ChristopheJ.Barbe´,*FrancineArendse,PascalComte,MarieJirousek,FrankLenzmann,
ValeryShklover,andMichaelGra¨tzel
InstituteofPhysicalChemistry,SwissFederalInstituteofTechnology,Lausanne,Switzerland
Duringratorytrochemicalathenewpastfiveyears,wehavedevelopedinourlabo-aprocess.typeofsolarThelightcellthatisbasedonaphotoelec-adsorbedmonolayerofdye(i.e.,aRutheniumabsorptioncomplex)isperformedthatby
istitaniumhasoxidechemically(TiOatthesurfaceofasemiconductor(i.e.,2)).WhenThetheabilitytotransferanelectronexcitedtobyaphoton,thedyetionelectricfromoffieldthatisinsidethematerialthesemiconductor.allowsextrac-Athethedyeelectron,toaredoxandmediatorthepositivethatchargeispresentistransferredinsolution.byrespectablephotovoltaicefficiency(i.e.,10%)isobtainedparticles.theusetureWeofmesoporous,willshowhownanostructuredthefilmsofanatasespecifically,influencesthephotovoltaicTiOresponse2electrodeofthemicrostruc-cell.Moresuchmalasprecursorwewillchemistry,focusonhowtemperatureprocessingforparametershydrother-fluencegrowth,tering,thefilmbinderporosity,addition,pore-sizeandsinteringdistribution,conditionslightin-solar-cellandefficiency.
electronpercolationandconsequentlyaffectscat-theI.
Introduction
R
ECENTLY,
nanocrystallinebecausecreasingTheseofamountofattentionmaterialsinthehavescientificattractedcommunityanin-(i.e.,unusualtheirpropertiesspectacularresultphysicalfromandthechemicalultrafineproperties.structureclassifiedagrainbulkinsizeof<50nm)ofthesematerialsandcanbecisely,andmenttheirpropertiestwocategories:smallcrystallitethatarerelativepropertiesthatarerelativetothesizecantoleadthetosurface.quantumMorepre-CdS),effectsbeenmagneticinsemiconductors1materialswith(siliconorcadmiumconfine-sulfide,refrigeration,usedformonodomaingrainsthathave2high-densityties.1,3orinformationstorageandmagneticgrain-boundary-to-volumeAnothercharacteristicmaterialswithofthesestrongmaterialspiezoelectricistheirproper-ofsynthesisductile4orsuperplasticratio,ceramicswhich5andenablesthefabricationhighmembranesof6ultradenseandmaterialsorcoatingstheandlow-temperaturehighlyporoustheinnanocrystallineusingthisveryelectrodes.7Byhighsurfacearea,whichisprovidedbyphotoelectrochemicalourlaboratoryaparticles,newprocess.typeweofhave8,9solarsuccessfullycellthatisbaseddevelopedonasolardifferentiated.cells,thelight-absorptionIncontrasttoconventionallayerThelightabsorptionandcharge-separationisperformedbystepsaarechemicallyofdyeoxide(TiOat(i.e.,thesurfaceaRutheniumofasemiconductorcomplex)thatisadsorbedmono-(excitation),(i.e.,titanium2)).Viaabsorptionofaphotonthedye
C.J.Brinker—contributingeditor
ManuscriptNo.191723.ReceivedJunegewandteSupportedbytheSwissNationalScience24,1996;FoundationapprovedandMarchtheInstitut31,1997.fu¨rAn-*Member,PhotovoltaikAmerican(Gelsenkirschen,CeramicSociety.
Germany).changesstatetiondizedband(S*).fromofThen,thethesemiconductoritelectronicinjectsangroundexcitedstate(S0)totheexcited(TiOelectronintotheconduc-2)and,thus,bandreductionand(S+).theTherecombinationoftheelectroninthebecomesconductionoxi-thatfromholetheonS+thestateoxidizeddyeismuchslowerthanthetheisinsolution.Thus,chargebytheseparationreducedstateisofefficient.themediatormaximumoxidizedredoxvoltagemediatorcorrespondsisreducedtoatthetheNext,differencecounterelectrode.Thesemiconductor.potentialrespectableTheofthesimplicitymediatoroftheandcellthedesign,Fermibetweenaslevelwelloftheastheitscostreductionefficiencyof(i.e.,10%–11%),promisesasignificantuseThiscastingofporousrespectablesolarnanostructuredphotovoltaicenergy.
filmsefficiencythatisobtainedviathewillthus,describenanocrystallinehowtheanataseTiOparticles.areInsynthesizedthefollowing,bytapewephotovoltaiccolloidalvestigateresponsesynthesisofand2thecell.filmelectrodemicrostructure—and,Moreprocessing—influencesthearea,conductor.thedyetherelationshipsadsorption,betweentheparticleprecisely,size,wethewillsurfacein-tributionWealsowillshowandthehowelectrontheinitialinjectionparticle-sizeinthesemi-which,therefore,inturn,influencesaffectsthetheelectrolytepore-sizedistributiontransportphenomenainthefilm,dis-efficiency.
theinternalresistanceofthecell,aswellasitsglobaland,II.
