JournalofPowerSources239(2013)572e583
ContentslistsavailableatSciVerseScienceDirectJournalofPowerSourcesjournalhomepage:www.elsevier.com/locate/jpowsourCycleandcalendarlifestudyofagraphitejLiNi1/3Mn1/3Co1/3O2Li-ionhighenergysystem.PartA:FullcellcharacterizationStefanK?bitza,c,*,JochenBernhardGerschlera,c,MadeleineEckera,c,YusufYurdagela,c,BritaEmmermacherd,DaveAndrée,TimMitsche,DirkUweSauera,b,caChairforElectrochemicalEnergyConversionandStorageSystems,InstituteforPowerElectronicsandElectricalDrives(ISEA),RWTHAachenUniversity,Jaegerstrasse17-19,52066Aachen,GermanybInstituteforPowerGenerationandStorageSystems(PGS)@E.ONERC,RWTHAachenUniversity,GermanycJuelichAachenResearchAlliance,JARA-Energy,GermanydZSWeZentrumfürSonnenenergie-undWasserstoff-ForschungBaden-Wuerttemberg,Helmholtzstrasse8,89081Ulm,GermanyeDeutscheACCUmotiveGmbH&Co.KG,NeueStrasse95,73230Kirchheimu.Teck(Nabern),Germanyhighlights??Extendedacceleratedagingtestsonlithium-ionbatteriesincludingstorageandcycling.??Detailedanalysisoftemperatureandvoltagedependenciesofcellaging.??Correlationofagingresultstopost-mortemanalysis.??Lifetimepredictionbasedonidenti?edmainagingphenomena.??AlternatingagingtestundervaryingSOC.articleinfoArticlehistory:Received24September2012Receivedinrevisedform12January2013Accepted12March2013Availableonline26March2013Keywords:CalendarlifeCyclelifeLithium-ionAgingNickelcobaltmanganeseLifetimepredictionabstractThisworkprovidesanagingstudyofagraphitejLiNi1/3Mn1/3Co1/3O2(NMC)Li-ionpouchcellwithanominalcapacityof10Ah.Bymeansofresistanceandcapacitymeasurementsthecell’sbehavioristrackedovertimeunderconsiderationoftemperatureandcellvoltageimpact.Testsdurationwasupto15monthseffectivetestingtime.Observedeffectsandpossibleagingmechanismsarediscussedconsideringtheresultsfromcapacityandresistancemeasurements.Thetestresultsareusedforacalendarlifetimeprediction.Inadditiontoadetailedcalendarlifestudy,alsocyclelifetestsarediscussedbrie?ytopointoutadditionalagingeffectsbasedoncycling.Thepapermayalsoserveasdatasourceonagingforfurtherworkonbatterylifetimemodelingandbatterydiagnostics.Selectedcellsweretakenfromtheagingtestsandwereusedforadetailedpost-mortemanalysis(seepaper“PartB:Post-MortemAnalysis”).ó2013ElsevierB.V.Allrightsreserved.1.IntroductionForareliableintegrationofbatteriesintoapplications,knowl-edgeaboutbatteryagingbehaviorandespeciallytheachievablelifetimeofthebatteryisindispensable.Thein?uenceofdifferent*Correspondingauthor.ChairforElectrochemicalEnergyConversionandStorageSystems,InstituteforPowerElectronicsandElectricalDrives(ISEA),RWTHAachenUniversity,Jaegerstrasse17-19,52066Aachen,Germany.Tel.:t492418099600;fax:t492418092203.E-mailaddresses:batteries@isea.rwth-aachen.de,kb@isea.rwth-aachen.de(S.K?bitz).0378-7753/$eseefrontmatteró2013ElsevierB.V.Allrightsreserved.http://dx.doi.org/10.1016/j.jpowsour.2013.03.045operatingconditionsonoccurringagingphenomenahastobewellunderstood,e.g.forthedevelopmentofoperatingstrategies,designofbatterysystemsandalsobusinessmodels.Sinceextendedbat-terytestsunderrealoperatingconditionsarequitetimeconsumingandthereforecostly,acceleratedagingtestsareconducted.Ifin-formationaboutoccurringagingphenomenaisdesired,typicallyapost-mortemanalysishastobeconducted[1].Thelackofpubli-cationsonextensiveagingtestsoflithium-ionbatterieshasbeenreportedasahindranceforthemodelingoflithium-ioncellagingphenomena[2].ItisthepurposeofthisworktoprovidegeneralinformationabouttheagingbehaviorofagraphitejNMCS.K?bitzetal./JournalofPowerSources239(2013)572e583573
