LASL Controlled Thermonuclear Research 1977

CORRECTION NOTICE (CPW-06, merged 2026-08-17): this document was mislabeled at ingest as a “Lockheed Martin CFR adiabatic compression patent.” It is actually the Los Alamos Scientific Laboratory (LASL) Controlled Thermonuclear Research annual progress report for 1977 (LA-7474-PR class). It contains NO patent numbers and no Lockheed Martin content.

Controlled Thermonuclear Research Program

January-December

REPRODUCTION COPY IS-4 REPORT SECTION

C3l

LA-7474”PR Progress Report

LASL

LOS ALAMOS SCIENTIFIC LABORATORY Los Alamos, New Mexico 87545 Post office Box 1663

The fourmost recentreportsinthisseries,

~ AffiimstiveAction/EquslOpportunityEmployex

unclassified,areIA-5656-PR, LA-6044-PR, LA4582-PR, and LA-7082-PR.

r.tus. completmrss. product. or Process disclosed. or rer.resents that It. u= would not mfrmm PW.MV

by the United S1.1.s Government. north. Umt.d Stiles Dtpattmenl empl.yce% nor lny of th,tr assume,

.s an account et work sponsored the Umled S1.1.s nor lny of the{. subco. tractors, or o,

Thiswork was supportedby the US Department

of Energy,Officeof Fusion Energy.

.ny )ccal hab,l, w or respo”sib,l, ty 1., y usefulness of any Inform.tie”.

OF ENERGY W-7408 -ENG. 36

DEPARTMENT CONTRACT

impl,.d. the .< CU”CY. .pp.

.mPloY. e,. makes .ny warranty,

Th,s tepott was ptep.md

Netthrr of %er~v.

thctr contz.ctwf.

UNITED

STATES

owned nghfs.

,X P,,CS o,

Controlled Thermonuclear Research Program

Compiled by

LASL

K. S. Thomas G. A. Sawyer

January-December

LA-7474-PR Progress Report UC-20 Issued: February 1979

This report presents the status of the Controlled Theaonuclear Research program at the Los Alamos Scientific Laboratory for calendar year 1977.

I. INTRODUCTION …

II. REVERSED-FIELD PINCH PROGRAM …

Introduction …

B. ZT-S …

c. ZT-40 … 10 References … 15

III. SCYLLAC FEEDBACK STABILIZATION EXPERIMENTS … 16

Introduction … 16

B. Experimental Arrangement … 16 c. Results … 19 D. Conclusions … 24 References … 25

IV. SCYLLA IV-P LINEAR THRTA-PINCH EXPERIMENTS… 26

A. Introduction … 26 B. Theta-Pinch Experimental Arrangement … 26 c. Experimental Results Without End Plugs … 26 D. End-Loss Calculations … 31 E. Experiments with Solid End Plugs … 32 References … 38

VI.

ABSTRACT

CONTENTS

STAGED THETA PINCH … 39

Introduction … 39

B. Thomson Scattering Apparatus … 39

c. Experimental Conditions … 40

D. Plasma Results … 40

E. Summary … 43

preferences… 43

FIELD-REVERSAL EXPERIMENT … 44

Introduction … 44

B. Diagnoatica … 44

c. Experimental Results … 47

D. Conclusions and Direction of Future Work … 54

References … 55

VII. TMPLOSION HEATING EXPERIMENT … 57

Introduction … 57

B. Summary …0 .

References … 60

EXPERIMENT …

… … … … … … … … … … … … … … … … … … … … … .

INJECTION EXPERIMENT …

Introduction …

Experimental Results …

Summary …

IX.

XI.

XII.

VIII.

FAST LINER

References

XIII .

XIV .

EXPERIMENTAL PLASMA PHYSICS …

Summary …

B. Plaems AC Resistivity Studies …,…

References …

HIGH-DENSITY Z-PINCH …

Introduction …

B. Description of Experiment …

c. Prototype Experiments …

Reference …

PLASMA DIAGNOSTICS …

Introduction …

Scanning Infrsred Heterodyne Interferometry …

Long-Pulse Ruby Laser for Holographic Interferometry …

Portable Thomson Scattering …

Space Resolved Thomson Scattering …

Computer Codes …

Laser-Plasma Interaction Studies …

Future Diagnostic Developments …

References …

TREORY …,…,…

Introduction …

B. MHD Studies …,…

c. Kinetic Effects …

D. Magnetoacoustic Heating …

E. Internsl Ring End Stopper for Open-Ended Devices …

)?. Numerical Simulation …

G. Fast Liner Dynamics …

H. Translations of Russian Reports and Preprints …

References …

COMPUTERS …,…

Introduction …

B. CTRUser Service Center …

c. Automated Data Processing and Computer Control …

102

106

109

111

113

XVI . SYSTEMS STUDIES … 130 Introduction … 130 Linear Theta-Pinch Reactor (LTPR) Studies … 130 Reversed-Field Pinch Reactor (RFPR) Studies … 132 Fast-Liner Reactor (FLR) Studies … 139 Other Non-Mainline Reactor Studies … 143 Hybrid (Fusion/Fiaaion) Economics Studies … 146 Synthetic Fuel Production by Fusion Power … 147 Plasma Systems Analysia … 149 Fusion Reactor Nucleonics … 156 Insulator and Ceramics Research … 162

References …

MAGNETIC ENERGY TRANSFER AND STORAGE … 117

Summary …

7-T, 20-MJ Superconducting TPFS Coil … Bipolar TPFS Coil Simulation Testa with the 300-kJ METS Coils … Power Supplies for TPFS …

TNS Studies …

TPFS Switch Development and HVDC Teat Facility …

Critical-Current Studies and Loss Measurements …

300-kJ, 10-kAMETS Storage Coils …

One-Component Module Superconducting Prototype METS System …

Homopolar Machine Development Program … Miscellaneous …

References … 129

xv.

XVII . ENGINEERING …

ZT-40 System Design …-…-…

B. ZT-40 Circuit-Plasma Simulation …

c. Component Development …

D. ZT-40 Engineering Prototype …

E. ZT-40 Machine Control and Data Acquisition Facilities …

F. Pulse-Power Technology and Component Development …

G. Advanced Pulse-Power Technology …

References …

XVIII. TRITIUM SYSTEM TEST ASSEMBLY …

Introduction …

B. TSTA Function …

c. Project Interface with the U.S. Fusion Program …

Reference …

PUBLICATIONS …

126

125

128

118

188

187

186

190

177

170

172

experiments continued on

the

point of view.

can operate over

IV-P linear theta pinch.

feedback system

The

Reversed-Field

end-plug end-stoppering experiments designed to

eliminate axial particle flow.

experiment. Efforts concentrated on the measure-

Activities in the Staged Theta Pinch concen-

ment of the radiation loss from impurities and on

trated on constructing, checking out, and col-

raising the electron temperature by operating at a

lecting data with a Thomson scattering apparatus.

lower filling pressure and by the use of low-

Data were collected at two axial positions, one

temperature discharge cleaning.

near the center of the main compression coil and

the 40-cm-bore ZT-40 experiment was begun. ZT-40

one near the end. At the center of the coil, data

has been designed to have great versatility and

were taken at three radial positions; at the end,

wide

data were taken at four radial positions.

(2.5 PS-1 ins). It will allow the study of the

surements were made as a function of time with

high-beta programmed reversed-field profiles as

initial D2 fills from 3 to 15 mtorr.

well as the generally lower beta configurations

Two small experiments were used to study the

obtained by self-reversal.

stability and confinement properties of the re-

properties of the high-beta equilibrium can be

versed-field configuration formed in a theta pinch

demonstrated, efforts will be made to determine

when a reversed bias field is used. This configu-

the optimum method for setting up and maintaining

unexpectedly favorable

the configurations from the reactor technology

characteristics which are of interest for in-

creasing energy confinement in open-ended systems

The final experiments on the Scyllac feedback

such as mirrors, theta pinches, and liners.

stabilization project have been completed, and the

Experiments on the Implosion Heating Experi-

experiments have been terminated. The last exper-

ment were ended in November 1977,and the personnel

iments differed from previous efforts in several

involved transferred to the Fast Liner Program.

important aspects. First, the equilibrium config-

The first part of the year was spent finishing a

uration was changed from the !.0,1 combination of stellarator fields to 9,q 1,2 stellarator fields.

detailed measurement of the magnetic field as a

function of time and radius using magnetic probes

Secondly, the helical plasma perturbation was much

and Faraday rotation. The major part of the year

larger on these final experiments than on the pre-

was spent taking temperature and density data

vious Scyllac experiments giving a plasma helical

using Thomson scattering at four radial positions.

radius of about 3 cm vs the value of 1.4 cm used

In the fall of 1977, approval was given to

in the earlier experiments.

begin a Fast Liner experimental program.

of the fields was increased from 0.62 m to 1.0 m.

experiments on the imploding liner will use three

experiments was also different from the previous

preparation studies will be conducted on the Gun

the

the Scyllac bank.

cases in that the output of the position detectors

was processed in such a manner as to drive the

Injection

power amplifiers in response to the model struc-

The initial measurements

ture Of the plasma instability rather than the

were of the properties of the gun plasma itself.

displacements as seen by the individual detectors.

These were followed by an investigation of plasma

Experiments continued in the 5-m-long Scylla

injection into a short (50-cm-long) solenoidal guide field having a typical strength of about

the

Pinch

shows

ration

Program,

Construction of

15-cm-bore ZT-S

I. INTRODUCTION

Once the desirable

range of risetimes

Ii.Dreicer, G. A. Sawyer, K. S. Thomas

capacitor racks of

throughout the year.

Injection Experiment.

Also the wavelength

final series of

high-frequency

The main experiments

primary

10 kG.

The

Gun

The

Mea-

stability

Experimental

resistivity,

Experiment

operated

Initial

Plasma

plasma

the

were an investigation of the characteristic end-

10ss times, study of the plasma flow processes

near the theta-pinch coil ends, and material

Plasma Physics Group are directed ‘toward the study

efforts

plasma

and

the

tus

was begun.

ZT-40 experiment.

CTR program at LASL.

the theory group.

device control.

scientific

computers.

Areas of

activity

used

for

heating, and heat flow in the presence of plasma

control and data acquisition system was completed

turbulence, electron drifts, and other effects

and the Scylla IV-P operating system was upgraded.

likely to be present in plasmas of current fusion

During the year the goal of the 14agnetio

Interest. Experimental studies, which were being

Energy Transfer and Storage program changed from

done on a Q-machine, were temporarily suspended on

development of the superconducting magnetic trans-

May 25, 19’77,to allow dismantling of the equip-

system

ment to make room for the construction of the

theta-pinch experiments to development of super-

Relocation of the experiment

conducting tokamak poloidal field systems.

into a new area was begun in late December, and it

emphasis in this program is directed toward future

is planned to be in operation by June 1978.

large tokamak experiments.

An experimental program was set up in May

The focus of all systems and design studies

1977, to investigate a very high density Z-pinch

of magnetically confined fusion concepts. was on

initiated by a small-diameter laser beam.

exploratory concepts.

1977, a prototype of one module of the power

design effort on the Reversed-Field Pinch Reaotor

supply was built and brought into operation.

concept waa completed, based upon a moderately

Studies of plasma initiation using a ruby laser

pulsed mode of operation.

will begin in early 1978.

second design study was begun that

The Plasma Diagnostic group continued work on

the

the spztlally resolved Thomson scattering appara-

associated with a larger, extended-burn operating

scanning

mode. Studies of both the Fast Liner Reactor and

interferometer. Work on several new diagnostics

Liner Magnetic Fusion concepts emphasized phyaica

Some general-purpose computer codes

and energy-balance constraints for fusion power,

were developed and a major effort continued on the

although crucial technological issues for both the

use of minicomputers for data acquisition and

concepts were addressed.

bicusp confinement (Tormac) was begun whioh fo-

The work of the theory group is closely

cused on elucidating viable physics operating

associated with and motivated by the experimental

points with total power, energy balance, and

first-wall loading being enforced as major con-

problems and requirements of the Scyllac experi-

Scoping studies of generalized hybrid

ment provided the major motivation for research in

program has ended and the CTR effort at LASL has

begun. Work on neutronics studies and insulator

However, now that the Scyllac

hydrogen) applications of fusion power were also

become directed toward alternate concepts, the

variety of theoretical research has increased.

The main activity of the engineering groups

inoluded MHD

development, design,

numerical simulations, and studies of kinetic

components for the new ZT-40 experiment.

effects and magnetoacousti.cheating.

areas of activity included the design of

The shortage of computer resources at LASL,

coupled with the increased utility of the PDP-10

switch life testing, and design and testing of

and MFE network, has made the USC facility a vital

advanced components which may be needed in future

link in the LASL MEW effort. The pDP-10 has been

problems,

The Tritium Systems Test Assembly, which was

accounting, and inventory control.

initiated in February 1977, is dedicated to the

large-particle simulation and MHD stability codes

development, demonstration, and

have been shifted to the network from the LASL

technologies related to the deuterium-tritium fuel

Preliminary

cycle for fusion reactor systems.

fer

and

During

storage

infrared

design, a

heterodyne

alternative or

investigated both

For several years the

research continued.

(fusion/fission)

high-density Z—pinch,

administrative

calculations,

experiments.

Many of the

straints.

was the

design

ZT-4O

the

and

The

use

for

future

A major

toroidal

technology

A study of

physios and

On the basis of this

(thermochemical

interfacing of

testing of

capacitor

ignition

synfuel

Other

and

the

and

D. A. Baker, G. P. Boicourt, C. J. Buchenauer, R. T. Buck, A. Buffs (Visiting Scientist, University dl Padova, Padova, Italy), L. C. Burkhardt, G. 1. Chandler, J. N. Di Marco, R. S. Dike, J. N. Downing, Jr., P. R. Forman, A. Haberstich, C. F. Hammer, K. W. Hanks, L. D. Hansborough, R. B, Howell, A. R. Jacobson, F. C, Jahoda, R, W. Kewish, Jr., R. Kristal, K. J. Kutac, J. W.,Lillberg, E. M. Little, M. D. Machalek, J. G. Melton, W. C. Nunnally, A. E. Schofield, K. S. Thomas, L. C. Wilkerson, R. W. Wilkins, and P. C. T, Van tierLaan

The

INTRODUCTION

compact torus.

prOfi.kW as we~

General.

confinement

desirable

high-beta

midpoint

filling

lower

ZT-S

time

the

its

and

effects.

experiment

the 40-cm-bore

II. REVERSED-FIELD PINCH PROGRAM

The scheduled date for

vacuum ultraviolet

temperatures for

reversed-field

high-beta

schedule.

nitrogen.

smaller

1000 A.

oxygen

ZT-S

Once

with

the

use

the

The

The measured electron temperatures are

toroidal reversed-field pinch program

low and field diffusion rates are high.

continues to be motivated by its promise of

candidate for keeping the temperature down is the

high-beta stable equilibrium configuration with

energy loss from impurity line radiation.

the possibility of ohmic heating to ignition with

general objective this past year have been to (1)

moderate magnetic fields. Further advantages stem

determine the nature and magnitude of the radia-

from the freedom of choice of aspect ratio that

tion loss, (2) operate at lower filling pressures

gives an escape from possible adverse trapped-

in an attempt to raise electron temperatures, (3)

particle effects and can lead to a simpler modular

design and test a discharge cleaning scheme, (4)

construction with easier access as compared to a

design, build, and test a power crowbar circuit to

lengthen useful time of the field.

The LASL program is currently centered on the

February 1978,the plan is to then upgrade the ZT-S

15-cm-bore ZT-S experiment and

cleanliness,

ZT-40 experiment presently under construction. In

programming modes, and to incorporate improved

1977, efforts on the ZT-S experiment have been

diagnostics. The experiment is to be used to give

concentrated on the measurement of the radiation

as much Information as possible in the areas of

loss from Impurities and in the study of possible

cleaner systems, impurities, discharge cleaning of

raising of electron temperature by operating a

alumina tori, and, if possible, programming at

pressure

elevated electron temperatures. This Information

low-temperature discharge cleaning.

is to give guidance for the LASL ZT-40 experiment

The ZT-40 experiment is presently about at

currently under construction.

construction

2 . Radiation Loss Measurements.

experiment has been designed to have great versa-

Nature and Level of Impurities. Spectro-

tility and can operate over a tide range of

scopic and mass analyzer measurements of the ZT-S

riaetimes (2.5 ps-1 ins). It will allow the study

discharge have shown that the main impurity ia

programmed

amounts

as the generally 10Wer beta

The impurity number density is in the

configurations obtained by self-reversal.

range from . 0.3$ to . 1% of the deukerium fill.

equilibrium can be demonstrated, efforts will

losses that can cause limitations on the electron

properties

Theoretical calculations’ predict serious energy

to determine the optimum method for setting up and

current

maintaining the configurations from the reactor

studied (JTotal = 1-3 kA cm2).

technology point of view.

beginning physics on ZT-40 is July 1979.

b. Absolute Measurements of the Radiation

Method. A survey of the impurity lines

was made spectroscopically to determine the spec-

A chief factor limiting the

tral range of the line radiation emitted from the

the

ZT-S plasma. The wavelength region containing the

experiment after favorable plasma-field profiles

substantial portion of radiation is between 200-

are produced is the degeneration of the stability

fast system that was sensitive to

of the configuration with time due to resistive

radiation

The

One

ease

changing

Starting in

assembled,

densities

carbon

being

was

and

tested, and arranged to view the ZT-S pinch

that

A schematic of the detector

largest uncertainty, about a factor of two,

arrangement is shown in Fig. II-1. Basically the

detecting system consists of a fast, linear photo-

The results indicate that,for sharp boundary

dlode of wide dynamic range (ITT FW 128).

plasma with a flat emission profile, a radiative

detector is sensitive to visible wavelengths and

loss of 5 to 8 MW per meter length is obtained

is used with a sodium salicylate phosphor wave-

during the - 30-ps stable time of the discharge.

intercepts

The total energy balance for a sample discharge is

emerging from the plasma. The fluorescence emis-

The two curves show the

sion spectrum of the phosphor varies by a factor

energy input and energy content per meter of

of 1.45 from 2500 A to 1200 A and then is flat

length (plasma + field energy) of the discharge as

The fluorescence spectrum centers

function of time as deduced from current,

at . 4200 A,well within the high-sensitivity range

voltage, and magnetic probe data. For comparison,

of the photodiode. This was the primary detection

the difference between the two curves, which

system used for results reported below.

represents the energy loss, is shown in Fig. H-3

The radiation from the plasma in the ZT-S

along with the radiation leas as deduced from the

discharge passes through the aperture in the alu-

above detector measurements.

mina plug inserted Into the pump port and then

loss curve is compared with the calculated ohmio

through two successive I-cm-diam apertures to the

heating for Te . 15 eV in Fig. 11-4. These curves

detector. The last two apertures are covered with

indicate that essentially all of the loss in due

a wire mesh grid to improve the detector shielding

to the radiation. In view of the factor of - 2 in

againat electrical noise.

uncertainty in the absolute radiation measurement

detector i.a amplified and then transmitted to an

mentioned earlier, this agreement may be fortu-

The analysis and interpreta-

the radiation contributes in a significant way to

itous. The measurement does indicate however that

that

length

shifter

down to 200 A.

through a pump port.

oscilloscope display.

corrections, (c)

(?) Results.

TmLm A j- _

Comparisons

energy.

factors

several

single

port,

with

the

show

This

photons

distributions

this determination.

shown in Fig. II-2.

The signal from the

the energy loss.

calibration

assumed

source

about

1’ ~.

the

into

this

introduces

The same radiation

TOTALINPUT ENERGY

SHOT 824706 I

tion of the results of the measurements involve

(a) the

detector, (b) the source and detailed solid angle

the determination of mean photon

Since the measurement was made through a

distribution are needed to make determinations of

the absolute value of the radiative energy loss.

assumptions

different

\ ”-‘EazEEIT

Fig. II-1. schematic of radiation detector.

Fig. II-2. Curves of energy input and content vs time in ZT-S.

TIME-psec (AFTER ONSET OF IZ]

TOTALENERGY LOSS

results are comparable to

RADIATIONLOSS

TIME-psec (AFTERONSET OF Iz)

Fiz. 11-3. Curves comvarine measu~~ed ra~~iation loss with that deduced from en~rgy balance shown in Fig. II-2.

150-

; z U -1 e E 100.- u s! a W & ~ -1 s . c!)so— In cj

This agreement is

For plasma

: u E

better

determining the ionization, excitation, and emis-

The conclusions of these studies are as

follows. For ZT-S operating at I . 60 kA and a

filling pressure of 20 mtorr, (a) the radiation

loss is ~ 50% of the energy loss in ZT-S, (b) the

radiation is primarily in the vacuum uv with a

mean energy of - 30 eV ( A.- 400A),

oxygen

theory.

SHOT’6A2470t

RAOfATIONLOSS

SliOT# 24706

sion processes.

impurities are

tively minor at late

OHMICHEATING aAsEO ON

The radiation

about 75 W.

10 mtorr.

15-20 eV,

and the

of 4 cm.

used h

Te.15eV

than expected in view Of

and

Good

Thom-

carbon,

(d) these

those predicted by

This problem has now

time in the discharges

The open circles are

The solid

35 kA/vs.

gives

The

Electron Temperature Measurement.

son scattering measurements of electron tempera-

ture at filling pressures of 18 and 34 mtorr have

been described in the previous LASL Controlled

Thermonuclear Research annual report.1 Measure-

ments at lower densities were then hampered by

instrumental scattering.

been resolved by the addition of two 6943-A notch

filters to the scattered spectrum detector. The

width at half maximum of each filter is 10.2 A.

The overall sensitivity of the detector Is reduced

by a factor of 2, whereas the instrumental scat-

tering signal is down by a factor of 15.

data have been obtained at filling pressures of

Data acquisition at 5 mtorr is limited

by the stability of the pinch rather than by the

sensitivity of the diagnostic.

The temperature measurements shown in Fig.

II-5 have been obtained with a filling pressure of

D2 of 10 mtorr, and a peak toroidal current of

60 kA with an initial rise of

initial bias field was set at 0.064 T and the

toroi.dalfield at the wall reversed to -0.079 T in

The postimplosion plasma radius as

estimated from pressure profiles was of the order

The square data points have been obtained

with a polychromator dispersion of 33 ~/mm and the

circles with a dispersion of 66A/mm.

circles are the result of a least squares fit to 7’

polychromator channels.

based on the came data, but using the two closest

channels only on each side of the 6943-A channel.

The difference between the two analyses is rela-

(> 10 P.S). At early times, however, the analysis 10Wer based

points

4 data

much

TIME-)iSeC(AFTER ONSET OF Izl

Fig. II-4. Curves comparing the measured radiation loss with the calculated ohmic input.

temperatures of

estimated for ZT-S from Thomson scattering and

spectral impurity ionization state measurements, a

calculation of the theoretical radiation loss rate

gives values of 8-10 MW per meter.

10SS rates were computed for asswed values of Te

using Ortolani!s POWRAD code2 based on McWhirter9s radiation loss equation3 with a multiplying factor Of 5 as suggested by the Maryland Group results.4

the uncertainties in the measurements

approximations made in the functions

!i2i!

TIME [~c)

l.! 40 % s ~ 30

: 10 . -1 ‘“o

~ 20 - g

of 27 eV at 25 VS.

temperature

the 7-point

of the main

at 25 US.

analysis.

Figure

“o

{

1.)

the

,. 3.0 cm

four.

r. 1.5cm

simple

Comparison

charge vessel.

3 .’ Motivation.

group at Grenoble

current I e windings, in this

current 1P Is the

the feedplates for

breakdown would

the first 150 us.

baking required.

ASSUMing

forms a

is the

not

of this result

Te measurement indicates an increase of the elec-

tron temperature at lower filling pressures. The

peak values of Te reported earlier’ were 12 eV at

34 mtorr and 15 eV at 18 mtorr.

Dischar$ze Cleaning Studies.

The radiation loss studies

indicate the need for determining if suitable di.a-

charge cleaning methods can be developed in order

to reduce the impurity level to acceptable limite.

A preliminary investigation was made to determine

the usefulness of the low-temperature discharge

technique5 for oxygen removal from an alumina dis-

Previous work in this area has

been limited to metal-walled systems.

known, however, from the results of the Petula

that the replacement of a

tungsten limiter with an alumina one can reduce

the oxygen in the discharge a factor of three to

low- temperature discharge

cleaning proves successful, it is planned to use

discharge cleaning pulses on ZT-40 prior to the

main discharge, thus minimizing the amount of

b. Discharae Cleaninu Circuit.

charge cleaning circuit consists of capaoitor and

switching circuit capable of delivering - 110 J

per pulse to a series connection of the toroidal

and poloidal field circuits.

2-kA screw pinch and is capable of

producing one to five pulses per second.

circuit can be quickly removed from the experiment

by the use of a pair of pneumatically operated

switches thus minimizing the time between the

cleanup discharges and the normal firing of ZT-S.

c. Discharue Characteristic.

forms for the discharge cleaning oircuit are shown

in Fig. II-6 for the low-pressure

voltage shown in the figure is the voltage across

It is

with the previous

This arrangement

these lowest

occur.

case.

wave

dis-

The

one quadrant of the torus. The

toroidal plasma current.

current in the toroldal field

case breakdown is delayed for

Plasma was generated over a range of 0.7 to

40 mtorr, 0.7 mtorr being a limit below which

TIME l~cl

l .

Fig. 11-5. Thomson scattering measurement of electron temper- ature at (a) O-, (b) 1.5-,(c) 3.0-,and (d) 4.5-cm radius. filling Peak toroidal current 60 kA, pressure 10

mtorr of D2.

and, perhaps, a better representation

body of the electron distribution than

II-5(a) shows a peak central value of

Te of 25 eV,which then decays gradually to 15 eV

The electron temperature at 1.5- and

s-cm radii levels off at about 15 eV.

full preionization, conservation of particles, and

a poloidal B of 0.6, one would expect a value of

Ti + Te of approximately 60 eV at peak current and

5op3/dw---

PRESSURE (tTIT)

‘vl7----

Fig. II-6. Voltage, plasma current, and input toroidal field current wave forms durLng a discharge cleaning pulse in ZT-S.

Fig. 11-7, Plasma energy deposition per pulse with probes in (0), out (V), and an independent check (0) of the probe~ut case.

4mT

the current feed terminals.

to contact the wall.

shown in Fig. 11-7.

using Dc 4861 A.

‘A

ing the discharge cleaning.

When B+ goes

19, 20, and 28.

crease In D20.

levels off.

Using the

The

A mass analyzer is

The ambiguity

The behavior

systems,

The

pressures uv radiation was required to initiate

and the background pressure is normally about

breakdown, and then substantial plasma currents

4 x 10-6 torr in the torus.

failed to flow during the first half-cycle.

attached to the pumping system just above the

Also from Fig. II-6 a dc offset can be seen

pumps so that it samples the gases pumped out dur-

in the plasma current lP. This iS due to traPPing

of the poloidal flux in the plasma, and leads

Results. The mass analyzer has been used

thereby to a time variation of q.

to observe the gases that are pumped from the

through zero, q is zero. When this happens there

torus before and during discharge cleaning.

is probably an instability and/or a rapid expan-

principal mass numbers observed are 14, 16, 18,

sion of the plasma radius which causes the plasma

These correspond to N, O or CH4,

through zero shows up as a voltage spike aoross

The result of q’s going

H20, HDO, D209 and CO or N2. between D20 and CD4 was resolved by comparing

results between hydrogen and deuterium discharges.

The energy input for different pressures is

The mass analyzer was used to monitor the D20

line during the discharge cleaning.

The electron temperature of the plasma during

of D20 during and after discharge cleaning is

discharge cleaning was measured by two methods.

shown in Fig. II-8 for 1, 10, and 20 mtorr. Mith

First the resistivity of the plasma was calculated

the turn-on of discharge cleaning there is an in-

from the voltage and current traces.

The amplitude of this signal de-

Spitzer equation and obtaining the plasma radius

creases rapidly in the first few minutes and then

from the streak pictures we find . 2-3 eV for the

When the discharge cleaning is ter-

filling pressures of 1, 10, and 20 mtorr.

minated, the D20 decreases immediately. When the

Second method was a line to continuum measurement

gas flow is terminated the D20 further decreases

From this we found 3.0 eV at

usually below the initial level present.

10 mtorr and 2.o eV at 1 mtorr.

These data indicate that, as previously ob-

d. VacuurnSvstem. The torus is pumped by a turbomolecularpump and an ion pump through a long

served by R. J. Taylor in metal-wall

oxygen is being converted to water by the low-tem-

Pipe and five 1.4-cm-diam tabulations. The con-

perature discharge and is then pumped out.

ductance for this system is . 17 g/s at the torus

f. Conclusions.

of the RFP discharge.

mass 28 (CO or N2).

ZT-S Upgrade.

of ZT-S.

cation.

I@ m tot,

O-ring seals.

impurity studies.

This was done without

A new segmented alumina

technology will be

for this purpose.

Uae of vacuum

operation has

charge.

ZT-S

The

new, oil-free, vacuum ayatem and the effeotive

pumptng speed will be increased from 17 to 60 kls..

The gas fill system will be entirely rebuilt to

permit accurate mixing of a variety of ga8es for

The reduction in the number of torus O-rings

will be made possible by the use of a glass seal-

ing technology developed at LASL for ZT-40. Four-

degree ceramic segments will be sealed in sets of

five and six at a temperature of 1260”C.

surfaces of the remaining olrcumferential ~ointa

will be coated with a thin layer of the same

glazing material to help ensure hermeticity of the

Viewing and pumping sections will be made of

quartz. This mixing of alumina and quartz torus

segments and the ensuing mismatch of expanaion

coefficients can be tolerated in ZT-S because of

its small size, and since there is no plan to bake

the discharge torus in situ.

The present 14-mm-diam pump perts (3 of whioh are either closed off or occupied by q agnetio

probes) will be replaced by a single 44.5-mm-dlam

side tabulation. Measurements on a replica of the

primary conductor indicate a toroidal field error

of - 35% for a 57.5-mm-diam hole. TO reduce this

perturbation, the new pump port will be fitted

with a movable conducting plug which will be moved

into position just prior to the firing of the dis-

One of the benefits of the improved vacuum

pressure In the discharge chamber.

is important since it affects the rate at which

layers of impurity build up at the wall of the

torus during the interval between gas fill and

firing of the experiment.

The base pressure in the present configura-

tion has been determined indirectly by interpreta-

tion of remote pressure gauge readings. A computer

code solving 20 coupled equations has been written

Known values of oonduotance,

pumping, and outgassing rates are entered in the

program and unknown values of these parameters are

adjusted to give the best possible flt with

The

Proceeding in this

This pressure

reduction

base

the

TIME IMINUIESI

Fig. II-8. The time history of D20 during a discharge clean- ing proceaa aa measured by a maas spectrometer for 3 different gas fill pressures.

(1) Water is generated by the 2- to 5-eV

plasma and is pumped from the system as has been

observed in metal tori.

overheating the polyethylene electrical insulation

(2) Five hours of discharge cleaning reduced

the outgassing rate a factor of two or more.

(3) In addition to water the main impurity

observed on the mass spectrograph was found to be

(4) A simple capacitive discharge circuit

allowing immediate switching between discharge

cleaning and normal high-voltage

been demonstrated to be practical for this appli-

a. Torus and Vacuum System.

experiment is scheduled to undergo major modifica-

tion starting In February 1978. The main goal of

this effort is to significantly improve the purity

torus will be installed with only 24 instead of

the present 90 Viton O-rings.

grease during the aasembly will not be permitted.

available pressure readings.

The present pumping stand will be replaced by a

fashion a base pressure of 4 x 10-6 torr has been

calculated for the present ZT-S configuration.

After adjustment of the codes for the new pumping

using only one laser, by means of Bragg diffrac-

configuration and the reduced number of O-rings,a

tion heterodyne techniques.

base pressure of 6 x 10-7 torr is predicted for

is in quadrature’ so that an accurate, unambigu-

the new torus, a factor of 6.5 lower then obtained

ous, and uniform record of interferometric phase

is obtained in each channel.

b. ImDroved Circuitry. Another facet of the

The optical layout is shown in Fig. 11-10.

ZT-S upgrade will be the use of a power crowbar

The scheme is basically a Mach-Zehnder arrangement

system, which together with a modification of the

with the acousto.-optic cell serving as a beam-

toroidal field banks, will allow more flexibility

splitter and with each of the five scene beams

in the setting up of the RFP discharges.

each traversing the plasma twice.

power crowbar has recently been installed and

not only splits the scene and reference beams but

typical current waveforms with passive crowbar

also shifts the reference beamsv infrared frequen-

only and with both passive and power crowbars are

cy by 40 MHz with respect to the frequency of the

Therefore ‘the

c. Improved Diagnostics. Whereas new Fara-

contains an oscillating component at 40 MHz due to

day rotation techniques are under development for

interference between a reference beam and its

ZT-40, the ZT-S experiment will have to rely on

corresponding scene beam. The plasma refractivity

internal magnetic probes for the measurement of

along the scene paths then phase-modulates this

magnetic field distributions. These data are of

utmost importance since they form the basis of our

The modulatlonldemodulation electronics is

diffusion, energy content, and ideal and resistive

All five receivers derive

MHD stability calculations.

their local oscillator signals from the 40-MHz

The present magnetic probes have a rather

oscillator which also drives the Bragg cell via a

outer diameter of 10 mm.

In Fig. 11-11 only one receiver

will be available on ZT-S with a jacket diameter

is shown; all five are identical.

of the order of 4 mm. The integrity of the vacuum

must be driven in burst mode because the required

will be ensured by the use of newly designed

RF power for reasonable diffraction efficiency in

bellow-sealed probe mechanisms.

the infrared is more than can be safely applied CW

The electron density distribution in ZT-S

to the cell. The quadrature signals sin+ and COS$

will be measured with a multichord interferometer

will be digitized by ADC~s and analyzed unambigu-

operating at a wavelength of 3.9 pm.

ously and sensitively by computer, facilitating

acquire complete data for Abel inversion on each

the calculation and display of radial electron

large

previously.

shown in Fig. II-9.

ously.

It CURRENT WITH PASSIVE CROWSAR

The

scene

beams.

Smaller probes

40-MHz carrier.

burst amplifier.

shown in Fig. 11-11.

Iz CURRENT ANO POWER CROWBAR

In order to

I 45*

power

detected

The Bragg cell

The phase detection

‘“K’P%“‘F7

I 45A

L o

Fig. II-9. Comparison of ZT-S current waveforms with and without power crowbar.

Fig. 11-10.

Schematic of the five-channel interferometer.

discharge, the Integrated electron density will be

measured along five parallel chords simultane-

A device has been developed to do this

Pw#xllo

the

h As

%AS !

charge.

at 85 MHz

quired due to

operation, Ls used.

modulated at 40 MHz.

111!1 40 MHz 093LLAKR

passband of

passband.

subtract

computer

receiver

the

The ZT-40 experiment is the

next step up from ZT-.Sin the Rsverae-FieLd Pinch

The initial objective of ZT-40

will be to demonstrate confinement time scaling

from the results of ZT-I and ZT-S.

confinement time scales as the square of the minor

bore, ZT-40 should be stable for at least 150 to

A second objeotive will be to operate in

a regime clean enough to overcome the impurity

radiation barrier and raise the temperature to Ti

used to determine the ranges of the current ;nd

electron densities which allow the production of

stable configurations. Another ob~ective will be

to extend the duration of the current to milli-

seconds and operate with reduced rates of rise of

current in order to establish the time scale over

which control of the stability can be effected by

programming of external currents.

will be used to investigate the self-reversal mode

of operation. A general goal of the project is to

determine which of the potential advantages of the

reversed-field pinch concept can be brought to a

practical realization with a direct application to

an energy-producing fusion reactor.

experiments on ZT-40 will begin in the calendar

J (5mai)

room-

ZT-40

200 Us.

strips;

Fd4crrm

however,

(XX+ (5 ACH)

sensitivity of

Objectives.

program at LASL.

= Te - 100 to 200 eV.

solution consists of

P . Description.

Random variations

The technique

year 1979.

plasma

light

from

Very

the

The

projected

If the

Finally ZT-40

In addition ZT-40 will be

The current risetime is

The ZT-40

plasma

1 MS)

first

that

the

The

can

ZT-40 has a 40-cm minor-

bore torus with plasma aspect ratio of 5.7.

torus is fed at. 12 points and has 12 magnetic

cores equally spaced about the torus.

facility ia shown in Fig. 11-12 and a sketch of

the front end is shown in Fig. II-13. The number

of feedplates was determined by the requirement

that the initial risetime of the current be the

same as ZT-S (2.5 HS).

lengthened to - 20 PS by removing feedplates and

shorting the respective toroidal feed points, and

by series-paralleling the poloidal field windings.

slow operation (40 US

attained by removing all feedplates and inducing

the toroidal current by means of toroidal windings

on the outside of the primary or by energizing the

field only with the power crowbar banka.

Fig. 11-11. Block diagram of the interferometer electronics.

density profiles within seconds of the plasma dis-

Several milliwatts of laser power are re-

temperature InAs photodiodes to 3.39-pm radiation

To achieve this, a Spectra

Physics model 125 gas lacer, modified for 3.39-pm

The longitudinal mode beats

are not detectable with the 10 MHz

photometric mixing on the detectors translates

this 85-MHz beat down to 45 MHz, within the

low

40-MHz

The

removing the mirrors from both ends of the model

125 laaer, and of using it as an amplifier.

laser is excited by a short oscillator whose mode

beat is Ln the hundreds of megahertz.

The apparatus has been tested successfully on

the bench with a resolution of ~ 1/50 of a fringe.

It is felt that the problems and their solutions

have been well identified.

In addition,it is intended to interface the

present Thomson scattering diagnostic with the

existing data acquisition system.

under consideration has been used successfully on

Soyllac. It has not yet been implemented on Z’f-S

because of the low level of radiation detected by

the scattered-light analyzer.

of the background signal make it difficult for the

scattered laser light. Part of the effort will be

to Improve the photon count at the detector.

Fig. 11-12. ZT-40 Facility.

single-feedplate design.

currents

10RO1DALCU[MNT SIKU

Figure

11-15.

will

lead

use

and

The

PROGRAMMED

f30(wALL)

-a’LL!a

Ei==+=l

TIME(ms)

Fig. 11-14. Numerical calculation of toroidal and poloidal currents vs time for a 12-feedplate design.

0.6

0.8

0.6

0.2-

0.4-

0.4 -

SLOW

FAST

o .005 .Olcsfm.02

i= o ii ~ g l-. ; a 4 -0.2 z

tions inside the torus.

the four stations.

Further experiments

The four stations

viscous flow predominates.

discharges.

(cryogenic

numerical

current

-0.6 o

pumps,

shows

-0.4 -

from

The

the

0.2

,*,

Be (WAU)

Q25

MODE

B@LL)q

0.3 0.35 0.4

The 900-!2 volume is

that are used

sublimation,

These

and

4.0

3.0

I 0.02 0.040.060.0s 2.0

TIME (ins)

Fig. 11-15. Numerical calculation of toroidal and poloidal currents vs time for a single-feedplate design.

removable plug to prevent large field perturba-

pumped at a combined pumping speed of 450 g/s for

The pumps

titanium

alr-bearing turbomolecular pumps which exhaust

directly to atmosphere) are completely oil free.

The initial roughing of the torus from atmospheric

pressure is done with a Roots-type vane pump in

series with an oil seal roughing pump.

systems are trapped and will be switched off the

experiment while still in the pressure range where

110 m Muot Ou 40 m MlNon UA

obtained

11-16

11-13. Artist!s sketch of the front end of ZT-40.

calculation for the 12-feedplate design and a

single-feedplate design are shown in Figs. 11-14

obtained with the power crowbar bank and the

The initial experiments on ZT-40 will utilize

a ceramic discharge tube.

with quartz and metal tubes are planned if their

cleaner

discharge tube is located inside a 2-cm-thick

aluminum shell which serves as the primary con-

ductor of a transformer with the plasma as the

secondary and also as the conducting wall for

stabilization of the plasma.

The vacuum system is designed as a four-point

pumping system on the torus.

will each have a single, large aperture with a

REVERSAL

reported in more detail in Sees. XII, XIV, and

The basic machine parameters for ZT-40 are

toroidal field, 16, capacitor banks were installed

Beginning in May 1977, the

in this area which remains is the fabrication and

experimental area was cleared.

structure was then installed.

installation of the blast walls.

b . Cavacitor Banks.

principal capacitor banks is shown in Fig. 11-12.

3 . Enaineerinu and Construction.

for the ZT-40 experiment was given in late 1976.

complete at that time, most of the final design

design, along with development and construction

primary

engineering and development groups as well as

numerous other staff in the Controlled Thermo-

nuclear Research program. A summary of activities

during 1977 Is given below.

AIDED AND SELF-REVERSAL

TIME

Fig. 11-16. Numerical calculation of toroidal and poloidal currents vs time for a single-feedplate design using the power the currents.

crowbar banks

drive

1.0

0.2

0.1

-..,~

‘“6~

0.4 -

0.5 -

0.3 -

-0.2}

-0.1-

n -1 w L

~__

Iii z c1 a z

given in Table II-I.

I I
I

~AIDED

Toroidal Current

Electron Density

Confinement Time

Current Risetime

ion temperature

Range of Plasma

Current.Density

Poloidal Field

Gas Fill (D2)

Mean Toroidal

Base Pressure

Major Radius

Electron,

Reversed Toroidal Field

(aesuming a2 scaling)

Initial Toroidal Field

the

(ins)

XVII.

-1

During 1$177, the SO-W

Although much

the capacitor,

a. Structural.

activities, was

remained to be done.

p . 5-30 mtorr ne = 1014-1016 cm-3

c = 150-200 US

-r= 2.5 ps to 1 ms

subassembly work.

P. - 10-8 torr

J < 2 kA/cln2

reliability.

150-600 kA

B.o.2-o.6T

B@R.<-0.5T

B6. <0.6T

R = 114 cm

a = 20 cm

current

number

begun.

area.

This

Minor Vacuum Chamber Radius

TABLE-II-I

Completion of the ZT-40

All

the

Approval

activity

design was

preliminary

The steel support

The only aotivity

Some of them are

Construction of the

The location of the

poloidal field, 14, and

construction crew for

lifetime and

satisfactory

Tests were

The mixer

feeds.

and cabled as far as the crowbar spark gaps.

start spark gaps,and cables were

tested before Installation.

bias capacitor bank was about half completed.

Design of the charge, air, and trigger systems was

completed and installation of these systems was

completion of the capacitor banks was largely

completed, partly as a result of a slowdown in

construction caused by a two-month delay in the

delivery of the new crane for the experimental

freed the

Before the design of the 50-kV start and

crowbar switches was finalized, extensive tests

These inoluded changes In geometry

and materials to see the effect of theee changes

on operating parameters and gap

collection of data on prefire rates.

also made on a mockup of the current mixer which

will be installed between the capacitor banks and

the current feeds to the front end.

will equalize voltages at the various current

feeds and will facilitate the changing of the

ZT-40 PHYSICS DESIGN PARAMETERS

The subassembly work required for the

Te - Ti . 100-200 eV

were conducted to determine their performance and

electrical design was developed for the mixer.

remain in operation even when the control computer

Final mechanical design was completed and parts

the

computer

personnel.

was begun.

were ordered.

priority user.

were conducted.

e. Vacuum Svstem.

and installed.

The design of

ZT-40.

c. Power SuDDliea. The design of the power

ceramic torus will be constructed from 60 straight

supply system for ZT-40 was finalized and instal-

sections. Each 30° sector will consist of four 6°

lation of components in the power supply room (see

sections which will be glass-sealed together to

Fig. 11-12) was 70% completed. After the instal-

form a 24° sector and one 6’ diagnostic/pump port

lation of the saturable reactor controllers was

Figure 11-17 shows the position of the

finished, initial performance tests of the units

various tabulations and sapphire windows. It also

A screen room for the power

projected location

supply controllers was built.

special features not readily available in coln-

procurement

mercial units so it was constructed by LASL

Glass seals for sealing the

sections together and for sealing in windows and

d. Control Svstem. Control and data acqui-

tabulations were developed by LASL Group CM8-6.

sition functions on ZT-40 will be provided by a

Two options for vacuum pump ports and diagnostic

Prime 400 computer. During 1977, the computer was

tabulations were developed. The simpler one,which

delivered and checked out and software development

does not require z metal-to-ceramic

In collaboration with representatives

chosen as the final design.

manufacturer,

methods developed were tested on 20-cm and 30-cm

programs were combined so that the Prime 400 can

sections.

now operate in a timesharing mode while at the

assembled 30-cm minor-diameter torus which has

same time always keeping the experiment as the top

been glass-sealed together into sectors is shown

The screen room for the computer

The figure also shows a section

and other data acquisition equipment was purchased

which has had holes for the sapphire windows

A light-coupled interface, which

will electrically isolate the computer from the

experiment, was designed and construction begun.

R . Metal Primary Shell and I Feedplates. The metal primary shell for the experiment was

Work was started on a time delay system for

designed and procurement was begun.

the

two

shows

During

sector.

is down.

existing

diagnostics.

minor-diameter

in Fig. 11-18.

This room had some

f. Ceramic Torus.

sections was begun.

Insulation of the I

VERTICAL PROBE [8+.0..5,

ground out.

,FARAOAY ROTATIONJ

ZT-40

)

The 40-cm minor-diameter

the

ceramic

seal was

the various

photograph of

The glass-sealing

r UV SPECTROMETER.

of location Location of

Electrical

PARTICLE DETECTOR

showing Diagram tabulations and sapphire windows. projected diagnostics is also shown.

Fig. 11-17. torus

feedplates is a difficult

controlling the capacitor bank and diagnostics. A

light-coupled 8-kV pulser was built and tested in

a test bay and on the electrical prototype (see

Sec. k). This system allows the pulser to “floatn

at the electrical potential of the system it is

triggering and eliminates feedback of electrical

noise through the pulser cables.

The vacuum system for

ZT-40 consists of four high-vacuum pump stands

which mount near the plasma chamber (Fig. 11-13)

and roughing and gas-fill systems. A major part

of the design and procurement of these systems was

completed. Turbomolecular, cryogenic, and titani-

um sublimation pumps were purchased and evaluated.

They were found to be satisfactory for use on

the protection and control

logic for the vacuum system was largely completed.

Protection interlocks will be hardwired so they

techniques.

secondary.

pressure.

tests.

cores

used

The

the

gap.

were

power

tests

initial

tests of

crowbar components.

After a series of

system was begun.

shown in Fig.

be evaluated.

anticipated.

Work with

operating

probes.

ranges,

During

second

option

uses

circuit parameters caused by mechanical design

constraints. The different operational modes were

examined in detatl with particular attention being

given to power crowbar requirements. The magnetio

field waveforms shown in Figs.

11-16 are a product of these studies.

3 . Power Crowbar.

will be connected into the I@ and 18 oi.rouitswith

a variable-turn ratio transformer.

transformer design,

parallel plate windings in order to reduce leakage

inductance, were made.

design would work, so design and procurement of a

prototype was started. The power crowbar oircuits

will also require the development of a switch with

high Coulomb capacity. Development was started on

an improved spark gap.

conducted.

design work was done on a mercury-wetted spark

This switch would not be available for

crowbar

required for operation of ZT-40 with magnetic

fields which last for milliseconds.

k. Electrical Prototvpe. An electrical pro-

totype of one-twelfth of one of the two hlgh-

voltage banks was constructed and checkout of the

trigger system performance, to examine available

It will be used to determine

and

determine how well electrical transients can be

suppressed using RLC shunts across the mixers.

Later the prototype will be used to test power

Diaunosti.cs. The potential diagnostics

for ZT-40 were Identified and the primary shell

and ceramic torus was designed to accommodate

them. The location of the various diagnostic

11-17.

design of both internal and external magnetio

Three methods of making Faraday rotation

measurements are being investigated.

only the most promising one will be used on ZT-40.

Four approached which use lasers to make deneity

measurement are being pursued. Considerable work

was done on two of them.

ferometer has worked well on the bench.

being installed on ZT-S and its performance will

A scanning interferometer which

may

but

operation

conceptual

which uses

Preliminary

Also, some

11-14, 11-15, end

The ZT-40 power crowbar

A model was built but no

The tests showed that the

Work was begun on the

A multichannel inter-

Probably

conduct

It is

tests

Photograph of 30-cm minor-diameter ceramic torus.

Fig. 11-18.

problem because of their complicated geometry.

Two options were investigated. The first has the

metal conductors molded inside an epoxy casing.

In this design the epoxy is cured under high

Tests on molded samples showed this

method leads to high electrical strength, at least

a factor of 2 higher than obtained by previously

The

combination of coated metal pieces, polyethylene

hata and sheet insulation.

tests on both approaches, the epoxy-molded design

was chosen for use on ZT-40.

Magnetic Cores. In ZT-40, magnetic cores

are used in the 14 circuit to increase the

coupling between the metal primary and the plasma

The cores used are tape-wound from

0.05-mm-thick Iron tape and are larger than cores

previously manufactured in this manner.

1977, all of the cores were wound and stacked.

Some of them were completed and personnel from

LASL visited the vendor and performed acceptance

Delivery of all the cores i~ expected in early

performed

i. Electric Circuit Simulation.

the oircuit analysia program continued during

19’7’7.The code was updated to reflect changes in

gives a denstty profile every S-10 ps has been

evaluated on the bench through mediocre windows

and worka well. Only conceptual etudies have been

S. Costa and S. Ortolanl, “Stu~y of the Impurity Radiation Losses in a Pinch Discharge,IVUpEE 7510rj, University of PadOva, Italy (December 1975).
R. W. P, McWhirter, wspectral Intensities!n in Plasma Diat?nosticsTechniques, Richard H. Huddlest.oneand Stanley L. Leonard, Eds, (Academic Press, 1965), Chap. V., pp. 201-263.
Newsletter, DMFE, Atomic Data for CTR, July
L. Oren and R. J. Taylor, ‘Trapping and Removal of Oxygen in Tokamaks,” Nuclear 1143-1151 (1977). Fusion~,

11A N(y/el

the Measurement of Plasma Density,” Los Alamos Scientific Laboratory report LA-6727-MS (1977).

c. J. Buchenauer and A. R. Jacobson,

‘Quadrature Interferometer for Plasma Density Measurements,n Rev. Sci. Inst. M, 769-774 (1977).

I!LASLControlled Thermonuclear Research

Program, January-December 1976,” Los Alamos Scientific Laboratory report LA-7082-PR (1978), p. 58.

other two

methods of using Thomson scattering to measure

electron temperature and density are being worked

on. One is a conventional arrangement which gives

data at, two spatial and one temporal point.

other would use a mode-locked ruby laser to give

density and temperature as a function of time.

Work also continued in other areas which would

extend the diagnostics currently used on ZT-S to

the

ZT-40.

RLTERENCES

performed on

The

Two

approaches.

60 A. R, Jacobson,

for

pp.

In~~erfe~~Orneker

Stabilization of an !t.0,1,2High .Eeta Stellara-

Br/Bo = bl sin(e-hz) + b2 sin(20-hz)

A vertical field given by bv . - b1b2/ha,

resulted in centered flux surfaces, was

The conducting wall shape is accurately

by the following paremetrical representation,

characterized by b, . 0.133 and b2 = - 0.0132 (a x

1 cm) with a helical wave number h . 0.06 cm-l and

. 104.7 cm.

the radial field parameter bl normalized to the compression field E. are defined

The final experiments on the Scyllac feedback

stabilization project have been completed, and the

experiments have been terminated.

Details of the experiments leading up to the

1=1,2 results are presented in a paper accepted

for publication in Nuclear Fusion titled “Feedback

tor.” Also, a letter published in Nuclear Fusion,

ItObservationof Propagating m=l Waves in Scyllac;”

contains pertinent results of

In preparation is a journal paper discus-

sing in detail the final results on the 1=1,2

Finally, some of the results

are presented in more detail in the 1977 Quarterly

work.

INTRODUCTION

Progress Reports.

sector experiment.

from 0.62 m to 1.0 m.

produced a

detectors.

pure

R. R. Bartsch, E. L. Cantrell, R. F. Gribble, G. Miller, K. J. Kutac, L. E. Handy, and R. E. Siemon

X = X.

a and

present.

First the

previous sector

wavelength A . &/h

III. SCYLLAC FEEDBACK STABILIZATION EXPERIMENTS

0.265, and X. = 1 cm.

shape of the plasma.

The compression coil

feedback driven

angle.

field

and

The last experiments differed from previous

efforts in several important aspects.

equilibrium configuration was changed from the

1=0,1 combination of stellarator fields to !2..1,2

stellarator fields. Secondly, the helical plasma

periments than on the previous Scyllac experiments

giving a plasma helical radius of about 3 vs the

value of 1.4 cm used in the earlier experiments.

Also the wavelength of the fields was increased

EXPERIMENTAL ARRANGEMENT

and

also

given

which

Plasma radius ia

Additional &=2 fields

The magnet.1.o

mounted

the

The experimental apparatus was essentially

magnetic field was larger In this configuration

the same as that described in detail in Ref. 1.

than in the !L=O,l configuration.

In this experiment the !.0,1 shaped compression

field was initially 16 kG with a risetime of 3 us,

coil was replaced by a compression coil with an

falling to 12 kG in 40ps.

1=1,2 shape and the sector length was extended to

were produced by equilibrium adjustment windings

be 7/15 of a full torus.

!=1,2 vacuum magnetic

windings

The feedback system in the final series of

rotating ellipse with major radius b(l + A2) and

experiments was also much different than In the

minor radius b(l - A2), the center of which is

previous cases in that the output of the position

helically displaced by b/A1.

detectors was processed in such a manner as to

The quartz discharge tube was formed into a

drive the power amplifiers in response to the

toroidally curved helix with a radius of 3.2 cm

modal structure of the plasma instability rather

corresponding to the expected equilibrium helical

than the displacements as seen by the individual

The positions of the dis-

minor axis of the b-m-radius torus, x is the hori-

zontal coordinate increasing toroidally outward, y

is vertical, and the parameter $ is an azimuthal

Values are b = 10.6 cm, Al = 0.235, A2 =

This describes a helically

bA1 cos(hz) + b[cos $- A 2 cos ($ -hz)]

y = bA1 sin(hz) + b[sin $+ A 2 s in ($ -hz)]

charge tube and conducting wall relative to the

minor axis of the torus are shown in Fig. III-1.

Because of the smaller average coil bore, the

perturbation was much larger on these final ex-

where z 1s the longitudinal coo~dinate along the

vertical motions.

amplifier with

charge tube

the mode

Ref. 1.

The

feedback system

was observed.

not orthogonal

modes was selected.

the sector plasma.

wavelength motions.

(10 in all), and

feedback system.

horizontal and

function of z.

the infor-

utilized

signal

twenty

16-

the

10 modes, it was necessary to rotate the coordi-

nates from the horizontal-vertical of the position

detectors 45 degrees to correspond to the orienta-

tion of the %=2 feedback,coil system. There were

44 modules, each driving one-half wavelength,

corresponding to 11 wavelengths in all and two

orthogonal force directions.

The longitudinal mode structure in an open

ended sector is not known unless the boundary

conditions at the ends are known.

modes that might be expected on physical grounds

were considered, corresponding (1) to the plasma

ends being fixed, and (2) the plasma having no

curvature at the ends (an elastic bar model).

second set of modes agreed best with the data in

that it allowed the full motion near the ends that

However, that these modes

leads to very undesirable fea-

tures, and thus a third set of orthogonal Fourier

These modes, illustrated in Fig. III-2, are

not intended to represent the time eigenmodes of

The function of the mode

analysis procedure is to interpolate the position

detector information as a function of z , providing

a short wavelength cutoff on the response of the

The mode analyzer approach thus

circumvents undesirable oscillations arising from

the system response to otherwise stable short

Orthogonal mode signals have

the convenient property of being the best fit

values (in the least squarea sense) for a fit of a

sum of modes to the position detector data as a

The layout of the L=1,2 sector

Fig. III-3. The location of the Thomson scatter-

ing apparatua, the excluded flux loop probe, and

the two orthogonal 16-channel luminosity apparatua

are indicated. The top viewing position detectors

were located at hz = O in every wavelength except

at z = O (the left end in Fig-.III-3) and z = 2 X

the front viewing position

detectors were located at hz = n every wavelength

except at z = 2.5 X (10 in all). Other diagnostic

lnformatLon was obtained from 5 streak cameras,

viewing the plasma in different locations, and a

coupled cavity interferometer which determined the

preionized plasma density.

The

are

Two sets of

is shown

segment photodiode position detectors2 viewing the

plasma motion from the top and front of the dis-

determine

These signals fed into a fast

analog processor which mode-analyzed

mation about plasma position as a function of

longitudinal position into five longitudinal mode

signals for the horizontal and vertical planes.

The mode-analyzer output consisted of the sum of

signal and its time derivative for

additional damping. The gain for each mode signal

and its derivative were individually adjustable.

Each power amplifier was driven by a summing

properly weighted mode

inputs to transform the mode signals, effectively

interpolating between position detectors. For all

Fig. III-1. Discharge tube position relative to conducting wall.

discharge tube in the same q anner as described in

DEFINITION OF MODES

MC41EO

MOOE 3

MOOE 2

MOOE 4

DISCHARGE

PLASMA

TUBE

librium was studied qualitatively with streak

cameras and quantitatively with uae of the feed-

back position detectors and the multichannel lum-

The streak photographs showed less dynamic

plasma behavior than in the 9..0,1experiment and a

surprisingly long time, even without feedback,

before strong wall hits were indicated by lighting

up of the discharge tube. A typical streak photo-

graph is shown in Fig. III-4.

k.O instability growth rate from

small-6 theory is Y . 0.27 W- 1 (vA = 17 cm/pa,

B = 0.7) and three growth times is about 10 US.

The streak photographs suggest that the plasma,

although m=l unstable, gently rebounda from the

The plasma behavior shown in Fig. III-4 was

approximately the same over a wide range of

filling pressures, from 12-96 mtorr.

sharp contraat to the 9..0,1 experiment where the

filling pressure strongly influenced the inward or

outward motion of the plasma.

limit was mainly due to the decrease of the plaama

wall.

RESULTS

The ml,

Z(A)

Toroidal

inosity apparatus.

THETA PINCH COIL

TOP VIEW

luminosity

Z=3.5A

‘s’”-+

+—

the

point

where

This ia in

The lower pressure

. The toroidal equi-

SHOT 7005 NO FEEDBACK

there

was

~—5o,s-+

Ffg. 111-4. Streak photographs of plaama without feedback.

“Modes controlled by feedback system.

Fig. III-2.

insufficient light for observation by position

detectors and streak cameras.

Fig. III-3.

Experimental arrangement.

-r---------

Fig. III-5. Position and trimming circuit waveforms.

outward, striking and rebounding from the wall at

TIME (@

pressures

tendency for the column to bifurcate, evidently

due to an m.2 instability.

instability was not sharply defined as a function

of pressure since above threshold the instability

had a slow growth. Below about 70 mtorr there was

no evidence of an m.2 instability. The filling

pressure threshold for the m=2 instability calcu-

sharp-boundary small-ti theory

8 mtorr, thus it is apparent that the theory is

very pessimistic, as has already been observed.3

As in the !..0,1 experiment 9..2windings were

used for adjustment of the equilibrium.

were driven with Jones-Weibe14 circuits that were

designed to provide a step waveform. However, in

this experiment such a waveform was not suitable.

initial tendency to move

If 2 step i?.2current beginning at main bank

time was applied with sufficient amplitude to

prevent this outward motion, the plasma would move

rapidly inward at late times. As an ad-hoc proce-

dure the slow ringing capacitor in the Jones-

Weibel circuit was not fired, changing the output

waveform from a step to a pulse.

initiated 3 us before the main bank.

the

(average over five central wavelengths) together

with the !?,.2field waveforms for the step and

pulse cases are shown in Fig. III-5.

Ideally the equilibrium adjustment waveform

would be such that a small change at late times,

after initial transients had died away, would

produce a change in the trajectory from inward to

outward. In the limited time available to perform

this experiment it was not possible to provide

system in the programmed mode to produce a current

However, using the feedback

The

high

from

lated

filling

plasma had

the locations hz = T.

amplifier system).

such a waveform.

trajections

plasma

-2

was

-1 -L -,

there

These

1- Z w Eo t!

The onset of the

i! ’

‘v”I

~J…

This pulse was

The average

deformation

horizontal

helical

plane

the

-2

-o

-0.005

superimposed views

wavelength apart,

plasma as a

aoos

the

pulse at late times, it was possible to experi-

Small-& diffuse profile equilibrium theory5

mentally demonstrate nearness to equilibrium.

tends to require slightly larger helical fields

The data are shown in Fig. 111-6,which gives

than observed in the experiment to give equili-

the average trajectory in the horizontal plane

brium (theory: b2 = 0.018, exp: 0.012 < b2

together with the total 1.2 field (equilibrium

0.016, for ~= 0.7, a = 1 cm).

adjustment plus programmed pulse from the feedback

Using a streak camera that

of the plasma one-half helical

TIME (pa)

Fig. III-6. Sensitivity of equilibrium.

(poorly

given

image

plasma

small-~

~ = 0.6).

involving 1=0 fields.

proportional to

with the wall.

occurred,

the 8=0

field.

wall

hit

l%=20mT,3.7m FROM EWO ----4=l,2ExP, m Ao0maJuFaa.Lk3-

Fig. III-8. Lxcluded flux for 9.=1,2compared with 1=0,1 fields.

A common feature of the exoluded flux be-

havior from the two experiments is a fall-off

beginning about 20 us. This is due to end-loss as

based on two-dimensional end-loss computations by

Brackbill.6 Other small effects (such as the fall

off of the main Bz field) are included by incor-

porating the particle loss due to magnetohydro-

dynamlc flow as calculated by Brackbill into a

one-dimensional calculation describing the radial

evolution of a theta pinch.7

show that particle loss begins after a time delay

approximately equal to the Alfven propagation time

(vA = ‘outside/p) while loss of temperature on

40 —

30 - -

(for

theory

-e 20 Q

visible

Fig. III-7)

Z=7.5A,8A

shown in Fig. III-9.

The excluded flux traces

SHOT 7347 TOP VIEW

transient behavior

observed with

length) and

streak

unit

The theoretical curve is

TIME (PS)

Fa=25mT,4.2m FROMENO

The calculations

SHOT 72!54

I !50

TIME (ps)

Figure III-9. Excluded flux decay compared with end-loss theory.

Fig. III-7. Streak photograph mehsuring helical shift.

function of time was studied. A streak photograph

1s shown in Fig. III-7. t4arkerscentered on the

indicated the plasma helical radius to be approx-

imately 3.2 cm which was the helical radius of the

quartz discharge tube and also the helical radius

sharp-boundary

were not strongly correlated with whether or not a

cameras. This is remarkable in that the excluded

flux is directly related to the plasma energy (BA$

energy per

therefore the time for radial energy losses is

long even when the plasma is in apparent contact

Because of this the excluded flux

depended little on whether or not the equilibrium

adjustment field or feedback were applied.

In Fig. III-8 is shown a comparison of the

excluded flux traces obtained in this experiment

with those obtained in the previous experiment

The excluded flux, in this

case, is quite different at the field maximum

(hz = O) and field minimum (hz = ~) locations of

strong

associated with g=O oscillations of the plasma is

also present. The shorter time scale erratic be-

havior of the excluded flux in the 9.=0case is due

to pick up of the time varying feedback magnetic

axis begins after a longer time equal to the cusp

where c is the sound speed.

between theory and experiment clearly shows that

the fall-off of excluded flux observed both in

this experiment and in the L=l,O experiment is due

Thomson Scattering.

with the same apparatus as in ,the 1,.0,1 experi- ment,8 was used to determine the electron tempera- ture. The results are given in Fig. 111-10. The

data points were obtained on different shots with

and without feedback (no difference was evident)

with the laser pulse occurring at different times.

Measurements at early time with the laser beam

displaced sideways showed a definite non-constant

The

measurements is at least partly due to the tem-

perature profile and the variation in plasma

position at the time the laser was fired. Also

’

1 -1/2

profile.

)

temperature

~+- ,

to end loss.

propagation time

Ioa m w 7a w

~EAK FIELD

Isa

LUMINOSITY PROFILES SHOT 7176,Z=4X 4-13PS ,At=lps

with the outer plasma cooling more rapidly than

the plasma on axis and this seems consistent with

the larger data spread at late times. Energy loss

due to radiation, not included in the calculation,

may also become important as the plasma tempera-

b, Luminosity Measuremen& .

16-channel luminosity devices were used to de-

termine the luminosity profiles of the plasma. At

the location of the luminosity apparatus (hz = O,

see Fig. III-3) the elliptically shaped plasma was

oriented so that it appeared narrow when viewed

The calculations show

There

the

spread

ture drops.g

two-dimensional

already discussed.

‘he good agreement

Thomson scattering,

shown in Fig. 111-10 is a

with the upper envelope of

measurements Vs

SIDEVIEW AY =.205cm

… … … . .

… … …

… … .

… …

… .. .

… …

… . .

… .

…

. .

the data. an nonion-skinu effec?t

theoretical curve besed

end-loss

seems to be agreement

calculations

Two orthogonal

TOP VIEW AX= .225Cm

… … …

… … . .

… … .

… …

… . .

… ..

…-.

… .

…

.,..

. .

Fig. III-11. Luminosity data showing profiles,

?!i2

TIME (#S)

Thomson scattering temperature time.

Fig. 111-10.

top and broad when viewed from the side.

ity signal available from the feedback position

are shown in Fig. 111-11.

luminosity profiles were

detectors, as seen in Fig. 111-13.

detectors’ spectral response was

xO)2/a2], assuming lumi-

650 nm and as in the I.=1,0experiment the absolute

n2), to determine the plasma radii al and

plasma luminosity was within a factor of two of

Results are given in Fig. 111-12. These data

are consistent with the sharp-boundary small-6

The plasma radius obtained from the luminos-

al = a(l - 62),

ity measurements together with the exoluded flux

82 = 2b2/[ha(2 - !3)1, only if a rather low beta

information allow the determination of the plasma

‘32 were

pressure, or f3,(~ - 0.1 +0.5

Assuming a Gaussian pressure profile with

however, the luminosity measurements did show a

radius a . 1 cm and using the measured values

regular oscillation of the peak luminosity, ap-

A@ = 20 kG-cm2, and E = 16 kG, the plasma beta

proximately in phase in the two views.

oscillation was also present in the peak luminos-

The electron and ion temperature equilibra-

;

a2.

g15

The

2.0 -

theory,

?-/…/

Raw data

nosity .

Gaussian

from the

Oscillations

(- exp [-2(x-

value is used, f3- 0.2.

Z 1.0-

0.5 -

/’

The

not

With

occurs

~sition

SHOT7176

observed.

k/d’fpwh+-

fit with a

value is 0.7.

:2 x 1016 cm-3.

a2 = a(l + 62),

and temperatures.

pure bremsstrahlung.

external current.

was about 5.5 ph.

resistance of

rises.10-12

15 kV-~

bO kV-M

111-14

field.

cases

for

the

)

TIME (/s)

Fig. 111-12. Plasma ellipticity from luminosity.

SKX VIEW

Tw VIEW

Fig. 111-13. Position detector peak luminosity signal.

l/e

350-

field

initially

The feedback

A@/(a2B), 6 > 0.2).

In a high-beta stellara-

The return wires had an

in the return

for f3 . 0,

resistors

(7 kV-M

the

and

tion time is a few microseconds at these densities

$ = o.7 the Plasma density is determined to be ne

Assuming Ti = ‘fe= 120 eV and

tor an axial electric field is produced by the

change In azimuthal flux between the outer bound-

ary of the plasma and the conducting wall that

magnetic

In the !.0,1 sector the ends were

electrically isolated to prevent the flow of

However in the 1.1,2 configuration, to avoid

a large induced voltage with the larger helical

fields, a current return was provided between the

metallic vacuum valves at the ends of the sector.

Normally a loop linking the main Bz field and a

1 Cl were placed

circuit to partially cancel the current and to

damp out the remainder.

inductance of 2.2 ph while the plasma inductance

The measured axial current is shown in Fig.

with

cancellation loop in the circuit and with them

shorted out. The azimuthal flux calculated from the sharp-boundary small-6 mode 11 is - 40 kV-~s

(independent of a and 13) from the %=2 field and

for $ = 0.65

for !3. 1, independent of a) from the &l

likely due

Fig. 111-15. Horizontal mode amplitudes controlled by feedback.

SttOT6754

TIME (/&)

Fig. III-14.

Axial current with and without cancellation circuit.

The cancellation loop reduced the current to

0.29 of its original value with a flux of

.-60 kV-Vs implying a azimuthal flux of 80 kV-ps.

The initial current of 12 kA (see Fig. 111-14) and

the circuit inductance of 7.7 ph gives a value of

92 kV-ps. This discrepancy between theory and ex-

l-\

12.5

m-l-..

assumptions of the theory.

z * HN

loo”1‘“1

I I./’*\

75 -;

periment is most

;./-,
J .P,

b2 - ~~+T1~~

back.

HORIZONTAL PLANE —2cm/DIV ----1kA/DIV 5us/DIV

NO FEEDBACK (7325)’ ,, ,,,

!,,!,,

“ODEO-4-

b,,,,,,,,, -‘G—’-::

MOOE I --- --- ---------

,!,

,,,,

l!,

!,,

!,,,,,,

!!,,!

the small-6

In Figs. 111-15 and

‘0”4-

‘“DE4z

Shown in Fig.

In the

,,,,,,

,!!!!

,,,

!,,,,

l,,

,,,

,4,

.—.

,,:

,,,

,,,,,

% ,,,

.’~y,;

,, !1111

FEEIX3ACK(7326)

,$,.,$$, \

FEEDB/XK (7326]

—

,!!!,!!

,,!!,.!

!!,!

!.!.

VERTICALPLANE

—2cm/Dlv ---lkA/DIV

5&s/Div

NO FEEDBACK (7325)’

,,!!,!!

-“‘x- ‘“DEG ‘“DE’4

Effect of Feedback.

111-16 are shown the mode signals (see Fig. III-2

for a definition of the modes) and feedback

currents for a typical shot with feedback.

The effect of feedback was only to stabilize

the plasma m.1 motion since the energy confinement

was dominated by end loss with or without feed-

All modes had a velocity gain ‘cl. 3 us,

( current of about 500 A/cm and about 50 G/cm for

)

and an overall gain for the ?.=2

the 9..2field (at a = 1 cm).

In Figs. 111-15 and 111-16, an improvement

for all modes with feedback is apparent.

independent determination of the effectiveness of

feedback is the sixteen-channel luminosity device

discussed in a previous section.

111-17 are two successive plasma discharges where

all experimental parameters are held constant

except that in case (a), the feedback system is

turned off while in (b) it is active.

first case (a), the plasma moves rapidly to the

wall of the discharge tube (indicated by the

dashed parallel lines). In case (b), the feedback

system controlled the instability and held the

Vertical mode amplitudes controlled by feedback.

Fig. 111-16.

4’-------------- ..

‘\J-8,~,.-”’;

l____ ------------------

2 -.-.-----;’.---------------

Y ! ‘w]

2 -----------------------------

WITH FEEDBACK

u i= ii) o Q

(al WITHOUT

-2

/..=”

t -4 ~ o

4 (b) [

feedback is plotted.

oscillations in this case.

shape as observed in Ref. 1,

aggravated by the weak damping of the helical

In no case was the confinement limited by the

presence of these small oscillations.

feedback, the average plasma excursions were quite

An unexpected osolllatory behavior with a

frequency approximately equal

oscillation frequency was present and in fact

varying the velocity gain setting, which mainly

governed the amplification of these oscillations,

Without feedback the helical

oscillations were observable with both the top

viewing position detectors and those looking from

the front implying that both spatial phases were

The tendency for the feedback system to

oscillate at the helical frequency is probably due

to direct coupling of the feedback L=2 field to

of axis with feedback.

(#S)

small.

present.

CONCLUSIONS

the plasma L.1

had little effect.

I FEEDBACK

improvement

general,

the

feedback

the

helioal

feedback

Many interes-

However, even without

performance.

Furthermore,

improved

Present

thermal

system

the

The

clearly limited by partiole loss and

conduction to the open ends of the sector.

The m=l instability which accompanies high-

beta stellarator equilibrium has been shown to be

amenable to feedback stabilization.

it has been demonstrated that the equilibrium

obtained with the 9.=1,2 stellarator fields is

qualitatively superior to the t-0,1 system in that

the transient behavior occurring as the plasms

assumes its equilibrium ahape is minimal. The use

of modal control in the feedback system resulted

its

technology exists that could reduce the amount of

feedback energy loss by a factor of 20.

Further experiments in a complete toroidal

geometry would be necessary to determine the upper

limits of confinement time.

The experiments using the modal

system were uniformly successful.

ting features of the system were unable to be ex-

perimentally studied due to time restrictions. In

TIM:

Fig. 111-17. Luminosity position data showing feedback control.

plasma near the center of the discharge tube.

Fig. 111-18, the quantity (~C2 dz/L)l’2 averaged

over 10 shots with feedback and 10 shots without

crease in the magnitude of the plasma excursions

experimentally obtained confinement times were

This shows a definite de-

stability of the plasma column in all cases.

TIME(ps)

Fig. 111-18. Averaged rms displacements with and without feed- back stabilization.

1. R. R. Bartsch, E. L. Cantrell, R. F. Gribble, K. A. Klare, K. J. Kutac, G. Miller, and W. E. Quinn, Nucl. Fusion (1978), to be published.

R. Gribble and G. Miller, Rev. Sci. Instrum.
J. Neuhauser, M. Kaufmann, H. Rohr, and G. Sdramrn, Nucl. Fusion lZ, 3 (1977).
G. Miller, Los Alamos Scientific Laboratory

unpublished data (1977).

R. E. Siemon, Appl. Optics J.3,697 (1974).
F. C. Jahoda, Los Alamos Scientific Laboratory,

private communication (1977).

E. Funfer, M. Kaufmann, W. Lotz, J. Neuhauser, G. Schranmn,and U. Seidel, “Linear 8=1 Stel- larator Experiments Using a Shaped Coil,” Garching Laboratory report IPP 1/130 (1973).

