E ff ects of High and Low Salt Concentrations in Electrolytes at Lithium − Metal Anode Surfaces Using DFT-ReaxFF Hybrid Molecular Dynami

《E ff ects of High and Low Salt Concentrations in Electrolytes at Lithium − Metal Anode Surfaces Using DFT-ReaxFF Hybrid Molecular Dynami》由会员上传分享，免费在线阅读，更多相关内容在学术论文-天天文库。

pubs.acs.org/JPCLLetterEffectsofHighandLowSaltConcentrationsinElectrolytesatLithium−MetalAnodeSurfacesUsingDFT-ReaxFFHybridMolecularDynamicsMethodYueLiu,QintaoSun,PeipingYu,YuWu,LiangXu,HaoYang,MiaoXie,TaoCheng,*andWilliamA.GoddardIIICiteThis:J.Phys.Chem.Lett.2021,12,2922−2929ReadOnlineACCESSMetrics&MoreArticleRecommendations*sıSupportingInformationABSTRACT:Duetocreatingapassivatedsolidelectrolyteinterphase(SEI),highconcentration(HC)electrolytesdemonstratepeculiarphysicochemicalpropertiesandoutstandingelectrochemicalperform-ance.However,thestructuresofsuchSEIremainsfarfromclear.Inthiswork,ahybridabinitioandreactivemoleculardynamics(HAIR)schemeisemployedtoinvestigatetheconcentrationeffectofSEIformationbysimulatingthereductivedegradationreactionsoflithiumbis(fluorosulfonyl)imide(LiFSI)in1,3dioxalane(DOL)electrolytesatconcentrationsof1M,4M,and10M.TheefficientHAIRschemeallowsthesimulationstoreach1nstopredictelectrolytes’deepproductsatdifferentconcentrations.ThesimulationfindingsshowthatthemostcriticaldistinctionbetweenHCanditslowconcentration(LC)analogueisthataniondecompositioninHCismuchmoreincompletewhenonlyS−Fbreakingisobserved.Theseinsightsareimportantforthefuturedevelopmentofadvancedelectrolytesbyrationaldesignofelectrolytes.ithium(Li)-ionbatteries(LIBs),servingasapromisinghasbeenconductedtoinvestigateanddevelopSEIinL15−18candidatetoreplaceconventionalenergystoragedevices,LMBs.StrategiesareproposedtocontroltheSEIforminghaveachievedcommercializationinmanyfields,suchasprocess,includingtheoptimizationofelectrolytecomposi-mobilephones,computers,andelectricvehicles.1−3However,tion,19,20thecreationofsolidelectrolytes,21andtheowingtographite-anodecoupledLi+interaction/deinteraction22,23constructionofanartificialSEIsheet.Forexample,highcathodeusedinLIBs,thespecificcapacityofLIBsisconcentration(HC)electrolyteswithLi-saltshavebeenDownloadedviaUNIVOFNEWMEXICOonMay16,2021at06:47:51(UTC).approachingthetheoreticalvalues(372mAhg−1),butitisdevelopedinrecentyearstoinhibitdendritegrowthandfarfrommeetingthegrowingdemandsforhigh-energyenhanceCEs.24AnHCelectrolytesystemusing7Mlithium4,5storage.High-voltage(>4.0V)lithium(Li)metalbatteriesbis(trifluoromethanesulfonyl)imide(LiTFSI)concentrationSeehttps://pubs.acs.org/sharingguidelinesforoptionsonhowtolegitimatelysharepublishedarticles.(LMBs),whenusingtheidealanodematerial,Limetal,asthedissolvedinthemixtureofdimethoxyethane(DME)andanode,havedrawnsignificantattentionowingtoitsultrahighdioxalane(DOL)inLi/SbatterieswassuggestedbySuo,25theoreticalspecificcapacityof3860mAhg−1andanextremelyshowingthattheHCelectrolytesystemwouldovercometwolowelectrochemicalpotential(−3.040Vvsstandardhydrogenkeytechnologicalproblemsatthesametimewiththe6−8electrode,SHE).Unfortunately,thehighreactivityofLipolysulfideshuttleandpoorlithiummetalanodestability.metalposesagreatchallengetotherealisticapplicationof26Qianetal.studiedtheuseofsaltlithiumbis(fluorosulfonyl)-LMBsduetothecontinuousreductionofthesolventsandLi-imide(LiFSI)indimethoxyethane(DME)athighconcen-saltsintheelectrolytewithelectrodesduringtheprocessof9,10trations(upto4M)andfoundthatdendritegrowthwaschargeanddischarge.Furthermore,theinevitabledendriteeliminated,whereashighCEs(>99%)werealsoachieved.KimgrowthofLiandelectrolytereductionbyLimetalleadstolow11,12Coulombicefficiencies(CEs)andseveresafetyconcerns.TheseimportantbarrierstoLMBsarethereforedesperatelyReceived:February5,2021requiredtobuildthenextgenerationofhigh-energystorageAccepted:March10,2021equipment.Published:March16,2021Itiswidelyacceptedthattheformationofasolidelectrolyteinterphase(SEI)playsanimportantroleindeterminingthe13,14stabilityandperformanceofLimetal.Extensiveresearch©2021AmericanChemicalSocietyhttps://doi.org/10.1021/acs.jpclett.1c002792922J.Phys.Chem.Lett.2021,12,2922−2929