cellforfurthermodelingstudiesandapplicationse.g.inbatterymanagementsystems.AlsomainfeaturesofatestmethodologyforacceleratedagingtestingarepresentedandamathematicalbaseforlifetimeprognosisandlifetimemodelingisprovidedbasedonRef.[3].Forseveralmaterialsextendedagingtestswereperformedinpreviousworks,e.g.forcellswithdifferentcathodematerialsvs.graphite,likeLiCoO2[4,5],NCA[6],NCO[7,8]andLFP[9,10].Onlyfewpublicationsexistaboutextensiveagingtestswithlithium-ionbatteriesandtestmatrixsizescomparabletoRefs.[9,10].ThoughalsotheNMCcathodeisagoodcandidatefore.g.electricvehicle(EV)applicationsandisprovidedbymanycellmanufacturersforthispurpose,onlyfewagingstudiesconcerningthismaterialcanbefoundinliterature.TheworkofBloometal.[11],analyzescycleandcalendarlifeperformanceofNMCatdifferenttemperaturesand60%stateofcharge(SOC).Eckeretal.[3]usedaNMCcathodebasedcellformodelingofcalendarlifeaging.Alargecalendarlifetestmatrixandaloadpro?lewereusedfordeterminationofagingparametersandmodelveri?cation.However,thefocusofEckeretal.isonmodelingandtheinvestigatedcellsweredesignedforhighpowerapplicationsuptoaC-rateof20inhybridelectricve-hicles.TheworkofBloometal.didnotstudytheimpactofcellvoltageonaging.IncontrasttoBloometal.andEckeretal.thecellsinvestigatedinthisworkarelargerandhavebeenbuiltinaprocesscomparabletocommercialbatchproduction.Theycanprovidecurrentratesandcapacitiesadequateforarealworldapplicationinabatterystoragesysteme.g.foranEV.Apartfromthesedifferences,somecellsofthisworkwereusedforadetailedpost-mortemanalysisinpaper“PartB:Post-MortemAnalysis”[1],allowingadeeperunderstandingoftheoccurringagingprocessesinsidethistypeoflithium-ionbattery.2.ExperimentalThetestsemployalithium-ionhighenergypouchcellwithstackedelectrodes,anominalcapacityof10Ahandanominalvoltageof3.6V.Theyweresemi-automaticallycustom-manufac-turedforthisstudy.TheanodeconsistsofgraphiteandthecathodeemploysLiNi1/3Mn1/3Co1/3O2(NMC)asactivematerialandaluminumfoilascurrentcollector.Theelectrolyteisconstitutedof1MLiPF6saltinanorganicsolvent.Inthefollowingteststhetemperatureisvariedasanimpactfactoracceleratingtheratesofbothelectrochemicalcellreactionsandparasiticsidereactionscausingcellaging.Correspondingtoothercalendarlifestudiesandstudiesonagingeffects[12,13],parasiticreactionsoftheelectrolyteyieldingelectrolytedecompo-sitionandpassivationlayergrowthatthesurfaceofoneorbothelectrodesareconsideredasdominatingdegradationprocess.Theratesofthesereactionsarestronglyrelatedtotheelectrochemicalstabilityofthedifferentcomponentsandadditivesconstitutingtheelectrolyteatvariousoperatingconditions.Basedontheseconsid-erations,atestmatrixasdepictedinTable1,hasbeendesigned.Thechoiceoftesttemperatureshasbeenmadewithrespecttospeci?