Limits to Electrical Mobility in Lead-Halide Perovskite Semiconductors - Xia et al. - 2021 - Unknown

《Limits to Electrical Mobility in Lead-Halide Perovskite Semiconductors - Xia et al. - 2021 - Unknown》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/JPCLLetterLimitstoElectricalMobilityinLead-HalidePerovskiteSemiconductorsChelseaQ.Xia,JialiPeng,SamuelPoncé,JayB.Patel,AdamD.Wright,TimothyW.Crothers,MathiasUllerRothmann,JulianeBorchert,RebeccaL.Milot,HansKraus,QianqianLin,FelicianoGiustino,LauraM.Herz,andMichaelB.Johnston*CiteThis:J.Phys.Chem.Lett.2021,12,3607−3617ReadOnlineACCESSMetrics&MoreArticleRecommendations*sıSupportingInformationABSTRACT:Semiconductingpolycrystallinethinfilmsarecheaptoproduceandcanbedepositedonflexiblesubstrates,yethigh-performanceelectronicdevicesusuallyutilizesingle-crystalsemiconductors,owingtotheirsuperiorcharge-carriermobilitiesandlongerdiffusionlengths.Hereweshowthattheelectricalperformanceofpolycrystallinefilmsofmetal-halideperovskites(MHPs)approachesthatofsinglecrystalsatroomtemperature.Combiningtemperature-dependentterahertzconductivitymeasurementsandabinitiocalculationsweuncoveracompletepictureoftheoriginsofcharge-carrierscatteringinsinglecrystalsandpolycrystallinefilmsofCH3NH3PbI3.WeshowthatFröhlichscatteringofchargecarrierswithmultiplephononmodesisthedominantmechanismlimitingmobility,withgrain-boundaryscatteringfurtherreducingmobilityinpolycrystallinefilms.Wereconcilethelargediscrepancyincharge-carrierdiffusionlengthsbetweensinglecrystalsandfilmsbyconsideringphotonreabsorption.Thus,polycrystallinefilmsofMHPsoffergreatpromisefordevicesbeyondsolarcells,includinglight-emittingdiodesandmodulators.nthepastdecade,organic−inorganicmetal-halideperov-mobilityanddiffusionlength.Theelectricalmobility(μ)isIskite(MHP)semiconductorshaveemergedaspromisingdefinedasthedriftvelocityattainedbyachargecarrierperunit1−4ofappliedelectricfield,whilethediffusionlength(L)isthematerialsforphotovoltaicapplications,owingtotheireaseDoflarge-scaledepositionandexcellentoptoelectronicproper-averagedistanceachargecarriermovesbetweengenerationties,suchashighcharge-carriermobility,5−7longcarrierandrecombination.DespitemanymeasurementsofμandLDdiffusionlength,8−10andcomposition-tunablebandgap.11,12inMHPs,nouniversalagreementonvalueshasbeenachievedDownloadedvia157.100.74.25onMay14,2021at06:26:03(UTC).Todatetheprimaryapplicationofthesematerialshasbeenevenforthewell-studiedCH3NH3PbI3(MAPbI3).Thephotovoltaics.Inparticular,single-junctionsolarcellsbasedondiscrepanciesmayinpartbeattributedtodifferencesinthethesematerialshaveshownrapidgrowthinsolar-to-electricalpurity,stoichiometry,andmorphologyofthesamples.Forpowerconversionefficiency(PCE)from3.8%toover25%,example,differentfabricationroutesofperovskites,suchas19−22whereasperovskite−silicontandemsolarcellshavereachedantisolventone-stepspin-coatingmethods,air-blading2324,25Seehttps://pubs.acs.org/sharingguidelinesforoptionsonhowtolegitimatelysharepublishedarticles.efficienciesofover29%.1,13MHPsinglecrystalshaveattractedtechniques,vapor-assisteddeposition,all-vacuumse-2627intenseinterestbecauseoftheirpotentialinfabricatingquentialdeposition,andvacuumcoevaporationmethods,photodetectors,14X-rayscintillatorsanddetectors,15−17ascanleadtodifferenttypesandconcentrationsofimpuritiesaswellasvastlydifferentgrainsizeswithinperovskitethinwellasapplicationsinphotocatalysisandphotoelectrochemical22,25fields.18Ontheotherhand,thecutting-edgephotovoltaicfilms.However,significantdiscrepanciesmayalsobe28deviceshavebeenmostlydevelopedonaplatformoftracedbacktodifferentwaysofmeasuringμandLD.InthisworkweshowthatbecauseMHPsarehighlyluminescent,thepolycrystallinethinfilmsbecauseoftheireaseoffabrication.reabsorptionofphotonsemittedfromthesamplecanleadtoaWhilesuchthinfilmshaveshownremarkableperformanceinsignificantoverestimateofLD,andthateffectisparticularlysolarcells,thequestionremainsastowhattheupperlimitofstronginmeasurementsonsinglecrystals.Thus,wehelpperformancemightbeforperfectlycrystallinethin-filmdevicesandwhetherthistransitionisrequiredtofurtherexpandtheapplicationsofMHPintoareassuchasopticalcommunica-Received:February25,2021tionsthatrequiresemiconductordevicesincludingopticalAccepted:March29,2021transmitters,modulators,anddetectorswithhighswitchingPublished:April6,2021speeds.Twoimportantfiguresofmeritforquantifyingtheintrinsicelectricalpropertiesofasemiconductorarecharge-carrier©2021TheAuthors.PublishedbyAmericanChemicalSocietyhttps://doi.org/10.1021/acs.jpclett.1c006193607J.Phys.Chem.Lett.2021,12,3607−3617

1TheJournalofPhysicalChemistryLetterspubs.acs.org/JPCLLetterFigure1.(a)PhotographoftheMAPbI3singlecrystal.(b)Top-downSEMimageofthesinglecrystal’s(100)facet.(c)XRDspectrumofthesinglecrystal’s(100)facet.(d)PhotographoftheMAPbI3thinfilm.(e)Top-downSEMimageofthethinfilm.(f)XRDspectrumofthethinfilm.Theindicesshowninpanelscandf,e.g.,(200)and(110),representthetypicallatticeplanesobservedinMAPbI3assignedaccordingtothe46previouslyreportedXRDpatternintetragonalphase.reconcilethewiderangeofvaluesforchargerecombinationbecauseoftheabsenceofgrainboundaries.Surprisingly,parametersandchargediffusionlengthspreviouslyreportedhowever,themobilityofMAPbI3singlecrystalshasbeenfromMHPsinglecrystals,aswellasdiscrepanciesbetweenreportedtocoveranevenwiderrange(0.7−600cm2V−1s−1)10,28,41singlecrystalsandpolycrystallinethinfilms.thanforthinfilms.Conventionally,μofasemiconductorisextractedfromInthisworkwestudytheelectricalpropertiesofthe9,29devicesviatheHalleffect,space-chargelimitedcur-prototypicalMHPMAPbI3asbothsinglecrystalsand9,10,30,319rent,time-of-flightmeasurements,orfieldeffectpolycrystallinethinfilmsandcomparedirectexperimental32,33transistorcharacterization.However,thestrongpolar-measurementsofμwithabinitiocalculationsoftransport34−36izabilityofMHPsowingtoionmigrationandthecoefficientsbasedontheBoltzmanntransportequation(BTE)complexityofmakingsuitablecontactstoMHPscanandGWquasiparticlebandstructures.Themeasuredtemper-complicatethecalculationofmobilityfromthedataacquiredaturedependenceofthesingle-crystalmobilityshowsclose43viathesetechniques.Asaresult,awiderangeofelectricalagreementwiththemobilitycalculatedbysolvingtheBTE.mobilitieshasbeenreportedforMAPbI3.AliteraturesurveyofAsignificantdifferenceinthemobilitytemperaturedepend-measuredmobilitiesforMAPbI3isgiveninTableS1,whereitenceisobservedbetweenpolycrystallineandsingle-crystalcanbeseenthatreportedmobilityvaluesforsinglecrystalsofmorphologies,whichweattributetocharge-carrierscatteringMAPbI3areparticularlyinconsistent.fromcrystallographicgrainboundaries.Indeed,byexpandingAnalternativeapproachistouseanoncontactmethodtoourBTEtheorytoincludesuchgrain-boundaryscatteringwedeterminetheintrinsicelectricalmobilityofamaterial,viaareabletoreplicateallourexperimentaldata.Furthermore,we37−40techniquessuchasmicrowaveconductivityandterahertzreconcilethelargevariationinthereportedvaluesofμ,LD,and6,7,41(THz)spectroscopy.Thesetechniquesusetime-varyingchargerecombinationconstantsforsingle-crystalMHPsbyelectricfieldstravelinginfreespaceorawaveguidecavitytoaccountingforthesignificantinfluenceofphotonreabsorption.perturbandprobetheresponseofchargecarriersinamaterial.Thus,wehaveperformedadirectexperimentalcomparisonofThisisadvantageousastheinfluenceofametalliccontactontheelectricalpropertiesofsingle-crystalandpolycrystallinethematerialisremovedandthehigh-frequencyelectricfieldsMHPs,whichhasallowedustodevelopaholisticmodelthatofmicrowaveorTHzprobesavoidthecomplicationspredictsdevice-specificparameterssuchasμandLDofMHPsassociatedwithionmigration,whichoccursonamuchlongerinarangeofmorphologiesoverawidetemperaturerange.timescale.THzspectroscopyhastheaddedadvantagethatUnlikemostconventionalsemiconductorssuchasSiandconductivitycanbeobservedonasub-100fstimescale,GaAs,wefindthatthetransitionfromsinglecrystalstoallowingcharge-carrierdynamicstobefollowedandpolycrystallinefilmsresultsinremarkablylittledegradationto42recombinationparameterstobeextracted.theelectricalpropertiesofMHPs,whichindicateshighlyNonetheless,evenwithintheTHzspectroscopyregime,thebenigngrainboundariesinthesematerials.charge-carriermobilityofMAPbI3thinfilmsatroomTodeterminethefundamentalupperlimittothemobilityoftemperaturehasbeenreportedtorangefrom8to35MAPbI3andhenceanswerthequestionsoftheinfluenceofcm2V−1s−1.5−7,28Owingtothelackofinformationaboutgrainboundariesandwhatmightbetheupperperformancethegrainsizeofthosefilms,itishardtoconductasystematiclimitofperfectlycrystallinethin-filmdevices,wechosetostudycomparisonbetweenthoseexperimentalresultsandtheoreticaltheelectronicpropertiesofsingle-crystalsamplesandcomparemodels.Therefore,onewouldexpectthatμmeasuredfromthemdirectlytothoseofhigh-performancepolycrystallinethinperovskitesinglecrystalswouldgivemuchcloseragreementfilms.MAPbI3singlecrystalsweregrownusingtheinversewiththeoryandbetweendifferentexperimentalstudies,temperaturecrystallizationtechnique(seeExperimentaland3608https://doi.org/10.1021/acs.jpclett.1c00619J.Phys.Chem.Lett.2021,12,3607−3617

2TheJournalofPhysicalChemistryLetterspubs.acs.org/JPCLLetterComputationalMethodsfordetailsofsamplegrowth).AphotographofoneofthecrystalsisshowninFigure1ashowinglarge(10mm×10mm)opticallyflatfacets.Ascanningelectronmicroscopy(SEM)micrographofthe(100)facetisdisplayedinFigure1b.Wealsogrewhigh-qualityMAPbI3polycrystallinethinfilmsonquartzsubstratesusingvaporcodeposition(seeExper-imentalandComputationalMethodsfordetailsofsamplegrowth).Itisimportanttostudythinfilmsthatarerepresentativeofthoseusedinhigh-efficiencydevices;thus,weusedthesamedepositionparametersutilizedtoproduce13single-junctionsolarcellswithover19%PCE.Oneofthehighlyreproducible600nmthickfilmsusedinthisstudyisshowninFigure1dandfeaturedanopticallyflatsurface,asistypicalofvapor-codepositedfilms.TheSEMimagedisplayedinFigure1erevealsadenseassemblyofgrains,whileamoredetailedanalysisofthegrainsizedistributionispresentedintheSupportingInformation(seeFigureS12),whichgivesameanlength-scaleof∼580nm.Itshouldbestressedthatthisvaluerepresentsanupperlimitofgrainsizeasinternal44misorientationandstrainarenotrevealedbySEManalysis,andsmallercrystalsmayalsobepresentbelowtheobservedsurface.ThissuggeststhatthegrainsizeobtainedfromtheSEMmeasurementplacesanupperlimitoftherealgrainsizeFigure2.PhotoconductivitydecaydynamicsofMAPbI3measuredatofpolycrystallinethinfilms.Anothermethodtoassesscrystalroomtemperature.(a)MeasurementofMAPbI3thinfilmperformedsizeistoexamineScherrerbroadeninginX-raydiffractionintransmissionmode.ReflectionmeasurementoftheMAPbI3thinfilmisshowninFigureS17oftheSupportingInformation.(b)(XRD)peaksresultingfromthefinitesizeofthesmallMeasurementofMAPbI3singlecrystalperformedonthe(100)facetcrystallites.Asexpected,theXRDpeaksmeasuredfromtheinreflectionmode.Theinsetsillustratetheexperimentalgeometrythinfilm(Figure1f)aresignificantlybroadenedcomparedwiththephotoexcitationandprobingTHzpulsesincidentnormallywiththoseofthesinglecrystal(Figure1c),withtheScherreronthesamplesurface.Thecirclesrepresenttheexperimentaldata,equationreturningacrystallitesizefortheMAPbI3thinfilmofandthesolidcurvesrepresentthetheoreticalfits.Thenumerical∼30nm.Thisvaluerepresentsalowerlimitofcrystalsizeasvaluesofphotoconductivity,Δσ,weredeterminedfromthemeasuredthecontributionsofstrainanddisordertobroadeninganXRD−ΔT/Tand−ΔR/RusingeqsS4andS20givenintheSupportingpeakwidtharenotincludedintheScherrerequation.45Thus,Information.thetruelateralextentofcrystalsinthepolycrystallinethinfilmsisexpectedtobe∼100nmandwithintheboundsof30−580thatofthesingle-crystalMAPbI3sample,itdoesnotshowthenm.threeorder-of-magnitudedropinmobilityseenbetweenTheelectricalmobilityandchargerecombinationdynamicssingle-crystalandpolycrystallineGaAs,47,48indicatingthatofthin-filmandsingle-crystalsampleswererecordedusingthegrainboundariesmaybemorebenigninMHPs.Moreover,techniqueofoptical-pump−terahertz-probespectroscopybecausetheTHzmeasurementsaresensitivetothesurface(OPTPS).Thesampleswerephotoexcitedbyshort(35fs)conditions,thepresenceofsurfacedefectsonthesinglecrystalpulsesofbluelight(photonenergy3.1eV,centralwavelengthcanresultinanunderestimatedmobilityvalue.Aswillbe400nm)andphotoconductivity-probedwithasubpicosecondshownlater,theexperimentallymeasuredmobilitiesofTHzpulse.PanelsaandbofFigure2showtheroom-MAPbI3singlecrystalarefoundtobesmallerthanthetemperaturephotoconductivityofthin-filmandsingle-crystaltheoreticalvaluescalculatedbytheBTE,whichisattributedtoMAPbI3,respectively,asafunctionoftimeafterphoto-theeffectofsurfacedefectsandimpuritiesinthesinglecrystal.excitationbylaserpulseswithfluencesrangingfrom4.6to71UnderstandingchargecarrierrecombinationdynamicsisμJcm−2.Thecombinedelectronandholemobility,μ+μ,atehimportantformodelingsemiconductordevices,forexample,roomtemperaturewasfoundtobe(33±2)cm2V−1s−1forallowingpredictionofsolarcellPCEs,laserthresholdcurrents,thethinfilmand(59±3)cm2V−1s−1forthesinglecrystalbyandtransistorswitchingtimes.ThethreeprimarymechanismsapplyingeqsS15andS23givenintheSupportingInformationbywhichelectronsandholescanrecombineinMHPsarebytothephotoconductivitydatarecordedimmediatelyafter(i)Shockley−Read−Hall(SRH),(ii)bimolecular,and(iii)photoexcitation(i.e.,attime=0nsinFigure2a,b).Adetailed49Augerrecombination.SRHrecombinationisusuallyaexplanationofhowthemobilitywasdeterminedfromtherawnonradiativeprocesswhichismediatedbydefectsortraps.experimentaldataisprovidedintheSupportingInformation.Thisprocessisparasitictotheperformanceofmanydevices,Becausethemobilitywasmeasuredatthepeakoftheforexampleleadingtoalossofcollectedphotocurrentinsolarphotoconductivitydecaycurve,priortothechargerecombi-cellsandhigherthresholdcurrentsinlasers.Thus,theaimisnation,diffusion,andphotonreabsorption,theexperimentaloftentominimizetheSRHrecombinationpath.TheSRHmobilityisindependentofthoseprocesses(however,whenprocessisproportionaltothedefectdensityandscaleslinearlyanalyzingthecharge-carrierdecaydynamicsaftert=0,itiswiththecharge-carrierdensity,soitisstronglyinfluencedbycrucialtotakeintoaccounttheeffectofsuchcharge-carrierthepurityofasampleandisthedominantrecombinationdiffusionandphotonreabsorption,aswillbediscussedlater).pathwayforlowcharge-carrierdensities.TheothertwoWhilethecharge-carriermobilityofthethinfilmisalmosthalfprocesses,bimolecularandAugerrecombinations,scale3609https://doi.org/10.1021/acs.jpclett.1c00619J.Phys.Chem.Lett.2021,12,3607−3617

