THERMALANDNUCLEARASPECTSOFTHEPd/D OSYSTEM 2 Vol. 1: ADECADEOFRESEARCHATNAVYLABORATORIES S.SzpakandP.A.Mosier–Boss,eds. Contributingauthors(inalphabeticalorder) Dr. PamelaA.Mosier–Boss CodeD363 SpawarSystemsCenterSanDiego SanDiego,CA92152–5000 (619)553–1603;FAX(619)553–1269;e–[email protected] Dr. ScottR.Chubb Code7252 NavalResearchLaboratory Washington,DC20375–5343 (202)767–5270;FAX(202)767–3303;e–[email protected] ProfessorMartinFleischmann,F.R.S. BuryLodge,DuckStreet Tisbury,Salisbury,WiltsSP36LJ UnitedKingdom (+44)1747870384;FAX(+44)1747870845 Dr. M.AshrafImam Code6320 NavalResearchLaboratory Washington,DC20375–5343 (202)767–2185;FAX(202)767–2623e–[email protected],mil Dr. MelvinH.Miles DepartmentofChemistry MiddleTennesseeStateUniversity Murfreeboro,TN37132 (615)904–8558;e–[email protected] Dr. StanislawSzpak 3498ConradAve SanDiego,CA92117 (858)272–9401 i ii FOREWORD Twelve years have passed since the announcement on 23 March 1989 by Professors FleischmannandPonsthatthegenerationofexcessenthalpyoccursinelectrochemical cellswhenpalladiumelectrodes,immersedinD O+LiOHelectrolyte,arenegatively 2 polarized.Theannouncement,whichcametobeknownas“ColdFusion,”causedfren- ziedexcitement. Inboththescientificandnewscommunities,faxmachineswereused to pass along fragments of rumor and “facts.” (Yes, this was before wide spread use oftheinternet. Onecanonlyimaginewhatwouldhappennow.) Companiesandindi- vidualsrushedtofilepatentsonyetto beprovenideasinhopesofwinningthegrand prize.Unfortunately,thephenomenondescribedbyFleischmannandPonswasfarfrom beingunderstoodandevenfactorsnecessaryforrepeatabilityoftheexperimentswere unknown.Overthenextfewmonths,thescientificcommunitybecamedividedintothe “believers”andthe “skeptics.”The“believers”reportedtheresultsoftheir workwith enthusiasmthatattimesoverstatedthesignificanceoftheirresults. Ontheotherhand, many “skeptics” rejected the anomalous behavior of the polarized Pd/D system as a matterofconviction,i.e., withoutanalyzingthepresentedmaterialandalwaysasking “where are the neutrons?” Funding for research quickly dried up as anything related to“ColdFusion”wasportrayedasahoaxandnotworthyoffunding. Theterm“Cold Fusion”tookonanewdefinitionmuchastheFordEdselhaddoneyearsearlier. By the Second International Conference on Cold Fusion, held at Villa Olmo, Como, Italy, in June/July 1991, the attitude toward Cold Fusion was beginning to take on a morescientificbasis. Thenumberofflash-in-the-pan“believers”haddiminished,and the “skeptics” were beginningto be faced with having to explain the anomalousphe- nomenon,whichbythistimehadbeenobservedbymanycrediblescientiststhroughout the world. Shortly after this conference, the Office of Naval Research (ONR) pro- posedacollaborativeeffortinvolvingtheNavalCommand,ControlandOceanSurveil- lanceCenter,RDT&EDivision,whichsubsequentlyhasbecometheSpaceandNaval Warfare Systems Center, San Diego (SSC San Diego); the NavalAir Warfare Center, WeaponsDivision,ChinaLake;andtheNavalResearchLaboratory(NRL).Theeffort’s basic premise was to investigate the anomalous effects associated with the prolonged iii charging of the Pd/D system and “to contribute in collegial fashion to a coordinated tri-laboratoryexperiment.” Each laboratory took a different area of research. At San Diego, our goal was to un- derstandtheconditionsthatinitiatetheexcessheatgeneration(theFleischmann–Pons effect)andthesearchforevidencethatindicatestheirnuclearorigin. Toeliminatethe longincubationtimes(oftenweeks),Drs. StanSzpakandPamBossdecidedtoprepare the palladium electrodes by the co-deposition technique. Initially, they concentrated ontritiumproductionandthemonitoringofemanatingradiation. Morerecently,they extendedtheireffortto monitoringsurfacetemperaturevia IRimaging techniqueand showedtheexistenceofdiscreteheatsourcesrandomlydistributedintimeandspace. This discoverymay proveto be a significant contribution to the understanding of the phenomenon. AtChinaLake,Dr.Milesandhiscollaboratorsshowedthatacorrelationexistsbetween therateoftheexcessenthalpygenerationandthequantityofheliuminthegasstream. SuchacorrelationisthedirectevidenceofthenuclearoriginoftheFleischmann–Pons effect. The research at NRL was directed toward the metallurgy of palladium and its alloys and the theoretical aspects of the Fleischmann–Pons effect. In particular, Dr. Imam preparedPd/BalloysthatDr. Milesusedincalorimetricexperiments.Itwasshownthat thesealloysyielded reproducibleexcessenthalpygenerationwithminimalincubation times(approximately1day). ThetheoreticalworkofDr. Chubbcontributedmuchto ourunderstandingoftheFleischmann–Ponseffect. AlthoughfundingforColdFusionendedseveralyearsago,progressinunderstanding thephenomenoncontinuesatamuchslowerpace,mostlythroughtheunpaideffortsof dedicatedinquisitivescientists. Inpreparationofthisreporttheauthorsspentcountless hours outside of their normal duties to jointly review their past and current contribu- tions, including the “hidden” agenda that Professor Fleischmann pursued for several yearsinthe1980swhenhewaspartiallyfundedbyONR.Specialthanksareextended to all scientists who have worked under these conditions, including those who con- tributedtothisreportandespeciallytoProfessorFleischmann. As I write this Foreword, California is experiencing rolling blackouts due to power shortages.Conventionalengineering,plannedahead,couldhavepreventedtheseblack- outs, but ithasbeenpoliticallyexpedientto ignorethe inevitable. We donotknowif ColdFusionwill betheanswerto futureenergyneeds,but we doknowtheexistence ofColdFusionphenomenonthroughrepeatedobservationsbyscientiststhroughoutthe world.Itistimethatthisphenomenonbeinvestigatedsothatwecanreapwhateverben- efitsaccruefromadditionalscientificunderstanding. Itistimeforgovernmentfunding organizationstoinvestinthisresearch. Dr. FrankE.Gordon Head,NavigationandAppliedSciencesDepartment iv SpaceandNavalWarfareSystemsCenter,SanDiego v vi TABLEOFCONTENTS 1. THEEMERGENCEOFCOLDFUSION ....................................1 S.SzpakandP.A.Mosier–Boss 2. EVENTSINAPOLARIZEDPd+DELECTRODESPREPAREDBY CO-DEPOSITIONTECHNIQUE ...........................................7 S.SzpakandP.A.Mosier–Boss 3. EXCESSHEATANDHELIUMPRODUCTIONIN PALLADIUMANDPALLADIUMALLOYS ...............................19 MelvinH.Miles 4. ANALYSISOFEXPERIMENTMC-21: ACASESTUDY PartI:DevelopmentofDiagnosticCriteria ..................................31 PartII:ApplicationofDiagnosticCriteria ...................................51 S.Szpak,P.A.Mosier–Boss,M.H.Miles,M.A.ImamandM.Fleischmann 5. ANOVERVIEWOFCOLDFUSIONTHEORY .............................91 ScottChubb APPENDIX:LISTINGOFPUBLICATIONS/PRESENTATIONS RELATEDTOCOLDFUSIONBYNAVYLABORATORIES STAFF .................................................................113 vii viii CHAPTER1: THEEMERGENCEOFCOLDFUSION S.SzpakandP.A.Mosier–Boss 1.0Introduction. In this chapter, we address briefly the events proceeding and followingthe 23 March 1989announcementthatnuclearreactionscouldbeinducedatroomtemperaturesand atmosphericpressurewhenelectrochemicallygenerateddeuteriumiscompressedinto thePdlattice. Inparticular,wediscusstheeventsthatledFleischmanntothisconclu- sion, his philosophy of research and the characteristic of the Pd/nH (n = 1,2) system that prompted him to initiate research into host lattice assisted nuclear reactions. An extensive discussion of these topics can be found in the recently published paper by Fleischmannentitled: ReflectionsontheSociologyofScienceandSocialResponsibil- ityinScience,inRelationshiptoColdFusion[1]. The announcement by Fleischmann and Pons that nuclear events can and do occur in thePd/D system whendeuteriumiselectrochemically compressedin thePd lattice wasatotallynewandcontroversialconcept,incompatiblewiththestandardteachings of nuclear physics. Aquestion that naturally arisesis what prompted Fleischmann to undertakethiskindofresearch.WasittheshortnotepublishedinNaturebyOliphantet al. in1934[2]whodemonstratedthatnuclearreactioncanoccurincondensedmatter, or was it something else? In what follows, we seek the answer in Fleischmann and hiscollaboratorsnumerouspublications/presentationsthatappearedafterthe23March 1989pressconference. InalecturegivenattheFirstInternationalConferenceonColdFusion(ICCF–1),Fleis- chmann[3]saidthefollowing:“Ourinterestinnucleationphenomenaandourknowl- edgeofthepredictionoftheformationofmetallichydrogen(anddeuterium)atextreme compressionsin UnitedStatesand Sovietworkduringthe mid 70swas, in fact, a key element in the initiation of this research project.” Toward the end of his lecture, he 1 remarked:“We, for our part, would not have started this investigation if we have ac- ceptedtheviewthatnuclearreactionsinhostlatticescouldnotbeaffectedbycoherent processes.” These quotes suggest that his interest in the Pd/nH system extended over a period ofyears prior to the 23March announcement andthat his research wascon- cernedwithfundamentalaspectsofsolidstatechemistryandphysics. 2.0Chronologyofevents. A brief chronology of events is as follows. Early in 1947, Fleischmann realized that thePd/Hsystemis“themostextraordinaryexampleofanelectrolyte”,i.e.,exhibiting behavior that could not be satisfactorily explained in terms of the Debye-Huckelthe- ory.Inthe1960s,hewasconvincedthatthecorrectapproachtothebehaviorofionsin solutionmustbeintermsofquantumelectrodynamics(QED).Inthelate60s,hecon- cludedthat“themeasurementsandinterpretationoffluctuationsinsmallsystemswas onepossiblerouteforprobingtheapplicabilityofQED,especiallytheapplicabilityto thebehaviorofcondensedmatter”[1,p. 27]. Facingoppositioninscientificcirclesto thisapproach,Fleischmanndecidedtofollowan“hiddenagenda”.Theunderlyinggoal ofsuchresearchwastoillustratetheneedtoapplyQEDreasoningwhenexaminingthe behaviorofcondensedmatteraswellasdemonstratingthatsucheffectscanbeprobed usingelectrochemicalprocedures(methods),sincethesemethodshavetherequiredac- curacyand sensitivity to probe such effects (e.g., the ability to measure small signals forsmallsystems,anincreaseinsensitivitybyusingmodulationmethods,etc.). While attheUniversityofSouthampton(1967–1980),heandhiscollaboratorsstudiedtheef- fects ofthevariousparametersonthebehaviorofthePd/nHsystemthatcould notbe predictedusingclassicalandquantummechanics. Concerning the emergence of cold fusion, we have to ask (i) how did cold fusion fit withFleischmann’sresearchplansand(ii)whydidFleischmannandPonsselecttoin- vestigatetheelectrochemicalcompressionofdeuteriumintoahostlattice?Theanswer tothefirstistodemonstratethattheQEDparadigmisthecorrectone. Theanswerto the second is a conclusionthat, to probe the Pd/H system, energybalance rather than momentumwillbeconsistentwiththe“hiddenagenda.”Experimentswereconducted toprobetheeffectsof(i)space,(ii)time,(iii)length,(iv)dimensionality,(v)number, and (vi) structure. The missing factor was (vii) energyand experiments on this were started attheUniversityofUtah. As such, coldfusion was, andis simply,a partofa widerprogramaimedatshowingthatelectrochemicalmeasurementscouldbeusedto probetheapplicabilityoftheQEDparadigm. In1983,collaborativeprojectswithProfessorS.Pons(UniversityofUtah)wereiniti- atedandaimedatansweringtwoquestions[1,p. 31]: (i)“wouldtheputativereactionsofD (cid:0) compressedintohostlatticesbedifferentfrom thereactionsinadiluteplasma(orreactionsofhighlyexcitedDinsolids)?”(i.e.,could nuclearreactionsbegeneratedwithinahostlattice? (ii)“couldsuchchangesinthereactionsbeobserved?” Toanswer thesequeries, two methodsofchargingthe metallatticeswereconsidered: 2 (i) Compression of D (cid:0) using applied electric fields (electro-diffusion) and (ii) com- pression using electrochemical charging. Of these, the latter providesthe easiest and efficient way to raise the potential energyof an extendedquantum system [1, p. 31]. Initially,calorimetricstudieswereselectedtoassessthemagnitudeofexcessheatgen- erationbynuclearevents. Furthermore,theisoperiboliccalorimetrywasthepreferred methodtoexplorethebehaviorofthechargedPd/Dsystembecauseitisthelowcostand “catchall”method.By1988,measuredratesofexcessenthalpygenerationwereshown tobeconsistentwiththoseobtainedfornuclearreactions.In1986,anuncontrolledheat releaseduetosystembeingdrivenintothe“positivefeedback”wasobserved.Withthe passageoftime,othertechniqueswereusedtoinvestigatethebehaviorofthePd/Dsys- temandtheorieshavebeenformulatedtounderstandthedynamicsofsuchsystems. In spiteoftheenormouspotentialforpracticalapplications,thedisseminationofrelevant informationislimitedtoaveryfewjournals.Tocomprehendthescaleofactivitiesfol- lowingthe23March1989pressconference,oneshouldreviewthematerialpublished intheProceedingsoftheInternationalConferenceonColdFusion,ICCF1–8. 3.0ThePd/nHsystem. What was it about the Pd/D system that prompted Martin Fleischmann to begin this research? ItappearsthatthestartingpointwastheworkofCoehn[6],doneinthelate 1920sandearly1930s,ontheelectro-diffusionofhydrogeninPdwires. Coehnfound thattheabsorbedhydrogen(deuterium)ispresentasachargedspecies,i.e.,itexistsin itsnuclear–notatomicstateandthattheNernst–Einsteinrelation,u D (cid:1) =FD D (cid:1) /RT,is obeyed. But, the existenceof D (cid:0) while in the Pd lattice in the presence of high con- centration of s – electrons should lead to the formation of D as dictated by the law 2 of mass action. Furthermore, the application of the Born–Haber cycle to the dissolu- tion of protons into the lattice is ca 12 eV. Such a large magnitude of the “solvation energy”impliesthattheprotonsitsindeepenergywellswhilehighmobilityputsitin shallowholes.Thus,toquote:“Howcanitbethattheprotons(deuterons)aresotightly boundyettheyarevirtuallyunboundintheirmovementthroughthelattice?”[5].Thus, Coehn’sobservation,whencoupledwiththequasi-thermodynamicanalysisoftheelec- trochemical potential, as defined by Lange [6] (µ D (cid:1) (cid:2) µ D (cid:1) (cid:3) ef ), posed a number of questions, amongthem: WhatisthenatureofthespeciesathighD/Pdatomicratios? WhatarethedynamicsofD (cid:0) undertheseconditions? Thesequestions,combinedwith experimental evidence (e.g., heat after death, electro-diffusion), led Fleischmann to considerthepossibilitythatnucleareventscanoccurinthehostlattice. ThecharacteristicsofthePd/nHsystemthatsetsitasidefromothermetalhydridesys- tems include (i) high concentrations of ionized hydrogen (deuterium), (ii) its (their) highmobility,(iii)highH/Dseparationfactoratequilibrium,(iv)largediffusioncoef- ficientswithinverseisotopiceffect,and(v)highelectrochemicalpotentialofdissolved hydrogen(deuterium). Eachofthesecharacteristicsisassociatedwithacertainaction (activity). Inparticular: (i)Ahighconcentrationofionizedspecieswithinthelatticeindicatesthatelectrostatic fieldswithintheunitcellforcethetransitionfromtheatomictonuclearstate.Thehigh solvationenergyimpliesthatdeepelectrostaticpotentialholesarepresent. 3 (ii)Thehighhydrogenanddeuteriummobility,acceleratedbyelectricfieldsindicates that nH, in their nuclear states, are immersed in a dense plasma of d–electrons; if so, thenwhydoeshighlycompressedatomichydrogennotform? (iii) The high H/D separation factor is consistent with a model based on delocalized classical oscillators having a high affinity for Pd. High affinities and high separation factorsimplyhighlydelocalizedwavefunctionsandshallowpotentialholes. (iv)Largediffusioncoefficients(D=10 (cid:4) 7cm2s (cid:4) 1)whereD (cid:0) D (cid:5) D (cid:0) H (cid:5) D (cid:0) indicatethe T presenceofshallowholeswhiletheinverseisotopeeffectimpliesthatdeuteriumhasa configurationspacedifferentfromthatofhydrogenandtritium. (v)Highchemical/electrochemicalpotentials,viatheirgalvanicpotentialf ,tendtopro- motetheformationoflargeprotonclusters. 4.0Theannouncementandestablishmentresponse Itisknown[1]thatFleischmannopposedthedisclosureoftheresultsofthisresearchin March1989;attheearliest,hepreferredautumnof1990.Thereasonsforhisopposition were(i)aprematuredisclosurewouldforcehimtoworkinarathernarrowsetoftopics whilehisinterestswereinexploringtheimplicationsofquantumfieldeffectsinnatural sciences,and(ii)theexpectedattitudeofindustry,wheretheoptionofcleanproduction oflowgradeheatwouldbecontrarytotheirshortandmedium-terminterests. Indeed,theresearchresultsofFleischmannandhiscollaboratorswerequestionedbe- causetheydidnotfit intotheaccepted viewsofthe D (cid:0) +D (cid:0) fusion path. Insteadof proceedingalongtheusualrouteofscientificinquiry,thecriticsdisregardedtheexperi- mentalresultsofmanyscientistsconsistentwiththemanifestationsofnuclearactivities in the Pd/D system. Fleischmann’sview(in 1989)that the establishment would seek tostoptheresearch,byridicule,disinformation,cuttingoffunding,andpreventionof publications was confirmed. Moreover, manyresearchers decided that it would be in their interest to report negative conclusions. This can be done by selecting bad data, byusinginadequateorflawedexperimentdesign, orbynotprovidingthe rawdatato preventfurtherevaluationoftheresults. Toillustrate,inthisreport,frequentreferences aremadetonon–authorizedchangesinproceduresorinterpretationemployedevenby collaboratinglaboratories(cf.Chapter4). 5.0Fleischmann’sphilosophyofresearch TheanswertotheoriststhatthePdhostlatticeassistednuclearprocessesarenotpossi- bleisobvious:experimentalevidencecarriesmoreweightthantheoreticalspeculations. In1991,Fleischmann[7],inhisaddresstotheRoyalInstituteofChemistry,stated: “It isthequalitativedemonstrationswhichareunambiguous;thequantitativeanalysesof the experimental results can be subject to debate but, if these quantitative analyses stand in opposition to the qualitative demonstration, then these methods of analysis must be judged to be incorrect”. It is quite remarkable that a similar view was ex- pressedseveraldecadesearlier(1943)bythenotedtheoreticalphysicist,MaxBorn[8], inhisaddresstotheDurhamPhilosophicalSociety,viz.,“Myadvicetothosewhowish 4 tolearntheartofscientificprophecyisnottorelyonabstractreason,buttodecipher thesecretlanguageofNaturefromNature’sdocuments,thefactsofexperience”. 6.0Summaryofevents. Toreiterate,asearlyas1960,FleischmannconcludedthatthebehaviorofH (cid:0) andD (cid:0) electrochemicallycompressedinto Pd-hostlatticescouldonly be understoodin terms ofquantumfieldtheory. This conclusionled Fleischmann, in1983, to twoquestions: (i)wouldthe nuclearreactionsofD (cid:0) compressedintohostlatticesbedifferentto the reactions in a dilute plasma? and (ii) would such effects be observed? The expected answers: YestothefirstandNotothesecond. Intheinterveningyears(1986, 1987), Fleischmanncollectedenoughevidence,e.g.,heatafterdeath,compressionbyelectro- diffusion,tochangetheanswerto(ii)fromNotoYes. Finally,inMarch1989,events forced Fleischmann and Pons to present their evidence of nuclear activities in Pd/D system. 7.0References 1. MartinFleischmann,AccountabilityinResearch,8,19(2000) 2. M.L.Oliphant,P.HarteckandLordRutherford,Nature,133,413(1934) 3. M.Fleischmann,AnoverviewofcoldfusionphenomenaICCF1 4. A.Coehn,Z.Elektrochem.,35,676(1929) 5. C.Bartomoleo,M.Fleischmann,G.Larramona,S.Pons,J.Roulette,H.Sugiuraand G.Preparata,Trans. FusionTechnol.,2623(1994) 6. E.Lange,Z.Elektrochem.,55,76(1951) 7. M.Fleischmann,Thepresentstateofresearchincoldfusion,ICCF2,p. 475 8. M.Born,Experimentandtheoryinphysics,Doverpubl.,NewYork,1953. 26,23(1994) 5 6 CHAPTER2: EVENTS INA POLARIZEDPd+D ELECTRODES PREPARED BYTHECO-DEPOSITIONTECHNIQUE. S.SzpakandP.A.Mosier–Boss 1.0Introduction. ThischapterreviewsourresearchactivitiesofthepolarizedPd/D Osystem.Incontrast 2 tothepioneeringworkofFleischmannandhiscollaborators,weconsideronlyevents at, and/or,within Pd electrodesprepared bythe co-depositiontechniquedevelopedin thislaboratory. Oureffortproceededalongtwopaths: (i)investigationofthermaland nucleareventsinthePdhostlattice[1–8]and(ii)examinationoftheroleoftheinter- phaseregion[9–13]. Thesepathswereundertakentoassesstheintensityofeventsand to providesome informationon the factors controlling the initiationandmaintenance ofexcessenthalpygeneration,i.e.,the“performanceenvelope.” Thescopeislimitedtoabriefdescriptionoftheexperimentalworkfollowedbyconclu- sions.Afulldescriptionoftheexperimentaltechniquesaswellasathoroughdiscussion isprovidedincitedreferences. 2.0Co-depositiontechnique. It is well known that the structure of electrodeposited metal is controlled by a num- beroffactors,amongthem(i)currentdensity(cellcurrent),(ii)concentrationofmetal ions (or its complexes), (iii) additives, and (iv) the structure of the substrate. One of the methods to examine the details of a deposit is the use of scanning tunneling mi- croscope. Recently, Naohara et al. [14] reported that during the electroreduction of PdCl2 (cid:4) complex, “the Pd deposition proceeds in a layer-by-layer growth mode.” If 4 theelectroreductionofthepalladiumcomplextakesplacein thepresenceofevolving hydrogen/deuterium,theabsorbedH/Daccumulatesintheregionsseparatedbythelat- tice defects where the b –Pd/D is formed, transforming the smooth Pd surface into a 7 modular–likestructure[15]. Fig. 1SEMphotographofco–depositedPd. The Pd+D co-deposition is a process where palladium and deuterium are simultane- ously deposited on a non-hydrogen absorbing metallic substrate, e.g., Cu or Au, at sufficiently high negative potentials from electrolytes containing palladium salts dis- solved in heavy water [1]. The surface morphology and bulk structure are controlled by the solution composition and cell current. As a rule, at cell currents close to the Pd2 (cid:0) +2e (cid:4)(cid:7) (cid:6) Pdlimitingcurrentdensity,“cauliflower–like”Pdfilmsareproduced. An SEM photograph, Fig. 1, shows the typical structure of an electrode prepared by co-deposition. Theindividualsphericalglobulesareofsubmicronsize. Characteristic features of the co-deposited films are (i) an almost instantaneous saturation of the Pd lattice [2] with D/Pd atomic ratios (cid:5) 1.0, (ii) high surface to volume ratio, and (iii) reproduciblebulkstructure. 3.0Thermalevents. The objectiveofthis research was, and still is, directed towardsdetermining the con- ditions maximizing excess enthalpy production. At the present time, a sustained low gradeheatsourcecanbemaintainedforconsiderableperiodsoftime[3].Weconsidered twotypesofmeasurements,viz. excessenthalpyandsurfacetemperaturedistribution. 3.1Excessenthalpy. The excess enthalpy/power production was assessed in two types of calorimeter de- 8 signs: (i)forshortdurationexperiments,acalorimeterwithanadiabaticenclosurewas employed,(ii)forlongtermexperiments,aFleischmann-Ponstypecellwasused[3].It isnoteworthythatcalorimeterswithadiabaticenclosuresofferadditionalinformation, viz. information on the effectof electrolyte temperatureon the processeffectiveness. Examplesofexcessenthalpyplottedagainstenthalpyinputforshorttimeexperiments areshowninFig. 2aandthatforlongtimeexperimentsinFig. 2b. Several points can be made: long charging times are eliminated and the rate of ex- cess enthalpy production is both cell current and temperature dependent with occa- sional bursts, points A, B,.., Fig. 2a and, most importantly, electrodes prepared by co-deposition yield reproducibly higher excess power than the commonly used solid electrodes,Fig. 2b. OneofthefeaturesofthePd/Delectrodespreparedbytheco-depositionprocessisthe generation of excess enthalpy at relatively low current densities (cell currents). This featuresuggeststhatanewclassofPd/Delectrodesshouldbeconsidered,amongthem, the fluidized bed electrode [16]. The behavior of copper fluidized bed electrodes has beeninvestigatedingreatdetail. Theseelectrodescanbeemployedinavarietyofcon- figurations,dependingonthelocationofthecurrentfeederelectrodesandthedirection ofcurrentandfluidflow. Itisnoteworthythatsuchelectrodeshaveverygoodheatand masstransfercharacteristics. 3.2Temperaturedistribution. The electrode surfacetemperaturedistribution can be monitored by infraredimaging. Usingthistechnique,thepresenceofdiscretereactionsitesrandomlydistributedintime andspace,Fig. 3aandsteeptemperaturegradients,Fig. 3b,areobserved. Thesefea- turesarecharacteristicoftheco-depositionprocess. Thesteep temperaturegradients, seenintheimages,indicatethattheheatsourcesarelocatedintheimmediatevicinity oftheelectrode/electrolytecontactsurface[3,4].Theaveragesurfacetemperaturesare ca 6oC above that of the solution. It is noted that the infrared imaging requires very closeplacementofthenegativeelectrodetothecellwalltominimizeattenuation. Thedisplayof“hotspots”andtheirinterpretationusingsimplifyingassumptionsmay defineanumberofnewexperimentswhich,inturn,couldthrownewlightonthe“cold fusion” mechanism(s). Employing the most drastic assumptions, it is concluded that the nuclear activities occur within the 1µm layer adjacent to the electrode/electrolyte contact surface. It is noted that this conclusion is in an agreement with the findings reportedbyBockrisetal. [17]. 4.0Nuclearevents. For the excess enthalpy generation to be of nuclear origin, there must be a resulting nuclear “ash” present. Cells for the simultaneous measurements of excess enthalpy 9 a) b) Fig. 2. Excessenthalpygenerationin(a)shorttimeand(b)longtimeexperiments. 10 Fig. 3. Surfacetemperaturebyinfraredimaging;(a)perpendicular,(b)parallelview. 11 and nuclear ash are difficult to construct and operate. Of possible nuclear events, we focusedoureffortsonmeasuringX-rayemanationandtritiumproduction. 4.1Emanatingradiation. Early in the investigation [1], we constructed a cell in which photographic film was placed in close proximity to a working electrode made of Ni screen onto which the Pd+D was co-deposited. After 24 hours exposure to the cathodically polarized elec- trode,thephotographicfilmwasdevelopedproducingtheimageshowninFig. 4a. To obtain spectral data, Fig. 4b, of the X-ray emissions required the use of background radiationshielding,theappropriateselectionofdetector(s)andcelldesign. Becauseof theverylowintensityoftheelectromagneticradiation,boththephotographicfilm[1] andthedetector[6]mustbeplacedascloseaspossibletotheradiationsource. Tosummarize,weofferthefollowingconclusions: (i)Reliablemonitoringofemanatingradiationrequiresadequateshielding,propercell design,andtheplacementofthesuitabledetector. (ii)CathodicallypolarizedPd/DsystememitsX-rayswithabroadenergydistribution withanoccasionalemergenceofrecognizablepeaks(e.g.,at21keV). (iii)TheemissionofX-raysappearstobesporadicandoflimitedduration. (iv)Thesurfacemorphologyinfluencesradiationemission,eg,co-depositedelectrodes exhibitshorterinitiationtimethansmoothsurfaces.Also,theadditionofBe2 (cid:0) ionsand ureaactivatetheX-rayemission. 4.2Tritiumproduction. Tritium production is determined by (i) comparing the computed and measured con- centrations of tritium, (ii) observation of the non-equilibrium distribution of tritium betweenthesolution andgasphasesand(iii)massbalance. Selectedexamplesoftri- tiumproductionanditsdistributionareshowninFigs.5aand5b.Itisclearthattritium releaseoccursviatwopaths,onefavoringtheelectrolytephase,Fig. 5a,theotherthe gas phase, Fig. 5b. [8]. Thepresence oftritium in the bulkmetal was observedonly uponadditionofsmallamountsofAl3 (cid:0) ionstotheelectrolytepriortoelectrolysis[7]. Thesporadicaswellaslowproductionrates,103to104 atoms/secondaveragedovera 24-hourperiod[7],demandaverycarefullydesignedsystemandsamplingprocedure, Figs. 6aand6b. Figure6ashowsthedesignofthecalibratedcell(a)andrecombiner (b) containing a suitable catalyst. Due to the sporadic occurrence of nuclear events, lowratesoftritiumproductionanderrorsintritiumanalysis,shortsamplingtimesare necessarybecauseaveragingoverlongtimeperiodsmayobscureitsdetection. Toreiterate,wenote: (i)Closed cells(i.e., cellswithrecombiningcatalyst)areconsideredsuperiortoclose systemarrangements(cf.Fig. 6a)forthedetectionoftritiumproductioninelectrolytic cells. But, a closed cell, by design, represents an integrating system, i.e., a system incapable of detecting time dependent tritium production rates. In contrast, a closed system arrangement, such as used in our laboratory, providesinformation on the rate 12 a) b) Fig. 4. Emanatingradiation,a)recordedonfilm,b)spectraldata. 13 a) b) Fig. 5. Tritiumproductionandreleasepaths. a) b) Fig. 6. Apparatusfortritiummeasurements,a)celldesign,b)samplingschedule. 14 andfrequencyofthe“burst-like”tritiumproduction. (ii)Theevidencefortritiumproductionaswellasitsproductionrateiscalculatedfrom the difference between the computed and observed concentration of tritium, the non- equilibriumdistributionandthetotalmassbalance. (iii) The production of tritium takes place within the interphase region. The surface morphologyaffectsthedistributionoftritiumbetweenthegasandelectrolytephases. 5.0Theinterphaseregion. Even a cursory examination of the thermal and nuclear activities indicate the impor- tanceofthe regionseparatingthehomogeneouselectrolyteandbulkmetalphases. In an attempt to determine the factors affecting the “performance envelope,” we under- tookanexplorationoftheinterphasestructureandprocessestherein. Inparticular,we discuss the structure of the interphase, the driving forces on loading/unloading, and developmentofthermalinstabilities. 