E. S. Weibel and I. R. Jones, Rev. Sci.
G. Miller, Phys. Fluids 18, 1704 (1975).

972 (1961).

D. C. Barnes and J. U. Brackbill, Nucl.

Sci. and Eng. @L, 18 (1977).

J. U. Brackbill, Los Alamos Scientific Labora-

tory, personal communication

P.C.T. Van der Lann, “Effects of Flux Con-

servation on the Field Configuration in Scyllac,” Los Alamos Scientific Laboratory report LA-6767-MS (1977).

REFERENCES

Instrum. x,

.!& 778 (1977).

(1977).

R. J. Commisso, C. A. Ekdahl, K. B. Freese, R. F. Gribble, K. F. McKenna, G. Miller, and R. E. SLemon

INTRODUCTION

axial particle flow.

radius Z 1 cm.

Six hundred

reactor.

devices.

age.

(a)

coil. Primary bank operation at 45 kV generates a

of 0.6 kV/cm at the inside wall of the

The ability of linear theta-pinch devices to

8.8-cm-i..d.quartz discharge tube and a peak com-

generate high-energy plasmas of fusion interest

pression field of 50 kG is obtained 3.0 pa after

has clearly been established. This fact combined

A Scyllac- type crowbar

with the additional advantages of simple geometry,

system extends the magnetic field in time with an

high plasma beta and density, demonstrated plasma

The experimental results

equilibrium and neutral stability properties, and

discussed below were obtained at 45 kV primary

ease of plasma heating make the linear theta pinch

bank voltage and with a theta-pinch fill pressure

an attractive alternate approach to a pure fusion

reactor and a prime candidate for a fusion-driven

In order to minimize plasma-wall interactions

fisslle fuel producer (hybrid). The fundamental

at the discharge tube ends and provide access for

problem encountered in the adaptation of

optical and internal plasma diagnostic, eaoh end

linear theta pinch to reactor applications is the

of the theta-pinch discharge tube is terminated,

loss of plasma energy containment due to particle

4.5 cm beyond the coil ends, within a quartz

streaming through the ends and axial electron and

The chambers are 20 cm in

ion thermal conduction along the open magnetic

diam, 50 cm long, and end-supported by 20-cm-diam

field lines. A significant reduction in the mag-

nitude of these losses is necessary for the suc-

Gas preionization ia accomplished, in the

cessful development of a linear theta-pinch fusion

theta-pinch mode, with a preionization capaoit.or

The main objective of the Los Alamos

bank consisting of ten 0.7-PF, 75-kV capacitors

Scientific Laboratory (LASL) linear theta-pinch

The preionization bank oscil-

program is the investigation of end-loss physics

lates at 500 kHz and produces a peak field of 2kG.

and the development of end-stoppering techniques

which enhance the plasma confinement in these

EXPERIMENTAL RESULTS WITHOUT END PLUGS

In order to determine the plasma parameters

The principal experiment in the program is

and end-loss characteristics, which establish ref-

the 5-m-long Scylla IV-P linear theta pinch. The

erence conditions for end-stoppering studies, ex-

peak theta-pinch plasma parameters are

periments were conducted with the theta-pinch ends

1016 cm-3, Te + Ti z 3.2 keV, $ Z 0.9, and plasma

End-on interferograms are used to

With the reduction of end losses

determine the time history of the plasma column

aa the ultimate objective, the following experi-

particle inventory, peak density, radiua, and, in

ments have been carried out on Scylla IV-P:

conjunction with the excluded flux measurement,

an investigation of the characteristic end-

the plasma beta and total temperature.

10ss times, and plaama flow processes near the

theta-pinch coil ends, and (b) material end-plug

Figure IV-1 presents a time sequence of end-on

end-stoppering experiments designed to eliminate

plasma column interferograms obtained with a 30-ns

Each interferogram was obtained on a

ruby

THETA-PINCH EXPERIMENTAL ARRANGEMENT

plasma discharge and the entire 8.8-cm diam of the

The Scylla IV-P theta pinch has a maximum

discharge tube was illuminated by the interfero-

energy storage of 2 MJ at 60 kV primary bank volt-

The fringe pattern on each

1.8-PF capacitors feed the

interferogram was digitized and reduced with soft-

500-cm-long, 11.2-cm-diam single-turn compression

ware developed for the CTR-PDP-1O computer.

the

vacuum E

metal vacuum tees.

expansion chamber.

L/R decay of 110 W.

of 10-mtorr deuterium.

discharge initiation.

IV. SCYLLA IV-P LINEAR THETA-PINCH EXPERIMENTS

meter laser light.

charged to 55 kV.

unobstructed.

holographic

pulsed

~etic

1.5 x

n q

interferometer.

separate

FieM,.

laser

20.0+s

Fig. IV-1.

end-on plasma column interfero-

7.5US

301U

115/Ls

Time sequence of grams.

profiles.

symmetric

fringes.

highly

riginal Data

Radius (cm )

Fig. Iv-2. End-on interferogram obtained at t = 3.5 ps and corresponding reduced fringe shift profile.

Effects

1.0

0.5

?5.-

For Refractive

Data Corrected

coil length (500 cm).

number of fringes.

nak(Te+Ti) = Ba ~

density, na,

instability

expression

27rrdr,

plasma

free.

(1)

1.5

the

all

was

1.5

2.5

interferograms an

From analysis

determined

from the

(2)

(3)

the above values of f3aand na yield a total plasma

‘a = 3.24 x 1017 fa/L cm-3 ,

The plasma parameters derived from analysis

of the interferograms are most. accurately deter-

where r~ Is the excluded flux radius (discussed

mined near peak field time when the plasma column

below), b is the tube radius and fa is the peak

and

Equation (1) is derived from

However, refractive bending of the interferometer

the defini.tlonof excluded flux and the pressure

laser light in the strong radial plasma density

balance equation, assuming that the plasma tem-

gradients existing near peak field time can com-

perature is independent of radius.

plicate the interpretation of the interferograms.

hand - analyzed

In order to assess these refractive effects and to

averaged L3a’of 0.90 ~ 0.02 was obtained. The peak

ponding corrected and uncorrected fringe shift

where L was taken to be equal to the theta-pinch

The maximum correction due to refrac-

From the hand-reduced in-

tive bendingl typically amounts to about

terferograms a peak na of (1.49 ~ 0.03) x 1016

cm’3 was obtained. From pressure balance,

determine the peak plasma parameters as accurately

as possible, several interferograms obtained at

t: 3.5 IIS were hand-analyzed; the interferogram

computer reduction program does not account for

ray-bending effects. Figure IV-2 shows a typical

near-peak field time interferogram and the corres-

Since the corrected fringe shift profiles for

the hand analyzed interferograms are not exactly

Gaussian, the peak plasma beta, ~a, was determined

by numerically integrating the exact equation

the

dimension of the plasma column is unknown.

cordingly, the plasma radius obtained from Eq. (5)

can consider only a best approximation. The time

the plasma radius obtained from

temperature at t z 3.5 PS of Te + Ti = 3.25 ~ 0.15

observed m . 1 wobble

The total electron inventory, Ne, was deter-

dition, the appropriate length, L, which should be

onsets - 5 ps after discharge initiation. In ad-

From analysis of both the corrected and

uncorrected fringe shift profiles, a maximum error

The plasma excluded flux radius, re, meaaured

in the total particle inventory of about 7% was

with the diamagnetic loop-probe ayatem located at

interferometer refractive

the center of the theta-pinch coil is shown in

bending effects. From the corrected fringe shift

Fig. IV-3(b). The presented curve is the average

profiles a partiole inventory at t : 3.5 lJs of

of 44 data shots, taken simultaneously with the

(1.96 fO.04) x 1019 was determined.

ccmpares well with the initial theta-pinch par-

one standard deviation from this

ticle fill of 2.1 x 1019 at 10-mtorr D2 fill Pres-

The magnitude of the error bars, es-

pecially at late times, results more from limita-

Time Historv of the Plasma Parameterq.

tions in the data acquisition system

The time history of the plasma parameters was de-

irreproducibility.

keV.

sure.

snalyais

attributed

Ne = 3.24 X 1017

mined from the equation

()Ne 1/2

indicated from the

column contains an

dients are large.

> 10 US (Fig.

tines t -

mentally.

this has

equation

however,

LnJJ

a=—

’

f(r)2nrdr .

interferograms,

the

(4)

used

ray-

plasma

end-on

average.

describe

shot-to-shot

This value

represent &

evolution of

(a) Time history averaged excluded

verified experi-

1.2-;

:i:m

been

%e .

02 -

0.4-

1 I

0.8-

(5)

0.8-

0.4-

0.6-

1.6-

*-,

termined using the results of the PDP-10 computer

of the end-on interferograms. This ap-

proach ia justified since the errors introduced by

refractive ray-bending effects are small and occur

only near-peak field time when the density gra-

The low density plasma “halo,”

observed to surround the central plasma column for

IV-l), was not included in

the interferogram computer analysis data used to

determine the plasma parameters. The halo plasma

is generally believed to originate from wall-

plasma interaction near the theta coil ends,

not

The plasma radius ia determined from the

where again Ne is the total number of particles

evaluated from the end-on .interferogramsneglec-

ting the plasma halo, and na is the peak density.

It should be noted that na can be higher than

interferograms if the plasma

axial curvature generated by

the

and

Ac-

than

from

bara

axial

error

time-dependent

inatability2 which

; l. I

1 1

0.8; .

)

1 I

(w)

(a)

l l

.{;.ea.

Fig. IV-3.

of plasma radius, a, and (b) flux radius, r .

Specifically, the excluded flux data waa recorded

ferogrsms (Fig. IV-4a), the

on a 6-bit SEC unit which, it has recently been

field Bz, and the calculated values of ~a. The

insufficient to

resulting time evolution of the central plasma

excursions in the excluded flux signal, particu-

the low-level (-l-bit magnitude) signals

Neglecting the data obtained during the first 6 US

existing at late times during the discharge. More

of the column life time, when the main magnetic

recent data have been recorded on an oscilloscope

field undergoes its maximum oscillation before the

in order to eliminate the problem.

crowbar becomes effective at t Z 6 ps, a charac-

Assuming a Gaussian density profile at all

teristic temperature decay time, TT, of 12.8 z 0.9

times during the plasma lifetime, the plasma peak

US is obtained from a least squares” fit to the

beta, ~a, is calculated at a given time using the

plasma radius obtained from the individual inter-

?. Plasma End-Loss Measurement&. As in pre-

averaged

‘4 ‘he time history of (3ais shown in

theta-pinch experiments, the

central plasma column particle containment time

was obtained from analysis of the end-on inter-

The time history of the total plasms tempera-

ferograms, neglecting the low-density plasma halo

ture is determined from pressure balance

surrounding the main column.

Eq. (3) using na evaluated from the end-on inter-

the central column electron inventory is presented

larly

and

the

m ‘1

1.6r.

found, ia

ferograms,

Fig. IV-4b.

(Fig. IV-3b).

‘(lb6

-… .. .*. 8” …

I0.8

. t.*- . .

, 1 1

(3=, and

0.6-""

beta,

02 -

0.4-

The initial period (t < 6 US) of

constant electron inventory, also observed in pre-

vious experiments,3-5 has been shown on Scylla

IV-P to result from radial confinement of the

plasma in the end expansion chambers;2 the axially

ejected plasma remains collimated (little radial

value

column

resolve

the large

temperature

in Fig. IV-6.

vious high-energy

data of Fig. IV-5.

*o + t=—

.. ..O

t>6.O PS

o.50~

iio

2.0-

l m

1.0—

”.”.

5.0

.. ,

1 1

.”.

shown

mSgnetic

measured

Fig. IV-5.

The time history of

Ttme(ps)

Time (US)

Fig. IV-4. (a) Time history of plasma peak density na, (b) peak (c) main magnetic field waveform, B=. -

Fig. IV-5. Time history of total plasma temperature deter- mined from pressure balance.

, 1 I , I I * 1

”.

… .

;. t I 1 1 1 , 1 t I 1

A ~nearScy[ac(5m) s ScyllaIV-3(3m) n Scylla I-C (Ire)

Fig. IV-6. Time history of the central plasma column electron inventory.

expansion) for tens of centimeters after flowing

out the ends af the theta-pinch coil.

.quares fit to the data of Fig. IV-6, for t ~ 6

e-folding

021

() 1/2

determination.

terms of thermal

12.5 ~0.5

yields

L ‘i

‘=7=

data

VS.

us,

transit

15 Tim. (ps)

‘x

(2)

5 -

4 -

2 -

3 -

2T7

~=L

6 -

A least

nat determined.

The solid

represent

end-loss

between

times as

curves

time

2.2

and

(6)

The data from

The data

particle

end-leas

thermal

transit

tiut’s.

where L Is the theta-pinch coil length, T the

total plasma temperature, and rlis a normalizing

From examination of Fig. IV-7 the fallowing

parameter, which is theoretically found to be a

major results can be identified, (a) Fo~ all the

function of the plasma 6; the possible dependence

experiments the plasma particle end-loss time :$

of rlon other plasma properties is unknown.

2.8

Fig. IV-7, ~ is plotted against ~.

(b) Experimentally, the normalized end-loss ::n~

In order to obtain a consistent

existing

comparlaon with the present experimental results,

the electron inventory data from previous experi-

theories which are discussed by l?reidberg and Weitzner.6 The x’s mark the resulte of a recent

ments495 have been re-analyzed so that only the

numerical magnetohydrodynamic treatment of the

following the Initial period of constant

end-loss prablem by Brackbill et al.7

inventory are used in the particle end-loss time

points are fram the Scylla IV-P, Scylla IV-3,4 and

Linear Scyllac5 experiments which generated high

collision dominated Scylla I-C experiment.8

determining the experimental points, the peak

values of B and total plasma temperature obtained

in each experiment were used.

Scylla IV-13 were excluded from Fig. IV-6 because

the magnetic field was nat crawbarred snd S waa

Fig. IV-7. Theoretical and computational resulte for rlv~ and experimental data points.

The particle end-lass time can be defined in

energy collisionless plasmas, and the low-energy

42 B2 +—=—,

2A2

f 5’3exp(2se/3) + exp(2si/3)1 ()

[

aPPears to be independent of the plasma p.

plasma area A is determined from the pressure

where $ . a ~1 - B. B. is the flux inside the

temperature

set of equations for se and s ,

and (12) constitute a closed

Figure IV-8 shows a comparison of the energy

line density calculated from the above equations

(using the correct l.nAee, !2nAie, Q,nAii factors)

with experimental observations for cases with and

The data were essentially fit

~is independent of the plasma col-

The

two-dimensional unsteady MHD code best describes

the observed experimental results.

longitudinal temperature profile in a theta pinch

is probably very near the form (1-(z/L)2)2’7 given

by the constant pressure model.9~10

calculation of theta-pinch heat loss is therefore

possible since the end-cooling term 2/aZ(KaT/az)

(uSing K . T5/2).

equations take their simplest form If entropies

rather than temperatures are used as the dependent

variables. If Se, Si are the entropy per particle

for electrons and Ions at r . 0 and z . 0 (i.e.,on

axis in the center of the machine), the equations

END-LOSS CALCULATIONS

It was pointed out

(d)

then

~ Ti/Te-l

dse —=.— dt

-4 KT/(7L2)

The parameter

lisionality, and

describing cooling are

‘eq(~s) = 6 x 10-3

(using lnA = 10),

~ Te/Ti-l

Tce(@

T eq

X=2

dsi

-G

(

(7)

(8),

The

(c)

the

that

al.,

Siemon

plasma.

A simple

Brackbill

Equations (7),

balance relatlon,

without end plugs.

solved numerically.

Te = (N exp(se)/A)2/3,

From Braginskiill

(11)

OPEN

(8)

(lo)

”., ~

1.0

(9)

‘i=~o ‘i, CLASS r = Zqls to= 9ps

Te(eV)3/2/n(1016 cm-3) ,

= 1.4 X 106 L(LU)2

n(1016cm-3)/Te( eV)5/2 ,

‘ci(ps) . 5 x 107 L(m)2

n(1016cm-3)/Ti( eV)5/2 .

Pressure balance is used to relate entropy

and temperature. If N(t) is the line density at z

= O and B(t) the external magnetic field, then the

, Tce )

)

where,e.g.,Tce = 21L2 n/(8 Kc).

using

(1.2)

obtained

which can be

Ti = (N exp(si)/A)2/3.

of excluded flux

LID PLUGS

‘i, CLASS

15,

L”%

‘i=&

’+.

‘J

”

{

TIME (ps)

Fig. IV-8.

Measured and calculated values times magnetic field.

ion thermal conductivity (uncertain

because of long mean free path), and particle end

loss (i.e.,N(t)) as adjustable parameters.

the plugged case N(t) was taken as constant and Ki

varied to give the fit shown in Fig. IV-8. In the

unplugged case, using the value of K~ determined,

and sssuming the simple form N(t) = NOexp(-(t -

tO)/T) (for t > to), the parameters to and T were

varied to obtain the fit shown.

EXPERIMENTS WITH SOLID END PLUGS

Exploratory experiments performed in 1976 on

the Scylla IV-P theta pinch using silicon dioxide

(quartz) end plugs12 demonstrated that plasma flow

past the plugs and out of the ends of the pinch

the plasma column was improved with the plugs

In addition, the stability of

Ablation of

slight, with only about 0.1$ of the plasma energy

invested in the ablation process. However, these

initial experiments indicated only a slight in-

crease in energy confinement.

plasma flow to the end-plug region and energy loss

by atomic processes in this region of ablated plug

material dominated the energy confinement at the

Recent experiments have been carried out with

end plugs constructed of lower Z materials in an

attempt to reduce the energy loss by atomic pro-

Tabulated in Table IV-I are all the ma-

terials used in the Scylla IV-P end-plug experi-

ments, along with their atomic numbers (Z), and

the ionization potentials from their ground-state

hydrogen-like (single-electron) configurations.

inserted.

using the

could be stopped.

center of the pinch.

LID (LITHIUMDEUTSRIDE)

BN (LkIRON NITRIDE)

S102(QuARTz)

cesses.

NAME

For

the

plug

surface was

end-plug region.

It appeared that

This hollow radial

surface of the plug.

The density very close to

sity profile, Fig. IV-10.

Side-viewing ruby laser in-

detailed characterization of the processes in the

Since the plasma column diam in Scylla IV-P la of

ferograms of the end-plug region exposed at two

different times in the discharge are shown in

inversion of these holograms that the high elec-

Fig. IV-9 for the boron nitride (BN) end plugs.

In addition, new diagnostics have been implemented

tron density resulting from ablation and subse-

of the column resulting in a annular radial den-

the surfaoe is characterized by a more nearly

terferometer was used to observe the high-density

quent ionization is concentrated at the periphery

in these experiments that have provided a olore

ablated plasma near the end-plug surface. Inter-

profile does not, however, extend axially to the

the order of 2 cm, it is evident from Abel

Gaussian radial distribution, Fig. IV-11, with

3.5psm!

— 1-

Side view interl’erograozsin the vicinity nitride end plugs.

r 2cm 1-

Fig. IV-9.

I.P. (HYDROGEN+KE)

l-;

2,G7(KEV)

of boron

2cm L

0,87 (KEV)

0.i2 (KEV)

0,01 (KEV)

I1.5ps

Plasma

5 cm

.—.

,’.

(KEV)

0,30

0,55

TAPLE IV-I

MATER[ALSUSELIFORENOPLUGS

ELEMENTS “Z”

BN Plugs

Fig. IV-10. (Upper left) Plot of the fringe count taken in a plane perpendicular to the plasma column axis. (Upper right) Fringe count corrected for the linear shift introduced by the background fringes. (Lower right) Abel inversion of the corrected fringe couit.

-. ,m+ ,

-..

>..

”-------

,— -“,,0 —.>!.”

5 x 1018 cm-3.

electron

15.5 Us

maximum

3,5 ps

’ ..-

,. : : :,.

.: ,,.,…

Boron Nitride Plugs

X=O @ Plug Face

‘fo= 4 x ‘0’6cm-’

At 3-5 US the line density is a substantial

fraction of that predicted by computer simulations

in which flow of the deuterium plasma to the plug

is stopped (also shown in Fig. IV-12), however, as

noted above, most of this density is near the edge

near the edge of the plasma column where the axial

pressure is less. The axial variation of the num-

ber of electrons per cm of column length, plotted

at various times in Fig. IV-12, decreases rapidly

with distance from the plug surface, with an

e-folding distance of about 1 cm or less.

formative time for the ablation layer is about

1-1.5 ps, and after about 10 ps, the electron line

density begins to decrease.

”

.—

—!.

l.m.

,= mm

,.. -‘mnanmm.ruu.

Z=OOI cm

EiFtl

i54 _

w ii

excess

-25

i I

)

The

(Malane,Marse)

X(cm)

Fig. IV-12. Axial electron line density inferred from side- view interferograms. Also shown is the line den- sity predicted by computer simulations.

Computation

FullIonization

Radius (cm)

Fig. IV-11. Radial density profiles obtained by Abel inversion of fringe counts in a plane 0.1 centimeters from the boron nitride plug surface.

densities

This suggests that although the

ablation is well distributed over the contact

region of the high-energy plasma column on the

plug surface, the flow of ablated material toward

the midplane of the theta pinch is less inhibited

the

main

body

10ss .

i.e.,no

perature of 100 eV.

geometry was measured.

2 cm L

“m

of the column and can have little effect on flow

Radial density profiles obtained from

Abel inversion of the interferograms, Fig. IV-14,

Furthermore, diamagnetic loop measurements of the

charaoterizad by a high electron density near

energy confinement at the pinch midplane were much

the column edge, however, with these plugs there

the game for these BN plugs as for the Si02 plugs,

ia now a significant increase of density in the

substantial improvement over

column.

lower Z of the BN plugs, atomic processes acted as

is very close to that predicted by Malone and

a major energy sink preventing the ablation and

Morse’s13 simulation of a plugged column with Z

ionization of sufficient plug material to impede

appropriate to these plugs, but again it must be

the dauterium plasma flow and associated energy

remembered that much of this plasma density is

Lithium deuteride (LiD) is the lowest Z solid

end-plug material available at room temperature,

Streak camera photographs taken in the region of

and the line radiation from this material should

the end plug indicate that a luminous front init-

be completely burned through at an electron tem-

ially propagates away from the surface with a

Side-viewing interferograms

1.4 X 106 cm/seo

obtained with the LID plugs (Fig. IV-13), indicate

IV-16). Later In time and at positions closer to

much higher density of ablated plasma. Indeed, ao

the midplane, spectroscopic measurements of con-

much material is ablated and ionized in the first 2 pa that the fringes in

tinuum radiation show a propagation away from the

plug with a velocity of 2-4 x 106 cm/see, Fig.

from the surface are obscured. It is also evident

IV-17. These velocities are less than half those

that, unlike the BN plugs, there is little radial

predicted for the ablation front in the computer

confinement of the ablated material near the plug

such

Fig. IV-15, at distancea far from the plug surface

outside of the outer radius of main plasma column.

the

open

surface.

interior

simulations.

velocity of about

the firat centimeter

the deuteri.um plasma.

Apparently even with the

-$- Plosma

vicinity

t~[+jjjq~

. rlm -

the

CI”-t.:.

SDectr~

The

line

could

density,

(see Fig.

propagation

’.-M-,.-

hi.th

,, ----

.—<

.A cl.?,.

-a e

o .

,,.,s-

-x

”

m,. ,M, $

Fig. IV-14. (UPPer left) Plot of the fringe count taken in a plane perpendicular to the plaama column axis. (Upper right) Fringe count corrected for the linear ehlft introduced by the background fringes. (Lower right) Abel inversion of the correoted fringe count.

the LiD plugs, diamagnetic loop-probe measurement

at the pinch midplane show a significant increase

identified spectroscopically with the PN plugs.

m2 cm

2.2 Jls

_l-

11.2JIS

Fig. IV-13.

Side-view lithium deuteride end plugs.

interferograms

EMISSION,

---2”0""

+""2-4’""s—

‘-m-z=

[/AS)

, ‘“2L.:

Z (cm)

4.5

N _

: \

-0

2.2JIS

o -

0.2

Zai -

Column ::~:123

…”… .:: … …”.

o~ o

electron

::.

COMPUTATION

Z (cm)

Fig. IV-15.

)

LiD

o—

Plugs

( Malone

ti Morse

TIME

RELATIVE

IC(8V/div)

CONTINUUM

IC(10V/div)

lC(70 V/d(v)

-t- 50 kG L

(Fig. IV-19), but

DiWsion

stabilizing

finement

inferred

using

plugs

have

that

‘“L

(p)

~ Cz+

Time

Fig. IV-16. Side-viewing streak camera photograph showing the axial propagation of a luminous front away from the lithium deuteride plug.

from Axial side-view interferograms. Also shown is the line density predicted by computer simulations.

density

line

Z= 84cm

lC, AT 522f3fi

30 \ 40 ’

Streak camera photo-

qualitative

different

materials

column

the

are

Fig. IV-17. Oscillograms of monochromator-photomultiplier con- tinuum monitor signals. The velocity of the luminous front inferred from the increase in these signals is indicated by the diagonal line.

in energy confinement over that obtained either in

open geometry, or with higher Z end plugs.

shown in Fig. IV-18, the energy confinement time

has been increased by a factor of three (from 9 to

29 ps) over the open geometry bY use of the LiD

plugs, which is significantly better than the

improvement observed using the higher Z end plugs.

‘his encouraging result is in excellent agreement

with computer calculations by Siemon and Miller

(Sec. D) of the expected energy confinement in a

column in which flow is stopped end the losses are

dominated by electron thermal conduction to the

ends and decay of magnetic fields.

Differences in energy con-

suggests that a comparison of measurements for

these experiments may yield clues to the improved

performance of LiD plugs.

graphs indicate that both high-and low-Z materials

influence

there

OPEN TE= 9.0 ~S

L__—_2

“o

TIME (~S)

FiE. IV-18. The energy line densi~y inferred from diamagnetic loop measurements in the theta-pinch midplane. The straight lines and time constants shown result from least squares fitting exponential decays to the data for times greater than six VS.

4.C

w z

IY w z Id

> 1- C5 z w n

differences in

with different

of the column

radiation.

observed,

early

time

the

The continuum radiation

energy loss mechanism

Finally, a

CAMERA

hLiDpLuGs

An apparent “bloomtng”

be an important

of radial den-

Comparison of

the performance

comparison

should not

the line

compared

theory.

plugs.

nearer

plugs.

which

plugs

much

that

with

(see

for

not

yet

characterizing the open geometry.

axially resolved continuum radiation measurements

for open-ended and plugged geometries show that

the continuum emission near the plug surface is

lower for LiD than for BN (Fig. IV-21). This is

probably attributable to the lower effective ionic

charge of LiD and the lower temperature with BN.

understood. Comparison of the energy line density

(measured at several axial positions with diamag-

netic loops) for high- and low-Z

Fig. IV-20) shows the strong cooling effect of the

higher Z plasma near the plugs. With LiD plugs,

on the other hand, the axial energy distribution

materials.

reason

the column behavior when plugged

STREAK

QUARTZ PLUGS

of LiD over higher Z

By using LiD end

l—w —l

STEREOSCOPIC

PHOTOGRAPHS

Fig. IV-19. Stereoscopic streak camera photographs taken with an open geometry and two different end-plug mater- ials.

sity profiles extracted from side-viewing inter-

ferograma (Fig. IV-22) indicates that there is a

much higher density of ablated material propa-

gating away from the plug surface with the LiD

These observations strongly suggest that

materials results, at least in part, from less

plasma energy deposition in atomic processes in

the ablated material, leaving sufficient energy

flux on the surface to efficiently ablate the ma-

terial required to retard the deuterium plasma

flow into this region.

In summary, the use of LiD end plugs gives a

dramatio increase in energy containment aa com-

pared with the higher atomio number Si02 and EN

For plug materials with Z greater than

three, atomic processes do not allow the buildup

of plasma density in the ablated layer necessary

for expected plug behavior.

plugs, we have succeeded in improving the energy

confinement time of a thermonuclear plasma by a

factor of three, in excellent agreement with

at late times with LiD plugs is

improvement

BORON NITRIDE PLUGS

LITHIUM DEUTERIDE PLUGS

3 -

1.0

0.5

__l

2,5

2.0

1.5

.- (/l z

: z—

4.8 ~S

2.4 US

: a -1 u a

CONTINUUM

Li D PLUGS

END PLUG
POSITION

20 ps

2.0

1.5

[

Z(cm)

AT 5226A

T= IOPS

Fig. IV-21 .

Measurements of continuum radiation near the end of the compression coil at 10 us after primary bank initiation.

AXIAL POSITION (m)

Fig. IV-20. Diamagnetic loop measurements of the plasma energy line density, (The axial coordinate zero corres- ponds to the end of the compression coil.]

RADIUS

Fig. IV-22.

profiles 2.2 centimeters from the

0 ”

“E 04

Radial density plug surface.

REFERENCES

(cm)

Z=2.2 from surface

E. M. Little, W. E. Quinn, and G. A. Saw]er, Mpla=rnaEnd ksses and Heating in the Low-Preaaure Regime of a Theta Pinch,” Phya. FLuids &, 1168 (1965).

R. P. Gribble, W. E. Quinn, and R. E. Siexr.r,, “Plasma Experiments With a Three- Meter ?- Pinch,” Phys. Fluids ~,

2042 (1971).

K. S. lhomas, E. W. Hsrri.e,F. C. Jahoda, G. A. Sawyer, and R. E. Siemon, “Plasma Experiments on the Linear Scyllac Theta Pinch,” Phys. Fluids~,

1314 (1974).

J. P. Freidberg and H. Weitzner, “fidloss from a Linear O-Pinch,” Nucl. Fusion~, (1975).

J. U. Brackbill, M. T. Menzel, and D. C. Barnes, “Numerical Studies of the Linear Theta Pinch,” Third Topical Conference on pulsed High-Beta Plasmas, Culham, England (1975).

K. F. McKenna and T. M. York, “End Loaa froc a Collision Dominated Theta Pinch Plasma,” Phya. Fluids ~,

1556 (1977).

A. W. DeSilva, “Thermal End Loss from a Plugged Theta Pinch,” Loa Alamos Scientific Laboratory report LA-6980.

T. K. Chu and L. C. Johnson, “Conduction Heat in Open Field Line Geometries,” Loss Scaling Phys. Fluids 20,1684 (1977).

S. I. Braginakii, “Transport Processes In a Plasma,” Reviews of Plasma Physics ~, 205 (1965).

R. J. Commisao, C. A. Ekdahl, K. B. Freeae, K. F. McKenna, and W. E. Quinn, “Solid-End-Plug Experiment on a 8-Pinch,” Phys. Rev. Lett. 39,137 (1977).

R. C. Malone and R. L. Morse, personal communication, 1977.

F. C. Jahoda and R. E. Siemon, “Holographic

Interferometry Cookbook,” Los A.lamesScienti- fic Laboratory report LA-5058-Ms (1972).

MI.&L controlled Thermonuclear Research

Program January—December 1976,” Loa A.Lames Scientific Laboratory report LA-7082-PR (1978), p. 67,87.

The Staged Theta Pinch (STP) is a 4.5-m-long,

22-cm-bore theta pinch designed to studY the

physics and technological problems associated with

separate capacitor banks for implosion

heating and adiabatic compression. The STP uses a

low-energy, high-voltage capacitor bank (PFN I) to

high-energy capacitor bank to provide a variable

compression.

low-energy, high-voltage capacitor bank (PFN II)

is available to shape the implosion magnetic field

and to assist in containing the plasma before it

contacts the wall of the discharge chamber.

expe~~iment produces high-temperatu~~e plasmas with

a much larger ratio of plasma radius to discharge

tube radius than conventional theta pinches which

utilize a single capacitor bank for both shock

heating and adiabatic compression.

the

and

using

plasma

produce

INTRODUCTION

amount of adiabatic

Theta-Pinch experiment and a

the end.

complete

the lens

windows

—

C. J. Buchenauer, J. N. Downing, A. R. Jacobson, E. M. Little, K. S. Thomas

Fig. V-1. Schematic layout of the Thomson scattering experi- ment.

The

COIL

~lkkm

second

POLWNWKR

the Staged

LIGHT ~T&TION

u RIINl#l

lower voltage,

STAGED THETA PINCH

five-channel polychrometer.

‘lowriderftwas stored

were part of the

plasma damage to

being focused by

paragraph.

Data were

The - 3-J

figure.

Pyrex

wby

OBM#ATION

POfi~~fPLASNA

each channel

system

along

nm,

The

was

The

the

was

‘Icrossn

steel

The five polychrometer channels had band-

widths of 692-688 nm, 688-681 nm, 681-670

670-659 nm, and 659-644 nm, respectively.

measured spectral response of

background, which was obtained from the average

background over several shots, was subtracted. It

description

vacuum chamber. The Pyrex windows were mounted at

summary of results

the Brewster angle to reduce reflections. (The

through 1976 was given in the last annUal rePOrt.l

windows are shown 90° from their actual mounting

The period covered by this report was sPent

position.) Scattered llght was collected with a

constructing, checking out , and collecting data

with a Thomson scattering apparatus.

aligned using a wheeled carriage (called the

collected at two axial positions, one near the

‘!lowridertr)which could be moved

center of the main compression coil and one near

discharge tube inside the vacuum chamber by means

At the center of the coil, data were

of external magnets. The same magnets caused a

taken at three radial positions; at the end, data

movable vane to be raised when the ‘lowrider!!was

were taken at four radial positions. Measurements

in its alignment position. When not in use, the

were made as a function of time with initial D2

stainless

fills from 3 to 15 mtorr. (At the end the range

the end of the vacuum chamber.

was 5-15 mtorr because of the rapid decrease of

The outputs from the five polychrometer chan-

density with time at this position.) The results

nels were displayed on oscilloscopes. The height

of these measurements, as well as a description of

of the signals from the channels used for a plaama

the experimental apparatus, are given below.

discharge was measured on a digitizer and a

B. THOMSON SCATTERING APPARATUS

experiment is shown in Fig. V-1.

laser beam Waa expanded before

shown in the left of the

were required because of

and beam dump when these

A schematic layout of the Thomson scattering

was then analyzed as indicated in the following

to compute the relative intensities which

TABLE V-I

would be observed for different electron tempera-

PLASMA PARAMETERS - STP

perature. The observed scattered intensities were

Electron temperature (eV) 137

fitted by computer to these calibration curves to

give a least squares fit.

Electron temperature (eV) 102

Table V-I gives expected plasma parameters for 5-,

plasma-circuit

15-mtorr fills.

tures with 694-rimexcitation and 90” scattering.

Curves were constructed from the computed relative

channel intensities using the normalization that

the sum of the squares equals unity for each tem-

EXPERIMENTAL CONDITIONS

All data were taken with voltages of 100 kV -

PFN I, 90 kV - PFN II, and 20 kV - Staging bank.

These conditions were chosen because they allowed

control of the plasma over a wide range of initial

The magnetic field waveform for

Io-mtoFr fill is given in Fig. V-2. The waveform

for other cases was approximately the same except

for the lowest fills where the first dip was

used

filled

fills.

fall of

7-, 10-, and

field.

’

-o

> ~

i?j

:G 4 -

(measured)

(calculated)

Fill (mtorr)

interactions.

The parameters,

Density (1015 cm-3)

Magnetic field (kG)

Ion temperature (eV)

the peak of magnetic

as measured on axis.

as follows: At all

PLASMA RESULTS

data were

fills at

(< 100 ns)

one-half

rapidly

its

Tlma (us)

except Te, were taken from data collected in the

Figure V-3 shows measured radial

density profiles taken at

The profiles become more diffuse at lower

fills and, although data were not available, the

trend probably continues at fills below 5 mtorr.

This trend may be related to some of the results

presented in the next section.

Fig. V-2. Magnetic field vs time.

6.0

108

416

-10

254

3.5

2.3

1.5

(Z . 0 cm Ls the center

This occurred at a

4-mtorr case where

During this period

po.Yition8,

approximately

positions

value

rose

R~DIUS (cm)6

Fig. V-3. Radial plaama density profiles for different D2 fill pressures.

Figures V-4 and V-5 show the time dependence

the electron temperature for different initial

the two different axial

Z = 38 cm and Z = 206 cm.

of the compression coil and Z . 229 cm is the

end.) All measurements were made on axis with the

exception of the Z . 38 cm,

included for R . 3.3 cm because of the

lack of data on axis. Electron temperature time

histories for other off-axis cases were the same

In many cases Te was measured out to 50 us

after discharge initiation.

the temperature showed a slow decay tO 0.5-O.7 Of

its peak value. The early time behavior of Te was

radial

peak value.

time which corresponded to the arrival of the im-

plosion density front. The temperature then rose

----+ -”—“-T-----i”—”..

Tim, [“1]

Fig. V-4. vs time for five D2 fills at Electron an axial position near the center of the compres- sion coil.

[

.,l

‘i

.” 0

temperature

discharge conditions.

slowly

the

/ A

‘1

[

. .

D2 Fill

perature

diffusion at early times.

initial condition.

tures could also

R. Siemon.

apparatus

the code

excluded

magnetic

could

field

flUx

not

Fig. V+.

fills at Electron temperature vs time for four D2 an axial position near the end of the compression coil.

compression

increased. In some cases at higher fills the peak

temperature occurred after peak magnetic field.

This is interpreted as an impurity effect because

its behavior was related to the cleanliness of the

Figure v-6 shows the dependence of T= on

initial D2 fill. The open circles are the temper-

ature measured at peak magnetic field. For those

cases where the peak temperature occurred later,

the peak temperature is plotted as dark circles.

As can be seen from the figure, the electron tem-

Tim.IA,,I

Field

Dashed

Mean Moximum Tmmperoture

Mean Temperotufe Peak Uagne?lc

Reasonable

electron

used.

and

the

Press ure(mTorr]

Fig. v-6.

Electron temperature vs initial D2 fill. line is prediction of a computer model.

increased rapidly for D2 fills below

5 mtorr. A correlation was found between the rate

of magnetic field diffusion as measured by the

temperature; the faster the field diffusion, the

higher the temperature. Higher electron tempera-

be produced by lowering the pre-

ionization capacitor bank voltage.

The predicted electron temperatures, assuming

heating of the electrons by the ions and classical

thermal conductivity out the ends of the experi-

ment, were calculated using a code written by

The temperatures predicted by this

code are plotted as a dashed line in Fig. v-6.

Below 5 mtorr the plasma parameters were not known

assumptions about the parameters would, however,

predict a slight decline in Te at lower fills.

The higher measured temperatures at lower fills

are, therefore, probably due to anomalous heating

of the electrons by the rapid magnetic field

To test the effect of

extra heating during the implosion, the code was

run with the measured electron temperature as the

The code predicted a rapid

two at Z = 38 cm and three

data

dependence

are plotted in Figs. v-7 and v-8. As can be seen

from Fig. v-8, the temperatures drop off rapidly

around R = 5 cm near the ends but not at the

center of the compression coil.

the radial scan was limited to 5 cm at the center

becauee of the geometry of the apparatus.)

The ratio of peak electron temperature at

Z . 38 cm and Z = 206 cm was calculated for 5, 7,

10, and 15 mtorr and the results were compared

theoretical model

result is shown in Fig. V-9.

plasma contact with the ends of the machine is

assumed

Fig. V-1. The horizontal error bars indicate the

range of possible contact points.

When the initial calibration of the poly-

chrometer was set up, the gain of the various

channels was adjusted to give an output propor-

tional to the channel width.

the

For

five

v-6.

Fig,

The

radial

study

cases,

with

to the

sufficient

aPProach

(< 5 WI)

Z = 206 cm,

uncertain.

Electron fills at an axial compression coil.

} I

temperature

temperature vs radius for three initial at a position near the end of the com- Coil.

7mTorr 10mTorr 15mTorr

Radius (cm)

Fig.

computed

Te.

were

given

W ~ O

values

-101234

ehown

point

available

The results

(The extent of

T. K. Chu.2 The

The position of

Therefore, the sum

Electron D2 fills pression

for two initial D2 near the end of the

[@lJ2 2/7

density.

plasma

Z=38cm

Z=206cm

v-8.

a mea8ure

plotted

!!!li#(:

sums

were

Fig. V-9. Axial temperature profiles - theory and experi- ment.

of the intensities of the channels was

When it was discovered that

the plasma column near the end had a different

density time dependence from what it had at the

center of the coil, the

RELATIVEAXIALF’OSITIONZ/256(cm)

7 mTOrr

Rodius(cm)

Fig. V-7. vs radius

position

The electron temperature measurements on the

STP experiment yielded the following result8:

The functional dependence of Te on plasma

Fig. V-11. Density vs time, as measured from the scattered light signal, for four initial D2 fills at a position near the end of the compression coil.

different from the one predicted by a theoretical

tures at lower fills can be explained by anomalous

model which asmimes that Te is dominated by

heating during the implosion phase.

classical energy input from the iona and clasai.cal

decay in temperature requires, however, either

thermal conductivity out the ends of the compres-

continued anomalous heating or

measured temperatures at the

conduction at the lower D2 fills.

higher initial fills may be an impurity radiation

Te as a function of radius decreases more

effect which would be expected to increaae with

slowly at the center of the machine than at the

higher plasma densities if the number of impurity

This is probably a result of radial heat

The higher measured tempera-

If electron heating is proportional to n2

This gives

lengthy

density

plotted

density

available at a particular D2 fill. The data for

Z = 206 cm are probably more accurate than for

Z . 38 cm because more care was taken to make the

and V-11 for data taken on axis.

dependence of the density was the same at other

noticeable for the Z . 206 cm, 5- and 7-mton’ fill

are

the

after

The

drop

cases.

time.

radii.

observed.

SUMMARY

density

The results

measurement

data meaningful

function of

included only when

sion coil. The lower

parameters (see Fig. v-6

t,o~o

Density light position

ions is constant.

vs time,

R.O, m one Z.38cm

signal,

four

for

and

R. Ocf” one Z.206r.m

40,

were

was

run

V-I)

data

were

Data

Table

time.

variations

The time

qualitative

particularly

in Figs. V-10

scattering measurement.

at lower D2 fills.

is in the center.

beta of the plasma,

to be an important

1684 (1977).

REFERENCES

at coil.

collisions)

(ion-electron

D2 fills

phenomenon

scattered

ends.

flow.

~ r,,, P,.,..,,

the

(-1.<,,

radial

Fig. V-10.

as measured from

initial

near the center of the compression

i I

The slow

~Flll P,,,,.,. {ml.., ]

reduced thermal

Because of the high

Los Alamos Scientific

is expected

Thomson

density

(1978).

this

heat

flow

give the observed results.

radial heat flow could

consideration.

The time dependence of plasma density is

different at the ends of the machine from what it

ends after peak field is particularly noticeable

The decrease in density at the

A detailed study

will require much more accurate

measurements than is possible with a

Thermonuclear Research Program, January-December 1976,n Laboratory

LA-7082-PR

report

T. K. Chu and L. C. Johnson, Phys. Fluids ~,

l-m* Id

K. S. Thomas, compiler, llLASLControlled

q q q staff on FRX-A have prevented rapid progress .x

However, full operation of this larZ~.

system is expected in the first half of 1978.

The only diagnostics used during 1976 that

useful were

B-field probe and the side-on streak camera.

Several new diagnostics have been added, and twc

of them were new developments. These diagnostics

are briefly described in this section and moat are

shown schematically in Fig. VI-2.

End-On Framing Camera.

single-frame Beckman-Whitley camera was used in

conjunction with an image intensifier (see Fig.

Two Corning Glass filters (#4-76, and

#3-70) were required to reduce the atomic end

molecular deuterium line radiation from the cold

plaama in the ends of the discharge tube outside

Later, the single-frame camera was replaced

with a multiframe Imacon camera. A 1-NHz framing

plug-in was modified to frame at the rate of 6 per

Two theta-pinch systems FRX-A and FRX-B are

being used to study the stability and confinement

the reversed-field configuration

VI-1.

the plasma is confined by closed poloidal field

lines supported by the toroidal plasma current.

If sufficiently stable, this RFC could be used to

increase the energy confinement in open-ended

plaama devices such as mirrors, theta pinches, and

The FRX-A and FRX-B facilities were produced

by modifying other experiments during the latter

half of 1976. ‘lhe 1976 Annual Report describes

these facilities and provides the historical moti-

vation for initiating the research.1

name of the experimental program has changed from

Field-Reversal Theta Pinch (FRTP) to the present

FRX.) The FRX-A system was completed, and initial

experiments performed in 1976.

produced the majority of the results presented in

‘IheFRX-B system was completed in

the spring of 1977, but certain deficiencies,

discussqd below, were discovered in the plasma it

produced. The concentrated efforts of the small

liners.

INTRODUCTION

properties of

(RFC) shown in Fig.

this report.

prove

R. K.

FRX-B .

VI-2) .

(Note: The

continued to

B. DIAGNOSTICS

‘his system also

In this configuration

VI. FIELD-REVBRSAL EXPERIMENT (FRX)

Linford, D. A. Platts, and I?.G. Sherwood

the compression coil.

[ SIDEON STREU CAMERA

POLYCHROMETER <

~-;

the

external

Initially,

-PLASMA JFEEDRATES

F MTER

:l!!l~---w

PRINGEINTERFERO- METER

L___21

PRAMINO CAMERA

51AIA

[

17vlEwlNG~Ts IN COIL

D ----

SPEC%%UETER

Cm, CX MEASUREMENTS TAKEN AT ANY AxIAL~ RAOIAL POSITION

Fig. VI-1. Reversed-field configuration.

Fig. VI-2. Diagnostic arrangement on FRX-A.

10 P.s. The previous filters were replaced by more

Internal B-Field Probe.

7-channel polychromator to measure the Doppler

6.35-mm diam, stainlesa tube waa positioned along

broadening and shift of the C V 2271-A line. The

the axis of the system (see Fig.

electrostatic shield were attached to the end of a

solid-copper-shielded coax cable, and the whole

insulated.

assembly (identical to the external probe) was

inserted into the stainless tube until the coil

reached the desired axial position. To prevent a

necessary to attach a series-tuned RLC network

between the stainlesa tube and ground.

network damped a 40-kV transient that occurred

during the initial breakdown of the preionized

J-Y Spectrometer. A

spectrometer has been used in conjunction with an

RMI 6256B photomultiplier to monitor the C III

2297-A and the C V 2271-A lines from the naturally

occurring impurity. As shown in Fig.

spectrometer is used to make shot-by-shot scans of

the axial and radial plasma profiles. It is alao

probe

the

was

probe

plasma.

structure

breakdown

efficient filters.

the internal field probe.

r NYLON INSULATOR

the next subsection.

STAINLESS STEEL VACUUM CROSS

L &~NJtmT;

rOUARTZ

The probe coil and an

function of time on a single shot. By moving the

VI-4 on subsequent shots,

various chords of the plasma can be scanned (l-cm

steps, l-cm resolution). Tf the plasma rotates as

rigid rotor, the component of velocity (or

Hence, the line shape is undistorted so

given

the position along

temperature as a function of radius and time can

The polycbromator is very similar to one

developed by Scott, et al.2 This one uses a

cylindrical quartz lens to increase the dispersion

so that an array of seven lP28/Vl photomultiplier

tubes could be used as the detector. The entrance

slit is adjusted to 40 pm by Che usual micrometer

optimum

intensity and resolution. The exit slit, however,

is fixed at 400 urnwith a 440-um-diameter quartz

J-Y

was

been

This

VI-2, the

adjustment

thin-wall,

insulation,

‘Ibis probe

be obtained.

quarter-meter

VI-3) to house

independent of

monochromator has

Polychromator. A

mirror shown in Fig.

accurate measurements

Doppler shift) along

seven channels produce

showing the the internal

STAINLESS STEEL TUEE FOR INTERNAL FIELD PROSE

used

the

and

that

Line

shape

carbon

between

rotation

compromise

the chord.

chord will

developing a

half-meter JarreJ.-Aah

This defocusing lens

PLASMA

COIL

.—

Fig. VI-4. A transverse plasma cr;ss section showing 2 of the 15 possible mirror postions for measuring ion temperature and rotation.

used to monitor the total light of the C V line

fiber cradled in the slit.

radiated from the same plasma location that is

being monitored by the polychromator described in

disperses the light so that when the tubes are

positioned 10° apart (staggered about an average

Fig. VI-3.

Axial cross section of the system position of the stainless tube for field probe.

DISCHARGE TUSE

COMPRESSION COIL

arc with a 12.4-cm radius) and when the grating ia

used in second order, the polychromstor resolution

proportional to the density.

is 0.355 A between channels. The angular blurring

due to the finite f number of the monochromator is

amplifier

C. J. Buchenauer,

The plans are to feed all seven outputs to

frequency vibrations are damped by using a rigid,

A/D converters and have a computer reduce the

compact construction and by using acoustically

At present, the outputs are sequentially

absorbing materials around the beam patha and

high-frequency cutoff of about 2 Hz.

sampled every 2 IIS by

A sample output of the device is shown in

msrker pulse followed by a sequence of seven

VI-7 along with the magnetic field trace.

pulses, which plot out the shape of the C V line.

By comparing these traces, several effects are

This pattern is repeated every 2 IIS during the

readily observable including the plasma formation

plasma discharge as is shown in Fig. VI-5.

the ringing preionization field, the fast

Feedback-Stabilized Fractional-Fringe

density increase during the implosion, and the

Interferometer. The density measurement

increase during

from the fractional-fringe interferometer shown in

The sharp density spikes near the end of

VI-2 were not very useful because these

the plaama life are caused by the rotational m-2

axial measurements included the density of the

instability discussed later in the report.

cold plaama in the ends of the discharge tube. To

sag in the base line that is evident at the end of

this problem, a

the plasma life is not yet understood, but it is

interferometer was designed by D.

under study. It does not occur if the machine ia

look transversely through the plasma diameter (see

fired without gas so that no plasma is produced.

Fig. VI-6). ‘Me apparent problem of a small phase

Pressure Probe. A pressure probe haa been

shift (< 1/8) for the 6328-A line (HeNe laser) was

turned to an advantage by using feedback to

stabilize the interferometer at the most sensitive

the end of the ayatem through a ball

the pressure of the plasma leaving

the compression coil region can be

function of

Fig.

avoid

data.

only 2.9°.

comparison of

The

Fig.

phase

phase.

slower

ayatem,

designed

feedback

obtained

position.

Platts to

The result is a

optical components.

a multiplexer /circuit

double-pass, Michelson

t..i ’ L

—<.--- -:—

joint ao that

inserted into

axially from

measured as

The

the

haa

and

then

axial

signal

output

The higher

compreaaion

speaker-coil

VI-6. interferometer design.

”- ” ” ”- … … … … … . .

radial

axial

’ .

”.*

and

…

.—

”

w ; . IIEiz?lll . —. iiils ..—.— ‘-3

-~’ .’”

. …=

…--

”

l.-

j&5PLITTER.

Fig. VI-5.

A the multiplexed polychrometor output (top trace) and the J-Y spectrometer output (bottom trace) for a chord at a 3-cm radius.

Fig . Transverse Michelson

Three major topics have been addressed during

discovered connection between impurities and two

modes of plasma behavior, which are typified by the disparate results of Bodin and Newton4 and Eberhagen and Grossmann.5 The second topic is the

confinement reported by Kurtmullaev et al.6 The

rotational m=2 instability, which presently limits

diagnostic study of the

position. The probe was designed and constructed

by T. M. York, based on previous work,3

aa some initial tests on the system with existing

probes. The final design, shown in Fig. VI-8(a),

2-mm-dism, O.S-mm-thick

probe

8-mm-diam quartz tube and the front is capped with

a 0.25-mm-thick quartz disc.

The probe output is sent through a

driver for impedance matching.

a shock tube is shown in Fig.

signal showa the step-function response of the

probe. Probe calibration indicates a sensitivity

The

uses

first

topic

crystal.

The entire

of 0.41 V/atm.

c. EXPERIMENTAL RESULTS

attempted duplication of

third

topic

the

~STAINLESS

CONDUCTIVE

LNIFORM

Fig. VI-8. Pressure probe (a) cross section, and (b) shock tube response.

the lifetime of the Rl?C. The experimental results

the SO-US of stable

OlSC

TUBE

~.25mm

QUARTZ

@UARTZ

the

‘he

~BEESWAX

line

L-lI-l""d

as well

VI-8(b).

is housed

accidentally

\ h?ANSD”CER

piezoelectric

A test trace from

behavior of the fractional-fringe the external mag-

mode” behavior.

the main bank.

reliability.

at 19.0 kV.

associated

turbine

vacuum

torr.

these

pump

With

summarized in this section.

three

TUBES

COAXJ

STEEL~

are

I cm

EPOXY

’\

subjects

The bearings on the

changed

from

oil

The

Initially FRX-A was constrained to operate in

the dirty mode because of limitations in the

vacuum system and main bank.

were

lubrication to greaae in order to improve the

‘Ibismodification also lowered the

baae pressure from about 4 x 10-7 to 5 x 10-8

As a result, the clean mode of operation

could be achieved at 19.5 kV 011 the main bank (ZI)

kV full charge), but the dirty mode waa produced

This discovery of the clean mode motivated a

series of improvements to the vacuum system and

These modifications decreased the

Impurity Effects. Two distinct modes of

plasma behavior have been observed in FRX-A.

only two parameters that appear to determine the

mode are vacuum cleanliness and the power input to

the plasma. This effect is shown qualitatively by

the graph in Fig. VI-9. The curve represents the

sharp boundary between the “clean mode” and “dirty

Fig. VI-7.

A comparison of the temporal transverse line density from the interferometer (top trace) with netic field (bottom trace).

critical voltage that separates the modes from the

comDared with earlier experiments.

19.2 kV to less than 10 kV.

correlation with impurity radiation barriers is

The horizontal axis is labeled ‘:powerinput,”

indicated, some difficult ouestions still remain:

RFC destruction in the clean mode is oaused by a

rotational m=2 instability at later times.

side-on-streak and end-on-framinu camera piotures

in Fig. VI-10 illustrate these differences. Note

the sudden annihilation of the central hole in the

followed immediately by

plasma luminosity and loss of axial confinement.

Note also the effect of the m=2 mode on axial

confinement in the clean mode.

There are strikine similarities between the

plasma behavior indicated in Fiu.

behavior observed in other experiments. The dirtY

mode is very similar to the results of Bodin and Newton,4 and the majority of the other e3rlY

experiments (before 1964). The clean mode results

EberhaQen and Grossman.5 Thus, these results

identical with

Drovide

long-lived results of more recent

Fig. VI-9. Effect of impurities on plasma behavior.

CLEAN Morx

cm MICRO INSTABILITY

POWER INPUT

DIRTYMOOE

I r< lop

transition from

shots.

The

are

help

FRX-A

almost

dirty mode

CY ROTATIONALM=2 r> 15ps

In contrast, the

clean mode is

VI-9.

(j.Js)

This

The

instead of ~lbankvoltage,!!because the capacitance

and inductance of the system have been varied to

demonstrate that the product B~ is the important

parameter. The boundary between the modes cannot

only be traversed “horizontally” by varying B~

(usually by the voltage) but also ‘vertically.”

dirty

observed each day during the first few clean-up

Unfortunately, the lack of a gas analyze?

prevents the quantitative labeling of the axes.

The differences between the clean and dirty

modes are partly summarized in Fiq.

behavior of the C III and C V radiation followinu

the implosion indicates that Te _ 10-30 eV in the

dirty mode and 70-150 eV in the clean mode.

large jump in Te arising from small changes in E&

is indicative that an impurity radiation barrier

may play an important role.

The other major difference between the modes

is that the RFC is terminated much sooner in the

dirty mode, and the field annihilation is caused

by an unknown instability. This mode is called a

‘microinstabilityfgin Fig. VI-9 simply because no

macroscopic distortion of the plasma is observed,

and therefore, the wavelength is probably small

compared with the plasma size.

the

The

from

link

those

between

Increased

Althouah a

reported bv

VI-10 and the

exoerimenta516

end-on-framinu

The

Shot Sequence

Fi&. VI-10.

Side-on-streak pictures and illustrating the difference between (a) the dirtY mode, and (b) the clean mode. The time of the framinu implosion pictures were taken at 1.6-ps intervals with first frame at t=2 ps.

corresponds

t=o.

axial-implosion-coil system, and (3) the operation

exhaust the possible combinations of first- or

of the main bank in the “second-half-cycle” mode.

second-half-cycle operation with a variety of

In spite of these similarities, the resulting

plasma properties are quite different as indicated

of the barrier field was varied between a single

in Table VI-I. The large difference in the ratio

sinusoidal pulse of 30-us duration to a typical

iS consistent with

crowbarred waveform, depending on the timing of

the crowbar. The timing of the barrier field was

However, the low plaama densities

varied from the beginning of the bias field to

are inconsistent with tbe filling pressures quoted

after the implosion, and the amplitude was varied

6 unless the percentage of ionization is

from zero to 400 G at a radius of J cm (950 G at 9

(1) What is the “micro-instability”and what is its

degree of preionization is achieved by either a

connection with radiation barriers? (2) Why did

successive bursts

theta-pinch PI, one at the beginning of the bias

TeZ 60 eV, and Bodin observe dirty mode behavior

field and one just before implosion.

Attempted Duplication of Kurtmullaev’s

Kurtmullaev system are now known.

Results. The FRX-A system operates in the same

attempt

additional differences between

filling pressure and B-field range as reported by

Kurtmullaev results were performed before this

Kurtmullaev6 and has a very similar main coil and

known.

preionization and bias fields were not described

listed

in the paper. ‘Ihemajor differences between the

barrier field was added to the system and the

Kurtmullaev system and FRX-A appeared to be: (1)

system was modified to operate second-half-cycle.

consisted of removing two of the three originally

the

clean

in Ref.

observe

barrier

of Te/T1

multipole

Rberhagen

belt-pinch

experiments

(< 1014cm-3).

at Te’-150 eV?

vacuum chamber geometry.

be less than in Ref.

FRX-A (CLEANUODfl

[lERHAOW: #WJLM

URL-REB 8YtTEM

IIURIMIMAIV

personal

C0&X14L

COMPARISON

Recent

l,Z-1.1

SYSTEM

low.

2.W4

31-60

0.1

experiments

The

two

the

The

low

(2)

mode

with

known

latest

field,

Z-pinch

density

initial

behavior

Numerous

differences

details of the

theta-pfnch and

information was

barrier-field conditions.

general, only

instability.

CAuww TtRMMnc4

differences

conductors.

This high

chamber

between

PARAMETERS

quartz

vacuum

1/SOSCAV

’. #v—

cm).

that

s&-la

16-21

WARlm

wrnJslOM

180

M-2

U.z

still

the

were

This

above.

attempt

hexapole

Thus two

performed

duplicate

FRX-A and

The temporal behavior

The end-on framing

Under the present

differences were

supports

the m=2

between

At the

exist

that

mis

the

barrier field at the time of implosion.

optimum pressure the RFC life time was still 20-25

us and the termination waa still caused by the

rotational m=2 instability.

camera also showed a hexagonal pattern in the

“halo” around the main plasma, but the main plasma

remained circular until the onset of

No more detailed results are given here

because of tbe lack of positive results.

lack of success may be associated with either the

experiments or because of the comparatively small

size of the barrier field.

design, the barrier field cannot be increased

substantially without the risk of crushing the

communication

minor

Kurtmullaev and A. R. Sherwood resulted in new

observed in the plasma behavior. The major effect

information. The fill pressure is now claimed to

was to increase the optimum fill pressure (< 40%)

preionization is essentially 100%.

effect increased according to the strength of the

6 (< 1 mtorr), and the

for maximum plasma lifetime. The strength of the

TABLE VI-I.

OF REVERSED -FI13LDPLASMA

n 2rr dr,

: -B(r=O) for B(r=O) CO

attempts to duplicate the Kurtmullaev experiment

major

understanding warrants the required effort.

focus of the present research

equal to half the toroidal asDect ratio.

effective minor radius a is defined bv

parameter K can also be shown to be approximately

experiment.7

control,

The modification of the barrier field and the

other

rotational

including finite-Larmor-radius (FLR] effects, has

accurately predicted the m=l mode [wobble] in the

substantial amount of effort.

lowering of the filling pressure would require a

major modification of the preionization system and

the development of new diagnostics compatible with

the lower density. In addition, there is not much

interest in a plasma with as hiuh a ratio of TelTi

as is obtained at these low densities.

the

and

IV-P

code

Scylla

removal

however,

Rot~

modifying

hopefully

considered;

instability.

theoretical work

will be made only

E ;+==—.— Zz

independent of radius.

rigid rotor.

-nE-l-n=-S)*

B= BmtanhK

sech2 K

where

term

each

n=n

‘ei

---1 R2

r2()

r2

R2 ()

‘i

the

The

m=2

The

the

The

the

are

one

then

RFC.

each

R —. a

study

beinu

where

requires

Future

recent

~ n(r=R)

where R

1 + tanh K

rotational

differences

Bm = B(r>>R)

nm 2nR(2a) 3

expected gain

instabilities,

J. P. Freidberg

For example, the

is to understand,

the temperature is

frequency

angular

choice

Thus

and

(2)

(3)

(1)

tbe

-1

ratio and degree of

diagnostic study described in this section will be

correlated with this theory in an effort to

understand the m.2 instability.

The theory assumes @at

1 ani

ZerBz ni M

two

(6)

(5)

The

(4)

If an

radius.

toroidal

parameters R and K

field reversal.

DensltV

several

independent of radius and that the rotation is

profiles are plotted in Fict. VI-11 for

This allows the equilibrium ion

values of the aspect ratio.

equation to be written in the followinu form.

The rigid rotor profiles are given by

NORMALIZED RADIUS ~

FiK. “1-11. Riuid rotor densitv profiles.

completely determine the profiles inoludinu aspect

-0.30 < a < 1.45 ,

profile, the pressure gradient drift velocity is

For the WC

temperature is independent of radius.

The stability analysis for an ordinary theta

8 x 105 K Ur (2 x 106 has been

pinch7 for f3.1indicates that the plasma is stable

to the m=2 for rotations given by

b, ~, The rotational velocity of the C V

In addition, the theory predicts that the real

measurements, all of the results can be summarized

part of the m.2 mode frequency at the critical a =

by the normalized rotation plotted in Fib?. VI-12*

This behavior is in qualitative ameement

RFC .

theta

IDE-Q

normal

pinch,

1.45 is

given by

where a = - .flli2*.

or end-on-framing pictures.

determine the value of Ur.

characterize

VI-10(b)).

the mode.

obtained

above.

plasma

major

Fig.

from

m.2.

the

‘r”

the

The

onset

accuracy

critical

Q*.

(?) isac~

possible

For ion temperatures Ti < 300

eV, the rotation appears to be rigid rotor and the

assumptions

theory.

consistently show that the C V does not rotate

significantly for the first 8-1o ps, but then it

the

(9)

(8)

(7)

with

been

that

until

scales

theory

Within

because

measured

expected

ions has

described

satisfies

observed.

previously

experiment

instability

accelerates

rigid rotor

instability.

and the ranKe

described above.

theoretical value

rotation uiven by Eq.

the observed ac N 0.4.

However, the end-on

fractional-fringe

rapid growth of

from a barely

discrepancy,

associated

general

can be

o-f o

0.4—

0.2—

this

with

and

the

—

the

for

m.2

the

The

the

with

results

critical

However,

velocity

Thus the

available

difference

normalized

polychromator

1.45 instead of

difference

TIME (#s)

Fiiz.VI-12. Normalized C V rotation as a function of time.

Although these critical values of a and ~r are for

qualitatively similar behavior will exist for the

may be associated with the difference between the

The following measuremetikshave been made to

possible

determine the values of Ur, .Q,and~*.

As stated above, the code is being modified to

‘The critical value of Ur

handle the RFC to remove that DOSSible source

rotational velocity between the C V and deuterium.

RFC and the standard theta pinch, or it miuht be

This numerical difference

interferometer, J-Y spectrometer, side-on-streak,

rotational velocities is briefly discussed here.

pictures give the best measure of Or at the onset

inaccuracy of

measurement arises from the very

The ellipticity grow

detectable value ( ..,1.2)to about 4 in a time

corresponding to half a revolution, t.?ln+.‘1 (see

This corresponds to a Rrowth

rate uZ wr/2, which is so fast that the mode

becomes nonlinear in the time needed to accurately

From the evidence so

far, it does not appear that Ur changes more than

50$ during this time, so that an approximate value

can be obtained from the duration of the first

period from any of the diagnostics mentioned

A complete parametric study has not been

made, but Ur normally increases with increasing B

by assuming that initially the

same rigid rotor profile. The

~ 1/2

between impurities and deuterium is more

300 eV the expected decrease of line width is

rotational

samplins?technique of the multiplexer). When T >

than the thermalization process.

drag force arising from a difference in the

Since 7E is known to be short compared with

rotational velocity causes the impurity to drift

the plasma lifetime, Eo.

radially in the -V B2 direction.

C V should be rotatinu at the same sDeed as the

continues until the narrowing density profile of

In other words, the polychrom~tor measurement

the impurity compensates for the higher Z, i.e.,

should yield the correct value of the critical Q.

until the Q* of the impurity and the deuterium are

(1) and Eq.

It is apparent from Eq.

independent parameters are

The relaxation time has been calculated by

assuming that the drag force is proportional to

evaluate Q*, i.e., Ti, B~Y temperature can be determined within 5fI0eV by the

the ratio of the azimuthal velocity difference to

measurements

the classical slowing-down time, (VBD-VW)/TS, and

Except for the previously mentioned C V ec!uilibra-

-2

()%

The

process

equal (see Eq.

T = 1.39 x 10

combination of

B field).

time TE.

consistent

ratio

these

rapid

Fig.

The

tion effects, the ion temperature is independent

of time and radius. The mamitude of Ti soales in

the expected ways with B and the fillina Dressure

and the observed ranue of values is 150 CTi <400

The value of BW is measured bv a calibrated

probe located outside the quartz tube at r.11 cm.

Because of the finite plasma lenath, this value of

B may be sliRhtly smaller than the field just

Most of the dats have been

between 4 and 6 ko.

observed at early times.

eV.

D+.

four

(lo)

(2)).

3h_f&

complex

This drift

result yields

polychromator

equilibration

The azimuthal

C V and D+ have the

outside the plasmz.

taken for values of B

temperature prevent

relaxation

the RFC.

This iS

error.

times

have

been

(11)

With

The

the

the profile parameter K.

definition of

above.

The ion

(8) that

described

R, and K.

necessary to

diaanosti.cs, the

(11) indicates that the

A rouqh estimate of the

described above.

by Eq.

aiven

best

(4)

The

The most difficult Darameter .to determine is

aspect ratio R/a can be made from the end-on

photographs, but the nonlinearity of both the film

and the luminosity dependence on the density and

an accurate determination.

approximately 1 for the observed plasmas (PD is

The result of such an estimate yields 2 < R/a C 3,

the deuterium thermal gyro radius in the external

or 0.8 < K < 1.4 by usina Eq. (6).

calculated values

provides another method of determining K.

confirmed experimentally. Polychromator data like

ratio of Bi/B~ is usually about 0.85 as determined

VI-5 show that the C V temperature is

internal probe

independent of time when T < 300 eV.

stainless tube that houses this probe was not left

with the calculated TE < b ps because

in the system after these measurements were made

data cannot be obtained from the polychromator for

because the hot plasma boiled substantial material

about 5 us after the implosion (because of a

off of the tube resultinu in a premature loss of

light increase and

Thus , the internal field measurements

where ‘rris in ps, !4in amu, T in eV, and 6 in cm.

The density gradient scale length 6 is essentially

existina

1.18 a (the minor plasma radius) for the aspect

method of determining the major radius R is from

ratios of interest (see Fig. VI-11). Usinp the M

the end-on-framina pictures. If the luminosity Is

and Z values of C V and D+ yields the following

assumed proportional to some power of n, then R

comparison between Tr and the energy equilibration

can normally be determined within less than 1 cm

were only taken for a limited number of plasma

However, the typical value of 0.85

corresponds to Ic=I.26and R/a=2.8, which supports

the values obtained from the end-on pictures (see

Fig. VI-10 and Fig. VI-11).

also

measuring the intensity of the C V 2271-A line as

a function of radius with the J-Y spectrometer

The results for two different

bias fields, but the same Bo, are plotted in FiK.

These curves appear to correspond with an

obtained by, visible light and internal field

1.3 instead of the 2 to 3 ranKe

profile

artificially broad because of a slightly hitzherTe

in the sheath region compared with the peak

This slightly higher Te could

result from the ohmic heatinR in the sheath.

Since Te is much less than the 305 eV needed to

For

C V

The

was

Would

large

result

VI-13.

VI-2).

profile

(see Fig.

intensity.

conditions.

measurement.

density region.

aspect ratio R/a-

The probable range of

theta-pinch model.

side-on-streak

compressed

same range

with Eq.

d. P&ma

obtained

are not

camera.

profile

camera

from

size

and

the

may

1iuht

appear

investigated

about 1.5 x 1015 when R ‘4

are compatible

interferometer

end-on-framing

implosion was

detectable C

Fig. VI-14.

coil region.

photographs

indicated

complete.

include

plasma

When

Some

cold

the

was

the

1S1

—

t = lops

l aA

0.3kG BIA! LOkG 61A

fractional-frinue

were made.

radiation

after the

density

exist.

mi~ht

scans

line

was

excite the 2271-A line, slight increases in Te

RADIAL POSITION (cm)

increases

example, if Te(r.R).lOO eV and

Fiz. VI-13. Radial profiles of the C V 2271-A radiation.

Te(r.Rfa)=130 eV, the fraction of the electrons

capable of excitation would be 10$ at r.R and 20%

at r.R~a, which is more than sufficient to explain

confirmed by C V scans such as the ones shown in

the broader profile Ln Fig. VI-13.

Since C V is only present in relatively hot

In spite of the general aareement between

plasmas (Te > 70 eV), the C V profiles would not

these three profile measurements, there is still a

that

substantial uncertainty in the value of K.

determine whether substantial density exists in

other diagnostic such as Thomson scattering is

the Term of cold plasma, C 111 scans

needed to accurately determine the radial profile.

(2297-A)

Q* usinu compatible

observed anywhere inside the coil region

combinations of the parameters Ti, Bm, R, and K is

Similar radial

7 x 105 < Q* < 4 x 106. This is essentially the

indicated substantial C 111 radiation oriuinatinu

as the observed critical value of Ur.

from the ends of the discharue tube outside the

Thus the observed values of u#p

(9), even though it is for the standard

transverse

developed,

10ss. When the measured values of

measurements were made throuzh a plasma diameter

WV ‘i) and ‘e relation, the resulting peak density nm is 2 to 3

are used in the pressure balance

at the mid-plane of the system (z=O). The results

confirmed the low density implied by pressure

times smaller than would be obtained if all the

balance. For a 5-mtorr fill, the peak density is

initial D2 fill inside the implosion coil were

cm. Unfortunately, an

The uncertainties in the radial profile

To better understand this apparent loss of

enough to explain it, and the axial

plasma, a pressure probe has been used to measure

streak

the flow of plasma out of the end of the svstem.

axial scan with the interferometer has not yet,

been made to further confirm the axial profile.

AXIAL IWSIT!ON (cm)

Fig. VI-14. Axial profile of the C V 2271-A radiation.

FIR. VI-15. A comparison of the pressure probe siznal obtained on axis 8 cm from the end of the coil (top trace) and the external magnetic field (bottom trace).

B’

!j=

-lo

the probe.

’:~-40 -20 -20

density measurement.

observed in Fig.

second burst.

Preliminary

ps) pressure

plasma.

Even

the

A..

into

s040

t. lo/18

plasma

behavior

operation.

l 0.3kGSIAS

l,OkG BIAS

destroys the RFC.

open-ended system.

of the plasma

measurements

pressure

profile.

these

These

mode

A typical trace with the probe on axis and 8 cm

from the end of the coil is shown in Fig. VI-15.

The slight sag in the base line is caused by a

plasma-generated acceleration or thermal strain of

The pressure associated with the PI

and the fast 2-ps spike at the time of implosion

is observed at all radii. However, the long (-10

signal just following the implosion,

and the second long burst following the m.2

breakup of the RFC are only observed for r < 3 cm

if the probe is within 6 cm of the coil.

jets of plasma spread radially and hit the wall

about 15 cm from the coil.

time-of-flight

indicate that the first long pressure burst is

oaused by a plasma of comparable temperature and

slightly larger mass than that which causes the

It also appears that the initial

loss occurs during the axial contraction phase of

though

measurements help explain the low plasma density,

they apparently conflict with the density increase

VI-7 during axial contraction.

One possible explanation is that ionization is

taking place during the axial contraction process.

More detailed density and pressure probe studies

should help to determine the origin

loss and resolve the apparent conflict.with the

two

modes

distinct

5 6 The “microinstabi.l-

typical parameters of

is needed to explain

on the tearina mode,

The sausace

cm/ps.

CLINCLUSIONSAND DIRECTION OF FUTURE WORK

A radiation barrier appears to divide the

This result provides a link between

the early short-lived experiments and the recent

longer lived experiments. Y

i.ty” that terminates the RFC in the short-lived

dirty mode needs to be identified and studied.

The longer lived clean mode is typified bv

almost constant plasma pressure being confined for

up to 25 ps, when an m=2 rotational instability

The

n - 1.5 x 1015 cm13, Te - 100 eV, and 150 < Ti <

400 eV correspond to a sound speed C~ z12

Thus even with the instability, the plasma is

confined for 6 times lonjzerthan it would be in an

More theoretical work

the absence of tearing and MHD instabilities. In

particular the recent work

which has included the effects of diffusion,8

f10w,9?10 and other n6nlinear effect~,ll,12 need

to be examined in liuht of the FRX-A results.