ExperimentalProcedure
(1)ColloidalSynthesis
asTypical(97%,follows.synthesisAquantityoftheTiO2nanoparticlescanbedescribedwisesolutionandAldrichChemical(125Co.,Milwaukee,mL)oftitaniumWI)isisopropoxideadded,drop-stantaneously.underatroomvigoroustemperature,stirring.toA750mLofa0.1Mnitricacidheatedtizationto80°CImmediatelyandstirredvigorouslyafterthewhitehydrolysis,precipitatefor8h,toachievetheformsslurryin-pep-isinto(i.e.,destructionoftheagglomeratesandredispersionfritfiltratetoprimaryremoveparticles).nonpeptizedTheagglomerates.solutionisthenWaterfilteredonaglassunderThegrowthtoadjustthefinalsolidsconcentrationtois∼added5wt%.totheheatedhydrothermaloftheseconditionsparticles,inupatotitanium10–25autoclavenm,isachievedthattemperaturefor12tationishdependentinthetemperatureisontherangeof200°–250°C;thedispersedoccurspulses).troducedAfterusingduringtheautoclaving,desiredandparticletheparticlessize.Sedimen-arere-twoatitaniumsonications,ultrasonicthecolloidalhorn(400suspensionW,15×2s(3MPa))intoaarotaryevaporatorandevaporated(35°C,30ismbarin-theSeveralvariationsfinalTiOof2thisconcentrationsynthesiswereof11designed,wt%.
tostudypH,influenceofprocessingparameters,suchasprecipitationonhydrolysisrate,autoclavingpH,andprecursorchemistry,ofthealkoxidepHmorphologyandthehydrolysisofthefinalrateparticles.duringTostudytheinfluenceshotbasicto(0.1dropwise)wasaddedMammonia)tousinganwateracidicdifferentspeedsprecipitation,(varyingfromtitaniumonesolution.(0.1MnitricTostudyacid),theneutral,influenceandof
pHautoclavedduringhydrothermalgrowth,precipitateswerepeptizedandments:13precursorinaatatatemperatureof250°Cindifferentenviron-triethylaminepH1innitricsolution.acid,atpHThe11inammonia,andatpHideCorp.,ortitaniumchemistrybutoxidewasstudied(bothpurchasedbysubstitutinginfluenceofthetitaniumfromtitaniumethox-reactingRonkonkoma,inthetitaniumalkoxideNY)fortitaniumwithaceticisopropoxideFlukaacidoracetylorChemicalbypre-prioramoisture-freetohydrolysis.environment10(i.e.,anargon-gasgloveacetonebox)(2)TiO2ElectrodePreparation
techniqueTheTiO2pasteisdepositedusingasimpledoctor-bladethatonsheetglass(NipponSheetGlass,Hyogo,Japan)(SnOhasbeencoatedwithafluorine-dopedstannicoxidegreen2)layer(sheetresistanceof8–10⍀/ᮀtemperaturelayeris∼100mthick.Thelayerisdried).Theinairresultingatroomofof∼50°Cforfor1510min.min,Then,followedthefilmbyistreatmentatatemperatureto20°–50°C/minroomtemperature.
andleftat450°Cforheated30minto450°CbeforeatcoolingaratetheTopolyethylenesynthesispreventofcrackingsinteredduringfilmsfilmthatdrying,are10whichmthusrendersMerck,50%Darmstadt,glycolGermany)(PEG,molecularisaddedweightinaproportion(MW)thickofpossible,of20000,0%–screw-threadedoftheTiO2weight.TheresultingpasteisstoredinaThisAnglassbottleuntildeposition.
ofstepadditionalconsistsstepofimpregnatingcanbeperformedtheTiOonthefiredelectrode.2filmwithawatertitanium450°C(concentrationtetrachloridefor30min.
of0.1(TiClsolutionM),4which)thathasisfollowedbeendissolvedbyafiringiniceat(3)ColloidsandElectrodesCharacterization
relationTheparticle-sizedistributionwashavennmInstruments,spectroscopyAustin,(ModelTX)BI2030ATstudiedthatwasequippedinstrumentusingphotoncor-with(Brook-aHoriba,laser)determinedTokyo,andcentrifugalJapan).†Thesedimentationsurfacearea(ModelofthecolloidsCapa700,488wasGeminiZeta(Modelpotential2327,usingMicromereticsanitrogenadsorptionapparatus(ModelstudiesondifferentInstrumentCorp.,Norcross,GA).ditionalZetasizer,MalvernInstruments,suspensionsMalvern,wereU.K.).performedAd-croscopycharacterizationmicroscope(TEM),whichwasincludedtransmissionelectronmi-Eindhoven,(ModelCM30ST,performedPhilipsResearchusingahigh-resolution(XRD),(Scintag,whichThewasNetherlands),andX-raydiffractometryLaboratories,resolutionTheelectrodeSantaClara,performedmicrostructureCA)usingusingCuapowderdiffractometerwasK␣studiedradiation.
usingmicroscopescanningelectronmicroscopy(SEM)field-emissionahigh-trodeadsorption–desorptionpore-size(ModeldistributionS-900,Hitachi,Tokyo,Japan).Theelec-mereticsapparatuswas(ModelstudiedASAPusing2010,anitrogenflectionsInstrumentCorp.).TheopticaltransmissionsandMicro-re-rangespectrometerof300–800oftheTiOnm2onfilmsanultraviolet–visiblewererecordedinlightthewavelength(UV–VIS)anintegrating(Varian,sphere.PaloAlto,CA)thatwasequippedwith(4)Solar-CellAssembly
inTheTiO2nanocrystallineelectrodedicarboxilate)asolutionofrutheniumdye(rutheniumwasimmersed(2,2Јbipyridyl-4,4overnightЈ-trationacetonitrileof3×2(NCS)10−4M2).thatThewaselectrodedissolvedwasinethanol)thenrinsed(concen-withsolutionpenetratedwasanddepositeddried.Oneontodroptheofsurfaceaniodine-basedoftheelectrodeelectrolyteandtrolytemethylsolutioninsidewasthecomposedTiO2filmofvia0.5capillarymmol/Laction.ofhomemadeTheelec-mmol/Lhexylpyridineofthatlithiumimidazolium,11wasdissolvediodide(LiI),20inacetonitrileandmmol/L500mmol/Lofiodine(Idi-2),40(alltheofchemicals
tert-butyl†Notcorrectedforlightscattering.
wereelectrodepurchasedelectrodewastoformthenfromourclippedFlukatestcell.