3TheJournalofPhysicalChemistryLetterspubs.acs.org/JPCLLetterquadraticallyandasthecubeofthecharge-carrierdensityn(t,z),inMAPbI3asafunctionoftime,t,afterphotoexcitationrespectively.Theseprocessesaremainlyinfluencedbyanddepth,z,fromthesamplesurface.Themeasuredunderlyingperiodiccrystalstructureandthusdependlessonphotoconductivity(Δσ)wasrelatedton(t,z)viaΔσ(t)=thepresenceofdefectsthantheSRHpath.42Bimolecularμe∫zmaxn(t,z)dz,wherezistheinverseoftheabsorptionz=0maxrecombinationistherecombinationofanelectron−holepair,coefficientofthesampleatthelaserwavelength(400nm).WeisgenerallyaradiativeprocessindirectbandgapMHPs,andmodeledboththethin-filmandthicksingle-crystaldatainshouldideallydominateifcreatingefficientLEDsorlasers.Figure2withtherateequationAugerrecombinationissignificantonlyatveryhighcharge-carrierdensitiesandisparasiticformostoptoelectronic2∂n∂n23devices.Asthedominantmechanismbywhichcharges=D2+−−Gk12nknk−3n(1)∂t∂zrecombineisstronglydependentoncharge-carrierdensity,howchargedensitychangesafterthesuddeninjectionofahighwherek1representsthemonomolecularchargerecombinationdensityofelectron−holepairsallowsthesignificanceofeachofrate,whichismainlyaresultofSRHtrap-mediatedthesethreerecombinationmechanismstobequantified.recombinationinMHPs;k2representstheradiativeHence,photoinjectingchargeandobservingchargerecombi-bimolecularrecombinationconstant;k3istheAugernationviathetimeevolutionofphotoconductivityisarecombinationconstant.Disthediffusioncoefficientpowerfultoolforquantifyingkeycharge-recombinationdeterminedbythecharge-carriermobility,andGisthe42parameters.charge-generationrate,whichincludesnotonlythechargeTheevolutionofphotoconductivityinpolycrystallineandcarriersgeneratedbyinitialphotoexcitationbutalsothesingle-crystalMAPbI3asafunctionoftimeafterphoto-chargesgeneratedbyphotonreabsorptionafterradiativeexcitation(andhencechargedensity)aredisplayedinpanelsabimolecularrecombination.50ThecirclesinFigure2representandbofFigure2,respectively.Observationofsuchdecaysintheexperimentaldata,whilethesolidcurvesareglobalfitsofphotoconductivityafterpulsedphotoexcitationallowschargeeq1.Fulldetailsofthemodelwhichincludesaone-recombinationconstantsfortheSRH,bimolecular,andAugerdimensionalfinite-difference-time-domainsolvercanbefoundmechanismstobedetermined.WhiletheserecombinationintheSupportingInformation.constantsareimportantbythemselvesfordevicemodeling,ThephotoconductivitydynamicsoverthetimeandfluencetheyalsoallowthediffusionlengthLDofchargecarriersintherangeoftheexperimentspresentedinFigure2areexpectedtomaterialtobedeterminedforanycharge-carrierdensityinthebedominatedbybimolecular(radiative)recombination,and42semiconductor.Asexpected,thedecayofphotoconductivitythelargecontrastinphotoconductivitydecaybetweentheforthin-filmMAPbI3isshowninFigure2aandisseentosinglecrystalandthinfilmdataatfirstsightindicatesalargedependstronglyonthephotoexcitationfluence,i.e.,thedifferenceintheirradiativerecombination.However,ondensityofphotoinjectedelectron−holepairs.Thisphenom-applyingthefullrate-equationmodelgivenineq1,itisenonhasbeenobservedpreviously,andrateequationswereclearthattheslowerphotoconductivitydecayofthethick5,7,50usedtoextracttherecombinationconstants.singlecrystalarisesprimarilyfromphotonreabsorption,withIncontrasttothethin-filmdata,thedecayofphoto-theunderlyingbimolecularrecombinationbeingremarkablyconductivityseeninFigure2bforthesinglecrystalappearssimilartothatofthethinfilm.Theextractedbimolecularquitedifferent,despitetheexcitationconditionsbeingrecombinationconstantofthesinglecrystalwask2,crystal=8.7identical.Suchbehaviorinsinglecrystalshasbeenobserved×10−10cm3s−1,whichisofthesameorderofmagnitudeasbeforeandwasattributedtosignificantlyhighervaluesofthatofthethinfilm,k=2.6×10−10cm3s−1(seeTableS29,102,filmcharge-carrierlifetime,longdiffusionlengths,andhenceintheSupportingInformationformoredetails).Thephysicalpotentiallymuchbetterdeviceperformance.However,weoriginofthesmalldropinthebimolecularrecombinationshowthatthefundamentalrecombinationparametersunder-constantinthethinfilmcomparedwiththesinglecrystalislyingthesedecaycurvesareremarkablysimilarbetweenthelikelytobeassociatedwithasmalldegreeofelectron−holethinfilmandsinglecrystaldespitethelargedifferencesseparationatgrainboundariesinthethinfilms.Thus,wefindbetweenthedatainFigure2a,b.Indeed,theextendedthatthebimolecularrecombinationinsingle-crystalMAPbIis3photoconductivitydecaycurvesofsingle-crystalMAPbI3consistentwiththethin-filmvaluemeasuredinthisstudyandshowninFigure2bariseprimarilyfromphotonreabsorption,alsowithpreviouslyreportedvaluesofotherMAPbI3thinaprocesswherethephotonsgeneratedintheradiativefilms.7,50bimolecularrecombinationprocessarereabsorbedbytheOneofthemostdramaticdifferencesbetweenthin-filmand13sample,givingrisetonewelectron−holepairs,thussingle-crystalMAPbI3hasbeenthemuchlargercharge-carrierextendingtheobservedphotoconductivitydecay.Theeffectdiffusionlength(LD)observedforthethicksinglecrystals,ofphotonreabsorptiononthecharge-carrierdynamics950whichisreportedtodifferbyafewordersofmagnitude.dependsstronglyonthesamplethickness.ThethickertheHowever,afteraccountingfortheeffectsonphotonsampleis,themorelikelythephotonsaretobereabsorbedreabsorptioninourmeasurements,wefindamuchsmallerbeforeescapingthematerial,therebyprolongingthedecayofdifferencebetweenLDvaluesforsingle-crystalandpolycrystal-photoconductivity.Consequently,photonreabsorptioncanlineMAPbI3.Wedefinethediffusionlengthastheaverageprolongthedecayofphotoconductivityinopticallythickdistancetraveledbyachargecarrierbetweengenerationandsamples.recombination/trappingatauniformcarrierdensitynintheTherefore,todetermineaccuratelythechargerecombina-absenceofanelectricfieldtionconstantsforthin-filmandsingle-crystalMAPbI3fromexperimentaldataitisnecessarytoconsiderhowphotonDreabsorptionandchargediffusionaffectthedata.WeusedtheLnD()=approachofCrothersetal.50tomodelcharge-carrierdensity,Rn()(2)3610https://doi.org/10.1021/acs.jpclett.1c00619J.Phys.Chem.Lett.2021,12,3607−3617