-cationsofthecellmanufacturer,regardingthethermalstabilityofTable1Calendarlifetestmatrix.Twodifferentkindsofcalendarlifetestswereperformed:OCestorageofcellsatopencircuitconditions,andCVeusingaconstantvoltagesourcetoascertainaconstantpotentialduringthetest.Numbersindicatetheamountofcellspertest.T?25??CT?40??CT?50??CT?60??C20%SOC,3.59V2xCV/2xOC3xCV50%SOC,3.72V3xCV3xCV/3xOC3xCV3xCV80%SOC,3.96V2xCV/2xOC3xCV90%SOC,4.06V2xCV2xCV100%SOC,4.19V3xCV2xCV/2xOC3xCVactivemassesandelectrolyte.Forachievingahigherdegreeofrepeatabilityinthetests,everyworkingpointofthematrixistestedasabatchconsistingoftwoorthreecells.Beforestartingthetests,thesebatcheswereselectedwithrespecttosimilaractualcapacitiesandresistancesforminimizingtheimpactsofstatisticalspreads.Duringthetests,thecellsmarkedwith“CV”arestoredinovensorclimatechambersatelevatedtemperaturesandatconstantvoltage.Theothersmarkedwith“OC”werestoredatopencircuitconditions.Forkeepingthevoltageconstant,everysinglecellisconnectedtoapowersupplyunit.Voltagesarederivedfromopencircuitvoltage(OCV)measurementsattesttemperatureofT?25??CasdepictedinFig.1.Thestateofcharge(SOC)anddepthofdischarge(DOD)inthisworkisrelatedtothenominalcapacityof10Ah.Inadditiontocalendarlifeagingtestsalsocyclelifetestswereperformedat1C-rateandT?40??Cwith10%DODaroundameanSOCof95%,5%,50%and50%DODaround75%meanSOCatthistemperature.Intestreachingeitherendofdischargevoltage(EODV)of3.0Vorendofchargevoltage(EOCV)of4.2Vthecyclingwasstartedatthosevoltageswithachargeordischarge.Inadditiontothesetests,cyclesbetweenEODVandEOCVweretestedatT?25??C,40??Cand60??Cat1C-rate.Thoughnotreachingthefullcapacityduetoohmic,kineticanddiffusionbasedvoltagedrops,thesetestsarereferredtoas“fullcycle”inthefollowing.OnecyclebetweenEOCVandEODVatT?40??Cleadstoadischargedcapacityofapproximately9.5Ahpercycledirectlyafterbeginoftest(BOT)and7.5Ahnearendoftest(EOT).Allcyclelifetestswereconductedwithaircoolingandheatsinksonbothsidesofthepouchcellsinclimatechambersorovens.Theambienttemperaturewascontrolledtokeepthesurfacetemperatureofthecellatthedesiredtesttemperaturewithamaximumdifferenceof2K.Toinvestigateagingandtodeterminetheactualcellstatus,areferenceparametertest(RPT)iscarriedoutatBOT,atEOTandinintervalsof6weeksinbetween.TheRPTisdepictedinFig.2.Itconsistsofastaticcapacitytestatstandardconditions(T?25??C)includinga1C-ratedischargeofretentioncapacitytoEODV,astandardcharge(1C-rateconstantcurrentchargetoEOCV,con-stantvoltagechargeatEOCVuntilcurrenthasdecreasedbelow0.05C-rate)anda1C-ratedischargeofthefullychargedcelltoEODVfordeterminingtheactualcapacityCact.Furthermoreitcontainsdeterminationofchargeanddischargeresistancesusingacurrentpulseatdifferentdepthsofdischarge.Incontextofthiswork,itemploysan18s4C-ratedischargepulsefollowedbya40srestperiodanda10s3C-ratechargepulsealsofollowedbya40srestperiod.Resistancesaredeterminedbetween90%and10%SOC.4.24.1Discharge4Charge
3.9V/VC3.8O3.73.63.53.4
0
2
4
6
8
10
discharged capacity/Ah
Fig.1.Opencircuitvoltage(OCV)plottedvs.dischargedcapacityfortheinvestigatedgraphitejNMCLi-ionhighenergycellatT?25??Cafterchargeanddischargeshowingasmallhysteresisatlowstatesofcharge.
