which in turn prevents proper calibration of first year of this project to analyze for 99Tc in aged debris LANL’s computational design tools. A capability to obtain a samples. Technetium is purified in high yields (80-100%) complete radiochemical data set for early U.S. tests would using extraction chromatography (TEVA resin) with good significantly improve modeling and forensic capability envi- decontamination from both Mo and Ru. This decontami- sioned for post-detonation nuclear forensics. nation is crucial to the success of the analysis because of potential isobaric interference from 98MoH+ and 99Ru. Publications Standard addition experiments with solutions of Peruvian Gai, W., C. Jing, A. Kanareikin, C. Li, R. Lindberg, J. Power, soil (NIST SRM 4355) and dissolved Trinitite suggest that D. Shchegolkov, E. Simakov, Y. Sun, C. X. Tang, and A. accurate determinations of 99Tc can be obtained. The as- Zholents. Assesment of Opportunity for a Colinear Dielectric Wakefield Accelerator for a Soft X-Ray FEL sociated measurement uncertainty (20-30%) is higher than Facility. 2014. In 36th International Free Electron Laser ideal and reflects inconsistencies in observed instrument Conference FEL 2014. (Basel, Switzerland, 25-29 Au- count rates that are not yet fully understood. Additional gust, 2014). , p. FRB02. online: JACoW. work is needed to identify the source of the variation and improve measurement precision. In the case of 95, 97Zr Huang, C., T. Kwan, D. Shchegolkov, and E. Simakov. Parti- 634

Page 635

cle-In-Cell Modeling of Dielectric Wakefield Accelera- tor. 2013. In 2013 North American Particle Accelerator Conference. (Pasadena, CA, Sept. 29 - Oct. 4th, 2013). , p. MOPAC23. online: JACoW. Shchegolkov, D. Y., and E. I. Simakov. Design of an emit- tance exchanger for production of special shapes of the electron beam current. 2014. PHYSICAL REVIEW SPECIAL TOPICS-ACCELERATORS AND BEAMS. 17 (4). Shchegolkov, D. Yu., E. I. Simakov, C. Jing, C. Li, and A. A. Zholents. Suppressing Parasitic Effects in a Long Dielec- tric Wakefield Accelerator. 2015. In Advanced Accelera- tor Conference Workshop. (San Jose, CA, 13-18 July, 2014). , p. Still did not publish, ugh!!!. Melville, NY: American Institute of Physics. Shchegolkov, D. Yu., E. I. Simakov, and A. A. Zholents. To- wards a practical multi-meter long dielectric wakefield accelerator: problems and solutions. To appear in IEEE Transactions on Nuclear Science. Shchegolkov, D., E. Simakov, S. Antipov, M. Fedurin, and C. Swinson. Beam Pulse Shaping Experiments for Uni- form High Gradient Dielectric Wakefield Acceleration. 2013. In 2013th North American Particle Accelerator Conference. (Pasadena, CA, Sept.29-Oct.4th, 2013). , p. MOPAC24. online: JACoW. Shchegolkov, D., E. Simakov, S. Antipov, and M. Fedurin. Dielectric wakefield accelerator experiments at ATF. 2015. In 2015 International Particle Accelerator Confer- ence. (Richmond, VA, May 3-8, 2015). , p. WEPJE006. online: JACoW. Shchegolkov, D., E. Simakov, and A. Zholents. Simulation studies of BBU suppression methods and acceptable tolerances in dielectric wakefield accelerators. 2015. In 2015 International Particle Accelerator Confer- ence, Richmond. (Richmond, VA, May 3-8, 2015). , p. WEPJE007. online: JACoW. Simakov, E., C. Huang, T. Kwan, and D. Shchegolkov. Shap- ing Electron Bunches for Ultra-bright Electron Beam Acceleration in Dielectric Loaded Waveguides. 2013. In 2013 North American Particle Accelerator Confer- ence. (Pasadena, CA, Sept. 29 – Oct. 4th, 2013). , p. TUMPA13. online: JACoW. Zholents, A., S. Doran, W. Gai, G. Ha, C. Li, J. Power, R. Lind- berg, N. Strelnikov, Y. Sun, E. Trakhtenberg, C. Jing, A. Kanareykin, D. Shchegolkov, E. Simakov, P. Piot, and S. S. Baturin. A collinear dielectric wakefield accelerator for a soft X-ray FEL facility. To appear in 2nd European Advanced Accelerator Concepts Workshop. (Isola d’Elba, September 13-19). 635

Page 636

Nuclear and Particle Futures Exploratory Research Final Report Beyond the Standard Halo Michael S. Warren 20130626ER Abstract much of the funding for FY15 was transferred to Prze- We describe a project that culminated in the capability mek Wozniak, of ISR-2, to finance LANL’s continuing role to do N-particle cosmological simulations for a record as an institutional partner in the Large Synoptic Survey number of particles (N ~ 1012) in a record (short) Telescope (LSST) consortium. Przemak provided the fol- amount of computer time. This capability was used to lowing motivation for continuing LANL’s connection with discover that dark matter distributions on galactic and the LSST: larger scales in the Universe are not uniform “halos” as Time-domain astronomy and the science of cosmic the standard approach predicts (or assumes), but are explosions are revolutionizing modern astrophysics with clumped in irregular structures. The complete results of profound impact on fundamental physics, as evidenced the calculation have been released publically, and pre- by the 2011 Nobel Prize for supernova (SN) studies lead- liminary interpretations of them have been published. ing to the discovery of Dark Energy. The Large Synoptic However, a decision in 2015 by the PI of the ER project Survey Telescope (LSST) selected as the top priority to take entrepreneurial leave changed the course of the ground-based project by the Astro2010 Decadal Survey research in its last year. The focus became LANL partici- of National Academies will be the first multi-Petabyte pation in the Large Synoptic Survey Telescope (LSST) con- time-domain survey. In 2014 NSF and DOE approved the sortium, which will probe four key areas: time-domain $600 Million construction budget for LSST. The project astrophysics, dark energy and weak gravitational lensing gained new momentum and is poised to conduct a 10- studies, Solar System science, and galactic structure and year sky survey with unprecedented sensitivity, resolu- evolution. tion, and coverage starting in 2022. LSST will open for exploration a parameter space that is three orders of Background and Research Objectives magnitude larger than available today in four key areas: This ER project addresses a crucial question in cosmol- time-domain astrophysics, dark energy and weak gravita- ogy about the nature of the distribution of dark matter tional lensing studies, Solar System science, and galactic in the Universe on galactic and larger scales. This is structure and evolution. necessary in order to interpret the results of dark mat- Scientific Approach and Accomplishments ter searches, such as the data coming from the Planck Satellite. Investigating these distributions in both coordi- A major milestone achieved during the second year nate and momentum (velocity) space requires doing the of the project was a complete trillion-particle simula- largest N-body particle simulations ever attempted, with tion performed on the Titan supercomputer (ORNL), N now of the order of a trillion particles. Preliminary which ran continuously for about a day and a half with non-linear gravitational collapse calculations with 100 no failures. No other cosmological simulation code has billion particles had revealed quite non- uniform struc- achieved this level of resolution with such short run tures in the dark-matter distributions in both spaces, in times. In the third year, the team focused on higher-res- sharp contrast to the uniform distribution assumed for olution simulations on the galactic scale, which reveals the “standard halo.” more about baryonic properties of the early universe. They also performed calculations to assist in interpreting During the late spring of this year, Mike Warren, the PI the possible dark matter signatures that will be com- of the project, decided to reduce substantially the time ing out of new observational facilities such as the Dark he spends at LANL, and on this ER project. Therefore, Energy Spectroscopic Instrument (DESI) at the Kit Peak 636

Page 637

Observatory. These calculations were done using a DOE Release. 2014. arXiv preprint arXiv:1407.2600. INCITE proposal grant on the Titan supercomputer at Oak Song, J., J. J. Mohr, W. A. Barkhouse, M. S. Warren, K. Ridge. These calculations and code (2HOT) clearly repre- Dolag, and C. Rude. A Parameterized Galaxy Catalog sent the state-of-the-art in N-body cosmological simula- Simulator for Testing Cluster Finding, Mass Estimation, tions. and Photometric Redshift Estimation in Optical and Near-infrared Surveys. 2012. The Astrophysical Journal. Impact on National Missions 747 (1): 58. The dark matter project is a flagship of the NPAC compo- nent of the Nuclear and Particle Futures Pillar. It is also Sutter, P., G. Lavaux, B. D. Wandelt, D. H. Weinberg, and M. a prime example of the Data Science at Scale area of the S. Warren. Voids in the SDSS DR9: observations, simu- lations, and the impact of the survey mask. 2013. arXiv Laboratory’s Information, Science, and Technology Pillar. preprint arXiv:1310.7155. These calculations were used to push the limits of - and uncovered scaling issues with - the Mustang supercom- Sutter, P., G. Lavaux, B. D. Wandelt, D. H. Weinberg, and puter Having a working code that can operate at the M. S. Warren. The dark matter of galaxy voids. 2014. capability limits of any new computer system, yet is able Monthly Notices of the Royal Astronomical Society. : to get answers in a reasonable length of time, not only is stt2425. critical to national needs and emerging issues, but also is Sutter, P., G. Lavaux, B. D. Wandelt, N. Hamaus, D. H. Wein- unique - not duplicative of capabilities at other national berg, and M. S. Warren. Sparse sampling, galaxy bias, labs, esp. NNSA. It establishes enabling capabilities critical and voids. 2013. arXiv preprint arXiv:1309.5087. to Lab strategies for emerging technical endeavors, espe- cially those related to “big data.” There is ample potential Sutter, P., G. Lavaux, N. Hamaus, A. Pisani, B. D. Wandelt, for a two- way transfer of ideas, and technology to national M. S. Warren, F. Villaescusa-Navarro, P. Zivick, Q. security missions (e.g. nuclear, global, energy). Mao, and B. B. Thompson. VIDE: The Void IDentifica- tion and Examination toolkit. 2014. arXiv preprint Wozniak summarizes the importance of the LSST connec- arXiv:1406.1191. tion as follows: Since 2006 LANL scientists (led by Przemek Wozniak, Thomas Vestrand, Chris Fryer, and Michael War- Teodoro, L. F. A., D. G. Korycansky, M. S. Warren, C. L. Fryer, G. Rockefeller, K. Zahnle, and V. R. Eke. A new suite of ren) are members of LSST science working groups focus- hydrodynamical simulations of the origin of the obliq- ing on studies of explosive variability, supernovae, and uity of Uranus. Invited presentation at 46th Lunar and structure formation in the Universe. We are recognized as Planetary Science Conference. (The Woodlands, Texas, intellectual and technical leaders in data intensive analysis 16-20 March 2015). and large-scale simulations. As a member institution of the LSST consortium, LANL is also represented by a voting Teodoro, L., D. Korycansky, M. Warren, C. Fryer, G. Rock- member on the Board of Directors of the LSST Corporation. efeller, and K. Zahnle. A Ten Million SPH Particle Simu- Supporting astrophysics and cosmology research at LANL lation of the Origin of Obliquity of Uranus. 2014. In and forging external collaborations with leaders in the Lunar and Planetary Institute Science Conference Ab- stracts. Vol. 45, p. 2542. field are both high priority areas for LDRD. Funding LANL participation in LSST is a great opportunity to advance Teodoro, L., M. Warren, C. Fryer, V. Eke, and K. Zahnle. A these goals and reinforce LANL leadership in time-domain One Hundred Million SPH Particle Simulation of the astrophysics and cosmology where LSST is expected to Moon Forming Impact. 2014. In Lunar and Planetary have a revolutionary impact. Institute Science Conference Abstracts. Vol. 45, p. 2703. Publications Hamaus, N., B. D. Wandelt, P. Sutter, G. Lavaux, and M. S. Turk, M. J., M. S. Warren, and S. W. Skillman. Building a Warren. Cosmology with Void-Galaxy Correlations. Community-Focused Data Platform. 2014. Urbana. 51: 2014. Physical review letters. 112 (4): 041304. 61801. Johnston, R., D. Bacon, L. Teodoro, R. Nichol, M. Warren, Warren, M. S.. 2HOT: An Improved Parallel Hashed Oct- and C. Cress. Reconstructing the velocity field beyond Tree N-Body Algorithm for Cosmological Simulation. the local universe. 2014. General Relativity and Gravi- 2014. Scientific Programming. 22 (2): 109. tation. 46 (11): 1. Warren, M. S., and B. Bergen. Poster: The Hashed Oct-Tree Skillman, S. W., M. S. Warren, M. J. Turk, R. H. Wechsler, D. N-Body Algorithm at a Petaflop. 2012. In High Perfor- E. Holz, and P. Sutter. Dark Sky Simulations: Early Data mance Computing, Networking, Storage and Analysis (SCC), 2012 SC Companion:. , p. 1442. 637

Page 638

Nuclear and Particle Futures Exploratory Research Final Report Coherent Diffractive Imaging of Ultrafast Ejecta Processes Cynthia A. Bolme 20130632ER Abstract ing this challenging project that involves material phase The physical processes leading to the formation of transitions and plastic deformation processes is the particles ejected from a shocked rough surface (i.e. dynamic evolution of the micro-jet that breaks up into ejecta) is not well understood. The scientific challenge particles ranging in size from 10s of microns to nanome- in studying the ejecta formation and evolution is obtain- ters. These particles can be solid or liquid and can be ing high resolution structural information leading to an reabsorbed, coalesce, or continue to break up depend- understanding of particle size and speed distributions. ing upon the surface geometry, the surrounding environ- There are many factors affecting ejecta growth and pro- ment, and the strength of the shock wave. Developing duction, especially when the pressure and temperature computational tools to understand and predict this conditions of the shock wave causes the material to go process requires sophisticated material models and is of through structural phase changes with the most signifi- critical importance to the mission of the laboratory. cant changes occurring when the material transitions Ejecta have been studied using laser velocimetry, proton between solid and liquid phases. Not fully understanding and X-ray radiography, and ultraviolet holography. The the physics of ejecta leads to the inability to accurate current techniques of proton radiography (pRAD) and predict ejecta formation. Diagnostics that are currently ultraviolet (UV) holography have imaged late time ejecta being used to probe the size and velocity distribution dynamics with spatial resolutions down to 10 µm and have produced incredibly useful data. However, these di- to 800 nm, respectively. These techniques have been agnostics have not been able to measure the small scale used to probe the particulate size and velocity distribu- extent of the ejecta distribution due to the limited spa- tion. However, due to the currently available resolution, tial resolution. This project sought to extend the spatial the sizes, velocities, and shapes of the micro-jet and its resolution of the measurement of ejecta particles to well subsequent particulates is unknown for features less below that which is currently achievable by performing than ~1 µm [1-5]. However, molecular dynamics (MD) coherent x-ray diffractive imaging with soft x-rays up to models of singly and doubly shocked metals show a 100 eV (resolution < 100 nm) on a high repetition rate, wealth of dynamic evolution on the sub-micron and in-house system laser system and down to a few nm sub-100 ps scale such as how and where bubbles (the resolution at X-ray free electron lasers. Here we sum- region of material feeding the spike growth) form, break marize the background of the problem of imaging ejecta up, and eventually evolve into spike growth, spike break while it is developing and evolving at shock velocities (1- up and the constituent particle size distribution [6]. In a 10 km/s), our efforts to develop single shot, nanometer recent paper, molecular dynamics simulations of copper scale X-ray imaging for ejecta, and the importance of this showed a power law decay in the particle size distribu- work to national missions. tion with respect to number of aggregates [7]. The smallest particle volumes observed were about 40 cubic Background and Research Objectives angstroms. For ejecta formation, molecular dynamics When shock waves break out from a roughened sur- simulations have an advantage over other simulations in face, the interaction of the shock wave with the free that they are more accurate in predicting surface ten- surface causes material to be ejected from the surface sion – an important part of understanding liquid ejecta in a process known as ejecta. Ejecta can form starting droplet evaporation and breakup. Thus we would expect from an instability in the shock front caused by surface that the molecular dynamics should produce a more re- roughness that evolves into micro-jets and eventually alistic particle size and shape. In addition to the sizes and into moving particulates. An important aspect of study- 638

Page 639

shapes of the particles, molecular dynamics simulations orders of magnitude lower than expected, but at the time also provide information about the structure (crystalline of this report, we have been scaling up the intensity of this or liquid) and the crystalline grain sizes that are predicted soft X-ray system and expect soon to achieve enough flux for the ejected material, and those predictions can be for single shot imaging. compared to measurements made with wide angle x-ray As mentioned before, the development of this unique sys- scattering by hard x-rays. While very powerful, the infor- tem has been attracting interest from outside collaborators mation gleaned from molecular dynamics simulations have including Prof. Brian Abbey’s group from La Trobe Univer- yet to be experimentally verified due to a lack of spatial sity in Melbourne, Australia who was a Physics Colloquium and temporal resolution in previously available diagnostics. speaker last fall. Furthermore, a recent LANSCE hire, Prof. To extend the available size and speed distribution mea- Edwin Fohtung, who is also a professor at New Mexico surements to capture the dynamics of smaller features and State University, has expressed interest in collaborating particulates, we proposed to develop new x-ray imaging with us and using this unique single shot imaging system. capabilities with which we would probe ejecta dynamics These two groups bring strong expertise in coherent dif- for the first few ns with resolutions of sub-50 fs and sub- fractive imaging techniques and algorithms and will greatly 100 nm. speed up the progress we can make on the imaging once the imaging experiments begin. The main objectives of this project were to develop an ultrafast laser to create a shock wave at a metal sample Computational progress has been made by performing surface and then to image dynamics with X-ray imaging. many simulations of ejecta formation in Al with vary- The samples have a periodically roughened surface, and ing depths of sinusoidal roughness on the shocked free the interaction of the shock wave with the structured surface. The calculations have shown that there is a larger surface creates instabilities in the material that will evolve effect of material viscosity for more shallow surface rough- in small jets of ejecta particles. This process was to be im- ness. The simulations have also shown that temperature aged using ultrafast x-rays either from a tabletop soft X-ray does not strongly influence the ejecta production in Al. source or from an intense accelerator based X-ray Free The most significant progress has been the successful field- Electron Laser (XFEL). In addition to ultrafast X-ray imaging, ing of nanometer scale imaging of shocked materials at we also have performed molecular dynamic simulations of the world’s first hard X-ray laser, the Linac Coherent Light this ejecta forming process that are well matched in length Source (LCLS) at the SLAC National Accelerator Laboratory and time scales to the experiments. in May 2015. We obtained beamtime at the Materials Scientific Approach and Accomplishments in Extreme Conditions (MEC) endstation at the LCLS for approximately 12 hours of experimentation. Beamtime This last year, our efforts were focused in three areas: 1) allocations for the LCLS are extremely competitive, so developing the single shot, tabletop soft X-ray source, 2) this was a great opportunity and it was the first time that conducting the first dynamic ejecta imaging experiments ejecta experiments have been performed at LCLS. We at the LCLS, and 3) performing MD simulations of ejecta fielded a low angle scattering detector for direct images of from Al surfaces using the best Al potential as determined the ejecta process alongside wide angle scattering detec- in FY14’s evaluation of the Al potentials. We continued to tors for measuring changes to the ejecta material phase. have problems with damaging the optical components in The intense, sub-100 fs X-ray pulses from the LCLS capture the laser system due to poor laser beam mode quality and the dynamic material motion in blur free images with high power. Team member Steve Gilbertson successfully approximately 500 nm resolution. Materials studied were incorporated a hollow-core spatial filter that improved the tin, cerium, and titanium. Analysis of this data is ongoing. laser beam mode and enabled us to obtain much higher These results will not only lead to high impact publications laser powers. With these changes, we have successfully but also to additional LCLS beamtime through strongly compressed the amplified 10 Hz laser for the high harmon- motivated user proposals. ic generation drive and imaging to 38 fs, better than the 50 fs we were attempting. We currently have up to 100 Impact on National Missions mJ of pulse energy in this short pulse and have been able The development of nanometer scale dynamic imaging is to generate intense soft X-ray pulses from argon filled gas addressing core mission problems and providing important cells. The soft X-ray photon energies have not been mea- validation of physics-based material models as the spatial sured but are expected to be centered at 40 eV, and we and temporal resolutions achievable with these imaging have produced about 10 million photons in a single 10 fs diagnostics are an order of magnitude better than current long pulse. This soft x-ray photon production is about 3-4 dynamic diagnostic alternatives. This project’s relevance 639

Page 640

intersects the Nuclear Performance, IS&T, and Materials 6. Dimonte, G., G. Terrones, F. J. Cherne, T. C. Germann, Grand Challenges, and it is directly linked with the pro- V. Dupont, K. Kadau, W. T. Buttler, D. M. Oro, C. Mor- posed LANL Signature Science Facility MaRIE (Mater-Radia- ris, and D. L. Preston. Use of the Richtmyer-Meshkov tion Interactions in Extremes). The strategic impact of this Instability to Infer Yield Stress at High-Energy Densities. project is the development of enabling capabilities in the 2011. Phys. Rev. Lett.. 107: 264502. following areas: in situ dynamics of materials in extremes 7. Durand, O., and L. Soulard. Large-scale molecular with x-ray probes, intense next-generation laser sources of dynamics study of jet breakup and ejecta production secondary radiation, and advanced research on real-time from shock-loaded copper with a hybrid method. 2012. reconstruction of phase-coherent diffraction data into 3D Journal of Applied Physics. 111 (4): -. images. Development of technologies like those in this project are crucial to providing otherwise unobtainable data for validation of codes on current materials problems of interest and will begin the decadal development of ad- vanced coherent X-ray imaging at MaRIE. Los Alamos is currently the world-expert in studying ejecta through both experiments and modeling. The majority of the ejecta research up to this point has been performed using codes and examining the material on the meso-scale. This work has expanded our expertise in ejecta into the nano-scale and will advance our basic physical understand- ing of the ejecta process by comparison with simulations that are based solely on atomic potentials. References

1. Zellner, M. B., M. Grover, J. E. Hammerberg, R. S. Hixson, A. J. Iverson, G. S. Macrum, K. B. Morley, A. W. Obst, R. T. Olson, J. R. Payton, P. A. Rigg, N. Routley, G. D. Stevens, W. D. Turley, L. Veeser, and W. T. Buttler. Ef- fects of shock-breakout pressure on ejection of micron- scale material from shocked tin surfaces. 2007. Journal of Applied Physics. 102 (1): -.
2. Grigsby, W., B. T. Bowes, D. A. Dalton, A. C. Bernstein, S. Bless, M. C. Downer, E. Taleff, J. Colvin, and T. Dit- mire. Picosecond time scale dynamics of short pulse laser-driven shocks in tin. 2009. Journal of Applied Physics. 105 (9): -.
3. Rességuier, T. de, L. Signor, A. Dragon, M. Boustie, and L. Berthe. On the dynamic fragmentation of laser shock-melted tin. 2008. Applied Physics Letters. 92 (13): -.
4. Li, Z., Z. Luo, Z. Liu, Y. Ye, Z. Li, J. Zhong, and J. Li. High- speed microjet particles measurement using in-line pulsed holography. 2010. Journal of Applied Physics. 108 (11): -.
5. Sorenson, D. S., R. W. Minich, J. L. Romero, T. W. Tun- nell, and R. M. Malone. Ejecta particle size distribu- tions for shock loaded Sn and Al metals. 2002. Journal of Applied Physics. 92 (10): 5830. 640

Page 641

Nuclear and Particle Futures Exploratory Research Final Report In Search of Light WIMPs Alexander Friedland 20130637ER Abstract matter that satisfy all experimental and observational Extremely low-background detectors designed for other constraints. experimental searches, for example neutrinoless double Scientific Approach and Accomplishments beta decay, can also be used to look for light (~1 GeV mass) dark matter. Theoretical possibilities for dark The MAJORANA experiment is designed to search for matter at this mass scale includes weakly-interacting neutrinoless double beta decay, a process in which two massive particles (WIMPS), these can be searched for neutrons decay into protons inside a nucleus emitting in experiments like MAJORANA. LANL has a substantial two electrons but no neutrinos. This process is, in the interest in the MAJORANA double-beta decay experi- standard picture, only possible for Majorana neutrinos, ment looking for neutrinoless double beta decays. The those that are there own anti-particle. same detector, and others like it, may be able to discover The current generation double-beta decay experi- WIMPs in a mass and coupling range not accessible to ment is called the MAJORANA DEMONSTRATOR, it is a other dark matter experiments. We also discuss pos- smaller scale experiment designed to understand what sible new scenarios for light dark matter which include is required, in particular in the area of reducing back- interactions with neutrinos to resolve problems with an grounds, to a full ton-scale detector that would be much overprediction of satellite galaxies if collisionless cold more sensitive. Part of the strategy for improving these dark matter is assumed. experiments is to reduce background and to reduce the energy threshold of the detectors. Achieving a sub-keV Background and Research Objectives energy threshold would allow the MAJORANA demon- Dark Matter is known to compose about 25% of the strator to also search for light WIMPs.[1] energy density of the universe, as measured by Planck, and previously COBE and WMAP and other experiments Figure 1 shows the projected sensitivity to light mass including big-bang nucleosynthesis. However we do not WIMPs for 100 kg-years of exposure with a 0.3 keV know the particle content of this dark matter, what is it energy threshold. As shown in the figure, at a mass of ~ exactly? Various large experiments are searching for 10 GeV present generation experiments are extremely this matter, typically looking for recoils of large nuclei sensitive. However below 1 GeV the cross sections have like Xenon to search for the presence of dark matter no stringent limits. The MAJORANA DEMONSTRATOR at the earth. Next generation experiments in the US could be much more sensitive in this mass range, and of include SuperCDMS at SNOLAB, the LUX-Zepelin (LZ) ex- course the full MAJORANA would be even more sensi- periment, and the Axion Dark Matter experiment ADMX- tive, nicely complementing the higher-mass searches of Gen2. LZ and SuperCDMS will look for WIMPs, while other experiments. A modified low-background broad ADMX-Gen2 will search for a different type of particle energy germaniam detector (MALBEK) has been de- called axions. ployed at the Kimballton Underground research facility in Virginia. To date, dark matter searches have not been very sensi- tive to light WIMPs, those of order 1 GeV or so. Also, As part of this project, theoretical particle physics sce- they have not clearly detected dark matter particles in narios with light WIMPs were also investigated. Previ- the heavier mass range. This project is aimed at studying ously it had been thought that heavier WIMPs were how searches for lighter WIMPs might be undertaken, more natural, and hence the concerted experimental ef- and searching for new theoretical models for cold dark 641

Page 642

fort to look in this regime where the recoil energy is larger. and in the right panel 1 MeV. The allowed parameters that It has been suggested that collisionless cold dark matter fix the observational issues with small-scale structure are over-predicts the number of satellite galaxies. This appar- indicated by the filled blue regions. The other lines rep- ent lack of structure at small scales can be reconciled with resent constraints from other observations, including the theory by postulating interactions between neutrinos and Lyman alpha forest (solid green curves) and the observed dark matter with a force carrier of mass of order MeV. An total amount of dark matter ( black curves). important feature here is that the dark matter couples to This investigation shows that new models of dark matter normal matter only in the form of neutrinos. This scenario interacting with neutrinos are possible, resolve some dis- can be tested in the IceCube experiment.[2] crepancies with observations of clustering at small scales, and are potentially testable at IceCube. The allowed regions of dark matter masses encompass the regions measured by traditional dark matter experiments and also the lighter masses that could be probed with MAJORANA. Impact on National Missions Dark Matter and neutrino physics are two of the highest priorities of the DOE Office of Science High Energy phys- ics program. This LDRD-ER project materially advanced the prospects for searching for light WIMPs, an important experimental possibility for future experiments including those primarily designed for double beta decay searches. A small test experiment is already being performed, and future larger-scale efforts are being studied. This LDRD project also proposed new theories for dark matter that incorporate interactions between neutrinos and dark mat- ter. These theories can experimentally tested in the future. The theories produced in this project can simultaneously solve some problems which result with comparing stan- Figure 1. Sensitivity of the MAJORANA DEMONSTRATOR dard collisionless dark matter compared with astrophysical compared to other experiments to WIMPs as a function of the observations. WIMP mass. Lower curves demonstrate increased sensitivity. The MAJORANA DEMONSTRATOR could be particularly sensitive to References light-mass WIMPs.

1. Giovanetti, G. K., and S. R. Elliott. A Dark Matter Search with MALBEK. 2014. Physics Procedia. 00: 1.
2. Friedland, A., and M. Shoemaker. Integrating In Dark Matter Astrophysics at Direct Detection Experiments.
3. Physics Letters B. 724: 183. Publications Cherry, John Francis Jr., Alexander Friedland, and Ian M. Shoemaker. Neutrino Portal Dark Matter: From Dwarf Galaxies to IceCube. 2014. Physics Letters B. Figure 2. Allowed regions of mass and coupling for dark mat- Friedland, A., and M. Shoemaker. Integrating In Dark Mat- ter with the inclusion of neutrino-dark matter interactions. The ter Astrophysics at Direct Detection Experiments. 2013. shaded region is the allowed region. The left panel shows results Physics Letters B. 724: 183. for a 10 MeV force carrier, and the right for a 1 MeV force carrier. See text for further details. Friedland, Alexander, Annelise Malkus, and Gail C McLaughlin. Matter-Neutrino Resonance Above Merg- Figure 2 shows the allowed regions of dark matter mass ing Compact Objects. 2014. Physical Review Letters. (vertical axis) versus coupling to neutrinos (horizontal axis) for two different mediator masses. In the left panel the Giovanetti, G. K.. A Dark Matter Search with The MAJORA- force carrier for this new interaction has a mass of 10 MeV NA Low-Background Broad Energy Germanium Detec- 642

Page 643

tor. Invited presentation at TAUP 2013. (Asilomar, CA, USA, 9 Sep. 2013). Giovanetti, G. K., and S. R. Elliott. A Dark Matter Search with MALBEK. 2014. Physics Procedia. 00: 1. Gupta, Rajan, Alexander Friedland, Michael Lawrence Graesser, Tanmoy Bhattacharya, Michael S. Warren, and Emil Mottola. Los Alamos HEP Theory. 2014. 643

Page 644

Nuclear and Particle Futures Exploratory Research Final Report Emittance-Reduction System for Future Accelerator Solutions Kip A. Bishofberger 20130688ER Abstract gineering and scientific goals. In every application, a de- More than any other measure, the beam brightness has mand is placed on the number of beam particles (elec- been the driving force for technological development trons, protons, ions, etc.) that hit some target or pass in beam physics. The brightness refers to density of through some structure. The “emittance” quantifies the particles, both in physical space and phase space. The amount of disorder within the beam (in each of three di- brighter the beam, the more structured the beam is mensions); a low-emittance beam, like a laser beam, can to enable generation of X-rays, electromagnetic radia- be focused very tightly or pass uniformly through a long tion, and unique particles. The opposite of brightness distance. Beams with high emittance cannot be focused is called “emittance,” representing the level of disorder as tightly and are not nearly as efficient in their utility. and temperature within the beam. Until recently, the For decades, emittances were considered constants- beam’s emittances were considered constants of lin- of-motion through linear forces; in addition, nonlinear ear motion, and nonlinear forces only increased them, effects can raise emittances, but no simple technique thereby reducing the beam brightness. could reduce them. Recently, unique beam systems However, within the past couple of years, a break- were discovered that could reduce one or more emit- through has developed, largely originating with scien- tances. These cases, led largely by LANL scientists, tists at LANL. The term “eigen-emittances” refers to the grew into a new framework upon which beam physics emittances that a beam would possess if there were is based. A new term “eigen-emittance” was coined no additional correlations. The new discovery is that, at LANL to describe intrinsic beam characteristics, and by generating a beam with specific initial correlations, methods to reduce beam emittances were formalized. the eigen-emittances can be substantially reduced, and This project used this new formalism to study a beam- through beam optics that unravel the correlations, the line device that converts transverse beam emittance final, observed emittances will take on these improved into longitudinal. Similar to converting a flashlight into values, thereby generating a significantly brighter beam a laser, the beam is capable of being ten-times brighter for accelerator applications. This project successfully through this conversion process, significantly increasing analyzed this technique and optimized an experimental its utility in nearly all applications. setup to demonstrate this breakthrough eigen-emittance manipulation scheme. Scientific Approach and Accomplishments The technology development capability supported by The primary focus on the project was to study and op- this project enhances our ability to design Engineered timize a unique beamline insertion device, known as an Systems. Additionally, compact, advanced accelerator XZ-FBT (referring to a Flat-Beam Transformer between technology supports a wide array of accelerator-based the x-plane and z-plane). The key piece of this device is solutions to National Security missions. Two examples in- an ultra-thin, wedge-shaped foil that the beam passes clude Global Security, such as Warfighter Support (high- through. This tactic generates a unique spatial relation- power Free-Electron Lasers) and Countering Weapons of ship on the particle distribution; a subsequent magnetic Mass Effect (active interrogation). field profile converts this relationship into a lower-emit- tance beam that has significantly higher brightness than Background and Research Objectives prior to the foil. Beam physics has been a useful tool for a variety of en- 644

Page 645

The majority of the project was devoted to a careful study of this manipulation system. In particular, the mechanism scales in a unique manner with beam energy and species. A model insertion device was optimized for protons and electrons at a variety of beam energies and initial emit- tances. Improvements in brightness were shown to be about than a factor of three in one dimension (a second insertion device would increase the brightness to nearly ten times total). Multiple test facilities were analyzed to attempt a physi- cal demonstration of the device. The Argonne Wakefield Accelerator (AWA) was selected as the ideal candidate to perform experiments, after they renovated their beamline to 75 MeV energy. An insertion device was constructed specifically for those beam parameters. At the end of the project, AWA was in the process of testing the critical foil experiments that underpin the success of the project. These tests indicate similar results to the simulations have indicated, but ongoing data analysis will provide the ex- perimental verification to determine the system’s viability. Impact on National Missions Major accelerator facilities at LANL include LANSCE and DARHT, and successful implementation of the proposed MaRIE signature science facility will require developing new accelerator capabilities. Additionally, several emerg- ing major Global Security work- for-other missions require state-of-the-art accelerator technology. The work in this project directly supports future light sources (such as that proposed for the MaRIE XFEL) and future linear colliders, which are major DOE Office of Sci- ence projects. In turn, light sources support the Materials: Discovery Science to Strategic Applications grand chal- lenge and linear colliders support the Beyond the Standard Model grand challenge. More directly, the technology development capability supported by this project enhances our ability to design Engineered Systems, another grand challenge. Addition- ally, compact, advanced accelerator technology supports a wide array of accelerator-based solutions to National Secu- rity missions. Two examples include Global Security, such as Warfighter Support (high-power FELs) and Countering Weapons of Mass Effect (active interrogation). 645

Page 646

Nuclear and Particle Futures Exploratory Research Final Report Reactor Power for Large Displacement Autonomous Underwater Vehicles Patrick R. Mcclure 20150035ER Abstract has to fit in the “as designed” vehicle or designed into This research project is an investigation into small reac- the vehicle without rendering it non-functional. The tors suitable for powering Autonomous Underwater final issue is the reactor must not present a proliferation Vehicles (AUVs.) The primary motivation was U.S. Navy concern given that it is unattended. moving toward a large fleet of AUVs. AUVs, much like The reactor needs to be self-regulating since by defi- drones in the U.S. Air Force, are envisioned taking over nition AUVs have no personnel in attendance. The or supplementing a variety of missions, including: threat self-regulating feature is common to the small reactors surveillance, situational awareness; anti-submarine designed by Los Alamos. Self-regulation is accomplished warfare; information operations; wreckage recovery; and by using small highly reflected reactors (creating a large time critical strike. negative temperature feedback in the reactor) that al- However since beginning the research, the Navy has ex- lows the reactor to adjust to the power demands of the plicitly stated to LANL that they are not interested at this system. The physics of this system causes the reactor to time in nuclear reactors for AUVs for political reasons. load follow the power conversion system. However, the general idea of a long-term power source Fitting a reactor into a small diameter cylinder that is for commercial AUVs is still viable and NASA has interest needed for the hydrodynamics of a sub is challenging. in AUVs for planetary exploration of the Saturn moon The core, its reflector, any reactivity make-up system and Titan, which has large methane seas. shielding has to fit into the AUV outer cylinder diameter. The goal of this proposal is to design a small reactor suit- In addition, care must be used to maintain some key able for AUVs. Nuclear power removes any time limita- aspects of center of gravity of the AUV. Finally, how the tion on a mission, extending it from weeks to years (up heat is rejected from the reactor must be evaluated and to a decade.) A reactor is best suited for larger AUVs. designed. This proposal targets development of a nuclear reactor LANL proliferation experts have provided guidance on capable of powering a large displacement (> 3 ft diam- how to lower the nuclear attractiveness of a reactor eter, > 20,000 lbs) AUV. The reactor needs several key concept. The first is to keep the enrichment below the attributes including 1) self-regulation, 2) small volume threshold for highly enriched uranium at 20% or at a dimensions and 3) it must be proliferation resistant. value below 50% for highly enriched. The second is to The last two attributes are the primary challenge for this keep the amount of U235 fuel low. In these concepts, concept. Proliferation resistant means that enrichment, ideal values are 6 kgs for 93% enriched or 50 kgs for fuel mass and fuel form must be managed carefully to 50% enriched and unlimited mass for 20% enrichments. address security concerns while keeping the reactor at These mass values favor a moderated reactor (slower an appropriate size. These issues are addressed in the neutrons and more uranium efficient) over a fast reactor research. (faster neutrons and less uranium efficient.) The third is to pick a form of fuel that is not a metal and preferably Background and Research Objectives a ceramic to slow down the conversion to metal if the Several issues are important if a nuclear reactor were reactor were compromised. The final is to pick a fuel to be used for unattended operations in the ocean. The form where the fuel is dispersed in another material that first issue is that the reactor needs to self-regulate and makes conversion to metal even harder. A measure of operate safely. The second would be that the reactor 646

Page 647

proliferation is the security category of the system using of the AUV, which is approximately 90 cm. The diameter the DOE graded safeguards table. has to accommodate the core, core reflector and any shielding needed in the radial direction. These attributes were examined for a class of reactors spanning a power range from 1 kWe (electric) to 150 kWe Weight/size: For the higher power ranges (125 to 150 (electric) for use either in a NASA AUV or a commercial kWe), using a moderator makes about a 10% decrease in AUV. The desired power level is in the low 10’s of kWe. the weight of the reactor core regardless of fuel choice. Given the challenges of trying to engineer the moderator, Scientific Approach and Accomplishments moderating the reactor was not seen as a good engineer- The research was performed using several LANL reactor ing choice for this power range. design tools and commercial design tools. These tools For the middle power cases (22 to 38 kWe) moderating included LANL developed tools MCNP-61, Monteburns2, the core can have some substantial benefit. The drop in Maggie, and the commercial software ANSYS3. LANL weight is significant and more importantly the drop in group NEN-5 has the capability to quickly model a reac- Uranium in the core is about 75%. A single UZrH fuel (fuel tor using pre-processers to quickly develop a set of radia- and moderator combined) was examined. UZrH is a fuel tion transport (MCNP-6) input decks that allow for a quick with a long history and pedigree in TRIGA4 reactors. The evaluation of the neutron physics of a particular design. downside to UZrH is that in order to limit hydrogen loss, Reactor burn-up (a measure of the fuel used) and shielding the reactor core must operate at a lower temperature that, are done by the computer codes Monteburns and Maggie, in turn, would limit efficiency. But the combined fuel-mod- respectively. LANL group AET-1 then uses ANSYS (a com- erator is quite good in lowering overall weight, although bination computer aided design, structural, heat transfer reduction in U235 was about the same as fuel and modera- analysis software) to analyze the reactor fit and heat rejec- tor being separate. tion of the proposed reactor concept. For the very low power cases (1 kWe), the use of a mod- LANL looked at multiple reactor designs ranging in power erator could approach the weight of an 93% enriched ura- from 1 kWe to 150 kWe and using a fuel that spanned a nium system and lower the U235 mass by as much as 80%. range of attributes in order to achieve specific non-prolifer- ation and size goals. All of these reactors are self-regulating Proliferation: All of the 20% enriched cases are Security due to a large negative reactivity coefficient. Small highly Category 4 (the lowest category.) But the use of 20% reflected cores that use heat pipes for cooling can be comes with substantial weight/size penalties. The metal designed to have large negative reactivity coefficients. The core cases show the tradeoffs between weight/size and physics of the design precludes inadvertently inserting better security. The 93% enriched case is small and light, positive reactivity into the core, thus making for a very safe but would be burdened as Security Category 1 material reactor. (the highest category.) The 50% enriched case is a good example of getting the enrichment down and decreasing The factors that were traded included: the U235 to a level that would be categorized as Security • Neutron Spectrum – Fast neutron spectrum versus Category 3. The less than 50% enrichment and less than a moderated neutron spectrum by adding a hydride 50 kgs of U235 would drastically decrease the security (Yttrium Hydride or YH) to the reactor. Moderating requirements. The 20% enriched metal case is Security (slowly down) the neutrons in a reactor core is one Category 4, but the size and weight would be unacceptable means of lowering the amount of uranium or lowering for many applications. the enrichment of the uranium. The moderated cases show that even the 93% enriched • Fuel Type – Conventional reactor fuel (uranium metal case can be lowered to a Security Category 3 by reducing and uranium oxide) versus a uranium oxide fuel dis- the U235 to less than 6 kg. These cases have a total weight persed in graphite. that is either better than or competitive with the HEU metal core cases. However, these designs do come with • Power Level – Varied over three power ranges. engineering issues. The moderated reactor cores have the following issues: • Peak Temperature – Some moderators and fuels do not perform in a high temperature environment (> 500 C) • Large temperature gradients between the fuel and and thus have lower power conversion efficiency. moderator that require very robust means of insulating the moderator. Constraints on design are based upon the upper diameter 647

Page 648

• The problem of hydrogen retention in the moderator during sequences that cause high temperatures in the moderator or over long periods of time (years.) • Manufacturing some of the moderated designs would be difficult given the interleaved moderator and fuel configurations. Figure 2. Shielding Study for Neutrons Dose for a Surfaced AUV. Figure 1. MCNP Model of Reactor for Titan Moon AUV Packaging and Shielding: Fitting the reactor into a small long cylinder proved challenging. The AUV small diameter and long length is required to provide the appropriate hydrodynamics. The small diameters in AUV made shield- ing the reactor problematic. Shielding and packaging studies were done using the water (on earth) or methane Figure 3. Titan Moon AUV Schematic of Reactor, Shielding and (on the moon Titan) on the outside of the AUV with other Power Conversion. internal shielding to get doses down to an acceptable level. This means the AUV would have to be shut down prior to surfacing, but it is an option over complete shielding of the reactor. An example of the shielding model for an AUV for the moon Titan is shown in Figure 1. The results of the shield- ing study for neutrons dose for a surfaced AUV is shown in Figure 2. The study showed that using water or methane could achieve the required shielding while the AUV was submerged. Total shielding can be done for an AUV but is large and heavy in comparison. An example of the packing of the reactor in the Titan moon AUV is shown in Figure 3. The packing of a reactor in a generic commercial AUV is shown in Figure 4. The results show that packing can be accomplished, but changes to Figure 4. Generic AUV Schematic of Reactor and Power Conver- the AUV outer structure are a possibility. Center of mass sion. on the AUV could be impacted depending on reactor loca- tion. These issues are all solvable by traditional engineer- Impact on National Missions ing methods. Small, unattended reactors for missions that require self- regulation, small size and proliferation resistance can play a role in national missions. The U.S Navy may change its 648

Page 649

stance on their use of nuclear power in AUVs. Agencies such as NASA and the NNSA have shown considerable interest in the use of reactors in AUVs. NASA is likely to purse the use of nuclear power in an AUV for its mission to the Saturn moon Titan. Fission power is an option over the use of traditional radioisotopes power sources. Recent news reports have indicated that other foreign gov- ernments are also pursuing nuclear powered AUVs. The technology is very likely to be developed by the U.S if its adversaries continue to develop its use. References

1. Pelowitz, D. B.. MCNP6 User’s Manual. 2011. Los Ala- mos National Laboratory, LA-CP-11- 1708.
2. Poston, D. I., and H. R. Trellue. User’s Manual, Version 2.0 for MONTEBURNS, Version 5B. 1999. Los Alamos National Laboratory, LA-UR-99-4999.
3. ANSYS Manual. 2003. ANSYS Inc..
4. TRIGA Reactor Fuel. 2015. General Atomic. Publications Fallgren, A., D. Poston, D. V. Rao, and P. McClure. The Use of Moderation in Small Reactors to Reduce Prolifera- tion Concerns and Security Costs . Presented at To be determined. (TBD, TBD). 649

Page 650

Nuclear and Particle Futures Exploratory Research Final Report Towards Generating Laboratory Gigagauss Magnetic Fields and Their Impact on ICF Dynamics Kirk A. Flippo 20150541ER Abstract EP in the near future. The two experiments performed This project aimed to build a basis for ultra-high mag- developed and calibrated a Faraday rotation diagnostics netic field generation and use in the formation of a and measured a laser generated ultra-high magnetic turbulent magnetic dynamo and develop a design for field using the Trident laser. an experiment to produce one. Turbulent dynamos Scientific Approach and Accomplishments are believed to power the brightest light sources in the universe such as gamma ray bursts, black hole jets, and We executed two experimental campaigns, one using the most powerful cosmic rays and it also underlies the the National High Magnetic Field Laboratory’s (NHMFL) physics of the universe from the Earth’s magnetic field Single Turn Magnet, which can produce 100 T in a 1 to the formation of stars and galaxies [1-7]. In addition micro second pulse, and another using the LANL Trident the ability to generate a dynamo and model it would laser system. We used the NHMFL facility to test and have implications for Inertial Confinement Fusion if the calibrate a Faraday rotation diagnostic using an insulat- models could be used to understand the spontaneous ing hollow corundum tube with a polarizer/analyzer creation and evolution of these fields and show that inside, which sits directly in the magnetic field. The they might impose significant changes to the physics of polarizer/analyzer consists simply of two thin-film polar- the implosion. izers and a crystal between them. When a magnetic field is present the crystal will rotate any light passing through Background and Research Objectives it due to the Faraday effect. The polarizers assure that The magnetic dynamo is a fundamental process in the light entering the crystal is polarized in one direc- plasma physics, taking kinetic energy and converting it tion and the second polarizer allows us to measure how to magnetic energy and is very important to planetary much rotation of the initial polarization has occurred. physics and astrophysics [8]. Yet, surprisingly, a first- The rotated light can be detected with either a fast diode principles understanding of turbulent dynamos and its or a streak camera to determine the amount of rotation saturation mechanism are missing and it has never been and thus the field. We were able to measure the rota- demonstrated experimentally [1]. tion from a 30T pulse at the NHFML, and on Trident we used the laser to produce a current to produce a high It was shown by Ref. [9] that magnetic fields of up to magnetic field, which was measured in the same fash- 3000T could be generated using 1000 J of laser energy ion. Figure 1 shows the experiment as performed and focused onto a specially designed magnetic field coil the data from Trident experiment. The inset shows the target. These measurements have not been confirmed measured 30 T field from the Trident shot using 200 J of independently yet, thus using this method is limited. We laser energy to drive the coil. To our knowledge this is proposed to test the field generation and then design an the first direct measurement of the field inside a laser experiment to use the applied field to produce ampli- driven coil. We designed at test four laser-driven coil tar- fication in a turbulent magnetic dynamo to study the gets during the Trident beam time, with several different process by which magnetic fields are amplified and can laser pulse shapes and durations. Several coils produced reconnect to produce energetic particles. ~10T and one produced ~30T. A publication on this work is in preparation [11]. The current project was limited due to time to two ex- periments and to simulation studies to design a magnet- In addition a suite of simulations have been performed ic dynamo experiment which could be fielded at Omega using the LA-COMPASS MHD code to refine and coalesce 650

Page 651

a turbulent dynamo design and to understand the evo- Impact on National Missions lution of a turbulent magnetic dynamo produced using The design of a platform to produce a turbulent magnetic a cylindrical target irradiated with 15kJ of laser energy. dynamo provides a tool and a path forward for the C-10 These simulations scoped out the required applied B-field program to consider a possible experiment to look into the strength and the expected plasma density and tempera- feasibility that spontaneous magnetic fields are one of the tures in our turbulent dynamo configuration. Figure 2 potential reasons for the poor performance of ICF cap- shows the simulations results of a designed experiment, sules, and magnetic fields can change the viscosity and the which should be capable to amplify an applied magnetic morphology of an ICF implosion well before the traditional field by a measurable factor of 3. A publication is in prepa- thinking using the plasma beta equal to 1, where beta is ration [12]. the ratio of plasma pressure (temperature) to magnetic field pressure. What is possibly more important is the kinematic beta, which is the ratio of the fluid flow (possi- bly turbulent) energy density to the magnetic field energy density, which can reach unity at magnetic fields orders of magnitude smaller than those of the plasma beta, and cause deviation of the bulk flow and morphology inside the ICF implosion and may be responsible for the residual kinetic energy that does not go into heating of the hot spot and producing fusion energy. Such a discovery would have huge implications on ICF design and ICF code develop- ment. References

1. Lathrop, D. P., and C. B. Forest. Magnetic dynamos in the lab. 2011. PHYSICS TODAY. 64 (7): 40. Figure 1. The experiment as performed and the data from Tri- dent experiment. Inset: the measured 30 T field from the Trident 2. Zhang, , R. E. Cohen, and Haule. Effects of electron cor- shot using 200 J of laser energy to drive the coil. relations on transport properties of iron at Earth’s core conditions. 2015. NATURE. 517 (7536): 605.
2. Li, , K. H. Yuen, Otto, P. K. Leung, T. K. Sridharan, Zhang, Liu, Tang, and Qiu. Self-similar fragmentation regulated by magnetic fields in a region forming massive stars.
3. NATURE. 520 (7548): 518.
4. Balbus, S. A., and J. F. Hawley. Instability, turbulence, and enhanced transport in accretion disks. 1998. RE- VIEWS OF MODERN PHYSICS. 70 (1): 1.
5. Xu, H. a. o., H. u. i. Li, D. C. Collins, Li, and M. L. Nor- man. TURBULENCE AND DYNAMO IN GALAXY CLUSTER MEDIUM: IMPLICATIONS ON THE ORIGIN OF CLUSTER MAGNETIC FIELDS. 2009. ASTROPHYSICAL JOURNAL LETTERS. 698 (1): L14.
6. Xu, H. a. o., Govoni, Murgia, H. u. i. Li, D. C. Collins, M. L. Norman, Cen, Feretti, and Giovannini. COMPARI- SONS OF COSMOLOGICAL MAGNETOHYDRODYNAMIC Figure 2. 3D simulation of the plasma flows inside a cylinder. GALAXY CLUSTER SIMULATIONS TO RADIO OBSERVA- (left) Density distribution (in 10 cm) in the [R,Z] plane showing TIONS. 2012. ASTROPHYSICAL JOURNAL. 759 (1). the formation of a central column of plasma. (right) Temperature (in keV) and vorticity distributions in the [R,f] plane at z=0.5. The
7. Bell, A. R.. Turbulent amplification of magnetic field central hot plasmas are formed from three radial inflows collid- and diffusive shock acceleration of cosmic rays. 2004. ing with each other and become turbulent around the central axis. MONTHLY NOTICES OF THE ROYAL ASTRONOMICAL 651

Page 652

SOCIETY. 353 (2): 550. 8. KRAUSE, F., and K. H. RADLER. THEORY OF THE GEO- MAGNETIC DYNAMO BASED ON MEAN FIELD ELECTRO- DYNAMICS. 1979. PHYSICS OF THE EARTH AND PLAN- ETARY INTERIORS. 20 (2-4): 158. 9. Fujioka, , Z. h. e. Zhang, Ishihara, Shigemori, Hironaka, Johzaki, Sunahara, Yamamoto, Nakashima, Watanabe, Shiraga, Nishimura, and Azechi. Kilotesla Magnetic Field due to a Capacitor-Coil Target Driven by High Power Laser. 2013. SCIENTIFIC REPORTS. 3. 10. Rasmus, A., A. Stier, K. A. Flippo, H. Li, S. T. Li, and C. Mielke. Direct Faraday Measurements of a Laser-Driv- en Field Coil. Phys. Rev. Lett (in prep). 11. Li, H., S. T. Li, A. Rasmus, A. Stier, and K. A. Flippo. Design of a Laser-Driven Turbulent Magnetic Dynamo. Nature Physics (in prep). Publications Li, H., S. T. Li, A. Rasmus, A. Stier, and K. A. Flippo. Design of a Laser-Driven Turbulent Magnetic Dynamo. Nature Physics (in prep). Rasmus, A., A. Stier, K. A. Flippo, H. Li, S. T. Li, and C. Mielke. Direct Faraday Measurements of a Laser-Driven Field Coil. Phys. Rev. Lett (in prep). 652

Page 653

Nuclear and Particle Futures Postdoctoral Research and Development Continuing Project Theoretical Investigation of Nucleon and Nuclear Structure at Very High Energies Zhongbo Kang 20130783PRD2 Introduction extract such information from data collected in the US The theoretical study and experimental exploration of and abroad. It also provides much-needed theoretical the internal structure of nucleons (protons and neu- guidance for and interpretation of the experimental trons) and nuclei are of fundamental importance to sci- results from the flagship nuclear physics program in US ence and have recently entered a new exciting phase. In and abroad. This project will also help firm the scientific the past decades an understanding of nucleons in terms case for a future Electron Ion Collider. of quarks and gluons (partons) has successfully emerged. Progress Progress has been made in constructing a “one-dimen- sional” picture of the nucleon, in the sense that we In FY15 we made significant progress toward the ulti- “only” know about the longitudinal motion of partons in mate goal of this project. We have written 9 papers, with fast moving protons and neutrons. However, the par- six published already: one in Physical Review Letters, ton’s motion that is perpendicular (transverse) to the two in Physics Letters B, two in Physical Review D, and proton momentum is still largely unknown and urgently one in Nuclear Physics A. The rest three are submitted to needed in order to construct the 3-dimensional image Physical Review D, and we expect them to be published of the proton. Since the transverse momentum compo- soon. We have delivered four invited talks at the interna- nent is usually much smaller than parton’s longitudinal tional conferences. We have also helped organize three component, it is critical to develop accurate theoretical workshops, with Dr. Kang being the co-chair for QCD tools to extract this information. We propose two unique Evolution Workshop 2015, in which our research direc- ways to attack this problem. One is through the trans- tion is one of the hot topics discussed. Dr. Kang has also verse spin dependence of experimental observables, as been invited to give six one-hour lectures in the HUGS transverse spin can correlate with the parton transverse 2015 Summer School organized by Jefferson Lab, with momentum. The other way is through a discovery of the lecture topic on “QCD Structure of the Nucleon and a novel nuclear dependence, as the small transverse Spin Physics’’ — one of the major research direction of motion can be amplified by the nuclear size. We will our project. develop the first solid theoretical framework that can In studying parton’s transverse motion in the nucleon, utilize both methods to pinpoint the partons’ transverse most of the efforts in the community have been fo- motion in both the proton and the big nucleus. cused on quark transverse momentum, while gluon’s transverse motion has to be further explored. With this Benefit to National Security Missions in mind, in FY15 we investigated for the first time the This project ties directly into the Laboratory mission role of gluons in generating the single transverse spin in scientific innovation and discovery. Understanding asymmetry. In this study, we understood better how the the structure of the nucleon is central to the nuclear gluon’s transverse momentum is correlated with the physics component of the National Science Foundation proton’s transverse spin. We derived the next-to-leading (NSF) and Department of Energy (DOE) Office of Sci- order transverse momentum-weighted Sivers asym- ence portfolios. Specifically, this project will provide new metry in semi-inclusive deep inelastic scattering (SIDIS) and unique 3 dimensional imaging information for the

• specifically contributions from the so-called three- nucleon structure at the ~20% level. It will also give es- gluon correlation functions. Such a study provides a sential insights into how the quark and gluon dynamics way to connect the quark and gluon’s transverse motion are modified in a large nucleus. It will produce theoreti- through the Quantum Chromodynamics (QCD) evolution cal tools for the community to accurately and reliably 653

Page 654

of the relevant quark-gluon and three-gluon correlation distribution in the large nucleus, which contains important functions. information on parton’s transverse motion. To further test QCD evolution and understand the spin Conclusion correlation in the fragmentation process, In FY15 we per- We will develop a new consistent theoretical formalism for formed global analysis of Collins spin asymmetry measured evaluating the cross sections in polarized reactions with in both SIDIS at HERMES, COMPASS, and Jefferson Lab, as protons and nuclei. We will also investigate the novel nu- well as the e+ + e- experiments at BaBar and Belle. In this clear dependence of experimental spin observables in both work we extracted the so-called quark transversity distri- electron-nucleus and proton-nucleus collisions. We will use bution and the Collins fragmentation function, both with these theoretical formalisms to interpret the experimental proper QCD evolution taken into account. data collected at the major experimental facilities in the US and abroad and extract the valuable information on the In studying the parton’s transverse motion in heavy nuclei, parton’s transverse motion. This will provide unique 3-di- in FY15 we concentrate on the role of parton multiple scat- mensional imaging information for the proton and nuclear tering in various physical observables in p+A and e+A colli- structure at high energies. sions. For example, we investigated the heavy meson pro- duction in p+A collisions at backward rapidity region, and Publications found the parton multiple scattering in this region is mostly Aschenauer, E. C., A. Bazilevsky, K. Boyle , R. Fatemi , C. incoherent and led to the interesting nuclear enhancement Gagliardi, M. Grosse-Perdekamp, Z. B. Kang, Y. Kov- compared to p+p collisions. Our predictions turned out chegov, and J. Lajoie. The RHIC Spin Program: Achieve- to be consistent with experimental observations at both ments and Future Opportunities . 2013. White Paper RHIC and LHC. We further extended such a study to e+A on RHIC spin physics to the Tribble Panel and Nuclear collisions, in which we performed a next-to-leading order Science Advisory Committee. computation of nuclear transverse momentum broadening Dai, L. Y., Z. B. Kang, A. Prokudin, and I. Vitev. Next-to- for single hadron production in SIDIS. leading order transverse momentum-weighted Sivers asymmetry in semi-inclusive deep inelastic scattering: Future Work the role of the three-gluon correlator. PHYSICAL RE- In FY16, we will continue to study the role of gluons in gen- VIEW D. erating the spin asymmetries. In particular, we will general- ize our current study on three-gluon correlation functions Echevarria, M. G., A. Idilbi, Z. B. Kang, and I. Vitev. QCD to Drell-Yan production and a color neutral scalar (for Evolution of the Sivers Asymmetry. 2014. PHYSICAL RE- example the Higgs) production in transversely polarized VIEW D. 89: 074013. proton-proton collisions. From this study, we will identify Gamberg, L., Z. B. Kang, A. Metz, D. Pitonyak, and A. Proku- the complete Quantum Chromodynamics (QCD) evolution din. Left-right spin asymmetry in ℓN↑→hX. 2014. equations for both quark-gluon and three-gluon correla- PHYSICAL REVIEW D. 90 (7): 074012. tion functions. At the same time, we will study the QCD evolution in other important spin observables, such as the Gamberg, L., Z. B. Kang, I. Vitev, and H. Xing. Quasi-parton Boer-Mulders function in Drell-Yan production. distribution functions: a study in the diquark spectator model. 2015. PHYSICS LETTERS B. 743: 112. To further our understanding of parton multiple scattering, in FY16 we plan to concentrate on the so-called small-x Gamberg, L., Z. B. Kang, and A. Prokudin. Indication on the process-dependence of the Sivers effect. 2013. Physical region. In this region the parton’s longitudinal momentum Review Letters. 110 (23): 232301. fraction is small and, thus, the transverse momentum component becomes the dominant degree of freedom. Gamberg, L., and Z. B. Kang. Single transverse spin asym- Furthermore, a new kind of divergence - so-called rapidity metry of prompt photon production. 2012. Physics Let- divergence - appears, which leads to the well-known small- ters B. 718 (1): 181. x evolution of parton densities. We plan to use the newly developed rapidity regulator technique to regularize and Huang, J., Z. B. Kang, and I. Vitev. Inclusive b-jet production in heavy ion collisions at the LHC . 2013. PHYSICS LET- isolate such a divergence, and perform the calculation for TERS B. 726: 251. the cross sections of virtual/real photon and jet production to the next-to-leading order. The experimental data will be Kang, Z. B.. QCD and RHIC spin physics. Invited presenta- available in the near future from both RHIC and LHC. We tion at 2012 Fall Meeting of the APS Division of Nucle- will compare our theoretical formalisms with the experi- ar Physics. (Newport Beach, CA, 24-27 Oct. 2012). mental data, and extract the so-called unintegrated gluon 654

Page 655

Kang, Z. B.. Polarized p+A, single spin asymmetries. Invited Kang, Z. B., E. Wang, X. N. Wang, and H. Xing. Transverse presentation at BNL-LANL-RBRC Joint Workshop on momentum broadening in semi-inclusive deep inelas- The Physics of p+A Collisions at RHIC. (Upton, NY, 7-9 tic scattering at next-to-leading order. PHYSICAL RE- Jan. 2013). VIEW D. Kang, Z. B.. Single transverse spin asymmetries in polarized Kang, Z. B., I. Vitev, E. Wang, H. Xing, and C. Zhang. Mul- SIDIS and pp scattering. Invited presentation at The 5th tiple scattering effects on heavy meson production in Workshop of the APS Topical Group on Hadronic Phys- p+A collisions at backward rapidity . 2015. PHYSICS ics. (Denver, CO, 10-12 Apr. 2013). LETTERS B. 740: 23. Kang, Z. B.. Forward physics from a theoretical perspective. Kang, Z. B., I. Vitev, and H. Xing. Transverse momentum- Invited presentation at STAR Meeting on eSTAR Letter weighted Sivers asymmetry in semi-inclusive deep in- of Intent, Forward-Upgrades and Results from U+U elastic scattering at next-to-leading order. 2013. Physi- Collisions. (Los Angeles, CA, 28-30 Aug. 2013). cal Review D. 87 (3): 034024. Kang, Z. B.. Double parton fragmentation function and its Kang, Z. B., I. Vitev, and H. Xing. Next-to-leading order for- evolution in quarkonium production . 2014. In QCD ward hadron production in the small-x regime: the role Evolution 2013. (Newport News, VA, 6-10 May 2014). of rapidity factorization . 2014. PHYSICAL REVIEW LET- Vol. 25, p. 1460040 . Singapore: International Journal TERS. 113: 062002. of Modern Physics. Kang, Z. B., I. Vitev, and H. Xing. Multiple scattering effects Kang, Z. B.. Unique opportunities in p+A collisions at RHIC on inclusive particle production in the large-x regime . and LHC. Invited presentation at APS Division of Nucle- 2013. PHYSICAL REVIEW D. 88: 054010. ar Physics 2014 Long-range plan: Joint Town Meetings on QCD. (Philadelphia, PA, 13-15 Sept. 2014). Kang, Z. B., I. Vitev, and H. Xing. Initial-state cold nuclear matter energy loss effects on inclusive jet production Kang, Z. B.. Nucleon spin: longitudinal, transverse, and evo- in p+A collisions at RHIC and LHC. PHYSICAL REVIEW C. lution. Invited presentation at 2014 RHIC and AGS An- nual Users Meeting. (Upton, NY, 17-20 June 2014). Kang, Z. B., R. Lashof-Regas, G. Ovanesyan, P. Saad, and I. Vitev. Jet quenching phenomenology from soft-collin- Kang, Z. B.. TMD evolution of Sivers asymmetry. Invited ear effective theory with Glauber gluons. 2015. PHYSI- presentation at Studies of 3D Structure of Nucleon. CAL REVIEW LETTERS. 114 (9): 092002. (Seattle, WA, 24-28 Feb. 2014). Kang, Z. B., S. Mantry, and J. W. Qiu. Probing nuclear dy- Kang, Z. B.. QCD evolution of TMDs: what works?. Invited namics in jet production with a global event shape . presentation at Indiana-Illinois Workshop on Fragmen- 2013. Probing nuclear dynamics in jet production with tation Functions. (Bloomington, IN, 12-14 Dec. 2013). a global event shape . 88: 074020. Kang, Z. B.. TMDs: Mechanisms/universality with ep and pp Kang, Z. B., X. Liu, and S. Mantry. The 1-Jettiness DIS event collisions. Invited presentation at QCD Frontier 2013. shape: NNLL + NLO results. 2014. PHYSICAL REVIEW D. (Newport News, VA, 21-22 Oct. 2014). 90: 014041. Kang, Z. B.. TMD evolution and global analysis. Invited Kang, Z. B., Y. Q. Ma, J. W. Qiu, and G. Sterman. Heavy presentation at The 6th Workshop of the APS Topical Quarkonium Production at Collider Energies (I): Factor- Group on Hadronic Physics. (Baltimore, MD, 8-10 April ization and Evolution. 2014. PHYSICAL REVIEW D. 90: 2015). 034006. Kang, Z. B., A. Prokudin, P. Sun, and F. Yuan. Nucleon tensor Kang, Z. B., Y. Q. Ma, J. W. Qiu, and G. Sterman. Heavy charge from Collins azimuthal asymmetry measure- Quarkonium Production at Collider Energies: Partonic ments . 2015. PHYSICAL REVIEW D. 91: 071501. Cross Section and Polarization . 2015. PHYSICAL RE- VIEW D. 91 (1): 014030. Kang, Z. B., A. Prokudin, P. Sun, and F. Yuan. Extraction of Quark Transversity Distribution and Collins Fragmenta- Kang, Z. B., Y. Q. Ma, and R. Venugopalan. Quarkonium tion Functions with QCD Evolution. PHYSICAL REVIEW production in high energy proton-nucleus collisions: D. CGC meets NRQCD. 2014. Journal of High Energy Phys- ics. 01: 056. Kang, Z. B., E. Wang, X. N. Wang, and H. Xing. Next-to- Leading QCD Factorization for Semi-Inclusive Deep Kang, Z. B., and B. Xiao. Sivers asymmetry of Drell-Yan pro- Inelastic Scattering at Twist-4 . 2014. Physical Review duction in the small-x regime. 2013. Physical Review D. Letters. 112: 102001 . 87 (3): 034038. 655

Page 656

Vitev, I.. Transv. Invited presentation at Transverse momen- tum-weighted Sivers asymmetry in semi-inclusive deep inelastic scattering at next-to-leading order. (Newport News, VA, 6-10 May 2013). 656

Page 657

Nuclear and Particle Futures Postdoctoral Research and Development Continuing Project Investigating Properties of Quark-Gluon Plasma Using Jets and Heavy Quark Production at RHIC Michael P. Mccumber 20140665PRD2 Introduction Benefit to National Security Missions Jets measurements feature prominently in the current This project is highly relevant LANL’s Scientific Discovery set of Large Hadron Collider (LHC) results. Measure- and Innovation mission, in the area of Nuclear, Particle, ments at Relativistic Heavy Ion Collider (RHIC) will pro- Cosmology, and Astrophysics. Quantum Chromodynam- vide insight into properties of the Quark Gluon Plasma ics (QCD), the Standard Model theory governing the be- (QGP) under different physical conditions than can be havior of quarks and gluons, predicts that a new phase explored with the LHC alone. The 400 person PHENIX of matter is present in the relativistic collisions of large collaboration is planning to propose the construction nuclei. Matter generated by these collisions is at such of a new $30M detector, sPHENIX, to DOE to provide a high temperature and density that quarks and gluons comprehensive set of jet measurements at RHIC and to are no longer imprisoned in hadronic states, but are prepare for the construction of an electron-ion collider instead free to interact with one another directly, a state (EIC). These physics goals are much beyond the original of matter known as Quark-Gluon Plasma (QGP). The design capabilities of the PHENIX experiment, which was new phase of matter is being actively studied at hadron designed two decades earlier. To achieve these measure- collider facilities such as the Relativistic Heavy Ion Col- ments, this LDRD-PRD project proposes to spearhead the lider (RHIC) at Brookhaven National Laboratory and the development and construction of a new RHIC detector Large Hadron Collider (LHC) at CERN. Research work on needed to bring a world-class experimental jet phys- the understanding of properties of Quark-Gluon Plasma ics program to LANL. The new detector is crucial to the will also contribute and lead to new resources of nuclear full exploration and understanding of the Quark-Gluon energy in the next generations. Plasma. Progress This LDRD-PRD project is focused on a new research area FY15 has been a crucial period for progressing the of Heavy Ion Physics (using particle jet production as a proposal for a new detector at RHIC, sPHENIX, toward probe to study the properties of Quark-Gluon Plasma). reality. Mike has been a leading physicist defending and We propose to lead the development of the future advancing this case. At the beginning of FY14 sPHENIX upgrade of the PHENIX detector, opening up a new had gone through an initial positive science review by research direction for LANL’s Nuclear Physics Program, the Department of Energy (DOE) and a committee of which will go beyond the current capabilities of the DOE- experts. In preparation for a final review, Mike and a funded program. The current DOE-funded Heavy Ion small team of collaborators created a conceptual design Physics program in P-25 has focused on higher energy for charged particle tracking device capable of precision muon detection through a magnetic spectrometer at Upsilon reconstruction, a mission need raised during the forward angle, and using a silicon vertex detector the earlier DOE review process. He also developed ad- to separate primary muon from secondary decays. Our ditional particle jet observable projections and assisted work in this project will go beyond this limitation by ini- with the development of a particle-flow jet algorithm for tiating a new set of detector systems (sPHENIX) which is sPHENIX. The result of these efforts was a much im- designed specifically for high energy jet measurement. proved sPHENIX proposal document (arXiv:1501.06197). Therefore, technical schemes and scientific goals of this This new proposal was reviewed by the DOE committee LDRD-PRD project are beyond existing DOE funded non- in April 2015, during which Mike defended the esti- LDRD projects. mations of jet backgrounds and rejection techniques, features of the sPHENIX that are key to its ultimate suc- 657

Page 658

cess for measuring modified jets in the collisions of large pre-shower detector installed this year. nuclei. The DOE review committee subsequently endorsed the sPHENIX project stating the science case as “clear and Impact on National Missions: convincing” and that sPHENIX would “answer [science Mike’s scientific research efforts in LANL’s P-25 group have questions] of great significance not only to QCD and heavy put LANL in the forefront as one of the world’s leading ion physics, but to the larger nuclear and particle physics group in the field of Experimental High Energy Nuclear communities.” This represents a major milestone for the Physics. The research direction that Mike is currently lead- project allowing the technical design phase of the experi- ing clearly opens up several new fronts and new opportu- ment to fully begin. nities for US’s Nuclear Physics program of the DOE’s Office of Science, and especially for LANL to play leading roles in Mike has continued to defend the sPHENIX program in sci- nuclear physics for many years to come, especially in the entific conferences, presenting the case at the APS Division proposed future sPHENIX upgrade project. LANL is very of Nuclear Physics meeting, the APS Hadron Topical Group fortunate to have Mike as a contributing and productive meeting. He was invited this year to present on sPHENIX employee. bottom-quark jet observables at the RHIC & AGS Users meeting. He’s been invited to present the sPHENIX science Future Work case at the International Conference for New Frontiers in This LDRD-PRD project focuses on a new research area of Physics in August 2015. These presentations have been Heavy Ion Physics (using particle jet production as a probe invaluable in creating broad community support for the to study the properties of Quark-Gluon Plasma) and lead- program and bringing in new collaborators on the sPHENIX ing the development of the future upgrade of the PHENIX project. This wide spread support was recognized during detector. Our research will open up new and exciting the US Nuclear Physics Long Range Planning process where research directions for LANL’s Nuclear Physics Program, the sPHENIX program was unanimously voted to be the which is already striving to move beyond the existing capa- future direction at the Hot QCD town hall meeting in Fall bilities of the currently funded DOE program. The current 2014. DOE-funded Heavy Ion Physics program in P-25 has been focusing on higher energy muon detection through a mag- Mike is continuing the planning effort for the charged netic spectrometer at the forward angle and using a silicon particle tracking for sPHENIX by developing a much de- vertex detector to separate primary muon from second- tailed technical design. He is pursuing support to ensure ary decayed muons. During FY15, our work on this project a tracking system to be constructed that can fully exploit surpassed this limitation by initiating a new set of detector the opportunities for heavy flavor jet identification at RHIC. systems (sPHENIX) which is designed specifically for high Mike is also participating in the technical design of the energy jet measurement. During FY16, our work will focus hadronic calorimeter detectors of sPHENIX by joining in on providing a more sophisticated physics simulation pack- detector workshops such a one held at Georgia State in De- age for bottom and charge quark jet physics, and making cember 2014. Associations with facilities and experiment- iterations on details of the hadron calorimeter and charge ers through the LANL will result in radiation testing of SiPM particle tracking detector designs. Therefore, the techni- prototypes for the sPHENIX program later this year. Other cal schemes and scientific goals of this LDRD-PRD project ongoing activities include the organization and forma- enable us to push the research boundaries beyond the tion of an sPHENIX collaboration and additional sPHENIX existing DOE-funded non-LDRD projects. simulation workshops to complete the tools need for the technical design effort. During the development of the sPHENIX design, the PHE- NIX experiment gathered interesting data during FY15. In February 2015 Mike completed a 3-year term as PHE- As a physics working group convener in PHENIX for hard NIX physics working group convener on hard-scattering in scattering physics, Mike will direct ongoing analysis and heavy ion collisions. In his role as convener he organized publication of the new research directions on topics includ- weekly physics working group meetings, prepared PHENIX ing modified jet fragmentation, heavy quark energy loss, results for presentation at APS, JPS, DNP, Hard Probes, and quarkonia production. This part of the research effort Quark Matter, and many other conferences, and shepherd- is not on the original PHENIX experiment’s physics plan ap- ed many analyses into publication in PRC, PRD and PRL. proved by DOE, therefore, Mike’s research work will push He now serves as on the PHENIX speakers bureau, which the physics frontier of PHENIX experiment much beyond is responsible for selecting the speakers to represent the the DOE funded projects. collaboration at conferences. He also continues to support PHENIX data taking and was responsible for successfully sPHENIX detector at RHIC: in FY16, we will provide leader- expanding the PHENIX data acquisition system for the new 658

Page 659

ship of the detector research and development needed for 2013. NUCLEAR PHYSICS A. 904: 154C. the construction of a jet-focused experiment at Relativistic McCumber, M. P.. High p(T): Energy Loss Physics at PHENIX. Heavy Ion Collider (RHIC) covering both mid- and forward- 2013. NUCLEAR PHYSICS A. 904: 154C. rapidity in preparation for an Electron-Ion Collider (EIC) detector. Nagle, J. L., Adare, Beckman, Koblesky, J. O. Koop, Mc- Glinchey, Romatschke, Carlson, J. E. Lynn, and McCumber. Conclusion Exploiting Intrinsic Triangular Geometry in Relativistic He-3 A key piece of physics necessary to obtain in the desired

• Au Collisions to Disentangle Medium Properties. 2014. data is the energy loss of massive quarks in the Quark- PHYSICAL REVIEW LETTERS. 113 (11). Gluon Plasma (QGP). Mike will take a leadership role in the first direct analysis of heavy quarks at forward rapidity using the Forward Vertex (FVTX) detector (a LANL-led DOE project) from the heavy ion data to be collected in 2014. This work presents a pivotal test of the models describing the interaction of fast moving quarks with the QGP and thus will prepare the field for the suite of measurements to come using the sPHENIX detector. Publications Adare, A. M., M. P. McCumber, J. L. Nagle, and Romatschke. Examination whether heavy quarks carry information on the early-time coupling of the quark-gluon plasma. 2014. PHYSICAL REVIEW C. 90 (2). McCumber, M.. PHENIX Future Plans and Prospects. To appear in International Conference on New Frontiers in Physics 2015. (Kolymbari, Greece, 23-30 Aug. 2015). McCumber, M.. Future b-jet Measurements with sPHENIX. Invited presentation at RHIC and AGS Users Meeting 2015. (Upton, New York, 9-12 Aug. 2015). McCumber, M.. Updated Jet Performance and Algorithm Approaches. Invited presentation at sPHENIX Department of Energy Science Review. (Upton, New York, 30 April 2015). McCumber, M.. An Opportunity for Forward Jet Single Spin Asymmetry Measurements at RHIC. Presented at 6th Workshop on APS Topical Group on Hadron Physics. (Balti- more, Maryland, 8-10 April 2015). McCumber, M.. fsPHENIX: Forward Jet and Drell-Yan Single Spin Asymmetries at RHIC. Presented at Joint Nuclear Phys- ics Division Meeting of the APS and JPS. (Waikoloa, Hawaii, 7-10 Oct. 2014). McCumber, M.. sPHENIX: An Upgrade Proposal from the PHENIX Collaboration. 2015. arXiv:1501.06197. McCumber, M. P.. Back-to-back pair suppression at large transverse momentum in root s(NN)=200 GeV Au + Au col- lisions at PHENIX. 2011. NUCLEAR PHYSICS A. 855 (1): 408. McCumber, M. P.. High p(T): Energy Loss Physics at PHENIX. 659

Page 660

Nuclear and Particle Futures Postdoctoral Research and Development Continuing Project New Tools to Probe Matter with an Electron-Ion Collider Christopher Lee 20140671PRD2 Introduction Benefit to National Security Missions Protons and neutrons (hadrons) were once thought to The proposed research is directly aligned with LANL and be indivisible elementary constituents of ordinary mat- national priorities in physics at a future electron-ion ter. One of the greatest discoveries of the 20th Century, collider (EIC), a machine that would probe the structure through electron-proton collisions, was that hadrons of the proton and the nature of the strong interaction in themselves are made of smaller “quarks.” We also unprecedented detail. The proton is a ubiquitous com- discovered that quarks are bound together by gluons, ponent of all ordinary matter in the universe, and the which transmit the force called the strong interaction. strong interaction is what binds the constituents of the Now in the 21st Century, the focus of US nuclear physics proton (quarks and gluons) together and protons and is to map out the structure of this dynamic system in un- neutrons to each other in nuclei. Understanding protons precedented detail at a new Electron Ion Collider (EIC), and the strong interaction is key to our fundamental which will collide high-energy electrons and protons to understanding of Nature. Central to this proposal are probe their internal structure. Crucial measurements will development and application of Soft Collinear Effective include the strength of the interaction between quarks Theory (SCET) to provide the high precision predictions and gluons, called the strong coupling, and the 3-D mo- necessary to interpret the unprecedentedly detailed mentum distribution of all the different kinds of quarks data on nucleon structure that will come from EIC. The and gluons within the proton. One treasure trove of in- proposed calculations will allow us to use EIC data to formation in EIC collisions will be hadronic jets, collimat- resolve current disagreements in measurements of the ed bunches of strongly interacting particles produced by strong coupling, a fundamental constant affecting all the collisions. In the last decade, a powerful theoretical studies of nuclear matter (e.g. LHC, RHIC). EIC physics tool called soft-collinear effective theory (SCET) has led and SCET are high priorities under the LANL Nuclear & to unprecedentedly precise predictions for jet produc- Particle Futures Pillar, under the research thrust “Ad- tion, including determination of the strong coupling to vancing our understanding of QCD and the fundamental incredible precision < 1% in electron-positron collisions. properties of nuclear matter.” The 2007 NSAC Long- However, the value extracted this way disagrees with nu- Range Plan for DOE identifies this research as highest merical simulations of strong interactions by more than priority: “The EIC embodies the vision of our field for three standard deviations, a puzzle urgently demanding reaching the next QCD frontier,” further emphasizing, resolution. “It is essential that theoretical support for EIC-related physics is maintained at a healthy level.” The proposed We will push the frontier of precision jet physics at the research will strengthen LANL and US leadership in this EIC using SCET. Together with the EIC, a concomitant exciting international endeavor. effort to improve accuracy of theoretical predictions is essential. Our preliminary work on 2-jet production in Progress electron-proton collisions has led to new methods to In November, together with Iain Stewart at MIT, Dae- compute jet cross sections to any desired accuracy. We kyoung and I published a paper in the Journal of High- will use this power to compute EIC jet cross sections to Energy Physics on “Analytic Calculation of 1-Jettiness in the highest precision to date, allowing us to use future DIS at O(alpha_s)” (JHEP 1411, 132) on the fixed-order EIC data to resolve the existing discrepancies in deter- computation in Quantum Chromodynamics (QCD) of the mining the strong coupling. 1-jettiness distribution in Deep Inelastic Scattering (DIS) to first order in the strong coupling constant alpha_s). 660

Page 661

This enables a more precise prediction of the 1-jetti- In parallel with the main thrust of his work on QCD predic- ness (tau_1) distribution at large values of tau_1, where tions for EIC physics, Daekyoung is also pursuing applica- fixed-order perturbation theory is more important that tions of effective field theory to the physics of bound states resummed perturbation theory. This is an important step in systems of particles with large scattering length, in par- towards predicting DIS jet cross sections to a precision suf- ticular in applications to dark matter physics. Daekyoung ficient to perform percent-level extractions of alpha_s from is seeking to answer the question of whether bound state experimental measurements. effects need to be accounted for in dark matter detection experiments. This work is outside the direct scope of this In April, together with PhD student O. Labun at Arizona, proposal, but constitutes a broader, high-impact applica- Daekyoung and I completed a paper preprint submitted to tion of similar effective field theory methods used in both arXiv (1504.04006) which is being accepted by Physics Let- parts of Daekyoung’s research. ters B for publication on “Equality of hemisphere soft func- tions for e+e-, DIS, and pp collisions at O(alpha_s^2)” on Conclusion two-loop, or second order in the strong coupling alpha_s, The goals of this project include predicting cross sections soft functions that contribute to event shape distributions to produce two hadronic jets at EIC to the highest precision in the three types of collisions. These functions describe to date, and so reduce theoretical uncertainty by up to an the effect of soft gluon radiation on jet cross sections in order of magnitude, and then to predict multi-jet produc- these collisions and are an essential ingredient to achieve tion in EIC collisions, which exhibit even greater sensitivity next-to-next-to-leading logarithmic (NNLL) and NNNLL ac- to the strong coupling, alpha_s. These results will make curacy in resummed perturbation theory for event shape possible the most precise extractions of alpha_s from DIS distributions. Since the e+e- (electron-positron) 2-loop soft data and shed light on the current tension among different function was already known, our proof of equality of the methods of extraction. alpha_s is a fundamental constant e+e-, DIS, and pp (proton-proton) soft functions immedi- of nature and affects all predictions at colliders involving ately fills our previous gap in knowledge of the soft func- the strong interaction. tions for DIS and pp. Publications Future Work Kang, D.. DIS Event Shapes at N3LL. Invited presentation We plan to predict cross sections in Deep Inelastic Scat- at DIS 2015: XXIII International Workshop on Deep- tering (DIS) of electrons and protons, differential in one Inelastic Scattering and Related Subjects. (Dallas, TX, of several event shape variables known as “N-jettiness”. 27 Apr - 1 May 2015). These measure the degree of collimation of final state had- Kang, D., C. Lee, and I. W. Stewart. Analytic Calculation of rons into N energetic “jets” in addition to radiation along 1-Jettiness in DIS at O(alpha_s). 2014. Journal of High- the beam direction. Energy Physics. 11: 132. We will use our results obtained in the last fiscal year Kang, D., O. Z. Labun, and C. Lee. Equality of hemisphere on two-loop soft functions together with progress in the soft functions for e+e−, DIS and pp collisions at literature on two-loop beam functions to complete our O(alpha_s^2). Physics Letters B. 748. prediction of DIS 1-jettiness cross sections to unprecedent- ed next-to-next-to-next-to-leading logarithmic (NNNLL) accuracy in resummed perturbation theory in Quantum Chromodynamics (QCD), a calculation which is already well under way. We will also compute numerically the fixed- order part of the distributions to two-loop order, going beyond the one-loop results we obtained last year. We plan to explore generalization of our proof of equality of two-loop soft functions in electron-positron (e+e-) colli- sions, DIS, and proton-proton (pp) collisions to three loop order. We will also pursue the computation of multi-jet cross sections with N > 1. With all these predictions in hand, we will begin to deter- mine the best strategy to apply them to the extraction of the strong coupling at an Electron Ion Collider (EIC). 661

Page 662

Nuclear and Particle Futures Postdoctoral Research and Development Continuing Project Electric Dipole Moments of Hadrons from Lattice Quantum ChromoDynamics Vincenzo Cirigliano 20140673PRD2 Introduction Benefit to National Security Missions The “Standard Models” of elementary particles and The experimental search for Electric Dipole Moments cosmology cannot account for a number of key features of the neutron and light nuclei is a recognized priority of our observed universe. Among these, most promi- of the DOE/SC Office of Nuclear Physics. This theoreti- nently the two standard models do not account for the cal project will elucidate the relationship between the excess of matter over antimatter in the universe (i.e. why experimental findings and the underlying interactions there are stars and stardust instead of just radiation, i.e. that break the so-called CP-symmetry, i.e. the symmetry photons). that interchanges matter and antimatter. The relation between experimental results and underlying sources of An important clue towards the resolution of this puzzle CP-violation is characterized by a handful of coefficients would arise by the discovery and characterization of new (matrix elements), that are currently uncertain at the physics that violates the symmetry exchanging particles order-of magnitude level. Our theoretical calculations and antiparticles (known as CP). A spectacular manifes- will reduce the current uncertainty to the level of 20- tation of this symmetry breakdown would be the exis- 30%, thus greatly sharpening the interpretation of cur- tence of a permanent Electric Dipole Moment (EDM), i.e. rent experimental searches, in which DOE/SC is making a deviation from spherical symmetry, in certain atomic considerable investments. nuclei and their constituents, proton and neutron. Progress The overall goal of this theoretical project is to quantita- Electric dipole moments (EDMs) of elementary particles tively improve the relation between the observation (or are an extremely sensitive probe of violation of time- non observation) of a nuclear EDM and the mechanism reversal (T), or, equivalently, of the product of charge and size of CP symmetry breakdown. The symmetry conjugation and parity (CP) beyond the Standard Model. breakdown happens at the level of elementary constitu- An observation of an EDM in the next generation of ents of neutron and proton, called quarks, that interact experiments will be a clear indication of new physics, with each other very strongly. Therefore, in order to due either to a minute value of the QCD $$\begin{gathered} bar \\ theta \end{gathered}$$ learn anything from EDM searches, one needs to per- angle, or to new particles and interactions appearing at form complex calculations of how CP symmetry breaking very high scales, much larger than the electroweak scale. interactions alter the neutron, proton, and their interac- However, the interpretation of results of EDM experi- tions in nuclei. We propose to perform such calculations ments, and the connection to complementary probes at using the only first-principle and improvable approach collider experiments, is hampered by the poor knowl- to deal with the theory of quark strong interactions, edge of the matrix elements of $CP$-odd operators known as “lattice Quantum ChromoDynamics (QCD)”. between nucleon, and nucleon and pion states. The key observation at the basis of this project is that the properties of QCD simplify the task of calculating In the past year, this project has focused on improving these CP-breaking interactions by relating them to shifts the tools for the determination of these nucleon matrix in the masses of neutron and proton, that are relatively elements. Their calculation from first principles in- simpler to calculate. Our calculation will reach a preci- volves physics at a scale at which QCD becomes strongly sion of 20-30%, a considerable improvement over the coupled, and is an inherently non-perturbative problem. current order-of-magnitude (factor of ten) uncertainties. Lattice QCD currently offers the most promising oppor- tunity to perform systematically improvable evaluations 662

Page 663

of the matrix elements. Operators induced by physics Future Work beyond the Standard Model (BSM), like the quark electric In the next fiscal year, we plan to complete some of the and chromo-electric dipole moments, are also starting to steps that are necessary to express the pion-nucleon cou- receive attention, with important contributions from the plings induced by the qCEDM in terms of the coefficients LANL lattice group. In order to make contact with phenom- of these operators at high energy. Currently, the lattice enology, and with possible manifestations of the effects of group at JLAB is pursuing the evaluation of the corrections CP-odd operators at high energy, it is necessary to convert to the nucleon and Goldstone boson masses induced by any results of lattice simulations to the same scheme that the quark chromo-magnetic dipole moment (qCMDM), the is used in perturbative calculations. This is a non-trivial chiral partner of the qCEDM. In order to use these results, task, which requires the definition of a renormalization the careful definition of a renormalization scheme that scheme which can be implemented both in numerical lat- does not violate the chiral Ward identity relating chromo- tice simulations, and in perturbation theory. We defined electric and chromo-magnetic operators is necessary, and such a scheme, and studied the mixing and renormaliza- this requires the extension of our recent work to include tion of flavor-diagonal dimension-5 T- and P-odd operators dimension-5 CP-even operators. In addition, it is extremely involving quarks, gluons, and photons, including quark important to assess the robustness of the relation be- electric dipole (qEDM) and chromo-electric dipole opera- tween spectroscopy and T-odd couplings in the case of the tors (qCEDM). Once the lattice matrix elements will be- qCEDM, come available, our result will immediately allow to bridge the gap between lattice and phenomenology. The results Another interesting direction that we intend to pursue of this investigation have appeared as a preprint. is the investigation of the complementarity of new phys- ics searches at collider and low energy experiments. The For systems more complicated than the nucleon, like deu- CP-violating operators that induce EDMs at low energy teron, 3-He, or 199-Hg, a very important role is played by also modify the interactions of the Higgs, Z and W bosons, another class of couplings, T-odd pion-nucleon couplings, in a way that can be directly tested at colliders. We have which mediate long-range T-violating nucleon-nucleon started to compare the bounds on CP-odd operators from interactions. In this case, a direct lattice QCD calculation is EDMs to bounds coming from the Higgs and vector bosons more difficult, and has not been attempted yet. However, production cross sections measured at the LHC. We plan in many interesting cases the approximate symmetry of to complete our analysis in the next few months, and to QCD under chiral transformations allows to relate the T- determine for which class of operators EDM and collider odd pion-nucleon couplings to modifications of the meson bounds are competitive, and how they can be improved in and baryon spectrum induced by the CP-even chiral part- the next LHC runs. ners of CP-odd operators. For example, the isoscalar T-odd coupling induced by the QCD theta term is related to the Conclusion proton-neutron mass difference induced by the quark In the next few years, several experiments with strong mass splitting. Lattice QCD is extremely well suited for LANL involvement will significantly improve the sensitiv- the calculation of spectroscopic properties of meson and ity to Electric Dipole Moments (EDM) of the neutron and baryons, and indeed available lattice evaluations of the nuclei, and, hopefully, discover new physics. Our results mass splitting allow to extract the pion-nucleon coupling will be essential in order to relate the experimental results with 10% uncertainties. Similar relations can be derived for to the underlying sources of charge conjugation and parity the qCEDM, and offer a very promising route to determine (CP) symmetry breaking (CP is the symmetry that relates couplings that at the moment are known with very large matter and antimatter). This will allow a more robust inter- uncertainties. The relations between baryon spectrum pretation of EDM signals, and a more direct connection to and pion-nucleon couplings have been proven at leading BSM models, with impacts in the fields of nuclear physics, order in the chiral expansion. With collaborators, we have particle physics, and cosmology. started an investigation on the robustness of these rela- tions, including subleading corrections in Chiral Perturba- Publications tion Theory, the effects of the strange quark and of the Chien, Y. T., V. Cirigliano, W. Dekens, J. de Vries, and E. decuplet baryons. We just completed the analysis for the Mereghetti. Direct and indirect constraints on CP- QCD theta-term, in which case we showed that the cor- violating Higgs-quark and Higgs-gluon interactions. rections are small, and the relation robust. Next, we plan Journal of High Energy Physics. to extend our discussion to the qCEDM. The results have Vries, J. de, E. Mereghetti, and A. Walker-Loud. Baryon appeared in a preprint. mass splittings and strong CP violation in SU(3) Chiral Perturbation Theory. 2015. Physical Review C. 92: 4. 663