ontoChemical).thetopAofplatinizedtheTiOcounter2working(5)PhotovoltaicCharacterizations
terizedThephotovoltaicpropertiesofthesolarsistedsolarofusingrecordingtwodifferentthecurrent–voltagetechniques.Thecellswerecharac-characteristicsfirsttechniquecon-1Sun),cellwhichinthecorrespondeddarkandundertoaansolarilluminationelevationofofAM42°1.5oftheto(orthe
TableI.SurfaceAreaandAverageParticleSizeofSamplesAutoclavedatDifferentTemperatures
Powder
Surfacearea(m2/g)
dBET†(nm)
AutoclavedAutoclavedat200°C145Autoclavedat210°C13010.8Autoclavedatat230°C250°C11512.013.6P25(Degussa)1085514.528.4
†ticlesAveragearespherical.
diameterback-calculatedfromthesurfacearea,assumingthatthepar-horizon.Anadditionalmeasurementwasperformedatalowerlightintensity(1/10Sun)toexaminethepotentialnonlinearityinthecellresponse.Atypicalcurrent–voltagecurveisshowninFig.1.Thiscurvecanbedescribedusingthefollowingsetofparameters:theshort-circuitcurrent(ISC),inunitsofmA/cm2),theopencircuitvoltage(V0C,inunitsofmV/cm2),theefficiencypercentage(),andthefillfactorpercentage(FF).ISCisthecurrentthatisrecordedwhenthevoltageiszero,andV0Cisthepotentialwhenthecurrentiszero.Soasnottooverwhelmthereaderwithcurrent–voltagecurves,theperfor-mancesofthecellswillbediscussedusingthepreviouslydescribedsetofparameters.Theefficiencyisgivenbytheratiooftheelectricalpowerthatisdeliveredbythecelldividedbythepowerofthelightthatisilluminatingthecell.
Thesecondtypeofphotovoltaictestthathasbeenperformedonthesolarcelliscalledincidentphotonconversionefficiency(IPCE).Itrepresentsthepercentageofincidentphotonsthatare
convertedtoelectronsatacertainwavelengthandisdefinedbytheformula
ISCIPCE͑͒=1240(1)
⌽whereisthewavelength(innanometers)and⌽istheinci-dentradiativeflux(inunitsofW/m2).Theexperimentaldetailsthatconcernthetwoexperimentalsetupsthatareusedforthesecharacterizationsaregivenelsewhere.12ͩͪIII.ResultsandDiscussion
(1)InfluenceofProcessingParametersontheFinalFilmMicrostructure
(A)InfluenceofHydrolysispHandRate:Nonoticeabledifference,withrespecttotheBrunauer–Emmett–Teller(BET)surfacearea,isobservedintheprecipitates,regardlessof
whethershot.taniumThistheeffectratiophenomenontitaniumprecursorisaddeddropwiseorinonethatispresentlyisexplainedused,whichbythevery-highwater:ti-appearsAlthoughofhydrolysisrendersthekinetictheprecipitatenegligible.
inanacidiclyzedtobebetterdispersedthanthatwhichenvironmenthasbeenvisuallyhydro-similarinthermalandneutralsizegrowth.wellThisdispersedorbasicsolution,thethreesamplesappearobservationafterpeptizationisconfirmedinacidbyandhydro-withanalysis,hydrothermalBETamedianwhichhasrevealedthepresenceoftheaggregatesparticle-analysisvalueshowsdiameterof0.15minthethreesolutions.larthesurfacearea,treatment∼115±atthat2amtemperaturethethreedifferent2/g,of230°C,colloids,haveasimi-afterthatcorrespondingfilmsareidentical.andThus,theSEMitcanmicrographsofhasaneithertheprecipitationpHnorthealkoxidebeadditionconcludedratedevised(B)noticeableRoleoftheinfluencePeptizationontheStep:finalelectrodemicrostructure.tion.todestroyagglomeratesthatformedThepeptizationduringprecipita-stepwasprovidedTheheating,theandbyenergythethethermalthatisneededforthedeagglomerationwasstabilizationenergywasthatensuredwassuppliedelectrostaticallyduringthebyshouldpresenceof0.1Macid.Infact,thepeptizationofthecolloidelectricbepointeasierthemorethepHvaluedeviatesfromtheiso-peptization.Ishows(pHofm2Thesurfacethat6.5–7).13Tablesubstantialareadecreasesgrowthfromalso∼297occursm2/gduringto165particle/gattheendofthepeptization,whichcorrespondstoaThissificationdecreasegrowthfrom5.4nmto9.4nm(inaveragediameter).nanoparticlesphenomenainsurfacetheduringpeptization.thatareaarelinkedalsocouldbeattributedtoden-AlthoughtothecrystallizationtheXRDpatternoftheoftospectrumanataseas-precipitatedpeakof(seethisFig.powderpowder7),theexhibitssomepeaksthatcorrespondshowsdifferentialthepresencethermalofanalysis(DTA)peakat∼350°Cthatcanbeattributedtocrystallization.anexothermic8whichistion.suggestsabsentintheDTAspectrumofthepeptizedsample,ThissignificantlyBecausemightlowerthethatdensitycrystallizationthanofthatamorphousisoccurringofanatase,TiOduringpeptiza-the2isparticleexpectedtobeticle(C)notGrowth:InfluencebeasimportantgrowthTheofAutoclavingasstatedpreviously.
autoclavingTemperaturestepisdesignedonPrimarytogrowPar-the
primarynismssultsdecrease(summarizedthatparticlesoccurfurtherunderbydissolution–reprecipitationmecha-inTablehydrothermalI)clearlyconditions.TheBETre-particlesgrowthorincrystallitesurfaceareagrowth.thatcanindicatethatthereisaSEMbeinvestigationstranslatedtotheprimaryallnantlytheprimary(seeFig.particles2).TEMareinvestigationcrystalline;(seethefacesFig.3)confirmshowsthisthatParticle-size(D)orientedInfluencealongarepredomi-toclavedanalysisofAutoclavingthe<101>direction.