Most of these effects tend to be stabilizing. The

MHD modes also need to be studied.

is stabilized by the appropriate diffuse

profile, which may be satisfied by the rigid rotor

The kink mode may have a substantially

reduced growth rate due to the highly elongated

cross section of the torus (see Fig.

either model were

can easily be shown that an infinitely long torus

particle confinement would delay the onset of the

the kink.

instability. For this reason, a careful analysis

geometry-modified growth rate may be too slow to

of the end loss, observed by the pressure probe,

see during the 25-va lifetime.

is imperative. As previously mentioned, this loss

alao be important. It is also possible that the

may result from poor preionization or from snme

the axial contraction,

anomalous process that allows particle loss while

bouncing, andfor rotation may have a stabilizing

the plasma is contracting to its equilibrium

effect. All of these possibilities need careful

length. The obvious remedies are to improve the

theoretical consideration. Experimental investi-

if that is at fault or add axial

gation may become more practical in the future if

implosion coils to drive the plasma more rapidly

the stable plasma lifetime can be extended long

toward its equilibrium shape.

enough to observe these modes.

changes would increase the similarity between the

The major thrust of the present research is

FRK systems and the Kurtmullaev experiment, thus

instability. The existing theory7 for rotational

If the line-tying argument is correct, then

increasing the motivation to make these changes.

hopefully,

for

It is being modified to describe the

open-field lines by some type of end plug may

deficiencies

limit the rotation to the stable regime.

are the lack of accurate information

approach was auccesafully used in Scylla IV-P to

stabilize the related m=l wobble, 13 and will be

the

The

two

and

motion

nnly

and,

RFC.

study

valid

about

modes

msjor

density

columns.

profiles.

diagnostics

associated with

is neutrally stable

in the opposite direction.

electric field (see Eq.

measurements are

correlation

scattering

velocity.

between

needed

should

solve

the

m=2

the

use

the

Thus

‘lhis

VI-l).

normal

control

boundary

problem.

existing

modifying

PI system

temperature

theta-pinch

FLR effects may

investigated in FRX-A.

T. M. York, Rev. (1970).

The second argument,

determine whether

experiment is

REFERENCES

the B

The

and the eventual addition of Thomson

electron

careful

the

theory and

necessary to clearly identify the mode.

If the mode is the rotational m=2, as it now

appears, then it can be stabilized by preventing

the plasma from accelerating past the critical

At least two plausible explanations

exist for the observed angular acceleration. One, due to Eberhagen and Grossmann,5 relies on plasma

leaving the confined region carrying away angular

momentum as a result of diffusing across

field in the presence of a radial E field.

reaction force causes the confined plasma to spin

due to A. H. Boozer, is very similar to the source

of rotation in a standard theta pinch. Line-tying

of the open field lines shorts out the radial

(l)) causing the plasma

on open field lines to rotate as in a standard

theta pinch. This rotating plasma then causea the

confined plasma to rotate by ion viscosity. More

careful theoretical consideration and diagnostic

either of these models is correct.

the

for

This

these

Both of

conditions

correct, increased

Phys . ~,

Instrum.

4~,

Controlled

Research

“LASL K. S. Thomas, Program, Compiler, Los Alamos Scientific Laboratory report LA-7082-PR (March 1978).

Thermonuclear 1976.”

January-December

F. R. Scott, E. N. Ducas, and R. G. Tuckfield, Jr., RSV. Sci. Instrum. 33, 1001 (1962).

Sci.

E. A. B. Bodin and A. A. Newton, Phys. Fluids ~, 1338 (1963).
A. Rberhagen and W. Grossman, Z. 130 (1971).

O.A. Zolotovaky,

A.G. Kslygin, 6. A.G. Es’kov, R.Kh. Kurtmullaev, Ya.N. Laukhin, Al. Malyutin, A.P. Proshletsov, V.N. Semenov, “Plasma Confine- ment in a Pulsed System with a Compact Toroidal Seventh European Conf. Configuration,” in Proc. Controlled Fusion and Plasma Physics, Lausanne, 1975 (CR.pp,1975) Vol.

I, p.

J. P. Freidberg and L. D. Pearlstein, to be published in Phys. Fluids.

D. Dobrott, S. C. Prager, and J. B. Taylor,

1850 (1977).

P. N. Hu and H. Grad, Tear’in& Evolution,” Bull. 1136 (1977).
A. I. Shestakov and J. Killeen, “The Effect of Equilibrium Flow on the Resistive Tearing Mode,t’ Bull. Am. Phys. Soc. ~,

1156 (1977).

A. Kalek, unpublished numerical calculations

inoluding plasma rotation.

Y. Y. Lau, personal communication.
R. J. Commi.sso, C. A. Ekdahl, K. B. Freese, K. F. McKenna, and W. E. Quinn, Experiment on a Theta Pinch,” Phys. ~,

137 (1977).

Phys. Fluids ~,

~m.

Rev.

Lett.

!@Solid-End-Plug

Phys. SOC* a,

‘Nonllnear Analytic

The

INTRODUCTION

behavior was

about 10° off the beam.

magnetic

probes

using

will

not

Implosion Heating

allowing measurements of either forward or back

In our case, near-angle scattering

40-cm-diameter, l-meter-long high-voltage linear

had some advantages over 90” scattering.

theta pinch designed to study the implosion phase

example, It was possible to use a viewing dump to

of conventional theta-pinch heating.

help eliminate stray light. In fact, we were able

sion coil was fed at four places by low-impedance

to measure Rayleigh scattering from N2, and thus

pulse-forming networks (PFNs) designed to produce

we were able to make absolute density measure-

a flat-topped current pulse. The e!fective imped-

Also , the small angle between the laser

ance of the four PFNs was approximately equal to

beam end the viewing system allowed us to measure

that of the imploding plasma, thus the circuit

both Tell and Tel depending on whether we did

strongly coupled

forward or back scattering,respectively. Finally,

dynamics. The E’FNcircuits and other details of

we dld not have to modify the driving coil for the

the apparatus have been described previously.1$2

viewing optics or the laser input ana we were able

This experiment was terminated near the end of the

to use good quality viewing optics.

year in favor of the Fast Liner Experiment.

The scattering volume was about 1 mm in

The experiment operated reliably in 1977 with

diameter and 40 mm long and it was located at the

no major changes in the hardware. The results of

midplane of the theta pinch.

the density measurements made with a four-beam

used for this scattering measurement was a Korad

fractional-fringe inter!‘erometer have been pre-

The beam was expanded to about

sented previously.3 This inter!‘erometerwas used

35 mm and focused with a 1.2-meter focal length

in almost all data-taking to check consistency of

lens system to give a focal spot of less than 1 mm

machine operation. The first part of the year was

in diameter. The pulse width (full width at half’

spent finishing a more detailed measurement of the

max) was about 20 ns. since electron temperatures

magnetic field as a function of time and radius

and

density implosim heating experiments,5 the poly-

Detailed descriptions of the probe and Faraday

rotation apparatus are available elsewhere3~4 and

temperatures from 2 eV to 2 keV for Tlland from

been

discussed here.

information about the magnetic field improves but

design because for temperatures less than 250 eV

does not substantially change the results which

T1 should be nearly equal to T,, at densities on

are ahown in Ref. 3. The major part of the year

To achieve this large

was spent taking temperature and density data

range of temperatures we used 10 channels of

using Thomson scattering at four radial positions.

detectors. As it turned out only the first eight

These radial positions are R = 16, 12, 8, and

chahnels were used and maximum temperatures (both

4.6 cm, where R . 0 corresponds to the axis of the

theta pinch. The operating conditions for most of

T1 and T,,) of around 300 eV were observed.

The ratio of each channel width to its difference from

these data correspond to a flat-topped 6.5-kG

driving field and 7.O-mT fill pressure. We also

the laser frequency was about 1:3. The background light was fairly large in the wider channels and

took data with filling pressures of 3.5 mT and

was changing rapidly as the sheath went by, but

14 mT, but only at R = 12 cm.

fortunately it was quite repeatable from shot to

The Thomson scattering apparatus was designed

shot. Because of this large amplitude fluctuation

such that the laser beam was parallel to the axis

it was not adequate to simply subtract the light

of the theta-pinch coil with the viewing angle

level which was present before the laser was fired

enter from either end of the vacuum chamber, thus

necessary to observe the discliaruewithout firing

The laser beam could

to account for the background.

was

ments.

Experiment

The implo-

the plasma

scattering.

K1500 ruby laser.

VII . IMPLOSION HEATING EXPERIMENT (IHX)

T. Jarboe, I. Henins, A. Sherwood, C. Swannack

250 eV to 250 keV for T1.

the order of 1015/cm3.

in excess of 1 keV have

The additional

rotation.

chrometer

designed

Faraday

was

For

The light source

This was an adequate

Instead it was

electron

measure

measured

lower

TIME (m)

Fig. VII-1.

The two methods agree to

could be a consequence of

was

This

then

ratio

density

Thomson

measured

Fig. VII-1.

scattering at

center of the gate.

of the measurement.

pressure condition.

of the theta pinch.

the measurement.

these positions.

each channel.

end density.

the laser to measure the ratio of the plasma light

at the time of the laser pulse (if the laser were

fired) to that which was present 70 ns earlier.

background level during the scattering time from

the light level meaaured earlier. The gates for

the measurement of the background and laaer pulse

were each 70 na wide. The jitter of the laser was

only about 10 ns, so the gates were triggered at a

aet time after a signal was sent to Q-switch the

ruby laaer. However, the time of arrival of the

peak of the signal from the laser output was com-

pared with the timing of the gate on each shot to

check that the light signal appeared near the

The results of the density measurements from

These data are for the 7.O-mT fill

The

assuming a constant density along the 90-cm path

within experimental errors, but the density from

scattering data appears somewhat higher.

be that the density in the center of the theta

pinch is higher than at the ends.

allowed, this possibility would have been checked

by doing scattering measurements out near the ends

At 16 and 4.6 cm Rayleigh

scattering was not done but the shape and timing

of the density deduced from the Thomson scattering

agree very well with the interferometer

The temperature data are shown in Fig. VII-2

for R . 16-, 12-, 8-, and 4.6-cm positions. The

temperatures were deduced in the usual manner by

finding the least squares fit of a Gaussian func-

tion to the time-integrated signal collected in

The large variations seen in the

data are mostly due to shot-to-shot variations in

Where

rapidly changing function of time a small vari-

ation in the time of arrival of the sheath could

account for some of the variations in temperature

It could also be that the sheath

region has macroscopic inhomogeneities which could

give a scatter in temperature and density. There

used

;

n- -~

4 —

the

are

s Y u,

, #

curve

solid

[

It may

is the

R . 8 cm

shown in

determine

interferometer,

> k3 _ m z w n z

w = F(o) -/

the temperature is a

than those at 7.0 mT.

fit when NRD = 0.080.

reasonably good

data at

time

The

Had

R=8cm

This

1000

in the

R=8 cm. dennlty

I , , 1 ,

the larger volume

Figure VII-3 shows a

The solid curves in

the numerical

that is used

-sn]2

of the nth

expression:

measured

signal

here

are

the

The temperatures at 3.5-mT

where F(x) ia the Gaussian function

in the fit, x is the frequency shift from the

laser frequency, .Sn iS

channel centered at Xn, and N is the number of

channels used in the ftt.

The results of the data taken at 3.5-mT and

14.O-mT fill pressures were not much different

fill were somewhat higher than the 7.O-mT data,

and the 14.O-mT data were somewhat lower but by

less than a factor of two in either case.

temperatures

agreement with

simulations of A. G. Sgro.6

Thomson scattering density measurements at The solid curve is the inter~erometrio measurement at the same radius.

are also some shot-to-shot differences

density aa measured by the interferometer, but

these variations are not nearly as large.

sampled (2-mm diameter and 900-mm length). All 01

the temperature data shown here fit a Gaunaian

with a normalized rms deviation (NRD) of less than

0.15. NRD is given by the following

f[F(xn) n=l

l l * $

TIME (ns)

A l

o:,,,,

l *

300

‘e

s al

‘0 l

IIl l\l 0

i 1 ”\

l lL”- “*A

G m 0.5 -

FR%JENCYOSil

~ o (n

-0

0.0

;

l l

[

20C

30C

J-

-1

R=8cm

I000

R=12cm

—*

100 —

I500

I i

200 —

: e i=

300—

R= 4.7cm

l e

TIME [ns)

1 R=16cm

1,,,,,

‘“wmARy

interferometer, a single-beam measurement,

vacuum Ee was 4 kV/cm, and plasma Ees of 1,5 klJ/cm

density and magnetic field as a function of time

does not contain end loss which makes the agree-

In this experiment the implosion and first

VII-2 are the results of these simulations

for the IHX experimental parameters.

ment somewhat surprising.

electron temperature as a

time was measured using

Detailed measurements

function of radius

Thomson scattering.

Faraday rotation

expansion were

studied most

single shot.

extensively.

calibrated

observed.

four-beam

The code

magnetic

TIME (ns)

probes,

T=218eV

radius

using

l m

were

made

Fig.

1500

-1

0.5

The

the

and

0.4

The

and

—1

TIME (ns)

Fig. VII-2. IHX temperature data at R=4.~-, 8-, 12-, and lb-cm. The solid curves are the results of numerical simulations by A. G. Sgro.

FT (:z3X10’4) .

Fig. VII-3. Typical scattered signals for a solid curve is the Gaussian fit.

2. J. E. Hammel, I. Henins, J. Marshall, and

A. R. Sherwood, llTheDesign of a High Voltage Generator for the LASL Implosion Heating Experiment,t’Proo. Fifth Symposium on Engineering Problems of Fusion Research, Princeton (1973) IEEE Pub. No. 73 CH0843-3-NPS, 1974, p. 664.

K. S. Thomas, compiler, ltLASLControlled Thermonuclear Research Program, January- December 1976,!1Loa Alamos Scientific Laboratory report LA-7082-PR (March 1978), p. 106.
T. R. Jarboe, I@Measurementof FaradaY Rotation in the Implosion Heating Experiment,‘fJ. APP1. phys. ~, (1977).
K. Hothker, !IplasmaDynamics and

Current-Sheath Structure In a Collision-Free Theta Pinch,” Nucl. Fusion J.&,253-261 (1976).

A. G. Sgro, to be published in Phy8. Fluids, 1978.

R. R. Bartsch, C. J. Buchenauer, E. L. Cantrell, J. N. Downing, F. L. Freeman, K. B. Freese, R. F. Gribble, 1. Henins, F. C. Jahoda, T. R. Jarboe, R. Kristal, G. Miller, W. E. Quinn, F. L. Ribe, A. R. Sherwood, R. E. Siemon, K. S. Thomas, ‘Plasma Experiments on the Staged Theta Pinch, the Implosion Heating Experiment, and Scyllac Feedback-Sector Experiment,gf~ Phvsics and Controlled Nuclea Fusion (IAEA, Berchtesgaden, 1976) vgl. II, pp. 193-200.

I. Henins, J. E. Hammel, T. R. Jarboe,

J. Marshall, and A. R. Sherwood, 1’ 1S d HiuhB eta PlasH fod and New York, 1976) PP. 203-207.

Thomson scattering and double foil soft x-ray

production

measured using a scintillator counter.

The implosion speed, the expansion speed, the

radiua of the sheath at. turnaround, the density

jump at the sheath, and the peak neutron produc-

tion rate agree quite well with that predicted by

the simple bounce model.7 The actual width of the

sheath and the electron temperature in the sheath

are not. predicted by this model, but the measure-

agree with calculations done by Sgro and

This temperature agreement is somewhat

losses

these models. These results show that within the

parameters of this experiment implosion heating is

an effective means of heating plasmas to kilovolt

temperatures and that the level of ion heating can

be predioted reasonably well by the simple bounce

ment

model.

others.

Neutron

REFERENCES

techniques.

surprising since end

were

rates

(Pergamon Press,

are not included in

Resear~

A. R. Sherwood, E. L. Cantrell, I. Henins, T. R. Jarboe, J. Marshall, C. E. Swannack

Fig. VIII-1. Geometry of the Fast Liner Experiment. Material end plugs are employed, and the Lmplosion is driven by an axial current carried within the liner itself.

field to inhibit thermal losses to the walls.

The basis for optimism in the Fast Liner

Be field, shown in the figure, has the advantage

Experiment stems from the scaling law for DT fills

gain

Early Ln the year a proposal to do an

imploding liner experiment at LASL wss prepared

eventually approved, and preparations for the

liner experiment began after the termination of

the IHX experiment in early December. Substantial

proposal

B. Suydam,

R. Miller, R. Moses, and others.

The general concept is to explore the l~ner

approach to fusion in the 106 cmls liner velocity

regime, where wall-confined plasma configurations

As shown in Fig. VIII-1, a

thin, nonrotating, cylindrical, metallic liner

driven by a magnetic field in the theta dtrection

is employed. The liner carries its own implosion

The plasma needs an embedded magnetic

of providing Insulation against both axial and

radial heat flow. The initial plasma must be warm

Theoretical guidance for the initial

plasma conditions comes from numerical models

which include liner physics and plasma losses.

The following parameters for the initial plasma

conditions are representative for reactor-like experiments: no = 5 x 10’7 cm-3, To = 200 eV, and Relaxed conditions would still allow Boo = 5 T.

DMFE.

and

the

current.

R. Malone,

R. Gerwin,

presented to

contributions to

are a possibility.

interesting experiments.

well as insulation.

and dense.

were

This

made

optimized

proposal)!

for maximum

R. Krakowski,

proposal’ was

VIII. FAST LINER EXPERIMENT

unit length (GJYm).

4Po

calculations

—E .2T1.’

detailed

concept.

Q’ = 7

liner.

the

(derived

Simple theoretical

the exact model

density of

this

where Q’ is the ratio of the thermonuclear energy

to the liner energy, P. is the initial

The production of the

the liner (cgs) and Et is the liner energy per

initial plasma is a challenging task and is one of

The coefficient in

the most important goals of the experiment.

equation depends somewhat on

The approach being pursued in this experiment

employed, and other factors of order unity are

thus involves inertial confinement of the plasma

introduced if one considers such effects as liner

particles and magnetic confinement of the energy.

compressibility. From this equatfon we see that

In this and other respects (e.g., BB liner drive

reasonably high Q’s might be obtained from the

field) it is similar to the LN-20 exeriment of

Scyllac bank (8 MJ) for a 10-cm-long aluminum

Alikhanov2 in the USSR, and it is a complementary

It is important to note, however, that

approach to slower liner compression experiments such as the NRL LINUS program3 and the experiments

this equation assumes that the plasma energy loss

rate is small compared to the rate at which energy

of Kurtmullaev.2 In the latter experiments mag-

is supplied from the liner.

netic fields are used for plasma containment as

estimates make this assumption plausible, but more

some experimental results are necessary to check the

(especially)

and

major

experiment, namely plasma preparation experiments

and liner implosion experiments. Initially these

two elements will be separate, but they will be

combined as soon as it is practical. The initial

goal of the plasma preparation experiments is to

produce a suitable initial plasma In an implodable

Here ‘suitable” depends on geometric

factors such as the length and diameter of the

liner, but in general what is desired is a plasma

with enough denstty, temperature, and embedded

magnetic field that it should be heated during the

The first plasma preparation

experiments wL1l be based upon coaxial plasma flow

systems, that is plasma guns or modifications

These experiments will be done on the

same faoility as was used for the Gun Injection

Experiment (300-kJ bank operated at 180 kJ).

the end of the year this facility was betng

modified for the new experiments.

begin, the exlstfng 30-cm-o.d. coaxial plasma gun

two

are

There

muzzle,

thereof.

geometry.

liner implosion.

added.

the

3-RACK

elements

At least to

It is planned

The initial

scientists,

inoluding

At the

The

will be used with geometrical changes at the

AND 7-RACK

The liner implosion experiments will be done

initially without plasma on three racks of the

Scyllac bank (1.75 MJ at 55 kV).

goals of these experiments will be to develop

diagnostics, to check the symmetry and velocity of

Fig. VIII-2. Initial and future liner implosion experiments on The initial experiments will the Scyllac bank. use three racks (12, 13, and 14), but with only small additions to the collector plates four more racks could be added.

the implosions, to learn how to make current

As this new experiment begins to get unaer

contacb to the imploding liner, to compare the

way it appears quite possible that collaborations

efficiency of coupling of energy from the bank to

will be established with other elements of the CTR

the liner with numerical calculations, etc. As is

community. Candidates for such collaborations are

illustrated in Fig. VIII-2, it happens that with

a) parallel plasma preparation experiments to be

only a minimal addition to the existing collector

done at other US institutions so that several

plates four more racks of the Scyllac bank can be

possible paths towards the required initial plasma

Assuming the seven racks can be run at

can be followed, and b) a joint experiment wLth

50 kv, about 3.5 MJ would then be available for

the Soviets in which cooperative liner implosion

the liner implosion experiments.

and liner/plasma experiments are oonduoted, and in

that collector plate additions to accommodate the

which the Soviets would conduot further plasma

four additional racka will be ordered, but the

preparation experiments following the approach

inttial experiments will be done on three racks

(out of a total of 15 for the whole bank).

used in the LN-20 experiment (roughly—plaama

injection from electromagnetic shock tubes). The

three racks chosen for the first experiments have

latter collaboration was initially suggested by

not been operated for 2-1/2 to 3 years.

E. P. Velikhov in May 1977.

end of the calendar year they were being checked

three

out and brought back into operable condition.

Moscow in October to review the Soviet liner

program and to explore the possibility of joint

RACK

LINER

7- EXPT

LINER

EXPERIMENTS

COLLECTORPLATE

fi!s!!lFUTURE

”~ PRESENT 3-

”:,= RACK EXPT

-(2

A team of five US

Vlslted

LASL,

from

A. D. Muzychenko, V. P. Novikov, V. V. Pichugin, V. N. .Semenov,G. E. Smolkin, E. G. Utyugov, and I. Ya. Shipuk. nStudv of Models of Liner Thermonuclear Systemsn (translated by D. L. Book), Proc. 6th Int. Conf. on Plasma Physics and Controlled Nuclear Fusion Research Berchtesgaden, FGR, October 1976, (IAEA, Vienna, to be published), Paper No. E-19-2.

A. E. Robson and P. J. Turchl, ‘The NRL LINuS Program,n Proceedings of the 3rd Topical Conference on Pulsed, High Beta Plasmas, Culham, 9-12 September 1975.

LASL-Kurchatov liner experiments.

were fruitful and the prospects for such joint

experiments appear promising.

A. R. Sherwood, B. L. Freeman, R. A. Gerwin, T. R. Jarboe, R. A. Krakowski, R. C. Malone, J. Marshall, R. L. Miller, B. Suydam, “Fast Liner Proposal,i’Los Alamos Scientific Laboratory report LA-6707-P (August 1977).
S. G. Alikhanov, V. P. Bakhtin,

Wm. M. Brusnikin, I. S. Glushkov, R. Kh. Kurtmullaev, A. L. Lunin,

REFERENCES

The meetings

and

space,

17 kG).

electrode

INTRODUCTION

linear theta pinch.

as plasma energy

rather broad.

fast valve.

of D2 gas.

During

the

2-m-long guide field coil was installed in place

As reported in the last annual report, this

of the shorter 50-cm one.

experiment was undertaken as part of the Scylla

measurements at three positions have been used to

IV-P program to develop a plasma injector for a

determine plasma stream veloclty and the trans-

The 1.5-m-long coaxial gun

verse energy of the 0.1 plasma stream.

has an inner electrode diameter of 10 cm and a

the magnetic field Inside the plasma at the

30-cm-diameter outer electrode.

midplane of the 10-kG guide field coil also has

valve admits a gas load at the middle of the

shown flux exclusion out to a 12-cm diameter,

which is comparable to that calculated for 6=1

oapaoitor bank with an approximate inductance of

plasma from diamagnetic measurements.

8 nH is used to power the gun. For all the exper-

The progress of the experiment was slowed

iments the bank has been charged to 45 kV,

because of manpower shortages in CTR-5. Work was

corresponding to a stored energy of 187 kJ.

suspended on the Gun Injection Experiment for

During the previous year the experiment was

about five months so that we could concentrate on

assembled and the first measurements were made.

measurements

The initial measurements were of the properties of

Heating Experiment and preparation of the propos-

the gun plasma itself. These were followed by an

als for the Fast Liner Implosion Experiment.

investigation of plasma injection into a ehort

the end of this year the Gun Injection Experiment

(50-cm-long) solenoidal guide

typical strength of about

has been closed down and the experimental faoility

will be used for plasma preparation studies aimed

The diagnostics employed included single

at providing a suitable preplasma for the Fast

beam and holographic interferometry, fast-framing

photography, neutron measurements, magnetic flux

exclusion measurements in the guide field, and

EXPERIMENTAL RESULTS

calorimetric measurements of the streaming plasma

Plasma Injection into a 2-m-long B. G-

energy both upstream and downstream of the guide

field coil. Two distinct gun operation modes were

Field Coils.

selected. In Mode 1 the gun is fired 300 PS after

coil assembly to replace the short 50-cm-long ‘see

50 CU13 atm of deuterium gas is injected by the

throughn coil was made in two l-m-long seotions by

The maximum gun current in this case

wrapping a single layer of insulated #2 AWG copper

is about 800 kA and the resulting plasma stream is

turn-to-turn

In the second, Mode 2, operation

23-cm-diameter stainless steel cylindrical form.

the gun is fired 36o US after injecting 63 cm3 atm

In this case the gun current

Then a layer of fiber glass and epoxy was applied to hold the windings in place. A stainless steel

1.3 MA and the plasma stream forms on axis in

can was welded over the windings to provide a

front of the gun with a diameter of only a few

case.

centimeters. Calorimeter measurements showed that

coaxially through a metal tube attached to the

over 80$ of the stored electrical energy emerges

coil case. One end of the coil is attached to the

from the gun. The tighter plasma

coil form which serves as the current return.

stream of the Mode 2 operation was better able to

The coil assembly is mounted 40 cm from the

penetrate a 10-kG B= guide field.

gun as shown in Fig. IX-I.

present year

gun has six turns spaced 1 cm apart to minimize

measurements were extended to measure the plasma

the B= decrease near the of the coi..I.There in a

stream divergence without a guide field, and a

2-cm gap between the two coils;

the

10 kG

185-uF

finishing

Scyllac-type

(maximum was

A fast-acting

field having a

I. Henins and J. Marshall

IX. GUN INJECTION EXPERIMENT

vacuum-tight coil

Liner Experiment.

calorimeter

reaches

cable

K et

with

2-cm

Ma n

the

Implosion

Probing of

Plasma diamagnetic

The end nearest the

A new guide field

spacing

current

extra

turn

The

fed

one

C.ll. c,ol P,.!, M.mbly

,00 ,50 200

s..!,

/c.b!*

C…w..w.

near

Tmfi

each end

10-kV capacitor bank.

measurements at each of

are 110 cm2, 50 cm2, and

. Figure IX-2

coil wall.

calculated

stream

f3=l,

shows

from

can

the

Fig. IX-1. Schematic diagram of coaxial gun experiment with 2-m-long Bz guide field coil.

the coils here provides an

average 2-cm turn spacing. The coil near the gun

has a total of 56 turns, and the other 52.

Each of the coils is energized by a 3000-PF,

The risetime of the Bz

field in the coils is 800 ps.

and the Bz are not in phase because the L/R time

of the stainless steel shell is about 550 ps.

Maximum Bz at 10 kV bank voltage is 16 kG.

b. Excluded FIUX Measurements. These

surements were made with three loops placed near

the inside walls of the Bz coils at 60 cm, 150 cm,

and 220 cm from the gun muzzle.

approximately 0.5 cm wide and 65 cm long and com-

pletely surrounds the plasma to eliminate off-axis

effects. A 20-cm-i.d., 200-cm-long, 0.5-cm-thick

quartz tube liner inside the loops protected them

from direct plasma bombardment.

During the short time of plasma injection the

stainless steel coil form acts as a good flux

conserver; therefore the total flux excluded by

the plasma can be derived from the measurement of

the change in magnetic field strength near the

If diamagnetic features of the plasma

stream can be identified, then plasma velocity in

progresses through the three loops.

measured

the

energy for the long delay mode (Mode 2) as the

plasma is injected into a 10-kG guide field.

maximum areas of the plasma, assuming it to be

the three loop positions

25 cm2. The time-varying

L..o,

M..,u,,nq

mea-

\ [Kl”dedil..

The coil current

the diamagnetic signals.

Each loop is

diamagnetic

excluded

derived

plasma

quartz

liner

flux

the

The

and

the

220 cm From

These values for

the quartz wall,

Extrapolation

coil wall.

The

Fig. IX-2. Diamagnetic energy and velocity of plasma injected Into a 10-kG guide field. Mode 2 gun operation.

areas at each loop position were multiplied by the

external magnetic field energy density to give the

diamagnetic plasma energy plotted in Fig. IX-2.

These curves, together with a velocity distribu-

tion can be further integrated over time to give

the total diamagnetic energy.

the data in Fig. IX-2 are 4.6 kJ, 2.1 kJ, and

0.6 kJ for the loops 60 cm, 150 cm, and 22o cm

from the gun muzzle, respectively.

of the plasma position to the gun muzzle indicates

that the plasma emerges at approximately 10 us

after the gun current starts. It appears that the

paramagnetic signal observed at later times, as

seen in Fig. IX-2, is caused by plasma bombardment

of the quartz liner and is confined to the region

near the entrance of the coil. Secondary plasma,

from evaporation of

apparently entrains magnetic flux and carries it

inward, so as to decrease the flux between the

inner

paramagnetic component was therefore not included

in the total energy calculation. When the gun is

operated in the short delay mode (Mode 1), a dia-

magnetic plasma appears to emerge from the gun as

a short pulse with a velocity distribution from

2 x 107 to 9 x 107 cmls as shown in Fig. IX-3. There

allow measurement of velocity throughout the duration of

well-defined

features

that

are

The peak values of the excluded flux areas at the three loop positions in

Fig. IX-3. Diamagnetic energy and velocity of plasma injected into a 10-kG guide field. Mode 1 gun operation.

measured by the diamagnetic loop at 150 cm, which

corresponds to a flux exclusion radius of about

70 cm2

130 cm2, 70 cm2, and 45 cm2.

3, Pla ma Stream Diverue

integrated total plasma diamagnetic energies are

mate of the plasma stream spread in the absence of

8.7 kJ, 6.5 kJ, and 4.8 kJ.

external magnetic fields can be obtained by ob-

the 8hort delay mode of gun operation produces

energy

more diamagnetic plasma which retains its dia-

calorimeter at various distances from

magnetism as it travels down a B= guide field.

these measurements a large

The velocities observed in the 2-m-long guide

diameter,

field are higher than the - 20 cm/vs velocities

mounted behind a 60-cm by 60-cm copper plate

olear; it may have something to do with the length

energy passing through the aperture waa measured

The reason for this is not

containing a 3(J-cm-diameteraperture. The plasma

: ”

: :

1.5-

-M s :

:s0 -

~ 0.5-

=10 - :’

this case are

near optimum. ~.

reported last year.

alumina tube.

loops,

stream

ment.

guide

field

the

Usually the duration of the diamagnetic signals

was shortened and the later parsmagnetio signal

‘imr@120~30

exclusion aa observed by the B= probe can be

The 6-cm-radius outer limit of complete flux

Typical smoothed magnetic probe data for

several radii are shown in Fig. IX-4. On axis the

signal shows complete flux exclusion lasting about

8 ps, after which the flux penetration (or dis-

appearance of plasma) takes place.

radii the maximum flux exclusion lasta only a

couple of microseconds, as is shown in the figure.

However, nearly complete flux exclusion ocoura out

60 cm

150 cm

The

how

the

much

5 cm.

muzzle.

serving

compared

was enhanced.

to a radius of 6 cm.

Thus it appears that

It appears that the

187 kJ (""- ""-’ ’ ”

Magnetic from gun

A magnetic field

calorimeter

accurate

(5O cm

make

the

and

Fig. field exclusion muzzle.

nce.

flux

91 cm

the gun

At larger

long) was

colleoted by

A rough esti-

exclusion area

IX-4. from plasma stream 150 cm

at distances of 30, 60, and 110 cm from the gun

Time (#see)

magnetic plasma by varying the gun delay and

amount of gas injected.

300-lJs delay with 1.4 x 106 Pa plenum pressure is

probe was inserted into the Bz guide field as

shown in Fig. IX-1. The axial position was kept

at 150 cm from the gun muzzle (the #2 diamagnetic

loop position) and the probe was moved radially.

The gun was operated with the short (300-iIS)

delay. The probe coil had an area of about 2 cm2

and was inserted in a 9-mm outside-diameter

Quartz envelopes were also tried

but they could not withstand the plasma bombard-

The insertion of the probe in the plasma

disturbed the signals in the diamagnetic

even upstream from the probe position.

measurement this allows.

muzzle. The gun input energy was kept constant at

An effort was made to maximize the dia-

Ios-vr oankccnargea to 45 KV). ’.-.—’

”

case

were

larger diameter coil, by increasing the current

density just at the end of the coil, and by

and 98 %8 kJ for the above calorimeter positions,

respectively. When the gun was operated with the

covering the end turns. Previous experiments with

the

smaller guns did not 8how this behavior. In those

110 *6, and 54 *3 kJ,

experiments the gun was of smaller diameter than

respectively. The uncertainties are the standard

the coil, whereas in the present the gun is larger

deviations calculated from four to five gun shots

in diameter than the inner liner surface by 50%.

for each calorimeter position. From these data it

that there is relatively more

is seen that for the longer delay and larger gas

high-energy plasma streaming directly at the end

loading the plasma stream is confined closer to

of the coil than before. Bombardment of the liner

the axis of the system than in the other case.

is undesirable because it introduces large amounta

When the calorimeter is placed behind the

of impurity by evaporation of the liner material

2-m-long guide field coil, about 40 kJ of energy

and because it can mechanically destroy the liner.

is collected in the Mode 2 operation of the gun,

The small diamagnetic energy of the injected

essentially independent of the strength of the

plasma (- 10% of the total plasma energy measured

magnetic field up to 10 kG.

calorimetrically) is unsuitable as a theta-pinch

more divergent and more diamagnetic plasma stream

adiabatic

of Mode 1 operation, however, does not flow down

with its low initial transverse temper-

The energy collected at

ature, it would require excessive compression to

the end of the guide field coil is only 2, 14, 22,

reach thermonuclear temperatures.

and 13 kJ with B= equal to O, 5, 10, and 15 kG,

velocity

the

For

long

delay

161 tlO,

respectively.

collected were

short delay (Mode

the guide field as well.

collected amounts of energy

translational kinetic energy.

other previous plasma guns.

demonstrate

experiment

insufficient

coaxial

SUMMARY

strated

pinch.

plasma

entrance

barrier

this

that

coil

for

the

for

the

high

because,

pinch.

energy

filling

amounts

(Mode 2),

reducing

streaming

thickness

subsequent

135 *8, 108 *5,

The energy of the

This probably means

section to the solenoid.

against one another.

plasma into a

parameters.

thermalize

reactors

suitable

the plasma

plasmas

would

baaed

to an

have

very

able

using

not

for

due

lead

would

passive

material

compression

In addition,

theta - pinch

successfully

8cattering

Coulomb

This coaxial plasma gun has been the largest

This approach has been

and highest voltage gun operated at LASL so far.

investigated by Skvortsov et al. at the Kurchatov

Its operation has been fairly typical compared to

Institute. Another possibility is injection from

We have again demon-

both ends of the theta pinch, allowing the two

plasma gun

efficient in converting electrical energy into

plasma energy. With the limited effort available

While gun injection of the precompression

were

injection

replace implosion heating, and thus eliminate the

theta pinch might

replacing the implosion heating phase of a theta

need for fast high-voltage power supplies, it

There were two principal deficiencies in

aPPears to be limited to relatively short systems,

the system investigated. One of these was severe bombardment of the entrance end of the fused

because of the finite axial velocity of the

injected plasma. This implies that, if this were

silica liner. The other was the small diamagnetic

to be used in a reactor, it would have to have

energy of the injected plasma relative to its

very effective end-stoppering, or else the plasma

Plasma bombardment of the quartz liner at the

presumably

of flux separating

from the liner and the end of the coil. It should

be possible to alleviate this problem by using a

high densities, as in an imploding liner system. It is not obvious how gun injection could be used in

compressed

very

conventional

excessive losses through the far end of the theta

The difficulty might be alleviated by

a long, gradually increasing field entrance

The

primary

present

SUMMARY

interest.

Such a frequency-independent reduction

efforts

of this cross section implles that any transport

Plasma Physics Group are directed toward the study

coefficient that is a function of the eleotron-lon

of high-frequency plasma resistivi.ty,plasma heat-

collision cross section could be modified by

ing, and heat flow in the presence of plasma tur-

strong electron oscillatory velocities induced,

bulence, eleotron drifts, and other effects likely

for example, in the heating of a tokamak using

plasmas of current fusion

high-frequency power, or in the heating of a long

linear plasma column by an intense C02 laser beam.

During the past year we have continued our

The above experimental studies, which were

study of the effect of a strong dc electron drift

being done on our Q-machine, were temporarily sus-

on the high-frequency plasma resistivity. We have

pended on May 25, 1977, so as to allow dismantling

taken a great deal more data and refined our tech-

of our equipment to make room for the construction

niques of data collection and analysis.

of a large toroidal reversed-field pinch (ZT-40).

findings confirm our preliminary results reported

Relocation of our experiment into a new area was

last yearl that the high-frequency plasma absorp-

begun in late December, and we plan to be in

tion becomes De~ative if the electron drift veloc-

operation by June 1978.

ity exceeds the electron thermal velocity.

necessary to postpone our planned studies of plas-

plan to extend this work in the future to the case

ma heat flow until the Q-machine is operating

where turbulence is simultaneously present in the

A considerable amount of time in the past

plasma, in which case an experimental test of the concept of the Plasma Wave Laser2 may be possible. We have continued to use our Dual-Mode Res-

year has been devoted to planning for the disman-

tling and relocation of our experiments.

Testing has continued on a new experiment to

onator to study the absorption experienced by a

study nonlinear interactions between a plasma and

weak test microwave field in the presence of tur-

a traveling electromagnetic wave.

bulence generated by parametric instabilities that

sis will be on the study of the beat-heating of a

are excited by a strong microwave driver field.

plasma using two electromagnetic wavea separated

Our recent test-field resistivity measurements as

in frequency by a natural-mode frequency of the

a function of the teat-driver frequency separation

plasma, and on the study of the decay of a single

support our earlier hypothesis that the enhanced

strong electromagnetic pump into two eleotron

Our

the

again.

section.

Experimental

EXPERIMENTAL PLASMA PHYSICS

H. Dreicer, Martin E. Banton, J. C. Ingraham, R. S. Masaey, F. E. Wittman, and B. L. Wright

ically in Fig. X-1.

plasma waves.

Here we ob-

continued,

magnetized

primarily

potassium

machine.

that

This delay has made it

Primary empha-

single-ended

Langmuir

column

(Te N

our

Q-

plasma

and the weak test signal in such a manner

PLASMA AC RESISTIVITY STUDIES

additional energy is transferred from the test

Our studies of the ac plasma resistivity have

field to the plasma through this beat oscillation.

study weak test-field absorption in the presence

We have also used the Dual-Mode Resonator to

This device provides a fully ionized

of a strong driver field, but In the absence of

~ 2250 K, n * 1-5 x 1010 cm-3, diameter - 2.5 cm,

parametric instability turbulence.

B=s 5 kG, length== 200 cm) on which to study plaa-

serve a reduction of the test-field absorption

ma reslstlvity using high-Q (s=20 000) microwave

that is Identical to our earlier reported measure-

resonators and movable high-frequency

ment of the reduction of the driver absorption.3

probes. The experimental setup is ahown schemat-

‘1’hiareduction is caused by the effect of the

large driver-induced oscillatory velocity of the

electrons on the electron-ion collision cross

test absorption occurs through nonlinear beating

between some portion of the unstable wave spectrum

~‘p

quantity is proportional to n2.

Fig. X-1. Experimental setup of Q-machine experiments.

The ac plasma resistivity is determined from

This has the advantage of reduced ohmic

due to the electron current flow, but it also

reduces the sensitivity of our measurement of the

plasma absorption in the resonator, since this

avoid ohmic heating is to use a short pulse length

for the electron current and to measure the micro-

wave absorption rate during this short time inter-

This also insures that the Buneman instabil- val. i ty4 cannot grow sufficiently to influence the

electron drift speed and the microwave absorption.

rate of the plasma. This absorption is determined

The price paid for gaining these advantages is the

from the measured rate of decay of microwave ener-

increased difficulty of accurately measuring the

gy stored within a high-Q resonator containing a

slope of the microwave energy decay from our

portion of the plasma column.

oscilloscope traces during such a short time in-

between this rate of decay and that obtained in

the absence of plasma yields a measurement of the

reported

microwave

10 Observation of Negative Inverse Brems-

velocity, VD, exceeds the electron thermal veloc-

measurements showing that if the electron drift

strahlung Absorption Due to Strong Electron Drift

&E!E2@. We have continued our study of the effect of a strong dc electron drift on the plasma ac

actually becomes negative, that is, Nevative In-

Bremsstrahlung Absorption occurs.

microwave absorption.

past year we have refined our analysis techniques

experiment a short (- 0.4-ps) positive voltage

and taken a great deal more experimental data

cold

which confirm our previous conclusions.

Q-machine (Fig. X-1) in order to induce a uniform

The refinements have included efforts to

electron drift in the plasma column. The electron

reduce any possible systematic errors that might

drift velocity is determined from simultaneous

occur from measurements of resonator decay rates

electron current

taken from the oscilloscope-trace photographs.

circuit external to the plasma, and the electron

Each oscilloscope trace picture was a composite of

density. Measurements of the radial profiles of

four traces: a baseline and three resonator de-

electron density and drift current density using

Two of the decays, spaced by about one cm

’

I.p

measurement of the microwave

P -

50c.

HOT PLATE

OUAL MODE RESONATOR

resistivlty or

plasma absorption rate.

pulse is applied to

measurements of

of the application

The change

independent

effects.

of plasma

probes

small

that

show

the

of radial position

~rd

the

verse

energy

ity,

have

COLO FIATE

plate

‘T =-

terval.

In this

absorption

Previously

SINGLE MOOE RESONATOR

The difference

Trace measurements

0.005 cm. —

it was necessary

in the plasma.

resolution

as a result

oscilloscope

variations

comparing

resonator

is meas-

velocity

position

perform

plasma.

cays.

drift

decay

with

the

our

the

and

the

heating

In the

absorption

preliminary

Another way to

recalibrated the

The traces were

made using an

photograph of

with and without

sweep-rate

arising

when

from

gain

face

were

and

absorption

first two decays, was for the resonator with plas-

of the electron

ma in the presence of electron drift. These four

ured using our high Q-microwave resonator tech-

traces were all made within a period of less than

niques. Since the electric field of the resonator

30 s so as to minimize effects of baseline and

mode is parallel to the strong dc magnetic field

gain drifts of the oscilloscope.

of the Q-machine, the absorption measured is a

positioned as close together as possible on the

result of electron-ion collisions alone (inverse

photograph so as to minimize errors

bremsstrahlung) rather than electron cyclotron

To achieve large electron drift velocities,

in the face of the limited electron current emis-

sion of the Q-machine hot plate,

to lower the electron density,

observations in the range 0.35

optical digitizing system having an equivalent

oscilloscope

In addition, we

on the photograph, were for the resonator without

plasma and the third decay, located between the

oscilloscope

rates

Fokker-Planck equation and assuming the electron velocity distribution was a drifting 14axwellian.7

As a check on our electron ohmic heating computa-

tions, we have performed microwave measurement of

sponds to the average of measurements taken from

the electron temperature during the period that

at least ten oscilloscope photographs, and the

follows termination of the electron current pulse,

vertical error bars are equal to the rms deviation

and these measurements support the computed elec-

of these measurements about the average.

horizontal error bars correspond to the variations

We plan in the future to study the effeot of

during the current pulse of the measured electron

microwave

current about its average.

computation

Oberman-Dawson formalism5 of the resistivi.tyof a

Our past studies of para-

drifting Maxwellian electron velocity distribution

metric instabilities near the plasma frequency

Weak Field AC Resistivl.tv Enhanced by

microwave power response of our fast (0.ol-ps

and including the effect of a do magnetio field

response time) microwave crystal detectors several

parallel to the drift.

times during each data run, and at each frequenoy

good agreement with the earlier results of Uuaha

for which data was being taken. (For example. the

empty resonator frequency was 1992 ~Z.

Fokker-Planck equation to evaluate the resistivity

utilized

= 0.35, the

Experiment and theory are seen to agree well,

especially at higher drift velocities where the

Fig. X-2 and indeed,it can be seen that Negative

negative absorption occurs.

Inverse Bremsstrahlung Absorption corresponding to

In order to compare theory and experiment we

a negative plasma resistivity does occur at larger

have corrected the theoretical computation for the

drift velocities. The vertical axis in the figure

ohmic heating of the electrons that occurs during

is the measured plasma ac resistivity normalized

the time of the absorption measurement. The elec-

to its measured value in the absence of electron

was

The horizontal axis is the electron drift

velocity, normalized to

of the electron drift.

‘P

Our

2,W2

plasma

drift.

2023 ~Z.)

experimental

velocity, vTO . -m,

j-.oo~10

theoretical

ao --------------

050 -

-c125 -

025 -

results

who

and

The

was

the

are

tron

With

shown

ohmic

prior to the application

having

plasma

heating

thermal

electron

Yoshida,6

frequency

electron drift

tron heating results.

The solid line is our

Each data point corre-

flasma Fluctuations.

turbulent plasma.

mechanisms.

wavelength

. O) and

treated

~1 +$2

Foruo~

based

level,

upon

have

long

the

Fig. X-2. Normalized ac plasma resistl.vityVS electron drift velooity; comparison of theory and experiment.

‘D/vT()

the

in a

using

computed

absorption

Boltzmann-

drifting Uaxwellian.)

(Our computation is in

which a microwave

absolute measure-

when the

exhibits

strongly

case

(ko~ O)

electromagnetic ‘pumpn field of frequenoy W. and

enhanced absorption rate as its energy is drained

by the unstable growth of eleotron plasma waves

((I),)and ion fluctuations (U2). The process is a

nonlinear one in which frequency and wave-number

matching conditions are satisfied (h),+ U2 = fI)O,

which can only occur

pump field amplitude, E, exceeds the threshold

ET s required to overcome natural damping

up, two distinct instabil-

ities are recognized: the parametric decay insta-

bility in which U2 corresponds to an ion acouatio

wave, and the oscillating-two-stream instability

in which W2 is ZerO. For the case Te x Ti and (JJO

’ u , our experimental’8-10 have verified details -P of these phenomena including

ment of instability thresholds in agreement with

theory, transient observations of pump depletion

considerable

resulting

enhancement of the weak-field ac resistivity of a

doubtless caused radical changes In the plasma~s

resistivity] of a microwave field at a frequency

other than that of the pump field that generated

them. The topic addressed by this question ia the

plasma by nonthermal fluctuations. The mechanisms

consistent with computed growth rates, verifica.

conclude immediately that

tion of appropriate structure in the electron

dominant. It is worth noting that the two regimes

fluctuation spectrum (corrected for ion drift),

in which we see the strongest effects are those to

and analysis of ‘Ibump-in-tailtffast electron beams

which conventional theories of steady-state weak

created by large amplitude electron plasma waves.

turbulence may not apply.

With the nonlinear mechanisms of pump-field

(a) the transient period of rapid growth of the

absorption well established, the question arises

fluctuations in which the parametric growth rates

whether such driven plasma fluctuations could also

affect the absorption rate

are comparable to the ion plasma frequency, w

pi’ and (b) the steady state maintained by a pump,

with

independent of

sampling field.

resistivity caused

interpretation

dielectric

involved.

measure.

plasmats

roughly

waves.

below,

the

field

the

and

pump

wel1

above

waves.

produced

acts as

broadening

threshold,

is therefore

saturation has

(or equivalent ac

dielectric behavior.

< m 2/&12 ~ -P difference,

microwave sources at

rates of both fields.

separations

I 86 MHz.

frequency

spectral

response

depends

quency.

There

with

Thus

that

data

(3.90

the

not

and

has

are

above

which

waves

We use

spectral

electron

nonlinear

described

turbulence

field to monitor

These regimes are

of + 15 MHz.

frequency

11.6,
27, —

times,

Those

and

can

—

test-driver

responsible can be nonlinear couplings similar to

The experiment is performed with a dual-mode

those of parametric instabilities but for which

microwave resonator (Fig. X-1) that is capable of

(because the fluctuation levels are externally

supporting two nearly degenerate modes (even and

maintained) the observed resistivity is associated

odd symmetry) with overlapping fields of the TMOIO

first-order current and

character and frequencies near 2000 MHz.

the amplitude of the weak ac

one of these modes as a strong driver field. When

Theoretical treatment of &his

the driver field is above threshold (E > ET) It

probleml~ suggests that electron (or ion) waves in

a plasma can enhance the resistivity of an exter-

generates and maintains electron and ion plasma

nal ac field, providing that the mode produced by

The other mode of the resonator is then

field-wave coupling lies in the resonant region of

used aa a weak, subthreshold ~

the ac resistivity at a somewhat different fre-

associated with ion (or electron) motion.

By switching off the external, pulsed,

the enhancement of weak-field ac resistivity by

suitable

plasma fluctuations involves matching conditions

measure and compare the simultaneous absorption

that can lead to frequency dependence from both

In accord with expecta-

the resonant behavior of the dielectric response

tions, the effective resistivity of the driver

and the spectral distribution of the fluctuations

field is governed by the nonlinear processes

In the case of the experiment described

field is found to be independent of test-field am-

studied previously; the resistivity of the test

plitude, but noticeably affected by the (driver

immediate: both strong electron wave.?and strong

produced) plasma fluctuations.

ion waves are generated by the pump field.

We have already reportedl the results of

are thus two major possibilities for the source of

experiments performed over the density range 0.65

the enhanced weak-field ac resistivity that we

One can argue that ion fluctuations are

‘test-fdriver’

unimportant because their parametric origin and

experiments investigated the absorption of test

the Manley-Rowe relations give them small ampli-

and driver fields both during the period of

tudes in comparison with the cogenerated electron

transient growth of parametric instabilities and

On the other hand, the enhancement of ac

during the subsequent saturated steady state, We

electron waves

have now extended our work to include test-driver

derivative

intensity.ll As the pertinent spectrum appears to

These more recent experiments have

be quite broad in some of our work, we camot

shown phenomena that are qualitatively similar to

Fig. X-5. Normalized reslstivlty of test (---) and driver ) fields corresponding to Fig. X-3 in order (— (top to bottom) of decreasing driver power.

,’

;’

,/’

—

0.4

TIME (ps)

fTEST ‘fORIVER

.’

/’

‘TEST=fDRIVER

”‘“6MHZ

,”,—L)RI I.A

,- ”, /’

” , -.’

;’,

‘TEST=fDRIVER

those found at + 15 MHz but that indicate a

decreasing influence of the plasma fluctuations on

the test-field resistlvity as its frequency be-

comes more distant from that of the driver. For

comparison, the ion plasma frequencY, upil~,

Thus, all our work has involved test

frequencies that are outaide the range of the

initial, unstable electron wave frequency: fe-wave

Figures X-3 through x-6 illustrate the fre-

quency dependence of the enhancement of test-field

resistivity during the growth of driver-produced

parametric instabilities.

—

30,

’ ‘driver - ‘ion-wave”

near 7 MHz.

.~ o

‘TEST

0.2

’”~

“o

.-’

0.- ‘u

0.4

0.8

0.6

~- U

‘EFF -qj-5

TIME

+11.6MHZ

and X-4

Figures X-3

test-fdriver

in the

for f

MHz

().M)

+27

0.6

0.8

I.0

0.2

—.

‘fDRIVER

0.4 TIME (ps)

Fig. x-4.

The five transient resiativlty measurements of Fig. x-6.

driver-field time histories used in the

The four driver-field time histories used transient resistivity measurements of Fig. X-5.

Fig. X-3.

. .

’.

,’

0.8

0.6

..-

----TEST

—DRIVER

’, ,,

+27MHz

----TEST

0.8

Fig. x-6. Normalized resistivi.tyof test (---) and driver ) fields corresponding to Fig. X-4 in order (— (top to bottom) of decreasing driver power.

show how the driver time history was controlled

. 11.6 and 27 MHz. The quantity

‘$l&, ‘:j_J-l.,

0.4

TIME(/K)

of the driver

rapidly

to the threshold level, ET, re-

decay away upon depletion of the driver field that

quired to excite instabilities, In each of the

several cases shown, the driver field in the res-

onator passes through three stages:

(a) During the application of an external,

pulsed, microwave source, the driver amplitude

rises rapidly to a maximum level controlled by the

time of termination of the pulse.

(b) The above-threshold field decays slowly

by means of electron-ion collisional absorption

(and resonator wall losses) while the unstable

plasma fluctuations continue to grow from noise.

the driver field occurs when the fluctuation

reach sufficient amplitude to draw substantial energy from it via nonlinear mechanisms.1

Figures x-6 and X-7 compare the absorption

rate exhibited by the driver field in these cases

with that shown slmultaneoualy by a freely decay-

ing test field, The quantity plotted is the ef-

fective resistivity (indicated here as an effec-

tive collision frequency, Veff)

plotted

is the square

tude, E, relative

enhanced resistivity of

Figs. X-5 and x-6.

one 27 MHz above,

‘test-fdriver

fields (vo).

enhancement

= + 86 ~Z.

enhancement

driver).

largest

because

the

one

—

fluctuations

Thus the maximum values attained

are determined more by the time history of the

driver field following pump ddpletion than by

inherent limits on the absorption process.

greater interest are the differences observed

between the times of onset of enhanced absorption

driver

explanation of the behavior of these differences

The onset of enhanced absorption

occurs when the pertinent fluctuations reach a

sufficient, critical amplitude, Kc, to compete

noticeably with electron-ion collisional absorp-

The absorption rate of the test field is

independent of test-field amplitude and therefore,

&c(test) is roughly constant for a given teat fre-

By contrast, our computations of para-

metric absorption ratesl show that, for the driver

(pump) field, &c(driver) is an increasing function

the driver field is increased, the difference

Thus as the average strength of

becomes

and

the

grow

rate

field

test

ampli-

tion.

quency.

absorption

is as follows,

generates them.

of driver power.

normalized to the

&c(driver)-&c(test)

fdriver by about 27 MHZ.

lence produced.

figures show

separations:

conditions.

increasing

cies grows

field is

likewise

figures.

becomes

enhance

between

onsets,

driver

delay

that

seea

that

The

the

first

rough

but then

fields.

positive.

driver

jump

t(driver)-t(test)

function

Figures X-7 and x-8 show such a comparison of

steady-state data with ftest both above and below

the normalized resistivity measured with either

field as a function of the driver field strength.

The quantity plotted is

characteristic

driver-field resistivity aa its amplitude exceeds

the instability threshold (E/ET . 1), but only at

When a sufficiently wide pulse of microwave

power is supplied by the external source, the

driver field in the resonator can be maintained at

a fairly constant level after pump depletion. The

result is a steady-state regime in which (for E

> ET) the unstable fluctuations have saturated.

By varying the driver power relative to threshold,

ities of test and driver fields under these

Again we compare the resistiv-

strength. This behavior is readily noticed in the

the

accompanied by a strong influence of the growing

fluctuations on the resistivity of the test field.

For a higher average driver power (faster growth

of unstable plasma waves) there is a greater

enhancement of the test-field resistivity with an

earlier onset in time (relative to that of the

Comparing the behavior of a test field

11.6 MHz above the driver frequency with that of

easily

resistivity

together. Note that we are here speaking of ftest

> ‘driver while the unstably growing electron waves these

fe wave < fdriver,

Satisfy

transient studies we found noticeable (factor of

of the test-field resistivity even

frequency

One should not ascribe particular Importance

to the heights reached by the absorption peaks in

The maxima in veff occur

electron-ion collisional value measured with weak

The plots ahow that the strongly

Because the fluctuation amplitude at all frequen-

with time during this period, the time

pronounced as the two frequencies come closer

(E/ET)2, we can control the level of plasma turbu-

I .0

(EIET)2

Fig. X-7.

0.1

enhancement

(Fig. X-9)

exception

notable

I .0

(E/ET)2

1 TEsT-fDRlvER+27

-t

6 ’

0.1

2 -

4 -

and driver (+)

vla four-wave processes

. 2). Such a Plot

enhancement of

resistivity.

test-fdriver

increased.

increases.

case

the

The

100

MHz

$1+

-1

+++ +++

This strong

One

Fig. x-8, Normalized resiatlvity of test (.) and driver (+) fields in the steady atate.

. -11.7 MHz which shows much more sensitivity than

the case ftest-fdriver = +11.6 ~Z.

influence seen on a test field just below the

driver in frequency is reminiscent of the less

saturated, more concentrated, parametric deoay

spectrum that occurs at lower driver powers.

is tempted to suggest that the enhancement of

test-field resistivity seen in these steady-state

experiments scales directly with the intensity of

electron waves at ftest and that we are seeing the

effects of symmetric spectral broadening (possibly

as the driver power Is

But such an interpretation discounts

the conventional theoretical result that it is

instead the derivative of the spectrum that should

determine the contribution of electron waves to ao

We are convinced that the enhancement of

weak-field ac resistivity that we have measured

can be attributed to the presence of fluotuationa

in our plasma. However, without further modeling

Normalized resistivity of test (*) fields in the 8teady state.

the higher powers does the test-field reaistivlty

show any enhancement. Presumably we see here the

effects of a broadening of the pertinent wave

spectrum (ion or electron) that occurs when the

fluctuation amplitudes are well saturated.

nearly identical appearance of these two plots

further suggests a frequency dependence of the

enhanced test-field resistivity that is symmetric

about the driver frequency.

In an attempt to document the above features

we can plot, as a function of ftest-fdriverv the

normalized driver power, (E/ET)2, required to

generate a fluctuation level that doubles the

test-field resistivity (veff/vo

indicates that

test-field reslstlvlty, as a function of ftest,

has symmetrio structure about ‘driver in the saturated steady state and that more driver power

(stronger turbulence) is needed to produce a given

I ‘test-fdriver1

the

We have already reported3 measure-

ments taken under conditions for which the plaama

was parametrically stable but for which the oscil-

latory velocity of an electron in the rf field, v~

. eEo/mti, was comparable to the random thermal

velocity, VT . [2kT/m]l’2. A pronounced reduction

observed

reduction attributed to the strong modification of

electron orbits as they interact with the screened

potentials of nearby ions.

dual-mode resonator, we have now applied simul-

taneous weak and strong fields at a frequency sep-

Figure X-10

the normalized collision rates of both the strong

field (crosses) and the weak field (apota) as

functions of the ac amplitude parameter, VE/VT,

associated with the strong field.

cation of electron orbits by the strong field is

seen to affect the absorption of the weak field so

that indeed both fields show identical reductions

collisional resistivity.

supports the idea that an intense ac field can

produce a genuine, frequency-independent reduction

of the electron-ion collision rate of the sort

resistivity enhancement.

frequencies

properties

transient

are

ion

the

was

instability.

of ac resistivity

aration of 28 MHz.

Because of the

In the proc-

fluctuation

We have

(e.g.,

I 0.5

From

the

1.0

0.s

)*“A

Fig. X-9. Dependence on test frequency of the driver field required to double the test-field resistivity,

of the processes involved, the data by themselves

do not suggest whether electron waves or ion waves

make a dominant contribution.

Implied frequency and wave-number matching re-

quirements it is probable that in either case the

levels, dielectric response) at both electron and

plasma

important.

experiment we have obtained a fairly complete

picture of the phenomena involved in both the

and the steady-state regimes.

documented their dependence on (a) the test and

driver frequency separation, (b) the field ampM.-

tudes, and (C)l the plasma density.

ess we have identified many general features that

can serve as tests of theoretical models for ac

this

case,

This

observation

The modifi-

shows a plot of

Through use of the

vE/vT

I 1.0

Modification of Electron -Ion @llision

Rate by Intense RF Fields. The dual-mode resonator

has also been used to confirm earlier work on

high-field effects that do not involve parametric

Fig. X-10. Normalized collision rates of weak (.) and strong (+) fields as functions of the strong-field ampli- tude.

zero

that

might

UCLA.

Jinear

affect

field.

magnetio

plasma source.

wave interaction.

investigated.

noted.

probe.

current

Two

transport properties as well.

other

Progress on an Experiment to Study Non-

Interactions Between Plasma and Traveling

Ilec tromaunetic Waves. Q-machine research, work is in progress on another

This is a cylindrical,permanent

magnet multipole-confined argon plasma device of the type developed by MacKenzie and coworkers13 at

Experiments presently planned for this

machine include an investigation of the decay of a

photon into two plasmons, (the “2 OIpinstability”)

and a study of heating a plasma by beating two

electromagnetic waves whose frequencies are sepa-

rated by a natural-mode frequency of the plasma.

These experiments involve traveling

netic waves, in contrast to the Q-machine studies,

and they take place in a region of essentially

100-cm-long cylindrical mesh waveguide, trans-

parent to plasma, is mounted coaxlally with the

plasma source in the field-free region of the

plasma, and is used to contain the plasma-micro-

Our research during the past year can be

divided into three categories:

(a) Characterization of the plasma In the

presence of the mesh waveguide.

(b) Investigation of the plasma noise near

up in the absence of microwaves.

of plasma on the propagation of subthreshold (for

the 2 up-instability) microwave signals.

different mesh

The original waveguide, made of

copper-plated copper wire, had a mesh spacing of

6.4 mm and a transparency of 62%. Representative

axial and radial profiles of density and tempera-

ture can be seen in Figs.

temperatures are obtained using the In(I) vs V

characteristic of a Langmulr probe near space

potential. Higher temperature tails were usually

The axial density profiles were measured

using ion saturation current to a plane probe, and

radial density profiles were inferred from elec-

tron and ion saturation current to a cylindrical

Although these data are for a discharge

of 1 amp at 40 volts with the mesh at

DISTANCE FROM MIDPLANE

6.4-mm MESH ~FILAMENT+

Fig. X-n.

Electron density and temperature on axis vs axisl position measured from mldplane for 6.4-mm mesh spacing. The product neT is also shown. Mesh at floating potential (= - SO V), discharge voltage = - 40 ~, discharge current = 1 A, gas Pressure 8 x 10-4 torr.

3 -

4 -

5 -

the

6 ------1

this

FILAMENT

achieve

addition

electromag-

collision-dependent

Electron midplane spacing.

waveguldes have

X-II and X-12.

&EsHR,Dlusl

I PROBE

been

The

I LENGTH

(cm)

‘NYU

ne x 10

cm-3

-lo

density and temperature at 30.5 cm. from 6.4-mm mesh

Vs Same discharge conditions as Fig. X-n.

position

radial

for

4 RADIUS (cm)

Fig. X-12.

radial profiles of density and temperature for the

Fig. X-14.

parameters were varied widely.

(=- 30 V),

potential

interior densities, as, of course, did higher cur-

However, sputtering became serious when

either the filament bias (discharge voltage) or

the mesh potential was more negative than - 40 V.

disadvantage of this mesh spacing

was its relative lack of transparency to the plas-

ma, reducing the plasma density inside the mesh to

the

investigation of the 2 up instability, undesirably

high discharge currents would be required to

achieve the necessary density.

For this reason, the mesh was clipped to give

12.7-MM mesh spacing and a transparency of 82$.

This resulted in much higher interior plasma

densities but also caused a greater disturbance of

microwave propagation in the waveguide. Axial and

same operating conditions as Fig.

and X-14.

Several checks on the density measurements

that

mesh

—

rents.

outside

The main

negative

floating

potential

seen in Figs. X-13

have been made.

interferometry

FILA~NT

20 - Ak

DISTANCE

l------+

o A O

12.7-mm

3.0 -

25 -

Q5 -

I .5 -

Electron density and temperature at midplane vs spacing. radial Discharge parameters same as Fig. X-n.

12.7- mm

position

mesh

for

(

I I

X-II

2 -

Alrr

the

For

gave

RADIUS

mesh.

these

three

higher

may be

It was found that

absence of plasma.

ments

using

10%.

mesh

also

the

are

(cm)

—

[

\
\

MESH RADIUS

(cm]

PROBE LENGTH

tip, amplified

to relate

difficult

Because the two phenomena to be investigated

excite plasma waves near up, noise around Up in

the absence of microwaves is being examined. This

noise is detected by a shielded probe having a

1.6-mm-long by 0.15-mm-diameter

20 db by a low-noise amplifier, and displayed on a

spectrum analyzer. At the time these measurements

were made, the ends of the waveguide had Mylar

windows which had become heavily coated with metal

quency, and fco is the cutoff frequency in the

This determines up and thus

ne. Because of the radial profile, however, this

ne is less than the density on-axis, which the

axial probe measures. The inter~erometer measure-

somewhat

precisely to probe measurements because of the

radial and axial profiles in density, but the

Density in the wavegulde can be

waveguide. Density was Inferred from the measured

inferred from microwave cutoff experiments, and

cutoff frequency by assuming the plasma uniformly

average density may be deduced from microwave

filled the waveguide and using the relations fc

measurements

= ~ fcoz + (IJ)p/21T)2,where fc is the cutoff fre-

FROM MIDPLANE

Fig. X-13.

Electron density and temperature vs axial position measured from midplane for 12.7-mm mesh spacing. Discharge parameters same as Fig. X-II.

Te(eV)

phase shift predicted for the profiles of Figs.

X-13 and X-14 agree with the measured shift within

MESH

FI LAMENT

ferometer measurements (not shown) alao tend to

controlled Thermonuclear Research

by sputtering. Several features of the noise were

these frequencies would be far below the wavegufde

(a] For the one case where noise spectrum

(511)0 W)

radial profiles have been meaaured thus far (dis-

have been done on the 12.7-mm mesh guide to

charge current m 1A], the noise level was much

establish the performance of the system in the

(10-100 times) inside the mesh than

Investigations at higher powers

will be carried out when our l-kW TWT amplifier

(b) The noise frequency was essentially the

system is in operation.

frequency corresponding to the peak density

than the l-kW level we will either resort to a

inside the waveguide over a wide range of density.

higher power microwave system, or utilize micro-

shows, as a function of discharge cur-

wave resonator techniques to enhance the field

rent., the noise frequency, the plasma frequency

levels of the traveling waves experienced by the

calculated from probe density measurements, and

the plasma frequency deduced from microwave cutoff

The latter tend to underestimate

the on-axia density, as mentioned before.

lie below the probe data by about 10$.

When the coated Mylar windows were removed,

the behavior of the noise changed completely. The

noise was much greater (- 10 times) just outside

the mesh than inside. That such a change at the

end-boundaries inside the mesh should produce this

change in noise distribution is somewhat sur-

prising unless the conducting ends formed a ‘tres-

onatorn for the waves associated with the noise.

It should be noted that electromagnetic noise at

plasma

noted.

greater

Fig. X-15

outside.

measurements.

,ofJll

,.11 /

,11

!’

I!LASL

Inter-

plasma.

cutoff.

Low-power

REFERENCES

linear regime.

I l CUTOFF MEASUREMENT . CYLINDRICAL

ION SATURATION

fp FROM NOISE

(1973).

PROBE

1 1 I

.AAA

,11’”

I I

IIty

, ,,

)

,.,

DISCHARGE CURRENT(A)

Fig. X-15. Plasma frequenoy (Hz) vs discharge current (amps). Discharge voltag~ -40 V, mesh bia, -40 V, neutral pressure 4 x 10- torr.

microwave

experiments

To reach powers higher

Research

Program, January-December 1976,n Los Alamoa Scientific Laboratory report LA-7082-PR (1978), PP.

110-126.

A. T. Lin, P. K. Kaw, and J. ?4. Dawson, ‘A Poesible Plasma Laser,!!Phya. Rev. Al., 2618 (1973).

J. H. Brownell, H. Dreicer, R. F. Ellis, and J. C. Ingraham, ItInfluenceof Intense AC Electrlc Fields on the Electron-Ion Collision Rate in a Plasma,!!Phys. Rev. Lett. ~,

1210 (1974).

o. Buneman, “Instability, Turbulence, and Conductivity in Current-Carrying Plasma,it Phys. Rev. Lett. ~, 8 (1958).

Dawson and C. R. Oberman,

J. M. IfHigh-FrequencyConductivity and the El!JiSSiOn and Absorption coefficients of a Fully Ionized Plasma,ftPhys. Fluids~, 517 (1961).

T. Musha and F. Yo8hida, ‘Negative Absorption Due to Coulomb Scattering of an

Electron Stream,fgPhys. Rev. ~A, (1964).

H. Dreicer, nElectron and Ion Runaway in a Fully Ionized Gas,!?Phys. Rev. ~, (1959).

H. Dreicer, D. B. Henderson, and J. C. Ingraham, llAnomalous~crowave Absorption near the Plasma Frequency,n Phya. Rev. Lett. 26, 1616 (1971).

H. Dreicer, R. F. Ellis, and J. C. Ingraham, I!Hot-Electronproduction and Anomalous Microwave Absorption near the Plasma Frequency,n Phys. Rev. Lett. X,

WLASL ~~ntr~~lled Thermonuclear Program, January-December 1975,” L08 Alamos

Scientific Laboratory report LA-6582-PR (1976), PP. 100-110.

s. Y. yuen, IfSaturationof parametric Decay

Instability by Nonlinear Bull. Sot.

Ichikawa,

!!High- of a Plasma with

Fluids J2, 2563

and K.

t!Magneti.c Uniform Collisionless Quiescent Plasmas,if Rev. Sci. Instrum. ~,

Containment

Phys.

and Y.

Nishikawa

conductivity

K. Frequency Fluctuations,g’

Am.

(1969).

Phys .

Multipole

Rudolf Limpaecher

Mackenzie,

of Large

726 (1973).

Four-Nave u,

cOUpllng,n (1977).

J. Hammel, C. Ekdahl, K. Hanks, L. Jones, W. Nunnally, D. Thomson

CTR

force

by a

effort

proposal

original

19’76. to

initiated

in January

cooperative

investigate

Divisions.

INTRODUCTION

Marx Power S-

erected six high.

of a 10-cm-long

of 12 Maxwell

pinch chamber

the average

pinch.

consists

coupled

should

An experimental program was set up in May of

for such an experiment

a very

small-diameter

the

Division. 1

between

The experiment will be used to investigate a

Z-pinch under the conditions of very high density

(n . 1020 to 5 X 1020 cm-3) and very small radius

(1 X 10-2 cm to 5 X 10-2 cm).

radius, the current density will be large and ohm-

ic heating should take the temperature to many

kilovolts for currents of 0.7 to 1.0 MA.

rent will need to rise very rapidly for the pinch

field to be able to maintain a small pinch radius.

An idea of the required current rate of rise,

under assumptions of a simple pinch model with

classical behavior is presented in Ref. 1. For a

pinch density of 1020 cm3 and a radius of 350 pm

rate of rise from initiation

be 5 X 1012 A/s or greater.

pinch 350 pm in radius

is about

above dI/dt implies a voltage of 500 kV across the

B. DESCRIPTION OF EXPERIMENT

source, a Marx generator has been designed to

operate at 600 kV with coupling to the pinch load

through a water transmission line.

consists of twelve 6-kJ Marx units fed in parallel

to a I-Q, 90-ns water transmission line.

pinch chamber is then closely coupled to the

transmission line in a manner which is similar to

the diodes in an e-beam machine. The Marx module

connected two in parallel at each level and

100-kV,

Plasma~tiation

The

and

The

task

CTR,

high

beam.

laser

Z-pinch

density

concepts

was made

alternate

experiment

The cur-

machine.

deliver 18 J in

With this small

XI. HIGH-DENSITY Z-PINCH

100 nH and for the

O.1-PF capacitors

breakdown occurs.

The inductance

The supply

90 kV/cm.

in a close

duration

to 200 kA

current

25 na.

about

30-ns

The

For

the

The modular design has an advantage in allow-

ing a one or more unit prototype to be built and

tested before final construction of the full

A one-module prototype has been con-

structed with a 5-fi, 60-ns line and test pinch

chamber. This prototype has been used to check on

the Marx triggering, voltage holdoff of the system

and also a full voltage test of the pinch voltage

breakdown in different geometries.

been operated at a full 600 kV into dummy loads.

Near full voltage, the erection time is about

80 ns with an observed jitter of about 30 ns. At

low voltages the jitter becomes larger.

chased for the initiation of the pinoh is a

Holobeam neodymium glass and is specified to

18 ns with a beam divergence of

3 mrad. The actual performance of the laser does

not meet specifications and is now being worked on

by Holobeam. When the laser becomes operational,

laser breakdown studies will be made prior to the

actual pinch experiments on the full machine.

PROTOTYPE EXPERIMENTS

The prototype is in operation now for prelim-

inary pinch experiments with currents up to 60 kA.

For the initial experiments the initiating pulse

will be from a ruby laser with an energy of 7 J in

The short-pulse breakdown voltage for H2 in

the pinch geometry was measured. At a H2 presnure electrode 8eparation the of 3 atm and a $cm breakdown was at about 300 kV for a pulse rising

in 180 ns. ‘rhebreakdown was apparently initiatea

at the point of highest stress which is 80 to

The first pinch experiment with the prototype

will be an initiation study using the ruby laser

on the present gas chamber which has 5-cm elec-

trode separation. It is desirable, if not essen- tial, to put the laser pulse in a time window of

The Marx has

. The laser pur-

spontaneous

before

The spark gaps are Physics

just

International 100-kV, 100-kA gaps.

laser pulse should be timed from Marx erection to

This means that the initiating

1. J. E. Hammel, ,,AnOhmically Heated High- Density Z Pinch,” Los Alamos Scientific Laboratory report LA-6203-MS (January 1976).

eliminate the jitter in erection time. This puts

stringent demands on the electronics since the

total time from erection to gas breakdown is about

Fast spark gap triggering for laser and

Imacon triggering is being tried.

developments are complete we hope to have our

first pinch experiments on the prototype.

24o ns.

REFERENCE

When these

F. C. Jahoda, W. T. Armstrong, P. R. Forman, K. B. Freese R. Kristal, and R. E. Siemon

The scanning infrared heterodyne interfero-

conclusion chapter are reproduced (Sec. G).

Three desirable diagnostics in an advanced

conceptual or early design phase for ZT-40 are

far infrared Faraday rotation for poloidal

field measurement, (2) short-pulse laser scatter-

for density profiles, and (3)

laser scattering for time history of electron

temperature (Sec. H).

A substantial fraction of the Diagnostics

Group effort centers on the uae of minicomputers

for data acquisition and for device control as

well as adequate electromagnetic shielding between

computer and plasma device.

reported in Sec. XIV.

