Characterization of CdS thin films grown by chemical bath deposition using four different - PDF document

CharacterizationofCdSthinfilmsgrownbychemicalbathdepositionusingfourdifferentcadmiumsourcesHaniKhallaf,IsaiahO.Oladeji,GuangyuChai,LeeChowDepartmentofPhysics,UniversityofCentralFlorida,Orlando,FL32816,USASOLMAXTECHNOLOGIES,1209WestGoreSt.,Orlando,FL32805,USAApolloTechnologies,Inc.,205WaymontCourt,Suite111,LakeMary,FL32746,USAReceived17April2007;receivedinrevisedform25November2007;accepted2January2008Availableonline9January2008Acomprehensivestudyoftheeffectofcadmiumsourcesonchemicalbathdepositedcadmiumsulfidethinfilmsisreported.Fourdifferent A vailable online at www.sciencedirect.com ThinSolidFilms516(2008)7306 www.elsevier.com/locate/tsf Correspondingauthor.Tel.:+14078232333;fax:+14078235112.E-mailaddress:chow@ucf.edu(L.Chow).0040-6090/$-seefrontmatter©2008ElsevierB.V.Allrightsreserved.10.1016/j.tsf.2008.01.004 Asummaryofoptical/electricalpropertiesandcrystallinityofCBD-CdSfilmsreportedintheliteratureisshowninTable1Table1indicates,twoattempts[11,18]tostudytheeffectofCdsourcesonCdSthinfilmpropertieswerereported,withtwodifferentCdsourcesbeingusedineachcase.TheobjectiveofthisworkistoprovideacomprehensivestudyontheeffectofCdsourcesonthephysicalpropertiesofCBD-CdSthinfilms.2.ExperimentaldetailsEachbathcontained100120mlofde-ionizedwater(resistivity18M-cm)keptunderstirringat70°C.Quartzsubstratesof3.8cm×3.8cmwereusedintheentirework.Thecleaningstepsofthesubstrateandthegrowthproceduresarereportedelsewheree.WiththehelpofTeflonholders,allsubstrateswerekeptverticallyinthesolutions.Allsubstrateswereheldinthebathpriortotheadditionofanyofthereagents.Singledipdepositions(for15minutes)aswellasmulti-dipdepositions(foursuccessivedepositionsfor812mineach)werecarriedout.Singledipdepositionswereusedsolelytostudythefilmthickness/growthratedependenceonCdsource.Multi-dipdepositionswereusedtoobtainthickerfilmsforfurthercharacterizations.Cadmiumsulfate,acetate,iodide,andchloride(0.2mmoleach)wereusedasCdprecursors,whilethiourea(0.4mmol)wasusedassulfurprecursor.Ammoniawasusedasacomplexingagent,similartotheworkofKitaevetal.al..Toensureastablecomplex,weemployedammoniumsaltasNHbuffer.So,whenwasusedasCdsource,(NHwasusedasabufferandwhenCdClwasusedasCdsource,NHClwasusedasabufferandsoforth.TheanionsfromthecadmiumandammoniumsaltsarebelievedtoplayaroleinthegrowthofCdSfilms.Thisroleissuspectedtobethatofacomplementarycomplexingagent.So,inadditiontoCd(NHcomplexinthesolution,wehaveCdanioncomplex.ThestabilityconstantsofthesecomplementaryCdcomplexesaswellasthatoftheCdCdaresummarizedinTable2Eachreagentwasdissolvedin1015mlofde-ionizedwaterbeforebeingaddedtothemainsolution.Aftertheadditionofallreagents,thefinalsolutionwas150ml.Thebathtemperaturewaskeptconstantduringtheentiredepositionprocess.Alpha-step500surfaceprofilometer(Tencor)wasusedtodeterminethefilmthickness.Speculartransmissionmeasurementshavebeendoneatroomtemperaturewithunpolarizedlightatnormalincidenceinthewavelengthrangefrom200to1000nmusingCary500(Varian)doublebeamUV/VISspectrophotometer.Specularreflectancemeasurementshavebeencarriedoutatanangleofincidenceof7°inthewavelengthrangefrom1200to350nm.Theopticalabsorptioncoefficientwascalculatedforeachfilmusingtheequation:ðÞðisthefilmthickness,andaretheintensityoftransmittedlightandinitiallight,respectively.Theabsorptioncoefficientisrelatedtotheincidentphotonenergyh isaconstant,istheopticalbandgap,andisequalto1fordirectbandgapmaterialsuchasCdS.Thebandgap Table1Asummaryofoptical/electricalpropertiesandcrystallinityofCdSfilmsreportedintheliteratureandthecorrespondingCdsourceusedCdsourceOtherreagentsusedinsolutionpH(°C)Crystalstructure-cm)Reference1.CdCl1180Hexagonal2.4510[11]Amorphous2.50102.CdClCl/TU11.540Cubic2.451010CdI2NH4OH/NH4I/TUN60Hexagonal2.62103.CdSO85Hexagonal2.4510104.CdSO60Mixed2.382.4510105.CdSOOH/TU1112702.4710106.CdAc2.3910107.CdAcAc/TU95090Mixed2.3510108.CdAcTEA+NHH9.CdAc2.58101010.Cd(NO/NaOH/TU7.813.520CubicCubicaAc=Acetate;(CHTU=Thiourea;SC(NH Table2StabilityconstantsofsomeCdcomplexescomplexesCd[L]nStabilityconstant(logscale) Table3FilmthicknessdependenceonCdsource(singledipdeposition,for15min,wascarriedoutforeachfilm)CdsourceThickness(nm)H.Khallafetal./ThinSolidFilms516(2008)7306 wasdeterminedforeachfilmbyplotting(versushthenextrapolatingthestraightlineportiontotheenergyaxis.ScanningElectronMicroscopy(SEM)micrographswereobtainedusingJEOL6400FSEMatanaccelerationvoltageof10kV.RutherfordBackScattering(RBS)measurementsweredoneusing2.25MeV-particlesIONIX1.7MUTandetron,withasurfacebarrierdetectorwhichhasanenergyresolution15keV(fullwidthathalfmaximumFWHM),positionedatascatteringangleof165°.X-RayDiffraction(XRD)wascarrieroutusingRigakuDXRDunit(with40kV,30mACuKradiation,=0.15406nm).Thesamplewasmountedat2.5°andscannedfrom25°70°instepsof0.02°withascanrateof1.2°min.Resistivity,mobility,andcarrierconcentrationwereevaluatedbyHalleffectmeasurementsatroomtemperatureinaVanderPauwfour-pointprobeconfiguration,usingindiumcontacts,inanautomatedHalleffectunit(QuantumTechnologyCorp.,BlaineWA,USA)withamagneticinductionof0.75T.3.Resultsanddiscussion3.1.ThicknessdependenceonCdsourceTable3showsthefilmthicknessdependenceontheCdsourceusedinthedepositionprocess.ThehighestthicknessobtainedwasinthecaseofCdSO,andtheleastthicknessobtainedwasintheCdIcase.Apparently,amongallotherCdsalts,CdIalwaysresultsinamuchthinnerfilm.ThisobservationwasinagreementwithwhatwasreportedbyKitaevetal.al.,andOrtega-BorgesandLincotLincot.However,inourwork,CdSO,notCdCl,resultsinthehighestfilmthickness.ThiscanbeunderstoodbyconsideringtheroleplayedbythecomplementaryCdcomplexesinthedepositionprocess.Table2showsthattheCd[I]complexhasamuchhigherstabilityconstantthanCd[Cl],Cd[CHCOO],orCd[SOcomplexes.ThismeansmuchslowerreleaseofCdionsand