wasperformedTemperatureonsamplesonAggregation:thatsentedgrowth,inatFig.different4,revealtemperatures.that,inparallelTheresults,tothewhichwerearepre-au-areanimportantaggregationoccurs.Althoughprimarytheparticlefinesoftemperatureaggregatesunderestimated,(averagebecausesizeofof∼light0.3scattering,m)withchangestheformationinthepore-sizeTheautoclavingisclearlytemperaturevisible.
alsohasaninfluenceonthewilldeposition,distributiontheconcentrationofthefinalTiOtheof2solidsfilm(seeinFig.thesuspension5).DuringcontactincreaseondryinguntiltheaggregatescomeintophysicalThen,shape.thewitheachotherandformathree-dimensionalnetwork.thatanareTheporesizesaredeterminedbytheaggregatesizeanddictatedlatterbyisdependentontheaggregationmechanismsfilmidenticalpension.willbecoatinggreatlyprocedure,thechemicalinfluencedtheenvironment.Therefore,forbypore-sizethedistributioninthepresentaggregatecaseInfact,betweenwehavetheobservedaverageporeadirectaggregatesizeinsus-sizecorrelationinthewillasbeinfluencedsize(seebyFig.many6).Ofothercourse,andtheaverageprocessingthepore-sizeparameters,distributionSometheamountsections.
oftheseoftopicsbinderwillorbethediscussedsinteringtemperatureinthefollowingandtime.suchsub-mation:(E)InfluenceofAutoclavingTemperatureonRutileFor-autoclavedXRDtures.in0.1analysiswasperformedonsamplesthatwerewerethatautoclavedTheresultsatareMnitricpresentedacidinsolutionFig.7.Onlyatdifferentthesamplestempera-thatScherrercorrespondeda∼relation,wetotemperatureofՆ240°Cexhibitedpeakscalculatedtherutilethephase.UsingtheDebye–larger200nmaddition,thaninthethe250°Crutilecrystallitesizetobeaveragesample,whichisanorderofmagnitudewhetherparticlesthenoprogressiveanataseobservationgrowthcrystallitesize(15nm).InwasviawasXRDobserved,orregardlessofwerecrystallites.alreadyappearedanordersuddenlyofmagnitudeatatemperatureSEM.largerthanofthe240°CTherutileanataseandtionthatgatesinvolvedofanataseAlltheseobservationssuggestthatthetransforma-sometorutileproceededbyacooperativetransitioncooperativebeforeshowntransitiontherutilelocaltransitionrearrangementsintheanataseaggre-issupportedoccurred.bytheThishypothesisofainsidelater,theinFig.larger8,whererutileclusterscrystals.ofmicrographsthatareMoreover,anataseparticlesasareburiedventsmodificationphase,thisaggregationoftheandtitaniumpreventsalkoxidetheformationbyaceticwewillshowofacidthepre-preventingevendrothermalaggregationatatemperatureoftheofanatase250°C;particlesthissuggestsrutileduringthat,thehy-byservationstep,theformationofrutileisprevented.Thisob-study(F)Influencealsoconfirmsofthecooperativenatureofthetransition.morphologytheinfluenceacidoftheparticles,ofpHpHduringduringtheHydrothermalStep:Tosampleshydrothermalgrowthonthein0.1(i)have0.1Mbeennitricautoclavedacidatatemperaturethatwereof250°Chydrolyzedfor12inhBETMsolution,(ii)ammoniaatpH11,and(iii)underresultstriethylamineatpH13.SEMmicrographs(Fig.9)andgestsronment.thatbasic(TableII)bothshowthattheparticleswerelargerOstwaldconditionsripeningthanunderacidicconditions,whichsug-thatconditionsareformedThisobservationinabasicsuggestsismoreimportantinabasicenvi-environmentthatthetransientspeciespH(acidicareconditions).morestableThisthanstabilitythosewhichunderoftheareintermediateformedhydrothermalatspe-
lowciesand,therefore,theincreaseinthesolubilityoftheTiO2,alsocouldbeduetoacomplexationbythecounterion(NH4+ortriethylamine)thatstabilizestheanionicspeciesthatisformedbydissolutioninamannerthatisanalogoustothatwhichhasbeenobservedwithsulfate(SO42−)orfluoride(F−)ions.14Inaddition,theredissolutionoftheTiO2increasesastheconcentrationofbaseincreases,inamannerthatissimilartowhathasbeenobservedinhydrothermalgrowthinacidicmedia.15(G)InfluenceofBinderAddition:Tostudytheinfluenceofbinderadditiononfilmporosity,differentamountsofcar-bowax(PEG,MWס20000)wereaddedtoasuspensionofP25(DegussaAS,Frankfurt,Germany)in0.1Mnitricacidsolution.P25waschosenbecauseofthepossibilitytoproduce
crack-freefilmsthatare10mthickwithoutanybinderad-dition.Pore-size-distributionanalysisandtotal-porositymea-surementswereperformedonthedifferentsamples,andtheresultsarepresentedinFigs.10and11.Forthesamplesthatcontained0,10,and30wt%ofbinder,theaverageporesizeremainedcenteredat30nm,althoughthedistributionbecamebroadertowardthelargerpores.Increasingtheamountofcar-bowaxincreasedthetotalporosity,asshowninFig.11,whichsuggeststhattheaverageporesizewasstilldependentontheaverageclustersize,whichwasconstantinthethreesamples.Forthesamplethatcontained50wt%ofbinder,theaverageporesizeincreasedto50nmandthedistributionwasmuchbroader,withporesaslargeas120nm,whichsuggeststhatweareinaregionwherethemorphologyoftheporousnetworkisnotgovernedbythepercolationoftheTiO2clustersanymore.Infact,aproportionofcarbowaxof50%oftheweightofTiO2representsabinder:TiO2volumeratioof19.