SCANNING INFRARED HETERODYNE INTERFEROMETRY

This project ia concerned with measuring

side-on plasma density profllea in the parameter

range where visible interferometry is not sensi-

tive enough (~ ndl < - 1017 m-2).

Obtaining sufficient phase and apatfal

resolution to enable accurate Abel inversions of

meter has been succeaafully bench tested wi$h

inclusion of a vacuum chamber in the plaama arm.

Resolution of 25 spots over a S-cm field of view

in 0.25 us with 0.1 fringe accuracy (~ ndl - 3 x 1015 ~m-2) haa been demonstrated.

awaita installation of flat windows for aide-on

viewing on Scylla IV-P.

larger field of view and capable of a trade-off of

more resolution apota for slower total scan time

for ZT-40 ia under construction (Sec. B).

The insertion of end plugs in Scylla IV-P

increased the electron density near the plugs eff-

iciently so that side-on ruby laser holographic

interferometry became a very useful diagnostic.

It haa also renewed intereat in a long-pulse ruby

laser as a source for multi-frame interferograms

from a single discharge. The activity under way

to achieve such a ruby laser with adequate phase

coherence is described (Sec. C).

A portable Thomson scattering apparatua is

being constructed in order to meaaure electron

several

INTRODUCTION

stand (Sec. D).

high-energy

temperature

calibrations (Sec. E).

ing

(1)

sity

Colorado.

The system

A second system with

XII. PLASNA DIAGNOSTICS

ent frequencies.

Obtaining

interest are

resolution

locations

standard

teated.

include

axial

over

The spatially reeolved Thomson scattering

apparatus ia being modified to increase signal-

to-noise ratio. This entails extensive expansion

of the automated read-out and computer assisted

and

The

mode-locked

thesis abstract

This activity is

During this year a ayatem

lhe concept waa de-

fractional - fringe

Of particular

mlti-fringe

heterodyne

range,

and

Some general-purpose computer codes relating

complete description and documentation for the

(1) to ray tracing including phase errore due to

existing system follows.

beam bending and (2) to impurity radiation as an

(stationary beam)

energy leas mechanism and as a temperature diag-

interferometry the deneity time history ia ob-

nostic, have been developed (Sec. F).

served aa a phaae distortion of the beat (heter-

Analyeis of previous data on C02 laser-theta-

odyne) signal between two optical beams of differ-

pinch plaema interaction haa been completed, and

a doctoral dissertation submitted to the Univer-

The virtues of the technique

self-calibration,

Scylla IV-P. A more compact version of the three-

interferometric data.

grating polychrometer ia combined with a special

multiple profiles during

ruby laser oscillator on a single

single discharge to avoid the ambiguities of

plasma irreproducibility. scribed previously.1

has been completely built and successfully bench

It ie ready for installation on Scylla

IV-P as soon as flat windows are available.

second system is under construction for ZT-40.

unambiguous sense determination. Also, the output

beams in this way, in identical fashion, effects

is in electronic, rather thsn photographic, form

an enormous reduction in apurioue differential

which permits more rapid and automated data analy-

phaae shift that might otherwise occur, as well as

optimizing signal-to-noise ratio.

By combining standard heterodyne interfero-

The angular scan off the rotating mirror ia

metry with rapid spatial scanning, the temporal

converted to a linear scan by placing the rotor at

(heterodyne) fringes are converted to epatial

the front focal plane of the scan lens. The two

fringes. By making the scan time short enough to

bearns scanning side by side are

“freeze” the plasma, multiple density profiles can

separated into scene and reference arms.

be obtained during a single discharge using a

field epot forms at the back focal plane, at which

The advantages of the

point the return mirrors are placed,

technique accrue mainly with the lower density

The return beams are super-

(area density < - 1017/cm2) where

posed on the beamaplitter and collected by the

highly developed photographic techniques available

focus lens, which focuses the entire scan field

in the visible region of the spectrum are not

Tbua the rotor is imaged onto

sensitive enough, and hence the infrared wave-

the detector, which insures a stationary apot on

The system layout is shown in Fig. XII-I.

time a new mirror

cw He-Ne laser beam at 3.39 pm ia focused inside a

rotating turbine aweeps by.

Iead-molybdate acousto-optic cell. This generatea

The scan field parameter

the detector during the scan.

a first-order diffracted beam, frequency shifted

Scene beam scan waveforms (acouato-

by - 100 MHz from the zero-order beam at the laaer

shown

lhe zero order ia used as the scene

resolution test grills in place of the object, for

single

sis.

detector

plasmsa

channel.

frequency.

lengths must be used.

SCAN LENS, 2m

R- REFERENCE BEAM

S-SCENE BEAM

f,a~10k~~

~~URB~~;E

150n (CYL.)

Im (CYL.)

S+R

the

are

every

system

off)

cell

0 ccurs

Optic

Fig. XII-2.

double pass.

onto the detector.

Scanning the two

~0+=.93 kHz

SPATIAL

BEAMSPLITTER

05ps/dw

OBJECT

’=‘2

(b)

(A)

Y *

(c)

(d)

BRAGG CELL,- 100 MHZ

beam and the first order is the reference beam.

The two beama are collimated and impinge side by

side on a turbine driven rotating mirror, which

provides the spatial scanning.

BEAMSPLITTER

zDETECTOR,Ge” Hg

~:C:S LENS,

~ &R

Fig. XII-1. System optical schematic.

the

scan

making

with

The far

facet of

different

A new mirror

then easily

are evident from

~0+= 9.5 kHz

50ns/dlv

5m ‘a)IElmi19

Fig. XII-2 . (a) Normal scene beam scans; (b) Resolution test grill in object plane, 10 mm/period (11 spots across field); (c) 5 mm/period (21 spots); (d) 2.2 Object plane is 17 cm from mm/period (50 spots). return mirror.

RESOLUTION

low and high rotor speeds.

5.5 cm, in a time aa short as l/4 PS.

modulation due to the test grill is observed at 21

spots, even at top speed.

is 25-30 spots, corresponding to a maximum scan

rate of — 100 spots/Us.

Heterodyning while scanning at low speed is

shown in Fig. XII-3. The heterodyne frequency is

100 MHz, and so appears to fill in the waveform

at the sweep speed used.

test grill [Fig. XIIT3(a)], demonstrates the pos-

sibility of obtaining virtually 100% contrast.

Figs. XII-3(b), 3(c), and 3(d) show resolution

test results in agreement with Fig. XII-2.

High-speed heterodyning with a test cell with

medium-quality quartz windows under vacuum in the

plasma arm is shown in Fig. XII-4. Three linear

phase objects are simulated by three different

tilts of the scene beam return mirror. The heter-

odyne frequency may not be constant across the

Fig. XII-4.

quartz

(b) High-speed (9.6-kHz) heterodyne scanning (a), Lower through trace - acoustio trace - heterodyne scan; upper si;filrd), (d) scene beam return mirror tilted ; (e), (f) mirror tilted - 1/4 mrad in

windows

under

PHASE OBJECT

(a)

FAST HETERODYNE

LINEAR

Strong

50 ns /div

SCANNING

opposite direction.

The scanning without

The scan field may be

‘lhe actual resolution

scan Fig. XII-4(b).

many points.

at low speed

accuracy.

profile.

REE

TILT)

20ns/div

(MIRROR

vacuum.

SCANNING,

POS. TILT

NEG. TILT

be~ter than that for

For this reason a

The valleya in

The

scene and reference optics.

reference exposure technique is used, whereby a

scan made with the object (data) compared to a

scan made without the object (reference).

difference in the two phaae plots givea the object

‘Ihe accuracy of this procedure may be

found by comparing scans 4(d) and 4(f) against

4(b), in Fig. XII-4 since the differences are

strictly linear phase profiles.

the fringe pattern have been digitized in each

case and the differences appear in Fig. XII-5.

Linear interpolation was used on the reference

the linear profile is s 0.1 fringe for some

‘Xhe maximum deviation from

points, the error being mch

The error includes sources such as

digitizer, scope, and camera optics, so that the

basic system is capable of better than 0.1 fringe

Multiple framing with the present system ia

limited to 100-us frame interval (single facet at

10 kHz). A rotor is on order (from Cordin), which

haa 12 facets and rotates at the same top apeed.

Fig. XII-3.

(a) Heterodyne

Lower trace: (fret-o.94 kHz) with quartz windows on vacuum beam; (b), (c), and (d) same chamber in resolution teat grills as in Fig. XII-2. Upper trace - scan at low sweep speed; middle trace - acoustic burst.

scanning

scene

SLOW HETERODYNE

scan, owing to slight phase differences between

[MIRROR

Also, a staclonary beam

of two, thereby simplifying the data acquisition,

—

OBJECT

POSITION-cm

LINEAR PHASE

will remain the same.

access is available.

beam if necessary)

phase detection

oscilloscope

In applying

(z20-cm

electronic

diam).

face.

the

system will be Incorporated on axis.

allow normalization of the measured profile in

case the density in the wings of the plasma pro-

file is not negligible compared to the peak.

LONG-PULSE RUEY LASER FOR HOLOGWPHIC

All previous ruby laser interferometry on

theta pinches has been end-on because of the sen-

at A = 694 nm one fringe

displacement is produced by J ndl = 3.2 x 1017 -2cm

[This fact was the motivation for the

hydrogen fluoride laser interferometry (X=2.9 P,

one fringe for ~ ndl = 7.7 x 1016 cm-2), success-

fully but painfully carried out on Scyllac in the

Even for this wavelength the sensitivity is

that, after considerable effort, it

manpower-ineffective

insertion

—

’\

TILT

the

was

POSITIVE

NEGATIVE

’\

With

past.

TILT, = 1/4 mr)

abandoned

so marginal

Scylla IV-P.]

INTERFEROMETRY

sitivity limitation:

storage, and analysis.

period of this report.

ferometry becomes

at least a factor

allow a reduction

ferometry.

to ZT-40,

initial

This

some

the

—

end

This

into

plugs

This will

diagnostic.

experimental results

Ex-

mis

Scylla IV-P, the density is greatly increased near

the end plug and side-on ruby holographic inter-

represents a happy circumstance in that the basics

the diagnostic were well developed before

suitable circumstances for ita application arose

and it moved from the conception of the idea in

the Diagnostics group to being a standard working

measurement on an experiment within the reporting

useful

The

are outlined in the Scylla IV-P section.

development has also stimulated renewed interest

in developing a long- (500-Pa) pulsed ruby laser

suitable for multiframe live fringe holographic interferometry,2 since the dynamics are rapid

time, data shots are limited, and in some cases

the window contamination problem is severe.

perimental effort waa prompted by comment from

some Soviet visitors to the effect that confocal

cavity ruby lasers yield quasi-de outputs with

about 10% amplitude modulation suitable for inter-

Figure XII-6 shows a typical output pulse

obtained from such a confocal cavity ruby laser.

The laser consisted of two mlrrora with 30% and

100% reflectivity each having a radius of curva-

Fig. XII-5.

phase

Figs. XII-4(d) heasured and -4(f) (data scans) using Fig. XII-4(b) as the reference scan.

profiles

for

When multiple framing is used, it is a simple

matter to get four sequential scans on a single

Therefore

configuration, one scope is required for each four

profiles during a discharge. Development work on

is under consideration.

The system as described above is suitable for

Scylla IV-P, even with only one frame per dis-

charge. It will be installed whenever flat window

concept

changes are needed to accommodate the larger scan

format. This will consist of X6 expansion optics

in the scene beam (and shared with the reference

to 33 cm, adequate to cover the zT-40 plasma

to increase the field from 5.5

If the scan time is kept con-

stant, then the number of spots across the field

However we will probably

capitalize on the much slower dynamics anticipated

in ZT-40, to allow a scan time of ~ 1 us.

will permit a maximum resolution of - 100 spots

across the field. It will also

in the heterodyne frequency by

Output from long-pulse intensity modulation.

Fig. XII-6.

The conclusions drawn from this series of

that while such lasers are adequate for

simple interferometers, they are not adequate for

holographic interferometry.

variable transverse mode structure between the

time-separated wavefronts that form the interfero-

grams prevented high-quality results.

of a pinhole into the laser cavity to achieve good

transverse mode selection resulted in good holo-

grams but changed the laser output from quasi-de

to irregularly spaced intermittent pulses that

synchronized

recording. This result led to the abandonment of

The quasi-de output from a confocal cavity

results from the random auperpoaition of many

longitudinal modes of equal frequency and random

‘he resulting nearly uniform intensity

does not provide the phase coherence needed in our

oscillation.

function

ably

due

~is

phase.

cannot

tests ia

flat anti-

application.

this approach.

ruby laser with modest

faces introducing more

If this is not done,

The laaer head must

During the set of

generator.3

ing wave

threshold

operation

voltage.

cavity

which

ruby

rod,

with

This

The

are

ture of 100 cm. The ruby was a standard 0.95-cm

has

diam by 10-cm Korad laser head with

single longitudinal mode operation which, however,

reflection coated end faces.

generally suffera from longitudinal mode hopping

The cavity length was varied and it was found

due to gain profile hole burning. TWO approached

that a cavity length of 109 cm produced the

to eliminate mode hopping even during long pulse

smallest degree of oscillations, but this ia far

being

from a criticsl adjustment.

method employs compensated phase modulation (CPM)

tests, it was discovered that one must skew the

and the second uses a ring laser.

laser head diagonally as much aa poaaible to make

The CPM method consists of driving two I(DP

the facea of the rod as far off from parallel with

crystals 180° out of phase, one at each end of the

the mirrors as possible.

many strong spurious reflections occur within the

‘l’hiseffectively moves

cavity and the output is dominated by very strong

standing wave

through the ruby rod, allowing more homogeneous

Positioning the laaer head other than in the

depletion of the active medium.

center of the cavity was tried. It was found that

method was reported for a confocal cavity with a

the farther off center the rod was placed, the

pumping voltage three times threshold and resulted

worse the oscillation in the output became, prob-

in reducing the modulation to less than 10%.3 Our

the rod

initial attempts at such experiments have not

reflections as the beam walk-off between the rod

proved as successful. We have applied 2 kV at 40

and end mirror becomes less.

MWZ to two longitudinal Pockels cells and as yet

be placed somewhere close to the center of the

have not noticed any decrease in modulation depth

cavity, but it is not a critical positioning.

of the random spiking.

Performance of the laser waa studied as a

A ring laser is currently under construction

pumping

to test the second approach.

voltage was found to be 2.8 kV and a typical

using a 10-cm length of SF-6 glass has been built,

operating voltage was established at 4.8 kV.

pumping voltage gives a useful pulse of about 250

determines the lasing direction. Having a travel-

conjunction with

precludes

IIS and contains 18 J energy.

with

camera

framing

Insertion

investigated.

In particular the

forced us to consider

A Faraday rotator

half-wave plate

The use of this

high-frequency

pattern back

possibility

and forth

the laser

first

hole

sine

wave

the

The

achieve constant amplitude. For thLs purpose part

the coil snd then moved into place, guaranteeing

of the laser output is directed to a high-voltage

that the laser snd detector see the same plasma

has the advantage that it may be aligned

‘I%e signal

on a Pockels

the

arrangement

The triple-grating polychrometer built for

this experiment is a modification of Siemon’s

design suggested by Greenwald snd Smith.6

in the literature.

new design requires one less mirror in the optical

path than Siemon’s design, and it is relatively

PORTABLE THOMSON SCATTERING

However, it does introduce

The interest in axial heat conduction in

some astigmatism and chromatic abberation into the

generated

problems can be reduced to

axially resolved electron temperature measure-

this, it is proposed that one

Figure XII-8 is a schematic of the Greenwald-

Thomson scattering experiment be set up at the

SmLth design. Esch stage of the polychrometer ia

center of the theta pinch and another portable ex-

essentially identical to the rest. Light entering

periment be moved along its length, each making

the device through the entrance slit, S1, is

simultaneous measurements of the electron temper-

collimated by lens LI and falls on the first grat-

ature. The fixed-location Thomson scattering mea-

The dispersed spectrum is then imaged

surement has been installed on the center section

onto a mask between the first and second stage.

of the theta pinch and construction has begun on a

dispersed light is recombined by the second

small, compact, portable laser-detector system

which can be moved to different axial locations.

Siemon’s

Figure XII-7 shows the layout of this new

Greenwald-Smith design is the way this second

It consists of a 5-J ruby laser (Korad

stage recombines the dispersed light.

K5Q) mounted in a stand below the theta-pinch

XII-9 shows a detail of the first and second

On top of this same stand is mounted s

stages of the polychrometer. Light of wavelength

triple-grating polychrometers to detect the spec-

k incident on the entrance slit enters at an angle

trum of scattered laser light. This arrangement

al, with respect to the normal of grating G1. The

grating Is used in a near Littrow mounting and the

light leaves the grating at angle 131where

feedback

the

has

r).

diode

cavity

TO do

114 A).

Negative

feedback

burning.

decrease

directly

similiar

ments.

system.

is placed

Scylla IV-P

has been reported

photodiode (ITT

(w ;;$.#

coil.

mm OUUP

1 I

PLA5MI

. .

.,,

—@

,…

The

and

for

will

from

this

cell

These

laser

using

a Kerr

output.

imaged

between

grating

volume.

ing G1.

cell within

be attempted

difference

requirement

optical system.

a tolerable level.

essier to fabricate.

mmmfcl

1 I

the

and

The

S2.

This

slit

onto

Figure

design

away from

(1)

&~=j—

TRIPU-G+AIINGFCiYWiWA1OR

RLS7LAWlSiSTEM

Fig. XII-7. Portable Thomson scattering set-up.

Fig. XII-8. Triple-grating polychrometer.

i.e., the focal length of the second-stage lens L2

compensates for the differences in dispersion by

effectively magnifying the mask image.

design constraints require fl = f2, thus making

Eq. (5) impossible. The solution to this problem

proposed by Greenwald and Smith ia to move the

field lens FLI so that the combination of FL1 and

L2 has an effective focal length f2 as ahown in

For fl = 381 nxn,fm = 189 mm, and

f2/fl = 0.918 at A - 6500 ~, the lens separations

are a = 15.5 mm and b = 364.1 mm. When the spec-

trum A(x) at the mask is fairly linear, wave-

lengths other than the central wavelength are also

recombined successfully.

only 300 ~,

problems. We however, wanted a wider band-pass,

and nonlinearities in

blurred recombined image at slit S2.

required is a lens combination that has a variable

Fig. XII-9. First two stages of triple-grating polychrometer.

and 9. is the line spacing of the grating.

linear dispersion at the msak is given by

For

dkl —.— dx

dAl —=~ fl d61

second stage should be

dk2 _.---- dx

f2 da2 ’~

~=dx

dA2

For

dA2

d~l

nSI

COSB1

The

(2)

‘2 ~=—

Cos 81

800 8,

Cos 13~

Fig. XII-10.

band-pass was

Table XII-I.

(3)

(4)

coa 62 .

solution ie found by requiring that angle a2 = B1 In our

and 132= al for the central wavelength A. system the mask transmits wavelengths from 6900 8

to 6100 ~, thus making the central wavelength 6500 For the second grating G2 to faithfully

recombine the light, the linear dispersion for the

we must have

(5)

Other

What is

k(x) produce a

and there were no

In Ref. 6, the design

‘he required lens

the dispersed light to be faithfully

focal length as a function of wavelength such that

recombined, the linear dispersion of the second

the two dispersions can be matched. The required

stage must be identical to that of the first. It

focal lengths vs wavelength are summarized

can be shown that when this is the case, the image

One way to approximate this is

of the recombined beam will fall on the entrance

tilt the field lens FL1 so that the combination

slit. lTIis is the Siemon design and requires a turning mirror to avoid this complication. In the

lenses FLl and L2 magnifies the different parts

the spectrum appropriately.

Greenwald-Smith design an alternative approximate

separations are also given in Table XII-I.

least squares linear fit was used to approximate

a(x) and determine the amount of tilt for FLI.

Fig. XII-10. The use of the field lens FL1 to control the dispersion produced by the collimating lens L2.

this

Being forced to return to a uniform master

scan pattern, another method was sought to combat

the dark-current buildup on the tube.

was cooled by passing chilled dry nitrogen through

its enclosure, and it was found that cooling the

greatly

This technique effectively eliminates

the dark-current problem and permits multiple read

Each one of the 10 000 pixels read must be

calibrated for accumulated charge vs integrated

The calibration data for each

pixel can in principle be stored on the PDP-10 and

used to normalize the data for analysis. However

this would require approximately 106 words of data

characterize the calibration curve of each pixel

by a few (- five) parameters, which would reduce

the amount of storage to - 50 000 words.

Also, the triple-grating polychrometer and

SIT tube detector must be jointly calibrated for

wavelength vs spatial coordinate, again for 10 000

TO simplify this procedure, a CAMAC-

motor

This method Improves the image quality at S2 to an

TABLE XII-I

LENS SEPARATION PARAMETERS FOR DIFFERENT WAVELENGTHS

SPACE RESOLVED THOMSON SCATTERING

The spatially resolved ~omson scattering ex-

periment successfully measured electron temper- 7 but the signal-tO-nOise ratio ‘as

poor due to the small number (600) of spatial

locations read on the SIT tube detector.

device has been taken off the Scylla IV-P experi-

ment to effect changes both in method and hardware

to increase the signal-to-noise ratio.

Cos 62

Cos 61 (m)

0.0

9.66

6500

354.6

19.32

352.1

349.5

344.4

346.8

x (v)

6700 - 9.69

6900 -19.43

x (mm)

atures in 1976,

acceptable level.

irregularities across

3D graphics routines on

used.

the

Presently, ways

the

16.2

tube

‘lhe

17.6

18.8

13.58

366.4

14.87

364.1

363.3

360.2

361.7

below

20.6°C

PDP-10,

storage.

current.

b (mm)

a (mm)

light intensity.

scans after an exposure.

100 and unnoticed.)

the PDP-10 computer.

COMPUTER CODES

tube, but

necessity.

controlled

storage.

stepping

general

amounts

points.

other

large

than

also

Some

The

the experimental results.

average over 16 points at each space-wavelength

position in order to reduce the noise level by a

factor of four. To accommodate this increase, the

Prime 100 computer memory was also Increased from

16k to 32k, and the time to read the data

increased to - 0.5 s. During this amount of time,

significant dark-current begins to build up on the

detector, thus limiting ita dynamic range. It was

found that certain sections of the detector had

severe dark-current problems

sections, so initially a scan pattern that reads

the more sensitive zones of the tube first was

‘Ibis tended to equalize the dark-current

produced anomalous readings at the zone boun-

daries. The problem was discovered by using the

the

had

dark

gone

effect

reduces

lhe tube

are being

sought to

This calibration

computer

storage

system

codes

data

been

for

has

and

transmission required by this experiment make the

use of a large computer facility like the PDP-10 a

With a minimum of —30 000 words of

data gathered for each shot analyzed, high-speed

data transmission will be required. The communi-

cations link between the Prime 100 and the PDP-10

has been upgraded to an eight-bit serial line,

which can eventually be operated at 9600 baud.

The 3D graphics capability on the PDP-10 will be

useful for the proper display and understanding of

It was decided to increase the number of

developed to automatically determine the mapping

points read from 600 to approximately 10 000 and

of wavelengths and positions.

data will occupy an additional 20 000 words of

(Before the data transfer link between the Prime

problems that are common elements of various diag-

purpose

nostics have been developed.

Ray Trace Code. Starting from Fermat’s

Impurity Radiation.

principle, 6 ~ nds = O, the eikonal equation for

are available, both having at their core the time

geometrical optics ray propagation can be derived

dependent solution of ionization and recombination

nate vector.

&(nfi)-gradn

contributions over

wavelength.

versions

fringe

step

the

Appleton

Laboratory

general analytic

As is well a 12 dependence ‘n

For the latter pur-

a factor of two.

trajectory and

temperature,

and peaking

rectilinear

available.

aurn the

assuming

actual

phase

code,

t Ime

ionizing

rates for the ground state populations of various

ionization stagea of impurities.

The first code is suitable for radiation loss

calculations and general predictions of temporal

temperature behavior as a function of impurity

concentration with ohmic heating and ion-electron

equilibration terms included, starting from speci-

fied initial values. Only oxygen is included as a

prototype for all low-Z impurities. No detailed

spectral information is derived.

plication is the prediction of Z’I’-4O heating

previously reported already.

recombination rate coefficients are taken from the

Tab1es

H. P. Summers, and the electron excitation ratea

of averaged value resonance lines are taken as the

expression of Seaton, 9 increased

by a factor of 5 to bring it more into line both

with recent detailed calculations on individual

lines and fragmentary experimental results.

The second code is motivated as a temperature

diagnostic. It iterates on solutions of the rate

equationa with successive guesses of the temper-

ature history, given the density history, until a

best fit ia obtained to matching the appearance

spectral lines for which experimental data are

This code uses analytical approxima-

tions of the various rate coefficients. The most

serious error is an overestimation of the dielec-

tronic recombination in a formulation given by

Landini and Monsignori-Foasi

not valid for ne > 1012 cm-3. However in a rising

situation for which the code is designed) the ion-

ization coefficient is of dominant importance.

For the latter the “exchange classical impact

parameter” formulation due to Burgess11 is used,

which is generally regarded as the most accurate

and agrees with extensive experimental data within

and

Graphs

A typical ap-

TWO related codes

8 The ionization and

10 that is generally

the low-Z impurity

specific

plasma

(the

all

where n is the spatially varying refractive index,

is the differential displacement along the

propagation direction,and r is the spatial coordi-

Several versions of one basic numsrical code

have been developed from this equation, given an

analytic cylindrically symmetric density distribu-

tion, for ray propagation with initial conditions

at specified displacements either parallel or

transverse to the cylindrical axis.

known the refractivity has

wavelength, and the motivation was to obtain a

generally applicable formulation for any of the

considerable number of laser wavelengths available

from the visible to the far infrared, a range

spanning more than three orders of magnitude in

The focus of experimental interest may be on

the angular or linear displacements of the rays

themselves (Schlieren and related methods) or on

the errors introduced by these effects on the

interferometric fringe measurements that assume

straight-line propagation.

pose, special double precision, small-integration-

the

count

compare these with relevant quantities to give

both total fringe count and the error that is made

propagation. To obtain these relevant quantities

the assumption is made that a lens positioned to

image the plasma center on the detection plane

equalizes the optical lengths of all straight-line

paths emanating in a cone from the object point.

The error count can have contributions both from

the geometrical differences between the actual

curved path and segmented rectilinear path and the

fact that any deflection takes the ray through

regions where the actual refractivity is less than

along an extension of the initial path.

survey

ment.

plaama.

reported

Abstract.

including a

fully were report12 Shortly

upper bound of nef S 24.

significant SBS gain

mismatching

nef ~ 14.

exceeded.

Measurement

Phase

The

near

possible

LASER-PLASMA INTERACTION STUDIES

on-axia density minfmum was observed at early

The experimental data obtained during the in-

times during the laser pulse. However, refractive

teraction of a 150-J C02 laser with the cold dense

containment was lost at late times due to the dif-

mode of the one-meter-long Scylla I-C theta pinch

fusive loss of the density minimum.

the previous Annual after termination of the experi-

“bleached” absorption to account for the observed

expected

During this year an extensive analysis,

A plasma absorptivity of —46% was

relevant

inferred from calorimetry measurements at

completed and resulted in the submission of a

These results confirm that

doctoral thesis (William T. Armstrong) to the Uni-

classical heating end refraction dominated the

versity of Colorado. The abstract and concluding

summery of the theais are given below verbatim.

For details reference is made to the thesis.

Conclusions. Stimulated B~illouin Scatter:

scatter (SBS) fa studied on an experiment in which

terized by a single pulse near the time of peak

Prompt stimulated Brillouin

SBS was observed with a signal charac-

a high-power, pulsed C02 laser irradiates an inde-

pendently produced, theta-pinch plaama.

The single pulse occurred in a time

not significantly affect laser heating of the

window of reduced axial temperature gradient.

c. SBS displayed the expected red-shift

The SBS signal consisted of a single, strong

in frequency corresponding to the ion-acoustic

pulse occurring within ~ 15 ns of the incident

beam’s peak power. The signal displayed some in-

theoretical spectral fit of

termittence and shot-to-ehot power fluctuation.

backscattered line shape agrees well with the

the

attributed to a strong axial temperature gradient.

The spectral fit of the data and en

A brief period of reduced temperature gradient,

absolute power estimate gave the SBS exponential

and corresponding enhanced SBS coupling, occurs

gain value, ne (number of e-foldings), to be 14-24

peak incident power.

above the initial noise level.

to this time period is

experimentally determined

consistent with the single-pulse characteristic

SBS

and power fluctuations of the SBS signal.

with the high-power

The SBS signal was red shifted by - 51 A

which corresponds to the ion-acouatic frequency.

‘Ihe constraints on SBS gain due to

A theoretical fit of the SBS line shape gave a

gradients can

lower bound for the number of exponentiations of

thresholds and intermittent behavior of the SBS

the backscattered signal from the noise level,

SBS power measurements established an

‘lhe SBS power waa limited by a long

With a gain length of

gain length, low plasma noise level, and short

0.9 cm, these limits on nef correspond to an in-

teraction length of 13 cm S gint ~ 22 cm.

i. More detailed temperature measurements

interaction length is consistent with the high-

are needed to quantitatively study SBS dependence

power focus region where the SBS threshold is

on the temperature gradient.

the

SBS

d. A

power.

SBS does

heating.

modeling

coupling

frequency.

theory, was

mtorr fill pressure.

laser-plasma interaction.

spectral dependence in

resolution is needed

measured temperature.

laser focus region.

The constraint of

corresponds

temperature

consistent

effects

Faraday

region.

signal.

This

‘his

thg

heating

required

Classical

account for the high

density, focus

‘ef length

theoretical

interaction

rotation on SBS are discussed.

of density profiles and temper-

ature histories permitted examination of laser

refraction, local heating, and net absorption.

Refractive containment of the C02 laser beam by an

Single-shot spectral data with greater

to study the

greater detail.

‘lhe most important feature of these results

These results confirm that claasical heating

is the long SBS gain length.

and refraction dominated the laser plaama interac-

limited by the long gain length and short region

tion, and SBS caused only a minor change in

in which threshold waa exceeded.

was apparently limited by the short time during

which reduced axial temperature gradients allowed

FUTURE DIAGNOSTIC DEVELOPMENTS

gain over a significant length.

There are currently three desirable diagnos-

tic developments in an advanced conceptual or

late times.

e. Further

laser pulse.

burn patterns.

a. Refractive

Laser Beam Propagation:

initial laser spike.

surements.

and

(B$

far

field

needed

lost at

the co~

transverse

The SBS energy

The SBS power was

transverse chords.

early design phaae.

heating efficiency.

refractive bending.

just one or two chords.

port

and,

view

al1

and

laser beam was observed at early times during the

The possibility exista to measure the poloidal

containment of

Far Infrared Faraday Rotation for ZT-40.

field in ZT-40 with a Faraday rotation measurement

Channeling of the laser beam in the

infrared laser beam

programmed plasma density minimum was inferred

For the 118-II wavelength of

from exit burn patterns.

CH30H, a central density of 5 x 1015 cm-3, and a

c. Refractive containment was

maximum poloidal field of 6 kG one obtains (a) up

d. Laser heating did not prevent diffu-

in the longitudinal component of the poloidal

sive loss of the density minimum.

field, (b) completely negligible rotation in the

study

and

substantiate filamentation structure In the exit

component), and (c) an only modest displacement

to about one radian linear polarization rotation

of the laser beam at the exit window

As in any Faraday rotation

Propagation through a long plasma was diffi-

measurement the density must be known independ-

Cult. Improved containment of the laser beam re-

ently in order to determine the field, and for a

quires a density minimum sustained through further

proper measurement it is necessary to measure

magnetic-field programming.

along several chords and use an Abel inversi.on-

Laser Beam Absorption and Plasma Heating:

In principle, however, the density can be gotten

a. Local absorption efficiency was high

from the same laser by interferometric techniques

consistent with classical theory for the

the

portion of the laser beam contained in the plasma.

information could be derived from madel fitting to

type unfolding to deduce a radial distribution.

Net absorption efficiency was low due

to a lack of refractive containment of the laser

Ideally one would use nonpolarizing, high-

beam at late times during the laser pulse.

transmission windows on the machine diagnostic

c. Temperature measurements at two axial

polarization analysia with

stations showed substantial heating due to the

independent external elements. However, given the

desirability of using only sapphire windows on

Classical modeling of the expected

ZT-40, the extreme birefringence and differential

heating is consistent with the temperature mea-

absorption of sapphire to the two components of

polarization can be used to advantage as having

e. Bleached absorption, i.e.,an absorp-

polarizer and analyzer fixed in place provided the

tion coefficient incorporating the dynamically

correct crystal orientation exists in the windows.

changing temperature, must be included in the

Calculations show that windows “creased” in the

analysis to account for the observed heating.

absence of plasma will increase the transmission

the

due to

on off-center

orthogonal Be

rotation,

large

some

UP to a factor of 5 for 90° pIasma rotation. Commercially cryogenic

available

lasers

and

detectors can measure the “crossed” transmission

extent

spatial

(- 1-2 ns).

techniques.

(e.g.,ZT-40),

then be differentiated.

to local density.

corresponds to

diam, D - 20

parameters:

anticipated

cm-3,

space

T: i

cm.

should be approximately 0.2°. Experimental tests

2 us results in a background (free-free continuum)

with a algnal-to-n.oiseratio of at least 100, so

that the measurement sensitivity of rotation angle

of these predictions, for windows actually sealed

into ceramic, will be made in the near future.

comes

Integrated over 800 V and

The scattered light from a 3 cm

the

Short-Pulse Laser Thomson Scattering. An

1.2 x 10-10 J (for 0.1 steradfan solid angle).

absolute density profile of a plasma, not requir-

Thus the peak plaama signal would be about 10

ing the inaccuracies of an inversion procedure,

times greater than the background.

can be achieved, in principle, with laaer scatter-

important is the signal-to-noise ratio, since it

ing using only a single receiver, provided the

light pulse is short enough so that its limited

fluctuation in background

provides

4.2 x 107 photoelectrons of signal and 4.2 x 106

resolution. An additional advantage of this tech-

nique is that much of the instrumental scatter at

ratio based on Gauasian statistics is better than

the entry and exit points to the vacuum chamber

This is certainly optimistic since

can be discriminated against by time-of-flight

plasmas generally exhibit more continuum than pure

interest.

background,

the

For application to plasmas of diam ~ 20 cm

fluctuation nature may well be nonthermal, a point

pulse width

which can only be decided by experiment.

appropriate. The entire plasma is imaged onto the

extent to which the signal-to-noise ratio is more

detector which is gated on for slightly more than

favorable than necessary ia the extent to which

the transit time of the pulse through the plasma

like laser energy can be compromised.

The plasma profile is obtained from

A O.I-J energy is considered readily achievable,

the waveform of scattered intensity va time (time

and even 1-J should be below damage threshold.

in the plasma).

The measurement would be flexible enough to

may be considered, depending on detector risetime,

specific

direct snd integrated.

reducing the optical bandwidth if the temperature

(Tr << scattered light pulse width) the instan-

The measurement would be calibrated by

taneous signal (suitably calibrated) corresponds

Rayleigh scattering from a high-pressure gas fill

In the integrated mode (Tr >>

as well as against Interferometric data along the

scattered light pulse width) tbe waveform is es-

laser line of sight. Output recording would be on

sentially the integral of the profile, and would

either a fast scope, or the Tektronix 7912 scan

2 x

the

and

the

slug

$0.1

that

plasma

needed

is the

tranait

spatial

receiver

parameter

‘R?o modes

photoelectrons

bremsstrahlung

of 1.2 x 10-11 J.

0.2 by 20 by 20 cm3.

buried in the noise.

In the direct mode

With a 2-mm beam

time - dependent

The spectral

We assume a

converter.

is lower.

scattered

optimized

Thomson

signala

density,

n - 5 x

mode.

will

with

for

The

the

slab

from

With

would

center

during

roughly

detailed

Even more

background

parameters,

signal-to-noise

scattering.

result

e.g.,

from

methods

such

not

As a specific example we use ZT-40,

Laaer Scattering with Sequential Mode-

electron

Locked Pulses. The prospect of having long-pulsed

electron temperature, Te - 200 ev,

(500-us) ruby lasers haa stimulated interest in

consider a ruby laser

emitting one J in a single switched-out

locked pulse of O.1-ns duration.

photoemisafve detector (photomultiplier or image

the densities contemplated for ZT-40 (5 x 1015 Cm-3)

it is expected that the difficulty in observing

dissector) with a quantum efficiency of 10% at 0.7

inadequate number of scattered photons, but rather

Pm and an integration time of 2 ns.

from an excess of light coming from the plasma.

bandpass of the system is chosen to collect at

This then leads to the concept of phase-locked or

least 95% of the scattered spectrum, which re-

signal-averaging techniques to extract the signal

quires about ~ 400 V under the assumed conditions

We have recently become

(Te = 200 eV, 90° scattering).

aware of successful experiments using

size in the plasma, the receiver field of view is

but operating on a vastly different time scale.L3

2 mm by 20 cm. Thus the background radiation seen

The technique as envisioned at present ln-

volvea taking a long-pulsed ruby laser and intro-

ducing an active mode-locking element within the

The anticipated output would be a

series of mode-locked pulses lasting for the full

duration of the laser pulse, ~ 500 lh.

averaging would be done over as many pulses as

obtain

necessary

perature.

laser cavity.

chain ruby laaer.

sufficient

periodic

locking.

us.

adequate

Signal

REFEFJNCES

signal-to-noise

Shorter times are

The output of the

produce

Such a

mode

lhis

(e.g., for 5 Ps) for repeated discrete sets of

pulses over the full duration of the laser pulse

train in order to give time-resolved electron tem-

If 100 J are sustained over 500 us in mode-

locked operation, this average power of 200 kW

will be sufficient for good time resolution. The

shortest usable pulse will probably be about 500

ps. ‘lhislimit is determined by the combined rise

and fall time of the photomultipliers used to

detect the scattered signals.

certainly available from mode-locked lasers and

faster detectors are commercially available, but

photomultipliers with the required gain are limlt-

ed to about this time duration. The maximum time

separation between pulses will be bounded by the

requirement to maintain a stable cavity configura-

tion. To limit the cavity dimensions to something

reasonable, this would suggest 20-30 ns between

pulses or a duty cycle of 1/40 to 1/60, implying

individual pulses of 8-12 MW.

detector would be sent to a superfast boxcar inte-

grator with an averaging time constant of several

The effective reduction of noise or equiva-

lent effective increase in power will be propor-

tional to the square root of the number of pulses

averaged. For 1 us this would be a factor of 5 to

7, or an effective laser power - 30 MN.

effective laser power ia in the range that yields

good results in current conventional scattering

experiments on ZT-S. Clearly there is a trade-off

between time resolution and effective laser power.

Initial efforta are aimed at seeing what can

be produced in terms of a mode-locked long-pulse-

A standing wave acousto-optic

cell has been developed by bonding a lithium

niobate transducer onto a quartz slab.

cell when antireflection coated should introduce

losses

LASL Controlled Thermonuclear Research Pro- gram Report January-December 1976,” Los A1.amosScientific Laboratory report LA-7082-PR (1978), p. 133-137.

F. C. Jahoda, “Submicrosecond Holographic Cine-Interferometry of Transmission Objects,‘t Appl. Phys. Lett. 14, 341 (1969).

V. V. Antsiferov, K. G. Folin, V. S. Pivtaov, and V. D. Ugozhaev, “Ruby Laser with a Spherical Resonator with Smoothing of Spatial Inhomogeneities,” Sov. Phy. Tech. Phy. Vol. 16, 2060 (1972).

F. R. Marshall and D. L. Roberts, “Use of Electro-Optical Shutters to Stabilize Ruby Laser Operation,!!pro=. IRE Vol.

R. E. Siemon, “Polychrometer with Extreme Rejection of Stray Light,” Appl. Opt., 13, 697 (1974).

—

M. Greenwald and W.I.B. Smith, “Triple Grating Polychromator for Thomson Scattering,” Appl. Opt. &

“LASL Controlled Thermonuclear Research Program Report January-December 1976,” L.Js Alamos Scientific Laboratory report LA-7082-PR (1978) p. 132.

“LASL Controlled Thermonuclear Research Program Report January-December 1976,” Loa Alamos Scientific Laboratory report LA-7082-PR (1978), p. 64.

R.W.P. McWhirter, in Plasma Diagnostics Techniques, R. H. Huddlestone and S. L. Leonard, Eds., (Academic Press, New York, 1965), Chap. 6, p. 269.

M. Landini and B. C. Monsignori-Fosai, “A Simple Formula for the Total Dielectronic Recombination Coefficient,” Solar Physics 20, 322 (1971).

H. P. Summers, “Ionization Equilibrium of Hydrogen-Like to Argon-Like Ions of Ele- ments,” Mon. Not. Rny. Aatron. Sot., 169, 663 (1974); A. Burgess, “Semi-Classical ~=ry of Electron-Atom Collisions,” AERE-R-4818, Sept. 14-16, 1964 (Meeting at Culham Lab), pp. 63-71, Proc. Symp. Atomic Collision Processes @Plasmas, 63.

MMSL Controlled Thermonuclear Research Program Report January-December 1976,” ~s Alamos Scientific Laboratory report LA-7082-PR (1978) p. 132.

H. F. Dobele, K. Hirsch, and M.von Hellermann, tttiscattering Diagnostics with Periodically Pulsed Lasers to Follow the Continuous Evolution of Time Dependent Plasma Parameters,” Preprint IPF 77-5, University of Stuttgart, April 1977.

(1962)

587 (1977).

Culham Report 4818, p.

D. C. Barnes, J. U. Brackbill, T. Cayton, R. Y. Dagazian, J. P. FreLdberg,

S. P. Gary, R. A. Gerwin, D. W. Hewett, D. S. Lemons, H. R. Lewis, P. C. LLewer (Univ. of Maryland), R. C. Malone, D. Montgomery (Univ. of Iowa and College of William & Mary), C. W. Nielson, W. B. Riesenfeld, J. J. Sanderson (Univ. of St. Andrews), J. L. Schwarzmeier (Univ. of Wisconsin), A. G. Sgro, B. Abraham Shrauner (Washington Univ.), K. R. Symon (Univ. of Wisconsin), L. Turner, G. Vahala (College of William & Mary), D.Wfnske

The work of the theory group is closely

first-order

associated with and motivated by the experimental

helical fields, each one characterized by t13+ hz.

The leading-order field is the 9-pinch field

For several years the

There can be an arbitrary number of helical fields

requirements

the major

with different t valuea, but they must all have

research in the theory group.

transverse fields are allowed in first order. The

the Scyllac program haa ended and the CTR effort

helical fields are determined from a potential

concepts, the variety of theoretical research has

second-order partial differential equation

directed toward alternate

satisfies

This variety is reflected in

following description of theoretical work carried

out in 1977 and it is expected to become greater

and

problems

increased.

INTRODUCTION

in the future.

experiment provided

at LASL has become

CTR program at LASL.

e-pinch field B(r)~z.

is very nontrivial.

MHD STUDIES

follows.

shifts.

exist.

turn

out

the

‘he

the

for

same

just

vf3.v+

which

pitch

vh+m.o

number

Scyllac

function

described.

motivation

XIII. THEORY

However, now that

lvo12=g(f3) - -+l+z

for small but finite B.

coefficients.

second-order

satisfied.

azimuthal

toroidal

fields,

where

with

the

The reason for this is as

Baaed on the sharp-boundary calcula-

tions, there is a strong suggestion that in a

diffuse system one should expand about a basic

Expansions of this type

incompatible

periodicity constraints and no such equilibria

The expansion described herel overcomes this

difficulty by considering the basic O-pinch field

to be of the form B(r,O)~z; that is, we must

consider the leading order system to consist of

noncircular flux surfaces with finite toroidal

Under this assumption diffuse high-B

stellarator equilibria can be found.

(1)

the

are

Also,

purely

fields

following

To calculate the

periodicity

imposes a

appearing

condition

system

second

and

(2)

Analytic and Numerical Studies of High-!3

Toroidal effects are assumed to

Stel-larators. An analytic formulation has been

enter first in second order.

derived for calculating equilibria in a diffuse

high-!3 stellarator configuration. Although ana-

(resulting from z independent terms) must be

lytic sharp-boundary equilibria have been known

‘This constraint

for some time, the extension to diffuse profiles

constraint between $ and B,

B(r,e) = 1 - B2(r,0)/B~

where g is an arbitrary function of B. Thus, the

problem of finding high-!3 atellarator equilibria

can be cast as two nonlinear, coupled partial

differential equations for the unknowns $ and 6.

These equations have been solved analytically

The result ia a diffuse

analog of the well-known sharp-boundary equili-

brium relation with no toroidal shift. For high-f3

equilibria, a numerical code has been written.

this approach the solution is Fourier expanded in

quasi-linear ordinary differential equations is

variable

is proceeding.

dipole current.

experiments on

stellarators,

equilibrium.

numerically

stability

tions,

brium

are:

and

demonstrate

obtained. It is possible to obtain solutions with

same pitch number h; and wall stabilized systems.

no toroidal shift when the appropriate helical

It has been conjectured that the addition of an

field product is applied.

L?=l,h configuration

results with earlier numerical and analytic work

instabilities.

straight systems with E=l,h and P.=3,3hhas been

Finally, a comparison has been made with the

studied, and, at least for small helical field

low-s stellarator. In both cases, helical fields

amplitudes, no stabilization by the t=3,3h fields

are required to provide closed flux surfaces.

However, in a high-~ stellarator the toroidal

desirable

drift force is balanced by the interaction of two

dependence of the equilibrium fields on the plasma

helical fields: 1, 1+1. In a 10w-8 stellarator a

One method by which this may be done is

single helical field suffices. Here, the toroidal

suggested by the much stronger dependence on B of

drift force is balanced, not by

the stellarator force produced by the interaction

fields, but by the interaction of a small applied

of 1=0 and 2=1 fields than by the interaction of

vertical field with the small induced toroidal

1=1 and 9.=2fields. By opposing these forces, the

6 dependence can be cancelled without canceling

Numerical solutions of the diffuse, high-f3

the stellarator force. A possible problem arises

stellarator equilibrium problem have been found by

because the 9.=0and 1=2 fields Interact to produce

solving finite-difference approximations to the

t=2,h=0 dependence.

nonlinear, time-dependent

studies of this interaction have been performed,

ideal, three dimensions.2 Such solutions have provided answers

magnetohydrodynamic

flow

deformation of the plasma is not bounded to small

the

to practical design questions associated with

amplitudes, but must be suppressed by an as yet

existing experiments; these design questions arise

undetermined =ternal intervention.

because, in contrast to tokamaks or even low-8

stabilization of

equilibria

unstable equilibria haa been

numerically correct helical field amplitudes. The

plasma to wall ratio of 1:3.

solutions also have been used to study the equili-

tions, a lower bound can be placed on the insta-

properties

bility growth times which are much longer than the

configurations. Equilibria have been computed for

existing experimental configurations with various

In summary we now have analytic and numerical

plasma profiles, helical field combinations, and

tools for studying high-~ stellarator equilibria

toroidal radii over a wide range of plasma 8.

and can systematically examine new configurations

From these results, the scaling of the helical

with more desirable equilibriumand stability pro-

field amplitude with the principal parameters has

been derived and summarized.

Simulation of Magnetohydrodynamic Insta-

been made between these scaling laws and recent

bilities. The 3-D MHD code used for studying

Scyllac where

high-f? stellarators also has been used to study

fields were made adjustable.

equilibria and stability of a straight, high-beta,

confirm the accuracy of the numerical calcula-

P = O pinch configuration, as well as internal

and

that

beta.

would

field

!=3,3h

indicate

equations

describing

force with

the helical

suppress MHD

has been observed.

Comparison of these

toroidal drift time.

These comparisons

Comparisons have

the equilibrium

Finally, wall

improvement

theories,

examined.

perties.

systems

exist

other

with

only

the

for

the

The

can

reduce

resulting

Numerical

elliptical

stability of

From the computa-

computed with a

In addition

numerical

otherwise

the

Other configurations which have been studied

aolutiona waa studied and the dependence of the

straight

solutions on varioua numerical parameters waa

combinations of helical fielda characterized by

t=l,h and 9.=3,3h;straight and toroidal systems

with applied fields g=O, t=l, and !.=2with the

kink modes of a straight screw pinch.

convergence

confinement resulting from the improvement of the

to comparing the numerical results with analytic

3. Rigid-Drift Magnetohydrodynamic Equilib-

cylindrical region with a

ria for Cylindrical Pinches. The rigid-drift equa-

conducting boundary.6 Canonical distributions have

tions of magnetohydrodynamic equilibria in cylin-

constructed from three rugged invariant

;

pan

When

r=m.

r-l,

Jz=n,

r = 1

Jo=rn;

field, current

drical geometry are:

density, respectively.

completely determ.tned by

parameters are specified:

current density.

reversed

magnetic

pressure

r=2,

and

the

fields

where ;(r), ;(r), p(r), and n(r) are the magnetic

density, pressure,

Such magnetohydrodynamic

equilibria have corresponding Vlasov equilibria.

(the isothermal

equations produce the Bennett profiles for the

case of a pure Z-pinch.3 The case of a pure 13

pinch with r = 1 has been solved by Morse and

Freidberg.4 In both Z- and e-pinch cases with

the pressure profiles drop to zero only at

For sn arbitrarily pitched current density

equations

analytically in terms of an infinite series of

hypergeometric functions.5 The

implies a local temperature that is proportional

to the local number density.

field

the

the local plasma beta on axis;

b. a quantftY related to the pitch of the

Particular profiles, plotted with the aid of the

symbolic manipulation system, MACSYMA, have shown

the possibility of occurrence of hollow profilea

independently. The pressure always falls to zero

at a finite value of the radius.

This work may have relevance to understanding

and manipulating ZT-40 equilibria.

Three-Dimensional Magnetohydrodynamic

Turbulence and Pinch Stability.

magnetohydrodynamic turbulence has been treated in

can

(3)

and

Al1

were

been

only

exist

below

fields

number

ratios.)

extremal.

expansion

case), these

two-dimensional

involve field reversal.

End-Loss Studies.

The shapes of the

J. B. Taylor.g

condition r = 2

simultaneously

Nondissipative

dissipated.

mysterious

Z-pinches

once two

profiles

whereas

solved

which

model

have

been

are

(magnetic helicity, cross heliclty, energy) and

two constraints have been Imposed to simulate

experimental conditions (fixed axial magnetic flux

and an integral related to the poloidal flux).

then

Chandrasekhar-Kendall eigenfunctions of the curl

operator.7 All quiescent initial conditions led tO

finite mean-square expectation values for

coefficients of

except those for which the magnetic helicity is

These are force-free states which may

(Such symmetric states

certain

Dual cascade situations sometimes can tend

toward these extremal states.

magnetohydrodynamlcs,

energy can be cascaded to higher wave numbers and

be dissipated there by resistivity and viscosity,

while the inversely cascaded quantity (mean square

vector potential) is not so readily dissipated.

In three dimensions, energy and cross helicity can

be cascaded to higher wave numbers and dissipated,

the magnetic helicity ia less readily

This basic process (first suggested

Bretherton and Hsidvogel for Navier-Stokes

fluids8) provides a mechanism for the relaxation

process to occur in three-dimensional (toroidal)

energy to helicity is reached, as conjectured by

the

studies of plasma end loss have been obtained hy

solving finite-difference approximations to the

time-dependent equations for ideal, magnetohydro-

dynamic flow in two dimensions.

found to compare well with new experimental data

on mass lifetime in straight pinches, and with old

data on the effect of simple mirrors on end loss

from the linear Scyllac experiment.

difference

two-dimensional results and those obtained with a

one-dimensional formulation which assumes radial

pressure balance has been examined.

the magnitude of those terms neglected in the

the

field

excess

velocity

expanded

current-to-flux

For example, in

rigid, perfectly

The results are

By comparing

numerical

Further

between

minimal

state

the

1. Gradient-driven microinstabilities.

electromagnetic heat-flux modes have considerably

out

left

field

magnetic

strength.

completely

approximation.

KINSTIC EFFECTS

multiple mirror.

momentum equation.

mirror confinement.

perpendicular shock.

this research were:

gradients, but

instability

secondll

fully

the

The

one-dimens%onsl approximation with thefr actual

Figure XIII-1 compares the growth

value in the two-dimensional aolutiona, the source

rate of these two modes aa a function of the

of the difference has been identified to be the

The

instability at

comparable with the axial pressure gradient, yet

has no threshold in the electrostatic and local

This term is found to be

evident; it is also clear that the universal mode

curvature

approximations and that further work on this m~e

is necessary to determine the conditions for its

The numerical method haa been extended to

presence in a peat implosion theta pinch.

incorporate the guiding center fluid equations,

instabilities driven by

and will be applied to the study of multiple

heat flux parallel to a magnetic field were alao

The firat step will be to

the subject of a comparative study. The threshold

recover both the guiding center and the magneto-

of the ion acoustic heat-flux instability, the

hydrodynamic mass lifetime scaling with mirror

mode considered moat likely to lead to a heat-flux

Next, energy anisotropy instabilities

limit in Te>>Ti laser fusion plaamaa, was studied

will be modeled by a simple relaxation process to

In the electrostatic (B = O) limit,

examine how mirror strength, mirror separation,

other instabilities (ion cyclotron, electron beam)

and relaxation time affect plasma end loss in a

are the first to go unstable at Te ~ Ti, but these

modes have relatively high thresholds and do not

seem likely to lead to an effective heat-flux

limitation.13 However, at nonzero 13and Te - Ti,

The

Two

the

term

family of

the axial

in detail.

fnatability.

one-dimensional

cyclotron drift

diamagnetic drift.

The first10 included

threshold is

propagation

restricted

arbitrary

density

0.04-

0.02-

0.08-

Q06 -

~ q

was

the

zero beta

threshold of the ion-

ion - cyclotron

1 /-l

I 0.50

‘d y

1.00

Figure XIII-1. The growth rate maximized over wavenumber, a , of the ~rift instabilities as a function density gradient drift apeed Vd for Te = Ti. Here Qi = ion-cyclotron frequency and Vi = ion thermal speed.

universal

and

comparative studies of the family of instabilities

driven by gradients perpendicular to a magnetic

field were carried out.

electromagnetic effects and

oblique to B for the high-frequency (IA)>>S2Ci)modes

associated with the relatively steep gradient of a

The important results of

(a) At Te>>Ti, the ion acoustic instability

is dominant, and its linear dispersion properties

are unaffected by nonzero Be and

(b) At Te - Ti and with density and magnetic

field gradients pointing in the same direction,

lowest

study

modified two-stream instability (with kz+ O rather

than the lower hybrid drift instability (k= - O).

electrostatic instabilities driven by

considered modes

frequency and generalized the work of Freidberg

and Gerwin12 to oblique propagation. The primary

new results here were (a) a generalization of the

Freidberg-Gerwin threshold criterion for the ion-

cyclotron drift instability to arbitrary Te/Ti

ratio, and (b) detailed studies of the threshold

and dispersion properties of the universal drift

the

mode

work

driven

Future

energy

maximal

could be

achieving

addition,

increased.

reduced and

confinement time.

distributions.

distribution

constants.

dispersion

relation

that the growth rate of

This

and

lower threeholde than the ion acouetic insta- bflitY;14 therefore future studiee till give ‘ore

expressions for the energy in the wave magnetic

field, the temperatures and the entropy of the

attention to electromagnetic modes.

system in terme of one unknown constant, T1/Tll,aa

simple

the syetem evolves from TL/Tl = TLo/T1o to TL/Tn =

one-fluid code has been written to model electron

1.0. One important consequence of the theory is a

temperature in a linear theta pinch and to study

simple expression for the maximum field energy.

the possible effecte of an anomaloue heat flux

verified

limit.15 The results indicate that near the end of

simulation of the Weibel instability. Extensions

a linear theta pinch the classical formula for the

of the theory to magnetized plasmas appear to have

heat flux breaks down, due to the temperature

These include: computing the

scale length becoming shorter than the mean free

absorption

path. If the flux is limited, in this region the

acoustically heated plasmas, understanding the

temperature gradient over most of the profile is

saturation level of the Alfven Ion Cyclotron in-

confinement

will

increase

excited

injection of neutral beame in mirror machines),

and determining the reduction of heat loss in

Current results indicate that

linear devices and the solar wind by high-beta

an increase of at least an order of magnitude with

electromagnetic heat-flux instabilities.

respect to classical is possible.

Linearized Analysis of Imhomogeneous Col-

A third type of instability studied was an electromagnetic mode driven by an ion beam. 16 ~is

lisionless Plasma Equilibria: General T%eory and

BGK P.quilibria. A generalized framework has been

developed for analyzing the linearized equetione

particles escaping in an axial direction from a

for perturbation of inhornogeneousplasma equili-

linear reactor. The most important reeult here is

bria in which there is a collisionles& species,

some properties of the solutions of the linearized

proportional to the 1/3 power of the beam density

equations have been described, and a basis has

in contrast to the 1/2 power for many thermo-

nuclear Instabilities driven by isotropic alpha

been given for numerical computations of the linearized properties of such equilibria.17 The

general approach has been applied successfully to

A Nonlinear Theory for Electromagnetic

some numerical studies of the stability of screw

Instabilities. In order to determine the anomalous

pinches within the framework of the Vlasov fluid

transport associated with a

mode118 and of the stability of Bernstein-Greene-

bility, an understanding of its nonlinear behavior

19 When studying such inhomoge-

is essential. A theory to describe the nonlinear

neous equilibria, it is useful to expand the

state of transverse electromagnetic instabilities

perturbation potentials in eigenfunctions of the

driven by anisotropic temperature distributions

field operator which appears in the linearized

has recently been developed.

equations and to define a dispersion matrix whose

theory is an assumed distribution function based

analytical properties determine the nature of the

on two single particle integrals of motion for a

solutions of the initial-value problem.

charged particle in an electromagnetic wave field.

also useful to introduce auxiliary functions to

function

perturbation

instantaneous state of the wave-plasma syetem as

functions, and to expand the auxiliary functions

it evolves in time and determines a nonlinear

in eigenfunctionss of the equilibrium Liouville

In the case of a zero frequency wave

functions, great freedom is achieved in the choice

due to the Weibel instability driven by an initial

operator, which

electron temperature anisotropy, Tlo/T io > these relations are particularly simple and yield

linearized equations. This freedom can be used in

some problems to define expaneion functions for

introducing

The

has

been

time

theory

thermal

stability

(which is

concentrate

collisionless

one-dimensional,

wide application.

thermonuclear alpha

this instability is

Kruskal equilibria

given microinsta-

The basis of the

eystern energy

operatore.

describes

replace

usual

field

the

two

the

rate

computer

magneto-

perpendicular

distribution

auxiliary

appears

It is

the

effective

truncate

the

The

some

theory.

general

“wobble”

possible

stability

Greene-Kruskal

independently by

these equilibria.

satisfactory

experiment

present

e pinch

solved

are

the

the potentials that are particularly suitable for

approximately

studying specific normal modes.

shorting. To get this agreement, we must assume

theory has

that the plasma has a smsll positive charge which

recently by using it as the basis for studying the

gives rise to an outward-pointing radial electric

large-amplitude

field (as arises from the ambipolar potential in a

equilibria.19

properties of the BGK equilibria were determined

Formulation of the Linearized Vlaaov

numerical

Fluid Model for a Sharp-Boundary Screw Pinch.

studies for comparison with results of the general

theoretical formulation has

choosing an auxiliary function which yields a

field operator particularly suitable for studying

analyzing linearized equations appropriate to a

straight, cylindrical, sharp-boundary screw pinch within the framework of the Vlasov fluid model.20

One of the eigenfunctions of

This formulation haa been applied to some computa-

the field operator is very similar to the unstable

tions relevant to finfte-ion-gyroradius stabi-

eigenmode of the equilibrium. This fact makes it

lization of MHD modes,18 and it is being used to

study magnetoacoustic heating in a colliaionless

severely. Excellent agreement with the numerical

plasma.21 Surrounding the plasma la a cylindrical

simulation results was obtained by using

conducting wall, and there is a nonconducting

truncated dispersion matrix.

the plasma and

Rotational Instabilities in a 0 Pinch. A

perturbation-dependent transfor-

theory has been derived which attempts to explain

mation of the phaae space and linearizing about a

instability

e pinches. The instability corresponds to a gross

which

linearized

m=l,n-O mode (m is the azimuthal mode number and n

Freidberg’s Vlasov fluid mode122 are put into a

the number of radial nodes) which rotatea with a

form which would be correct for a hypothetical

nonzero real frequency in the laboratory frame.

problem in which the plasma boundary is a rigid

Although rotation has long been thought to be the

cylinder. The effects of the impulsive electric

driving mechaniam for the instability, none of the

field at the actual perturbed boundary are taken

existing rotational instability theoriesprovides a

into account in the zeroth-order atate.

explanation.

m=l,n=O mode is not found to be unstable in most

boundary are continuity of the normal component of

theories; in those in which it ia, the growth rate

~ and vanishing of the normal component of the net

and real frequency predicted are far from the ex-

perimentally measured values.

finite Larmor radiua equationa for a rotating

heating

calculation, the high-f3,

D. MAGWETOACOUSTIC HSATING

numerically

method for preferentially heating plasma ions when

appropriate boundary conditions with particular We find attention focused on the m=l,n=O mode.

an appropriate dissipative mechanism exists at a

magnetoacoustic

that this mode is unstable but that its maximum

Experimental evidence for the existence of a

growth rate occurs not at k2=0

dissipative mechanism in a collisionless plaama is

neutrally stable) but at some finite k2>0.

Grossmann,

threshold for this mode ia lower than that of any

Neuhauser of collisionless damping of magneto-

mode in the system and in this sense m=l,n=O is

acoustic waves in a high-beta plasma.23 Such a

the worst mode. However, typical growth rates are

scheme may prove to be practical for

considerably smaller than m=2,n=0 growth rates

reversed-field pinches; and, for a linear pinch,

when both modes are unstable. Good agreement with

heating

obtained

limitation on end-atoppering schemes which have

the

for

the

was

The

the

been

found

atate

linear

matrix

method

observed

stability

dispersion

Bernstein-

introducing

Illustrated

zeroth-order

perturbation,

corresponding

vacuum between

mirror machine).

particle-simulation

observations by

current density.

magnetoacoustic

Magnetoacoustic

(where it is

convenient

conditions

particular

conditions

asauming

boundary

subject

heating

The

the

for

end

been

derived

depends

complete

equations

the wall.

Keufmann, and

frequency.

attractive

perturbed

resonant

impose

plasma

would

not

The

the

contribute in leading order.

judicious choice of

description of the plasma which contains kinetic

parameters, msgnetoacoustic heating may allow the

effects along the magnetic field lines but retains

use of lower values of induced W

only fluid-like effects transverse the the field

than are required when implosion heating is the

finite-ion-gyroradius

The

been

With

physics

relevant

proposed.

oscillations

these studies have

sole method of heating.

of magnetoacoustic waves.

the assumption

treatment of

displacement

Vlaaov fluid

negligible

computed.

model

Both

sea

the

plasma

the

two

this

shift

field;

lines;

Most of

amplitude

dissipation.

electric fields

magnetoacoustic

plasma column.

high-temperature

the sharp boundary.

Vlaaov fluid

physical

to 0.7.

on the

proper

model

order

based

take

the

The

two

the

For

Several aspects of resonant magnetoacoustic

applying this description to a pinch configuration

high-beta,

is the identification of ion Landau damping as a

plasmas have been considered, including energy

balance, generalization of Poynti.ng’sTheorem to

viable dissipative mechanism for converting magne- toacoustic wave energy into ion thermal energy.25

accommodate plasma motion, and various dissipation

The Landau damping causes a phase shift in the

mechanisms for collisionless plaamas; also, the

response of the plasma relative to the phase of an

poaaible use of magnetoacoustic resonances as a

imposed oscillation of the confining magnetic

plasma diagnostic has been considered.

responsible

involved the solution of

At a magnetoacoustic resonance the

linearized MHD or finite-ion-gyroradius equations;

general screw-pinch equilibria have been con-

associated rate of dissipation both peak.

sidered for modes with finite axial wavelength. A

total rate of energy dissipation is obtained by

numerical scheme for integrating the singular

calculating the Poynting vector external to the

ordinary differential equations that occur with

l%at the energy dissipated

diffuse profiles has been devised.

converted completely into ion thermal energy la

proved by use of a thermal transport equation

heating of a screw pinch in a collisionless plasma

derived from the ion Vlasov equation. In order to

regime has been studied theoretically by

obtain the shortest doubling time, T, for the

approaches and the results are being applied to

plasma energy with the least induced RF electric

cases of experimental interest. Heating times and

field, E, it appears desirable to operate at a

associated induced RF electric fields have been

ratio of wave phase velocity along the magnetic

approaches are

field to ion thermal velocity approximately equal

Vlasov fluid mode122 in which collisionlesa ions

example, T -

10-%.

and masaleaa, fluid electrons are treated under

obtained with E Z 80 volts/cm.

of quasi charge neutrality and

Inclusion of finite-ion-gyroradius ef-

current.

fects. In the other approach, the effect of finite

approaches differ with respect to their ordering

ion gyroradius on magnetoacoustic heating of a

of small parameters and with respect to their

sharp-boundary screw pinch is being studied.

the boundary region between

account

plasma and a surrounding vacuum.

boundary, the starting equations are those derived

Landau damping as a dissipation mechanism.

by Lewis and Turner for applying the Vlasov fluid

In one approach, the techniques of earlier analy- ~f Turner24 have been combined with a new

model to a sharp-boundary screw pinch.18 Within

the pinch the collisionless Boltzmann equation for

ordering scheme, one which is adapted to the study

The basic smallness

analytically for small values compared to unity of

parameter is the ratio of ion gyroradius to plasma

k rL and @/uci, where k is the largeat relevant

radius; in terms of this parameter, the following

quantities are asaumed to be of order unity: local

‘avenumber’ ‘L is the ion gyroradius, ~ is the frequency, and Wci is the ion cyclotron frequency.

plasma beta, the ratio of plasma radius to any

Finite-ion-gyroradius effects were kept to

scale length of the wave, and the ratio of wave

and first order and the solution is valid up to

phaae velocity along the magnetic field to the ion

Because this approach uses a

thermal velocity. This ordering scheme applied to

more general ordering, it alao should be possible

providea

to determine whether dissipative mechanisms other

The

the

and

the

for

not

effects

response

A baaic result of

has been

solved

sharp

been

zero

has

the

was

aniams.

Finally,

developed

linear theta pinches.

pinch on the other.

perpendicular

heating.27

constitute

resonant

that

and

than Landau damping are operative for magnetO-

varied

acouatic heating in a collisionless plasma.

end-stoppering efficiency of the arrangement. The

differential equation for ~ (1(1) = X(0)) con-

current in the internal rings flows in a direction

taining finite gyroradiuss effects was derived by

opposite to that of the primary coil and la

substituting the solution of the collisionless

induced by pulsing the external solenoid.

Boltzmann equation into the transverse force-

configuration

balance equation (the component perpendicular to + B of the Maxwell $x; equation).

solution of the equation for ; was used to

possesses a separatrix surface whose section is

shown by the dashed line.

calculate the heating of the pinch in the presence

tending to minus infinity on a circle inside each

of an oscillatory driving electric field on a cy-

The specific plasma volume ~d!t/B (roughly

lindrical surface in the vacuum region surrounding

average

the pinch.26 When there is a finite longitudinal

directions away from the separatrix surface.

wavelength (kZ+O), significant additional heating

integral is taken along a line of force, di being

can occur as a result of finite ion gyroradius.

the length increment along this line.

Computations relevant to specific experimental

configurations are being carried out.

plasma

Relaxation processes aa dissipative fiech-

configuration as it entera the trap.

azimuthally symmetric

low-beta argumenta, but there is no reaaon to

axially uniform magnetoacoustic heating of colli-

believe that they will be unfavorably modified in

sionless theta-pinch type plasmas was studied

Radomtsev has made simple argumenta

early in the year, and a phenomenological model

about the stability of B = 1 plasma in the region

curvature between

turbulent-induced relaxation between the parallel

ion

dissipation

From the collisionlesa single-particle point

of view, it is the isobars of B that play the

Numerical

important role in confinement.

integrating the nonlinear partial differential

cuting adiabatic motion tend to be reflected at

equations of the model have been examined.

the points of intersection of the field lines with

the

all

can

and

how

equal

ring.

micro-

showed

of bad

mirrors

pressure

magnetic

high beta.

distribution

temperatures

An analytical

separatrix, the

potential function29U5

XIII-2,where two ring

Schematic of

mechanism

solenoid.

lxxDaxxl~

/~\ ,

methods

could

such

-”.

The

for

—

-1~

tends

and

the

The

all

B2)

the

peak

which

rings

stable

optimize

settling

generated

These are

The plasma

increases in

On this line, the

~dg/B assumes a minimum

0/’(3’1h!f;lccc ,

Particles exe-

SOLENOID WITH MIRRORS

/ \ \

EPARATRIX

LINE

,0-

INTERNAL RING END STOPPER FOR OPEN-ENDED DEVICES

In open-ended magnetic confinement systems,

it is desirable to reduce end loss of plasma

because it constitutes the severest mechanism-

limiting confinement. A scheme was proposed which

could be employed in end-stoppering such devices

and could be particularly suited to the needs of

The configuration that is proposed here is basically a multiply connected magnetic trap.28,29

It ia illustrated in Fig.

conductors are shown inserted in a solenoid with a

strong mirror at one end and connecting to a theta

Mirror coils can also be

inserted in the regions between the rings.

cross-section area, the spacing, the total number

of rings, and the strength of the solenoid and

0/’;, (.

=------’../

Figure XIII-2.

magnetic field geometry of IRES.

the

end

works

points

Figure

space.

stably.

motion.

isobars,

velocity.

reflection

particular

are employed.

function of CR and n.

of the end mirror(s).

0.4

0.1

0.2

0.8

I.0

0.6

depending

The existence of a circular region

inside each ring where the magnetic field strength

They MHD stabilize the magnetic mirrors

is negligible causes a breakdown of the adiabatic

present in their vicinity; and

Nonadiabaticity in

they introduce nonadiabatic behavior in a

produce scattering of the particles in velocity

small but sizeable region inside the plasma which

This may result in enhanced trappfng of

tends to shorten the spacing required between

particles which are knocked out of the loss cone

A device of the type being

proposed here can be easily ehaped into a cone by

Denoting the fraction of particles trapped

employing rings of diminishing diameter

because of the presence of a single ring to the

XIII-4) providing for easy external mechanical

total number of particles going through the ring

support of the forces exerted by the plasma wind

region in the absence of the ring by CR, the

on the rings and the end coils.

reduction loss to fR = (l+ER)‘n,where n is the number of rings. For simplicity, we have also assumed that only one

internal ring confiner end (CIRCE) could also be

viewed as a magnetically shielded solid end plug

when it is looked at through the end of the theta

final end mirror does the trapping and that it

A formal study of high-beta equilibria in

XIII-3

It is seen that if CR =0.3,

internal conductors has been

It will lead to a more realistic

the end loss can be reduced by 75% if five rings

stability

properties of internal ring end stoppera.

It follows from these considerations that the

the

ring

such

ahowa

pinch.

conductors

particle

location

internal

fR as a

function:

examination

proportional

underway also.

the presence of

multiple mirrors.

this region will

NUMERICAL SIMULATION

understand.

2 RINGS

5 RINGS

hybrid

I RING

Upon

/+OLENOIO

0.3

0.4

RING7

and

(Fig.

twofold

perform

confinement

This conical

Two-Dimensional Finite-Electron-Mass Hy-

brid Simulation. The finite-electron-mass hybrid

plasma simulation model of the lower hybrid drift

(LHD) instability haa encountered some numerical

difficulties that have proved quite difficult to

comparison of full PIC and

simulation of the same problem, it ia

apparent that the physics required to simulate the

COILS

EXTERNAL SUPPORTS

Figure XIII-3. Percent reduction in end loss versus fraction of particles trapped by single end mirror as a result of nonadiabatic scattering by the rings.

Figure XIII-4. Schematic of magnetic field geometry of CIRCE. One half-section.

0.2

“‘FZ?JZEI

Figure XIII-6. Hybrid simulation of the time evolution of B= with parameters the same as the full-particle all simulation shown in Fig. Comparison of the resulta of the two simulation techniques reveals very similar wavelength and growth rate for the developing flutes.

necessary in order to eliminate the random PIC

fluctuations already mentioned and has

frequency (w E 240 wee) oscillation of the in-

identification of

Finally, to insure that the instability could

included in the model, a one-dimensional stablity

analysis of this model has been completed.

” = O, the model contains only

Alfv6n and ion acoustic modes, as expected. The

dispersion relation for ” = O is

The instability arises from

homogeneous equilibrium.

LHD instability is retained in the hybrid model.

Figure XIII-5(a) shows B= contours of this insta-

bility at various times and y-averages of these

came contours are shown in Fig. XIII-5(b). Figures

XIII-5(a) and XIII-5(b) were obtained from a PIC

simulation and Figs. XIII-6(a) and XIII-6(b) show

the comparable results of the hybrid simulation.

Also evident is some enhanced diffusion (see Fig.

XIII-6(b)) on the high magnetic field aide of the

sheath. This diffusion has proved to be numerical

in origin, and consequently, the code as it now

atanda cannot be used for nonlinear studies. The

underlying this phenomenon

are quite complex and, in addition, have been

obscured by a combination of random particle-like

fluctuations due to the PIC treatment of the ions

and recently discovered lowfrequency oscillations

of the inhomogeneous equilibrium.

It is now believed that a numerical insta-

bility has been identified which gives rise to the

between

equations, which

numerical turbulence in the electrostatic field.

It is believed that this turbulent field then,

through greater resistivity, produces the observed

Progressing to this point has required the

further expanaion of this model to include the

capability of treating the ions as well aa the

electrons as an Eulerian fluid.

the

interaction

electrostatic

enhanced diffusion.

numerical mechanism

diffusion.

field

T=50

T=100

not

the

and

T=50

fluid

T=100

result

produces

electron

quently allowed

the experimentally

neutron yields was

This haa been

find that for

O<y<20cwe

neutron

pulses

T=150

T=200

from

some

Figure XIII-5. Full-particle simulation of the time evolution of the msgnetic field component B . The x-y contour -? of B= for O <-x < 10 cu are shown in (a) while t%eey-averaged profil~ of B= is shown in (b).

