STUDYOFORGANIC-INORGANICHETERO-INTERFACESANDELECTRICAL TRANSPORTINSEMICONDUCTINGNANOSTRUCTURES By SeanRobertWagner ADISSERTATION Submittedto MichiganStateUniversity inpartialentoftherequirements forthedegreeof Physics-DoctorofPhilosophy 2015 ABSTRACT STUDYOFORGANIC-INORGANICHETERO-INTERFACESAND ELECTRICALTRANSPORTINSEMICONDUCTING NANOSTRUCTURES By SeanRobertWagner Astheelectronicsindustrycontinuestoevolveandmovetowardsfunctionalelectronic deviceswithincreasingcomplexityandfunctionality,itbecomesimportanttoexplorema- terialsoutsidetheregimeofconventionalsemiconductors.Organicsemiconductingsmall moleculeshavereceivedalargeamountofattentionduetotheirhighdegreeofy, theoptiontoperformmolecularsynthesistomodifytheirelectronicandmagneticproperties, andtheirabilitytoorganizeintohighly-orderedfunctionalizednanostructuresandthin Beingabletoformcomplexnanostructuresandthinwithmolecularprecision,while maintainingtheabilitytotunepropertiesthroughmointhemolecularchemistry couldresultinvastimprovementsinconventionaldevicearchitectures.However,before thisisrealized,therestillremainsatlackofunderstandingregardinghowthese moleculesinteractwithvarioussubstratesurfacesaswellastheirintermolecularinteractions. Theinterplaybetweentheseinteractionscanproducedrasticchangesinthemolecularorien- tationandorderingatthehetero-interface,whichcanthetransportpropertiesofthe molecularthinandultimatelymodifytheperformanceoftheorganicelectronicdevice. Thisstudyfocusesonthegrowthdynamics,molecularordering,andmolecular orientationofmetalphthalocyanine( M Pc)molecules,particularlyonSi,asubstratewhich isnotoriouslytoformanorganizedorganicthinonduetothesurfacedangling bonds.Bydeactivatingthesebonds,theformationofahighlyorderedorganicmolecular thinbecomespossible.Combiningscanningtunnelingmicroscopy,scanningtunneling spectroscopy,low-energyelectronanddensityfunctionaltheorycalculations,the growthevolutionof M Pcmolecules( M =Zn,Cu,Co)fromthesinglemoleculelevelto multilayeredsonthedeactivatedSi(111)-Bsurfaceisinvestigated. InitialtestsarecenteredaroundthermallyevaporatedZnPc.Thesemoleculesdisplaya highly-ordered,close-packed,tiltedwhichfromanyknownbulkpacking motif.TheZnPcmoleculesareabletorapidlyontheSisurfaceandpreferentially nucleateatSistep-edges.Thisisfollowedbytheformationofhighly-orderedanisotropic stripestructureswhichgrowacrosstheSiterraces,i.e.anisotropicwgrowth.The wgrowthmodefurtherimpactsthegrowthbyreducingtheallowedsymmetryofthe moleculardomainssuchthatthinwithanexclusivein-planemolecularorderingare formed.Additionally,theZnPctiltedpackingmotifstabilizesthemolecularallowing ittomaintainthispackingformultilayereddespitethedecreasingsubstrate Thestrengthofthe M Pc-substrateinteractioncanbemobychangingthecentral transition-metalionwithinthemolecule.Throughselective p - d orbitalcouplingbetween M Pcmoleculesandthesubstrate,thedegreeoforbitalcouplingcaninducemo inthemolecularorderingandorientationof M Pcmoleculesattheinterface. Thesecondaryfocusofthisstudyistoinitiatepreliminaryexperimentationtowards understandinghoworderedorganicmolecularthincanbeappliedtosilicon-based devicesthatcouldhaveatimpactontheelectronicsmarket.Sinanomembraneis alow-dimensionalnanomaterialwithelectronicpropertiesthatarehighlysensitive totheinterfacecondition.Bymergingtheknowledgeof M PcthingrowthonSiwith Sinanomembranetechnology,possibilitiestowardsmodifyingthetransportpropertiesof nanomaterialsthroughengineeringtheorganic-inorganichetero-interfacecanbeexplored. Thisthesisisdedicatedtomylovingwife, CourtneyMarieMich iv ACKNOWLEDGMENTS TheworkdetailedinthisthesiswassupportedbytheU.S.DepartmentofEnergy(DOE) ofScienceEarlyCareerResearchProgram(GrantNo.DE-SC0006400)throughthe ofBasicEnergySciences. Therearealargenumberofindividualswhohaveplayedasigtroledirectlyinthe workscontainedinthisdissertation,aswellasinmylifethroughoutitsduration.Iwould liketotakethisopportunitytoacknowledgethosewhohavebeenthereformeeverystepof theway. First,Iwouldliketothankmyadvisor,ProfessorPengpengZhang,fortakingmeon asstudentduringthegroup'sinfancyandallowingmetoexperienceallofthechallenges associatedwithestablishingaresearchlaboratory.DuringmytimeinherresearchgroupI wasgiventheopportunitytoplayalargeroleinlayingthegroundworkformultipleprojects, aswellashelpingtobringthemtofruition.Thishasprovidedmewithnumerousskillsand experienceswhichIwillrememberfortherestofmysciencareer. OvertheyearsIhavehadtheopportunitytoworkcloselywithavarietyofent groupmatesandformasenseofcomraderyinthelaboratory.Iwouldliketoacknowledge ChuanpengJiang,DanielEnderich,Dr.JiaguiFeng,XiaoyuLiu,NonThongprong,Shreya Nad,Lo-HanYuan,andDr.JiebingSun.Iwouldalsoliketonotehowenjoyableithasbeen tohelptrainAndrewTanwhojoinedtheresearchgroupduringthetimethisdissertation waswritten.Iwishhimthebestinhisfuturework. IfeelprivilegedtohavebeenabletoworkwithProfessorRichardLuntandhisresearch groupintheMolecularandOrganicExcitonicsLaboratoryintheChemicalEngineeringand MaterialsSciencesDepartmenthereatMichiganStateUniversity.ProfessorLuntexposed v metosomenewtypesofanalysistechniqueswhichhavebeenquiteusefulthroughoutmy studies.Wehavesharedmanyfruitfuldiscussionsovertheyearsandhisgrouphasbeen particularlyhelpfulinpurifyingthesmallorganicmoleculesusedthroughoutmystudiesas wellassomeevaporationprocesses.Particularly,IwouldliketoacknowledgeChrisTraverse forallofhishelponthewidevarietyoftestsandforthegoodcompanywhiledoingso.I wouldalsoliketothankMargaret(Peggy)Young,Padmanaban(Paddy)Kuttipillai,and JohnSuddardforallthefunconversations. IwasfortunatetobeabletocollaboratewithDr.MinaYoon,Dr.ChangwonPark,and Dr.BingHuangatOakRidgeNationalLaboratory.Theywereabletoprovideacomputa- tionalinterpretationofourexperimentresultsusingdensityfunctionaltheory,discussedin Chapter8 .Throughthesecalculations,thefundamentalmechanismunderlyingthenature ofthemolecule-substrateinteractionandthegrowthdynamicscouldbeunderstood. IwouldliketoacknowledgeProfessorXianglinKe'sresearchgroupforallowingmetouse thePPMSequipmentandfortheirhelpwiththeSinanomembranetransportmeasurements. Particularly,IwouldliketothankDr.TaoZouforallofhisadviceandhelpthroughoutthe measurementprocess. Throughoutmytimespentperformingexperimentsontheultra-highvacuumsystemin thelaboratory,iftherewereeveranysystemissuesortechnicalquestionsIneededaddressed thesupportatOmicronNanoTechnologyGmbH(OxfordInstruments)wasalwaysable toprovideadvice.Particularly,IwouldliketoacknowledgeDaveWyniaforallofhishelp overtheyears. IwouldliketothanktheinthephysicsandastronomyMachineShopfortheir technicalsupportandhelpconstructinganyextraequipmentthatmyexperimentsmay haverequired.ThankyouTomPalazzolo,TomHudson,JimMuns,andRobBennett. vi Dr.RezaLoloeeisoneindividualIwillneverforget,andIencourageallofthenew experimentalcondensedmatterphysicsstudentstogettoknowhim.Whetherheisgiving adviceonyourcurrentexperimentproblems,usinghismagictouchtomiraculouslymake anyexperimentalapparatuswork,orsimplyaskingyouhowyourdayisgoing,healways awaytomakeyourdayjustalittlebitbrighter.Thebasementwouldbeavery tplacewithouthim.Thankyouforyourendlesshelp,andforyoursoundadvice overtheyears. NolesscriticaltomysuccessisDr.BaokangBi,whomanagesandmaintainstheclean roomoperationsforthephysicsdepartment.Hehasbeentherethroughoutthemanychanges insamplepreparationproceduresthatIhaveencounteredovertheyearsandconstantly hisownadvice.Hisquickwitandjovialdemeanorhavemadethecleanroomavery pleasantenvironmentformywork. IwouldberemissifIdidnotacknowledgethekindnessIhavereceivedfromProfessor StuartTessmer,aswellastheentireTessmerresearchgroupoverthemanyyearsatMichigan