4TheJournalofPhysicalChemistryLetterspubs.acs.org/JPCLLetterFigure3.(a)Charge-carriermobilityofMAPbI3asafunctionoftemperature,wheretheblueandredcirclesrepresenttheexperimentaldataofthinfilmandsinglecrystal,respectively.Eachdatapointwasmeasuredrepeatedlythreetimes,fromwhichtheerrorbarwasdeterminedbythestandarddeviation.TheorangelinerepresentstheabinitioBTEcalculationofintrinsicphonon-limitedmobilitywhereallthephononmodesareincluded,whilethegreenlinerepresentstheBTEcalculationincludinggrain-boundaryscatteringwithcrystalsizeg=100nm.(b)MeanfreepathofpureMAPbI3singlecrystalintheorthorhombicphase.(c)IntrinsicmobilityofMAPbI3singlecrystalobtainedbysolvingtheBTEintheself-energyrelaxationtimeapproximation.whereR(n)=n2k+nk+kisthetotalrecombinationratestatesfromdensity-functionalperturbationtheory32153−55andD=μkBT/eisthediffusionconstant.Thisfunctioncan(DFPT)wherethescatteringrateswereobtainedthrough56thusbedeterminedusingexperimentallydeterminedvaluesofamaximallylocalizedWannierfunctioninterpolationofthe57−59themobilityμandrecombinationconstantsk1,k2,andk3.first-principleselectron−phononmatrixelements.DetailsUsingthemobilityvaluesmeasuredatt=0,k2valuesextractedofthecalculationsaregivenintheExperimentalandfromthephotonreabsorptionmodeldescribedbyeq1andk1ComputationalMethodsandinapriorworkwherewe60valuesobtainedfromtime-resolvedphotoluminescencereportedtheaveragedmobilityμ=(μe+μh)/2.However,in(TRPL)measurements(seeFigureS13intheSupportingourTHzphotoconductivitymeasurement,theobservedInformation),thediffusionlengthsforthethinfilmandthemobilityisinfactthesumofelectronandholemobilities.singlecrystalare1and2.83μm,respectively,foracharge-Therefore,usingtheroom-temperatureelectronandholecarrierdensityconsistentwith1sunillumination(1kWm−2mobilitiescalculatedforperfect,defect-freesingle-crystalAM1.5-filteredlight).Lisplottedasafunctionofcharge-MAPbIwhereμ=33cm2V−1s−1andμ=50cm2V−1s−1,D3ehcarrierdensityforthin-filmandsingle-crystalMAPbI3inatotaltheoreticalelectricalmobilityμ=μe+μh=83FigureS15intheSupportingInformation.Asdiscussedinthecm2V−1s−1isobtained.ThisvalueshouldrepresentanupperSupportingInformation,ifphotonreabsorptionisneglectedinlimitofanymeasurementonanimperfectrealcrystal,whichthedeterminationofk2,thenLDisartificiallylonger,asitwillcontainimpurities,defects,andsurfaces.Thus,thevalueofincludesonaveragemorethanonegeneration−recombinationmobilitymeasuredforourrealcrystal(59±3)cm2V−1s−1eventandwillbedependentonthethicknessofthesample.agreeswellwiththevaluecalculatedusingthefirst-principlesSuchoverestimateofthesinglecrystal’sdiffusionlengthistechnique.Wenotethatourcomputedmobilityisinlinewithparticularlysignificantatlowvaluesofk(<107s−1)wherethetheonecomputedinref61butlowerthanthatofrefs62−661bimolecularrecombinationbecomesmoreprominent.There-whichneglectedtheroleofmultiphononFröhlichcouplingasfore,aninaccuratedeterminationofthebimoleculardiscussedinref67.Thisiscrucialbecauseitwasshownthatatrecombinationconstantwillleadtoanunrealisticallylongleasttwosetsofmodescontributetothecharge-carrier60diffusionlengthforthesinglecrystal.Theseresultsindicatethemobilityofhalideperovskites.importanceofstatingcharge-carrierdensityaswellastakingSpecifically,wefindthatcharge-carrierscattering(andhenceintoaccountphotonreabsorptionwhenreportingdiffusionmobility)inMAPbI3isprimarilyinfluencedbythreelengthsforMAPbI3anddirectbandgapsemiconductorslongitudinaloptical(LO)phononmodes:(i)aPb−I−Pbgenerally.Meanwhile,itisworthnotingthatsomeotherbendingwitha4.3meVenergy,(ii)adominantPb−Ifactorsmayalsogiverisetothedifferentdiffusionlengthsstretchingmodeat14.4meVthataccountsforhalfoftheobservedinsinglecrystalsandthinfilms,suchascapacitivescattering,and(iii)abroadcontributionaround21meVcharginganddischargingeffects,defects,andimpuritiesoriginatingfromthelibrationalmodesoftheCH3NH3presentinthesamples.molecules.Thus,includingmultiplephononmodesinInordertogainabetterunderstandingofthefundamentalcalculationsofcharge-carrierscatteringreconcilestheover-limitstointrinsicelectricalmobilityandcharge-carrierestimateofthetheoreticalmobilityinMAPbI3comparedwithdiffusionlengthinMAPbI3,wesolvedfortheorthorhombicexperimentalvalues.43phaseofMAPbI3theabinitioBTEintheself-energyTotestthevalidityofourmultiple-phonon-modeBTE51relaxationtimeapproximationincludingspin−orbitcouplingmodelwecomparedmeasuredandcalculatedvaluesforeffectsandwithelectronicstatescomputedwithaneigenvalueelectricalmobilityofMAPbI3overawidetemperaturerange.52self-consistentmany-bodyGWmethod,vibrationaleigen-Figure3adisplaysthemeasuredelectricalmobilityofsingle-3611https://doi.org/10.1021/acs.jpclett.1c00619J.Phys.Chem.Lett.2021,12,3607−3617

5TheJournalofPhysicalChemistryLetterspubs.acs.org/JPCLLettercrystalMAPbI3(redcircles)overatemperaturerangeof75−wherewkisthek-pointweight,fnktheFermi−Dirac310Kincomparisontoourmultiphonon-modeBTEoccupationfunction,vnkthecarriervelocity,andτnkthecalculations(orangeline).Astemperatureislowered,theelectron−phononlifetimeforastateofbandnandmomentummobilityincreases,whichisconsistentwithreducedoccupancyk.Themeanfreepathaverageforelectronsandholesisshownofphononmodesandhencereducedelectron−phononinFigure3b.Forcomparison,themeanfreepathcalculatedforFröhlichcoupling.themostrelevantenergy(3/2)kBTisgivenintheSupportingExperimentalmobilitydataareoftenfittedwithatemper-Information(seeFigureS9).ThegreenlineinFigure3ashowsaturepowerlawrelationtohelpdeterminethephysicalorigintheabinitiocalculatedμfilm,withcrystalgrainsizegastheonly28,65ofscattering.However,simplepowerlawmodelsofLO-freeparameter.Agrainsizeg=100nm,whichasdiscussedphononcouplingassumeonlyoneLOphononbranchandearlierisreasonableforoursamples,wasfoundtobeinsimplifieddispersionrelations,whichareclearlynotapplicableexcellentagreementwithexperimentalmobilitydataforforMAPbI3whichhasmanyatomsinitsprimitiveunitcelltemperaturesabove30K.Below30Kexcitonformation,basisandhenceadensephononspectrum.Thus,theonlywaywhichisnotincludedinthetheoreticalmodel,isexpectedtotoproperlyaccountforthetemperaturedependenceofdominateandisconsistentwiththereducedexperimental71electricalmobilityinMHPsistoconsidertheoccupancyandmobilityatlowtemperatures.TheeffectofincreasingandFröhlichcouplingofallsignificantphononmodesasafunctionreducinggrainsizeonmobilityisdisplayedinFigureS10inoftemperature,aswehavedonewithourBTEmodel.theSupportingInformationwithmobilitybeingmostsensitiveThetemperaturedependenceofelectronandholemobilitiestograinsizeattemperaturesbelow200K.Thus,ourBTEcalculatedbysolvingtheBTEforapureMAPbI3crystaltheory,whencorrectedforgrain-boundaryscatteringprovideswithoutanyimpuritiesisshowninFigure3c,whiletheanexcellentpredictionofelectricalmobilityinMAPbI3forcombinedelectronandholemobility(orangeline)ispolycrystallineandsingle-crystalmorphologiesoverawidecomparedwiththeexperimentalmobilitydata(redcircles)temperaturerange.Weemphasizethatthissettlesalong-inFigure3a.Weexpectthatthecalculatedmobilitywillstandingdebateintheliteraturewherebyacousticscatter-72,7374representanupperlimitforthemobilityofanyrealcrystal,ing,singleopticalmodescattering,ionizedimpurity757664owingtothepresenceofimpuritiesinrealcrystals.Indeedthescattering,piezoelectricscattering,orpolaronicscatteringexperimentalsingle-crystaldata(redcircles)agreeswellwithwereallmentionedascontributingtothecarriermobilityintheshapeofthetheoreticaldataandisboundedbyit.TheMAPbI3.Herewehaveshownthatonlymultimodeoptical-overestimateissomewhathigheratlowertemperatures,wherephononscatteringforasinglecrystalaugmentedbygrain-ionizedimpurityscattering,whichisnotincludedintheBTEboundaryscatteringforpolycrystallinethinfilmissufficient.calculations,becomesmoresignificant.Inconclusion,weperformedacomprehensiveexperimentalIncontrast,thevaluesofexperimentalmobilityareandtheoreticalstudyoftheelectricalpropertiesofbothsingle-significantlylowerforthepolycrystallinethinfilm(bluecirclescrystalandpolycrystallineMAPbI3.WereconcilethelargeinFigure3a)comparedwiththesinglecrystal,andimportantlydiscrepancyinpreviouslypublishedvaluesofkeyfiguresofthediscrepancybecomeslargeratlowertemperature.Wemeritsuchasmobility,diffusionlength,andrecombinationhypothesizethatthisdifferenceisrelatedtothepresenceofparametersbyincludingtheeffectsofphotonreabsorption.Ingrainboundariesinthepolycrystallinethinfilmsandtestthisparticular,wefindthatneglectingphotonreabsorptionwhenhypothesisbymodelinggrain-boundaryscatteringwithinourmodelingthicksinglecrystalsleadstoasignificantover-BTEframework.estimateofcharge-carrierdiffusionlength.OurexperimentalTogaininsightintotheinfluenceofgrainboundariesonthedataagreeextremelywellwithabinitioBoltzmanntransportelectricalmobilityofpolycrystallineMAPbI3weimplementedcalculationswhentheFröhlichinteractionofmultiplephonon68themodelofMayadasandShatzkeswherethemeanfreemodesareincluded.Thisresultexplainstheoverestimationofpathofchargecarriersislimitedbytheextentofgrainmobilityinpreviouscalculationswhereonlyonephononboundaries.ThemodelisanextensionoftheBoltzmannbranchwasincluded.TheBTEmodelprovidedexcellenttransporttheorytoincludereflectionofthechargecarriersatagreementwithsingle-crystalmobilitydatawithoutanyfittingthegrainboundariesofpolycrystallinethinfilmsandcanbeparameters.However,themeasuredmobilityofpolycrystalline68writtenasthinfilmsdeviatedfromthesingle-crystaldataandBTEmodel.ÄÅÅÉÑÑWefoundthatgrain-boundaryscatteringaccountedforthisμ=−μαÅÅÅÅ13+33α23−+αlnijj11yzzÑÑÑÑdeviationandincludedthiseffectinourBTEmodel.WhilethefilmphononÅÅ2jkαz{ÑÑ(3)mobilityofpolycrystallinefilmsisprimarilylimitedbygrain-ÇÖboundaryscatteringatcryogenictemperatures,wefoundthatwhereα=R(λ/g)/(1−R);λisthecarriermeanfreepath,Rroom-temperaturemobilityisdominatedbyLO-phonontheprobabilityofreflectionatthegrainboundary,andgthescattering.Assuchwefindthemobilitydroppingonlyfrom59cm2V−1s−1inthesinglecrystalto33cm2V−1s−1forgrainsize.μfilmandμphononrepresenttheelectricalmobilitycalculatedbytheBTEmodelwithandwithoutthepolycrystallineMAPbI3indicatingthebenignnatureoftheseconsiderationofgrain-boundaryscattering,respectively.grainboundariesatroomtemperature.Followingpreviouswork,wesettheprobabilityofreflectionOverallthisstudyprovidesacompletepictureoftheatR=0.5,69,70andthecarriermeanfreepathiscomputedabfundamentalelectricalpropertiesofthemodelMHPMAPbI3initiobydirection-averagingthecarriermeanfreepathinbothsingle-crystalandpolycrystallinemorphologies.Weweightedbyatransportoccupationthusunifyapparentlycontradictorypreviousexperimentalandtheoreticalstudies.Ourresultsindicatethatpolycrystallinethinα∑wf(1−||f)vτfilmspossessperformancesimilartosinglecrystalsattheusualα,nkknnkknnkkλtransport=operatingtemperatureofelectronicdevices.Thisispromising3(∑nkwfknnkk1−f)(4)forfutureapplicationsofMHPthinfilmsinhigh-speeddevices3612https://doi.org/10.1021/acs.jpclett.1c00619J.Phys.Chem.Lett.2021,12,3607−3617

6TheJournalofPhysicalChemistryLetterspubs.acs.org/JPCLLettersuchastransistors,emitters,modulators,anddetectors,aswelldetailedderivationoftheelectricalmobilityisgivenintheasforupscalingfuturegenerationsofsolarcellsandlightingSupportingInformation.panels.ThermometryandTemperature-DependentPLSpec-troscopy.TheverylowthermalconductivityofMAPbI3has■beenreportedtobebelow0.5W(mK)−1,79−81whichisEXPERIMENTALANDCOMPUTATIONALMETHODShundredsoftimessmallerthanthatofconventionalsemi-82conductorssuchasGaAs.Therefore,extremecaremustbeMAPbI3Single-CrystalFabrication.TheMAPbI3perov-takenwhenmakingtemperature-dependentmeasurements,skitesinglecrystalswerepreparedbyinversetemperatureparticularlyifsamplesareexposedtolocalizedheating,suchascrystallization.30,77MAPbIprecursor(1.25molL−1)was3vialaserexcitation.Forthin-filmMAPbI3depositedonapreparedbyaddingPbI2(461mg)andmethylammoniumquartzsubstrate,thelowthermalconductivityisnegatedbytheiodide(CH3NH3I,159mg)intoγ-butyrolactone(0.8mL),material’sclosecontactwiththequartzsubstrate,whichhasaheatedat60°Cfor2hwithstirring.Theprecursorswerehighthermalconductivityoverawidetemperaturerange.filteredwithsyringefilters(0.22μmporesize).TheobtainedUnfortunately,forthecentimeter-sizedMAPbIsinglecrystal3solutionwastransferredtocleancontainers,whichwerekeptinourstudythelowthermalconductivitymakestheonastablehotplateandgraduallyheatedto120°Candkeptexperimentsextremelychallenging,andgreatcarewastakenforanother6h.Crystalswereformedatthebottomofthetoensurethattherecordedtemperaturewasactuallyitslatticecontainers.Finally,thecrystalswerecollectedanddriedat60temperatureatthepositionwherethemeasurementswere°Cinvacuumovenfor12h.made.Therefore,wedevelopedaPL-correctedTHztechniqueMAPbI3Thin-FilmFabrication.TheMAPbI3thinfilmstomeasurethemobilityoftheMAPbI3singlecrystalwithwerepreparedintwosteps.(1)Cleaningofsubstrates:z-cutaccuratetemperaturedetermination.Toimprovethethermalquartzsubstrateswerecleanedwithhellmanexsolution,contactbetweenthesinglecrystalandthecoldfingercryostatfollowedbyathoroughrinsewithdeionizedwater.The(OxfordInstruments,MicrostatHe),whichwasusedtochangesubstrateswerethenwashedwithacetone,isopropanol,andthecrystaltemperatureinourOPTPSmeasurement,weethanol.ThereafterthesubstrateswereplasmaetchedinO2forinsertedasapphiresubstrateatthefrontofthesinglecrystal,10min.(2)ThermalcoevaporationofMAPbI3:theMAPbI3sothattheheatgeneratedbythepumpbeamcouldbewasfabricatedusingthermalevaporationasreporteddissipatedmoreefficiently,whichenabledustocoolthecrystal13previously.Inbrief,MAIandPbI2wereplacedinseparateto75KandobserveaclearphasetransitioninitsPLspectrumcrucibles,andthesubstratesweremountedonarotatingat160K.Inthemeantime,todeterminethecrystalsubstrateholdertoensurethatauniformfilmwasdeposited.temperaturemoreaccurately,wemeasuredthecorrespondingThetemperatureofthesubstrateswaskeptat21°CPLspectrumatdifferenttemperaturesusingbothcoldfingerthroughoutthedeposition.Thechamberwasevacuatedtoandgas-exchangecryostats.Inthegas-exchange-cryostatsetupreachahighvacuumof10−6mbarbeforethePbIandthe2(OxfordInstruments,OptistatCF2)whichisseparatefromtheMAIwereheated.ThesubstrateswerethenexposedtotheOPTPSsetup,becausethecrystalwasimmersedinheliumgas,vapor.TheratesofboththeMAIandPbI2weremonitoredtherewasnothermalcontactissueandthetemperatureusingaquartzcrystalmicrobalance.Thethicknessoftheregisteredbythesensorinthegas-exchangecryostatwastheperovskitethinfilmwassetbycontrollingtheexposuretimeoftruetemperatureofthecrystal.Therefore,wewereabletousethesubstratestothevapor.thePLspectrummeasuredbythegas-exchangecryostatasaOptical-Pump−THz-ProbeSpectroscopy(OPTPS).referencetocorrectthetemperaturemeasuredbytheTheOPTPSwasutilizedtomeasurethephotoconductivitycoldfingercryostat.AdetailedPL-facilitatedtemperature-andtheelectricalmobilityofMAPbI3thinfilmsandsinglecorrectionprocessisgivenintheSupportingInformation.Incrystals.TheTHzpulsewasgeneratedbyaTHzspintronicthecoldfinger-cryostatsetup,thePLspectrumwasgenerated78emitterbecauseoftheinversespinHalleffect.Anamplifiedfromexcitationbythepumpbeam(400nm,35fs)usedintheultrafast(35fs)laserbeamwithanaveragepowerof4WandOPTPSsetupandcollectedbyafiber-coupledspectrometercentralwavelengthof800nmwassplitintothreearms:a(HoribaScientific,iHR320)anddetectedbyaCCD(Horibaprobe(THz)beam,agatebeam,andapumpbeam.TheTHzScientific,SiSymphonyII).Inthegas-exchange-cryostatsetup,pulsewasdetectedbya0.1mmthick(110)ZnTecrystalthePLspectrumwasgeneratedfromexcitationbyatogetherwithaWollastonprismandapairofbalancedpicosecondpulseddiodelaser(PicoHarp,LDH-D-C-405M)photodiodesviaelectro-opticsampling.Thepumpbeamwasatcentralwavelengthof398nmwiththesignalsubsequentlyconvertedfrom800to400nmbyaβ-barium-borate(BBO)collectedandcoupledintoadifferentspectrometer(PrincetoncrystaltophotoexcitetheMAPbI3thinfilmsandsingleInstruments,SP-2558)anddetectedbyaniCCD(Princetoncrystals.Underphotoexcitation,thephotoinjectedchargeInstruments,PI-MAX4).carriersgiverisetoareductionoftheTHztransmission,ComputationalMethodsBasedonDensity-Func-whichisproportionaltophotoconductivityandusedfortionalTheory.WeperformedDFTcalculationsusingextractingthecharge-carriermobility.Whilethephoto-pseudopotentialsandplanewaves,asimplementedinthe55conductivityofMAPbI3thinfilmwasmeasuredintrans-QuantumESPRESSOpackage.Weusedthelocaldensitymissionmode,thephotoconductivityofthesinglecrystalwasapproximation(LDA)withnorm-conservingpseudopotentials83measuredinreflectionmodeinsteadbecauselittleTHzsignalfromthePseudoDojorepository.Weusedfullyrelativisticcouldtransmitthroughthethickcrystal.Aschematicofthepseudopotentialswhichincludetheeffectofspin−orbitOPTPSsetupintransmissionandreflectionmodesisshownincouplingaswellassemicoreelectronsinthecaseofPb.WeFigureS1intheSupportingInformation.Allmeasurementsusedaplane-wavekineticenergycutoffof100Ryandthewererepeatedthreetimesateachtemperature,fromwhichthefollowingorthorhombiclatticeparameters:a=8.836Å,b=46uncertaintywasdeterminedbythestandarddeviation.A12.581Å,andc=8.555Å.Wecalculatedphononsusing3613https://doi.org/10.1021/acs.jpclett.1c00619J.Phys.Chem.Lett.2021,12,3607−3617

7TheJournalofPhysicalChemistryLetterspubs.acs.org/JPCLLetter53,54DFPTwith4×4×4k-pointsand2×2×2q-pointsAdamD.Wright−DepartmentofPhysics,Universityofgrids.WecorrectedtheDFTbandstructuresviatheOxford,ClarendonLaboratory,OxfordOX13PU,U.K.;52quasiparticleGWmethodusingtheYambocode.Weorcid.org/0000-0003-0721-7854employedahigherplane-wavekineticenergycutoffof150TimothyW.Crothers−DepartmentofPhysics,UniversityofRy;weevaluatedtheexchangeself-energyandthepolar-Oxford,ClarendonLaboratory,OxfordOX13PU,U.K.izabilityusingcutoffsof80and6Ry,respectively,andMathiasUllerRothmann−DepartmentofPhysics,Universityperformedthesummationsoveremptystatesusing1000bandsofOxford,ClarendonLaboratory,OxfordOX13PU,U.K.forthecalculationofthepolarizationandtheGreen’sfunction.JulianeBorchert−DepartmentofPhysics,UniversityofThefrequencydependenceofthescreenedCoulombOxford,ClarendonLaboratory,OxfordOX13PU,U.K.;interactionwasdescribedviatheGodby−Needsplasmon-orcid.org/0000-0001-7973-6907polemodel84usingaplasmon-poleenergyof18.8eV.BecauseRebeccaL.Milot−DepartmentofPhysics,UniversityoftheDFTgapofleadhalideperovskitesisverysmallbecauseofWarwick,CoventryCV47AL,U.K.;orcid.org/0000-spin−orbitcoupling,71wewentbeyondtheGWapprox-0003-0865-297000imationbyincludingself-consistencyontheeigenvalues.WeHansKraus−DepartmentofPhysics,UniversityofOxford,appliedself-consistencybyusingthestrategyofref60,whichOxfordOX13RH,U.K.includesawavevector-dependentscissorsoastoobtainQianqianLin−KeyLabofArtificialMicro-andNano-accurateeffectivemasses.TheBrillouinzonewassampledviaaStructuresofMinistryofEducationofChina,Schoolof4×4×4unshiftedgrid,andtheterminationschemeofref85PhysicsandTechnology,WuhanUniversity,Wuhan430072,wasemployedtoacceleratetheconvergencewithrespecttoP.R.China;orcid.org/0000-0002-6144-1761thenumberofemptystates.Wecalculatedtheelectron−FelicianoGiustino−DepartmentofMaterials,UniversityofphononmatrixelementsandscatteringratesusingtheEPWOxford,OxfordOX13PH,U.K.;OdenInstituteforcode,58inconjunctionwiththewannier90library.56WeComputationalEngineeringandSciencesandDepartmentofincludedspin−orbitcouplinginallcalculations.WestartedPhysics,UniversityofTexasatAustin,Austin,Texas78712,fromthe2×2×2gridofphononwavevectorsandUnitedStates;orcid.org/0000-0001-9293-1176interpolatedonafinegridcontaining100000pointsandLauraM.Herz−DepartmentofPhysics,UniversityofOxford,followingaΓ-centeredCauchydistributionweightedbytheirClarendonLaboratory,OxfordOX13PU,U.K.;Voronoivolume.Weneglectanharmoniceffectswhichcouldorcid.org/0000-0001-9621-334X61beimportantatroomtemperatureandabove.Completecontactinformationisavailableat:■https://pubs.acs.org/10.1021/acs.jpclett.1c00619ASSOCIATEDCONTENTNotes*sıSupportingInformationTheauthorsdeclarenocompetingfinancialinterest.TheSupportingInformationisavailablefreeofchargeathttps://pubs.acs.org/doi/10.1021/acs.jpclett.1c00619.■ACKNOWLEDGMENTSDetailsofTHzspectroscopysetup,derivationsofThisworkwasfundedbytheEngineeringandPhysicalcharge-carriermobilityfromOPTPSmeasurements,SciencesResearchCouncil(EPSRC).M.B.J.thankstheanalysisofgrain-boundaryscatteringeffectonelectricalAlexandervonHumboldtFoundationforsupport.S.P.mobility,analysisofphotonreabsorption,anddetailsofacknowledgesthesupportfromtheEuropeanUnion’sHorizonPL-facilitatedtemperaturecorrectiontechnique(PDF)2020ResearchandInnovationProgramme,undertheMarieSkłodowska-CurieGrantAgreementSELPH2DNo.839217.F.G.’scontributionwassupportedaspartoftheComputa-■tionalMaterialsSciencesProgramfundedbytheU.S.AUTHORINFORMATIONDepartmentofEnergy,OfficeofScience,BasicEnergyCorrespondingAuthorSciences,underAwardDE-SC0020129.MichaelB.Johnston−DepartmentofPhysics,UniversityofOxford,ClarendonLaboratory,OxfordOX13PU,U.K.;■REFERENCESorcid.org/0000-0002-0301-8033;(1)Kojima,A.;Teshima,K.;Shirai,Y.;Miyasaka,T.OrganometalEmail:michael.johnston@physics.ox.ac.ukHalidePerovskitesasVisible-LightSensitizersforPhotovoltaicCells.J.Am.Chem.Soc.2009,131,6050−6051.Authors(2)Loi,M.A.;Hummelen,J.C.HybridSolarCells:PerovskitesChelseaQ.Xia−DepartmentofPhysics,UniversityofOxford,undertheSun.Nat.Mater.2013,12,1087.ClarendonLaboratory,OxfordOX13PU,U.K.(3)Snaith,H.J.Perovskites:theEmergenceofANewEraforLow-JialiPeng−KeyLabofArtificialMicro-andNano-StructuresCost,High-EfficiencySolarCells.J.Phys.Chem.Lett.2013,4,3623−ofMinistryofEducationofChina,SchoolofPhysicsand3630.Technology,WuhanUniversity,Wuhan430072,P.R.China(4)Kim,H.S.;Im,S.H.;Park,N.G.OrganoleadHalidePerovskite:SamuelPoncé−DepartmentofMaterials,UniversityofNewHorizonsinSolarCellResearch.J.Phys.Chem.C2014,118,Oxford,OxfordOX13PH,U.K.;TheoryandSimulationof5615−5625.(5)Wehrenfennig,C.;Liu,M.;Snaith,H.J.;Johnston,M.B.;Herz,Materials(THEOS),ÉcolePolytechniqueFédéraledeL.M.Charge-CarrierDynamicsinVapour-DepositedFilmsoftheLausanne,CH-1015Lausanne,Switzerland;orcid.org/OrganoleadHalidePerovskiteCH3NH3PbI3−xClx.EnergyEnviron.Sci.0000-0003-1159-83892014,7,2269−2275.JayB.Patel−DepartmentofPhysics,UniversityofOxford,(6)Wehrenfennig,C.;Eperon,G.E.;Johnston,M.B.;Snaith,H.J.;ClarendonLaboratory,OxfordOX13PU,U.K.;Herz,L.M.HighChargeCarrierMobilitiesandLifetimesinorcid.org/0000-0001-5132-1232OrganoleadTrihalidePerovskites.Adv.Mater.2014,26,1584−1589.3614https://doi.org/10.1021/acs.jpclett.1c00619J.Phys.Chem.Lett.2021,12,3607−3617