5.1Structureoftheinterphase. Thelayerseparatingtheelectrolyteandbulkmetalhomogeneousphasescontainspar- ticles that interact with particles in neighboring phases. If the number of interacting particlesislargecomparedtothetotalnumberofparticles,thenthislayerisdefinedas non-autonomous. Evidently,thePd/D Ointerphaselayerhasanon-autonomouschar- 2 acter. ThecomplexstructureofthePd/D Ointerphaseandtheoperatingforcesacting 2 duringloadingand/orunloadingcanbebestvisualizedbyconsideringthesequenceof eventstakingplace[4]. Theseeventsareasfollows: (cid:8) b (cid:9)(cid:11) (cid:10) (cid:8) l s (cid:9)(cid:12) (cid:10) (cid:8) l m (cid:9)(cid:11) (cid:10) (cid:8) m (cid:9) (1) where(b)isthehomogeneoussolutionphase, (cid:8) l s (cid:9) and (cid:8) l m (cid:9) arethesolutionandmetal sides of the non-autonomous interphase and (m) denotes the bulk metal, Fig. 7. The solutionsidecomprisesoftwolayers:thereactionlayer(r)andtheabsorptionlayer(a) whilethemetalsideconsistsoftheabsorption(ab)andionization(io)layers.Thus,for aPdelectrodeincontactwithanelectrolytecontainingdissolvedD inD Oacidified 2 2 with DCl or D SO , the distribution of components is as follows: Pd, e 2 4 (cid:4) and D (cid:0) in themetallicphase,Pd, D,D (cid:0) ande (cid:4) intheinterphase,andD , D 2 (cid:0) , Cl (cid:4) /SO2 (cid:4) inthe 4 electrolytephase. Evidently,notallphasescontainthesamecomponents. 5.2Drivingforces. The dynamics of the interphase during loading/unloading is driven by forces arising fromchemicalpotentialgradients. Theuseofthechemicalpotentialdifferenceasthe drivingforceforthetransportofspeciesbetweentwophasesissubjecttotheapplication of the Gibbs–Duhem equation which states that local equilibrium must be assumed. Buttransportacrosstheinterphaseaswellasotherprocessesputthesysteminanon- equilibrium state. A non-equilibrium system in local equilibrium can be modeled by segmentingthesysteminindividuallayers,eachinmechanicalandthermalequilibrium 15 C (r ) (ad) (ab) (r ) 1 2 (cid:13) l 1 (cid:14) (cid:13) l 2 (cid:13) (cid:14) l (cid:14) (cid:14) I (cid:14) j 1 (cid:14) j 2 (cid:13) j 3 (cid:14) j 4 (cid:14) j el (cid:14) j ch (cid:14) j d Fig. 7. Structureoftheinterphase. with a stopped transport/reaction, i.e., wherethe Gibbs–Duhem equation isvalid. By reassemblingthesystemandassuminglocalequilibria,thenon-homogeneousnatureof thesystemisrestored[4]. Chemical/electrochemicalpotentialsinasystemcontainingchargedparticlesinthermal andmechanicalequilibriumisgivenbyµ i (cid:2) ¶D G ¶ ni (cid:15) p (cid:16) T (cid:16) ni (cid:18) (cid:17) . Whenthissystemisplaced nj in an external electric field, y , the potential energy of charged particles becomes a functionofposition,thesystembecomesnon-electroneutralanditschemicalpotential becomesµ i (cid:2) ¶D G ¶ ni (cid:15) p (cid:16) T (cid:16) ni (cid:18) (cid:17) nj (cid:16) y . Thus,anychangein p (cid:19) T (cid:19) n andy hasadirecteffecton j thedynamicsoftheelectrode/electrolyteinterphase. 5.3Developmentofthermalinstabilities. Evensmallchangesinsystemvariablesareexpectedtohaveaneffectonthedynamics of the interphase. To demonstrate, a single grain when viewed under a microscope equipped with Nomarski optics shows preferred sites for deuterium to enter, Fig. 8. The associated volume changes within the l layer produces motion in the l layer m s which can be displayed by interference fringes. Obviously, at high current densities, the formation, growth and detachment of evolving deuterium bubbles would have a profoundeffectontheoverallprocessesinboththesolutionandthemetalsideofthe interphase. The nature of the driving forces and the experimental evidence suggest that excess enthalpygenerationcanbeexpressedasafunctionofexternallyappliedfield,y ,(over- potential,h ),surfacecoverage,q ,concentrationofabsorbeddeuterium,c andconcen- D trationofreactivestates,c ,i.e.,D H s (cid:2) F (cid:8) p (cid:19) T (cid:19) y (cid:19) q (cid:19) c D (cid:19) c s (cid:9) . Thetime/spacedependent location ofshortdurationofdiscretereactionsites furthersuggeststhat derivativesof thesevariablesareinvolved. Suchfunctionaldependenceandthehighlynonlinearbe- 16 Fig. 8. EffectofDabsorption. Left,changesinelectrode(byNomarskioptics),right, inelectrolyte(recordedbyinterferometer). haviorleadstothedevelopmentofthermalinstabilitieswhich,inextreme,canleadto electrodemelting[5]. 6.0Concludingremarks. The Pd electrodes prepared by the co-deposition technique show (i) excellent repro- ducibility, (ii) an increase in the excess enthalpy production with the increase in cell currentandelectrolytetemperature,and(iii)theheatsourcesarelocatedincloseprox- imitytotheelectrode/electrolytecontactsurface. The searchfor the evidenceof the Pd latticeassisted nuclear eventsrequires well de- signedcellsandastrictadherencetoexperimentalprotocols.Thus,X-raydetectionne- cessitatesshieldingandplacementofthedetectorincloseproximitytothePdelectrode while tritium production must be based on a complete mass balance. Short sampling timesarerequiredtodetectlowratesofproduction. Note: To our knowledge, electrodes prepared by co–deposition technique were em- ployed by Hodko and Bockris [18] and Miles [10]. In both cases, remarkable repro- ducibilitywasdemonstrated. 7.0References 1. S.Szpak,P.A.Mosier–BossandJ.J.Smith,J.Electroanal. Chem.,302,255(1991) 2. S.Szpak,P.A.Mosier–BossandJ.J.Smith,J.Electroanal. Chem.,379,121(1994) 3. S.Szpak,P.A.Mosier–BossandM.H.Miles,FusionTechnology,36,234(1999) 4. P.A.Mosier–BossandS.Szpak,IllNuovoCiminto,112A,577(1999) 5. S.SzpakandP.A.Mosier-Boss,PhysicsLettersA,221,141(1996) 6. S.Szpak,P.A.Mosier–BossandJ.J.Smith,PhysicsLettersA,210,382(1996) 7. S.Szpak,P.A.Mosier–Boss,R.D.BossandJ.J.Smith,FusionTechnology,33,38 (1998) 8. S.SzpakandP.A.Mosier–Boss,FusionTechnology,34,273(1998) 17 9. S.Szpak,C.J.GabrielandJ.J.Smith,J.Electroanal.Chem.,337,273(1991) 10. S.Szpak,P.A.Mosier–BossandS.R.Scharber,ibid.,337,147(1992) 11. S. Szpak, P. A. Mosier–Boss, C. J. Gabriel and S. R. Scharber, ibid., 365, 275 (1994) 12. S.Szpak,P.A.Mosier–BossandJ.J.Smith,ibid.,379,121(1994) 13. S.Szpak,P.A.Mosier–Boss,S.R.ScharberandJ.J.Smith,ibid.,380,1(1995) 14. H.Naohara,S.YeandK.Uosaki,J.Phys. Chem.,102B,4366(1998) 15. T. Ohmori, K. Sakamaki, K. Hashimoto and A. Fujishima, Chem. Letters (The ChemicalSocietyofJapan),p. 93(1991) 16. M.Fleischmann,privatecommunicationtoF.Gordon,15Nov.00 17. J. O’M. Bockris, R. Sundarasen and Z. Minevski, Extended Abstracts, Electro- chemicalSociety185thMeeting,SanFrancisco,CA,May1994 18. D.HodkoandJ.O’M.Bockris,J.Electroanal.Chem.,353,33(1993) 19. M.H.Milesprivatecommunication,1999 18 CHAPTER3: EXCESSHEATANDHELIUMPRODUCTIONINPALLADIUM ANDPALLADIUMALLOYS. MelvinH.Miles 1.0Introduction. The research effort at the Naval Air Warfare Center, Weapons Division, China Lake, proceededalongthreemainpaths:(i)thedevelopmentofaccuratecalorimetricmethods fordetectingexcessheatgeneration,(ii)samplingoftheelectrolysisgasesfordetermin- ing helium production, and (iii) monitoring the electrolysis cells for radiation effects. Thereviewofourresearchispresentedintwoparts. Thefirstpartcoversresearchac- tivities at China Lake during 1989 to 1995 that became part of an official U.S. Navy programtitledAnomalousEffectsinDeuteratedSystems,fundedbytheONRin1992. The second part reviews experiments conducted by the author at the New Hydrogen EnergyLaboratory(NHE),Sapporo,Japan,duringOctober1997toMarch1998. The researchatNHEfocusedonproducingthefollowing:(i)excessheatinChinaLaketype cellsusingpalladiumandpalladiumparticles,(ii)excessheatusingpalladiumalloysin the Fleischmann–PonsDewar typecells, and (iii) excessheat using the co-deposition methodintheFleischmann–Ponscells. 2.0ExcessHeatProduction. ThemainsignatureforfusioninthePd/D OsystemreportedbyFleischmannandPons 2 is excess heat production. Their announcement in 1989 excited the world because it offeredthepossibilityofunlimited,almostfree,non-pollutingenergy.Ifcoldfusioncan berenderedreliableandscaled-up,thenitwilllikelybeoneoftheimportantscientific discoveriesofthe20thcentury.Unfortunately,theDepartmentofEnergy(DOE)panel, intheirreportonColdFusion,publishedinNovember1989,statedthattheChinaLake studiesalongwiththoseofCaliforniaInstituteofTechnology(Caltech),Massachusetts InstituteofTechnology(MIT),andHarwellshowednoexcessheatproduction. 19 Indeed, the initial calorimetric measurements at China Lake showed no measurable excess heat generation, and this was reported at the Santa Fe, New Mexico meeting, 22–25May1989andlaterpublished[1].However,earlyintheresearch,werecognized thattwofactorsplayadecisiveroleintheinitiationandmonitoringoftheexcessheat production,viz.,themetallurgyofthePdanditsalloysandacorrectcalorimeterdesign. 2.1ChinaLakeIsoperibolicCalorimeter. Various designs were investigatedconsisting mostly of open, isoperibolic systems. It was foundthat the decreasing levelof the electrolyte, asD O was electrolyzedto D 2 2 andO gases,wasidentifiedasamajorerrorinthecalorimetry[1]. Tominimizethis 2 problem,theelectrochemicalcellwasplacedintoasecondarycompartmentfilledwith H O,andthetemperaturewasmeasuredwithinthesecondarycompartmentinsidethe 2 calorimeter, Fig. 1. With this design, the electrochemical cell served basically as an electrical heater for the secondary compartment. These experiments showed that the ratioofHeatOut/HeatInwas1.00 (cid:20) 0.04[1]. Fig. 1BasicfeaturesoftheChinaLakeisoperiboliccalorimeter. 2.2Johnson–MattheyPalladium. All the early China Lake studies that showed no excess heat production used cath- odespreparedfrompalladiumwire(Wesgo)ofunknownorigin. Itwaslaterfoundthat the Wesgo palladium shows very poor loading characteristics. After a few months, a largediameter(d=0.63cm)palladiumrodwasreceivedfromJohnson–Matthey. Two segments from this rod were studied in two similar calorimetric cells in experiments thatstarted inSeptember1989. Afterabout10daysofelectrolysis, bothexperiments 20 showedexcessheatproductionthatwaswelloutsidethecalorimetricerror[3]. Welater turnedtheseexperimentsoffforafewweeksandthenre-startedtheelectrolysis. Once again,excessheatwasmeasured[3]. These results were presented at the First International Conference on Cold Fusion (ICCF–1)inSaltLakeCity,Utah,28–31March1990[4].Itisnoteworthythatbothour resultsandtheFleischmannandPonsresultsindicatethatdaysofelectrolysisarenec- essarybeforeanyexcessheatappearsandthatratherlargecurrentdensities(exceeding 100mA/cm2)arerequired. ThesametwoinitialsamplesoftheJohnson–Mattheypalladiumrodwerelaterused,cf. 4.1,intheexperimentsthatyieldedhelium-4intheelectrolysisgas[5]. TheJohnson– Mattheypalladiumprovedtobereliableforproducingexcessheatinourexperiments. Theseexperimentsdemonstratedtheimportanceofthemetallurgicalaspectsofthepal- ladium. 2.3NavalResearchLaboratoryMaterialsProgram. In 1992, our research activities merged with those of NRL in a program funded by ONR.Itwasrealized,bythistime,thatthepropertiesofthepalladiumwereacritical experimental parameter. Consequently, a major program was undertaken to produce palladium materials that yielded excessheat andto identify the critical parameters of suchmaterial. In1994,Imam[6]producedthreecompositionsofanewpalladium–boronalloymate- rialwithnominalconcentrationsof0.75,0.50,and0.25weightpercentboron.Analyses showedthatthethreealloycompositionsactuallycontained0.62,0.38,and0.18weight percentboron. Twodistinctphasesofthesamecubicstructurewerefoundinallthree compositionsofthealloy. UnlikepreviousNRLmaterials,thesenewPd-Balloysproducedexcessheatinalmost everyexperiment[7]. TheonlyPd-Bsamplethatfailedtoproducetheexcessheatwas onethathadalarge,folded-overmetalregionduetotheswagingofthisrodthatacted as a long crack [7]. Although we had achieved a major Navy goal for this program, i.e.,productionofpalladiummaterialsthatreproduciblyyieldedexcessheat,theONR sponsoredprogramwasterminatedafewmonthsafterthereportofexcessheatforthe Pd-B alloys, and no further research was conducted at China Lake after June 1995. Thissamepalladium-boronmaterialproducedsignificantexcessheatinanexperiment inJapanusingtheFleischmann–PonsDewartypecalorimetry(cf. 5.3). The question naturally arises regarding why the Pd-B alloys proved so successful in producing excess heat. Possible explanations include the fact that the added boron significantly increases the hardness of the palladium and the presence of boron also greatlyretardstherateatwhichdeuteriumescapesfromthepalladiummetal[7].There arealsoproposalsoffusionreactionsthatinvolveboron. Themostlikelyexplanationforthebeneficialeffectoftheaddedboronisthatitmini- 21 mizestheactivityofoxygeninthepalladiumbyconvertingittoB O .ThisB O floats 2 3 2 3 tothesurfaceandisremovedduringthemoltenphaseofthePd–Bpreparation.Thisex- plainsthelowerboronconcentrationsinthefinalmaterial. TheJohnson–Mattheypro- cessthatproducedgoodmaterialsforexcessheatgenerationreportedlyusedacracked ammonia atmosphere, i.e., N and H . Here again, oxygen would be removed from 2 2 the palladiumbyits reaction with hydrogento produce water. Perhapsthisisthe key for reproducible excess heat effects: palladium that is relatively free of oxygen. The co-deposition method developedby Szpak and Mosier–Bosswould also produce pal- ladium that isfreeof oxygencontamination. One canspeculatethat the deuteriumin the lattice reacts with the oxygen impurity to form D O and that this breaks up the 2 palladium–deuteriumlatticestructure. 3.0RadiationMeasurementsatChinaLake. Werealizedourlackofexpertiseinradiationmeasurementsandneverplannedonthis becominganyofficialpartofourresearchprogramatChinaLake. However,radiation monitoringwasrequiredbythesafetypersonnel;hence,wepurchasedsomeequipment includingGeiger–Mueller(GM),sodiumiodide(NaI)andneutrondetectors,aswellas ScalarRatemetersformonitoringanypossibleharmfulradiation. 3.1DentalFilmExposure. The excess heat measurements for the two Johnson–Matthey palladium cathodes at ChinaLakeledtofurtherexperimentsusingthesesametwoelectrodes. Theseexperi- mentsweredesignedtotestfor excessheat, X-raysbydentalfilmexposure,neutrons by gold activation, radiation by GM detectors, and helium-4 by the sampling of the electrolysis gases [5]. Evidence was found for everything except for neutrons [5, 8]. ExposureofthedentalfilmX–rayswasobserved[5].Thefilmpositionedtheclosestto thepalladiumcathode(CellA)showedthegreatestexposure[5]. 3.2MeasurementsUsingGMandNaIDetectors. AnomalouslyhighradiationcountswereobservedusingseveraldifferentGMdetectors as well as NaI detectors during electrolysis experiments with palladium cathodes in heavy water [7]. These high radiation counts were often observed in co-deposition experiments where palladium metal is deposited from a D O solution onto a copper 2 cathodeinthepresenceofevolvingdeuteriumgas.Theradiationcountsreachedvalues ashighas73s abovenormalbackgroundcounts. Theradiationwould appearwithin afewhoursintheco-depositionexperiments. Incontrast,theappearanceofradiation requireddaysofelectrolysisforthepalladiumrods[9]. Theemissionoflowintensity X-raysfromsimilarPd/DsystemswasreportedbySzpaketal. [10]. 3.3NeutronMeasurements. Neutron emissions from these experiments are very low and difficult to detect. Our measurementswerestrictlyforsafetyconcerns. WeusedaLudlumModel15neutron survey meter that was placed close to the water bath containing the electrochemical 22 cells. We used only an audio detection signal and neverrecorded any neutroncounts versus time for either the experiments or the background. Our one experiment using neutronactivationofindiumandgoldfoilsmountedatthesurfaceoftheelectrochem- icalcellsshowsthatanyneutronproductionwouldhavetobelessthan105persecond [5]. 4.0HeliumMeasurements. Two major theories had predicted that helium-4 would be the main fusion product in thePd/Dsystempriortoourexperimentalmeasurementsandhadalsopredictedthatthe helium-4would be present in the electrolysisgases. The first, byPreparata [11], was based on Quantum Electrodynamics (QED) while the second, by Chubb and Chubb [12],wasbasedonIonBandStates. Itisnotedthatthefirstsolidevidenceforhelium production was reported by us. Following our initial measurements of helium-4 pro- duction in the Pd/D system, a number of other laboratories have verified this result. Verystrongevidenceforhelium-4productionisfoundintherecentworkofArataetal. [13]andMcKubreetal. [14]. 4.1SamplesCollectedinGlassFlasks. Ourinitialreportofhelium-4productionduringexcessheateventsinD Oelectrolysis 2 experimentswaspublishedinMarch 1991[15]. Intheseexperiments,theflowofthe electrolysis gases was directed through a 500 mL glass flask and then through an oil bubblertotheoutsideatmosphere. Apositivepressurewasmaintainedwithinthesys- temtominimizeanyatmosphericcontamination. Theseexperimentsbegan3October 1990 and ended 25 December 1990. The system was thoroughly flushed with boil- offN gaswheneveraglassflaskwasreplacedorwhenD Owasadded. Thecollected 2 2 electrolysisgassamplesweresenttotheUniversityofTexasforheliumanalysis.Based ontheseexperiments,helium-4isthemajorproductwhenexcessheatoccurs[15]. A major criticism of these results was the possibility of atmospheric helium-4 con- tamination, especially due to the known diffusion of helium through glass. It was preciselybecauseoftheseconcernsthatweconductedcontrolexperimentsperformed using H O+LiOH in place of D O + LiOD. These control studies gave no evidence 2 2 of helium-4 production [5, 7, 8, 15]. Our first D O + LiOD electrolysis gas sample 2 (10/17/90–A) also served as a control since there was no significant excess heat and nohelium-4detected[5,7,11,16]. Ourcontrols,therefore,coveredtimeperiodsboth beforeandaftertheexcessheatexperiments; thisrefutesargumentsbycriticsthatwe weresimplygettingbetteratkeepingouthelium-4[16]. 4.2SamplesCollectedinMetalFlasks. Thesehelium-4experimentswererepeatedusingmetalflasksforcollectingelectrolysis gassamplestoruleoutthepossibilityofheliumdiffusionthroughglass.Allexperimen- tal conditions were intentionally kept the same except for the use of the metal flasks. Thehelium-4measurementsforthemetalflasksampleswereperformedattheU.S.Bu- reauofMines,Amarillo,Texas,laboratorythatspecializedinthesemeasurements.The 23 endresultwasthesameasbefore.Theelectrolysisgassamplescollectedinmetalflasks duringexcessheatproductionalsocontainedexcesshelium-4[7,17]. Furthermore,the rate of helium production could nowbe established at 1011 to 1012 atoms per second per watt of excess power[12, 17]. This is the correct magnitude for typical deuteron fusionreactionsthatyieldhelium-4asaproduct. 4.3SummaryofHeliumMeasurements. A total of thirty-three experiments were conducted that involved the measurement of helium-4 in the electrolysis gas. In experiments producing excess heat, 18 out of 21 also produced helium-4. Two experiments using a Pd–Ce cathode produced excess heat, but no helium-4 was detected[7]. The explanationis that the helium-4remains trappedinthisalloy. Thethirdexperimentinvolvedaflawedexcessheatmeasurement due to an unusually low D O level in the cell [7]. For all 12 experiments where no 2 excess heat was produced, there was no evidence for helium-4 production [7]. The probabilityoffindingthecorrectrelationshipbetweenexcessheatandhelium-4in30 outof33experimentsisaboutoneinamillion[7]. Theprobabilityofalsoobserving thecorrectmagnitudeofhelium–4production(1011to1012atomspersecondperwatt ofexcesspower)ineachexperimentduetorandomerrorsisaveryunlikelysituation. 5.0ResearchatNHE. No further research in the Pd/D system was done at China Lake after the ONR fund- ingendedinJune1995. However,aNewEnergyDevelopmentOrganization(NEDO) appointmentbecameavailabletoworkattheNewHydrogenEnergy(NHE)laboratory in Sapporo, Japan, from late October 1997 until the end of March 1998. Numerous calorimetricdatawerecollected. Muchofthisdatastillawaitsextensiveanalysis. 5.1ChinaLakeCalorimetryatNHE. ThefinaltwocoldfusionexperimentsatChinaLakein1995involvedtestsof1.0mm diameterJohnson–Mattheypalladiumwire. Oneexperimentproduced200mWofex- cesspowerwhiletheotherdidnot[7]. Thesesametwocells,electrodes,andcalorime- terswereusedagainatNHEinJapan.Theonlymajorchangewastheuseofaluminum foilratherthanwaterinthesecondarycompartmentsurroundingthecells.Thischange made the calorimetric system much more sensitive to the detection of excess power ( (cid:20) 5mWversus (cid:20) 20mW).Onceagain,thepalladiumwirethatproducedexcessheat inChinaLakeproducedsignificantexcessheatatNHEinJapan. Theotherpalladium wirealsoperformedasbeforeandproducednomeasurableexcessheateffects. These resultshavebeenrecentlypublished[18]. 5.2CellsUsingPlatinumandPalladiumparticles. TheseexperimentsinChinaLakecellsweredesignedtogivedynamicelectrolysiscon- ditionsbyusingsmallpalladiumandplatinumparticles. Theseparticleswereactually miniature cylindrical rods with the dimensions of 0.6 to 0.65 mm diameter and 0.65 24 to0.70mmlength. TheChinaLakecalorimetrywasusedtotestplatinumparticlesin CellAasacontrolwhilepalladiumparticleswereinvestigatedinCellB[19]. Figure 2showstheelectrochemicalpoweralongwiththeoutputpowerforthecellcontaining palladiumparticles. 3.00 Cell B − DC and Pulse Power February 11 − March 9, 1998 2.50 Pd 2.00 1.50 1.00 0.50 0.00 00 5000 10000 15000 20000 25000 30000 35000 40000 Time (minutes) )W( rewoP PB(el) P−out(B) PB Pulse Eq. Figure2. Theelectrochemicalpowerandoutputpowerforthepalladiumparticlesin CellB. Afterabout a weekof electrolysis, theoutput powerbeganto exceedthe input power tothecell. Thisexcesspowerwasnearly100mWfordirectcurrentelectrolysis. The electrochemicalinput powerwasswitched to pulse powerat 20490 minutes. This in- volvedapeakvoltageofnearly100V, apeakcurrentof6A,apulsewidthof1.0µs, a pulse frequencyof 5 kHz, and an average electrolysis current of 0.012 A [19]. As shown in Fig. 2, largeramounts of excess powerexceeding 200 mW were observed. Thecellcontainingplatinumparticlesgavenoexcesspowerforeitherdirectcurrentor pulseelectrolysis[19]. Thesmallmetalparticlesjostleaboutduringelectrolysis;hence,newsurfaceareasare continually exposed to the metal/electrolyte interface where electrolysis occurs. This experimentwasdesignedto givea fluidized bedelectrolysiseffect, butthe metalpar- ticlesweretooheavy. Themanytinypalladiumparticlesmaketheseexperimentsless sensitivetothevariablesthatproduceexcessheatinonepalladiumrod,butnotinan- othersimilarrod. 5.3PalladiumAlloyCathodesinFleischmann–PonsTypeCells. The three Dewar-type electrochemical cells used at NHE were silvered in their top portionssothatheattransferisconfinedalmostexclusivelytoradiationacrosstheun- 25 silveredpart. Thepalladiumcathodesselectedforthefirstcalorimetricstudiesinthese cellswerePd-Ce-B,Pd-B(0.5weight%boron),andPd-Ce.Theseexperimentsrequire accuratedeterminationoftheradiativeheattransfercoefficientandthewaterequivalent ofthecell.Theapproximatemethodsusedfortheanalysisarediscussedelsewhere[20]. TheseapproximatemethodsshownomeasurableexcesspowerforthePd-Ce-Bcelland significantexcesspowerforthePd–BandPd–Cecells[19,20]. Figure3presentsthe excesspowerforthePd–BexperimentusingtheFleischmann–Ponscalorimetry. Fig. 3. ExcesspowermeasurementsforthePd-0.5BcathodeinCellA-2. The data set from this Pd–B experimenthas beenexamined in detail by Fleischmann and his results are presented elsewhere in this report. This independent evaluation of the rawdata by Fleischmann showsthe same general trends asFig. 3, but the excess powerissignificantlyhigher.Comparisonsofthesetwomethodsshowthattheradiative heattransfercoefficientusedforFig. 3is4.64%toosmall(8.112x10 (cid:4) 10W/K4 versus 8.5065x10 (cid:4) 10W/K4). TheNHEmethodusedforthisexperimentaswellasforother experiments[21]isfoundtobecompletelyinvalid[21]. Aninterestingfeatureofthis Pd–Bstudyistheearlyonsetoftheexcessheateffect. ExcesspowermeasurementsforthePd–CecathodeintheFleischmann–Ponstypecell ispresentedinFigure4. ThisPd–CematerialwasobtainedfromFleischmannandgavesignificantexcessheat in a previousstudyat ChinaLake [7]. Theradiativeheattransfer coefficientusedfor theresultsshowninFig. 4was8.000x10 (cid:4) 10W/K4. Anindependentevaluationofthe rawdataforthisexperimentbyFleischmannisinprogress. 26 Figure4. ExcesspowermeasurementsforthePd-CecathodeinCellA-3. 5.4Co-DepositionExperimentsinFleischmann–PonsTypeCells. Themethodofdepositingpalladiumfrom solutiononto a coppercathodein thepres- enceofevolvingdeuteriumgaswasfirstreportedbySzpaketal. [22]. Fortheexperi- mentsatNHE,amodifiedplatingsolutionwasusedconsistingof0.025MPdCl ,0.15 2 MND Cl,and0.15MND ODinD O[19]. Nolithiumsaltswereused. Themixing 4 4 2 ofthesechemicalsproducedanorangesolutionandtheformationofaprecipitate.This precipitatewaslikelyPd(OD) duetotheratherhighinitialpHofthesolution(pH=9 2 to10). Threeco-depositionexperimentswereconductedatNHEusingtheFleischmann–Pons calorimetric cells. The initial current was 0.006 A in each cell. The deposition of palladiumontothecoppercathodewasvisiblewithinafewminutes,andthecopperwas completelycoveredbyadarkpalladiumdepositwithin30minutes. After24hours,the platingsolutionwasnearlyclearandgassingwasreadilyvisibleatthePd/Cucathode. Thecurrentwasthenincreasedto0.100Aineachcell. Onthesecondday,thesolution had turned to a pale yellow color. The current was then increased to 0.200 A, but a chlorineodordevelopedintheroom; hence, thecurrenthadtobereducedto0.020A fortheweekend.Thefollowingweek,thecellcurrentswereincreasedto0.100A,then 0.200A,andfinallyto0.400Awithoutanyfurtherproblemswiththechlorineodor. Theexcesspowerforthesethreeco-depositioncellsisshowninFigure5. Excesspowerisgeneratedineachcell. Duringthelast2daysofthisexperiment,about 400mWofexcesspowerwaspresentinbothCellsA-1andA-3whileabout100mW ofexcesspowerwaspresentinCellA-2.Theseresultshavebeenincludedinarefereed 27 Figure5. Excesspowermeasurementsinco-depositionexperiments. publication[23]. Therawdatasetsfortheseexperimentsstillawaitcompleteanalysis. BasedonotherindependentanalysisfortheseFleischmann–Ponscells,evenlargerex- cess heat effects are likely. The co-deposition results for Cell A-2 were evaluatedby Fleischmann,andresultsreportedattheMarch2001meetingoftheAmericanPhysical Society [24]. There is clear evidence for positive feedback effects in this experiment [24]. 28 6.0ConcludingRemarks. ThisnewfieldofanomalouseffectsinthePd/Dsystemhasenduredadifficult12-year survivalstruggle. Manyscientistswhohavepersistedwiththisresearchhaveseentheir careers placed in jeopardy. Nevertheless, no scientific publications have clearly dis- provedanyclaimsofexcessheat,heliumproduction,radiation,ortritium. Incontrast, similar results for this research have been reported by many laboratories around the world. Unfortunately,thisnewfieldwasdismissedfromthescientifictablein1989by ridicule rather than bythe properapplication of the scientific method. Arecent book hasclearlydocumentedthisstruggle[25].Intheend,thescientifictruthaboutthisfield willprevail. 7.0References. 1. D.E.Stilwell,K.H.Park,andM.H.Miles,J.FusionEnergy,9(1990)333. 2. ”Cold Fusion Research. A Report of the Energy Research Advisory Board to the United States Department of Energy”, John Huizenga and Norman Ramsey, co- chairmen,November1989,p. 12. 3. M.H.Miles,K.H.Park,andD.E.Stilwell,J.Electroanal.Chem. 296,241(1990). 4. M.H.Miles,K.H.Park,andD.E.Stilwell,Proc.oftheFirstAnnualConferenceon ColdFusion,(ICCF–1),1990,SaltLakeCity,Utah,pp. 328-334. 5. M. H. Miles, R. A. Hollins, B. F. Bush, J. J. Lagowski, and R. E. Miles, J. Elec- troanal. Chem.,346,99(1993). 6.D.D.Dominguez,P.L.Hagans,andM.A.Imam,TechnicalReport–NRL/MR/6170- 96-7803,Jan. 1996. 7. M.H.Miles,B.F. Bush,andK.B.Johnson,TechnicalReport–NAWCWPNSTP 8302.,Sept. 1996. 8. M.H.Miles,B.F.Bush,andJ.J.Lagowski,FusionTechnol.,25,478(1994). 9. M.H.MilesandB.F.Bush,Proc. ofICCF-7,1998,Vancouver,Canada,p. 236. 10. S.Szpak,P.A.Mosier-Boss,andJ.J.Smith,PhysicsLetters,A210,382(1996). 11.G.Preparata,Proc.oftheFirstAnnualConferenceonColdFusion,1990,SaltLake City,Utah,p. 91. 12. T.A.ChubbandS.R.Chubb,FusionTechnol.,20,93(1991). 13. Y.ArataandY.C.Zhang,Proc. ofICCF-8,2000,Lerici,Italy(pp. 11–16). 14. M. McKubre, F. Tanzella, P. Tripodi, and P. Hagelstein, Proc. of ICCF-8, 2000, Lerici,Italy(pp. 3–10). 15. B.F.Bush,J.J.Lagowski,M.H.Miles,andG.S.Ostrom,J.Electroanal. Chem., 304,271(1991). 16. M.H.Miles,J.Phys.Chem. B,102,3642(1998). 17. M.H.Miles,K.B.Johnson,andM.A.Imam,ProceedingsofICCF-6,1996,Lake Toya,Hokkaido,Japan,p. 20. 18. M.H.Miles,J.Electroanal.Chem.,482,56(2000). 19. M.H.Miles,NEDOFinalReport,NHELaboratory,Sapporo,Japan,1998. 20. M.H.Miles,Proc. ofICCF-8,2000,Lerici,Italy,(pp. 97–104). 21.T.Saito,M.Sumi,N.Asami,andH.Ikegami,Proc.ofICCF-5,1995,Monte-Carlo, Monaco,p. 105. 29 22. S. Szpak, P. A. Mosier-Boss, and J. J. Smith, J. Electroanal. Chem., 302, 255 (1991). 23. S.Szpak,P.A.Mosier-Boss,andM.H.Miles,FusionTechnol.,36,234(1999). 24. M.H.Miles,S.Szpak,P.A.Mosier-BossandM.Fleischmann,Proc. oftheAmer- icanPhysicalSociety,March18-22,2001,Seattle,WA. 25.C.Beaudette,ExcessHeat,WhyColdFusionResearchPrevailed,OakGrovePress, LLC,SouthBristol,Maine,2000. 