‘1~nd

T=200

T=150

subse-

a low-

XIII-5.

physical process still

Good agreement with

With

Thus

most

the

(4)

For configurations for which ” ~ O, electron

cyclotron motion must be followed but the relevant

frequency for time-step restriction is determined

by the components of ~ parallel to {.

the stability analysis has verified and exceeded

earlier claims concerning the maximum sllowable

time step—further encouraging our belief in the

ultimate utility of the model.

lD Hybrid Simulation Results.

a. THOR Experiment. Several more simulations

of the implosion phase of the Maryland Toroidal

Device THOR were performed.

observed field profiles and

found. The observed sequence

correlated

t (p)

the boundaryfieldswhich generated %11(C)

outer edge of the plasma.

each

than

much

When

smaller

meaningful

comparison

temperatures.

Krall has been

compressed phase

particle time step.

total Z flux) may be

ZT-40 should be made.

We have begun

programming

systematic

applied.

theory.

proper

6-1.

indicated by the calculations.

raya corresponding to a plasma of 5 kev were found

experimentally,which correlated in the calcula-

tions with a 3-to-5 keV low-density region on the

mimics the experimental x-ray detector is being

included in the code and ehould allow a more

C@aailinear Resistivity. The quasilinear perpendicular resistivity30 model of Liewer and

alternate option to the quasiempirical Chodura

reaistivity.31332 ‘l’hismodel requires a time step

with

included in

required

dynamics routines (which consume moat of the time

per cycle); therefore we are currently attempting

to subcycle several field advances into one

the running time of the code in the quasilinear

mode should be increased by a factor of 3 to 10.

c. Pitch Programming of zl’-4o.

stable q(r) profiles which satiafy the global

stability conditions - < 0.5, $= > 0 ($ = is the

found from equilibrium

to consider how

profiles may be set up dynamically. One way is by

q(rwall, t).

address the question of how to map q(rwall, t)

into q(r,~), the equilibrium q profile.

mapping depends on both plasma motion and field

diffusion; indeed, some diffusion is required for

q to penetrate the plaama. The low density region

near the wall is represented by a force-free,

wall-emitted plasma in which anomalous radial

electron thermal conduction and other processes

are assumed to keep the plasma isothermal and less

than 100 ev. Figure XIII-7 illustrates one example

of the transformation. In this case $= > 0, but

Note that q(r) near the center falls from

its programmed value due to resiativity.

higher q is desired in this region qwall must be

larger at earlier times, or a bias field must be

profilee on ZT-S is called for. These should then

be compared with experiments, and predictions for

this is accomplished

= ~

0.2

the

0.1—

code

0.0 —

-0.1L

0.0 -

Mercier

0.1 r

particle

Furthermore, x

several bounces

the experimental

A diagnostic that

over-compressed.

that used in

continuing.

initiated.

atable

Vlasov

-(J,~

If a

such

This

Thus

for

was

the code.

ions

t—

represented

This work

agreement

is still

wall(t)

was

are

7.7

d. APProach to Equilibrium and Anomalous

Cross-field Thermal Conduction. Late-time (- 3 aa)

data from the Staged-Theta-Pinch experiment were

used to test the approach of the code to an equi-

librium. Unfortunately the calculated equilibrium

Better

achieved when the electron temperature was limited

to 100 eV and smoothed. ‘l%ismay be indicative of

anomalous cross-field thermal conduction and/or the

decrease of turbulence and anomalous reaistivity

(observed at the University of Maryland) below

Cylindrical Hybrid Code Development. De-

velopment of a 2-D cylindrical (r-z) plasma simu-

lation technique which treats the electrons as a

massless, neutralizing fluid has recently been

3.85

R(cm)

Figure XIII-7.

The q(r) profile resulting from the given q profile.

in the nonradiative limit.

aPPIY this model first to the 8-pinch end-loss

standard PIC techniques and all components of both

Currently we are attemptin8 to

the electric and magnetic fields are retained but

treat this region aa a vacuum.

as follows. We assume denstty N, + Ji, and magnetic field ~ profiles. ‘e = O limit of the electron momentum equation gives the self-consistent electric field

ion current

Now the

problem, but a great variety of problems can be

The compressible fast cylindrical liner model

addressed once the basic code is written.

of Shearer and Condit, namely the impulse-momentum

The essential features of the algorithm are

approach,33 was extended to more correctly treat

eNc

i----

:ex;

v (NTe)

E=---

quaai-neutral limit

current, respectively.

useful as N + O.

update B

interest

B=-

cVXE,

CTR

n(:i+:e)

(6)

the

from

plasma

cylindrical

irrotational

density region.

Our intentions are to

G. FAST LINER DYNAMICS

are still ignored.

Faraday’s law to

many problems

is that (5) is

possessed an

modulus).

the bank.

regions

driving

(5).

have

not

the

The

the

This

liner

liner.

circuit

loaalesa

undriven liner

The results

(The

the

convergence

extended model was then used to obtain analytical

representationa for the optimized neutron yield

adiabatically compreaaed

and analytical representations for

corresponding initial and final conditions.

analytical results agreed very well with

CHAMISA code, and a paper by R. A. Gerwin and

R. C. Malone deriving and presenting these resulta

has been written for submittal to Nuclear Fusion.

As an example, one of the results ia that the

liner turn-around radius is a factor of order

unity times the geometric mean of the initial

liner radius and thickness.

More recently, the impulse-momentum approach

has been further extended to include the effect of

electrical

magnetically imploded in a Z-pinch mode.

earlier work asaumed that the

initial kinetic energy.) In

earlier work, the results could be obtained by

examining the dynamics near turn-around, so that a

high-pressure approximation was appropriate (final

plasma pressure larger than initial liner bulk

In the new work, the resulta have had

to be extended to include arbitrary pressures. A

result based on dimensional analysis is that, for

fixed final plasma temperature (- 10 keV), and for

fixed values of certain dimensionless ratioa that

are appropriate to the problem, the system gain

(ratio of fusion energy to capacitor bank energy)

scalea as the square root of the initial charge on

Thus, one has to optimize the device by

adjustment of these dimensionless parameters.

code is being written to do this.

will then constitute an upper bound on the gain of

a real liner-plasma device, since plasma loasea

During 1978, considerable effort was directed

to the development of the faat liner implosion

code, CRAMISA. At the end of 1978, the code had

where Te and Je are the electron temperature and

Te will be advanced in

time from an initial value and Je can be easily

generated from Ampere’s law given :L in the

electron and ion currents cancel.

in which

The quasi-psrticle “ions” are given initial

positions and velocities using weight factors so

that by integrating over velocities the desired

initial densities and ion currents are approxi-

mately recovered. The ions can now be advanced in

time by particle-in-cell techniques, and the newly

advanced density and ion currents calculated.

The remaining requirement to complete the

time step ia the recalculation or updating of the

electric and magnetic fields ao that they are

consistent with the new density and ion currents.

This is accomplished by using

in whtch ~ is again produced by

The most immediate problem

In that

indeed

comparatively low density, it is desirable to

the algorithm automatically detect and provide a

field solution from Maxwell”s equationa in the low-

capabilities: ID

in Russian conferences on plasma diagnostics and

hydrodynaatics; (2)

covering

superconducting magnet technology. As a result,

much Russian material of programmatic interest waa

phases, ionization, and electron degeneracy; (3) a

furnished to LASL scientists, and a considerable

resiativity of aluminum and copper covering all

English either in its entirety or in title and

fraction of this material was translated into

phasea; (4) nonlinear magnetic field diffusion (Be

in cylindrical geometry only);

translations were

thermal

following preprints and reports.

(6) bremsstrahlung loaa; (7) nondepleting thermo-

nuclear burn; (8) full coupling of liner to an

“A study of the Dynamics and Stability of a Liner,

circuit

Electromagnetically Accelerated on the LN-20

(9) boundary conditton options which

Device,” S.G. Alikhanov, V.P. Bakhtin, V.G. Belan,

various physical entities (e.g., an adiabatic

V.M. Brusnikin, 1.S. Glushkov, N.A. Karimova,

plasma, an incompreaaible liner, or Be or Bz

A.D. Muzychenko, Yu.M. Senatorov

adjacent to a perfect conductor), and thereby

permit significant savinga in computer time.

“Acceleration of Liners to High Velocities on the

Development of this code is continuing in

LN-20 Device,” S.G. Alikhanov, V.P. Bakhtin,

response to the needs of the Fast Liner Experiment

G.Yu. Lopatovsky, A.D. Muzychenko, V.A. Sibilyev

and

new

and

the

ion

Gerwin

driving

at LASL.

analytic

implicit

external

electron

following

treatment

equation-of-state

confinement systems.

experiments.

implosions

documented

obtaining

complete

feedback

provide

tions

LASL.

older

have

been

from

With

data

and

the

frustrated by

two

for

all

the

only);

tabular

simulate

Complete

(z-pinch

material

electrical

Lagrangian

conduction;

thermodynamic

abstract form.

(5) cross-field

experimental

difficulty

capacitor

Efforts

plasma

1iner

fast

bank

the

made

The code has been used extensively to make

“A Study of the Outlook and Optimization of

comparisons with the analytic implosion model of

Energetic ElectrodynamicsPlasma Guna,!t

benchmarks

V.I. Vasilyev, A.M. Zhitlukhin, V.G. Solovyeva,

reactor-studies codes. More recently the code was

Yu.V. Skvortsov, N.M. Umrikhin

used to model the full time history of liner

two

“A Theta Pinch with External Plasma Injection,”

configurations planned for use in experiments at

V.P. Alipchenkov, V.I. Vasilyev, I.K. Konkashbaev,

The results have been used as guides in

1.S. Landman, L.B. Nfkandrov, Yu.V. Skvortsov,

preparing for the initial faat liner experiments.

F.R. Ulinich, S.S. Tserevitinov

Comparisons have been made of code calcula-

Russian

“Construction of the MK-200 Installation,”

The agreement waa quite good for

R.G. Bikmatov, N.V. Goryacheva, A.M. Zhitlukhin,

(circa 1967) experiments which are well

A.I. Kolchenko, A.P. Lototsky, A.D. Kiakin,

literature.

A.M. Tikhonov, S.A. Sergeyev, Yu.V. Skvortaov,

comparison with the more recent LN-20 experiments

V.M. Strunnikov, S.S. Tserevitinov

H. TRANSITIONS OF RUSSIAN REPORTS AND PREPRINTS

“A Study of the Neutron and X-Ray Output from the

During 1977,Russian Exchange Workshops were

MK-200 Plasma Gun,” R.G. Bikmatov, V.I. Vasilyev,

held at LASL on the prospects of high-beta systems

V.V. Gavrilov, N.V. Goryacheva, A.D. Kiskin,

stabilization

N.M. Umrikhin, A.I. Yaroslavsky

In addition, a USSR-LASL

exchange workshop took place in the USSR on fast

“Passage of a Plasma along a Magnetic Field,”

liner implosions, a meeting of the Joint Fusion

A.M. Zhitlukhin, I.V. Ilyushin, B.Ya. Lyubimov,

Power Coordinating Committee was held at LASL In

Yu.V. Skvortsov, V.G. Solovyeva

midyear, and LASL CTR Staff Members participated

parameters and results from Russian preprinta.

Yu.V. Skvortaov, N.M. Umrikhin

eet

“Calculation and optimization of Plasma Guna,”

Translations of titles and abstracts (where

“Experiments with a High-Current Pulsed Arc,”

available) were made of the following reports and

S.G. Alikhanov, I.K. Knnkashbaev, B.S. Estrin

“Transport of Plasma Blobs with Energies of 100

“Production of Megagsuss Fields by the Method of

kJ,” A.M. Zhitlukhin, V.N. Lyaahenko,

Magnetic Implosion,” S.G. Alikhanov, V.G. Belan,

S.A. Ravichev, Yu.V. Skvortsov, V.M. Strunnikov,

G.I. Budker, A.I. Ivanchenko, G.N. Kichigin

“Collision of Two Dense Plasma Streams,”

“Ohmic Heating of a Dense Hydrogen Plaama,”

V.M. Alipchenkov, R.G. Bikmatov, V.I. Vasilyev,

S.G. Alikhanov, E.L. Boyarintsev,

N.V. Goryacheva, A.M. Zhitlukhin, I.V. Ilyushin,

I.K. Konkashbaev, V.A. Kornilov

A.D. Kiskin, I.K. Konkashbaev, V.N. Lyashenko,

L.B. Nikandrov, S.A. Ravichev, Yu.V. Skvortsov,

“Investigation of a Fast Pulsed Discharge in

V.G. Solovyeva, V.M. Strunnikov, A.B. Titov,

Hydrogen,” S.G. Alikhanov, E.L. Boyarintsev,

V.A. Kornilov, T.S. Melnikova

“Processes Occurring the the Gap of a Coaxial

S.G. Alikhanov, G.N. Kichigin, I.K. Konkashbaev

Plasma Gun Fed from an Inductive Energy Store,”

E.A. Azizov, I.V. Kochurov

“The Generation of Intense Toroidal FielciaThrough

Magnetic Flux Compression,” S.G. A.likhanov,

“A Pulsed Thermonuclear System with Dense Plasma,”

“A Study of Plasma Stream Parameters in Plasma

Guns,” V.I. Vasilyev, A.M. Zhitlukhin,

A.P. Lototsky, V.G. Solovyeva, Yu.V. Skvortsov,

“Radiative Instability of a Bounded Plasma,”

N.M. Umrikhin

S.S. Tserevitinov

V.A. Sibilyev

I.Ya. Shlpuk

preprints:

V.G. Belan

S.G. Alikhanov

“Comprehensive Experiments on Plasma Compression

Injection,” Yu.V. Skvortsov

in a Z-Pinch Liner System on the LN-20 Device,”

S.G. A1.ikhanov,V.P. Bakhtin, V.M. Brusnikin,

“Experiments on Heating of a Dense Plasma by a

1.S. Glushkov, G.Yu. Lopatovsky, A.D. Muzychenko,

Strong, Fast-Rising Magnetic Field in a Straight

“A Theta Pinch with Liner and External Plasms

Theta Pinch,” V.A. Burtsev, A.P. Zhukov,

B.V. Lyubin, V.A. Kubasov, V.N. Litunovsky,

“Technical Design of the LN-20 Device,”

V.A. Ovayannikov, V.G. Smirnov, V.P. Fedyakova

S.G. Alikhanov, V.G. Belan, 1.S. Glushkov,

A.I. Kolchenko, I.K. I@nkashbaev, V.A. Lsnis,

“Psssage of a Plaama Through a Magnetic Nozzle,”

A.D. Muzychenko, E.N. Polyanskaya

A.M. Zhitlukhin, I.V. Ilyuahin, B.Ya. Lyubimov,

Yu.V. Skvortsov, V.G. Solovyeva

“A Straight Z-Pinch-Liner Thermonuclear System,”

S.G. A1.ikhanov,V.P. Bakhtin, V.M. Brusnikin,

“The Angara-5 Accelerator Complex,” E.P. Velikhov,

1.S. Glushkov, A.D. Muzychenko

V.A. Glukhikh, O.A. Gusev, G.M. Latmanizova,

S.L. Nedoseyev, O.B. Ovchinnikov,

“A Study of Model Thermonuclear Liner Systems,”

A.M. Pasechnikov, O.P. Pechersky, L.I. Rudakov,

S.C. Alikbanov, V.P. Bakhtin, V.M. Brusnikin,

M.P. Svinyin, V.P. Smirnov, V.I. Chetvertkov

1.S. Glushkov, R.Kh. tirtmullaev, A.L. Lunin,

A.D. Muzychenko, V.P. Novikov, V.V. Pichugin,

“A High-Current Electron Accelerator Producing

V.N. Semenov, G.E. Smolkin, E.G. Utyugov,

Microsecond Beam Pulses,” V.A. Glukhikh,

O.A. Gusev, M.P. Svinyin, O.P. Pechersky,

A.S. Perlin, E.M. Mellekh, E.P. Pavlov,

“Numerical Modeling of the Compression of a

Magnetic Field by an Imploding Liner,”

R.A. Volkova, V.M. Goloviznin, F.R. Ulinich,

G.M. Latmanizova, V.S. IWznetsov, M.I. Avramenko,

Generator for Feeding a High-Current Electron

V.B. Msrkov, L.A. Vykhodtsev, N.G. Beruchev,

Accelerator,” V.A. Glukhikh, B.G. Ksra.sev,

V.V. Kharitonov, V.N. Skripunov

“A Fast-Acting Foil Disconnect Switch,”

“Homopolar Generators with Superconducting

A.B. Andrezen, V.A. Burtsev, V.M. Vodovozov,

Excitation Winding,” V.V. Kharitonov,

V.A. Dubyansky, N.P. I?gorov,V.N. Litunovsky,

V.A. Ovsyannikov, A.B. Produvnov, V.F. Prokopenko,

“A Study of the Electrical Explosion of

Cylindrical Foils in Air,” V.A. Burtsev,

V.A. Dubyansky, M.P. Kssatkina, A.B. Produvnov

“Aspects of a Pulsed Thermonuclear Reactor with

Imploding Liner,” G.A. Baranov, V.A. Glukhikh,

O.A. Gusev, V.A. Zheltov, E.N. Izotov,

Yu.K. Kslmykov, A.B. Komin, M.V. Krivosheyev,

V.G. Kuchinsky, I.F. Malyshev, A.B. Mineyev,

“A Two-Stage Electroexplosive Current

Interrupter,” V.A. Zheltov, A.V. Ivlev,

A.S. Kibardin, A.V. Kromin, V.G. Kuchinsky,

“An Inductive Energy Store with Electromagnetic

Current Multiplication,” A.V. Ivlev,

A.S. Kibardin, A.V. Kronin, V.G. Kuchinsky,

Yu.A. Morozov

I.V. Shestakov

S.A. Kolyubakin

M.A. Gotovsky, V.V. Kantan

K.M. Lobanov, I.F. Mslyshev

G.V. Trokhachev

REFERENCES

A.P. Favorsky

A.Yu. Koretsky

“Pulsed Energy Sources Based on Inductive Energy

Stores,” V.A. Glukhikh, O.A. Gusev, A.I. Kostenko,

B.A. Larionov, N.A. Monoszon, A.M. Stolov,

“Research and Development of a Homopolar Generator

with Superconducting Excitation Winding,”

V.N. Artemov, V.A. Glukhikh, K.K. Dyschenko,

V.V. Ivanov, B.G. Karasev, A.Yu. Koretsky,

I.F. Malyshev, N.A. Monoszon, V.N. Skripunov,

G.V. Trokhachev, G.F. Churakov, V.V. Kharitonov

“A Shock-Excited Hompolar Disc Generator,”

E.A. Bezgachev, V.A. Glukhikh, V.V. Ivanov,

B.G. Karasev, I.F. Malyshev, V.V. Kharitonov

“Development of a Shock-Excited Homopolar

Beta Plasmas,

(to

J. P. Freidberg and L. D. Pearlstein, “Rotational Instabilities in a Theta Pinch,” submitted to Phys. Fluids (1978).

D. C. Barnes and J. U. Brackbill, Nucl. Sci. 18-32 (1977). and Engr., ~,

W. H. Bennett, Phys. Rev. 45, 89 (1934).

R. L. Morse and J. P.Freidberg, Phys. Fluids 13, 531 (1970).

L. Turner, to be published.

D. Montgomery, L. Turner, and G. Vahala, “Three-Dimensional Magnetohydrodynamic Turbulence in Cylindrical Geometry,” to be published in Phys. Fluids.

S. Chandrasekhar and P. C. Kendall, Astrophys. J. ~,

457 (1957).

F. P. Bretherton and D. B. Haidvogel, J. Fluid Mech. ~,

129 (1976).

J. B. Taylor, in Pulsed ~ ed. by D. E. Evans (Pergamon Press, Oxford, 1976) 59-67.

Don S. Lemons and S. Peter Gary, “Current Driven Instabilities in a Laminar Perpendicular Shock,” J. Geophys. Res. be published) (1978).

S. Peter Gary and J. J. Ssnderson, “Density Gradient Drift Instabilities: Oblique Propagation,” Phys. Fluids (to be published) (1978).

J. P. Freidberg and R. A. Gerwin, Phys. Fluids ~,

1313-1315 (1977).

S. Peter Gary, “Electrostatic Heat Flux Instabilities,” J. published) (1978).

Plaama Physics (to be

S. Peter Gary, “Ion-Acoustic-Like Instabilities in the Solar Wind,” J. Geophys. Res.

(to be published) (1978).

L. Turner, Proceedings of the Annual Controlled Fusion Theory Conference (May 4-6, 1977, San Diego, California), paper D20.

H. R. Lewis, Proceedings of the Annual Controlled Fusion Theory Conference (May 4-6, 1977, San Diego, California), paper H8.

T. E. Cayton, R. A. Gerwin, and L. Turner, Proceedings of the Annual Controlled Fusion Theory Conference (MSy 4-6, 1977, San Diego, California) paper PB8.

S. I. Braginskii and B. B. Plasma Physics and the Problems of CTR, Vol. III, p.

B. B. Kadomtaev in Plasma Physics and the Problems of CTR, Vol. IV, p. 417.

P. C. Liewer and N. A. Krall, Fhys. Fluids 16, 1953 (1973).

R. Chodura, Nucl. Fuaion~,

A. G. Sgro and C. W. Nielson, Phys. Fluids ~,

J. W. Shearer and W. C. Condit, “Magnetically Driven Metal Liners for Plasma Compression,” in Energy Storage, Comvressiont~ ing, W. B. Bostick, V. ~rdi, and O.S.F. Aucker, Eds., (Plenum Publishing Co., New York, 1976).

S. Peter Gary, A. G. Sgro and A. W. DeSilva, Mhial Density profiles and Nonclassical

Thermal Conduction in Linear Theta Pinches,” Phys. Fluids (submitted for publication) (1977).

S. Peter Gary, “lhe Electromagnetic I-onBeam Instability and Energy Loss of Faat Alpha Particles,” Nuclear Fusion (to be published) (1978).

H. Ralph Lewis and Keith R. (submitted for publication in Journal of Mathematical Physics).

H. R. Lewis and L. Turner, Nuclear Pusion 16, 993 (1976).

Schwarzmeier, H. R. Lewis, B.

J. L. Abraham Shrauner, and K. R. published.

Ralph Lewis and Leaf Turner, to be

H. R. Lewis, Proceedings of the Annual Controlled Fusion Theory Conference (May 4-6, 1977, San Diego, California), paper H8.

J. P. Friedberg, Phys. Fluids Q (1972).

Kaufmann, J. Neuhauser, 462 (1973).

L. Turner, Phys. Fluids 20, 654 (1977) and Phys. Fluids ~,

662 (1977).

H. published.

w. Grossmann, M. Nuclear Fusion~,

Symon,

126 (1976).

Symon, to be

Switch-

55 (1975).

Kadomtsev in

G. 1, Chandler, R. Conrad, N. E. Greene, F. C. Jahoda, K. A. Klare, J. W. Lillberg, L. W. Mann, C. W. Nielson, S. M. Ross, R. W. Wilkins

The

c-l).

used for

scientific

ment (Sec.

INTRODUCTION

B-4) , applications of

experiment minicomputers

per discharge (Sec.

tion system (Sec.

experimental

improved by

sional data

(Sec. c-6).

Hardwar~.

algorithm

completed

multiple

handles

higher

Scylla

c-5).

speed

data,

IV-P

The

program, and adds an additional 32 lines for

The shortage of computer resources at LASL,

coupled with the increased utility of the PDP-10

expanded to handle a total of 64 lines at a cost

and MFE network, has made the USC facility a vital

of approximately $21k.) This terminal handler is

link in the LASL MFE effort. The PDP-10 has been

also expected to cut response time on the PDP-10.

problems,

former terminal concentrator requires the

accounting, and inventory control.

PDP-10 CPU to perform all character handling one

large particle simulation and MHD stability codes

The new device contains a small

computers. The yearly progress will be discussed

functions and receives from and transmits to the

have been shifted to the network from the LASL

in five phases: hardware enhancement (Sec. B-1)

software enrichment (Sec. B-2), network develop-

b) A disk controller (DEC designation RH1O )

B-3), experimental interfacing (Sec.

was ordered so that the traffic to and from the

the

three disks on the system could be separated.

network, and future plans for augmentation of the

the

computing facilities (Sec. B-6) .

response time for time-sharing users much faster.

The preliminary design of the ZT-40 control

c) Memory Multiplexer (DEC designation MX1O)

and data acquisition system is outlined (Sec.

was needed to give the terminal line-concentrator

operating system was

memory access. The existing memory has only four

upgraded and routinely handles 60k bytes of data

c-2).

needed

data

(The

own

its

The

and

USC

This

use.

at a

time.

computer

processor

improvement

Many of the

time-sharing

administrative

XIV . COMPUTERS

PDP-10 in block mode.

Automated two-dimen-

A link between the

channel, and the

plexor handles up

thus allowing the

to share a port

interface, the

is in progress

advanced with

frame grabber

microcomputer

new products.

graphics and

capability

expansion.

equipment

digitized

handling

papers.

PDP-10

which

ports

those

d) A

uses

disk

for

and

the

and

are

which

can

these

system

handles

should make

concentrator

for the network

(Motorola 6800)

The multi-

developed

channel,

tape

data

The

the

and

central processor.

to eight channels on one port

channels with the least traffic

and also providing room for

improved with the aid of a computerized calibra-

microprocessor

acquisition has

for

dual-headed printer (QUME) were purchased so that

a sufficient character set could be put on line to

do the word processing required for technical

Polaroid streak pictures and work

The second head provides Greek letters

implementing a digital video

and mathematical symbols in addition to the normal

alpha-numeric symbols on the first head.

software

CTR USER SERVICE CENTER

PDP-10 to drive the printer.

items

Although all of the items mentioned above

purchased this year are listed below together with

were ordered early in the year they did not arrive

a description of how each Item is expected to

until February 1978. In a field that is improving

increase the capacity and capability of the USC:

technically very rapidly it is unfortunate that

a) A terminal line concentrator (Digitial

one has to plan an acquisition so far ahead.

Equipment Corporation designation DN87)

is wise to attend new product displays to aid in

lines

provides

this planning but we have learned the hard way

that delivery dates often slip badly especially on

networking to minicomputers in the experimental

operational (Sec. c-3). The Sigma 2’s have been

terminals, serial CA!4AC

links, and a new graphics display system (Sec.

c-4) . Transient recorder linearity has been much

was incorporated into the text editor and improved

line speed was originally llmited to 1200 baud but

The word-proceeding formatter

error correction protocol is also provided, The

the new terminal line concentrator will permit It

a. Equation formatting for technical papers

to be increased to S1600baud.

formatter.

CTR-Division

was made easier.

Laboratory standards.

in the following ways:

much easier to maintain.

asynchronous llnes as

asynchronous llnes.

consultant).

these sites.

resulting

methods.

This

high

for

our

PDP

This

system

making

facility

represents a

PDP-10 FORTRAN

3, Network.

subroutines

procedures.

installed.

Facility.

language

designed

overhead

The new

Library

severe

There

have

was

and

b. Double-column formatting was implemented.

reorientation of

o. Figure captions and figure placement in

thinking on how to handle data streams from the

relation to the text were made to conform with

minicomputers. It was originally thought that,the

d. Footnote handling was made easier for the

11/40 to the PDP-10, could be shared by

DA28, which handles traffio from the MFE network

another PDP 11/40 passing experimental data.

e. Internal changes were made which de-

has become apparent that the software complexity

creased the editor response time as much as ten

of such a system would require one St+aff Member

times on some commands and made the editor code

working nearly full time for maintenance since the

software would have to be modified for every DEC

f. Commands were added which made FORTRAN

monitor and network change.

programming, debugging, and execution easier and

DA28 appeared to be cost effective for us, it is

inadvertent file deletion was made more difficult.

aPParent that the MFE network staff would nok be

The word-processing system is now used by all

able to support such a change.

secretaries to do everything but

nonstandard, was

small memos or lett,ers. Each secretary in the

effective for us. We plan to support higher baud

Division has a word-processing terminal which

rat,es than 9600 by buying synchronous lines for

produces exportable copy and it has been estimated

the DN87 which can pass up to 40 JOO baud.

that a trained secretary can produce a technical

The PDP-10 utility library has been improved

paper in one-third the time required by former

substantially throughout the year.

Papers produced by word-processing are

national Mathematical and Statistical Library was

also less likely to have errors and more likely to

purchased to supplement our library of subroutines

have revisions that improve content or conciseness

translated from the Los Alamos Central CompuLing

because of the ease of modification.

Systematic testing of Lhe accuracy of

The LASL USC text editor word processor has

been exported to the other major nodes of the MFE

routinee for simple functions is now under way

network and has been found to be equally useful at

because users have noted discrepancies between the

The access to the PDP-10 from experimental

The system software is currently up to date

data-acquisition minicomputers has been enhanced

with the latest DEC monitor and network sofkware

by a special PDP-10 monitor software module added

this year (in collaboration with a DEC software

written

module

input-output, byte manipulation, and operational

PDP-10 solutions and solutions on other computers.

eliminate some of the drawbacks associated with

connecting the minicomputers to the PDP-10 by

The MFE network has developed

The normal monitor treats

into a stable and reliable computing system for

terminal lines with

the local users. Due to the current saturation of

software

LASL computing facilities, most users have found

restrictions on the nature of the data.

that the MFECC network provides better turnaround

monitor module bypasses the terminal handler with

and produces rapid, high-quality graphios ouLpuL

its high overhead and establishes a protocol with

via the Versatec printer/plotter.

byte count and check sum which allows passing of

microfiche is still produced at LLL and arrives in

arbitrary streams of eight-bit data bytes.

the mail a week after it is run. The addition of

data

lines

handling

not cost

the USC staff

To go it alone,

Although sharing the

Unfortunately

The Inter-

several

machine

handle

obher

been

tape

subroutines and

In order to satisfy the

for

the

large core (LCM) limitations and the small LASL

time allocation, so these codes are st;ll run at

Tape access time for archival files has

SLOPE (a simulator for the CDC operating system)

participated in the development of an Impurity-

and DBCTRL (an improved interactive debugging

modifications

package) wibhln the past year has made the system

Equilibrium-Stability code so that it could be

somewhat more flexible and easier to use, as well

applied to ZT-40 and the High Beta Tokamak.

as provided a greater degree of compatibility with

difficulty in running large codes at night due to

give users an easy method for generating 6800

microcomputer has been purchased and installed to

been improved by the new FILEM (file manager) and

processor is scheduled to be replaced in fiscal

archival

It is hoped that the new processor

diminish even more with the new mass storage

will be four times as fast as the KI currently in

system that is being acquired for the network.

use. The equipment being purchased in fiscal year

WiLh the addition of a CRAY computer to the

1978 and 1979 will be compatible with migration to

network by early summer, it is expected Lhat

the new configuration. The data channels will be

Ex~~erimental Interfacin~~.

on Scylla IV-P is connected to the PDP-10 via a

upgraded to a more compact, reliable type that

1200-baud line that employs the monitor module

requires less maintenance, and higher density

discussed under software. This line is used for

disks will be considered if user demand warrants

replaced with a type that can access more than

256k of memory, the tape units will be replaced

with higher density units, the memory will be

time

LASL computing facilities.

For

the

LASL .

demands

network

increase.

shipping shot

retrieval time

equilibrium codes

The USC now has

background or at

spring of 1978.

baud if needed.

been prepared.

analysie.

editing,

network

night.

users

word

data

are

for

and

it.

code

will

code.

demands

A cross

compiler

continue

Radiation

increased

year 1980.

storage and

The PRIME 300

files should

Future Plans.

We have found some

the PDP-10

1, ZT-UQ.

integrated

Operating

AUTOMATED

procured

particle

is near

logging

CONTROL

tasks,

System

staff

data

Time

DATA

four

300,

both

true

USC

and

MHD

has

(ZT-40 control and data acquisition minicomputer)

control and data-acquisition system was completed.

and the Sigma 2 .on ZT-S, will be available in the

A prime 40() computer system was ordered and

makes it possible to boost all these lines to 96OO

speed advantage over the Scylla IV-P Prime

The new terminal concentrator

delivered for this task. It has a factor three to

greater capacity for eventual extension of

not experienced programmers, elementary manuals on

multi-user time-sharing (RPRIMOS IV) in Real

on. Since many of our users are

fast memory and disk size, and capability for

how to use the PDP-10 and the MFE network have

(RTOS)—i.e., while the

104 users and

mode. The present hardware consists of 64k 16-bit

Control Program is operating in a top-priority

saturation during prime time. Often we have all

words of main storage, two 3M word disk drives,

eight dial-up lines in use and as many as 40 users

two Tektronix 4014 terminals, TI 733 terminal for

on Lhe PDP-10. The users are split about equally

Versatek

between network and PDP-10 stand-alone users. The

printer-plotter. A CAMAC branch driver has been

generally running

simulation, reactor design, or MHD codes that are

interface to the experiment.

large production problems. The PDP-10 problem mix

The computer will be located in a screen room

processing, and

which will also house much of the data acquisition

Some 2-D MHD time-dependent and

equipment. As shown In Fig. XIV-I, the Computer,

run

The

via the CAMAC branch driver, communicates over a

parallel highway with several CAMAC crates within

MHD

the

Usc ,

central

the Motorola

electrostatic

COMPUTER

into

RTOS

for

the

AND

archival

subsequent data analysis on the PDP-10. The PRIME

100 has been tested on the same type of line. IL

is passing data to the PDP-10 for analysis of

experimental diagnostics. Ports for the PRIME 400

PROCESSING

Preliminary design of the ZT-40

aPRluc

F18EROPTIC LtNK

-+________

CAUAC ,” CRATCOU,

________—-J-

ntcys

ANALO0

DIOITAI.

~-__---

InANalcNr

$UPPRSSSORS

CONTROL CENTER

Ilourlold DEVICE,

*OSITION S8MXIRS

VOL1bCE DIVIDERS

L-q:

the screen room.

machine operation.

operation

valves,

wiring,

crowbar

excePt

etc.)

room,

gaps.

fast

the

.-,

Erc

OIGITAL

ANALOG

$UWLIE8

-1

PDP-lo

RAMPS10 POWCP

Primets

Thomson

Mol.b?mN 0EVICE8

SWITCHES COAITAC1OM

Biomation

equipment.

scattering

latest Prime

A few modules are

DRIVERS l Tnols!cN1 5uPPnEaxWS I

optical coupling, the

(operating andlor

ASCII-formatted

isolating

include

support

control

shot.

CAMAC

shown

the

Scvlla IV-~. The operating system on the

Scylla IV-P Prime 300 computer was upgraded to Lhe

revision

controller converted from an 8-sector to 32-sector

format, for compatibility with currenljreleases of

Prime software and with our other Prime computers.

time-sharing system,

implemented to permit multiple users when

control program is not runninz.

Data acquisition devices currently being read

by the computer include 52 Biomation 610’s, two

8100’s, 80

scalers, and 62 ADC channels of the single-point

diagnostice—nearly

bytes of data per shot. A small auxiliary (6 ft

by 5 ft) screen room ia being installed, which,

through optical coupling and serial CAMAC link to

the computer, will accommodate as many as 100

additional diagnostic channels. The two Biornation

8100’s were placed in tandem to extend the time

scale of the coupled cavity data from 40 to 80 IIS

whfle maintaining 0.02-I.Isresolution.

the Experiment Interface Computer (EIC) PDP-10

link concept has been abandoned. The new system

Is based on a high-speed serial terminal-type link

with CAMAC teletype driver modules rather t,hana

very high speed parallel link.

1200-baud line speed all the data from a Soylla

IV-P shot is shipped In two to four minutes. When

the planned installation in early 1979 of9600-

baud input port,s is complete, less than half a

minute will be required and adaptation of Lhe

shipping program to the Real Time Operating System

will cause this to occur automatically after eaoh

SEC

t,he

60 000

610”s,

PRIMOS

neutron

the disk

111, was

(No. 11) and

As disoussed above,

At the present.

files can be shipped in

arbitrary

The same

support

number

remote

for

either direction so that program development as

well as data analysis can be run on either end of

the link according to convenience.

system is operational for the Prime 100.

undergone several important improvements. These

Our two Sigma 2’s have again

terminals to edit and control programs, aerial

Figure XIV-1. Conceptual design of ZT-40 control system. Dotted immediately line above power supply room.

encloses electronics located

Many of the Individual CAMAC

modules will be data acquisition devices connected

to the external world through patch panels and

signal conditioning units.

dedicated to conbrol functions either for machine

diagnostics

particular, as shown, one module communicates with

the fast Liming pulse generators,which in turn,

through fiber optios links, provide triggering

signals both for the capacitor banks that energize

ZT-40 and various diagnostic equipment. Two other

modules, also through fiber optics links, serve

the less aritical timing control and monitoring

functions of the diagnostics equipment and the

Finally, there

Fig. XIV-I, a CAMAC module serial highway driver

that is optically coupled to the machine control

center located in nine shielded relay racks near

t,hepower supplies. This cont,rolcenter contains

additional CAMAC crates and interface circuitry

designed to implement all the control functions

monitoring capacitor banks, interlocks, vacuum

cfrouitry from voltage transients generated by

More details of the screen

triggering

while

the

control center to machine interface, and the gap monitor development are dascr~bed in Sec. XVII.

instrumentation

terminals, and a new display system:

and

can

the

send

well

only

treats

Terminals.

recognizes

Serial CAMAC.

distances with

card reader and teletype.

defective devices.

noise-sensitive

plots on

standard

Scyllac

been

link

was

the

The system monitor has been

smaller version of it (calendar plotting removed)

revised to permit connection of any number of

The output character and

terminals just as if they were the system console

control strings intended for storage displays,

teletype. These terminals are interfaced through

such as the Tektronix 4010; were converted by

CAMAC modules having first-in first-out buffers as

monitor overlay into commands for the Grinnell

Systems unit. The result, although software-speed

communicate to a core resident program Which

limited, is faster than a terminal operating at

The old routines for standard graphs

necessary

characters in a manner similar to the PDP-10.

remain available where the 8k word saving in core

This program queues for~ground tasks. The monitor

usage of this more limited package is necessary.

structure itself limits the system to a single

background user at. any one time.

Trangient Recorder~.

the Standard

the new editor reported last year, this terminal

transient recorders initiated examination of their

capability has eliminated nearly all ueage of the

A method to quickly test linearity of a

Serial CAMAC drivers have

It uses a repetitive linear ramp waveform input,

large number of units was required and devised.

been implemented through Individual modules in a

histogramming

standard (parallel highway driven) CAMAC crate.

frequency of occurrence of each discrete output

With the use of special adapters (U-ports) they

voltage level. With removal of certain resistors,

receive data

one

glaring breaks in the frequency spectrum (output

levels that almost never occurred) were also

direction. The original motivation was to support

removed. The method measured the dynamic gain and

Biomation 61o transient digitizers at the RFX

also revealed that the divider string for the

experiment located over 500 feet from the computer

comparators has too little current, resulting in

screen room. The standard 200-s

low frequencies and hence

24-bit CAMAC operation was achieved.

serious error If dc calibration is used.

one serial CAMAC driver replaces the previous

deficiency is in the process of being corrected

with installation of a special Zener diode.

Biomations and another links the electronics chop

cumulative error from the average gain measures

the integral nonlinearity, i.e., deviation from

tesbing of new equipment or trouble shooting of

The circuit changes have

improved this from more than f 2.0 to about ~ 0.4

c, New Display. The old four-channel video

in units normalized to 64 (six bits) full scale.

disk system used to display waveforms, text, and

problems

interfaced Biomation 610 transient recorders have

increasingly unreliable.

resulted in a simpler and more reliable handshake

electronic system to perform the same function has

procedure Ln the interface while reducing the

purchased from Grinnell Systems. It has 512 X 512 display points for each of four screens. It

commands needed to read each data word from four

Because the dynamic shift register that

interfaced through CAMAC and is capable of

refreshes the Biomation memory operates on a

generating alphanumerics, vectors, boxes, more.

In the past the Sigma 2!s have had their own

limited plotting routines. For compatibility with

the increasingly popular plotting package PLOT 10,

faster cycle time (2 us) than the data can be

transmitted over a CAMAC line, consecutive words

are read from the output register at 512 ps

intervals, i.e.,after one complete 256-word memory

refresh cycle, when the next consecutive word

used on the PDP-10, It was decided to make a

again appears in the output register.

computer

and

each

pair

wire

They

number

design.

control

readout,

computer

9600 baud.

serializing.

linearity of

Together with

time period per

over considerable

for the Sigma 2~s.

A more modern digital

straight line ramp.

general -purpose

decreased

Solutions

Currently

monitors

gain at

to one.

reading

become

ZT-S

and

has

for

the

61o E

Engineering

Doubts about the

CAMAC

with

This

The

the

As customary with first models from a

fications are sought to do the entire analysis on

grabber,

(VIPS), was

can

eight-bit information (256 gray levels) either for

one 512 by 512 picture or for four-256 by 256

manufacturer, improvements have been necessary and

implemented both by ourselves and EG&G.

camera has been interfaced to an Imacon streak

Aut_mated 2-D Dat~.

camera output and the frame grabber memory has

oipation of obtaining a TV frame grabber, an algo-

been interfaced through CAMAC LO the Sigma 2.

rithm was completed by ,Group M-8 for determining

memory test program was written which writes the

streak pioture trajectories and trace widths from

checke

digitized Polaroid pictures.

pattern lengths and random access. Failures after

transferred to the MFECC network CDC-7600 and

modification of t,he unit and replacement of some

debugged, and a substantial inventory of Scyllac

marginal components are down to about one bit per

data shots accumulated, awaiting further detailed

The

Inc.

frame

sYatem

analysis.

pictures.

Videology,

processing

Stored frames can be read on Lape in the same

format used for the digitized Polaroid piotures

for transmission to the network 7600, enabling

comparisons to be made as both line detection al-

gorithms and spatial and intensity calibrations

Eventually, sufficient simpli-

and

from

then

video

image

store

local

The TV

In enti-

acquired

digitize

frame and

100 frames.

acquisition.

minicomputer

are developed.

This algorithm was

immediately

it,

data

upon

including various

studies are being performed for several conceptual

Successful development and testing of a

J. D. Rogers, E. M. Honig, C. R. King, J. D. G. Lindsay, G. A Miranda, P. Thullen, H. Vogel, R. W. Warren, D. M. Weldon, J. M. Weldon, D. H. Whitaker, K. D. Williamson, Jr., J. J. Wollan

the

will

power

placed

(TPFS).

SUNNARY

directed

objective

development

toward TNS

switching, and

tokamak TNS machines.

engineering design

(~12. lkA, ~2.72T,

Westinghouse

400-kJ,

25-kA

coil

for

one

During the year the goal of the group changed

from development of the superconducting magnetic

energy transfer and storage (METS) system for use

in future theta-pinch experiments to development

of superconducting tokamak poloidal field systems

The emphasis on the TPFS program is

requirements.

program ia

integrated system demonstration by

nominal 50-kA, 20-MJ, 7-T superconducting coil

swung in a 2-s bipolar mode.

program of the past few years as a basis, emphasis

areas - superconducting

three

supplies.

The METS program wI1l be concluded with a

demonstration METS-SFTR coupled superconducting

prototype system which should be operable in late

1978 or early 1979. This system is to be composed

superconducting

storage coil and all associated charging and

switching equipment. The coil will be charged in

300 s to a field level of 2.5 T at 25 kA through

continuous duty leads and discharged in 0.7ms

through special coaxial pulsed-duty leads.

energy transfer is efficient, being determined by

the storage coil inductance, transfer capacitance,

and load inductance resonant circuit.

Major accomplishments during the past year

include the following items:

Initiation of a preliminary conceptual

superconducting TPFS coil.

Successful swinging of the 300-kJ, METS,

the

1.1 T/s).

Successful demonstration of commutated dc

machines as pulsed power supplies for the TPFS.

mln

The

With

major

switch

hybrid

coils,

provide

Near

the METS

addition,

Successful

technology

1982 of a

continuously,

Successful

level of 30 kA.

Successful

500-MJ homopolar.

resistance of 15 nS2.

superconducting magnet

xv. MAGNETIC ENERGY TRANSFER AND STORAGE

conservation program.

Coil Design. A

withstanding 60 kV.

transposed cable.

superconducting

negotiated for

20-MJ COfi.

Proposals

suppliers

complete.

bipolar

20-MJ,

50-kA,

cable.

energy

were

7-T,

TPFS

mode

for

lhe

Completion of several studies for the

tokamak design group of General Atomic including

the conceptual design of a 3-MA breaker and a

Completion of the interrupter facility

upgrade permitting a 60-kA, 60-kV test capability.

interrupters to 37 kA and 38 kV.

testing

operation

vacuum interrupters in parallel at a total current

completion

for

evaluation test facility.

development

cryogenic disconnect capable of carrying 5 kA at a

capable

interrupting

Successful operation of Helix Corp.,

700-W, 4.5-K refrigeration system.

Instigation of a Laboratory-wide helium

7-T, 20-MJ SUPERCONDUCTING TPFS COIL

engineering destgn of a 7-T, 20-MJ superconducting

energy storage coil has been initiated. The first

magnetics and superconducting cable designs are

Energy loss calculations for a

cosinusotdal magnetic field swing from +7 T to

-7 T were made for a 50-kA, multistrand Roebel

A cryostable 13:1, Cu to NbTi

sub-bundle was designed.

calculations were made to assure stability.

similar design was also carried out for a 5.5-T,

received

development

Ttro contracts are being

7.6-m lengths of prototype cable

and

the

use

the

two

this

25 kA

3-pin

field

axial

cable

17.8-cm

current

carrying

6-T, dipole,

Minimal heat transfer

preliminary

industrial

conceptual

50-kA

from

NbTi

1-s

’

Magnetic Corporation of America (MCA) working with

coil was

first

~12.1-kA bipolar operation without going normal.

together and insulated with double nylon wrap.

Peak swing in the magnetic field on the winding

Two configurations will be made.

from + to -2.72 T with an average rate of

the sub-bundle will have two superconducting and

These numbers are essentially

limits of the test facility and not of the coil.

superconducting and three copper strands. six of

Energy loss during the operation was 0.19% of the

these first sub-bundles combine around a stainless

peak energy stored in the coils or 0.095% of the

cable to form the second sub-bundle. Twenty-four

total energy transferred”out of and back into the

of these are then wrapped around a strap to make

This loss is equivalent to vaporizing 0.34

the cable. The second cable, which ia to be made

liquid helium

Intermagnetics General

single,

and varioua lengths of subcable components for

One such cable is to be built by

Preliminary analysis

with

four

starts

copper

testing.

strands;

six-around-one

outer periphery.

the first sub-bundle.

Study. Fabrication

300-kJ METS COILS

Corporation for

and

with

was

the

three

coil.

value

(IGC),

larger,

second,

negative

soldered

liters of

sub-bundle

polyami.de-

Corporation

magnetically

In the first,

superconducting.

The Westinghouse

change of 1.1 T/s.

hysteretic in nature.

wfthout going normal.

with the 300-kJ coil.

became very marked.

made.

first

both

the

7- and 9-T, 20-MJ, 50-kA prototype superconducting

TPFS coil designs, costs, and schedules.

imide-filrinsulated strand for the equivalent of

with a monolithic conductor and its performance

‘IIIisstrand has a complex

was not expected to be as good aa the Westinghouse

copper, copper-nickel matrix with a copper-nickel

coil. The coil could be swung ~3 kA at a rate of

spider segmenting the bulk of the copper in the

about 0.34 T/s without going normal.

The LASL coil is a fully potted coil made

the current swing to ~4 kA for the same time

interval of 3.5 s caused a sizable increase in the

induced stress distribution in the proposed 20-MJ

losses indicating the coil was probably going

TPFS coil has been completed. Various mechanical

normal. At a awing of ~5 kA the increased boiloff

models representing coil construction have been

from the dewar caused by the energy transfer

considered. Several analytical models represent-

In previous stesdy-state,

ing stress diatributiona within the coil have been

nonpulsed tests, the coil had operated at 12.5 kA

applied to the mechanical models and the less

accurate discarded. Stress magnitudes and spatial

Using Rotating Machinery. A used 500-kW

variations have been determined. Peak magnitude

motor generator set, consisting of a synchronous

are on the order of 20 MPa.

motor and a commutated dc generator was acquired.

7- and 9-T Industrial TPFS Coil Design

The general characteristic of the dc machine make

testing

it suitable for reversing the current in the

prototype TPFS coil are planned to be done in

existing 300-kJ METS COil.

house. However, an assessment will be made of the

operated with the dc generator as a motor, and

time and cost involved for having an industrial

preliminary measurements of its characteristics

supplier build such a prototype coil. To this end

The set cannot conveniently be uncoupled so

a contract was placed with Westinghouse Electric

the synchronous motor will serve as a flywheel.

industrial study of

Further measurements will be made before operation

the

still

coils

“swung”

and was

Increasing

principally

has been

l%e

set

POWER SUPPLIES FOR TPFS

c. BIPOIAR TPFS COIL SIMULATION TESTS WITH THE

Commutated DC Machine Feasibility Tests.
UsinE Conventional Power SupDlies. The

pulsed power supply. This machine, rated at 2 A

LASL and Westinghouse 300-kJ energy storage coils

was capable of supplying 20 A into a direct short

have been tested to determine how fast the coil

circuit at the terminals wit}.no apparent damage.

current could be swung from a positive to a

These teats showed the feasibility of this mode of

A 0.19-kW (0.25-hp) dc motor was operated as a

operation of commutated dc machines.

2. TPFS, 20-MJ Homopolar.

Parameters supplied for the studies set the

a. Westlnghouae. The design of a homopolar

peak current and voltage at about 3 MA and 1 kV,

machine suitable for reversing current in the

20-MJ TPFS coil was completed by Westinghouse. It

‘lhe stored energy in the

is based on the HETS 1O-MJ work done previously.

ohmic-heating coils was estimated to be 500 MJ and

This machine uses superconducting field coils as

change to be 34 Vs.

2-s

the

It differs by having a six-

turn drum-type rotor. ‘lherotor has a solid iron

the plasma resistive losses. The

core, which permits the increased energy storage.

the burn phase.

The machine has a terminal voltage of 2 kV, a

3-MA Circuit Breaker for the TNS Doublet.

current of 62.5 kA and a hslf -cycle

A conceptual design of a 3-MA breaker having a

reversal) time of one second.

recovery voltage of several kV was

b. University of Texas at Austin. The

Current interruption is achieved by a set of low-

Energy Storage Group at The University of Texas at

impedance circuit breakers, each with a capacity

Austin began work on a 20-MJ homopolar machine as

as shown in Fig. XV-1.

an alternative to the design of Westinghouse. The

hammer” effect will be used to actuate or open the

machine will use superconducting field coils but

Transformer oil will be used both as the

an attempt is being made to simplify the overall

actuating fluid and the arc-quenching fluid. A con-

mechanical construction details.

tinuous flow of fluid will pass through the hollow

Traction Motora. It was determined that

electrodes prior to actuation.

traction motors can serve as a current-reversing

open under the influence of the Inertial force OE

power supply for the TPFS 20-MJ coil.

the oil column (water hammer) following closure of

motors are series dc commutated machines used to

It is anticipated that

drive diesel-electric locomotives.

loads on mechanical components will

the HETS machine did.

obtain 50 kA at 2.51 kV.

Several separate but

E. TNS STUDIES

attractive.

doublet,

the TNS

Rebuilt

tested

20 MJ.

store

with

100 M

20 Vs

shock

switch.

amount,

supplies

Traction

(current-

respectively,

As a current-

remainder powers

the total flux

loading and 8 Vs

ohmic-heating phase.

pilot and main valves.

Bmu—mmc”

capability,

motors

‘lhe

cost

the

interrelated studies

address

the

for

plaema

startup

of this

inductive

completed.

The “water

The switch will

1.19

reversing power supply 30 to 50 motors will be

connected in a series-parallel arrangement to

In this configuration

they will have a capacitance of 6.35 F and will

traction

1.1 to 1.5 cents per J, and thus are economically

‘IWO units will be purchased and

determine over- current

capability for parallel operation and suitability

for operation in the separately excited mode.

switching problems, and to consider the nature of

the plasma shutdown at the end of the burn cycle.

Fig. xv-1. Hydraulically actuated oil circuit breaker for 100-kA intermittent duty and approximately 2-kA, 500-V interrupting capacity.

were undertaken in conjunction with the Center for

Electromechanics of The University of Texas at

Austin in support of the tokamak TNS design group

of General Atomic at San Diego, California.

primary emphasis of the work has been to examine

the feasibility of a homopolar machine as an

SCALE1.Z.3

energy transfer and storage element in the circuit

developed

the

medium;

Shutdown Problems in LarRe Tokamaks. The

two types of shutdown considered were the normal

shutdown at the end Of the burn phase

manageable in such a configuration.

concluded that the moving brush feature waa not

the water-hammer application, the design is being

useful for controlling the burn period.

considered for patenting.

showed that some of the same principles in arc

results and details of reported.1,2

any

other

500 MJ.

breaker

for use

standard

3-MA bus.

homopolar

disc-type

extinction were

current transfer

transmission project.

to collect the current.

cents per J.

differential

functions.

magnetic

current

within

Austin

Texas

flux

for

2 s

The arc voltage drop in oil is higher than in

shutdown) and a shutdown caused by a disruptive

instability (hard shutdown).

particularly useful for parallel operation and

a cursory literature eearch was undertaken and

methods for controlling the thermal wall loading

counterpulse circuit is proposed consisting of a

were investigated. Since shutdown computer codes

capacitor bank for each of the parallel breaker

units and a saturable reactor core enclosing

startup codes and shutdown codes were

investigated, along with program changes needed to

500-MJ Homopolar for General Atomic.

change a startup code to a shutdown code.

some

major

a. Machine Design. The

machine,

General Atomic Company (GA), was completed.

machine used two counter-rotating discs to store

coils

literature

Superconducting field coils supplied the

was carried out.

main field and copper-graphite brushes were used

plasma-wall interface can be quite

For example, material evaporated

can form a virtual limiter or shield

It was shown in this study that the coat of

such a machine would be in the range of 1 to 2

protecting the wall from major damage.

A parallel study of a 500-MJ

gradients which occur during the interaction can

copper-iron homopolar done by The University of

produce currents whose associated magnetic field

showed

also helps shield the wall.

copper-iron machines will be equivalent to that of

Combination of Blip and Vertical Field (VF)

superconducting machines.

Coil Mounted Outaide the Toroidal Field (TF) Coils

report and an oral presentation were given to GA.l

in Tokamaks With and Without Iron. Since a large

Circuit Analysis of a TNS Doublet Relative

amount of energy is needed if the plasma

to a Variable Capacitance Homopolar Machine.

initiated by inserting a “blip” resistor into the

homopolar machine functioning as a capacitor in an

LC circuit was studied where the current reversed

effort was made to reduce the energy by using a

from negative

aet of unbiased blip coils.

continuing to increase for the 28-s burn period.

designed to generate zero field in the plaama

The differential equation for the series LCR

region and supply the voltage pulse.

was integrated in current for two given flux

coupling from the VP coil circuits, controlled

circuit with a variable brush homopolar machine

functions describing the variable position of the

required

brushes. To answer the question whether there is

correct VP currents. It was concluded, therefore,

any continuous flux function that can provide for

that the blip coils and their driver circuits may

a monotonically increasing current over the period

best be combined with those for the VP circuits.

that includes current reversal and burn, the same

to distribute

equation was

current among the individual turns forming the

not

are

lhe

for

the

and,

hard

hence,

400-kV

design

induced

between

erosion.

voltages

European

undertaken

Because of

widespread,

complicated.

interactions

from the wall

European oil

occur at the

shutdown, the

equilibrium-field

homopolar machine.

A literature survey

bias - current -carrying

A preliminary written

to positive before

supplies are

integrated

virtually

possible

vertical

excluded

energy

It is

power

study

blip

this

that

cost

the

was

the

and

are

the

The

high

large

(soft

For a

‘Ihermal

problems

plasma-wall

differences

first - wall

ohmic-heating

Phenomena which

For a soft shutdown

the calculations were

the plasma region is

supplied by magnetic

The blip coils are

to maintain

Since most

any given

poloidal

field

the

ohmic-heating coil, an

given monotonically rising

field coil so

From

from

and

view

pulse.

operating

operation are

the operating range.

the poloidal-flux path.

energy requirements.

completed.

upgrade

Supply

300 MJ

power

peak

The

and

field

generated by the current in the coil.

initiation by inducing a pulsed voltage in the

incorporated were fiber-glass laminated capacitor

plasma region may thus be obtained by modulating

racks, optical coupling for meter signals and

the current in the equilibrium field coils. In the

high-speed trigger signals, pneumatic coupling for

TNS case, it is expected that a flux swing of

slower control eignals, and a fiber-glass laminated

40 Vs over the range -2 T…2 T may be obtained

the charging equipment.

without varying the plasma major or minor radii.

The great advantage from a practical point of

also

these

of having the VP-and-blip coils mounted

upgraded facility is shown in Fig. XV-2.

outside the TF coils is bought at the expense of a

reliability

large increase in field energy, unless effective

attention during design and construction of the

use is made of iron to reduce the reluctance of

facilities. Safety to the operator is enhanced by

the VP-and-blip field, the VP coils are used as

In addition to providing

and

operation,

drivers for reversing the flux corresponding to

separation and electrical isolation of control

-2 T at the start and >2 T at the end of the

equipment from high-voltage equipment.

constant-current

is enhanced by reduction of electrical

conducting bias coil is wound on the center post

pickup through the use of single-point grounding

iron to provide the bias field of -2 T. Analysis

and optical and pneumatic coupling of controls.

complicated by

The dual facility was constructed to allow

presence at varying levels of saturation because

independent operation of the two facilities with a

the inductance matrix varies significantly over

30-kA, 60-kV capability or combined operation with

a 60-14, 60-kV capability.

the

and

the

Some

ility

Safety

super-

iron’s

Plasma

control

include

possible

original

automatic

interlocks

otherwise.

facility was

balcony to hold

development program.

accomplished

time-

using

about

Some

for

3 T

and

The

design

Reliab-

special

physical

devices,

received

grounding

improvements

improvements.

retrofitted where

The facilities were

have been investigated over the last year.

work was done to specify the power supply and

OH-coil requirements of a TNS reference design

generated by ORNL. The ohmic-heating coil of this

reference design has a peak stored energy of about

half-biased operation. Other possible designs for

TNS would be smaller with correspondingly smaller

OH coils operating at higher fields. The effect

requirements

dependent inductances caused by the change of

plasma current distribution in flux conserving

tokamaks was investigated. Generally the effects

are small and tend to reduce the ohmic-heating

essentially duplicating the original facility. In

Fig. XV-2.

general, the same component design was used unless

Upgraded HVDC test facility.

safety and reliability considerations suggested

F. TPFS SWITCH DEVELOPMENT AND HVDC TEST FACILITY

Interrupter Facility Upgrade. The

upgrade of the HVDC Interrupter Test Facility was

was

General TNS Studies. Problems associated

successfully tested to these limits and can now be

with the ohmic-heating circuits of the TNS tokamak

used to provide new test data for the switch

‘1 1S3

‘2

Simplified circuit diagram for khe HmC facility.

Fig. XV-3.

diagram of each of the interrupter facilities.

Figure XV-4 showa the current and voltage wave

forma and contact separation, S, of the teat

shows

/-~

xV-3

Figure

~: c1

‘4 r-’—l

Typical interrupter test.

are

test

shunt

shown

‘1

circuit

sequence

capacitor

simplified

an LC circuit.

Imax” must be

peak recovery voltage.

facilities.

varied are

withstand.

sequence.

the bases

recovery

voltage,

together

Second ,

achieve

in R1.

energy

point

1 to

tied

One

the

are

the

switch, typically a vacuum interrupter, VI.

the ‘closing times

awitchea in the test facility: S1, S2, S3, and S4.

begins

capacitor bank Cl and counterpulae capacitor bank

C2 to predetermined voltagea of npposite polarity

while switches S1 through S4 are open.

current through storage coil L1 and teat switch VI

is initiated by closing switch S1. L1 and Cl form

When the peak current, Imax, la

reached, switch S2 closes automatically to crowbar

cl.

exponentially decaying LR circuit involving Ll and

the total loop resistance.

slowly enough to allow the test switch VI to open

while its current is still a large percentage of

To achieve interruption the switch current

driven through zero after the swftch

contacts have opened sufficiently to withstand the

switch S3 to connect the counterpulse bank acrosa

switch VI and saturable reactor L2.

creates an LC circuit which includes C2 and the

stray inductance around the counterpulse loop.

The current in the interrupter ia driven rapidly

toward zero until L2 comes out of saturation to

hold the current near zero. After interruption L1

and C2 form an oscillating LC circuit with the

energy stored in LI being transferred to C2,

charging it to a peak voltage, Vmax. ‘Ihiavoltage

is also the peak recovery voltage the switch must

After the energy is transferred from

C2 back into Ll, switch S4 is closed to dlsaipate

Some of the test parameter

This

teat and

event

methods

reactance, and repetition rate.

facilities is shown in Fig.

allows independent charging and triggering of the

storage and counterpulse capacitor banka of both

Only two connections are required.

First, the ground plates are tied together at

of the two saturable reactora

arrangement, both facilities feed current to the

Also

storage

charging

The teat

Closing S3

‘lhia createa

of the varioua

‘he current decays

l’hia is done by closing

common system ground.

counterpulae current,

paralleling the

completes the

This method

which can

saturable

timing,

xv-5.

point

With

this

test

for

Fig. XV-4.

current and voltage waveforms from an

switch independently with the total test current

5 to 7 smaller modules in parallel. The modules

being the sum of the two individual test currents.

and the techniques used to parallel them would, in

Interruption is achieved when the switch current

the same as those already

ia driven to zero and held there by the combined

the duplicate

action of the counterpulse banks and saturable

capacitors and most other components for the 2-MJ

system are already available.

Parallel System Operation. System A and

such as the storage coils, have been designed and

its duplicate, system B, were operated together up

The capacitor racks are

The operating mode was first employed

layout

in which the two systems were interconnected only

modifications are being studied.

at the interrupter under test as in Fig. XV-5. No

Existing coil modules used in the switch test

sort were

facility were analyzed to determine the cause of

systems were then connected at the tops of the

cracktng in the epoxy potting.

and Lm,

applied to tbe design of new coils for use with

facilities directly. l%is simulates more ideally

The new modules will be

the whole interrupting system which might be, for

of similar construction to the existing double-

example, applied to an ohmic-heating system.

this second mode interactions were seen between

complete magnetic force restraint system to allow

the two counterpulse systems, which showed up as a

operation at currents up to 60 kA, which is twice

energy

the rating of the original design.

Modifications of the systems have been undertaken

Saturable Reactors. Tests were performed

to reduce these interactions.

2-NJ System. Circuit

on silicon steel, ferrite, and Permslloy saturable

the

started on the 2-NJ system, which is to be used to

interruption of their flux rating, tape thickness,

test interrupter systems above 100 kA.

Interruption was found to be

was made to construct the system by

independent of all of these variables within wide

reactors of both facilities.

any

fast

“ringing”

to 60 kA.

problems of

saturable reactors, LU

Circuit diagram test facility.

$-’

for

The

has

most

been

them.

being

System

built.

design

between

spiral pancakes, but will

reactors

developed

encountered.

respects, be

to parallel bOth

are being purchased.

the 2-MJ capacitor bank.

and air-gap size.

the bottle; the

interrupters

12.7-cm-diem

conducted.

assembling

A decision

determine

limits.

14 kA.

Tests

other

with

was

The

—

arc

and

The

upon

effect

facility

facilities.

Some components,

This analysis was

incorporate a more

conventional

Conventional

10-cm-diam

improved

field

were

also

Axial Field Vacuum Interrupters Two

special 17.8-cm-diam interrupters were purchased

from Westinghouse for use with axial magnetic

fields. One, PRI, had the field coils internal to

other, PRII, needed external

coils. PRI waa tested to 37 kA at 38 kV without

failure. ‘his was the limit of the test facility.

Higher current tests were postponed until the

duplicate facility became operational.

PRII was tested at various fields to 29 kA.

At low fields the, maximum interruptible current

At the highest field used, about

0.05 T, the extrapolated maximum current would be

well in excess of the system’s limit.

sizes

axial

performance

Fig. XV-5. parallel operatiQn of the HVDC

interrupter performance

about 50% in a field of 0.1 T.

interrupters was not aignfficantly improved. The

20-kA

30 kA.

opening.

homopolar

a series of - 10 arcs.

chattering of

interruption;

1/2 v,

than

the

(f)

the

17.8-cm-diam interrupters were not investigated

Fast-Actuator Studies. There are impor-

pending availability of the duplicate facility.

tant advantages to be gained if vacuum inter-

Parallel Operation of Vacuum Interruptere.

rupters can be opened rapidly.

Two parallel-connected 17.8-cm-dlam interrupters

provides the possibility of opening under no-

were tested for maximum interruptible current. A

conditions thus

current-sharing transformer was sometimes used.

erosion rates of the contacts due to arcing. The

Without the transformer a resistor as large as

1 mfl placed in series with each interrupter was

mechanical

achieve

unable to force current sharing, leading to a

1.6 mm/me,which is relatively slow.

failure to interrupt at a total current well below

An actuator which uses repulsion coils was

With the transformer, current sharing was

obtained and tested. At rated conditions opening

usually very good and interruption at 30 kA, the

velocities of 8 mm/ms were achieved.

system limit, was successful.

unit which utilizes a combination of repulsion

flux rating of the transformer was inadequate, the

coils and pneumatic systems ia on order and will

current sharing was imperfect, and interruption

be tested early in 1978.

wae not successful. With the availability of the

Gas Breakers for Tokamak OhmieHeating

duplicate facility, a larger transformer will be

Duty. The current interrupting capacity of air

expected rating of about 45 kA.

application in tokamak ohmic-heating circuits.

used to test the interrupters to their maximum

Persistent Currents in Interrupters.

electrodes of the vacuum interrupters prior to

generator

breakers

waa

suitability

Facility modifications were made to tie in the

for their large current interrupting

Investigations were made of contact

Interactions With Other Laboratories snd

heating, melting, welding, popping, and the effect

With Industry. Tests were performed on saturable

of all of these on interruption. An actuator was

reactors for Westinghouse Electric Corporation and

modified to allow higher contact-closing forces,

for the Oak Ridge National Laboratory.

which reduces the contact resistance and most of

involved

the above undesirable effects. It was found (a)

ohmic-heating circuit of Ormak Upgrade and needed

that under most circumstances, i.e., whenever the

help with the design of the interrupter circuit.

voltage acroas the closed interrupter waa greater

performed

contacting

commercially available interrupters of GE (Model

electrodes melt; (b) if the electrodes are not

PV07) and ITT-Jennings (Model RP-728A).

immediately opened, such melting leads to welding;

terrupters failed to perform as well as expected.

the electrodes are opened before

The GE sample developed a vacuum leak; the ITT

current is reduced, welding does not occur; (d)

interrupter could not interrupt large currents.

interruption is not affected by the melting until

New Interrupters have been procured for retests.

12t equals or exceeds 4 x 109 A2S; (e) “popping,”

extensive series of tests was performed

contacts caused by

re-evaluate the erosion of

magnetic repulsion of opposing electrodes, can

interrupter.

cause welding problems but does not interfere with

their actuator and with

very

supervision with respect to the method of mounting

forces (J36kg to 227 kg)will be needed if currents

and actuating the interrupter. The results were

of 50 to 100 kA are to be carried by one device;

the same as those of the earlier tests, i.e.,

and (g) if welding occurs and the weld is broken

erosion is much more rapid than for other similar

mechanically, the roughness produced interferes

with interruption until it can be smoothed away by

and

SF6

and

the

used

blast

current

preheat

capacity

breakers

attention

routinely

circuits.

Particular

conventional

Occasionally the

contact-closing

performed with

interrupters.

entities

with GE

special

points

large

Tests

Pv08

were

the

for

was

for

paid

systems

reviewed

A second

generator

actuation

ohmic-heating

Fast actuation

opening velocities

resulting in minimal

designing

Both in-

largest

These

their

teats

their

lheae

were

with

the

A continuous dialog with Ebasco-Grumtnanhas

The matrix reaiativity at

been held with regard to tests of vacuum inter-

approximately one twentieth that of

Special components have

Theoretical

been ordered for these tests, but they hsve not

indicated that the energy losses would be similar

been started becauae of changea in the plans of

mixed-matrix

dynamic

should

almost

assembly

complete.

Ebasco-Grumman.

rupters for use on TFTR.

operational early in 1978.

multifilament NbTi

CTR-9 personnel.

than 2.4 x 105 A/cm2 at 3 T.

0.36 mm,

5500 A.

tested

wires

Other

were

0.51

mm,

and

the

All

and

the

haa

with

line

been

293 K.

copper

14 T/a.

tested,

4.O K was

equivalent

(ICC) coil

to perform

were made on

experimentally ,

motor-generators.

0.05 wt% Iii matrix.

significantly higher.

superconductor ordered.

unequivocal concluaiona.

individual stranda.

300-kJ coil cable.

groups of wires.

broken wires.

the normal

impossible

diameter.

measured

(rather

obvious

within

cable.

‘l’he

than

made

part

were

5 T.

wae

not

the

G. CRITICAL-CURRENT STUDIES AND LOSS MEASUREMENTS

programmatic

Dipole Magnet. The new 102-mm-bore, 6-T

preliminary work was begun to convert the loss

dipole magnet for the high-current test facility

‘Ihe winding

aPParatus oPerating parameters tO a maximum field of 7 T with a bipolar sweep on the order of

finished, the magnet fully aesembled, the transfer

A 7-T, 60-kJ coil was designed and the

line and cooling tubea installed, the eddy current

‘he method of swinging

shield mounted, the shunta connected, and

the coil has not yet been decided.

magnet suspended from the dewar lid. The magnet

under way to evaluate the possibility of using

completed,

Initial investigations indi-

cated that capacitors would be too expensive.

Critical- current measurement

samplea of the conductor from the atart of the

300-kJ, 10-kA METS STORAGE COILS

winding, the break between the two halves of the

Diagnostic Tests for the ICC Coil.

dipole, and at the end of the winding.

an effort to determine the source or sources of

samples met the specifications.

the failure of the Intermegnetics General Corp.

Critical-Current Measurement Below 1 kA.

to specifications, two

Short- sample critical-current measurements were

series of tests were run, one on the 300-kJ coil

made on a series of intermediate resistivity,

and the second on a small test coil made from the

evaluation of the losses of the conductors. me

On the 300-kJ coil the terminations were

critical-current densities at 3 T were 1.7, 2.0,

unsoldered, the 319 strands all separated and

and 2.2 x 105 A/cm2,respectively,for wires of 0.25

extensive testing done on individual wires and

comparison, TGC mixed-matrix wire, for which this

arose was that there were many interwire shorts

was to be a possible substitute, had a jc of more

miscellaneous wires

and not turn to turn in a single strand winding;

internal in a multiatrand, fully transposed cable

including some wire for the Q-10 SMES program.

Because

that

Critical-Current Measurements Above 1 kA.

automatically the source of the problem. Some of

Both samples of the cables for the HETS test coil

the tests did ahow, however, that a resistively

‘l’hemoat damaging fact which

approximately

extrapolation of the data to 6.4 T indicated that

voltage drop would be established across the short

both cables met their specifications of 2800 A and

if there also existed high-resistance contacte at

the termlnatimn and, in addition, if there were

Additional measurements were made of SMES

The broken wires were verified; it

conductors by Q-10 personnel with assistance of

was impossible, however, to determine termination

Loss Measurementa. Hysteretic and pulsed

The variety of test results

contact resistances at the terminations of the

but

had

were

losses

changes,

analysia

Tests are

conductors,

inductively) induced

shorts were

totally

shorta

reach

were

any

the

energy loss measurements were made on a series of

intermediate resistivity matrix wiree manufactured

by MCA. The wires had 156 NbTi filamenta in a Cu

A small magnet was

these items are in a highly developed atate and

excess 1O-W cabled conductor used in the original

are being designed for final hardware fabrication.

IGC 300-kJ cOil. The purpose of the coil was to

METS Prototype Coil. A

investigate, In particular, the affect that no

25-kA superconducting coil ia being designed and

potting might have had on the 300-kJ coil.

built by Westinghouse Electric Corporation.

coil reached a maximum current of T 10 kA, about a

coil is similar to the highly aucceaaful 300-kJ

factor of three higher than the 300-kJ coil.

coil made by Weatinghouae.

waa designed, however, to go to T 18 kA.

shorted, nonbroken, low terminal resistance wires

to the cable did not degrade the coil performance.

It was assumed that the coil did not reach design

tested with a 0.7-ma discharge time and a 60-kV

Subsequently the coil will be tested in a slower 1- to 2-s bipolar field and energy transfer to simulate tokamak ohmic-heating

current because of wire motion since thfs coil waa

performance. The conceptual engineering design on

Unfortunately no definitive tests

the coil ia complete and detailed engineering has

Mixed-matrix copper, copper-nickel NbTi

superconductor has been ordered from IGC for the

were run to establish this conclusively.