(H)InfluenceofSintering:Filmswerefiredat400°,450°,500°,and550°Catarateof5°C/mininairtostudytheinfluenceoftheheat-treatmenttemperatureonthefinalfilmmorphology.Deformationofourborosilicateglasssubstratesat600°Climitedourheattreatmenttoamaximumtemperatureof550°C.Adecreaseinsurfacearea(seeTableIII)revealedthatsinteringwasoccurring,evenattheselowtemperatures;thiswasconfirmedbydilatometry,whichshowedthatpelletsthatweremadebydrypressingtheTiO2nanoparticlesstartedtoshrinkinthetemperaturerangeof∼350°–400°C.Thepore-sizemeasurementthatwasperformedonthefilmsshowedthatporecoarseningwasoccurringduringsintering(seeFig.12(a)).Theaverageporesizeincreasedfrom15nm,atatemperatureof400°C,to20nm;morespecifically,thenumberofsmallpores(i.e.,<10nminsize)decreasedsubstantiallywhile,si-multaneously,thepore-sizedistributionexpandedtowardthelargerpores.Stabilizationofthisporecoarseningwasobservedafteratemperatureof500°Cwasattained(i.e.,thepore-sizedistributiondidnotchange).
Whenthefilmswereintroduceddirectlyatatemperatureof500°Candfiredfortimesthatrangedfrom30minto5h,tostudytheinfluenceofthesinteringtimeonthefinalfilmmi-crostructure,asimilarphenomenawasobserved(seeFig.
Fig.8.SEMmicrographsofafilmsynthesizedfromacolloidautoclavedat260°C.
12(b)).Asthesinteringproceeded,porecoarseningoccurredsimultaneously.
(2)InfluenceofTiO2ElectrodeMicrostructureonthePhotovoltaicProperties
(A)InfluenceofPore-SizeDistribution:Tostudythein-fluenceofthepore-sizedistributiononthephotovoltaicre-sponseofthesolarcells,twofilmsthathaveasimilarsurfacearea(∼55m2/g)butdifferentpore-sizedistributions(seeFig.13)havebeenusedasphotoelectrodesforoursolarcellstocharacterizetheirphotovoltaicproperties.Theircurrent–volt-agecurveshavebeenrecordedunder1Sunand1/10Sun.TheresultsaresummarizedinTableIV.Theelectrodewiththesmallerporesexhibitsanimportantnonlinearity(i.e.,tobelinear,thecurrentat1Sunshouldbelargerthanthecurrentat1/10Sunbyafactorofexactly10).Underreducedillumination
TableII.SurfaceAreaandAverageParticleSizeforColloidsPeptizedinAcidbutAutoclavedat250°Cat
DifferentpHValues
Powder
Surfacearea(m2/g)
dBET†(nm)
Ina0.1Msolutionofnitricacid(pH1)Inasolutionofammonia(pH11)Ina0.1Msolutionoftriethylamine(pH13)
1087642
14.52037
†Averagediameterback-calculatedfromthesurfacearea,assumingthatthepar-ticlesarespherical.
(1/10Sun),thetransportkineticsarefastenoughtoregenerate(i.e.,reduce)thedye;however,underfullillumination,tentimesmoredyemoleculesareactive,andthetransportofI3−/I−ionstoandfromthecounterelectrodeisnotfastenoughtofullyregeneratethedyemolecules.Thisobservationcouldbeexplainedbythefactthat,inporeswithadiameterof4nm,3nmareoccupiedbythedyemolecules(moleculardiameterof15Å(1.5nm))thatareadsorbedontheporewalls,whichleavesanapertureofonly1nmforthediffusionoftheelec-trolyte.ThisdistanceisverysimilartothesizeoftheI3−ion,especiallyifoneconsidersitssolvationshell,andFick’slawofdiffusionisnotvalidanymore.Thediffusionkineticsinthe
TableIII.InfluenceofSinteringConditionsonSurfaceAreaofFilms
Filmconditions
Surfacearea(m2/g)forvarioussinteringtemperatures400°C450°C500°C550°C
Surfacearea(m2/g)forvarioussinteringtimes
30min1h2h5h
Sinteredby5°C/minand30min
Fireddirectlyat500°Cinair
12610410188
12111110386
TableIV.ComparisonofPhotovoltaicPropertiesofTwoTiO2ElectrodeswiththeSameSurfaceAreabutDifferent
Pore-SizeDistributions
ISC(mA/cm2)
V0C(mV)
Efficiency(%)
Electrodewithaverageporesizeof4nm
eculesperunitvolumeoffilm,which,intermsofelectrodemicrostructure,istheproductofthesurfaceareaandtheden-sityandcanbeexpressedbythesurfaceareaofTiO2perunitvolumeoffilm.
(3)RoleofTiCl4Post-Treatment
TheTiCl4treatmentincreasestheinjectionofelectronsintotheTiO2(seeFig.14)and,thus,thecurrentthatisdeliveredbythesolarcell.Twohypothesescanexplainthisimprovementoftheinjectionaftertreatment:(i)smallparticlesarenucleatedonthesurfaceofourelectrodeand,thus,thesurfaceareaandtheamountofdyethatisadsorbedincreases,or(ii)theelectronpercolationintheTiO2mesoporousfilmisimproved.