Page 664

Nuclear and Particle Futures Postdoctoral Research and Development Continuing Project Multi-wavelength Studies of Explosive Astrophysical Transients Przemyslaw R. Wozniak 20140674PRD3 Introduction Benefit to National Security Missions High energy transients producing photons in the GeV- Tilan Ukwatta will perform research to establish the TeV energy range represent a largely unexplored phase origin of very high-energy photons in γ-ray bursts and space in transient phenomena. To study these extreme explain the broad-band spectra of explosive astro- phenomena one needs a sensitive (large) detector, wide physical transients. To accomplish these goals, he will field-of-view and near continuous operations such as the cross-correlate real-time data streams from the RAPTOR High Altitude Water Cherenkov (HAWC) observatory. As- telescopes at LANL with data from the Swift satellite and trophysical transients such as gamma-ray Bursts (GRBs), HAWC (High Altitude Water Cherenkov) detector. For this Active Galactic Nuclei (AGN), or high-mass X-ray binaries purpose Tilan will utilize and extend LANL’s unique capa- can emit across the entire electromagnetic spectrum bilities in the area of telescope networking and continu- from radio to high-energy gamma rays. ous sky monitoring. He will also develop new transient search and classification algorithms and software for Rapid multi-wavelength observations are key to under- massive time-domain sky surveys. standing the underlying mechanisms of particle ac- celeration and localizing emission regions. Swift is the This research is well aligned with the long-term vision for only X-ray observatory that can respond quickly enough Global Security and Threat Identification and Response to HAWC transients and cover the HAWC error box to to develop cutting edge Information Science and Tech- observe any contemporaneous lower-energy counter- nology capability for national security. This include real- parts before they fade away. RAPTOR telescopes at LANL time, multi-source data fusion tools which have long- are uniquely equipped to rapidly (< 6 sec) cover the term implications for nuclear nonproliferation and other HAWC error box in the optical and catch the associated intelligence, surveillance, and reconnaissance missions, afterglow emission. Our approach is to automatically and wide-field sensor approaches relevant to space situ- alert the most significant transients detected by RAP- ational awareness. TOR, Swift, or HAWC, and immediately follow them up. For this purpose we will utilize our approved Swift Guest Progress Investigator Target of Opportunity program and the RAP- In the first year of the project we developed a transient TOR network. The fast slewing capability of RAPTOR will management system for High Altitude Water Cheren- enable follow-up of sub-threshold triggers from HAWC. kov (HAWC) observatory that accepts gamma-ray burst In the case of a detection, we will use Swift to perform (GRB) alert notices and sends them to the online data additional observations in X-rays. analysis system. The system stores GRB alerts and filters GRBs observable by HAWC. It also tracks transient search To accomplish these goals we will develop new software results from HAWC. This system provides an easy to use and the instrument networking framework for real-time web interface allowing observers to view and respond to correlations of data arriving simultaneously from multi- HAWC transients. HAWC started full detector operations ple instruments. Another important advance will be new in March 2015 and is in the process of verifying proper transient search and classification algorithms for massive detector operations and data quality. HAWC transient time-domain sky surveys such as the Palomar Transient searches have begun collecting transient alerts and we Factory (PTF) and the Large Synoptic Survey Telescope are developing real time alert distribution system for (LSST). In the future, this technology will enable similar HAWC and RAPTOR based on live network socket con- studies on much larger data sets. nections implemented using zeromq software library. 664

Page 665

Tilan Ukwatta is leading a research group investigating We are developing machine learning techniques to iden- physical properties of evaporating primordial black holes tify high-redshift (or high-z) GRBs using promptly avail- (PBH). The group includes two high energy theorists, able multi-wavelength data from the Swift observatory. statistician, a postdoc and a graduate student (Michigan This work includes both classification and regression and State University and University of North Florida). The main in the accompanying paper we introduces the new term result in the past year is a detailed study of the sensitiv- “machine-z” which is the redshift derived from machine ity of HAWC to detect high-energy bursts from PBHs and learning techniques to complement two other redshift new methods that can be used to detect those exotic measurements: photo-z and spectroscopic z measure- transients. We have already published one paper in As- ments. Unlike the alternative methods, our algorithm is troparticle Physics journal titled “Milagro limits and HAWC capable of catching almost all high-z GRBs on average with sensitivity for the rate-density of evaporating Primordial limited follow-up resources. Black Holes” (LA-UR-15-20829) and another paper is “Ob- Tilan Ukwatta is working with Intermediate Palomar Tran- servational Characteristics of Evaporating Primordial Black sient Factory (iPTF) team to discover and study superno- Holes” (LA-UR-15-23058) is almost ready to be submitted vae. He already discovered three supernovae as a scanner. for publication. In addition, we have presented our work at Tilan also performs regular shift duty as the Swift “Burst the “5th Fermi Symposium” and American Physical Society Advocate (BA)” and Swift “Burst Alert Telescope (BAT) Burst Meetings. We will also present our work at the 34th Inter- Scientist (BBS)”. national Cosmic Ray Conference in July 2015. For his work Tilan Ukwatta received The HAWC Excellence We developed supernova.lanl.gov, an online database of Award with the citation “Awarded in recognition of ex- time-resolved spectra and multi-color light curves from ceptional and sustained contributions to HAWC Remote simulations of supernova explosions. This system allows Monitoring”. users to compare their supernova data with supernova ex- plosion models developed by the LANL supernova theory Future Work group. System has ability to analyze both spectra and light curves obtained with a variety of broad-band filters com- The main goal for the project is to develop a preliminary monly used by modern sky surveys. version of the framework to cross-correlate transient events from RAPTOR, Swift, and HAWC. Triggering criteria Tilan Ukwatta began a new collaboration with Kevin Hurley for Target of Opportunity (ToO) observations of HAWC from University of California, Berkeley to develop a novel events must be designed, coded, and integrated with event method to constrain distances of gamma-ray bursts (GRBs) messaging and database systems. New data ingest and using multi-satellite triangulation with the Interplanetary event translation routines must be developed for anticipat- Network (IPN). So far we have constrained the minimum ed transient events from all three instruments. We plan to distances of 49 short duration GRBs that could in fact be perform cross-correlations of data streams from RAPTOR, PBH bursts. Our results indicate that using this technique Swift, and HAWC to provide enough triggers for executing we can set upper limits on the rate of PBH that are two ToO Swift observations. We will also develop a software orders of magnitude better than the current best limit. pipeline to perform a blind search for intermediate dura- A corresponding paper titled “Investigation of Primordial tion HAWC transients lasting days to weeks. Black Hole Evaporation Rates from a Sample of Interplane- tary Network GRBs” will be submitted to the Astrophysical Another goal is to perform iterative probabilistic classifi- Journal in the following weeks. cation of transients based on heterogeneous data such as time-resolved multi-color photometry, spectra, and A paper on redshift and duration dependent clustering in possible X-ray/gamma-ray detections. We will develop a the GRB sky distribution is in the final stages. This investi- mathematical toolbox based on Bayesian networks that gation found that previous claims of clustering at redshift can go beyond classification problems constrained to a z ~ 2 are most likely due to statistical fluctuations and not major class of transients such as Supernoave and instead a physical overdensity. However, we found more evidence handle a diverse set of transients. for clustering of very short GRBs with duration less than 100 milliseconds. The title of the paper is “Investigation Finally, we will model and interpret interesting transients of Redshift and duration Dependent Clustering in the observed over the course of the year. We will fit the ob- Gamma-ray Burst Sky Distribution” (LA-UR-15-24559) and served light curves and time-resolved spectra with blast we are planning to publish it in the Monthly Notices of the wave models to identify the relevant emission components Royal Astronomical Society (MNRAS) journal. and constrain their parameters. Using the relative strength and timing of those components we will constrain possible 665

Page 666

emission sites and shock propagation scenarios. Zepeda, and Zhou. Milagro limits and HAWC sensitiv- ity for the rate-density of evaporating Primordial Black We will finish developing and tuning our “machine-z” Holes. 2015. ASTROPARTICLE PHYSICS. 64: 4. redshift (distance) estimator for gamma-ray bursts (GRBs). Abeysekara, A. U., Alfaro, Alvarez, J. D. Alvarez, Angeles, We plan to implement the technique in the Swift auto- Arceo, J. C. Arteaga-Velazquezd, Avila-Aroche, H. A. A. mated pipeline so machine-z values will be available to the Solares, Badillo, A. S. Barber, B. M. Baughman, Bautis- GRB community for all GRBs newly discovered by the Swift ta-Elivar, J. B. Gonzalez, Belmont, Benitez, S. Y. BenZvi, satellite. Berley, Bernal, M. B. Rosales, Braun, R. A. Caballero- Lopez, K. S. Caballero-More, Cabrera, Carraminana, We will also search HAWC data for hypothetical TeV Castaneda-Martinez, Castillo, Cotti, Cotzomi, de la transients caused by Primordial Black Holes (PBH) that Fuente, De Leon, DeYoung, Diaz-Azuara, Diaz-Cruz, R. evaporate by Hawking radiation using techniques that we D. Hernandez, J. C. Diaz-Velez, B. L. Dingus, Dultzin, M. developed and published last year. A. DuVernois, R. W. Ellsworth, Fernandez, D. W. Fiori- no, Fraija, Galindo, Garcia-Torales, Garfias, Gonzalez, L. Conclusion X. Gonzalez, M. M. Gonzalez, J. A. Goodman, Grabski, The overarching goal of this project is to perform state Gussert, Guzman-Ceron, Hampel-Arias, J. P. Harding, of the art real-time searches and joined follow-up cam- Hernandez-Cervantes, C. M. Hui, Huentemeyer, Imran, paigns of flaring GeV-TeV emitters using RAPTOR, Swift Iriarte, Karn, Kieda, G. J. Kunde, Langarica, Lara, Lara, R. J. Lauer, W. H. Lee, Lennarz, H. L. Vargas, E. C. Lin- and HAWC instruments. For this purpose we must develop ares, J. T. Linnemann, Longo, Luna-Garcia, Marinelli, new software tools and the instrument networking frame- L. A. Martinez, Martinez, Martinez, Martinez-Castro, work for real-time cross-correlations of heterogeneous Martos, J. A. J. Matthews, McEnery, E. M. Torres, data streams. Our joined RAPTOR/Swift transient localiza- Miranda-Romagnoli, Moreno, Mostafa, Nava, Nel- tions and HAWC observations are likely to deliver the first len, Newbold, Noriega-Papaqui, Oceguera-Becerra, ground-based detection of a GRB. Combined with follow- D. P. Page, Patricelli, Pelayo, E. G. Perez-Perez, Pretz, up studies and distance measurements, these observations Ramirez, Renteria, Riviere, Rosa-Gonzalez, Ruiz-Sala, will cover photon energies from eV to TeV, providing strong E. L. Ruiz-Velasco, Ryan, J. R. Sacahui, Salazar, Salesa, constraints on emission mechanisms that power these Sandoval, Santos, Schneider, Silich, Sinnis, A. J. Smith, extraordinary objects. K. S. Woodle, R. W. Springer, Suarez, Taboada, Tepe, P. A. Toale, Tollefson, Torres, Tinoco, T. N. Ukwatta, J. F. Publications V. Galicia, Vanegas, Vazquez, Villasenor, Wall, Weisgar- ber, Westerhoff, I. G. Wisher, Wood, G. B. Yodh, P. W. Abdo, A. A., A. U. Abeysekara, Alfaro, B. T. Allen, Alvarez, J. Younk, Zaborov, Zepeda, and Zhou. VAMOS: A path- D. Alvarez, Arceo, J. C. Arteaga-Velazquez, Aune, H. A. finder for the HAWC gamma-ray observatory. 2015. A. Solares, A. S. Barber, B. M. Baughman, Bautista-Eli- ASTROPARTICLE PHYSICS. 62: 125. var, J. B. Gonzalez, Belmont, S. Y. BenZvi, Berley, M. B. Rosales, Braun, R. A. Caballero-Lopez, K. S. Caballero- Abeysekara, A. U., Alfaro, Alvarez, J. D. Alvarez, Arceo, J. C. Mora, Carraminiana, Castillo, G. E. Christopher, Cotti, Arteaga-Velazquez, H. A. A. Solares, A. S. Barber, B. M. Cotzomi, De la Fuente, De Leon, DeYoung, R. D. Her- Baughman, Bautista-Elivar, S. Y. BenZvi, Bonilla Rosales, nandez, K. L. Diaz-Cruz, J. C. Diaz-Velezi, B. L. Dingus, Braun, K. S. Caballero-Mora, Carraminana, Castillo, Cot- M. A. DuVernoisi, R. W. Ellsworth, D. W. Fiorinoi, Fraija, ti, Cotzomi, de la Fuente, De Leon, DeYoung, Diaz Her- Galindo, Garfias, M. M. Gonzalez, J. A. Goodman, Grab- nandez, B. L. Dingus, M. A. DuVernois, R. W. Ellsworth, ski, Gussert, Hampel-Arias, J. P. Harding, Hays, C. M. D. W. Fiorino, Fraija, Galindo, Garfias, M. M. Gonzalez, Hoffman, C. M. Hui, Huintemeyer, Imrani, Iriarte, Kam, J. A. Goodman, Gussert, Hampel-Arias, J. P. Harding, Kieda, B. E. Kolterman, Kunde, Lara, Lauer, W. H. Lee, Huentemeyer, C. M. Hui, Imran, Iriarte, Karn, Kieda, Lennarz, H. L. Vargas, E. C. Linares, J. T. Linnemann, G. J. Kunde, Lara, R. J. Lauer, W. H. Lee, Lennarz, Leon Longo, Luna-Garcla, J. H. MacGibbon, Marinelli, S. S. Vargas, J. T. Linnemann, Longo, Luna-Garcia, Malone, Marinelli, Martinez, Martinez, Martinez-Castro, J. A. J. Marinelli, S. S. Marinelli, Martinez, Martinez, Martinez- Matthews, McEnery, E. M. Torres, A. I. Mincer, Miran- Castro, J. A. J. Matthews, Mendoza Torres, Miranda-Ro- da-Romagnoli, Moreno, Morgan, Mostafa, Nellen, Ne- magnoli, Moreno, Mostafa, Nellen, Newbold, Noriega- methy, Newbold, Noriega-Papaqui, Oceguera-Becerra, Papaqui, T. O. Oceguera-Becerra, Patricelli, Pelayo, E. Patricelli, Pelayo, E. G. Perez-Perez, Pretz, Riviere, G. Perez-Perez, Pretz, Riviere, Rosa-Gonzalez, Salazar, Rosa-Gonzalez, Ruiz-Velasco, Ryan, Salazar, Salesa, San- F. S. Greus, Sandoval, Schneider, Sinnis, A. J. Smith, doval, P. M. S. Parkinson, Schneider, Silich, Sinnis, A. J. K. S. Woodle, R. W. Springer, Taboada, Tollefson, Tor- Smith, Stump, K. S. Woodle, R. W. Springer, Taboada, res, T. N. Ukwatta, Villasenor, Weisgarber, Westerhoff, P. A. Toale, Tollefson, Torres, T. N. Ukwatta, Vasileiou, I. G. Wisher, Wood, G. B. Yodh, P. W. Younk, Zaborov, Villasenor, Weisgarberi, Westerhoff, D. A. Williams, Zepeda, and Zhou. SEARCH FOR GAMMA-RAYS FROM I. G. Wisher, Wood, G. B. Yodh, P. W. Younk, Zaborov, 666

Page 667

THE UNUSUALLY BRIGHT GRB 130427A WITH THE HAWC GAMMA-RAY OBSERVATORY. 2015. ASTROPHYS- ICAL JOURNAL. 800 (2). Sonbas, , G. A. MacLachlan, K. S. Dhuga, Veres, Shenoy, and T. N. Ukwatta. GAMMA-RAY BURSTS: TEMPORAL SCALES AND THE BULK LORENTZ FACTOR. 2015. AS- TROPHYSICAL JOURNAL. 805 (2). Ukwatta, T. N., D. Stump, J. H. MacGibbon, J. T. Linnemann, S. S. Marinelli, T. Yapici, and K. Tollefson. Observational Characteristics of the Final Stages of Evaporating Pri- mordial Black Holes. To appear in The 34rd Interna- tional Cosmic Ray Conference (ICRC2015) . (Hague, Netherlands, 30 Jul - 6 Aug 2015). Ukwatta, T. N., D. Stump, J. H. MacGibbon, J. T. Linnemann, S. S. Marinelli, T. Yapici, and K. Tollefson. Sensitivity of HAWC to Primordial Black Hole Bursts. To appear in The 34rd International Cosmic Ray Conference (ICRC2015). (Hague, Netherlands, 30 Jul - 6 Aug 2015). Ukwatta, T. N., D. Stump, J. T. Linnemann, J. H. MacGibbon, S. S. Marinelli, T. Yapici, and K. Tollefson. Primordial Black Holes: Observational Characteristics of The Final Evaporation. Astroparticle Physics Journal. Ukwatta, T. N., and P. R. Wozniak. Investigation of Redshift- and Duration-Dependent Clustering of Gamma-ray Bursts. To appear in Monthly Notices of the Royal As- tronomical Society. 667

Page 668

Nuclear and Particle Futures Postdoctoral Research and Development Final Report 3D Turbulent Magnetic Reconnection Experiments and Simulations Scott C. Hsu 20120768PRD3 Abstract es, dusty plasmas are composed of particles (the dust The collective behavior, like transport, of strongly grains) that can have very large charges – perhaps as coupled plasma under various external fields is a key high as tens of thousands. For this reason, dusty plasmas problem in studying plasma physics processes, although have further exotic behavior; for example, they can crys- it has not been understood. In the past two decades, tallize and form matter more similar to liquids and solids. the collective behavior of strongly coupled dusty plasma Since 1990s, the collective behaviors of dusty plasma without an external field has been studied well. Al- without external fields have been studied well, not only though external fields drastically change the motion of in theories but also in many experiments. However, the each individual particle, the effects of these external properties of such plasmas under external fields are cur- fields on the collective behavior of strongly coupled rently not well understood yet. dusty plasma are still unknown. Using Langevin molecu- Because of the properties of dusty plasmas, we can- lar dynamical simulations, we have been studying the not use standard plasma physics techniques, such as collective behaviors of strongly coupled dusty plasma particle-in-cell or hydrodynamics methods. As such, we under various external fields. In our study, we focus on employ simulations that treat each dust grain explicitly. two types of external fields, which are (1) perpendicular To compare and motivate experiments in the response magnetic fields, and (2) one-dimensional electric sub- of various fields, we simulate the behaviors of two strates. In the simulation with perpendicular magnetic dimensional dusty plasmas with different external fields. fields, we discover that the motion of two-dimensional For example, we can add a force term corresponding to dusty plasma tends to be more superdiffusive under either a magnetic field or a periodic substrate. Here, we stronger magnetic fields. We also find that the shear are interested in some basic properties of such plasmas viscosity changes when the perpendicular magnetic field and will focus primarily on diffusion and waves in these changes, while the changing trend is different for differ- systems. We want to have a big picture of the effects ent dusty plasma conditions. More simulations with dif- of external fields on the collective behaviors of strongly ferent dusty plasma conditions are still needed to study coupled dusty plasmas. this changing trend systematically. In the simulation with a one-dimensional periodic electric substrate with vari- Scientific Approach and Accomplishments ous widths, we find that the diffusion of two-dimension- Our research performed here is primarily computational, al dusty plasma oscillates periodically when the width of with theoretical support. Since standard plasma physics the electric substrate changes. We also investigate that techniques cannot be used in dusty plasmas, we employ the wave properties change dramatically with this one- simulations that treat each dust grain explicitly. We use dimensional electric substrate. Further simulations with the Langevin molecular dynamical simulation to mimic more dusty plasma and electric substrate conditions are the motion of thousands of dust particles within the two still needed to complete this project. dimensional plane. In the equation of motion of each simulated particle, we include various forces, like the Background and Research Objectives particle interaction and gas frictional damping, which are The plasma state of matter fills most of the universe, the same in the dusty plasma experiments. We integrate including the interstellar medium and stellar interiors. the equations of motion for all particles, step by step, Because plasmas are strongly affected by electric and to obtained all particles’ positions and velocities. With magnetic fields, they exhibit quite complex behavior. these particles’ positions and velocities, we can do vari- While the particles in most plasmas have modest charg- 668

Page 669

ous structural and dynamical analyses. to the shear viscosity of this system. Four curves of black, red, blue, and green correspond to the zero, smaller, bigger, and big- In the first project, we added an additional Lorentz force gest perpendicular magnetic fields, respectively. For this dusty term corresponding to a perpendicular magnetic field in plasma condition, these curves indicate that the shear viscosity is the equation of motion. From the particles’ positions and increased as the perpendicular magnetic field increases. velocity, we obtain many kinds of dynamics measures, like In the second project, we add a force term corresponding the mean-squared displacement, velocity autocorrelation to a periodic electric substrate in the equation of motion function, vibrational density of states, phonon spectra of for all particles. A typical plot of trajectories of all dust par- this system. These dynamics measures indicate that, the ticles is shown in Figure 2. To mimic the effects of different two dimensional strongly coupled dusty plasma tends to substrates, we change the width of this periodic electric be more superdiffusive under stronger perpendicular mag- substrate in our simulations. Then, to study the diffusion of netic fields. These results have been published [1]. dusty plasma under this one dimensional periodic electric Besides the diffusion, we have also calculated the shear substrate, we calculate the mean squared displacement viscosity of this system, which has never been studied due to the motion of all particles in the two directions. before. We use the Green-Kubo relation to calculate the Figure 3 shows the obtained mean squared displacement shear viscosity here. In the Green-Kubo relation, the vis- under one dimensional electric substrate with different cosity is the integral of an autocorrelation function of the widths. From Figure 3, we find that the diffusion of this sys- shear stress of the total system. We find that the external tem oscillates periodically as the function of the substrate magnetic field changes the behavior of this autocorrelation width. Although the electric substrate is aligned regularly function significantly, so that the integral of this func- in one direction, the diffusion in both directions shows the tion has also been changed due to this external magnetic same general oscillation trend, which indicates that the field. As shown in Figure 1, the integral of this autocor- motion in two directions is coupled with each other. relation function with bigger magnetic fields (green and blue curves) is also bigger than those with smaller or zero magnetic fields (red and black curves). However, in other dusty plasma conditions, our simulation results indicate a different trend: bigger magnetic fields would reduce this integral. Further simulations with different dusty plasma conditions are still needed to systematically study this changing trend under external magnetic fields. Our current conclusion is that the magnetic field modifies the shear vis- cosity of this system significantly. Figure 2. Trajectories of simulated dust particles within the two-dimensional plane with a one-dimensional periodic electric substrate. Different colors correspond to different times. Al- though most particles stay within the electric substrate potential well, a few of them can jump across the substrate. Figure 1. Autocorrelation functions of the shear stress, calcu- lated from the two-dimensional dusty-plasma simulations with perpendicular magnetic fields. From the Green-Kubo relation, the time integral of this autocorrelation function corresponds 669

Page 670

ing such novel transport processes is crucial for under- standing diffusive mixing in a variety of applications, such as material mixing in experiments performed at a variety of Department of Energy facilities (e.g., Omega, Z, NIF, and LCLS). Figure 3. The mean square displacement of simulated dust particles in two directions as functions of the electric substrate width. The red symbols correspond to the diffusion along the sub- strate, while the black symbols correspond to the diffusion across the substrate. The diffusion in both directions oscillates periodi- cally as the function of the substrate width, and they oscillate Figure 4. The longitudinal wave spectra of two-dimensional roughly in the same phase. This result indicates that the motion dusty plasma without electric substrate. These spectra and the in the two directions is coupled with each other. related dispersion relation have been studied widely in the past in dusty plasmas, from both experiments and simulations. Besides the diffusion, we have also studied the wave spectra in this system. We find that the wave properties also change due to external force fields. Figure 4 shows the longitudinal wave spectra of 2D dusty plasma in this sys- tem without any electric substrate, while Figure 5 shows the modified longitudinal wave spectra, corresponding to the motion in the y direction, due to the applied electric substrate. From these plots, we can see that the one di- mensional electric substrate modified the wave spectra sig- nificanly lot, causing the spectra to split to two branches, and one branch seems to show some optical mode feature. Similar features have also been observed in the transverse wave spectra. Further studies with more simulations, with various dusty plasma and electric substrate conditions, are still needed to confirm our conclusions. Impact on National Missions This work explores the cutting edge area of what is re- Figure 5. The longitudinal wave spectra of two-dimensional ferred to as anomalous properties; that is, properties that dusty plasma with a one-dimensional periodic electric substrate. are not well described by standard models. For example, This electric substrate causes the longitudinal spectra to slit standard model for diffusion is based on Fick’s Law, which to two branches. The fundamental reason for this split is still leads to the standard diffusion equation. However, here unknown. Further studies with more simulations, with various we discover that, in the two dimensional strongly coupled dusty plasma and electric substrate conditions, are still needed to dusty plasma with very strong magnetic fields, the dif- confirm our conclusions. fusion tends to be superdiffusive. Besides diffusion, the shear viscosity behavior is also unknown before our inves- References tigation. The study of the dusty plasma behavior under one Feng, Y., J. Goree, B. Liu, T. P. Intrator, and M. S. Murillo. dimensional electric substrate is a new topic. Understand- Superdiffusion of two-dimensional Yukawa liquids due to a 670

Page 671

perpendicular magnetic field. 2014. Physical Review E. 90: 013105. Publications Feng, Y. a. n., Goree, B. i. n. Liu, T. P. Intrator, and M. S. Mu- rillo. Superdiffusion of two-dimensional Yukawa liquids due to a perpendicular magnetic field. 2014. PHYSICAL REVIEW E. 90 (1). Feng, Y., J. Goree, and B. Liu. Longitudinal viscosity of two- dimensional Yukawa liquids. 2013. PHYSICAL REVIEW E. 87 (1): -. Goree, , B. i. n. Liu, and Y. a. n. Feng. Diagnostics for trans- port phenomena in strongly coupled dusty plasmas. 2013. PLASMA PHYSICS AND CONTROLLED FUSION. 55 (12). Intrator, T. P., Dorf, Sun, Feng, Sears, and Weber. Laboratory observation of magnetic field growth driven by shear flow. 2014. PHYSICS OF PLASMAS. 21 (4). Intrator, T. P., Sun, Dorf, J. A. Sears, Feng, T. E. Weber, and H. O. Swan. Flux ropes and 3D dynamics in the relax- ation scaling experiment. 2013. PLASMA PHYSICS AND CONTROLLED FUSION. 55 (12). Sears, J., Y. Feng, T. P. Intrator, T. E. Weber, and H. O. Swan. Flux rope dynamics in three dimensions. 2014. Plasma Physics and Controlled Fusion. 56: 095022. 671

Page 672

Nuclear and Particle Futures Postdoctoral Research and Development Final Report Boosting New Physics Discoveries with Jet Substructure Christopher Lee 20130794PRD2 Abstract the resulting particles that are produced. The most com- High-energy physics is now in one of the most exciting mon type of SM final state contains collimated streams eras in history as the Large Hadron Collider (LHC) col- of hadrons called jets, produced by radiation from lides protons at unprecedentedly high energies in an high-energy quarks or gluons. On occasion, new, non-SM attempt to create new particles beyond the Standard particles may be produced and decay into jets. A very Model (SM) that could help explain unsolved mysteries difficult needle-in-the-haystack problem is to distinguish of Nature such as the composition of dark matter, the jets produced by new particles from the vastly larger origin of mass, and the huge hierarchy of energy scales number of jets produced by ordinary SM processes. This between gravity, electromagnetism, and nuclear forces. makes high-precision predictions of the SM jet back- Many of these new particles decay to jets of hadrons, ground essential. Unfortunately, jet cross sections are strongly interacting particles made up of confined quarks among the least precisely understood due to the often and gluons. These jets can thus contain signals of new prohibitively difficult nature of the calculations. physics beyond the SM. However, the SM itself produces In this research we aimed to vastly increase the po- an enormous background of jets that can swamp out the tential precision of theoretical jet cross section predic- signals of new physics. This project has developed new tions. The theory of Quantum Chromodynamics (QCD) measures and strategies to probe jet substructure that is highly successful in describing the strong interaction can help distinguish between signal and background jets. amongst quarks and gluons that make up jets, protons, It has developed new, modern theoretical tools that can and neutrons. Using the theory to predict properties of be applied to predict jet and subjet cross sections to the jets precisely, however, is still a formidable calculational high precision and accuracy required to extract signals challenge. In this project, we developed several new cal- of new physics from LHC data. These new predictions culational and strategic tools to predict jet cross sections will serve as input to Monte Carlo event generators that more precisely and mine the structure jets for maximal simulate large numbers of particles at the LHC and that information about their origin. are widely used in data analyses at the LHC. All together, these complementary tools and developments will serve Specifically, we developed strategies to characterize as a comprehensive strategy to identify new particles and probe the substructure of jets, which can serve as produced at the LHC through their decays to hadronic a discriminant between signal and background jets. We jets. applied our expertise in theoretical methods to resum the effects of infinitely many quarks and gluons radiated Background and Research Objectives from jets to predict jet cross sections to high precision The Standard Model (SM) of particle physics is the most and accuracy. We developed new methods to predict precise and successful theory in the history of science, cross sections dependent on jet sizes or radii, never but it still cannot explain dark matter or the huge hier- before accurately predicted. We improved the theoreti- archy of energy scales (from nuclear forces to gravity) cal knowledge that can be put into Monte Carlo event found in nature. Much excitement today comes from generators that simulate large numbers of particles in predictions of new models of elementary particles that high-energy collisions. All of these tools together greatly could help solve these mysteries and be detected at the advance our ability to find evidence of new physics at Large Hadron Collider (LHC). the LHC inside hadronic jets. The LHC collides protons at high energies and records 672

Page 673

Scientific Approach and Accomplishments quark jets, often an important component of signals for Our research accomplishments during the period of this heavier BSM particles decaying to tops [3]. project are in three categories: 1) advanced methods to mine more information from jets by considering multiple jet formation histories; 2) a new effective field theory that can predict the dependence of jet cross sections on the size or radius of jets; and 3) new predictions for a wide variety of jet shape variables in hadron collisions. Qjets (“Quantum” Jets) Postdoc A. Hornig has been developing a method called “QJets” that is designed to optimize the information one obtains when making jet measurements at particle collid- ers such as the Large Hadron Collider (LHC). He invented the method earlier with collaborators S. Ellis (Washington), Figure 1. QJets moves the distribution of masses of QCD D. Krohn and M. Schwartz (Harvard), and former LANL background jets from higher values mu_j to lower value mu_a. postdoc T. Roy [1]. Traditional methods rely on a single This removes much of the background from the mass window in interpretation of how quarks and gluons shower from one which one is searching for jets constituting a signal (for heavy to many particles (and eventually hadrons) to form the jets of SM or BSM origin). [2] jets we measure. By utilizing the fundamentally quantum mechanical nature of this process and allowing for multiple such interpretations, much can be gained. For example, the spread in the masses one obtains, a measure of the ambiguity of a jet’s mass (dubbed the Qjet “volatility”), can be used to distinguish background QCD jets from Stan- dard Model and Beyond the Standard Model signal jets. In addition, the statistical nature of signal searches can be enhanced. During the course of this project, Hornig worked with Ellis, Krohn, and Roy to perform detailed analysis of the statis- tics of QJets and evaluate its quantitative power to im- prove signal to background ratios. The improvement arises because assigning multiple interpretations to a jet done in a particular manner moves the QCD background to smaller mass, while only slightly distorting the signal distribution (negligible on the scale of a jet mass window typically used in experimental searches). This phenomenon is illustrated in Fig. 1. Hornig and collaborators showed that significant improvements in sensitivities to signals and more precise Figure 2. Performance of various algorithms and variables measurements of jet observables can be achieved by using for tagging jets produced by top quarks at the LHC. The QJets QJets. This work was published in the Journal of High-Ener- algorithm (light blue solid line) performs the best of all methods gy Physics [2]. in signal efficiency and reducing backgrounds. [3] Thanks to these demonstrated statistical improvements, New Effective Theory for Jet Radii: QJets has now been incorporated into independent codes All searches for hadronic jets require the use of a jet used by the wider theoretical and experimental commu- algorithm to define which particles in the final state of a nities analyzing LHC data for signals of new physics. The collision belong inside a jet or not. The simplest way is to HEPTopTagger2 program codes an array of tools used to define cones of radius R and define particles to be in a jet mine LHC data for signals of heavy top quark jets and BSM- if they fit inside the cone. An event is a 2-jet event, for produced jets. The latest version of the code now includes example, if most of the energetic particles fit inside two QJets and has demonstrated, as shown in Fig. 2, that it has cones of radius R. A small amount E of the total energy Q the best performance of all observables and algorithms in in the final state can still lie outside the cones. improving signal-to-background ratios in searches for top 673

Page 674

Predicting the dependence of jet cross sections in QCD on Finding ways to characterize this substructure in ways that the parameters R and E/Q accurately has been a long- are theoretically predictable to high accuracy is, however a standing problem. For relatively large values, the problem challenge. is not so hard, but for smaller value, large logarithms of these parameters appear in perturbative predictions for cross sections dependent upon them, spoiling the conver- gence of the perturbative expansions in the strong cou- pling which is supposed to be small. In the past 15 years, Soft Collinear Effective Theory (SCET) has been developed and proven to be a powerful tool providing technology to resum such logarithms to all orders in the strong coupling for other jet observables, restoring convergence and reli- Figure 3. Impact of new effective theory with soft-collinear ability of the perturbative expansion. This had never been degree of freedom on jet thrust. The distribution in jet thrust in done beyond the leading accuracy, however, for jet radii R electron-positron collisions converges better from one order of or the energy fraction E/Q outside jet cones. accuracy to the next when the soft-collinear (s-c) mode is intro- duced into the theory and used to resum large logarithms of the In this project, A. Hornig and C. Lee together with LANL jet radius (left). Without the mode, large unresummed logs of postdoc Y.-T. Chien succeeded in augmenting SCET with give a significantly larger theoretical uncertainty (right). [4] additional degrees of freedom that are able to probe the dependence of cross sections on the jet boundaries, A set of observables that smoothly probe the internal isolate dependence on the scales of the soft radiation of structure of a jet is the set of angularities. These sum over energy E outside the jets, the hard energy Q inside the the particles in a jet with a particular weighting function jets, and the scales ER and QR of soft and hard radiation, of their energy and their angle from the jet axis. Smoothly respectively, that probes the jet boundaries. Degrees of varying this function gives an infinite set of different angu- freedom probing ER had never been introduced into SCET larities. Comparing distributions of final states in different before. We called this the new soft-collinear mode. We angularities reveals much about the structure of the jet. showed that introducing this mode allows resummation Angularities are also predictable to high precision in QCD of all logarithms of R in jet cross sections, allowing for the perturbation theory. This was explicitly demonstrated for vastly improved accuracy of predictions of the dependence jets in electron-positron collisions by Hornig, Lee, and their of these cross sections on R. This is a major advance in collaborators in earlier work [5,6]. the theory of jets in QCD and SCET, and opens the door to a wide array of improved predictions of many jet observ- In this project, A. Hornig together with collaborators Y. ables. This work has been submitted to the arXiv and has Makris and T. Mehen (Duke) has modified definitions of been accepted for publication in Phys. Rev. D [4]. the jet angularities to be appropriate for jets produced in proton-proton (pp) collisions such as at LHC. They extend- In Fig. 3 we illustrate the improved convergence of re- ed the technology and computations in SCET to describe summed perturbation theory for the prediction of a jet dependence on jet angularities and algorithms used in cross section as a function of the “jet thrust,” a measure pp collisions. They demonstrated the ability to predict jet of how collimated are the particles in a jet. With our new angularities at LHC to NLL and NNLL accuracy and provide soft-collinear mode, the resummed perturbative prediction the basis to go to higher accuracy and predict other jet converges noticeably better from one order of accuracy shape observables at LHC. One of the first predictions for a (NLL, next-to-leading-log) to the next (NNLL, next-to-next- jet angularity cross section in pp collisions to these orders to-leading-log) than without it, in which case the theoreti- of accuracy is shown in Fig. 4. This opens the door to cal uncertainties are considerably larger. This is because of precision jet and subjet studies at LHC that will be crucial uncontrolled dependence on the jet radius R in the latter in separating jets constituting signals for new physics and case, which we have captured accurately in the first case. backgrounds from SM QCD. A publication on this work is very near completion and about to be submitted to JHEP Jet Shapes for the LHC: [7]. While many jet cross sections simply look for the pres- ence of jets, looking inside at their more detailed structure BOOST Workshops on Jet Substructure reveals much information about their origin and evolution, The study of jet substructure to mine more information in particular, whether they are more likely to have come about jet origins and distinguishing signals and back- from the decay of a heavy particle in the SM or BSM, or grounds has been such promising, fruitful, and active from simple, standard, light quark/gluon dynamics in QCD. endeavor in the past several years that a series of annual 674

Page 675

workshops called BOOST bringing together theorists and in and from jets. We have developed and applied tools for experimentalists has been established. A. Hornig has been resummation that can make possible unprecedented preci- a leader in this endeavor. He is a co-author on a series of sion in predicting jet cross sections for probing QCD and the very highly cited reports arising from these workshops, new physics at the Large Hadron Collider (LHC). In concert not only summarizing but also actively evaluating the new we invented and developed new strategies for finding ideas and experimental results in this field each year and and characterizing jet substructure and interpreting those pointing towards the most promising avenues for research characteristics as signatures for new physics. Many of the in the coming years. His leadership and expertise were same tools we have developed can be applied to better called upon by the community to be a principal organizer predictions and understanding of physics at the Relativistic and contributor of the report of the BOOST2013 workshop. Heavy-Ion Collider (RHIC) and future Electron-Ion Collider This report is no simple proceedings of the workshop but (EIC), all high-priority facilities in US Nuclear Physics and in rather an original synthesis and analysis of the ideas and which LANL maintains extensive involvement and leader- results presented and discussed there and requiring signifi- ship. cant new research time. The work was peer-reviewed and The advances made by Dr. Hornig under this LDRD pro- published in 2015 [8]. posal open several new paths of research that support our goals in T-2 sponsored by DOE Office of Science, Offices of Nuclear Physics and High-Energy Physics, and also by the DOE Early Career Research Award given to Christopher Lee in 2015. We have extended Dr. Hornig’s appointment here for a third year, through FY 2016, to continue his research under the aegis of DOE funding. His expertise is invaluable to the advancement of our research towards several of the highest priority goals of US Nuclear and High-Energy Phys- ics efforts. References

1. Ellis, S. D., Hornig, T. S. Roy, Krohn, and M. D. Schwartz. Nondeterministic Approach to Tree-Based Jet Sub- Figure 4. Angularity distribution for jets produced in hadron structure. 2012. PHYSICAL REVIEW LETTERS. 108 (18). collisions at LHC to NLL (red) and NNLL (blue) accuracy. This is the first prediction of a jet angularity distribution at LHC to these 2. Ellis, S. D., A. Hornig, D. Krohn, and T. S. Roy. On Sta- orders of accuracy [7]. tistical Aspects of Qjets. 2015. Journal of High-Energy Physics. 01: 022. Impact on National Missions
2. Kasiecka, G., T. Plehn, T. Schell, T. Strebler, and G. P. Two key missions of the DOE Office of Science in High- Salam. Resonance searches with an updated top tag- Energy and Nuclear Physics are 1) to improve our under- ger. 2015. Journal of High-Energy Physics. 06: 203. standing of Quantum Chromodynamics (QCD), the theory of the strong interactions between quarks and gluons,
3. Chien, Y. T., A. Hornig, and C. Lee. A Soft-Collinear constituents of all ordinary matter; and 2) to search for Mode for Jet Cross Sections in Soft Collinear Effective new particles and forces beyond the Standard Model that Theory. To appear in Physical Review D. will explain the origin of matter, dark matter, and masses of elementary particles. 5. Hornig, A., C. Lee, and G. Ovanesyan. Effective pre- dictions of event shapes: factorized, resummed and Understanding properties of jets of hadrons produced by gapped angularity distributions. 2009. Journal of High- energetic quarks and gluons are key to achieving these ob- Energy Physics. 05: 122. jectives. They reveal the behavior of QCD itself and contain evidence of new particles that decay to jets. 6. Ellis, S. D., A. Hornig, C. Lee, C. K. Vermilion, and J. R. Walsh. Jet shapes and jet algorithms in SCET. 2010. This project has tackled problems at the forefront of QCD Journal of High-Energy Physics. 11: 101. perturbation theory and strategies to search for signatures of new physics in the substructure of jets. Physically ac- 7. Hornig, A., Y. Makris, and T. Mehen. Jet shapes in SCET curate predictions of jet cross sections require summing ar- at the LHC for dijet events. Journal of High-Energy bitrarily many soft and collinear quark and gluon emissions Physics, preprint LA-UR-27299. 675

Page 676

8. Adams, D., A. Hornig, and others. Towards an Under- standing of the Correlations in Jet Substructure. 2015. European Physical Journal. C75 (9): 409. Publications Adams, D., A. Hornig, and others. Towards an Understand- ing of the Correlations in Jet Substructure. 2015. European Physical Journal. C75 (9): 409. Chien, Y. T., A. Hornig, and C. Lee. A Soft-Collinear Mode for Jet Cross Sections in Soft Collinear Effective Theory. To appear in Physical Review D. Ellis, S. D., A. Hornig, D. Krohn, and T. S. Roy. On Statistical Aspects of Qjets. 2015. Journal of High-Energy Physics. 01:

Hornig, A., Y. Makris, and T. Mehen. Jet shapes in SCET at the LHC for dijet events. Journal of High-Energy Physics, preprint LA-UR-15-27599. 676

Page 677

Science of Signatures

Page 678

Science of Signatures Directed Research Continuing Project Optical and Laser Spectroscopy of Th-229 Electronic and Nuclear Transitions for the Development of Solid State Nuclear Quantum Sensors Xinxin Zhao 20140011DR Introduction impact in navigation, quantum information, and threat Laser spectroscopy and atomic clocks are at the heart reduction. It addresses LANL Science of Signatures goals: of several very important applications. For example, to revolutionize measurement and discovery signatures. remote sensing and measurement techniques rely on It also supports the LANL Materials for the Future Sci- converting an observable, such as the composition of a ence Pillar by creating novel materials with controlled Martian rock or the position of a missile or smart phone, functionality that advance our knowledge in intrinsic and into frequency or time signatures that can be measured engineered defects and enable observation of the emer- with the utmost precision and/or accuracy. All atomic gent phenomena of laser-nuclear interactions. A global clocks have a high quality atomic oscillator, but they dif- network of nuclear clocks and the related gravity sen- fer in the type of oscillator that is used, how the atoms sors will be a powerful system for fundamental science, (ions) are isolated from the environment, and the way national defense, and threat reduction. LANL is one of electronic transition frequencies are detected. Atomic only a few institutions worldwide that can assemble clocks have led to many scientific and technological a tightly coupled effort combining experimental capa- advances including the Global Positioning System (GPS) bilities in actinides nuclear chemistry (relevant to the for navigation. weapons program), precision laser spectroscopy, atomic, nuclear and material sciences. LANL has an opportu- The Th-229 nucleus holds the promise of similarly pro- nity to be a world leader in this groundbreaking field. found impact by using nuclear rather than atomic states. Besides the advanced scientific applications mentioned Achieving laser interaction with a nucleus for the first above, thorium isotopes are also a nuclear “timer” for time would dramatically advance the field of “nuclear special nuclear materials. The ultra-sensitive detection quantum science and technology” and have transforma- of thorium isotopes developed under this DR will have tional impact on fundamental science and sensing. This important impact in nuclear forensic applications. project builds on our recent breakthrough in Th-229 iso- mer research and aims for the first-ever demonstration Progress of direct laser excitation of the nucleus. This project is Some tasks were delayed due to funding cuts but we high risk, but the potential payoff is very high because have made exciting progresses as summarized below the demonstration of nuclear laser spectroscopy would and the project was rated outstanding at its midterm literally be a “quantum” leap over Mössbauer spectros- review. copy and enable the realization of a nuclear clock with unmatched precision. If successful, this project would We searched for the transition from 156nm-250nm by usher in a new era of nuclear quantum science with collecting thorium-229 recoils into MgF2 plate, and mea- broad implications from fundamental science to impor- suring the subsequent light emission using a monochro- tant new applications. Scientists from four divisions (C, mator. Results so far are very encouraging but additional P, MST, and T) will contribute their diverse expertise and measurements are required to confirm the discovery. collaborate on these experiments. A paper will be updated for resubmission to Nature or Physical Review Letters. The challenge for claiming the Benefit to National Security Missions discovery is that the observed signal is very weak due This project lays the foundation for nuclear clock and to poor recoil efficiency of the U-233 sources. We have nuclear quantum technology with transformational studied the source problem and plan to address it in 678