30 CHAPTER4: ANALYSISOFEXPERIMENTMC–21: ACASESTUDY. S.Szpak,P.A.Mosier–Boss,M.H.Miles,M.A.Imam,andM.Fleischmann PARTI:DEVELOPMENTOFDIAGNOSTICCRITERIA. I/1.0Introduction. Inthetext,frequentreferenceismadetothe“ICARUSmethodology”and“experiment MC-21”. ThetermICARUSisanacronymforIsoperibolicCalorimetry: Acquisition, ResearchandUtilitiesSystem.Itisadocumentspecifyingcelldesign,operatingequip- ment,experimentalprotocol,anddataanalysis. ExperimentMc–21identifiesanexper- imentalrunconductedbyM.H.MilesattheNewHydrogenEnergy(NHE)laboratories inSapporo,Japan,whileonleavefromtheNWCChinaLake. This chapter contains two parts. the first deals with the development of diagnostic criteriafortheassessmentofexcessenthalpygenerationbasedonthemodellingofthe isoperiboliccalorimetersusedandleadingtothedefinitionofanumberofversionsof theheattransfercoefficient. Theseheattransfercoefficientsdefinethebehaviorofthe calorimetric systems. The second part contains the application of these criteria to a specificrun,e.g.,thatofexperimentMc–21. I/2.0Symbolsused. C p (cid:16) –heatcapacitanceoftheD Ovapor.[J(gMole) g 2 (cid:4) 1K (cid:4) 1] C p (cid:16) –heatcapacitanceofliquidD O.[J(gMole) l 2 (cid:4) 1K (cid:4) 1] E c (cid:8) t (cid:9) –cellvoltageattimet. [V] E th (cid:16) –thermoneutralpotentialatthebathtemperature.[V] b F –Faradayconstant. [coulombs(gMole) (cid:4) 1] H (cid:8) t (cid:21) t i (cid:9) i (cid:2) 1 (cid:19) 2–Heavysideunityshiftfunction. (cid:22) H (cid:8) t (cid:21) t i (cid:9) (cid:2) 0fort (cid:23) t; i H (cid:8) t (cid:21) t i (cid:9) (cid:2) 1fort (cid:5) t]. i D H –rateofevaporativecooling. [W] ev D H net (cid:8) t (cid:9) –rateofnetenthalpyinputatthetimet. [W] 31 I –cellcurrent.[A] k i (cid:16) j (cid:16) –heattransfercoefficient.[WK l (cid:4) 4]. L–latentheatofevaporation.[J(gMole) (cid:4) 1 M–numberofmoleofD Oatt 2 (cid:2) 0. P–vaporpressureatthecelltemperature. [bar] P (cid:24) –atmosphericpressure. [bar] P –pressureofdeuterium. [bar] D2 Q f (cid:8) t (cid:9) –generationofexcessenthalpyinthecellattimet. [W] t –time. [s] Dq –temperaturedifferencebetweenthecellandthewaterbath. [K] q –bathtemperature. [K] b µ–chemicalpotential. [J] µ–electrochemicalpotential. [J] t –time. [s] f –Galvanipotential. [V] F –proportionalityconstantrelatingconductiveheattransfertotheradiative I/3.0Calorimetry:thegoverningequation. Atlowtointermediatecelltemperatures(i.e., 30 (cid:25) C (cid:23) q (cid:23) 80 (cid:25) C), thebehaviorof the calorimeters,showninFig. 1,ismodelledadequatelybythedifferentialequation: (cid:26) C p (cid:16) dDq M l dt (cid:27) (cid:2) (cid:8) (cid:22) E c (cid:21) E th (cid:9) I (cid:28) (cid:3) (cid:22) Q f (cid:8) t (cid:9)(cid:29) (cid:28) (cid:3) (cid:22) D QH (cid:8) t (cid:21) t 1 (cid:9)(cid:30) (cid:21) D QH (cid:8) t (cid:21) t 2 (cid:26) (cid:9)(cid:31) (cid:28) (cid:21) 3I 4F (cid:26) P (cid:8) t (cid:9) P (cid:24) (cid:21) P (cid:8) t (cid:9) (cid:27) (cid:8) (cid:22) C p (cid:16) g (cid:21) C p (cid:16) l (cid:9) Dq (cid:8) t (cid:9) (cid:3) L l (cid:28) (cid:27) (cid:21)" ! k # R (cid:8) (cid:22) q b (cid:3) Dq (cid:8) t (cid:9)$ (cid:9) 4 (cid:21) q 4 b (cid:28)&% (2) where terms in square brackets indicate that the time rate of change in the enthalpy contentofthecalorimeterequalsthesumoftherateofenthalpyinputduetoelectrol- ysis, rate ofexcessenthalpygeneration, thecalibration pulseless therate ofenthalpy removalinthegasstream,andtherateofheattransfertothewaterbath,giveninEq. (1)bytherateofradiativetransferalone. InarrivingatEq.(1),wehavemadeanumberofapproximations,1themajoronebeing therepresentationoftheheattransferterm (cid:21) k0q 3 R b (cid:22) 1 (cid:21) g t (cid:28) (cid:26) (cid:8) q b (cid:3) Dq (cid:8) t (cid:9)’ (cid:9) 4 q 3 b (cid:3) 4FDq (cid:8) t (cid:9) (cid:27) (3) bythepurelyradiativetermwithanappropriateincreaseoftheradiativeheattransfer coefficienttok # . Aswehaveshownelsewhere[1],thisleadstoasmallunderestimate R oftheheatoutputand,therefore,toasmallunderestimateofQ .2 Atthefirstlevelof f 1Extensivediscussioncanbefoundine.g.,ref.[1]. 2WehavesoughtthroughouttoensurethatallapproximationsshouldleadtounderestimatesofQf. 32 Fig.1. SilveredDewarcalorimeter. approximation,wecanneglecttheresidualsmalltimedependenceoftheheattransfer coefficient. WiththecalorimeterssuppliedwiththeICARUSsystems,theconductivecontribution to heat transfer was very small. In fact, if this term was “lumped” into the radiative termbyallowingforasmallincreaseintheradiativeheattransfercoefficient: radiativeheattransfer (cid:2) (cid:8) k (cid:16) # 0 R (cid:9)( (cid:22) 1 (cid:21) g t (cid:28)(cid:29) (cid:8) (cid:22) q b (cid:3) Dq (cid:9) 4 (cid:21) q 4 b (cid:28) (4) then the values of the “pseudo-radiative” heat transfer coefficient derived (k (cid:16) # 0 R (cid:9)) (cid:22) 1 (cid:21) g t],wereclosetothosecalculatedfromtheStefan-Boltzmanncoefficientand theradiativesurfacearea.3 4 IfthetimedependenceoftheheattransfercoefficientisnotincludedexplicitlyinEq. (3),then radiativeheattransfer (cid:2) (cid:8) k # R (cid:9)) (cid:8) (cid:22) q b (cid:3) Dq (cid:9)(cid:31) (cid:28) 4 (cid:21) q 4 b (cid:28) (5) wheretheradiativeheattransfercoefficient(k # )nowshowsaweaktime–dependence. R 3Typicalvalues: 0.72 * 10 + 9WK + 4 , (k -. 0 R / [1–g t] , 0.76 * 10 + 9WK + 4. However,forthecellused intheexperimentMc–21,this“pseudo-radiative”heattransfercoefficientis0.85 * 10 + 9WK + 4sothatthe conductivecontributionwasevidentlyincreased. Wehavetoassumethatthisincreaseintheheattransfer coefficientmusthavebeenduetoa“softening”ofthevacuumintheDewarcalorimeters. 4Anincreaseinthe“pseudo-radiative”heattransfercoefficientcannormallyonlybeobservedifthecells are“overfilled”withD2Oduringtheperiodicreplenishmentofthecells.This“overfilling”ofthecellsleads totheapproachoftheelectrolytelevel tothebaseoftheKel-Fplugsealingthecellstherebyincreasing theconductive lossesthrough thetop ofthecell. Thiseffect (whichleadsto a4to 5%increase inthe “pseudo-radiative”heattransfercoefficient)canbeobservedintheresultsforday61ofexperimentMc21. 33 I/4.0Excessenthalpygeneration. Whether or not a particular cell generated excess enthalpy is determined by energy balance. We,therefore,needtoexamineallthetermsinEq. (1). I/4.1Enthalpyremovalbygasstream. Incalculatingtherateofenthalpyremovalbythegasstream,Eq. (5), (cid:8) 3I 0 4F (cid:9)) (cid:22) P 0 (cid:8) P (cid:24) (cid:21) P (cid:9)(cid:31) (cid:28)(cid:29) (cid:8) (cid:22) C p (cid:16) g (cid:21) C p (cid:16) l (cid:9) Dq (cid:3) L (cid:28) (6) we have always assumed that the partial pressure of D O in this gas stream can be 2 calculatedusingtheClausius–Clapeyronequationwiththelatentheatofevaporation,L, beingthatattheboilingpoint. Evaporativecoolingonlybecomesanimportanttermat temperaturesclosetotheboilingpoint(atcaDq (cid:5) 70 (cid:25) C)wherethesetwoassumptions are justified. At lowto intermediate temperatures, D H ev (cid:8) t (cid:9) is a minor correctionterm so that errors due to the two assumptionsintroduce second order small quantities. In particular,theerrorsintroducedbyneglectofthetemperaturedependenceofthelatent enthalpyofevaporationare (cid:23) 0.1%underallconditionsofoperationofthecells. Itis also important to bear in mind that such errors are further reduced for all evaluations ofthe“true”heattransfercoefficients,astheseevaluationsarebasedondifferencesin temperatureinducedbythecalibrationpulses(orondifferencesintemperatureinduced by “topping” up of the cells or perturbations of the current density; such methods of calibrationarenotconsideredinthisreport). TheparametersrequiredforthiscalculationwerecontainedindatafilesoftheICARUS- 1 and ICARUS-2 software and were identical for both systems5. The Handbooks [2, 2A] contained specific instructions that some of these parameters would need to be changed(here,q andP b (cid:24) ;seebelow)aswellasinstructionsastohowsuchchangesin theparameterlistingweretobecarriedout.6 The first of these changes is the adjustment of the boiling point to the value which 5Thevaluesinstalledintheprogramsassuppliedwere: Cp . l=84.349JMol 1 1K + 1 Cp . g=44.500JMo1 + 1K + 1 q b=374.570K + 1 F=96484.56CMole + 1 R=8.314410JMole + 1K + 1 L=41,672.600JMole + 1 P=1.003Ats 6However,itappearsthatvaluesoftheratesofevaporativeenthalpylossclosetothosegivenintheNHE. Analysesmaybecalculatedforlowtointermediatetemperaturesusingtheparameterlistingsuppliedwiththe instruments,i.e.,thechangesrequiredwerenotmade.(Italsoappearsthatthelatententhalpyofevaporation wasnotcorrectedforchangesintemperature.) Theconsequenterrorsaresufficientlysmallthattheydonot invalidatetheanalyses.However,thevaluesoftheratesofevaporativeenthalpylosscontainedinthe(k - R / 11- spreadsheetsoftheNHEanalysesfortemperaturesclosetotheboilingpointcannotbecalculatedwithany valuesoftheparametersclosetothosecontainedinthelistingsuppliedwiththeinstruments.Thismatterhas notbeeninvestigatedfurtherasitisinanyeventnecessarytomakethreefurtherchangesifexperimentcycles closetotheboilingpointaretobeevaluated. 34 appliestotheambientatmosphericpressure.7 ThevaluesofP (cid:24) thathavebeenusedin thepresentinterpretationhavebeenobtainedfromtheSapporoAirport. Furthermore, it has been assumed that the pressure in the cell is the same as the ambient pressure, although it may well be that the pressure in the cell was somewhat higher than this value. Thesecondchangeisthatitisnecessarytotakenoteofthefactthattheboilingpoint correspondstothatoftheelectrolytesolutioninthecell. Ithasbeenassumedthatthis correction is given by that for an ideal solution Dq bp (cid:2) (cid:8) R 0 L (cid:9) (cid:8) q 2 lnx bp 1 (cid:9) where x is the mole fraction of the D O in the electrolyte. 1 2 Itwillbeevidentthatthiscorrectionbecomesespeciallyimportantonday68whenthe cellcontentsare driventodryness. Inthat case, the boilingpoint must beadjustedat eachmeasurementintervalastheD Ocontentofthecelldecreases. Thevaluesofthe 2 boilingpointsappropriatefortheinterpretationoftheexperimentaldataforday68are discussedfurtherinvol. II/5.0 I/4.2Rateofreflux. The third change again applies specifically to day 68, namely, an allowance for the effect of reflux in the cell. In order to evaluate the effects of reflux, we need to take note of the fact that the vapor space in the cell is filled predominantly by D Oas the 2 cell is driven to dryness. Thus, even at the start of day 68, the mole fraction of D O 2 inthevaporspacewasca0.85forthisexperiment. Inconsequence,heattransferfrom the vapor phase to the walls of the Dewar (to providethe enthalpy input required by radiationacross thevacuumgap)was dominantlyfromthe D Ocontentofthe vapor. 2 Wealsoneedtotakenoteofthefactthatthecontributiontotheheatcapacitanceofthe vaporphaseinthevicinityoftheboilingpointduetotheD Ocontentofthisphaseis 2 (cid:21) d (cid:8) LP 0 P (cid:24) (cid:9)$ 0 dDq (cid:2) L2 0 Rq 2 e bp (cid:4)2 LDq 3 Rq 2 bp 4 (7) whereDq isthetemperaturedisplacementfromtheboilingpoint. Thisheatcapacitance isca67timeslargerthanthatofequivalentgasspacefilledwithoxygenandhydrogen and,therefore,ca380timeslargerthantheheatcapacitanceduetothesegasesforthe actualworkingconditionsatthestartofday68. Thismarkedincreaseintheheatcapacitanceofthepartofthecellfilledwithgasand D Ovaporhastwoconsequences.Inthefirstplace,theheattransferacrossthevacuum 2 gap must be maintained at the same value as that which applies to the liquid phase.8 Secondly, the radiative output across the section of the Dewar Cells filled with vapor must be balanced by the condensation of an amount of vapor sufficient to supply the radiativeenthalpy. 7ItshouldbenotedthattheICARUS-2systemwassuppliedwiththemeansforthecontinuousrecording ofthebarometricpressure,butthisfacilitywasevidentlydisabledfollowingtheinstallationofthesystem. 8Independentcalibrationsshowthattheheattransfercoefficientforcellsfilledwithairareabout0.75of thevaluesofthesecoefficientsforcellsfilledwithliquid[3,4].Itfollowsthereforethatthemarkedincrease intheheatcapacitanceofthecellsfilledwithD2Ovaporattemperaturesclosetotheboilingpointmustlead tothemaintenanceoftheheattransfercoefficientatthevaluewhichappliestocellsfilledwithliquid. 35 Afirstapproximationfortherateofrefluxinthecellis rateofreflux (cid:2) (cid:8) k # R (cid:9) f 12 1 (cid:8) q (cid:9) SD M 0 LD tM0 (8) where D M is amount of D O evaporated in each measurement interval, D t. Equation 2 (7) representsan upper limit for thisextent of reflux since we are neglectingthe heat transfertothewallsbythedeuteriumandoxygeninthegasspaceaswellastheeffects ofthereheatingofthisgasspacebytheliquidinthelowersectionofthecalorimeter.9 ItwillbeevidentthatanalysesbasedontheuseofEqs. (5–7)canonlybeapproxima- tions. Twoofthemostobviousdeficienciesaretheuseofdilutesolutiontheoryinthe interpretationandtheneglectofhydrostaticpressureontheboilingpointsusedinthe Clausius–Clapeyronequation.Itfollows,therefore,thatapartoftheanalysesoftheraw datafortheepisodesofcellsbeingdriventodrynessshouldbebasedonassumptions whichareindependentoftheuseofEqs(5–7). Thesemattersareconsideredfurtherin vol. II/5.0. I/5.0HeatTransferCoefficients: Definitionandevaluation. The heat transfer coefficients will be described by the suffices used previously, i.e. (cid:8) k # R (cid:9) i 5 j (cid:16) where i l (cid:2) 1 (cid:19) 2 (cid:19) 3 denotes differential, backward integration and forward inte- gration, j isdefinedatappropriatepointsbelowandl (cid:2) 1 (cid:19) 2denoteslowerboundand truecoefficients,respectively. Thesimpleststartingpointistoassumethatthereisno excess enthalpy generation in the system i.e., Q f (cid:8) t (cid:9) (cid:2) 0 in Eq. (1) and to evaluate a lowerboundheattransfercoefficient(i.e.,acoefficientwhichassumesthattherateof excessenthalpygenerationiszero)atatimejustbeforetheendofthecalibrationpulse, t (cid:2) t : 2 (cid:8) k # R (cid:9) 1 (cid:2) (cid:22) E c (cid:8) t (cid:9)(cid:30) (cid:21) E th (cid:16) b (cid:28) I (cid:21) D H ev (cid:8) t (cid:9)(cid:30) (cid:21) C M p (cid:8) dDq 0 dt (cid:9) f 1 (cid:8) q (cid:9) (9) where f 1 (cid:8) q (cid:9) (cid:2) (cid:22) q b (cid:3) Dq (cid:8) t (cid:9)(cid:29) (cid:28) 4 (cid:21) q 4isthetemperaturefunction. This,Eq.(9),wasthefirstheattransfercoefficientusedinourinvestigations;hence,the designation(k # R (cid:9) . Itshouldbenotedthatthisdesignationshouldreallybechangedso 1 astobeconsistentwiththedefinition(8),butthiswillnotbedoneprincipallybecause thedefinition(10)wassubsequentlyextendedtoanypartofthemeasurementcycle,the coefficientbeingdesignated(k # ) .10 R 11 Having obtained (k # ) , it is frequently desirable to establish the 11-point averages R 11 9ThegroupattheNHElaboratoriesattemptedtodeterminethevaluesofD Mdirectlybyaddingacon- densation section to the cells. It was difficult to see how anybody could convince themselves that such measurementscouldgivemeaningfulresults. Onewouldatbesthavederivedinformationaboutthereflux ratio,aquantitywhichdoesnotgiveanyusefulinformationabouttherateofexcessenthalpy generation. Theonlyuseableinformationisthedetectionofthetimeatwhichthecellsaredriventodryness. However, thistimecanbedetermineddirectlyfromrawdatabynotingthefallinthecellcurrentorbydirectvisual observation. 10Weshouldperhapschangethisdesignationof(k - R / 101 todenotei 6 1,differential; j=0,anypartof themeasurementcycle;l 6 1,lowerbound;butthedescription(k - R )11 willberetainedasithasbeenused extensivelyinearlierreportsandpapers. 36 (k # ) soastodecreasethenoise.11 Suchaveraginggivesca26independentvaluesfor R 11 measurement cycles lasting 1 day, or better ca 52 values for the recommended 2-day cycles. Inturn,itisusefultoevaluatethe6-pointaveragesof(k # ) whichhavebeen R 11 designatedas(k # ) . Itisnotusefultoextendthisaveragingbeyond6points,because R 11 anysuchextensionmakesthesystematicerrors(duetotheresidualdecreaseof(k # ) R 11 withtime)largerthantherandomerrors. I/5.1DeterminationofC M. p ItisapparentfromEq. (10)thatweneedaccuratevaluesofC M tomake(k p # ) gen- R 11 erallyuseful.12 AfirstapproachtothedeterminationofthevalueofC Mforanygiven p cellistorearrangeEq. (8)tothestraightlineform y (cid:2) mx (cid:3) c (10) i.e., (cid:22) E c (cid:8) t (cid:9)(cid:30) (cid:21) E th (cid:16) b (cid:28) I (cid:21) D H ev (cid:8) t (cid:9) f 1 (cid:8) q (cid:9) (cid:2) C M p (cid:8) dDq 0 dt (cid:9) f 1 (cid:8) q (cid:9) (cid:3) (cid:8) k # R (cid:9) 1 (cid:16) j (cid:16) (11) 1 andto derivethen approximatevalues ofC M from the slopes of the plots in regions p wherethetemperatureisvaryingrelativelyrapidlywithtime. Wecandistinguishfour suchplotsdesignatedbytherelevantderivedheattransfercoefficients(k (cid:16) # 0 R (cid:9) ,(k 151 (cid:16) # 0 R (cid:9) 161 (k (cid:16) # 0 R (cid:9) ,and(k 171 (cid:16) # 0 R (cid:9) accordingtowhetherthefittingofEq.(10)iscarriedoutattimes 181 somewhatabovetheorigin,attimessomewhatabovet (thetimeofapplicationofthe 1 calibrationpulse),attimessomewhatabovet (thetimeofcessationofthecalibration 2 pulse),orbythecombinationofthelasttwotimeregions,seeFig. 2.13 Itshouldbenotedthat(k (cid:16) # 0 R (cid:9) cannotbeevaluatedsystematicallyforexperimentMc– 151 21becauseoftheirregularscheduleoftheadditionofD O(seevol.II/1.3).Evaluations 2 of(k (cid:16) # 0 R (cid:9) ,(k 161 (cid:16) # 0 R (cid:9) and(k 171 (cid:16) # 0 R (cid:9) fortheimportantdatasetforday3aremarkedlyde- 181 gradedduetotheearlyonsetofpositivefeedback,seevol. II/1.3. Theprocedurebased on Eq. (10) has limited precision because of the need to differentiate the inherently noisy experimental data. It is therefore necessary to carry out the fitting procedures overextendedregionsoftheabscissae,(dDq 0 dt (cid:9)7 0 f 1 (cid:8) q (cid:9) ,sothatthedataareinevitably affectedbytheonsetofthepositivefeedbackdetectedfortheoperationofthecellon thatday. Inthisconnection,itshouldalsobenotedthatseparateinvestigationshaveshownthat (dDq 0 dt (cid:9) isbestestimatedbyusingthesecondordercentraldifferences(i.e.,thechords 11Otheraveragescanbemadebuttheuseofthe11-pointaveragehasbeenfoundtobeespeciallyuseful. 12ItisapparentthatthegroupatNHEretainedthevalueofCpMspecifiedintheparameterlistingrather thantodeterminethecorrectvalueandtosubstitutethiscorrectedvalueinthelisting. 13However, thereisameasureofambiguityabouttheinterpretationofthevaluesof(k -. 0 R / 1 . j . 1 derived, whichisdiscussedinvol.IIofthisreport. 37 Fig. 2. Evaluationof(k # R (cid:9) andC MaccordingtoEq.(10). 181 p ofthecurves).Moreaccuratevaluescouldbederivedinprinciplebyusinghigherorder differences. However,inpractice,therepeateddifferentiationoftheexperimentaldata (implicit when using higher order differences) leads to an increase in noise if we use differenceshigherthanthesecondorder.14 IntheabsenceofsufficientlyprecisedeterminationsofC M,theevaluationsmustnec- p essarilyberestrictedtoregionsoftimewherethecontributionofthetermC M p (cid:8) dDq 0 dt (cid:9)7 0 f 1 (cid:8) q (cid:9) is adequately small. In that case, it is adequate to use a “guesstimate” ofC M. This p matter (including the evaluation of a “guesstimate” of C M) is considered further in p vol. II/2.0. Itis nextnecessarytoevaluateatruedifferentialheattransfer coefficient. Thesimplestprocedure,giving(k # R (cid:9) neartheendofthecalibrationperiodattimet= 2 t ,isobtainedbyincludingthecalibrationpulse,15D Q: 2 (cid:8) k # R (cid:9) 2 (cid:2) D Q (cid:3) (cid:22) E c (cid:8) Dq 2 (cid:19) t 2 (cid:9)8 (cid:21) E c (cid:8) Dq 1 (cid:19) t 2 (cid:28) I f 2 (cid:8) q (cid:21) (cid:9) D H ev (cid:8) Dq 2 (cid:19) t 2 (cid:9) (cid:3) D H ev (cid:8) q 1 (cid:19) t 2 (cid:9)8 (cid:21) C M p (cid:8) (cid:22) dDq 0 dt (cid:9) Dq 2 (cid:16) t2 (cid:21) (cid:8) dDq 0 dt (cid:9) Dq 1 (cid:16) t2 (cid:28) f 2 (cid:8) q (cid:9) (12) 14Objectionshaveoftenbeenraisedtotheprocedureswhichwehaveadoptedbasedonthefactthatwe havenot“binnedthedata,”i.e.,wehavenotsignalaveragedbeforethedataanalysis.However,“binningof thedata”mustalwaysbeapproachedwithgreatcaution:oneshouldonly“bindata”or“bincoefficients”if thesedataorcoefficientsaretobeexpectedtobeconstantovertheaveraginginterval.Thisisnotthecasefor 9 k - R / 11unlesstheeffectsofthetermCpM 9 dDq : dt / havebeentakenintoaccount.Oncethisisdonewecan, ofcourse,binthecoefficientsaswehavedoneinderiving 9 k - R / 11and(k - R / 11[aswellas 9 k - R / 181]. 15 9 k - R / 2wasthesecondheattransfercoefficientusedinourinvestigations. 38 wherewenowhave f 2 (cid:8) q (cid:9) (cid:2) (cid:22) q b (cid:3) (cid:8) Dq 2 (cid:19) t 2 (cid:9)(cid:29) (cid:28) 4 (cid:21); (cid:22) q b (cid:3) (cid:8) Dq 1 (cid:19) t 2 (cid:9)(cid:29) (cid:28) 4 (13) Inordertocarryoutsuchevaluations,itisusefultoconstructA.4-orA.3-sizedplotsof therawdataandthentoobtainappropriateaveragesbyusingatransparentruler. This typeofanalysis usedto be a generallyaccepted approachbut thenfell intodisrepute. However,themethodologyisnowagainacceptedgivingso-calledrobustestimates. I/5.2Precisionandaccuracy–differentialcoefficients. It may be noted that the errors in (k # R (cid:9) are measures of the accuracyof the true heat 2 transfercoefficientastheestimatesaremadeintermsoftheknownJouleenthalpyinput tothecalibrationheater. Errorsin(k # R (cid:9) or(k 1 # R (cid:9) aremeasuresoftheprecisionofthe 11 lowerboundheattransfercoefficientsasthereisnoindependentcalibrationandthere maybeexcessenthalpygenerationinthesystem. Itisimportantthat(k # R (cid:9) and(k 11 # R (cid:9) 2 aretheleastpreciseandleastaccuratecoefficientswhichcanbeobtainedfromtheraw data. Statements that the errors are larger than this (e.g., see [5]) simply show that mistakes have been made in the data analysis procedures and/or the execution of the experiments. We have always insisted that the construction and evaluation of plots of the raw data is an essential prerequisite of the more elaborate data evaluation procedures. For one thing,itshowswhetherthenoiselevelsintheexperimentsweresufficientlylowtojus- tifymoredetailedevaluationsandalsopointstomalfunctionsintheexperiments.Italso shows immediately whether the q (cid:21) t and E c (cid:21) t transients have relaxed sufficiently to permit the evaluation of (k # R (cid:9) 1 and(k # R (cid:9) . Furthermore,itgivesimmediateindicationsofthepresence(orabsence)of 2 positivefeedback.Ashasbeenpointedoutrepeatedlyallcalibrationproceduresrequire thattherateofexcessenthalpygeneration,Q f (cid:8) t (cid:9) ,beconstantduringthecalibrationpe- riods. Thesemattersareconsideredfurtherinthemaintext,vol. II/2.0andvol.II/3.0. Havingobtainedthetrueheattransfercoefficientatasinglepoint(usuallyneartheend of the calibration pulse, t (cid:2) t ) it is important to ask: what is the true heat transfer 2 coefficient, (k # R (cid:9) , at any other time? We can make such an evaluation within the 12 duration t 1 (cid:23) t (cid:23) t of the calibration pulse simply by using Eq. (11 ) giving (k 2 # R (cid:9) 12 ratherthan(k # R (cid:9) . NotealsothatEq. (11)canberearrangedtothestraightlineform 2 D Q (cid:3) (cid:22) E c (cid:8) Dq 2 (cid:19) t (cid:9)(cid:30) (cid:21) E c (cid:8) Dq 1 (cid:19) t (cid:9)(cid:31) (cid:28) I (cid:21) D H ev (cid:8) Dq 2 (cid:19) t (cid:9) (cid:3) D H ev (cid:8) Dq 1 (cid:19) t (cid:9) f 2 (cid:8) q (cid:2) (cid:9) C M p (cid:8) (cid:22) dDq (cid:9)7 0 dt (cid:9) Dq 2 (cid:16) t (cid:21) (cid:8) dDq 0 dt (cid:9) Dq 1 (cid:16) t (cid:28) f 2 (cid:8) q (cid:9) (cid:3) (cid:8) k #. 0 R (cid:9) (14) 162 which is applicable at times close to and abovet . It is evident, therefore, that such 1 plots can also be used to obtain estimates ofC M, but the accuracyof such values is p inevitably much lower than the precision of those obtained by the application of the 39 corresponding expression for the lower bound heat transfer coefficient, (k # R (cid:9) , Eq. 161 (10). Nevertheless, Eq. (11) is useful because it allows the removal of the effects of the water equivalent, C M, on the true heat transfer coefficient, (k p (cid:16) # 0 R (cid:9) , simply 162 byextrapolatingtozerovalueoftheabscissa. However,thetimecorrespondingtothis pointwillnotbeaccessibleexperimentallyforcalibrationscarriedoutwithacalibration pulseof6-hourdurationforpolarizationscarriedoutatlowcellcurrents(althoughthis timeisprobablyclosetot (cid:2) t ).16 2 In the regions in which there is no application of a heater pulse, i.e., for 0 (cid:23) t (cid:23) t 1 andt 2 (cid:23) t (cid:23) T,thetrueheattransfercoefficientcanonlybeobtainedfromtheheating and cooling curves,i.e. the driving force is the change in the enthalpycontent of the calorimetersratherthatD Q. ItisnowsensibletocastEq.(11)intheform 16Asimilarcommentappliestothedeterminationof 9 k -. 0R / 161:thetimeatwhich 9 dDq : dt 6 0 / willusually beaccessibletoexperimentsinwhicht1 6 9hours. However,nosuchpointcanbedefinedfor 9 k -. 0R / 171so thatthisdeterminationismathematicallyquestionable.Thisisthereforeequallytruefor 9 k -. 0R / 181,although theseextrapolationsarecertainlysoundfromanoperationalpointofview. 40 C M p (cid:8) (cid:22) dDq 0 dt (cid:9) Dq 2 (cid:16) t (cid:21) (cid:8) dDq 0 dt (cid:9) Dq 1 (cid:16) t (cid:28) f 2 (cid:8) q (cid:21) (cid:8) (cid:9) (cid:2) k (cid:16) # 0 R (cid:9) 152 (cid:3) (cid:22) E c (cid:8) Dq 2 (cid:19) t (cid:9)(cid:30) (cid:21) E c (cid:8) Dq 1 (cid:19) t (cid:9)(cid:31) (cid:28) I (cid:21) D H ev (cid:8) Dq 2 (cid:19) t (cid:9) (cid:3) D H ev (cid:8) Dq 1 (cid:19) t (cid:9) f 2 (cid:8) q (cid:9) (15) (samefor(k (cid:16) # 0 R (cid:9) .) 172 Ifthesystemisfunctioningcorrectly,thenitwillbefoundthattheL.H.S.ofEq. (14) is essentially constant (although this constancy can only be probed over a short time range). ThesecondtermontheR.H.S.ofEq. (14)willbemuchsmallerthantheterm ontheL.H.S.,i.e. itisinthenatureofacorrectiontermtogivepoint-by-pointvaluesof (k (cid:16) # 0 R (cid:9) or(k 152 (cid:16) # 0 R (cid:9) .Itwillbeevidentthattheaccuracyoftheseversionsofthetrueheat 172 transfercoefficientislimitedbytheaccuracyoftheestimatesofC M. Thisparticular p partofthemethodologyisthereforeonlyusefultoserveasacheckontheoperationof the cells and methods of data evaluation. Furthermore, it is not possible to apply Eq. (14)systematicallytothetimeregion0 (cid:23) t (cid:23) t forexperimentMc–21inviewofthe 1 irregularscheduleofadditionofD Otothecell.17 2 The assumptionunderlying thispart ofthe account presented in thisreport is that we canonlydetermine(k # R (cid:9) withinthedurationofthecalibrationpulset 12 1 (cid:23) t (cid:23) t ,Fig. 2 3, and, ataloweraccuracy,(k (cid:16) # 0 R (cid:9) and(k 152 (cid:16) # 0 R (cid:9) inregionsadjacenttotheoriginand 172 fortimesadjacentandabovet respectively.However,thisconclusionisincorrect. We 2 needtomaketheadditionalassumptionthattherateofanyexcessenthalpygeneration isconstantduringanyparticularcalibrationperiodinordertodetermine(k # R (cid:9) . 12 This meansthat we can only obtain a single valueofthis heattransfer coefficientper calibrationperiodand,consequently,asinglevalueof[(k # R (cid:9) –(k 12 # R (cid:9) ]. Twoimportant 11 points followfrom this conclusion. In the first place, the precision of (k # R (cid:9) must be 12 very nearly equal to the precision of (k # R (cid:9) . Secondly, and related to the first point, 11 we see that if we extend the assumption that the rate of excess enthalpy production is constant during the period t 1 (cid:23) t (cid:23) t to saying that it is constant for the whole 2 measurement cycle, 0 (cid:23) t (cid:23) T, then it is immediately possible to derive (k # R (cid:9) over 12 thewholeofthiscycle. Thus,ifthedifferencebetweenthetrueandlowerboundheat transfer coefficients can be established at any one time [say, D (cid:8) k # R (cid:9) at time t ], then t 2 (cid:22) k # R (cid:8) t (cid:9)(cid:29) (cid:28) atanyothertimet willbegivenby 12 (cid:22) k # R (cid:8) t (cid:9)(cid:31) (cid:28) 12 (cid:2) (cid:22) k # R (cid:9) (cid:8) t (cid:9)(cid:31) (cid:28) 11 (cid:3) D (cid:8) k # R (cid:9) f 1 t2 (cid:8) q (cid:9) t2 f 1 (cid:8) q (cid:9) (16) t1 The ratio f 1 (cid:8) q (cid:9) t2 0 f 1 (cid:8) q (cid:9) is of order unity, which implies that the shift 2 (k # R (cid:9) l2 (cid:21) (cid:8) k # R (cid:9) isalwaysclosetothatatthecalibrationpoint. 11 Equation (15) shows that the precision of (k # R (cid:9) is very nearly equal to the precision 12 of (k # R (cid:9) .18 It follows that changes in the rates of excess enthalpyproduction can be 11 17Ashasbeennotedpreviously(cfvol.II),wehavebeenunabletocombinedataintheregionsjustabove t1andt2togiveasimpleequationleadingto(k -. 0 R / 182. 18ThevalidityofEq.15wasestablishedatthetimeofconstructionoftheICARUS-2system. 41 Fig. 3. Schematicofmethodologyusedincalorimetercalibrating. establishedatthesamelevelofprecisionasthatof(k # R (cid:9) . Thesamecommentsapply 11 totheprecisionofthetrueheattransfercoefficient,(k # R (cid:9) relativetothatofthelower 22 boundheattransfercoefficient,(k # R (cid:9) ,whichisdiscussedbelow. Inconsequence,the 21 changesintheratesofexcessenthalpyproductioncanbeestablishedwithrelativeerrors (cid:23) 0.01%,andtheseerrorsdeterminethelevelofsignificancewithwhichsuchchanges canbediscussed. Ofcourse,theaccuracyofthetrueheattransfercoefficientsremains determinedbytheerrorsofdifferencessuchasthatof[(k # R (cid:9) 12 (cid:21) (cid:8) k # R (cid:9) ]. 11 It is important here to stress once again that any attempt to calculate the variation of rates of excessenthalpygeneration within the measurementcycles must also paydue regard to the fact that it is not possible to calibrate the systems if the rate of excess enthalpy generation varies with time. It is also important that this comment applies equally to any calorimetric system which we might wish to use. If the rate of excess enthalpy generation does, in fact, vary with time, then D (cid:8) k # R (cid:9) must be derived from separate experiments. This is the situation which applies to experiment Mc–21 as is discussedin vol. II/2.0andvol. II/3.0. Thecommentsmadein thispartofvol. I/5.0 shouldbereadinconjunctionwithvol. IIofthisreport. Thediscussionoftheaccuracyoftrueheattransfercoefficientsversustheprecisionof thelowerboundheattransfercoefficientspromptedoursearchformethodsthatwould increaseboththeprecisionandaccuracy.Thereasonforthelimitedprecisionof(k # R (cid:9) 11 andaccuracyof(k # R (cid:9) ismainlyduetotheneedtodifferentiatenoisyexperimentaldata 12 setsinordertoderiveC M p (cid:8) dDq 0 dt (cid:9) . 