TableVshowsthattheTiCl4treatmentdecreasesthesurfaceareaofthefilms.Therefore,thefirsthypothesiscanbere-jected.Inaddition,theTiCl4treatmentdecreasestheaverageporesize(seeFig.15)andtheporosity(seeTableV).AllthisdatasuggeststhatthetitaniumcomplexesthatarepresentintheTiCl4solutioncondenseattheinterparticleneck.Thishypoth-esisisconsistentwiththepore-sizereduction,thesurface-arealoss,andthedensificationthatareobservedinthefilmsaftertheTiCl4treatment.Inaddition,theincreaseoftheneckingbetweentheparticleswillfacilitatethepercolationoftheelec-tronsfromoneparticletotheother,whichwilllowerthere-combinationprobabilityandleadtoaglobalincreaseofthecurrent.More-sophisticatedanalysissuchasfemtosecondfluo-rescencespectroscopywouldbeneededtoprovethatthere-combinationrateisactuallyaffectedbythistreatment.(4)InfluenceofPrecursorChemistryonFilmMorphology
Itiswellknowninsol–gelchemistrythattheprecursorchemistryinfluencesthehydrolysisandthecondensationki-neticsand,thus,thefinalmaterialmorphology.Inthepresentcase,wehaveattemptedtwodifferentapproachestocontrolournanoparticlesizeandmorphology.Thefirstapproachcon-sistedofsubstitutingtheisopropoxygroupswithmore-hydro-lyzable(ethoxy)groupsandless-hydrolyzable(n-butoxy)groups.Inbothcases,thealkoxidesareintheformofoligo-mers,whichiscontrarytotitaniumisopropoxide,whichisamonomer.16Thepresenceofoligomersisknowntoslowthecondensationreactions.Thesecondstrategyconsistedofsub-stitutingoneisopropoxygroupwithaless-hydrolyzablegroup(i.e.,acetateoracetylacetone).Experimentsthatwerecon-
1/10Sun1Sun1/10Sun1Sun
1.49.2550620560630
4.723.384.95
Electrodewithaverageporesizeof20nm
1.312.1
electrolytebecomethelimitingstepinthecurrentproduction,which,ofcourse,hasadramaticinfluenceontheefficiency,whichdecreasesfrom4.7%to3.4%.Inthesamplethathasthelargerporesize,nononlinearityisobservedandtheefficiencyisthesameat1/10Sunor1Sun.
Therefore,thepresenceofsmallporesslowsthediffusionintheelectrolyteand,thus,affectsthephotovoltaicresponseofthesolarcell,especiallyunderhighillumination.Ofcourse,otherparameters,suchastheviscosityoftheelectrolytesol-ventandtheconcentrationofiodine,willaffectthetransportkinetics.
(B)InfluenceofSurfaceAreaandPorosity:ThecurrentthatisproducedbythesolarcellsisdirectlylinkedtothenumberofdyemoleculesthatareadsorbedontheTiO2elec-trode.Therefore,thehigherthesurfacearea,thehigherthecurrentthatisgeneratedbythesolarcell;thisisoneofthekeyreasonsforusingTiO2nanoparticles.
Theporosityoftheelectrodealsodrasticallyinfluencesthephotovoltaicproperties.InFig.14,wehaveobservedthatTiO2electrodesthathaveaconstantfilmthickness(∼10m)andaconstantsurfacearea(∼55m2/g)butanincreasingporositygeneratelesscurrent.Thisphenomenonisexplainedbythefactthat,whentheporosityincreases,themassofTiO2persquarecentimeteroffilmdecreases;thus,thetotalTiO2surfacepersquarecentimeteroffilmdecreases.Iflesssurfaceisavailable,thentherearelessdyemoleculesthatareadsorbedpersquarecentimeteroffilm;thus,thecurrentdecreasesastheporosityincreases.Ifthecurrentdecreases,thentheefficiencyofthesolarcellautomaticallydecreases.Infact,thedeterminingfig-ureofmeritforthephotocurrentisthenumberofdyemol-
ductedmodifiedbysols.
precursorsLivageetal.17ledtoshowedmorethatmonodispersethehydrolysisandofsmallersuchsentedTheBETtheinTableresultsVI.ofSubstitutingthecorrespondingethoxyorpeptizedsolsarepre-ately.isopropoxycondensationTheeffectgroupofthedoessubstitutionincreasethesurfacen-butoxyareagroupsmoder-forwater:alkoxidereactionthatratio(i.e.,are50maskedmolofbecauseonthehydrolysisandtheHofthevery-highoreffectacetylhasbeenmodification,isacetoneusedinoursynthesis.Modification2Opertitaniumbyaceticatom)acidmoreimportanthasadrasticintheeffectcaseonofthethesurfaceacetylarea.Thenotfortunately,ashydrolyzablebecauseasthethetitaniumtitaniumacetateacetylacetone(CH(acac-Ti)acetonateis3COO-Ti).thermaltheacetylacetonedecomposesduringthehydro-Un-acid-modifiedTherefore,growth,wewhichwillconcentrateleadstoaheavilyontheaggregatedstudyofthesol.acetic-isTiO2.TableVIshowsthathydrothermalgrowthobservedmoreimportantisopropoxide;inthecolloidinthethatmodifiedhasbeencolloidsynthesizedthanfromthatwhichisand,conditions.therefore,thiscanisduedissolvetothemorefactthattitaniumeasilytheparticlesaresmallerOstwaldfirmedripeningThesmallereffect.theThisparticles,highergrowththemoreunderratepronouncedhydrothermalhasbeenthethismodifiedcolloidbySEMthat(seehasFig.been15).synthesizedAnotherremarkablecon-fromthefeatureofparisonisopropoxide,withprecursorthatwhichisthathasitisbeenalmostsynthesizednonaggregated,acetic-acid-fromtitaniumincom-locatedandatthesurfacewhichsuggestsoftheparticlesthattheactacetateasagroupsthatareThepreventlastcharacteristicaggregationduringofthisthecolloidhydrothermaldispersing(butnotthestep.
agentleastim-
portant)electrodesisusingarethatmuchthephotovoltaicbetterthanthatperformanceswhichhasofbeentheresultingterizationtheincreased(e.g.,classicalTEM)routeisunderway(seeFig.16).todetermineMore-detailedprocessedwhethercharac-towhichadifferentperformanceswillenablecrystallographicaredueabetterinjection.