StateUniversity.Stuartsparkedmyinterestinacademicresearch,particularlyusing scanningprobemicroscopytechniques,andtheskillsIobtainedthroughhisguidancehave helpedmebecomeasuccessfulresearcher.Iwouldliketothankallofthepast,current,and honorarygroupmembersthathaveplayedatroleinmyresearchcareer:Dr.Josh Veazey,Dr.MeganRomanowich,Dr.SanelaLampa-Pastirk,Dr.KathyWalsh,IanDayton, MattDeNinno,Dr.MorewellGasseller,Dr.ShannonNicley,andDr.JohnRaguse DuringmytimeinthephysicsandastronomydepartmentIhavegottentoknowalarge numberofwonderfulpeople{Dr.ZachMeisel,Dr.Joe(Danny)Kimball,JessieMuir,Dr. BrendonMikula,Dr.onBarthelemy,AndyParpart,Dr.JohnNovak,AlexBaumann, Dr.JaylanJones,BradSchoenrock,ProfessorDonaldMorelli,Dr.EricGingrich,Dr. vii YixingWang,MarkOlson,Dr.MazinKhasawneh,Tung-WuHsieh,ConnorGlosser,Dr. EricMacaulay,MatthiasMuenks,ProfessorNelsonSepulveda,AlexCramer,JosephGlick, JenniferGlick,BethanyNiedzielski,VictorAguilar,Dr.SteveQuinn,AimeeShore,Luke Titus,Dr.ChengSun,Dr.Yung-HsiuTang,MengzuZhu,KevinBerry,andmanyothers.I greatlyenjoyedallthemomentswewereabletospendtogetheratMichiganStateUniversity. Also,IwouldliketothankDebbieBarrattandCathyCordsforalloftheirhelpoverthe years.Whetheritwascourseworkrelatedormakingsuretherightthingsgetorderedfor thelab,theywerealwaystheretolendahand. ThereareafewnamesinparticularthatIwouldliketorecognize{StephenDeCamp,Dr. WilliamMartinez,andAndrewChegwidden,allofwhomhavebecomeveryclosefriends. Thanksforallofthefuntimesandforbeingthereformethroughoutthiswholeprocess. Also,KimCrosslan.Thankyouforyourunendingsupport,thelaughs,andalwayshelping toliftmyspiritswhentimeswererough.Iwilltrulymissourconversations.Icouldnotask forabetter\physicsmom". Last,butbynomeansleast,Iwanttothankmyfamily{mywifeCourtney,mymother andfather,Matt,Lauren,Katie,Chris,Whitney,Chris,Jim,andKathie,fortheirencour- agementandsupport.IthasbeenalongroadandIwouldneverhavebeenabletodothis withouteachandeveryoneofyou.Tomywife:Thankyouforputtingupwithallofthelong workdays,thelatenights,andthemissedmeals.Youhavebeenmyconstantthroughout thisentireexperienceandIamforevergratefulforyourloveandsupport. Thankyou,everyone. viii TABLEOFCONTENTS LISTOFTABLES .................................... xii LISTOFFIGURES ................................... xiii Chapter1Introduction ............................... 1 1.1Motivation .....................................1 1.2ContentsandLayoutofThisWork .......................5 Chapter2OrganicMolecularThinFilmGrowth ............... 9 2.1InorganicFilmGrowthNearEquilibrium ....................10 2.2KineticProcessesinInorganicFilmGrowth ..................14 2.3ClassicalNucleationTheory ...........................20 2.4KineticGrowthModeTransitionsforInorganicGrowth ............23 2.5BetweenOrganicandInorganicThinFilmGrowth ........25 2.6OrganicMolecularGrowthonVariousSubstrates ...............27 2.6.1MetallicSurfaces .............................27 2.6.2InsulatingSurfaces ............................29 2.6.3SemiconductingSurfaces .........................30 2.6.4Summary .................................31 Chapter3ExperimentalandComputationalMethodologies ........ 33 3.1ExperimentEquipment ..............................34 3.1.1Ultra-HighVacuumSystem .......................34 3.1.2AtomicForceMicroscopeandInvertedOpticalMicroscopeSet-Up ..38 3.1.3PhysicalPropertyMeasurementSystem ................42 3.2ExperimentalTechniques .............................43 3.2.1ScanningProbeMicroscopy .......................43 3.2.1.1ScanningTunnelingMicroscopy ................43 3.2.1.2ScanningTunnelingSpectroscopy ...............51 3.2.1.3tialConductanceMapping ..............54 3.2.1.4AtomicForceMicroscopy ...................54 3.2.1.5KelvinProbeForceMicroscopy ................62 3.2.2Low-EnergyElectron .....................65 3.2.3VanderPauwTransportMeasurement .................69 3.2.4HallMeasurement .............................73 3.3ComputationalTechniques ............................76 3.3.1EpitaxialRelationwiththeSubstrate ..................76 3.3.1.1GeometricPhaseCoherenceModel ..............79 3.3.1.2SimulatedMoirePattern ....................80 3.3.2SimulatedLEEDPatterns ........................81 ix 3.3.3DensityFunctionalTheory ........................82 Chapter4BackgroundofMaterials ....................... 85 4.1IntroductiontoSemiconductorPhysics .....................85 4.2SiliconSurfaces ..................................90 4.2.1TheSi(111)7 7SurfaceReconstruction ................91 4.2.2TheDeactivatedSi(111)-B p 3 p 3 R 30 SurfaceReconstruction ..95 4.2.3TheHydrogenPassivatedSi(111)-HandSi(001)-HSurfaces ......99 4.3Silicon-on-InsulatorTechnology .........................103 4.3.1SiNanomembranes ............................106 4.3.2FabricationofSiNanomembranes ....................109 4.4MetalPhthalocyanineMolecules .........................111 Chapter5AnisotropicStep-FlowGrowthofZincPhthalocyanineonthe DeactivatedSi(111)-BSurface .................... 116 5.1ZnPcMolecularPackingandEpitaxialRelationonDeactivated Si(111)-B .....................................117 5.2ObservationsandDiscussionoftheAnisotropicStep-FlowGrowthMode ..123 5.3InterruptedAnisotropicStep-FlowGrowthMode ...............126 5.3.1DeactivatedSi(111)-BSurfaceDefects .................126 5.3.2Ehrlich-SchoebelBarrier ....................129 5.3.3AdditionalActivationBarriersandNucleationatDomainBoundaries 132 5.4Conclusions ....................................134 Chapter6Out-Of-PlaneGrowthStudyofZincPhthalocyanineonthe DeactivatedSi(111)-BSurface .................... 136 6.1Low-EnergyElectronMeasurements ................137 6.1.1Sub-MonolayerGrowthofZincPhthalocyanineandSimulatedLow- EnergyElectronDPattern ...................137 6.1.2MultilayeredGrowthofZincPhthalocyanineandSimulatedLow-Energy ElectronPattern .......................142 6.2ComparisonwithScanningTunnelingMicroscopyMeasurements .......146 6.3Conclusions ....................................150 Chapter7TemperatureDependentGrowthofMetalPhthalocyanineson theDeactivatedSi(111)-BSurface ................. 152 7.1GrowthModeTransitionsofZincandCopperPhthalocyanine ........153 7.2SmoothingofZincPhthalocyanineFilmsatElevatedTemperatures .....162 7.3Conclusions ....................................165 Chapter8MetalPhthalocyanineGrowthonDeactivatedSi(111)-BMe- diatedbySelectiveOrbitalCoupling ................ 167 8.1GrowthComparisonbetweenZinc,Copper,andCobaltPhthalocyanine ...168 8.2MolecularBindingMechanismtotheSi(111)-B ................174 8.3Transition-MetalIonInducedModulationinthePotentialEnergyLandscape 178 x 8.4Conclusions ....................................184 Chapter9TuningSiNanomembraneTransportPropertiesbyanInter- facialMetalPhthalocyanineThinFilm .............. 186 9.1GrowthofMetalPhthalocyanineFilmsonHydrogenPassivatedSiSurfaces .187 9.1.1GrowthontheSi(111)-HSurface ....................187 9.1.2GrowthontheSi(001)-HSurface ....................191 9.2WorkFunctionMeasurementsofPhthalocyanineonPassivatedSiNanomem- branes .......................................193 9.3TransportMeasurementsofSiNanomembranes ................199 9.4Conclusions ....................................200 Chapter10ConclusionandFutureProspects .................. 202 10.1SummaryofResults ...............................202 10.2FutureWork ....................................207 Appendices ........................................ 211 AppendixATransportApplicationInsertProbeDesign ...............212 AppendixBOtherContributedWorks ........................219 BIBLIOGRAPHY .................................... 