Page 679

the next few months while taking more data with existing neous crystallization (SC) that performs slow cooling of a sources. A separate paper on preparing and charactering ~200 mg melt drop. An undoped crystal was successfully the U-233 sources is also in preparation. prepared this way, and we have tested its optical back- ground at longer wavelengths (450 – 1800 nm). The SC The research team also performed absorption and laser growth system will now be moved into a radiological area induced fluorescence spectroscopy on MgF2 samples for growth runs with thorium. irradiated by thorium-229 and/or alpha particles with ~1011-12 atoms implanted into MgF2 crystals. The alpha Finally, we have investigated the prospect of super-radi- radiated MgF2 plates serve as control disks to rule out ance in thorium-229-doped materials. Even at moderate spectroscopic features from color centers caused by doping concentration, thorium emitters can find them- radiation damage or other impurities. According to our selves within the space that favors the cooperative emis- theoretical predictions, the most likely charge state for the sion of radiation called super-radiance. Super-radiant implanted thorium-229 atoms is +4. However, there may coupling may off-set the formidable challenges posed by be some thorium atoms in lower charge states because the weak photon coupling for manipulating Th-299 nuclei. the implantation is an equilibrium state. The lower charge These calculations did not, however, take into account the states will have low-energy excited electronic states in the inhomogeneous broadening resulted from environmental visible and near-IR portion of the electrometric spectrum, effects that would not affect the superradiant dynamics at so they can be detected if exist. Experimental results so early times but would degrade the coherence of the beam far support the theoretical prediction that the Th+4 are the and require a higher concentration of thorium to observe most abundant in the crystal. We are finishing up these super-radiance. Better understanding of these effects measurements and preparing a paper for publication. would guild the experiment for the laser search. Several options were examined to cover a wide range of Future Work laser wavelengths and we have constructed and charac- • Improve the method to clean up U-233 solution terized prototypes to generating the required light for chemically for molecular deposition of U-233 sources. laser excitation of the isomeric transition. For the shortest Complete and publish the study and characterization wavelengths, a possible scheme involves CW 5th harmonic of U-233 recoil sources. generation of a Ti:Sapphire laser by mixing the 4th har- • Complete and publish the ongoing nuclear optical monic of a Ti:Sapphire laser with the unconverted funda- spectroscopy measurement of thorium-229 isomeric mental wavelength using Sum-Frequency Generation. We transition. Continue to improve the measurement have completed and optimized the first doubling stage resolution and accuracy. with power output sufficient for the subsequent stages. • Complete and publish optical absorption and laser For the long wavelength range, a two stage 4th harmonic induced fluorescence measurements of thorium elec- generation can be used. First and second stage doubling tronic transitions. cavities have been constructed using 920 nm Ti:Sapphire • Build the lasers needed to drive thorium-229 nuclear laser, with a first stage output of 200-300mW from a 1W isomeric transition. Setup the laser excitation/mono- input. We have also setup a new UHV and DAQ system that chromator detection system. Minimize the PMT back- takes advantage of the high resolution of a new monochro- ground and optimize the system. mator. We are characterizing a new PMT to reduce the • Setup and optimize crystal growth system with tho- background and improve the sensitivity for future mea- rium-232, and make thorium-229 doped crystals. surement. • Perform theoretical study to better understanding of nuclear super-radiance effect. To maximize thorium-229 ion density in a crystal, we have • Calculate the charge state and the electronic transi- partnered with the University of Pisa (Italy) who are lead- tions energies of thorium ions doped in LiCaAlF6. ing experts in micro-pulling-down fluoride crystal growth. They have fabricated an exact copy of their high-perfor- Conclusion mance growth furnace for delivery to LANL in the next few We expect following results with the ultimate goal of months. We have secured a new lab space for this equip- achieving direct laser excitation of the nuclear transition: ment, completed the Radiological Engineering review, • Improving the accuracy of direct optical spectroscopic completed electrical/venting modifications, and installed a measurements to within 0.01 nm or better. This is a chiller system. In parallel to the high-quality crystal growth prerequisite for the subsequent laser search to suc- effort, we have explored a “quick and dirty” approach of ceed. growing a thorium-229 doped LiSrAlF6 crystal via sponta- 679

Page 680

• Identifying the electronic transitions of the Th ions in the host crystal. This is an unexplored area of research and will require close collaboration between theory and experiment. • Preparing high-quality single crystals doped with Th- 229. • Developing a vacuum ultraviolet laser and demonstrat- ing direct excitation of the Th-229 nuclear transition. Publications Ellis, J. K., Wen, and R. L. Martin. Investigation of Thorium Salts As Candidate Materials for Direct Observation of the Th-229m Nuclear Transition. 2014. INORGANIC CHEMISTRY. 53 (13): 6769. 680

Page 681

Science of Signatures Directed Research Continuing Project Remote Raman-LIBS Spectroscopy (RLS) Signature Integration Samuel M. Clegg 20140033DR Introduction detect actinides, similar RLS instruments and theoretical Distinguishing man-made and natural sources of ma- methods could be used to remotely detect chemical and terials such as actinides can be accomplished from the biological weapons as well as high explosives. RLS can be detection of the molecular and elemental compositions. also used as an in situ analytical geochemical instrument However, the detection of both molecular and elemen- to identify source of soil contamination (i.e. harmful tal composition is a challenge for any analytical method metals) and terrestrially sequestered carbon. Finally, and it is rarely possible remotely. An integrated Raman this RLS instrument will be the foundation for future and Laser-Induced Breakdown Spectrometer (LIBS) planetary science mission. For the weapons program, we instrument is uniquely capable of remote molecular and anticipate a new analytical tool for materials inspection. elemental quantitative analysis. This proposal will com- plete the fundamental physical studies required to real- Progress ize the full potential of this novel integrated approach. The new transmission spectrometer that enables the The most significant scientific challenge of this project is collection of both Raman and Laser Induced Breakdown the development of first principal molecular spectrosco- Spectroscopy (LIBS) spectra has been completed. This py and plasma physics theoretical models to accurately transmission spectrometer was designed by LANL and is assess and predict Raman and LIBS spectra, respectively. now realized due to this LDRD program. We will also integrate these theoretical and experimen- tal spectra into a single multivariate matrix from which a The ultraviolet (UV) and violet (VIO) spectrometers that self-consistent molecular and elemental quantitative de- record LIBS data over the 240 – 340 and 380 – 470 nm scription will be developed. In order to fully realize the regions will be completed by the end of FY15. potential of this integrated technique, we will prototype the first integrated Raman-LIBS remote sensing instru- All three spectrometers use miniaturized intensified ment capable of detecting actinides such as uranium. charge coupled device (ICCD) detectors that were de- Finally, the theoretical models and integrated analysis signed and built on this LDRD program. These miniature methods will be validated against carefully designed ICCD detectors enable the time-resolved LIBS and Raman remote sensing experiments under the most challenging analyses that are critical to understanding the underly- environmental conditions. Consequently, this Science ing physics. They use components that make them more of Signatures project will satisfy all three themes: new spectroscopically sensitive than commercially available signature detection, revolutionize measurement, and ICCD detectors. Finally, these detectors are capable of forward deployment. many applications beyond the defined Raman and LIBS requirements including passive infrared spectroscopy. Benefit to National Security Missions This LDRD DR Science of Signatures project will focus on These intensifiers used in the ICCD detectors require a the detection of actinides which is an element of LANL’s 6 kV high voltage power supply (HVPS) that can operate core mission. The team will use Raman and LIBS spec- under reduced pressures. Such a power supply is not troscopy (RLS) to distinguish anthropogenic and natural commercially available. The engineering team on this actinides within complex geological materials as well project successfully designed a HVPS that meets these as the advanced theoretical methods developed in this requirements. project. While the strategic focus of this proposal is to 681

Page 682

The team successfully produced a theoretical LIBS spec- Conclusion trum of a complex basalt sample. The actual spectrum was The overall goal of this project is to develop the theoretical reported in the 2012 ChemCam calibration paper by Wiens foundations that are capable of predicting and quantitative et al. because it reports the intensity in photons whereas interpretation of Raman and LIBS spectra. These theoreti- most other published LIBS spectra report intensities in “ar- cal models will be tested and validated with the first fully bitrary units.” The theoretical and experimental spectra are integrated remote Raman – LIBS instrument used to probe in excellent agreement and theoretical compositions are actinides doped in complex geological samples. The theo- proportional to the relative elemental abundance. This is retical methods will revolutionize the analysis of Raman the first major accomplishment towards the development and LIBS spectra within these communities and the instru- of a quantitative theoretical interpretation of LIBS spectra. ment demonstration will lead to novel national security This work is in press in Spectrochimica Acta B (Colgan et and planetary science capabilities. al). Publications We recently identified experimentally observed oxygen Clegg, S. M., Wiens, A. K. Misra, S. K. Sharma, Lambert, emission lines that the theoretical calculations indicate are Bender, Newell, Nowak-Lovato, S. u. e. Smrekar, M. not energetically possible by standard LIBS ablation ener- D. Dyar, and Maurice. Planetary Geochemical Investi- gies. We determined that these oxygen lines are actually gations Using Raman and Laser-Induced Breakdown produced from the breakdown of the atmosphere rather Spectroscopy. 2014. APPLIED SPECTROSCOPY. 68 (9): than the sample. This suggests that the breakdown of the 925. atmospheric molecules produces atoms that are in highly Colgan, , E. J. Judge, D. P. Kilcrease, and J. E. Barefield II. excited states. This observation will be confirmed with the Ab-initio modeling of an iron laser-induced plasma: time-resolved RLS spectrometers early next FY. Comparison between theoretical and experimental atomic emission spectra. 2014. SPECTROCHIMICA We also completed an initial study on the Raman and LIBS ACTA PART B-ATOMIC SPECTROSCOPY. 97: 65. detection of biological materials. This work was presented at an SPIE meeting and published (Anderson et al.). Colgan, , E. J. Judge, H. M. Johns, D. P. Kilcrease, J. E. Li- arefield II, McInroy, Hakel, R. C. Wiens, and S. M. Clegg. In addition to the two published papers discussed above, Theoretical modeling and analysis of the emission spectra of a ChemCam standard: Basalt BIR-1A. 2015. we have submitted three papers and are preparing six SPECTROCHIMICA ACTA PART B-ATOMIC SPECTROS- manuscripts for submission before the end of the FY. COPY. 110: 20. Future Work Hartig, K. C., Colgan, D. P. Kilcrease, J. E. Barefield II, and Jo- The project will complete one task and continue to ad- vanovic. Laser-induced breakdown spectroscopy using vance two tasks in FY16. The integrated Raman-Laser- mid-infrared femtosecond pulses. 2015. JOURNAL OF APPLIED PHYSICS. 118 (4). Induced Breakdown Spectroscopy (LIBS) instrument will be constructed by the end of FY15.The integrated spectrom- eters will be fully characterized as soon as possible in FY16 such that it will be used to conduct experiments necessary to the theoretical calculations. This characterization will describe the component level throughput to produce a detailed end-to-end photon budget and intensities that are reported as photons. The LIBS plasma spectroscopy theoretical calculations will continue to integrated ele- ments with higher atomic masses into the complex LIBS spectra. The Raman theoretical analyses will continue to develop the fundamental methods needed to accurately simulate Raman spectra of pure minerals and validate with the RLS instrument. Finally, we will continue experiments on complex actinide containing samples using standard laboratory hardware as an initial validation of the theoreti- cal calculations. 682

Page 683

Science of Signatures Directed Research Continuing Project Explosives Signatures for Detection: Nonlinear GHz to THz Responses David S. Moore 20140049DR Introduction of materials signatures of explosives within its global This project aims to fill a major gap in our improvised security mission. The deep fundamental understand- explosive device (IED) detection arsenal, by detecting ing of energy absorption and subsequent responses are the explosive itself using penetrating but non-ionizing applicable to MaRIE’s “Decadal Challenges for Predicting electromagnetic radiation. This project will develop and Controlling Materials Performance in Extremes” for new experimental and theoretical capabilities to exploit the design and control of energy release in explosives. newly-discovered nonlinear coupling of GHz-to-THz Work under this proposal will develop a new framework electromagnetic radiation to bulk explosives and the de- for a unified description of explosives’ hot spots coupled tection of the alternative signatures that are generated. to the local mechanical, thermal and electromagnetic The characteristic that we exploit is the intrinsic prop- signatures. Understanding and controlling the material erty that defines an explosive - its metastable chemical functionality of defects and crystalline interfaces is im- energy that can be quickly released on command (by portant for hot spots and their signatures and underpins shock, friction, or spark). We will model the expected the Materials for the Future pillar with the focus area signatures across relevant length scales from molecular Defects and Interfaces, as well as the priority area 2 for to bulk levels, quantify the complex permittivity versus advanced (THz) spectroscopies. The explosive expertise temperature and amplitude, demonstrate piezoelectric developed in this project is of course highly relevant to and pyroelectric coupling to convert electromagnetic to the weapons program. ultrasonic energy in situ, and evaluate electro-mechano- chemistry effects. The measurements will be guided and Progress interpreted using electromagnetic theory coupled to • Let a subcontract (Interagency Agreement) to NIST- strain and heat diffusion, with the ultimate goal to devel- Boulder for the measurement of complex permit- op fundamental principles that define the processes and tivity from 220 to 500 GHz; samples sent to or are signatures, leading the design of a prototype detection currently being prepared for this collaboration. system with area-scanning capability from safe stand-off • Made THz measurements of pressed powder sam- distances. The capability is not intended to be used in ples of TNT, HMX, PETN, RDX, and CompB as well as isolation, but rather as a tool in the toolset – a tool with those materials in polytetrafluoroethylene (PTFE) at drastically improved detection capabilities, viz. pen- 10% and 70% concentration. etration through clothing, camouflage, or packaging by • Obtained micro-CT and THz spectra of the explosives using GHz-to-THz frequency radiation, which is a unique simulants 1,3- and 1,4-dinitrobenzene (DNB); theory electromagnetic spectral region with sparse application and modeling extended to calculate THz spectra of to explosives detection. 1,3- and 1,4-DNB to compare to the experimental THz spectra (linear low power regime); theoretical Benefit to National Security Missions methods used involve extending electronic structure This project represents a transformational approach to methods to a supercell approach for calculations at uncovering the direct explosive signal type applicable to the lower THz frequencies where possible motions the Discover Signatures and Revolutionize Measurement are longer range and intermolecular rather than components of the LANL Science of Signatures pillar. The intramolecular; a paper on this work was submitted. overarching goal of the project directly supports LANL’s • Theory and modeling methods are being used to long-term objective of discovery of the next generation calculate THz spectra of the explosives and will be 683

Page 684

compared to the experimental THz spectra; these strated using KBr dilution. We can find no scientific models are being extended to include the nonlocal literature with data on the variation of THz spectra in temperature fields caused by energy localization and/ mixtures of neat explosives or HME precursors with or chemical energy release. common materials or fuels, which is of paramount • Methodology was established to input micro-CT importance for agencies investigating use of linear re- image-derived FEA models into ABAQUS and COMSOL; sponse THz spectral methods in screening applications. code was developed to simulate RF electromagnetic • We (in collaboration with Mitchell Wood, Purdue Uni- field effects on the FEA models of PBX9501 and PBXN- versity) adapted the molecular dynamics THz simula- 110 and we have quantified the heterogeneous dis- tion methods pioneered by CalTech to investigate sipation of energy (this is currently in 2-D, but is being anharmonic frequency mixing and Raman processes in extended to 3-D); code was also developed to extract RDX and PETN in the THz spectral region. A paper is in the corresponding heterogeneous temperature fields preparation. (again currently in 2-D to be extended to 3-D). • Developed a theory of explosions in solids where the Future Work parameters of the system exhibit frozen, sample spe- Pursuant to the mid-project review, we have revised the cific spatial fluctuations. In this case, the crucial feature tasks to the following: of the system is the existence of long tails in the distri- bution functions of the parameters of the system. We The team will focus on the following explosive materials: discovered the existence of a phase transition in the PETN, AN (with and without fuel), HMX and/or RDX, single system. Above the critical temperature, an explosion of crystals of the above with engineered defects (microma- one spot induces an avalanche of explosions in which chined void structures and deposited metamaterial struc- the total number of explosions diverges at the critical tures). point. We confirmed existence of this phase transi- tion performing numerical simulations. A paper is in Task 1: Elaborate and extend the 95 GH excitation up and preparation. down conversion emission experiments in the TA-35 an- • Constructed a waveguide based system to measure echoic chamber at MPA-CINT. upconverted non-harmonic RF emissions under RF stimulation; this effort is being replaced by task 1 in Task 2: Reproduce the time-dependent reflectivity experi- the technical description. ments at or near 11 GHz with sugar and HMX, but with • Used a Raman spectroscopy system to measure the control of water content (by storage of samples in dessica- change in Raman spectrum induced by RF stimulation tors). at 2.54 GHz, but the first results were negative. • Used the Raman spectroscopy system to look for Task 3: Complete measurement of the complex permit- changes in Raman spectrum with static electric fields; tivity in the 200-350 GHz regime using open Fabry-Perot in this case, the samples were held in a planar geom- resonator methods via collaboration with NIST Boulder etry double plate capacitor (1 mm sample thickness), (subcontract let June 30, 2014 for two years) capable of electric fields up to 2 MV/m, but no changes were observed. Task 4:Complete GHz and THz linear dielectric permittiv- • Measured THz spectra (low power linear response) of ity experiments, including 90/10 PTFE/material samples homemade explosive precursor oxidizers, again finding and 30/70 PTFE/material samples. Samples at these same several with large absorbance at the 1 mm sample ratios will also be sent to NIST Boulder (see task 3). thickness; dilute samples in PTFE were pressed and their THz spectra are being obtained. Task 5: Calculate heterogeneous dissipation of energy and • We measured THz spectra of a TNT-CL20 co-crystal, conversion to heterogeneous thermal distributions using and compared to the TNT THz spectra (new peaks in Abaqus and/or COMSOL through input of micro-CT derived co-crystal), also as a function of temperature to be- FEA models of explosive/air systems. yond the TNT normal boiling point (showing co-crystal is not just a mixture of TNT and CL20). A paper is in Task 6: Continue MD simulation of nonlinear generation of preparation. new frequencies, extending towards lower frequencies by • Investigated masking of our modification of THz spec- enlarging the simulation size and time, including geometric tra of explosives and HME precursor oxidizers by fuels, boundaries. We will also compute frequency conversion in matrix materials, and common interference materials; reflection and transmission modes. the first example of this kind of masking was demon- 684

Page 685

Task 7: Complete the THz and near IR studies on 1,3-DNB and submit a journal paper. Task 8: Extend THz and NIR measurements and theory/ simulation to binder effects starting with measurements on HTPB. Task 9: Manufacture samples (micromachined single crys- tals) for the Keith Nelson group at MIT for high power THz experiments. Samples include PETN and TATB. Conclusion The approach described in this project utilizes technologies in frequency regimes capable of penetrating non-metallic packaging, clothing, or camouflage to interrogate bulk explosives, filling a capability gap in the current suite of counter-IED technologies, which includes X-ray imaging, metal detection, trace analysis, intelligence, and persistent surveillance. It will jump-start new modeling capabilities at LANL enabling theory of elasto-electric coupling with chemical energy release at hot spots in explosives. The model development at multiple scales with the combina- tion of coupled matter-energy interactions ranging from the nano to meso scale brings a fresh and new approach to this long-standing problem. Publications Chellappa, R., D. Dattelbaum, M. Graf, D. Moore, and Z. Liu. High pressure vibrational spectroscopy of dinitroben- zene isomers. Chemical Physics Letters. Chellappa, R., D. Dattelbaum, N. Mack, T. Ahmed, M. Graf, A. Azad, D. Moore, and Z. Liu. Vibrational anharmonic- ity in 1,3-dinitrobenzene. To appear in Journal of Physi- cal Chemistry A. Chowdhury, D. R., X. Su, Y. Zeng, X. Chen, A. J. Taylor, and A. K. Azad. Excitation of dark plasmonic modes in sym- metry broken terahertz metamaterials. 2014. Optics Express. 22: 19401. Sorensen, C. J., and D. S. Moore. Radio-frequency electro- magnetic emissions from materials under high-fre- quency mechanical excitation. To appear in American Physical Society Topical Conference on Shock Compres- sion of Condensed Matter-2015. (Tampa, 14-19 Jun. 2015). Wood, M. A., D. A. R. Dalvit, and D. S. Moore. Nonlinear electromagnetic interactions in energetic materials. Physical Review Applied. 685

Page 686

Science of Signatures Directed Research Continuing Project Chemical Signatures of Detonation Born From Extreme Conditions (U) David Podlesak 20150050DR Introduction models critical to improved weapons simulation and The ability to infer information about high explosive results will be relevant to products Equation of States (HE) type and design is of high value. Traditional nuclear (EOS) used in DOE/NNSA defense and nuclear nonprolif- forensic investigations provide little or no insight into the eration programs. As well, the integration and analysis dynamics of HE detonation. Post-detonation and real- of heterogeneous data to link detonation products with time signatures of explosive test programs and unique initial detonation conditions using modern data analytic materials identifiers are needed, but can only come from techniques will provide a tool useful to multiple agen- a fundamental understanding of the physical processes cies involved in nuclear nonproliferation. The develop- that lead to their formation and evolution in time. Ex- ment of techniques to interpret heterogeneous data sets plosives detonation is an exothermic process whereby will also benefit other nuclear and non-nuclear forensic metastable, complex molecules are converted to simple, programs. stable molecules and solid carbon, with the type of solid carbon formed depending on the type (composition) of Progress explosive. Recovery studies indicate that several allo- We continue to develop the capability to use detonation tropes of solid carbon may form including amorphous products as novel chemical signatures of explosive com- graphite, “onion-like” carbon, and detonation nano- position, performance, and state. Successful execution diamonds (DND). More recently, carbon graphite-to- of this program will provide models critical to improved diamond ratios have been correlated to explosive type, weapons simulation and results will be relevant to with high-performance, aromatic-ring based explosives products Equation of States (EOS) used in DOE/NNSA producing the largest quantities of DNDs. We propose defense and nuclear nonproliferation programs. As well, to establish an interdisciplinary effort to understand the integration and analysis of heterogeneous data to how solid carbon forms and evolves during detonation, link detonation products with initial detonation condi- develop new models to describe its evolution, and link tions using modern data analytic techniques will provide in situ measurements of signature formation with post- a tool useful to multiple agencies involved in nuclear detonation characteristics. A key feature of this effort nonproliferation. The development of techniques to will be discovery of signature formation in real-time interpret heterogeneous data sets will also benefit other behind the detonation front using in situ time-resolved nuclear and non-nuclear forensic programs. We have x-ray scattering at the Advanced Photon Source and made considerable progress to date. Please see specific U.S. free electron laser beam lines. The formation and notes on progress for each main task below. evolution of solid carbon (and fluids) will be modeled from formation through chemical equilibrium, and the Task 1: Continue in-house detonation experiments to concert of signatures from post-detonation residues will obtain detonation product samples and use state-of- be analyzed in the context of statistical theories to glean the-art analytical techniques to determine chemical requisite information pointing to intent. compositions of detonation products including solids and gases. We have conducted >20 detonations, both Benefit to National Security Missions steady and overdriven conditions, at LANL firing facilities We will establish detonation products as novel chemical with three HE formulations (Composition B, PBX 9501, signatures of explosive composition, performance, and PBX 9502) in ambient conditions and in an Ar atmo- state. Successful execution of this program will provide sphere. Detonations and analyses are ongoing. Carbon 686

Page 687

soot was collected from all detonations and analyzed with Task 5: Complete development of a new analytical meth- multiple techniques including electron microscopy, X-ray odology for pyrolysis-fluorination of actinide-oxides that diffraction (XRD), X-ray photon spectroscopy (XPS), Raman can be used to quantify O and H isotope effects in these spectroscopy, and isotope ratio mass spectrometry (IRMS). materials. We have innovated modifications to a pyroly- Initial results will be presented at the 19th Biennial Confer- sis reactor, and made proof of concept measurements of ence on Shock Compression of Condensed Matter (SCCM- O-isotope ratios in surrogate oxides (Eu2O3) by high-tem- 2015) June 14-20, 2015. Results will also be published in perature pyrolysis-fluorination using PTFE as the F source. the associated conference publication. Results will also This approach that will enable faster, safer and simpler be presented at the AVS 62nd International Symposium & analysis of actinide oxide samples as small as ~100µg. We Exhibition October 18-23, 2015. continue to improve this system and will now carefully conduct experiments to demonstrate the robustness of Task 2: Continue to develop x-ray scattering techniques the technique for actinide oxides, quantify any systematic to record evolution of carbon (clustering, morphology, fractionations induced by the methodology or impacts polydispersity) in high explosives in real time to provide of analytical conditions and validate the new method by a scientific basis of post-detonation signature formation. comparison with conventional laser-assisted fluorination We continue to collaborate with LLNL on developing this analyses. capability including improving the design of the detonation vessel, vacuum system, and issues with timing. To date, Future Work we have collected some of the first X-ray scattering data in Our goal is to continue to develop the capability to identify real-time for high explosive (HE: Composition B and PBX the intent/design of a detonation test using detonation 9501) detonations in the United States. Early results that products. describes clustering characteristics of Composition B will be combined with static X-ray scattering measurements Task 1: Continue in-house detonation experiments to ob- of soot collected from detonations at LANL for a time-line tain detonation product samples and use state-of-the-art description of carbon clustering. Results will be presented analytical techniques to determine chemical compositions at the 19th Biennial Conference on Shock Compression of of detonation products including solids and gases. Condensed Matter (SCCM-2015) June 14-20, 2015 and will also be published in the associated conference publication. Task 2: Continue to develop x-ray scattering techniques We have added two postdocs, Rachel Huber and Bryan to record evolution of carbon (clustering, morphology, Ringstrand, to work on both Task 1 and Task 2. polydispersity) in high explosives in real time to provide a scientific basis of post-detonation signature formation. Task 3: Continue development of carbon clustering phys- This builds on ongoing and developing capabilities at the ics capturing the fluid/solid mixture and carbon surface Dynamic Compression Sector (such as vessels, firesets, and chemistry. This effort is critical in concert with the x-ray x-ray optics). scattering data for defining how detonation products form and evolve; e.g. signature formation. We continue Task 3: Continue development of carbon clustering phys- development of carbon clustering physics, specifically by ics capturing the fluid/solid mixture and carbon surface combining theoretical models of X-ray scattering during chemistry. This effort is critical in concert with the x-ray detonation with empirical measurements. scattering data for defining how detonation products form and evolve; e.g. signature formation. Task 4: Continue to combine multi-phenomenological data such as the combination of pXRD analysis with XPS Task 4: Continue to combine multi-phenomenological analysis of detonation soot to develop a methodology for data such as the combination of pXRD analysis with XPS abductive reasoning from the detonation products to the analysis of detonation soot to develop a methodology for initial detonation conditions. This effort is ongoing with a abductive reasoning from the detonation products to the collaboration established between LANL and North Caro- initial detonation conditions. lina State University (NCSU), specially the Consortium for Nonproliferation Enabling Capabilities (CNEC). This col- Task 5: Complete development of a new analytical meth- laboration includes sharing of an existing and open data odology for pyrolysis-fluorination of actinide-oxides that set that can be used by an NCSU student so that progress can be used to quantify O and H isotope effects in these on the combination of heterogeneous data sets moves materials. forward more efficiently. 687

Page 688

Conclusion We will establish detonation products as novel chemical signatures of explosive composition, performance, and state. To this end, we will combine novel methodologies to understand how solid carbon forms and evolves during detonation, develop new models to describe its evolution, and link in situ measurements of signature formation in real time behind the detonation front using time-resolved x-ray scattering at the Advanced Photon Source with state- of-the-art analysis of post-detonation gases and solids. Publications Firestone, M. A., D. M. Dattelbaum, D. W. Podlesak, R. L. Gustavsen, R. C. Huber, B. S. Ringstrand, E. Watkins, B. Jensen, T. Wiley, L. Lauderbauch, R. Hodgin, M. Bagge- Hansen, S. Seifert, and T. Graber. Structural evolution of detonation carbon in Composition B-3 by X-ray scat- tering. To appear in Proceedings of the 19th Biennial APS Conference on Shock Compression of Condensed Matter. (Tampa, FL, 14-19 Jun). Podlesak, D. W., R. C. Huber, R. S. Amato, D. M. Dattel- baum, M. A. Firestone, R. L. Gustavsen, C. E. Johnson, J. T. Mang, and B. S. Ringstrand. Characterization of detonation soot produced during steady and overdriv- en conditions for three high explosive formulations. To appear in Proceedings of the 19th Biennial APS Con- ference on Shock Compression of Condensed Matter. (Tampa, FL, 14-19 Jun). Willey, T. M., m. Bagge-Hansen, L. Lauderbach, R. Hodgin, D. Hansen, C. May, T. van Buuren, R. Gusavsen, E. Wat- kins, M. Firestone, D. Dattelbaum, B. Jensen, T. Graber, S. Bastea, and L. Fried. Measurement of carbon con- densates using Small-Angle X-ray Scattering during det- onation of high explosives. To appear in Proceedings of the 19th Biennial APS Conference on Shock Compres- sion of Condensed Matter.. (Tampa, FL, 14-19 Jun). 688

Page 689

Science of Signatures Directed Research Continuing Project Integrated Biosurveillance Benjamin H. Mcmahon 20150090DR Introduction be to examine how the widespread co-morbidities of Our study will apply three state-of-the-art diagnos- HIV, malaria, Salmonella, Streptococcus, Staphylococ- tic technologies to a high disease burden population cus, and tuberculosis combine to cause 1) poor outcome in western Kenya. We will perform multiple types of for individuals and the community (30% mortality rate, epidemiological analysis necessary to derive understand- at age 5), 2) emergence of antibiotic resistant forms, ing from the novel data. The samples are derived from which spread across the globe, and 3) presence of novel patients who frequently have co-infections (tuberculosis, pathogens. Salmonella, Streptococcus, and Staphylococcus) and are associated with a 12 year longitudinal study performed This project will lay the foundations to achieve our long- in a pediatric malaria clinic in the Saiya District of Kenya term technical goals of situational awareness for global by our collaborator, Prof. DJ Perkins of the University of pathogen circulation and emergence. New Mexico. The three diagnostic technologies are as follows. Benefit to National Security Missions This project directly addresses the National Security mis- Rapid Biomarker Detection sion in the area of biological threat reduction by charac- Approaches will be developed to detect proteins, lipids, terizing emerging diseases in the high-disease-burden and carbohydrates from human or animal samples. Re- region where they emerge. Providing this understanding agents for tuberculosis, Staphylococcus, Streptococcus, impacts LANL’s ability to respond to sponsor requests to and Salmonella will be applied and assessed. predict the course of emerging diseases or bio-terrorism events or understand how climate change and demo- Oligonucleotide-based diagnostics for pathogen charac- graphic changes impact infectious diseases, which still terization cause the majority of deaths in developing countries. This will include 48x or 96x multiplexed, PCR-based Furthermore, this project demonstrates how to system- nucleic acid assays to species- strain- and antibiotic resis- atically characterize these diseases as they emerge, in- tance- type for Streptococcus, Salmonella and Staphylo- situ, and how to integrate the resulting biosurveillance coccus. data. This capability is essential to accurately predict risk from novel engineered pathogens, as well as the risk to Sequencing of non-host RNA US personnel as they operate in the developing world. Novel sample preparations will be developed and automated to provide 10 million pathogen RNA se- Methodologically, this project integrates signature dis- quences from each sample. These data will be analyzed covery with measurement in patients, and deployment to observe unexpected pathogens, expression level of of assays and models - an important template for the virulence factors, validity of nucleotide diagnostics from SoS strategy. Biosurveillance has been a key SoS endeav- previous task, and strain-type relative database and or at LANL, and senior lab leaders organized a biosur- other samples. veillance deep dive in Sep 2013. At this event, external experts (programmatic, academic, industrial) critically Our epidemiological model will predict outcomes for reviewed LANL strategy and investment in BSV, and our both individual patients and the population of Siaya proposed concept was extensively appreciated as in- (800,000 residents), as a function of investments, treat- novative, challenging and critical. Our effort requires the ments, and mitigations. Our focus in the modeling will 689

Page 690

systematic integration of IS&T and experimental capabili- patients. This study will also demonstrate deployment of ties at LANL (Co-Design), which is not readily achieved in the tuberculosis, Salmonella, Staphylococcus, and Strep- an academic setting. This integration is not limited to BSV, tococcus assays, tailored PCR assays of these and other and can be applied to other threats (nuclear, chemical, en- pathogens, and provide both metagenomic and diagnostic vironmental) in the future. Similar tools can be applied to samples. 2) prepared a set of spiked samples of all sample nuclear non-proliferation or any other “big data” national type that can be related to the level of pathogen in the security problem. patient blood. Progress Public presentations of project work included two poster presentations at the Tropical Medicine meeting in New Overall, the project has made good progress towards its Orleans in November, 2014. Two posters were presented objectives. The following are technical highlights accom- at DTRA’s S&T meeting in St. Louis in May, 2015. Two plished in each of the five tasks, in preparation for prospec- invited talks were given at the Biosurveillance integration tive data collection in Year 2. conference in Washington, DC, in June, 2015. Project work was featured in Mukundan’s Frontier’s in Science public Biomarker task, led by Mukundan: 1) Prepared antigens lectures in March and April, 2015. The project was briefed (proteins, lipids, sugars and DNA) for three of the focal to DHS’s APEX program in biosurveillance in June, 2015. pathogens (Salmonella, Staphylococcus, and Streptococ- cus) from patient isolates for assay development and Four follow-on projects were awarded in Year 1: screened them against all commercially available anti- bodies for detection assays. 2) Developed assays for two The DTRA projects awarded to Fenimore and Vuyisich on Salmonella signatures and one Staphylococcus biomarker. the basis of ideas and data developed in the pilot project were awarded supplemental projects. Fenimore’s to pro- PCR assays task, lead by Doggett: 1) Developed the experi- vide epidemiological modeling support for CBEP-sponsored mental approach for testing blood spots, enabling hypoth- workshops in Africa and Asia, and Vuyisich’s to apply RNAS- esis-testing on retrospective samples. 2) Designed PCR equencing as an infectious disease diagnostic in the US, in probes and begun validation assays on spiked samples of collaboration with the New Mexico Department of Health. epidemiologically relevant questions on strains, virulence DARPA awarded $150ktoMukundanandHengartnertomarkers,andantibioticresistanceidentifiedincombinedevaluateits‘ChikungunyaChallenge’andprovide‘lessonsgenomic/clinicalanalysis.learned’reportforuseindevelopingUSgovernmentbio-surveillancecapability.DHSawarded$100k to McMahon Sequencing task, lead by Vuyisich: 1) Completed optimiza- and Deshpande to articulate the case for global applica- tion and validation of depletion of host reads from blood tion of multiplexed antibody-based diagnostics in order to samples, 2) Characterized, minimized, and mitigated cross- reduce the risk of $6Bn mitigations efforts such as the one talk in RNASeq diagnostics due to bar-coding of samples. required for the Ebola outbreak in West Africa over the 3) Identified virulence markers, antibiotic resistance genes, past 1.5 years. and relevant strains from which PCR assays, above, should be derived. Collaborations were developed with DJ Perkins, Director of the Center for Global Health at UNM, Dr. Ravi Durvasula, Modeling task, lead by McMahon: 1) Worked closely with Chief of Infectious Disease Medicine at UNM, and Richard UNM collaborators to apply new statistical methods to Hatchett, Deputy Dir. of BARDA, Dylan George, currently longitudinal (across time) data from 18,000 patient visits, at OSTP, and others involved in implementing the national to accurately identify significant genetic, immunological, biosurveillance strategy. clinical, and demographic correlates for outcome in pedi- atric infectious disease patients. 2) Further developed and transitioned our (fairly complex) epidemiological model of Future Work nine co-circulating pathogens to an interactive web-based Year two of our project will focus on collection of prospec- tool, enabling our UNM and Kenyan collaborators to relate tive clinical data in Siaya, Kenya, building off of a visit to the model to the clinical study design. Siaya by McMahon, Mukundan, and Noormohamed that should transpire at the end of Year 1. This visit will begin With the UNM/KEMRI subcontract, lead by DJ Perkins, we the process of translating the diagnostic assays and model- have: 1) Designed a prospective clinical study to provide ing capability to forward deployment, as well as facilitate diagnostic samples enriched in pathogens and character- the collection of samples to be shipped back to UNM and ize the burden of virulent pathogens in less-seriously-ill LANL for continued assay development and validation, 690

Page 691

in keeping with our original project plan. Although we McMahon, B., P. Fenimore, S. Del Valle, N. Hengartner, R. feel well-prepared for this effort, it will involve consider- Ribeiro, and J. Hyman. Modeling the impact of spatial able logistical complexity, including international human heterogeneity, behavior change, and mitigations on the current Ebola epidemic. 2015. LAUR 14-27813. subjects approvals, shipping of infectious disease samples, coordination with academics, clinicians, and goverment personnel in Kenya, data- and sample-tracking, and com- munication of results in publications and conference pre- sentations. These logistical issues are formidable barriers to other US Government biosurveillance efforts, and our experience gained here should enable continued develop- ment of competitive proposals to several sponsors. More specifically, our technical efforts will include: • Validation of tuberculosis and Salmonella waveguide- based biomarker assays on clinical samples shipped back from Kenya, together with development of Staphylococcus and Streptococcus assays. • Systematic application of PCR assays identified in Year 1. • Application of RNASeq using end-to-end process de- veloped in Year 1 on selected clinical samples, shipped back from Kenya, together with validation and devel- opment of analysis pipeline for RNASeq data. • Transitioning of interactive epi-model to use by Ke- nyans in real-time interpretation of diagnostic results and further clinical study designs. In addition, considerable effort will be required in bringing technical work from Year 1 to successful publication. Conclusion We will demonstrate integrative biosurveillance in the high disease-burden region of Siaya, Kenya. This will in- volve biomarker discovery, assay development, and assay deployment of three complementary infectious disease assays on a human population. Direct biomarker detec- tion assays will provide results at the point of care, nucleic acid detection with PCR is a mature technology which will build on established reference laboratory techniques, and RNA sequencing will provide a broad range of genomic information of a variety of pathogens. Overall information integration and process optimization will result from both statistical analysis and development and application of realistic epidemiological models. Publications Brown, M., L. Moore, B. McMahon, D. Powell, M. LaBute, J. Hyman, A. Rivas, M. Jankowski, J. Berendzen, J. Loep- pky, C. Manore, and J. Fair. Constructing rigourous and broad biosurveillance networks for detecting emereg- ing zoonotic outbreaks. 2015. PLoS ONE. : e0124037. 691

Page 692

Science of Signatures Directed Research Continuing Project Signatures of Change - Habitat Earth Reinhard H. Friedel 20150647DR Introduction Benefit to National Security Missions This DR proposes to expand the scientific understand- The science of signature and the means to detect and ing of fundamental physical processes that are critical interpret these signatures is directly applicable to the to maintenance of habitat earth homeostasis with the detection needs for nonproliferation and counter pro- long-term objective of achieving sufficiently detailed liferation community, space weather and space events, knowledge to identify the tipping points that can push remote sensing and detection of chemical, biological, habitat earth out of its homeostatic equilibrium. This nuclear, radiologic, or explosive threats, climate impact scientific goal is achieved by promoting and coordinat- and treaty verification, and cosmology/astrophysics. ing basic research based on the science of signatures to Signature discovery and alternate signatures provide the gain understanding of the structure, and evolution of new methods to detect and understand these areas of the earth, the solar system and the Universe in which national need. habitat earth resides, ultimately relevant to understand- ing future changes as they might perturb habitat Earth Progress homeostasis. Specifically, we propose to address the This project has been in effect about nine months. Four following scientific areas: refereed publications have been published with more in progress. Also, 27 separate projects are underway in • Geoscience, with the goal of advancing our ca- the disciplines of Geophysics, Climate, Space and Astro- pabilities to measure and understand significant physics. A call for proposals for new task starts in FY16 natural and human-induced changes within their has been issued after ensuring alignment with Labora- background context using a portfolio of theoretical, tory National Security strategies. A total of 38 proposals modeling, and experimental methods; were received and are now under evaluation for scien- • Climate with the goal of scientific advances that tific merit and alignment with Laboratory strategy. integrate theory, models, simulations, experiments, sensing and observations to push the frontiers of Future Work predictability to better prepare for impacts (extreme The scientific goal is to develop signatures or the means weather, tipping points, sea level rise), and plan for to measure signatures that provide insight into changes improved domestic, climate-friendly energy and re- on earth and space. This goal is achieved by promoting silient infrastructures for current and future climate and coordinating basic research on the understanding states. of the structure, and evolution of the earth, the solar • Space Physics with the goal of advancing our under- system and relevant to understanding future changes as standing of the space environment from the Sun to they might perturb habitat earth. The tasks will be to the Earth and beyond – with the particular goal of fund Los Alamos postdoctoral research associates and understanding how the space environment affects collaborative research between Los Alamos scientists space platforms that support security and quality of and university researchers. life in our increasingly technological society; • Astrophysics which includes astro-biology, planetary Conclusion studies, exo-planets, extinction events, other impor- tant astrophysical processes; and The impact is the enhancement of University-Laboratory • An emphasis on signature science and the overlap relations by fostering collaborations between academic sciences between these fields. campus faculty, staff, students, and Los Alamos National 692

Page 693

Laboratory staff; providing Los Alamos National Laboratory programs with systematic infusion of new ideas, people, and contact with the larger university community; direct support of Los Alamos postdocs; fostering top-quality research at Los Alamos National Laboratory in the more “basic” or “fundamental” aspects of fields that map into existing and/or emerging mission thrust areas of the Labo- ratory; and harnessing LANL’s supercomputing and multi- scale measurements and modeling capabilities. Publications Nonlinear elasticity and hysteresis: Fluid-solid coupling in porous media. 2015. Anthony, R. E., R. C. Aster, Wiens, Nyblade, Anandakrish- nan, Huerta, J. P. Winberry, Wilson, and Rowe. The Seismic Noise Environment of Antarctica. 2015. SEIS- MOLOGICAL RESEARCH LETTERS. 86 (1): 89. Ferdowsi, , Griffa, R. A. Guyer, P. A. Johnson, Marone, and J. a. n. Carmeliet. Three-dimensional discrete element modeling of triggered slip in sheared granular media. 2014. PHYSICAL REVIEW E. 89 (4). Harding, J. Patrick, C. L. Fryer, and S. Mendel. Explaining TeV Cosmic-Ray Anisotropies with Non-Diffusive Cos- mic-Ray Propagation. 2015. High Energy Astrophysical Phenomena. 1: 26. Hertzfeld, U. C., E. C. Hunke, B. McDonald, and B. Wallin. Sea ice deformation in Fram Strait— Comparison of CICE simulations with analysis and classification of air- borne remote-sensing data. To appear in Cold Regions Science and Technology. McDowell, N. G., N. C. Coops, P. S. A. Beck, J. Q. Chambers, Gangodagamage, J. A. Hicke, Huang, Kennedy, D. J. Krofcheck, Litvak, A. J. H. Meddens, Muss, Negron- Juarez, Peng, A. M. Schwantes, J. J. Swenson, L. J. Ver- non, A. P. Williams, Xu, Zhao, S. W. Running, and C. D. Allen. Global satellite monitoring of climate-induced vegetation disturbances. 2015. TRENDS IN PLANT SCI- ENCE. 20 (2): 114. Paris, M. W., E. B. Grohs, G. M. Fuller, and C. Kishimoto. Toward a Unitary and Self-Consistent Treatment of Big Bang Nucleosynthesis. 2015. Los Alamos National Laboratory for public release. Ranasinghe, N. R., A. C. Gallegos, A. R. Trujillo, A. R. Blanch- ette, E. A. Sandvol, Ni, T. M. Hearn, Tang, S. P. Grand, Niu, Y. J. Chen, Ning, Kawakatsu, Tanaka, and Obayashi. Lg attenuation in northeast China using NECESSArray data. 2015. GEOPHYSICAL JOURNAL INTERNATIONAL. 200 (1): 67. 693

Page 694

Science of Signatures Directed Research Final Report High Performance Atom-Based Sensors for Fields and Rotations Malcolm G. Boshier 20130058DR Abstract simpler technology demonstrated in the laboratory, but The National Security community needs improved sen- which has not been easy to use in the field. The four sors for magnetic fields, gravity, acceleration, and rota- main research objectives and their outcomes are as fol- tions that advance the state of the art in performance, lows. size/weight/power requirements, and ruggedness. To