42 I/5.3Precisionandaccuracy–integralcoefficients. Ifwewishtoavoidthenumericaldifferentiationoftheexperimentaldatasets,thenwe can rely instead on the numerical integrations of these data and compare these to the integralsofthedifferentialequationrepresentingthemodelofthecalorimeters.Forthe backwardintegralsstartingfromtheendofthemeasurementcyclesatt (cid:2) T,weobtain (cid:8) k # R (cid:9) 21 (cid:2)= < t D H T net (cid:8) t (cid:9) dt < t f T 1 (cid:8) q (cid:9) dt (cid:21) C M p (cid:22) Dq (cid:8) t (cid:9)(cid:30) (cid:21) Dq (cid:8) T < (cid:9)(cid:31) (cid:28) t f T 1 (cid:8) q (cid:9) (17) dt whilethecorrespondingequationforforwardintegrationfromthestartofthemeasure- mentcycleis (cid:8) k # R (cid:9) 31 (cid:2) < tD H 0 net (cid:8) t (cid:9) dt < t f 0 1 (cid:8) q (cid:9) dt (cid:21) C M p (cid:22) Dq (cid:8) t (cid:9)8 (cid:21) Dq (cid:8) 0 (cid:9)(cid:29) (cid:28) f 1 (cid:8) q (cid:9) (18) dt Here, the suffices 21 and 31 denote respectively backward integration, lower bound andforwardintegration,lowerbound. (k # R (cid:9) and(k 21 # R (cid:9) arethecorrespondingintegral 31 heat transfer coefficients defined at time t. We have to take note of the fact that care is needed when integrating the terms f 1 (cid:8) q (cid:9) and net enthalpy input, D H net (cid:8) t (cid:9) , around the discontinuities at t (cid:2) t and t 1 (cid:2) t (also the times t 2 (cid:2) 0 and t (cid:2) T if the range of the integrations is extended).19 It may be noted that the only straightforward way in which we can integrate around the discontinuities at t (cid:2) t andt 1 (cid:2) t is by means 2 ofthetrapeziumruleandthisisthemethodwhichhasbeenusedintherecalculations presentedinthisreport.Ifthetimesofapplicationandcessationoftheheatercalibration pulsescorrespondexactlytot andt respectively,thenwecancarryouttheintegrations 1 2 aroundthediscontinuitiesbyinsertingextradatapointsatthesetimes. Itappearsthat the data sets in experiment Mc–21 satisfy this criterion although this is not generally trueforallexperimentscarriedoutwiththeICARUS-2system;lackofsynchronization ofthecalibrationpulseswitht andt appearedtobegenerallytrueformeasurements 1 2 with the ICARUS–1 systems. In that case, it is necessary to determine these times separately (this can be done adequately from the q (cid:21) t plots) so as to establish the integrationintervalsanditisthennecessarytoinsertfouradditionaldatapoints.20 Theadequacy(orinadequacy)oftheparticularintegrationprocedurescoupledtothead- equacyofthechosenintegrationintervalisrevealedmoreclearlywhenwecometothe useofEqs. (16)and(17)todetermineC M andtocarryoutextrapolationstoremove p the effects of the second term on the R.H.S. of these equations on the corresponding 19Atdifferenttimes,thetrapeziumrule,Simpson’sruleorthemid-pointrulehavebeenusedtocarryout theintegrations. Oftheserules,onlythemid-pointruleisstrictlyspeakingcorrectinthatitagreeswiththe mathematicaldefinitionofanintegral. Itisquitegenerallyassumedthatintegrationscarriedoutusingthe trapeziumorSimpson’srulewillconvergeontothecorrectalgebraicresultiftheintegrationintervalismade adequatelysmall,butthisdoesnotnecessarilyfollow. Thisisamatterwhichneedstobeinvestigatedfor eachparticularcase. 20Theevaluationsof(k - R / 21and(k - R / 22(seebelow)andof(k - R / 31and(k - R / 32(seealsobelow)weretohave beencarriedoutusing(k - R / 21and(k - R / 31spreadsheetsproducedbythesoftware.Aswehaveneverhadaccess tothesespreadsheets(if,infact,theywereeverproduced),wecannotnowestablishwhethertheintegrations aroundthediscontinuitieswerecarriedoutcorrectly,althoughwebelievethattheymusthavebeeninerror. Inanyevent,alltheintegrationsusedintheevaluationsdescribedinthisreporthavebeencarriedoutusing therawdata. 43 heat transfer coefficients. The procedure set out in the handbook for the ICARUS-1 System [2, 2A] restricted the integrations to the time region of the application of the heatercalibrationpulse. Forbackwardintegration,weobtain < t D H t2 net (cid:8) t (cid:9) dt < t f t2 1 (cid:8) q (cid:9) dt (cid:2) C M p (cid:22) Dq (cid:8) t (cid:9)(cid:30) (cid:21) Dq (cid:8) t 2 < (cid:9)(cid:31) (cid:28) t f t2 1 (cid:8) q (cid:9) dt (cid:3) (cid:8) k (cid:16) # 0 R (cid:9) (19) 261 whileforforwardintegration,weobtain < t D H t1 net (cid:8) t (cid:9) dt < t f t1 1 (cid:8) q (cid:9) dt (cid:2) C M p (cid:22) Dq (cid:8) t (cid:9)(cid:30) (cid:21) Dq (cid:8) t 2 < (cid:9)(cid:29) (cid:28) t f t1 1 (cid:8) q (cid:9) dt (cid:3) (cid:8) k (cid:16) # 0 R (cid:9) 361 > (20) Equation(18)canbeusedtoderiveaccuratevaluesofC Mwhilethereissomeminor p degradationwhenusingforwardintegration,Eq. (19). TheapplicationofEq. (18)to thedatasetswasthetargetmethodologyoftheICARUSsystems(e.g.,see[2,2A])and the derivedlower bound heat transfer coefficient, (k (cid:16) # 0 R (cid:9) , wasdescribed as (k 261 # R (cid:9) in 21 the Handbookand the associated correspondence. We have since then used the more extendeddescription,(k (cid:16) # 0 R (cid:9) ,todenotethefactthatwith j 261 (cid:2) 6,wearecarryingoutthe evaluationinthetimeregiont 1 (cid:23) t (cid:23) t . Thesametypesofevaluationmaybeusedto 2 derive(k (cid:16) # 0 R (cid:9) ,(k 251 (cid:16) # 0 R (cid:9) ,and(k 271 (cid:16) # 0 R (cid:9) aswellas(k 281 (cid:16) # 0 R (cid:9) ,(k 351 (cid:16) # 0 R (cid:9) and(k 371 (cid:16) # 0 R (cid:9) .Itisonly 381 necessarytostarttheintegrationsfromtheappropriatetimeswhichalsogivethestarting values of q for the R.H.S. of the relevant equations. Of these sets of estimates, that leadingto(k (cid:16) # 0 R (cid:9) isespeciallyusefulandthisparticularfitalsogivesgoodestimates 281 ofC M. However,itshouldbenotedthatitisnecessarytousecareinapplyingthese p procedures to the data for day 3 of experiment Mc–21 because of the early onset of positivefeedback,seevol. II/2.0. Inordertoobtainthetrueheattransfercoefficientsitisnecessarytocombinetheinte- gralsoftheenthalpyinputsinEqs. (18)and(19)withthermalbalancesmadeatoneor a series of points. This can be done in a numberof ways and itis important that this partoftheevaluation[2,2A]waschangedinthesummerof1994followingthereceipt of the first two sets of data collected by NHE. Attention will be confined here to the procedureoriginally suggested in the Handbookfor the ICARUS-1 system [2, 2A].21 We makea thermal balance justbeforethe applicationofthe calibrationpulseand, if thesystemhasrelaxedadequatelyandifdq 0 dt (cid:2) 0,thenifweconsider(k # R (cid:9) , 32 0 (cid:2) D H net (cid:8) t 1 (cid:9)) (cid:22) t (cid:21) t 1 (cid:28) (cid:3) Q f (cid:22) t (cid:21) t 1 (cid:28)? (cid:21) (cid:8) k # R (cid:9) 32 (cid:8) (cid:22) q b (cid:3) Dq (cid:8) t 1 (cid:9)’ (cid:9) 4 (cid:21) q 4 b (cid:28)(cid:31) (cid:22) t (cid:21) t 1 (cid:28) (21) combinationwithEq. (14)eliminatestheunknownrateofexcessenthalpygeneration. Weobtain (cid:8) k # R (cid:9) 32 (cid:2) < t D H t1 net (cid:8) t (cid:9) dt (cid:21) D H net (cid:8) t 1 (cid:9) (cid:8) t (cid:21) t 1 < (cid:9) t f t1 2 (cid:8) q (cid:9) dt (cid:21) C M p (cid:22) Dq (cid:8) t (cid:9)(cid:30) (cid:21) Dq (cid:8) t 1 < (cid:9)(cid:31) (cid:28) t f t1 2 (cid:8) q (cid:9) dt > (22) 21Itisinanyeventnecessarytochangethemethodologyoftheevaluationinviewoftheearlyonsetof “positivefeedback”(II/3.0). 44 Thecorrespondingequationfor(k # R (cid:9) followsfromEq.(21)onreplacingt byt . (see 22 1 2 Eq. (22)below).22 Itisconvenientalsotorewritethederivedequationfor(k # R (cid:9) inthe“straightlineform” 22 < t D H t2 net (cid:8) t (cid:9) dt (cid:21) D H net (cid:8) t 2 (cid:9) (cid:8) t (cid:21) t 2 < (cid:9) t f t2 2 (cid:8) q (cid:9) dt (cid:2) C M p (cid:22) Dq (cid:8) t (cid:9)(cid:30) (cid:21) Dq (cid:8) t 2 < (cid:9)(cid:31) (cid:28) t f t2 2 (cid:8) q (cid:9) dt (cid:3) (cid:8) k (cid:16) # 0 R (cid:9) (23) 262 (k # R (cid:9) and(k 22 (cid:16) # 0 R (cid:9) weretheversionsofthetrueheattransfercoefficientthatweused 262 inourinvestigationspriortotheconstructionoftheICARUS–1system. Aswedidnot wish to discuss the differences between (k # R (cid:9) , (k 32 (cid:16) # 0 R (cid:9) , (k 362 # R (cid:9) and (k 22 (cid:16) # 0 R (cid:9) , and, as 262 weexpected(k # R (cid:9) toconvergeonto(k 32 # R (cid:9) forthespecified2-daymeasurementcycles 22 (withintheerrorlimitsspecifiedfortheICARUS-1system)wealsolabelled(k # R (cid:9) as 32 (k # R (cid:9) 23. 22 It should be noted that the extrapolation (21) automatically removesthe effect of the termC M p (cid:22) q (cid:8) t (cid:9)@ (cid:21) q (cid:8) t 2 (cid:9)(cid:29) (cid:28)A 0 < t f t2 2 (cid:8) q (cid:9) dt onthetrueheattransfercoefficient. This applica- tionofEq. (21)(andof(k # R (cid:9) evaluatedclosetothemid-pointt 22 (cid:2) t )wasoneofthe 2 major objectivesfor our methodologybecauseC M is theleast accurateparameterin p theanalysis. WhileitisalsopossibletowriteEq. (21)intheform(22)togive(k (cid:16) # 0 R (cid:9) ,thismethod 362 ofanalysisisnotusefulastherangeoftheextrapolationrequiredistoolong[2,2A](see alsovol. II).Forthisreason,itwasrecommendedintheICARUS-1Handbook[2,2A] that(k # R (cid:9) shouldbeevaluatedattimesclosetot 32 (cid:2) t usingvaluesofC Mdetermined 2 p fromapplicationsofEq. (19). However,inviewofthe errorsin thedeterminationof C M,thesevaluesof(k p # R (cid:9) areinevitablylessaccuratethanthoseof(k 32 # R (cid:9) or(k 22 # R (cid:9) 262 (seealsovol. II/3.0).24 WeshouldobservefurthermorethatEq.(21)issoundlybased(inamathematicalsense) in that the extrapolation to [Dq (cid:8) t (cid:9)B (cid:21) Dq (cid:8) t 2 (cid:9) ] = 0 gives the value of (k (cid:16) # 0 R (cid:9) at a well 262 definedtime,t (cid:2) t . Thisisequallytrueofallofthecoefficientsbasedonforwardor 2 backwardintegration;however,thestartingpointsfortheseintegrationswillusuallybe chosentobet (cid:2) 0,t (cid:2) t ort 1 (cid:2) T andthedefinitionoftheheattransfercoefficientsat thesepointsisnotgenerallyuseful.Theexceptionhereisthelowerboundheattransfer coefficient,(k (cid:16) # 0 R (cid:9) whichisalsodefinedattimet 261 (cid:2) t . Weobservealsothat(k 2 (cid:16) # 0 R (cid:9) 261 and(k (cid:16) # 0 R (cid:9) arethemostpreciseandaccuratevaluesofthelowerboundandtrueheat 262 transfercoefficientswhichcanbederivedwiththemethodologyaspresentlydeveloped. 22WenoteagainthatthegroupatNHEdidnotfollowtheinstructionintheICARUS-1Handbook[2,2A] tousemeasurementcyclesof2-daydurationand,forthereducedtimescalesof1-daycyclesinparticular,it isnecessarytoincludethetermCpM 9 dDq : dt)inthethermalbalances,Eq.(18).However,thegroupatNHE continuedtousetheoriginalformoftheequation.ItalsoappearsthatNHEdidnotfollowtheinstruction[2, 2A]toevaluate(k - R / 32attimesclosetot2.Thismatterisdiscussedfurtherinvol.II/3.0. 23Weretainedthedesignation22asaflagtoindicatethebackwardintegrationmethodologywasthepri- maryobjectiveforaccurateevaluations 24Wenoteherealsothatcaremustbetakenincarryingouttherequiredlinearregressionfittingprocedures asisillustratedinvol.II/3.0. 45 Furthermore,(k (cid:16) # 0 R (cid:9) =(k 261 # R (cid:9) att 11 (cid:2) t and(k 2 (cid:16) # 0 R (cid:9) =(k 262 # R (cid:9) att 12 (cid:2) t ,sothatthebest 2 valueofD (cid:8) k # R (cid:9) thatcanbeobtainedforuseinEq. (15)is t D (cid:8) k # R (cid:9) t (cid:2) (cid:8) k # R (cid:9) 12 (cid:21) (cid:8) k # R (cid:9) 11 (cid:2) (cid:8) k (cid:16) # 0 R (cid:9) 252 (cid:21) (cid:8) k (cid:16) # 0 R (cid:9) (24) 251 This sound basis of the heat transfer coefficients derived by forward and backward integration should be contrasted with the corresponding position for the differential heattransfercoefficientswhichhasbeendiscussedabove. I/6.0Timedependenceoftheheattransfercoefficients. Inthefinalpartofthissection,weneedtoconsidersomewhatfurtherthetimedepen- dence of thevariousforms of the heattransfer coefficient(compare [6]). We observe firstofallthatweareinterpretingherethesystematicvariationsoftypically0.4%ofthe differentialor0.2%oftheintegralcoefficients.Theonlyreasonwhyweareabletoin- vestigatesystematicvariationsofsuchsmallquantitiesistheveryhighprecisionofthe methods of data evaluation. We observe secondly, that as the differential coefficients areevaluatedatlocaltimes,theywillshowtheweaktimedependence: (cid:8) k # R (cid:9) (cid:2) (cid:8) k (cid:16) # 0 R (cid:9)) (cid:22) 1 (cid:21) g t (cid:28) (25) (c.f. Eqs. (3)and(5)). Inthedefinitionoftheintegralheattransfercoefficientsgiven in this section (k # R (cid:9) has been regarded as being constant whereas the investigation of the differential heat transfer coefficientsshowsthat we should really include the time dependence,Eq. (24),i.e.,wemustuseEq. (3)intheintegrations. Integrationofthis equationgives (cid:8) k (cid:16) # 0 R (cid:9) (cid:26)D C f 1 (cid:8) q (cid:9) dt (cid:21) g t C f 1 (cid:8) q (cid:9) dt (cid:3) g CE C f 1 (cid:8) q (cid:9) dt dt (cid:27) > (26) Ifwenowregard f 1 (cid:8) q (cid:9) asbeingconstantthroughoutthemeasurementcycle(whichis aroughapproximationforthecaseofthelowerboundheattransfercoefficients),then theintegralbecomes (cid:8) k (cid:16) # 0 R (cid:9) f 1 (cid:8) q (cid:9) t (cid:22) 1 (cid:21) g t 0 2 (cid:28) > (27) ItfollowsthattheheattransfercoefficientsgivenbyEqs(16)and(17)aregivenby (cid:8) k # R (cid:9) 21 (cid:2) (cid:8) k (cid:16) # 0 R (cid:9) 21 (cid:22) 1 (cid:3) g (cid:8) T (cid:21) t (cid:9)$ 0 2 (cid:28) (28) and (cid:8) k # R (cid:9) 31 (cid:2) (cid:8) k (cid:16) # 0 R (cid:9) 31 (cid:22) 1 (cid:21) g t 0 2 (cid:28) (29) withinthelimitsofthisapproximation. (k (cid:16) # 0 R (cid:9) and(k 21 (cid:16) # 0 R (cid:9) arerespectivelythevalues 31 of(k F R (cid:9) att 21 (cid:2) T andof(k # R (cid:9) att 31 (cid:2) 0.Itfollowsthattheslopesoftheplotsof(k # R (cid:9) 21 and(k # R (cid:9) versustimeareroughlyonehalfoftheplotof(k 31 # R (cid:9) versustime. 11 46 Equation(24) also showsthe way in whichwe can test whether the characteristicsof theDewarcellscanbedescribedbyasingle,time–independentheattransfercoefficient. Thus,evaluationof(k # R (cid:9) accordingtoEq. (16)givesustheheattransfercoefficient 21 (cid:8) k # R (cid:9) 21 (cid:2) (cid:8) k (cid:16) # 0 R (cid:9) 21 (cid:26) 1 (cid:21) g t (cid:3) g < t T < t f T 1 (cid:8) q (cid:9) dt dt < t f T 1 (cid:8) q (cid:9) dt (cid:27) (30) sothatthetimeindependentheattransfercoefficient,(k (cid:16) # 0 R (cid:9) isreadilydetermined.The 21 factthatheattransferfromthecellscanberepresentedbysuchasingletimeindepen- dent heat transfer coefficient has been demonstrated several times (e.g. see Fig. 51 of vol. II). Indeed, such representations are the basis of our statement that the inte- grallowerboundheattransfercoefficientscanbedeterminedwithaprecisiongivenby relativeerrorsoflessthan0.01%.25 The variations of (k # R (cid:9) , (k 11 # R (cid:9) and (k 21 # R (cid:9) with time show that this time dependence 31 of the heat transfer coefficients must be taken into account in evaluation of the rates ofexcessenthalpygenerationaimingatthehighestachievableaccuracy. Ifthisisnot done,thenthevaluesoftheheattransfercoefficientsatthemid-points,t (cid:2) t ,shouldbe 2 used. Inthatcase,thevaluesoftheratesofexcessenthalpygenerationcalculatedwill beslightlytoosmallfort (cid:23) t andslightlytoolargefort 2 (cid:5) t .However,thetotalexcess 2 enthalpycalculatedforacompletemeasurementcyclewillbeapproximatelycorrect.26 Wemustalsonotethatthedifferentialheattransfercoefficient,(k # R (cid:9) ,mustbeusedin 12 theevaluationsoftheratesofexcessenthalpygenerationandtheintegralheattransfer coefficientsintheevaluationoftheexcessenthalpy(includingthetotalexcessenthalpy forcompletemeasurementcycles). Inparticular,theuseof(k # R (cid:9) intheevaluationof 22 theratesofexcessenthalpygenerationwillunderestimatethesequantities. I/7.0RemarksconcerningICARUS-1dataevaluationproceduresandexperimen- talprotocols. ThemodellingoftheICARUS-1typecalorimeters,Fig.1,hasbeeninvestigatedrepeat- edlybymeansoftheevaluationofdatasetsforappropriateblankexperiments(usingin themainPt-cathodespolarizedinD O–basedelectrolytes).Theobjectiveherehasbeen 2 the definitionofthe appropriateinstrumentfunction, whichcanbeaccurately defined byEq. (1). Thenextstepin thisinitialphaseoftheworkhas beentodefineasetofheattransfer coefficients that characterize the behavior of the calorimeters and to investigate their precision and accuracy leading up to their use in evaluating the raw data sets of the experimentalmeasurementcycles. Therawdatausedintheseinvestigationshavebeen boththosefortheappropriateblankexperimentsandthosegeneratedbysimulationsof thecellbehavior.Anillustrationofthisphaseoftheinvestigationhasbeengiveninvol. 25Vol.IIcontainsextensivediscussionsoftheerrorsofthevariousheattransfercoefficientsandthecause oftheseerrors. 26Thiswillexplainbothourstrategiesfordeterminingtheheattransfercoefficients(whichgivethevalues att 6 t2aswellasgivingafurtherreasonforchoosing2-daymeasurementcycleswitht2correspondingto theendofday1.) 47 II(seealsoe.g.,[3,4,6]). Theoutcomeoftheseinvestigationshasbeenthedemonstrationthatitisusefultode- termine first of all the time dependence of the differential lower bound heat transfer coefficient,(k # R (cid:9) ,aswellasofthederivedmeans, 11 (cid:8) k # R (cid:9) and 11 (cid:8) k # R (cid:9) . However,these 11 coefficientshavealimitedprecisionbecausetheirevaluationrequiresthedifferentiation of the inherentlynoisy experimentaldata. Preciseand accurate evaluationsare there- forebestbasedontheintegrallowerboundheattransfercoefficient,(k # R (cid:9) ,andthein- 21 tegraltrueheattransfercoefficient,(k # R (cid:9) ,aswellasonthevalues(k 22 # R (cid:9) and(k 251 # R (cid:9) 252 derived in the extrapolation procedures. These extrapolation procedures lead both to the elimination of the effects of the water equivalent, C M, on their values as well p as to reasonably accurate determinations ofC M. The differences (k p # R (cid:9) – (k 22 # R (cid:9) or 11 (k # R (cid:9) –(k 252 # R (cid:9) betweenthe“true”andlowerboundheattransfercoefficientscanthen 251 beusedtodefinethedifferentialtrueheattransfercoefficient,(k # R (cid:9) . 12 Ithasbeenfoundthattheprecisionandaccuracyoftheintegralheattransfercoefficients issohigh,thatitispossibletoinvestigatetheirsystematicvariationswithtime(typically thesystematicvariationsofjust0.4%oftheirnumericalvalues).Furthermore,itispos- sibletoreducesuchdatatoasingle,timeindependentheattransfercoefficient,e.g.,of (k (cid:16) # 0 R (cid:9) with relative errors below 21 0.01 %. This result is hardly surprising. The physics of the calorimeters are quite simple (they are ideal well-stirred tanks) and the errors are mainly those set by the temperaturemeasurements. Itisalsorelativelystraightforwardtospecifythechanges which would need to be made to reduce the errors further – say to 0.001% – if that shouldeverprovetobenecessaryordesirable. Althoughtheprecisionandaccuracyoftheheattransfercoefficientsbasedonthefor- wardintegrationprocedures,(k # R (cid:9) and(k 31 # R (cid:9) ,wasknowntobelowerthanthosebased 32 onthebackwardintegrations,(k # R (cid:9) and(k 21 # R (cid:9) ,theICARUS-1methodologywasnev- 22 erthelessbasedonsuchforwardintegrations[2,2A]becausesuchforwardintegration was easier to implement and to combine with the evaluations of the data sets. It was anticipatedthattheextensionofthemeasurementcyclesfrom1to2daysand,inpar- ticular of the calibration periods from 6 to 12 hours, would allow the determination of(k # R (cid:9) and(k 31 # R (cid:9) withtherequiredandspecifiedprecisionsandaccuracies[2,2A]. 32 These changes in the measurement cycles were also expected to facilitate other parts oftheinvestigationsuchasthedeterminationofthetrueheattransfercoefficient(k # R (cid:9) . 2 Theproductionofplotsoftherawdataandtheinspectionoftheseplotsleadingtothe graphicalevaluationof(k # R (cid:9) and(k 1 # R (cid:9) weretobethefirststepinthedataprocessing. 2 Unfortunately,theprotocolslaiddownintheHandbookfortheICARUS-1system[2, 2A] were not followedin the experiments carried out by the Group at NHE. Further- more, following the receipt of the first set of data for experiments carried out in the SapporoLaboratories, itbecame clearthat thereweretiming errorsin the ICARUS-1 system. These timing errors did not affect the determination of (k # R (cid:9) and (k 21 # R (cid:9) . It 22 wasthereforerecommendedthattheprotocolsetdownintheHandbook[2]bestrictly adheredto,thatthepreliminaryevaluationsshouldbebasedon(k # R (cid:9) ,(k 1 # R (cid:9) ,and(k 2 # R (cid:9) , 11 andthatthefinalevaluationshouldbebasedon(k # R (cid:9) ,(k 21 # R (cid:9) ,(k 22 # R (cid:9) ,and(k 251 # R (cid:9) . It 262 48 isevidentthattheseinstructionswereignored. The development of the various aspects of the data analysis described in vol. I/7.0 isillustratedinpartbytheanalysisofExperimentMc–21describedinvol. II/3.0. In- evitably,thisillustrationisincompletebecauseoftheveryearlydevelopmentofpositive feedbackinthisexperiment. 49 50 PARTII:APPLICATIONOFDIAGNOSTICCRITERIA. In part II, we illustrate how the use of a faulty methodology (i.e., the non–standard ICARUS methodology), used by the New Hydrogen Energy (NHE) group, led to an incorrect evaluation of data. An experiment, designated here as Mc–21, provides the required data for the correct/incorrect application of diagnostic criteria developed in partI. II/1.0Preliminarydescriptionsandevaluations. II/1.1Experimentalset-up. ThecellusedintheexperimentwasoftheICARUS-1typewiththe99.5%Pd+0.5%B electrode in the form of a rod (4.7 x 20.1 mm), Fig. 1. The electrolyte was 0.1 M LiOD/D O.Thecellwasinsertedintowaterthermostatswhosetemperaturewasinde- 2 pendently controlled by Techne TE–8A stirrer/heater/regulatorunits. The water ther- mostatswereinturnmaintainedinaroomwhosetemperaturewascontrolledtowithin (cid:20) 20ofthatofthethermostats.27 TheexperimentwascarriedoutusinganICARUS-2typeelectrochemicalpolarization, control, anddata acquisition system. The electrochemicalsystem consisted ofan Hi- TekDT2101potentiostatwiredupasagalvanostat.Thesepotentiostats/galvanostatsare capableofdeliveringcurrentsof (cid:20) 1Aatoutputvoltagesuptoca (cid:20) 100V.Aseparatepo- tentiostat/galvanostatwasusedtodeliverconstantcurrentstotheresistiveheaterusedto calibratethecell. Thesystemwascontrolledbya486dataacquisitioncomputerwhich also controlled an Hewlett Packard 44705A multiplexer and data acquisition system. ThisdataacquisitionsystemwasonanIEEE-GPIBbussothatitwouldbeanticipated thattherewouldnothavebeenanytimingerrorsintroducedintothemeasurements. II/1.2Experimentalprotocols. 27Therearemisleadingstatementsaboutthisaspectoftheexperimentdesign. Thisdesignfollows the commonstrategyofusingtwothermalimpedancesinseries,astrategywhichisrequiredforexperiments aimingathighaccuracy. 51 TheprotocolusedfortheexperimentMc–21wasasfollows: (i)theelectrodewasfirstofallpolarizedfortwodayswithoutanyapplicationofcali- brationpulses; (ii)onthethirdday(andonallsubsequentdaysincludingdays68and69whenthecell hadreacheddryness)calibrationpulseswereapplied; (iii)changesofcurrentdensityweremadefrequently.Thesechangesofcurrentdensity areshowninFig. 4; (iv)thecellwastoppedupwithD Owheneverthiswasjudgedtobenecessaryatthe 2 starttimeofallthemeasurementcycles;thecellwasthenlefttoequilibratefor9hours followedbytheapplicationofcalibrationpulses of6hourduration; thecellwasthen againlefttoequilibrateforafurther9hoursbeforereachingthenextdayoftheexperi- mentalsequence; (v)asisevidentfrom(iv),thedurationofthemeasurementcycleswas24hours; Fig. 4. Celltemperature(A)andcurrentdensityprofiles(B).Thedottedlines delineatetheregionsfortheexpectedonsetofpositivefeedback(A)andexcess enthalpygeneration(B). ThisprotocoldifferssubstantiallyfromthatspecifiedfortheoperationoftheICARUS- 1and-2systems,whichwasasfollows[2,2A]: (ia) the electrodeswere to be polarized for 4 days (i.e., two measurementcycles, see (va)below)withoutanyapplicationofcalibrationpulses; (iia)onthe5thday(i.e.,forthethirdmeasurementcycle)andfor9furthermeasurement cycles, calibration pulses were to be applied as specified in (iva) below; this was to be followedbytwo furthermeasurementcycleswithoutthe applicationof calibration pulses and, in turn, by 10 further cycles with calibration pulses. A total experiment 52 durationof48dayswasthereforespecifiedfortheinitialphaseofthework. (iiia) the initial experiments were to be carried out at a single low current density, typically (cid:23) 250mAcm (cid:4) 2;inlaterexperimentsasingle,low,currentdensitywastobe appliedforvariousinitialdurationsfollowedbyaraisingofthecurrentdensitytovalues typically (cid:5) 1 Acm (cid:4) 2; this protocol was in broad accord with that used in previous investigations[7, 8]; changes of current density were to be made at the beginning of measurementcycles. (iva)cellsweretobetoppedupatthestartofeachmeasurementcycle;thecellswere then to be left to equilibrate for 12 hours and calibration pulses of 12 hour duration werethentobeapplied;thecellswerethenagaintobelefttoequilibrateforafurther 24hourssoastoreachthestartofthenextmeasurementcycle. (va)asisevidentfrom(iva),thedurationofthemeasurementcycleswastobe48hours. We now consider further the major differences between the operation of experiment Mc–21andtheconditionsusedinpreviouslyreportedinvestigationse.g,[1,7,8].Apart from the frequent changes of current density, Fig. 4, we can see that these current densities were mostly in the vicinity of the threshold value required for the onset of thephenomenonofexcessenthalpygeneration[1]. Furthermore,thecelltemperatures weremostlybelowthelevelrequiredfortheonsetofpositivefeedback,Fig. 4[9,10, 11],andwhichleadstoamarkedincreaseintheratesofexcessenthalpygeneration.28 The conditions in the cell therefore remained in the vicinity of the region of onset of positive feedback and, under these conditions, we would not expect to see a marked buildupintherateofexcessenthalpygeneration. Consideration of Fig. 4 also allows us to decide on the measurement cycles likely to provide examples of “Heat after Death” (objective (v) of this investigation). As waspointedoutintheoriginalinvestigation[3, 4]itwould beexpectedthat thisphe- nomenonwouldbeobservableunderseveraldistinctconditionswhichinclude (i) Cell full: cell operated at intermediate temperatures; cell current then reduced in stages (ii)Cellempty: cellallowedtoboildry;cellthenmaintainedattherailvoltageofthe galvanostat (iii)Cellempty: cellallowedtoboildry;celldisconnectedfromthegalvanostat. Consideration of the hard copy of the data sets shows that condition (ii) applies to part of day 68 of the sequence measurement cycles (see II/7.0) while condition (iii) applies topart ofday 69ofthis sequence(see II/8.0). Considerationof Fig. 4 shows that condition (i) is likely to apply to several of the measurement cycles. The effects wouldbeexpectedtobemostmarkedforpartsofdays25and26(reductionofthecell currentfromabovetobelowthethresholdforexcessenthalpygeneration;reductionin cell temperaturefrom abovethe levelfor the onset of positivefeedbackto belowthis level). Attentionisconfinedinthisreporttothisparticularday(seeII/9.0)althoughit isevidentthatthereareseveralfurtherregionsoftimewhichmightwellgiveexamples 28Ithasbeenarguedthatthisphenomenonislinkedtotheneedtoachievehighlevelsofloadingofthecath- odebyD (cid:1) ,whichisprobablyassociatedwithachangefromexo–toendothermicabsorption.Analternative explanationisthatthesephenomenaarelinkedtotheformationofanewphase,theg –phase. 53 of“Heat-after-Death”followingscenario(i). II/1.3Furtherdifferences. In this section we should also consider a further differencebetween the protocols for experimentMc–21andthoseusedinearlierstudies,namely,theschedulesofaddition ofD Otomakeupthelossesduetoelectrolysis.Thevolumeoftheelectrolyteinacell 2 inanhypotheticalexperimentcarriedoutfirstatacellcurrentof200mAfor29days followed by a cell current of 500 mA and with a daily schedule of additions falls by some1.21cm3 betweenthetwotimeregions. Wecanestimatethatthiswouldcausea decreaseofthemeanvalueof(k # R (cid:9) byca0.15%orof(k 12 # R (cid:9) byca0.075%.Suchsmall 22 changes are close to the error limits quoted for the instrumentation and can normally beneglected. However,themagnitudeofthechangesareabovetheerrorlimitswhich canactuallybeachieved(e.g.,seevol. IIandI/3.0)andshouldbetakenintoaccountin evaluationscarriedoutatthemaximumachievableprecisionandaccuracy. Figure5showstheeffectsofthescheduleofadditionsasactuallyusedinexperiment Mc–21. TheexactvaluesofD Oaddedwererecordedthroughoutthisexperiment. It 2 canbeseenthattheexpectedchangesin(k # R (cid:9) liebetween–0.15%and+0.3%,changes 12 whichshouldcertainlyagainbetakenintoaccount. Fig. 5. Thechangesinthevolumeofelectrolyte. Thepolarizationswerecarriedoutat thecurrentdensitiesshowninFig. 