Enerw Loss Measurements. The

loss measurement data for the four 300-kJ coils

methods.

not potted.

measurement apparatus.

one-component ayatem

The fast-pulsed

METS SYSTEM

systems,

storage

tested.

capable

system

The leases were measured

the

lhe

coil.

energy

Adding

started.

following

voltage rise.

constructed from the

superconducting, the

a multicontact switch.

semigalling switch.

assembled and

diameter pin.

cylindrical

one-module

bushings

After

488MJ

delivering

complete

amount of information for ongoing inductive energy

hava been analyzed and reported.3 Reference 3

CryoRenic Disconnect. After

summarizes the resulta and describes the leas

several different materials it was decided that

the 50-50 Pb-Sn alloy was the best suited for the

by both electrical and calorimetric methods to

contact surfacea. Plating or coating of the Pb-Sn

verify the reaulta with two independent measuring

alloy onto annealed OFHC copper seemed to offer

Comparison was also made with short

advantages.

sample measurements taken with the astatically

wound pick-up coil method.

low resistance at cryogenic temperature and does not

very

copper

have the abrupt transition to the normal atate aa

I. ONE-COMPONENT MODULE SUPERCONDUCTING PROTOTYPE

do the superconductors. The Pb-Sn alloy coating

has a very low contact resistance, especially fn

energy storage program was initiated to provide a

copper forces current to distribute more evenly in

superconducting inductive

the semigalling mode, and the resistance of the

theta-pinch experiment. An optimization study led

selecting the materials,

to a modular design of superconducting coils, each

effort was spent on the method of constructing a

to deliver 400 kJ of energy to create a 5.5-T

The testing included ways of

field on the fusion plasma. To bring the program

putting on a uniform coat of the Pb-Sn alloy,

to a logical conclusion in view of the de-emphaais

methods of controlling the galling action of the

of theta-pinch research and to obtain a maximum

switch without seriously impairing the operating

haa

not

‘l’he

2.5-T,

400-kJ,

Although

evaluating

The coil ia to be

This switch carried a

the main

15.9-mm-

The one-component system to be aasembled for

connected by

the 400-kJ coil test will use a hybrid external

All parta are made of copper,

switch to isolate the coil from the charging power

coated with Pb-Sn alloy. The two bushings are the

supply, a cryogenic disconnect to withstand the

contacts and remain in a fixed position; the pin

60 kV occurring during energy transfer, and a

the movable element and forms the brfdge

helium-vapor-cooled coaxial lead to conduct the

between the two contacta by sliding through one

current during the energy transfer. All three of

bushing into the other.

lifetime, methods of stitch connection to the superconducting coil, and switch construction that

would minimize any motion of the connecting leada.

A switch waa constructed conaiating of two

It was able to carry a

current of over 5000 A at a resistance of 15 nfl.

No degradation was observed after 500 cycles of

iS under way

switch

External Breaker. To

cryogenic disconnects to open under zero current

conditions, the current in the series charging

current of 1.5 kA with a total resistance of less

after 1500 cycles of operation.

Tests on the above switch revealed the higher

resistance portions of the switch and calculation

were done to optimize the deeign to reduce the

resistance in these placee. The final design uses

the cups and a piece of

multifilament superconductor soldered into a hole

length of each

with

superconducting wire in each finger).

final design, the superconductors from each cup

will pass through the mounting plate and attach

closely to the superconducting cable from the

per

cup

the

Work

coil.

similar

than 50 n

construct

operation.

(12 fingers

running through

three-pin version

capability of 25 kA.

thinner fingers on

constructed and tested.

To meet

charging

external

placed

these

the

’

the

METS

with

this

allow

finger

In the

current

switch was

design and

a 1.27-mm-dfam

High- current switch.

of the external breaker.

erosion, shown in Ffg.

interrupter, and

Several different

preliminary data

The external

breaker will

begin soon.

resistance.

pneumatic

required.

surface

bypass

switch

before

bypass

though

vacuum

switch

also

the

was

dewar.

the hybrid

‘lhe contact

completed.

collected.

Recovery

would

The

were

interrupted.

Engineering drawings of

maintenance

Fig. XV-6.

component

consist of a vacuum interrupter and bypass switch

voltages were as high as 51 kV.

connected in parallel to form a hybrid switch.

Erosion of the copper contacts in the bypass

A full-scale prototype of the bypass switch

investigated.

has been built and tested. Figure XV-6 shows the

XV-7, after 100 transfer

bypass switch and actuator in the open position.

operations at 26 kA Indicatea that 103 to 104

A 6-!4, 600-v switch by ITE Imperial was modified

operations could probably be accomplished by the

to withstand high voltages by using nonmetallic

structural members and immersing the unit in oil.

New electrodes with a silver-tungsten

The switch was successfully hi-potted to 90 kV.

instead of copper have been ordered.

Tests show that the bypass switch can handle the

These should provide 5 to 10 times better erosion

current requirements because

increased heat transfer from the contacta due to

Testing has shown that the hybrid vacuum

the oil. The opening time from contact break to

interrupter-bypass switch meets the requirements

full open is about 30 ms. A 1.03-MPa

actuator is used to operate the switch.

final bypass switch have been

The crucial transfer and interruption steps

Construction of the complete external breaker will

breaker must carry the five-minute charging ramp

switch worked about as expected. About 30 kA was

to 26 kA, interrupt it, and withstand up to 60 kV.

carried by the bypass switch, transferred to the

of the complete external breaker were teated in

the interrupter test facility.

loop will be interrupted by a circuit breaker

modes of operation were investigated and extensive

requirements the

Fig. XV-7. Contact erosion on high current component of the hybrid switch.

Corp.

refrigeration

during the year.

dewar were achieved.

computer control.

Specifications

testing.

testing

400-kJ

dewar.

for

the

The

J. HOMOPOLAR MACHINE DEVELOPMENT PROGRAM

1O-MJ HETS Machine. The detailed design

of the 1O-MJ, 30-ms HETS homopolar machine was

completed by Westinghouse and preliminary copies

of the final report issued. A program was begun

and pursued at Westinghouse to test brushes under

conditions of speed and electrical loading similar

to those found in the HETS machine.

‘WO magnetic characteristics were derived for

optimizing the magnetic circuit of the machine.

One is the linkage flux vs the cross-sectional

area of the flux return and the other the total

flux vs the coil current density.

terminated due to lack of

funding and a relevant application.

HETS Load Coil. As part of the design of

the HETS 30-ms machine test facility, a high-Q

1.5-MA test coil was designed. The coil was to be

capable of operation at 300 K and 80 K.

the

formed.

demonstrate

accomplished

of the program was

Discussions were held

fully epoxy potted.

K. MISCELLANEOUS

Number of Turns

Characteristic

Inductance, VH

Time Constant

Inefficiency

manufacture

(No Bus), %

(No Bus), S

Resistance

Laboratory

load coil

funding.

general

(No Bus),

switch

LASL.

were

and

lJfi

Square,

potential

lhe remainder

program

funds

10.5

8.28

24.2

were

made

was

5.81

0.7

8.28

0.7

700-W, 4.5-K Refrigeration System.

insulated, No. 7 aluminum conductors were to be

Startup and successful operation of the Helix

formed into Roebel bara of which the coil was

aystern was

Tooling and model bars were made

Refrigeration rates of 880 W in

construction

the experimental dewar and 924 W in the 1000-gal

manufacturers. General characteristics and linear

250 t/hr was measured in the 1000-gal storage

‘l’heprogram was terminated due to loss of

As a liquefier, a rate of

dimensions for the coil are listed in Table XV-I.

A GHe collection system was installed along

the outside west wall of SM-253 to accommodate

boiloff from the 3800-! and 500-L storage dewars

Helium Conservation. A

and the LN2 supply line was extended from SM-253

established to implement helium conservation at

to SM-322 to facilitate automation of the helium

overhead

recovery-purification system.

available to purchase capital equipment items such

Micro Nova computer components were ordered

for conversion of the refrigerator from manual to

Load Coils. A aet of high.Q

TABLE KV-I

modules wound of Litz wire was designed for use in

1O-MJ LOAD COIL CHARACTERISTICS AND DIMENSIONS

0.060-s Swing 2.0-s Swing

coil

coils are

their

composed. A contract for the construction was let

to Elms Engineering Co.

with the manufacturer and minor modifications made

to ease construction and improve performance.

refrigerator.

conserving an important natural resource, the

program will result in cost savings to those users

of liquid helium having recovery systems.

SCEPTRE. Several persons from other LASL

Divisions indicated their interest to use the

circuit analysia code SCEPTRE for problems that

SCEPTRE has been made

available in public utility files on LASL”S NOS

complement the existing AFWL reports describing

guide

as dewars and recovery compressors—$l5k in 1977

Recovery systems will be

problems of designing poloidal field coils for

established at lsrge liquid helium user sites

tokamaks was organized by, coordinated by, and

within the Lsborstory. The recovered helium will

Rspresentatives of all major US

then be transported to CTR-9 where it will be

laboratories

purified and reliquefied using the Helix Corp.,

Proceedings were prepared and submitted to DOE for

and

4.5-K

users

systems

1000-w,

the code.

and $50k in

NET2 is unable to solve.

H. Vogel, P. Thullen, and D. Weldon, al., “TNS Doublet Tokamak Ohmtc-Heating supply Study,” Los Alamos Scientific Laboratory report LA-7053-MS (November 1977).

H. Vogel, “Current control by a Homopolar Machine with Moving Brushes,” Los Alamos Scientific Laboratory report LA-7090-MS (January 1978).

J. D. G. Lindsay and D. M. Weldon, “Loss Measurements in Superconducting Magnetic Energy Storage Coils,” Los A.lamesScientific Laboratory report LA-6790-MS (May 1977).

was

written

tokamak

addition

research

REFERENCES

distribution.

held at LASL.

TPFS Workshop. A

workshop

the

attended.

et Power

R.A. Krakowski, R.L. Miller, R.L. Hagenson, R.W. Moses, G.E. Gryczkowski,*

A.S. Tai,** G.E. Cort, R. Bartholasew, J.W. Barnes, J.O. Jacobson, J.H. Pendergrass, K.E. Cox, D.J. Dudziak, P.D. Soran, D.W. Muir, F.W. Clinard, Jr., J.M. Bunch, G.F. Hurley, and J.G. Hoffman

field) to temperatures of -l keV;

and has

INTRODUCTION

Fast-Liner Reactor

the basia of this RFPR

support of all of

also commenced.

constraints.

summarized.

would be

Scoping

hybrid

have

The focus of all systems and design studies

plasma is subsequently compressed adiabatically to

of magnetically confined fusion concepts has been

ignition temperatures (-5 keV), and a burn cycle

on alternative or exploratory concepts.

occurs along a plasma radius/temperature

design effort on the Reversed-Field Pinch Reactor

jectory determined primarily by the dynamics of an

(RFPR) concept has been completed, and is based

energetic, high-beta plasma.

upon a moderately pulsed mode

The loss of plasma particlea and/or energy

design, a second design

would make the achievement of this burn cycle on a

study haa comnenced that investigates both the

reasonable time scale, while

physics and technology associated with

ing an energy-balance or Lawson-like criterion,

extended-burn operating mode.

technologically difficult for an open-ended 1inear

(FLR) and

Firat, the end-loss problem

Fusion (LMF) concepts have emphasized physics and

ameliorated by use of extremely

energy-balance constraints

Second,

although crucial technological issues for both the

ends could be replaced by solid or gaseous end-

FLR and LMF concepts have been addreased. A

plugs, which would maintain

of toroidal hi-cusp confinement (Tormac) waa begun

(i.e.,contain the particles), but would

focused on elucidating viable physics

energy too rapidly, unless the plasma column were

operating points with total power, energy balance,

shown later, this approach

and first-wall loading being enforced as major

leads to the following condition on

studies

(axial) field B(T) and plasma length P.(m) for

(fusion/fission)

ideal ignition (i.e.,no loss of the 3.5-MeV alpha

chemical hydrogen) applications of

particles): B2!?,= 1.5(10~ T2m.

The results of

classical axial electron

studies of plasma systems analyses performed in

conduction and predicts “acceptably” short

the above - mentioned reactor

reasonable power levels and lengths) devices only

for high fields and plasma densities; the instan-

studies are also reported.

neutronics studies and insulator

flux at

ultimately eatabliahes an upper

allcwable plasma density,4 not

LINEAR THETA-PINCH REACTOR (LTPR) STUDIES

magnet design and energy transfer/storage (ET5)

The heating and (radial) confinement princi-

constraints which accompany high fields.

ples for the Linear Theta-Pinch Reactor

Although ongoing research may ultimately lead

similar to those envisaged for

to solutions to these technology problems and

toroidal Reference Theta-Pinch Reactor1’2

permit uae of the favorable B2!. scaling, the

were it not for the rapid loss of plasma energy

approach adopted for the purposes of

from the open ends of the former scheme. Hence, a

pre-ionized DT gas is heated by a fast

wherein the end-loss particles and energy emanating

implosion (-1 kV/cm azimuthal or ‘D” electric

are directed by a

Staff Member

Wisiting Michigan, Ann Arbor, Michigan, *visiting Staff Member Institute of Technology, Lausanne, Switzerland.

curvature conduit to a second, parallel device.

The plasma column within the REP

necessarily be in “toroidal” equilibrium, it may

contact and in part be supported by a structural

for

XVI.

study

col~ns

power,

fusion

Linear

longer,

(>11)~).

A major

Magnetic

device.

of operation.

SYSTEMS STUDIES

also very long.

Studies of both the

from the Swiss Federal

frcan the University of

condition is based on

taneous radiation heat

Finally, progress in

fusion power

invokes the

generalized

from a LTPR

reentrant

(.-l-pa)

(thermo-

research

Synfuel

general

(LTPR)

(RTPR)

the

and

can

tra-

long

plasma

conduct

the confining

this preheated

the free-streaming

simultaneously meet-

the plasma pressure

This Lawson-like

small-radius-of-

the first wall

endplug (REP)

this study

to mention

region may

limit on

scheme,

(i.e.,

not

the

wall and, therefore, will be

intermittent toroidal equilibrium may

tablished in the REP region that is similar but

not as effective as that envisaged

end loss would

periodic plasma dump to the end-region

not

is assumed to

majority of the 3.5-MeV alpha particles, which

vital to the successful operation of

within

iwition will be subsequently addressed.

The REP idea is sufficiently recent to pre-

clude detailed study of either the end-region

operating-point

search, or a detailed LTPR design based

the applicable RTPR

have been used for this study,and the design point

should be viewed as interim. The major

this study is to develop a

technologically credible design point which

lengths of at most a few hundred meters,

lating power fractions below 0.25, and plasma

densities (ccsnpression fields) and

times that are technologically feasible.

The interim LTPR design configuration con- sists of two parallel, linear sections (f,= 150m)

two semicircular reentrant

sections of 5-m radius, giving a

The reactor is segmented into

approximately 160 independent modules, nominally

2 m in length. As depicted in Fig. XVI-L, each

module contains blanket, magnetic coil,and coolant

system components. The modules are suspended in a

vacuum/shielding trough such that pumping and

fueling access to the plasma chamber is obtained

through the gaps between adjacent modules .

modules used in the REP sections of

differ somewhat fran this design in order

expected,

remova1 requirements, although details of

the LTPR design have

Many

wall .

physics

The REP

resolved.

physics, a

RTPR;2 the

This issue has

particle confinement

field transport losses.

devices like the LTPR.

joined at the ends by

ccanpensate for the

length of 331.4 m.

composed of -100

aspect of

resolved.

dependent of the magnet

Each 2-m-long

coil

leads

column

are

the

es-

the

yet

for

Plasma

goal

been

S. Electrical

gives

fully

first

‘-l I

alpha-

context

ignited

confine

recircu-

thereon.

parameter

correspond

The issue of

to odlalwfic compressl~ \

subject to cross-

design resultsl’2

the other hand an

self-consistent and

/’

the essential plasma

radius first wall.

total reactor

the LTPR may

features of

confinement

increased

systems.

ba1ante

blanket

inboard

module

These

code,

heat

this

yet

The

gas

The

thick, and is structurally and electrically in-

azimuthal segments, ia 0.35 m

diem,

/+’

6.Llflilum

coolont ducts

(first-

( Im first-wall 0.4m thick)

(LTPR) module.

I 4.ACIA3,,C

room-temperature

(Table XVI-III),

the first-wall

is driven by

generated by

dynamically

engineering

envisaged.

‘orl,!ihe

design

that

and

the

The

Fig. XVI-1.

Cutaway view of a wall) Linear Theta-Pinch Reactor This module is similar to that envisaged Toroidal Reference Theta-Pinch Reactor (RTPR).

2-m long, 0.5-m-radius

edges of these blanket segments provide the 0.5-m-

Outboard of the blanket is the

implosion heating coil, which

capacitive energy store.

adiabatic compression coil, driven by homopolar

motor generators, provides the subsequent, slower

compression to ignition with a field of 8 T.

Table XVI-I through Table XVI-III sumaarize

and

the LTPR as presently

plasma, energy-balance,

design parameters have been

three-particle (ions, electrons, alpha particles),

time-dependent, one-dimensional (axial) pressure-

lDRBURN~

computes the LTPR energy balance

all plasma parameters, as well

thermal response. On the basis of the detailed

computation model, the merits of exchanging axial

heat-conduction losses for cross-field transport

losses in the REP are quantified.

‘w’ I ’“”e’s’""

Litlium-cooled

blonkst

Coil

IHC(m)

ACC(m)(c)

Radius of reentrant sections, R(m)

Length of straight sections, g(m)

TABLE XVI-I INTERIM LTPR DESIGN FEATURES

Shock Magnetic field, BSH(T)

Maximum compression field, BO(T)

Ccxspressionfield rise time, TR(ms)

Caspression field flat-top time, T~(ms) 300.0

Fusion neutron energy worth, EN(MeV/n)(a) 15.66

ETS efficiency (for TR = 30 ins),~ETS

Thermal conversion efficiency, ~~

Auxiliary pcwer fraction, Am

f(b)

thin

very

This

Based on neutronic calculations.for the 0.35-m- blanket thick blanket. captures 86% of the fusion neutron energy, the remaining 14% or 11.84 MJ/m being removed from the implosion and adiabatic compression coils aa low-grade (- 300 K) heat; this compares to 7.36 M.J/mlost as low-grade heat as a result the of joule losses. blanket to .-0.40 m will lead to a recovery of the fusion energy (i.e., the nearly 100% of to the thermal output 11.84 M.J/mwould add the increased coil radii transfer and loss optimization of this trade-off is

increaae the ETS

= 96.29 M.J/m),

thickening

Although

energy. k presently in progress.

(a)

IH coil thickness, Ab

AC coil thickness, Ab

First-wall radius, b(m)

Blanket thickness, Ah(m)

of ignition in LMF

On the basis of

alpha-particle

(Sec. XVI

XVI-2 ;

:Jgld

(b)

(c)

significantly limits the range of

found.

decreased thermal conduction relative

classical value ia also ahown on Fig. XVI-2.

that in operating a LTPR, a

veraus

startup

followed

acceptable burn, and the required

(final) B21 valuea are generally above

c. REVERSED-FIELD PINCH RSACTOR (RFPR) STWIES

Dynamic plasma and energy-balance models were

used to perform a parametric ayatems analyais of a

wide range of RFPR burn cycles using the engineer-

the

Major system dimensions (first-wall radiua, coil

sizes and configurations, blanket location and

thickness), initial plasma conditions (temperature,

ratio, filling pressure), initial toroidal

field, magnitude and wave

toroidal current,and the total burn

varied over a wide parameter range. The magnitude

of the first-wall temperature rise and associated

thensal stresses, both of which were alao computed

aa a function of time, and the total plasma beta

BZ2

5.0

8.0

classical

0.35

0.02

0.0s

0.04

0.50

0.40

0.95

30.0

1.18

biaa

150.0

dynamic

trajectory

ergonically

according to

reflects the

should be noted

compression must

given on Fig. XVI-2.

ing Q-value QE as

ignition is to be

constraints.

treated

= 9.29

versus

which

were

termined by detailed MHD

B2L

The

the

and

the

The

were

that

time

form

those

object

insure

theories

influence

adiabatic

temperature

B2?. where

function.7’8

slowing-down

stability computations.

startup, compres-

few aeconda by

of TB and TR

R(-loo Iss),

and used as

the trapped

for other

homopolar

computed

time T~

motorl

major

The

in-

Fraction of total electrical power used for lighting, auxiliary control, etc.).

(pumps,

plant

needs

continually monitored

toroidal current 12 was

Tqtal volume to outer radius of ACC m’h.

creaaed sinusoidally with a rise time T

is consistent with

generator supply for the poloidal coil system.

the lDRBURN design calcu-

12 was held constant (crowbarred) for a

lation, alpha-particle confinement has emerged as

(- 1 s) and then sinusoidally decreased with

a crucial issue. Consequently, an

analytic study

systems has been conducted,

fall time equal to TR. The Value

were selected to assure, conservatively, that the

using a simplified alpha-particle thermalization

poloidal beta Be at the quench time (~ 2TR + TB)

model described in Sec. H.2. This analytic model

did not exceed a critical value (@Oc = 0.5) de-

integrates all reactivities across an appropriate

axial temperature profile and defines an ignition

quenched, expanded plasma and

condition (alpha-particle heating equals the sum

field were asaumed to rest againat the first wall

of radiation and axial conduction leases) in the

and to cool over a period of a

form of B21. Treating the alpha-particle trapp-

clasj3ical cross-field conduction.

ing efficiency fa analytically but conaiatently

first-wall thermal history reflects all stages of

H.2), results in the

this proposed RFPR bum cycle:

temperature ignition relationship depicted in Fig.

sion, bum, expanaion, and quench.

thermalization

fusion reactor models, notably tokamak systems,

THERMONUCLEAR BURN CONDITIONS FOR i = 150-m> REP LINEAR

TIiETA-pINcHREACTOR (LTPR) INTERIM DESIGN pOINT

Initial DT filling pressure (297 K), PA(mtorr)

Implosion-heating electric field, EO(kV/cm)

Implosion-heating voltage, V(kV)

Normalized plasma radius at shock, XSH

Imploded electron temperature, TeSH(keV)

Imploded ion temperature, TiS (keV)

Imploded ion density, niSH(m )

Beta of imploded plasma, ~.

Postimplosion plasma energy, WINTo(MJ/m)

implosion-heating energy, WSH(MJ/m)

Implosion-heating efficiency, WINTo/WSK

Implosion magnetic field, BSH(T)

well understood.

Achieving MRD

length.

1.45(10)21

10.0

-9

1.0

(a)

0.8

8.0

0.9

1.18

0.29

2.58

0.11

1.33

30.0

0.70

222.1

300.0

TABLE XVI-II

Burn Conditions

Initial Burn Conditions

Pressure balance

the Bessel

turbulence

2.7(10)21

described

homopolar

205.0

0.12

0.27

0.99

0.06

24.4

0.81

26.1

5.2

5.0

and

the

(c)

-3,

(b)

0.9

1.26(10)22

2.75(10)21

full voltage

function

startup,

result.

during

model.

Adiabatic compression field rise and fall times, TR(ms)

Adiabatic compression field flat-top time, T~(ms)

Adiabatic compression field, BO(T)

Plasma radius at start of burn, ai(m)

Plasma radius at end of burn, a~(m)

Ion temperature at start of burn,[kTi]i(keV)

Ion temperature at end of burn, [kTi]f(keV)

Electron temperature at start of burn,[kTe]i(keV)

Electron temperature at end of burn, [kTe]f(keV)

Alpha-particle temperature at end of burn,[kTa]f(keV)

Alpha-particle confinement efficiency, ~

Ion density at start of burn [ni]i(m-3)

Ion density at end of burn [ni]f(m

Beta at start of burn, ~i

Beta at end of burn, %f

Fractional fuel burnup, fB

Integrated Lawson parameter, (nT)L(s/m3)

the physical processes associated with both

similar to that exhibited by tokamaks may

turbulent startup and the quench phases are not

assumed

however, and the approximate field profiles are

startup phase seems unlikely,

generator is applied at

stability during the initial

(a) On the basis of pressure balance and magnetic flux conservation the plasma The

beta is allowed to vary consistently throughout the LTPR burn. dynamic burn code lDRBURN reduces the fusion power density accordingly.

(b) Estimated by the ratio of linear section length to

total circumferential

lGNITION:fa

TO(kcV)

XVI-2.

TABLE XVI-111

INTERIM LTPR ENBRGY BALANCE RESULTS

. .

Cycle time (for IW = 2 MW/m2), T=(e)

Implosion-heating energy, WSH(MJ/m)

Postimplosion plasma energy, WINTO(MJ/m)

Postquench plasma energy, WINTF(MJ/m)

Bremsstrahlung energy, WBR(MJ/m)

Reentrant section thermal loss, W

Fusion neutron energy, WN(MJ.fm)

Fusion alpha energy, Wa(MJ/m)

Direct conversion energy, WDC(MJ/m)

Direct conversion efficiency,~~~)

ACC transport losses, WT(MJ/m)

Blanket eddy current losses, WED(MJ/m)

Auxiliary energy needs, WA(MJ/m)

Maximum stored energy,WB(MJ/m)

Total recoverable thermal energy, Wn(MJ/m)

Total recirculating energy, Wc(MJ/m)

Total electrical output, WET(MJ/m)

Net electrical output, WE(MJ/m)

Recirculating power fraction, E

Thermal power density (MWt/m3)(d)

LTPR length, k(m)

ETS losses, WETS(MJ/m)

Engineering Q-value, QE

Total thermal power (MWt)

Net plant efficiency llp=

Gross power output (MWe)

Net power output (MWe)

toroidal field coil

initiation

current in

defined by

(4-5 kV)

?*;’

10s

Fig.

and

4.8

(c)

~G m ”

0.21

7.98

6.08

0.77

0.49

7.36

0.50

2.47

5.92

0.29

2.58

10s ~ o

(b)‘L

9.44

30.54

38.52

96.29

121.6

19.01

84.59

12.12

V/m-mtorr

(MJ/m)(a) 1.86

c : FREE STREAMING

the poloidal beta Be

at a minimum bum

established by

for a 50-50% DT

simultaneously

the high-beta

(PFC), and

initiation

injection,

either be

0.1 s.

Above a

by gas

must

0.85

0.32

the

The

MND

the

~(1+)

(b) Based on a thin 0.35-m blanket that is 86% efficient for capturing the fusion neutron energy.

plasma against the confining magnetic field.

(d) Based on total volume enclosed by magnets.

;

i li

B=o,e

the

above

P,= PBR+PC

pressure

pressures

line-radiation power

(by decreasing the

and ia increased

fueI mixture

(cold-gas

maintain

‘G~J

the

to the poloidal field coil

reaction rate reduced by other means

the toroidal current rises in

(TFC)

Qc throughout the burn cycle. Early quench is used

burnup)

activated to generate the bias field Bzo.

fuel mixture

tokamak startup code’” mode 1a

plasma current), whereas 90%-10% DT

phaae until the electron temperature has reached

controls ~e by burnout. Generally, the use of a

0.1 keV. A minimum electric field of 5 V/m-mtorr,

10% tritium fraction achieves a near msxim~

the resistive voltage produced by

time. Future studies will model

the plasma, is required for plasma

anomalous cross-field transport mechaniama as a

at initial filling pressures greater

means to control the poloidal beta.

Dependence on the center temperature of ignition conditions in the presence of classical axial (electron) thermal conduction to cold endpIug.s for a Linear Magnetic Fusion (LFM) device expressed as B% consistent, claasical treatment of the fraction ~ of alpha particles trapped in the plasma column of length 9.(m).

with

than 0.5mtorr and a plasma radi.ua oflm.

electron runaway limit corresponds to a maximum

requires

0.I-O.2 mTorr. To minimize

low temperatures (0.01 keV),

typically begins at 0.5-1 mtorr

injection to full density during

-0.1-s toroidal current rise above 0.05 keV.

pIasma temperature of O.L keV

zero-dimension plasma model cmnputea the RFPR burn

(crowbarred) at its maximum value for a burn time

and ohmic heating rapidly leads to plasma

‘B’ ignition. The vigorous burn associated with

ignited plasma (-250 MWt/m3 of plasma) drives

~a) Based on classical cross-field diffusion in

dynamica. The toroidaI current is held constant

the reentrant sections.

fl~re2tj3:2

“prematurely” quenched or

The burn-cycle parametric studies varied the

first-wall radiua rw, fixed the blanket thick-

ness at 0.4 m, situated the room-temperature TFC

cases are summarized

times for the more

operating mode

immediately outside the blanket with a

compared to the 6-a values found for the specific

of 0.15 (rl.25)m,and outside the TFC waa

90%-10% case given in Table XVI-IV.

the 0.6- (1.0-)m-thickroom-temperature PFC system

The RFPR-I design uses a

All joule losses accounted for a copper conductor

ciency of 40% to produce 750 MWe

filling fraction of 0.7 and parallel plate lead

plant efficiency of 30-33%. The design summarized

losses. k example of composite results for the

herein ia baaed primarily on the 50%-50% DT

50%-50% DT “premature-quench” mode

mixture case, although both cases are similar.

is depicted in Fig. XVI-3 in the form of r w

plan view of the reactor system containing key

the maximum toroidal current density for a given

subsystems is given in Fig. XVI-4.

Similar relationships have been

heat-removal and thermal conversion systems are

similar to those described in

operation which produced Q= values

transport the heat

the reactor to the single turbinelgenerator set.

mentally achieved toroidal current density

temperature of 820 K

20 MA/m2 and limiting the first-wall temperature

sodium exit temperature of 810 K

rise to 100-150 K, the interim design points for

heat exchangers. The remainder of the plant

similar to that envisaged for the Liquid-Metal

value of QE.

generated for the 90%-10%DT

those given on Fig. XVI-3.

for the 50%-50% (9O%-1O%)DT

2”0

Selecting an experi- 11 of

FUEL

RFPR

casea.

located

50%-50”J%D-T

thickness

30% above

fuel mixture

of operation

A lithium exit

“burn-out” mode

tional 50%-50% DT

in Table XVI-IV. Bum

1250-MWt coolant loops

the 50%-50% and 90%-10% DT

heat transport system with a

I—_“-lml

0=.5 10

WN.E-.

MAXIMUM AVERAGE TOROIDAL CURRENT DENSITY,‘jz(MA/m2)

Fig. XVI-3.

first-wall radiua on the average Dependence of fixed (maximum) toroidal current density for a recirculating engineering Q-value Pinch power fraction) for the dashed curves represent Reactor (RFPR-1). lines of constant first-wall surface temperature rise.

Reversed-Field

(inverse

The

,ROWIING

Plu4PS

Two

The

fuel

frmn

conven-

Ref. 12.

produces a

are 1 s as

at an overall

from the Li-Na

lithium-sodium-ste-

thermal cycle effi-

750 MWe

(net)

I—R

SAU.ERY

Plan and elevation views Reversed-Field Pinch Reactor (RFPR-1).

of a

Fig. XVI-4.

90%-10%DT

5.82(10)20

2.64(10)21

fractional burnup

first-wall radius (m)

minimum plaama compression

ion density at ignition (1/m3)

peak toroidal plasma current (MA)

energy transfer/storage

efficiency

initial plasma energy (MJ/m)

time-i tegrated Lawson parameter

(s/m )

initial tritium atom fraction

average toroidal current de sity at

minimum compression (MA/m )

ambient D-T filling pressure (mtorr)

‘T

‘A

‘i

‘if

‘2 J z

‘kH

Symbol

(dT/dt)Q

‘ETS o ‘INT

w c1’

‘INT

YOND

~TS

?3R

%Ht.1

‘N

ion temperature prior to quench

(keV)

electron temperature prior to

quench (keV)

post-burn plasma reactivity

(keV/s)

thermal conversion efficiency

0.4

0.5

2.0

0.4

2.0

0.95

18.5

16.3

14.4

20.0

40.0

0.086

0.109

Definition

50%-50% DT

8.41(10)20

8.52(10)20

TABLE XVI-IV SUMMARY OF TYPICAL RFPR PARAMETERS

345.4

358.7

342.7

137.1

279.6

0.036

13.3

6.23

0.67

37.1

17.9

18.9

5.38

1.72

54.7

5.11

34.7

cyclotron energy into first wall

(non-reflecting) (MJ/m)

fusion neutron energy (18 MeV/n)

total alpha-particle energy

plasma ohmic-heating energy

thermal conduction energy

(MJ/m)

direct-conversion energy (MJ/m)

energy requirements of toroidal

field system (M.J/m)

energy requirements of poloidal

field system (NJ/m)

total magnetic energy in ETS

system (MJ/m)

transport losses in coil and

leads @J/m)

eddy current losses in coil, leads, and blanket (MJ/m) field in plasma at end of burn

(MJ/m)

ETS transfer losses

(l-nETs)WBo (MJ/m)

total recoverable thermal energy

bremsstrahlung energy (MJ/m)

final plasma energy (NJ/m)

gross electric energy qW

(MJ/m)

2.0

0.4

30.0

15.0

1.35

0.1

0.184

23.8

15.3

-0.41

0.4

0.95

0.058

11.0

19.9

16.9

317.7

62.1

5.85

0.24

21.3

7.3

205.0

212.3

11.5

0.79

19.6

10.6

385.5

154.2

TABLR XVI-IV cent’d.

‘B

‘E

I w

Arc

’( c

Symbol

lift the 2.1-m module.

resulting in a torus of

presently in operation.

components.

seal .

‘THT

The

engineering Q-value

total circulating energy (M.J/m)

recirculating power fraction ~/QE

net electric energy

overall p ant efficiency ~TH(l-F)

(1-’) ’!T ‘W’mz

major radius (m)

cycle time (s)

burn time (s)

14~~m~utron

total thermal power (MWt)

system gross electric power (MWe)

system circulating electric

power (MWe)

net system electric power (MWe)

blanket thickness (m)

poloidal (toroidal) coil

thickness (m)

thermal power densit

the blanket (wt/my)in ‘he thermal power density averaged

over total volume.encompassed by the PFC (MWt/m~)

90%-10%DT

0.4

1.87

6.44

10.5

10.0

0.33

0.19

28.6

5.39

2.00

1.14

8.65

10.0

0.30

0.25

33.8

4.06

125.6

103.3

Definition

50%-50%DT

wall loading

0.60 (0.15)

module with the vacuum

toroidal sections,

is similar to the

fossil-fuel plants

to form a vacuum

circular sections.

in a single -turn

reactor floor.

also acts as

are located

each module.

a bus which

the iron-

energized

machines.

the iron

is used

reactor

radius.

major

7.18

6.63

2.71

1.66

755 K

are

the

two

The

the

The

The blanket/first-wall

20-30 homopolar

maximum steam temperature which

design temperatures of many

shown in Fig. XVI-4,

composed of thirty 2.1-m-long

core structure. Outside the modules

the 4-m-long poloidal field coils and

a vacuum pump and

on rails and

the reactor

generators

homopolar

leads is

lithium

the 15

semi-

used

and

the

isometric view of a RFPR segment, given in Fig.

floor level, and separate connections are made

10-m

coolant pipes encircle the torus at

XVI-5, shows the placement of

first wall

vacuum wall where the flanges at the ends of

The TFC is connected in parallel to

also encircles the machine

center-grounded

2.1-m-long modules are welded

machines. The PFC consists of wedge-shaped seg-

The lithium-cooled blanket and TFC

ments, where 1/15 of the conductors in each wedge

encircled by a structural ring which

are energized at each 4-m-long section, resulting

transformer coil when

assembly and vacuum ducting for each module

sections are connected. Each of the PEG

mechanically separate and supported by

then connected to two center-grounded homopolar

Referring to Fig. XVI-5,

associated ducting is provided for each 2.1-m-long

tunnel located under

inlet and outlet

Fast-Breeder Reactor. The plant operates at

cores, which are divided into top and bottom

energize the coils would be mounted

1.00 (0.25)

2-m MODULEN $IFTINSUK

require considerable development.

reactor subsystems such as

The technology associated with the large number of

remote manipulations of massive equipment will

Fig. XVI-5. Isometric view of 2-m- long Reverse-Field pinch Reactor (RFPR-1) modules (six modules are shown) and associated poloidal field coil assemblies and iron-core pieces.

could easily be decoupled

permanent lead structure.

is envisaged for the banks

the additional difficulty

rays and slow neutrons.

7Li), Nb-lZr

apectrum

blanket

RFPR

tritium

handling/

lating power.

bias field Bzo.

Other required

burn period, TB.

toroidal field. At

similar arrangement

from the massive but

of vacuum pumps with

of remote maintenance.

air-core system is most

hibitively high today.

The efficiency of

The balance of

can interrupt

mates13 are

represents a

ability of

enriched)

resolved.

addressed

-300-400

the TFC

overall

1- to

hard-

which

(95%

that

with

the

and

system12

blanket in capturing neutron energy is to be near

100% resulting in a total thermal energy

18 MeV/neutron. Division of the blanket into 30

azimuthal sections reduces the eddy-current losses.

plant appears “standard” except for the Na/Li

primary heat exchangers and Li pumps.

The 0.4-m-thick lithium-cooled (99% enriched

for tritium production, leading to an

tritium breeding ratio of 1.10, and a combination

of iron and boron carbide for attenuation of gamma

cryogenic facility for the superconducting homo-

polar energy stores, plant control system, and

associated maintenance machinery will be

in future RFPR design studies.

the PFC is a basic LC or tank circuit wherein the

lead for neutron multiplication in the

The electrical circuit for both

regions, stagnant

time -TR12 connects the homopolar machine

full speed and voltage to the TFC.

rises to a maximum, producing the initial toroidal

The plasma is preionized, and

the analagous switch on the PFC system is closed.

The toroidal current 12 rises to a maximum Value

as the toroidal field continues to “ring” induc-

tively, ultimately yielding the desired reversed

time TR

is closed, and the current Iz and reverse field

aremaintained at a near constant value during the

The homopolar machine remains

at rest during this period.

the “crowbar” switch is opened, and the respective

TFC and PFC currents are transferred back

homopolar machine, which motors up to 100% speed

and voltage (neglecting losses) at which

brushes are lifted. The above cycle is repeated

at the desired pulse rate. Losses incurred during

the transfer are made up by bringing the homopolar

generators up to full speed, using plant recircu-

Solid-state switching may be

TFC and PFC systems. The rating of coszsercially

4-6 kV, and many such switches must be paralleled

to give the desired current rating.

this kind of switching element, however, is pro-

The present price esti-

104 $/MVA with

on the order of 5(10)4 MVA.

alleviated by using hydraulic

-100 kA

switches are needed; the reli-

these switches must

An iron core ideally couples the PFC current

and plasma current, although

920(690)wb for the 50%-50%(90%-10%)

mixture cases results in 230(170)m2 of

large economic penalty.

attractive when

current is changed syu.metrically from minus

plus during startup and returned to a minus value

during the quench period. In this case the coil

to the

time the

The current

used for both

Following the burn

a “crowbar” switch

This problem may

(oil) breakers14

The cost of

requirements

ultimately

iron and

the PFC

Using

total

fuel

cost

and

may

homopolar machine acts as a capacitor. Closure of

current cannot be extracted until after the quench

a switch (homopolar brushes) in the TFC circuit at

(2-3 s) and should be left in the coil; using the

2-nzs-thickniobium structural walls uses stagnant

< $2000. This approach is favorable considering

processing, He-gas coolant system for the coils,

available solid-state devices is about 200 kA

transfer capacitor to

endplugs; thermal conduction to the liner and

current from plus to minus

endplugs would be inhibited by an azimuthal field

store equal to half

caae.

the ohmic losses in the

(-2.5-m-radius spherical containment vessel), high

The FLR would provide a relatively small

power.

conducting coils.

mode of operation.

magnetic field would

mode greatly increases

8yUIUetriCally swing the

homopolar generator as

would require an energy

needed by the iron-core

plasma particle pressure

(ZlfJ24 m-3) plasma (r

fusion energy during a

t---l

directly by

the liner

coil when compared to an iron-core system, because

of current flow during the reactor dwell time.

(-5-10 MWt/m3)

These considerations have lead to future studies

anticipated.

of iron-core and air-core systems using super-

problems andfor uncertainties

addition

preparation end injection into the liner, ii)

significantly the RFPR burn cycle, ongoing design

and/or

atudiea are baaed upon economic/cost

destroyed components, iii) blast

system optima. Generally, future RFPR

pulsed energy switching and transfer to the liner

studies will focus onto this long-pulse (-30 s)

(reversible recovery of pulsed energy is

envisaged), and v) rapid liner and leads replace-

ment . These problem areaa are being asaessed and

FAST-LINER REACTOR (FLR) ST~IES

The fast-liner concept was subject to prelim-

Plasma-Liner Simulation. The FLR

inary examination as a potential source of

nuclear burn code described in Section H.3 waa

This approach to plasma heating and

Thermal conduction and

confinement is shown schematically in Fig. XVI-6.

field diffusion are computed for

An axial current is driven through a thin

direction using an implicit Lagrangian method.

metal cylindrical shell to produce a

simple analytical model approximates axial thermal

nal magnetic field. For reactor applications the

conduction to the liner endplugs.

speed (.w104m/s) onto a warm

tially reduces thermal loasea when the global beta

implode the liner at high

magnetic field embedded in the plasma substan-

=0.2m,

Bremastrahlung radiation

compressed plasma would reach thermonuclear tem-

losses are included, and an

perature in _20 pa and release 2-4 GJ of D-T

is employed.

_2-pe

to the driving circuit is not yet considered.

that

fusion

source.

density

oriented

altering

(- 311nl)

recycling

economical

conceptual

quantified.

0.20-0.30 is

recirculating

1arge exter-

overall power

Operating in this

(250-500eV), dense

embedded in the plaama.

used for this analyais.

‘lo 13 T, where this choice of

compressibility model’6

PLASMA INJECTION AND BIAS- FIH):URRENT

liner parameters.

burn period.

proportional

resistivity.

liner/plasma

!3lo = 0.5.

Fig. XVI-7.

at the wall

!Z=0.2 m).

conditions

is - 0.5.

supported

104 m/6.

material

‘o ‘lo;

treated

= 0.2m,

Two

The

and

The

‘s

-2

o”

the

not

are

Major

power

radial

energy

plasma

thermo-

fraction

engineering

replacement

An , azimuthal

confinement, iv)

‘RRo = 1.7 GJ/m, . s

‘he ‘nitial in

field corresponds to

approximate liner

independent variable,

compressible

‘lo = are

represent

selected

Coupling

initial

models

and

Figure XVI-7 illustrates the computation of

scientific Q values (fusion energy relative to the

initial liner energy) for a variety of plasma and

initial liner velocity,

‘lo’ and the initial liner thickness becomes a variable

liner energy is held constant along each curve

‘n ‘his ‘ay

The liner material is copper without

data in Fig XVI-7

= 1 25 x 1024m-3


CONDUCTOR

Fig. XVI-6.
Schematic diagram of essential elements of a Fast-
Liner Reactor (FLR).

These parameters were

produce a Q greater than 10 for a

liner model, including plasma losses, and

other plasma/liner

would

surface

+?

Field 8 heat diffusion
Bremsstrahlung

1Compressible liner

! 1

I

I

I

Incam ressible liner
No p asma losses

1

I

1

I

4

P

IOJ

INITIAL

impressible liner

'lnce 'XRo
independent of

'lo
proportionality is not

liner can only be

duced, WF%PFTd,

Q WF/W=o"l/vlo.

compression

increases

at peak

clearly

Vlr)r'

ered.

of Q

'lo'

is

in

I

I

I

1

1

1

2

[05

the

vlo(m/s)

Second,

velocity

rlo = 0.2 m

T:= 0.5 keV

30% of Wno

wall in both

high velocities.

tively thick, a

w ~RO =1.68GJ/m

sufficient to raise Q.

j?= O.2m
nmx[.25 x 1024m-3

liner PFan/4 r~fn~~v&gt;

lossless plasma and

the liner of 80%, a

'f 'lo
maximum

= rlo/vlO, it is

(VI = O).* The

followed for

To = 500 eV,

is reduced T

dependence

Fig. XVI-7

estimate

'XRo .._

cases,

gain,

lIVIO

0.24.

lIVIO

pro-

the

the

low

for

'lo

in

'd

By

d

deposited in and near

The compressible liner and lossless plasma

model exhibits three interesting features.

when the liner has a lcw velocity and is rela-

large fraction

consumed in liner compression.

plasma energy is not compensated by

dwell time (i.e.,high burnup) and Q

in the region of maximum Q approximately

ia used to compress the

increasing the average liner density by a

of 2.5. Although the final plasma energy and peak

fusion power are reduced as compared to the in-

compressible case, the increased dwell time is

liner model approaches the incompressible caae at

Plasma losses are considered in the last

example; these include thermal conduction to the

the radial and axial directions,

bremsstrahlung,and magnetic diffusion.

velocities the dwell time is relatively short and

only a small fraction of plasma energy is lost.

Consequently little effect on Q

increases and

reduce the plasma energy at peak compression.

This accounts for the large reduction of Q by

losses at low velocities.

Many physical effects must still be

rated into the FLR plasma model, such as joule

losses in the liner, alpha preesure, and turbulence.

An interim reactor example is taken from Fig.
24 -3
m

= 1.7 G.T/m,

XVI-7:

B~N"L= 13T,

= 1.2 x 104m/s, and Ao=

'lo
Q = 11. For an energy transfer efficiency

thermal-to-electrical energy

conversion efficiency of 40%, and a neutron energy

multiplication of 1.25 the liner Q-value of Q - 11

leada to a plant recirculating power fraction of

2. Blast Containment. Approximately 20% of

the fusion energy from the -2-tIs burn would be

of

'XRo

's 'lo

factor

liner,

First,

At high

results.

the large

is reduced,

losses substantially

Third, the compressible

's
Here, the reduced

blast data are

no = 1.24 x 10

rlo =0.2m,

incorpo-

several

produce

hundred

?=O.2m,

leading

liner

alpha

21mn,

with

the

the

by

to

to

,

LINER VELOCITY,

Fig. XVI-7.
Effects of various loss processes on FLR Q-value
(neutron energyiliner kinetic energy) using single-
fluid MHD model
liner compression.

and analytic approximation for

represented here: the lossless plasma with both

cmnpressible and inccnnpressibleliner.

To understand qualitatively

dependence, the curve for a

incompressible liner in Fig. XVI-7 is first consid-

For this case the kinetic energy of the

transferred to the plasma.

and the initial plasma conditions are

the plasma radius, temper-

ature,and density are also independent

because of thick liner effects.

plasma radius at peak compression

me
by 1% as vlo goes from 10 m/s
a result of a change in the fractional DT burnup
fran 20% to 3%.

to ~oin::as::

particles, and this energy coupled

initial kinetic energy of the liner would

an explosion
kilograms of high explosive;17

equivalent

to

fusion power per meter of

(EN + Es), therefore, is independent of Vlo.

defining a liner "dwell time"

possible to approximate the energy per meter

and to

The

seen

all

&gt;10 m/s. Although Q monotonically decreases as

VT)rii

=!&]~lINNER

OUTER
SURFAC

SURFACE

100

o

G-m.

where R is the shell radius and AR

the containmentrequirements for the FLR.

theoretical analyses to predict

energy W is assumed to be thermalized in an ideal

gas . The circumferential stress

a spherical pressure vessel is computed using a

where E is Young's modulus, V is Poisson's ratio

and or is the radial stress.

agree well with experimental data for vessels with

that have elastically responded

vacuum to detonation equivalent to

In the presence of air at atmo-

spheric pressure, the blast-induced stresses were

about four times the analytic predictions.

The computer program PAD18

model shock spectra resulting from energy releases

of the magnitude expected for the FLR.

one-dimensional Lagrangian

model that computes motion of

transfer and thermal conduction,

however, are not yet considered.

FLR

for cmnputations

W = 1.42 GJ, was initially deposited in a

=

E

is

in

,

]/E

In the simplest containment

R = ().5 m

explosive.

[(l-V)GG-WK

~= Wj4TR2AR

being used with

The corresponding strain

thin-shell approximation

vacuum blast containment.

served as input to PAD

system is based upon a

allowed to yield.

acceptable levels.

of mass M with

exceeded in a

Radiative heat

preliminary

When

The

I

,

. .

in

w.

(1)

(2)

model

3(L
z

the blast

This code

. . ..A,

1 kg of high

that would result

its thickness.

These expressions

is being used to

constant
for
fg &gt;&gt; fv. For

at a frequency of f&lt; M112.

that is independent of AR

vessel oscillates at a

represent M = 25 and

and circumferential

the explosive gas

The maximum vessel

the vessel wall

the vessel was

with energy W.

functions of

Based upon

represents

stress-to-

container.

operating

and the

the gaa

density

ccsnpu-

energy,

can be

sphere

copper

small

time.

blast

solid

point

that

The

not

The

the

is

at

of

M

0246

under the thin-shell

Since the shell

a quiescent

fv=- 475 Hz,

stress

nearly

nearly

XVI-8B

It is

where

zero

and

The

is

of

is

M

the density of

strain is plotted as a

situated at the vessel center;

function of explosive mass M in Fig. XVI-9.

destroyed liner and leads material.

observed that the maximum

blast scaling with experimental data, the radius

values

of the containment sphere equals 2.6 m,

vessel wall thickness equals 0.15 m.

this situation the gas pressure

oscillates about

and modulus for the containment vessel was

ultimately damped to the pressure of quiescent gas

for 304 stainless steel, but

Meanwhile, the vessel moves

the yield stress

as a harmonic oscillator from a condition of

computational result, AR

hoop stress to a maximum stress. The average hoop

appropriately scaled to reduce the

stress will support the pressure of

Results of the PAD

gas of energy W as given above.

tations are shown in Fig. XVI-8, where

oscillates harmonically from zero to a maximum,

pressure at the inner wall

the peak stress is approximately twice the average

t (ins)
Fig. XVI-8.
One-dimensional Lagrangian hydrocode (PAD)18 anal-
ysis of FLR blast containment in a 2.6-m- radius
0.15-m-thick evacuated sphere made
of 304 stain-
less steel. Hydrostatic pressure is given as a
function of
the
vessel, u (t), for explosive energy
(1.42 GJ)
of 25 and
depositionrinto liner and leads massea
200 kg, respectively.
stress,
the inner and outer vessel
u (t), is given at
S:rfaces

time at the inner surface of

Circumferential

Figures XVI-8A and

200 kG, respectively.

frequency of

approximateion. The reverberating gas oscillates

and the response of the spherical

stress at the inner and outer walls are plotted as

___

W/27rR AR

I

M (kg)

Fig. XVI-9.

I

250

This

--T-

I

.___.

500 -

750 -

100or

fz~fv,

stress.

studies of mitigator.

the blast chamber

shots) at 800 K.

1016 iS

15;

of PAD is now being used

fluidized

function

tritium breeding

experimentally that some solids (vermiculite) can

reduce the effect of a blast by a factor of 2 or

I

I

a

a

&

In

If

I 50

the

this

when

0.05

fails

- O.[O d

more.15

than the

~--------=

benefits of

LiA102 could

helium) and a

- 0.15-
E

perhaps includes a

be eaaly recycled.

blast chamber with

3. Current Leads.

mechanism at the wall

'bo.9 Lio.1"
with

to improve on previous

a peak microstrain of

the leads cost could be

It has been shown

interpreted for

can be com-

Or . 20 Mpa

immersing

lithium-

19 as

in

To

fluidizing gaa for heat transfer would be accrued.

Another mitigating concept would

a mixture of gas

lithium-bearing liquid such

This approach hae

simple shock models17

amplification of the blast

believed that an improved and proper model

(mechanical) shock absorption

could greatly reduce the

containment requirements predicted thus far.

driving current of -200 MA at a voltage of

kV. Any structure within 1-2 m of the liner would

be destroyed each shot.

liner must be manufactured from material

A preliminary optimization study was made

coaxial current leads for use in the FLR.

eralized, coaxial lead configuration for use in

the FLR which was subjected toa

figuration, cost and energy optimization.

were assigned to both metal

components, and energy losses resulting from joule

heating of the conductor, inductive energy storage

in the leads,snd mechanical energy transfer to the

conductor by magnetic forces were included in this

optimization program. It was found that 20-30% of

associated with

chanical energy imparted to the conductor and that

generally the coaxial lead structure yielded an

optimized configuration that may be too expensive.

reduce this cost, an alternative lead

arrangement was devised as ahown in Fig. XVI-10.

In this leads configuration, currents flew alter-

nately to and from the lines on adjacent conduc-

tora,and insulation is woven between conductors of

opposite polarity. The impulse mamentun

to this structure is typically 5% of that encoun-

tered for coaxial conductors. The kinetic energy

lost to the interwoven leads structure is negli-

gible. This end other leads concepts are being

studied in an effort to reduce

A

of

is

It

as

or

~0

of

the

and

the

bed

gen-

that

show

that

been

great

added

using

Hence,

studied

effects.

that can

fill the

(probably

similarly,

the leads and

The FLR may require a

and insulator lead

simultaneous con-

imparted

the me-

overall

costs

the

Maximtnncircumferential stress in a
2.6-m -radiua
0.15-m-thick sphere and minimum th~~kness required
The explo-
to limit strain to
sive energy (1.42 GJ) is deposited in a mass M,
of destroyed leads
representing a combined mass
and liner.

1.02 x 10

.

approximation

explosive gas and shell crone into resonance at

as observed in Fig. XVI-8B.

case the maximum stress is 77% higher

value given by the above-mentioned approximation

Based on fatigue data,

liner blast containment,

acceptable for a 10-year life (2.5(10)7

By taking V= 0.29,

and E = 160 GPa, the maximum allowable stress and

minimum thickness as functions of M

puted; this dependence is shown in Fig. XVI-9.

This work provides a preliminary estimate of

the containment requirements for a FLR operating

in an evacuated sphere. The computer program PAD

has been modified to operate in the FLR regime,

and analytic comparisons have checked the accuracy

of PAD. It would be preferable to operate a FLR

with a mitigating and tritium-breeding material

the newly modified version

Two basic concepts are under consideration

for blast mitigation in addition to containment in

a vacuum. First is the possibility of

the liner aasembly in a fluidized bed of

bearing pellets or flakes.

resistivity coefficients

Radiation losses by

level of simplicity.

muthal magnetic

to 10% of

me approximations described above provide a

simple and rapid FLR modeling code. Comparison of

LNRBRN to the detailed hydrodynamic code21 cEAMISA

shows good agreement. Within

physics model, the LNRBRN code provides a good

description of the FLR, will continue to be used

for reactor modeling, and will be

continued evaluation againat the CHAMISA

system as refinements are made.

E. OTHER NON-MAINLINS RRACTOR STODIES

Fusion Burner Studies.

have presented a challenging idea for the

G" ~
operation of a steady-flow reactor that does not

rely on magnetic field shaping or compression for

ignited reactor has been pro-

posed to operate in a steady-state condition, much

concept is very

cold fuel would be injected at the end of a 1inear

solenoid. Somewhere along the machine

is somehow ignited, and approximately one-half of

the alpha particles generated by

travel "upstream" along the magnetic

heat the incoming, cooler plasma.

No flow limitations exist along any section

of the plasma where the net heating is zero.

condition can be insured after the point where

plasma reaches ignition by increased radiation or

reduced alpha-energy deposition. From the point

of view of reactor feasibility, it is of interest

to determine the length of

required to insure that the plasma may

self-heated and sustained at ignition. The

flm conditions are determined by

and momentum.

obtained and used to assess the reactor potential

of a steady-state fusion burner.

depicts the fusion burner of length L, mass

specified inlet and exit

spatially dependent energy releases.

Figure XVI-12 gives the numerically

the exit Mach

Ma. and the steady-state mass

1.

its operation. h

like a gas burner. The

solutions to the energy

mined relationship between

conservation of mass

only by increasing the

end. As the exit Mach

increaaed convective

the zero-gradient

an acceptable

The computer

Braginskii.

FLR plasma

and
20 *

~a~ytic

plasma

model

field

azi-

also

are

the

M,

of

An

--

et
--

code

simple;

Hasegawa,

field and

the plasma

fusion would

subjected to

the limits of the

the ignited region

is increased, the

Figure XVI-11

the laws of

conditions,

flow rate

indeed be

Numerical

equation

supplied

deter-

number

choke-

(y=o)

have

been

This

fl~

and

for

the

of

M

condi ion at the exit

&#125;
number

transport can be

source strength by means

LEADS STRUCTURE FOR FLR
Fig. XVI-10.

interleaved conductor lead
Cross section of
structure. Current alternates to and from the
Magnetic forces
liner in alternate conductors.
react on these conductors radially outward with
transferred to
only 5
coaxial leads.

the momentum

operating costs while maintaining

4. Liner Burn Code (LNRBRN).

program LNRBRN was developed to model

and liner processes. Thermal conduction and

diffusion are computed numerically for the radial

direction using an implicit Lagrangian method

from

simple analytic model approximates axial thermal

conduction to the endplugs of a liner.

field embedded in the

substantially reduces thermal losses when

global beta approaches unity at the wall initially.

bremsstrahlung

considered. At peak cmupression plasma pressures

are high enough to increase the liner density by a

factor of 2 or more. The impulse momentum approx-

imation wasused16 to give ~

liner ccsapressionand dynamics as related to the

plasma. The liner electrical resistivity is not

considered, but eventually will be incorporated.

be maintained, and a steady-state solution without

and/or refueling is not

Figure XVI-12 also shows the dependence of

the output power as measured by Jak(W/m).

expected, increasing Ma.

power to cmnpensate for the increased convective

loss from the hot end. Aa

drops helm.r- 0.2, the major

duction at the cold end, which

flow condition.

effects become important

values, the fusion rate needed

zero-gradient condition at y =

again, showing the minimum Jaf

.-.0.04, no amount

source

steady-state operation is not

external heat addition or DT

'"ok
M.pevo
2-ii~v'

co

Fig. XVI-11.
Schematic diagram used to examine the steady-state
The normalized axial
solenoidal fusion burner.
coordinate system is located at the
(hot) exit
end (y = O) and the total alpha particle, brems-
strahlung, radial diffusion, conduction, and flow
(convection) losses are exoressed Der unit cross-
sectional area ma
as Ja, iBR, JDF; J c, and 'o'
respectively.

For Mao&gt;0.5

vective loss represents the major energy demand on

the alpha-particle source, whereas

conduction loss through the cold end dominates.

For Mach numbers below --0.04 the burnup of D-T

fuel along the device length becomes so great that

the zero-gradient condition at the hot end cannot

JN1

@=T/To

(J17/2+Jdi

increased M$ values.

.01
,.4

t
,45

the

(JBR+JDFlf

con-

possible.

for Mao&lt;0.2

falls belcw

dependent of the

external heat addition

possible to keep the net

the product
flow rate
zero-gradient
The associated
total alpha-partic e
?!
Ja(W/m )
The

Ma- - M!t
the ele~tron-
conduction
classical
the

2.
23 of a

certain parameter

hi-cusp Tormac

included were

steady-state

Kunke125

of 1977.

for the

Tormac

phase

area

The

t
,:6

;7

)

Mf(k.J/SM

on

Fig. XVI-12.
Dependence exit math number Ma
of fusi~ burner length L(m) ana mass
M(kg/s m ) necessary to achieve a
exit condition (no heat input).
fusion pcwer, expressed as
power per unit cross -sectional
times the device length L(m) is also shown.
curves denoted by "y=" indicate the
relationship for the condition when
electron collision time equals the
where
time, which
conduction is expected to break down at
designated axial position y (re: Fig. XVI-11).

criterion

is a

As

is

of

very

The

the

Ma.

low

for

When

more

value.

burnup

fusion

fusion

O must

As Ha.

burner"

demands

increase

refueling.

to maintain

is roughly in-

positive, and a

possible without

loss beccsnes con-

the exit Mach number

the cusp confine-

DOE/DMFE

energy terms in

and energy

important ,

Concepts.

balance,

Studies.

toroidal

Fusion

energy

first

More

Also

the

the

the

The

the

of

Ma. value of 0.1 appears to represent a minimum

power output at To = 10 KeV.

Although the steady-state "fusion

offers possibilities for a stable, very simple

fusion reactor, classical thermal conduction along

the axis of the device leads to large devices that

develop enormous thermal powers. Unless a reduc-

tion in thermal conductivity relative to

classical value amounts to a few orders of magni-

tude, the self-sustained fusion burner will more

than likely remain a large power producer.

Tormac Reactor

systems study of

cosssenced in the last quarter

simple reactor model

balance differs little from that used by Brown and

Evaluation of Alternative

purposes

Emphasis is placed upon the sensitivity of the

Tormac reactor performance to

ment, electron-to-ion temperature ratio and plasma

beta. The algorithm used in this study is more

general and the parameter variations are more

extensive than those presented in Ref. 25.

all major

plasma

ranges.

although some of these terms may be negligible in

however, explicit and self-consistent expressions

Maxwellian D-T plasma was

Although

component (ions), the effect of differences in

electron and ion temperatures on pressure balance

and radiation losses was included by the parameter

i = TelT.. The plasma was

internal structure, but was described by an effec-

tive minor radius rp(m) and major

Ion and electron cusp losses PL(Wt) and

fraction I-fa of the total alpha-particle power

Pa(Wt) lost frmn the plasma (primarily through the

cusps) were assumed to be deposited to the thermal

cycle, as were the neutron (18.9 MeV/fusion) power

bremsstrahlung

and the cyclotron radiation Pm(Wt).

are given for all aspects of

balance, in order that differences in resulta can

be more readily explained. Whenever possible, less

sensitive parameters are chosen to agree with

those used in Ref. 25. The questions of startup,

steady-state operation, and radial profiles are

the subject of ongoing study.

Figure XVI-13 depicts the steady-state Tormac

converted with efficiency rlTE

PET(We) =ll~P~.

the total electrical power

'ceded '0
drive an unspecified energy injector (e.g., neutral

'INJ'?NJ

'lINJ 'aa ass"'"ed '0 con-
tribute the recirculating power needs; the power

p INJ(Wt) ia simply that needed to maintain
Tormac plasma in a thermal steady state. That is,

model.

reactor

PN(wt), the

thermal power,

PTH(wt) = PL + (l-fu)Pa +

I 1+1'lTH

(l-f=)P. v///A

I-zP,

r
4%%%

LPE

of

//

I

""

'N+ PBR+PCY

rw(m) as an effective first-wall radius, rp(m)

which is identical to the plasma energy balance

used in Ref. 25. Hence, the fraction of the total

electrical power generated by

recirculated to assure a

either in terms of an

I/C, or

the net
elec-
Either QE 'r
been parametrically evaluated as a

= PET(l+).

'E

fuction of ion temperature T(keV), with the first-

neutron

constraint.

as an effective plasma radius,and M

blanket energy multiplication relative to

(M= 18.9/14.1 = 1.34 for

by

to

one

the

the

wall

E=(P

pINJ

power

power

trical

(wt)=PL

PBR(wt)

14.1-MeV

radiation

described

The total

electrical

PE(We) have

+PBR+Pm-faPa

that must be

radius R(m).

~wlrl~NJ)IPET

steady-state,

serving as a

was assumed to be

the plasma energy

assumed to have no

beama) at efficiency

steady state is given by

engineering Q-value, QE =

All results are expressed

14.1-MeV fusion neutron

value, ~, used here.

will be addressed.

this simple model.

steady-state

considered),

the cases

~(MW/m2)

rlr.
pw

maj or

, ,.,1-m

and

(3)

~.

I

I

'

= (PN/MY(2T)2rwR

A

i

I

I

I

,-

,

(5)

(4)

the

only

thermal

current

Iw(NW/m2)

Ddsiguating

the Tormac reactor

outer neutron wall

future studies

Additionally,

the closed-

to distin-

the Tormac

equal
,

(A 'R/rp)

was made

to the

the

(6)

the

,

No attempt

guish between the inner vs

current, which for a small aspect ratio

cusp can differ appreciably from the averaged

Reference 23 presents a complete parameter

study of the Tormac reactor concept based upon

reactor performance (e.g.,size and efficiency) to

The sensitivity of

the sheath physics is noted, and

will explore more this aspect.

question of startup, transport within

field plasma, heating, and steady-state operation

PLASMAON

CLOSED FIELD L

5ss h

Schematic diagram
Tormac reactor energy balance.

idealized,

Fig. XVI-13.

[Cb3s=0.6

c~/cp=

I

cp/cg=
A=A*=O.15
77=77*.O.4

F. HYBRID (FUSION-FISSION) E~NOMICS STUDIES

A majority of fusion-fission design studies

blanket studies or specific fusion-driven con-

cepts. The economic performance of a given hybrid

detailed plasma, neutronic, and engineering models

of a specific, rather narrcw class of systems.

order to examine more generally the dependence of

fusion-driver performance (as measured

engineering Q-value, QE) on

(as measured

[CV], and the &

and

parameters (energy cost

capital costs of hybrid and burner reactors C

F'
~tc.) a simple but comprehensive expression has

handling

been derived that relates [CV] and M

constraint.27 Figure XVI-14

schematically the flow of fissile fuel and elec-

trical energy upon which the simple economic model

The return on investment over

periods T and @ are Aand

the hybrid and burner reactors. This costlrevenue

derived relationship between

evaluated parametrically,

constraints imposed upon [CV] vs M

blanketa have been taken into account.

relation between [CV] and M

focused

M,

fue1

*
P'

and C

economic

is based.

generally have

version ratio,

multiplication,

of the hybrid blanket

system, therefore, is

Figure XVI-15

k.-%"wm

~

I

C

~

CF

by

*P

In

2.5

Cf'

the

A*,

the

and

CP'

time

cost

fuel

upon

onto

costs

based

energy

either

depicts

economic

detailed

relatively

generalized

for a given

by the con-

respectively, for

the characteristics

similar27 evaluations,

depending strongly on

of [CV] and p = c /c

parameters indicated.

neutronic analysis

M = 1 point as Q

for the range

for unenriched

have been made.

graphically

neutronic

range, of

[CV]*.

[CVI

The

and

and

and

the

was

the

b)

of

a)

27

E

M

I

M

t$lc;=

cflcp=o.3

For a relatively wide

appear economically

Hence, externally

shown on Fig.

large values

be driven

the pure-

operating

that are

this and

[CV] = O,

[CV] and

p values

point

rate

the

of

to

M

Fig. XVI-15.

Cost-constrained relationship between the fusion-
fission conversion ratio [CV] and energy multipli-
fusion-driven engineering
cation M for a range of
the converter-burner conversion
Q-values Q with
fuel-to-power cost ratio
ratio [cv]~ = 0.6 and a
c /c
fp

= 0.3.

on the basis of

the following conclusions

Fusion-fission systems which must

by external power sources (i.e.,Q = QE M&lt; 1) are

f p"

fusion-fission blanketa subjected

the required

values have not been found, and the high

required appear unrealistic.

driven symbiotes (Q&lt;l, M=l)

curves of constant QE

XVI-15, as expected, converge to the

collapse to this pure-fusion

increases,

Fig. XVI-14.
Schematic representation of fusion-fissionlburner-
converter model used to evaluate cost-constrained
flow of fissile fuel and energy_. Quantities with
(*) superscripts refer to burner-converter para-
meters, whereas quantities without
superscript
refer to the fusion-fission system.

the magnitude of P = cf/

CP
fusion system becomes attractive and preferable.

Generally, for QE?13

Given that R(kg/MWt y)&lt;l at best, aymbiotes (i.e.,

M = 1) are feasible only for [CV] values

illustrates

economically possible only for very

e&gt;,

unacceptable for the assumed conditions.

d)

c)

of

systems

reasonable

becomes high

values of p,

system would be

energy of 2.5 FNJty,

will generate a

R = 0.4 kg/MWt y,

the coat of the

the cost of N =

(cp/c&.3).

P&/Pm

the

by

e)

considerably below unity and for

SYNTHETIC FUEL PRODUCTION BY FUSION POWRR

Unless the burner-converter system has a very low

fusion power for

conversion ratio [CV]* and the fuel cost becomes

generation of hydrogen gas by means

an appreciable fraction of the total energy cost,

temperature (~1500 K) thermochemical cycle and a

A study of the feasibility of utilizing the

and in view of conclusion a), pure symbiotes do

not appear feasible for this combination of neu-

in the last quarter of 1977.

tronic and economic reasons.

For the same values of

rules for this study include a) a

cylindrical blanket geometry will be used

increasing [CV]* or decreasing

the constraints imposed by a specific

required fusion-fission system efficiency QE

driver will not be specified), b) the blanket must

higher and higher values.

[BR] &gt; 1, c)

drivers are economic only if the fissile fuel cost

capacity will be in the range 2-5 GWt, d) no net

and the burner-converters in them-

electricity will be generated, e) a pure dc hybrid

selves are poor generators of fissile fuel.

electrolysis)

other words, for high values of

chemical cycle will be used that has scsaebasis

economic considerations force the

experimental fact. This synfuel study is divided

fusion-fission system to generate more

a) overall energy balance

power in order to remain competitive (a constraint

economics, b) neutronics analysis

imposed throughout this analysis).

ing, delineate low- and high-temperature regions,

Fissile fuel produced by

etc.), c) thermal/mechanical design of blanket,

attractive, according to

and d) (chemical) process design for the

fissile fuel costs expected for the coming years,

if the fuel-to-energy cost ratio

1. Energy Balance and Economics.

low ~ 0.3). The realization of this condition

fied but general fusion/thermochemical hydrogen

depends on technology advancement (high QE

model has been developed to examine the trade-offs

M) and the difficult requirement on capital cost

between tritium breeding (asaumed to occur in a

low-temperature (-300 K)

P

G.

In

to

the

the

(i.e.,

either

forces

(QE&gt;2).

electrical

efficiency

[CV] and M,

fusion-driver

[CV]* and low

low-temperature

into four tasks:

is sufficiently

the fusion-fission

thermochemical cycle.

Low-efficiency fusion

have a breeding ratio

unique features of D-T

the fiasile
28 of

and electricity markets.

P;H/Pm = 2.5 R/

a given ~lanket

low-temperature

low-temperature

fissile fuel

likely value

spread over

accompany-

systeme.

"real"

energy

energy

heat

mix

and

or

of

*

c

depicts this model, which

A

in

and

the

(i.e.,

fusion

thermo-

simpli-

(ex-core)

the plant

generalized

of a high-

(tritium breed-

The general ground

electrolysis coannenced

P3(recirculated)) of

tritium, hydrogen,

of ground rule

tritium-breeding

the econcmnic

(i.e., the

etc.) and

examines

high- to

region

in an

this

the

the

of

of

is

In

Given that R(kg/MWt y)&lt;l and that a kilogram

blanket), electricity production, and production

of fiaaile fuel represents a potential thermal

of high-temperature (.-1500 K) process heat

the fusion-fission system

outer blanket region. Figure XVI-16 schematically

in situ
----
comparable to the energy content of

quantity

of

trade-offs associated with overall system perfor-

fuel produced. For the more

mance (i.e., fusion-driver efficiency l-c, frac-

in situ
----

tion f of fusion energy deposited in the outer

generation equals the "virtual" energy production

of the bred fissile fuel. Since

(1-[cvI*), however, the power pm(~t)

high-temperature blanket versus

*
ratio [BR], thermochemical efficiency n,

ing the production of RP~(kg/y)

market prices associated with

the fusion-fission system is multiplied

amplified by a factor of 2.5 R/(l-[cV]*) when

d) cited above, the parameter k in Fig. XVI-16

this fuel is consumed by a burner-converter system

set to zero (i.e., no electricity production).

with an intrinsic conversion ratio [CV] . In effect

conjunction with specified capital costs indicated

fusion-fission system is

on Fig. XVI-16, the energy "spectrum"

burner-converter

between P4(high-temperature), P5(low-temper-

leaka El,

In view

ature), P*(electrolysis) and

design and ~he match

(-800-900 K)

SYNFUEL ENERGY FLOW DIAGRAM

YDROGEN MARKn

Gav:!%k+c;%"]k'c

'=

&E1-~/T6

VACUUM
Fig. XVI-17.
Schematic diagram of synfuel fusion blanket used
high-
to
temperature process heat generation and tritium
breeding. Generally, only the "multiplier" region
was varied both in composition and dimension.

neutronics

associated

atudy

with

C(*)

C*(*)

Ct(+)

yoyT#

CAPIPIL

T3HAN:C);;G -

ELECTRICITY
GENER. COST

dependence of f on

fusion-driven

varying both

i

K= (C@)~@~

0.001 J

configurations, as

cal geometry with a

obtained only at

specific thermo-

considered, all

ature (- 1500 K)

The effect of

balance task.

the outer,

Figs. XVI-19

thickness

favorable

(1400-1500

ite).

[BR].

the

key

the

3.

K)

.

vOID

D),

/Fe

the energyfeconomics

or Li A102 was

Concepts.

depicted

thermal/

.
'120'

stream.

regions

perform

Figs.

was

The

the

in

of

As

the expense of less fusion energy deposited into

high-temperature

blanket. Relationships similar to that depicted

on Fig. XVI-18 are used by

The power density in the outer

graphite region is used to

mechanical design studies.

Thermochemical Blanket

previously noted all thermochemical blanket models

assume a simple one-dimensional (radial) cylindri-

low-temperature (N800-900 K)

tritium-breeding inner region and a high-temper-

outer region (typically graph-

Either Li(metal),

considered for the tritium breeder, and in some

cases Pb was proposed for a neutron multiplier.

Cooled structural first walls and thermal insula-

tion represent essential elements for all blanket

illustrated schematically in

Although direct neutron heating of

products in the thermochemical cycle

blankets

XVI-19A-D ultimately generate a high-temperature

helium process heat

a

Fig. XVI-16.
Power flew diagram used to model the energy split
for
hydrogen
inner blanket produces tritium and
plant. The
process heat, whereaa
low-temperature (-800
the
high-temperature
blanket
This system interacts
(-1500 K) process heat.
economically with
tritium, energy, and hydrogen
markets as indicated.

K)
produces

thermochemical

outer

"spectrum" to the load line of

chemical cycles is examined.

2. Neutronics. The relationship between

fraction f of the total fusion energy passed

through the low-temperature breeder region to the

high-temperature process-heat blanket region and

the tritium-breeding ratio represents a crucial

parameter for the energy/economics balance task

described above. Figure XVI-17 depicts a simple

neutronica model used to examine numerically the

relationship between f and

the (neutron) multiplier

and composition on the energy EN deposited in

the outer blanket and [BR] was ccmputed fora range

of blanket compositions. Figure XVI-18 gives

[BR] for a more

Pb-multiplier case. The trend indicated on Fig.

XVI-18 is similar for the other blanket systems

studied; increased [BR] can be

I
0.3
[BR], TRITIUM IWEECWG RATIO

I
0.8

I
0.6

I
0.7

I
0.5

Fig. XVI-18.