1. To realize a waveguide atom interferometer, in which meet this challenge, we have used unique LANL technol- matter waves confined in a waveguide trap are split ogy for manipulating ultracold atoms to demonstrate into two sections that use matter wave interference two new types of sensor. The first, called a waveguide to measure acceleration or an energy difference. atom interferometer, is able to measure tiny energy We were able to build this device with our atom differences and will be an extremely accurate device for manipulation technology and measure an energy gravimetry. The second sensor is called an Atom-SQUID. difference equivalent to a change in temperature of We have shown how to build an Atom-SQUID, and also a few billionths of a degree. provided the first demonstration that they can be sensi- tive detectors of rotation. In addition, we have taken
2. To realize a rotation sensor based on an atomic ana- atomic magnetometer technology from the laboratory log of the Superconducting Quantum Interference to the field with the construction of a battery-powered, Device (SQUID). During this project we demonstrat- human-portable atomic magnetometer. Finally, we de- ed the world’s first DC Atom-SQUID, and we then veloped the theory of a new type of force sensor based showed that when it is rotated its properties change on a circuit containing an Atom-SQUID. The results of just as theory predicts. this project have informed Defense Community science panels about the potential of cold atom sensors, been 3. To build a portable atomic magnetometer to enable leveraged to start a Strategic Partnership project, and this technology, which had previously only been started a new research direction of using atomic magne- applied in the laboratory, to be used in the field. tometers as compact low-frequency antennas. During this project we developed a magnetic field sensor with the state of the art sensitivity, and then Background and Research Objectives engineered it into a self-contained, human-portable, The National Security community needs improved sen- battery-powered instrument. sors for magnetic fields, gravity, acceleration, and rota- tions that advance the state of the art in performance, 4. To investigate new theoretical approaches to sensing size/weight/power requirements, and ruggedness. The with quantum systems. In the course of developing overarching goal of this project is to develop new atom- the theory for our Atom-SQUID device, our team based sensors for these fields and motions. The gravity, discovered a new approach to sensing forces with a acceleration, and rotation sensors exploit the quantum variation of that device. mechanical result that when atoms are made sufficiently Scientific Approach and Accomplishments cold they behave as waves. These “matter waves” can create interference patterns, somewhat like the colored The technical approach to the cold atom sensors applied fringes seen in a soap bubble, which are exquisitely a unique LANL atomic physics capability, the “Painted sensitive to interactions or forces felt by the atoms, such Potential”, which allows for arbitrary manipulation of as gravity or rotation. The magnetic field sensor uses a matter waves [1] (Figure 1). In this system, atoms in a vacuum cell are cooled to near absolute zero, forming a 694

Page 695

so-called Bose-Einstein condensate (BEC), using standard techniques from the field of laser cooling and trapping. The cold atoms are attracted to the center of a focused laser beam, in much the same way that a sweater with a static charge attracts lint. The Painted Potential system moves the focused laser so fast that the atoms experience the time-averaged intensity distribution (exactly like a laser light show). With suitable computer programming we can trap the BEC atoms in any shape that could be drawn on a sheet of paper, and then move them around as we please. Figure 1. Schematic diagram of the Painted Potential system. To realize the waveguide atom interferometer, one needs to start with a single BEC and then divide it into two pieces very gently, so as not to disrupt the matter waves in each half. Previous experiments have accomplished this by ef- fectively moving apart two traps that were initially over- lapped. This approach unavoidably disturbs the BEC, which in turn limits the measurement time and hence the sensi- tivity of the device. We used the flexibility of our Painted Potential system to demonstrate a much gentler approach to splitting the BEC (Figure 2). The BEC is first created in a U-shape (Figure 2a). We then slowly move the connecting arm in the U away from the two legs, which very gently turns a single BEC to two identical BECs (Figure 2b). The Figure 2.(b) Image of the BEC after the arm of the U has been phase of each BEC then evolves at a rate which is propor- moved away to leave two BECs. (c) Interference fringes formed tional to its energy. After allowing this evolution for some when the Painted Potential is turned off and the BECs expand time, the trapping laser beams are turned off. Freed from and overlap. (d) Data showing that the measured interferometer confinement, the two BECs expand to overlap with each phase difference changes as expected when an energy differ- other, where they form matter wave interference fringes. ence is imposed on the two BECs. Inset shows how the spread in We record the density variation of the fringe pattern by phases increases with the time between splitting and measure- taking an image with a video camera (Figure 2c). A differ- ment. The blue dashed line indicates the point where it is no ence in energy between the two BECs results in a differ- longer possible to determine the interferometer phase difference. ence in phase which shifts the fringe pattern. We imposed changed, showing that the device operates as an atom a known energy difference by changing the intensity of interferometer and that it is sensitive enough to measure the painting laser when it illuminated one of the two BECs. an energy difference equivalent to a few billionths of a de- Figure 2d shows our measurement of the expected shift in gree in temperature. The inset to Figure 2d shows how the the fringe pattern as the energy difference is spread in the phase difference measurements increases 695

Page 696

as the time between splitting and interfering is increased. with rotation frequency, just as theory predicts. Refining The increase is due to the combination of unavoidable ran- our simulations of this behavior will allow us to design an dom fluctuations in the number of atoms in each BEC with optimal Atom-SQUID for rotating sensing and predict its the fact that the rubidium-87 atoms we are currently using performance. The observation of JJ behavior was pub- repel each other. That effect adds an additional random lished in Physical Review Letters [3]. A paper reporting our energy difference which currently limits the measurement observation of the effect of rotation on the Atom-SQUID is time and hence the sensitivity. In a BEC of potassium-39 in preparation. atoms, it turns out that applying a magnetic field with the right strength tunes the interactions to zero, which would remove this limitation on performance. A paper describing our new type of atom interferometer and its performance is in preparation. The Atom-SQUID device is based on a BEC created in a toroidal Painted Potential. This toroidal BEC is very much a quantum mechanical object, as shown by our observation that it can only be made to rotate at certain discrete rates [2]. The toroidal BEC becomes a SQUID when two thin barriers are painted on the toroidal potential (that is easily accomplished by reducing the intensity of the attractive laser beam when it passes over the desired barrier loca- tions) (Figure 3a). We performed experiments to demon- strate that our barriers behave as what are known as ideal Josephson Junctions (JJs) [3]. The JJ has the property that a small flow of atoms can pass through it (by quantum mechanical tunneling), but as the flow is increased a point is reached where the barrier becomes impenetrable. That flow rate corresponds to the so-called critical current. We were able to make an atomic current flow through the barriers by moving them towards each other. When they are moved slowly, atoms tunnel from one side to the other, keeping the density the same in both segments of the ring (Figure 3b). However, when the barriers are moved fast, the tunneling switches off and the atoms are compressed to high density in one side of the barriers and expanded to low density in the other (Figure 3c). Locating the transition point between these regimes measures the critical current. The critical current behavior that we measured was found to be in excellent agreement with our computer simula- Figure 3. Images showing atoms in the Atom-SQUID. Red = high tions of the system. The main reason for our interest in the density, blue = low density. The diameter of the Atom-SQUID is 8μm. (a) The BEC initially formed in the Atom-SQUID, with the Atom-SQUID is that theory predicts that the Atom-SQUID Josephson Junction (JJ) barriers opposite each other. (b) The final critical current should be a periodic function of rotation distribution of atoms when the JJs have been slower than the frequency. This is appealing because the analogous ef- critical velocity, which allows atoms to tunnel through the JJs and fect in the standard superconducting SQUID allows those maintain constant density on each side of the JJs. (c) The final devices to function as extremely sensitive magnetometers. distribution of atoms when the JJs have been moved faster than We have just become the first group to observe the effect the critical velocity. No tunneling is possible, so the atoms in one of rotation on an Atom-SQUID. We repeated the proce- side are compressed to higher density and those in the other side dure to measure the critical current, but this time with the are expanded to lower density. Painted Potential rotating about the center of the Atom- Work by a number of groups around the world has estab- SQUID. Figure 4 shows the result. Here the critical point lished the atomic magnetometer as the most sensitive is located by measuring how many atoms must be added instrument for measuring magnetic fields in the labora- to the SQUID before it switches into the tunneling regime. tory. Since there are mission-relevant applications of The graph shows this critical atom number oscillating 696

Page 697

magnetometers in the field (e.g. using the magnetic field Finally, members of our team developed theory for the emitted by power lines to determine the nature and status devices described above and also performed numerical of equipment being powered by those lines), our third simulations. An unexpected discovery from this work was objective was to take this technology out of the laboratory. a new concept for a force sensor. The device would include We first designed a magnetic field sensor head with state an Atom-SQUID connected inside a larger loop (a geometry of the art sensitivity over a wide range of magnetic field which is easily realized with the Painted Potential). If an frequencies. This work was reported in Applied Physics atom current is established in the larger loop, quantum- Letters [4]. Next, the engineers in our team designed and mechanical constraints on the atom velocity mean that built a battery-powered portable control and data acquisi- shrinking the dimensions of that loop will increase the tion system for this head, using a combination of custom current, until it reaches the critical current of the Atom- circuit boards and compact commercial data acquisition SQUID. We have found that the loop size at which the hardware. The resulting unit (Figure 5) is easily carried by critical point is reached is very sensitive to any forces act- one person and can operate off rechargeable batteries for ing on the atoms in the Atom-SQUID. This work, along with several hours. supporting simulations, is being written up for publication. Impact on National Missions Improved rotation sensors, gravity sensors and magne- tometers are of broad interest to many agencies in the Intelligence and Defense Communities. Device applications include respectively inertial navigation, detection of under- ground structures and of oil or mineral deposits, and the use of power lines for communication and for characteriza- tion of activities in inaccessible facilities. The devices based on quantum technologies that are the focus of this project offer potential gains in Size, Weight, and Power (SWaP) and in sensor performance. The work performed in this project has been presented, at their invitation, to several science panels investigating future technology needs and opportunities for various parts of the Defense Community. Some of our work has already been leveraged to establish a new Work For Others Figure 4. Measurements showing that the critical atom number program. The success of the portable atomic magnetom- in the Atom-SQUID varies periodically when the device is rotated. eter component has enabled a new research direction to This demonstrates that the Atom-SQUID can act as a rotation quantify the performance of atomic magnetometers as sensor. compact low-frequency antennas. The project has also established Los Alamos National Labo- ratory as a world leader in the emerging field of Atomtron- ics, which seeks to build analogs of optical and electronic circuits based on matter waves flowing in waveguides. This new technology holds great promise for realizing better sensors and for providing new approaches to signal processing. References

1. Henderson, , Ryu, MacCormick, and M. G. Boshier. Experimental demonstration of painting arbitrary and dynamic potentials for Bose-Einstein condensates.
2. NEW JOURNAL OF PHYSICS. 11. Figure 5. The battery-powered, human-portable atomic magne- 2. Ryu, , K. C. Henderson, and M. G. Boshier. Creation of tometer constructed as part of this project. matter wave Bessel beams and observation of quan- tized circulation in a Bose-Einstein condensate. 2014. 697

Page 698

NEW JOURNAL OF PHYSICS. 16. 3. Ryu, , P. W. Blackburn, A. A. Blinova, and M. G. Boshier. Experimental Realization of Josephson Junctions for an Atom SQUID. 2013. PHYSICAL REVIEW LETTERS. 111 (20). 4. Savukov, , Karaulanov, and M. G. Boshier. Ultra-sensi- tive high-density Rb-87 radio-frequency magnetom- eter. 2014. APPLIED PHYSICS LETTERS. 104 (2). Publications Blinova, A. A., M. G. Boshier, and E. Timmermans. Two po- laron flavors of the Bose-Einstein condensate impurity. 2013. Physical Review A. 88: 053610. Campo, A. del. Shortcuts to adiabaticity by counter-diabat- ic driving. 2013. Physical Review Letters. 111: 100502. Campo, A. del, I. L. Egusquiza, M. B. Plenio, and S. F. Huel- ga. Quantum Speed Limits in Open System Dynamics. 2013. Physical Review Letters. 110: 050403. Ryu, C., K. C. Henderson, and M. G. Boshier. Creation of matter wave Bessel beams and observation of quan- tized circulation in a Bose–Einstein condensate. 2014. New Journal of Physics. 16: 103406. Ryu, C., P. W. Blackburn, A. A. Blinova, and M. G. Boshier. Experimental realization of Josephson junctions for an atom SQUID. 2013. Physical Review Letters. 111: 205301. Savukov, I., T. Karaulanov, and M. G. Boshier. Ultra-sensitive high-density Rb-87 radio-frequency magnetometer. 2014. Applied Physics Letters. 104: 023504. Torrontegui, E., S. Ibanez, S. Martinez-Garaot, M. Modug- no, A. del Campo, D. Guery-Odelin, A. Ruschhaupt, X. Chen, and J. G. Muga. Shortcuts to Adiabaticity. 2013. ADVANCES IN ATOMIC, MOLECULAR, AND OPTICAL PHYSICS, VOL 62. 62: 117. Zurek, W. H.. Quantum Darwinism, classical reality, and the randomness of quantum jumps. 2014. Physics Today. 67 (10): 44. Zurek, W. H.. Probabilities from entanglement and the origin of quantum jumps. 2013. In 25th Solvay Confer- ence on Physics. (Brussels, October 2011). , p. 54. Sin- gapore: World Scientific. Zwolak, , C. J. Riedel, and W. H. Zurek. Amplification, Re- dundancy, and Quantum Chernoff Information. 2014. PHYSICAL REVIEW LETTERS. 112 (14). 698

Page 699

Science of Signatures Directed Research Final Report Battlefield MRI Michelle A. Espy 20130121DR Abstract highly shielded environment have also begun to be tack- Magnetic Resonance Imaging (MRI) is the best method led. In the course of this project, we sought to accom- for non-invasive imaging of soft tissue anatomy, saving plish the engineering to integrate these capabilities and countless lives each year. But conventional MRI relies move ULFMRI from a proof-of-concept in our laboratory on very high fixed strength magnetic fields > 1.5 T, with to a functional prototype that will exploit the inherent parts-per-million homogeneity, requiring large and ex- advantages of the approach, and increase accessibility. pensive magnets. The overwhelming technological trend Background and Research Objectives has been toward ever higher magnetic fields because the signal (sample magnetization), and the sensitivity of Magnetic Resonance Imaging (MRI) is the best method the Faraday detectors traditionally used scale with the for non-invasive imaging of soft tissue anatomy, saving applied magnetic field (readout frequency). Thus the countless lives each year. But conventional MRI relies benefit to signal-to-noise scales as the magnetic field on very high fixed strength magnetic fields > 1.5 T, with squared. Thus, bigger fields have been widely accepted parts-per-million homogeneity, requiring large and ex- as the only way to obtain quality images. But there is a pensive magnets. The overwhelming technological trend price: high field (HF) MRI is a first-world method that can has been toward ever higher magnetic fields because only be used in highly controlled settings in well-funded the signal (sample magnetization), and the sensitivity of medical centers where such large magnetic fields can the Faraday detectors traditionally used scale with the be generated and do not pose a hazard. MRI machines applied magnetic field (readout frequency) [1]. Thus, weigh many tons, cost $3-5million,andtypicallyrequirebiggerfieldshavebeenwidelyacceptedastheonlyway100-1000litersofcryogens.Thehighmagneticfieldsre-toobtainqualityimages.Butthereisaprice:highfieldstrictaccessotherwaysaswell:HFMRIisnotavailablein(HF)MRIisafirst-worldmethod[2,3]thatcanonlyberuralsettings,andisnotdeployabletoemergencysitua-usedinhighlycontrolledsettingsinwell-fundedmedicaltionsorbattlefieldhospitals.Thecostofaconventionalcenterswheresuchlargemagneticfieldscanbegener-HFMRIsystemiswellbeyondwhatcanbeaffordedforatedanddonotposeahazard.MRImachinesweighbillionsofpeopleinpooranddevelopingcountries.Sub-manytons,cost$3-5 million, and typically require 100- jects with unknown medical histories (e.g. unconscious, 1000 liters of cryogens. The high magnetic fields restrict soldiers with possible shrapnel injuries, etc.) cannot be access other ways as well: HFMRI is not available in rural imaged due to the potential hazards of heating or mov- settings, and is not deployable to emergency situa- ing of metal in the body. Even non-ferrous metal sig- tions or battlefield hospitals. The cost of a conventional nificantly distorts a HFMRI, precluding imaging of those HFMRI system is well beyond what can be afforded for with medical implants. billions of people in poor and developing countries. Sub- jects with unknown medical histories (e.g. unconscious, Recent demonstrations have shown that MRI can be soldiers with possible shrapnel injuries, etc.) cannot be performed at much lower magnetic fields (100 μT – 100 imaged due to the potential hazards of heating or mov- mT, the ULF or ultra-low field regime). Through the use ing of metal in the body. Even non-ferrous metal sig- of pulsed pre-polarization and SQUID detection, much nificantly distorts a HFMRI, precluding imaging of those of the signal loss can be mitigated. We and others have with medical implants. shown ULFMRI of the human hand, knee, arm, and brain, including imaging in the presence of (and even Recent demonstrations have shown that MRI can be through) metal. The difficulties of operation outside of a performed at much lower magnetic fields (100 μT – 100 699

Page 700

mT, the ULF regime). Through the use of pulsed pre-po- ity ULF MRI images of the human brain recorded, over a larization [4] and SQUID detection [5], much of the signal very large (200 cm3) field of view. This result is summa- loss can be mitigated. We and others have shown ULFMRI rized in the brain images shown in Figure 1, acquired in our of the human hand [6], knee [7], arm [8, 9], and brain [10, V1 system. As our ultimate aim is unshielded operation, 11, 12], including imaging in the presence of (and even we also demonstrated a second-generation V2 system that through) metal. The large remaining challenges were to performs ULF MRI outside of heavy magnetically-shielded improve upon and integrate the existing methods into a enclosure, using field and adaptive noise cancellation functional prototype, while moving outside of the heavy methods. These results are summarized in Figure 2. The magnetic and/or radio-frequency shields in which ULF MRI results from both V1 and V2 were validated against a com- systems typically operate [13]. In the course of this project, prehensive modeling capability that was developed for this we accomplished the engineering to demonstrate ULF MRI project, and provides a quantitative method for designing that exploits the inherent advantages of the approach, and assessing the image quality from ULF MRI systems. and shows true potential to increase accessibility of MRI This result is summarized in Figure 3, showing excellent outside of traditional settings. agreement between our images (Figure 1) and the model (Figure 3). We also addressed the issue of flux trapping in Our research objectives were to develop a portable the SQUID gradiometer coils after exposure to the large “Battlefield” MRI machine based on SQUID (superconduct- pre-polarization field. At the end of the project we have ing quantum interference device) sensor technology and demonstrated two capable ULF MRI systems. ULFMRI techniques developed at LANL. Our motivation was the lack of MRI technology available in any combat support hospital, despite of the promise of the method for early diagnosis of traumatic brain injury. The device was intended to provide a diagnostic quality image in a field deployable package. Specific objectives were to increase the pre-polarization field used to enhance the MRI signal, advance the SQUID detector tolerance to the changing magnetic field environment associated with MRI, without increasing the noise floor, and implement a novel “adap- tive” open-coil shield to replace the tons of metal typically required in conventional MRI. In the course of the project we believe we have demon- strated all these goals and shown the proof-of-concept for an imager that employs very low magnetic fields, which can be completely turned off during transport or when not in use. Most importantly, because the device operates Figure 1. Horizontal (top down) image of the human brain ac- in a fundamentally different regime than a high field MRI quired with the Los Alamos National Laboratory ULFMRI system machine, it will be able to perform unique imaging tasks developed with this project. The MRI was acquired at 100 uT that traditional MRI cannot, such as imaging in the emer- (about twice the earth’s magnetic field) and 10,000 times less gency room or in the presence of metal. Battlefield MRI than a conventional MRI. will bring the power of MRI to settings where it is presently The V1 shielded system: A photograph of the V1 system not possible is shown in Figure 4. The sensor system in the V1 ULFMRI system consists of 7 second-order axial gradiometers with Scientific Approach and Accomplishments pick-up loop diameter and baseline ~90 mm. The gradiom- Summary: Our scientific approach toward demonstration eter intrinsic field resolution is 0.5 fT/√Hz. Johnson noise of a fieldable prototype was two-pronged. A first version from the RF shield increases this to ~1.3–1.5 fT/√Hz. To (V1) system operating inside a magnetically shielded room shield the SQUIDs from the high magnetic fields from the (MSR) was engineered for the highest possible signal and pre-polarization coil we use shields made of a eutectic lead lowest possible noise. The V1 system was also used to bismuth alloy. Importantly for this project, we implement- develop and demonstrate highly stable and low noise mag- ed a method of 2nd feedback [14] that was used to reduce netic field generation; pre-polarization; SQUID feedback low frequency transients and keep the SQUIDs within their and noise reduction; and “fast” imaging, pulse sequences. dynamic range after pulsing the magnetic fields. Using this system we have demonstrated the highest qual- 700

Page 701

The polarization coil is a 510-turn copper coil. It typically When performing MRI, the noise rises to ~ 1.6–2 fT/√Hz operates at 100 A for a field at the center of 100 mT. The near 8 kHz where the images are acquired. The 1/f nature polarization field could ramp to 100 mT over a 200 cm3 of the noise even out to kHz is a phenomenon seen during volume and could be removed in < 50 ms. More impor- pulsing, whereas the noise is flat if the system is not active. tantly, the shape of the pre-polarization field (Bp) was We infer that this noise is associated with flux trapping in optimized to minimize the effects of transients and magne- the niobium wire. Flux trapping in the niobium wire tradi- tization of the MSR during and after the dynamic magnetic tionally used in SQUID gradiometers has been the primary field changes associated with MRI. This reduced the “dead limitation to increasing Bp beyond 100 mT in both our, time” after Bp removal, which in turn provided higher and our colleagues approaches to ULFMRI. To address this signal-to-noise in the images. The coil is liquid nitrogen issue we demonstrated a method of gradiometers made of (LN) cooled and consumes 1.6 L/min of LN at 100 A. While tantalum, which can be easily and rapidly heated to expel this coil was used for the results presented below, we also flux. The noise from Bp pulsing was thus reduced. This ap- completed development and testing of a water-cooled coil proach will become increasingly important as Bp is raised capable of achieving 250 mT. from 100 mT to 250 mT, and the technique has resulted in a patent application presently under LANL legal review. Figure 2. Simulated ULFMRI based on a model developed under this LDRD project. The parameters are the same as those in the experimentally acquired image shown in Figure 1. However, the brain used in the model is different. The similarities in signal-to- noise and image quality are quite similar, giving us confidence in Figure 3. MRI for a gelatin brain phantom acquired in the un- the model’s ability to predict diagnostic ability. Also shown is a shielded MRI (V2). The small simulated stroke inclusion is noted spherical signal from water for reference. by a circle. The magnetic fields required for the remaining MRI The ULF MRI system described above operated inside a ~ sequence are provided by the coil set shown in Fig. 4. Pa- 3×3×4 m3 magnetically shielded room (MSR) made of two rameters of the coils and amplifiers are provided in Table II. pairs of alternating layers of aluminum and mu-metal. This Our data acquisition (DAQ) was a commercial off-the-shelf room is in part a legacy of our combined ULF MRI and MEG a National Instruments (NI) 4497 PXI card. To galvanically efforts; such shielding is required in MEG to remove very isolate the DAQ reference path (zero volts, or ground) from low frequency noise. However, it is heavy and inhibits por- that of the SQUIDs we used audio transformers. The SQUID table or inexpensive systems. Thus, one additional goal of electronics were grounded to a single point on the MSR our work was to reduce or eliminate the MSR entirely. To where the RF shield on the cryostat and Bp coil were also that end, we set up a second (greatly simplified) ULF MRI grounded. system operating outside of any MSR. Critical to maintaining the SQUIDs within their dynamic The V2 unshielded system: A photograph of the V2 system range was the development of an adaptive field compen- is shown in Figure 5. The ULF MRI working in unshielded sation coil set (Comp coils in Figure 4) which was able to environment consisted of gradient and measurement dynamically cancel out the magnetic field transients associ- field coils as described in [15], and cryostat and 7-channel ated with eddy currents in the MSR. Although the final goal insert as used in [10, 11]. Three pairs of square Helmholtz is operation outside the MSR, compensation coils are still coils, surrounding the gradient coil system, were added a critical demonstration because they can be implemented for cancellation of the Earth’s magnetic field. The cancel- to maintain a stable magnetic field in the presence of dy- lation coils were powered by regular power supplies and namically varying fields present in an open environment. adjusted until a fluxgate in the sample space showed zero 701

Page 702

fields in three directions. The measurement field, which simulate the signal and images from the brain we used the was on all the time, was generated by a power supply with Montreal Neurological Institute digital brain model “Colin a large capacitor across its terminals to reduce the noise. 27”. The gradients and spin-flip field were generated by battery- powered current generators receiving input voltages from an NI-6733 card. The pre-polarization coil was the same as used in the shielded system described above, with field generated by three sealed 12 V lead-acid batteries in series and ramped down through banks of solid-state switches both in parallel and series to handle both the required volt- age and current. Using this system we were able to demon- strate MRI (Figure 3) of tissue phantoms taken unshielded and in noisy, active laboratory environment [13]. Figure 5. Photograph of the unshielded (V2) ULFMRI system. Impact on National Missions This proposal directly addresses the issue of improved ac- cessibility to healthcare, the ability to provide care at the point of trauma (theaters of war, emergency or disaster response), our fundamental understanding of the human brain, and global health equality. All of these aspects are of specific concern to National security. Customers from DoD, DARPA, and NIH continue to spe- cifically request proposals in support of the issues listed above. The LDRD support for this effort represented a critical and timely “next step” in LANL’s development of a world-leading capability in ULFMRI and applications. We arguably maintain at LANL the world’s expertise in ULFMRI Figure 4. Photograph of the shielded (V1) ULFMRI system. methods. Developing the ULF MRI model: We have developed a The work accomplished under this grant also enhances predictive model of our ULFMRI systems. The excellent LANL’s technical capability to serve the Nation in other agreement between data and this model can be seen in areas of R&D including Science of Signatures and Biosci- the comparison between Figure 1 and 2. Using this model ence. We anticipate a broad range of additional applica- we have shown good contrast for a 1 cm inclusion with the tions from basic research (NIH, e.g. ULFMRI combined with relaxation properties of 50% blood and 50% brain tissue, functional methods, fundamental understanding of brain, indicating sensitivity to bleeding in the brain [13]. A more novel contrast at ULF for new diagnostics; global security, complete discussion can be found in [16]. To simulate the e.g. disaster response, non-invasive screening). We also magnetic fields produced at the sample, the response of anticipate significant potential for transition to industry magnetization in the sample, and the signal detected in partners. In fact, negotiations are actively in place at the our sensor coils we use a simulation written in MATLAB, time of this writing for follow-on work with a company to based on the Biot-Savart-LaPlace formulation of magne- which two patents have already been licensed. Outside tostatics and the reciprocity principle. Spin evolution and of the battlefield application, the most compelling need NMR signals are computed using the Bloch equations. To for this technology is in developing countries. Sharing 702

Page 703

our prosperity in health care with the world is a national Igor M. Savukov, Michelle A. Espy, John C. Mosher, security interest as well as a humanitarian imperative [17]. John J. Gomez, and Robert H. Kraus Jr.. Microtesla MRI This work belonged squarely within the National Labora- of the human brain combined with MEG. 2008. Journal tory where the fundamental groundwork could be done, of Magnetic Resonance. 194 (1): 115 . decoupled from immediate need for a profit. 11. Magnelind, P. E., J. J. Gomez, A. N. Matlashov, T. References Owens, J. H. Sandin, P. L. Volegov, and M. A. Espy. Co- Registration of Interleaved MEG and ULF MRI Using a 7

1. Clarke, John, Michael Hatridge, and Michael Mößle. Channel Low-SQUID System. 2011. Applied Supercon- SQUID-Detected Magnetic Resonance Imaging in ductivity, IEEE Transactions on. 21 (3): 456. Microtesla Fields. 2007. Annual Review of Biomedical Engineering. 9 (1): 389.
2. Vesanen, P. T., J. O. Nieminen, K. C. Zevenhoven, J. Dabek, L. T. Parkkonen, A. V. Zhdanov, J. Luomahaara, J.
3. Ubelacker, S.. More MRI cash helping rich more than Hassel, J. Penttilä, J. Simola, A. I. Ahonen, J. P. Mäkelä, poor, study finds . 2009. The Globe and Mail. and R. J. Ilmoniemi. Hybrid ultra-low-field MRI and
4. Health Care Resources: Medical Technology. 2014. magnetoencephalography system based on a commer- Organization for Economic Co-Operation and Develop- cial whole-head neuromagnetometer. 2013. Magnetic ment. Resonance in Medicine. 69 (6): spcone.
5. Macovski, A., and S. Conolly. Novel approaches to low- 13. Espy, M., P. Magnelind, A. Matlashov, S. Newman, cost MRI. 1993. Magnetic Resonance in Medicine. 30 A. Urbaitis, and P. Volegov. Toward High Resolution (2): 221. Images With SQUID-Based Ultra-Low Field Magnetic Resonance Imaging. 2013. Applied Superconductivity,
6. Myers, W., D. Slichter, M. Hatridge, S. Busch, M. IEEE Transactions on. 23 (3): 1603107. Mossle, R. McDermott, A. Trabesinger, and J. Clarke. Calculated signal-to-noise ratio of MRI detected with 14. Matlashov, A. N., L. J. Schultz, M. A. Espy, R. H. Kraus, I. SQUIDs and Faraday detectors in fields from 10$mu$T M. Savukov, P. L. Volegov, and C. J. Wurden. SQUIDs vs. to 1.5 T. 2007. Journal of Magnetic Resonance. 186 (2): Induction Coils for Ultra-Low Field Nuclear Magnetic
7. Resonance: Experimental and Simulation Comparison.
8. IEEE Transactions on Applied Superconductivity.
9. Zotev, V. S., P. L. Volegov, A. N. Matlashov, M. A. Espy, J. 21 (3): 465. C. Mosher, and R. H. Kraus. Parallel MRI at microtesla fields. 2008. Journal of Magnetic Resonance. 192 (2): 15. Espy, M, M Flynn, J Gomez, C Hanson, R Kraus, P Mag-
10. nelind, K Maskaly, A Matlashov, S Newman, T Owens, M Peters, H Sandin, I Savukov, L Schultz, A Urbaitis, P
11. Espy, M., M. Flynn, J. Gomez, C. Hanson, R. Kraus, P. Volegov, and V Zotev. Ultra-low-field MRI for the detec- Magnelind, K. Maskaly, A. Matlashov, S. Newman, M. tion of liquid explosives. 2010. Superconductor Science Peters, H. Sandin, I. Savukov, L. Schultz, A. Urbaitis, P. and Technology. 23 (3): 034023. Volegov, and V. Zotev. Applications of Ultra-Low Field Magnetic Resonance for Imaging and Materials Stud- 16. Espy, M., P. Volegov, and I. Savukov. Chapter 8 - Hard- ies. 2009. Applied Superconductivity, IEEE Transactions ware Developments: Detection using Squids and on. 19 (3): 835. Atomic Magnetometers. 2016. In Mobile NMR and MRI . Edited by Johns, M.. p. tbd. London: Royal Society of
12. Seton, H C, J M S Hutchison, and D M Bussell. A 4.2 K Chemistry. receiver coil and SQUID amplifier used to improve the SNR of low-field magnetic resonance images of the 17. Bonventre, E. V., K. H. Hicks, and S. M. Okutani. U.S. human arm. 1997. Measurement Science and Technol- National Security and Global Health: An Analysis of ogy. 8 (2): 198. Global Health Engagement by the U.S. Department of Defense-A Report of the CSIS Global Health Policy
13. Mossle, M., W. R. Myers, S. K. Lee, N. Kelso, M. Hat- Center-Working Draft. 2009. ”. Washington DC: Center ridge, A. Pines, and J. Clarke. SQUID-detected in vivo for Strategic and International Studies. MRI at microtesla magnetic fields. 2005. IEEE Transac- tions on Applied Superconductivity. 15 (2): 757. Publications Espy, M.. Applications of SQUID Detected MRI. Invited
14. Zotev, Vadim S., Andrei N. Matlashov, Petr L. Volegov, presentation at International Society for Magnetic 703

Page 704

Resonance in Medicine. (Salt Lake City, Utah, 4-7 April, 2013). Espy, M. A., P. E. Magnelind, A. N. Matlashov, S. G. New- man, H. J. Sandin, L. J. Schultz, R. Sedillo, A. V. Urbaitis, and P. L. Volegov. Progress toward a deployable SQUID- based ultra-low field MRI system for anatomical imag- ing. 2015. IEEE Transactions on Applied Superconduc- tivity. 25 (3): pg 1601705. Espy, M., P. Magnelind, A. Matlashov, S. Newman, A. Ur- baitis, and P. Volegov. Toward High Resolution Images With SQUID-Based Ultra-Low Field Magnetic Reso- nance Imaging. 2013. Applied Superconductivity, IEEE Transactions on. 23 (3): 1603107. Espy, M., P. Volegov, and I. Savukov. Chapter 8 - Hardware Developments: Detection using Squids and Atomic Magnetometers. 2016. In Mobile NMR and MRI . Edited by Johns, M.. , p. tbd. London: Royal Society of Chemistry. Kraus, R. H., M. Espy, P. Magnelind, and P. Volegov. Ultra- Low Field Nuclear Magnetic Resonance: A New MRI Regime . 2014. Magnelind, P E, Y J Kim, A N Matlashov, S G Newman, P L Volegov, and M A Espy. Toward early cancer detection using superparamagnetic relaxometry in a SQUID- based ULF-MRI system. 2014. Superconductor Science and Technology. 27 (4): 044031. Matlashov, A., P. Magnelind, H. Sandin, M. Espy, A. Ander- son, and H. Mukundan. SQUID instrumentation for early cancer diagnostics. 2013. In Superconductive Electronics Conference (ISEC), 2013 IEEE 14th Interna- tional. , p. 1. Matlashov, A., P. Magnelind, S. Newman, H. Sandin, A. Urbaitis, P. Volegov, and M. Espy. Multi-channel SQUID- based Ultra-Low Field Magnetic Resonance Imaging in Unshielded Environment. 2015. In International Su- perconducting Electronics Conference 2015. (Nagoya, Japan, 6-9 July 2015). , p. 1. IEEE/CSC & ESAS Super- conductivity News Forum: IEEE Xplore. Volegov, P. L., L. J. Schultz, and M. A. Espy. On a ghost ar- tefact in ultra low field magnetic resonance relaxation imaging. 2014. Journal of Magnetic Resonance. 243: 98. 704

Page 705

Science of Signatures Early Career Research Continuing Project Laser-Driven Neutron Source for Detection of Nuclear Material Andrea Favalli 20140580ECR Introduction throughput interrogation of transportainers at ports, as Los Alamos National Laboratory has pioneered a new well as in treaty verification. Many more applications short duration yet extremely intense source of neutrons and extension to other signatures can be expected, for using laser-driven targets, with world record beating example the replacement of nuclear reactors used for flux. This offers the potential for an entirely new way of neutron production at universities. nondestructively assaying (NDA) nuclear materials for safeguards and other purposes. Potential advantages of Progress this neutron source include the tailoring of the neutron Detecting shielded nuclear material in transport remains energy and angular distribution for particular applica- a challenging task. The need is for a fast, movable, and tions. operationally safe neutron source featuring energy tunable and high directional neutron fluence, overcom- The ultimate goal is to develop and optimize the process ing the limitations of present methods, which typically of neutron production for safeguards and homeland se- produce isotropic monoenergetic neutrons. Recent ex- curity applications. Pursuant to that goal, this project will periments show that a short-pulse laser-driven neutron perform the measurements needed to establish yield, source possesses these features, allowing for the inter- temporal and energetic characteristics of the neutron rogation of packages with variable shielding conditions beam for different configurations of target/converter and increased signal for the interrogation while improv- (by changing converter/target distance as well different ing operational safety. materials): this information is needed for Monte Carlo simulations of future instruments, and is required to ad- The goal of this project is to develop a laser-based dress the detection of nuclear material by a laser-driven method of neutron production for active detection of neutron source. Initially using depleted uranium (DU) special nuclear material. this work will use induced fission signatures: specifically neutrons following neutron fission will be detected in a During the past year the project has tackled, mainly high efficiency neutron counter. through experimental activities, the following goals: Benefit to National Security Missions Target/converter study, improvement and optimization This research directly supports the nuclear security mis- The short-pulse high-power laser is focused on a foil of sion. The demonstration of this novel technology and deuterated plastic, creating an accelerated deuteron capability will make Los Alamos a world leader in the ac- beam directed towards a neutron converter. The neu- tive interrogation of nuclear materials using laser driven trons produced from various nuclear reactions have an neutron sources. The existence and demonstration of array of energies dependent on the source properties. this type of capability is expected to lead to work for Neutron production from beryllium and copper convert- existing and new sponsors in safeguards and homeland ers was investigated using MCNP Monte Carlo transport security. The assay of fuel debris following an accident, code. Based on such simulations, we redesigned a flex- such as that at Fukushima, is one example where a ible, modular beryllium converter, composed of disks of bright source of this type combined with the measure- different thickness, surrounded by a tungsten cylinder to ment methods discussed might have important practi- enhance neutron production. cal advantages. Another example could be a very high 705

Page 706

Diagnostics and methods for neutron-beam Future Work characterization The goal of our LDRD ECR project is to develop a laser- Neutron beam characterization during the past year based method of neutron production for active detection showed a significant advance using our new diagnostics of special nuclear material. The project will end on March tools in development. We have been developing a new 2016 ( FY16) and for FY16 it will focus on the following prototype of a neutron-time-of-flight detector (nTOF) sys- tasks: tem, necessary to measure the neutron energy spectrum produced. A nTOF measurement of an intense pulse of • Data analysis and modeling of the experimental cam- neutrons using photomultiplier and plastic scintillator pres- paigns conduct in the project. This is expected to bring ents several challenges related to PMT non-linearity when a better understanding of the physics, and the technol- the intensity of the light pulse is too high. In our LDRD proj- ogy challenge, of all the process from neutron produc- ect, we have been able to develop a prototype with LANL tion to interrogation. custom-made components/devices, that was tested during • Analysis of deuteron angular and energy distribution our experimental campaign conducted at Trident, showing after the target (inside the target chamber): very good response linearity. We also worked on develop- • Analysis of neutron angular distribution (outside the ing prototypes of an on-line monitor based on detecting target chamber) source neutrons using a 3He proportional counter embed- ded in a high-density polyethylene moderator. We have • Analysis of neutron spectrum from nTOF detector developed front-end electronics to couple with the detec- • Analysis of prompt neutron signal (current measure- tors. MCNP modelling was also used here for the study of ment) for neutron interrogation the interaction of neutrons with the nTOF plastic detector, • Analysis of delayed neutron signals (high efficiency as well as in the design of the 3He based detectors. detectors) for neutron interrogation Active interrogation of special nuclear materials Conclusion For the laser-driven interrogation, we explored an experi- mental setup where high efficiency well counters were The expected results are the knowledge of the key param- used both in the so-called fast and thermal mode, where eters for the performance of active neutron interrogation either fast neutrons or thermal neutrons interrogated based on short-pulse-laser-driven neutron sources. A bet- the nuclear material, respectively. We developed a set of ter understanding of the physics of all the processes from analysis codes to extract and analyze the data. neutron production to interrogation will be a key benefit of the project. The results will be disseminated by presen- We conducted a three week experimental campaign at the tations at international conferences and through peer- Trident laser facility in the past year, where the innovations reviewed papers. The know-how developed will be the introduced above were tested. Here we report the high- guidelines for the engineering of the approach proposed light of the scientific results obtained: and will be used to write proposals to sponsors such as NA- 22, NA-24, DHS and DoD for field applications. • Record neutron beam intensity at the Trident laser facility, 1.1x1010 n/sr in one shot (surpassing the cam- Publications paign in May 2012). Favalli, A., F. Aymond, J. Bridgewater, S. Croft, O. Deppert, M. Devlin, K. Falk, J. Fernandez, D. Gautier, M. • Good reproducibility of shot-by-shot neutron yield, Gonzales, A. Goodsell, N. Guler, C. Hamilton, B. with an average yield of about 5x109 n/sr, a yield com- Hegelich, D. Henzlova, K. Ianakiev, M. Iliev, R. Johnson, parable with previous 2012 neutron record production D. Jung, A. kleinchmidt, K. Koehler, E. McCary, S. in Trident. ptimized neutron flux diagnostics improved Palaniyappan, I. Pomerantz, M. Roth, P. Santi, T. considerably the knowledge of the neutron energy Shimada, M. Swinhoe, T. Taddeucci, and G. Wurden. spectrum, especially in the region around 1 MeV, not Nuclear material detection by one-short-pulse-laser- observed before. This should been seen as an impor- driven neutron source. 2014. LA-UR-14-28697, IEEE tant improvement allowing detailed characterization of Nuclear Symposium, 2014-11-09 (Seattle, Washington, the laser-driven neutron source we have been devel- United States). oping in Trident. Favalli, A., and M. Roth. Active interrogation of sensitive • First-time demonstration of detection of 235U in sam- nuclear material using laser driven neutron beams. ple of enriched uranium by short laser-active neutron 2015. LA-UR-15-23312, International Workshop on interrogation with only a single shot. Physics of High Energy Density in Matter, 2015-01-25 706

Page 707

(Hirschegg, Austria). Goodell, J. J., A. Favalli, M. T. Swinhoe, and J. S. Hendricks. Initial MCNP Simulations of Neutron Production at the Laser-Driven Neutron Source. 2014. Poster Presentation, LANL Student Symposium, 2014-08- 05/2014-08-06, Los Alamos (NM, US). Wurden, G., A. Kleinschmidt, A. Favalli, J. Fernandez, D. Cort, S. Frydrych, A. Tebartz, O. Deppert, J. Hornung, S. Jahn, D. Schaumann, V. Bagnoud, F. Wagner, B. Zielbauer, and M. Roth. Short-pulse thin-foil neutron generation experiments using the PHELIX laser. 2015. LA-UR-15-21249, 42nd European Physical Society Conference on Plasma Physics 2015, 2015-06-22/2015- 06-26 (Lisbon, Portugal). 707

Page 708

Science of Signatures Early Career Research Continuing Project Deployment and Installation Technologies for Distributed Measurement Systems in Inconvenient/Hazardous Environments David D. Mascarenas 20140629ECR Introduction the sensor itself is not damaged. This research aims to develop modular, scalable tech- nologies/capabilities that allow us to deploy and install Benefit to National Security Missions large-scale, distributed measurement systems in an This work will strongly support LANL’s mission in Global automated, precise, reliable, cost effective manner while Security, Nonproliferation, and Environmental monitor- minimizing intrusion into the environment. This capabil- ing. It directly address the Science of Signatures pillar ity will be easy to employ from a variety of platforms with respect to Forward deployment and it supports including commonly available aerial robotics. Emplacing the IS&T pillar by providing new data streams for model measurement systems in the appropriate environment is validation and uncertainty quantification for global generally one of the most challenging and costly aspects climate modeling and energy security. It is inspired by a of any experimental or field research program, but to crosscutting approach to global security mission needs. date this aspect of technology development has been The work focuses on capability development for envi- widely ignored. Manual installation of sensors is still the ronmental sensing and nonproliferation. This work is in most common technique to deploy distributed measure- particular meant to address needs concerning the loose ment systems. This is particularly problematic when sen- nuke problem. The team recognizes that in order for this sors must be located in inconvenient environments such effort to have a high impact we must engage with both as at heights, in rivers, or in Chem/Bio/Rad hazard areas. the LANL community/customers and the broader sensor community to build useful deployment techniques. We The reason for the lack of work in this area is that the will hold a mid-project workshop to get feedback from development of a lightweight, intelligent remote sen- the LANL sensor community and global security program sor placement capability that can place sensor nodes managers on how the work can interface with customer- accurately and reliably without damaging them in the driven sensor projects. The goal is that this will lead to process is a challenging multi-disciplinary, cyber-physical follow-on collaborations and customer-driven work. system research effort. It requires taking into account dynamics, sensor measurement uncertainty, computer Progress vision, projectile dynamics, and embedded systems. The 2015 fiscal year has resulted in many accomplish- ments for the remote sensor placement project. First, This research effort intends to make Los Alamos a tech- there has been some delays in getting testing approvals. nology leader in this largely unexplored field by develop- So in order to ensure progress continued in spite of the ing new engineering tools and capabilities that facilitate delays, work has continued on technologies that will the large-scale deployment of wireless sensor networks facilitate remote sensor placement. One of the main in inconvenient environments. These techniques will be problems that needs to be explored going forward is designed with discreetness in mind so that they mini- navigating an aerial robot inside buildings in order to de- mally disturb the environment and avoid influencing ploy sensor nodes. This is an important mission for first the subjects we are trying to study as little as possible. responders as well as for security professionals. In order We will develop the high-finesse engineering tools to to address this problem, the team has been developing enable a capability to intelligently, remotely deploy sen- techniques using emerging silicon retina technology to sor nodes in such a way that when they arrive at their solve important problems such as rapidly detecting the installation site they have enough energy to be installed presence of transparent barriers. This problem is impor- correctly, but they have a low-enough energy to ensure 708