4. The schedule of additions leads to an important conclusion. We find that by day 67, the total volume of D O added was 262.5 ml whereas the total volume electrolyzed 2 54 was 253.3 ml. It is evident that the volume of D O is 3.6% larger then the volume 2 electrolyzed;therefore,therecouldnothavebeenanyrecombinationofthedeuterium andoxygenproducedbyelectrolysis. Thisisinagreementwithearliermeasurements [1]andnumerousmeasurementsbyotherauthors. The horizontal lines in Fig. 5 delineate the volumes of D O belowand above which 2 wewouldexpecttheelectrolyteleveltofallbelowthebaseofthesilveringintheupper partofthecell,Fig. 1,ortoapproachthebaseoftheKelFplugatthetopofthecell. It canbeseenthattheelectrolytelevelremainedwithinthespacedefinedbythissilvered portion throughout the measurement cycles. However, we can see that at long times the electrolyte level must have approached the base of the Kel F plug at the start of severalofthemeasurementcyclesfollowingthetoppingupofthecells. Intheworkat IMRA-Europe,itwasestablishedthatsuchoverfillingofthecellsleadstoananomalous increaseofthepseudo-radiativeheattransfercoefficientby4to5%ofthevalueswhich applyatthemeanlevel. Thisincreasein(k # R (cid:9) isalmostcertainlyduetoanincreasein theconductivecontributionthroughtheKelFplugtotheoverallheattransferfromthe cell29(cf. Eq. 4). II/1.4Temperature/potential–timeprofiles. We also make a number of preliminary assessments of the form of the temperature- time and cell potential-time series for day 3, i.e., the third measurement cycle of the experimentMc–21,Figs. 6–8. Thedataforthisdayareofspecialimportancebecause thegroupatNHEhavequoteda valueofthetrueheattransfercoefficientasgivenby theirmethodofevaluationforthisday. Thisvalueofthetrueheattransfercoefficient wasthenusedintheevaluationofallthemeasurementcycles.30 We can seethat we can immediately drawa number of importantconclusions. Thus, Fig. 6givesaplotofthetemperatureofthewaterbathversustimeforthefirst32,400s ofthemeasurementcycle(theperiod0 (cid:23) t (cid:23) t precedingtheapplicationoftheheater 1 calibration pulse) while Fig. 7 gives plots of the cell temperature versus time for the sameperiodandforbothpositionsinthecellwherethetemperaturewasmeasured,see Fig. 1. Itisevidentthatthenoiselevelofthemeasurementsinthewaterbath(s =0.0088K, mean=295.198K)ismuchhigherthanthatofthemeasurementsofthecelltempera- ture,Fig. 7. Thisdifferenceistobeexpectedbecausethewaterbathiscontrolledbya singlethermalimpedancewhereasthecelliscontrolledbytwoimpedancesinseries.At 29Thistypeofbehaviorappliestoday61,whichisameasurementcycleforwhichwecangetimportant confirmatoryevidenceofthetrueheattransfercoefficientwhichappliestotheoperationofthecell(seeFig. 20,sectionII/4.0). Figure20showstheexpectedincreasein(k - R / 11attimesclosetothetoppingupofthe cell. 30TheevaluationgivenbyNHEisconsideredfurtherinsectionII/2.0whilesectionII/3.0givestheappli- cationoftheICARUSmethodologytothisparticulardataset. 55 Fig. 6. Thetemperatureofthewaterbathforthefirst32400softhefirstmeasurement cycle(i.e.,theperiod0 (cid:23) t (cid:23) t ). Meantemperature:295.198K,s 1 (cid:2) 0.0088K. Fig. 7. Thecelltemperatureatthetwomeasurementlocationsforthefirst32400sof thefirstmeasurementcycle(0 (cid:23) t (cid:23) t ). Meantemperaturedifference=0.0045K;s = 1 0.0027K. 56 Fig. 8. Therawdata–celltemperatureandpotentialasafunctionoftime–forthe thirdmeasurementcycle. 57 thesametime,thenoiseinthetemperatureofthewaterbathismuchhigherthanthatin our original measurements (s = 0.003 K)[6] and, in our experience, such an increase is due to inadequate con- troloftheroomtemperature. It is evident that the noise in the measurements of the temperature of the water bath is one factor which limits the precision of the lower bound heat transfer coefficients, (k # R (cid:9) ,viaitseffectonthetemperaturefunction, f 11 1 (cid:8) q (cid:9) .31 Itcanbeseenthatthevariationwithtimeofthecelltemperaturemeasuredatthetwo positionsinthecellissystematic,Fig. 7. Moreover,itisclearthatthereisasystematic difference in temperature between the two positions which must be due to either one ortwo errors in the calibration.32 Forthese measurements, we obtainmean [q 1 (cid:21) q ] 2 = 0.0045 K and s (cid:22) q 1 (cid:21) q ] = 0.0027 K (subscripts 1 and 2 denote the short and long 2 thermistors). The mean gives an indication of the accuracy in (k # R (cid:9) which we can 11 expecttoachieve.Theerrorca0.05%issomewhatabovethetargetfortheprecisionof themeasurements,errors (cid:23) 0.01%,whichishardlysurprising. Thestandarddeviation gives double the value of the expected standard deviation for the measurements with onethermistor. We canseethatthisvalue,ca0.00135K,willnotaffecttheaccuracy ofthe determination ofanyversionof thetrue heattransfer coefficient. However,we shouldnotethatitisevidentlydesirabletocalibratethethermistorssothatwecanmake thetemperaturemeasurementstowithin (cid:20) 0.001K. Differencesintemperaturebetweenthosegivenbytheshortthermistorandlongther- mistorwillbeconsideredfurtherinsectionII/6.0dealingwithday68asthecellisbeing driventodrynessandinsectionII/8.0dealingwith“Heat-after-Death”onday69. Fi- nallyweconsidertheplotsoftherawdataforday3,Fig.8.Wecanseeimmediatelythe inadequacyof restrictingthecalibration pulseto 6 hours becausethe temperaturehas notrelaxedtoequilibriuminthistimeperiod.33 However,inthisparticularcase,there isanevidentcomplicationbecauseoftheveryearlyestablishmentofpositivefeedback. Thiseffectcanbeseenmostdirectlyfromthedelayedrelaxationofthetemperatureto thebaselinefollowingthecessationoftheheatercalibrationpulse(thebaselineisgiven bytheextrapolationoftheq (cid:21) tseriesobservedbeforetheapplicationofthecalibration pulse). Evidently,theraisingofthecelltemperaturebythecalibrationpulsehasledto anincreaseinthethermaloutputfromthecellwhichpersistsfollowingthetermination 31Thevalues =0.0088KisoutsidetherangespecifiedfortheICARUS–1systemifmeasurementsare madeatlowcelltemperatures.Bycontrast,thetrueheattransfercoefficientsarenotaffectedbysuchfluctu- ationsbecausethetemperaturefunction f2 9 q / isdeterminedbythecelltemperaturealone. 32Differencesintemperatureduetoinadequatemixinghavefrequentlybeeninvokedinargumentsabout theperformanceofICARUScalorimeters. However,inadequatemixingwouldnotgiverisetoasystematic andtimeinvariantdifferenceintemperaturebetweenthetwopositions. Moreover,suchdifferencesintem- peraturewouldnotbeexpectedbecausethethermalrelaxationtime,t 6 CpM : 4 9 k - R / q 3,isoftheorder5000 swhereastheradialandaxialmixingtimesareca3andca20asrevealedbytracerexperiments.Smalldif- ferencesinthecelltemperaturecanonlybeobservedinthevicinityofelectrodesandcalibrationresistor,i.e., withinthePrandtlboundarylayers.However,theirvolumesarenegligiblysmallcomparedtotheelectrolyte volume. 33Withathermalrelaxationtimeof5000s,thetemperatureperturbationwillonlyhavereached98.67%of itsfinalvaluewithinthecalibrationperiod. 58 ofthecalibration,i.e.,aformofpositivefeedback.34 Thecalibrationofsuchasystem canobviouslyonlybeachievedwithmanyrestrictionsandwithgreatdifficulty.35 TheinterpretationofFig. 8willbeconsideredfurtherinsectionsII/2.0andII/3.0. II/2.0TheNHEInterpretationofexperimentMc–21. As has already been pointed out, the NHE interpretation of experiment Mc–21 rests on the determination of the true heat transfer coefficienton day 3 of the measurementcycles. Apartfromthecitationofthevalueofthiscoefficient(0.793504 G 10 (cid:4) 9 WK (cid:4) 4) in the header for the spreadsheet for day 1, the information givenby N.H.E. is contained in a set of spreadsheets which appear to be related to the (k # R (cid:9) 11 spreadsheetsoftheICARUSmethodologyforanalyzingthedata.Wehavetotakenote ofthefollowingobservations: (a)itisnotclearhowthevalueofthetrueheattransfercoefficientwasdeterminednor whichofthedefinitionsoftheheattransfercoefficientsmayhavebeenused. However, itislikelythatthiswasthecoefficient(k # R (cid:9) anditwillbeassumedherethatthiswas 32 the case, i.e., we will assume that the values of the excess enthalpies were based on calculationsusingthesinglevalue(k # R (cid:9) =0.793504 32 G 10 (cid:4) 9WK (cid:4) 4. (b) it is also not clear to what extent the values of the true heat transfer coefficient and of the excess enthalpies may have been affected by the value C M = 490 JK p (cid:4) 1 used in the calculations. Values as high as this applied to cells used prior to 1992 and the Handbooks for the ICARUS systems contained instructions for changing this (andother)parameter(s),dependingonthevaluefoundusingthemethodsofevaluation outlinedintheHandbook[2,2A].Itshouldbenotedthatthe“guesstimate”ofthewater equivalentofthecellis: C M isapproximatelythesumofthecontributionofD Oin p 2 the electrolyte and of the glass in the inner cell wall = (422 + 31) JK (cid:4) 1 = 453 JK (cid:4) 1. Theremainingcomponentsofthecell(LiOD,metals,glassframing,heater,thermistor, a portion of the Kel–F plug) will contribute only a small additional term toC M. It p follows, therefore, that observations of C M far above or below 453 JK p (cid:4) 1 indicate malfunctionsofthemethodsofdataevaluation. (c) as has been noted elsewhere (see section I/4.1), the values of the rates of evapo- rative cooling cannot be calculated using the instructions given in the handbooks for the ICARUS-1and -2systems[2, 2A].Thedifferencesare not importantatlowtem- peratures(suchasthosewhichapplytoday3ofthemeasurementcycles)butbecome significant at temperatures close to the boiling point. However,at such elevated tem- peraturesother factorsneglectedin thecalculationscarriedoutbyNHEbecome even moreimportant. (d) it is apparent that the enthalpy inputs given in the NHE spreadsheets have been 34SucheffectscanbeseenintherawdataofsomeoftheexperimentscarriedoutbythegroupatHarwell. 35Thisisafeaturewhichwillbecommontoallcalorimetricsystemsusedtoinvestigateathermalsource subjecttopositivefeedback. Itislikelythattheneglectofthisfactisresponsibleformuchoftheconfusion intheresearchoncoldfusion. 59 calculatedusing1.54Vasthethermoneutralpotentialwhereasmostotherauthorshave used the value 1.527 V. The circumstances leading to our choice of the value 1.54 V havebeendescribedelsewhere.36 (e) the most serious shortcoming of the NHE calculations is that the input due to the calibrationheaterhasbeenenteredaszeroratherthantheactualvaluegivenseparately as0.25000W.IntheprocedureusedbyNHE[5]thelowerboundheattransfercoeffi- cient,(k # R (cid:9) ,iscalculatedwiththisassumedzeroenthalpyinputanditisthenassumed 11 that the magnitude of the enthalpy input can be recovered together with any rate of excessenthalpygenerationbyusingthisderivedlowerboundheattransfercoefficient together with the true heat transfer coefficient, (k # R (cid:9) , and f 32 1 (cid:8) q (cid:9) . Let us assume first of all that such a procedure is correct. Then we can see an immediate disadvantage ascomparedtothemethodoutlinedfortheICARUSsystemsinthatweareunableto determinewhether (k # R (cid:9) duringthe period ofthe application ofthe calibration pulse 11 in t 1 (cid:23) t (cid:23) t is the same as for t 2 (cid:23) t , or t 1 (cid:5) t .37The data derived, e.g., see Fig. 9, 2 are certainly further degraded by using incorrect values ofC M. However, in actual p fact, the procedureused by NHE is invalidas has been pointed out in a report andin subsequent correspondence.38 It is difficult to see why the straightforwardprocedure outlinedintheHandbooksfortheICARUS-1system[2,2A]wasnotfollowed.39 We can only concludefrom these data that the evaluationsareincorrect based onthe fol- lowingevidence: (f) it is impossible for the true heat transfer coefficient, (k # R (cid:9) , to be smaller than the 32 lower bound heat transfer coefficient, (k # R (cid:9) , because the lower bound value is based 11 on the assumption that there is a zero rate of excess enthalpy generation in the cell. ThetypeofdifferenceseeninFig. 9couldonlyariseifthecellwereendothermicand the endothermicity has already been fully taken into account using the thermoneutral potential. Anyadditional endothermicitythereforerequires that the celloperatesas a spontaneousrefrigeratorandthisviolatesthesecondlawofthermodynamics. (g)the pronouncedvariationofthelowerboundheattransfercoefficient,(k # R (cid:9) , with 11 timefollowingtheapplicationoftheheatercalibrationpulseatt (cid:2) t anditscessation 1 36Thewaterthermostatssurroundingthecellswererunat300Cinourearlywork.In1988,weattempted toallowforthisshiftinthereferencetemperatureaswellasthefactthatelectrolysistakesplacefrom0.1M LiODinD2OandnotD2Oitself. Whilethethermoneutralpotentialiscertainlynot1.527V,itiscloserto thisvaluethanto1.54V. 37Moreexactly,whetherthevalueof(k - R / 11plottedversustimefallonacommonstaightlineasshownin e.g.[6]. 38ThemethodproposedbyNHEcanonlygivethecorrectresultprovidedthereisazerorateofexcess enthalpygenerationfortheperiodt , t1 beforetheapplicationofthepulse(aswellasfort H t2following theterminationofthepulse)whilethecalibrationpulseitself(duringt1 , t , t2)leadstothegenerationof excessenthalpy. 39InthemethodoriginallyproposedbyNHEthelowerboundheattransfercoefficientdeterminedbefore theapplicationofthecalibrationpulsewasusedinattemptstoderivetherateofexcessenthalpygeneration duringtheapplicationofthispulse. Itisnotsurprisingthatsuchamethodcanonlygivethecorrectresult providedthereisazerorateofexcessenthalpygenerationfortheperiodt , t1beforetheapplicationofthe pulse(aswellasfort H t2followingtheterminationofthepulse)whilethecalibrationpulseitself(during t1 , t , t2)leadstothegenerationofexcessenthalpy. 60 Fig. 9. Thelowerheattransfercoefficient,(k # R (cid:9) ,asafunctionoftimeforthethird 11 measurementcycleasdeterminedbytheanalysisprovidedbytheNHElaboratories. Theverticallinesdelineatetheperiodofapplicationofthecalibrationpulse, t 1 (cid:23) t (cid:23) t . Theamplitudeofthecalibrationpulse,D Q=0.2500W,hasbeenexcluded 2 inthecalculationof(k # R (cid:9) duringtheperiodt 11 1 (cid:23) t (cid:23) t andithasbeenassumedthat 2 C M=490WK p (cid:4) 4. att (cid:2) t impliesattheveryleastthattherawdatahavebeenevaluatedusinganincorrect 2 valueofthewaterequivalent,C Mofthecell. p (h) the excess enthalpy given by the NHE evaluation is apparently negative both for t (cid:23) t andt 1 (cid:5) t whichisfurtherillustrationoftheapparentviolationofthesecondlaw 2 ofthermodynamics. (i) it has been maintained [5] that the NHE evaluation recovers the magnitude of the heatercalibrationpulse,D Q,duringitsperiodofapplication,t 1 (cid:23) t (cid:23) t ,togetherwith 2 anyrateofexcessenthalpygeneration.40 Figure 10 shows that this is incorrect: The values of the rates of enthalpy generation (whichhereincludetheenthalpyinputtothecalibrationheater)arelessthanD Qinthe periodt 1 (cid:23) t (cid:23) t if we take Q = 0 as the base line. If we fix the baseline at the 2 excess levelofnegativerateofexcessenthalpygenerationfort (cid:23) t ,thenQ 1 excess (cid:5) D Qduring the periodofthe calibrationpulse,t 1 (cid:23) t (cid:23) t . We concludethatthe evaluationgiven 2 byNHEisinvalidandthatitislikelythatthisevaluationissubjecttoseveraldistinct errors. 40Itsohappensthatthereissomevaliditytothisconclusionduetotheinfluenceofpositivefeedback. 61 II/3.0TheICARUStypeinterpretationofexperimentMc–21. As a first step, we correct the (k # R (cid:9) -spreadsheet by including the magnitude of the 11 calibrationpulse,D Q,inthedefinitionofthelowerboundheattransfercoefficient.The values of 109 (cid:8) k # R (cid:9) in the region t 11 1 (cid:23) t (cid:23) t can now be shown together with those 2 for t (cid:23) t andt 1 (cid:5) t on a graph using a single scale for the ordinate, Fig. 11. While 2 we cannot be certain whether or not an incorrect choice ofC M can explain the fall p of(k # R (cid:9) intheregiont 11 (cid:5) t (butclosetothistime)ortherisefort 1 (cid:5) t (butcloseto 2 thistime), itis clearthat (k # R (cid:9) dropsmarkedlyin the regiont 11 1 (cid:23) t (cid:23) t comparedto 2 the values fort (cid:23) t andt 1 (cid:5) t . Such a drop in (k 2 # R (cid:9) can only be due to the neglect 11 ofthebuildupoftherateofexcessenthalpygeneration duringt 1 (cid:23) t (cid:23) t . Itfollows 2 thattheincreaseintemperatureduetothecalibrationpulseincreasestherateofexcess enthalpy generation. In fact, experiment Mc–21 shows a very early establishment of positive feedback as is indeed evident from the plot of the raw data, Fig. 8. It is very important that the presence of positive feedback can be established by a simple examination of a (k # R (cid:9) –spreadsheet constructed according to the instructions in the 11 ICARUS-1Handbooks[2,2A]. ItshouldbenotedthattheamplitudeofthecalibrationpulsewouldhavehadtobeD Q =0.2763Winordertobringthevaluesof(k # R (cid:9) intheregiont 11 1 (cid:23) t (cid:23) t tothelevelof 2 theregressionlinewhichappliestothedatafort (cid:23) t andt 1 (cid:5) t . SuchachangeinD Q 2 isbeyondallpossibilities. Thenextstepistoprepareamodified(k # R (cid:9) –spreadsheetwherewecorrecttheenthalpy 11 inputs(see(c),(d)and(e)insectionII/2.0)andpresentthedatainaformsuitablefor theapplicationofEq. (10). Inviewoftheearlyinterventionofpositivefeedback,we wouldonlyexpecttobeabletoapplyEq.(10)attimesclosetot whereweseethatthe 1 trueheattransfercoefficient,(k # R (cid:9) mustbeatleast0.83808 12 G 10 (cid:4) 9 WK (cid:4) 4 whilethe waterequivalent,C Misoftheorderof454JK p (cid:4) 1(inagreementwiththe“guesstimate,” seesectionII/2.0andtherighthandpartofFig. 12). Theinfluenceofpositivefeedbackonthefailureofsimplemethodsfortheevaluation ofthelowerboundandtrueheattransfercoefficientsaswellasofthewaterequivalent ofthecellisalsoshownclearlybyattemptstoderive(k # R (cid:9) (whichrelyonthecom- 181 binationofdataforthetimeregionst 1 (cid:23) t (cid:23) t andt 2 2 (cid:23) T;seepartI).Thisevaluation has been found to be especially useful in the analyses of data sets for blank experi- ments(e.g. seevol. II).Figure11illustratesthatweareunabletoobtainasatisfactory interpretationofsuchdataforexperimentMc–21. Figure13givestheplotofthedataversustimeandalsoshowsthevariationof109 (cid:8) k # R (cid:9) 11 withtimepredictedusingthevaluesfort (cid:23) t andtheknownbehaviorestablishedwith 1 appropriateblankexperiments,e.g.,[6]. AsinthecaseofthedatainFig. 11,wecan see that the temperature rise induced by the calibration pulse leads to a decrease in (k # R (cid:9) whilethecoolingconsequentontheterminationofthepulseleadstoanincrease 11 in (k # R (cid:9) . These changes can only be due respectivelyto an increase and decreasein 11 therateofexcessenthalpygeneration,whichcannotbetakenintoaccountinderiving thevaluesof(k # R (cid:9) ,i.e.,theeffectsofpositivefeedback. 11 62 Fig. 10. Therateofexcessenthalpygeneration,Q W,asafunctionoftimeforthe f thirdmeasurementcycleasdeterminedbythegroupatNHElaboratories. 63 Fig.11. Thelowerboundheattransfercoefficient,(k # R (cid:9) ,asafunctiontimeforthe 11 thirdmeasurementcycleasdeterminedbytheICARUSsystemsprocedurewiththe inclusionofthecalibrationpulse,D Q=0.2500W,inthecalculationof(k # R (cid:9) during 11 theperiodt 1 (cid:23) t (cid:23) t . IthasbeenassumedthatC M=490JK 2 p (cid:4) 1. Figure 13shows that we still observediscontinuities in the lower bound heat transfer coefficient,(k # R (cid:9) att andt . However,itisevidentthattherecanbenomechanism, 11 1 2 which could account for such changes which must thereforebe due to an error in the analysis. The most obvious error is the use of an incorrect value of C M (see (b), p II/ 2.0). The analysis of the time dependence according to Eq. (10) in the region t (cid:5) t (butadjacenttot )indicatesthatthecorrectvalueis450JK 1 1 (cid:4) 1. Theheattransfer coefficients109 (cid:8) k # R (cid:9) and109 11 (cid:8) k # R (cid:9) arebasedonthisvalueofC M. Figure13shows 11 p a plot of (k # R (cid:9) values versus time and we can see that the discontinuities in the heat 11 transfercoefficientatt (cid:5) t andt 1 (cid:5) t (butadjacenttothesetimes)arenoweliminated. 2 However,asexpected,theeffectsduetopositivefeedbackaremaintained. Figure 14 shows that there is indeed only a small rate of excess enthalpy generation fort (cid:23) t whiletheapplicationofthecalibrationpulseleadstoabuild-upofthisrate, 1 which again decreases for t (cid:5) t (there is a small long-term increase in the rate of 2 excessenthalpygenerationfort (cid:5) t ). Figure15showsasimilarcalculationbutusing 2 theNHEmethodology(notethedifferenceinscalesofthey-axesinFigs. 14and15). Weagainseeanearzerorateofexcessenthalpygenerationfort (cid:23) t ,whilefort 1 (cid:5) t 1 butadjacenttot ,wenowseethestepduetothecalibrationpulse,D Q=0.2500W.In 1 theregiont 1 (cid:23) t (cid:23) t ,wethenseethebuild-upintherateofexcessenthalpygeneration 2 due to positivefeedback. Att (cid:2) t but adjacenttot , we againsee a step in the total 2 2 observedrateofexcessenthalpygeneration. Asexpected,thisstepagaincorresponds 64 Fig.12. Evaluationof (cid:8) k # R (cid:9) andC MaccordingtoEq. (10). 1811 p Fig. 13.Thelowerboundheattransfercoefficient,(k # R (cid:9) ,asafunctionoftime. Third 11 measurementcycle;D Q=0.2500W;C M=450JK p (cid:4) 1. 65 Fig. 14.Therateofexcessenthalpygeneration,Q,asafunctionoftimeforthethird measurementcycle. (k # R (cid:9) =0.85065 12 G 10 (cid:4) 9WK (cid:4) 4 andthevaluesforthelower boundheattransfercoefficientshowninFig. 12. 66 to the expected value D Q = 0.2500 W; at longer times, we see the gradual decrease of the rate of excessenthalpygeneration due to the removalof the effects of positive feedback. Itcanbeseenthatacomparisonoftheplotsof(k # R (cid:9) and(k 21 # R (cid:9) versustime,Fig. 16, 31 withthecorrespondingplotsforblankexperiments,e.g.,see[6],showsveryclearlythe interventionofpositivefeedbackduetothesuperpositionofthecalibrationpulse.Ifwe focusattentionfirstofallonthebehaviorof(k # R (cid:9) fort 31 (cid:23) t ,thenweseetheexpected 1 smalldecreasewithincreasingtime.41 Fort (cid:5) t weseeamorerapiddecreasedueto 1 theonsetofpositivefeedback. Theeffectsofthispositivefeedbackdecreasefort (cid:5) t 2 sothatweobserveasmallincreaseof(k # R (cid:9) withincreasingtimeinthisregion. 31 The variation of (k # R (cid:9) with time can be interpreted in a similar way provided one 21 bears in mind that there is now no region in time in which the integrals used in the calculation of the heat transfer coefficient are independent of the effects of positive feedback. Theinfluenceofpositivefeedbackontheintegralsusedintheevaluationof (k # R (cid:9) explainswhywecannotobtainasatisfactoryevaluationofthetargetvalueofthe 21 lowerboundheattransfercoefficient,(k # R (cid:9) . Wewouldonlyexpecttobeabletoapply 251 theICARUSmethodologyinaregionoftimewheretheinfluenceofpositivefeedback can be expected to be adequately small, say, in the region 72,300 to 75,300 s of the measurementcycle. Theestimatesofthelowerboundheattransfercoefficient,(k # R (cid:9) 261 andofC M,are0.81821 p G 10 (cid:4) 9WK (cid:4) 4and475.3JK (cid:4) 1. Thecommentswhichhavebeenmadeabouttheevaluationoftheintegralheattransfer coefficients using the whole measurement cycles apply equally to the evaluations ac- cordingtotheinstructionsandsoftwareintheICARUSsystems[2,2A].Theprecision of (k # R (cid:9) and (k 31 # R (cid:9) is low because of the intervention of positive feedback and the 351 consequentneedtorestrictattentiontotheregiont (cid:5) t butclosetot . Thisisequally 1 1 trueoftheaccuracyof(k # R (cid:9) and(k 32 # R (cid:9) . 352 However,theevaluationsofthesecoefficientsisinstructivebecauseitisvirtuallycertain thatthevalueofthetrueheattransfercoefficientquotedbyNHEiseitherthevalueof (k # R (cid:9) ataparticulartimeorelse(k 32 # R (cid:9) evaluatedoveraparticularrangeoftime. We 352 therefore have to investigate whether we can modify the approach so as to allow the determination ofthis true heattransfer coefficient. We haveto note that it is unlikely that we would be able to find a generally valid procedure because it is in general not possibletocalibrateclosedloopsystemssubjecttopositivefeedback. However,forthe particularexampleofday3ofexperimentMc–21,wecanseethattheeffectsofpositive feedbackarerelativelysmalland,moreover,confinedinthetime-domain,Fig. 14. We canthereforeincludetheobservedvaluesoftheratesofexcessenthalpygenerationin theevaluationoftheintegraloftheenthalpyinputandusethismodifiedintegraltore- evaluate(k # R (cid:9) and(k 22 # R (cid:9) . Figure17illustratesthisevaluation. Itcanbeseenthatwe 252 41Thevaluesforthefirst20to30pointsmustbeexcludedasthebenefitsofusingtheintegralcoeffi- cientsareonlyestablishedwithincreasingtime. Similarly,thefirst20to30pointsmustbeexcludedifthe interpretationisbasedonbackwardintegration,i.e.,ifweconsider(k - R / 21. 67 Fig.15. Therateofexcessenthalpygenerationasafunctionoftime. Third measurementcycle;(k # R (cid:9) =0.85065 12 G 10 (cid:4) 9WK (cid:4) 4,D Q=0.2500W 68 Fig. 16. Thevariationof(k # R (cid:9) and(k 21 # R (cid:9) withtimeforthewholeofthethird 31 measurementcycle. doindeednowobtainasatisfactoryfittoEq. (22)whichexplainsthechoiceof(k # R (cid:9) 252 =0.85065 G 10 (cid:4) 9WK (cid:4) 4 andC M=450JK p (cid:4) 1 forthefurtherevaluationofthedata. In view of the fact that this evaluation of the true heat transfer coefficient, (k # R (cid:9) , 252 requires the development of a special approach, it is necessary (and advisable) to in- vestigatewhetherthevalueobtainedcanbeconfirmedbyothermeansusingdifferent partsoftheexperiment(i.e.,othermeasurementcycles).Suchconfirmationscanbeob- tainedusingthemeasurementsonday61andthefirst57hoursofdays1and2. These confirmationsareoutlinedinsectionsII/4.0andII/5.0respectively. II/4.0ApplicationoftheICARUSTypeInterpretationtothedataforday61. The earlyinterventionof positivefeedbackrequires us to modifythe ICARUS evalu- ationstrategiesinordertoachievethe calibrationofthesystem, i.e.,to determinethe valueofthetrueheattransfercoefficient. Itisthereforeimportanttofindconfirmatory evidencethatthisheattransfercoefficientisindeedca0.85065 G 10 (cid:4) 9WK (cid:4) 4 asgiven attheendoftheprevioussection. Evidencepertinenttothisconclusionispresentedin thepresentsectionaswellassectionII/5.0. Wenoteinthefirstplacethevaluesofthetotalexcessenthalpyforeachdayofopera- tioncalculatedusingthetrueheattransfercoefficient,(k # R (cid:9) =0.79350 32 G 10 (cid:4) 9WK (cid:4) 4 asgivenbytheNHEevaluationaswellasthosecalculatedwithtrueheattransferco- efficient, (k # R (cid:9) = 0.85065 x 10 252 (cid:4) 9WK (cid:4) 4, as determined in section II/3.0 using the modified ICARUSmethodology. These values areplotted in Figs. 18and 19respec- tively. We can see immediately that the evaluation given by NHE must be incorrect becauseweobtainnegativeexcessenthalpiesforsomeofthesedayswhichcontravenes 69 Fig. 17.Evaluationoftheintegralheattransfercoefficient109 (cid:8) k # R (cid:9) /WK 252 (cid:4) 4,and waterequivalent,C M/JK p (cid:4) 1 forthethirdmeasurementcyclewithcorrectionforthe effectsofpositivefeedback(seeAppendix). the second lawof thermodynamics(cf. section II/2.0). On the other hand, the evalu- ationbasedontheheattransfer coefficientgivenbythe modifiedICARUSevaluation schemeonlygivesaveryslightlynegativeexcessenthalpyforday61. Itisthereforereasonabletoassumethattherateofexcessenthalpygenerationonday 61isclosetozero. Theevaluationofthelowerboundheattransfercoefficient,(k # R (cid:9) , 11 mustthereforebeclosetothevaluesofthetrueheattransfercoefficient,(k # R (cid:9) . Figure 12 20 gives a plot of the relevant data compared to the plot which we predict using the value(k # R (cid:9) = 0.85065 12 G 10 (cid:4) 9 WK (cid:4) 4 and the variation of(k # R (cid:9) with time givenby 11 therelevantblankexperiments[6andvol. II].It canbeseenthattheobservedvalues of(k # R (cid:9) areincloseaccordwiththosewhichwewouldpredictontheassumptionthat 11 thereisonlyalowrateofexcessenthalpygenerationonthatday. Itcanbeseenthatthereisonlyoneregionoftimeinwhichthereisamarkeddeviation from the predicted behavior, namely, for 0 (cid:23) t (cid:23) 10 (cid:19) 000 s. In this region, (k # R (cid:9) is 11 markedly larger than the expected value and, moreover, decreases rapidly with time to these predicted values. It has already been noted in section II/1.3that the cell was overfilledwith D O at the start of this particular day (see Fig. 5) so that the level of 2 electrolyte would have been expected to approach the base of the Kel-F plug sealing the topof thecell. Separatemeasurements haveshownthat thepseudo-radiativeheat transfer coefficient increases by ca 5% over the expected value presumably because of an increase in the conductive contribution through the top of the cell. It is likely, therefore, thatthedeviationseeninthistimerangecanbeattributedtotheoverfilling ofthecell. 70 Fig. 18.Theexcessenthalpyasafunctionoftimeusingtrueheattransfercoefficient, (k # R (cid:9) =0.79350 12 G 10 (cid:4) 9WK (cid:4) 4 asgivenbytheanalysisofthegroupatNHE laboratories. Fig. 19. SameasinFig. 18evaluatedusing(k # R (cid:9) =0.85065 12 G 10 (cid:4) 9WK (cid:4) 4 asgiven bytheICARUSsystemanalysismodifiedtoaccountforpositivefeedback. 71 Fig. 20. Thevariationwithtimeofthe11-pointaverageofthelowerboundheat transfercoefficient,109 (cid:8) k # R (cid:9) forday61. Thefulllinegivesthevariationwithtime 11 fortherelevantblankexperiments[6andvol. II]. Fig. 21.Thelowerboundheattransfercoefficient,(k # R (cid:9) ,andtherateofexcess 11 enthalpygeneration,Q,forthefirst57hoursofoperation. 