222 xi LISTOFTABLES Table8.1: M Pc- M Pcand M Pc-Sibindingenergy(BE) .............182 xii LISTOFFIGURES Figure2.1:SchematicofVolmer-Webergrowth ..................11 Figure2.2:SchematicofFrank-vanderMerwegrowth ..............12 Figure2.3:SchematicofStranski-Krastanovgrowth ................13 Figure2.4:Schematicofadatomadsorptionandprocessesonasurface 16 Figure2.5:SchematicdiagramoftheEhrlich-Schoebelbarrier(ESB) .....18 Figure2.6:SchematicillustrationoftheZenoinorganicandinorganic growth ..................................19 Figure2.7:Schematicdisplayingatwo-dimensionalnucleus,approximatedinsize byacircleofradius r ..........................21 Figure2.8:Schematicofttypesofkineticlimitedgrowthmodesofadatoms depositedonasurface ..........................24 Figure3.1:ImageoftheUHVsystemutilizedfororganicmolecularthin growthstudiesinvestigatedusingSTMmethodsandLEED .....34 Figure3.2:ImageoftheSTMscanningstagecontainedwithintheUHVsystem 36 Figure3.3:ImageofatypicalmechanicallycutPt/IrSTMtipconnectedtoatip carrierforuseintheUHVsystem ...................37 Figure3.4:ImageofthecombinedinvertedopticalmicroscopeandAFMset-up 38 Figure3.5:ImageofanAFMcantlieverinthecantileverholder ......39 Figure3.6:tviewsofthesealednitrogenenvironmentwhichcanbeat- tachedtotheAFMset-upallowingforAFMstudiesinnitrogen ..40 Figure3.7:Imageofthenitrogengloveboxthatisusedtopreparethesealed nitrogenenvironmentattachmentforAFMstudies ..........41 Figure3.8:ImageofthePPMSset-uputilizedforperformingtransportmeasure- mentsonSinanomembranedevices ..................42 xiii Figure3.9:Diagramrepresentationofdirectedelectrontunnelingthroughavac- uumbarrierbetweenthesampleandthetip .............44 Figure3.10:Graphicalrepresentationoftheelectronwavefunctiontunnelingthrough avacuumbarrierofwidth d betweenthesampleandthetip ....47 Figure3.11:Cross-sectionalschematicdrawingofthescanningpiezoelectricele- mentcommonlyutilizedinSPMexperiments .............49 Figure3.12:SchematicdiagramdepictingtheoperationofSTMwhileinconstant currentmode ...............................50 Figure3.13:SchematicdiagramdepictingtheoperationprincipleofAFM ....55 Figure3.14:ExampleplotoftheLennard-Jonespotentialfortwointeractingatoms 56 Figure3.15:PlotdepictingthechangeintheinteractionforcebetweentheAFM tipandthesampleasafunctionoftheseparationdistance z during astandardforce-distancecurve .....................59 Figure3.16:SchematicsoftheenergylevelalignmentforthesampleandAFMtip duringtheKPFMmeasurementprocess ................63 Figure3.17:SchematicdiagramdepictingtheEwaldsphereconstructionforan electronbeamwhichisincidentnormaltothesubstratesurface ...66 Figure3.18:SchematicdiagramoftheLEEDopticsusedfordetectingacted electrons .................................68 Figure3.19:Circuitdiagramsoftheerentwiringrequiredfor vanderPauwtransportmeasurements .................70 Figure3.20:Graphicalrepresentationoftherelationshipbetweenthevoltageratio Q andthegeometricfactor f involvedinvanderPauwtransport measurements ..............................71 Figure3.21:SchematicdiagramdepictingtheHallandmeasuredHallpo- tential .............................74 Figure3.22:Circuitdiagramsofthetwomainwiringtionsrequiredfor Halltransportmeasurments ......................75 Figure3.23:Schematicrepresentationofthetpossibleepitaxialrelations betweenanatomicoverlayerandtheunderlyingsubstrate ......77 xiv Figure3.24:Exampleimageofamoirepatternproducedbytwoperiodicstruc- tureswhichhavebeenrotatedrelativetooneanother ........80 Figure4.1:Energydiagramofdonorandacceptorlevelsthatcanbepresentin theSibandgapformedbetweenthevalenceandconductionbands .86 Figure4.2:MeasuredLEEDpatternandsimulatedpatternonthe Si(111)7 7surface ..........................92 Figure4.3:TypicalSTMtopographyimageofSi(111)7 7 ...........93 Figure4.4:NormalizedSTSspectratakenonthecornerad-Siintheunfaulted regionsoftheSi(111)7 7surface ...................94 Figure4.5:SimulatedstructurediagramofthedeactivatedSi(111)-B p 3 p 3 R 30 surface ...............................95 Figure4.6:STMtopographyimageofapristineregiononthedeactivatedSi(111)- B p 3 p 3 R 30 surface ........................96 Figure4.7:MeasuredLEEDpatternandsimulatedpatternonthe deactivatedSisurface ..........................97 Figure4.8:STMimageofatomicdefectsthatarecommonlyobservedonthe Si(111)-Bterraces ............................98 Figure4.9:NormalizedSTSspectratakenonad-SisitesofthedeactivatedSi(111)- B p 3 p 3 R 30 surface ........................99 Figure4.10:MorphologyofthehydrogenterminatedSi(111)surfaces .......100 Figure4.11:MeasuredLEEDpatternandsimulatedpatternonthe smoothSi(111)-Hsurface ........................101 Figure4.12:MorphologyofthehydrogenterminatedSi(001)surfaces .......102 Figure4.13:Schematicdiagramoftheseparationbyimplantationofoxygenmethod forpreparingSOIwafers ........................104 Figure4.14:SchematicdiagramoftheSmart-CutmethodforpreparingSOIwafers 105 Figure4.15:Exampleschematicdiagramofthesurfacetransferdopingmechanism inducedbyclosely-lyingmolecularorbitalstotheSinanomembrane bandedges ................................107 xv Figure4.16:Schematicdiagramsofthetpossiblemethodsforintroducing aninterfacialdipolelayeronthesurfaceofSinanomembranes ...108 Figure4.17:ImagesofafabricatedSinanomembranedevice ............111 Figure4.18:Simulatedimageofafree-standing M Pcmoleculeshowingthecom- pleteatomicstructure ..........................112 Figure5.1:ZnPcpackingandenhancedmoiresignaturefromSTM 118 Figure5.2:DensityofstatesofZnPconthedeactivatedSi(111)-Bsurface ...119 Figure5.3:Simultaneoustopographyandtialconductancemappingof ZnPconSi(111)-B ............................120 Figure5.4:TheapparentheightdeterminedfromSTMofZnPconthedeacti- vatedSi(111)-Bsurface .........................121 Figure5.5:ZnPcepitaxialrelationshipandsimulatedmoirepattern .......122 Figure5.6:AnisotropicwgrowthofZnPconSi(111)-B ..........124 Figure5.7:AtomicdefectlinesonthedeactivatedSi(111)-Bsurface .......127 Figure5.8:ofSidefectlinesonthegrowthofZnPc ...........128 Figure5.9:ofSistepedgedefectsonthegrowthofZnPc .......129 Figure5.10:Schematicdiagramsofthetactivationbarriersforadmolecules duringthegrowthofZnPconthedeactivatedSisurface .......131 Figure5.11:CoalescingPcmoleculardomainsanddomainboundaryformations 132 Figure5.12:NucleationofZnPcdomainsatdomainboundariespriortomonolayer completion ................................133 Figure6.1:TypicalLEEDpatternofsub-monolayerZnPconthedeactivatedSi surface ..................................138 Figure6.2:LEEDpatternofZnPcexhibitingtheidealanisotropicwgrowth 141 Figure6.3:LEEDpatterncomparisonbetweensub-monolayerandmultilayer ZnPconSi(111)-B ............................142 xvi Figure6.4:Schematicreal-spaceunitcellrepresentationsofthelayerand multilayeredZnPc ............................143 Figure6.5:LEEDpatternsofmultilayeredZnPcanddecreasingsatellitepeak intensity .................................144 Figure6.6:LEEDpatterncomparisonwithFFTsobtainedfromSTMimagesof ZnPc ...................................146 Figure6.7:STMimagesofmultilayeredZnPconthedeactivatedSisurface ...148 Figure6.8:Side-viewschematicoftheZnPcpackingcoonthedeac- tivatedSisurface ............................149 Figure6.9:SummaryschematicofZnPcgrowthonthedeactivatedSi(111)-B surface ..................................150 Figure7.1:STMimagesofZnPcgrowthmodechangeswithincreasingtemperature 154 Figure7.2:Plotoftheandnucleationcontributionstotheoverallnu- cleationprobabilityoftwo-dimensionalclustersonthesurfacewith respecttothesubstratetemperature ..................157 