1TheJournalofPhysicalChemistryLetterspubs.acs.org/JPCLLetter27etal.conductedexperimentsbyusinganultrahighLiFSI-Todescribetheelectronexchangeandcorrectionenergies,basedelectrolyteresultingincomparableCEsandcyclethePerdew−Burke−Ernzerhof(PBE)functionalwithinthestability.Theoretically,althoughmanystudiesonHCgeneralizedgradientapproximation(GGA)isusedinthiswork43electrolyteattheatomiclevelhavebeenreported,thetoconductAIMDsimulation.Inaddition,weemploythecomprehensivedegradationreactionofelectrolyteonaLiGrimmeD3correctiontodescribetheLondondispersion28−344445metalanodewithSEIformationisindebate.Molecularintegrations.A1×1×1Monkhorst−Packk-pointmeshdynamics(MD)simulationshavebeenknowntobeoneofthewassettosampleBrillouinzoneintegration,andtheprojector46mostpowerfulmethodsforexploringcomplicatedreactionaugmentedwave(PAW)methodasimplementedisusedtoprocessesbytrackingMDtrajectories.Forexample,Camacho-considerelectron−ioninteractions.Asforplane-wavebasis28expansion,wechosea400eVenergycutoff.ThethresholdforForerocomparedthebehaviorof1Mand4MconcentrationelectrolyteswithLiFSIandLiTFSIinDMEtheelectronicstructureconvergenceoftheself-consistentfieldwassetto10−4eVwithaGaussiansmearingwidthof0.2eV.solventusingtheabinitioMD(AIMD)process;thesetheoreticalfindingsindicatedthatthepossiblecomponentsTable1showsthereactionenergiespredictedwithQMandoftheSEIlayerdependonthechemicalstructureofthetheReaxFFmethod.TopreparetheQMtrainingset,theFSIelectrolyte,andLiFwasobservedasasubstanceproducedbythecompletedecompositionofLiFSI.Inaddition,theTable1.RelativeReactionEnergies(inkcal/mol)forQMsacrificialanionreductionprocesshasbeenexplainedbyandReaxFFMethod30Sodeyamaandhisco-workers,showingthatTFSIanionsrelativereactionenergies(kcal/mol)prefertobereducedbyacceptingtheelectronatthehighconcentration(HC)conditioninordertopreserve2electronreactionQM(B3LYP/6-311+g(d,p))ReaxFFreductivesolventdecompositionusingDFT-MDsimulationsHN(SO2F)2HN(SO2F)SO2+F83.991.9withexplicitsolventsandΔSCFthatconsiderstherelaxationHN(SO2F)2HNSO2F+SO2F54.847.7ofexcesselectronsafterthereductivedecompositionofthe30electrolytes.anionisneutralizedbyaddingoneHatom.Thegas-phaseOwingtothehighcomputationalcostofAIMD,thebonddissociationenergieswithQMresultsfortheS−Fandsimulationtimescaleislimitedtodozensofpicoseconds(ps),N−Sbondsare83.9and54.8kcal/mol,whilethewhilenanoseconds(ns)tomicroseconds(ms)arerequiredtocorrespondingvaluesare91.9and47.7kcal/molwiththeclarifytheSEIfilmformationprocess.ReactiveforcefieldReaxFFmethod.AlthoughthebondenergyofS−Fisstronger35−38(ReaxFF),derivedfromquantummechanics(QM)thanthatofN−Sinthegasphase,itisnotnecessarythattheresults,havebeenwidelyusedtosimulatecomplexmultiphaseN−SbondmustbebrokenbeforetheS−Fbond.AsshowninchemicalreactionswithmuchmoreaffordablecomputationalFigure1a,theS−FbonddissociationcurvesshowthattheHcost.IthasbeendemonstratedthatinstandardReaxFF,theatomisbondedwiththeFatom,indicatingthattheFSIionlackofexplicitconsiderationoftheelectronleadstoanwoulddecomposeintotheFionwithareactionenergyof64.438incorrectdescriptionoftheelectrochemicalreaction,whichkcal/mol,whichopensthepossibilityofS−Fbondcleavages.requiresmoreelaboratetreatmentoftheelectroninadvancedInordertoinvestigatetheinitialreactionprocesstoreduce39ReaxFF,suchaseReaxFF.Thus,asinglesimulationapproachthedeteriorationofthelithium-electrolytesystem,HAIRcannotmeettheneedforlong-termMDsimulationswithsimulationsareperformedbasedonthetheoreticalmodel,reasonableprecisionandcost.whichcontainsaLi-slabwithamixedelectrolyteofDOLandInthiswork,ahybridscheme,hybridabinitioandreactiveLiFSI;thedetailedinformationisshownintheSupportingmoleculedynamics(HAIR),thatcombinesAIMDandReaxFFInformation.Figure2showsthesequenceofFSIdecom-40MDisproposed.TheAIMDpartoftheHAIRmethodcanpositionduring0−6.0ps,atapproximately0.5ps,theLi-FSI-describethelocalizedelectrochemicalreactionsaccurately,DOLpairisobservedintheperiodicboxthroughLi−OwhileReaxFFMDcouldacceleratechemicalreactionsandcoordination.ThefollowingdecompositionstageoftheFSImasstransferwithamuchmoreaffordablecostwhilekeepinginitiatesanS−FbondbreakasexpectedbyQMfindingsandtheQMaccuracywhentheforcefieldparameteriswell28previouswork.Afterthat,theN(SO2)2fragmentundergoestrained.reductivereactionbyLi0toformLiO,andthentheN−S2ToapplytheHAIRscheme,wefirsttrainedthemissingbondisbrokenintoNSO2andSOataround5.6ps.InordertoforcefieldparametersofFSI-anions.RelevantReaxFFvalidatetheaccuracyoftheHAIRsimulation,a3.5pslongparametersareoptimizedbyfittingQMcalculationsstartingAIMDsimulationisperformedandmoredetailscanbefound38fromtheparametersdevelopedbyIslametal.TheoptimizedintheSupportingInformation.AsshowninFigure3,theinitialReaxFFparametersanddetailedresultsareshowninthereactionofFSIpredictedwithAIMDisS−FbondcleavageatSupportingInformation.DuringtheHAIRsimulations,the1.0ps,whichconfirmstheHAIRresults(1.5ps).S−FbondAIMDandReaxFFsimulationsareperformedwithViennaAbbreakingisinducedbythedecompositionprocessofFSIand41InitioSimulationPackage(VASP5.4.4)andLarge-scaleprecededbyN−Sbondcleavage,indicatingnovariationsAtomic/MolecularMassivelyParallelSimulator(LAMMPSbetweenDFT-basedAIMDandHAIRresults.422018)software,respectively.MDsimulationsareconductedThepurposeofincludingAIMDistodescribechemicalalternativelyusingtheNVTensembleat300Kwith1fsandreactionsmoreaccurately.AlthoughthetrainingsetofReaxFF0.25pstimestepsforAIMDandReaxFFMDsimulationswhileisderivedfromQMcalculations,ReaxFFitselfcannotfullyensuringefficientconvergeforcollisionsandsmoothreactions.guaranteetheaccuracyindescribingelectrochemicalreactions,Inthiswork,10-timeaccelerationischosenonliquidespeciallyconsideringthattheworkfunction(theelectron’selectrolyte,inwhichdiffusionisfast,whichmeansthatthechemicalpotential)significantlychangesduringthereactions.moleculardynamicssimulationsstartwiththeAIMD(0.5ps),Therefore,itisstillnecessarytoincludetheAIMDinthefollowedbyReaxFFMD(5ps),andcontinuedalternatively.HAIRscheme.Indeed,criticalreactions,suchasS−FandN−S2923https://doi.org/10.1021/acs.jpclett.1c00279J.Phys.Chem.Lett.2021,12,2922−2929

2TheJournalofPhysicalChemistryLetterspubs.acs.org/JPCLLetterFigure1.BonddissociationcurvesarepredictedbyQMandtheReaxFFmethod:(a)S−Fbondand(b)N−Sbond.TheblacklineisfortheQMresults,andtheredlineisfortheReaxFF.Thecolorsarelithiuminpurple,oxygeninred,carboningray,fluorineincyan,sulfurinyellow,nitrogeninblue,andhydrogeninwhite.time,noDOLdecompositionwasobservedinMC.ThedifferenceininitialreactionsismoresignificantwhencomparedwiththecasesofLCandHC.AsshowninFigure4c,onlyS−Fbond-breakingwasobservedinHC,whileallthebackboneofFSIanionsstillexistsintheelectrolyte.Asexpected,noDOLdecompositionwasobserved.Withtheincreaseofsaltconcentration,LiFSIdecomposedtoconsumefreeLi0topreventDOLfromreducing,whichhasbeenproposedasthe“sacrificialreactionmechanism”bySodeya-30ma.InordertofurtherinvestigatethecomponentsandstructureoftheSEIsheet,thetimescaleHAIRsimulationswithvariousconcentrationsareexpandedto605ps,followedby400psReaxFFMDand10psAIMDsimulations.InFigure5,thesesnapshotsexhibitsignificantdifferencesatdifferentconcen-trationsafter1nssimulation.At1Mand4Mconcentrations,theSEIlayerwasformedwithclustersandcavities,introducingFigure2.SequencesofFSIdecompositionwereobtainedfromHAIRsimulationsfortheDOL/LiFSImixturebetween0and6.0ps:(a)0.0aheterogeneousSEIlayer.Whileahomogeneousandstableps,(b)0.5ps,(c)1.5ps,(d)5.52ps,(e)5.6ps,and(f)6.0ps.Thelayer,consistingofatypicalinorganiclayerbytheorderofLi−colorsarelithiuminpurple,oxygeninred,carboningray,fluorineinO,Li−S,andLi−Fcoordination,isobservedattheHC(10cyan,sulfurinyellow,nitrogeninblue,andhydrogeninwhite.M)concentration.AlayerofSEImajoringinLixFisobservedafter1.015nsHAIRsimulationthatconsistsof0.055nsAIMDbondcleavagesofFSIareobservedintheAIMDloop,whichisand0.96nsReaxFFMD.SuchLixFlayerisgenerallybelievedbasicallyinlinewithourexpectations.Thatis,AIMDtoplayacrucialroleinstabilizingthelithiumanode.Intheguaranteestheaccuracyofthechemicalreactions,whileworkofOspina-Acevedoetal.,theyalsoreportedtheReaxFFspeedsupthemasstransfer.formationofasimilarLixFlayeratlowerconcentration(2M),largersimulationsize(∼20nm3),andlongertimescaleTheconcentrationeffectofFSIinSEIformationis49investigatedbycarryingoutHAIRsimulationsatdifferent(∼20ns).concentrationsof1M,4M,and10M(ultrahighFigure6showstheradialdistributionfunction(RDF)and47theintegratednumberofdifferentbondsafterlong-timeconcentrationsintheexperiment).Figure4showsthesnapshotsofHAIRsimulationat275psofsimulationswithsimulationswithLC,MC,andHCconcentrations.Apparently,differentconcentrations.ThesimulationresultsofLCaresimilarpeaksataround2ÅareobservedforLi−FandLi−OshowninFigure4a,whichimpliesthecompletedecompositionbondsatdifferentconcentrations,whichimpliesthestableofFSIanionintoLiSN,SO,LiF,andLi2Oproducts.productsLiFandLi2OareformedduringMDsimulationstoMeanwhile,thereductionreactionofDOLisalsoobserved,contributetotheinorganiclayerofSEIfilmasmentionedby0Ospina-Acevedo.49ItisaremarkablefactthattheLi−FbondwhichisreducedbyLiviaring-openingprocesstoformC2H4,LiOCHO,CH2O,LiOCHCH2.SimilarsimulationresultsRDFshowssignificantdifferencesfor1M,4M,and10M48werealsoreportedbyYunetal.Additionally,thesimulationLiFSIsystems,andseveralsharpenedpeaksareseenbetween3resultsofamediumconcentration(MC,4M)areshowninand8Åforthe1MLiFSIsystem,whilesmoothcurvesareFigure4b.ThemajordifferencebetweenMCandLCisthatshowninMCandHCconcentrations(4and10MLiFSIthedecompositionofFSI-isincompleteinMC.Atthesamesystems).Obviously,intheHCconcentrations,freeLi0are2924https://doi.org/10.1021/acs.jpclett.1c00279J.Phys.Chem.Lett.2021,12,2922−2929

3TheJournalofPhysicalChemistryLetterspubs.acs.org/JPCLLetterFigure3.ReactionpathwayobtainedfromHAIR(blue)andAIMD(red)simulations.Figure4.SnapshotsfromMDsimulationat275psfor(a)1MLiFSI/DOL,(b)4MLiFSI/DOL,and(c)10MLiFSI/DOL.Thecolorsarelithiuminpurple,oxygeninred,carboningray,fluorineincyan,sulfurinyellow,nitrogeninblue,andhydrogeninwhite.consumedbythedecompositionfragmentsofLiFSIwitha(LiF,Li2O,Li2S)andtheorganiclayer(LiCHO,LiOCHlargeamountofLiFformed,butfortheLCsystem,onlyservalCH2,etc.),whichmainlyderivesfromthereductionofLiFSIpeakscouldbefoundastheinorganicparts(LiF).AccordingandDOL.Concerningthe4MLiFSI/DOLsystem,theSEItoFigure6c,d,effectsonthedecompositionprocessesofLiFSIfilmincludesinorganicpartLiF,Li2O,Li2S,andN2gas,withdifferentconcentrationsarealsoverifiedbytheRDFindicatingsalt-derivedSEIformationaspredictedbyprevious28,30analysesforLi−SandS−Obonds,in1Mand4Mwork.Withregardtotheultrahighconcentrations,moreconcentrations,theobviouspeaksarefoundinLi−SRDFatstableproductsareclarifiedinSEIinorganicparts,suchas2.4ÅwhilenopeaksareshowninS−Obond.IncontrasttoLi2CO3andLi2SO3,whichareidentifiedtobethemainparts32the1Mand4MLiFSIsystems,aclearpeakisidentifiedat1.7oftheSEIlayerasexperimentally.Additionally,owingtotheÅfortheS−Obond.TheseRDFresultsindicatethattheincompletedecompositionofTFSIions,LiFispriortobeingLiFSIaredecomposedincompletelywithnoLiSformedingeneratedwiththeconsumptionoffreeLi0,whicharethemain2the10MLiFSI/DOLsystem,showingagoodagreementwithcompositionsoftheSEIlayerformedatHCelectrolyte,whilethesnapshotsat275pswithdifferentconcentrations.moreLi2OareobservedintheMCsystem.AstheincreaseofTofurtheridentifythedifferentcomponentsofSEIfilmatconcentrations,wealsofoundLiFshowacontinuousincrease,differentconcentrations,Figure7showsthemainproductsandnoLi2SwasseenintheHCsystemaspredictedbyRDFobtainedfromlong-timesimulations(morethan1ns),andanalyses.otherproductsarelistedinTable2.In1MLiFSI/DOLForthesakeofexploringtheLiFafter1nslongsimulations,electrolyte,theSEIfilmiscomprisedoftheinorganiclayerX-rayphotoelectronspectroscopy(XPS)isemployed.A2925https://doi.org/10.1021/acs.jpclett.1c00279J.Phys.Chem.Lett.2021,12,2922−2929

4TheJournalofPhysicalChemistryLetterspubs.acs.org/JPCLLetterFigure5.SnapshotsfromMDsimulationafter1nsfor(a)1MLiFSI/DOL,(b)4MLiFSI/DOL,and(c)10MLiFSI/DOL.Thecolorsarelithiuminpurple,oxygeninred,carboningray,fluorineincyan,sulfurinyellow,nitrogeninblue,andhydrogeninwhite.Figure6.Radialdistributionfunctionandintegratednumberofbondsfor(a)Li−F,(b)Li−O,(c)Li−S,and(d)S−Oafter605psHAIR,400ReaxFF,and10psAIMDsimulations.Theblacklineisfor1MLiFSI/DOL;theredlineisfor4MLiFSI/DOL;andthebluelinefor10MLiFSI/DOL.thousandstructuresfromthelast5psAIMDsimulationsarearecalculatedbyinitialstateapproximationusingVASPatthe50selectedtosimulatetheXPSspectra,andthebindingenergiesPBE-D3level.Statisticaldistributionsarefittedwiththe2926https://doi.org/10.1021/acs.jpclett.1c00279J.Phys.Chem.Lett.2021,12,2922−2929

5TheJournalofPhysicalChemistryLetterspubs.acs.org/JPCLLetterFigure7.Mainproductdistributionsafter1nslong-timesimulationsatconcentrationsof(a)1MLiFSI/DOL,(b)4MLiFSI/DOL,and(c)10MLiFSI/DOL.Table2.ProductsObtainedfromHAIRMDSimulationafter1nsfor1MLiFSI/DOL,4MLiFSI/DOL,and10MLiFSI/DOLSystemsmainproducts1MLiFSI/DOLLiH,LiF,Li2O,LiOH,Li2S,LiNS,C2H5F,LiCHO,LiOCHCH24MLiFSI/DOLLiF,Li2O,Li2S,N210MLiFSI/DOLLiF,Li2O,LiNSO,LiSF,Li2CO3,Li2SO3,LiOH,LiNSO2,LiH,HF,H2O51GaussianfunctionasshowninFigure8.ThebindingenergyofS−F,H−F,andLi−Fis657.2eV,655.5eV,and654.1−654.4eV,respectively,whichindicatesbindingenergyshiftofLi−FtoS−Fis−2.95±0.15eV,showingagoodagreement52withbindingenergyshift−2.9eVpredictedexperimentally.Inaddition,anobviousincreaseofintensityforLiFisseenalongwithaconcentrationincreaseastheproductdistributionispredicted.Insummary,ahybridAIMD-ReaxFFscheme(HAIR)thatcombinesQMandMMreactivedynamics,isemployedtoinvestigatethereductionreactionandSEIformationatdifferentconcentrations(1M,4M,and10MLiFSI/DOLFigure8.XPSspectraofF1sof(a)1MFSI-DOL(b)4MFSI-DOL,electrolytesystems).ToachievetheReaxFFMDpartinHAIRand(c)10MFSI-DOL,respectively.simulations,parametersaredevelopedbytrainingQM-basedcalculations.ThereductionreactionofLiFSIobtainedfromconcentrations,including1M,4M,and10MLiFSI/DOLHAIRsimulationsisinitiatedbyS−Fbondcleavageandelectrolytesystems.Accordingly,ourtheoreticalresultsfollowedbyN−Sbondbreaking,whichisnotonlysupportedindicatethatthecompletedecompositionofFSIionandbyQMcalculationsbutalsoverifiedbyshortAIMDDOLsolventcontributestotheinorganicandorganicpartsofsimulations.OnthebasisofthereliableinitialreactionstheSEIlayerintheLCsystem.Instead,inHCelectrolyte,FSIusingtheHAIRmethod,MDsimulationsareextendedto1nsionsdecomposedincompletelyviainitialS−FbondbreakinglongtoclarifytheeffectsonthedegradationreductionoftoconsumefreeLi0,whichpreventsthesolventDOLfromelectrolyteandcompositionsoftheSEIlayerwithdifferentsaltreducingsacrificially.FurtherRDFandXPSanalysisimplythe2927https://doi.org/10.1021/acs.jpclett.1c00279J.Phys.Chem.Lett.2021,12,2922−2929