72 II/5.0Apre-ICARUSevaluationofthetrueheattransfercoefficient. It is possible to find a further value of the true heat transfer coefficient (k # R (cid:9) by ap- 12 plying a method used in 1992 [7, 8]. It was shown at that time that the lower bound heattransfercoefficient,(k # R (cid:9) ,decreasesmarkedlyfromtheexpectedvalueduringthe 11 initial stages of the measurement cycles. The full line in Fig. 21 shows the expected variationwithtimefor thepresentexperiment. ValuesofQarebasedontheassump- tion that (k # R (cid:9) is given by the regression line. The horizontal line shows the value 12 ofQbasedontheassumptionthatthecurrentefficiencyforthechargingtheelectrode at t = 130500s is 100% and that the heat of absorption of deuterium in the lattice is 40 kJ/mole. In this case, the decrease, att= 130500s, is due to the completion of the exothermicabsorptionofdeuteriuminthelattice. Itwouldbeexpected,therefore,that thelowerboundheattransfer coefficient,(k # R (cid:9) , wouldrisemarkedlytotheexpected 11 true value as this process is completed with the provisothat we can observe a period ofoperationduringwhichthereiszeroexcessenthalpygeneration. Itfollowsthatwe canderiveavalueofthetrueheattransfercoefficient,(k # R (cid:9) ,fromthemaximumofthe 12 lower bound heat transfer coefficient, (k # R (cid:9) , which is observedwith increasing time. 11 Figure21showstherelevantdataforthefirst57hoursofoperationofexperimentMc– 21(i.e.,uptothetimeofapplicationoftheheatercalibrationpulseonday3). Thefull line shows the expected variation of (k # R (cid:9) with time based on the value of (k 11 # R (cid:9) at 12 t (cid:2) t onday3, (i.e., (k 2 # R (cid:9) = 0.85065 12 G 10 (cid:4) 9 WK (cid:4) 4, the assumptionofzeroexcess enthalpygeneration,(i.e.,(k # R (cid:9) 11 (cid:2) (cid:8) k # R (cid:9) )andtheknownvariationof(k 12 # R (cid:9) withtime 11 establishedwithblankexperiments[6andvol.II].Itcanbeseenthat(k # R (cid:9) doesindeed 11 risetothepredictedlevelsasthechargingoftheelectrodeiscompleted. Figure21alsoshowsthederivedratesofexcessenthalpygenerationbasedontheex- perimentalvaluesof(k # R (cid:9) andtheassumptionthatthetrueheattransfercoefficientis 11 givenbytheregressionline.Itcanbeseenthattheexperimentalvaluesareinreasonable accord with the assumption that the charging of the cathode is ca 100% efficient and that the heat ofabsorptionis ca 40 kJMol (cid:4) 1. Figure 21furthermoreshowsthat there isasmallbuild-upofexcessenthalpygenerationonday3followingthecompletionof thechargingprocess(compare[7,8]). II/6.0Day68: theperiod0 (cid:23) t (cid:23) 21,300sduringwhichthecellisdriventodryness. We consider next the penultimate day of the investigation of experiment Mc–21; the cellisdriventodrynessduringthe firstpartofthismeasurementcycle. We candraw anumberofimportantconclusionsfromtherawdataalone. Wenoteinthefirstplace thatthetemperaturegivenbythelongthermistorisnowslightlyhigherthanthatgiven bytheshortthermistorwhereastheoppositeistrueformeasurementsmadeatlowtem- peratures. Atfirstsightsuchachangemightbeattributedtoagenuineeffect,namely, the increase in the enthalpy input in the bottom part of the cell (containing the Pd-B cathode). However, such an interpretation is unlikely because the temperature differ- encebetweenthetwothermistorsisessentiallyconstantfor,say,20,000seventhough the enthalpy input increases by a factor of three. It is more likely therefore that this particulartemperaturedifferenceisa furthermanifestationof errors inthe calibration 73 ofthethermistors. Thetemperaturedifferencesbetweenthetwothermistorsareappreciablylargerforthe last four data acquisition points, and this difference is especially marked for the last point,0.590K.Suchadifferenceistobeexpectedbecausethelongthermistorisnow intherelativelyconcentratedLiODsolutionwhiletheshortthermistorisinthevapor phase.However,wealsohavetonotethatthetemperatureatbothpositionsisabovethat oftheboilingpointofpureD O.Evidently,wehavetotakeintoaccounttheincrease 2 oftheboilingpointwiththeelectrolyteconcentrationastheD Oisprogressivelyevap- 2 orated(seesectionII/1.4). However,wealsohavetotakenoteofthefactthatthevapor phasecanbesuperheated(albeittoonlyalimitedextent).42 Ifwedonottakeaccount oftheincreaseoftheboilingpointwithconcentration,wearriveattheimpossibleresult ofnegativeenthalpiesofevaporationwithincreasingtemperatureasshownbytheNHE evaluation. We alsohavetousethecorrectatmosphericpressureinthecalculationof therateofevaporativecoolingandweneedtochangethethermoneutralpotentialand thewaterequivalentofthecellintheNHEevaluation. Asthewaterequivalentofthe cellonlyleadstoasignificanttermC M p (cid:8) dDq 0 dt (cid:9) intheinitialstagesforday68,ithas been assumed thatC M is unchanged throughout the stage leading to evaporation to p dryness(however,seefurthercommentsinsectionII/7.0). This calculation is similar to one which has been described previously (cf. vol. II) except that the published version included comments on the time dependence of the rateofexcessenthalpygeneration. Itisquiteobviousthattherate ofexcessenthalpy generationmustincreasewithtimebecausetheinitialrateonday68islessthan1W. It is importantthereforeto tryto establish the variation ofthe rate ofexcessenthalpy generation with time, if only to make a connection with the initial rate of“Heat after Death”observedafterthecellhasreacheddryness(seesectionII/7.0).Inordertoderive thisvariation,wehavetoincludeanestimateoftherateofrefluxinthecellandthispart ofthecalculationwillfollowtheschemeoutlinedinsectionI/4.0. Wecanseethatthe negativevaluesoftheenthalpiesarenoweliminatedastheD Ointhecellismaintained 2 bytheamountofreflux. Thetotalamountevaporatedisalsoinreasonableaccordwith theamountofD Oinitiallyinthecell. Itisimportanttorealizethatwehaveassumed 2 thatthewholeoftheheattransferfromthecellintheregionfilledwithvaporleadsto recondensation,i.e., wehaveoverestimatedthereflux andunderestimatedtheamount evaporated.Weshouldalsonotethatthecalculationisimprovedsomewhatifweallow forthefactthattheboilingpointreachesalimitduetothelimitedsolubilityofLiOD inD Oattheboilingpoint(thisaspectisnotillustratedinthisreport). 2 Although the calculation as outlined givesa reasonable interpretation of the behavior of the cellas thecontents are drivento dryness(eliminationof negativeenthalpiesof evaporation), we neverthelessstill derivenegativerates of excessenthalpygeneration atlongtimes. Thisisundoubtedlyduetoremaininginaccuraciesinthecalculationof the rates of evaporative cooling. At the present time it is best to restrict attention to 42HeattransfertothewallsoftheDewarcellismaintainedbythevaporphaseattheveryleastifthis phaseisfilledwithD2Ovaporattemperaturesclosetotheboilingpointoftheelectrolyte.Theheattransfer coefficientforthecellfilledwithvaporwillbeca5%abovethevalue0.85065 * 10 + 9WK + 4. 74 theearlierpartoftheperiodleadingtoevaporationtodryness,say,tot (cid:23) 18,000s. The rate of excess enthalpygeneration reaches ca 9.3 W at this time, or, say, 25 Wcm (cid:4) 3. It is important to realize that similar orders of magnitude are obtained even with the interpretationgivenbyNHE,i.e.,theestimateisrobust. II/7.0Day68: Theperiod21,300s (cid:23) t (cid:23) 86,400sfollowingevaporationtodryness. AshasbeennotedinsectionII/1.2oneoftheobjectivesofthepresentinvestigationhas beenthesearchforthepresence(orabsence)oftheeffectsof“HeatafterDeath.”The period following the evaporation to dryness on day 68 is an example of the protocol originallydescribedascase(ii)[3,4] (ii) Cell empty: cell allowed to evaporate to dryness; cell then maintained at the rail voltageofthegalvanostatwiththeexceptionthatthecelldidnotreachboilingcondi- tionsduringtheperiodleadingtodryness. The original investigation was divided into two parts: (i) the investigation and inter- pretationofthecoolingcurvesfollowingevaporationtodryness;(ii)theevaluationof thermal balances in the corresponding period. Attention here will be confined to the secondoftheseapproaches. The values of the rates of excess enthalpy generation have been based on true heat transfercoefficient,(k # R (cid:9) ,observedforthecellfilledwithelectrolyte,i.e.,0.85065 12 G 10 (cid:4) 9WK (cid:4) 4,whichwillcertainlyapplytoinitialstageoftheobservationof“Heatafter Death” when the cell is filled mainly with D O vapor. However, calibrations of cells 2 filledwithair[3,4]haveshownthattheheattransfercoefficientfallstoabout0.75of thevalueforthecellsfilledwithelectrolyte. Thevaluesoftheratesofexcessenthalpy generationhavethereforebeencalculatedusing(k # R (cid:9) =0.65x10 12 (cid:4) 9WK (cid:4) 4. The initial rate of excess enthalpy generation is approximately the same as the rate reachedduringtheperiod0 (cid:23) t (cid:23) 21,300sasthecellisbeingdriventodryness,Fig. 22.Suchacorrespondencewould,ofcourse,beexpectedifexcessenthalpygeneration takesplaceinthebulkofthemetalphase. We note also that the rate of excess enthalpy generation is about 10 times that of the rateofenthalpyinputduringthisperiodof“HeatafterDeath.” II/8.0Day69: Theperiod2400s (cid:23) t (cid:23) 32,400s. This period is of special interest in the operation of the cell becausethe cellwas dis- connected from the galvanostat at 2400 s so that the enthalpy input was zero during theremainingperiodofoperation. Inanysearchfortheeffectsof“HeatafterDeath,” the protocol there should be described at case (iii)[3, 4] Cell empty: cell allowed to evaporate to dryness; cell disconnected from the galvanostatwith the exceptions that thecelldidnotreachboilingconditionsduringthe periodleading todrynessandthat theapplicationofcase(iii)wasprecededbyaperiodcoveredbycase(ii)asdescribed insectionII/7.0. 75 Fig. 22.Comparisonofspecificratesofexcessenthalpygeneration,Wcm (cid:4) 3,onday 68duringtheperiod0 (cid:23) t (cid:23) 21,300sandtheinitialperiod21,300s (cid:23) t (cid:23) 30,300sof observationof“HeatafterDeath.” 76 The examination of the behavior of the cell has been restricted here to the time t (cid:23) 32,400 s as the usual calibration pulse was applied at t = 32,400 s. The Joule heat 1 injectedbythecalibrationsystemisdevelopedinasmallvolumesothatthiscalibration cannot be usedto derivethe trueheat transfer coefficientofthe cell for the particular operatingconditions.43 ThecoolingcurveforthisperiodofoperationisplottedinFig. 23. Itcanbeseenthatalthoughthetemperaturedifferencesbetweenthecellandwater batharesmall,theyareneverthelesssignificant. Fig. 23.Thecoolingcurveonday69followingthedisconnectionofthecellfromthe galvanostat. InspectionofFig.23showsthattheremustbeasourceofenthalpyinthesystem:firstly, becausetherateofcoolingatshorttimesistooslowtobeaccountedforbythecooling of a calorimeter with a water equivalent of 28.3 JK (cid:4) 1 and any conceivable value of theheattransfercoefficient;secondly,becausewecandetectatleastoneperiodduring whichthecellcontentsarereheated. The analysis of the cooling curve according to the method originally outlined [3, 4] usingtheequation ln (cid:8) (cid:22) 1 (cid:3) y (cid:9)7 0 y (cid:8) 1 (cid:3) y 0 (cid:9)(cid:31) (cid:28) (cid:3) tan (cid:4) 1 (cid:8) 1 (cid:3) y (cid:9)(cid:30) (cid:21) tan (cid:4) 1 (cid:8) 1 (cid:3) y 0 (cid:9) (cid:2) 4 (cid:8) k # R (cid:9) q 3t b 0 C M p wherey (cid:2) (cid:8) Dq 0 q (cid:9)7 0 q ; y b 0 (cid:2) Dq 0 0 q andDq istheinitialtemperaturedifference.Figure b 0 24 shows a plot of the experimental data; the full line shows the predicted behavior 43Ashasbeennoted,thecalibrationusedinanearlierinvestigationwerederivedbyusingaheaterspiral spanningthewholevolumeofthecell,i.e.,heatwasapplieduniformlythroughoutthisvolume. 77 usingC M =28.3JK p (cid:4) 1 and(k # R (cid:9) =0.65x10 (cid:4) 9 WK (cid:4) 4. Thedeviationsfromthisplot duetoenthalpygenerationaresimilartothosepreviouslyobserved[3,4]. Fig. 24.TheanalysisoftheinitialportionofthecoolingcurveshowninFig. 23.The fulllineshowstheRHSoftheequationplottedwithC M=28.3JK p (cid:4) 1,(k # R (cid:9) =0.65 12 G 120 (cid:4) 9WK (cid:4) 4 andq =295.204K. b Wecanalsomakethermalbalancesateachpointofthecoolingcurvesusingparticular valuesofthewaterequivalentandtrueheattransfercoefficient.ThosebasedonC M= p 28.3JK (cid:4) 1 and(k # R (cid:9) =0.65x10 12 (cid:4) 9 WK (cid:4) 4 giveinitialratesofenthalpygenerationca 0.5 W. Unfortunately,the thermal balances in the period preceding the disconnection ofthecellfromthegalvanostat(i.e.,thelastpartofcase(ii),sectionII/7.0)cannotbe madewithsufficientaccuracytoallowacomparisonoftheratesofenthalpygeneration at the end of the period followingcase (ii) and the beginningof the period following case(iii)(c.f. comparisonoftheratesattheendoftheperiodleadingtoevaporationto drynessandthebeginningoftheperiodfollowingcase(ii),sectionsII/6.0andII/7.0). II/9.0Days25and26: TheperiodDay25+76,300s (cid:23) t (cid:23) Day26+22,300s. As has already been noted in section II/1.2, there were frequent changes of current densityinexperimentMc–21. Considerationofcase(i)oftheconditionslikelytogive demonstrationsofthephenomenonof“HeatafterDeath”[3,4]: i) Cell full: cell operated at intermediate temperatures; cell current then reduced in stagesshowsthatthechangeofcurrentclosetothestartofday26ofthemeasurement cyclesislikelytoprovidethebestexampleofthisparticularcase,seeFig. 4. 78 There are two principal reasons that indicate this was the case. In the first place, the currentdensityattheendofday25isabovethethresholdvaluerequiredfortheobser- vationofthephenomenon[1]whileonday26itisbelowthisthresholdvalue.Secondly, thecelltemperatureonday25isabovethatwhichhasbeenobservedtobeimportant fortheonsetofpositivefeedback[7,8,11]whereasonday26itdropsbelowthisvalue. Wewouldthereforeexpecttoseeamarkeddecreaseoftherateofexcessenthalpygen- erationatthestartofday26fromthevaluewhichappliesattheendofday25tothat whichappliestowardstheendofday26.44 Thedatacoveringmeasurementsinthelaststagesofday25andthebeginningofday 26areusedtodefinethelowerboundheattransfercoefficient,(k # R (cid:9) .Wenoteherethat 11 we haveusedthevalueC M = 475JK p (cid:4) 1 inviewoftheevidentoverfillingofthecell onday25,seeFig.5.TheratesofexcessenthalpygenerationderivedareplottedinFig. 25. Wecanseethewelldefinedfallintherateofexcessenthalpygenerationwhich,as intheotherexamplesof“HeatafterDeath”discussedinthisreport,isconsistentwith a diffusionalrelaxationtime. We canseefromthe plotinFig. 25that thisevaluation predictsanegativerateofexcessenthalpygenerationonday25. Aswehavenotedelsewhereinthisreport,suchnegativeratesviolatethesecondlawof thermodynamicsandarecertainlyduetotheuseoftheincorrectvalueofthetrueheat transfercoefficient,(k # R (cid:9) ,givenbytheNHEanalysis. Nevertheless,wecanseefrom 12 Fig. 26thatwecandetecttheeffectsof“HeatafterDeath”onday26evenwhenusing thisfaultyanalysis.Furthermore,theincreasingvaluesofthelowerboundheattransfer coefficient (k # R (cid:9) on that day demonstrate the presence of a rate of excess enthalpy 11 generationwhichdecreaseswithtime. If we use the value of (k # R (cid:9) given by the correct ICARUS methodology, we obtain 12 therates ofexcessenthalpygenerationshownin Fig. 26. Itisimportant, however,to drawattentiontoaremainingdifficultyintheinterpretation,namely,thattheinitialrate of excess enthalpy generation on day 26 is larger than the final rate on day 25. The discrepancy would be diminished if the water equivalent were even higher than 475 JK (cid:4) 1 or if we increased (k # R (cid:9) in view of the evident increase in the D O content of 12 2 thecell,Fig. 5. Itdoesnotseempossiblethoughtoeliminatetheeffectcompletelyby anysensible choiceofthevaluesofC M and(k p # R (cid:9) so thattheeffectmaybereal. If 12 thisisso,thentheestablishmentof“HeatafterDeath”and/orpositivefeedbackwould be more complicated than is indicated by the state variables alone. For example, the time derivatives may also be involved [9, 10]. It is evident that much further work is required on these particular aspects. This work would be justified not only from the objective of clarifying the science involved, but, also, because the judicious use of positive feedback and “Heat after Death” offers us the prospect both of increasing thepowerdensityand,atthesametime,ofincreasingtheenergyefficiency. Itshould be noted that if we exclude the enthalpy input due to the cooling of the cell, the rate ofexcessenthalpygenerationintheinitial stagesofday26isapproximatelyequalto 44Excessofenthalpygenerationwasobservedonday3ofthemeasurementcycleatacurrentdensitybelow thethresholdvaluewhilepositivefeedbackwasestablishedatatemperaturebelowthisfurtherthreshold.We can,therefore,onlyregardthecriteriausedtosearchforcategoryofthephenomenonof“heatafterdeath”as rather“broadbrushindicators.” 79 Fig. 25Thespecificrateofexcessenthalpygeneration,Wcm (cid:4) 3,forthelastpartof operationonday25andthefirstpartonday26. 80 Fig. 26.Thespecificrateofexcessenthalpygenerationforthelastpartofoperation onday25andthefirstpartofoperationonday26.Evaluationgivenbythegroupat theNHElaboratories. 81 theenthalpyinput,i.e.,apowergainofca100%whereasitapproachesca1000%for the initial stagesof “Heatafter Death”according to case(ii)[3, 4], section II/7.0, and infinity for the example of case (iii), section II/8.0. It appears that if the cooling of such cells is prevented (effectively by raising the temperature of the heat sink), then enthalpy generation may be maintained for long durations (ca 1 week) at very high energy efficiencies [13]. It is evident that this aspect of the work requires intensive furtherinvestigation,particularlywithregardtoattemptstoraisethepowerdensityof suchdeviceswhilemaintainingthehighenergyefficiency. II/10FurtherCommentsandConclusions. ExperimentMc–21exhibitsallthekeyfeatureswhichhavebeenfoundinearlierwork. Theseareinthemain: (i)excessenthalpygenerationintheearlystages(t (cid:23) 2days)duetoabsorptionofdeu- teriuminthelatticefollowedby (ii)abuildupoftherateofexcessenthalpygeneration (iii) the development of positive feedback, i.e., the increase in the rate of excess en- thalpygenerationwithincreaseoftemperature (iv)amarkedincreaseintherateofexcessenthalpygenerationattemperaturescloseto theboilingpointoftheelectrolyte (v)avarietyofexamplesofthephenomenonof”HeatafterDeath,”i.e.,amaintenance ofelevatedratesofexcessenthalpyproductionfollowingreductionofthecurrentden- sitywhichmaybeaccompaniedbythecompleteevaporationoftheelectrolyte. At the same time there are some marked differences between experiment Mc–21 and the earlier investigations: the effects of some of these differences can be explained in terms of the earlier results while some of the results are surprising. The majordifferenceisthatthemeasurementcycleshadtobecarriedoutatratherlowcur- rent densities (low for the observation of the phenomenon) in view of the relatively highsurfaceareaoftheelectrode(itisnecessarytolimitthepowerinputtothecellto satisfy the design criteria of the calorimeter). As the rate of excess enthalpy genera- tionincreasesmarkedlywiththecurrentdensity[1],thevaluesachievedinexperiment Mc–21werenecessarilylimited(theratesincreasedtoca25Wcm (cid:4) 3 onday68prior toevaporationtodryness). Asecondaryconsequenceofthelowcurrentdensitieswas that the electrode was polarized in the vicinity of the region for the onset of positive feedbackformostoftheexperimentduration(seeFig. 3). Theuseofsuchconditions isknowntolimittheratesofexcessenthalpygeneration,and,inthelimit,maydestroy thephenomenon[9,10]).45. Themajorunexpecteddifferencehasbeentheobservationofthedevelopmentofposi- tivefeedbackataveryearlystageoftheexperiments(day3),atalowcurrentdensity andatalowtemperature. Itisobviouslyveryimportanttoestablishwhetherthisearly establishmentof positivefeedback is a property of Pd/B alloys (such as the electrode usedinexperimentMc–21). 45Possiblybecauseofthecrackingoftheelectrodesduetotherepeatedloadinganddeloading. 82 Fig. 27. TheICARUS–14Calorimeter. Amajorfeatureoftheinvestigationof“HeatafterDeath”inexperimentMc–21isthe demonstrationthattheratesofexcessenthalpygenerationbeforeandaftertheonsetof thephenomenonareprobablyidentical. Suchanidentitywouldbeexpectedifexcess enthalpy generation takes place in the bulk of the electrode, but these effects clearly require further investigation. It is also apparent that these processes relax with a dif- fusional relaxation time and prolonged maintenance of the effects evidently requires specialconditions(increaseofthetemperatureoftheheatsinks)[13]. TheinvestigationofexperimentMc–21hasdemonstratedyetagainthatcertainmeth- odsofdataevaluationareunsoundand/orinaccurateorimprecise(comparee.g.,[6,9, 10]).Furthermore,itisessentialtoavoidtheeffectsofpositivefeedbackasitisimpos- sibleingeneraltocalibrateclosedloopsystemssubjecttosuchfeedback. Calibrations canonly be achievedif theeffectsare nottoo marked,ashas beenthecasefor day3 ofexperimentMc–21. Unfortunately,itisalmostcertainthattheinvestigationscarried outbyNHEhavereliedpreciselyonsuchunsoundandinaccuratemethodsofcalibra- tion and the effects of positive feedback have been ignored. However, this neglect is probablyquitegeneraland,nodoubt,accountsformanyofthecontradictoryresultsin thisfieldofresearch. Itshouldbenotedthatmuchofthepointlesscontroversyinthis field could have been avoided if it had been possible to replace the ICARUS-1 to -3 Calorimeters, Fig. 1, by the ICARUS-4 version,(later reclassified as the ICARUS-14 Calorimeter), Fig. 27. Whileitis not certainthat thisparticularredesign would have eliminatedtheweaktimedependenceoftheheattransfercoefficientsobservedwiththe ICARUS–1Calorimeter,it islikely thatthiswould havebeentrueandthat thesesys- tems could have been developed so that all measurements could have been evaluated 83 withasingle,predeterminedvalueofthetrueheattransfercoefficient. Finally,itisimportanttonotethatithasbeenpossibletoachieve: (vi)asatisfactoryinterpretationofevaporationtodryness(day68). This interpretation has had to take into account: the actual barometric pressure, the change of the boiling point of the solution with increasing electrolyte concentration (saturation of the electrolyte – not discussed in the present report), and changes in the reflux ratio.46 However,prolongedinvestigationsof boiling conditions[7, 8] will clearlyrequirethe designandapplicationofdualcalorimeterssuchastheICARUS-9 version [4, 8]. It is also important to determine whether the marked increase of the rates of excess enthalpy generation at temperatures near the boiling point are depen- dent on the establishment of boiling conditions or are simply due to the increase in temperature. Whileitiscertainlydesirabletodeveloppressurizedsystemstoincrease the boiling point, significant increases in the boiling point could also be achieved by usingconcentratedelectrolytesolutions. Theuseofsuchelectrolyteswouldallowthe extensionoftherangeofapplicabilityoftheICARUS-1calorimeters. Finally,wecannotethattheinterpretationofthisexperimentgivesagoodillustration of the need to evaluate all such measurements as individual case histories: the state of development of research in this field in 1993 (when the first ICARUS system was constructed)wascertainlynotatthepointatwhichsuchinterpretationscouldbecarried out as a matter of routine. Furthermore, the instrumentation also required a number of additionaldevelopmentsto facilitateanysuch attempts atroutine evaluations. The ICARUS-14system(thendescribedwiththelabelICARUS-4)wastobethenextstep, but,ashasalreadybeennoted,thismodificationcouldnotbeaccomplished. II/11.0References. 1. M.Fleischmann,S.Pons,M.W.Anderson,L.J.LiandM.Hawkins,J.Electroanal. Chem.,287,293(1990) 2. ICARUS–1,Documentversion1.0(Dec. 1993). 2A.ICARUS–2,Documentversion2.0(Feb.1995)47 3. S.PonsandM.Fleischmann,ICCF–4,p. 8,1994 4. S.PonsandM.Fleischmann,Trans.FusionTechnology,26,87(1994) 5. T.Saito,M.Sumi,N.AsamiandH.Ikegami,ICCF–5(1995) 6. M.Fleischmann,ICCF–7,p.119(1998) 7. M.FleischmannandS.Pons,ICCF–3,p.47(1993) 8. M.FleischmannandS.Pons,PhysicsLetters,A176,118(1993) 9. M.Fleischmann,S.Pons,M.LeRouxandJ.Roulette,ICCF–4. p. 1(1994) 10. M.Fleischmann,S.Pons,M.LeRouxandJ.Roulette,Trans. FusionTechnology, 26,323(1994) 11. M.Fleischmann,ICCF–5,p. 140(1995) 46Itisunlikelythatthevariationofthedistillatewithtime(asdeterminedintheNHEinvestigation)could beusefullyinterpreted. 47Documents,ref.2and2Aarenotavailableinopenliterature.ContactDr.Milesforfurtherinformation. 84 12. G. Mengoli, M. Bernardini, C. Maduchi and G. Zannoni, J. Electroanal. Chem., 444,155(1998) 13. M.FleischmannandS.Pons,unpublished,August1994. 14. T.Roulette,J.J.RouletteandS.Pons,ICCF–6,p.85(1996) 85 TABLE:Evaluationofheattransfercoefficients Thecombinedabridged (cid:8) k # R (cid:9) and 21 (cid:8) k # R (cid:9) –spreadsheetspreparedaccordingtothein- 31 structionsinthe ICARUS-systemsHandbooks(restriction oftherange oftheintegra- tionstotheregionofapplicationofthecalibrationpulset 1 (cid:23) t (cid:23) t ).Thethirdmeasure- 2 mentcycleofexperimentM–21.Theevaluationof (cid:8) k # R (cid:9) 31 (cid:19) (cid:8) k # R (cid:9) 351 (cid:19) (cid:8) k # R (cid:9) 32 (cid:19) (cid:8) k # R (cid:9) 362 (cid:19) (cid:8) k # R (cid:9) 22 and (cid:8) k # R (cid:9) . Modificationoftheprocedurefortheevaluationof 262 (cid:8) k # R (cid:9) and 22 (cid:8) k # R (cid:9) to 262 takeaccountoftheeffectsof”positivefeedback”andevaluationofthesecoefficients. Column1: Theelapsedtimes/s(fromthestartofthemeasurementcycle). Column2: 109C M p (cid:8) q (cid:21) q 0 (cid:9)7 0 < f (cid:8) q (cid:9) dt /WK (cid:4) 4. Here,q 0 (cid:2) 300 > 3175K,theaverageof the11measurementsprecedingtheapplicationofthecalibrationpulse. Column3: 109 < (cid:8) input (cid:9) dt 0 < f (cid:8) q (cid:9) dt /WK (cid:4) 4 Column4: 109 (cid:8) k # R (cid:9) /WK 31 (cid:4) 4. Column 5: 109 (cid:8) k # R (cid:9) /WK 31 (cid:4) 4: correlation coefficientC M/JK p (cid:4) 1. The arrowsindicate therangeofthefittingprocedure. Column6: 109C M p (cid:8) q (cid:21) q 0 (cid:9)$ 0 D < f (cid:8) q (cid:9) dt /WK (cid:4) 4 [evaluationof (cid:8) k # R (cid:9) and 32 (cid:8) k # R (cid:9) ] 352 Column7: 109D < (cid:8) input (cid:9) dt 0 D < f (cid:8) q (cid:9) dt /WK (cid:4) 4 [evaluationof (cid:8) k # R (cid:9) and 32 (cid:8) k # R (cid:9) ]. 352 Column8: 109 (cid:8) k # R (cid:9) /WK 32 (cid:4) 4. Column9: 109 (cid:8) k # R (cid:9) /WK 352 (cid:4) 4; correlationcoefficientC M/JK p (cid:4) 1. Thearrowsindicate therangesofthefittingprocedures. Column 10: 109C M p (cid:8) q (cid:21) q 0 (cid:9)7 0 D < f (cid:8) q (cid:9) dt /WK (cid:4) 4 [evaluation of (cid:8) k # R (cid:9) and 22 (cid:8) k # R (cid:9) ]. 252 Hereq 0 (cid:2) 303 > 074K,theaverageofthelast11measurementsduringtheapplicationof thecalibrationpulse. Column11: 109D < (cid:8) input (cid:9) dt 0 D < f (cid:8) q (cid:9) dt /WK (cid:21) 4. Column12: 109 (cid:8) k # R (cid:9) /WK 22 (cid:4) 4. Column13: 109 (cid:8) k # R (cid:9) /WK 252 (cid:4) 4:correlationcoefficientC M/JK p (cid:4) 1. Thearrowsindicate therangeofthefittingprocedures. Column 14: Modification of column 11 to take account the effects of “positive feed- back”. Column15: Valuesof109 (cid:8) k # R (cid:9) takingintoaccounttheeffectsof“positivefeedback”. 22 Column16: 109 (cid:8) k # R (cid:9) /WK 252 (cid:4) 4; correlationcoefficienttaking intoaccounttheeffects of“positivefeedbackC M/JK p (cid:4) 1. 86 87 88 89 CHAPTER5: ANOVERVIEWOFCOLDFUSIONTHEORY. ScottChubb 1.0Introduction. NavalResearchLaboratory(NRL) involvementin coldfusion (CF) startedwhenTal- botChubbandScottChubbstartedtodevelopatheoryoftheanomalousheatingeffect [1]. The basis of this theory involves known phenomena (associated with wave-like behavior)thatoccurwhenhydrogen(H)anddeuterium(D)interactwithPd(andother transition metal) lattices. In particular, at an early stage, Talbot and Scott Chubb ob- servedthatwellknowneffectsassociatedwithHinmetalsystems,aswellasthewell documented literature concerning this area could imply that the hypotheses involving highenergy,closeproximityeffectsthatwerecommonlythoughttoberelevant,atbest, couldbeonlytangentiallyrelatedtotheexcessheateffect;atworst,theseideascould belargelyirrelevant. Inmostcases,theassociatedpicturereflectsanintuitivescenariothatisbasedoncon- ventional nuclear fusion, where a classical/semiclassical model applies, involving a collision betweentwo, clearly distinguishableparticles, at an isolated location in free space. Althoughthisphysicalmodelisperfectlysatisfactoryforthiskindofsituation, itomitsimportantdetailsinvolvingcoherenteffectsinsolidsthatareknowntobees- peciallyimportantatlowtemperature. Thispicturealsoimplicitlyrequiresthathighmomentumparticleseitherbepresentor becomeinvolvedinsuchawaythatradiation,atcopiouslevels,bereleased. Because, infact,itisnowknownthatappreciablelevelsofradiationarenotinvolved,itisclear thatthissemiclassicalpictureatbestisonlytangentiallyinvolved. Atworst,theasso- ciated picture oversimplifies the associated situation to such an extent, that it, in and of itself, can be viewed as providing a hidden barrier for understanding the relevant physics. Becauseoftheinherentlimitationsofsuchabarrier,itisconvenienttoview 90 thisconventionalpictureofexistingphysicsasabox. Withinthiscontext,itisuseful to examine precisely what is known about this box and how it relates in more gen- eral terms to less conventional pictures, which although consistent with conventional Physics,areviewedtodayasbeingoutsidethebox. Withthisinmind,inthenextsection,were-examinetheconventionalpictureoffusion, basedonitspre-definedframework. (Werefertothisframeworkastheboxassociated with conventional fusion, and to the underlying theory behind the framework as its organizingprinciples.) In the followingsection, inherent oversimplificationsof these principlesareidentified. Inthesamesection,throughtheconventionallawsofphysics, we explain how it is possible to move beyond the boundaries of the box, associated with these organizing principles. In the third section, we provide some history asso- ciated both with theidentification of these boundariesandwith attempts to overcome them. This provides a useful context for identifying well formulated theories, from thosethatmustbeviewedasbeinginamoreprimitivestate. Specifically,althoughthe associatedphenomenahavebeenillusive,withtimenotonlyhavethemostimportant effects been identified, three theories (by Chubb and Chubb [2], Hagelstein [3], and Kim [4]) provide a common theoretical framework, involving many-particle interac- tions (many-body physics), that are based on well formulated physical ideas that are consistentwithknownphysicallaw,andthesetheoriesnotonlyprovideausefulframe- work forexplainingmanyofthese effectsbutformakingnewpredictionsconcerning theirbehavior. Inparticular,afterpresentingcriteriaforidentifyingthemostusefultheories,weexam- inethreeofthemostwelldevelopedtheories(whichinclude[2–4],aswell,asafourth theory by Preparata [5]). Each of these theories is sufficiently well developed that it providesaprocedureforconstructingarealisticreactionrateexpressionthatexplicitly illustrateshowreactionsmightoccur,basedonknownphysicaleffects,insucha way thatexcessheatcouldbeproducedthroughafusionreaction,withoutneutrons,tritium, and radiation. Three of these theories[2 through 4] also make use of a basic, known idea,i.e.,coherence,asitrelatestomany-bodyphysics,toaccountforthis.