Figure7.3:Pcnucleationatthesurfacestepedgesorinthemiddleoftheterrace 158 Figure7.4:Graphicalrepresentationoftherelationshipbetweenthe barrierandvdWbindingenergyformolecularnucleationformation 159 Figure7.5:STMimagesofCuPcgrowthmodechangeswithincreasingtemperature 160 Figure7.6:EvidenceofZnPcsub-monolayersmoothingbysubstratetem- peraturecontrol .............................163 Figure7.7:EvidenceofZnPcmultilayeredsmoothingbysubstratetemper- aturecontrol ...............................164 Figure8.1:Largescalecomparisonofmoleculardomainformationconsistingof ttypesof M Pcmolecules ....................169 Figure8.2:Comparisonofmolecularpackinganddefectswithindomainscon- sistingofttypesof M Pcmolecules ..............170 xvii Figure8.3:Shiftsin M Pcmolecularpackingandapparentheightofmoleculesin aortiltedorientation .....................171 Figure8.4:Individual M Pcmoleculesintheorientation ........173 Figure8.5:Schematicsofthe M Pcorbitalhybridizationmechanism .......176 Figure8.6:Chargedensitydmapsof M Pcmoleculesadsorbedonthe deactivatedSi(111)-Bsurface ......................177 Figure8.7: M PcvdWinteractionwiththedeactivatedSisurfaceandrotational potentialenergylandscapecomparison ................179 Figure8.8:inthepotentialenergylandscapesfort M Pcmolecules adsorbedonthedeactivatedSisurface .................180 Figure8.9:Comparisonbetween M Pcgrowthonametallicsurfaceandthede- activatedSisurface ...........................183 Figure9.1:MorphologyofCuPcdepositedonthesmoothSi(111)-Hsurface ..188 Figure9.2:MorphologyofCuPcdepositedontheroughSi(111)-Hsurface ...189 Figure9.3:MorphologyofCuPcdepositedontheroughSi(001)-Hsurface ...192 Figure9.4:SimultaneoustopographyandsurfacepotentialmappingoftheSi(001)- Hnanomembranesurface ........................194 Figure9.5:Simultaneoustopographyandsurfacepotentialmappingof1.5nm CuPcontheSi(001)-Hnanomembranesurface ............195 Figure9.6:Simultaneoustopographyandsurfacepotentialmappingof3.0nm CuPcontheSi(001)-Hnanomembranesurface ............196 Figure9.7:Simultaneoustopographyandsurfacepotentialmappingof6.0nm CuPcontheSi(001)-Hnanomembranesurface ............197 Figure9.8:SummaryplotofthechangeintheworkfunctionofSinanomem- branesforallmoleculesandthicknessesmeasuredbyKPFM ..198 Figure9.9:Exampleschematicrepresentationofworkfunctionshiftinducedby dipolelayerattheinterfacebetweenametallicsurfaceandaof organicsmallmolecules .........................199 xviii FigureA.1:Three-dimensionaldesignoftheapplicationprobeinsertendusedfor transportmeasurements .........................215 FigureA.2:TopoftheapplicationprobeinsertwithSMAcablefeedthroughcon- nectorsandaKFconnectorusedtosealtheprobeonthePPMS 216 FigureA.3:Imageofthecablingwhichspiralsdownfromtheonthetop oftheprobeforelectricalconnectionstothesample .........216 FigureA.4:Aside-viewimageofthebottomoftheapplicationprobeinsertwhere thesampleconnectionismade .....................217 FigureA.5:Otherside-viewimageofthebottomoftheapplicationprobeinsert wherethesampleconnectionismade .................217 FigureA.6:Imageoftheadjustmentschemeforchangingtheorientationofthe samplemountingplatform .......................218 xix Chapter1 Introduction 1.1Motivation Throughpreviousdecades,theelectronicsindustryhascontinuallypushedformethodsthat cancreatefunctionalelectronicswitheverdecreasingsize,butalsoincreasingfunctionality andcomplexity.Thisledtotheadventofstrategies,suchasphoto-lithography,stamping, andfocusedbeamlithography,todevicearchitectures[ 1 , 2 , 3 ].Overtheyears,these strategiesprovedsuccessfulandassuch,havebeentothepointthattheyarecapable ofproducingdevicesdowntothenanoscale.Thesestrategies,calledtop-downmethods, relyonimposingpatternsorstructuresonthesubstratesurfaceinordertofabricatea device.Top-downmethodshaveenabledthemassproductionofnanoscaledevicesandthese methodscontinuetoserveasthebasisfornanodevicedevelopment.However,theinherent fundamentallimitationsofcommontop-downtechniqueshavepreventedthesemethodsfrom meetingtheeverdecreasingsizedemandsoftheelectronicsmarket[ 4 , 5 , 6 ].Tocircumvent theselimitations,strategieswhichaprecisecontrolofmatterdowntotheatomiclevel arerequired. Analternativetothetop-downmethodologyistoinvokeastrategywhichcentersaround formingthedesirednanostructuresoutofthesmallestbuildingblocksavailable,namely atomsandmolecules.Inthisbottom-upapproach,atomsand/ormoleculesaredirectly appliedtosurfacesorinterfacesofinterest,andtheadsorbedspeciesareabletoself- 1 organizeintothedesiredstructure,precisecontrolattheatomiclevel.Inthis scenario,thenanostructureorproducedisgovernedbyacompetitionbetweenthermo- dynamicandkineticprocessesduringthegrowth.Bycontrollablytippingthisbalance,the atomic/moleculargrowthcanbemanipulatedsuchthatitispossibletoaccessavarietyof growthmodesforformingofvaryingmolecularorderingandorientation,aswellas amultitudeoftnanostructures[ 5 , 7 , 8 , 9 , 10 ].Mergingthebottom-upapproach withexistingtop-downstrategiespavesthewayforformingincreasinglycomplexandnovel nanoscaledevices. Inordertoproducecomplexthinandnanostructures,itisnecessarytoexplore alternativematerialsthatcanahigherdegreeofcomplexityandfunctionalitythan conventionalsemiconductingmaterials,whilestillbeingintegrableintocurrentdeviceplat- forms.Organicsemiconductingsmallmoleculeshavereceivedalargeamountofattention duetotheirhighdegreeofility[ 11 , 12 ],theirabilitytoorganizeintofunctionalized nanostructures[ 5 , 9 , 13 ],andthroughmolecularsynthesis,tunetheirelectronicandmagnetic propertiesonthemolecularlevel[ 14 , 15 , 16 ].Theuseoforganicmolecularthinisal- readybeginningtoemergeintoday'selectronicsmarketintheformoforganiclight-emitting diodes(OLEDs),organictransistors(OFETs),andorganicphotovoltaics(OPVs) [ 17 , 18 , 19 , 20 ].However,despitesomeearlysuccessesofthesetherestillremainsa tlackofunderstandingregardinghowthesemoleculesinteractwithmetallic,semi- conducting,andinsulatingsurfacesaswellasthenatureoftheintermolecularinteractions [ 7 , 21 , 22 , 23 ].Theseinteractionscanthemolecularorientationandorderingatthe hetero-interfacewhichcanthetransportpropertiessuchaschargeinjection,charge transfer,excitonetc.[ 8 , 10 , 20 ],drasticallychangingtheperformanceoforganic electronicdevices[ 24 , 25 , 26 ]. 2 Ofthemanyorganicsemiconductingsmallmolecules,metalphthalocyanines( M Pc)are amongthemostfrequentlystudiedmoleculesinorganicmolecularthingrowth.Avari- etyofstudiesinvolving M Pcshavealreadybeenperformedoninsulating[ 27 , 28 , 29 , 30 , 31 , 32 , 33 ],semiconducting[ 34 , 35 , 36 , 37 , 38 , 39 , 40 , 41 , 42 , 43 ],andmorecommonly,metallic surfaces[ 23 , 44 , 45 , 46 , 47 , 48 , 49 , 50 , 51 , 52 , 53 , 54 , 55 , 56 , 57 , 58 , 59 , 60 , 61 ].Applications of M Pcthinareevenbeingexploredontopologicalinsulatorsaswell[ 62 , 63 ].These moleculesarewell-knownsmallmetalorganiccomplexeswhichexhibitextended ˇ electron systems,owingtotheirsimplisticsquare-planargeometry,anddisplaybulksemiconducting properties.Atportionofthisdissertationfocusesonthegrowthdynamics,molec- ularordering,andmolecularorientationof M PcmoleculesonSisurfaces.Siisasubstrate whichisnotoriouslytoformanorganicmolecularthinonduetothesurface danglingbonds[ 36 ].Throughsurfacedeactivationandterminationmethods,the ofthesebondscanbenulmakingtheformationofhighlyorderedorganicmolecular thinmsonSipossible[ 37 , 38 , 39 , 40 , 41 , 42 , 43 ].Thecombinationofscanningtunneling microscopy(STM)andscanningtunnelingspectroscopy(STS)providesanidealtoolforin- vestigatingthephysicalandelectricalpropertiesatthe M Pc-Siinterfaceonasub-molecular scale.Asthe M PcnanostructuresextendacrosstheSisurfacetoformacompletemolecular