8TheJournalofPhysicalChemistryLetterspubs.acs.org/JPCLLetter(7)Milot,R.L.;Eperon,G.E.;Snaith,H.J.;Johnston,M.B.;Herz,FilmsforSolarCellsPreparedbySingle-SourcePhysicalVapourL.M.Temperature-DependentCharge-CarrierDynamicsinDeposition.Sci.Rep.2016,6,29910.CH3NH3PbI3PerovskiteThinFilms.Adv.Funct.Mater.2015,25,(26)Hsiao,S.Y.;Lin,H.L.;Lee,W.H.;Tsai,W.L.;Chiang,K.M.;6218−6227.Liao,W.Y.;RenWu,C.Z.;Chen,C.Y.;Lin,H.W.EfficientAll-(8)Stranks,S.D.;Eperon,G.E.;Grancini,G.;Menelaou,C.;VacuumDepositedPerovskiteSolarCellsbyControllingReagentAlcocer,M.J.;Leijtens,T.;Herz,L.M.;Petrozza,A.;Snaith,H.J.PartialPressureinHighVacuum.Adv.Mater.2016,28,7013−7019.Electron-HoleDiffusionLengthsExceeding1MicrometerinAn(27)Momblona,C.;Gil-Escrig,L.;Bandiello,E.;Hutter,E.M.;OrganometalTrihalidePerovskiteAbsorber.Science2013,342,341−Sessolo,M.;Lederer,K.;Blochwitz-Nimoth,J.;Bolink,H.J.Efficient344.VacuumDepositedp-i-nandn-i-pPerovskiteSolarCellsEmploying(9)Dong,Q.;Fang,Y.;Shao,Y.;Mulligan,P.;Qiu,J.;Cao,L.;DopedChargeTransportLayers.EnergyEnviron.Sci.2016,9,3456−Huang,J.Electron-HoleDiffusionLengths>175μminSolution-3463.GrownCH3NH3PbI3SingleCrystals.Science2015,347,967−970.(28)Herz,L.M.Charge-CarrierMobilitiesinMetalHalide(10)Shi,D.;etal.LowTrap-StateDensityandLongCarrierPerovskites:FundamentalMechanismsandLimits.ACSEnergyLett.DiffusioninOrganoleadTrihalidePerovskiteSingleCrystals.Science2017,2,1539−1548.2015,347,519−522.(29)Stoumpos,C.C.;Malliakas,C.D.;Kanatzidis,M.G.(11)Even,J.;Pedesseau,L.;Jancu,J.M.;Katan,C.ImportanceofSemiconductingTinandLeadIodidePerovskiteswithOrganicSpin-OrbitCouplinginHybridOrganic/InorganicPerovskitesforCations:PhaseTransitions,HighMobilities,andNear-InfraredPhotovoltaicApplications.J.Phys.Chem.Lett.2013,4,2999−3005.PhotoluminescentProperties.Inorg.Chem.2013,52,9019−9038.(12)Even,J.;Pedesseau,L.;Katan,C.;Kepenekian,M.;Lauret,J.-S.;(30)Saidaminov,M.I.;etal.High-QualityBulkHybridPerovskiteSapori,D.;Deleporte,E.Solid-StatePhysicsPerspectiveonHybridSingleCrystalswithinMinutesbyInverseTemperatureCrystal-PerovskiteSemiconductors.J.Phys.Chem.C2015,119,10161−lization.Nat.Commun.2015,6,7586.(31)Rizvi,S.;Mazhari,B.AnImprovedMethodforExtractionof10177.(13)Patel,J.B.;etal.LightAbsorptionandRecyclinginHybridMobilityfromSpaceChargeLimitedCurrentinOrganicSemi-conductorFilms.J.Appl.Phys.2017,121,155501.MetalHalidePerovskitePhotovoltaicDevices.Adv.EnergyMater.(32)Chin,X.Y.;Cortecchia,D.;Yin,J.;Bruno,A.;Soci,C.Lead2020,10,1903653.IodidePerovskiteLight-EmittingField-EffectTransistor.Nat.(14)Lian,Z.;Yan,Q.;Lv,Q.;Wang,Y.;Liu,L.;Zhang,L.;Pan,S.;Commun.2015,6,7383.Li,Q.;Wang,L.;Sun,J.L.High-PerformancePlanar-Type(33)Zhang,M.J.;Wang,N.;Pang,S.P.;Lv,Q.;Huang,C.S.;Photodetectoron(100)FacetofMAPbI3SingleCrystal.Sci.Rep.Zhou,Z.M.;Ji,F.X.CarrierTransportImprovementof2015,5,16563.(15)Wei,J.H.;Liao,J.F.;Wang,X.D.;Zhou,L.;Jiang,Y.;Kuang,CH3NH3PbI3FilmbyMethylamineGasTreatment.ACSAppl.Mater.Interfaces2016,8,31413−31418.D.B.All-InorganicLead-FreeHeterometallicCs4MnBi2Cl12Perov-(34)Unger,E.L.;Hoke,E.T.;Bailie,C.D.;Nguyen,W.H.;skiteSingleCrystalwithHighlyEfficientOrangeEmission.MatterBowring,A.R.;Heumüller,T.;Christoforo,M.G.;McGehee,M.D.2020,3,892−903.HysteresisandTransientBehaviorinCurrent-VoltageMeasurements(16)Shrestha,S.;etal.High-PerformanceDirectConversionX-RayofHybrid-PerovskiteAbsorberSolarCells.EnergyEnviron.Sci.2014,DetectorsBasedonSinteredHybridLeadTriiodidePerovskite7,3690−3698.Wafers.Nat.Photonics2017,11,436−440.(35)Meloni,S.;etal.IonicPolarization-InducedCurrent-Voltage(17)Peng,J.;Xia,C.Q.;Xu,Y.;Li,R.;Cui,L.;Clegg,J.K.;Herz,L.HysteresisinCH3NH3PbX3PerovskiteSolarCells.Nat.Commun.M.;Johnston,M.B.;Lin,Q.CrystallizationofCsPbBr3Single2016,7,10334.CrystalsinWaterforX-RayDetection.Nat.Commun.2021,12,1531.(36)Lan,D.ThePhysicsofIonMigrationinPerovskiteSolarCells:(18)Wang,X.D.;Huang,Y.H.;Liao,J.F.;Wei,Z.F.;Li,W.G.;Xu,InsightsintoHysteresis,DevicePerformance,andCharacterization.Y.F.;Chen,H.Y.;Kuang,D.B.SurfacePassivatedHalidePerovskiteProg.Photovoltaics2020,28,533−537.Single-CrystalforEfficientPhotoelectrochemicalSynthesisof(37)Hutter,E.M.;Eperon,G.E.;Stranks,S.D.;Savenije,T.J.Dimethoxydihydrofuran.Nat.Commun.2021,12,1202.ChargeCarriersinPlanarandMeso-StructuredOrganic-Inorganic(19)Burschka,J.;Pellet,N.;Moon,S.J.;HumphryBaker,R.;Gao,Perovskites:Mobilities,Lifetimes,andConcentrationsofTrapStates.P.;Nazeeruddin,M.K.;Grätzel,M.SequentialDepositionasARouteJ.Phys.Chem.Lett.2015,6,3082−3090.toHigh-PerformancePerovskite-SensitizedSolarCells.Nature2013,(38)Reid,O.G.;Yang,M.;Kopidakis,N.;Zhu,K.;Rumbles,G.499,316−319.Grain-Size-LimitedMobilityinMethylammoniumLeadIodide(20)Jeon,N.J.;Noh,J.H.;Kim,Y.C.;Yang,W.S.;Ryu,S.;Seok,S.PerovskiteThinFilms.ACSEnergyLett.2016,1,561−565.I.SolventEngineeringforHigh-PerformanceInorganic-Organic(39)Semonin,O.E.;etal.LimitsofCarrierDiffusioninn-TypeandHybridPerovskiteSolarCells.Nat.Mater.2014,13,897−903.p-TypeCH3NH3PbI3PerovskiteSingleCrystals.J.Phys.Chem.Lett.(21)Xiao,M.;Huang,F.;Huang,W.;Dkhissi,Y.;Zhu,Y.;2016,7,3510−3518.Etheridge,J.;Gray-Weale,A.;Bach,U.;Cheng,Y.B.;Spiccia,L.A(40)Kim,D.H.;Park,J.;Li,Z.;Yang,M.;Park,J.-S.;Park,I.J.;FastDeposition-CrystallizationProcedureforHighlyEfficientLeadKim,J.Y.;Berry,J.J.;Rumbles,G.;Zhu,K.300%EnhancementofIodidePerovskiteThin-FilmSolarCells.Angew.Chem.,Int.Ed.2014,CarrierMobilityinUniaxial-OrientedPerovskiteFilmsFormedby53,9898−9903.Topotactic-OrientedAttachment.Adv.Mater.2017,29,1606831.(22)Nie,W.;Tsai,H.;Asadpour,R.;Blancon,J.C.;Neukirch,A.J.;(41)Valverde-Chávez,D.A.;Ponseca,C.S.;Stoumpos,C.C.;Gupta,G.;Crochet,J.J.;Chhowalla,M.;Tretiak,S.;Alam,M.A.;Yartsev,A.;Kanatzidis,M.G.;Sundström,V.;Cooke,D.G.Intrinsicetal.High-EfficiencySolution-ProcessedPerovskiteSolarCellswithFemtosecondChargeGenerationDynamicsinSingleCrystalMillimeter-ScaleGrains.Science2015,347,522−525.CH3NH3PbI3.EnergyEnviron.Sci.2015,8,3700−3707.(23)Ding,J.;Han,Q.;Ge,Q.Q.;Xue,D.J.;Ma,J.Y.;Zhao,B.Y.;(42)Johnston,M.B.;Herz,L.M.HybridPerovskitesforChen,Y.X.;Liu,J.;Mitzi,D.B.;Hu,J.S.FullyAir-BladedHigh-Photovoltaics:Charge-CarrierRecombination,Diffusion,andRadia-EfficiencyPerovskitePhotovoltaics.Joule2019,3,402−416.tiveEfficiencies.Acc.Chem.Res.2016,49,146−154.(24)Yin,J.;Qu,H.;Cao,J.;Tai,H.;Li,J.;Zheng,N.Vapor-Assisted(43)Poncé,S.;Li,W.;Reichardt,S.;Giustino,F.First-PrinciplesCrystallizationControltowardHighPerformancePerovskitePhoto-CalculationsofChargeCarrierMobilityandConductivityinBulkvoltaicswithover18%EfficiencyintheAmbientAtmosphere.J.SemiconductorsandTwo-DimensionalMaterials.Rep.Prog.Phys.Mater.Chem.A2016,4,13203−13210.2020,83,036501.(25)Fan,P.;Gu,D.;Liang,G.X.;Luo,J.T.;Chen,J.L.;Zheng,Z.(44)Jariwala,S.;Sun,H.;Adhyaksa,G.W.;Lof,A.;Muscarella,L.H.;Zhang,D.P.High-PerformancePerovskiteCH3NH3PbI3ThinA.;Ehrler,B.;Garnett,E.C.;Ginger,D.S.LocalCrystal3615https://doi.org/10.1021/acs.jpclett.1c00619J.Phys.Chem.Lett.2021,12,3607−3617