Interrelationship between the fraction of
the
total fusion neutron energy deposited into the
outer blanket region
high-temperature (-1500 K)
on the tritium breeding ratio
[BR] for a Pb-
Multiplier/lOO%-Li-breeder caae.

might not be a limiting factor, and, assuming the

decomposition to take place in times less than one

minute, a decompose might operate in a "falling

I

I

O

o

"

i

I

I
0.4

% I

5; 0"7

g~:
an
IL*"

:"'''''7';

I
I
0.1 0.2

i
kg
OR 0.3

&gt;P
*hi 0.8-
ax
w=

(energy requirement

not
Bi203.x S03 cycle29,30

GeneraLly, vacuum tritium

materials engineering of

fully established.

neutronic,

fraction

1/3 Bi203.3S03= 1/3 Bi203 + S03

(1200-1500 K)

Figure XVI-20 depicts a preliminary process

flow diagram proposed for the H2S04/Bi203.3S03

decomposition

air, the sulfate decomposition rate

governed by heat

particles. The dominant mode

temperatures above 1000 K

If bismuth sulfate (B~03.3S03)

adequately dispersed, heat

thermochemical cycle. Baaed

results,31~32

I

I

I

I

I

I

1.2

SO

The

the

the

-\\

I
1.1

I
1.0

I
0.9

bed"

mode.

The SO

be used

cles were

should be

sulfate in

radiation.

transfer at

decomposing

experimental

temperature);

oxide (Bi203).

Krikorian,33 for

S03 = S02 + l/2 o

be dropped through a

'resh '2s04
occurs

ities associated with

PLASMA SYSTEMS ANALYSIS

achieved at 1500 K

Specification of

Since detailed

chosen, even

The bismuth

mechanical,

versions of

parameter.

electrical

pressures.

following

specific

barriers

to form

blanket

L12S04J

cannot

energy

until

(7A)

used

to a

and

are

all

H.

a

the

by

to

of

as

on

the

then

zinc

(7D)

(7C)

parti-

sulfate

of heat

transfer

transfer

reported

particles

is thermal

Equilibrium

(750-1200 K)

Pechonakii'a

700 K) would

to ensure that

'1203"3s02
as bismuth

studies. This area

99.8% at moderate

contact with

is returned

electrolyzer

confinementt

physically

to obtain

should be

A major

general

gaseous

system

activ-

con-

the

The

is

fusion energy deposited in the cooled first-wall

bismuth

and inner tritiuim-breeding regions is used to

(100-200Pm in size and preheated to

generate electricity for in-plant use, although a

portion of the lower temperature heat may

'e
particles emerge from the decompose

decompose.

to complete the thermochemical-cycle Load line

possibility of a high-temperature topping cycle

versus

2'
be passed through a catalyat bed

3'

and 02 product gaaes might

for electricity generation by the high-temperature

the S03 is decomposed.

process heat stream is also considered an option.

the S03 to S02 and 02

between the process heat regions and the tritium

regions, energy transport occurring between the

separated from the product gases and

two primarily by thermal radiation.

4. Thermochemical cycles.

the blanket "energy spectrum" (e.g., the split

(chemical) reactor in which

'O1utiOn 'rem

bismuth sulfate.

to within

oxide

between thermal energy versus temperature and the

products are separated, the S02 is recycled, and

thereof used for

the 02 is discarded after purification.

production) cannot be made

thermochemical cycle is adopted.

unknown in the above process is the rate at which

bismuth sulfate decomposes at elevated temper-

thermal-hydraulic,

ature; experiments are planned at LASL

these data, but the synfuel studies will treat the

proceed until a thermochemical cycle is selected,

decomposition reaction rate as a major

The

is used:

S02 + 2L120= H2S04 + H2 (conventional 300 K

electrolysis)

systems studies are not specifically associated

H2S04(aq) + 1/3 Bi203 = 1/3 Bi203.3S03 + H20(350 K)
(7B)

with a given concept, but instead are more

and supportive of most design

A number of computational and modeling

the magnetic

an interim or preliminary cycle was

though the experimental baais for this cycle is

RADIATOR

RADIATOR

,

A

SIMPLE

~--~R&;&!Ts:NT/

t
LOW-TEMPERATURE
HEAT STREAM
(&lt;loc@K)

EX-REACTOR

D)

MULTIPIER

t

B

I

I
t

BLANKET

EXTENDED

HIGH -TEMPERATuRE
PROCESS HEAT
(Ha,1500 K)

LOW-TEMPERATURE
HEAT STREAM
(Li m HO,&lt; IOOOK)

LOW-TEMPERATURE
HEAT STREAM
(Pb, &lt;lCOOK)

Figs.XVI-19A-D

FLUIDIZED

BLANKET

-BED

&
t

/

D

t

BLANKET

HIGH -TEMPERATURE
PROCESS HEAT
STREAM (He,1500K)

P CESS H AT
ST EAM(He,1500

~HIGH-TEMP

BLANKET

ATURE

1-

v~

K)

Schematic diagrams of a number of potential synfuel fusion blanket concepts:
A)
B)
C) Lithium reflux blanket, with radiation-heated process heat

Simple radiation blanket with bank of process heat tubes
Radiation blanket with radiator surface extended by heat pipes

tubes located

external to the blanket
Packet (fixed) or fluidized-bed blanket

REFLUX

c

REJECT HEAT
FROM H@FLUIDIZING GAS

2H@O@yH@4 w34'~~' iIi20j*I@4~~3WH#*

81SWIHOXIOE
REACTION

For some applications, the analytic represen-

tation ia preferred. Considered here are expres-

Fig. XVI-20.
Flow diagram for a hypothetical Bi203.xS03 thermo-
chemical hydro~en cycle that is to be matched with
a given synfuel fusion blanket design.

for three fitting coefficients A, B, C.

two-parameter

(8) have been

such expressions came frrsa the published liter-

derived independently and

Table XVI-VI presents the

values of the coefficients and an indication of

the temperature ranges for which

~

p~

--

so*

S02,02

I&lAol_

HtPROEHJCTIOH,

(ELECIFIOLYZER)

summarized.

TKRMALWX)NMEHT)

D.c.
.
.
?1

SEPAIIATIOMWORK
I's
THERMALEOUVALHI1,

known reactivity &lt;ov&gt;.

rENESCY,P,

ELECTRICITY ~
I
H20 llYOR06fllCEKiWM

network.36 The ORNL

to the US

values.

SoJ

of

-1

the

are

are

&lt;OV&gt;=

EIIERCY,P2

~THERMAL

DESCUMSITIO#

-Eliio@so,

(C=O)ofEq.

dependence.

+
USIIUTHSIJLFATE

[cl between

into fusion

sions of the form

ni2033s$--lliio3+3s03

expressions which

AT-2/3 [ l+ CT4/3]

is introduced here.

ature, and a fourth was

I
SIIIFURTRIOXIDE
+
OECOHFOSITIM
S03+S$+'401 so~,o~

of an alpha particle born

passes. For a mirror

RaNl - exp (l-A/~a).

approximated by

Planck mode139

(MFE) computer

incremented by

table span-

standard.

densities

mirrored

interest

spanning

ends of

nuc1ear

Hence,

Table

used.

it a

~.40

~021

the

and

the

its

Ra

on

2)

is

an ORNL

of activity is termed "Plasma Systems Analysis,"

the analytic expressions and

and iterns which fall into this category

standard Table XVI-V values is less than a

maximum allowed, expressed as a nominal percentage

1. D-T Fusion Reactivity Data. Averaging of

(here 10% and 25%) of the standard value.

the microscopic D-T fusion cross section over the

2. Alpha-Particle Thermalization Model.

Maxwellian ion velocity distribution expected in

fusion-product alpha particles, born

magnetically confined plasmas yields the well-

3.52 MeV, will thermalize by Coulombic drag with

Incorporation

the plasma electrons and to a lesser extent with

ion temperature dependence of &lt;uv&gt;

the plasma ions, and thus should provide a plasma

reactor computer codes ia generally accomplished

heating mechanism which may lead to ignition.

by either 1) numerical interpolation of &lt;uv&gt;(T)

determine the rate of energy deposition, a Fokker-

using a finite table of discrete values, or

using en analytic representation to give a best

range-energy relation for a plasma of thermo-

least squares fit to the table of discrete &lt;UV&gt;

(T= 10keV)

-1023 m-3

As an example of the former approach, Tab1e

in Fig. XVI-21. The alpha-particle ranges exceed

XVI-V provides the reactivity table used in the

the length of any acceptable linear device, and

LASL fusion reactor design codes. This table is a

reduce the range to a few kilometers by

modification, in the sense that temperatures below
~ kev are inc1ude&34,35 of

the plasma density requires large confining fields.

A possible solution to this problem involves

ning the range 1-1000 keV, which has been provided

magnetic mirrors to reflect alpha particles at

fusion research community

linear reactor to allow multiple

National Magnetic Fusion Energy

simi1ar reactivity compilations,

general (US) availability makes

The ORNL table uses values of T

table is repreaentative of
37,38

that an alpha particle will

ratio

the effective

linear reactor of

0.2 keV for 1&lt;T&lt;1O, 2 keV for 10~&lt;100, and 20 keV

the infinite series

for 100&lt;T&lt;1OOO. A linear interpolation scheme is

tation, k/2 + !?PR+ !LP~+ .... which

employed in the ORNL subroutine package.

to k = (1/2) [l/(l-PR) + PR/(1-PR)21.

XVI-V is less finely resolved, requiring fewer

tive alpha-Darticle energy confinement efficiency

values for the same range, but uses a cubic spline

can be defined, where Aa is alpha-particle range;

interpolation scheme to give a smoother functional

is calculated dynami-

cally by the lDRBURN code and is used

,

To

at

D-T

the

(8)

Ua =

Three

chosen

Several

versions

identified.

exp(BT-1/3)

the discrepancy

the center of a

"
1s
'R
length

the probability

length k, may

alpha-particle

to estimate

increasing

reflected

represen-

converges

several

effec-

shown

path

with

the

The

be

An

is

be

to

at

~

(a) To be read as 7.00(10)-41

Maxwellian Ion Temper-
ature Ti (keV)

D-T Rea tivity

&lt;CJv&gt; (m /s)

0.05
0.1
0.5
0.8
0.9
1.0
2.0
3.0
4.0
5.0
6.0
7.0
8.0
9.0
10.0
20.0
30.0

I

111

Iv

v

are briefly described.

activity of magnetic

a. RFPR.

3.

ANALYTIC D-T FUSION REACTIVITY EXPRESSIONS OF THE FORM:

(-25)

(-27)

(-23)

-2/3

&lt;Gv&gt; = AT

TABLE XV I-V

TABLE XVI-VI

ature Ti (keV)

FITTING COEFFICIENTS

[ 1 + cT4/3] exp (13T-1/3)

D-T FUSION REACTIVITY TABLB

D-T Rea tivity Maswellian Ion Temper-
&lt;OV&gt; (m /s)

40.0
50.0
60.0
70.0
80.0
90.0
100.0
200.0
300.0
400.0
500.0
600.0
700.0
800.0
900.0
1000.0

7.00 (-41)(a)
3.00(-36)
6.00(-29)
2.00
4.00
6.27
2.83
1.81 (-24)
5.86
1.35
2.53
4.14
6.17
8.57
1.13 (-22)
4.31
6.65

the alpha-

3.68(-18)

5.00(-18)

7.80(-18)

9.43(-18)

9.46(-18)

integral

codes.

model.

-19.94

-19.94

-20.70

-21.11

-20.94

-4.55(-3)

5-24

The

5-16

1-15

5-38

of

.--

An

0.0

0.0

0.0

0.0

A

B

c

?.93(-22)
8.54
8.76
8.76
8.64
8.46
8.24
6.16
4.90
4.13
3.63
3.28
3.02
2.83
2.68
2.55

as radial diffusion

and is the theta-

Bremsstrahlung and

Alpha-particle

introduction

Full engi-

transfer,

thermal-

ener-

The

T(keV) R4NGE

IEI&lt;1O% \cl&lt;25%

2-3 and 2-6 and
28-37

20-43

4-29

3-21

1-20

3-44

gy/particle losses can be modeled by

the fraction of all alpha particles that are lost

reversed-field pinch plasma.

as a function of burn time, whether

cyclotron radiations as well

particle thermalization model is analytic or based

and anomalous losses are included.

on a Fokker-Planck model.

neering energy balance, first-wall heat

Burn Code Development.

and economics models are included in RFPR.

confinement systems studies

alpha thermalization is treated by a Fokker-Planck

is the development and use of a variety

time-dependent thermonuclear bum

major code systems that have been developed and

tion of an earlier approach

used to model the dynamics of reactor-like plasmas

pinch counterpart of the RFPR code. End-loss

b. DTBURN . This code system is an adapta-

This pressure-balance code uses

of an ad hoc loss term to this zero-dimenaional,

analytic field profiles to model the dynamics of a

pressure-balance code.

three-species (electrons, ions, alpha particles)

ization is modeled by a Fokker-Planck cmnputation.

EXPRESSION

THERtIALIZATlONOISTANCE(m]

0-PARTICLE RANGE-ENERGY RELATION

Fig. XVI-21.
Alpha-particle (3.5-MeV) range-energy relationship
for a D-T background plasma. Pressure balance for
a given beta value gives the required confining
field B for the various densities.

fields (RFPR, screw pinches, Tormac, high-field

e. LNBURN. The coupled dynamics of a high-

beta plasma and an imploding, cmopressible liner

is modeled with the LNBURN system (Sec. XVI.D.)

The Lagrangian MES code treats only simple parti-

cles, has all major loss terms (radiation, axial

and radial conduction, etc.), and uses an analytic

mode116 to compute the liner compression.

continually

against a complex MHD physics code, CHAMISA.

Both the transient and quasi-

steady-state behavior of a neutral DT

the edge of a thermonuclear theta-pinch plasma ia
NGBD 42,43

modeled by the neutral-gas-blanket code

This MHS code system treats a wide

charged and neutral particles interacting with

c.

3.52M#

This

lDRBURN.

Since radial effects are

Planck computations.

terms) procedure.

MHSBRN

both

more

f.

is

generally

code

NGBD.

three-

devices

Tokamak).

averaged,

linear DT

fast-running

modified to model

theta-pinch plasma.

particle and energy fluxes.

plasma was concluded,

dynamically modeled.

theta-pinch or a

single- particle

processes are

to include

The NHSBRN

minimized.

energetic

inertial

analytic

wall

from

The

speciea code models the dynamics of a

plasma in both the radial and axial dimensions.

This code system will be

neutral-gas blanket phenomena

time-dependent,

three-particle (iona, electrons, alpha-particles)

associated with the RFPR studies.

lDRBURN is used primarily as an axial burn code to

4. Neutral-Gas Blanket Theory. Research

compute the plasma energetic, first-wall response

the use of a neutral gas to cool a quenched RTPR

and overall energy balance of linear

and produced a numercial

(primarily linear theta-pinch reactors) that are

code, NGBD, that models both the transient for-

subject to various end- stoppering schemes (solid

mation and quasi-steady operation of a neutral- gas

endplugs, reentrant endplugs, etc.). An

blanket. The tranaient formation of a neutral-gas

alpha-particle thermalization model is used, which

blanket at the plasma perimeter is examined to

has been previously calibrated by separate Fokker-

determine the conditions under which a neutral-gaa

layer can be stably formed. The main

d. MHSBRN . This code system (actually two

formation is that the neutral gas together with

separate, but nearly identical systems) computes

the cold plasma adjacent to the wall

the burn dynamics of either a

Z-pinch plasma in the radial direction using a

full magnetohydrostatic (i.e.,MHD with no

sufficient density and thickness to shield the

charge-exchange

species; during the formation wall damage is to be

is a

penetration

time-dependent code with alpha-particle heating

neutral-gas "slaba" into the plasma

back fitted to siroulate,onthe basis of separate

reaction rate of incoming diatomic and monoatomic

Fokker-Planck computations, the associated dynam-

gas with the plasma ions, electrons,and alphaa are

ics. This code is used to examine the effects of

radial profiles end field diffusion on the results

Hydrodynamic and radiation

simultaneously

obtained from the previously described pressure-

first-wall (physical) sputtering and evaporation

rates are computed on the basia of the computed

balance, zero-dimensional burn codes.

system is presently being modified

ion-electon-alpha interactions as well as to join

After the tranaient formation of

theta- and Z-field profiles in order to model

gas blanket a quasi-steady analysis determines

accurately high-beta systems with helical

whether or not the plasma energy can be extracted

with minimal wall damage. Since heat flow through

a

into

This

gas with

calibrated

spectrum of

of monoenergetic,

the neutral-

concern for

can acquire

calculated,

neutral

and

the

and

I

loss rate by

should be possible.

~t~d ~t#i~-~fter

154

.6

o

I

the neutral gas and enhanced bremsstrahlung radi-

keV neutrals from the dissoci-

ation are dominant cooling mechanisms during the

ation of diatomic ions primarily contribute

quasi-steady-state stage, the rate of plasma ener-

diatomic

gy loss should be determined primarily by the rate

exchange with the hot plasma in the core con-

of neutral penetration into the plasma. This neu-

tributes to the spectrum above 10'2 keV.

tral penetration is determined by the two compet-

illustrates

ing processes of neutral diffusion and neutral-

yield and sputtering rate

plaama interaction. Control of the plasma energy

a
time for the case given in Fig.

'1203

as

controlling the neutral diffusion

XVI-22. The sputtering rate peaks at 25 ps, after

rate, which in turn is determined by

which the neutral-gas blanket begins to moderate

neutral density in the neutral-gas

the outwardly directed monoatomic neutrals. Figure

XVI-23 also gives the first-wall evaporation rate,

monoatomic neutral particle spectral flux at

which is small compared to physical sputtering,

first wall 35 us after injection for a diatomic

for a diatomic neutral density

neutral density
of
flux for energies less than 5 x 10-4

consists of a 0.3-Ims layer of

l-sIs layer of niobium and

from diatomic neutral charge-exchange with the low-

cooled by liquid lithium at 1100 K. The wall

temperature plasma adjacent to the wall.

and integrated loaa for both charge exchange and

I

I

.

-4

The

the

the

The

keV

gives

Between

results

~021 m-3

particle

reservoir,

function of

'2 and 5(10)

Figure XVI-22

the diatomic

Figure XVI-23

the spectrum.

The first wall

I(J41

alumina, backed by a

no =102'/m3
t =35ps

_FRANCK-CONDON
DISSOCIATION
COMPONENT

as : ;Yft:!?q
no
getic plasma.

HOT PLASMA
EXCHANGE

__:
_EVAPORATION

CHARGE-
COMPONENT

b=O.5m
noxlo m

I
CHARGE:

Wall erosion

/','
//
II

EXCHANGE

PLASMA

0

'COLD

n
-1
w
F

/+
//

21 -3

/1
Y

do -

142 -

107-

Id' -

,1 I_

~
z
g

10°

of

--//

log

108

'&lt;

1/

Is

,1

//'

l\

:/

~

/'

l/

I

I

TIME

SPUTTERING

/

I

[

/

/

%

~--

to

of

for

the

,.19''"

dose

charge-

neutral

~021 m-3

sputtering

- [0'7 ~
m
z
E

- 10'6 P
&

a-

10's :
z
a

13 i
a
&gt;
w

(/AS)

-10

14 ~

i=
a

- #

F-N

U
i-

do

0

12

'-

~

:

Fig. XVI-22.
A typical first-wall particle energy spectrum com-
the admittance of neutral gas

adjacent to an energetic plasma.

Fig. XVI-23.

rates and integrated yields computed
time after the admittance of
of neutral gas adjacent to an ener-

5.

o'

The

(for

power

A1203 ,

Modeling of

values of n

low density.

below -'1021111-3.

standing and high

which generally will

simple geometry, a high

of Linear Magnetic Fusion

penetration into the hot

Energy dose and
first wall as a
density.

&JTRi~

,d'~lo'

, ~19

2[

evaporation are shown on Fig. XVI-24 as a function

scaling on the basis of either free-streaming

of neutral-gas density no.

sputtering increases rapidly for no
~022 ~-3

conduction predicts device lengths

(parallel-field) thermal

as a result of increased neutral atom

fusion powers that generally are excessive from

the reactor viewpoint. It is possible, however,

bremsstrahlung energy dose dominates

to reduce the size of machines while

whereas sputtering dominates at

favorable overall energy

composite evaporation

balance, by exploiting the B21 scaling predicted

indicates sputtering and evaporationare very smal1

by classical theories of end flow, where B

atomslm

confining field and L is the total length of

4.5(10)-15 m or 1.1(10)-2Um/y at a O.1-EZ pulse

In order to achieve ignition

rate and an 80% duty factor). The major

conditions with total alpha-particle confinement,

instead, becomes one of thermal cyclic fatigue

(10)6 T2m

require that no be kept

axial thermal conduction

Plasma/First-Wall

E = 1000 m, then fields in excess of 30 T will be

determines the energy containment time. Hence, if

actions in High-Field LMF Reactors. A

The resulting high

the significant advantages of plasma stability,

first-wall constraint is

(LMF) concepts exhibit

will cause a high bremsstrahlung radiation flux at

level of physics under-

imposed. One approach to this problem is to relax

to

at

by

The

high

Wall

core.

curve

exceed

Inter-

plasma

erosion

concern,

material

density.

B2k must

Classical

required.

less than

permitting

wide range

evaporative

corresponds

the first-wall

plasma column.

ature of - 7 keV when

neously maintaining a

the first wall, and a

particle loss or axial

therefore, becomes one of

transient description of

the thermonuclear burn

complex and coupled

evaporation model

plasma, and the

nuclear plasma,

viability of

evaporation

Although

dynamic

wall.

(m-3)

1(34

of

the

and

work

This

total

is the

temper-

impurity

simulta-

at plasma

constraints,

plasma densities

(lithium) neutral-

comparable to

time required

a detailed,

interaction

diffusion,

question,

#provide

similar

between

becomes

would

major

of a

The

the

be

of

a

thermal-mechanical

ablation

into the plasma chamber.

examines the interaction between a burning D-T

plasma and an evaporating (liquid-lithium) first

first-wall cooling during the short, intense bum,

the evaporated lithium will also cool the ignited

evaporating first-wall and the burning thermo-

question. The penetration of &.vaporatedlithium

into the high-beta, dense plasma will occur by the

therefore,

processes

ionization, and charge-exchange; these processes

occur on a time scale which may

time.

competing reaction and

diffusion rates and the rapidity of the inevitable

plasma quench relative to the burn

to achieve a net energy balance. A simple surface

is coupled with

the

gas/plasma interaction to form the basis

simple LMF reactor model with which to examine the

the evaporating first-wall scheme.

The computational model and techniques are

IC?2

DENSITY

Fig. XVI-24.

integrated erosion yields at the
function of
initial neutral gas

I. FUSION REACTOR NUCLEONICS

triangular spatial mesh, neutron transport code,

Several tasks are included in the category of

is designed specifically for the fusion cotnsunity,

to those used to describe neutral-gas blanket

the national fusion energy development program.

(See.XVIH.4).

In addition, shielding and blanket analysis for

As illustrated in Fig. XVI-25, the reactor

the Intense Neutron Source Facility and consul-

performance, as measured by

tation on other programs are provided.

grated Lawson parameter, n T, increases

system nucleonics calculations for

6(10)21 s/m3 at about 45 T.

heated, small tokamak reactor concept (RICCATRON)

this point nT gradually begins to decrease.

have continued in support of an INESCO Corporation

increaae in n'rwith B is considerably below that

conceptual study. Progress in each of these tasks

expected without lithium evaporation but neverthe-

less an appreciable gain in n~ is observed.

Nucleonica,

dimensional transport code development and appli-

cations, quantitative nuclear data

computer code and data

streaming

reactor systems studies.

42,43

Fusion

Reactor

phenomena

maximum of

deterministic

'MAX
To = 5 keV

@o=o.999

b=O.5m

c1
z
a

PEAK

I

156

4

c

I

=300ms

a

to

a)

The

The

and

file

two-

After

tool.

fusion

design.

methods,

including

assessment,

standardization,

1. TRIDENT-CTR.

(triangular) band

which was written

is summarized below.

the space-time inte-

These efforts support

Lawson
field
(LMF)
first

the band, the number of

constant source, and

autcssatic mesh

generated by

FIELD,B(T)

The TRIDE~

estimate.

at a

I

I

be

d)

e)

b)

c)

An

1

time- ~nd space-integrated

Dependence of
parameter as a function of peak compression
to dense Linear Magnetic Fusion
applied
an evaporating liquid-lithium
device with
wall.

I

COMPRESSION

Fie. XVI-25.

so

an

two-

ohmic-

Blanket

small core

time is in

finite-element

generator that performs

The SN conatanta can

The option takes as

(EQN sets47),

specification,

is defined

specified

starting

included

read

flux

be

dimensional (rectangular and cylindrical geometry)

TRIDEtTl?-CTR45 is a

and was written and partially tested during 1977.
TRIDENFCTR was developed fras a previous code,46

for use in fission reactor

following modifications have been

made to TRIDENT for use as a fusion reactor design

The storage structure has been modified

that on a CDC 7600 computing system only the

information required to solve one

memory. This modification increases the maximum

allowable problem size from about 3000 triangles

in TRIDEtW to 10-15 000 triangles in TRIDENT-CTR.

restriction that six triangles

share each interior vertex has been relaxed; this

modification has made the specification of spatial

mesh leas rigid and easier to establish.

triangulations of the spatial mesh is available in

TRIDEN17-CTRas an option.

input on a per-band basis the number of zones in

trianglea per zone, and

the zone boundary coordinates. A zone

as a region of constant material

constant

A group-dependent SN order may

and used in TRIDENT-CTR.

in, or a combination of both.

the code

Options to allow interior boundary aourcea

(right, left, top, and bottom) have been

in TRIDENT-CTR. This option allows the replace-

ment of a distributed neutron source, isolated in

a central vacuum region, and a deterministically

the

CTR .

upon use of

surface (i.e., first wall).

large tokamak problem

r(plasma) Ar(Cu)
--

22.5

22.5

22.5

22.5

(a)

10

----

10

0

calculated boundary source at the first material

related work, sets of one-dimensional scoping cal-

culations for possible plasma, coil, and blanket

configurations were performed, and two-dimensional

The changes embodied in items a) and b) have

computations were initiated. Following correction

increased the flexibility and scope of TRIDENT-CTR

of errors in the MATXS (30 x 12) library, predic-

over TRIDE~ with a minimal cost in increased com-

mutational time. The code has been

(collapsed) VITANIN-C50

(30 x 12) libraries

(RIGGATRON).

various one-dimensional RIGGATRON configurations

tion of the RIGGATRON involved the calculations

Results of that comparison are

and storage of over 3 million

compiled in Table XVI-VII, which

provided a realistic, large-scale test of TRIDE~-

ratios (T) and leakages (L) for various plasma,

coil, and blanket thicknesses.

2. RIGGATRON Blanket Nucleonics Study.

the three cross-section sets is + 5%.

and two-dimensional neutronlgasrsa-ray transport

quent analysis, the NATXS (30 x 12) neutron/gamsa-

calculations have been performed for the INESCO

ray cross-section set will be used.

RIGGATRON, a tokamak device using ohmic heating to

Additional one-dimenaional calculations were

induce deuterium-tritium fusion.

performed on the RIGGATRON assuming a void between

tested and evaluated various cross-section sets

coil and blanket in order to increase the tritium

for consistency and accuracy and

these computations are

neutronlgamna-ray set for future analysis.

0.225 m, coil of thickness 0.10 m,

(30 x 12)

given in Table XVI-VIII for a

In

The

One-

solu-

coupled

MATXS49

tested on a

TABLE XVI-VII

were compared.

finally settled

Earlier efforts

flux moments and

(101 x 12) and NJOY51

tions of the MATXS3 CTR

breeding. The results of

CROSS-SECTION COMPARISONS

~7(Li7)

T5(Li6)

0.888

0.257

0.880

0.249

0.927

0.252

1.119

1.081

0.760

0.128

0.123

0.801

0.126

0.757

0.138

0.778

0.182

0.769

0.182

0.914

0.932

0.884

0.637

0.640

0.132

0.638

0.773

0.955

0.135

0.182

1.096

_ T

(a)

T+L

--.

-

-

L

--

-

-

-

-

0.960

0.951

1.145

1.179

1.129

for

void

thickness

In subse-

lists breeding

Agreement between

plasma of radius

CTR
101 x 21
NATXS

CTR
101 x 21
MATXS

CTR
101 x 21
NATXS

CTR
101 x 21
MATXS

NJOY
30 x 12

7% H20
Coil

30 x 12

30 x 12

30 x 12

30 x 12

30 x 12

30 x 12

30 x 12

no H20

NJOY

NJOY

Coil

NJOY

all dimensions in centimeters

T6 = $i tritium breeding ratio
T7 = Li tritium breeding ratio
T=T6+T
L = leakag~ per fusion neutron

Ar(Li)

100

COMPARATIVE BREEDING AND LEAXAGE FOR RIGCATRON WITH
AND FULLY DENSE COILS WITH/WITHOUT VOID BETWEEN COIL

0 (7%)
D BLANXET(=)

la) all d menslons in centimeters

T6 = 7Li tritium breeding ratio
T7 = Li tritium breeding ratio
T=T6+T7
L = leakage per fusion neutron

of 2.0 m, and blanket thickness of 1.0 m.

comparison includes errors associated with

quadrature set and P3 cross-section

discrete-ordinates approximation and the statis-

were employed. It is seen that the addition of

tical error in the Monte Carlo results.

the void between coil

cases agreement is to within 4%

approximately 7%-9% increases in breeding ratios

breeding and leakage. The higher breeding value

for the two cases considered.

frmn the Monte Carlo calculation is inferred to be

For a more realistic one-dimensional model

caused by the resonance self-shielding effect, and

graphite reflector (0.25 m) was placed outside the

further study of this effect is being undertaken.

blanket, similar to the two-dimensional mockup

3. Connnitteeon Computer Code Coordination

the RIGGATRON shown in Fig. XVI-26.

(Cccc). During the past year neutron and photon

gamna-ray heating was alao cmnputed using the

multigroup cross sections in the proposed Cccc

MATXS kerma factors with

format called NATXS were

copper coils. The results of these calculations

(TRANSX) for processing the MATXS

for coil thicknesses of 0.07 m

also provided. With this code,neutron, photon, or

values for neutrons and

coupled cross-section sets can be easily generated

r(plasma) Ar(coil)

&

10

22.5

22.5

in Table XVI-IX. The Q

standard deviation of

given in MeV.

~

0.182
0.159
0.884

0.248
0.205
1.096

:

a

l&

of

and

An S8

blanket

300

Ar(void)

produces

expansion

300
o

100

100

0

Ar(blanket)

TABLE XVI-VIII

0.643
0.739
.-

0.778
0.878
--

0.757
0.874
--

0.123
0.129
--

0.135
0.139
--

0.880
1.003
--

tritium breeding

coolant in the

Neutron and

Table XVI-X

and 0.10 m

7% D20

appear

4 1365

E
s
Q
x
&lt;

Shews

f
l
~

g
z
z

1%.

The

sI

92.5%L17+7.5%L$

the

T+L
--

In all

tritium

for both

no H20
coil

7% H20
coil

1.129
1.20
--

0.960
1.037
--

RADIAL DIRECTION (cm)

libraries was

produced.

1745 200

used for

code

A

LI6

Two-dimensional RIGGATRON blanket model
neutronics computations.

Fig. XVI-26.

ganxnasin the copper, lithium, and graphite are

As an additional check on the cross sections

as well as the discrete-ordinates calculational

model, a continuous energy Monte Carlo calculation

(MCNP code) was performed. This Monte Carlo cal-

culation duplicated the one-dimensional discrete-

ordinates calculation for a 0.225-m-radius plasma,

0.07-m-thick copper coil, and a l-m- thick natural

Li blanket. All calculations of

and neutron leakage were performed to a relative

results of comparisons among various cross-section

sets, where the MATXS library was processed from

ENDF-IV. These calculations show direct compar-

ison of the MATXS library with a calculation using

its progenitor ENDF-IV point-wise library.

COMPARISON OF MONTE CARLO RESULTS WITH DISCRETE-ORDINATES CALCULATIONS FOR

Cross-Section Set Leakage

from Monte Carlo

r(plasma)Ar(coil)Ar(void)Ar(blanket)Ar(graphite)&
----

.

~Cu

(a)alldimensionsincentimeters
T6 = 6Litrititmbreedingratio
T7 = 7Litritiumbreedingratio
T= T6+T 7
L . leakageperfusionneutron
Q inunitsofMeV

10

'8 '3'

200

0N33TRAN

22.5
22.5

Monte Carlo, MCNP

CalculationalMethod

linking I.ASL/LTSS with

to user specifications.

CTR-related analyses.

cant. An

vectors.

ONETRAN

'8 '3'

1.o810.229 1.31 0.004 1.31L 1.09 6.11 7.69 1.05 0.03 0.16
0.9690.1391.108 0.003 1.111 1.34 8.55 6.22 0.7s 0.02 0.13

MATXS

.

T

L

--

~

T

&

LASL/TD

ENDF-IV

Pointwise

17--

30-Group

3.74

25

0.275

0.277

0.267

LOO

1.097 4

1.054 5

30-Group

TABLE XVI-X

TABLEXVI-1X

%Difference

from Monte Carlo

(200-cmVOIDAND25-c.CRAPHITEREFLEGP3S)(a)

.07-m COIL, 7 vol% D20 and l-m BLANXET (nLi)

CXMPARATIVSBREEDINGMD HSATINGFOR RIGGATRONWITH7XD20GOILS

machine-independent,

haa been modified

to LTSS at LLL,

TRANSX-MATXS

systems. A

individual

Livermore

Progress

signifi-

response

but who

easily.

do not

1.056 5

used

they

3.00

the

for

The

can

in-

in

to

be

Z Difference

G

!&

~Li

-3.73

-3.91

n
9Li

system-independent,

in question.

routines,

code has

computer

complex

(CGS)

Using

sub-

This

for

the

Individual reaction cross

for users who have had some contact with FORTRAN,

sections, such as tritium production and gas pro-

understand

duction (if present in the ENDF source library),

tape conversion process for moving

are readily available through the TRANSX-MATXS

files back and forth from LASL/LTSS to LLL/LTSS

link. These reaction cross sections can be

has been formalized, and a document explaining how

cluded in the cross-section tablea that are used

to do that has been circulated within LASL.

in various transport codes, or

the tape procedure, the MCNP Monte Carlo

extracted from TRANSX aa

been standardized to be identical both on LLL/LTSS

and I.ASL/LTSS,allowing updates to be made more

Considerable testing of

system has occurred.
MATXS cross-section sets,
such as those used in the RIGGATRON studies,52

standardization of graphics

which allows switching from one machine to another

have been generated and are being

with a minimum of difficulty, has also been made.

Toward this end the TRIDENT-CTR code on LASL/LTSS

4. Adoption of LASL C&des to the Magnetic

Fusion Energy Computer Network.

uae

device-independent,

the Lawrence

routines that can easily be modified

Laboratory (LLL) MFE network has been

particular system or machine

introductory manual

change should make the graphics transportable with

which is aimed at the beginning user, has been

a minimum of effort. At LASL these routines were

released in final form. The document is intended

written using the Crsnmon Graphics System

tion method

method was

calculations.

routines with

neutrons in the MeV

blenket/shield

The methods

material.

reactor

shield.

media.

6.

respect to the SENSIT code;

sensitivities to the production cross sections of

similar scheme was used on LLL/LTSS using the

sensitivity to either the

TV80LIB graphics package.

energy or angular distribution of

5. Use of Monte Carlo Computations In the

particles, however, could not be given.

Study of Deep Penetration Problems. A Monte Carlo

the classical sensitivity

developed to

theory, which allows one to calculate sensitivity

information on the forward and adjoint

profiles for secondary energy and angular distri-

providing a means to calculate contribution density

butions, has been developed and is being incorpo-

via forward Monte Carlo. A

between surface integrals of

analysis code system (LASS).53

sensitivity and uncertainty

current and the volume integral of the current

provides a new approach to detector

adequacy of existing neutron and gamma-ray cross

This approach proves useful

sections needed to perform the nuclear design

deep-penetration problems. Calculations made

calculations for TNS have been assessed.

such problems in the analog Monte Carlo mode

assessment was performed

indicated up to 90% time saving for the contribu-

established sensitivity and uncertainty analysis

as compared to the regular analog

techniques as demonstrated in previous conceptual

method. The two methods are now being tested with

designs of TFTR and EPR. The ERRLIB cross-section

various biasing schemes.

error data base developed for the TFl?R and SPR

for the contributionMonte Carlo are very different

expanded and used extensively

than those used in regular Monte Carlo.

assess the data accuracy requirements for TNS

inary results indicate that advantages similar to

applications. The sources of error estimates for

those for the analog case may be obtained for the

local

biased case, especially for highly

nuclear systematic, and (3)

ing ENDF/B formatted "error files."

Sensitivity Profiles for Secondary Energy

effort has been devoted to a collaborative effort

and Angular Distributions.

with ORNL to improve and generalize the ENDF/B

tration transport calculations the integral-design

Parallel developments have also

or response parameter of interest is determined

occurred in the error processing capabilities of

mainly by secondary particles which are produced

by interactions of the primary radiation with the

Multigroup error estimates for

If the primary radiation consists of

range, then the secondary

into the ERRLIB format for test purposes.

neutrons from inelastic scattering

design considered was

(n,2n) reactions, and the ganma rays from

General Atomic Company and is

reactions are of primary concern in most

continually through close cooperation with

A

a

7.

in

for

the

flux,

ERRLIB

Prelim-

response

generate

include:

scattering

correlation

Quantitative

contribution

studies was

An extension of

such secondaries.

rated in the LASL

simple transformation

Biasing considerations

the LASL-developed NJOY

statistical analysis

from the ENDF/B-V

structural metals.

error formats.

calculations.

design team.

code.

deep-pene-

first TNS

radiation

spectra.

In most

In many

simi1ar

fusion

events

(njy)

These

(1)

An

or

B

Data

to

The

(2)

This

using

Assessment.

assessments,

quantitatively

these secondary

the rapidly develop-

to scope the uncer-

cross-section processing

as important as

discrepancies

library were

Considerable

important

processed

provided

updated

neutron

Li and

14-MeV

being

1arge

the

the

The

by

a

secondaries can be considered additional radiation

attempt was made

sources within the blanket or shield.

tainties in secondary energy by performing

reactor shielding calculations the biological doae

rate at a given point of interest is dominated by

inelastic-scattering spectra for all

the contribution from secondary gauma rays, and in

Large differences (typically

fusion reactor nucleonics calculations important

50%) are found between materials that should have

design parameters (tritium breeding,

damage, and induced radioactivity) depend strongly

indicate that for some design calculations un-

on (n,2n) and (n,3n) reactions in the blanket and

certainties in spectra will be

of standard cross-section

uncertainties in the cross sections themselves.

of

These

sensitivity analysis can be used to calculate the

These data are

above 500 keV to include new measurements from the

fast-neutron energy spectrum typical of the Oak

representation

radiation-damage-resistant

LLL . Particular emphasis was placed on accurately

experiments and are summarized in Table XVI-XI.

The ENDF/B formats are presently not

lation. An evaluation of selected neutron-induced

Cross sections and angular distributions for a

support of the LASL materials program for the

incorporated into the ENDF/B-V evaluation frmn

sulators for fusion applications, the

cross sections for 27 hydrogen-

production reactions have been

representing the neutron emission spectra fras

(n,2n) reactions, as measured recently at IASL.

In addition, the total cross section was

fission

9.

8.

of

analyze

planned

14N(n,a)

14N(n,p)

13C(n,a)

12C(n,a)

Reaction

(ORR) was

development

Cross-Section

Gas-Production

Ridge Reactor

spectrum-averaged.

spectrum-averaging.

25 Mg(n,p)

'atSi(n,a)

'atMg(n,p)

'atMg(n,a)

27Al(n,p)

28Si(n,p)

29Si(n,p)

170(n,na)

24Mg(n,p)

29Si(n, )

28Si(np)

170(n,a)

160(n+2)

Si(n,p)

Mg(n,a)

1)
(:)

Nat

used to perform

correlations in neutron-emission

Be reactions. For this reason a special

Evaluation

used to incorporate the

Cross

cross sections for 'Be has been

TABLE XVI-XI.
GAS-PRODUCTION CROSS SECTIONS FOR C, N, O, Mg, Al and Si

A

N@

In

to

1.81

(a)

was

for

in-

and

the

and

and

+ 0.5

9.65(-2)

3.58(-2)

8.59(-3)

Calcu-

needed

reactor

helium-

neutron

accurate

evaluated

Sections.

and IASL56.

irradiation

1.27(-3)(b)

Thermal
Cross Section (b)

Spectrum Aver-
aged Cross Section (b)

.235 + .01

+ .1(-3)

1.10(-2)

1.05(-1)

1.59(-3)

1.35(-3)

3.48(-3)

1.47(-3)

1.39(-3)

2.83(-2)

4.94(-3)

8.07(-3)

3.46(-3)

4.70(-3)

6.24(-3)

.08(-3)

6.68:

2.9

-3

.

of

from

data.

format

suited

updated

spectra

9Be(n,2n)

energy-angle

completed and

Method

ENDF/B

Estimated

ENDF/B

ENDF/B

ENDF/B

Calculated

Calculated

ENDF/B

ENDF/B

Experimental

Estimated

Estimated

from ENDFIB

ENDF/B

ENDF/B

ENDF/B

Experimental
Spectrum
Experimental

Experimental

Spectrum
Estimated

Cross sections estimate~3to be less than 10
TO be read as 1.27 x 10

barn are deleted.

series of (n,n') level excitation energy bins

of radiation damage in order to assist in inter-

suitably represent the LASL54 experimental data

pretation of radiation-induced changes in physical

at 5.9, 10.1, and 14.2 MeV were derived.

properties. In response to a request fran OFE/

the evaluated excitation data actually represent

DOE, an irradiation test plan was

(n,2n) reactions, a multiplicity factor of 2

concurrent displacement-damage

specified in the evaluation. The level excitation

generation effects in a model ceramic. To

smoothly interpolated and extra-

the OFE Fusion Materiala planning effort reports

polated to other incident neutron energies, so

were prepared assessing insulator and

that the energy from threshold to 20 MeV

problems and describing a program plan for their

Because of the special data format used, it

Effects.

was necessary to modify the NJOY code to permit

refractory ceramics (six oxides and three ni-

new

cm2(En &gt; 0.1 MeV) at 925 and

trides) were irradiated to doses of

processed into the I.ASL 30-neutron by

Swelling measurements have been made

materials, with results which may be summarized aa

Under funding from the Division of Physical

Research (DPR) of DOE, development of the nuclear

lMgA~04, Y3A15012, Si20N2, and SiAION exhib-

theory code GNASIi57was continued, and a variety

ited low swelling (less than 0.5 vol%)

of cross-section calculations were performed that

Y203,Y203-l%Zr02 showed moderate swell-

fusion reactor interest.

charged-particle emission spectra were

for 15-MeV neutrons on stainless steel 316,

and reaction cross sections, including

.Only Al O

vol% (in the 3-10% range).

and BeO swelled more

(n,n'), (n,p), (n,a), (n,pn), (n,2n),

With the exception of the last two materials,

(n,na), and (n,cm) reactions, were

tested ceramics behaved well

isotopes of

swelling criterion at high damage levels under the

for several zirconium isotopes.

incorporated in ENDF/B-V

Thermal diffusivity tests have been completed

Reviews of calculations and capabilities of

on samples irradiated earlier in EBR-11 to a

GNASH code and general data needs were

(En&gt; O.l MeV)

sented59-61 between 10 and 40 MeV.

Earlier work showed significant degradation (8 to

Also under DPR funding, development of the

68%) at rocm temperature for eleven ceramics after

EDA R-matrix code progressed, and analyses

this moderate-dose irradiation.

several mass systems that included reactions of

have been concluded by determining the change in

fusion reactor interest were

analyses

for

These

D(d,p)T, T(d,n)4He, 3He(d,p)4He, and

and '2°3-1%zm2"
reduction in this property of

are of

covered.

functions were

group structure.

ccxnplete, and the

multigroup processing.

to 20 MeV for all five

6Li(n,a)T reactions.

D(d,n)~e,

ceramics.

will be

studies

include

J.

1.

ia

is

is

has

Study

been

This

that

(n,y),

Because

" si3N4'

12-gamma

follows:

solution.

Structural

evaluation

modification

Neutron-induced

ing (0.5 to 1.5 vol.%)

calculated

scattering theory, and

thermal diffusivity

conditions utilized.

of 2.8x 1021 n/cm2

These results

calculated up

respectively.

titanium and

evaluations.

continued.

irradiated

microscopy

important

electron

(n,np),

These

dence

pre-

the

for

the

of

of

and

gas-

Nine

1100 K

ceramic

support

than 1.5

on these

candidate

l-2x1022n/

in EBR-11.

developed to

The measurements

polycrystalline

compared to

'2°3
a

materiala

and 33%,

single-

Thermal

showed

A1203.

phonon

1015K.

With

With

dose

25%

for

the

the

at

Temperature dependence of the degradation in

INSULATOR AND CERAMICS RESEARCH

thermal diffusivity was measured

Activities during 1977 included studies of

structural effects (swelling, thermal diffusivity,

crystal '1203' 'hich
300 K, and compared with

'hoWed a 45% 'eduction at
the temperature depen-

microhardness, and fracture toughness) and elec-

unirradiated

trical effects (multiple-pulse dielectric break-

diffusivity for both materials decreased

down strength) in unirradiated and

increasing" temperature, Consistent

Transmission

studies have been conducted to evaluate the nature

the irradiation decrement

"L_JJ

12C0
eao
ANNEALING TEMPERATURE ("C)

o

t
ALPHA+DELTA

XVI-28 .
volume

and
Thermal
function of
occurrence of phases plotted as a
Al1
isochronal annealing
points are the average of two samples except where
circles
Scatter fell
1 is indicated.
except where error bars are shown.

temperature

shrinkage,

within

(1 h).

restoration of the alpha phase and closure of

large improvements in

diffusivity (Figs. XVI-28 and XVI-29), but

microcracking accompanying the phase

retards this recovery until very high

temperatures are reached.

probably result in

bond

of such a materials
applications will

also decreased, in agreement with

dependence of scattering from defect sitea (Fig.

XVI-27). The observed reduction in degradation to

indicates that problems associated

with an irradiation-induced decrease in thermal

diffusivity should be lessened at higher operating

Thermal diffusivity of unirradiated plasma-

cooperative project with

also

Technolgy group. This material has been proposed

for applications where neutron flux is low but

thermal throughput is significant.

sivity equal to approximately one-tenth that of

conventional alumina, because of

transition phases and porosity.

isothermal anneals up to 1925 K

I

t

1.0

Q9 -

was

0.8 -

'1203

'12°3

sprayed

15% at 723 K

temperatures.

that "as-sprayed"

a7 -
'\

Q15~

Q6 -

09 -

Q2--

Q5 -

Q3 "

300

has

ALPHA+ ETA

I

I

I

I

a

a

in

the

that

found

diffu-

showed

It was

thermal

temperature-

investigated

Isochronal and

LAM., Materials

the presence of

pores result in

Fig.
diffusivity,

applications. The use
in high
thermal flux

function of
irradiated TYCO

of the ceramic/metal

during annealing would

1!
okI

700 800

mations

600

I

0'1

I

T( K)

the

ALPHA

that

thermal

transfor-

Changes in

shrinkage plotted
the

system
require

Symbols have

reactor

failure

density

fusion

T&E (:

-1

163

I

in

I

I

I

Fig. XVI-27.

Relative thermal diffusivity as a
temperature for unirradiated and
and LINDE sapphire:

filled circles--unirradiated LINDE, C-plane
filled squares--unirradiated TYCO, R-plane
open circles--irradiated LINDE, C-plane
open squares--irradiated TYCO, R-plane

I

AN;EALI:G

Fig. xVI-29.

Thermal diffusivity and volume
versus annealing time at 1625 K.
same significance as in previous figure.

directly related to brittle fracture strength and

which allows strength comparisons if the flaw size

distribution is known. These data show that both

development of special techniques, such aa spray-

ing onto a heated substrate to optimize

aa-sprayed ceramic structure.

Mi.crohardneasand indentation fracture test-

ing of ceramics irradiated to moderate doses in

EBR-11 has begun. In this work room-temperature

microhardneaa is obtained from dimensions of

indent, and fracture toughness is determined by

aPPIYing fracture mechanics techniques to inter-

pretation of crack patterns around the indenta-

Preliminary results for YAG

presented in Figs. XVI-30 and XVI-31.

XVI-30 shows room -temperature hardness for the

three ceramics in unirradiated and

This parameter is an indicator of the

INDENTATIONLENGTH (mm)

Fig. XVI-31.

Fracture toughness as a
diagonal
dimension of indentation for three single-crystal
ceram"ics in the
unirradiated
irradiated
condition.

presents valuea of fracture toughness, which

than is flow stress, a

plastic

and

to

form.

tion.

resistance

spinel (MgA1204),

hardness and fracture

irradiation.

,25,0

slopes of

negative

sense

18

~;=

A

$

Table XVI-XII compares changes

diffusivity, hardness, and fracture toughness at

200-g load, and volume changea for spinel, YAG,

and sapphire. These changes are interrelated and

UNIRRAOIATEOYAG,SPINEL,8SAFIWIRS! q OA
YAG,SPltEL,a SAWHIRElRRAOIATEO
T02.8x102'n/cm2AT 1015K:m0A

E-=k:

I

1-

of

is

are

the

the

I

3.2-

0.s

flow.

j ,2-

~ 1.6-

Figure

(A1203)

sapphire

sapphire

metals.62

toughness

irradiated

(Y3A15012),

N-
? 3.0 -
1=

Figure XVI-31

the properties

mutually consistent. No

Calculations based on

the damage clusters.

investigated by TEM

but they may

conductivity

irradiation;

observed in

suggests a

defects.

indicate

thermal

large

after

that

The

A

s

L

I

I

_O

in

the

and

I
Is

I

defect

thermal

electron

structure

phenomenon

function of

Y
25xti3

observed in

shown that irradiated

small loops or point

speculated

microscopy

decrease

crystais

evidence

further

found.

large

point

show

YAG

up

in

or

of

evidence of substantial

the thermal conductivity

change have shown that large concentrations of

point defects must also be present.

radiation damage in the form of damage clusters.

The exact nature of these clusters is not known,

alao

consist of

defect aggregates. It could also be

that local atomic rearrangement (e.g., cations on

the wrong sites) or metamictized regions make

The

(Table XVI-XII)

significant concentration of

increased upon irradiation, while

remained nearly unchanged in YAG and spinel.

the hardness curves increased in a

following

can be associated with

transmission

(TBM). TEM studies have

indents gave lower hardness readings for each

material, with the effect being more marked

This behavior may

sapphire contains a fine dispersion of small pores

with possibly a background dislocation network

which is presumed at least partly to result frcss

yield stress is affected more by radiation damage

intersecting loops cauaed by radiation damage.

1UNIRRADIATEOYAG,SPINEL,8 SAF?HIREq o A
'"i-

A

1 (mm)

The small changes in properties of spinel are

Fig. XVI-30.
Microhardness as a function of diagonal dimension
of indentation for three sirmzle-crvstal ceramics
in the unirradiated and irradiated condition.

concentrations of defects from either TEM

thermal conductivity measurements could be

q
chaf'&;::&y~

- 8/0.056

- 62~0.028

- 45/0.11

PROPERTY CHANGES FOR SINGLE-CRYSTfi SPI~L, YAG,AND SAPPHIRE AFTER

Thermal Diffusivity Change/Start"ng Value

Fracture Toughness

Value

Interpretation of the behavior of YAG

Changes in fracture toughness may

straightforward. Hardness and fracture toughness

in more than one way. Decreasing plasticity would

changes were small, and only a slight change in

expected to decrease fracture toughness by

indentation size dependence of these parameters

reducing the plastic work

The lack of

elastic modulus would change the fracture tOugh-

change argues against damage which

Modulus

Studies of

on yielding in copper

tion~4ar1e~e1;;;e;s11ci~~1e

that the unagglomerated vacancies did not affect

account for the observed increase.

yielding while vacancy clusters did cause harden-

likely cause for increased fracture toughness is

ing. The present results are consistent with

which could effectively increase

presence of point defects in the substructure

fracture toughness by blunting and/or lengthening

but not consistent with the presence of agglun-

2. Electrical Effects. Studies are under way

(111)

MsA1204

Material

vacancies

was noted.

'1203
(0001)

'3A15012
(111)

dislocation motion.

location line once it has

change of

+

be

-3

have

13.9

'alue

ever,

shown

+ 6/1.0

- 4/1.3

+ 2/15.3

+ 2/14.6

Hardneas

hardness

~,~m~?~g

+ 17/2.3

(at 200 g)

Change/Sta

is not as

substantial

TABLE XVI-XII

could affect

the effect of

nfcnl at 1015 K

%/GN m
(at 200 g)

IRMDIATION TO 2.8 X 10

ness in the same sense.

'1203'
duration. Tests

insulator will lead to a

to determine whether

since if the pores

cm at 104 pulses

the pore lattice,

very different

sub-breakdown

using pulaea

follow the

the crack.

only the

strength

vo1tage

small

Pro-

in-

the

(an

to

on

be

of

o

The

mos t

+ 2.5

+ 0.1

term.