Page 709

tant because indoor areas where samples may need to be sonality research, this is the only frame-work for artificial collected and sensor nodes may need to be deployed often personality synthesis that does not use a trait-based ap- feature windows that can hamper efforts to effectively proach. In this sense it is quite innovative. navigate an aerial robot. To date a prototype intelligent transparent barrier sensor has been developed and un- Lastly, in order to mitigate any risk associated with not dergone static testing. The results of this work have been obtaining the ability to fly R&D aerial robots in FY 2016, written up and submitted to the IEEE sensors journal. The the PI is considering developing a small, simple, ground title of the paper is “A Multi-modal, Silicon Retina Tech- robot that can demonstrate the concept. It is expected nique for Detecting the Presence of Reflective and Trans- this would not take long given David’s previous experience parent Barriers.” In FY 2016 the team is planning to deploy with ground robots. Such a robot may also be used as a the intelligent transparent barrier sensor onboard an aerial platform to demonstrate the utility of ground robots in robot for dynamic testing once flight approval is obtained. emerging applications. Technical staff in the ISR division have indicated to the Future Work team that there is a need for using aerial robots to perform The main technical milestones for the 2016 fiscal year are environmental radiation surveys. The team has worked as follows: with Tyler Trombetta of the US Navy to develop a system that can fly onboard an aerial robot to collect radiation • Finalize approvals to test the remote sensor placement measurements and take photos of an environment for device. later analysis. This system is currently undergoing final • Document the work and write it up for peer-reviewed integration and the hope is that it will be testing in FY2016 journals. once flight approvals are obtained. • Present our work at applicable robotics/wireless sen- sor networking conferences. In FY2015 the team was made aware that there is interest both commercially, and in the Federal Highway Administra- Conclusion tion to use aerial robots for remote inspections. To date A successful completion of this project will result in solid significant work has gone into using aerial robots for visual first step towards a modular remote sensor deployment surveys, but this only addressed half of the procedure typi- and installation capability that can be used by the wider cally used to inspect structures. Structural inspectors will sensor networking community. The most important out- often perform tap tests and clean structures with brushes put will be an intelligent, multi-shot, remote sensor place- in order to remove debris and rust that inhibits the inspec- ment device capable of being employed from commonly tion. To date there is very little work concerned with using available aerial robotic platforms. We will work toward de- an aerial robot to physically interact with a structure to veloping modular deployment packages and characterize a facilitate remote sample collaction. To fill this gap the PI is wide array of sensor package attachment mechanisms that currently developing an instrumented tap-testing hammer can be used to connect sensor packages to a wide variety that can be used to perform remote tap-tests. In addition of materials. Finally, multiple demonstrations of the uses work is currently underway to develop motorized tools of this capability will be executed. that can be attached to an aerial robot. The research cur- rently is focusing on outfitting the quadrotor with brushes, Publications drills, and saws. Green, A., K. Klein, I. Acevedo, D. Kraus, C. Farrar, and D. Mascarenas. A Multi-modal, Silicon Retina Another area that is important for facilitating remote sen- Technique for Detecting the Presence of Reflective sor placement is developing human-machine interaction and Transparent Barriers. To appear in IEEE Sensors technologies to facilitate remote sensor deployment. The Journal. PI has spent the last four years developing a theory for synthesizing artificial personalities. Artificial personality Mascarenas, D., A. Green, T. Trombetta, and C. Farrar. A synthesis is viewed as one tool that could have significant Multirotor-based Approach For Tap-testing Difficult- impact on human-robot teaming and human interfaces to to-access Structures. To appear in 10th International a variety of cyber-physical systems. A paper on this theory Workshop on Structural Health Monitoring. (Stanford, CA, Sept 1-3). has been submitted to the Neural Information Processing conference for review. The title is, “A Jungian Type-Based Mascarenas, D., L. Ott, A. Curtis, S. Brambilla, A. Larson, Framework for Artificial Personality Synthesis.” Based on S. Brumby, and C. Farrar. Video: remote sensor a recently (2014) publish literature review on artificial per- placement. 2014. In MobiSys ‘14- Proceedings of 709

Page 710

the 12th annual international conference on Mobile systems, applications, and services. (Bretton Woods, NH, 16-19 June 2014). , p. NA. Bretton Woods, New Hampshire: ACM. Mascareñas, D., L. Ott, A. Curtis, S. Brambilla, A. Larson, S. Brumby, and C. Farrar. Demo: A remote sensor placement device for scalable and precise deployment of sensor networks . 2014. In MobiSys ‘14- Proceedings of the 12th annual international conference on Mobile systems, applications, and services. (Bretton Woods, NH, 16-19 June 2014). , p. NA. Bretton Woods, New Hampshire: NA. Yang, Y., A. Cattaneo, and D. Mascarenas. Potential Structural Health Monitoring Tools to Mitigate Corruption in the Construction Industry Associated with Rapid Urbanization. To appear in 2015 International Conference on Sustainable Development (Winner of Best Paper Award). (New York, Sept 23-24). 710

Page 711

Science of Signatures Early Career Research Continuing Project Trojan Horse Drug Development Approach: Targeting Gene Dosage Control to Induce Bacterial Suicide Sofiya N. Micheva-Viteva 20150664ECR Introduction vanced therapeutics. DoD, DHS, and NIH are interested Multi-drug resistant pathogens present an escalat- in discovery of novel antimicrobial therapies. ing threat to public health. In this project we seek to discover a novel class of antimicrobial therapies that Progress can restrict the evolution of drug resistant pathogenic Our research is focused on the identification of structural bacteria. To achieve this goal, we propose to elucidate a elements that define the stability and function of bacte- poorly understood mechanism for regulation of protein rial ribonucleic acid (RNA) macromolecules. In our mod- turnover in the bacterial cell. Messenger ribonucleic el highly conserved RNA structures regulate the activity acids (RNA) translate the genetic information stored on of RNase E, an enzyme that is essential for the survival of deoxyribonucleic acid (DNA) molecules into amino acid bacterial cells. RNase E has a key role in the generation sequence in the proteins. Gene expression is regulated of mature ribosomal RNAs and regulates the stability both on transcriptional (DNA) and post-transcriptional of messenger transcripts (mRNA) encoding for proteins (RNA) levels. We will address a fundamental science with potential cytotoxic activity. To test our model and question of how bacteria use higher order molecular to develop an assay for identification of small molecule structures within RNA transcripts to control gene expres- inhibitors of RNase E function we generated an E.coli sion. Applying biochemical analytical tools and molecular bacterial strain that produces high levels of recombinant dynamics simulation models we will identify two and RNaseE consisting of 400 amino acid polypeptide that three dimensional RNA structures with regulatory role in harbor the enzyme catalytic and RNA substrate binding transcript stability. Furthermore, we will perform a high domains. We obtained 90% pure recombinant protein throughput screen of small molecule libraries to discover and verified its function using E.coli 9S RNA as substrate. compounds that specifically bind to the RNA regulatory In bacterial cells 5S RNA, a structural part of the larger structural elements and alter macromolecular dynam- (50S) ribosomal subunit, is synthesized as native tran- ics. We aim to identify chemical scaffolds that cause script (9S) that is cleaved by RNase E to produce a ma- deregulation of protein dosage control. We will exploit a ture ribosomal RNA. Due to the vital role of ribosomes mechanism of gene regulation that is exclusively found in protein synthesis, inhibition of 5S maturation process in bacteria and lower eukaryotic cells to develop broad- will lead to cell death. We established an assay where in spectrum antimicrobial drugs with minimum toxicity vitro synthesized 9S rRNA was cleaved by the recombi- to the human host. Discovery of small molecules that nant RNase E peptide. Using Northern blot analysis we specifically bind to higher order RNA structural elements demonstrated that ofloxacin, a fluoroquinolone antibi- is the grand challenge of this project. otic, can inhibit 9S rRNA cleavage by RNaseE in vitro with half maximal inhibitory concentration IC50=60μM. This Benefit to National Security Missions result has several key implications: 1) we have found This research directly supports LANL’s mission in national an alternative molecular target for antimicrobial agent security for biothreat reduction. Discovery of a novel known to inhibit DNA gyrase activity; 2) to date, no drug class of antibiotic drugs that can restrict the evolution of has been reported to in vitro inhibit RNaseE activity in drug resistant species directly addresses the “National the μM range; 3) we have identified a small molecule Strategy for Combating Antibiotic-Resistant Bacteria” inhibitor with well-established antimicrobial activity that and is aligned with the LANL focus area in Complex Natu- can serve as a positive control for our high throughput ral Systems and the new strategy for development of ad- drug screening assay. 711

Page 712

To prove that RNase E activity is regulated by RNA ele- that identifies structurally flexible regions in the backbone ments with evolutionary conserved higher order molecular of RNA. We will perform foot-print analysis to identify RNA structures we tested the ability of E.coli-derived enzyme to sequences that directly interact with the RNase E protein. process RNA substrate from genetically unrelated bacteria We will use both well-established (5S rRNA) and newly Burkholderia thailandensis. From our previous studies we discovered in our laboratory RNase E substrates, including have identified the RNA transcript encoding for soluble B.thailandensis SLT transcript. The results from this assay lytic transglycosylase (SLT) protein synthesis as a putative will guide our work on structure probing analysis and will substrate for RNaseE. In the first six months of our project lay the groundwork for future 3D RNA structure analysis we developed both in vivo and in vitro assays to prove the via nuclear magnetic resonance (NMR). Further, we will role of E.coli-derived RNaseE in the catalytic cleavage of apply two analytical assays based on fluorescence anisot- SLT RNA from B.thailandensis (B.th). We showed that high ropy and Fluorescence Resonance Energy Transfer (FRET), levels of full length B.th SLT RNA could accumulate in E.coli to measure differences in the size and structure of the RNA rne-3071 strain, producing temperature sensitive RNase E molecule alone and in the presence of RNase E protein protein, only upon shift to 45°C of bacterial growth. At this with small molecule inhibitor. We will use these analyti- temperature the RNaseE catalytic activity is suppressed cal assays to screen a small molecule library consisting of and we registered that full length SLT RNA accumulation 4,000 compounds in a high throughput mode to discover correlated with 2-fold increase in the percentage of dead novel class antimicrobials that inhibit RNase E binding to cells at 3h post slt gene activation. SLT hydrolyzes pep- RNA substrates. We will perform both in vitro and in vivo tidoglycan (PG) in the bacterial cell wall to facilitate the studies to validate the efficacy of compounds identified as cell division however overproduction of a PG-degrading leads from the high throughput assay (HTS) as drugs with enzyme is known to induce cell lysis. We synthesized B.th antibacterial properties. SLT RNA in vitro and demonstrated that the E.coli-derived RNase E recombinant protein can induce cleavage of full Conclusion length SLT RNA more efficiently than RNA with 300 nt We will (1) elucidate a key mechanism of microbial cell truncation at the 3’ end. Based on molecular modeling we survival driven by RNA instability, (2) complete a high have predicted that B.th SLT RNA has a complex secondary throughput screen of small molecules in a fluorescence- structure consisting of stem-loops and a very pronounced based assay to find compounds that inhibit RNA degrada- Rho-independent transcription terminator-like structure at tion, and will (3) validate the small molecule antibiotic the 3’end. Our biochemical assays indicate that there are activity in live bacterial cells. We will pursue a thus far several RNase E cleavage sites located closer to the 5’ of unexploited strategy for developing of an entirely new SLT RNA, however the 300nt fragment at the 3’ terminus class of antibiotics. Discovery of therapeutics that can re- appeared to be essential for efficient cleavage to occur. We strict the emergence of drug resistant pathogenic bacteria are currently solving the SLT RNA structure via biochemi- will have a high impact on drug development and national cal probing, known as SHAPE. Our results so far support security. a hypothesis that RNase E binding to the RNA substrates is regulated by structural elements located downstream Publications of the cleavage site. Currently all molecular assays used Stubben, C. J., S. N. Micheva-Viteva, Y. Show, S. K. to study the catalytic activity of RNase E employ short Buddenborg , J. M. Dunbar , and E. Hong-Geller single stranded RNA oligonucleotides which we believe . Differential expression of small RNAs from do not represent the in vivo substrates of the enzyme. We Burkholderia thailandensis in response to varying are developing a florescence resonance energy transfer environmental and stress conditions. 2014. BMC (FRET) assay to screen drug library for identification of Genomics. 19 (15:385): doi: 10.1186/1471. compounds that will specifically inhibit RNAse E cleav- age of RNA with secondary structure found in the native substrates. Future Work During the Fiscal Year 2016 we will continue structure probing analysis to determine the nucleic acid architecture that regulates RNA stability through direct interaction with RNase E, an enzyme with endonuclease cleavage activity that plays a key role in bacterial RNA processing and stabil- ity. We will apply SHAPE, a chemical modification analysis 712

Page 713

Science of Signatures Early Career Research Continuing Project Hand-held Laser-Ultrasound Two-Dimensional Scanner Eric B. Flynn 20150673ECR Introduction Materials Storage: Rapid, standoff inspection of material Develop and demonstrate a hand-held laser ultrasound storage containers for corrosion, cracking and the pres- scanning device that provides rapid, convenient, stand- ence of material contents such as fluid build-up. A com- off inspection. The form-factor of this device will be pact scanner could be mounted on robotic inspection similar to that of a hand-held barcode scanner. The systems (aerial or ground) and would enable inspection device will be able scan at a rate of up to 250 scan lines of hard-to-reach places, such as underground storage per second. In this way, the device will generate a visibly and between shielding and encasement layers. continuous horizontal scan line that the operator can manually sweep over the inspection region. Stockpile Stewardship: Provide a unique ability to detect interlayer disbonds and to map the transmission of The proposed hand-held scanner has only become real- acoustic/ultrasonic energy among assembled compo- izable with our recent demonstration of a novel ultra- nents. The latter may serve as a form of fingerprinting sonic inspection technique based on spatial-frequency for identifying when something may have unexpectedly response, referred to as acoustic wavenumber spectros- changed as a result of aging or exceeding environment copy (AWS). The premise of the technology is the mea- specifications. surement and analysis of a system’s spatially-distributed response to a steady, single-tone excitation. Nonproliferation: Ultrasonic fingerprinting (barcoding) of accountable systems and storage containers. In-the-field With AWS, we demonstrated the ability to scan the ul- detection of tampering without the need for tamper trasonic steady state response of a structure at up to five evidence devices. square meters per minute at a distance of three meters with a scan-line spacing of one mm. This is over 100 Manufacturing: Non-disruptive, layer-by-layer quality times faster than conventional ultrasound and translates assurance of additive manufacturing processes (3D print- to a linear speed of approximately 83 meters per second, ing). A sufficiently small and fast scanning system could which means a 30 mm long scan line can be recorded in be built into assembly lines to provide quality control 3.6 milliseconds, making a hand-held device possible. inspection of all manufactured parts (100% QC). AWS relies on the spatial frequencies, or wavenumbers, DoD and Aerospace: Maintenance of military air, ground, of the response. In a bounded medium, the wavenum- and naval assets accounts for the majority of system ber of a wave at a given frequency is a function of the lifecycle cost. New, not well understood, lightweight geometry and material properties of that medium. By composite materials magnify this problem. A rapid quick making direct estimates of local wavenumber on a pixel check capability could drastically improve the combat by pixel basis for a known excitation frequency, AWS readiness and reduce the maintenance costs of systems can make estimates of local properties such as thickness utilizing these advanced materials. or cracking. In AWS, the excitation source is fixed at an arbitrary location, and the location does not affect the Biomedical: Rapid assessment of the elasticity of dis- imaging. This insensitivity to excitation source is also a eased or burned tissue (far-reaching). key result for the proposed project. Benefit to National Security Missions 713

Page 714

Progress state, single tone out-of-plane velocity response measure- This project began in February, 2015. This progress report ments from a range of 1 meter that are comparable to reflects the first four months of the project. commercial off the shelf laser Doppler vibrometer systems. A single tone generation package that is smaller than 10 Develop and demonstrate a first version of the coregis- cm3 in size was designed, built, and tested. This is well tration algorithm. The goal is to demonstrate the ability within the desired specifications of “briefcase-sized”. The to produce a spatially accurate 2D response map from a maximum output is currently 20 Watts, and we do not manual, hand-held scan. A laser Doppler vibrometer will anticipate pushing it to 50 Watts to be an issue. The pack- be used in place of the final system for this demonstration age currently requires a warm-up time of about 20 seconds in the first year. to allow for the frequency of the tone to stabilize. We are developing ways to decrease this time, and modifying the Conclusion signal processing algorithms to compensate. Develop and demonstrate a scanner that is sufficiently compact to be carried by a person and operated by a single Also, we built and tested an initial prototype of the optics hand. The scanner will provide raw steady-state response package. It can measure ultrasonic frequencies up to 100 measurements with more fidelity and at faster speeds that KHz at a range of up to 50 cm without the need for a bulky present commercial off the shelf laser Doppler vibrometry Bragg cell. The prototype currently uses a HeNe laser, and technology. It will seamlessly measure transitions between we are in the process of designing a more compact photo- scanned components in an assembly and automatically diode as a replacement. The focusing optics are being in- map scan lines to physical space. corporated to extend the range to the required one meter. Alternatives to the Bragg cell are being explored to achieve Enabled technologies would include mobile robotic inspec- similar improvements in signal-to-noise ratio. tion platforms, “ultrasonic fingerprint” readers, tamper- indicating scanners, inspection of hard-to-reach compo- The electronics package has been designed and parts are nents, inspection of component assemblies, emergency currently being ordered. Testing will begin in July, 2015. and quick-check inspections, and in-line manufacturing Initial proof of concept work has been performed towards quality control. the co-registration algorithm. We have demonstrated the ability to automatically synchronize to an unknown excita- Publications tion frequency, an important step towards full automated Flynn, Eric B., Anthony J. Haugh, and Sheri B. Lopez. Small co-registration. Full testing and algorithm development will defect detection through local analysis of acoustic begin in June, 2015. spatial wavenumber. 2015. In International Workshop on Structural Health MOnitoring. (Palo Alto, 1-3 Sep. Future Work 2015). , p. 326. Lancaster: DEStech Publications, Inc. Design, build, and test the ultrasonic generation package of Gannon, Adam M., Elizabeth M. Wheeler, Kyle J. Brown, the system. The ultrasonic generation package includes a Eric B. Flynn, and William J. Warren. A high-speed dual- compact ultrasound generator customized for the applica- stage ultrasonic guided wave system for localization tion and a resonant-tuned excitation transducer. The goal and characterization of defects. 2015. In International is to generate 50 watts of ultrasonic power with a device Modal Analysis Conference. (Orlando, 2-5 Feb. 2015). , that can be carried over one shoulder. p. 123. New York: Springer International Publishing. Lee, , C. M. Cho, C. Y. Park, Chung Thanh Truong, H. J. Shin, Design, build, and test the optical breadboard prototype of Jeong, and E. B. Flynn. Spar disbond visualization in the optics package. The optics package includes the laser in-service composite UAV with ultrasonic propagation source, the Doppler system, and the scanning mirrors. imager. 2015. AEROSPACE SCIENCE AND TECHNOLOGY. 45: 180. Design, build, and test the working version of the electron- ics package. The role of the electronics package is to map the raw photo-diode signals into estimates of response amplitude and phase through analog and digital signal processing. The goal with the optics and electronics package in the first fiscal year is to demonstrate the ability to make steady- 714

Page 715

Science of Signatures Postdoctoral Research Final Report Remote Whispering Applying Time Reversal Brian E. Anderson 20150688ECR Abstract communications in highly reverberant environments The purpose of this project was to explore the use of (i.e. racquet ball courts, parking garages) for Morse-code time reversal technologies as a means for communica- communications [4-5] and for imaging purposes [6-7], tion to a targeted individual or location. The idea is to but these studies have not utilized typical room environ- have the privacy of whispering in one’s ear, but to do ments. This research aimed to develop covert extraction this remotely from loudspeakers not located near the techniques of the calibration signal required for success- target. Applications of this work include communicat- ful TR communications and to determine the range of ing with hostages and survivors in rescue operations, applicability and limitations of using TR to communicate communicating imaging and operational conditions in private, audible speech in various environments. deep drilling operations, monitoring storage of spent It was found that private communications are possible nuclear fuel in storage casks without wires, or monitor- in a typical conference room. Speech-like sounds could ing clandestine activities requiring signaling between be heard at all locations in the room but observers of a specific points. This technology provides a solution in live demonstration of the focusing could not understand any application where wires and radio communications the speech unless they were at the target location. It are not possible or not desired. It also may be configured was also found that several types of signal processing to self calibrate on a regular basis to adjust for changing methods could be used to further mask the intelligibility conditions. These communications allow two people to of the speech without destroying the ability to commu- converse with one another in real time, converse in an nicate to the desired focal location. Proof of concept of inaudible frequency range or medium (i.e. using ultra- communicating to a target inside a room with the equip- sonic frequencies and/or sending vibrations through a ment outside the room was also accomplished. structure), or send information for a system to interpret (even allowing remote control of a system using sound). One major focus of this work was to develop covert means of extracting the calibration signal necessary for The time reversal process allows one to focus energy to successful communications in clandestine operations. a specific location in space and to send a clean transmis- One of the methods proposed and studied was to record sion of a selected signal only to that location. In order ambient noise present in the room of interest and record for the time reversal process to work, a calibration signal that noise at the location of both where the transmit- must be obtained. This signal may be obtained experi- ting sources would be placed and at the intended, target mentally using an impulsive sound, a known chirp signal, location. or other known signals. It may also be determined from a numerical model of a known environment in which the Scientific Approach and Accomplishments focusing is desired or from passive listening over time to A set of proof of principle demonstrations of private ambient noise. speech communications was conducted using four loudspeakers, a microphone, and a laser vibrometer Background and Research Objectives setup in a conference room (see Figure 1). To start with, Time Reversal (TR) is a technique that provides spatially successful private communications were demonstrated localized energy focusing and has been shown to provide by placing the loudspeakers inside the room along with secure, optimized underwater acoustic communica- using a microphone for the calibration. A demonstration tion between two locations [1-3]. There have been a of simultaneously focusing two speech signals to two few studies that have explored the use of airborne TR 715

Page 716

different locations was also done (see the first video demo is that the loudspeaker and microphone could not be co- at Ref. [8]). Further, we showed that a masking white noise located. Thus one might expect a delay in the reference could be added to the TR broadcast to mask the speech calibration signal, which is indeed seen when comparing at most locations in the room but allow the target loca- peaks in Figure 3 (arrows added to identify delays). The tion to still understand the speech (see the second video respective delays of each arrival of sound (peaks in the demo at Ref. [8]). Private communications were also made signal) will depend on the path orientation relative to the with the loudspeakers placed outside the room (along a path distance between the loudspeaker and microphone wall with several windows) and by using a microphone for that should be collocated. Qualitatively, the extracted calibration (see the third video demo of focused pulses at calibration signal from ambient noise matches the direct Ref. [8], listen for the pronounced pop at the microphone reference measurement well, though the true quality of location). Then communications were demonstrated using the ambient noise extraction cannot be determined due to loudspeakers and a laser vibrometer in the same room (see the imperfect nature of the direct reference measurement the fourth video demo at Ref. [8]). In this case the laser (source and receiver could not be co-located). was used to remotely obtain the calibration signal to focus sound that one could hear at or near the location where the laser was shining. Figure 2. Photographs of the reverberation chamber used for ambient noise extraction of the calibration signal. (a) Setup used to extract the calibration signal from ambient noise. (b) Setup used to extract the calibrations signal directly between the microphones. Figure 1. Photographs of the conference room setup used for demonstrations of audible private communications. (a) Photo of two of the loudspeakers used and the microphone placed at a seat location. (b) Photo of a small laser spot shining on a white board above a chair with an outline of a hostage drawn on the board (note that the microphone was placed at the same loca- tion in order to record what one would hear at that location. To explore the idea of using ambient noise to extract the calibration signal, measurements were conducted in the facilities at Brigham Young University (BYU) by the project’s Principle Investigator (PI). Two microphones were setup in a large reverberation chamber to record 40 seconds of white noise that was broadcast from a loudspeaker (see Figure 2). The time-synchronized recordings of the noise were cross correlated. The cross correlation response has two halves which, according to ambient noise signal ex- Figure 3. Sample calibration signals extracted from ambient traction theory, should be identical if the technique is per- noise and using a direct measurement setup. The ambient noise forming well. These two halves were summed to provide signal was extracted using the setup pictured in Figure 2(a), the calibration signal. To determine the quality of the ex- while the direct reference signal was extracted using the setup tracted calibration signal from ambient noise, a loudspeak- pictured in Figure 2(b). er was placed at one of the microphone locations and the calibration signal between the two microphone positions The inability to co-locate the source and receiver does not could be obtained as a direct reference calibration signal. only pose a problem with determining the quality of the Figure 3 shows the comparison of these two results. One proposed ambient noise signal extraction, but also for the issue with obtaining the direct reference calibration signal practical implementation of the extracted signal in applica- 716

Page 717

tions. Further work is needed to determine how the source Impact on National Missions and receiver may be co-located or whether a reversible Secure acoustic communications are important to global source/receiver transducer may be used. Once this issue is security (GS) needs and this work has generated discus- resolved, one can explore the quality of the ambient noise sions with GS program managers, as Dept. of Defense signal extraction versus the length of recording time, the customers are looking for technologies to allow them to number of noise sources used, and/or the environmental communicate with acoustic or vibration signals in situ- conditions of the room (i.e. room size, and acoustical prop- ations where wires and radio communications are not erties). While the ambient noise calibration signal extrac- feasible or not desired. tion technique showed promise, it was not conclusive and further work needs to be done. This work provides a significant research advance for en- ergy resource extraction technologies. Our experiments on In FY13, the PI received LDRD Reserve Funding to study vibrational transmissions using time reversal have shown time reversal vibrational communications. The work com- communication rates of at least 20 kbit/s and perhaps pleted provided proof of concept but it was not carried even 1 Mbit/s. Additionally, this research provides a means through to the point where a journal aricle could be pub- to communicate monitoring information through sealed lished. The experiment utilized an 8.10 m long pipe system nuclear reactor vessels and storage casks for safe opera- with 12 junctions and a 3 m portion encased in concrete. tions and reliable storage of spent fuel. A single shaker source was mounted to one end of the pipe system with a triaxial accelerometer on the other References end. In this project, the PI conducted improved measure- 1. Anderson, B. E., M. Griffa, C. Larmat, T. J. Ulrich, et.al. ments and presented the work at an Acoustical Society Time reversal. 2008. Acoustics Today. 4 (1): 5. of America conference and submitted a journal article for peer review in the Journal of the Acoustical Society. Figure 2. Parvulescu, A., and C. S. Clay. Reproducibility of signal 4 shows how individual communication signals may be transmission in the ocean. 1965. Radio and Electronic independently transmitted to any of the 3 channels of the Engineers. 29: 223. accelerometer, allowing at least three channels of inde- 3. Song, , and Badiey. Time reversal acoustic communi- pendent, simultaneous communications between a single cation for multiband transmission. 2012. JOURNAL source and a single receiver. OF THE ACOUSTICAL SOCIETY OF AMERICA. 131 (4): EL283. 4. Candy, J. V., A. W. Meyer, A. J. Poggio, and B. L. Guidry. Time-reversal processing for an acoustic communica- tions experiment in a highly reverberant environment. 2004. JOURNAL OF THE ACOUSTICAL SOCIETY OF AMERICA. 115 (4): 1621. 5. Ribay, G., J. de Rosny, and M. Fink. Time reversal of noise sources in a reverberation room. 2005. JOURNAL OF THE ACOUSTICAL SOCIETY OF AMERICA. 117 (5): 2866. 6. Harker, B. M., and B. E. Anderson. Optimization of the array mirror for time reversal techniques used in a halfspace environment. 2013. Journal of the Acoustical Society of America. 133 (5): EL351. 7. Mimani, , C. J. Doolan, and P. R. Medwell. Multiple line arrays for the characterization of aeroacoustic sources using a time-reversal method. 2013. JOURNAL OF THE ACOUSTICAL SOCIETY OF AMERICA. 134 (4): EL327. Figure 4. Sample signal transmissions from one source to one 3-axis receiver. Each subplot represents the signal transmitted to 8. Anderson, B. E.. Video demonstrations of time reversal each axis of the 3-axis receiver. The transmitted signals are the acoustic communications in a conference room. 2013. ASCII representations of the letters “t”, “l”, and “m”. LAUR-13-27841. 717

Page 718

Publications Anderson, B. E., T. J. Ulrich, P. Le Bas, and J. A. Ten Cate. Three dimensional time reversal communications in elastic media. Journal of the Acoustical Society of America. 718

Page 719

Science of Signatures Exploratory Research Continuing Project Time Resolved Phonon Spectroscopy for Cryogenic Bolometer Readout John J. Goett III 20140200ER Introduction veloping the capability to monitor materials for all decay As the focus of experimental research moves to rare products in a single setup will streamline procedures for events beyond the scope of the standard model, it has researchers working in environmental science and the become increasingly apparent that new technologies non-proliferation/ treaty verification fields. Perhaps the with improved background discrimination and recon- greatest relevance to non-proliferation programs will struction capabilities are needed to advance the field. lie in the separation of alpha peaks without resort to Here we outline the first steps of an aggressive push destructive assay. The detectors might also have higher into detector development that seeks to fill the gap that efficiency but maintain the excellent low-energy x-ray has left particle instrumentation largely divorced from resolution of smaller bolometers. advances in materials science and chemistry. Though bolometers have been successfully operated as sensitive Progress thermometers, the physical phenomena underlying their In year two we have made progress in each of our effectiveness is heat propagation via phonon interac- planned goals. We now have a laboratory space to sup- tions within the bulk of the absorbing crystal. The details port cryogenic work with an operating cryostat, along of phonon propagation have not been exploited in with the necessary electronics and control systems in bolometric techniques. This proposal details the prereq- place to take bolometric measurements. In collaboration uisites for a program in phonon spectroscopy in radia- with our colleagues at NIST, we have determined the for- tion detectors with a fast readout device. Demonstrating mula for AlMn transition sensors, and have a clear path our ability to measure these underlying processes will forward for growth on the crystal bulk. We believe we open the doors to a precision era in low-energy nuclear are on track to begin crystal measurements in late 2015 physics. In particular, as DOE-NP determines the best ap- and early 2016. proach for a large-scale, double-beta decay experiment, the development of this technology will increase the We took delivery of the SQUID electronics and control LANL portfolio of capabilities applicable to that program. hardware in the middle of the first quarter. Our initial Benefit to National Security Missions effort was focused on operating the warm electronics in a ‘dummy-mode’ where a transistor array was set up to A high resolution calorimeter with good discriminating simulate a SQUID circuit. In this manner, we developed a ability and high throughput has applications that extend slow control application for adjusting the voltage or cur- beyond a fundamental physics program. By constructing rent bias without risk of damaging the sensitive devices. a well counter from high purity macro-bolometers with phonon readout one could perform non-destructive as- In tandem with the software development, we resumed say of materials extending beyond gamma spectroscopy. commissioning and refurbishment of the dilution refrig- At present, a comprehensive assay of a material requires erator unit. We ran extensive vacuum checks and elimi- choosing between time consuming counting with a nated a number of very small leaks discovered early in variety of techniques ranging from high purity germa- testing. In addition we designed and constructed a large nium detectors and barrier silicon detectors which force anti-vibration stand suitable for small bolometric mea- a compromise between either small active area/long surements. We produced an IWD for the recharge of the count times or high background; or destructive methods compressor and general cryostat operation and dem- such as mass spectroscopy, neutron activation analysis onstrated a successful cool-down to 4K. Following the or more comprehensive radio-chemical techniques. De- 719

Page 720

cool down, we designed, fabricated and installed custom cable feedthroughs to route signals from the cold stage to the warm electronics. Following the installation we were again successful in cooling to 4K and demonstrated good temperature stability for an 11 day run. This alleviates concerns that the cable mass we were routing to support the multi-sensor measurement would introduce an unac- ceptably large heat load on the crystal. In collaboration with our colleagues at NIST, we character- ized a first run of AlMn transition edge sensors in a variety of geometries grown on a conventional silicon substrate. The initial tests indicated a slow superconducting transition at a higher than expected temperature. With discussion and a variety of tests, we concluded the difficulty was due to a thick shadow mask, which introduced thickness varia- tions( and thus current irregularities) in the AlMn along the edge of the mask. This was simply solved by introducing a new shadow mask, which eliminated the variations. Initial measurements were completed in Boulder, and confirma- tory measurements here at LANL are imminent. The final two elements needed to operate the TESs and SQUIDS at base temperature here at LANL are supercon- ducting shielding for the SQUIDS and a nitrogen cold trap to purify the helium mash before circulation. Designs for both these parts have been completed, and at the time of this report, are out for bid on fabrication. Future Work This work is planned to take place over three years. During years 1 and 2 we will prepare a facility where bolometric measurements can be made, and with our collaborators develop the necessary expertise to grow AlMn transition edge sensors directly onto a crystal bulk. In the 3rd and 4th quarter we will characterize initial runs of AlMn sensors and exercise our full DAQ and slow control chain. In the third and final year of our program, funds will be al- located to a production run of AlMn transition sensors and a measurement campaign with multiple sensors deposited on the crystal bulk. These data will serve the dual purposes of demonstrating the utility of the focusing phenomenon in background discrimination and validating the simulation framework developed in years 1 and 2. Conclusion During the 3-year span of this proposed effort, we plan to determine the effectiveness of using multiple sensor readout of a bolometer for energy deposit localization and assess its utility for background reduction. 720

Page 721

Science of Signatures Exploratory Research Continuing Project Measuring Winds in the Stratosphere Using Passive Acoustic Sensors Omar E. Marcillo 20140237ER Introduction using low frequency sound from ocean storms. If suc- The fundamental goal of this project is to test the hy- cessful, we could measure the winds in the stratosphere pothesis that measurements of low-frequency sound anywhere over land at low cost with implications for a from ocean surface wave collisions (microbaroms) number of existing programs at LANL. The joint DOE/NSF recorded by multiple ground-based acoustic sensors Community Earth System Model (CESM) is one of three at distances of several thousands of kilometers can be US IPCC-class climate models for which LANL personnel used to accurately measure winds in the stratosphere. in T-3 and CCS-2 currently play an active research role. In a pilot study we have already demonstrated that our The CESM model includes an atmosphere configuration methodology can estimate winds in the stratosphere that extends to the thermosphere. Of particular interest using transient sources. Following these results we have for the stratosphere dynamics are effects of the ozone identified a fundamental opportunity: to apply this layer, which plays a key role in radiative interactions as inversion methodology to the continuous and ubiquitous well as gravity wave drag (a sub-grid scale phenomena acoustic background noise would allow us to measure that plays an important role in atmospheric dynamics). the evolution of stratospheric winds everywhere without Furthermore, the availability of direct wind measure- the need of an active source. ments in the tropics may be particularly important, due to the known inaccuracies of the thermal wind equation Three fundamental aspects are required to migrate from for the region. Our results also have implications for a transient to a continuous source for the inversion: NASA programs and warfighter support and have po- • infrasound noise from microbaroms is smaller in tential implication for improving GPS accuracy through amplitude and peaks at lower frequencies than improved correction for stratospheric effects. Improved signals from discrete explosive sources; we need to specifications for atmospheric winds also have an im- refine our data analysis methodology, portant implication for Global Security: because strato- spheric winds influence the propagation of infrasound • we need to validate our inversion results using an signals, our results can improve infrasonic location and independent technique, and yield estimation for nuclear tests as well as improve • we need to formally quantify uncertainty of our predicted dispersion of radioactive aerosols, which can measurements and predicted wind values. This proj- reach the stratosphere. ect will develop and validate a stochastic method based on inverting low frequency acoustic signals from microbaroms, recorded on multiple ground Progress sensors, to continuously track wind in the strato- The last 12 months of this project have been very suc- sphere. We will validate our technique by co-locating cessful as we finalized the deployment of the sensor a specially-designed infrasound network with the arrays for the NORwegian StratospherE (NORSE) ex- only LIDAR system capable of reliably recovering periment in northern Norway during April, 2015. This wind and temperature from the stratosphere and successful deployment involved great logistical efforts mesosphere, which is located in Northern Norway. from Los Alamos National Laboratory and the Alomar Observatory in Norway to install and maintain four infra- sound arrays for 12 months in a very harsh environment. Benefit to National Security Missions This deployment and data acquisition was the main task The anticipated result from this project is a robust, vali- for year 2 of the project. We have been receiving data dated method for measuring winds in the stratosphere 721

Page 722

from the network every three months and performed data network. By April 2016, we will complete the collection of preparation and quality control. Using data obtained dur- one year of data from our Norway deployment. We will ing the initial phase of the Norway deployment (2 weeks write a report on the quality of the data and the sea- in November, 2014), we have written and submitted a sonal variability of the characteristics of the microbaroms paper to the EOS Transactions AGU Journal focusing on the signals. We also plan to complete a modeling study of deployment of the network and its potential for atmo- the propagation of infrasound over long distances from spheric studies: “The NORwegian StratospherE (NORSE) microbaroms sources, using both geometric and full-wave experiment: Turning Low-Frequency Coherent Acoustic methods, that was begun in year one. Noise into Signal” by Arrowsmith, Marcillo, et al. This article has helped us publicize our efforts so other institu- The main goal for year three is to apply our inversion tions (e.g.: NORSAR, the Norwegian seismic center) are methodology to estimate winds in the stratosphere and aware of our network design and preliminary results. Also, validate the results against the measurements from the LI- using the data from the initial phase of the Norway deploy- DAR facility. We will first apply a refined time-series analy- ment we were able to identify and study acoustic signals sis on each array and measure the inter-array coherency of generated by a rocket launched in close proximity to one microbaroms for multiple propagation distances. With the of our network nodes. We studied the related infrasound, coherency results, we will apply our inversion methodol- modeled its propagation, and submitted a scientific article ogy to estimate stratospheric winds. These estimations will that is currently under review at the Journal of Acoustical be validated using the direct measurements of wind from Society of America Express Letters. The article highlights the ALOMAR LIDAR facility. We expect at least one journal the potential and implications of obtaining information paper on the data analysis and the inversion. Also, North- about rockets (e.g.: trajectory, number of stages and type ern Norway features a variety of continuous (aurora) and of engine) using acoustic observations: “Analysis and discrete (rockets launches and mining explosions) acoustic modeling of infrasound from a four-stage rocket launch” by sources that we expect to record with our network. We Blom, Arrowmisth, and Marcillo. We have also refined the anticipate having at least one additional publication on the inversion framework in order to apply the method to both identification and characterization of these other sources. transient and continuous infrasound. An article discussing the extended framework and a theoretical evaluation of Conclusion its stability is currently being revised for publication in the The main expected result is a validated and operational Geophysical Journal International: “A statistical framework ground-based technique that is capable of providing for monitoring variations in the stratospheric winds using measurements of wind in the stratosphere from measure- infrasonic signals” by Blom, Marcillo, and Arrowsmith. ments of the continuous and ubiquitous acoustic back- ground noise. This technique can be exported anywhere We also completed the analysis of the data of our first net- on Earth and does not require a known source, enabling us work prototype deployed in Albuquerque, NM. With this to track the wind continuously. We envisage this research dataset, we were able to identify and model infrasound to provide material for several peer-reviewed publications. produced by nearby wind farms in NM. We found that As our main research result will be truly innovative and these infrasound signals can propagate several 10s of kilo- transformational, we expect our research to be suitable meters in the lower atmosphere under certain conditions. for a high-profile broad-audience journal, i.e., Nature or We wrote a scientific article for the Journal of Geophysical Science. Research Atmospheres on the analysis of this source and its potential for applying the same methodology we are Publications exploring to extract winds in the stratosphere to retrieve Arrowsmith, M., S. Arrowsmith , and O. Marcillo . The boundary layer and tropospheric winds and temperature: NORwegian StratospherE (NORSE) experiment: Turning “On infrasound generated by wind farms and its propaga- Low-Frequency Coherent Acoustic Noise into Signal. To tion in low-altitude tropospheric waveguides” by Marcillo, appear in EOS Transactions AGU. Arrowsmith, et al. This result has very important implica- tions for near-surface atmospheric interactions and bound- Blom, P., S. Arrowsmith, and O. Marcillo. Analysis and mod- ary layer physics. This paper was accepted for publication eling of infrasound from a four-stage rocket launch. Journal of acoustical society of America. and will be published in the next few months. Marcillo, O., S. Arrowsmith, P. Blom, and K. Jones . On in- Future Work frasound generated by wind farms and its propagation In year three, we will complete the research for this project in low-altitude tropospheric waveguides. 2015. Journal by finishing the data collection and analysis of our Norway of Geophysical Research: Atmospheres. : 1. 722

Page 723

Science of Signatures Exploratory Research Continuing Project Matter Wave Circuits Changhyun Ryu 20140362ER Introduction and quantum information processing are of interest to The goal of this project is to create the de Broglie wave external sponsors in the intelligence and defense com- analog of an integrated optical circuit, a device that munities. might be called a matter wave circuit. In this technol- ogy, coherent atomic matter waves from an integrated Progress source are guided and manipulated in confining poten- During the first year of the project, we realized vari- tials in much the same way that laser light propagation ous circuit elements for matter waves such as a straight is controlled in a wave-guide photonic circuit. The push waveguide, a circular bend, a closed stadium waveguide, to develop matter wave circuits is motivated by several and a Y-junction. In addition, a coherent splitting of mat- important differences between matter waves and light ter waves through a Y-junction was demonstrated for waves: atomic velocities can be controlled over a wide the first time. To continue the success of the first year, range; atoms can be brought to rest and trapped; atoms we need to accomplish two tasks during the second year. feel gravity and electromagnetic fields; atoms can inter- First, a new Bose-Einstein condensate (BEC) setup for act strongly with each other; and atoms can be imaged 39K needs to be built because many applications in mat- with high efficiency. These characteristics enable matter ter wave circuits require varying interaction strengths, wave circuit applications to sensing (particularly inter- which can be achieved thorough magnetic field tuning ferometry) as well as to quantum information process- with 39K. Second, simulation of matter waves in various ing, quantum atom optics, and emulation of transport waveguides is necessary to design complicated matter problems in condensed matter systems. wave circuits and to find the right parameters for the experiment. During the last year, we have been working This promise motivates our proposal to realize the first on these two main focuses. matter wave circuit. We will confine de Broglie waves from a Bose-Einstein condensate (BEC) in waveguides To achieve a 39K BEC, we need several key elements: and manipulate them with optical elements, all created a vacuum chamber and pump capable of reaching the using LANL’s unique “Painted Potential” capability. This pressure of 10-11 torr, a laser system for cooling 39K technique creates arbitrary and dynamic time-averaged atoms, and various electro magnets with magnetic field optical dipole potentials compatible with the very low up to 500G. After careful design, the vacuum chamber energy of a BEC. was put together and sources for various alkali metal atoms were installed. After a few weeks of baking of the Benefit to National Security Missions chamber, the vacuum inside is now 4×10-11 torr, low Atoms lie at the heart of some of the best technologies enough for the experiment. The lasers for the cooling of for measuring time, detecting rotations, and processing 39K atoms were purchased and tested, including cou- information on quantum computers. The atom optical pling of various beams into single mode optical fibers technology that we will develop during this project sup- to minimize fluctuation of the beam position. Several ports Laboratory missions in Sensing, because the atom electro magnets were made and the magnetic field was circuit technology is relevant to rotation sensing, and measured to be in agreement with the calculation. in Information Science and Technology, because mat- ter wave technology in general is relevant to quantum After successful completion of these steps, we started information processing. These applications in sensing tests to achieve a 39K BEC. Because of the very low 723

Page 724

pressure needed for a BEC production, the source chamber strengths for various matter wave circuit experiments. and the BEC chamber have to be separated to ensure large Simulation of various waveguides was done to design differential pressure. In the source chamber we achieved a future experiments and find the right parameters for the magneto-optical trap (MOT) of 39K atoms, with the num- experiment. ber of atoms as high as 3×109. From this MOT, atoms were pushed to the BEC chamber with a push beam. The MOT During the final year of the project, we plan to focus on at the BEC chamber was also achieved. Currently both studying physics of matter wave cavities. Here two mirrors MOTs and transfer process are being optimized. We expect for matter waves will be placed to form a cavity for atoms. this process will be done within a month and the BEC of Although there is some theoretical work, no experimental 39K atoms should be possible within the next 3 months. demonstration of matter wave cavity has been done. To study this, it is essential to be able to control interaction Another major focus was the simulation of matter waves in strength. From the simulation study, we found that the various waveguides to determine experiment parameters coupling efficiency gets much worse with interaction. To for 39K atoms. One of the most interesting circuit elements increase the coupling of atoms into a cavity, it is necessary is a cavity for matter waves. Although theoretical study to have zero interaction which is possible with 39K atoms was done before, there has been no experimental demon- through magnetic field tuning. Another interesting regime stration of a matter wave cavity. Since this is unexplored to study is the large interaction limit with a cavity. Here it is before, it is necessary to study the dynamics of matter predicted that atoms leaving a cavity will have non clas- waves with a cavity to understand the basic physics and to sical distribution and we can reach this regime with the find the correct parameters for the experiment. The first same magnetic tuning technique. With these studies, we problem we studied was the coupling of matter waves into will demonstrate the experimental realization of matter a cavity. Here we found that without interaction, a high wave cavities for the first time and this will open the new coupling efficiency is possible. Also by varying the length of research direction of matter wave cavity physics. the cavity, the variation in coupling which is clear signa- ture of a cavity dynamics identical to optical cavity was Conclusion shown. Another simulation study we did was the study of The overall goal is to demonstrate for the first time analogs shortcuts to adiabaticity. Here the question is how fast we of optical fibers and integrated laser sources with cold at- can change a trap for atoms without creating excitations oms. Specific technical goals include: (1) launching coher- compared to the adiabatic process. With this technique, it ent matter waves into a single mode of a non-trivial wave- is possible to move and expand matter waves with much guide geometry at controllable speed, (2) demonstrating faster speed and the simulation was done to find the cor- the matter wave analog of a distributed Bragg reflector, rect parameters for that. It was shown that with our set up and using two such optics to create a matter wave resona- various processes including expansion and movement of a tor, (3) demonstrating a “Y” beamsplitter by adiabatically BEC can be sped up. deforming a single waveguide into two guides. Success will deliver new levels of control over atoms to important The construction of a 39K BEC set up and simulation of a applications in sensing and information processing. matter wave cavity and shortcuts to adiabaticity will be foundation for the work which will be done in the third Publications and final year of the project. During the third year, we Campo, A. del, M. G. Boshier, and A. Saxena. Bent wave- will focus on matter wave cavities. A 39K BEC with vary- guides for matter-waves: supersymmetric potentials ing interaction strength will be used to study matter wave and reflectionless geometries. 2014. Scientific Reports. cavity dynamics. The demonstration of matter wave cavity 4: 5274. will be a truly exciting development and our work will start a completely new research direction. Ryu, C., and M. G. Boshier. Integrated coherent matter wave circuits. 2015. New Journal of Physics. 17 (9): 092002. Future Work In the first year of the project, we realized several basic matter wave circuit elements such as a straight guide, a circular bend, a stadium waveguide, and a Y-junction. Another important achievement is the coherent splitting of matter waves through a Y-junction. In the second year, we built a Bose-Einstein Condensate (BEC) set up for 39K atoms. This will make it possible to tune the interaction 724