574S.K?bitzetal./JournalofPowerSources239(2013)572e583
Fig.2.Testprocedureofthereferenceparametertest(RPT)appliedatbeginoftest,endoftestandevery6weeksinbetween.
InthisworkthetwokeyvaluesdescribingthecellstatusaretheactualcapacityCactandthe10sdischargeresistanceRactat50%SOC.Theresistanceiscalculatedbythedifferencebetweenthevoltagebeforethe4C-ratedischargepulseandafter10sduringthepulse,asdepictedinFig.3,dividedbythecurrentduringthedischargepulse.Allvaluesinthisworkaremeanvaluesofthetwoorthreetestedcells.1Everycellwasalsocharacterizedviaelec-trochemicalimpedancespectroscopy(EIS)ateachRPT.Adetailedanalysisoftheimpedancespectraobtainedduringtheagingtestswillbegiveninaseparatepublication.3.Resultsanddiscussion3.1.VariationoftemperatureincalendarlifetestsForoutliningtemperatureimpactondegradation,actualca-pacityandresistanceareplottedvs.timeforstorageataconstantcellvoltageof3.72V,whichisequivalentto50%SOC,andfourambienttemperatures(T?25??C,40??C,50??C,60??C).Theirevo-lutionsovertime,whicharedepictedinFig.4(a)and(b),arenormalizedtoconditionsatBOT.Thediagramsshowaveragevaluesofthecellsconstitutingatestscenarioandalsothevaluesforeachindividualcell,toillustratethevarianceinbetweenthespecimen.Asexpectedanincreaseofstoragetemperaturecausesasigni?cantaccelerationofcapacitydeclineandresistanceincrease,whichisespeciallyhighforT?60??C.Theresistanceincreaseshowsasimilarbehaviorasthecapacitydecline,exceptforthetestatT?40??C.Theresistanceincreaseappearslowerforthistempera-turecomparedtotheincreaseofcapacitydegradationbetweenT?25??CandT?40??C.Partofthisphenomenonhastobeaccountedtotheproblem,thattheresistanceofthecellsisrathersmall(Ractisapproximately4.7mUatBOT).Furthermoreevenslightchangesintemperatureof2KfromRPTtemperatureofT?25??C,asde?nedinSection2,showachangeoftheresistancevalueofapproximately5%.Thusresistancevaluesaremorepronetomeasurementerrors.Anywaythecontrolofthetemperaturechamberswasnormallymoreexactthanthespeci?ed2K.Never-thelessitcannotbeexcludedthatsingledatapointsmightshowdeviationbecauseoftheseproblems.Identi?cationofmainagingmechanismsrelatedtosinglepro-cessesishardlyeverpossiblewithoutcarryingoutpost-mortemanalysis,especiallyforcheckingtheelectrochemistryofthesingleelectrode.Theinteractionsbetweenthecomponentsofafullcellsystemaretoocomplex,toallowaseparationintosingleagingeffectsbyRPTasusedinthiswork.Resultsfrompost-mortemanalysis[1]indicatethattheobservedcapacityfadeismostlyduetolossofactivelithiumattheanode,probablyduetocontinuoussolidelectrolyteinterface(SEI)formation.Changesinporosityandweightofthesamplesrevealtheevolutionofdepositionproductsintheporousstructureoftheanode,indicatingelectrolytedecomposition.Usinginductivelycoupledplasmaopticalemissionspectroscopy(ICP-OES)showed,thatmostofthelithiumlostinthecathodewasdepositedduringagingontheanode[1].FindingmostofthelostreversiblelithiumontheanodesidealsopointstoaSEIdepositionreactionasmainlithiumlossmechanism.This?tsquitewelltotheobservedphenomenaintheRPTs.Nearlyallagingcurvesinthisworkshowastrongerdegradationinthebeginningandanapparentlinearbehaviorafteralongerperiodoftime.UsuallythisbehaviorisaccountedtothegrowthoftheSEIontheanode.Sincethepotentialoflithiatedgraphiteelectrodesisoutsidethestabilitywindowoftheelectrolyte,astableSEIiscrucialforcell3.853.8cell voltage/V3.753.73.653.63.55
ΔUdischarge,10 s3.502040t/s
Fig.3.Cellvoltagevs.timeduringpulsecurrenttest.Therhombimarkthevoltagevaluesusedforcalculationofthe10sdischargeresistance.AllshownresistancesRactinthisworkarerecordedat50%SOCandt?10s.
60801001ThespreadbetweentheparametersofeachcellcanbeseeninFig.4(a)and(b).S.K?bitzetal./JournalofPowerSources239(2013)572e583575
[a]10,95Cact/CBOT[b]1,91,81,71,61,51,41,31,21,110100200300t/days40050% SOC CV 25°C25°C, each cell50% SOC CV 40°C40°C, each cell50% SOC CV 50°C50°C, each cell50% SOC CV 60°C60°C, each cell0,90,850,80,750,750% SOC CV 25°C25°C, each cell25°C, eachcell50% SOC CV 40°C40°C, each cell40°C, eachcell50% SOC CV 50°C50°C, each cell50°C, eachcell50% SOC CV 60°C50%SOC CV60°C60°C, each cell60°C, eachcell0100200300t/days400500Ract/RBOT500Fig.4.ActualcapacityCact(a)anddischargeresistanceRact(b)vs.timeatdifferenttemperaturesand50%SOC.Allvaluesarenormalizedtoconditionsatbeginoftest(BOT).Averagedvaluesandvaluesforeachindividualcellareshown.
performance[13e16].Bygrowinginthickness,theSEIisprotectingtheelectrolytemoreeffectivelyfromfurtherreduction.There-actantsforsidereactionsaresupposedtodiffusethroughtheSEI,leadingtoasquarerootoftimeshapeddegradationduetoSEIthicknessincrease[17].Sincethelithium-ionshavetopassthroughtheSEI,aresistanceincreaseisobservedrelatedtothickness.3.2.VariationofcellvoltageincalendarlifetestsFigs.5and6depicttheevolutionofactualcapacity(a)andresistance(b)vs.timeattesttemperaturesofT?40??CandT?60??CwithvariousSOC.ThecharacteristicsofthecapacityfadeatCVinFig.5(a)appeartohaveamorecomplexdependencyonvoltagethanontemperature.Infullychargedstate(4.18V,100%SOC)atT?40??Cthecellsageveryfast,especiallywithinthe?rstfewweeksofstorage,losingmorethan20%ofcapacityinlessthanayearundertest.Avisualinspectionofcellsstoredat4.18Vindi-catedamoderategasevolutionwithinthepouchpacks.Cellvolt-agesequaltoorlessthan4.06V(90%SOC)showastabilizationofthesystem.Between50%SOCand80%SOCthereisonlyadiffer-enceofapproximately3%incapacitydecreaseafter170days.Un-fortunatelythe80%CVendedafteraperiodof170daysbecauseofacelldamage.Thecapacitydecreaseofthe90%SOCtestlookssimilartothetestat80%SOC,butinthe?rstRPTadifferentagingbehaviorisobserved.ThereasonmightbeadifferentBOTtime,sinceboth90%SOCtestswerestartedapproximately?vemonthslaterthantheothertestsdescribedinthiswork.Thecellsusedinthedelayedtestswereusedforonlyfewprecedingtestcyclesandthenstoredatroomtemperature.AtBOTtheyshowednosigni?cantchangeinresistanceandcapacityvaluescomparedtothoseafter?rstcheckup,afterthecellsweresuppliedtothetestfacilities.The20%SOCtestshowsacompletelydifferentbehaviorwithanincreaseofcapacityatBOT,keepingtherelativecapacityabove100%for300daysoftesting.ThecomparisonbetweenOCandCVinFig.5showsadifferentagingbehavioronlyat100%SOC.Theothercapacitycurvesshownosigni?cantdifferencebetweenOCandCVstorageconditions.Afteraperiodofapproximately200days,thecapacityoffsetseemstoremainnearlyconstant.The100%OCtestwasaccidentlysetto20%SOCafterthefourthRPT(200days)andseventhRPT(361days)for6weeksrespectively.Inbothcasesacapacityrecovery(2.9%and1.5%respectively)isobserved.Lookingatself-dischargeofthecells,the100%OCtestshowsadropofvoltagebetweentheRPTsof100e180mV.Forcomparison,thevoltagedropofthe20%SOCtestbe-tweeneachRPTduetoself-dischargeisabout20e30mV.Itseemsthevoltageof4.18Vat100%SOCleadstopotentialsnearthesta-bilitywindowoftheelectrolyte.Ifthevoltageiskepthigh,asinaCVtest,thebatteryvoltageandthesidereactionratearekepthigherincomparisontoanOCtest.Byself-dischargetheOCtestscandroptopotentials,wheretheelectrolyteismorestable,decreasingthecapacityfades.AtlowerSOCtheelectrolyteisstableanyway,sonotonlytheself-dischargerateislow,butalsothein-?uenceofcellvoltagedecreaseisnegligible,leadingtosimilarbehaviorofOCandCVtests.[a]1.051Cact/CBOT[b]1.8T= 40°C100% SOC CV90% SOC CV80% SOC CV50% SOC CV20% SOC CV100% SOC OC80% SOC OC50% SOC OC20% SOC OC1.71.6Ract/RBOTT= 40°C100% SOC CV90% SOC CV80% SOC CV50% SOC CV20% SOC CV100% SOC OC80% SOC OC50% SOC OC20% SOC OC0.950.90.850.80.750100200300t/days4001.51.41.31.21.110100200300t/days400Fig.5.ActualcapacityCact(a)anddischargeresistanceRact(b)vs.timeatdifferentSOCandT?40??C.Solidlineswith?lledmarkersshowtestsatconstantvoltage(CV).Dashedlineswithhollowmarkersshowtestsunderopencircuit(OC)conditions.Allvaluesarenormalizedtoconditionsatbeginoftest(BOT).
576S.K?bitzetal./JournalofPowerSources239(2013)572e583
[a]
10.9Cact/CBOT[b]
8T= 60°C76Ract/RBOTT= 60°C100% SOC CV100% SOC, each cell90% SOC CV80% SOC CV50% SOC CV20% SOC CV0.80.70.60.50.40100200300t/days400500100% SOC CV90% SOC CV80% SOC CV50% SOC CV20% SOC CV543210100200300t/days400500Fig.6.ActualcapacityCact(a)anddischargeresistanceRact(b)vs.timeatdifferentSOCandT?60??C.Allvaluesarenormalizedtoconditionsatbeginoftest(BOT).(b)Showsinadditiondatapointsforeachcellat100%SOC,exposingstrongdeviationsinbetweenthethreecells.