6TheJournalofPhysicalChemistryLetterspubs.acs.org/JPCLLetterdifferencesofSEIcompositionsforLC,MC,andHCNaturalScienceFoundationofJiangsuHigherEducationelectrolytes,astheresultofthesignificantreactionmechanismInstitutions(GrantSBK20190810),theJiangsuProvinceHigh-isaffectedbysalt-concentrations.ThemainproductsoftheLCLevelTalents(GrantJNHB-106),andthePriorityAcademicsystemcontainLiF,Li2O,Li2S,LiCHO,andLiOCHCH2ProgramDevelopmentofJiangsuHigherEducationInstitu-whilemoreinorganicproducts,suchasLi2CO3andLi2SO3,tions(PAPD)forfinancialsupport.H.Y.thankstheChinaLiF,andLi2O,areidentifiedinHCelectrolyte,andtheproductPostdoctoralScienceFoundation(Grant2019M660128)fordistributionsshowthatLiFshowsacontinuousincreasealongfinancialsupport.Thisworkwaspartlysupportedbythewiththeincreaseofconcentration,andLi2SdoesnotexistinCollaborativeInnovationCenterofSuzhouNanoScience&HCsystemsduetotheincompletedecompositionofLiFSI.Technology.ThesetheoreticalresultsdemonstratetheSEIformationswitchesfromthesolvent-derivedlayertosalt-derivedSEI■REFERENCESlayerbysimplyincreasingsaltconcentrations,whichsupply(1)Tarascon,J.M.;Armand,M.IssuesandChallengesFacingnewinsightsforfurtherdevelopmentsofadvancedelectrolytes.RechargeableLithiumBatteries.Nature2001,414,359−367.(2)Dunn,B.;Kamath,H.;Tarascon,J.-M.ElectricalEnergyStorage■ASSOCIATEDCONTENTfortheGrid:ABatteryofChoices.Science2011,334,928−935.*sıSupportingInformation(3)Yang,Z.;Zhang,J.;Kintner-Meyer,M.C.W.;Lu,X.;Choi,D.;TheSupportingInformationisavailablefreeofchargeatLemmon,J.P.;Liu,J.ElectrochemicalEnergyStorageforGreenGrid.https://pubs.acs.org/doi/10.1021/acs.jpclett.1c00279.Chem.Rev.2011,111,3577.(4)Goodenough,J.B.;Park,K.S.TheLi-IonRechargeableBattery:Detailsofreactiveforcefieldparametersoptimization,APerspective.J.Am.Chem.Soc.2013,135,1167−1176.forcefieldparameters,modelofthelithium-electrolyte(5)Cheng,X.B.;Zhang,R.;Zhao,C.Z.;Zhang,Q.TowardSafesystem,shortAIMDresults,andXRDpatternsfromMDLithiumMetalAnodeinRechargeableBatteries:AReview.Chem.Rev.simulationsfor1MLiFSI/DOL,4MLiFSI/DOL,and2017,117,10403−10473.10MLiFSI/DOLsystemsafterlong-timesimulations(6)Xu,W.;Wang,J.;Ding,F.;Chen,X.;Nasybulin,E.;Zhang,Y.;(morethan1ns)(PDF)Zhang,J.G.LithiumMetalAnodesforRechargeableBatteries.EnergyEnviron.Sci.2014,7,513.(7)Lin,D.;Liu,Y.;Cui,Y.RevivingtheLithiumMetalAnodefor■AUTHORINFORMATIONHigh-EnergyBatteries.Nat.Nanotechnol.2017,12,194−206.CorrespondingAuthor(8)Zheng,J.M.;Engelhard,M.H.;Mei,D.H.;Jiao,S.H.;Polzin,B.TaoCheng−InstituteofFunctionalNanoandSoftMaterialsJ.;Zhang,J.G.;Xu,W.ElectrolyteAdditiveEnabledFastChargingandStableCyclingLithiumMetalBatteries.Nat.Energy2017,2,(FUNSOM),SoochowUniversity,Suzhou215123,China;17012.orcid.org/0000-0003-4830-177X;Email:tcheng@(9)Huang,C.;Xiao,J.;Shao,Y.;Zheng,J.;Bennett,W.D.;Lu,D.;suda.edu.cnSaraf,L.V.;Engelhard,M.;Ji,L.;Zhang,J.;etal.ManipulatingSurfaceReactionsinLithium-SulphurBatteriesUsingHybridAnodeAuthorsStructures.Nat.Commun.2014,5,3015.YueLiu−InstituteofFunctionalNanoandSoftMaterials(10)Shi,F.;Pei,A.;Boyle,D.T.;Xie,J.;Yu,X.;Zhang,X.;Cui,Y.(FUNSOM),SoochowUniversity,Suzhou215123,ChinaLithiumMetalStrippingBeneaththeSolidElectrolyteInterphase.QintaoSun−InstituteofFunctionalNanoandSoftMaterialsProc.Natl.Acad.Sci.U.S.A.2018,115,8529−8534.(FUNSOM),SoochowUniversity,Suzhou215123,China(11)Scheers,J.;Fantini,S.;Johansson,P.AReviewofElectrolytesPeipingYu−InstituteofFunctionalNanoandSoftMaterialsforLithium−SulphurBatteries.J.PowerSources2014,255,204−218.(FUNSOM),SoochowUniversity,Suzhou215123,China(12)Ding,F.;Xu,W.;Chen,X.;Zhang,J.;Engelhard,M.H.;Zhang,YuWu−InstituteofFunctionalNanoandSoftMaterialsY.;Johnson,B.R.;Crum,J.V.;Blake,T.A.;Liu,X.;etal.Effectsof(FUNSOM),SoochowUniversity,Suzhou215123,ChinaCarbonateSolventsandLithiumSaltsonMorphologyandCoulombicLiangXu−InstituteofFunctionalNanoandSoftMaterialsEfficiencyofLithiumElectrode.J.Electrochem.Soc.2013,160,A1894−A1901.(FUNSOM),SoochowUniversity,Suzhou215123,China(13)Xu,K.ElectrolytesandInterphasesinLi-ionBatteriesandHaoYang−InstituteofFunctionalNanoandSoftMaterialsBeyond.Chem.Rev.2014,114,11503−11618.