Page 725

Science of Signatures Exploratory Research Continuing Project Chemical Shift Signatures of Nuclear Material: 235U and 239Pu NMR Spectroscopy Michael T. Janicke 20140396ER Introduction Benefit to National Security Missions The crux of this research is the development of low fre- Plutonium has one of the most complex chemistries of quency Dynamic Nuclear Polarization Enhanced Nuclear any element. The shear diversity and unique chemical Magnetic Resonance spectroscopy (DNP-NMR, imagine properties that Pu exhibits are ideally suited for Nuclear a MRI instrument for chemistry studies with 1000 times Magnetic Resonance (NMR) and for this reason this increase in signal to noise). DNP-NMR spectroscopy will project aims to develop a new chemical signature, the be a tool for detection of 235U and 239Pu chemical shift 239Pu NMR chemical shift. Chemical shift is the change signatures, monitoring the origin, fate and transport of in NMR signal frequency from a known standard set at 0 these contaminants, and the bioavailability of such spe- ppm arising from local magnetic fields produced by the cies in the environments. Much of this project will focus electron distribution. It is a direct measure of the bond- on 239Pu but work will be easily extended to the more ing environment around specific nuclei and gives rise to challenging 235U in relation to using NMR as the most the sensitivity of NMR chemical shift spectroscopy to steadfast method for differentiating and quantifying chemical species. Analytical methods such as elemen- molecular species. tal analysis or detection of fission products (neutron, gamma, or beta) do not reveal the chemical species or This proposed research project will demonstrate that its origin. Complex samples with multiple species cannot we can increase the sensitivity without increasing the be analyzed using traditional methods, as they will inevi- sample size or experiment time and get the equivalent tably have overlapping results that require precise use of sensitivity to a commercial multi-million dollar instru- model compounds and accurate computational methods ment for a fraction of the cost. We will accomplish this for deconvolution. NMR presents us with the most accu- by using DNP-NMR. At LANL we have demonstrated use rate and quantitative method where differences in spe- of DNP at very low magnetic fields (10 mT) without cool- cies will result in distinct chemical shifts. But due to the ing the samples to liquid helium temperatures via the complexity of 239Pu NMR and plutonium chemistry, the Overhauser effect. In this approach, electrons spin trans- NMR signature of 239Pu was only published in May of fer to nuclei is enhanced by microwave irradiation. Using last year. This same basic technique can be expanded to this method we achieved ~ 100-fold polarization en- include additional chemical, nuclear, and environmental hancement, followed by NMR signal readout at very low threats that have chemical signatures and NMR active el- frequencies. A published system also based on relatively ements including 13C, 15N, 19F, and 31P and numerous low magnetic field and without cooling the samples to heavy metals. Thus this project supports our national liquid helium temperatures has been demonstrated security mission from nuclear weapons to threat reduc- for the relatively easy 1H isotope for roughly $100,000. tion and environmental monitoring/cleanup. However this system cannot measure low frequency nu- clei such as 235U and 239Pu. With this low-cost instru- Progress ment, the research group is achieving 1H enhancements At the end of the first year, the ultra low field (ULF) of greater than ~ 100, as well. Combining the DNP with dynamic nuclear polarization (DNP) nuclear magnetic our demonstrated expertise at ultra-low field we antici- resonance (NMR) platform had been commissioned pate for 235U and 239Pu the signal improvement will be with promising initial measurements of a water signal >1000-fold compared to routine NMR techniques. (task 1). Utilizing an air cored electromagnet, a DNP enhancement factor (that is, signal increase) of 20 was 725

Page 726

seen with water, leading to signal to noise ratio (SNR) of species such as 75As or 199Hg. The new probe boosts 200, and a spectral resolution of 70 Hz. While promising signal by improving the filling factor of the NMR coil sub- for a prototype, higher SNR and better spectral resolution stantially, as well as the efficiency of the DNP process by are needed to perform spectroscopic measurements on optimizing the resonator design. Measurements with low solutes. Improvements to the system were undertaken sensitivity nuclei will commence following the completion to achieve the year two goals. A conventional high field of the upgrade. This work fulfills both task 1– acquiring measurement was attempted on plutonium (IV) chloride signals from low sensitivity nuclei with adequate resolution sample, but no 239Pu NMR signal was observed in year 1. for signature development, and task 2 – Improvement of Continuing efforts on Pu (IV) and Pu (III) samples have also the DNP process through refinement of the probe hard- not been successful. ware. The two tasks for year 2 are as follows: Using the expertise we have gained in the optimization of the ULF system, we will begin to implement DNP at

1. Validate a high field 239Pu NMR spectrum in the ULF intermediate field with an iron-cored electromagnet and system, developing chemical signatures X-band (~10 GHz) cavity resonator for DNP. We possess
2. Develop the DNP protocols (polarizing agent optimiza- the magnet and electronics necessary to perform NMR in tion, probe refinements, etc) this regime. In pursuit of these goals, we improved the performance Future Work of the ULF DNP NMR system substantially. We increased In the first year we will have completed our first task, the spectral resolution by a factor of 70 by upgrading the construction of our platform for detecting Nuclear Magnet electromagnet and through the addition of shimming coils Resonance (NMR) signature of chemical species (non-ra- to produced higher magnetic field uniformity, as well by dioactive surrogates) with similar properties to 239Pu. We reducing the size of the sample. In addition, we were able are on track for this. to increase the SNR from 200 to 500 for a water sample. In the second year of the project we will focus on Tasks 2 These improvements are critical for the accomplishment of and 3 as defined in the proposal. Task 3 will extend into the the year 2 tasks. Although measurement of a 239Pu signal third year of the project. seems less likely in the absence of a result at high field, we are in the midst of measuring the NMR signal from a Task 2: Test and optimize our NMR instrument for 239Pu number of low sensitivity nuclei such as 199Hg and 125Te using a known PuO2 sample. We will validate and extend to fulfill task 2. This will pave the way for measurements results with conventional high-field NMR (available at LANL of a wide variety of difficult, low sensitivity nuclei that are or PNNL) and our new system (i.e. methods for handling generally not accessible to a fieldable system that is avail- short T1; methods for increasing temperature from 4K able off the shelf. to room temperature). The advantage of this approach is significant reduction in magnet cost, complexity, and ex- As an intermediate step to task 1, we were able to explore pense. Further, the compatibility with microfluidics means if the current ULF DNP NMR instrument could be used to sample preparation and delivery can be done “on board” measure J coupled spectra as a possible novel signature and without significant handling and sample minimization. for chemical threat agents. We have measured the ULF High field NMR validation months: 1-6; Signature exten- DNP NMR spectrum of isopropyl methylphosphonic acid sion/ measurements in new system Months: 12-24. (IMPA) in deuterated water, resolving the J couplings, which are field independent couplings that split the NMR Task 3: Develop protocol for dynamic nuclear polarization spectrum according to the number of bonds between enhanced NMR (DNP-NMR) for sensitivity increase. This the nuclei. IMPA is of particular interest because it is a task includes instrument development and identification of hydrolysis product of sarin gas; the potential to field this free radical species for polarization transfer (ie. 2,2,6,6-Tet- system makes such measurements especially interesting in ramethylpiperidin-1-yl)oxyl referred to as TEMPO, a water the context of rapid diagnostics after a chemical weapons soluble, stable free radical benchmarked in many DNP- event. Discussions are ongoing with programmatic spon- NMR studies) Months: 18-30. sors to build a program around this interesting method. Conclusion We are currently in the midst of a probe upgrade that will The lower bounds of the frequency range for metals might boost SNR enough to measure signals from dissolved metal 726

Page 727

not be reached; however, there are other organic com- pounds containing 13C, 15N, 19F, and 31P that will be made accessible for DNP-NMR analysis that do not have these complications and have equal if not greater need for detection and forensics for chemical and biological agents. In other words, if we are not successful in analyzing metal speciation, the instrument will already have succeeded in measuring important organic compounds. Additionally, if we meet the challenges above, we will have an R&D 100 candidate, commercialization and licensing possibilities, and a portable instrument. 727

Page 728

Science of Signatures Exploratory Research Continuing Project Solid-State Gamma-Ray Detectors Based on Quantum Dots Jeffrey M. Pietryga 20140406ER Introduction realizing performance on a par with current, much more The threat posed by the proliferation of illicit Special expensive and less scalable technologies such as single- Nuclear Materials (SNM) is especially grave as they carry crystal gamma scintillators. the potential for disaster in dangerously small packages. Hidden SNM can be found and identified by detection Benefit to National Security Missions of the characteristic gamma radiation it emits, but this The guiding goal we have chosen for this work is to type of screening becomes quite complicated in real make gamma-radiation detectors that are comparable scenarios. Detectors at high-traffic border crossings, for to gamma scintillators in energy-resolving performance, example, must be able to distinguish SNM from a range but much cheaper and easy to scale to portal-relevant of benign gamma-ray sources that are a part of ordinary detector sizes. The tie to nonproliferation-involved agen- commerce by determining the energy of gamma-rays cies, as well as to any other agencies (Defense or Intel- with high precision, while sampling a large area at a very ligence) interested in detecting the presence of nuclear high rate of throughput. Unfortunately, current tech- materials, is clear. The flexibility of our new detector nologies cannot meet this need. class would easily make them adaptable to needs in nuclear reactor/facility monitoring, and radioactivity-re- We propose to develop a new class of solid-state gam- lated environmental remediation efforts. It is quite likely ma-ray detectors based on cheap, solution-synthesized that specific needs of space science (gamma spectrosco- semiconductor nanocrystals that will offer a new stan- py) and satellite surveillance could ultimately be met by dard in combined performance, low-cost and adaptabil- the proposed class of detectors. Such sensors could also ity. be of interest to materials scientists for use in various types of radiography, and to the medical field as cheap, In this project, we will fabricate gamma-sensitive diodes superior replacements for current detectors used in based on semiconductor nanocrystals synthesized by x-ray, nuclear, and positron emission tomography (PET) low-cost, scalable chemical methods. Although these imaging techniques. Finally, because this research will nanometer-sized bits of semiconductor materials exhibit advance the understanding of achieving high conductiv- extremely useful size-tunable electronic and optical ity in nanocrystal films, it will also likely have incidental properties, for most of their history nanocrystals had benefits to current efforts in high-efficiency solar cells been considered ill-suited to applications in electronic and solid-state lighting based on nanocrystals. devices (like diodes) because of the difficulty in achiev- ing high electrical conductivity when they are assembled Progress into thin films. Recent developments have cast that Our three-year goal is to create a new class of solid-state perception aside, and there are now numerous ways to gamma-ray detectors based on colloidal semiconduc- make device-ready conductive films of nanocrystals that tor nanocrystals, or quantum dots (QDs). The principal exploit their unique properties without sacrificing their activity of this project is leveraging and expanding upon advantages in cost and scalability. We propose to extrap- the past decade of advances in the fabrication of highly olate this recent progress to an unprecedented degree conductive films based on colloidal QDs. Successful fabri- to enable a new class of gamma-ray detectors. We will cation methods not only achieve continuous, crack- and then probe their response to high-energy radiation, and pin-hole-free films, but also result in strong coupling optimize both the nanocrystals and the architecture to of QDs by removal of native organic ligands. The ligand maximize energy resolution, with the ultimate goal of 728

Page 729

shell surrounding the QDs are a byproduct of their colloi- organic ligands, and are smooth and crack/pin-hole free. dal synthesis allows for solution-processibility, but serves Evaluations of these films for conductivity and radiation re- as a barrier to the QD-QD coupling necessary for charge sponse are already underway. We also believe this achieve- transport. As described in previous reports, straightfor- ment will be received very enthusiastically by the larger QD ward application of known film deposition methods, which device community, and we will be submitting a manuscript are often iterative (i.e., require a sequence of thin deposi- on its development and use very soon. tions to build up to a final desired thickness), may not be appropriate for gamma detecting devices, which require Over the past year, this project also partially supported film thicknesses orders of magnitude higher those used for, the development of a new technique for analyzing car- e.g., light-emitting diodes or solar photovoltaics. rier transport in QD films, which we refer to as ultrafast transient photocurrent. In essence, it tracks the decay of Initial studies in the first year bore this out. Although conductivity of a thin film after excitation by a laser pulse response to X-ray excitation was observed in a simple with ~50 ps resolution. A manuscript describing the tech- photoconductive device based on a film ~3-5 microns nique in detail, and its use to solar relevant phenomena in thick, there were serious limitations to the “layer-by-layer” QD thin films, has been submitted for publication. Support method used. The most important were that film produc- from this project has also allowed us to revive our dormant tion was extremely tedious and prone to error (requiring capability in atomic-force microscopy, which allows us many 10s of sequential dip-coat or spin-coat steps to be to evaluate film morphology and thickness in a relatively performed manually), actual carrier mobility was less than straightforward, non-destructive manner. Both methods that seen in thinner films, and physical stability (specifically are critical for relatively rapid analysis during our continu- resistance to flaking or crumbling during manipulation) ing film optimization efforts. prevented pressing to even thicker films. In addition, be- cause this approach accomplishes QD layer deposition and Future Work ligand-removal in separate steps, it is incompatible with The next year will see us pushing the envelope of achiev- advanced methods like electrophoretic deposition. able film thicknesses using our new one-step deposition approach, including in conjunction with electrophoretic Accordingly, this most recent year saw a concerted and deposition. This single method has been determined to be ultimately successful effort at developing a single-step the most promising and flexible among all potential alter- method for replacing QD ligands without sacrificing solu- natives, so this will be the focus of the remainder of this tion dispersibility, which had been accomplished in a project. A key part of studies going forward will be careful handful of examples for other materials but never for lead evaluation of the effect of ligand-removal on the stoichi- chalcogenide QDs. The key breakthrough was to combine ometry of the QDs, that is, the atomic ratio of cations new methods for synthesizing lead chalcogenide QDs with (lead) and anions (chalcogenides) in the QDs after deposi- greater amenability toward ligand exchange, and the iden- tion. Nominally, the ratio should be 1, but because of the tification of the perfect solvent for the process. The solvent flexibility of a material that is dominated by surface atoms, needs to be polar enough to dissolve inorganic compounds it can vary substantially from this ideal. This is known to that can displace ligands (e.g., potassium iodide) and impact not only carrier mobility, but also carrier density in to keep the QDs dispersed after exchange, coordinative QD films, both of which need to be optimized for use in a enough to promote exchange chemistry without heating solid-state device. This will make use of our team-owned and yet aprotic and non-oxidative toward the frustratingly energy-dispersive x-ray spectroscopy (semi-quantitative) sensitive surfaces of lead chalcogenide QDs, and posses- and inductively-coupled plasma emission spectroscopy sive of a boiling point low enough to allow dip- or spin- (quantitative) instruments. By choosing different inorganic coating of QD films directly from solution. We found that exchange compounds, it should be possible to control exposure of lead chalcogenide QDs synthesized using lead the stoichiometry. The precise effect of stoichiometry on chloride to a 2,6-dimethylpyridine solution of any of a wide carrier transport properties will be examined by conven- range of ionic compounds results in quantitative ligand tional field-effect transistor device studies, as well as by removal and transfer of the QDs to the polar phase, often using our new ultrafast transient photocurrent technique. within seconds. Moreover, unlike any previous method, the Soon after, their response to alpha, beta, X- and gamma QDs in polar solutions retain their efficient fluorescence, rays will be evaluated. As described in previous reports, demonstrating that the treatment does not create surface this seeming departure from our motivation of gamma-ray defects that will ultimately impact charge transport in sensing (which we maintain) is actually a good way to al- films. Films deposited by spin-coating from these solutions low studies to take place before achieving optimized films show no detectable traces of the presence of the original thick enough to fully stop gamma rays (on the millimeter 729

Page 730

scale). We believe that this approach will best enable us to advance the development of practical QD-based gamma- sensors, to the point of observing and quantifying energy resolution. Conclusion In order to create the first gamma-sensing diodes based on nanocrystals, we will need to make very thick conductive nanocrystal films in order to capture radiation effectively. Then, we will extensively characterize how simple devices based on these films respond to radiation of a range of en- ergies, which would be a seminal accomplishment for this completely new category of sensing technology in itself. Finally, we will optimize the nanocrystals and the device structure with the goal of achieving performance compa- rable to common gamma scintillators, and potentially to other types of radiation detectors used in medical imaging or materials science. 730

Page 731

Science of Signatures Exploratory Research Continuing Project Signatures of Reactor Operations from Plutonium Production samples (U) Anna C. Hayes-Sterbenz 20140433ER Introduction Benefit to National Security Missions This is a project is to quantify a novel scheme for us- Nuclear non-proliferation represents a major component ing fission product ratios to determine detailed reactor of LANL’s mission, and is aimed at reducing the threat operations used to irradiate nuclear material. We are posed by nuclear weapons proliferation or the illicit proposing to analyze a very unique set of irradiated re- trafficking of nuclear materials through the long-term actor samples and to use our measured fission product development of new and novel technology. The new di- ratios to validate the scheme. The samples in question agnostic signatures for reactor operations proposed here were part of a plutonium production-testing program. directly address several components of this mission. This Our goal is to show how fission product ratios can be work is of potential interest to non-proliferation pro- used to determine the reactor flux, the irradiation time, grams. as well as the number of times the re-actor was shut off during irradiation. As detailed below, the key ra- Progress tios of interest are the xenon isotopes and their decay This year we analyzed several samples from the reactor. products, 135Cs/137Cs, 103Ru/106Ru (and their decay Because the samples are old (over 20 years old), many products 103Rh/106Pd), and 85Kr/84Kr. The project is of the short-lived fission products have decayed away, a joint experimental and theoretical one. The experi- necessitating the use of long lived or stable isotopes ments will concentrate on measuring the relevant fission for determining reactor conditions. Xenon isotopics product ratios using analytic chemical techniques and and daughter cesium isotopes (136Xe/134Xe and mass spectroscopy. The theoretical program will derive 137Cs/135Cs) can provide information on the reactor semi-analytic prescriptions for extracting the signatures flux, while ruthenium isotopics (106Ru/103Ru or their for the reactor operation parameters. The results will be progeny 106Pd/103Rh) in conjunction with Pu isotopics com-pared with the reactor operations declared by the can be used to infer irradiation time. Krypton isotopics facility that provided the samples. The coupling of the (85Kr/84Kr) can be used to determine cooling time. two components (experimental and theoretical) of the project will provide the necessary quantitative validation Historical methods for collecting these volatile fission that is needed for practical use of the scheme in scenar- products (VFPs) for isotopic determination are labori- ios ranging from pre-detonation forensics to monitoring ous and/or hazardous, thus we sought to streamline the of reprocessing facilities. collection and analytical process. Using simple extraction chromatography methods followed by Inductively cou- The ability to extract reactor flux, irradiation times, and pled plasma mass spectrometry (ICP-MS) and thermal shut downs from fission products would: ionization mass specroscopy (TIMS) analysis we deter- mined Cs, Ru (and daughter Rh/Pd), and Pu isotopics • Greatly narrow the possible reactors used to irradi- for a set of archived reactor targets that were subjected ated fuel, in the case of seized nuclear material. to low burn up irradiation. We compare our results to • Determine the grade of plutonium and identify reactor models to evaluate the usefulness of VFPs for undeclared nuclear blankets, in the case that fission determining reactor operations at low burnup. products were released in fuel reprocessing. • Provide new signatures for verifying reactor opera- Separate aliquots from the primary dissolution were tions, in the case of nuclear safeguards. taken for U and Pu isotopic and assay determination. 731

Page 732

Chemical separations were performed using a macropo- Future Work rous anion-exchange (60–150 mesh, chloride form) resin The goal of the project is to measure the ratios of fission using a variety of mineral acids HCl and mixed HCl-HI to se- products, as well as the uranium and plutonium content, quentially elute Am, Pu, and U, thus eliminating isobaric in- of a number of reactor-irradiated samples that were part terferences between Pu and Am, and Pu and U The total U of a plutonium production program. In FY15 we analyzed 3 and Pu elemental contents and the blank were determined of the available samples, and found that all ratios that de- using standard isotopic dilution techniques with 233U and pend on the total neutron flounce agreed well with experi- 244Pu enriched spikes. Following chemistry, elutions were ment. However, for those observables that are dependent dried, and loaded onto filaments for TIMS analysis. on both the fluence and flux there was poor agreement between experiment and theory. To address this problem Age corrected uranium and plutonium isotope composi- in FY16 we plan to extend our reactor burn code to include tions as well as total Pu content are shown in Table 2. The neutron production and depletion of all isotopes of inter- samples are slightly depleted in 235U relative to natu- est and their precursors. ral uranium. Total Pu contents are low (30 to 159ppm) and the Pu is almost entirely (>99%) 239Pu. There is a On the experimental side of the project we propose to de- strong negative correlation between 235U/238U and velop a gas handling system that will allow us to measure 240Pu/239Pu and a positive correlation between Pu con- the Kr and Xe fission fragment products in several samples. tent and 240Pu/239Pu. Projecting the The U 235U/238U The system for this is designed to prevent (or at least mini- and 240Pu/239Pu correlation to 240Pu/239Pu = 0 yields mize) noble gas losses. In addition, we plan to measure a 235U/238U of 0.0724, very close to natural U (0.0725). the daughters of the ruthenium isotopes, the rhodium and These data indicate that the starting fuel was likely natural palladium isotopes. Analysis of the Rh and Pd isotopes will U which was subjected to very low burnup. require correcting the measured yields in the samples for the direct fission held of these, as well as their production Results from ruthenium isotopic analysis are given in Table via the beta decay of the Ru isotopes. 3. The cumulative yields of 102Ru and 104Ru are vastly different between 235U fission and 239Pu fission. For In FY16, we also plan to finalize and publish our analysis of the reactor targets, the 101Ru/104Ru and 102Ru/104Ru the first three samples studied to date. values are consistent with burnup of mostly 235U. Samples 3 and 4 plot close to pure 235U fission. Samples 1 and 2 Conclusion have higher concentrations of 104Ru relative to 101Ru and We will analyze 15 reactor-irradiated samples that are rep- 102Ru than samples 3 and 4, indicative of more fission resentative of a range of reactor fluxes, irradiation times, contribution from 239Pu. Sample 2 plots off of the tie-line and shutdown times. The grade of plutonium will be deter- between 235U fission and 239Pu fission. This suggests that mined for each sample and the relationship between the this sample contains some natural Ru. fluence and the Pu grade mapped out. We will expand on our theoretical methods for deducing reactor fluxes from The samples range from 137Cs/135Cs of 1.30 to 2.34. the 136/134Xe and 135/137Cs ratios to include burn sce- These suggest the samples experienced very low burnup. narios in which isotopic ratios have not necessarily reached There is a positive correlation between 137Cs/135Cs and equilibrium because of frequent reactor shutdowns. We 240Pu/239Pu, which is consistent with increasing burnup. will develop new techniques for extracting irradiation Record indicate that samples 1 and 2 were irradiated for times from the xenon, krypton and ruthenium ratios, and 85 hours and samples 3 and 4 were irradiated for 80 hours. derive the necessary inversion algorithms. Thus, with the flux and irradiation time known (i.e. flu- ence) we can determine the number of reactor shutdowns Publications from 137Cs/135Cs . It appears that the reactor was shut Byerly, B., L. Tandon, A. Hayes, P. Martinez, N. Xu, F. down 3 times during irradiation of sample 2. Sample 3 Stanley, R. Keller, M. Schappert, M. Thomas, and K. does not match this model, which could be due to it being Spencer. Determination of flux and irradiation time in a different irradiation location within the reactor and in archived low-burnup uranium ratgets. To appear receiving less fluence than recorded or obtained from the in MARC (Methods and Aplications of Radioanalytical archives. Samples 3 and 4 were irradiated over the same Chemistry) 2015. (Hawaii, 12-17 April, 2015). time period. They appear to have experienced 5 reactor shutdowns during irradiation, with sample 4 potentially having experienced less fluence than records show. 732

Page 733

Science of Signatures Exploratory Research Continuing Project Mapping Relativistic Electron Precipitation: Where and When? Steven K. Morley 20150127ER Introduction Benefit to National Security Missions Earth’s upper atmosphere is a major sink for highly ener- This project targets understanding where, when and getic electrons in the Van Allen radiation belts. Electron why highly energetic electrons are lost from the Van precipitation has been shown to affect telecommunica- Allen radiation belts. In order to answer these ques- tions and atmospheric chemistry, and plays a critical role tions we will develop a new technique for combining in determining radiation belt dynamics; it is therefore a point measurements from a constellation of satellites to key process to understand for space weather modeling. give a remote sensing capability. From this we will map Current detection of electron precipitation into the at- regions of electron precipitation and infer key properties mosphere is limited to direct measurements at locations of the waves driving the electron loss. The specific work occupied by satellites. A major drawback is that for the proposed will enable future development of models of large volume of space occupied by the Van Allen radia- electron precipitation and EMIC wave activity and prop- tion belts we only measure this region sparsely and so erties. Such a capability would directly address the goals directly measuring this precipitation is done with satel- of the NASA Living with a Star program: to improve our lites at very low altitudes. Most satellite instrumentation understanding of how the sun impacts the environment on these low altitude missions cannot give sufficient in which satellites operate. While this proposal does not information to describe the changes in the main Van Al- directly address this, these models are known require- len belt electron population, nor do they provide infor- ments for ongoing Space Weather and Space Situational mation about the processes driving the precipitation. Awareness modeling efforts. The datasets generated will enable further basic research and applied space weather We propose a new approach that will allow the first research. specification of the spatial and temporal extents of elec- tron precipitation from point measurements at high alti- Progress tude. In essence, we propose to combine measurements The magnetic ephemeris data required to identify the from different points on the paths that electrons take magnetic conjunctions have been calculated for a signifi- when drifting around the Earth in the Van Allen belts cant time span for many satellites. Software implement- (drift orbits) to provide a novel remote-sensing capabil- ing the novel binary tree-based technique for efficiently ity. Specifically, we will use the LANL particle instrumen- identifying these conjunctions has been completed and tation on the Global Positioning System (GPS) constella- tested for a small sample of satellites over a limited time tion, and at geosynchronous orbit, to identify differences range. This work is being presented at the International in the electron population along a partial drift orbit, thus Association of Geomagnetism and Aeronomy (IAGA) “remote sensing” the loss of energetic electrons. By general assembly - a major international conference. using the measurements of the full electron population, and some modeling of the inner regions of geospace, Preparation of the data required for the “phase space we will also infer key properties of the waves believed to density differencing” analysis has been performed for precipitate the electrons into the atmosphere. several satellite-years of LANL geosynchronous particle data and processing is ongoing. An improved method for This approach will overcome current sampling limi- calculating the phase space density data from the Van tations to diagnose electron losses and answer the Allen Probes, using additional low-energy measurements question: Where, and under what conditions, do highly to constrain the energy spectrum, has been implement- energetic electrons precipitate into the atmosphere? 733

Page 734

ed and 4.5 satellite-years of data have been processed. gions in the Van Allen radiation belts from which electrons Validation of these data is underway, including estimation are being lost. Our primary objective is to demonstrate of the uncertainties on the data (which will allow us to that point measurements from a constellation of satellites determine the sensitivity of our technique in detecting real can be combined to remote sense, and hence map, regions loss of electrons from the Van Allen belts). of loss. Our secondary objective is to use these data to infer key properties of a type of electromagnetic wave re- Cross-calibration efforts are underway between the pri- sponsible for driving some of the losses. This will produce mary data sets: LANL geosynchronous electron data are new physical understanding, enable scientific studies not being validated against Van Allen Probes data; electron previously possible and provide critical inputs for radiation data from the LANL particle instruments on the Global Po- belt models. sitioning System (GPS) are also being validated against Van Allen Probes. Preliminary on-orbit comparisons between Publications GPS and Van Allen Probes show outstanding agreement Halford, A. J., B. J. Fraser, and S. K. Morley. EMIC waves and and a manuscript is in preparation for submission to the plasmaspheric and plume density: CRRES results. 2015. Space Weather journal. Preliminary comparisons between JOURNAL OF GEOPHYSICAL RESEARCH-SPACE PHYSICS. geosynchronous satellite data and Van Allen Probes are 120 (3): 1974. generally encouraging but have shown a need to incorpo- rate recent data, that better characterizes the Synchronous Orbit Particle Analyzer (SOPA) instrument response at dif- ferent energies, so that the measured fluxes are consistent between both platforms. As part of the National Science Foundation’s Geospace En- vironment Modeling workshop, PI Morley is a focus group convener and has led efforts to define a number of can- didate events for coordinated community modeling. One category of the candidate events provides ideal conditions for testing the proposed methodology and the community modeling will provide analysis against which the results can be validated. This gives the project a head start on the tasks for the second fiscal year, and provides specific intervals for which the remaining data processing and preparation can be targeted. The class of events is rapid depletions of the Van Allen radiation belt population that occur outside of geomagnetic storms – these events have rarely been reported in the literature and so a manuscript describing these events using data from many satellites is underway. Future Work The goals for the second fiscal year of this project are: to finalize the data and model preparation the underly the phase space density differencing technique; develop the PSD differencing technique and software to perform rou- tine analysis; select key intervals for preliminary analysis; generate time-dependent maps of inferred non-adiabatic processes. Manuscripts on the cross-calibrations and preliminary analyses will be prepared and submitted for publication. Conclusion We will develop and test a new approach to locating re- 734

Page 735

Science of Signatures Exploratory Research Continuing Project Exploiting Cross-sensitivity by Bayesian Decoding of Mixed Potential Sensor Arrays Rangachary Mukundan 20150236ER Introduction Benefit to National Security Missions Finding specific chemical signatures amongst a back- The concept described in this project has the potential ground of ordinary substances – a trace explosive signa- to provide low-cost, robust, easily deployable gas-phase ture, an illegal narcotic, a chemical agent intended to kill quantitative discriminatory sensing capabilities that our soldiers on the battlefield, or civilians in our cities do not exist today. The broad implications of this work – remains a daunting task. Dogs can seem to handle directly support the Revolutionize Measurements aspect this proverbial “needle” quite well, but what man’s of the Science of Signature pillar, and the applications best friend does so well has eluded the sensor industry. specifically targeted by the proposed work directly Science’s best solution to date is devices that use large address key Energy and National Security concerns. numbers of inexpensive sensor elements (typically resis- There are presently no low-cost and durable sensors to tive or polymer) in conjunction with pattern recogni- control and monitor the vehicle emissions systems on tion methods: commercial instruments based on these today’s Selective Catalytic Reduction (SCR) and Exhaust principles are available. However, there are significant Gas Recirculation (EGR) pollution control technologies shortcomings with the performance and capabilities of comparable to the automotive lambda sensor. Although these sensor constructs, as well as significant limitations exact details are understandably kept secret by the perti- to the pattern recognition methods used to extract a nent Federal Agencies, the explosives screening tech- meaningful signal from the data because these devices nologies employed at critical facilities and used on the target generic applications requiring significant calibra- battlefield are routinely rendered ineffective because of tion to obtain qualitative results. The sensors at the the presence of complex, natural and human-created heart of these systems are prone to drift and irreversible background interferences. The proposed research has poisoning, blocking their use in simple, yet harsh, ener- the ability to create a unique capability at Los Alamos gy-related tasks such as monitoring vehicle emissions or National Laboratory (LANL) by supporting the develop- smokestack pollutants. We propose to use a special class ment of inexpensive and portable systems for use both of ceramic solid-state electrochemical sensors that are by civilian first responders and as dedicated screening intrinsically inexpensive, durable, and stable. Nobody systems at airports, federal buildings, cargo containers, has ever attempted to create a sensor array using these etc. We will develop the program with collaborators in devices before. Their robustness opens the possibility of Materials Physics Applications (MPA-11) and Weapons detecting a great number of gas chemistries under con- (WX-9) and also work with Program Managers from ditions that would quickly destroy other types of elec- Applied Energy (M. Fox) and Global Security- Emerging tronic noses. Our highly innovative and original solution Threats (P. Knepper) to pursue additional funding from to the specificity problem is to create a new gas sensor federal agencies interested in Energy Security (Vehicle technology that marries unique sensor design created Technologies (VT)), and explosives (e.g. the Defense at LANL with a recently—developed, advanced Bayesian Threat Reduction Agency (DTRA) and the Department of inference treatment of the physical model of relevant Homeland Security (DHS)) sensor-analyte interactions. Our approach is designed to quantitatively decode complex mixtures with fidelity Progress greater than present technology can deliver. Our meth- The first stage of this project involves the collection ods will solve the major shortcomings of the electronic of a data set from LANL to feed into the proof of con- nose concept. cept model development work at Rutgers University. A 735

Page 736

subcontract was placed in March to Rutgers University to set was in sufficient to predict the NO/NO2 ratios with initiate the model development work. LANL performed confidence. The model is currently being extended to use extensive modification to an existing sensor test sta- the “biased” data in order to provide further fidelity to tion to enable the collection of the appropriate volume the predictions. These initial results were presented in an and quality of data needed for the model development invited talk titled “Quantitative Decoding of Complex Gas work. A multi-mass flow controller manifold was built and Mixtures for Environmental Monitoring Using Mixed Po- automated with a lab-view based control and data collec- tential Sensor Arrays” at the 227th meeting of the Electro- tion program. This system was successfully commissioned chemical society by Dr. Cortney Kreller. within the first 2 months of this project and is capable of automated data collection from multiple sensors in the The progress to date is on track to meet the first milestone presence of 4 distinct test gases. of this project by the proposed deadline of December 2015. When the modeling and experiments are completed LANL initiated the sensor testing with the low-risk activity with the current sensor combination we expect the com- of looking at the concentration of analyte species where putational framework to provide feedback to the design background interferents are known and quantifiable — of the physical mixed potential electrochemical Sensor relevant to vehicle emissions monitoring (Energy Security). (MPES) array by predicting optimal sensing characteristics For this test Nitrogen oxide (NO) was chosen as the analyte and the number of sensor elements required for robust species with Nitrogen dioxide (NO2), Ammonia (NH3) and operation. We will then prepare that MPES configuration Propane (C3H8) serving as the background interfering with the appropriate number and type of sensing element species. The first set of experiments involved the use of and evaluate it in the low risk application (Energy). We will two distinct mixed-potential thin film sensors, one hav- then extend that work to low concentrations of analyte ing a Lanthanum chromite (LSC) sensing electrode and species belonging to a chemical signature of interest in the the other using gold (Au) as the sensing electrode. Both presence of a large and varying background (Homeland Se- sensors had a Platinum (Pt) reference electrode and yttria curity). We envision as our end goal a versatile, standalone stabilized zirconia (YSZ) electrolyte. These sensors were device that after a single mission-specific calibration in a specifically chosen as prior work had determined that the laboratory setting may be deployed in both remote and Au electrode had preferentially sensitivity to NH3 while the harsh conditions to provide quantitative analysis of chemi- LSC electrode had preferentially selectivity to either C3H8 cal signatures against large and unknown environmental or NOx depending on the mode of operation. The two backgrounds. operational modes selected were “un-biased”: where the voltage was measured at zero current, and “biased”: where Future Work the voltage was measured at a fixed positive current bias. We will first demonstrate proof-of-concept of the mixed- potential electrochemical sensor (MPES) array by fabricat- The Au and LSC based sensors were tested in our newly ing a prototype sensor array and developing/validating modified test stand using various binary pairs of analyte/ predictive algorithms for the quantitative assessment of interferent gases at various mixing ratios (a). Five distinct individual constituent concentrations in a complex mixture as varying from 0 to 4 were selected for the 3 binary pairs under a laboratory setting.We will initially fabricate sensor (NO/C3H8, NO/NO2 and NO/NH3). A total of 125 combina- arrays by sputtering working electrodes that are known to tions of these mixing ratios were measured for 12 distinct have widely different selectivity to Nitrogen oxides (NOx), concentration of NO leading to 1500 distinct data points ammonia (NH3), carbon monoxide (CO), and non-methane being recorded for the Bayesian model validation. The data hydrocarbons (NMHCs). The electrode compositions to be was recorded simultaneously on both sensors at each gas used will include strontium doped lanthanum chromites compositions in order to accelerate data collection. This and strontium doped lanthanum manganates along with data was then analyzed by Dr. Morozov at Rutgers. Both platinum pseudo reference electrode onto the heater linear and non-linear models were considered, with the platform. The yttria stabilized zirconia (YSZ) electrolyte non-linear model providing the best fit to the data. This is layer will be electron beam (e-beam) evaporated. We will not surprising since the non-linear model accounts for the measure the voltage response of each individual sensor saturation of the sensor response at high analyte concen- element of this array to a wide range of ratios of target trations. analytes, α=[Ca]/[Cb] (where [Ca] denotes the concentra- tion of analyte a), as a function of operating temperature. The preliminary model was trained with just the “un- We have at our disposal a number of variables to optimize biased” data set and was able to predict the as within 20% the sensing characteristics of our array for a specific ap- when using just the NO/C3H8 mixtures. However this data plication. In addition to electrode composition and tem- 736

Page 737

perature, if needed, we can also control electrolyte thick- ness and morphology (porosity), for each sensor element of the array. The dataset described above will be used to determine model parameters for our previously developed Bayesian algorithm that determines concentrations of all constituents in an unknown mixture from the output of a cross-specific sensor array. The Bayesian analysis will provide quantitative guidelines for designing optimal sets of sensors by enabling us to 1) identify the optimal de- pendence of individual sensor response on the relative amount of compounds in the mixture and 2) determine the number of sensors in the array required to successfully discriminate the target species from one another, as well as against additional background interferents. Conclusion Proof-of-concept will be demonstrated by fabricating a prototype sensor array and developing/validating predic- tive algorithms for the quantitative assessment of individ- ual constituent concentrations in a complex gas mixture. After successful demonstration of proof-of-concept we will work closely with our partners at ElectroScience Laborato- ries (ESL) to fabricate for the very first time Mixed potential Electrochemical Sensor (MPES) array prototypes using their proven High temperature co-fired ceramic (HTCC) manufacturing methods. We will then demonstrate the ability of these sensor arrays and signal processing tech- niques to analyze individual gas components in two differ- ence applications; viz: 1) vehicle emissions monitoring and high-explosive/energetic-materials detection. Publications Kreller, C. R, A. Nadiga, S. C. Brown, J. M. Reynolds, D. Spernjak, F. H. Garzon, E. L. Brosha, A. V. Morozov, and R. Mukundan. Quantitative Decoding of Complex Gas Mixtures for Environmental Monitoring Using Mixed-Potential Sensors. 2015. In 227th ECS Meeting. (Chicago, IL, 24-28 May. 2015). , p. Abstract MA2015. Chicago: ECS Meeting Abstracts. Reynolds, J. M., S. C. Brown, E. L. Brosha, R. Mukundan, F. H. Garzon, and C. R. Kreller. Electrochemical Character- ization of Electrode Materials for Mixed-Potential Sen- sors. 2015. In 227th ECS Meeting. (Chicago, IL, 24-28 May. 2015). , p. Abstract MA2015. Chicago: ECS Meet- ing Abstracts. 737

Page 738

Science of Signatures Exploratory Research Continuing Project Measurement of Extinct Radionuclides in Historic Nuclear Debris (U) Warren J. Oldham 20150298ER Introduction Benefit to National Security Missions A nuclear explosion produces an intense burst of energy A major limitation of current radiochemical diagnos- and associated radiation driven by the violent assembly tics is the need to determine the concentration of key of a critical mass of fissioning nuclear material. Details of signature radionuclides with short halt-lives. Within a the nuclear event have traditionally been inferred from month after a nuclear detonation, these radionuclides measurements made during and immediately after the are largely extinct. If successful, the proposed work will detonation. In the case of radiochemical measurements, be a critical step towards a new capability to extract this many of the nuclear products that have been shown to crucial information from nuclear debris at later times. be of most diagnostic value exist for but a moment in Such a capability would have direct and lasting impact time, typically on timescales of hours to weeks. on two key DOE/NNSA missions: stockpile stewardship and post detonation nuclear forensics. A demonstrated The key innovation of this proposal is the realization that capability to reanalyze archived debris samples from his- short-lived fission products and key nuclear activation toric nuclear tests will improve the technical foundation products that form the basis of modern technical evalua- that supports Stockpile Stewardship and will strengthen tion can in fact be indirectly measured over vastly longer our Nation’s commitment to deterrence without nuclear timescales from years to decades, and indeed centuries, testing for decades to come. The limitations of histori- after the detonation. We borrow a key insight in how to cal measurements impact nuclear forensics as well. accomplish this task from similar approaches that have Older U.S. tests have relevance to low-technology threat been successfully used in the field of cosmochemistry space. Because many of these tests were fielded in the to measure extinct radionuclides in extraterrestrial late 1940’s and early 1950’s, the radiochemical data for samples. Chemical fractionation between Pu and U that these events is very sparse and/or of low quality. This is known to occur during a nuclear event will cause iso- leads to large uncertainties in evaluation of performance topic perturbation in 237Np and 239, 240Pu concentra- and yield, which in turn prevents proper calibration of tions resulting from decay of short-lived U isotopes 237, LANL’s computational design tools. A capability to obtain 239, 240U. The proposed sampling and measurement a complete radiochemical data set for early U.S. tests techniques will be used to determine diagnostically sen- would significantly improve modeling and forensic capa- sitive and now extinct 237, 239, 240U isotopes in historic bility envisioned for post-detonation nuclear forensics. debris samples. Progress A significant challenge of the proposed research will be This research aims to develop analytical techniques to to develop indirect methods to measure peak-yield fis- characterize the details of a nuclear detonation based on sion products to determine total fissions in a given sam- chemical analysis of aged nuclear debris. The analytes of ple. In order to reconstruct this information we propose interest are short-lived radionuclides that have decayed to measure perturbations in stable element isotopic beyond detection, but leave signatures of their original composition to infer the original presence of short-lived presence in the isotopic composition in archived debris fission products. This idea depends on the presence of samples. We term these signatures, extinct radionu- stable multi-isotope elements for which one isotope is clides. The goals of the project are to recover two types the end-point of a fission decay chain and another is of information, (1) short-lived uranium isotopes (237, blocked from fission decay by a different stable element. 239, 240U) and (2) fissions. For each of these goals good 738