thinlow-energyelectron(LEED)canbeusedtodetectthein-planestruc- tureoftheorganizedorganicmolecularforcomparisonwithSTMobservationsandcan evencontinuetoprovideinformationattlylargemolecularthicknesseswhere STMcannolongerbeused.Densityfunctionaltheory(DFT)calculationscanbeusedto aidthedescriptionoftheobserved M Pcgrowthphenomenabysheddinglightonthefun- damentalmechanismunderlyingthenatureofthe M Pcinteractionwiththesubstrate,and howthisinteractionwillthegrowthdynamics.Throughthecombinationofallof 3 thesetechniques,thegrowthevolutionof M Pcmolecules( M =Zn,Cu,Co)fromthesingle moleculeleveluptomultilayeredmoleculargrownonthedeactivatedSi(111)-Bsurface canbestudied. Overall,theseresultsprovideacomprehensiveviewofthegrowthevolutionof M PconSi andstrategiesforsteeringthemoleculargrowthtowardsforminglong-rangehighly-ordered organicmolecularthinonatechnologicallyrelevantsurfacelikeSi.However,itis paramounttounderstandhoworganicmolecularthincanbeintegratedintoexisting silicon-baseddevices,possiblyallowingforatimpactonthecurrentelectronics market.Currently,thesemiconductorindustryiscapableoffabricatingdeviceswithinthe nanometerregime.Atthislengthscale,thesurfacesandinterfacesbegintoplayacriticalrole indeterminingpropertiesandfunctionsofnanomaterials.Thispreventstheindustryfrom simplyscalingdowntypicaldevicearchitectures.Forexample,whenconventionalmetal- oxide-semiconductortransistorsarescaleddowntothenanometerregime,this givesrisetoshort-channelpreventingnormaltransistoroperation.However,through mointhedevicearchitecturetoincorporatealow-dimensionalnanoscalesemi- conductoronaninsulatinglayer,i.e.ultra-thinsilicon-on-insulator(SOI),theshort-channel couldberemoved[ 64 ].ThesuccessofSOIinthemicroelectronicsindustryhallmarks thepromisingapplicationsoflow-dimensionalnanoscalesemiconductors.Forthisreason, thereisgrowinginterestintheuseoftwo-dimensionalnanomembranes,one-dimensional nanowiresandnanotubes,andzero-dimensionalquantumdots,inordertotakeadvantage oftheiruniquetransportandquantumtproperties[ 65 , 66 , 67 ].Forallofthese materials,theinteractionsatthesurfacesandinterfacescandominateanyinherentbulk properties,duetothelargesurface-andinterface-area-to-volumeratio.Forexample,the existenceofsurfacestates,danglingbonds,defects,andchargedimpuritiesattheinterface 4 mayleadtoFermilevelpinning[ 68 ],dopingofnanostructures[ 69 , 70 ],trappingofmobile carriers[ 71 , 72 ],chargere-distribution,orchargescattering[ 73 , 74 ].However,itispossible toutilizethesetotailorthepropertiesofsemiconductingnanostructures[ 75 , 76 , 77 ]. OnesuchnanomaterialistheSinanomembrane,alow-dimensionalmaterialthatdisplays notabley,stretchability,stackability,andcompliancetotheunderlyingsubstrate [ 78 , 79 , 80 , 81 , 82 , 83 , 84 , 85 , 86 ].Additionally,Sinanomembraneshavebeenpreviously showntohaveelectronicpropertiesthatarehighlysensitivetotheirinterfacecondition [ 87 , 88 , 89 , 90 ].Bycouplingthingrowthof M PconSisurfaceswithSinanomembrane technology,itispossibletoexplorehowmooftheorganic-inorganichetero-interface ofasilicon-basednanodevicecanitstransportproperties.Throughtheuseof Kelvinprobeforcemicroscopy(KPFM)anddevicetransportmeasurementsofSinanomem- branesinterfacedwith M Pcmolecularmocationofthesetransportpropertiesis investigated.Takentogether,itiscriticaltoutilizetheiofsurfacesandinterfaces onsemiconductingnanostructuresfortheirapplicationsinnanoelectronics,molecularand biologicalsensors,andenergyharvestingdevices. 1.2ContentsandLayoutofThisWork Thinmformationishighlysensitivetoatomicandmoleculargrowthprocesses.Thus,in ordertoformanorderedthinwithmolecularprecisionitisnecessarytomaintaina highdegreeofcontrolovertheenvironmentalconditionsinwhichthematerialisdeposited. In Chapter2 ,thetheoryandnecessaryconsiderationsforthingrowthofinorganic material,andmoreimportantlyforthisstudy,smallorganicmoleculesaredetailed.Previous observationsofrecentstudiesregardingorganicmolecularthingrowthonmetallic, 5 insulating,andsemiconductingsurfacesareincludedtoassessthecurrentstandinginthe existingresearch. Priortoexploringthingrowthoforganicsmallmolecules,itisimperativetohavea solidunderstandingofthettechniquesthatcanbeappliedtostudysuchsystemswith thenecessaryprecision.In Chapter3 ,theexperimentalandcomputationmethodologiesthat areutilizedthroughoutthestudiescontainedhereinaredescribedsothatonecaninterpret theresultspresented.Thisincludesabriefdescriptionoftheprimaryexperimentalset- upsthathavebeenusedaswellastheircapabilities,informationregardingscanningprobe microscopytechniques,low-energyelectron(LEED),andthenecessarytransport measurementsforSinanomembranemeasurements.Theoreticalcalculationsalsoplayeda keyroleinsomeofthedataanalysis,andthesemethodsarealsodetailedwithinthischapter. Withtheaddedmolecularthinanalysisandinterpretationofthesurfacesandin- terfacescanquicklybecomeconvoluted.Thus,beforemoleculargrowth,thevarioussurface andinterfacesthatwillbeutilizedmustbethoroughlyinvestigatedtoprovideaworking knowledgeofthesystemsbeingused. Chapter4 providesbackgroundinformationregarding thesurfacemorphologyandelectronicstructureofthetSisurfacereconstructions utilizedthroughoutthisstudy,aswellasinformationregardingmethodsforsurfacepassi- vationanddeactivation.DetailspertainingtothepreparationofthesetSisubstrate surfacesareincluded.Sinanomembranematerialsareintroduced,includingfabricationand theirapplicationsincurrenttechnology.Abriefhistoryandpertinentbackgroundinforma- tionregardingthe M Pcmoleculeisdiscussed,aswellaswhythisprototypicalmoleculehas beenextensivelyutilizedinotherresearchstudies.Detailsregardinghowthesemolecules arepreparedpriortoandduringthemoleculardepositionprocessareincluded. Chapter5 through Chapter8 detailtheresultsofthegrowthevolutionof M Pcmolecules 6 ( M =Zn,Cu,Co)startingfromthesinglemoleculeleveluptomultilayeredgrown onthedeactivatedSi(111)-Bsurface[ 39 , 40 , 41 , 42 , 43 ]. Chapter5 specdetailsthe molecularpackingandthegrowthevolutionofZnPcatsub-monolayercover- ageswiththesubstrateheldatroomtemperature.Theobservationoftheanisotropic wgrowthmodeforan M Pcmoleculeonasemiconductingsubstrateisshownhere aswellaswhatfactorscandisruptthisgrowthmodeobservation. Chapter6 detailsmulti- layeredgrowthevolutionofZnPcusingcombinedSTMandLEEDobservations. Chapter7 illustrateshowthesubstratetemperatureduringmoleculardepositioncanbeused toalterthemoleculargrowthinordertooptimumgrowthconditions.In Chapter8 , strategiesformodifyingthemolecule-substrateinteractionandconsequentlythemolecular packingareexploredbysamplingthemoleculargrowthof M Pcmoleculeswitht coordinatedtransition-metal(TM)ions.ThroughDFTcalculations,thefundamentalmech- anismunderlyingthenatureofthemolecule-substrateinteractionandhowthisin thegrowthdynamicscouldbeunderstood. Chapter9 discussespreliminaryresultsofutilizing M Pcmolecularstotunethe electricalpropertiesofSinanomembranes.Thegrowthof M Pcmoleculesonthy- drogenpassivatedSisurfacesisdiscussed.TheoptimumhydrogenpassivatedSisurface forproducingasmoothmolecularisselectedandusedforSinanomembranemeasure- ments.ChangesintheelectricalpropertiesoftheSinanomembraneduetothepresenceof themolecularthinattheinterfacearetrackedthroughtheuseofKPFManddevice transportmeasurements. Allresultsareconcludedin Chapter10 .Inthischapterfutureworkandoptionalresearch routesarealsodiscussed.Lastly, Appendix A includesdetailsrelatedtothetransportappli- cationinsertprobethathadtobedesignedandbuiltforSinanomembranemeasurements. 7 Additionally, Appendix B referencesotherworksthatIplayedarolein,butdonotneces- sarilypertaintothestudiesincludedinthisdissertation. 