also occur

~~~ct~rr~~~~

Changing the

decreases, how-

Volume
Change
(%)

few - microseconds

Three samples,

application

fatigue"

effect).

on dis-

ceramic

tech-

-105

romn

to

to

at

of

a

a

repeated

pulses

decrease in dielectric

"electrical

Measurements to date have been made

temperature along the c-axis of single-crystal

property changes. The large increaae in hardness

twelve unirradiated samples

may be accounted for by the background dislocation

showed a gradual decrease in dielectric strength

tangles alone (or possibly by the ccmbined effect

from a one-pulse value of -1.8 MV/cm to ..01.4WI

of tangles and pore array). The latter may act

(Fig. XVI-32).

impede dislocation movement by pinning the dis-

however , exhibited no fatigue effect up

reached a pore.

pulses. Statistical calculations based

blems with this model exist,

location density, as revealed by etch-pit

are considered to act as obstacles, not

niques, indicate that the likelihood of a sample

yield stress but also the rate of strain hardening

not showing electrical fatigue correlates roughly

should be increased; this expectation may

with the likelihood of dislocations being absent

consistent with the increase in the indentation

from the breakdown area of a dimpled sample.

size dependence of microhardness with irradiation.

(accounting for the thermal conductivity decrease)

erated vacancy clusters or small loops. The

fracture toughness would

similar absence of change in hardness.

The behavior of sapphire was

from either YAG or spinel, since

'or '1203 all
properties changed and considerable visible and

inferred damage is available to account for the

XVI-32). The number of pulses for these tests

the form of dislocation loops or defect clusters.

lMgA1204 shows no resolvable defect aggregates.
laggregated damage in all other ceramics takes

Breakdown measurements have also been made

the defects and the observed macroscopic swelling

after EBR-11 irradiation

are described. TEM results obtained to date may

swelling. Two samples irradiated at 925 K

lvoids are formed in A12~ at several elevated

pulsed dielectric strengths of 1.8 and 2.8 MV/cm;

temperatures and in Zr~-6% Y2~

for this ceramic; however, no

could be attributed to radiation damage under the

3. Nature of Radiation Dama=e in Ceramica.

Emphasis has been placed on TEM examination of

ceramics irradiated to 2.8 x 1021 n/cm2

These studies have been

out in conjunction with

Results obtained may

those from earlier atudiea at similar doses but

,give

of the aggregated defects formed in ceramics irra-

diated at elevated temperature. In Table XVI-XXII

dose which resulted in 3 to 6 vol%

1.9 NV/cm, while two samples could not be broken

broke down at 1.6 and

I

EFFECTIVE NO. OF PULSES TO BREAKOOWN

XVI-32.

Dielectric strength of sapphire at rocau temper-
ature as a function of effective number of vo1tage
pulses to breakdown. Arrowed data points repre-
sent samples which d~d not break down at the indi-
cated electric field.

~ ~

'

-

-

l

E

I

Jy=*=oF:--_L__

down at fields of 2.2

o UNIRRADIATED
l IRRAD AT 925 K
A IRRAD. AT I ICS3 K
I

varied between 102 and

ii
~z

Fig.

-3
E

.---- ----

:;::2, a

&
EZ2
l--

w

!#
%.1
10'

two irradiated at 1100 K

0

I

I

$-

'@

to

on

and

A1203

103.

(Fig.

I

showed

I
I

signifi-

2.3 MV/cm

1-2 x 1022

be summarized as follows:

consultant.

MeV) at 1015 K.

Results to date

temperature range.

W room temperature

different temperatures

test conditions utilized.

I
SINGLE-CRYWAL

Defect Structure
clusters
voids
voids
voids
loops
voids
clusters
clusters
clusters & loops

Irradiation
Temp., K

875

1025

875

650

1025

1025

loops
clusters
clusters
none attributable
to irradiation

TABLE XVI-XIII

clusters(a)

clusters

clusters

clusters

Resolved

10.I5

1015

Ceramic

'1203
(single cry.
and polycry.)

Y20
(po?ycry.)

Y20 -10% Zr02
(po?ycry.)
MsA1204
(single cry.)

'3A15012
(singly cry.)
Si.3N4
(polycry.)
SiAION
(polycry.)

si2C'N2
(polycry.)

~a ) At limit of detectability

L. W.

Prof.

Hobbs,

concen-

carried

(En&gt; 0.1

be added to

over a narrou

an overall picture

obvious degradation

1.4
1.9
1.4
0.2
1.8
0.0
0.3
0.0
-0.3
0.0
0.0
0.1

0.3

0.0

0.5

0.1

shcw considerable scatter but indicate a

Unlike metals, ceramics may contain a high

cant electrical fatigue effect

tration of point defects after irradiation at

DEFBCT

;TRUCTURES AND SWELLING VALUES FOR IRRADIATED CERAMICS

Swelling
Vol%

~

ceramics will undergo (n,p) and (n,ci) reactions

with thermal neutrons to produce the gases of

contain input from ceramic workers at LASL, BNWL,

An irradiation program plan has been

GA, LLL, and ORNL, and represent a consensus from

developed to use the mixed-spectrum Oak Ridge

the field and the fusion materials community. The

Reactor (ORR) to induce concurrently displacement

plan is intended to serve as a guide and basis for

damage, but the absence of high-energy neutrons

Thus such a study will allow H and He generation

such as those produced in a D-T reaction, prevents

interest

the concurrent transmutation-induced formation of

H and He gases. Since these gasea can affect the

nature of radiation damage, their presence is

the effort by the Materials and Radiation Effects

required in order to simulate more closely fusion

Branch of OFE to create an overall plan for fusion

elevated temperatures. This behavior can lead to

0 content of 50 isotopic percent

lattice dilation and significant swelling and may

(the maximum attainable), roughly two-thirds of

account for the 1.3 vol% change

the desired He level (also 394 appm) will be
14C

from

Fission Reactor Irradiation Testing. All

neutron irradiation tests conducted to date by

conditions can be

this project have made use of the EBR-11 reactor.

irradiated

of

4.

thus

SiAION.

Certain

Samples

isotopes

interest.

This material

irradiation at 650 K.

reactor damage effects.

isotopic percent

dose at 1100 K).

irradiation at

(an average

fusivity.

reactor.

N(nth,p)

With a

SiAION

EBR-11

with

the

14

a

reactor materials development, an assessment was

of insulator and ceramic problems and a

5.

of

to

two

the

two

The

made

with

last

gases

common

suffer

With a

ceramic,

elements

reaction.

primarily

attained,

REFERENCES

displacement

to have high

'n '1203 after

simultaneously.

Other Activities.

future activities.

separately or together.

program plan was proposed.

3 x 1021n/cm2

content

6 dpa

njcm2

99.5

4.

5.

6.

7.

of

1.

2.

3.

to

The resulting reports

As a contribution to

the

added

being

achieved

O(nth,a)

R.A. Krakowski, F.L. Ribe, T.A. Cou~-tas,and
A.J. Hatch, "An Engineering Design Study of a
Reference Theta-Pinch Reactor (RTPR)," Los
Alamos Scientific Laboratory report
LA-5336/ANL-8019 (March 1974).

R.A. Krakowski, R.L. Miller, and R.L.
Hagenaon, "Operating Point Considerationa for
the Reference Theta-Pinch Reactor (RTPR),"
Proc. Technol. of Controlled Nuclear Fusion
~, 359 (September 21-23, 1976).

J.P. Freidberg et al., "Review of the Linear
Theta Pinch (LTP) Concept," Los Alamos
Scientific Laboratory unpublished data (1977).

G. Sawyer (cd.), Proc. of Workshop on End-
Stoppering of Linear Magnetic Fusion Systems,
Santa Fe, NM, October 12-14, 1977.

R.L. Miller and R.A. Krakowski, "A Linear
Magnetic Fusion (LMF) Reactor Design Code:
lDRBURN," Los Alamos Scientific Laboratory,
to be published (1978).

R.L. Miller and R.A. Krakowaki, 'tAlpha-
Particle Trapping and Thermal Conduction
Constraints for the Ignition of Linear
Magnetic Fusion Reactors," submitted to Nucl.
Fusion (1978).

R.L. Hagenaon, "A Toroidal Fusion Reactor
Design Based on the Reversed-Field Pinch,"
dissertation, Iowa State University (1978).

damage and gas formation in a model

is known

strength in the unirradiated condition, and LASL

measurements on ARL SiAION have shown a good di-

electric strength ( .-330 kV/cm at room temperature

and N240 kV/cm at 875 K for pulsed voltages).

1015 K

results in a modest 22% reduction in thermal dif-

Swelling after EBR-11 exposure is low

of 0.5 vol% under the above condi-

tions, 0.7 vol% when irradiated to 2 x 10

at 925 K, and 0.2 vol% when irradiated to the same

A significant advantage of this ceramic for

ORR studies is that four different damage states

can be achieved, depending on isotopic content.

Calculations show the following:

nitrogen

15N (the maximum attainable)

in a year will sustain a damage dose of

with low concurrent gas formation.

N content of 6 isotopic percent,

394 ppm of H will be generated, primarily from

C reaction, thus achieving the

dpa/H ratio expected at the first wall of a fusion

R.L. Hagenson, R.A. Krakwski, and K.I.
Thamasaen, "A Toroidal Fusion Reactor Based
on the Reversed-Field Pinch (RFP)," IAEA
Conf. and Workshop on Fusion Reactor Design,
Madison, Wisconsin (1977).

Thomassen et al. "Conceptual Engineer-
K.I.
ing Design of a One-GJ Faat Discharging Hao-
polar Machine for the Reference Theta-Pinch
Fusion Reactor,!,~pR1-ER-246, project 469.
Electric Power Institute (1976).

9.

8.

16.

19.

27.

24.

29.

28.

23.

25.

26.

5 (1976).

33.

34.

35.

36.

32.

30.

31.

Fusion Forefront ~,

R.A. Krakowski, R.L. Miller, and R.L.
Hagenson, "Preliminary Energy Balance Con-
siderations for the Toroidal Magnetic (Bi-)
cusp : To~ac,t! LOS
Laboratory report to be published (1978).

Alamos Scientific

A.II.Boozer and M.A. Levine, "Particle Trapp-
ing in Magnetic Line Cusps," Phys. Rev. Lett.
~;

21, 1287 (1973).

I.G. Brown and W.B. Kunkel, "Tormac Reactor
Scaling," personal conxnunication,Lawrence
Berkeley Laboratory (May 12, 1977).

J.F. Decker, "An Evaluation of Exploratory
Fusion Concepts," to be published, 1978
(also, "A Status Report on the Current DMFE
Review of Alternative Magnetic Confinement
Systems," Trans. Am. Nucl. Sot. ~, 41
(1977)).

R.A. Krakowski and A.S. Tai, "A Simple Eco-
nomics Parametric Analysis of Fiasile-Fuel
Production by Fusion-Fission Reactors," Proc.
2nd DMFE Fusion-Fission Energy Systems Review
Meeting," Washington, DC, November 2-3, 1977.

Shang-Fen, G.W. Woodruff,and N.J. McCormick,
"A High-Gain Fusion-Fission Reactor for
Producing Uranium-233," Nucl. Technol. ~,
392 (1976).

K.E. Cox and M.G. Bowman, "The LASL Program
on Thermochemical Processes for Hydrogen
Production: Statua on October 31, 1977,''Loa
Alamos Scientific Laboratory unpublished data
(1977).
M.G. Bowman, "Thermochemcial Production of
Hydrogen from Water," International Symp. on
Energy Sources and Development," Barcelona,
Spain (October 9-21, 1977).

V.V. Pechovskii, "Thermochemical Decompo-
sition of Zinc Sulfatej" J. Inorg. Chem. USSR
(Engl. Trans.) ~,

1467 (1957).

V.V. Pechovskii, "Decomposition of Zinc and
Cobalt Sulfates in a Current of Air and
Sulfur Dioxide," J. Appl. Chem., USSR (Engl.
Trana.) 2,

(1958).

O.H. Krikorian and H.H. Otsuki, "Revised
Flowsheet and Process Design for the ZnSe
Thermochemical Cycle," Proc. of the DOE
Chemical Energy Storage and Hydrogen Energy
Systems Contracts Review Meeting,
Valley, Md. (November 1977).

J.L. Tuck, "Thermonuclear Reaction Rates,"
Los Alamoa Scientific Laboratory report
LAMS-1640 (1954).

F.L. Ribe, T.A. Oliphant, Jr., and W.E.
Quim, "Feasibility Study of a Pulsed
Thermonuclear Reactor," Los Alsmos Scientific
Laboratory report LA-3294-MS (1965).

3, 7 (August, 1977).

Hun t

10. K.I. Thomassen et al., "Ohmic Heating Systems

Study for a Tokamak EPR," Los Alamos
Scientific Laboratory Unpublished data (1977).

11. D.A. Baker, J.N. DiMarco, and A. Haberstich,

personal communication, Los Alamos Scientific
Laboratory, (1977).

12. G.E. Cort and R.A. Krakowski, "Heat Transfer

in the Lithium-Cooled Blanket of the Refer-
ence Theta-Pinch Reactor," to be presented at
Sixth International Heat Transfer Conference,
Toronto, Canada (1978).

13. F.M. Heck, E.I. King, Jr., and R.E.

Stillwagon, "Experimental Power Reactor Ohmic
Heating Energy Storage Study,"
Westinghouse Electric Corporation Fusion
Power Systems Department report WEPS-TME-038
(1976).

14. H.F. Vogel, personal communication, Los

Alamos Scientific Laboratory (1977).

15. R.A. Krakowski, R.W. Moses, R.L. Miller and

R.A. Gerwin, "Fusion Power From Fast Implod-
ing Liners," Workshop on Fusion Reactor
Technology, Madison, WI October 1977 (to be
published in Nucl. Fusion), Los Alamos
Scientific Laboratory report
(October 1977).

R. Gerwin and R. Malone, "Theory of Plasma
Heating by Fast Liner Implosion," submitted
to Nucl. Fusion (1977).

17. R.A. Krakowski and R.W. Moses, "Energy-

Balance and Blast Containment Consideration
for FLR "Los Alamos Scientific Laboratory
unpublished data (1976).

18. W. Fickett, "PAD, a One-Dimensional Lagrang-

ian Hydrocode," Los Alamos Scientific
Laboratory report I.A-591O-MS(April 1975).

L.F. Coffin, "Fatigue at High Temperature-
Prediction and Interpretation," Proc. Instn.
Mech. Engrs. (London) ~,

109 (1974).

20. S.1. Braginskii, "Transport Phenmnena in a

CcnnpletelyIonized Two-Temperature Plasma,"
Soviet Physics JETP ~ (33), 358-369 (February
1958).

21. A.R. Sherwood et al., "Fast Liner Proposal,"
Los Alamos Scientific Laboratory report
IA-6707-P (August 1977).

22. A. Hasegawa, T.Hatori, K. Itah, T. Ikuta, Y.

Kodama, and K. Nozaki, "Concept of a Fusion
Burner," Nucl. Fusion~,

40.

38.

44.

41.

43.

42.

39.

37.

50.

49.

48.

51.

46.

45.

47.

W.C. 1401kenhauer, Ed., "The Pacific Northwest
Laboratory Annual Report on Controlled Thermo-
nuclear Research Technology," BNWL-1685
(November 1972).
G.H. Miley, H. 'l'owner,and N. Ivich, "Fusion
Cross Sections and Reactivities," University
of Illinois report COO-2218-17 (1974).

D.V. Sivukhinj Reviews of Plasma Physics,
M.A. Leontovich Ed., Consultants Bureau, New
York!, 93 (1966).

S. Glasstone and R.H. Lovberg, Controlled
Thermonuclear Reactors, (Van Nostrand
Reinhold Co., NY, 1961), p. 345.

T.A. Oliphant, "Fuel Burn-Up and Direct Con-
version Energy in a D-T Plasma," Proc. BNES
Conf. on Nuclear Fusion Reactors, UKAEA
Culham Laboratory (September 17-19, 1969), p.
309.

G.E. Gryczkowski, "The Application of Neu-
tral-Gas Blankets to Theta-Pinch Reactors,"
Thesis, University of Michigan, to be
published (1978).

T.A. Oliphant, G.E. Gryczkowski, and T.
Kamnash, "Transient Charge-Exchange Effects
in a Neutral Gas Blanket," Nucl.Fusion ~,
263 (1976).

R.A. Krakowski, "A Survey of Linear Magnetic
Fusion Reactor Concepts," 3rd ANS Topical
Meeting on ControlLed Thermonuclear Fusion,
May 8-11, 1978, Santa Fe, NM (to bepublished).

T.J. Seed, "TRIDENT-CTR Code Manual,''Los
Alamos Scientific Laboratory, to be published
(1978).

T.J. Seed, W.F. Miller, Jr., and F.W.
Brinkley, Jr., "TRIDEN1': A Two-Dimensional
Multigroup, Triangular Mesh Discrete
Ordinates, Explicit Neutron Transport Code
Los Alamos Scientific Laboratory report
LA-6735-MS (March 1977).

B.G. Carlson, "Tables of Symmetric Equal
Weight Quadrature EQn Over the Unit
Sphere,!!Los Alamos Scientific LaboratorY
report IA-4734 (July 1971).

W.F. Miller, Jr., D.J. Dudziak, R.D. O'Dellj
and M.F. Gcmez, "Transport and Reactor
Theory, October 1 - December 31, 1977," Los
Alamos Scientific Laboratory report
LA-7131-PR (February 1978).

C.I. Baxman and P.G. Young, "Applied Nuclear
Data Research and Development, January 1 -
March 31, 1977," Los Alamoa Scientific
Laboratory report LA-6893-PR (July 1977)

R.W. Roussin, Oak Ridge National Laboratory,
personal cousaunication(September 1977).

2,

57.

56.

55.

54.

53.

52.

61.

62.

63.

64.

58.

59.

60.

II

B.R. Wienke, Donald J. Dudziak, and G.E.
Bosler, "RIGGATRON Nucleonics: One-
Dimensional Analysis," Los Alamos Scientific
Laboratory report LA-7183-MS (1978).

Donald J. Dudziak, S.A.W. Gerstl, and D.W.
Muir, "Application of the Sensitivity and
Uncertainty Analysis System LASS to Fusion
Reactor Nucleonics," Proc. Specialists
Meeting on Differential and Integral Data
Requirements for Shielding Calculations,
Vienna, Austria, 12-16 Oct. 1976.

D.M. Drake, G.F. Auchampaugh, E.D. Arthur
C.E. Ragan, and P.G. Young, "Double
Differential Beryllium Neutron Cross Sect
at Incident Neutron Energies of 5.9, 10.1
and 14.2 MeV," Los Alamos Scientific
Laboratory report LA-6257 (1976).

R.A. Schrach, R.B. Schwartz, and H.T. Heaton3
II, "Total Cross Sections of Silicon and
Beryllium," Bull.

(1971).

Am. Phys. SOC.

~,

G.F. Auchampaugh, S. Plattard, R. Extermann,
and C.E. Ragan, 111, "MeV Neutron Total Cross
Sections of Be,
of the International Conference on the Inter-
action of Neutrona with Nuclei, Lowell,
Masa., p. 1389 (1976).

B, and

P.G. Young and E.D. Arthur, "GNASH: A Pre-
equilibrium, Statistical Nuclear Model Code
for Calculation of Cross Sections and
Emission Spectra,!!Los Alamos Scientific
Laboratory report LA-6947

E.D. Arthur and P.G. Young, "Calculations of
15 MeV Neutron-Induced Charged Particle
Spectra on Stainless Steel Type 316," Trans.
Am. Nucl. SOC. ~,

E.D. Arthur and P.G. Young, "Cross Sections
in the Energy Range 10 to 40 MeV Calculated
with the GNASH Code," Symp. on Neutron Cross
Sections from 10 to 40 MeVj Brookhaven
National Laboratory-NCS-50681, p. 467 (1977).

Leona Stewart and Edward D. Arthur, "Neutron
Evaluation at High Energies - Problems and
Prospects," Symp. on Neutron Cross Sections
from 10 to 40 MeV, Brookhaven National
Laboratory-NCS-50681, p. 435 (1977).

P.G. Young, '!DataEvaluation and Nuclear
Model Codes, "Workshop Chairman's report at
Symp. on Neutron Cross Sections from 10 to
40 MeV," 13rookhavenNational
Laboratory-NCS-50681, p. 41 (1977).
D. McLean, Mechanical Properties of Metala,
(John Wiley and Sons, NY 1962),P. 367.

J. Galligan and J. Washburn, "Effect of Va-
cancy Clusters on Yielding and Strain Harden-
ing of Copper," Phil. Msg. ~, 1455-66 (1963).

ons

(1977).

C," Proc.

(1977).

C.I. Baxman, G.M. Hale, and P.G. Young,
"Applied Nuclear Data Research and Develop-
ment, January 1 - March 31, 1976," Los Alamos
Scientific Laboratory report LA-6472-PR
(August 1976).

D.H. Bowen and F.J.P. Clarke, Chemical and
Mechanical Behavior of Inorganic Materials,
A.W. Searcy, D.V. Ragone, and U. Colombo,
Eda., (Wiley-Interscience, NY,1970) pp.
585-613.

G. P. Boicourt, R. T. Buck, G. I. Chandler, R. S. Dike, C. F. Hammer, K. W. Hanks

L. D. Hsnsborough, R. W. Kewish, Jr., K. J. Kutac, J. G. Melton,

and W. C. Nunnally.

The design, simulation, and development of

the ZT-40 system and components continued during

1977. The basic system consists of 12 identical

circuits, one of which is shown in Fig.

All of the circuits are connected in parallel at

the circular mixer structure that surrounds the

plasma torus. The toroidal plasma experiment has

12 poloidal current feed slots and 12 seta of

toroidal field coils. Each of the poloidal cur-

rent feed elota snd each of the 12 sets of

toroidal field coils will initially be connected

in parallel at the mixer output.

system is designed to permit many series-parsllel

combination of the poloidal current feed slots or

the toroidal field coils, all of which are to be

driven from the mixer output.

The load flexibility permits the current

rieetimes to be varied over a very large range,

and thus permits a large range of conditions in

the experiment. The capacitor bank parameters for

the ZT-40 system are listed in Table XVII-I.

The continuing design of the ZT-40 system in

1977 included the design of the transient "snub-

mixer

A.

ZT-40 SYSTEM DESIGN

Ril
D1'--Q+

circuits, the

ber"

ZusnAN(

TABLE XVII-I

ZT-40 EXPERIMENT

CAPACITOR BANK PARAMETERS

Energy
Voltage
No. capacitors
No. start switches
No. crowbar switches

XVII-1.

However, the

Poloidal Bias Bank

Poloidal Main Bank

XVII. ENGINEERING

Toroidal Main Bank

teristic impedance by

circuit and

system

Energy
Voltage
No. capacitor
No. start switches
No. crowbar switches
Poloidal Power Crowbar Bank

Energy
Voltage
No. transformers
No. capacitors
No. start ignitrons
No. crowbar ignitrons
Toroidal Preionization Bank 1

Energy
Voltage
No. capacitors
No. start ignitrons

Toroidal Prefonization Bank 2

Energy
Voltage (two stage Marx)
No. capacitors
No. Marx ignitrons

Energy
Voltage
No. capacitors
No. start switches
No. crowbar switches
Toroidal Power Crowbar Bank

Energy
Voltage
No. transformers
No. capacitors
No. start ignitrons
No. crowbar ignitrons

275 J
5 kV
12-1.85 PF
12-Size A

1.5 MJ
10 kV
180-170 VF
36 Size D
36 Size D

1.0 MJ
10 kV
12-20:2
120-170 VF
24-Size D
24-Size D

0.67 MJ
50 kV
288-185 UF
288-Fig. XVII-11
144-Fig. XVII-12

0.67 MJ
50 kV
288-1.85 IJF
288-Fig. XVII-11
144-Fig. XVII-12

227 kJ
30 kV
144-14 PF (?20 kV
48-Size D

3.1 MJ
10 kV
12-14:2
480-170 PF
36-Size D
36-Size D

the snubber resistance,

XVII-1.

The

hardware, and the power crowbar circuits, all of

which are illustrated in Fig. XVII-1.

'lhesnubber circuit was designed to limit the

switching transients due to impedance mismatch at

the coaxial cable terminations. The baaic snubber

electrical circuit ia shown in Fig.

cablea are terminated in their composite charac-

Fig. XVII-1.
ZT-40 circuit-one feedslot.

A series capacitor, Ce, ia ueed to prevent

chosen to supply the sustaining voltage of

large energy drain after the first few hundred

The inductance, Ls, shown in Fig.

because

could be

XVII-1 is inherent in the system and is minimized

Electrolytic capacitors

by paralleling many RLC circuits. The value of C~

were too costly and bulky because of their voltage

tests on a prototype system.

have low total leakage inductance as seen from the

the system using NET-2 indicated that the optimum

secondary terminals in order

location for the snubber circuit was at the mixer

energy delivered to the low-inductance load.

is determined by simulation and by experimental

mixer

I&.

The

between

compromise

nanoseconds.

malfunctions.

tional RLC units.

the capacitor banks.

crowbar switch.

XVII-1.

bank.

Fig .

can

output. The transients from the cable between the

mixer output and the load are not suppressed, but

there is insufficient room at the load for addi-

The mixer system for ZT-40 was designed to

"mix" or to average the individual load currents

and voltages when a portion of the capacitor bank

voltages at any feed slot and minimizes current

differences in the parallel loads.

system was designed using a scale model mixer

structure and SCR-switched capacitors to simulate

The final design was a

reliable

fabrication techniques and mixing performance.

The model tests indicate that only a 5% maximum

difference in load currents will occur for a

localized malfunction of 1/12 of the capacitor

A cross section of the mixer is shown in

Essentially, the mixer

consists of two, parallel flat-plate transmission

lines that encircle the load torus. The load cur-

rent traverses the mixer structure radially when

the capacitor banks operate normally.

event of a capacitor bank malfunction, the current

travel circumferentially to

individual load currents and voltages.

The power crowbar (PCB) system for ZT-40 was

designed to sustain the initial load currents of

107A within *7Z for 250 W,

XVII-2. In general, the 1$ and the Ie systems are

very similar, and thus a basic PCB circuit for

both systems is shown in Fig.

capacitor bank establishes the desired initial

current in the load and is shorted out with the

A low-leakage inductance trans-

former in series with the load supplies a voltage

to cancel the circulating current voltage drop and

sustain the current. The-low leakage transformer

o

o

I

was

%!

plate

about

*7%
I

prevents

excessive

collector

The mixer

capacitors

1000-1500 v

high-density

The simulation of

step-down transformer.

and current limitations.

I

XVII-3. The 50-kV

shown in Fig.

HIGH
VOLTAGE

equalize

In the

system

the

as

I

I

I

a

250 U

via

used

10-kV,

PC8 SOURCE

economical

to maximize the

lhe transformer nust

LOAO
INDUCTANCE

I

Fig. XVII-3.

Basic ZT-40 power crowbar circuit.

Fig. XVII-2.
General ZT-40 current waveforms.

The low-leakage transformer and the crowbar

XVII-5. The Matched Mode utilizes

ewitch are the main components to be developed for

the separate biae bank of Fig. XVII-1 to provide

simulation was used to determine the required

The Padua Mode circuits are shown in Fig.

rameters for these components. The parameters

XVII-4 and representative waveforms are shown in

the PCB system and the required parameters for

Fig. XVII-6. The Padua Mode can use the main 10

PCB transformer and the PCB crowbar switches

bank to provide an additional fast-rising bias

B. ZT-40 CIRCUIT-PLASMA SIMULATION

The Slow Mode I method of operation i8 also

Computer simulation of

termed the Self-Reversal Mode.

oircuit, including plasma dynamics, was a major

tor banks are not used in either circuit. The 10

The plasma-circuit interaction

circuit PCB system is also shorted and the biae

has a great influence on the overall system. The

bank is now used for reversing the field. The 10

simulation provided circuit component requirements

circuit utilizes the preionlzatlon (PI) oircuit to

information

initiate the plasma and the 10 PCB syetem to drive

physics-related results.

the plasma current. The Slow Mode I circuits are

operating

shown in Fig. XVII-7, and a typical waveform is

lated: the Matched Mode (MM), the Padua Mode

(PM), Slow Mode I (SMI), and Slow Mode II (SMII).

Slow Mode II is similar to the Matched Mode

These modes are compared in Table XVII-III with

and ts also termed the Aided Self-Reversal Mode.

respect to their basic time constants and maximum

The basic circuit that is used for Slow Mode II

listed in Table XVII-II.

as

as

and

four

well

modes

ourrents.

operating

Basically,

the ZT-40 PCB system.

activity in 1977.

No. required
Leakage inductance
Secondary current
Pulse duration
Primary voltage
Turns ratio
Core flux

Peak current
Average current
Inductances
Coulomb capacity
No. required

No. required
Start-size D
CB-size D

High Voltage Crowbar

Coumonent Reouirem~

Energy
Voltage

Capaoitor Bank

Switches

Low-Leakage Transformer

a slow-rising initial toroidal field.

The main bias bank is not used in the

representative waveforms of Fig. XVII-6.

is

are

pa-

the

for

the

the

were

modes

about

simu-

field.

possible

operation

shown in Fig.

entire ZT-40

shown in Fig. XVII-8.

Computer circuit-plasma

4 nH
500 kA
1 ms
10 kV
20:2
0.125 '?s

4 nH
750 kA
1 ms
10 kV
14:2
0.187 VS

200 kA
50 kA
50 nH
50 c

200 kA
50 kA
30 nH
50 c

in future simulations.

core saturation.

shown in FIR.

shown in Fig.

the Poyntlng

3 MJ
10 kV

1 MJ
10 kV

2X12
2X12

3x12
3x12

XVII-4

flux

XVII-10

XVII-9

I.e

.

to

shown

in

Fig.

The ~

XVII-4.

The 50-kV capaoi-

and the 10 loads,

possible and

determine

operation.

series-parallel combinations to produce very large

load inductances for SMI and SMII

The plasma was modeled using a snow-plow

formulation for the majority of the simulations to

date. A more appropriate, diffuse boundary plasma

formulation is being programmed and will be used

The power crowbar transformer utilizes an

iron core to increase the transfer efficiency.

The transformer operation using the flux model

the required core flux and the oircuit effeot of

was simulated to determine

In all the oircuit-plasma ZT-40 simulation

studies, the Poynting flux of the plasma system

was studied. The computer simulation program used

optimum bank switching times and to correlate

future ZT-W operation with other experiments.

The circuits for the Matched Mode, the Padua

Representative waveforms for Slow Mode 11 are

Mode, and Slow Mode II are shown in Fig.

and representative Matched Mode waveforms are

respectively, are arranged in several possible

TABLE XVII-II

ZT-40 POWER CROWBAR SYSTEM

AND COMPONENT REQUIREMENTS

Power Crowbar Bank Switches

ZT-40 MODE CHAIWCTERISTlCS FOR NOMINALTRAPPEDFIELD OF 0.6 TESLA

(OPERATING POINT)

r RISE
10 BANK (MAX)

[#BANK (MAX)

Iz BANK (PAX)

COMPUTING RUNS

Bz

BIAS

'RISE

'FLAT

'FLAT

'!:?;:;!

'RI SE

'REVERSAL

'yw)

COMPONENT DEVELOPMENT

start switch of Fig.

paragraphs

following

c.

0,6 T
840, US

0,344 MA

-005 T
---

20,3 YS
MA
3.

w

006 T
18.3 W
387 M

Ms

T

37

ys

56

ooO

0.6-

435,

--2000

-15000

-40000

006 T

4.0 ps

T
#s

3.44 MA

7.0 MA

77, us

0.353MA

-300. L&

005 T
---

TABLE XVII-III

---
---
---
---

0054 T
1160. /lS

0,6 T+0,5T
3.4 /Js

006 T
2.6 JIS
8,2 MA

0,59 MA
J30000 MS

0,0 T
778 JJs
---

2,3 jlS
26, M
-300, #s

006 T
1.9 ps
1205 M
-200s MS

1977.

02 r']-

-0.6
o

area

Slow

wide

Mode

and

the

The

The

-04 -

I

1

-0,6

I
so

04 -

0.2-

-04

0.0

150

II

I

Lil/

TIME [#S)

TIME(@

too

300

400

I20

I10

The component development program for ZT-40

was another major activity during

indicate

encompassed by this development program.

1. ZT-40 Field Distortion Start-Gavu

ZT-40 application because a wide voltage operating

XVII-11 was chosen for the

Fig. XVII-5.
Representative Matched Mode waveforms.

Matched Mode, Padua Mode,
circuits.

Fig. xVII-4.

MAONETICFIELD PRNILES

I

0.003
TIME

Fig. XVII-8.
Representative Slow Mode I waveforms.

MAGNETICFIELDPROFILES

I

[

TIME [#ts)

MAGNEllC FIELG PRQFIIES

I

I

ae~

0, z -

0.4 -

0.6 -

on,

I

-02 F

I

I

1
so

1s0

t

-081
o

-0.8

-06I

ZT-40

ZT-40

.,

\

I

I

I

"

I

I

\f

E~

I

I

I

I

I

I -

0.1

0.3

-0.3

O.m

0.4

0.001

0.002

0.2 \-

I

I

I

~ -o.

-0.2 -

0.0 --

J
I 00

0.5-

OEl

400

250

300

350

I
4s0

'"'~
0.6-

'4

!40.6

I
QOO066

I
00C062

-0.6
0.00078

Y
1-
;

0.oo1o

PR?fRY

-02 -

-04 -

a2 -

0.4 -

"'~

Qo

0.4

Q6

1

ao

1

I

Fig. XVII-6.
Representative Padua Mode waveforms.

TIM~(@

SLOW MODE I

1, -BZ CIRCUIT

Fig. XVII-7.
Slow Mode I circuits.

SLOW MODEl

lZ-B, CIRCUIT

I

I

I

I

I

I

#

0.006

0.007

Lloo5

0.004
(S)

CS.0030 Qoo40

I
QOCQ96

I
0JX)090

I
0DO094

'e

/

I

--.o..~

0.0020

K
o

-C.4

TIME (S)

TIME (S)

Fig. XVII-9.
Representative Slow Mode II waveforms.

tmrhrb'"g

B(T)
Lx IO*(H)

0.6

1.0

1.4

1.8

/

/

/

-

B

L

'"8

&gt;~

-----

-400

-800

I
-1200

electrodes with

average jitter

distortion

10-50 kV

spacing

switch

range

main

for

and

was

was

an

of

a

voltage

Fig.

-- ---4

-- MODEL

H (AMP-TURNS)

--- MEASURED

(SHIFT= 136 A-T)

since reduced the

damage is 1.25 cm.

modification.

interconnect

for nominal

inductance

evaluated

seems to

follows.

lifetime

desired.

10-50 kV

(f) the

scheme

shown

about

range

field

5 us

caps

moly

and

and

The

the

in

of

to

of

Fig. XVII-10.
Transformer simulation flux model.

thoroughly with the following results: (a) the

(a) The electrode spacing that permits

tested

test. The results of the development test are as

press-formed

operation from 10 to 50 kV at

operate about 5000 shots at a current of 100 kA

pressures and minimizes erosion and spark plug

before destruction; (b) the nylon insulators are

viable until iO 000 shots when they begin

(b) The average operational

paasive

track; (c) spun moly electrode caps can extend the

approxtiately 4000 shots at 50 kv, 60 kA, 7.2

electrode life to 20 000 shots; (d) the optimum

with the 1.25-cm spacing and a trigger circuit

reasonable air pressures is 0.635 cm; (e) the best

3. ZT-40 Power Crowbar Transformer.

switching performance occurs with negative capaci-

initial tests of the first prototype power crowbar

tor charge and a negative trigger pulse;

transformer were completed in 1977.

3 kV/ns, 100-kV trigger pulse planned for ZT-40

indicating the transformer winding geometry is

start switch produces an average delay of 50 ns

Fig . ZVII-13.

of

indicate that a transformer with a total leakage

operating conditions; (g) the prefire rate ia

L nH,

between 0.1 and 0.4%; (h) a brass modification of

secondary with the primary shorted, is possible

the electrode geometry has

with a single-turn secondary.

prefire rate to less than 0.1% up to 25 000 shots.

initial primary to secondary insulation scheme

_

-p

OIL lNSU_ATION

XVII-12

TRIGGERCABLE

SPUNMOLYBDENUM

test and

TuNGSTENALLOYOISK

After the initial

F

In addition, the

reasonable air

'lhe winding

A schematic

measured

initial

results

primary

winding

crowbar

The

The

the

the

gap

in

is

c,

/

ZT-40 field distortion start switch.

XVII-11.

2. ZT-40 Passive CB Switch. Life

modification studies of the ferrite-loaded ZT-40

passive crowbar spark gap shown in Fig.

were completed in June 1977. The pasaive crowbar

switch waa tested at voltages of 60 kV with a peak

current over 200 kA, an average maximum current of

140 kA, and a C load of 17 C.

teat, the average maxfmum current was reduced to

60 kA and the voltage reduced to 50 kV for life

function satisfactorily.

and

4. ZT-40 Power Crowbar Spark Gap. The first

prototype of a high-voltage, high-Coulomb spark

gap for ZT-40 power crowbar duty was procured in

'he test arrangement for the switch was

fabricated and the gap installed. Initial teats

will begin in 1978. The power crowbar spark gap

is shown in Fig. XVII-14.

5. ZT-40 Trizger Initiation System.

trigger initiation system for ZT-40 is diagramed

The digital time-delay system

in the screen room controls the timing of various

events in the ZT-40 circuit.

translates the timing pulse from the screen room

to the experimental area using an injection laser

cable

air-powered,

receiver-amplifier generates a 5-kV pulse that in

turn triggers a 200-kV pulse generator, which

initiates the various high-voltage systems of

ZT-40. The fiber-optic cable and the air-powered

receiver permit the pulaer system to be completely

Fig. XVII-12.
pasaive crowbar

ZT-40

K?CABLES
FROM BANK~

tracking.

,(

Further

1977.

to

be

to

and

lhe

CABLE

1977.

design

a CABLES TO

in Fig.

switch.

modified

XVII-15.

isolation.

fiber-optic

suppression.

r TRIGGER
lQAD7

"'"R'-

nROIKWOU,IOF 3T

The Marx

The Marx

switch.

stage

was

The

=7

PRIMARYWINLNNG

is

The

for

a small,

generator

light-pulse

'he pulser system

isolation and EMI

coupled by

the

\

insulation technique need

decoupled from ground, thus affording complete

prevent tolerance build-up problems and voltage

transformer

200-kV pulse

restricted by limited personnel availability in

eight-stage, Marx circuit, termed a micromarx.

switches are

ultraviolet light from the first-stage triggered

circuit incorporates barium

titanate capacitors for a total stored energy at

Fig. XVII-13.
ZT-40 power crowbar transformer winding geometry.

Fig. XVII-14.
ZT-40 power crowbar switch.

SHIELD,TO Sl(AOOw
lNswlrm FROMARC

AIRM.E'f
PORT--

XVII-15.

initiation

30-kV charge of 9 joules.

rises to approximately 200 kV In 20 ns with a

The risetime can be reduced to

lIJns tith an internal series pulse sharpening

'IheMarx electrical circuit is shown

XVII-16, and the generator Is shown in

The fiber-optic coupled pulser

system and the micromarx generator were operated

for 15 000 shots during the crowbar switch test.

k

I

J

I

~

___

Fig.

USER

ZT-40

Fig.

trigger

---- ____

F.O.ORIVES

SCREENSOOMT

in Fig.

OPTICAL
kly~TO~

XVII-17.

spark gap.

DIGGT:L
DELAY
SYSTEM

jitter of *1O ns.

OUTPUT
VOLTAGE
T200 kV,
20 ns

.---------

I
l-+

*--,

&gt;

.20MPRESSE0m INPUT

--

D.

system.

RG-8CASLE

T~\OTQEMR

RO-!SCABLE

;L3#
M&Rx

installed,

system was

initiated in 1977.

The micromarx output

sYatem perfor~nce,

standard CAMAC

determined by

subsystems

insulation

operation,

FACILITIES

PEAKING GAP

hardware

circuit

value

cab1e

via

and

and

or

pF

psig

lQwT&og;:;.

CHARGE VOLTAGE
lNPUT,f30kV

RESISTORS:lOM~
CAPACITORSV&J

AIR PRESSURE:50

-OPTIONAL

Fig. XVII-16.
ZT-40 micromarx electrical circuit.

was

This

system

control

h.

generator.

evaluation was

stripback, (2) snubber

(3) trigger

equipment.

spark gap

assembly

monitor

The

ZT-40 micromarx 200-kV pulse

Fig. RVII-17.

ZT-40 ENGINEERING PROTOTYPE

A major effort in hardware

consisting of l/24 of the ZT-40 circuit.

A prototype was constructed

a

fabricated, and checkout began in December 1977.

The prototype will be used to test the following

components: (1) mixer

components,

(4) master trigger spark gap

(5) fiber-optic

system, and (6) power crowbar system.

E. ZT-40 MACHINE CONTROL AND DATA ACQUISITION

The ZT-40 experiment will be controlled and

data will be accumulated with a Prime 400 computer

interface

arrangement of the computer control system is a

major concern requiring very careful specification

and placement of the screen room, the signal

conduits, the screen-room penetrations, the con-

trol and monitoring interface hardware, and the

ayatem components. The ZT-40 experiment haa been

very thoroughly planned as a system to minimize

ground loopa and to reduce the noise level to a

the signal-to-noise ratio

required in the screen room.

q
(cid:228)
ia

an
accomplished
room
parallel, aerial highwaYn fiber-

via

This type of fiber-optic control

Fig. XVII-18.
ZT-40 control and data acquisition system block
di.agram.

screen room

Erik A. Lfndgren.

consists of a layer of 24-gauge galvanized steel

layer of 24-gauge silicon steel.

location of all diagnoatio conduits, charge leads,

carefully determined to minimize ground loops and

Ground loops were also mlnl-

mized in the screen room by decoupling individual

control rack power with shielded isolation trana-

fomnera. In addition, the electrical input power

is filtered and the air-conditioning system is

isolated and conducted through the screen room

walla via wave-guide-below-cutoff structures.

The control function of the computer located

by

the

The

and

leads,

and a

control

optic link.

manufactured

associated noise.

in
screen
18-channel IIbyte

1?. PULSE-POWER

development,

DEVELOPMENT

component

areas

The

in

l&Aka4-ul ~' '---"""""

I

la,,"

-'

'"..

~

a

is

-w, -A,

The

The

has

room

wall

been

Each

planes

ground

double-wall

..--.--------_------
.~

-- .
I
/ l-=-r-l=

and advanced pulse-power

capacitor and

COMPONENT

component

ignitron

Z-pinch

system

system

AND

and

and

The

(3)

-------

--

,

.

.

I

I

.lOcu,..

--,"

I

I

__________

The development effort

switch

The pulse-power engineering development for the

ZT-40 system is detailed in the ZT-40 engineering

section of thie report.

included the following areas:

ZT-40 System Development

a. Basio System Circuit

b. Power Crowbar Circuit Design

c. Snubber Circuit Design

d. Mixer System

ZT-40 Circuit-Plaama Simulation

a. Hatched Mode study

b. F'aduaMode study

c. Slow Mode 1 study

d. Slow Mode II study

e. Plasma Model

f. Poynting Flux studies

g. Transformer Core Model

ZT-40 Component Development

a. Field distortion start

b. Passive crowbar switch

c. Power crowbar transformer

d. Power crowbar switch

e. Trigger initiation system

(1) digital time delay

(2) fiber-optic air-powered pulaer

200-kV micromarx pulse generator

f. Prototype facility

scheme was chosen becauae it will permit system

shutdown In a 200-ma cycle time and still be cost

effective. The fiber-optic highway is required to

operate at 5 megabaud over a 40-m path.

ayatem developed at LASL operates at 22 megabaud

with a signal-to-noise ratio of about 70 for a bit

error rate of 10-32 over the required dietance.

The fiber-optic system can alao be used ae an

analog link with a 3-dB bandwidth of 8 MHz.

A fiber-optic spark gap monitor syetem was

also developed to determine the firing time of the

systemls 288 crowbar spark gaps. The system looks

at the spark gap light via a fiber-optic cable.

electronic system records a current pulse

corresponding to the spark gap light time in a

given window with a duration of 25-100 ns.

data are stored and read out at a later time by

the computer system. A block diagram of the ZT-40

system is shown in Fig. XVII-18.

TECHNOLOGY

The pulse-power engineering development group

at LASL was involved in the following four major

1977: ZT-40

high-density

development,

switch life testing,

technology development.

2. HLgh-DensLty Z-Pinch System and Component

The High-Density Z-Pinch (HDZP) ex-

periment at LASL was designed to produce a time-

varying current waveform greater than that shown

any

high-pressure gas load. The experimental load is

a small-diameter (100-pm) current filament between

20-cm-diameter coaxial return conductor. The cur-

rent filament is to be initiated with a 20-J,

Q-switched Nd:glass

One version of the load geometry Is

The load inductance is on

the order of 100 nH and the Initial I from Fig.

XVII-19 is on the order of 5 x 1012 A/s. Thus the

load voltage at t=O must be on the order of 500

When the pulse-power source inductance ia

considered, the power supply voltage must be on

the order of 1 MV and the resulting maximum cur-

rent is on the order of 1 MA.

A study was undertaken in July 1977 to deter-

mine the system nwst suited for the desired HDZP

A low-inductance Marx system, a

Marx-charged water-capacitor system, and a low-in-

ductance Marx-charged water

system were investigated and evaluated.

the systems were simulated using NET-2 to deter-

operattng

at

in

kV.

two

Fig .

20-ns

spaced

XVII-19

electrode.

electrodes

Development.

power supply.

shown in Fig. XVII-20.

'"l-----l

realistic

o-l

1-Z
:

v

mine

,
0.2

0.I

TIME

o.5-

z
~

-

Fig. XVII-20.
HDZP load geometry.

'Ihe low-inductance

a

a

v
E

in

one

with

time

laser

small,

through

cm-apart

circuit shown in Fig.

water-insulated

requires a 1-W

transmission

performance.

six-stage,

The water

parameter

circuit.

voltage

modules

jitter.

in Fig.

All of

XVII-19;

HDZP.

voltage

Fig .

line

open

0.5

output

I
0.4

but

The

and

The

Marx

*t

0.6

RES

Marx

syatern

GROUND
ELECTRODE

LASER
INITIATION
WINDOW

energy-storage

inductance.

'he HDZP

because

schemes

*2O ns

system

space

Marx

was

two

use

of

for a 100-nH source in-

ductance. In order to achieve the re~uired induc-

tance, many Marx circuits must be paralleled and

triggered with very low jitter into an initially

other

intermediate

systems, require less Marx voltage and triggering

simultaneity, and permit more Nsrx

The system using the water-insulated transmission

line produced the best waveform with respect to

considerations, the Marx-charged water-capacitor

XVII-21 was chosen for

capacitor system produces a

risetime across the gas load that is

approximately four times slower than the water

load system using the same Marx bank.

600-kV

designed using commercial components.

Marx system consists of 12 parallel Marx circuit

that are to be triggered with

The Marx module circuit diagram is shown

XVII-22 and a side view of the module is

I
0.3

(P9w)

Fig. XVII-19.
HDZP minimum current waveform.

RESISTORS /

Fig. XVII-23.

configuration.

Fig. XVII-21.

Basic HDZP circuit.

R3

I-iTl

R2

~

R2

MARX SWITCHES

shown in Fig.

TRIGGER
INPUTS

R5

L

I

\

I

II

w

I

Marx

The

module

zmR=

XVII-IV.

/-SfWRKGAPS&gt;

OUTPUT
VOLTAGE

C=cl
Rl=
R2=
R3=
R4.
R5=

TWO, O.1-uF, 20-nH,

CHARGE
VOLTAGE
oH"

capacitance

Inductance

Parameter

Physics

Current

Voltage

Energy

Marx

--

Fig. XVII-22.
HDZP Marx module circuit diagram.

XVII-23.

100-kV, 50%-reversal capacitors are used for each

module stage. At 1 MA total output current, each

capacitor is required to yield 42 kA. A capacitor

reversal of 50% was required to prevent excessive

capacitor damage due to fault modes.

International T670 100-kV, 100-kA, 60-nH spark gap

switches are used between capacitor stages. 130th

components, the spark gap switch and the capaci-

tors, will be operated near their maximum ratings.

Future systems of similar nature will require the

'%kEiTElR2

energy

water-insulated

The large

achieve

600 kV

order

75 nH

72 kJ

I MA

in

to

development of better components. The parameter

of the HDZP Marx system are listed in Table

intermediate,

storage line was designed, as illustrated in Fig.

XVII-24, using flat aluminum plates because the

capacitance can be varied eaaily.

diameter of comparable coaxial water capacitor was

incompatible with the extremely small size of the

gas load. The energy can be stored close to the

load when a flat-plate capacitor is used and thus

reduce the source inductance. The maximum value

of the water capacitance was made equal to the

efficient Marx energy transfer. The capacitor in-

ductance is approximately 40 nH, which is only a

factor of 2 better than the Marx system alone, but

different operating modes can be obtained by

TABLE XVII-IV.

HDZP MARX PARAMETERS

MARX TRIGGER SYSTEM

CHARGE RESISTORS

w f'

Fig. XVII-25.
HOZP Marx trigger system.

Fig. XVII-24.
HDZP system diagram.

varying capacitance and load current initiation

times in order to uae the capacitor aa a peaking

circuit and vary the initial current slope over a

The Marx trigger system nmst insure erection

of all Marx modules within ~20 ns. The water ca-

I

OIL INSULATION

~MARX MCCWLE

wide range.

modification

teated in

vo1tage

electrode.

level.

indicate

riaetime

stage

stage

those

1977.

Marx

module

the

The

seen

Only

will

very

of

two

the

by

/

F

1 1-

~MARx SWITCHES

NOT SHOWN

INPUT
TRIGGER

TRIGGER MARX

WATER INSULATEO
ENERGY STORAGE

u :g&J

stages simultaneously.

through the trigger Wrx

and jitter in

12-parallel,

The peaking

of 15 ns.

gered by

Fitch.1

provide

XVII-25

XVII-25

200-kV

trigger

trigger

SWI t ch

36-w

delay

Marx

by

the

of

a

pacitor appears as a very low impedance to the

triggered. The first trigger Marx switch is trig-

parallel Marx modules and facilitates triggering.

an external 200-kV micromarx; then,

Several methods for triggering the Marx system

switching transient of the first gap 1s used to

were evaluated with the NET-2 circuit analysis

trigger the second trigger Marx gap with a delay

code. The trigger system shown in Fig.

The second trigger Marx stage switch

waa then selected, designed snd fabricated, and

cannot overvolt because transient overvoltages are

triggering scheme is a

severely damped by the composite coaxial cable im-

suggeated

pedance (3 W) across the stitch. Thus, the two-

Basically, a fast-erecting trigger Marx is used to

stage trigger Marx erects sequentially and very

initiate shorted cable trigger generators at each

reliably. The 12 coaxial cables from trigger Marx

cables, are charged to 100 kV, and when shorted by

attached to trigger Marx Switch 1 such that the

The

switch,

Switch 2 can be shortened with respect to those

trigger pulses arrive at the first two main Marx

open-circuit voltage pulse for triggering the main

Initially, the trigger

Marx switches. The peaking gaps on Fig.

cables will be the same length producing a 15-na

isolate the apark gap trigger electrode potential

delay between Stage 1 and Stage 2 switches. The

from the cable potential until the fast-rising

trigger Marx system also provides isolation from

trigger pulse breaks down the gap.

ground for the shorted cable pulse generators

gap also serves to decrease the trigger pulse

stage capacitors.

Evaluation of the high-voltage components and

geometries using computer solutions of Laplace's

be used because prototype tests of one Marx

equation was a major activity during the last

Marx erection time after the second Marx switch is

of the HDZP system were evaluated, and mechanical

quarter of 1977. The voltage stress of many parts

Marx

atages

little

L

T

t--

MARX

PEAKING
GAPS 1

q
*

The

test

system.

system.

acceptable levels.

inductance capacitor.

failure mode.

ignitrons and

continued

program

Many

used

The

for

D

designs modified to reduce dielectric stress to

coil, which has accumulated 40 0000 ahota at a

stored energy of 85 kJ, failed and was repaired

A prototype RDZP system consisting of one

and reinstalled, cauaing a two-month delay in

Marx module and a 60-na, 6-w water line wsa

testing. The test program also evaluated methods

constructed during the laat quarter of 1977.

of making reliable, high-current connections to

control and power supply system was designed and

the ignitrons and headers.

fabricated, and initial operation has begun.

The 60-kV, 2.8-PF, 20%-reversal high-density

The prototype HDZP system will permit full

test bay was operated a total of only 14 000 shots

voltage simulation of the final HDZP eyatem at

personnel shortage.

1/12 the peak current. High-pressure hydrogen gas

reached end-of-life at 49 000 total shots.

breakdown studies and laser channel formation

of the test shots were used to test a spark gap

studies will be carried out on the prototype

trigger circuit modification.

From an engineering viewpoint, the Marx

module operation can be monitored, evaluated, and

low-inductance

high-voltage

modified to produce a reliable unit for the larger

The standard 60-kV, l.85-PF capacitor

has a series inductance of approximately 24 nH, of

3. Capacitor and Ignitron Switch Life Test-

which approximately 20 nH ia the header. Two ca-

program

for

high-voltage coaxial cable, and ignitrona at LASL

fabricated by Maxwell using a polypropylene paper

ia unique nationally. The teat program involves

system and a standard paper system. During 1977

life testing of components and examination and

the capacitor waa assembled and the inductance

evaluation of the malfunctioning units to deter-

meaaured at approximately 12 nH.

mine the reason for failure.

the failure mode can be eliminated by changing the

design or the manufacturing process.

by Maxwell

be initiated

and

are then again life tested to determine the next

components are listed in Table XVII-V.

teat

constant effort over many months and even yeara.

G. ADVANCED PULSE-POWER TECHNOLOGY

capacitors that have been developed and

group

teated at LASL are the basic components of many

evaluate technology applicable to magnetic fusion

pulse-power systems nationwide, including those

experiments. During 1977, several concepts were

laser fusion, E-beam

investigated for future applications.

isotope separation, and weapona simulation.

LASL-CTR haa pioneered the 60-kV, Scyllac-

During calendar year 1977, the LASL

technology.

the

ayatema for a 60-kV capacitor bank are a major

high-density capacitors, 60-kV high-density ca-

portion of the overall system and are responsible

pacitors, size D ignitrona, and a 160-kV low-

for most of the system malfunctions. New concepts

for both the atart trigger system and the crowbar

The 10-kV test bay is used to life test size

trigger ayatem were investigated.

10-kV high-density capacitors

Triggering spark gap switches with a laaer

simultaneously. The 10-kV, 170-PF, 20%-reversal

pulse ia a well-developed technology pioneered by

capacitors were developed because the lowest cost

the Air Force Weapons Laboratory.2 However, the

per joule (0.03 cents/joule) in a capacitor occurs

number of optic components and the beam transport

at 10 kV. The test bay operated for 180 000 shots

ayatem required to switch many gapa simultaneously

in 1077 during which two of the ten test capaci-

become excessive and expensive.

tors failed. During the same test period, five of

development of fiber-optic cables for communica-

the size D ignitrons used to switch and crowbar

tion has presented another method of transporting

the capacitors failed. In addition, the test load

high-power laser pulsea to trigger many parallel

a

A

to

will

test

life

for

The

due

for

160-kV

pacitor

windings

designed

systems.

be impregnated

Vew units

capacitors,

In many instances,

The pulse-power

capacitor bank

fusion, laser

evaluation

requirea

program

parameter

10-kV

type

test

will

of

The

a

a

the

for

and

data

test

was

One

were

Moat

this

in 1978.

heating

Leboratoriea

capacitor

capacitor

implosion

capacitor

The capacitor

The continuing

trigger

seeks

LASL

above

The

at

to

TAHLE XVII-V

COMFONRNT TEST DATA

~,,:

HIGH
VOLTAGE

Fig. XVII-26.
Laser fiber-optic spark gap trigger system.

The 60-kV capacitor-bank crowbar gaps are

required to short out the capacitors at peak cur-

rent or switch at zero voltage.

system requires a very complex trigger system that

supplies about 300-500 joules to each gap.

trigger system components have a limited lifetime

and thus limit the crowbar system lifetime.

160 kV LI Capacitor

aTests still in progress.

Component/Parameter
10 kV HD Capacitor

Vo1tage
Reversal
Capacitance
Inductance

Voltage
Current
Reversal
Capacitance
Average life
Maximum life
Inductance

Voltage
Peak current
Coulombs
Average life
Maximum life
60-kV HD Capacitor

Voltage
Current
Reversal
Capacitance
Average life
Maximum life
Inductance
Size D Ignitron

is very feasible.

fiber-optic cable

the fiber cable.

early in 1978.

stimulated

limited

cab1e564,5

cable.

previous

fiber

short

0.33

by

is

%3

A=I.06

TuNGSTEN ELECTRODE

not life tested
160 kv
85%
0.3 PF
12 n~

0.1- 10mJ
ENERGY REQUIRED

~lf:~PLE
T
FIBER
OPTIC
CABLES

10 kV
100 kA
1000 c
30 0001
180 0001

60 kV
60 kA
20%
2.85 UF
32 0001
49 0001
30 nH

10 kV
5kA
5%
170 PF
120 0001
50 00001
40 nH

EEE==

conduct a peak

A laser fiber

the fiber is

fiber-optic

Brillouin

density

power

cable

and

in

or many series (Marx) spark gaps.

The technology of solid-state diodes has

optic for a spark gap trigger (LFOTS) system is

advanced in recent years into the range of cur-

diagramed in Fig. xVII-26. In order to trigger a

rents required of the 60-kV capa~itor crowbar

spark gap operating close to self-breakdown, only

switch. For example, a 7.5-cm silicon diode wafer

0.1 to 1 m.1 of energy is required at the spark

has a one-half cycle (60-Hz) surge current of 30

gap.2 The power denstty when focused on the spark

kA with a reverse holdoff voltage of 2000 V.

gap target electrode is the major parameter to be

60-kV crowbar switch operates with a duty cycle of

m~~imized, and thus, the power density in the

1 pulse per 5 minutes in a transformer oil medium.

the limiting factor

A new package arrangement for low-duty cycle

determining the amount of energy transferred using

operation and cooling in oil will be Investigated

the target electrode with a lens after leaving the

for air operation and repetitive duty, and thus is

The laser pulse is focused on

in 1978. The current industrial package is made

'l'he empirical

too bulky and expensive for 60-kV crowbar duty,.

threshold for low jitter and delay switching is

The 60-kV crowbar diode must withstand a 60-kV

GW/cm2.3 The power density h

Raman

reverse voltage for several microseconds and then

current of 60 kA for several

scattering to a maximum of about 0.8 GW/cm2 in

hundred microseconds. The reverse voltage parame-

studies and laser triggered spark-gap wurk thus

concept was initiated in 1977, and a preliminary

indicate that a laser fiber-optic trigger system

conceptual design of a 60-kV crowbar diode will be

This concept will be tested

completed in 1978 if manpower permits.

ter requires stacking many diode wafers.

+

&

LOAO

The

The present

This

The

The technologicalchallenge of sustaining or

power crowbarring fast-rising magnetic fields for

many hundred microseconds into the millisecond

range has been a major activity in 1977. A basic

circuit

and large I at t=O but have a Coulomb limit due to

electrode erosion. On the other hand, ignitrons

have large Coulomb capacity due to the renewable

mercury electrodes but have rather high induc-

tance, operate at only moderate voltages, and I is

Fig. XVII-27. The development of a high-Coulomb,

restricted at t=O. Thus

high-voltage, low-inductance crowbar switch shown

in Fig. XVII-27 has been of major concern.

low-inductance apark gap is being developed as the

power crowbar switch for ZT-40.

gaps are probably limited to about 100°C at 100

kA, or I-insconduction time. In order to sustain

desirable characteristics of a spark gap and an

ignitron into a mercury-wetted apark gap waa

initiated in 1977. Initially, a research contract

to a university was proposed to determine the

switching and voltage hold-off phenomena of a com-

However,

the many

delayed the program and finally required that the

mechanical switch must be closed in parallel with

investigations be done in house. These investiga-

the fast-switching power crowbar switch.

tions will start in 1978. A conceptual design of

desirable to use simple mechanical switches (i.e.,

hydraulic or pneumatic), which close in about 3-5

ms, because fast-closing mechanical switches (&lt;3

are complex, require large amounts of power,

and must be timed accurately.

spark

1977, the

(RFIHc), or

In order to use simple mechanical switches, a

using techniques applicable to a ceramic substrate

a

for a fusion-type reactor.

The RFIHC waa manufacture

approximately s me at 100 hi must be developed to

region between

experiment, but

ms)

into

with

power

SWitch

bridge

fields

crowbar

crowbar

the Coulomb

EN::GY
SOURCE

&gt;.t =0

START

I,B

!

-q-

TR

o

I

t = TR"

a

A

a

ia

of

in

for

time

used

It iS

During

posite

range,

switch.

implosion

diagramed

spark gap

conduction

millisecond

Scyllac-type

Fig. XVII-28.

heating coil

mercury-wetted

However, apark

received from CMB-6.

CROWBAR --
ENERGY

TRIGGER
ELECTRODE/

Spark

PLASMA

POWER

LOAD

GU

I
I
I

r~

-t

L

a

of

in

in

is

the

fed

gap

ahown

future

budget

because

heating

Marshall

radially

implosion

coil, was

limitations

The RFIHC was to be

a program to combine the

CATHODE
CONNECTION

COOLING
COILS

@NS"LATOR

E

START UP ENERGY

POWER CROWBAR ENERGY

Fig. XVII-27.
Basic power crowbar circuit.

switches and simple mechanical switches.

gaps operate at high voltages with low inductance

Fig. XVII-28.
Conceptual mercury-wetted switch.

ANODE

cONNEcT'O~

Fig. XVII-30.
Peaking gap systems.

correct breakdown characteristics

the design.

Scyllac program termination, It will not be used.

Measurements of the coil parameter

were conducted using small signals to verify in-

ductance calculations and to wrap up the RFIHC

The trigger systems used in fusion power

supplies routinely use series, peaking spark gaps

to isolate various circuit voltages and to reduce

trigger pulse voltsge risetimes. The peaking gaps

are irradiated with an initial arc to reduce

jitter. However, the initial arc has been created

with a small center pin trigatron configuration as

shown in Fig. XVII-30(a). The trigger pin assem-

bly distorts the electric field between the main

electrodes, reducing the main gap self-breakdown

voltage and also wearing out prematurely. A new

"inside-out" trigatron configuration shown in Fig.

XVII-30(b) was designed to overcome the problems

with the standard trigatron. The new peaking gap,

XVII-31, reduces a 200-kV pulse

rise time of 30 ns to 7 ns with less than 2-ns

load.

electrode is now more massive to reduce wear, and

the desired interelectrode capacitances requfred

The

a

into

SO-W

jitter

Fig. XVII-29.

shown in Fig.

development program.

wb~"

r

I

is

in

(a)

The

LOAD

-@'Np"T

shown

RFIHC

irradiating

and operation

to provide the

are inherent in

bouTPuT

IRRADIATION
ARC

OUTPUT/
ELECTRODE

EPOXY INSULATOR

EPOXY COATING

CONDUCTORS

??
u

4P!.m

a

&nrrWIDEGAPS

!?2-mm.WIDE CONDUCTORS

9GCOPPER

*

Fig. XVII-29.
Radially fed implosion heating coil.

V).\ +

AIR INPUT

+

(b)

~AO

#'N"T

OUTPUT

EOIATE
ODE

ODE

AIR OUTFUT

Fig. XVII-31.
Inside-out trigatron peaking gap.

Another area of intereat in all fusion ex-

periments ia monitoring spark gap operation and

A small development program was

2. A. H. Guenther and J. R. Bettis, "Everything
You Ever Wanted To Know About Laaer Triggered
Switching," Air Force Weapons Laboratory, Kirtland
Air Force Baae, NM.

initiated in 1977 to use the apark gap arc light

as a monitor signal and transport this signal via

to

The baaic development was a window in

the apark gap to observe the light.

1. R. A. Fitch, "Marx and Marx-Like High-Voltage
Generatora," Maxwell Labs, Inc., IEEE Trans. on
Nuclear Science NS18~,

(1971).

analysia.

REFERENCES

stitching times.

fiber-optic bundlea

A. H. Guenther,

Laser-Triggered

Rates,"

4. R. G. Smith, "Optical Power Handling Capacity
of Low Optical Fibers aa Determined by Stimulated
Raman and Brillouin Scattering," Appl. opt. ~
2489-2494 (1972).

5. S. M. Jansen and M. K. Barnoaki, "Stimulated
Raman
Proc.
Topical Meeting on Optical Fiber Transmission II,
Feb. 22-24, 1977 Williamsburg, VA, Paper TuD7.

Multimode Fibera,"

Scattering in

8, (1970).

and

for

the

room

screen

Repetition

J. R. Bettis

3.
second-Jitter
Moderate
Elec., ~,

IEEE

Switching
J.

"Subnano-
at
Quant.

long-term reliability

The Tritium System Test Assembly (TSTA),

project was initisted at LASL in February 1977.

demonstrate long-term safe handling of

This project is funded by the Office of Fusion

tritium with no major releaaes or incidents; and

as

Energy
(MDF)
project. A detailed proposal,1 including cost and

Fabrication

Device

7. investigate and evaluate the response of

the fuel cycle and environmental packages to

schedule projections was generated during the

normal, off-normal, and emergency situations.

In June 1977 the re-

'he objectives of the TSTA Project are to

sults of this study were presented at a MDF review

fabricate, construct, and

at DOE Headquarters. The project waa recommended

operation the Tritium Systems Test Assembly at

for approvsl by the DOE appointed review panel at

envelopes

that time. The Office of Fusion Energy acting on

schedule and cost baselines.

this recommendation approved the TSTA project in

The TSTA will consist of a large gas loop,

'he TSTA project was approved

Fig. XVIII-2, which can simulate the proposed fuel

with a baseline Total Estimated Cost (TEC) of

cycle for a fusion facility. The loop, as shown,

Of this, .\$4.8M is for LASL operations,

Include any specific fuel injection

\$6.9 M for hardware and equipment procurement and

system, but will be sufficiently versatile so that

installation and \$1.5 M is contingency funding.

such systems can be added as the design require-

The baseline schedule projection calls for TSTA to

be on-line and operating by FT-1981.

designed to handle up to 360 moles of DT per day.

a

A.

Major

\$13.2M.

INTRODUCTION

early July 1977.

February-May 1977 period.

handling system used

B. TSTA FUNCTION

power reactors;

Power Reactor

develop

(EPR).

and

in

1.

2.

6.

5.

LASL

within

design,

approved

does not

components;

demonstrate

James L. Anderson

XVIII. TRITIDM SYSTEM TEST ASSEMELY

ments are better defined.

for TNS and EPR systems.

impurity simulation)

approximately 150 g.

environmental and

assembly include:

Tranafer pumping

Vacuum pumping

Fuel storage

principal

Emergency

controls

mixing

later

these

Fuel

room

The

and

of

l

l

l

l

l

l

l

l

of

of

into

bring

The gas loop will be

technical

(including

effluent

injection

and

To accomplish this at

This flow will provide cycle operating experience

on a scale that is equal to or greater than the

The Tritium Systems Test Assembly (TSTA),

full-scale fuel cycles currently being addressed

monstration, and interfacing of technologies re-

TSTA will require an on-site tritium inventory of

Fig. XVIII-1, is dedicated to the development, de-

lated to the deuterium-tritium fuel cycle for

fusion reactor ayatems. The first such system to

'he assembly to do this will consist of a

be built till be a tokamak TNS (The Next Step)

number of integrated subsystems as follows:

deuterium-tritium burning machine. This TNS will

be followed by a TNS-upgrade or an Experimental

devices will build and expand on the tritium

Hydrogen isotope separation

. Fuel cleanup (ash and Impurity removal)

program objectives for TSTA can be stated as:

demonstrate the fuel cycle for fusion

Interfaces with external systems (neutral

beam, coolant breeding blanket systems)

personnel protective systems;

Essential functions associated with the overall

3. develop, test, and qualify equipment for

tritium service in the fusion program;

Tritium waste treatment

4. provide a final facility that can be used

cleanup

for demonstration and as an example that could be

directly copied by industry;

Either

TNS.

test

.----

- ' '----

:

_--

J-..

1--+.

.

.

.

.7

--

...

--.

_.\

... ..

~&gt;.

..--J _

---.

e-
*
~

CLEANUP

_
--------

._.--

mlRITIUIA MONITOa!N@

INWNTORY CONTROL

STORACC

4

l

l

'"

EMERGENCY TiUilUM
GI.EANUP

.

-.

-:.

~.

.._

-----

.Y.

I

_--p-*

-

~

/

.-

A

P-

!:!:

"""-

----

111111-

.----

-""N

Fig. XVIII-l

Artist'sconcept of TSTA.

. I
VACUUM TANK

ISOTOPE SEPARATION
SYSTEM

TRITIIM-WASTE'
TREATMENT-
I

Instrumentation

design of TNS.

ISOTOPE

and

w.4RAm0N

c.

Personnel safety

EXPERIMENTAL

CONTAMINATION

.

.

.

.

i

..-

.._

.--

-

*

---

STUOIES

System maintenance

Computer control and simulation

Analytical system (i.e., gaa analysis),

quality assurance procedures.

PROJECT INTERFACE WITH THE U.S. FUSION PROGRAM

The development of TSTA at this time will

permit design snd evaluation of the required fuel

cycle well in advance of the final design of a TNS

tritium system. This timely accumulation of data

will insure that the design and development of the

tritium system will not be a pacing item in the

lhe development of the tritium

technology required at TSTA for the fuel cycle en-

vironmental packages will not require any large

scientific or technical breakthroughs. There cur-

rently exists within DOE, and specifically within

the Los Alamos Scientific Laboratory,a consider-

able amount of experience and expertise in tritium

technology. This experience and expertise will be

used at TSTA to insure that the design of the

I

Fig. XVIII-2

Main process loop and auxiliary systems.

Waste diapoaal

Radiation monitoring

facility will reflect the most current information

and philosophy pertaining to tritium technology.

While the design of any specific subsyetem

may not require a significant technology advance-

ment, the integration of all of these subsystems

into one package is a significant advance. There

currently exists nowhere within the EOE complex a

facility which integrates all of these aubsyatems

into a continuously operating computer controlled

system. The flow rate and the total quantity of

tritium to be handled in a fusion reactor on a

routine basis necessitate the establishment of

TSTA as a demonstration facility. Table XVIII-I

inventories and release goala for five fusion

facilities. This comparison serves to emphasize

the

gives a

comparison of

obtained at TSTA.

TABLE XVIII-I

TRITIUM INVENTORIES IN FUSION

T2 Release

Fuel Feed

(kg/day)

0.001

0.5-1.5

0.25-0.5

1-2

2-6

1000-MWe net output

T2

4-6

EPR

TNS

0.2

(kg)

0.25

TSTA

6-12

TFTR

0.005

aDEMO

Energy

aDEMO =

tritium

Facility

Inventory

projected

REFERENCE

1

be

1-1o

5-1o

Goal

(Ci/day)

FACILITIES

a significant

the need for the timely development of TSTA. The

(ETM) program, will

Toksmak Fueion Test Reactor, TFTR, is currently in

milestone. The TSTA will serve as a base-line fa-

the early stages of construction at Princeton, NJ.

cility that will provide a large data base that

All of the other projected facilities in the table

can be ueed to establish future guidelines and re-

will follow TSTA and will utilize the information

quirements for fusion facilities.

A very significant goal at TSTA will be to

demonstrate that these large quantities of tritium

1. J. L. Anderson and R. H. Sherman, "Trftium

can be handled safely on a routine basis. The es-

Systems Test Assembly Design for Major Device

tablishment of environmental and safety packages

Fabrication Review," Los A.lamesScientific

for tritium systems within the Office of Fusion

Laboratory Report LA-6855-P (June 1977).

J. M. Bunch

D. C. Barnes

D. C. Barnea

W. T. Armstrong

E. D. Arthur
P. G. Young

D. C. Barnes
J. U. Brackbill

C. J. Buchenauer
A. R. Jacobson

R. D. Bengtson
S. A. Eckstrand
A. G. Sgro
C. W. Nielson

R. R. Bartsch
E. L. Cantrell
R. F. Gribble
K. A. Klare
K. J. Kutac
G. Miller
W. E. Quinn

R. J. Commiaso
C. A. Ekdahl
K. B. Freese
K. F. McKenna
W. E. Quinn

E. L. Cantrell
G. Miller
W. E. Quinn
W. Rfeaenfeld

F. W. Clinard
D. L. Rohr
W. A. Ranken

J. M. Bunch
J. G. Hoffman
A. H. Zeltman

R. Y. Dagazian
R. B. Paris

R. J. cOUIMiSSO
H. R. Green

R. Y. Dagazian

T. E. Cayton

T. E. Cayton

JOURNAL PUBLICATIONS

"Plasma Rotation During Implosion in a Theta-Pinch"

"Computation of Magnetohydrodynamic Flow in a
Magnetically Confined Plasma"

"Quadrature Interferometer for Plasma Density
Measurements"

"A Vacuum Field Solver for an Arbitrary Three-
Dimensional Mesh"

"Calculation of 15-MeV Neutron-Induced Charged-
Particle Spectrum on Stainless Steel Type 316"

"Stimulated Brillouin Backacatter Detection from a
Theta-Pinch Plasma"

"~ Approxtite Model for Toroidal Force Balance
in the High-Beta Stellarator"

"Feedback Stabilization of a 1=0,1,2 High-Beta
Stellarator"

"Mollwo-Ivey Relation Between Peak Color-Center
Absorption Energy and Average Ion Spacing in
Several Oxides of Group 11 and 111 Metals"

"Energy Anisotrophy Instabilities in the Vlasov-
Fluid Description of High-Beta Plasmas"

"Experimental Study of the Post-Implosion Phase
of a Theta Pinch"

"On the Nature of Features Seen by TEN in Fast
Neutron Irradiated AI.203"

"Stationary Convective-Like Modea in a Plasma
Slab with Magnetic

"The Laboratory Spring Mass Oscillator: An
Example of Parametric Instability"

I!InternalRing End Stopper for Open Ended Devicee"

"Observation of Propagating m=l Waves in Scyllac"

"Neutron Irradiation Damage In Stabilized Zr02°

"Solid-End-Plug Experiment on a Theta Pinch"

Shear"

884,

724,

J. Comput. Phya.

submitted to Phys. Fluids

submitted to Nucl. Fusion

submitted to J. Appl. Phys.

Phya. Rev. B @
(1977)

Phys. Rev. Lett. ~,
(1977)

Nucl. Sci. Eng. 64, 18
(1977)

Trans. Am. Nucl. Sot. 26,--
503 (1977)

Rev. Sci. Instrum. 48, 769
(1977)

J. Am. Ceramic Sot. 60, 287
(1977)

to be published in J. Nucl.
Mater.

Phys . Rev. Lett 39, 137
(1977)

Nucl. Fusion Letters ~,
1191 (1977)

Am. J. Phys. 4S, 723 (1977)

submitted to Nucl. Fusion

submitted to Phys. Fluids

Phys. Fluids ~,

Phya.. Fluidk ~,

44 (1977)

(1977)

R. Y. Dagazian

A. Dubi
Y. S. Horowitz

A. Dubi
Y. S. Horowitz

A. Dubi
Y. S. Horowitz

J. P. Freidberg
D. W. Hewett

J. P. Freidberg
R. A. Gerwin

J. P. Freidberg
R. A. Gerwin

A. Dubi
S.A.W. Gerstl
D. J. Dudziak

A. Dubi
Y. S. Horowitz
H. Rief

D. M. Drake
G. F. Auchampaugh
E. D. Arthur
C. E. Ragan
P. G. Young

R. L. Hagenson
R. A. Krakowski
K. I. Thomassen

S.A. W. Gerstl
D. J. Dudziak
D. W. Muir

S. P. Gary
A. G. Sgro
A. W. DeSilva

J. P. Freidberg
L. D. Pearlstein

S. P. Gary
J. J. Sanderson

S. P. Gary
W. C. Feldman

I. Henins
M. S. Kelly

R. Gri.bble
G. Miller

S.A.W. Gerstl

S. P. Gary

S. P. Gary

"Nonlinear Effects on Free Boundary Modes in Slab
Geometry"

submitted to Phys. Fluids

"Double Differential Beryllium Neutron Cross
Sections at Incident Neutron Energies of 5.9, 10.1,
and 14.2 MeV"

Phys. Rev. C ~,
(1977)

"A Note on the Interpretation of Conditional
Monte-Carlo as a Form of Importance Sampling

submitted to SIAM

"Lower Hybrid Drift Instability at Low Drift
Velocities"

Phys . Fluids ~,
(1977)

"Stability of a Closed-Line Scyllac, Revisited"

Nucl. Fusion=,

"Simultaneous Two Sample Activation Measurement
Immune to Detector Inter-calibration"

Int. J. Appl. Radiat. and
Isotopes ~,

"A Semi-Analytic Solution to the Transport
Equation in Spherical Geometry"

"Monte Carlo Aspects of Contributions"

to be published in Nucl.
Sci. Eng.

to be published in Nucl.
Sci. Eng.

"Calculation of the Total Flux at a Point by the
Monte Carlo Track Rotation Estimator"

submitted to Nucl. Sci.
Eng.

1792

443 (1977)

291 (1977)

(1977)

778

191

Trans. Amer. Nucl. Sot. ~,
47 (1977)

Rev. Sci. Instrum. ~,
(1977)

"Finite-Gyroradius Stabilization of Diffuse High-
Beta Stellarators"

submitted to Phys. Fluids

"Rotational Instabilities in a Theta Pinch"

submitted to Phys. Fluids

"Solar Wind Heat Flux Regulation by the Whistler
Instability"

J. Geophys. Res. 82, 1087
(1977)

"Electrostatic Heat Flux Instabilities in the
Solar Wind"

submitted to J. Geophys.
Res.

"Axial Density Profiles and Nonclassical Thermal
Conduction in Linear Theta Pinches"

submitted to Phys. Fluids

"Density Gradient Drift Instabilities: Oblique
Propagation at Zero Beta"

submitted to Phys. Fluids

"The Electromagnetic Ion Beam Instability and
Energy Loss of Fast Alpha Particles"

submitted to Nucl. Fusion

"Experience with Quantitative Data Assessment, and
Results for Fusion Reactors"

Trans. Am. Nucl. Sot. ~,

"Comments on the Concept of Spatial Channel Theory
Applied to Reactor Shielding Analyais"

Nucl. Sci. Eng. ~,
(1977)

"Plasma Position Detector Using a Segmented Photo-
diode"

Rev. Sci. Instrum. @
(1977)

"The Reversed-Field Pinch Reactor"

"Optically Isolated High Voltage Trigger System"

T. R. Jarboe

J. G. Hoffman

J, G. Hoffman

A. R. Jacobson

F. Hohl
S. P. Gary

H. R. Lewis
L. Turner

D. S. Lemons
S. P. Gary

D. S. Lemons
S. P. Gary

D. W. Hewett
C. W. Nielson

Y. S. Horowitz
B. Ben Shahar
S. Mordechai
A. Dubi
H. Pinto

M. G. Olsaon
E. T. Osypowski
L. Turner

D. Montgomery
L. Turner
G. Vahala

L. D. Pearlstein
J. P. Freidberg

R. L. Miller
R. A. Krakowski

K. F. McKenna
T. M. York

D. W. Muir
W. A. Reupke

R. C. Malone
R. L. Morse

H. R. Lewis
K. R. Symon

D. W. Muir

G. Miller

"A Multi-Dimensional Quasi-Neutral Hybrid Simu-
lation Code"

submitted to J. Conput.
Phys .

"Some Observationa of the Interaction of Alumina
with a Deuterium Plasma

to be published in J. Nucl.
Mater.

"Refractory Oxides for Fusion Reactor First Walls -
the Effects of the Reducing Environment"

to be published in Rev. Int.
des Hautes Temperatures et
des Refractaires

"Electron Kinematic in a Plasma Focus"

Phys. Fluids ~,

Thermoluminescence in LiF Inducedby Monoenergetic,
Parallel Beam, 13.8 MeV and 18.0 MeV Diffracted
Neutrons: The Intrinsic TL Response per Absorbed
Neutron"

Radiat. Res. 69, 402 (1977)

"A Heterodyne Quadrature Interferometer for Simul-
taneous Measurements of Plasma Density Along Sev-
eral Chords"

submitted to Rev. Sci.
Instrum.

"Measurement of Faraday Rotation in the Implosion
HeatLng Experiment"

J. ApP1. Phys. ~,
(1977)

"Electromagnetic Effects on the Modified Two-
Stream Instability"

J. Geophys. Res. 82, 2337
(1977)

"Current Driven Instabilities in a High Beta,
Quasi-Perpendicular Shock"

submitted to J. Geophys.
Res.

"Formation of the Linearized Vlasov-Fluid Model
for a Sharp-Boundary Screw Pinch"

submitted to Phys. Fluids

"Linearized Analysis of Inhomogeneous Plaama
Equilibria: General Theory"

submitted to Phys. Fluids

"Material End Plugging of Straight Theta Pinches"

Phys. Rev. Lett. 39, 134
(1977)

"End Loss from a Collision Dominated Theta Pinch
Plasma"

Phys. Fluids ~,
(1977)

"Long Wavelength m=l Magnetohydrodynamics of a
Theta Pinch"

Phys. Fluids 20, 928 (1977)

"Thermal Conduction and Alpha-Particle Constraints
for the Ignition of a D-T Linear Magnetic Fusion
(LMF) Reactor"

submitted to Phys. Fluids

"Three-Dimensional Msgnetohydrodynamic Turbulence
in Cylindrical Geometry"

submitted to Phys. Fluids

683 (1977)

296

"Data Covariance Estimation Methods for Sensi-
tivity-Based Data Aaaessment"

Trans. Am. Nucl. Sot. 26,

(1977)

"Data Adjustment in the Fusion Regime: Experience
and Future Directions"

Trans. Am. Nucl. Sot. 27,--
884 (1977)

"Resolution of an Ambiguity in Alternative Soft-
Pion Approached to TN + nnN Near Threshold"

Phya. Rev. Lett. ~,
(1977)

"Finite Larmor Radius EquatLons in an Arbitrary
Near Theta Pinch Geometry"

submitted to Phys. Fluid8

"Stability of BGK Equilibria"

submitted to Phys. Fluids

"0.54-MJ Superconducting Magnetic Energy Transfer
and Storage"

to be published in Adv. in
Cryogenic Eng. ~

J.

Blev.ins

L. Turner

L. Turner

A. G. Sgro

E. L. Simmons
D. J. Dudziak
S.A.W. Gerstl

J. L. Schwarzmeier
H. R. Lewis
B. A. Shrauner
K. R. Symon

J. D. Rogers
D.
J.D.G. Lindsay
G. A. Miranda
C. E. Swannack
D. M. Weldon
J. J. Wollan
C. J. Mole
E. Mullan
P. W. Eckels
H. E. Hailer
M. A. Janocko
S. A. Karpathy
D. C. Litz
P. Rerchner
Z. N. Sanjana
M. S. Walker

L. R. Veeser
E. D. Arthur
P. G. Young

D. Winske
P. C. Liewer

P. G. Young
E. D. Arthur

662 (1977)

654 (1977)

"Calculations of the 'Effectsof Incomplete Pre-
ionization in High Voltage Theta Pinches"

submitted to Phys. Fluids

"Nuclear Design Sensfvity Analysis for a Fusion
Reactor"

Nucl. Technol. ~,
(1977)

"Vlasov-Fluid Theory of Short-Wavelength Instabil-
ities of a Sharp-Boundary Screw-Pinch"

"Finite-Larmor-Radius Stabilization fn a Sharp-
Boundary Vlaaov-Fluid Screw-Pinch"

"Cross Sections for (n,2n) and (n,3n) Reactions
Above 14 MeV"

Phys. Fluids ~,

Phya. Fluids ~,

Phy&. Rev. C 16, 1792 (1977)

"Particle Simulation Studies of the Lower Hybrid
Drift Instability"

submitted to F'hys.Fluids

*v59c0.+n Calculation Up to Neutron Energies of
40 MeV"

Trans. Am. Nucl. SOC. ~,
503 (1977)

D. J. Dudziak

A. R. Jacobson

A. R. Jacobson

L. C. Burkhardt

J.D.G. Lindsay
D. M. Weldon

P. E. Armstrong
R. A. Krakowsi

J. D. Rogers
H. F. Vogel
D. M. Weldon
R. W. Warren

1).A. Baker
J. N. Di Marco
P. R. Forman
J. A. Phillips

T. J. Seed
W. F. Miller, Jr.
F. W. Brinkley, Jr.

A. R. Sherwood
B. L. Freeman
R. A. Gerwin
T. R. Jarboe
R. A. Krakowski
R. C. Malone
J. Marshall
R. L. Miller
B. Suydam

E. L. Simmons
S.A.W. Gerstl
D. J. Dudziak

R. L. Whitman
F. C. Jahoda
R. P. Kruger

H. Vogel
T. Thullen
D. Weldon

P.C.T. Van der Laan

R. E. Siemon

R. W. Warren

H. Vogel

"Proposed Tokamak Poloidal Field System Development
Program"

TRIDENT: A Two-Dimensional Multigroup, Triangular
Mesh Discrete Ordinates, Explicit Neutron Tranaport
Code"

"A Design for Coordinated Measurements of Faradsy
Rotation and Line-of-Sight Electron Density Using
Heterodyne Technlquea"

LABORATORY REPORTS

"Fast Liner Proposal"

"Toroidal Z-Pinch Demonstration"

"Thermal Shock Experiment (TSEX)"

"Imploding-Liner Reactor Nucleonic Studies"

"Preliminary Engineering Calculations for a
Toroidal Reverse-Field Pinch Reactor"

"A Novel Interferometer for the Measurement of
Plaama Density"

"Loss Measurements in Superconducting Magnetic
Energy Storage Coils"

"Experiments with Vacuum Interrupters Used for
Large DC-Current Interruption"

"Automated Computer Analysis of Plasma Streak
Tracea from Scyllac"

"Current Control by a Homopolar Machine with
Moving Brushes"

"Cross-Section Sensitivity Analyses for a
Toksmak Experimental Power Reactor"

"TNS Doublet Tokamsk Ohmic-Heating Power Supply
Study"

"A Summary of Scyllac Results"

LA-6707-P

LA-6811-P

LA-6735-M

LA-6875-MS

LA-6790-MS

LA-6861-MS

LA-6727-MS

LA-6948-MS

LA-6815-MS

LASL-77-3
Mini Review

LA-6909-MS

LA-6961-MS

LA-6767-MS

IA-7053-MS

LA-7090-MS

LA-6942-MS

LA-7125-MS

"Effects of Flux Conservation on the Field Config-
uration in Scyllac"

"Material End-Plugging for the Scylla IV-P Linear Theta Pinch"

"Design and Fabrication of a Rsdlally Fed Implosion Heating Coil"

"Final Analysis of the Engineering Data on the Scyllac Feedback
Stabilization Experiment"

"Circuit and Plasma Simulation for the Design of ZT-40"

"Spark Gap Overpreaaures in the 0.2 Megampere Range"

"Scylla IV-P Computer Based Control and Data Acquisition System"

SNENTH SYMPOSIUM ON ENGINEERING PROBLR-ISOF FUSION RESEARCH

L. D. Hansborough
T. R. Cole
R. J. collllllisSO
K. D. Williamson

L. D. Hansborough
J. M. Dickinson
J. G. Melton
W. C. Nunnally

E. M. Honig

G. P. Boicourt

R. W. Hanks
T. R. Cole

L. C. Burkhardt
R. S. Dike

R. Conrad
J. W. Lillberg
R. W. Wilkins

K. J. Kutac
R. W. Kewiah
G. Miller
R. F. Gribble

C. A. Ekdahl
R. J. COmmiSSO
K. B. Freese
R. Kristal
K. F. McKenna
R. E. Siemon
W. E. Quinn

J. G. Melton
R. S. Dike
K. W. Hanks
W. C. Nunnally

W. C. Numally
A. T. Brousseau

W. C. Nunnally
G. P. Boicourt

W. C. Nunnally

M. D. Mschalek

R. W. Warren

H. 1?.Vogel

OCTOBER 1977

KNOXVILLE, TENNESSEE

"A Toroidal Helical Quartz Forming Machine"

"Dual 30-kA, HVDC Interrupter Test Facility"

"Gas Breakera for Tokamak Ohmic-Heating Duty"

"Design of the ZT-40 Power Crowbar System"

TROY, NEW YORK

MAY 1977

"Development of Switching Components for ZT-40"

IEEE 1977 INTERNATIONAL CONFERENCE ON PLASMA SCIENCE

"Air Driven Fiber Optic Coupled Pulsed System for ZT-40"

"Power Crowbar Component Requirements for High-Beta Fusion
Reactors"

"Reduction of Plasma End Loss With Solid End Plugs in the Scylla
IV-P Linear Theta Pinch"

"Recent Work on Normal and Superconducting Inductive Energy
Storage Switching at the Efremov Institute, Leningrad, USSR"

"Vacuum Interrupters Used for the Interruption of High D.C.
Currents"

"1.5-MA, 1O-MJ Load Coil for a Pulsed Homopolar Nachine"

EIGHTH EUROPEAN CONFERENCE ON CONTROLLED FUSION AND PLASMA PHYSICS

"Feedback Stabilization Experiments Using !.=2Equilibrium
Windings in Scyllac"

"Plasma Parameters and End-Loss Measurements in the Scylla
IV-P Linear Theta Pinch"

R. K. Linford
D. A. Platts
E. G. Sherwood

K. F. McKenna
R. J. Commiaao
C. A. Ekdahl
K. B. Freese
R. Kristal
R. E. Siemon
W. E. Quinn

H. F. Vogel
P. l%ullen
K. D. Williamson, Jr.

R. W. Warren

H. H6thker
J. Downing

R. R. Bartsch
E. L. Cantrell
R. F. Gribble
K. B. Freese
L. E. Handy
R. Kristal
G. Miller
W. E. Quinn

D. A. D'Ippolito
J. P. Freidberg
J. P. Goedbloed
J. Rem

H. R. Lewis
L. Turner
M. Menzel

K. F. McKenna
R. J. Commisso
C. A. Ekdahl
K. B. Freeae
R. Kristal
W. E. Quinn
R. E. Siemon

K. R. Symon
J. L. Schwarzmeier
H. R. Lewis

M. E. Battat
D. J. Dudziak

M. E. Battat

SEPTEMBER 1977

PRAGUE, CZECHOSLOVAKIA

"Interruption of High DC Currents"

"Field-Reversal Theta Pinch Experiments"

"Shielding Standards - A Case History"

"Stability of BGK Equilibria"

KNOKVILLE, TENNESSEE

APRIL 1977

"Experimental Study of Magnetic Field Diffusion in a
High-8-Plasma of a 7-m 8-Pinch at Low Densities"

"Finite-Gyroradius Effects in Screw Pinches"

"Experiments on End-Loss Phenomena and Material End Plugs
in a 5-m Linear Theta Pinch"

"Shield Analyses for Intense 2.2-pJ (14-MeV) Neutron Sources"

"High-Beta Tokamake Surrounded by Force-Free Fields"

FIFTH INTERNATIONAL CONFERENCE ON REACTOR SHIELDING

"Applications of the Monte Carlo Track Rotations Estimator"

"Sensitivity Profiles for Secondary Energy and Angular
Distributions:

"Calculation of Toroidal Fusion Reactor Blankets by Monte
Carlo"

"Cross Sections In the Energy Range from 10 to 40 MeV
Calculated with the GNASH Code"

"Review of MFE Neutron Cross-Sections Needs - Neutronics
Viewpoint"

"Neutron Evaluation at High Energies - Problems and Prospects -
Review"

S.A.W. Gerstl

J. L. McDonald
E. D. Cashwell
C. J. Everett

E. D. Arthur
P. G. Young

D. J. Dudziak
D. W. Muir

L. Stewart
E. D. Arthur

norowitz

A. Dubi
Y. s.

D. C. Barnes
J. V. Brackbill
W. Schneider

R. L. Hagenson
R. A. Krakowski
K. I. Thomassen

R. A. Krakowski
R. W. Moses
R. L. Miller
R. A. Gerwin

R. A. Krakowski
A. S. Tai

R. L. Miller
R. A. Krakowski

J. D. Rogers
D. J. Blevins
J.D.G. Lindsay
G. A. Miranda
C. E. Swannack
D. M. Weldon
J. J. Wollan
C. J. Mole
E. Mullan
D. W. Deis
P. W. Eckels
H. E. Hailer
M. A. Janocko
S. A. Karpathy
D. C. Litz
P. Rerchner
Z. N. Sanjana
M. S. Walker

NAY 1977

UPTON, NEW YORK

OTHER CONFERENCES

SYMPOSIUM ON NEUTRON CROSS SECTIONS - 10-40 MeV

"Analytic and Numerical Studies of Diffuse, Three-Dimensional
Equilibria in Scyllac," Third International (Kiev) Conference
on Plasma Theory, Trieste, Italy, April 19?7.

"A Toroidal Fusion Reactor Based on the Reversed-Field Pinch
(RFP)," IAEA Conference and Workshop on Fusion Reactor Design,
Madison, Wisconsin, October, 1977.

"Fusion Power from Fast Imploding Linera," IAEA Conference and
Workshop on Fusion Reactor Design, Madison, Wisconsin, October
1977.

"A Simple Parametric Analysis of Fiasile Fuel Production by
Fusion-Fission Reactors," Second DMFE Fusion-Fission Energy
Systems Review Meeting, Washington, D.C., November 197?.

"Magnetic Fusion Alternate Concepts Heating Requirements,"
DOE/DMFE Plasma Heating Development Requirements Workshop,
Gaithersburg, Maryland, December 1977.

"Pulsed Superconducting Inductive Energy Storage Coils,"
Sixth International Conference on Magnet Technology,
Bratislava, Czechoslovakia, 1977.

"Magnetics and Superconducting Coil Design of a 10-mJ, Fast-
Discharging Honopolar Machine," Sixth International Conference
on Magnet Technology, Bratislava, Czechoslovakia, 1977.

"Vibrations of Electromagnetic Shields in Superconducting
Alternators," World Electrotechnical Congress, Moscow,
Russia, June 1977.

Printing Office:

G. J. Shevchuk
P. Thullen

H. F. Vogel
M. D. Driga

*U.S.

Government

.

1979-

677-013/13

001432s

0s147s
076-100
101-125

40!42s

451-475
476.S011

126-1SO
151-175

201-?2s
226-?s0

2S1-275
176.300
301-32s
326.350
35I .37s

I 0.7s
11.00
11.7s
12.00
, ~.~o

7.2s
M.a2
9.00
9.2S
9.50

4.00
4.50
5.25
6.00
6.50

in tl:c Unilcd State. of Ameritz

14.00
14.50
1s00

SSI-57S
576-600

S28S Purl ROYAI Road

sol .s25
5~6.s50

13.00
13.~s

US Department

[email protected]

!nfornution

of COtnnlmc

Anil.!hlc

Twhniul

.Scwicc

Prtntw!

tbtiowl

22!61

from

(AI-up

I S.25
15.50
16.25
16.50