IntheresistancecurvesinFig.5(b)aslightlydifferentbehaviorcanbeobserved.WhiletheOCandCVtestbehaviorat100%SOCand20%SOCcanbeexplainedbytheaforementionedmodel,theOCtestsatT?40??Cat80%SOCand50%SOCshowsomeresistancescatteringandahigherresistanceincreasethantheCVtests.Thisisquiteunexpected,butinthe?nalanalysisnoalternativeexplana-tionswerefoundbasedonthetestdataobservedinthisstudy.TheseOCvs.CVresistanceagingeffectshavetobestudiedinmoredetailinfurtherstudies.Furthermoretheresistancesforboth20%SOCand100%SOCtestsshowadifferentbehaviorcomparedtocapacity.Althoughacapacityincreasecanbeobservedat20%SOC,theresistanceincreasesinthistestscenario.Duringthechangeto20%SOCinthe100%SOCOCtest,thecapacityrecovers,butalsoheretheresistancedoesnotdecreaseatthesametime(Fig.5(b)).Neverthelesstheresistanceincreaseissmallestat20%SOCCVandhighestat100%SOCCV.The100%SOCOCtestmightshowaslightin?uenceofthetwotimeperiodsat20%SOC,butatleastthein-?uenceisquitesmallandwithintherangeoftheerroroftheresistancemeasurement.Altogethertheresistanceincreaseseemstobelesssensitivetothecellvoltagelevelthanthecapacityloss.ThecapacitytestsatT?60??C(Fig.6(a))behavequiteasex-pected,sincetheiragingbehaviorisincreasedinrateduetothehighertemperature.Whilethe100%SOCtestshowsanextremedecreaseofcapacityagain,80%SOCand90%SOCbehavenearlyidentical.TheresistancesinFig.6(b)correspondquitewelltothecapacitydecrease.The20%SOCtestshowsnocapacityincreaseanymore,butanunusualshapeforcapacitydecline,whichisalsodiscussedinmoredetaillaterinSection3.5.3.3.CyclelifeTheresultsofthecyclelifetestsanalyzedinthisworkaredepictedinFigs.7and8.Alltestsrunat1C-rate,leadingtoathroughputofapproximately20,000dischargedAhafter200days,whichequatesto2000equivalentfullcycles.ThetestsdiscussedinthisworkareshowninTable2.Withinthisstudyseveraladditionalcyclelifetestshavebeenperformed,buttheywillbediscussedinaseparatepublication.Comparedtothecalendarlifetestsasigni?cantin?uence,especiallyofthefullcycles,canbederived.InFig.7(a)the10%DODcyclesshowlittletonoaddedin?uenceoncapacitydecreasecomparedtothecorrespondingcalendarlifetests.Incontrast50%DODandfullcyclesdiffersigni?cantlyfromcalendarlifeaging.FromtheresultsdepictedinFig.7(b)nocleardependencyfortheresistancecanbederived.AllcycletestsatT?40??CshowasimilarbehaviorwithonlyslightdependencyonmeanSOCandnocleardependencyonDOD.Fig.8(a)showsthein?uenceoffullcyclesatdifferentcelltemperaturesoncapacity.IndifferencefromRef.[11]allcyclelifetestsshowahighercapacitydeclinethanthestoragetestswiththe[a]1.051Cact/CBOT[b]
1.8
T= 40°CFull Cycle, 50% SOC50% DOD, 75% SOC10% DOD, 95% SOC10% DOD, 50% SOC10% DOD, 5% SOC100% SOC CV80% SOC CV80% SOC OC50% SOC CV20% SOC CV1.71.6Ract/RBOT1.51.41.31.21.110
100
200300t/days
400
T= 40°C
Full Cycle, 50% SOC50% DOD, 75% SOC10% DOD, 95% SOC10% DOD, 50% SOC10% DOD, 5% SOC100% SOC CV80% SOC CV80% SOC OC50% SOC CV20% SOC CV
0.950.90.850.80.750100200300t/days400Fig.7.ActualcapacityCact(a)anddischargeresistanceRact(b)vs.timeforcyclelifetestsat1C-rateatT?40??C.TestswereperformedatdifferentDODarounddifferentmeanSOC.Dashedlinesshowthecyclelifetests.SelectedcalendarlifetestsequalorneartothemeanSOCaredepictedforcomparisonanddrawnascontinuouslines.Allvaluesarenormalizedtoconditionsatbeginoftest(BOT).