9TheJournalofPhysicalChemistryLetterspubs.acs.org/JPCLLetterMisorientationInfluencesNon-RadiativeRecombinationinHalide(66)Lacroix,A.;DeLaissardiere,G.T.;Quémerais,P.;Julien,J.-P.;Perovskites.Joule2019,3,3048−3060.Mayou,D.ModelingofElectronicMobilitiesinHalidePerovskites:(45)Hargreaves,J.S.J.SomeConsiderationsRelatedtotheUseofAdiabaticQuantumLocalizationScenario.Phys.Rev.Lett.2020,124,theScherrerEquationinPowderX-rayDiffractionasAppliedto196601.HeterogeneousCatalysts.Catalysis,Structure&Reactivity2016,2,(67)Schlipf,M.;Poncé,S.;Giustino,F.CarrierLifetimesand33−37.PolaronicMassEnhancementintheHybridHalidePerovskite(46)Baikie,T.;Fang,Y.;Kadro,J.M.;Schreyer,M.;Wei,F.;CH3NH3PbI3fromMultiphononFröhlichCoupling.Phys.Rev.Lett.Mhaisalkar,S.G.;Graetzel,M.;White,T.J.SynthesisandCrystal2018,121,086402.ChemistryoftheHybridPerovskite(CH3NH3)PbI3forSolid-State(68)Mayadas,A.F.;Shatzkes,M.Electrical-ResistivityModelforSensitisedSolarCellApplications.J.Mater.Chem.A2013,1,5628−PolycrystallineFilms:theCaseofArbitraryReflectionatExternal5641.Surfaces.Phys.Rev.B1970,1,1382−1389.(47)Wolfe,C.M.;Stillman,G.E.;Lindley,W.T.ElectronMobility(69)Steinhögl,W.;Schindler,G.;Steinlesberger,G.;Engelhardt,M.inHigh-PurityGaAs.J.Appl.Phys.1970,41,3088−3091.Size-DependentResistivityofMetallicWiresintheMesoscopic(48)Yang,J.J.J.;Dapkus,P.D.;Dupuis,R.D.;Yingling,R.D.Range.Phys.Rev.B:Condens.MatterMater.Phys.2002,66,075414.ElectricalPropertiesofPolycrystallineGaAsFilms.J.Appl.Phys.1980,(70)Sun,T.;Yao,B.;Warren,A.P.;Barmak,K.;Toney,M.F.;51,3794−3800.Peale,R.E.;Coffey,K.R.SurfaceandGrain-BoundaryScatteringin(49)Herz,L.M.Charge-CarrierDynamicsinOrganic-InorganicNanometricCuFilms.Phys.Rev.B:Condens.MatterMater.Phys.MetalHalidePerovskites.Annu.Rev.Phys.Chem.2016,67,65−89.2010,81,155454.(50)Crothers,T.W.;Milot,R.L.;Patel,J.B.;Parrott,E.S.;Schlipf,(71)Davies,C.L.;Filip,M.R.;Patel,J.B.;Crothers,T.W.;Verdi,J.;Müller-Buschbaum,P.;Johnston,M.B.;Herz,L.M.PhotonC.;Wright,A.D.;Milot,R.L.;Giustino,F.;Johnston,M.B.;Herz,L.ReabsorptionMasksIntrinsicBimolecularCharge-CarrierRecombi-M.BimolecularRecombinationinMethylammoniumLeadTriiodidenationinCH3NH3PbI3Perovskite.NanoLett.2017,17,5782−5789.PerovskiteisAnInverseAbsorptionProcess.Nat.Commun.2018,9,(51)Poncé,S.;Margine,E.R.;Giustino,F.TowardsPredictive293.Many-BodyCalculationsofPhonon-LimitedCarrierMobilitiesin(72)Karakus,M.;Jensen,S.A.;D’Angelo,F.;Turchinovich,D.;Semiconductors.Phys.Rev.B:Condens.MatterMater.Phys.2018,97,Bonn,M.;Canovas,E.Phonon-ElectronScatteringLimitsFree121201.ChargeMobilityinMethylammoniumLeadIodidePerovskites.J.(52)Sangalli,D.;etal.Many-BodyPerturbationTheoryPhys.Chem.Lett.2015,6,4991−4996.CalculationsUsingtheYamboCode.J.Phys.:Condens.Matter(73)Yi,H.T.;Wu,X.;Zhu,X.;Podzorov,V.IntrinsicCharge2019,31,325902.TransportAcrossPhaseTransitionsinHybridOrgano-Inorganic(53)Gonze,X.;Lee,C.DynamicalMatrices,BornEffectiveCharges,Perovskites.Adv.Mater.2016,28,6509−6514.DielectricPermittivityTensors,andInteratomicForceConstants(74)Filippetti,A.;Mattoni,A.;Caddeo,C.;Saba,M.;Delugas,P.fromDensity-FunctionalPerturbationTheory.Phys.Rev.B:Condens.LowElectron-PolarOpticalPhononScatteringasAFundamentalMatterMater.Phys.1997,55,10355−10368.AspectofCarrierMobilityinMethylammoniumLeadHalide(54)Baroni,S.;DeGironcoli,S.;DalCorso,A.;Giannozzi,P.CH3NH3PbI3Perovskites.Phys.Chem.Chem.Phys.2016,18,PhononsandRelatedCrystalPropertiesfromDensity-Functional15352−15362.PerturbationTheory.Rev.Mod.Phys.2001,73,515−562.(75)Zhao,T.;Shi,W.;Xi,J.;Wang,D.;Shuai,Z.Intrinsicand(55)Giannozzi,P.;etal.AdvancedCapabilitiesforMaterialsExtrinsicChargeTransportinCH3NH3PbI3PerovskitesPredictedModellingwithQuantumESPRESSO.J.Phys.:Condens.Matter2017,fromFirst-Principles.Sci.Rep.2016,6,19968.29,465901.(76)Lu,Y.B.;Kong,X.;Chen,X.;Cooke,D.G.;Guo,H.(56)Pizzi,G.;etal.Wannier90asACommunityCode:NewPiezoelectricScatteringLimitedMobilityofHybridOrganic-FeaturesandApplications.J.Phys.:Condens.Matter2020,32,165902.InorganicPerovskitesCH3NH3PbI3.Sci.Rep.2017,7,41860.(57)Giustino,F.;Cohen,M.L.;Louie,S.G.Electron-Phonon(77)Peng,J.;Cui,L.;Li,R.;Xu,Y.;Jiang,L.;Li,Y.;Li,W.;Tian,X.;InteractionUsingWannierFunctions.Phys.Rev.B:Condens.MatterLin,Q.ThickJunctionPhotodiodesBasedonCrushedPerovskiteMater.Phys.2007,76,165108.Crystal/PolymerCompositeFilms.J.Mater.Chem.C2019,7,1859−(58)Poncé,S.;Margine,E.R.;Verdi,C.;Giustino,F.EPW:1863.Electron-PhononCoupling,TransportandSuperconductingProper-(78)Seifert,T.;Jaiswal,S.;Sajadi,M.;Jakob,G.;Winnerl,S.;Wolf,tiesUsingMaximallyLocalizedWannierFunctions.Comput.Phys.M.;Kläui,M.;Kampfrath,T.UltrabroadbandSingle-CycleTerahertzPulseswithPeakFieldsof300kVcm−1fromAMetallicSpintronicCommun.2016,209,116−133.(59)Giustino,F.Electron-PhononInteractionsfromFirstPrinciples.Emitter.Appl.Phys.Lett.2017,110,252402.Rev.Mod.Phys.2017,89,015003.(79)Pisoni,A.;Jacimovic,J.;Barisic,O.S.;Spina,M.;Gaál,R.;(60)Poncé,S.;Schlipf,M.;Giustino,F.OriginofLowCarrierForró,L.;Horváth,E.Ultra-LowThermalConductivityinOrganic-MobilitiesinHalidePerovskites.ACSEnergyLett.2019,4,456−463.InorganicHybridPerovskiteCH3NH3PbI3.J.Phys.Chem.Lett.2014,(61)Mayers,M.Z.;Tan,L.Z.;Egger,D.A.;Rappe,A.M.;5,2488−2492.Reichman,D.R.HowLatticeandChargeFluctuationsControl(80)Heiderhoff,R.;Haeger,T.;Pourdavoud,N.;Hu,T.;Al-Khafaji,CarrierDynamicsinHalidePerovskites.NanoLett.2018,18,8041−M.;Mayer,A.;Chen,Y.;Scheer,H.C.;Riedl,T.Thermal8046.ConductivityofMethylammoniumLeadHalidePerovskiteSingle(62)Wu,K.;Bera,A.;Ma,C.;Du,Y.;Yang,Y.;Li,L.;Wu,T.CrystalsandThinFilms:AComparativeStudy.J.Phys.Chem.CTemperature-DependentExcitonicPhotoluminescenceofHybrid2017,121,28306−28311.OrganometalHalidePerovskiteFilms.Phys.Chem.Chem.Phys.(81)Gold-Parker,A.;Gehring,P.M.;Skelton,J.M.;Smith,I.C.;2014,16,22476−22481.Parshall,D.;Frost,J.M.;Karunadasa,H.I.;Walsh,A.;Toney,M.F.(63)Wright,A.D.;Verdi,C.;Milot,R.L.;Eperon,G.E.;Pérez-AcousticPhononLifetimesLimitThermalTransportinMethyl-Osorio,M.A.;Snaith,H.J.;Giustino,F.;Johnston,M.B.;Herz,L.M.ammoniumLeadIodide.Proc.Natl.Acad.Sci.U.S.A.2018,115,Electron-PhononCouplinginHybridLeadHalidePerovskites.Nat.11905−11910.Commun.2016,7,11755.(82)Nolas,G.S.;Goldsmid,H.J.ThermalConductivity;Springer,(64)Frost,J.M.CalculatingPolaronMobilityinHalidePerovskites.2004;pp105−121.Phys.Rev.B:Condens.MatterMater.Phys.2017,96,195202.(83)vanSetten,M.J.;Giantomassi,M.;Bousquet,E.;Verstraete,M.(65)Herz,L.M.HowLatticeDynamicsModeratetheElectronicJ.;Hamann,D.;Gonze,X.;Rignanese,G.M.ThePseudoDoji:PropertiesofMetal-HalidePerovskites.J.Phys.Chem.Lett.2018,9,TrainingandGradingA85ElementOptimizedNorm-Conserving6853−6863.PseudopotentialTable.Comput.Phys.Commun.2018,226,39−54.3616https://doi.org/10.1021/acs.jpclett.1c00619J.Phys.Chem.Lett.2021,12,3607−3617

10TheJournalofPhysicalChemistryLetterspubs.acs.org/JPCLLetter(84)Godby,R.W.;Needs,R.J.Metal-InsulatorTransitioninKohn-ShamTheoryandQuasiparticleTheory.Phys.Rev.Lett.1989,62,1169−1172.(85)Bruneval,F.;Gonze,X.AccurateGWSelf-EnergiesinAPlane-WaveBasisUsingOnlyAFewEmptyStates:TowardsLargeSystems.Phys.Rev.B:Condens.MatterMater.Phys.2008,78,085125.3617https://doi.org/10.1021/acs.jpclett.1c00619J.Phys.Chem.Lett.2021,12,3607−3617