(FUNSOM),SoochowUniversity,Suzhou215123,China;(14)Li,S.;Jiang,M.;Xie,Y.;Xu,H.;Jia,J.;Li,J.DevelopingHigh-orcid.org/0000-0002-8241-6231PerformanceLithiumMetalAnodeinLiquidElectrolytes:ChallengesMiaoXie−InstituteofFunctionalNanoandSoftMaterialsandProgress.Adv.Mater.2018,30,1706375.(FUNSOM),SoochowUniversity,Suzhou215123,China;(15)Yamada,Y.;Takazawa,Y.;Miyazaki,K.;Abe,T.Electro-orcid.org/0000-0002-9797-1449chemicalLithiumIntercalationintoGraphiteinDimethylSulfoxide-WilliamA.GoddardIII−MaterialsandProcessSimulationBasedElectrolytes:EffectofSolvationStructureofLithiumIon.J.Center,CaliforniaInstituteofTechnology,Pasadena,Phys.Chem.C2010,114,11680−11685.California91125,UnitedStates;orcid.org/0000-0003-(16)Yoshida,K.;Nakamura,M.;Kazue,Y.;Tachikawa,N.;Tsuzuki,S.;Seki,S.;Dokko,K.;Watanabe,M.Oxidative-StabilityEnhance-0097-5716mentandChargeTransportMechanisminGlyme-LithiumSaltCompletecontactinformationisavailableat:EquimolarComplexes.J.Am.Chem.Soc.2011,133,13121−13129.https://pubs.acs.org/10.1021/acs.jpclett.1c00279(17)Ueno,K.;Yoshida,K.;Tsuchiya,M.;Tachikawa,N.;Dokko,K.;Watanabe,M.Glyme-LithiumSaltEquimolarMoltenMixtures:NotesConcentratedSolutionsorSolvateIonicLiquids?J.Phys.Chem.BTheauthorsdeclarenocompetingfinancialinterest.2012,116,11323−11331.(18)Yamada,Y.;Yaegashi,M.;Abe,T.;Yamada,A.ASuper-■ConcentratedEtherElectrolyteforFast-ChargingLi-IonBatteries.ACKNOWLEDGMENTSChem.Commun.2013,49,11194−11196.T.C.andM.X.thankstheNationalNaturalScience(19)Zhang,X.Q.;Chen,X.;Cheng,X.B.;Li,B.Q.;Shen,X.;Yan,FoundationofChina(Grants21903058and22003044),theC.;Huang,J.Q.;Zhang,Q.HighlyStableLithiumMetalBatteries2928https://doi.org/10.1021/acs.jpclett.1c00279J.Phys.Chem.Lett.2021,12,2922−2929

7TheJournalofPhysicalChemistryLetterspubs.acs.org/JPCLLetterEnabledbyRegulatingtheLi+SolvationinNonaqueousElectrolyte.(39)Islam,M.M.;Kolesov,G.;Verstraelen,T.;Kaxiras,E.;vanAngew.Chem.,Int.Ed.2018,57,5301−5305.Duin,A.C.T.eReaxFF:APseudoclassicalTreatmentofExplicit(20)Jie,Y.;Ren,X.;Cao,R.;Cai,W.;Jiao,S.AdvancedLiquidElectronswithinReactiveForceFieldSimulations.J.Chem.TheoryElectrolytesforRechargeableLiMetalBatteries.Adv.Funct.Mater.Comput.2016,12,3463−3472.2020,30,1910777.(40)Liu,Y.;Yu,P.P.;Wu,Y.;Yang,H.;Xie,M.;Huai,L.Y.;(21)Wang,C.;Fu,K.;Kammampata,S.P.;McOwen,D.W.;Goddard,W.A.;Cheng,T.TheDFT-ReaxFFHybridReactiveSamson,A.J.;Zhang,L.;Hitz,G.T.;Nolan,A.M.;Wachsman,E.D.;DynamicsMethodwithApplicationtotheReductiveDecompositionMo,Y.;etal.Garnet-TypeSolid-StateElectrolytes:Materials,ReactionoftheTFSIandDOLElectrolyteataLithium−MetalInterfaces,andBatteries.Chem.Rev.2020,120,4257−4300.AnodeSurface.J.Phys.Chem.Lett.2021,12,1300−1306.(22)Wu,C.;Guo,F.;Zhuang,L.;Ai,X.;Zhong,F.;Yang,H.;Qian,(41)Kresse,G.;Hafner,J.AbinitioMolecular-DynamicsSimulationJ.MesoporousSilicaReinforcedHybridPolymerArtificialLayerforoftheLiquid-Metalamorphous-SemiconductorTransitioninGerma-High-EnergyandLong-CyclingLithiumMetalBatteries.ACSEnergynium.Phys.Rev.B:Condens.MatterMater.Phys.1994,49,14251−Lett.2020,5,1644−1652.14269.(23)Cheng,X.B.;Zhang,R.;Zhao,C.Z.;Wei,F.;Zhang,J.G.;(42)Plimpton,S.FastParallelAlgorithmsforShort-RangeZhang,Q.AReviewofSolidElectrolyteInterphasesonLithiumMolecularDynamics.J.Comput.Phys.1995,117,1−19.MetalAnode.Adv.Sci.2016,3,1500213.(43)Perdew,J.P.;Burke,K.;Ernzerhof,M.GeneralizedGradient(24)Markus,I.M.;Jones,G.;Garcia,J.M.InvestigationofApproximationMadeSimple.Phys.Rev.Lett.1996,77,3865−3868.ElectrolyteConcentrationEffectsonthePerformanceofLithium(44)Naserifar,S.;Oppenheim,J.J.;Yang,H.;Zhou,T.;Zybin,S.;OxygenBatteries.J.Phys.Chem.C2016,120,5949−5957.Rizk,M.;Goddard,W.A.AccurateNon-BondedPotentialsBasedon(25)Suo,L.;Hu,Y.