8 Chapter2 OrganicMolecularThinFilmGrowth Thischapterdetailsthemechanismswhichdictatethingrowthofsmallorganicmolecules. However,beforediscussingorganicmoleculargrowth,itisnecessarytoprovideagen- eralizedbackgroundofthingrowthsuchthatconceptscommonlyappliedtoinorganic growthcanaidintheobservationsoforganicmoleculargrowth.Both Section2.1 and Section2.2 introducethekeymechanismsinvolvedingeneralizedatomisticandmolecular growthofagivenmaterial. Section2.1 focusesonexpectedthinmorphologywhen grownatequilibriumconditions,while Section2.2 discusseskineticprocessesthatcanoc- curduringthindeposition.Ultimately,thesemechanismsdeterminehowthegrowth isexpectedtoevolvethroughoutthedepositionprocess. Section2.3 discussesthenucle- ationprocessofdepositedmaterialfromatheoreticalframework.Thiscanbeutilizedto understandthechangesintheobservedgrowthwhichappliestobothinorganicandorganic thingrowthprocesses. Section2.4 providesinformationregardingtypicallyobserved kineticinorganicgrowthmodesandhowtheycanbemobythesubstratetemperature duringthedepositionprocess. Section2.5 discussestheerencesbetweentypicalatomistic inorganicgrowthandorganicmoleculargrowthaswellaswhatadditionalfactorscanplay atroleinthegrowthevolutionprocess.Finally,previousobservationsoforganic moleculargrowthontsubstratesisdiscussedin Section2.6 toassessthecurrent standingoforganicmoleculargrowthintheexistingresearch. 9 2.1InorganicFilmGrowthNearEquilibrium Growthofthincomposedofatomsdepositedfromvaporphaseonacrystallinesurface isgovernedbythecompetitionbetweenkineticandthermodynamicgrowthconsiderations. Thus,inordertoobtainathinwiththedesiredsurfacemorphology,anunderstandingof thecompetingfactorsbetweenthethermodynamicandkineticgrowthregimesisrequired.In thethermodynamiclimit,thegrowthoccursinastateofequilibrium,wherethemorphology isdeterminedbytheminimizationofthesurfacefreeenergyofthecombined system.However,growthistypicallynotperformedexactlyatthethermodynamic equilibriumcondition.Asaresult,kineticprocesses,suchasadatomusion,nucleation, andadatomdesorption,modifytheexpectedequilibriumgrowthstructureand,ultimately playadeterminingroleinthemorphology.[ 91 , 92 , 93 , 94 , 95 , 96 , 97 , 98 ].However, beforeexploringthevariouskineticprocessesindetailaswellashowtheymodifythe morphology,itisnecessarytodiscusstheexpectedequilibriumgrowththat canbeobserved. Attheequilibriumconditionthefreeenergyisminimizedbyconsideringthecombined energiesassociatedwiththesystem.Therelevantenergiestoconsiderin theminimizationprocessincludethesurfaceenergybetweenthesubstrateandvacuum, substrate ,thesurfaceenergybetweenthebeingformedandvacuum, film ,andthe energyattheinterfacebetweenthetwomaterials, interface .Fromtheseenergiesaswell asthecontactanglebetweenthesubstrateandthe ˚ ,itispossibletoformulatean expressionwhichdictatesthegrowthatequilibrium: substrate = film cos( ˚ )+ interface (2.1) 10 Figure2.1: SchematicofVolmer-Webergrowth .Thistypeofequilibriumgrowthis commonlyreferredtoasthree-dimensionalislandgrowth,evidentfromtheschematicside- viewwithincreasingcoverage.Islandformationsoccuronthesubstratesurfacewithout formingacompletemonolayer. Thisresultsinthreetenergyrelationshipsgivenbythefollowingexpressions: substrate < film + interface (2.2) substrate > film + interface (2.3) substrate = film + interface (2.4) From Equation2.2 ,thelow-energysurfaceisthesubstrate.Thus,inordertominimizethe freeenergyofthesystem,thedepositedwillformthree-dimensionalislands,leavingareas 11 Figure2.2: SchematicofFrank-vanderMerwegrowth .Thistypeofequilibrium growthiscommonlyreferredtoaslayer-by-layergrowth,evidentfromtheschematicside- viewwithincreasingcoverage.Subsequentlayersinthedonotbegintoformuntil thepreviouslayeriscomplete. ofthesubstrateexposed,i.e.Volmer-Webergrowth(see Figure2.1 ).Ontheotherhand,if theconditionisthatof Equation2.3 ,thecombinedsurfaceenergyoftheandinterface providethelow-energysurfaceforthesystem.Inthisscenario,thewillprefertowetthe substrate,followingalayer-by-layergrowthmorphology,i.e.Frank-vanderMerwegrowth (see Figure2.2 ).Ofcourse,itispossibletoperformgrowthstudiesinvolvingsubstratesand consistingofthesamematerial.Thiscorrespondstotherelationin Equation2.4 ,which alsoproducesFrank-vanderMerwegrowth.However,themajorityofthinstudiesmake useoflmsconsistingofamaterialthatfromthesubstrate.Asaresult,therecanbe somelatticemismatchbetweenthetwomaterials.Thelatticemismatchwillcausestrainto buildupintheasthecoverageincreases,inducingachangein film .Inthisscenario,the 12 Figure2.3: SchematicofStranski-Krastanovgrowth .Thistypeofequilibriumgrowth iscommonlyreferredtoaslayer-plus-islandgrowth,evidentintheschematicside-viewwith increasingcoverage.Anumberofwettinglayersareformedonthesubstrate surfacefollowedbythree-dimensionalislandgrowth. freeenergyminimizationinitiallyfollowstherelationin Equation2.3 ,producingalayer-by- layergrowthmorphology.However,afterenoughstrainhasbuiltupinthethegrowth ismoreaccuratelydescribedby Equation2.2 .Thismeansthatonceanumberoflayers havebeenformed,thegrowthtransitionstoformingthree-dimensionalislandsontopofthe alreadyexistinglayeredstructure.Thistypeofgrowthiscommonlyreferredtoaslayer- plus-island,i.e.Stranski-Krastanovgrowth(see Figure2.3 ).TheFrank-vanderMerwe, Volmer-Weber,andStranski-Krastanovgrowthmorphologiesrepresentthethreedit typesofequilibriumstructuresallowedinthethermodynamiclimit[ 93 , 94 , 95 , 96 , 97 , 98 ]. 13 2.2KineticProcessesinInorganicFilmGrowth Inorderforcrystallizationofamaterialtooccur,theremustbeadrivingforceforthe materialtotransitionfromeitherthevapororsolutionphasetoacrystallinestructure.This isdoneinaccordancewiththelawsofthermodynamics,suchthat,byhavingthematerial formacrystallinestructure,itresultsinadecreaseinthefreeenergyofthewholesystem, promotingthecrystallizationprocess[ 91 , 92 , 93 ].Thedecreaseinthefreeenergyinthe systemisaconsequenceofthechangeinthechemicalpotential, .Inthecaseofmaterial crystallizationfromavapor,whenthevaporpressureofthematerial, P ,ishigherthanthe vaporpressureatequilibrium, P e ,crystallizationwilloccuronthesubstratesurface[ 91 ]. Fromthis, canbeexpressedasthefollowing: = k B T log P P e (2.5) where k B isBoltzmann'sconstantand T isthetemperature.Thiscanbefurther byrelatingthechangeinthevaporpressuretothesupersaturationlevel, ˙ ˙ = P P e P e (2.6) resultinginthefollowingexpression: = k B T log(1+ ˙ )(2.7) Duringtheactualcrystalformationprocessofatomswhichareintroducedtothesub- stratesurface,thecrystalgrowthisratelimitedbythekineticprocessesthatoccuratthe interface[ 91 , 92 , 93 , 97 , 98 ].Initially,atomsfromthevaporadsorbonthesubstratesurface. 14 Theadatomsresideonthesurfaceforagivenamountoftime, t ds .Theresidencetime islargelydeterminedbytheenergybarrierassociatedwithdesorptionoftheadatomfrom thesurface, E desorb ,andthefrequencyatwhichtheadatomattemptstodesorbfromthe surface, .Typically, isconsideredtobethevibrationfrequencyoftheadatom.From this, t ds canbewrittenas t ds = 1 exp E desorb k B T (2.8) Duringthistime,theadatomisabletoonthesurface,providedithasenoughenergy toovercometheactivationbarrierforuntilitisdesorbedfromthesurface.Ifthe adatomsareabletooveratlength,theycouldencounterterracevacancy sites,stepedgekinksites,oreventwo-dimensionalclusterscomposedofotheradatomswhich havenucleatedonthesurface,asdepictedin Figure2.4 [ 91 , 92 , 93 , 94 , 95 , 96 , 97 , 98 ].These allprovidelocationsinwhichtheadsorbatecouldnucleate,i.e.crystallizeonthesubstrate. Iftheadatomonthesurfaceanddoesnotencounteranappropriatenucleationsite, itwilldesorbfromthesurface. Inorderforadsorbatestoonthesubstratesurface,theymustbeabletoovercome theactivationenergyrequiredfortheadatomtomovetoanadjacentlatticepointonthe surface, E ds .Thecot( D ds )isasthefollowing: D ds = a 2 exp E ds k B T (2.9) where a isthesubstratelatticeconstant.Bycombining t ds and D ds ,theaverage lengthoftheadsorbate ds canbecalculatedusingtherelationship 15 Figure2.4: Schematicofadatomadsorptionandprocessesonasurface . Givenalargeenoughlength,adsorbedadatomscanto(i)surfacevacancy sites,(ii)stepedgekinksites,(iii)otheradatomsforclusterformation,or(iv)desorbedfrom thesurface. ds = p D ds t ds (2.10) resultinginthefollowingequation: ds = a exp E desorb E ds 2 k B T (2.11) Thus,adatomsthatareatadistanceof ds frompossiblenucleationsiteswillcontribute totheatomicgrowthprocess,whereasadsorbatespositionedfartherthan ds frompossible nucleationsiteswillnotcontributetothecrystalgrowthandinsteaddesorbfromthesurface, asshownin Figure2.4 . 