Page 739

progress has been accomplished. In order to determine the terference from 98MoH+ and 99Ru. Standard addition ex- short-lived uranium isotopes, a collection of samples must periments with solutions of Peruvian soil (NIST SRM 4355) span a suitable range of volatility as measured by sample and dissolved Trinitite suggest that accurate determina- specific 236U/238Pu ratios. Although only one unfraction- tions of 99Tc can be obtained. The associated measure- ated 236U/238Pu value is actually “true” of the device, ment uncertainty (20-30%) is higher than ideal and reflects chemical distortion that occurs during debris condensation inconsistencies in observed instrument count rates that serves to generate a range of samples that can be relative- are not yet fully understood. Additional work is needed to ly enriched (volatile) or depleted (refractory) in uranium identify the source of the variation and improve measure- compared to a reference 238Pu isotope. The proposed ment precision. In the case of 95, 97Zr determination, a regression technique requires that 236U/238Pu span a suf- novel technique using high precision isotopic measure- ficient range within the sample set to allow a robust linear ments of stable molybdenum is being pursued. A sequen- fit. tial two-column ion-exchange protocol was developed and optimized. Purified Mo concentrates have been isolated In the case of analyses of Trinitite, eight individual samples with good decontamination against potential interferences from archived LANL collections have been completed. The from Zr, Ru, Cr, Mn, and Fe. expected trends are observed, however the goodness of fit over a limited range in 236U/238Pu is not sufficient to The purification protocol performed well in an initial Mo obtain a high-quality measure of 237, 239, 240U. A new set analysis of a dissolved debris sample, which revealed of Trinitite samples obtained from a commercial rock and significant deviation from the natural isotopic composition. gem supplier has been procured. These samples appear as An optimized measurement routine using the Neptune light green vesicular samples, distinctly different from the Plus multicollector ICP-MS is currently being developed. black glass found in the LANL collection. It is hypothesized These high precision measurements will be used to quan- that the light green vesicular samples will be more volatile titatively determine perturbations in 95Mo/96Mo and compared to the black glass samples and will provide the 97Mo/96Mo isotopic ratios. Future analyses using an range in 236U/238Pu that is needed. Initial analyses of isotopically enriched 96Mo spike will be interpreted for archived core samples from an underground nuclear test absolute fission concentration. have also been completed. For this test, archived samples that transit seven locations through the underground Future Work environment are available for analytical study. Duplicate Chemical fractionation between Pu and U can be exploited analyses of the first two locations have been completed. to determine the original concentration of 237,239,240U Five more duplicate analyses will provide data to assess produced in a nuclear explosion. The sampling require- the 236U/238Pu volatility index and the feasibility of the ments as well as the scope and limitations of the method technique to measure 237, 239, 240U in underground need to be defined. This research will employ a case study debris samples. approach to help determine these boundaries. During this FY, work will focus on analysis of archived debris samples Two independent approaches are being pursued to deter- from the Trinity test and from an underground nuclear mine fissions in historical debris samples. These include test. The resulting data will be evaluated through a system- direct measurements of 99Tc and indirect measure- atic uncertainty analysis to understand how uncertainty in ments of 95Zr and 97Zr through isotopic perturbation of both measurements and in linear regression techniques stable molybdenum isotopic ratios (e.g. 95Mo/96Mo and ultimately defines the analytical determination of extinct 97Mo/96Mo). Each of these analytes is a peak yield fission radionuclides. Final data will be evaluated in the context product that can, in principle, be used to determine total of nuclear debris diagnostics using the modern Isotope fissions in the sample. The first phase of this project has Production Code and compared to historic assessments. focused on development of chemical methods used to The second goal is to develop indirect methods to measure isolate pure 99Tc and molybdenum concentrates that can peak-yield fission products to determine total fissions in a be assayed using both single and multi-collector ICP-MS given sample. Our first choice to accomplish this task is to instrumentation. In the case of 99Tc determination, a determine the concentration of extinct 95, 97Zr isotopes method has been developed during the first year of this through perturbations on stable molybdenum isotopic project to analyze for 99Tc in aged debris samples. Tech- composition. The short-lived Zr isotopes decay to stable netium is purified in high yields (80-100%) using extraction 95, 97Mo. Adjacent fission decay chains (mass 92, 94 and chromatography (TEVA resin) with good decontamination 96) are all blocked by stable Zr isotopes such that 92, 94, from both Mo and Ru. This decontamination is crucial to 96Mo are unperturbed by fission decay. A method to ob- the success of the analysis because of potential isobaric in- tain high precision isotopic analysis of purified Mo samples 739

Page 740

will be developed using multi-collector inductively-coupled plasma mass spectrometry (MC-ICP-MS). During this FY focused effort will be spent on developing and validat- ing the chemical and instrumental methods. Beginning in the second year of this project we should be in a position to apply the Mo isotopic analysis approach to determine extinct 95, 97Zr fission product isotopes in archived debris samples. The same solutions used to determine actinide composition will be held until the extinct fission product measurement techniques are ready. Conclusion Debris from historic nuclear test rests on the ground sur- face and within underground cavities at all nuclear testing sites. This material could be used to determine device de- sign and performance but is currently of limited diagnostic value because the short-lived fission products and nuclear activation species have long decayed away. The goal of this project is to develop new measurement and assessment techniques to reconstruct diagnostic information as the stable decay end-members. The success of the proposed effort will be a critical step towards a new capability with direct impact on nuclear forensics and on science based stockpile stewardship. Publications Hanson, S. K., C. R. Waidmann, J. L. Miller, H. D. Selby, and W. J. Oldham. Modern radiochemical measurement and evaluation of 1950’s era debris from the Nevada Test Site (U). To appear in Defense Research Review. 740

Page 741

Science of Signatures Exploratory Research Continuing Project Ultra-sensitive Parallel Micro-imaging with Atomic Magnetometer Igor M. Savukov 20150300ER Introduction of micro-chips; DNA detection and other applications of It is widely recognized that the next frontier for mag- magnetic nano-particles; and brain science and cancer netometry is biological and neuroscience applications, detection and studies. where impact will result from improvements in both resolution and sensitivity. We propose to meet this need Progress for higher resolution and higher sensitivity magnetom- A postdoc was hired and trained in safety, atomic mag- etry with a novel approach: combining an ultra-sensitive netometers, magnetic field modeling, sensitive measure- atomic magnetometer (AM) with flux concentrators ments, and some other fields. (FCs). The technological breakthrough will result from advances in atomic magnetometers and from the novel The research team designed, modeled, and constructed FC-AM concept. After developing the FC-AM device, we a ferrite flux concentrator (FC). Its resolution and expect- plan to demonstrate its performance by applying it to ed trade-off in sensitivity was verified. We found agree- important problems in science, security, medicine, and ment between theory and experiment. industry. A small-size atomic magnetometer (AM) (1 cm outside The proposed research program will have two broad dimension) was setup and tested. Its bandwidth, which research goals. The first goal focuses on technology: (i) is important parameter affecting the ultimate sensitiv- demonstrate the FC-AM principle, (ii) characterize its ity, was measured and found to be within the expected sensitivity and resolution, and (iii) demonstrate that FC- range. AM can operate in multi-channel configuration. The sec- ond goal focuses on applications in bio-security (detec- Atomic magnetometer was inserted into the flux con- tion of magnetic nano-particles, neuroscience, detection centrator and its sensitive operation was confirmed. of bio-magnetic fields of a cluster of neurons), in energy We expected some effect of the flux concentrator on research (fuel cell ion-transport imaging), and in industry the operation of AM, namely the broadening of the AM (non-destructive evaluation, NDE). magnetic resonance which could potentially substan- tially decrease the sensitivity. We solved the problem by Benefit to National Security Missions adding compensation gradient coils. Now the AM oper- The work directly addresses the need for sensitive mag- ates inside the gap of the flux concentrator the same netometers with high resolution. One class of NIH appli- way as in the ideal shielded environment. cations is based on detection of magnetic nano-particles with our micro-magnetometer: applications in detection At this point, we are planning to replace the current AM cancer cells, infections and cancer therapy. The second with a new-generation AM from QuSpin company that class is based on detection of magnetic field of neurons. will improve sensitivity by at least an order of magni- tude. With this new AM, we planning to make demon- Other potential applications include magnetometers, strations of sensitive detection of FC+AM single-channel used for fuel cell diagnostics, which is relevant to re- system. This is our major milestone. duction in CO2 emission and hence climate impact; pathogen detection based on magnetic nano-particles; We also focused on improving the resolution of the tip checking for pathogens at our boarders; authentication of the FC. Initial ferrite FC was extended with mu-metal 741

Page 742

tips, since ferrite material cannot be sharpened enough, research, non-prolifiration. In addition, We will test non- and with this approach we are reaching our proposed tar- destructive evaluation methods. Finally, if time permits, in get resolution of 0.1 mm. Various tests of resolution were collaboration we will work on detection of magnetic field conducted with micro-coils as the sources of magnetic field of a single neuron or small clusters of neuron. The prelimi- at small scale. Further improvements in resolution will re- nary data will serve as the basis for proposals, for example quire the use of micro-fabrication techniques and magnetic within President Obama BRAIN initiative, or DHS (biosecu- micro-particles. rity), or DOE for fuel cell research. Because we are now waiting for QuSpin company to manu- Publications facture for us a new-generation AM and we confirmed that Espy, M. A.. Hardware Developments: Detection using FC-AM approach works, we started work on an AM multi- SQUIDs and Atomic Magnetometers, Mobile MRI/ channel system. We are preparing 16-channel AM system Chapter 7. To appear in Mobile NMR and MRI: Devel- that is based on a large pancake Rb cell, 16 fiber-coupled opments and Applications. Edited by Price, B.. photo-diodes, and 16-channel DAQ interface. Karaulanov, T.. Spin-exchange relaxation-free magnetom- eter with nearly parallel pump and probe beams. To We are preparing a paper for publication where we dem- appear in Physical Review A. onstrated a few fT/Hz1/2 sensitivity with a large Rb cell, which will be used later for multi-channel detection. Savukov, I. M.. SERF and derivative high-density alkali-met- al magnetometers. To appear in Smart Sensing, Mea- An additional postdoc was appointed as Director’s Fellow, surements of Instrumentation. Edited by Grosz, A.. and will work in conjunction with this project. Although his project has many independent goals, there are many common goals, such as sensitive micro-imaging and its ap- plications. Detection of magnetic micro-particles with the system is in preparation stage and will be demonstrated after the new AM will be procured. Overall, substantial progress was made. We were able to demonstrate a single channel FC-AM operation, although we continue the work to improve sensitivity to make our results even more exciting for the scientific community. Future Work A single-channel atomic-magnetometer and flux-concen- trator system will be constructed to evaluate experimental- ly the resolution-sensitivity tradeoffs. A single-channel AM is already available for the project. The flux concentrator will be constructed from 500-micron ferrite rods sharp- ened to 100 micron, available from Ceramic Magnetic, Inc. Characterization of the sensitivity and resolution will be conducted and work will start on developing new applica- tions. Conclusion After the FC-AM devices have been tested and character- ized, we will work on developing novel applications. We will demonstrate detection of magnetic nano-particles functionalized for specific molecules. Nano-particle ap- plications are relevant to LANL missions in security, energy 742

Page 743

Science of Signatures Exploratory Research Continuing Project Segregated Fuel-Oxidizer Propulsion for CubeSat Deployment Bryce C. Tappan 20150323ER Introduction little or no oxygen, a segregated tandem system can be Small, utilitarian “cube satellites” (cubesat) that can be designed such that the energetic material will provide easily deployed into low-earth orbit and have high-func- fuel from their decomposition that will be oxidized in tionality are increasingly more popular. These cubesats a separate chamber by reaction with a solid oxidizer. ride along with much larger expensive satellite launches Because the fuel and the oxidizer are separated until on a “bus” and the customer is limited to a pre-deter- combustion of the fuel is initiated, and are both are rela- mined orbit. As of yet, no propulsion system exists that tively (or completely) insensitive to shock, the chance of has sufficient impulse for orbital plane adjustment as accidental detonation or initiation of the rocket is dra- well as the required factors of safety necessary to piggy- matically reduced. These technologies will allow for both back on a launch that may surpass $1B in cost. Here we a higher energy and much safer solid propellant propul- propose to develop a propulsion package based on the sion system than is currently offered the state-of-the-art. LANL segregated fuel-oxidizer rocket system, which could These aspects are particularly relevant to cubesat pro- provide both the impulse and safety required for orbital pulsion, as there are currently no high specific impulse plane adjustment for these small satellites. The main (Isp) options that qualify as safe. Cubesats are a second- thrust of the research will be to design and test innova- ary payload and the orbit is dictated the primary pay- tive form factors utilizing two different motor concepts load, thus the ability to change orbitology from 300km that will provide sufficient impulse for orbital correction. to 500km is huge. Further points that are significant in Energetic fuels that provide copious amounts of high- cubesat propulsion include the ability to build/maintain temperature H2 gas can be incorporated into segregated a constellation, station keeping, the ability to extend the systems that flows to a separate chamber that provides life of a very low orbit <300km and even interplanetary extremely high impulse by reaction with a solid oxidizer. space travel to asteroids or the moon. This will greatly Because the fuel and the oxidizer will remain separated expand the functionality of existing cubesats. Also key until decomposition of the solid fuel, and both are rela- in this development is the ability to deorbit, such that tively (or completely) insensitive to shock, the chance today’s cubesat will not be tomorrows space junk. of accidental detonation or initiation of the rocket is dramatically reduced. Modifications of the system that Progress will also be studied will incorporate a solid oxygen gen- Significant progress has been made in the primary goal erator, much like those used for emergency oxygen on of decreasing the thrust of the propulsion system. While airplanes, to provide gaseous O2 that can react with fuel high levels of thrust can be advantageous in atmospheric gases or solid fuels in a separate chamber, analogous to applications, in space applications only overall impulse a liquid/liquid rocket chamber or hybrid rockets. Thus, is important, and lower thrust, while preserving overall three engineering concepts will be available to ensure energy, is desired for spacecraft stability during ma- success of the project. neuvers. We were able to achieve this reduction by a two approaches; a modification in the fuel and oxidizer Benefit to National Security Missions chemistry to provide a slower, but stable burning rate, The segregated fuel-oxidizer propellant system is a major as well as development of a new internal geometry in break-through in solid rocket propulsion in terms of the fuel-oxidizer grains. With the change in fuel burning safety and energy. Because of the development of high- rate, and the level at which the oxidizer decomposition nitrogen/high hydrogen energetics at LANL that contain is catalyzed in this system, we were able to reduce the 743

Page 744

thrust in core-burning geometries by ~3x. By producing an that provides gaseous oxygen to a mixing chamber with end-burning fuel grain (much like the regression of a ciga- the fuel product gases, analogous to a liquid rocket engine. rette) we were able to reduce the thrust further by ~10x. The same oxygen generator will be studied to provide Furthermore, because of the inclusion of higher levels oxygen to a solid fuel grain, analogous to a classic hybrid of aluminum powder in the fuel, this advantage of lower rocket engine, and the reverse of our primary concept. thrusts also comes with high specific impulse. Motors have The primary approach motor array will be fixed to a cubic been successfully tested at 1” diameter geometries, and geometry, and the second two systems will be designed we are currently scaling the system to 2” diameter. Initial as back-up technologies in the event of insurmountable observations, however, indicate that some formulation problems with the primary approach. The goal of all three optimization will be necessary to obtain optimal system systems is to provide consistent thrust, with a high-level impulse as well as acceptable combustion stability. Further of safety and better overall performance (specific impulse) work this FY has also focused on scaling down the system than current proposed propulsion for cubesats. to 12.7 mm, and development of an x-ray transparent motor casing at this size to dynamically analyze the motor Well timed, reliable ignition systems will be a key com- performance during combustion. ponent to the success of the primary concept, so this will be one of the primary tasks for the first year. The goal will Significant calculations have been performed using the be to produce igniters that can function in less than one thermal equilibrium codes Cheetah (LLNL) and NASA CEA, second from the sending of the electric pulse, as well as both provide important guidance in motor performance, providing the sufficient heat to ignite the motor’s fuel. Also although there is required effort to reconcile the data key to the first year of the project will be the miniaturiza- between the two models. Cheetah will always do the tion of existing segregated fuel-oxidizer motors to fit the calculations by resizing the nozzle until the pressure of the space demands of the LANL cubesats. The goal here is a combustion gasses in the nozzle is the same as the ambi- 12.7 mm diameter motor. ent conditions. The Cheetah program has the advantage of a containing a more advanced library of condensed A new spacecraft concept will be explored requiring a phase species when considering aluminum reactions, single, larger diameter motor. but NASA CEA has a number of advantages. Both models converge on the overall fuel to oxidizer balance, but NASA Conclusion CEA indicates that aluminum doesn’t increase Isp or the The primary metric of success will be the measured system characteristic exhaust velocity, indicating that the opti- impulse of the propulsion system. These data will be used mum Aluminum content is ~10 wt% in fuel (about 3 wt% in to calculate the different orbits achievable from the pro- overall propellant), compared to the 30 wt% indicated by pulsion systems. Thrust and pressure measurements will the Cheetah code. This combined modeling efforts was ad- be performed on individual motor sections as well. Inde- vantageous this FY in guidance to optimize the formulation pendent, reliable, ignition control will be demonstrated while decreasing the number of expensive tests required. so that motor arrays can be utilized to provide a constant This is helpful as the high Al content is likely cause of the predictable direction of flight. A thrust stand with at least serious combustion instability we are observed in some of four separate load cells will be developed to verify uniform the rocket tests. Intermediate levels of aluminum additives thrust of the motor arrays. have thus been developed and are expected to provide op- timal impulse as well as combustion stability. Furthermore, alternative additives (aluminum hydride and aluminum- lithium alloy) have been investigated and show promise in modification of specific impulse, combustion stability as well as reduction in two-phase flow loss from the motor. Future Work Three different motor concepts will be considered, and the most practical will be selected and developed. The initial, and primary concept, will be arrays of right cylindri- cal motors with independent ignition control such that symmetrical sub-arrays can be fired at different times. The alternative designs will include the substitution of the flow through oxidizer with a solid pyrotechnic oxygen generator 744

Page 745

Science of Signatures Exploratory Research Continuing Project Practical Antennas from Disruptive Technology John Singleton 20150337ER Introduction Benefit to National Security Missions The project will design and construct two practical, Work on SAPC antennas will lead to intellectual property optimized, economical antennas to give improved relevant to radar (other federal agencies, DOT), remote performance compared to current state-of-the-art. The sensing and surveillance (ultra-long-range transmitters antennas to be built are called Superluminal Accelerated running on very low power) and secure communications Polarization Current (SAPC) antennas, and are based on (relevant to other federal agencies), pivotal to LANL core a new principle: moving a polarization current super- missions. The current proposal represents a high-payoff luminally (faster than the speed of light in a vacuum) quantum leap in the theory, computation and modeling and accelerating it. The radio waves emitted by the of SAPC antennas (information techology). This work polarization currents in SAPC antennas have properties could give back to the US a lead in antenna technology, analogous to the “sonic booms” of acoustics (emitted by reducing dependence on imported components from accelerating supersonic airplanes) in that they can have e.g. China (relevant to DOC). With topology optimization relatively high intensities at large distances from the applied to communications backhaul, we estimate that antenna. A technology-demonstrator SAPC antenna built factor 10-100 power savings could be made for all dis- at LANL has already shown advantages over conven- tances 1-100 km. This could lead to the replacement of tional technology for transmission distances > 10 km; in the world-wide telecommunications backhaul network the current project we hope to improve such antennas (relevant to DOC, infrastructure etc.), with huge reduc- using the numerical technique of topology optimization. tions in energy consumed; the lower transmitted powers This technique varies the shape, size and configuration needed ameliorate environmental concerns (relevant to of a component to give the best possible performance; environmental and Energy Impacts). it tries out many more variables and variations than are possible for a human designer to achieve in a lifetime. Accelerated superluminal polarization is thought to play Topology optimization has recently been used to give a role in the emission of radiation by pulsars, quasars dramatic weight savings combined with improved drag and gamma-ray bursts and in other astronomical phe- and lift in aeronautical components; however, it has only nomena. Therefore a deeper understanding of the elec- had limited application in antenna design. Advances in trodynamics of superluminal polarization currents is rel- computing mean that the time is ripe to apply topology evant to Cosmology and Astrophysics, as well as to DOE optimization to electromagnetic problems; and that the (SC) and NASA. Ultra-long-range, low-power communica- novel SAPC antennas are suitable disruptive technology tions are relevant to Space Science (communication with that could thus be brought to market. deep-space probes) and to NASA. Finally, the broadening of the application of topology optimization to antennas SAPC antennas for two different applications will be is relevant to information science and technology. optimized and built. The first will be for low-power, ultra-long-range communications. With optimization, Progress we estimate that factor 10-100 power savings could be The first stage in optimizing the design of the SAPC made for distances 1-100 km. The second SAPC antenna antennas was to adapt numerical calculations of the would focus radiation at a distance, demonstrating the emission from a superluminal polarization current to the feasibility of future medical, radar and directed-energy various possible antenna geometries: (i) circular with applications. radially applied electric field, (ii) linear with arbitrary 745

Page 746

acceleration and (iii) arc with vertically applied electric tion for SAPC antennas which is being studied for possible field. These numerical models have all been coded and de- future patents. bugged, and the results compared with experimental data from earlier “proof-of-principle” SAPC antennas. To check During the optimization process, a completely new ge- the robustness of the numerical method, each type of ometry of antenna (analogous to the synchrotron particle calculation was coded in two independent ways using dif- accelerator) was discovered; though not immediately ferent computer languages and operating systems. Within applicable as a communications antenna (the beam is of each implementation, different methods of numerical an unusual form) this may have applications in radar (the integration and differentiation (finite element evaluation of device has a 360 degree view without mechanical scan- the “curl” operation) were checked for accuracy, especially ning). The new design is being reserved for possible future at the distances required to model communication per- improvement and patenting. formance: the phases of signals from an antenna measur- ing 1/3 of a metre across must be accurately evaluated at Coding associated with the proposed, more sophisticated transmitter-to-receiver distances of up to 100 km. All of topology optimization to be used with the second antenna this was successfully accomplished. design is under way and proceeding according to expecta- tions. As part of this exercise, the numerical “algebra” nec- The simple optimizations outlined in the proposal have essary to deal with dielectric antenna elements of arbitrary taken the above results and used them to design a linear shape has been developed. antenna (type (ii) above) for communications purposes. In contrast with previous “proof-of-principle” SAPC anten- Future Work nas, the volume of the dielectric is much larger compared Optimization of the first antenna design will be carried out to the emitted radiation’s wavelength, and the elements in three stages. (i) Starting with simple shapes, the dielec- are sparser (i.e. there are fewer elements per unit length). tric parameters and shape, and the speed, acceleration Both of these refinements have resulted in an additional and spatial modulation of the polarization current within benefit: a design that should be economical to mass-pro- will be optimized for directivity, and distance-dependent duce on a base made from standard 16 mil circuit board. gain. This represents a relatively small number of design Components for the new antenna are either already made, variables, so a genetic algorithm (GA) will be employed. or are being fabricated and ordered; for flexibility, we are The primary challenge is the computational cost in deter- using mechanical phase shifters attached to each element mining the antenna performance. This will be addressed to vary the antenna’s emission pattern before fixing the using a multilevel radial basis function (mRBF) surrogate design of the passive feed network that would animate the model, avoiding the need to analyze every candidate de- polarization current in a mass-produced device. The chas- sign. (ii) The means of animating the polarization current sis, phase shifters, cables and several other components – electrodes on/in the dielectric driven by a passive feed are presently under test in the TA-35 anechoic chamber. network – will be included. We will employ a multiobjec- The design is modular, so that the current antenna ele- tive GA, NSGA-II to explore optimization strategies. (iii) ments can be replaced in future; refinements resulting Having a coarsely optimized design, topology optimiza- from the more sophisticated optimization routines that will tion will interrogate a wider design/shape space using the be applied to the proposed second antenna may result in SLP level set method. Here, revolutionary improvement further improvements. is expected; the previous stages are necessary to verify the method and the solutions. The number of variables All of these achievements form the basis of an invention will be ~1000 and the use of GA is no longer tractable. disclosure for a future patent. Based on the encouraging The optimum solutions from previous stages will provide results, a previous patent on SAPC antennas has also been benchmarks for verification of the level set optimization. extended to give further international coverage. Based on the analysis and experiment, we will select final optimum designs for the first antenna. Extensive testing of dielectrics (including a novel metama- terial comprising micron-scale copper loops dispersed in a We demonstrated an economical SAPC concept employ- foam) has been, and is, also under way. Based on the need ing passive feed networks and electrodes on commercial for a large volume polarization current suggested by the 30 mil circuit board, plus layers of easily shaped dielec- models, we have selected a plastic dielectric (rather than tric. Trial antennas, antenna elements and passive feed the alumina used in previous designs) for the preliminary networks will be built on this principle and measured to design of the proposed second antenna. Data from the validate numerical and optimization predictions using metamaterial have suggested a further possible applica- the anechoic chamber at TA-35. Optimization of dielec- 746

Page 747

tric properties will be studied in the numerical work; to validate this, some trial antennas may employ metamateri- als fabricated in MPA-CINT, tailored to fine tune dielectric properties. The first optimized antenna will be constructed based both on the outcome of the numerical optimization and fabrication/experimental experience with the trial antenna components. Conclusion The goal of the project is to build two antennas that com- bine and synthesize two recent developments: the novel Superluminal Accelerated Polarization Current (SAPC) antennas, developed at LANL and topology optimization, a numerical technique that has produced best possible designs for aeronautical components, but as yet has been little applied to antenna design. The first antenna, de- signed for communications, could lead to dramatic power savings for transmissions over distances of 1-100 km, lead- ing to the replacement of the worldwide telecommunica- tions backhaul network ($billions), with huge reductions in energy consumed. The second antenna will be used to demonstrate focusing of radio waves at a distance. 747

Page 748

Science of Signatures Exploratory Research Final Report Next Generation Earth Models Monica Maceira 20120047ER Abstract travel time or surface wave dispersion, for example) and this results in incomplete retrieval of earth properties We address the fundamental problem of creating an due to sensitivity limitations inherent in each of the data accurate image of the Earth, currently a barrier to un- types. Under this project we have addressed this difficul- derstanding the Earth’s present state and evolution, as ty through simultaneous inversion of multiple datasets well as a limitation in our ability to monitor the globe for each of whose sensitivities fill the resolution gaps of one nuclear explosions and other seismic events of interest. of the other data types, resulting in improved resolution. Our prime goal was to develop an efficient method to Prior efforts have primarily focused on integrating two, derive the Earth’s velocity structure from a combination and in rare cases three - and focusing only on restricted of four geophysical data types - body waves travel times, locations-, parameters into the inversion procedure. surface wave dispersion, receiver functions, and gravity We successfully developed a methodology and code anomalies. Because of the complementary sensitivities to incorporate the simultaneous joint inversion of four of each of the data types employed, our multi-param- consistent and complimentary datasets, and applied it to eter inversion is able to produce more accurate Earth large data sets over a broad region. models than currently available and will form the basis for the next generation of Earth models. We applied our Scientific Approach and Accomplishments method to the western US, a perfect test bed for this Data Gathering and Processing: The availability of Trans- effort, where unprecedented data sets are now available portable Array (TA) data provided a tremendous oppor- from the geophysics community. Our research effort has tunity, as data of this high a density over such a broad been a combination of code development, data gather- region had never before been available. The TA is a ing and processing, theory and method development, rolling array of 400 seismometers deployed on a 70-km- and extensive inversion of the multiple data types. This spacing semi-regular grid moving from the west coast research has created unique expertise at the Laboratory of the US towards the east coast. When these stations and enhances the Laboratory’s capability to address core are combined with other local networks in the west, the missions in Nuclear Nonproliferation Programs. data coverage is exceptional. We gathered and pro- cessed body wave travel time data for all these stations. Background and Research Objectives Initial body wave travel time tomography results were When an earthquake or underground explosion occurs, published in J. Geophys. Res.– Ref. [1]. seismic waves propagate from the source out through the Earth and can be recorded at a distance. Parameters Receiver functions (RFTNs), which can be thought of as derived from these waves, including their travel time, the impulse response of a near-vertical wave beneath dispersion, and amplitude, are used to infer the struc- a seismic station, are available only from the TA and ture of the Earth through which the waves have passed. other selected stations through the Earthscope Auto- When sufficient data are available, tomography can be mated Receiver Survey website [EARS]. We downloaded used to invert seismic parameters for Earth’s velocity these RFTNs for about 1500 stations in the western US, structure. sort them into bins by ray parameter and azimuth, and examine them for consistency and quality. Discarding the The science behind creating images of the structure of poor quality ones, we stacked receiver functions for the the Earth is well developed. Current practice, however, remaining 1400 stations across all azimuths to produce is to perform tomography with a single data type (either an average RFTN for each station in our region. This 748

Page 749

careful processing of the EARS RFTNs was presented at the as a function of ray crossing, with shallow -crustal- regions Seismological Society of America Annual meeting – Ref [2]. suffering from a lack of resolution due to the scarcity of ray In the current implementation of our inversion algorithm, crossing there. In comparison, gravity surveys have tradi- we can only use one station per cell, so we choose the best tionally offered optimal resolution at crustal depths. To in- RFTN closest to the center of each cell, resulting in only crease the usefulness of gravity data and to avoid mapping 482 RFTNs for the inversion target area of western US. We broad, possibly dynamically caused features in the gravity then decided to compute our own RFTNs for the study field into seismic speed variations, we explored high-pass region. We used all the TA stations for which stable receiv- gravity filtering. Filtered gravity anomalies are expected er-function estimates were possible, and included several to possess highest resolving power at short wavelengths portable deployments in Canada. We began with all events and thus enhance resolution of lithospheric structure. We (during station operation times) with magnitude mb > 5.7 modified the joint inversion code to include the filtering of recorded at epicentral distances from 30-100˚ and calcu- the gravity anomalies, as well as the corresponding grav- late receiver functions at 1096 stations using the time- ity partial derivatives, which has made the computer code domain iterative deconvolution algorithm. We obtained mathematically rigorous (Figure 1). Initial results with this over a million RFTNs which, once interpolated (see below), newly developed technique were accepted for an oral pre- provided us with the needed coverage. sentation at AGU Fall meeting – Ref [3], the most impor- tant conference in geophysics. As part of the systematic determination of earthquake moment tensors for North American earthquakes, col- TRADITIONAL APPROACH LANL APPROACH (only gravity data filter) (equalizing Ax=b) laborator Bob Herrmann (SLU) uses multiple filter analysis to obtain the fundamental mode Love and Rayleigh wave spectral amplitudes and group velocities. As of November 08, 2011, there were over 1227817 surface wave disper- sion measurements available [NATOMO]. We used Bob’s short-period observations and blended them with the intermediate and long-period observations from a global model from G. Ekström. We blended the data across the period range from 25 to 50 seconds, resulting in surface wave group velocities in the period range from 7 to 250 seconds. Gravity data from several space and land missions provide high-resolution coverage of the western US and are freely Figure 1. The effect on estimated shear velocity of filtering both available off the web [EGM2008]. These can be sampled the gravity and partial derivatives within the joint inversion code. at any regular or non-regular interval. To avoid mapping Results for two E-W cross sections (at 38.5°N and 28.5°N) are broad, possibly dynamically caused features in the western shown. Left column is the initial model. Gravity data is filtered in US gravity field into density and shear-wave speed varia- both the center and right column, but gravity partial derivatives are not filtered in the center column; they are filtered in the right tions, we high-pass wavenumber filtered the surface-grav- column. The results shown in the right column represent the ity measurements with a simple box-car filter with a width preferred methodology of four cells (~440 km). For data sets as large as those required for this project, Method and Code Development: Throughout the exten- combining data can require a laborious examination of the sion of this project, we have improved several aspects of receiver functions or sophisticated stacking approaches the inversion methodology and the associated software that may or may not be effective, depending on the ray- code. In order to process the millions of travel times avail- parameter and azimuthal coverage in the receiver-function able for the study area, we improved the efficiency of our collection. Anyone who has examined receiver functions body wave travel time predictions by replacing our finite carefully is aware that the signals are often variable and difference travel time predictor with a ray bending approxi- at times it is difficult to be sure that an underlying simple mation. Our ray bending is several orders of magnitude component of the signal exists. Small additions to even a faster to compute than finite difference with only small simple crust-mantle-transition-dominated receiver func- losses in accuracy. Given the errors associated with our tion can introduce vertical structural complexity that is as travel time data such losses are completely acceptable. intriguing as it is difficult to validate. Another issue that Body wave tomographic resolution increases with depth has been difficult to address is the difference in spatial 749

Page 750

sampling of the receiver-function and surface-wave data. ferent scales. We first applied it to Parkfield, a small - 15km Surface waves produce broad, low-wavenumber con- x 15km – region around the San Andreas Fault Observa- straints on lateral variations in the structure but receiver tory at Depth, California. The contributions of the different functions are sensitive to the higher wavenumber features data types to the inversion were controlled by the relative of the structure. For vertical resolution, the combination is weighting of the respective equations. We found that the ideal. However, when we look at how each signal averages trade-off between fitting the different data types, con- the lateral structure, the sensitivity difference is not ideal. trolled by the weighting, defines a clear optimal solution. To explore and solve these issues, we used our computed Preliminary results were accepted for an oral presentation TA RFTNs from the western US and constructed a receiver- at AGU Fall meeting – Ref [8], and the final model was pub- function wavefield interpolation scheme that helps to lished in Pure and Applied Geoph. – Ref [9]. equalize the lateral sensitivity of the receiver functions and We also jointly inverted 14 years of local earthquake body the surface-wave dispersion, and that greatly simplifies the wave arrival times from the Alaska Volcano Observatory complexity in the receiver functions (Figure 2). Details of catalog and Rayleigh wave dispersion curves based upon this novel technique were accepted for a presentation at ambient noise measurements for local seismic velocity SSA annual meeting – Ref [4], the most important confer- models and hypocentral locations at Akutan and Makushin ence in seismology, and it is under review for publication in Volcanoes, an area of approximately 35km x 50km in GRL, a high-impact journal in geophysics – Ref [5]. Alaska. We found that the velocity structure and relocated seismicity of both volcanoes are significantly more com- plex than many other volcanoes. Seismicity is distributed among several areas beneath or beyond the flanks of both volcanoes, illuminating a variety of volcanic and tectonic features. The velocity structures of the two volcanoes are exemplified by the presence of narrow high-velocity features in the near surface, indicating likely current or remnant pathways of magma to the surface. A single broad low-velocity region beneath each volcano is slightly offset from each summit and centered at approximately 7 km depth, indicating a potential magma chamber, where magma is stored over longer time periods (Figure 3). These Figure 2. Time slices from the western TA receiver-function results were accepted for a presentation at SSA annual wavefield. Each circle represents a seismic station at which meeting – Ref [10], and are under review for publication in we have computed a stacked receiver function. The color J. Geophys. Res. – Ref [11]. indicates the amplitude of the receiver-function stack (left) or the interpolated value of the receiver function wavefield (right) at a time of 3.8 s. Inset shows the stack of all receiver functions and the red line shows the location of the time slice, which was chosen to show the structure near the time of the Ps converted phase arrival from the crust-mantle transition region. Conventional seismic tomography methods typically produce smooth models, but these models may be inap- propriate when subsurface structure contains discontinu- ous features such as faults or fractures indicating that tomographic models should contain sharp boundaries. For this reason, we developed a novel tomography method that uses a modified total-variation regularization scheme incorporated with a priori information on interfaces to preserve sharp property contrasts and obtain accurate Figure 3. Cross sections of the V model and relocated seismicity inversion results. Our new approach was also accepted for through the caldera of Akutan Volcano and 5.5 km east of the a presentation at the SSA annual meeting – Ref [6], and it caldera of Makushin Volcano. Seismicity color indicates the year is under review for publication in GJI – Ref [7]. in which the earthquake occurred. The inverted triangle indicates the center of each caldera. No vertical exaggeration. Application cases: We successfully applied the new and modified joint inversion technique to several regions at dif- 750

Page 751

The region of interest, western US, was our last application ment, carbon sequestration, and used fuel disposition and of the method due to the large amount of data and size of salt repository. The expertise and methods developed here the area – 3000 km x 3000 km. Seismic wave speed varia- are transferable and DOE/Fossil Energy now funds us to tions in the resulting 3D model – Ref [5] - correlate strongly use joint inversion techniques to model the state of stress with expected geologic variations and illuminate broad around and away from the borehole. scale features of the western US crust and upper mantle. The model is smooth, self-consistent and provides a good References starting point for more detailed investigations (Figure 4).

1. Steck, L., M. Begnaud, W. S. Phillips, and R. Stead. Tomography of crustal P and S travel times across the western United States.
2. Steck, L., M. Maceira, R B. Herrmann, C. J. Ammon, and R. Stead. Crust and upper mantle structure of the western US from simultaneous inversion of surface wave dispersion, gravity, and receiver functions. 2012. In Seismological Society of America Annual Meeting. (San Diego, 17-19 April, 2012). , p. 380. El Cerrito: SSA.
3. Steck, L., M. Maceira, R. B. Herrmann, and C. Am- mon. Structure of the lithosphere and asthenosphere beneath the western US from simultaneous multi- parameter inversion. Presented at American Geophysi- cal Union Fall Meeting. (San Francisco, 9-13 December, 2013).
4. Maceira, M., C. Ammon, C. Chai, and R. B. Herrmann. Application of advanced multivariate inversion tech- Figure 4. Shear-velocity slices through the 3D model. Dark lines niques to the western U.S. 2014. In Seismological So- show Fenneman’s physiographic boundaries. Warm colors show ciety of America Annual Meeting. (Anchorage, 30 April relatively slower regions, cool colors indicate relatively faster regions. The anomalies show a combination of variations in - 2 May, 2014). Vol. 85, p. 502. El Cerrito: SSA. sedimentary basin and crustal thickness with lateral variations in
5. Chai, C., C. Ammon, M. Maceira, and R. Herrmann. shear-wave speed within the crust and mantle. Inverting interpolated receiver functions with surface- Impact on National Missions wave dispersion and gravity - Application to the west- ern US and adjacent Canada and Mexico. To appear in This research has enhanced the capability of the Labora- Geophysical Research Letters. tory through unique expertise in imaging Earth structure. This expertise has applications to Nonproliferation R&D
6. Lin, Y., E. Syracuse, M. Maceira, C. Larmat, and H. under Nuclear Nonproliferation, where accurate Earth Zhang. Joint inversion for Kilauea volcano with an models are needed to locate, identify and determine yield edge-preserving constraint. 2014. In Seismological So- for seismic events of interest such as underground nuclear ciety of America Annual Meeting. (Anchorage, 30 April explosions. In current operational practice, location accu-

• 2 May, 2014). Vol. 85, p. 502. El Cerrito: SSA. racy has reached a bounding limit, due to a lack of detailed models of the Earth’s crust. We are receiving additional

7. Lin, Y., E. Syracuse, M. Maceira, H. Zhang, and C. Lar- funds from GNDD R&D program to prove the efficiency mat. Double-difference travel-time tomography with of our new methodology and contribute to reducing that edge-preserving regularization and interface prior. To bounding limit. We were also granted a new proposal appear in Geophysical Journal International. through BAA jointly sponsored by DOE/NNSA/NA-22 to ap- ply our new advanced multivariate technique to regions of 8. Zhang, H., C. Thurber, M. Maceira, and P. Roux. Joint interest for monitoring. inversion of body-wave arrival times and surface-wave dispersion data for three-dimensional seismic veloc- High-resolution Earth models are also required for char- ity structure around SAFOD. Presented at American acterization and monitoring within several other areas of Geophysical Union Fall Meeting. (San Francisco, 9-13 LANL mission space including geothermal energy develop- December, 2013). 751

Page 752

9. Zhang, H., M. Maceira, P. Roux, and C. Thurber. Joint dispersion, gravity, and receiver functions. 2012. In inversion of body-wave arrival times and surface-wave Seismological Society of America Annual Meeting. (San dispersion for three-dimensional seismic structure Diego, 17-19 April, 2012). Vol. 83, p. 380. El Cerrito, California: BSSA. around SAFOD. 2014. Pure and Applied Geophysics. : DOI:10.1007/s00024. Steck, L., M. Maceira, R. Herrmann, and C. Ammon. Struc- ture of the lithosphere and asthenosphere beneath
10. Syracuse, E., M. Maceira, and H. Zhang. Joint inversion the western US from simultaneous multi-parameter in- of seismic and gravity data for velocity structure and version. Presented at American Geophysical Union Fall hypocentral locations at Akutan and Makushin volca- Meeting. (San Francisco, 3 - 7 December, 2012). noes. 2014. In Seismological Society of America Annual Meeting. (Anchorage, 29 April - 2 May, 2014). , p. 500. Syracuse, E., M. Maceira, H. Zhang, and C. Thurber. Seis- El Cerrito: SSA. micity and structure of Akutan and Makushin Vol- canoes, Alaska, using joint body- and surface-wave
11. Syracuse, E., M. Maceira, H. Zhang, and C. Thurber. tomography. To appear in Journal of Geophysical Re- Seismicity and structure of Akutan and Makushin search. Volcanoes, Alaska, using joint body- and surface-wave Syracuse, E., M. Maceira, and H. Zhang. Joint inversion tomography. To appear in Journal of Geophysical Re- of seismic and gravity data for velocity structure and search. hypocentral locations at Akutan and Makushin volca- noes. 2014. In Seismological Society of America Annual Publications Meeting. (Anchorage, 30 April - 2 May, 2014). Vol. 85, Chai, C., C. Ammon, M. Maceira, and R. Herrmann. Invert- p. 500. El Cerrito, California: SSA. ing interpolated receiver functions with surface-wave dispersion and gravity - Application to the western US Zhang, H., C. Thurber, M. Maceira, and P. Roux. Joint in- and adjacent Canada and Mexico. To appear in Geo- version of body-wave arrival times and surface-wave physical Research Letters. dispersion data for three-dimensional seismic veloc- ity structure around SAFOD. Presented at American Lin, Y., E. Syracuse, M. Maceira, C. Larmat, and H. Zhang. Geophysical Union Fall Meeting. (San Francisco , 9 -13, Joint inversion for Kilauea volcano with an edge-pre- December, 2013). serving constraint. 2014. In Seismological Society of America Annual Meeting. (Anchorage, 30 April- 2 May, Zhang, H., M. Maceira, P. Roux, and C. Thurber. Joint in- 2014). Vol. 85, p. 502. El Cerrito, California: SSA. version of body-wave arrival times and surface-wave dispersion for three-dimensional seismic structure Lin, Y., E. Syracuse, M. Maceira, H. Zhang, and C. Larmat. around SAFOD. 2014. Pure and Applied Geophysics. : Double-difference travel-time tomography with edge- DOI:10.1007/s00024. preserving regularization and interface prior. To appear in Geophysical Journal International. Maceira, M., C. Ammon, C. Chai, and R. B. Herrmann. Ap- plication of advanced multivariate inversion techniques to the western U.S. 2014. In Seismological Society of America Annual Meeting. (Anchorage, 30 April - 2 May, 2014). Vol. 85, p. 502. El Cerrito, California: SSA. Steck, L., M. Begnaud, S. Phillips, and R. Stead. Tomogra- phy of crustal P and S travel times across the western United States. 2011. Journal of Geophysical Research. 116 (B11304): doi:10.1029/2011JB008260. Steck, L., M. Maceira, R. B. Herrmann, and C. Ammon. Structure of the lithosphere and asthenosphere be- neath the western US from simultaneous multi-pa- rameter inversion. Presented at American Geophysical Union Fall Meeting. (San Francisco, 9-13 December, 2013). Steck, L., M. Maceira, R. Herrmann, C. Ammon, and R. Stead. Crust and upper mantle structure of the west- ern US from simultaneous inversion of surface wave 752

Page 753

Science of Signatures Exploratory Research Final Report Wide Field-of-View Plasma Spectrometer Ruth M. Skoug 20130564ER Abstract and space environment measurements for both scien- For 50 years, Los Alamos has flown plasma spectrom- tific and national security applications. eters for national security payloads as well as NASA Background and Research Objectives scientific missions. The space plasma environments measured by these payloads, such as the Earth’s magne- Throughout the space age, researchers have tried new tosphere, are tremendously complex, with large spatial ways to better measure and understand space plas- domains that can evolve within seconds or minutes. mas [e.g. 1], the bulk of which lie in an energy range of Since the dawn of the space age, researchers have several eV up to several 10s of keV with density ranging sought to measure and understand these plasmas, the from 0.01 to 100 ions or electrons per cubic centime- bulk of which lie in an energy range of several eV up to ter. One measurement technique employed for nearly several tens of keV. Nevertheless, numerous questions 50 years in space is electrostatic energy analysis us- concerning the nature and dynamics of space plasmas ing nested spherical or cylindrical plates with different remain unresolved, and breakthrough understanding will biases [e.g. 2; 3; 4]. The voltage difference between the require faster, higher accuracy plasma measurements. two plates generates a radial electric field, such that a charged particle of the appropriate charge, energy, and To address these needs, we have, through the Wide incident direction can pass between the plates and be Field-of-View Plasma Spectrometer LDRD-ER project, detected at the exit of the electrostatic analyzer. A full developed a fundamentally new type of space plasma energy measurement of the plasma is derived by step- spectrometer (named the 2PS Spectrometer, for its near- ping the applied voltage over a range that matches the ly 2-pi steradian field of view) with a wide field-of-view targeted plasma energies. We note that a typical detec- using fewer resources than traditional methods. The tor is a channel electron multiplier (CEM) or microchan- enabling component is analogous to a pinhole camera nel plate (MCP), both of which are routinely flown on having an electrostatic energy-angle filter at the image LANL space instruments. Los Alamos has flown this type plane. We have demonstrated the instrument concept of analyzer on both national security and NASA missions through a combination of electro-optic simulations and [e.g. 5; 6; 7]. laboratory measurements using a series of prototype instruments. We have begun studies to optimize the A weakness of a cylindrical or spherical section electro- instrument for use in space plasma applications. In par- static energy analyzer is its limited field-of-view (FOV). ticular, we have found that appropriate energy and angle These instruments measure a narrow slice of the plasma resolution can be obtained for very narrow filter plate distribution function at a given time. To fully characterize channels, and that the required voltage can be reduced the plasma, such an instrument is typically mounted so by appropriate shaping of the pinhole plate and filter that the fan is co-planar with the spacecraft spin vector, plate. We have begun to explore fabrication techniques and the plasma is fully sampled over one spin. However, that will permit development of such devices. plasma dynamics and the physical processes that drive them can occur on time scales much faster than a typical We are exploring funding for continued development of spacecraft spin of 10-20 sec, making time resolution the the 2PS instrument, both for basic NASA science mis- primary limitation of this type of space plasma spec- sions and for future national security payloads. We trometer. anticipate that the 2PS instrument concept will enable Los Alamos to continue to be a leader in space plasma The last high-impact innovation in plasma spectrometry 753