(Preparata’s theory[5]involvescoherenceinanon-standardform,thatmightormightnotbeappli- cabletoconventionalmany-bodyphysics). Thethreetheories[2–4]alsodealinfundamentalways,withthereleaseofmomentum, coherently from one location, to many locations, in a way that not only is consistent withtheknownlawsofQuantumMechanics,butthataccountsforthereasonthatthe standard, two-particlepictureisdeficient. Eachofthese theorieshasbeenformulated usingawellformulatedmathematicalmodel. Forthisreason,eachofthesetheoriesis knowntoapplywhenparticular,welldefinedmathematicallytestablehypotheses,and limits hold. This is by no means true for many of the more speculative theories that havebeenpresentedpreviously.Finally,thethreetheories[2through4]involvemodels that includecoherentcouplingto the solidthat canbe generalized, withinthe context of known physics, in such a way that they can be readily used to investigate other Low Energy Nuclear Phenomena. With this point in mind, in Section 5, additional informationaboutthecommonfeaturesofthetheoriesisdiscussed. Thefinalsection 91 providesaseriesofconclusionsabouttheexistingstateofaffairs,andpotentiallessons thatmightbelearnedasaresultoftheadjudicationprocess. 2.0 Inside and Outside the Box and the Organizing Principles of Conventional Fusion. Logicalthoughtrequiresrules.Inphysics,thelogicalrulesfollowfromNewton’slaws of motion, Maxwell’s equations, quantum mechanics, and relativity. Because these rules provide a framework, often they can be self-limiting. For example, sometimes physicistsmisinterprettherules,simplybecausetheyareconditionedtolookatthemin aparticularway. Theybecomeusedtoaparticularworldview. Theworldviewcanbe thoughtofasakindofboxthatdefinesacomfortzone. Often,theboxistiedtotheway wehavelearnedaparticularsubject. Differentpeopleviewtheboxindifferentways. Kuhn[6]referstoit,abstractly,asitrelatestoscience,asaparadigm. Othershavenot beenasopenminded[7]. 92 Figure 1 showsa pictorialrepresentation ofconventionalfusion reactionssuperposed on an idealized representation of the box, associated with what is commonly viewed as conventional(labeled inside the box)andunconventional(labeled outsidethe box) science. In this schematic, all reactions originate from a configuration in which two deuterons (shown as proton/neutron pairs) overlap with each other in a manner that formsaconfiguration(shownin thecenteroftheplot)thatresemblesanexcitedstate ofa4Henucleus. Thetwodominantreactions(D+D (cid:6) 3He+n,andD+D (cid:6) 3H+p)that occurinfreespaceareessentiallyblindtothepresenceoftheelectromagneticinterac- tion(EMI).Forthisreason,itispossibletotreatthesereactionswithinaframeworkin whichthedependenceofthereactiononelectromagneticinteractionsisindependentof itsdependenceonthenuclear(strongforce)interaction. Thismeansthatinthesereac- tions,theassociatedwavefunctionsdescribingtheinitialandfinalstatesdonotcouple thenuclearandelectromagneticinteractions. Asaresult,thegeneralreactionrateex- pressioneffectivelyprecludesthestrongforcefromtalkingtotheelectromagneticforce, by construction. The figure schematicallyillustrates this point through the labels (ig- nore E.M.), nextto thearrowsthat areshownin therightportionofthe figure. Also shown is the remaining fusion reaction (D+D (cid:6) 4He). This reaction occurs rarely in conventionalfusion.Forthisreason,inthefigureitisshownasoccurringatthebound- aryofthebox. Asecondreasonwehavedrawnitattheboundaryisthatitviolatesa paradigmthatmanynuclearphysicistsbelievetobevalid: inconventionalfusion, the strongandelectromagneticinteractionsremainuncoupled. Forthisreason,itiswidely believedthatthefinal(D+D (cid:6) 4He)reactionshouldrarelyoccurandthetworemaining reactions should occur with roughly the same probability. However, the D+D (cid:6) 4He reactiondoesoccur,andthereasonthatitisnotfrequentlyobservediswellunderstood: it violates energy and momentum conservation unless a high energy momentum g (cid:21) rayisemitted,andtheassociatedEMIinvolvesacomplicated(quadrupolar)coupling between nucleon spins (that occurs as a second order electromagnetic process). Two importantpointsareasfollows: (i)althoughthisfinalreactionoccursinfrequentlyrel- ativetotheothers,whenitoccurs,thenuclearandelectromagneticinteractionsdotalk toeachother,and(ii)itoccursrarelybecausetheassociatedprocessesinvolveoverlap betweentwoparticlesatasinglelocation. 2.1MotivationalPhysicsforGettingOutsidetheBox. Partoftheconfusionwiththeboxassociatedwithconventionalnuclearphysicsinvolves thedefinitionofmomentump: forasinglechargedparticle,pdoesnotequalmass(m) timesvelocity(v); therules ofthe boxare: for a particlepossessing chargeq, mv=p- q/cA,whereA(thevectorpotential)isassociatedwiththeelectromagneticinteraction, and c is the speed of light. Althoughthis rule is based on classical physics, howand whereitappliesseemstohavebeena sourceofconfusion. Therulefollowsfromthe boxdefinedbyclassicalphysics. (Falseassumptionsaboutthisrulenotonlyappearto haveledtoconfusionaboutColdFusionbuttomoreseriousproblems.) Anexampleof theimportanceofthisdistinctionoccursinthep (cid:6) 0limit, whenmanyparticlesshare a common density r . When this occurs, mv, which is proportional to the current J 0 (provided r is uniformly constant [8]), becomes proportional to A. But A, which is 0 definedbythestaticwaveequation( (cid:21) (cid:209) 2A (cid:2) 4p J 0 c),thenobeysaHelmholtzequation 93 [8] ( (cid:21) (cid:209) 2A (cid:2) (cid:21) 4p q2r A 0 0 mc2 (cid:9) that results in A asymptotically vanishing beyond a critical coherence length, where J approaches a constant value. This occurs even in theabsenceofanapplied electromagneticfield(EMF). Theresultingpicture explains thephenomenonofsuperconductivity.Italsoexplainshowasp (cid:6) 0,superconductivity not only is present, but because the current vanishes at some boundary, surrounding theregionwheresuperconductivityoccurs,theeffectsofboundariesmayresultinthe expulsion of magnetic flux when p=0 (the Meissner effect) or flux quantization [8], whenpdoesnotvanishbuttakesonvaluesthatareconsistentwiththeassociatedrules (definedbythebox)associatedwiththerequirementsofquantummechanics[8]. Thebasisofbothphenomenaisthatpdoesnotequalmv. InsituationswheretheDe- Brogliewavelengthsofparticlesbecomesufficientlylarge,particlesbecomewavelike. Inthiskindofsituation,theaveragevalueofthegradientofthephaseoftheassociated collection ofwaves(whichis describedbythe many-bodywave function)defines the momentum. Theimportantpointisthatthephaseofthemany-bodywavefunction,as opposedtoaquantityrelatedeitherdirectlytothecurrentortomass G velocitydefines howthemomentumbehaves.Whenp (cid:6) 0,thisquantitycanbeaffectedinwaysthatare non-localincharacter.Thismayoccurbecausenon-localchangesinAcansignificantly alter the valueof the phase. Becausea priori, it is notpossible to predict if a solid is atrestorinmotion,forexample,itscenter-of-masswavefunctioncanbealteredbyan arbitrarycomplexnumber. Thisintroducesthepossibilityofanarbitrarygaugetrans- formationinthedefinitionoftheAthatappliesinsideandoutsideasolid. Becausein thep (cid:6) 0limit,itbecomespossibletodetermineifthesolidisinmotionoratrest,the associatedarbitrarinessingaugeisremoved. Notonlydoesthismeanthattheassoci- atedgaugesymmetrybecomesbroken,butphysicaleffects(forexample,theexpulsion ofmagneticflux,orspontaneouslatticerecoil[asintheMossbauereffect])canoccur. The resulting coherence can be viewed in different ways, within the framework (the box)associatedwithaparticulardiscipline. Insimilarways,effectsofperiodicorderandothersymmetriescanbecomeimportant in situations in which the wave-like character associated with large DeBroglie wave- lengthsbecomesimportant. Theimportantpointisthatbecausemomentumisassoci- atedwithwave-likebehavior,itcanchangesuddenly,inunexpectedways,onarbitrar- ilyshorttimescales. Thesechangescanresultininstantaneouschangesinwhichlarge amountsofmomentumcoherentlyareshiftedto manyparticles,andviceversa. How orifthisoccursisdictatedbythedynamicsofthemany-bodysystem. 3.0SomeHistoryofTheoreticalDevelopmentandSomeUsefulCriteria 3.1AHistoricalDevelopment Ingeneralterms,oversimplificationhasplaguedCFandCFtheory,bothinthepast,and atthepresenttime. Inparticular,ataveryearlypointintheadjudicationprocess,the overlysimplified picture of Fusion, associated with the box, described in section 2.0, undermineddiscussion of Cold Fusion (CF) claims to such an extentthat the box, its products,andtheassociatedcontext,obfuscatedidentificationoftherelevantproducts 94 andprocess. This created such confusion that the resulting uproar caused a serious breach in the conventional scientific process [9]. From this starting point, for quite a while, it be- came virtually impossible to obtain a reasonably unbiased assessment of the existing theoretical situation. This occurred not only in the conventional review process in mainstreamscientificmeetings(wherediscussionsaboutCFandCFtheoriesremained largelynonexistentuntil1996),butalsoinlessconventionalsettings(includingthefirst fiveInternationalConferencesonthesubject). It also affected not only how theories were adjudicated, but how various reviews of theorieswereprepared.Inparticular,becauseoflackofinvolvementofoutsidereview- ers, theoretical ideas of marginal utility not only have been proposed, but published reviews of these ideas have appeared that neither have been objectively reviewed or assessedbasedonobjectivecriteria. Allofthishasoccurredprimarilybecauseoflack of funding and interest, and even rudimentary knowledge(in some cases [10]) of the relevantfacts.Furtheraggravatingthesituationhasbeenalanguageproblem:thefield, which was misnamed from the beginning, attracted many individuals with different backgrounds,areasofexpertise,andevendifferentintuitivenotionsaboutwhatconsti- tutesameaningfuldefinitionoftheory. Afundamentalreasonforthisisthatconsiderableattentionwasfocused,fromthebe- ginning,onmarginaleffects(involvinghighenergynuclearproducts).Asaresult,many oftheintuitivetheoreticalideasassociatedwithNuclearandHighEnergyPhysicswere applied. Unfortunately, because the associated effects have proven to be marginal at best (if applicable, at all), the associated intuitive ideas have been a source of confu- sion. Forexample,itwaswidelyassumedthattheseeminglyobviousideathathighmomen- tum particles are required by CF should be invoked. A less obvious intuitive notion thatappearstohavebeenapotentiallymoreserioussourceofconfusionistheopinion that one or several guiding principles, associated with either one, or a small number of particular forms of particle/particle reaction, or particular forms of interaction, are responsible for all of the observedphenomena. In particular, initially, logic based on reduction(orreductionism)toa singleformofreaction(or smallsetofreactions)led anumberoftheoriststospeculatethatsomenewformofparticle(Rafelski[11],Teller [12]),orinteraction(MayerandReitz[13],Vigier[14],Mills[15])couldbeinvoked.In subtlerways,thisreductionistprinciplehaspersisted,eventothepresenttime(Kozima [16]). A theoretical construction involving this Reductionist philosophy can be appropriate and useful when it is possible to identify how momentum is distributed. It can (and probablydoes)causeconfusionwhenmanyparticlesinteractwithlowmomentum.For thisreason,theintuitiveideathatsuchaconstructshouldbeapplicablemayobfuscate therelevantphysics.Specifically,insituationswherethisphilosophyhasbeenusedasa guidingprinciple,notonlyhasconfusionresulted,but,inanumberofcases,arguments 95 aboutterminologyandmeaninghaveresultedthathavehadacounterproductiveeffect oncommunication.Importantreasonsfortheassociateddeteriorationindialoguepartly reflect the very different perspectivesbetween theoretical practices followed by High EnergyandNuclearPhysicists,asopposedtothosethatareusedbyChemistsandSolid StatePhysicists. Anadditional,potentiallymoresignificantreasonforthisdeterioration,however,may reflectamorefundamentalaspectoftheproblem:relianceontheReductionistphiloso- physeemstobequiteappropriateintheexperimentsinvolvingcollisionsbetweenpar- ticlespossessinghighmomentum(HM),butprobablydoesnotapplyingeneral. When HMparticlesareused,clearlydefinedexperiments,involvingwelldefined,controllable variablescanbeconducted. Relianceonthisphilosophycanbecomeinappropriatein lowertemperatureenvironments,associatedeitherwiththegroundstateornearground state configurations. This is becausein these kinds of configurations, frequently,it is difficultto defineeither the experimentalsituationin preciseterms or to identify pre- ciselythevariablesthatgoverntheunderlyingdynamics. AnotherwayofphrasingthispotentialproblemisthatbecauseintheReductionistphi- losophyanattemptismadetoidentifyaparticularformofinteraction,itispossibleto misidentifytherelevantphysicssimplyasaresultofoversimplification. Inparticular, thiskindofapproachcansimplyfailtoincorporatetheeffectsofmany-particleinter- actionsthatareknowntooccuratlow/moderatemomentum. Thefactthatitisentirely possiblethattheseinteractionsareresponsibleforthecomplicatednatureoftheunder- lying phenomena suggests that a more useful approach involves a less restrictive set ofassumptionsthantheonesthatresultfromapplyingaReductionistphilosophy. For example,ininvokingthisReductionismconstruct,Teller[12]pointedoutthatnotonly isitnecessarythattheassociatedtheorybeconsistentwithallknowneffects, butthat insuringthatthisoccursisadifficulttask. Themuchsimpleridea,thatmomentacouldbesharedbymanyparticles,atonce,ina well defined way in solids, or through related, coherentphenomena, involvingmany- bodysystems,notonlyisaconsiderablymoreworkablehypothesis,butthisideawas suggestedearlyinthedebatebySchwinger[17],andothers[2]. Theimportantpointis thatboththeReductionismapproachandtheappealtothenotionofcoherencerelyupon knownstrategiesforovercomingseeminglyimpossiblecircumstances. Unfortunately, neithertheideaofidentifyingawaytoovercomeexistingtheoreticallimitations,orthe underlyingspiritthatisresponsibleforadoptingthesekindsofstrategiesseemstohave beenfullyappreciated. (Inparticular,Schwinger[17]wascriticized[18,1,10],based ona detailedargumentthatfocusedontheparticularmechanismthatheproposedfor coherence that assumed that the argument required that large changes in momentum occur at a particular point. Teller was criticized for his language. Neither criticisms paidattentionto underlyingmotivation: a meansofgoingoutsidethe box, associated withconventionalfusion.) Despitetheseproblems,bothSchwinger[17]andTeller[12]recognizedanimportant point. Conventional thinking about fusion has limitations. These were ideas. They 96 wereandremainimportant. Buttheyarenottheories.Ideascanleadtotheories.Ideas, bythemselves,aremerelyideas. Anunfortunateproblemisthatalthoughwelldefined theories that are consistent with the knownlaws of physicsdo exist, the largerscien- tific community appears to be ignoring them. Partly becauseof this fact, evenwithin the CF community,confusion [1, 10, 18, 19] existsabout what constitutes (or should constitute)atheory. 3.2CriteriaforaUsefulTheory In1990,Preparata[19]proposedaseriesofmiraclesthatinhisviewanyCFtheoryhad toaccountfor,inordertoforitbeconsidered“valid”.Giventhelackofcommunication that waspresentatthetime, andtheassumptionthat intuitivenotionsassociatedwith high energyphysics provided a useful starting point for understanding CF, this state- mentwasuseful. However,withhindsight,Iwouldsuggestthisviewreflectedmorea fundamentalproblemassociatedwiththerelevantperspectiveatthetimethanwiththe relevant physics. Specifically, Preparata defined the problems that would be relevant providedCFmimicsHotFusion. InthecontextofHotFusion,overcomingtheseprob- lemsseemedtobemiraculousbecauseofaverybasicassumption: forCFtooccuras itoccursinHotFusion,itisnecessaryformomentumfromasmallnumberofparticles tobeimparted,allatonce,ataspecificlocation. Thisisaperceivedproblemthatmaybeirrelevant,providedinstantaneously,momen- tumistransferredeitherfroma smallnumberofparticlestomanyparticles(asin the Mossbauereffect),orbetweenmanyparticles(asinalaser).Infact,bothhowPreparata identifiedandhowhedealtwiththisproblem,reflectsa moregeneraldifficultyinthe associateddebate: a propensityfor overrelianceonspecific, detailedviewsofthe rel- evant theoretical framework, without identifying a set of widely accepted organizing principles. Itcanbearguedthatamoregeneralprincipleprobablyapplies:thepossibilityofcoher- enttransferbetweenmanyparticlesofmomentumtomanylocations,atonce.Although inhighenergyphysicsthisideaisforeign,(asillustratedbytheexamplesmentionedin section2.0),itiswellknowntooccurinthelowtemperature(smallmomentum)limit inwhichtheidentitiesofindividualparticlescanbecomelost. Given the dynamics at the time, Preparata’s efforts were admirable. In fact, he was quitecorrectinidentifyingaparticularsetofideasthatbotherhighenergyphysicists. He was also quite correct in identifying a particular concept that could eliminate the associated problem: coherent coupling between an electromagnetic field and a solid. Inaddition,heidentifiedaparticularformofcoupling,involvingthepossibilityoflow momentumfluctuationsthathesuggestedcouldprovidesuchacoupling. Althoughtheideaofphotoninducedcoherenceinvolvinglowmomentumfluctuations isausefulstartingpointforpotentiallydescribingtheassociatedphenomena,thereare two seriousfailingsinhistreatment: (i)heassumedanoversimplified(semiclassical) couplingbetweenthephotonsandthesolid,involvingapictureinwhichdiscretepar- 97 ticlesareinvolvedintheinteractionatisolatedlocations,and(ii)moreimportantly,he assumedthathisveryspecificmodelwouldbecomewidelyaccepted. ThereasonforsinglingoutthissecondproblemisassociatedwithwhatIwoulddefine asthemostimportantgoalofanytheory: 1. Foratheorytobeuseful,itmustbeaccepted. Inordertosatisfythisassumption,itfollowsthat: 2.Foratheorytobeuseful,itmustbebasedonorganizingprinciplesthatareconsistent with the predominant language and theories that are present at the time the theory is formulated. Toinsurethatbothoftheseassumptionsaresatisfied, 3. A theory must be reducible to mathematical expressions that are useful to experi- mentersandarebasedonknownresults,derivedfromtheorganizingprinciplesassoci- atedwithknowntheory,asacceptedbythewiderscientificcommunity. AlthoughPreparataidentifiedfailuresintheexistinghighenergyphysicsparadigmas- sociatedwithpossiblelowenergynuclearreactions,hehaddifficultyhavinghistheory acceptedbecauseitwasnotbasedonwidelyacceptedorganizingprinciples.Incontrast tothisproblem, althoughSchwingeridentifiedwidelyacceptedorganizingprinciples, andusedtheseprinciplestodefineausefulmathematicalframeworkforanalyzingthe associatedeffects,histheorywasnotacceptedbythehighenergyphysicscommunity (includingPreparata)becausethiscommunityfoundthathisorganizingprincipleswere foreign. Unfortunately,because some of the mathematical details associated with his particularmodelcouldbequestioned,evenbysolidstatephysicists,afterhisdeath,ar- gumentswerepresentedthatquestionedthevalidityofhistheory,basedonveryspecific aspectsofhismodel[18,1,10]. Inbothcases,thetheoryfailedtosatisfyrequirement 3. (Preparata failed because he based his theory on organizingprinciples that are not widely accepted. Schwinger failed because, although his organizing principles were sound, they were not recognized as being relevant and because a detailed analysis of theassociatedmathematicalexpressionscouldbequestioned,basedonknownresults.) AnadditionalreasonbothPreparataandSchwingerhaddifficultyinhavingtheirtheo- riesacceptedisthattheexperimentalsituationwaspoorlydefinedinitially.Since1995, this situation has changed. In particular, it is now known that high energy particles essentiallyarenotinvolvedintheassociatedphenomena. Althoughvariouslowlevelbyproductsarefoundtobeproduced,inthemostwellstud- iedcase(involvingPd/D),thedominantbyproductisHelium-4,which,inmostcases, is released either in regions near the surfaces, interfaces, or cracks of the associated materials,orintheout-gases,locatedoutsidethematerials. Italsoisnowwidelyrec- ognizedthatmaterialpreparationseemstobeveryimportantininitiatingtheeffect,that Helium-3alsocanbefrequentlyproduced(butthatthisisnotthedominantbyproduct), 98 andthatincasesinvolvinganomalousheatinNi-basedsystems,averydifferentform ofreaction,initiatedwithsignificantamountsofHpresent,occurs. TheseexperimentalresultssuggestthreeadditionalrequirementsforavalidCFtheory involvingPd/D: 4.AnappropriatePd/Dtheorymustexplainwhyhighmomentumparticlesarevirtually neveremitted. 5. Itmustexplainwhycouplingcanbematerialspecific. AnappropriateCFtheoryassociatedwithNishould 6. Eitherexplainorprovideamechanismfor explainingwhytheNienvironmentpo- tentiallycanresultinformsofCFthatareverydifferentthaninPd/D. In fact, in the context of many-body physics, based on a well defined reaction rate expression,itispossibletosatisfyallsixofthesecriteria,providedtheassociatedtheory addressesanadditionalrequirement. 7. Thetheoryshouldexplainhownucleardimensionandatomicdimensionprocesses canbecoupledwithoutrequiringthereleaseofhighmomentumparticles. Therearetwoadditionalcriteriathatobviouslymustbesatisfied. 8. HowtoovercometheCoulombbarrier. 9. The theory should also provide a framework for understanding when high energy particlesarereleased. 4.0UsefulTheories Inthelastsection,asetofcriteriaforidentifyingmorematuretheoriesfromthosethat mustbeviewedasbeingincompletehasbeenprovided. Giventhelimitationsofwhat canbe presentedin an article ofthisscope, only those theorieswillbe examinedthat satisfythesecriteria. Thisdoesnotmeanthatothercreativeideasdonotexistconcern- ingtheassociatedphenomena.(Literallyhundredsofideasaboutthesubjecthavebeen suggested.) Information about the associated materialcan be obtained elsewhere (for example,inthereviewbyStorms[20]). Toreiterate,foratheorytobecomeacceptable(inaworkableperiodoftime),itmust useexistingphysics,andtherulesassociatedwithexplanationsofexistingphenomena. In this regard, it should be emphasized that beyond the well accepted rules of high energyphysics,andconventionalnuclearphysics,thereareadditionalconstraintsthat areappropriate. Forexample,quantummechanics(QM)isnotalocalizedsubject. The experimentercanaffectoutcomes,andbecauseofthisfact,certainpremisesbasedon assumptionsaboutlocalitysimplyareinappropriate. 99 Finally, theories do have organizing principles; in particular, any viable theory must possess a limit whereit is rigorously valid, i.e., provableby experiment. Giventhese assumptions,itsimplyisnottruethatcriteriaforassessingthevalidityoftheoriescan be stated in terms of input and output information associated with predictions about experiments.Itmusthavesometangiblerelationshipwithexistingphysicaltheory,and itmustberelatabletoexperiments. Seriousimplicationsfollowfromtheseassumptions.Forexample,itissimplyincorrect to believe that a theory is credible that is not related to known phenomena. For this reason,thepremisethattheoriesthatrelatepurelytoColdFusionisnotvalid. Instead, avalidtheorymustbebasedonorganizingprinciplesthatcanbeshowntohavesome validityoutsideColdFusion. Forthisreason,anumberofthemoreexotictheories(Mills,Matsumoto,etc,forexam- ple)donothavecredibility. Furthermore,QM,andthewellknownrulesforreactions associatedwithQM,should(anddo)providetheguidingprinciplesthatshouldbeused for assessing the validity of a proposed theory. Unfortunately,outside of efforts by a handfulofpersistenttheorists,thiskindofapproachhasnotbeenused. Thesealternativeeffortssimplymustbeviewedasbeingincomplete. Forthisreason, theories that purport that theyillustrate the phenomena as occurring (for example, by overcoming the Coulomb barrier) without showing how the results relate to reaction rates, orrelatedquantities,simplymust beviewedasbeingin a primitivestateofde- velopment,andshouldnotbetakenasseriouslyasthosethathavedonethis. Asmentionedinsection2.0,animportantsourceofconfusioninCFhasresultedfrom preconceivedideasaboutthepossibleinteractionsthatmaycouplethedifferentlength scales associated with nuclear processes and atomic scale processes. In point of fact, althoughinmostinstancesinconventionalfusion,thesescalesremainsofarapartthat theyeffectivelydon’ttalktoeachotherbefore,duringoraftertheassociatedprocess, becausetheelectromagneticinteractiondoespenetratetoalllengthscales,itdoespro- vide a means for coupling to occur between the two sets of processes. Because the electromagnetic interaction is involved in a nonseparable way with the nuclear inter- actioninoneformofreaction(D+D (cid:6) 4–He),experimentalevidenceexiststhatshows thatthetwoformsofinteractioncanbecomecoupled. Anumberofindividuals(Schwinger[17],ChubbandChubb[2],Preparata[5,19])did recognizeatanearlystagethatthetwoformsofinteractioncouldbecoupled,provided aformofcoherenceisinvolved. SchwingerandChubbandChubbrecognizedthatthe underlying rate expression could be significantly altered as a result of this. Preparata triedtoworkwiththeexistingrateexpression(inwhichtheGamowFactorisexplicitly included)whilemodifyingtheunderlyingpotential. An important distinction evolved as a consequence. The underlying wave functions and wave function fields associated with the charged particles provided the vehicles for describing the associated processes in the theories by Schwinger and Chubb and 100 Chubb;intheworkbyPreparata,greateremphasiswasplacedonthenatureofcoher- ence through processes that emphasize the behavior of photons and not on subtleties associated with the manner in which light can couple to chargedparticles coherently. (Specifically,forexample,histheorydoesnotincludeimportanteffectsthatarepresent atlowtemperatureandmomentumassociatedwiththemannerinwhichchargedparti- cles,bythemselves,canbecoupledthrougheffectsassociatedwithparticleexchange.) Forthereasonsoutlinedinsection2.0,atlowenergies(andmomenta)theseeffectscan beveryimportant. Schwingerrecognizedthisfact. Manyofthetextbooksonstandard many-bodyphysics,whicharethebasisofknowledgeofmany-bodyphysics,formost physicists, are based on the Greens function ideas associated with statistical physics that came out of Schwinger’s work. Preparata’s approach is more closely related to formulations (associatedwith higher energies)where these kindsof subtletiesare not important. Neitherapproach, apriori, shouldbeviewedasbeingsuperiortotheother. However, thereisaveryimportantdistinctionbetweenthewave-likeformulation(usedbyChubb andChubbandbySchwinger)andtheonedevelopedbyPreparata. Thefieldoriented pictureincludestheknown,important, nonlocaleffects, discussedinSection2.0, that occurastheDeBrogliewavelengthsofalargenumberofparticlesbecomelarge;inthe pictureproposedbyPreparata,thisphysicsisabsent. Preparata used an alternative organizing principle to introduce coherence: coherent fluctuations involvingphotons withchargedmatter. Adistinguishingfeaturebetween thetwoapproachesisthatthewavepictureisguaranteedtoincludewellknowneffects (Meissnereffect,superconductivity,etc.) throughawellknownlanguage(QM/Many– bodyphysics)inthelimitofvanishingtemperature,asaconsequence. Whilethealter- native(plasma)picturesuggestedbyPreparatadidrequire(andhasrequired)thatanew languagebedeveloped.Asithasbecomeapparentthatinalargenumberofsituations, there simply are no high energyparticles, it has been clear that the kinds of forms of coherence associated with low momenta (large DeBroglie wavelengths) are probably involved. In parallel to the developments associated with theories by Chubb and Chubb, and Preparata, Hagelstein developed a series of different theories. Each of these, in one wayoranother,invokeddifferentformsofcoherence. Initially,hefeltthatimplicitly,in theevaluationofrateexpressions,incorporationofCoulombeffectsweresuchaserious impediment,thatitwasnecessarytoinvokeanewformofinteraction(involvingneutral particles[neutronhopping],forexample)tocircumventtheassociateddifficulties.Note thatinthiscontexthedidnotrely(andhasnotrelied)onaformulationinwhichtherate expression uses the Gamow factor, and thus (in common with Schwinger and Chubb andChubb),hasnotconstrainedthestrongandelectromagneticforcestobeseparable intheevaluationofrateexpressions.