orientationtothefilmofmorphologythesetheparticleonlyfaces,or(5)PhotovoltaicInfluenceResponseofScatteringofSolaronCells
thebeenTheandautoclavedelectrodesatthattemperaturesarepreparedՅ230°CusingarecolloidsfullythathaveautoclavedthosethatTheabovehavethisbeentemperaturemadefromarecolloidstransparent,translucentthathavebeenbeenelectrodesbecauseautoclavedthatatare250°Csynthesizedgiveafromtheparticlesorthatopaque.haveincreasesofthus,thetheirpathabilitylengthtoscatterbetterphotovoltaicresponse,oflight.Thescatteringoflightandimportantinjectincreasesanitsprobabilitythetointeractphotonswithinsideadyethemoleculecelland,dyeforelectrontheredlightinto(wavelengththesemiconductor;of∼700thisnm),isespeciallyteredisthewithnotverythisefficient.processtemperature:Certainlimitationswhereour(i)havebeenencoun-is(updifficultcolloidtothatreproducehasbeenautoclavedatathetemperaturesizedistributionof250°Cofnmto30wt%),whichandisphotoactive(ii)thecolloidatwavelengthscontainssomeofrutile>400commercialToandcontrolshouldthebenumberavoided,offorscatteringlong-termcentersstabilityinthereasons.teringcolloidparticlesanatasewaspowdermixed(Fluka)withourthatsynthesizedcontainedtransparentlargelayer,scat-aanatasethathadawasbroadautoclavedparticle-sizeatadistributiontemperaturethatofwas230°C.centered
ThisTableVI.InfluenceofPrecursorChemistryonSurfaceAreaandParticleSizeoftheCorrespondingPeptizedSols
Precursor
Surfacearea(m2/g)
dBET†(nm)
TitaniumisopropoxideTitaniumethoxideTitaniumbutoxideTitaniumisopropoxide
modifiedwithaceticacidTitaniumisopropoxide
modifiedwithacetylacetone
165186175217293
9.58.48.97.25.3
†Averagediameterback-calculatedfromthesurfacearea,assumingthatthepar-ticlesarespherical.
at∼300nmandasurfaceareaof8m2/g.TheFlukaparticles,although99%anatase,werecoatedwithaverythinlayerofsilica(SiO2)orsiliconederivatives;thiswasdiscoveredbystudyingthezetapotentialevolution,relativetopH.Theiso-electricpointoftheFlukaparticleswas2.9(insteadof6.5–7foranatase),whichcorrespondedtothatofSiO2andwaslatterconfirmedbyelectronspectroscopyforchemicalanalysis(ESCA).ThepresenceofSiO2atthesurfaceoftheparticlesposedtwoproblems.Thefirstproblemwasthattheparticlesdidnotsinterattemperatures<1000°C;therefore,thelargeparticleswereembeddedandheldbythenanoparticlesthatsinteredaroundthem(seeFig.17).NofilmswithadecentmechanicalresistancecouldbeobtainedwithՆ30wt%Flukaparticles.At30wt%,wereachthepercolationthresholdandthelargeparticlesarecomingintocontactwitheachother;becausetheydonotsinter,themechanicalstabilityofthelayerislost.ThesecondconsequenceofthepresenceofaSiO2layeratthesurfacewasthattheseparticleswereinactive,withre-specttotheinjection,whichexplainswhy,inadditiontothefactthatmoreFlukaparticleslowerthesurfaceareaofthefilm,theoptimumcompositionwas85%ofnanocrystallineparticlesthathadbeenautoclavedatatemperatureof230°Cand15%Flukaparticles.
Acomparisonofspecularanddiffusetransmissionsbetweenasolarcellwithatransparentelectrodeandasolarcellthatincorporatesscatteringcenters(seeFig.18)clearlyillustratestheeffectofthescatteringparticles.Increasedadsorptioninthesamplethatcontainsthescatteringparticlesleadstobetterinjection,especiallyintheregionof>600nm.Tounderstand
theinfluenceofotherimportantparameters,suchasthesizedistributionofthescatteringparticlesandthethicknessofthelayer,wearecurrentlydevelopingamodelfortheopticalprop-ertiesofthesolarcell.
IV.
Conclusions
WehaveshownthattheTiO2electrodemicrostructurecanbecontrolledbycontrollingdifferentprocessingstepsduringthesynthesisofthenanoparticles.ThesizeoftheprimaryparticlesandthesurfaceareaarecontrolledbyparameterssuchasthetemperatureandpHduringthehydrothermalstep.Theaverageporesizeisdependentontheaveragesizeoftheag-
gregatesthatareformedduringthepeptization.Thepore-sizedistributioncanbealteredbyaddingvariousamountsofbinderorbyvaryingthesinteringtemperatureandtime,thuspromot-ingporecoarsening.Theformationofrutileduringthehydro-thermalstepcanbeavoidedbymaintainingatemperatureof<240°Corbypreventingaggregationbymodifyingthepre-cursorswithaceticacidorsomepolymericdispersingagent,suchaspoly(acrylicacid)(PAA).
Bycontrollingthepore-sizedistribution,wecanensureagooddiffusionoftheelectrolyteand,therefore,increasethelinearityandfillfactorofthecell.Iftheporesaretoosmall(i.e.,Յ4nm),adegradationintheperformanceofthesolarcellisobserved.Bycontrollingthedensityandthesurfaceareaofthenanocrystallinefilm,wecancontroltheamountofdyethatisadsorbedintheelectrode.Thisquantityisdirectlylinkedtotheelectroninjectionandthecurrentthatisproducedbythecorrespondingsolarcell.ThecurrentcanbefurtherenhancedbytreatingtheTiO2electrodewithTiCl4,whichresultsininterparticleneckgrowthandfacilitatestheelectronpercola-tionthroughtheTiO2.