-S.;Li,H.;Armand,M.;Chen,L.ANewClassofPeriodicQuantumMechanicsCalculationsforUseinMolecularSolvent-in-SaltElectrolyteforHigh-EnergyRechargeableMetallicSimulationsofMaterialsandSystems.J.Chem.Phys.2019,151,LithiumBatteries.Nat.Commun.2013,4,1481.154111.(26)Qian,J.F.;Henderson,W.A.;Xu,W.;Bhattacharya,P.;(45)Monkhorst,H.J.;Pack,J.D.SpecialPointsforBrillouin-ZoneEngelhard,M.;Borodin,O.;Zhang,J.G.HighRateandStableIntegrations.Phys.Rev.B1976,13,5188−5192.CyclingofLithiumMetalAnode.Nat.Commun.2015,6,6362.(46)Kresse,G.;Joubert,D.FromUltrasoftPseudopotentialstothe(27)Kim,H.;Wu,F.;Lee,J.T.;Nitta,N.;Lin,H.-T.;Oschatz,M.;ProjectorAugmented-WaveMethod.Phys.Rev.B:Condens.MatterCho,W.I.;Kaskel,S.;Borodin,O.;Yushin,G.InSituFormationofMater.Phys.1999,59,1758−1775.ProtectiveCoatingsonSulfurCathodesinLithiumBatterieswith(47)Fan,X.;Chen,L.;Ji,X.;Deng,T.;Hou,S.;Chen,J.;Zheng,J.;LiFSI-BasedOrganicElectrolytes.Adv.EnergyMater.2015,5,Wang,F.;Jiang,J.;Xu,K.;Wang,C.HighlyFluorinatedInterphases1401792.EnableHigh-VoltageLi-MetalBatteries.Chem.2018,4,174−185.(28)Camacho-Forero,L.E.;Smith,T.W.;Balbuena,P.B.Effectsof(48)Yun,K.-S.;Pai,S.J.;Yeo,B.C.;Lee,K.-R.;Kim,S.-J.;Han,S.HighandLowSaltConcentrationinElectrolytesatLithium−MetalS.SimulationProtocolforPredictionofaSolid-ElectrolyteInterphaseAnodeSurfaces.J.Phys.Chem.C2017,121,182−194.ontheSilicon-BasedAnodesofaLithium-IonBattery:ReaxFF(29)Kamphaus,E.P.;Angarita-Gomez,S.;Qin,X.;Shao,M.;ReactiveForceField.J.Phys.Chem.Lett.2017,8,2812−2818.Engelhard,M.;Mueller,K.T.;Murugesan,V.;Balbuena,P.B.Roleof(49)Ospina-Acevedo,F.;Guo,N.;Balbuena,P.B.LithiumInorganicSurfaceLayeronSolidElectrolyteInterphaseEvolutionatOxidationandElectrolyteDecompositionatLi-Metal/LiquidElectro-Li-MetalAnodes.ACSAppl.Mater.Interfaces2019,11,31467−31476.lyteInterfaces.J.Mater.Chem.A2020,8,17036−17055.(30)Sodeyama,K.;Yamada,Y.;Aikawa,K.;Yamada,A.;Tateyama,(50)Köhler,L.;Kresse,G.DensityFunctionalStudyofCOonY.SacrificialAnionReductionMechanismforElectrochemicalRh(111).Phys.Rev.B:Condens.MatterMater.Phys.2004,70,165405.(51)Qian,J.;Baskin,A.;Liu,Z.;Prendergast,D.;Crumlin,E.J.StabilityImprovementinHighlyConcentratedLi-SaltElectrolyte.J.AddressingtheSensitivityofSignalsfromSolid/LiquidAmbientPhys.Chem.C2014,118,14091−14097.PressureXPS(APXPS)Measurement.J.Chem.Phys.2020,153,(31)Wang,Y.;Liu,Y.;Tu,Y.;Wang,Q.ReductiveDecomposition044709.ofSolventsandAdditivestowardSolidElectrolyteInterphase(52)Zhang,P.;Zhu,J.J.;Wang,M.;Imanishi,N.;Yamamoto,O.FormationinLithium-IonBattery.J.Phys.Chem.C2020,124,LithiumDendriteSuppressionandCyclingEfficiencyofLithium9099−9108.Anode.Electrochem.Commun.2018,87,27−30.(32)Wang,A.;Kadam,S.;Li,H.;Shi,S.;Qi,Y.ReviewonModelingoftheAnodeSolidElectrolyteInterphase(SEI)forLithium-IonBatteries.npjComput.Mater.2018,4,15.(33)Camacho-Forero,L.E.;Smith,T.W.;Bertolini,S.;Balbuena,P.B.ReactivityattheLithium−MetalAnodeSurfaceofLithium−SulfurBatteries.J.Phys.Chem.C2015,119,26828−26839.(34)Bertolini,S.;Balbuena,P.B.BuildupoftheSolidElectrolyteInterphaseonLithium-MetalAnodes:ReactiveMolecularDynamicsStudy.J.Phys.Chem.C2018,122,10783−10791.(35)vanDuin,A.C.T.;Dasgupta,S.;Lorant,F.;Goddard,W.A.ReaxFF:AReactiveForceFieldforHydrocarbons.J.Phys.Chem.A2001,105,9396−9409.(36)Chenoweth,K.;vanDuin,A.C.T.;Goddard,W.A.ReaxFFReactiveForceFieldforMolecularDynamicsSimulationsofHydrocarbonOxidation.J.Phys.Chem.A2008,112,1040−1053.(37)Bedrov,D.;Smith,G.D.;vanDuin,A.C.T.ReactionsofSingly-ReducedEthyleneCarbonateinLithiumBatteryElectrolytes:AMolecularDynamicsSimulationStudyUsingtheReaxFF.J.Phys.Chem.A2012,116,2978−2985.(38)Islam,M.M.;Bryantsev,V.S.;vanDuin,A.C.T.ReaxFFReactiveForceFieldSimulationsontheInfluenceofTeflononElectrolyteDecompositionDuringLi/SWCNTAnodeDischargeinLithium-SulfurBatteries.J.Electrochem.Soc.2014,161,E3009−E3014.2929https://doi.org/10.1021/acs.jpclett.1c00279J.Phys.Chem.Lett.2021,12,2922−2929