(TheGamowtheoryassumesseparabilitybetween electromagneticandstronginteractions.) Animportantpointisthatinhispresenttheory,he(Hagelstein)hasincludedeffectsthat 101 implicitlyinvolvecouplingthroughtheCoulombinteraction(throughphonons). This hasbroughthistheorymoreinlinewithsomeoftheideassuggestedbySchwinger,and ChubbandChubb. Animportantdifferencebetweenthepictures,however,isthatthe effectsofcoherence,asmanifestedinthelargeDeBrogliewavelengthlimitphenomena (inwhichmomentump (cid:6) 0,foralargenumberofparticles)arenotdirectlyincludedin histheory.Thus,asinPreparata’stheory,heassumestherelatedp (cid:6) 0,coherenteffects (suchassuperconductivity),associatedwithT (cid:6) 0,arenotrelevant.Also,atthepresent time,histheorydoesnotincorporateboundaryeffectsorfinitecrystalsizeeffects. From an early stage, the focus of Yeong Kim’s work has been to develop a multi– nucleon theory that goes beyond the Gamow-like rate expression of conventional fu- sion. More recently, in examining problems that are involved in optically trapped (bosonic) atoms, it occurred to him that similar kinds of ideas could be used in the deuteron fusion problem in condensed matter. Because this framework is associated withcoherencethrougheffectsthatbecomeimportantatlargeDeBrogliewavelengths, Kimdoesdirectlyusethekindsofp (cid:6) 0effectsthatChubbandChubbinclude,which areomittedbyHagelsteinandPreparataandDelGuidice. 5.0CommonFeaturesofDevelopedTheories. Recently,asomewhatsurprisingdevelopmentoccurred.Three[2-4]ofthesefour[2-5] theoriesadoptedsimilar(many–bodyphysics)formulationsinwhichcoherencefollows eitherfromaparticularformofinteraction(asin[4])orfromacombinationoffactors involvingpossibleformsofmany-bodyinteraction,andparticleindistinguishability(as in[2,3]). Asaresult,plausibleexplanationsarebeginningtoemergeforanumberofimportant phenomena. Specifically, consistent with the idea that for a theory to be believableit shouldrelatetoanexistingtheory(asoutlinedinSection3.0),agreementbetweenthe theories appears to reflect: (i) use of a sufficiently sophisticated, and universally ac- ceptedformofmathematicsthatexplainshownuclearscaleandatomicscaleprocesses canberelatedtoeachotherwithouthighmomentumparticlesbeingreleased;(ii)useof reactionrateexpressionsthatincludecoherent,nonlocaltransferofmomentum,involv- ingmanyparticles; and(iii) relianceupon aformulation thatincludesa largenumber ofcharged,indistinguishableparticles,expressedintermsofastandard,common,well acceptedconcept:themany-bodywavefunction,associatedwiththeQMofcondensed matter physics. Chubb and Chubb have done this by explicitly illustrating how their atomicscaleIonBandStatetheorycanbegeneralizedtoincorporatenuclearscalepro- cesses through a generalization of standard multiple scattering theory techniques [21, 22]. In the process, they explain how a non-separable coupling between nuclear di- mension andatomic dimensionscale canoccur in the wave functions associatedwith nucleus/nucleusseparation,inanon–localfashion. Byadoptingan explicit formof representation for distinguishingbetweenthe coordi- nate dependencies involving the short ranged (nuclear) degrees of freedom and those thatcoupletotheelectromagnetic(longerranged)force,Hagelstein[3]hasdeveloped 102 a similar ideathat generalizes theresonant groupstructure ideaproposed byWheeler [23]inthe1930s. Infact,astrikingsimilarity,forthecaseassociatedwithD-Dfusion, occurs in Hagelstein’s representation of the relevant wave function and the compara- blechoiceusedbyChubbandChubb,onceitisrecognizedthatthecorrelationfactors g(r1,r2)thatareusedbyChubbandChubbtodescribethedependenceontheseparation variable (r1-r2) between deuterons located at r1 and r2, are equivalentto the channel factorsFj(inwhichthesubscript“j”referstor1-r2)definedbyHagelstein.Fromthese observations,threedistinguishingfeaturesbetweenthesetwo[2,4]theoriesfollow: 1. ChubbandChubbpointoutthatapproximatelydiscontinuouschangesinthegradient ofg(r1,r2)(thatarenotincludedexplicitlyincludedinHagelstein’schannelfactors,Fj) illustratethepossiblecoupling(throughtheassociatedmany-bodyproblem)thatallows for transferofmomentumto occurnon-locally,eveninthe T (cid:6) 0limit, 2. Chubband Chubbillustrate,explicitly,therelationshipbetweenthesediscontinuitiesandcoherent (latticerecoileffects)inwhichmomentumcanbetransferredfromaparticularlocation, to many locations, instantaneously; and 3. although Hagelstein does not particularly identifythispossibility,hepointsoutthatmanynucleonscanbecomecoupledtogether, simplyasaconsequenceoftheexistenceoftheassociatedrelationship.Hagelstein,fur- ther,explorestheimplicationsofthiscoupling,explicitly,throughcoherentmomentum transfer between nucleons to and from a coherent (or nearly coherent) set of optical phonons. He also uses the associated ideas to provide a possible explanation for the emissionofhighmomentaparticlesfromdeuteratedTifilms. Kim [4] also adopted a picture, based on many-body physics. Beginning from the common starting point [2-4] (involving the complete many-body wave function), he has drawnthiskind of connection byincorporating an approximateform for a poten- tialmany-bodyinteractioninvolvingbosons(borrowedfromhisopticalatomtrapping theory)so that itcould beused inthe Cold Fusionproblem. An intriguing difference between Kim’s approach and the approaches followed by Hagelstein and Chubb and Chubb is that he does not use the idea that deuterons (or other nuclei) are interacting with a well defined lattice. In place of this idea, he starts from the assumption that undersuitableconditions,acollectionofdeuterons,interactingwithasolid,mightbe- comeeffectivelytrappedinamannerthatresemblestheopticaltrappingofalkaliatoms thatformsthebasisofatomLasers,andtherelatedformsofneutralatomBoseEinstein Condensates. Although,superficially,thisideamightseemsomewhatforeign,inthelimitofperfect periodicorder,atsufficientlylowtemperature,theimplicationsofthisidea,andthose associatedwiththeionbandstatetheoryproposedbyChubbandChubbbecomeiden- tical. (Specifically, the D (cid:0) ions that occupyion band states in the Chubb and Chubb theoryforma BoseEinsteincondensate,atvanishingtemperature.) Apotentiallyim- portant new idea, associated with Kim’s work [4, 24] that he has applied to the CF problem, is the development of an effective two body Hamiltonian, from the exact many-body system, which can be used to determine a separable form for the ground state wave function of a many-body Bose system. Since the associated Hamiltonian is robust, the resulting expression for the wave function, may be applicable in many differentsituations. AsecondintriguingpointisthatbothKimandChubbandChubb, 103 independently,haveconcludedthatundersuitablecircumstances,coherenceassociated withthemechanismthatisresponsibleforperfectBoseEinsteincondensation(associ- ated withthe limitin whichmanyparticles, coherently,approacha state inwhich the momentum of each particle approaches zero) provides a potentially important source for coherent, nonlocalmomentum transfer that can be usedto accountfor the lack of highmomentumparticlesinCFreactions. Insummary,allthreetheoriesarenowbasedonanorganizingprincipleassociatedwith condensedmatterphysicsthatexplicitlyincludesaprocedureforincorporatingnuclear effects. And, although initially, two of the theories (Kim and Hagelstein) focused on nuclearscalephenomena,whilethethird(ChubbandChubb)focusedonatomicscale processes,allthreenowincludeeffectsthatcoupleatomicandnuclearscaleprocesses inamannerthatisconsistentwiththecriteria,outlinedinSection3.0. Althoughthefinerdetailsassociatedwiththeoreticalframeworksthathavebeenusedto coupletheverydifferentlengthscalesaredifferent,ineachcase,coherencethatresults through couplingto the electromagnetic field, providesthe dominantform ofinterac- tion. The associated coupling is expressed in most general terms, using the multiple scattering theory [21, 22], discussed by Chubb and Chubb. In particular, within this framework,anexactrateexpressionisderivedthatrelatesallpossiblemany-bodycol- lisions associated with a particular reaction to discontinuous changes in momentum. Thisexpressionillustratesexplicitlyhownon-localmomentumtransfercanoccur,co- herently,instantaneously,insuchawaythatitbecomesobvioushowhighmomentum particles (through the accumulation of large amounts of momentum at isolated loca- tions)canbeavoided. Thetheoryalso canbereadilygeneralizedtoincorporatearbi- trary forms of interaction. Within the context of this theory [2], it is also possible to recoverallofthepreviousresultsoftheChubbandChubbtheory[25]. IncontrasttotheworkbyChubbandChubb,whichhasfocusedprimarilyonD+Dfu- sion,andissuesassociatedwithnon-localmomentumtransfer,Hagelstein,ontheother hand,hasattemptedtodealwithamoregeneralsetofnuclearreactions.Forthisreason, whileChubbandChubbhavebeenconcernedprimarilywithnon-localformsofinter- action,andquestionsrelatedtomaterialsproperties,andsolidstateeffects,Hagelstein hasfocusedmorecloselyonquestionsassociatedwiththereleaseofhighenergyparti- cles,latticeimperfections,andtheinfluencesofinjectingnucleiintothehost.Fromthis startingpoint,hehasidentifiedanumberofpotential,triggeringphenomenaassociated withtheemergenceoffastparticlesandunconventionalnuclearbyproducts. Althoughthe recentfocusofKim’swork hasbeenoncoherentdeuteronfusion, from Bose condensedstates, he has also investigatedthe possibility of novel,nuclear reac- tions. Animportantpointabouthismostrecentworkisassociatedwiththeeffectsof finitesize. Inparticular,hepredictsoptimalreactionrates,withinaparticularscenario, basedonestimatesofparametersthatheinfersfromexperiments. Table1listssomeoftheorganizingprinciplesandcommonfeaturesofthethreetheories describedinthissection,aswellasthecomparablefeaturesassociatedwithPreparata’s 104 theory. In this table, under the label, organizing principle, by coherence, and the la- bels, Low P and High P, refer to the question of whether or not the theory applies in thelowmomentum(P)(LowP)limit, associatedwithlowtemperature,orinthehigh momentum (P) (High P) limit, or in both limits. Source refers to the effective form ofinteraction(orintermediateparticle)thatisresponsibleforthecoherence. Thedes- ignationNuc/EMSeparabilityinRateExpressionreferstothequestionofwhetheror notseparabilityinthecoordinatedependenciesbetweennuclear(Nuc)andelectromag- netic (EM) interactions is assumedin the associated wave functionsand reaction rate expression,andtoidentifytheorganizingprinciple(forexample,many-bodyphysics) that treats thecouplingbetweennuclearandelectromagnetic interactions. Thestarin thefinalcolumnreferstotheideathatrecoilmomentum(associatedwiththepossible nuclearreaction)canbeincorporateddirectlythroughnon-localtransferofmomentum fromthebulktothesurfaceregion. Table1: Organizingprinciples/ideasinColdFusiontheories Theory Coherence Nuc/EM Rate LowP HighP Source Separability Expression Chubb & Yes Yes E.M. No Many-Body Chubb Interaction (all Physics of it) / Particle Statistics I Hagelstein No Yes Phonons No Many-Body Physics Preparata No Yes “Photons” Yes Semi-Classical (Gamow) I ? Kim Yes Yes E.M. No Many-body Interaction (all Physics of it) / Particle Statistics I 6.0SuggestionsforTestingTheories/FutureTheoreticalWork. Anumberofpredictionshavecomeoutofeachofthethreetheories[2-4]discussedin thelastsection.Animportantpointtokeepinmindisthat,implicitly,thesepredictions, in most cases, seem to apply most rigorously within the contextof a particular set of circumstances. Forexample,ChubbandChubbsuggestedmanyyearsago[25]thata bosons-inandbosons-outruleshouldapply,providedparticularconditionsaremet. In thecontextofthemultiplescatteringtheorypresentedin[2],thelimitationsofthisrule wereidentified[26]: theruleappliesrigorouslyinthelowtemperaturelimit,provided suitablylarge, orderedcrystals are used. Similarly, Hagelsteinhas identifiedlimitsin which optical phonons, vacancies, and other effects can significantly enhance fusion (orothernuclearprocess)rate. Apotentiallyimportantpointisthatthewelldeveloped 105 theories [2–4] have evolved to the point that they include well defined mathematical expressions that relate particular reaction rates, rigorously, to known situations that apply under specific conditions. The significance of this point is that in each case where a reaction rate expression has been derived, given the uncertainties of existing experiments, it is probably important to attempt to match a particular experimental studysothatitmimicstheconditionsassociatedwiththeparticularexpression. Forexample,ChubbandChubbpredictedoptimalcrystalsizes(withcharacteristiclin- eardimensionsofapproximately0.1µm3)forproducinglargeamountsofexcessheatat elevatedtemperatures. Themotivatingargumentassociatedwiththisestimatewasthe requirementthatcrystallineorderbemaintainedandtheheliumbyproductbeexpelled. Atreducedtemperatures(forexample,below200degreesK),theargumentsassociated with thesepredictionsdonot hold. Thisis becauseasthetemperatureisreducedand crystallineorderisincreased,otherfactorsassociatedwithsurfacepreparation,periodic order,andloading,becomemoreimportant. Theimportantpointisthatundercertaincircumstances,onetheoreticalpredictionprob- ably willbe more usefulthan another. Confirmationof this point is important, in my opinion, because if it is confirmed that different theories are valid in different situa- tions, the speculationthat a singlemechanismis responsiblefor allCF phenomenais probablywithoutmerit. Withthispointinmind,itseemsappropriatetoemphasizean obviouslessonthathascomeoutofthelastdecadeoftheoreticalwork:inorderforthe- oriststoformulateameaningfultheory,theyrequiredetailedinformationaboutmaterial preparationandrelatedfactors(crystallinequality,andsize,aswellastemperature,and loading, for example). In the future, theorists certainly would benefit from measure- mentsassociatedwiththesekindsoffactors,aswellasthroughadditionalinformation documentingcorrelationsbetweenthesekindsofvariablesandpotentialtriggeringphe- nomena. Finally,itisusefultoidentifyanumberofLessonsLearnedassociatedwiththeinter- playbetweentheoryandexperiment. Inthiscontext,wewouldliketonoteinpassing threeapparentsuccessstoriesthathaveoccurred:(i)IndependentpredictionsbyChubb andChubb(C&C)andPreparata(P)thathighloadingisbeneficialindeuteriumfusion inPd/D(observedinvariousplaces),(ii)comparablepredictionsbyC&CandPthatHe- 4shouldbethepredominantbyproduct(observedbyMilesetal.,andBushetal.),and (iii)thesuggestionbyBhaktaRath(basedontheideaputforthbyChubbandChubb thatsmallcrystalsinaporousmediumcouldbebeneficial)thatPd/Bcouldpotentially provideausefulcompositematerialforproducingexcessheat(whichhasbeenverified byImamandMiles,asdocumentedinthisreport,cfchapter3). Animportantlesson, to date, however, is that although theory predicted these successes prior to the actual experiments,theory hasbeenlargelyignored. In particular,only in thethirdexample (involving PdB) was theory actually used to guide the associated experiment. Given theapparentlackofconsensusabouttheorythathasexistedinthepast, itisplausible thatadegreeofskepticismabouttheoreticalpredictionshasbeenwarranted. However, the situation has evolved considerably since the initial days of CF. In the future, one wouldhopethatexperimenterswouldmorecloselymonitorandtestthepredictionsof 106 themorematuretheories. There is a final, more general lesson, associated with the manner in which CF has been judged by the scientific community and with the potential role of theory in this process. In particular, an extremely naive, overly simplified picture of the relevant physical situation was adopted by most physicists, based on an idea (associated with conventional nuclear physics) that was doomed to fail, from the outset. When most experimentersfailedtoreproducetheeffect,thisextremelynaivepicturenotonlywas usedasformaljustificationforassumingthephenomenondidnotexist,buttopreclude thenotionthatadifferent,moresophisticatedtheorymightbemoreappropriate.Onthe otherhand,althoughmostphysicistsfollowedthisroute,afewmorecreativephysicists thoughtofalternativeideas,whichnotonlywerediscardedbymostphysicists,but,in somecases,wereoutwardlyscorned. Itseemsappropriate,giventhefactthatnotonly didthefieldnotdie, butthatviabletheoreticalexplanationsfor whatisinvolvedhave evolved, to ask a fundamental question about the impact of naive skepticism on the adjudication process. In particular, it is clear that in some cases, creative ideas about CFwerestifledtosuchadegreethatnonexperts[7,10,27]notonlyhavebeenallowed toopenlyridiculeandattackthem,buttodosowithoutallowingtheresponsibleparties torespond. Rhetorically,onemightask,Isn’titalwaysusefultolookatacreativeidea with an open mind, especially when an undercurrent of skepticism is present? More poignantly,onemightask,Howcansimilarfailuresbeavoidedinthefuture? At this time, it is clear that creative theories, based on mainstream thinking do exist. Inparticular,threehavebeenidentifiedthatshouldbetestedandapplied. Despitethe factthatthesetheorieshavebeendeveloped,onlylimitedworkinthisareaisgoingon atthepresenttime. Thissituationmustandshouldchange. Hopefully,thisarticlewill haveapositiveimpactinchangingthissituation. References. 1. T.A.ChubbandS.R.Chubb,FusionTechnology,17,710(1990) 2.S.R.ChubbandT.A.Chubb,TheoreticalFrameworkforAnomalousHeatand4–He inTransitionMetalSystems,Proc. ICCF8. 3.Y. E. Kim andA. L. Zubarev,Ultra Low–EnergyNuclear Fusion of Bose Nuclei in Nano–ScaleIonTraps,Proc. ICCF8. 4. P.L.Hagelstein,AUnifiedModelforAnomaliesinMetalDeuterides,Proc. ICCF8. 5. G. Preparata, Trans Fusion Technol., 26, 397 (1994); QED Coherence in Matter, WorldScientificPublishingCo.,Singapore;chap. 8,pp. 153–178(1995). 6.T.S.Kuhn,TheStructureofScientificRevolutions,Univ.ofChicagoPress,Chicago, (1962). 7.R.L.Park,VoodooScience:TheRoadfromFoolishnesstoFraud,OxfordUniversity Press,Oxford,2000. 8. R.P.Feynman,R.B.LeightonandM.Sands,TheFeynmanLecturesonPhysics,Ad- disonWesleyPublishing,Inc.,NewYork,1965. 9. S.R.Chubb,AccountabilityinResearch,8,(2000). (http://www.gbhapus.com/journals/149/149-top.htm).C.Beaudette,ExcessHeat:Why 107 ColdFusionResearchPrevailed,OakGrovePress,LLC,ME,2000.availablethrough http://www.infinite-energy.com 10. D.R.O.Morrison,PhysicsToday50,14(1997). 11. J.Rafelski,M.Sawicki,M.GajdaandD.Harley,FusionTechnol.,18,136(1990). 12. E. Teller, in Proc EPRI-NSF Workshop, Ed. Schneider, Electric Power Research Institute,PaloAlto,CA,pp.1-2(1989). 13. F.J.MayerandJ.R.Reitz,FusionTechnol.19,552(1991);ibid. 20,367(1991). 14. J.P.Vigier,Proc. ICCF3. 15.R.Mills,TheGrandUnifiedTheoryofClassicalQuantumMechanics,BlackLight- Power,Inc.,Princeton,N.J. 16. H. Kozima, Discovery of the Cold Fusion Phenomenon: Development of Solid StateNuclearPhysicsandtheEnergyCrisisinthe21stCentury,OhtakeShuppanInc., Tokyo,1998. 17. J.Schwinger,J.,NuclearEnergyinanAtomicLattice,inProc. FirstAnnualCon- ferenceonColdFusion(ed.Will,(ICCF–1)SaltLakeCity,UT,1990);Z.Phys.D:At., Mol. Clusters,15,221(1990); Z.Naturforsch.,45A,756(1990);Prog. Theor. Phys., 85,711(1991);SpringerProc. inPhysics57(EvolutionaryTrendsinthePhysicalSci- ences;Eds: M.Suzuki,R.Kubo,1991),171. 18. M.Rabinowitz,Y.E.Kim, V.A.ChechinandV.A. Tsarev,TransFusionTechnol., 26,3(1994). 19. G.Preparata,FusionTechnology,20,82(1991). 20. E.Storms,InfiniteEnergy,6,54(2000). 21. A. Gonis and W.H. Butler, Multiple Scattering in Solids. Springer Verlag, N.Y. 2000). 22. J.Korringa, Physica, 13,392(1947);. W. KohnandN.Rostoker,Phys. Rev.,94, 1111(1954). 23. J.A.Wheeler,Phys. Rev.,52,1107(1937). 24. Y.E.KimandA.L.Zubarev,J.Phys.B:At. Mol. Opt. Phys.,33,1(2000). 25. S.R. Chubb and T.A. Chubb, AIP Conference Proc. 228, 691 (1991) (eds. S. E. Jones, F. Scaramuzzi and D. Worledge), Amer. Inst. Phys. (New York); T.A. Chubb and S.R. Chubb , Fusion Technology, 20, 93, (1991); S.R. Chubb and T. A. Chubb, DistributedBosonicStatesandCondensedMatterFusion,NRLMemorandumReport 6600 (Documents, Code 2627, Naval Research Laboratory, Washington, DC, 20375- 5321,1990);T.A.ChubbandS.R.Chubb,NuclearFusioninaSolidviaaBoseBloch Condensate,NRLMemorandumReport6617(Documents,Code2627,NavalResearch Laboratory,Washington,DC,20375-5321,1990). 26.S.R.ChubbandT.A.Chubb,TheoreticalFrameworkforAnomalousHeatWithout HighEnergyParticlesfromDeuteronFusioninDeuterium/TransitionMetalSystems, talkpresentedatWinter2000AmericanNuclearSocietymeeting,8Nov. 2000,Wash- ington, DC.; S. R. Chubb and T. A. Chubb, 2000 Trans. Amer. Nuc. Soc., 83, 362 (2000). 27. D.Lindley,Nature,344,375(1990);Minutesofthe189thMeetingofWashington PhilosophicalSociety,(WashingtonPhilosophicalSociety,Washington,D.C.),31Jan. 1992. 108 109 APPENDIX:LISTINGOFPUBLICATIONS/PRESENTATIONS RELATEDTO COLDFUSION ThislistingislimitedtothecontributionsfromtheU.S.NavyLaboratories. Itcontains publications, theoretical and experimental, covering topics directed toward better un- derstanding of the Fleischmann–Pons effect and reflecting scientific interests of their authors. ContributionsfromtheSpaceandNavalWarfareSystemsCenter,SanDiego,San Diego, CA 92152–5001(formerly: NavalCommand, Controland OceanSurveil- lanceCenter,RDT&EDiv.,SanDiego,CA) Journalpublications 1. S.Szpak,P.A.Mosier–BossandJ.J.Smith,OnthebehaviorofPddepositedinthe presenceofevolvingdeuterium,J.Electroanal.Chem.,302,255(1991) 2. S.Szpak,C.J.Gabriel,J.J.SmithandR.J.Nowak,ElectrochemicalchargingofPd rods,J.Electroanal. Chem.,309,273(1991) 3. S.Szpak,P.A.Mosier–Boss,S.R.ScharberandJ.J.Smith,ChargingofthePd/nH system:roleoftheinterphase,J.Electroanal. Chem.,337,147(1992) 4. S.Szpak,P.A.Mosier–Boss,C.J.GabrielandJ.J.Smith,Absorptionofdeuterium inpalladiumrods: modelvs. experiment,J.Electroanal. Chem.,365,275(1994) 5. S.Szpak,P.A.Mosier–Boss,R.D.BossandJ.J.Smith,Commentsontheanalysis oftritiumcontentinelectrochemicalcells,J.Electroanal. Chem.,373,1(1994) 6. S.Szpak,P.A.Mosier–BossandJ.J.Smith,DeuteriumuptakeduringPd–Dcodepo- sition,J,Electroanal.Chem.,379,121(1994) 7. S.Szpak,P.A.Mosier–Boss,S.R.ScharberandJ.J.Smith,Cyclicvoltammetryof Pd+Dcodeposition,J.Electroanal.Chem.,380,1(1995) 8. S. Szpak, P. A. Mosier–Boss and J. J. Smith, On the behavior of the cathodically polarizedPd/Dsystem:Searchfortheemanatingradiation,PhysicsLettersA,210,382 (1996) 110 9. S.SzpakandP.A.Mosier–Boss,OnthebehaviorofthecathodicallypolarizedPd/D system:aresponsetoVigier’scomments,PhysicsLettersA,221,141(1996) 10.S.Szpak,P.A.Mosier–Boss,R.D.BossandJ.J.Smith,OnthebehaviorofthePd/D system:Evidencefortritiumproduction,FusionTechnology,33,38(1998) 11. S.SzpakandP.A.Mosier–Boss,OnthereleaseofnHfromcathodicallypolarized 1 palladiumelectrodes,FusionTechnology,34,273(1998) 12. P.A, Mosier–BossandS.Szpak,ThePd/nHsystem: Transportprocessesandde- velopmentofinstabilities,IlNuovoCimento,112A,577(1999) ProceedingsofICCF 1. S.Szpak,P.A.Mosier–BossandJ.J.Smith,Reliableprocedurefortheinitiationof theFleischmann–Ponseffect,Proc.ICCF–2(1991) 2. S. Szpak, P. A. Mosier–Boss, Comments on the Methodology of Excess Tritium Determination,Proc. ICCF–3(1992) ContributionsfromtheNavalAirWarfareCenterWeaponsDivision,ChinaLake, CA Journalpublications 1. M. H. Miles and R. E. Miles, Theoretical Neutron Flux Levels, Dose Rates, and MetalFoilActivation,J.Electroanal. Chem.,295,409(1990) 2. M.H.Miles,K.H.ParkandD.E.Stilwell,ElectrochemicalCalorimetricEvidence forColdFusioninthePalladium–DeuteriumSystem,J.Electroanal. Chem.,296,409 (1990) 3.B.F.Bush,J.J.Lagowski,M.H.MilesandG.S.Ostrom,HeliumProductionDuring theElectrolysisofD OinColdFusionExperiments,J.Electroanal. Chem.,304,271 2 (1991) 4. M.H.Miles,R.A.Hollins,B.F.Bush,J.J.LagowskiandR.E.Miles,Correlation ofExcessEnthalpyandHeliumProductionDuringD OandH OElectrolysis,J.Elec- 2 2 troanal. Chem.,346,99(1993) 5. M.H.Miles,B.F.BushandD.E.Stilwell,CalorimetricPrinciplesandProblemsin MeasurementsofExcessPowerDuringPd–D OElectrolysis,98,1948(1994) 2 6. M. H. Miles, B. F. Bush and J. J. Lagowski, Anomalous Effects InvolvingExcess Power,RadiationandHeliumProductionUsingPalladiumCathodes,FusionTechnol- ogy,25,478(12994) 7. M. H. Miles andB. F. Bush, Heat andHelium Measurements in Deuterated Palla- dium,Trans. FusionTechnology26,156(1994) 8. M.H.Miles,Replyto“ExaminationofClaimsofMilesetal. inPons–Fleischmann TypeColdFusionExperiments,J.Phys.Chem. B102,3642(1998) 9. M. H. Miles, Calorimetric Studies of Pd/D O + LiOD Electrolysis Cells, J. Elec- 2 troanal. Chem.,482,56(2000) 10.M.H.MilesandK.B.Johnson,ElectrochemicalInsertionofHydrogenintoMetals andAlloys,J.NewEnergy,inpress 111 ProceedingsofICCF 1. M.H.Miles,K.H.ParkandD.E.Stilwell,ElectrochemicalCalorimetricStudiesof theColdFusionEffect,Proc.ICCF–1(1990) 2. M.H.Miles,G.S.Ostrom,B.F.BushandJ.J.Lagowski,HeatandHeliumProduc- tioninColdFusionExperiments,Proc.ICCF–2(1991) 3. M.H.MilesandB.F.Bush,SearchforAnomalousEffectsInvolvingExcessPower, HeliumandTritiumDuringD OElectrolysesUsingPalladiumCathodes,Proc. ICCF 2 –3(1993) 4. M.H.MilesandB.F.Bush,CalorimetricPrinciplesandProblemsinD OElectrol- 2 ysis,Proc.ICCF–3(1993) 5. M. H.Miles, TheExtractionofInformationfrom anIntegratingOpenCalorimeter inFleischmann–PonsEffectExperiments,Proc. ICCF–5(1995) 6. M. H. Miles and B. F. Bush, Radiation Measurements at China Lake: Real ODR Artifacts,Proc. ICCF–7(1998) 7. M.H.Miles,CalorimetricStudiesofPalladiumCathodesUsingFleischmann–Pons DewarTypeCells,Proc. ICCF–8(2000) ContributionsfromtheNavalResearchLaboratory,Washington,DC. Journalpublications 1. T.A.ChubbandS.R.Chubb,Bloch—SymmetricFusioninPdD ,FusionTechnol- x ogy,17,710(1990) 2. T. A. Chubb and S. R. Chubb, Cold Fusion as an Interaction between Ion Band States,FusionTechnology,20,93(1991) 3. S.R.ChubbandT.A.Chubb,IonBandStateFusion:Reactions,PowerDensityand theQuantumRealityQuestion,FusionTechnology,24,403(1993) 4.S.R.ChubbandT.A.Chubb,TheRoleofHydrogenIonBandStatesinColdFusion, 26,414(1994) 5. S.R.ChubbandT.A.Chubb,TheoreticalFrameworkforAnomalousHeatwithout HighEnergyParticlesfromDeuteronFusioninDeuterium–TransitionMetalSystems, Trans. Am. Nuc.Soc. 83,362(2000) 6. S.R.Chubb,Reviewof“ExcessHeat: WhyColdFusionResearchPrevailed”,Fu- sionTechnology,39,288(2001) 7. P.L.Hagans,D.D.DominguezandM.A.Imam,SurfaceCompositionofPdCath- odes,ProgressinNewHydrogenEnergy,vol. 1p. 249(1996) ProceedingsofICCF 1. S.R.ChubbandT.A.Chubb,QuantumMechanicsof“Cold”and“Not–So–Cold” 112 Fusion,Proc.ICCF–1,119(1990) 2. S.R.ChubbandT.A.Chubb,AnExplanationofColdFusionandColdFusionBy– Products,BasedonLattice-InducedNuclearChemistry,Proc. ICCF–2(1991) 3. S.R.ChubbandT.A.Chubb,IonBandStateFusion,Proc.ICCF–3(1992) 4. T.A.ChubbandS.R.Chubb,TheIonBandStateTheory,Proc.ICCF–5(1995) 5. S.R.ChubbandT.A.Chubb, HiddenResultsoftheIonBandStateTheory,Proc. ICCF–6,Proc.ICCF–6(1996) 6. T.A.ChubbandS.R.Chubb, RadiationlessColdFusion: WhySmallCrystalsare Better,N RequirementandEnergyTransfertoLattice,Proc.ICCF–6(1996) cell 7. S. R. Chubb and T. A. Chubb, Periodic Order, Symmetry and Coherence in Cold Fusion,Proc.ICCF–7(1998) 8. S.R.ChubbandT.A.Chubb,ReallyCold,ColdFusion,Proc. ICCF–7(1998) 9. T. A. Chubb and S. R. Chubb, Deuteride Induced Strong Force Reactions, Proc. ICCF–7(1998) 113 Initialdistributionlist SpawarSystemsCenterSanDiego,SanDiego Dr. FrankE.Gordon,Code03 Dr. RandallH.Moore,Code Dr. PamelaA.Boss(10),Code PatentCounsel,Code0012 Library,Code NavalAirWarfareCenter,ChinaLake Dr. MelvinH.Miles(10).Code4T4220D Dr. RobinA.Nissan,Code4T4200D Dr. GeoffreyA.Lindsay,Code4T42200D Dr. JeffreyJ.Davis,Code4T4330D OfficeofGeneralCounsel,PatentOffice,Code772000D TechnicalLibrary,Code4TL000D NavalResearchLaboratory,WashingtonDC Dr. TimothyCoffey,Code1001 Dr. BhaktaRath,Code6000 Dr. ScottR.Chubb(10),Code7252 Dr. AshrafImam,Code6320 OfficeofNavalResearch Dr. FredE.Saalfeld,TechnicalDirector DefenseTechnicalInformationCenter Alexandria,VA22304–6145 NavyAcquisition,ResearchandDevelopmentInformationCenter(NARDIC) Arlington,VA22244–5114 114 ProfessorJ.O’M.Bockris 4973AftonOaksDrive CollegeStation,TX77845 Dr. TalbotA.Chubb ResearchSystems,Inc. 5023North38–thStreet Arlington,VA22207 ProfessorJohnDash PhysicsDepartment PortlandStateUniversity P.O.Box751 Portland,OR972078–0751 ProfessorPeterL.Hagelstein MassachusettsInstituteofTechnology Dept. ofElectricalEngineeringandComputerScience Room36–225 77MassachusettsAve Cambridge,MA02139 Dr. EugeneMallove NewEnergyResearchLaboratory P.O.Box2816 Concord,NH03302–2816 Dr. MichaelMcKubre SRIInternational 333RavenswoodAve MenloPark,CA94025 Dr. MichaelMelich NavalPostgraduateSchool 1224MeigsDrive Niceville,FL32578–3018 ProfessorGeorgeH.Miley FusionStudiesLaboratory UniversityofIllinois 103S.GoodwinAve Urbana,IL61801 Dr. DavidJ.Nagel TheGeorgeWashingtonUniversity 2933KStreet,Suite340J Washington,DC20052 115 Dr. RobertJ.Nowak DARPA 3701N.FairfaxDrive Arlington,VA22203 ProfessorRichardA.Oriani 112AmundsonHall UniversityofMinnesota 421WashingtonAveSE Minneapolis,MN55455 Dr. JerryJ.Smith U.S.DepartmentofEnergy CodeSC–13/GTN 19901GermantownRoad Germantown,MD20874–1290 ProfessorLouisD.Smullin MassachusettsInstituteofTechnology Cambridge,MA02139 Dr. LowellWood HooverInstitutionHT–1004 StanfordUniversity Stanford,CA94305–6010 116