16 Adatomisoneofthemostcriticalkineticmechanismswhichultimatelydeter- minesthegrowthprocess[ 91 , 92 , 93 , 94 , 95 , 96 , 97 , 98 ].Thediscussionabovefocuses onageneralizedactivationenergy, E ds ,foranadatomtoononeofthe substrateterraces.However,inadditiontosimpleadatomonaterracethereare othermechanismsatplayduringthegrowthprocess.Theadatomsarecapable ofup,down,oralongthesubstratestepedgesaswellasaroundtwo-dimensional islandsthatnucleateonthesubstratesurface.Additionally,adatomscanattachordetach fromsitesonthesubstrateorontwo-dimensionalislandsthathavenucleatedonthesur- face.Eachoftheseprocesseswillhavetheirownactivationenergyassociatedwiththem, and,dependingonthebarrierheight,thiscanresultindramaticchangesintheobserved growth[ 91 , 92 , 93 , 94 , 95 , 96 , 97 , 98 ].Forexample,thebarrierthatisassociatedwiththe downwardmasstransportatthestepedgesduetothelossofcoordination[ 99 , 100 , 101 ], i.e.theEhrlich-Schoebelbarrier(ESB)(see Figure2.5 ),canhaveadramaticon thegrowth.IftheESBisstrongenough,itcanrestrictadatomstoagiventerraceor previouslyformedtwo-dimensionalislands[ 102 , 103 , 104 , 105 , 106 ].Thisgivesrisetothe formationofdeeptrenchesthatcannotbealsoknownastheZenoingrowth (see Figure2.6 )[ 105 , 106 , 107 ].Themagnitudeofthesebarriersdependsontheinteractions betweenthesubstratematerialandthedepositedatoms.Onewaytohelpovercomethese activationbarriersistomodifythesubstratetemperatureduringthegrowthprocesswhich providestheatomswithenoughthermalenergytopromoteThisiseludedtoin Section2.3 anddiscussedfurtherin Section2.4 .However,ifanadatomisnotgivenenough timetoonthesurfacebeforeinteractingwithotheradatoms,thebarriers arenolongertheonlypointofconcerninthegrowthprocess. Consequently,theabovediscussionimpliesthedepositionrateofincidentadsorbates 17 Figure2.5: SchematicdiagramoftheEhrlich-Schoebelbarrier(ESB) .Thegraph illustratesthepotentialenergylandscapeforasingleatomonalatticewithastep edgepresent.Regardlessofwhethertheatomisontheupperorlowerterrace, thesurfacepotentialremainsthesame.Iftheatomapproachesthestepedgefrom(a),it willexperienceareductionincoordinationwiththesurfaceresultinginanunstableatom positionandanenergeticbarrierfordoingso.Iftheatomapproachesthestepedgefrom (b),itwillexperienceanincreaseincoordinationwiththesurface,resultinginamorestable andacorrespondingenergeticminimum.Thesechangesinatomcoordination withthesurfacegiverisetotheESBfordownorupagivenstepedge. isacriticalfactorinthegrowthprocess.Intheidealscenario,thecrystallizationprocess iscarriedoutattheequilibriumcondition.Inthiscase,evenatvaporsupersaturation conditions,thenumberofatomsthatadsorbonthesurfaceisbalancedwiththenumberof atomsleavingthesurface.Thus,aparticulardensityofadatomsisabletobemaintained onthesurface,andtheyareabletoaround,asshownin Figure2.4 .Withalow enoughdensityofadsorbatesonthesurface,adatomsareabletothefulldistance ds totheappropriatenucleationsite[ 91 , 92 , 93 , 94 , 95 , 96 , 97 , 98 ].However,ifnew materialisdepositedataratefasterthanthealreadyadsorbedadatomsareableto orthedensityofadsorbatesistlyhigh,theywillnotbeabletooverthe 18 Figure2.6: SchematicillustrationoftheZenoinorganicandinorganic growth .Duetothetlylargebarriersfordownwardand/orupwardmasstransport, incidentatomsormoleculesarerestrictedtothedomainstheylandon,resultinginthe formationofdeeptrencheswithinthe fulldistance ds andwillhaveanincreasedprobabilityofcollidingwithotheradsorbates priortotheappropriateenergeticminimumnucleationsites.Ifenoughadatomsare abletocongregatetogether,thiscanresultintheformationofatwo-dimensionalnucleus onthesubstratesurfacewhichcandisruptthecrystallizationthatwouldbepossibleat theequilibriumcondition.Forthisreason,itisnecessarytocharacterizethefrequencyof two-dimensionalnucleusformationandtheimpactthatitcanhaveonthecrystalgrowth [ 91 , 92 , 97 ]. 19 2.3ClassicalNucleationTheory Forclarity,itshouldbenotedthatthetheorydiscussedhereinislargelybasedonho- mogeneousnucleation,whichcanserveasastartingpointforinterpretingmorecomplex nucleationprocesses.Whenformingatwo-dimensionalnucleusonthesurface,adatomsare bothcontinuouslycollidingwitheachotherandalsoattachingtoeachother,causingmany adsorbatestocollectandformanucleus.Smallnucleiformedmightloseatomsduetocol- lisionswithadsorbateswhicharedislodgedfromtheformingnucleus;thiscouldeventually resultinthedisappearanceofthenucleusentirely.However,ifalargernucleusisformed,the probabilitythatthenucleussurvivesincreases.Thenucleusisconsideredacriticalnucleus whentheprobabilityofthatnucleustoincreaseordecreaseinsizeisthesame.Whenthe nucleusisformed,alloftheatomsontheinteriorofthenucleusareconsideredtohave crystallized;thoseattheedgeofthenucleusarenotbecausetheyareexposedtoprocesses thatareexternaltothecluster,suchasadsorbatecollisions.Thechangeinthechemical potentialisequaltothefreeenergydecreaseperatom.However,thereisalossintheenergy duetotheedgeatoms[ 91 , 92 , 97 ].Theenergylossperunitlengthalongtheedgeofthe nucleusisas .Assumingthatthesizeofthenucleuscanbeapproximatedbya two-dimensionalcircleofradius r (see Figure2.7 (a)),thechangeinthefreeenergy G for thewholesystemwhenanucleuswithradius r isformedcanbecalculatedasthefollowing: G ( r )= ˇr 2 n 0 +2 ˇr (2.12) where n 0 istheareadensityofatomsinthenucleusformed;making ˇr 2 n 0 approximately equivalenttothenumberofatomsinthenucleus.Ifacriticalnucleusofradius ˆ isformed, using Equation2.12 themaximumchangeinthefreeenergy G canbedetermined,as 20 Figure2.7: Schematicdisplayingatwo-dimensionalnucleus,approximatedinsize byacircleofradius r .(a)Schematicdisplayingtheradiusofatwo-dimensionalnucleus thathasformed.(b)Graphicalrepresentationofthechangeinthefreeenergywhenacluster ofaspradiusisformed.Ifthenucleushasaradiusof ˆ orlarger,itsfreeenergywill bemaximized,increasingtheprobabilitythatthenucleuswillsurvive. shownin Figure2.7 .Consequently, ˆ canbewrittenas ˆ = n 0 (2.13) Using Equation2.7 , ˆ takesthefollowingform: ˆ = n 0 k B T log(1+ ˙ ) (2.14) Pluggingthisexpressionfor ˆ into Equation2.12 givesthemaximumfreeenergychange 21 valueas G = ˇ 2 n 0 k B T log(1+ ˙ ) (2.15) TheBoltzmannfactorcanbeusedtoexpresstheprobabilityofforminganucleuswhich survivesonthesurfacewithradius r ( N p ).Thisisgivenbythefollowingrelation: N p =exp G ( r ) k B T (2.16) Theprobabilityofformingastablenucleus, J p isproportionalto N p ( Equation2.16 )and theprobabilityforanindividualadsorbatetotoapreexistingnucleus, D p ,i.e.the exponentialportionof Equation2.9 .Thisyieldsthefollowingexpressionfor J p : J p / D p N p = ) exp E ds k B T exp G ( r ) k B T (2.17) However,nucleiof r<ˆ willnotsurvivethegrowthprocess.Thus, J p canbemo toincludecriticalnucleiformationwhichyieldsamaximumchangeinthefreeenergy,i.e. G ( r )becomes G .Thus, J p canbewrittenasthefollowing: J p / exp E ds k B T exp " ˇ log(1+ ˙ ) k B T 2 # (2.18) whereistheedgeenergyofthecriticalnucleus,as 2 = 2 n 0 (2.19) Fromthisequation,itispossibletodetermineifthenucleationprocessisdominatedby 22 relatedprocessesorbythefactorswhichcontributetothenucleationprocess, suchastheedgeenergyofthecriticalnucleusorsupersaturationlevel[ 91 , 92 , 97 ].The inthetemperaturedependencefor D p and N p alsoshowsthatthesubstrate temperatureduringthematerialdepositioncanhaveanotableimpactonthenucleation process.Bycombiningtheconsiderationsforformingnucleionthesubstratesurfacewith theunderstandingofadatomprocesses,itispossibletointerpretgrowthphenomena commonlyobservedforinorganicgrowthevolution. 