Fig.1.SpeculartransmissionspectraofCdSfilmsgrownusingfourdifferentCdsources;CdSOm),Cd(CHm),CdClm),and Fig.2.SEMmicrographofCdSfilmsgrownusingfourdifferentCdsources.H.Khallafetal./ThinSolidFilms516(2008)7306 consequentlyamuchthinnerCdSfilmwhenCdIwasused.ThesameargumentcanbeusedtoexplainwhythehighestthicknesswasobtainedwhenCdSOwasused.Actually,theorderinwhichthestabilityconstantdecreasedwasexactlythesameorderthefilmthicknessincreased.IntheKitaevetal.al.case,theClconcentrationwastoolowtomaketheCd[Cl]stable.Whereasinourcase,theCd[Cl]complexhasahigherstabilityconstantthantheCd[CH,ortheCd[SOcomplexeswhichcausedthefilmthicknesstobelessthanthatoftheCd(CHortheCdSO-basedCdSfilm.Itshouldbenotedthat,comparingthesestabilityconstantstothatofthecomplex(Table2),theroleofthecomplementaryCdcomplexesinthegrowthprocessisveryimportant.3.2.OpticalpropertiesFig.1showsopticaltransmissionspectraofallfourfilms.Allfilmshavehightransmission,withthetransmissioninthecasebeingbetterthanthatoftheotherthreefilms.Thiswasactuallyexpected,sincetheSEMmicrographsshowninFig.2,showedtheCdCl-basedfilmtobemuchsmootherandmoreuniformthantheotherthreefilms.Thesurfaceroughness,duetocoveragebycrystalliteovergrowth,causeslightscattering,whichinturnlowersthetransmission.AsindicatedinthecaptionsofFig.1,there'snoappreciabledifferenceinfilmthicknessforallfourfilms.Fig.3showsspecularreflectanceforthetwofilmswithhighestandlowesttransmissions.ReflectanceofCdCl-basedfilmishigherthanthatofCdSO-basedfilm.So,althoughnodiffusetransmission/reflectancemeasurementswerecarriedout,webelievethattheSEMmicrographsaswellasthereflectancemeasurementsaresufficienttoconcludethatsurfacemorphologyisresponsibleforthehighertransmissionobservedintheCdClAnotherobservationaboutthesetransmissionspectraisthatandCd(CHCOO)-basedfilmssharethesameabsorptionedge.Aredshift(towardslongerwavelengths)intheCdClbasedfilmcaseandablueshiftinthecaseoftheCdSO4-basedfilmwereobserved.Thisreflectedontheircorrespondingopticalbandgap,asshowninFig.4.Theopticalbandgapofthe-basedfilmisthehighest(2.36eV)whilethebandgapoftheCdCl-basedfilmisthelowest(2.25eV).TheCdIandCd-basedfilmshaveanintermediatebandgap(2.31eV).ThisbandgapdependenceonCdsourceagreeswithwhatRamietal.[11]andNakanishiandItoItoreportedearlier.3.3.CrystalstructureXRDpatternsofthefourfilmsareshowninFig.5RegardlessoftheCdsaltused,allfilmswerecubicwithamain(111)reflectionandtwoweaker(220)and(311)peaksthatconfirmedthecubicnatureofallfilms.AfourthpeakwasdetectedintheCdCl-basedfilmcase,whichwasfoundtobethe(200)peak,whichisalsoacharacteristicpeakofcubicCdS.AsummaryofXRDdataforstandardcubicCdSpowderpowderandallfourfilmsisshowninTable4.Theobservedrelativeintensitiesimplythatallfilmsarepolycrystallinewithpreferred(111)orientation.However,whentherelativeintensitiesarecarefullyinvestigated,it'sobviousthatthedegreeoftexturealongthe(111)orientationincreasesintheorder:CdCl,CdI,CdSO.AsshowninTable5,therelative