Theadditionoflargerparticles(300–400nm)tothenano-particlesincreasesthescatteringofthefilmandincreasestheadsorptionintheredportionofthespectrum,whichincreasesboththeinjection,especiallyinthewavelengthregionof>600
December1997NanocrystallineTitaniumOxideElectrodesforPhotovoltaicApplications3171
nm,oftransparencytheandscatteringthecurrentalsothatisiscriticaldeliveredbythesolarcell.ControleffectFinally,theischoicerequiredof(i.e.,titaniumactiveforsomeapplicationswhereprecursorswindow).
hasasubstantialtureonboththecrystallitesizeandtheelectrodemodifyingand,consequently,thefinalphotovoltaicproperties.microstruc-Bymodifyticles;tionthisthethetitaniumisopropoxidewithaceticacid,weresults,condensationafterreaction,thusproducingsmallerpar-ide),ofpreventedwhichlargerarecrystalliteshydrothermaltreatment,intheproduc-verylightly(comparedwiththepureisopropox-Althoughprecursorthebythereasonspresencearenotofisopropoxyaggregated.groupsTheaggregationatthesurface.isformances.leadsproach,Wetohopeelectrodesyetfullyelucidated,themodifiedthat,bywithusingimprovedphotovoltaicper-matelyhigh-powerbreakwecanapplications.thefurtherthistypeofmolecularap-10%efficiencyimproveourbarrier,photoelectrodeswhichisneededandulti-forAcknowledgments:
ConradinvoltaicfrommeasurementsVonPlanta,forTheauthorswouldliketothanktheircoworker,
andfruitfulDr.PauldiscussionsBowenandandProfessortechnicalhelpHeinrichfortheHoffmanphoto-powder-characterizationthePowderTechnologyauthorsequipment,LaboratoryaswellatEPFLasforforstimulatingprovidingaccesstotheirthehigh-resolutionareindebtedtransmissiontoR.Wessickenelectronforthemicroscopyhelpinrecordingdiscussions.(HRTEM)andpatterns.
interpretingTheReferences
1(1995).A.P.Alivistos,‘‘SemiconductorNanocrystals,’’MRSBull.,20[3]23–322chiometryJ.C.Parker(1992).ofNanophaseandR.W.MetalSiegel,Oxide‘‘OpticalMaterials,’’DeterminationNanostruct.oftheMater.,Oxygen1,53–57Stoi-3mentJ.S.MaterialsinForesiMicrocrystallineandT.D.Moustakas,‘‘Piezo-resistanceandQuantumConfine-ResearchSocietySilicon’’;Symposiumpp.77–82Proceedings,inLightVol.Emission256.EditedfrombySilicon,S.S.
Iyer,PA,1992.R.T.Collins,andL.T.Canham.MaterialsResearchSociety,Pittsburgh,4CeramicsT.G.Nieh,andCeramicJ.Wadsworth,Composites,’’andF.Wakai,Int.Mater.‘‘RecentRev.,Advances36,146–61inSuperplastic(1991).5‘‘Preparation,M.M.Boutz,R.J.OldeScholtenhuis,A.J.YttriaMicrostructuralControlandSuperplasticityWinnubst,andofA.NanostructuredJ.Burggraaf,BritishStabilizedCeramicsTetragonalProceedings,ZirconiaVol.51.Ceramics’’;R.Freer,London,pp.75–86U.K.,inNanoceramics,1993.6CeramicQ.XuandZrOMembranes:andM.A.Anderson,Preparation‘‘Sol–GelandRoutetoSynthesisofMicroporousSoc.,Characterization77[7]1939–45ofMicropourous(1994).TiO27SwitchingA.Hagfeldt,2Xerogels,’’J.Am.Ceram.N.Vlachopoulos,andM.Soc.,141[7]in82–84Nanocrystalline(1994).OxideSemiconductorGraetzel,Films,’’‘‘FastJ.ElectrochromicElectrochem.8BasedB.O’ReaganandM.Graetzel,‘‘ALow(1991).onDyeSensitizedColloidalTiOFilms,’’Cost,NatureHigh(EfficiencyLondon),353Solar,737–39Cells29Liska,M.K.Nazeeruddin,A.Kay,I.Rodicio,R.Humphry-Baker,E.Mu¨ller,P.byChem.Charge-TransferN.Vlachopoulos,andM.Graetzel,‘‘ConversionofLighttoElectricitySoc.,115,6382–90Sensitizers(1993).onNanocrystallineTiO2Electrodes,’’J.Am.1023J.Livage,‘‘Synthesis,StructureandApplicationsofTiO2posiuminBetterUlrich.MaterialsProceedings,CeramicsResearchVol.Through73.Society,EditedChemistryPittsburgh,byC.II,J.MaterialsBrinker,ResearchGels’’;pp.717–PA,1986.D.E.Clark,SocietyandSym-D.R.11GraP.Bonhoˆte,A.-P.Dias,N.Papageorgiou,K.Kalyanasundaram,andSalts,’’¨tzel,Inorg.‘‘Hydrophobic,Chem.,35,Highly1168–78Conductive,(1996).Ambient-TemperatureMoltenM.12crystallineC.VonPlanta,‘‘Photo-electricalCharacterizationofDyeLausanne,Switzerland,Cells’’;Ph.D.1996.Dissertation.SwissFederalInstituteSensitizedofTechnology,Nano-13als’’;P.seeMcFaydenRef.5,pp.and175–85.D.Fairhurst,‘‘ZetaPotentialsofNanoceramicMateri-14Paris,J.-P.France,Jolivet,1994.DelaSolutiona`l’Oxide;pp.193–95.Intereditions/CNRS,15J.-P.Jolivet,ibid,pp.99–100.16Press,G.NewW.SchererYork,1990.andC.J.Brinker,Sol–GelScience;pp.43–49.Academic17ficationsC.Sanchez,Processing.ofTitaniumF.Babonneau,S.Doeuff,andA.Leaustic,‘‘ChemicalModi-Wiley,NewPrecursors’’;York,1988.pp.77–87inCeramicUltrastructureandᮀ
因篇幅问题不能全部显示，请点此查看更多更全内容