Fig.3.SpecularreflectanceofCdCl-basedfilm(0.20m)andCdSOfilm(0.18 Fig.4.OpticalbandgapcalculationsofCdSfilmsgrownusingfourdifferentCd Fig.5.XRDpatternofCdSfilmsgrownusingfourdifferentCdsources.H.Khallafetal./ThinSolidFilms516(2008)7306 intensitiesofboth(220)and(311)peaksareclosesttothatofpowderCdSwhenCdClisbeingusedasCdsource.BothpeaksaremoresuppressedwhenCd(CHCOO)andCdIbeingused,andalmosttotallysuppressedwhenCdSOisbeingusedwheretheirrelativeintensitiesdecreasetolessthan3%.ThismayexplaintheopticalbandgapvariationwithCdsource.AsshowninTable5,asthe(111)peakbecomesmorepredominantandtheotherpeaksgetmoresuppressed,thebandgapincreases.That'swhyCdCl-basedfilmhastheleastbandgap(2.25eV)andCdSO-basedfilmhasthehighestbandgap(2.36eV)whichistheclosest,amongstallfourfilms,tothebandgapofsinglecrystalCdS(2.42eV)eV).Theothertwofilmsareexpectedtohavethesamebandgapsincerelativeintensitiesofboth(220)and(311)peaksinCd(CHCOO)basedfilmareveryclosetothoseofCdI-basedfilm.Theintermediatevaluesoftheserelativeintensitiesmayalsoexplaintheirintermediatebandgap.SuchbandgapdependenceonfilmcrystallinityhasbeenobservedforCuInSeaswellasCuGaSepolycrystallinethinfilms;Chichibuetal.al.foundthatfilmswithapredominant(112)orientationandveryweakintensitiesfromotherdiffractionpeakshaveahigherbandgapthanfilmswithcomparativeintensitiesof(112)and(220)or(204)diffractionpeaks.TheyattributedthisbandgapdecreasetolatticedefectsinthelatterfilmsthatmaygiverisetoplasmascreeningofCoulombinteractionsduetodegradationoffilmquality.TheyreportedsimilarbandgapdependenceonfilmcrystallinityforCuGaSeTheaveragecrystallitesizeshowninTable6wascalculatedusingtheDebyeScherrerformulaormula.ThenoticeabledifferenceingrainsizebetweentheCdSOfilmandtheotherfilmsmaybeattributedtothefactthattherearetwodifferentdepositionprocessesinCBDthatcompetewitheachotherother;clusterbyclusterdepositionandionbyiondeposition.Now,sinceusingCdSOresultedinamuchfastergrowthratethantheotherthreefilms,thismayindicatethattheclusterbyclusterdepositiondominatedthedepositionprocessandconsequentlyresultedinamuchlargergrainsize.However,inthecaseoftheotherthreefilms,ionbyiondepositiondominatedthedepositionprocessandasaresultamuchsmallergrainsizeandthinnerfilmswereobtained.3.4.StoichiometryUsingRBSmeasurements,thestoichiometryofallfilmswerestudied.SimulationusingRutherfordUniversalManip-ulationProgram(RUMP)P)wasimplementedtofindthebestpossiblematchtotheRBSspectra.AsshowninFig.6,theS:Cdratiowasfoundtobe(1.00:1.00)inthecaseofCdI-basedfilmaswellasCdCl-basedfilm.However,morecadmiumwasdetectedintheothertwofilms.WefoundtheS:Cdratiotodecreaseto(1.00:1.06)whenCd(CHCOO)wasused.Thisratiodecreasedfurtherto(1.00:1.09)whenCdSOwasused.ThisissummarizedinTable7.TheseratiosappeartoagreewiththeCd-complexstabilityconstantssummarizedinTable2Apparently,thesmallerthestabilityconstantofthecomple-mentarycomplexthemoretheCdionsinthegrowthsolutionthatendupasexcessCdintheCdSfilm;whereas,whenthestabilityconstantishighthereleaseofCdandSiswellcontrolledleadingtoahighlystoichiometricCdSfilm.3.5.HalleffectmeasurementsTable7showstheHalleffectmeasurementsconductedforthefourfilms.TheyareinagreementwiththeRBSmeasurements.Thecarrierconcentrationdecreases(andaccordinglytheresistivityincreases)intheorderCdSO,CdCOO),CdCl.AccordingtotheRBSresults,theS: Table4AsummaryofXRDdataforstandardcubicCdSpowderpowderandCdSfilmsgrownusingfourdifferentCdsourcesSample2(nm)()I/ICubic(powder)26.5470.33550(111)100.030.7480.29055(200)21.644.0400.20545(220)46.752.1630.17521(311)33.254.6700.16775(222)4.564.0420.14528(400)5.8-basedfilm26.6400.33434(111)100.044.4350.20372(220)2.952.2800.17484(311)2.7-basedfilm26.5800.33509(111)100.044.1800.20483(220)18.452.2400.17497(311)13.5-basedfilm26.8000.33239(111)100.030.8370.28973(200)9.344.0790.20528(220)28.952.4800.17422(311)19.4-basedfilm26.9390.33070(111)100.044.3230.20421(220)16.452.6400.17373(311)11.8 Table5AsummaryofrelativepeakintensitiesandopticalbandgapforCdSfilmsgrownusingfourdifferentCdsource)Cubic(powder)basedfilmbasedfilmbasedfilmbasedfilm(%)I/I(%)I/I(%)I/I(%)I/I(111)100100100100100(220)46.728.918.416.42.9(311)33.219.413.511.82.7Opticalbandgap(eV)2.252.312.312.36 Table6GrainsizedependenceonCdsourceCdsourceGrainsize(nm)H.Khallafetal./ThinSolidFilms516(2008)7306 Cdratiodecreasesinthesameorder.WebelievetheexcesscontentofCdmeansthateitherinterstitialCdionsorsulfurvacanciesexistinthefilm,actingasdonorsandresultinginanincreaseinthecarrierconcentrationaswellasaconsequentdecreaseintheresistivity.ThemobilitymeasurementsshowtheCd(CHCOO),CdClbasedfilmstohaveverysimilarmobility.However,theCdSObasedfilmwasfoundtohaveamobilitythat'stwotimeshigherthantheotherthreefilms.ThiscanbeunderstoodbyconsideringthegrainsizemeasurementsshowninTable6;thegrainsizeoftheCdSObasedfilm(145nm)ismuchlargerthanthatoftheotherthreefilms(1316nm)whichexplainsthehighermobility[29].Inthemeantime,thegrainsizeoftheotherthreefilmsisalmostthesamewhichmayexplainwhythesefilmshaveveryclosemobilityvalues.ThemobilityvaluesobtainedinthisworkareinagreementwithwhathasbeenreportedearlierintheliteratureforpolycrystallineCdSthinfilmsfilms.4.ConclusionAcomprehensivestudyoftheinfluenceofCdsourcesonelectrical/opticalpropertiesaswellasthickness,structure,surfacemorphology,andstoichiometryofchemicalbathdepositedCdSfilmsispresented.FilmthicknesswasfoundtodecreaseintheorderCdSO,Cd(CH,CdCl,CdIHowever,thebandgapwasfoundtodecreaseintheorder,Cd(CHCOO),CdCl.Allfilmswerefoundtobecubic,regardlessoftheCdsaltused.ThegrainsizedecreasesintheorderCdSO,CdCl,Cd(CH,CdITheRBSdatashowedthattheusageofCdClandCdIresultsinhighlystoichiometircfilms(S:Cdratio=1:1).MoreCdwasdetectedwhenCdSOandCd(CHwereused.TheS:CdratioandcarrierconcentrationwerefoundtodecreaseintheorderCdSO,Cd(CH,CdI/CdCl.CdCl-basedfilmswerefoundtohaveabettertransmissionandmuchsmoothersurfacesthanotherfilms.UsingCdSOasCdsourceleadstothehighestgrowthrate,bandgap,carrierconcentration,andmobility.AcknowledgementsWethankK.ScammonandM.KlimovoftheAdvancedMaterialsProcessingandAnalysisCenter(AMPAC),Univer-sityofCentralFlorida,fortheirhelpwiththeXRD,RBS,andfilmthicknessmeasurements.Thisworkwaspartiallysup-portedbyApolloTechnologies,Inc.andFloridaHighTech.CorridorCouncil. Fig.6.RBSspectrumandRUMPsimulationofCdSfilmsgrownusingfourdifferentCdsources. Table7HalleffectandRBSmeasurementsCdsourceS:CdCarrierconcentration/V-S)1.00:1.091.68×108.01×101.00:1.068.83×103.38×101.00:1.009.74×102.96×101.00:1.008.19×103.88×107311H.Khallafetal./ThinSolidFilms516(2008)7306 [1]S.Mokrushin,Y.Tkachev,Kolloidn.Z.23(1961)438.[2]G.Kitaev,A.Uritskaya,S.Mokrushin,Russ.J.Phys.Chem.39(1965)1101.[3]I.Oladeji,L.Chow,C.Ferekides,V.Viswanathan,Z.Zhao,Sol.EnergyMater.Sol.Cells61(2000)203.[4]M.Contreras,M.Romero,B.To,F.Hasoon,R.Noufi,S.Ward,K.Ramanathan,ThinSolidFilms403/404(2002)204.[5]A.Davis,K.Vaccaro,H.Dauplaise,W.Waters,J.Lorenzo,J.Electrochem.Soc.146(1999)1046.[6]M.Contreras,B.Egaas,K.Ramanathan,J.Hiltner,A.Swartzlander,F.Hasoon,R.Noufi,Prog.Photovolt:Res.Appl.7(1999)311.[7]X.Wu,J.Keane,R.Dhere,D.Dehart,D.Albin,A.Duda,T.Gessert,S.Asher,D.Levi,P.Sheldon,Proceedingsofthe17thEuropeanPhotovoltaicSolarEnergyConference,Munich,Germany,October22262001,p.995.[8]J.Dona,J.Herrero,J.Electrochem.Soc.139(1992)2810.[9]H.Metin,R.Esen,Semicond.Sci.Technol.18(2003)647.[10]I.Oladeji,L.Chow,J.Electrochem.Soc.144(1997)2342.[11]M.Rami,E.Benamar,M.Fahoume,F.Chraibi,A.Ennaoui,SolidStateSci.1(1999)179.[12]P.Nair,J.Campos,M.Nair,Semicond.Sci.Technol.3(1988)134.[13]P.O'Brien,T.Saeed,J.Cryst.Growth158(1996)497.[14]I.Kaur,D.Pandya,K.Chopra,J.Electrochem.Soc.127(1980)943.[15]M.Nagao,S.Watanabe,Japan.J.Appl.Phys.7(1968)684.[16]G.Kitaev,S.Mokrushin,A.Uritskaya,Kolloidn.Z.27(1965)51.[17]R.Ortega-Borges,D.Lincot,J.Electrochem.Soc.140(1993)3464.[18]T.Nakanishi,K.Ito,Sol.EnergyMater.Sol.Cells35(1994)171.[19]C.Guillén,M.Martínez,J.Herrero,ThinSolidFilms335(1998)37.[20]T.Chu,S.Chu,N.Schultz,C.Wang,C.Wu,J.Electrochem.Soc.139(1992)2443.[21]M.Nair,P.Nair,J.Campos,ThinSolidFilms161(1988)21.[22]A.Reádigos,V.García,O.Gomezdaza,J.Campos,M.Nair,P.Nair,Semicond.Sci.Technol.15(2000)1022.[23]L.G.Sillén,A.E.Martell,StabilityConstantsofMetal-IonComplexes,BurlingtonHouse,London,1964.[24]InorganicCrystalStructureDatabase,NationalFachinformationszentrum(FIZ)Karlsruhe;file080-0019,1997.[25]S.M.Sze,PhysicsofSemiconductorDevices,JohnWiley&Sons,NewYork,1981.[26]S.Chichibu,T.Mizutani,K.Murakami,T.Shioda,T.Kurafuji,H.Nakanishi,S.Niki,P.Fons,A.Yamada,J.Appl.Phys.83(1998)3678.[27]E.F.Kaelble,HandbookofX-rays,McGraw-Hill,NewYork,1967.[28]L.Doolittle,Nucl.Instr.MethodsB15(1986)227.[29]I.Mártil,G.González-Díaz,F.Sánchez-Quesada,J.Vac.Sci.Technol.,A2(1984)1491.[30]I.Mártil,G.González-Díaz,F.Sánchez-Quesada,ThinSolidFilms114(1984)327.H.Khallafetal./ThinSolidFilms516(2008)7306