2.4KineticGrowthModeTransitionsforInorganic Growth Asdiscussedin Section2.3 ,ifthedepositionrateisslowerthantheadsorbate rate,thegrowthofatomicstructureswilloccurclosetothermalequilibriumconditions.In doingso,thisensuresthatadsorbateswillhavethepossibilitytobeabletofullyexplore thesubstratesurfaceandsuchthattheyformtheirenergeticminimumstructure.If theinteractionsandlatticematchingarebalancedcorrectly,thiscouldleadtothepreferred equilibriumgrowthmodebehaviordescribedin Section2.1 .However,theobservedgrowth structureisalsohighlydependentonkineticgrowthprocessesinvolvingadatom whichcanbelimitedbythetemperatureofthesubstrateduringdepositionandthe sionbarriers[ 93 , 94 , 95 , 96 , 97 ].Typicallyinthecaseofinorganicgrowthonasurface, lowsubstratetemperaturesmaketheincidentadatomsunabletolargerdistances, resultinginthemcongregatingtoformsmallclusters.Thisdependsonthecriticalnucleus sizeanditsassociatededgeenergy.Insomeinstances,thetemperaturemaybelowenough thatadatomscannotovercomethebarriersonthesurface,resultinginadatoms 23 Figure2.8: Schematicofttypesofkineticlimitedgrowthmodesof adatomsdepositedonasurface .(a)Randomroughgrowthwhereadatomsformsmall clustersduetolimited.(b)Two-dimensionalislandgrowthofadsorbedadatoms. (c)Stepwgrowthmodewhereadsorbedadatomsareabletotothesurfacestep edgestonucleate. nucleatingwherevertheyaredepositedonthesurface.Thiswillyieldtheappearanceof randomroughmaterialgrowthasseenin Figure2.8 (a)whichwillcontinuetobepresentas thethicknessincreases[ 93 , 94 , 95 , 96 , 97 ]. Asthetemperatureofthesubstrateincreases,adatomsareabletodoverlarger distancesandmorereadilyovercomeanybarriers,allowingtheadatomstosample thetavailablesitesfornucleationwithinthelengthatthistemperature. Theseadatomscannucleateatsurfacevacancysites,stepedgesites,ortheywillbegin 24 tocongregatetoformstablenuclei,asshownin Figure2.8 (b).Nearthelowerendof thistemperatureregime,manysmall,stabletwo-dimensionalnucleicanbeformedonthe substrate.Increasingthesubstratetemperaturefurthercanresultintheformationoflarge stablenuclei.Whenthishappens,theatomswithinnearbysmallnuclei( r ˆ )will dissociateandtothelarger,stablenucleus[ 93 , 94 , 95 , 96 , 97 ].Thisisancalled Ostwaldripening.Thus,anincreasingsubstratetemperatureduringthedepositionprocess willresultinatransitionfromsmalltwo-dimensionalislandswithahighnucleationdensity tolargetwo-dimensionalislandswithalownucleationdensity. Eventually,whenthetemperatureofthesubstrateisincreasedhighenoughthatthe sionlengthbecomescomparabletotheterracewidthofthesubstratesurface,adatoms/admolecules areabletosampletheavailableadsorptionsitesduringtotheenergeticmin- imumsitesatthestepedgespriortoaccumulationontheterrace.Thisresultsin wgrowth[ 108 , 109 ],asshownschematicallyin Figure2.8 (c). 2.5BetweenOrganicandInorganicThin FilmGrowth Thecontrolledcrystallinegrowthoforganicsemiconductorshasbeenoflong-standinginter- estinanattempttoanalogousgrowthmodestoinorganicdeposition[ 110 , 111 , 112 ],as discussedin Section2.4 .However,thereareseveralkeybetweenorganicmolecu- larandinorganicatomicgrowth.Firstly,foraninorganiccrystal,theatomstypicallyexhibit strongcovalentorionicbondstoformthecrystallattice.Ontheotherhand,fororganiccrys- tals,theinteractionsbetweenmoleculesaretypicallycharacterizedbyweak,vanderWaals interactions[ 113 , 114 ].Additionally,ininorganicepitaxialthingrowth,theepi-layeris 25 latticematchedtothesubstrate,promotingahighdegreeofregistrationbetweenthetwo. Theepitaxialrelationwiththesubstrateisdistinctlytinorganicmoleculargrowth. Organicmoleculesconsistofacollectionofatoms,producingananisotropicobjectwhichis tlylargerthananindividualadatom.Forthisreason,organicmoleculesareunable tolatticematchwiththesubstratesurfaceinthesamewaythatcanoccurforinorganic atomicgrowth,resultingincomplicatedepitaxialrelationshipswhichwillbediscussedin detailin Section3.3.1 and Section5.1 . Thegrowthmodelingofsmallorganicmoleculescanalsobecomequitecomplexduetothe factthattheydisplayanincreaseintheirinternaldegreesoffreedomcomparedtoatoms. Organicmoleculesarenotconsideredrigidobjectsincomparisonwithindividualatoms. Therefore,thesemoleculescanexhibitrotationaldegreesoffreedomaswellasbending orstretchingmodes.Theseadditionaldegreesoffreedomopenupavarietyoft channelsforthemoleculesonthesubstrate,complicatingmoresimplistic processconsiderationsintheinorganicgrowthcase[ 114 , 115 , 116 , 117 , 118 , 119 ].Dueto thesizeandanisotropicnatureoforganicmolecules,theelectronicpropertiesoforganic thinarealsohighlysensitivetotheorientationofthemolecules.Sincethestrengthof theinteractionsinvolvedinthegrowthprocesscanalterthemolecularorientation,thiscan introduceadditionalcomplicationsintheorganicmoleculargrowthprocessifaparticular orientationisrequired[ 10 ]. Despitetheseaddedcomplications,itisstillpossibletoincreasethelikelihoodthatthe moleculeswillhaveenoughenergytoovercometherelatedactivationbarriersby increasingthesubstratetemperatureduringtheorganicmoleculardeposition.Fororganic molecules,therangeinwhichthetemperaturecanbemoistlyreducedin comparisonwithinorganicgrowthduetomoleculedecompositionifthetemperatureisel- 26 evatedhighenough[ 31 , 112 , 120 , 121 , 122 ].Thoughthegrowthprocessandevolutionis morecomplicatedfororganicmoleculesthaninorganicatoms,thecommonalitiesallowfor thecreationoforganicthinutilizingsimilarstrategies. 2.6OrganicMolecularGrowthonVariousSubstrates Attemptstowardsanalogousorganicmolecularthingrowththatsharessome similaraspectstoinorganicatomisticgrowthhavebeenalong-standinggoalinthe oforganicmolecularelectronics.Inordertoensuregeneralizedintegrationoforganicthin intoexistingtechnology,thesewillhavetobeincontactwitheitherametallic, semiconducting,orinsulatingsurface.Tothisend,avarietyoforganicmoleculargrowth studieshavebeenperformedonthesesurfaces.Thefollowingsubsectionswilldetailsomeof themajorresultsfromthesestudies. 2.6.1MetallicSurfaces Metallicsurfaceshavebeenamongthemostextensivelystudiedsurfacesfororganicmolec- ulargrowthbecausemetalliccontactsarecommonlyutilizedindevicearchitectures[ 23 , 44 , 45 , 46 , 47 , 48 , 49 , 50 , 51 , 52 , 53 , 54 , 55 , 56 , 57 , 58 , 59 , 60 , 61 ].Particularlyforanorganic moleculardevice,theconductionattheinterfacebetweenthemoleculesandthemetallic contactscanplayasigntroleintheoverallelectronicpropertiesofthedevice,resulting inconductionpathwaysthatcandominatedevice[ 24 , 25 , 26 ].Thus,understand- inghowthemoleculesgrowandinteractwiththesubstratesurfacehasbeenofparticular interestinordertoproducemoretdevicearchitectures. Fromtheoreticalcalculationsandexperimentalobservations,metallicsurfacestendto 27 interactstronglywiththeorganicmoleculesdepositedonthesurface.Thisisbecause, onnoblemetalsurfaces,themolecule-substrateinteractionismediatedthroughorbitalhy- bridizationwiththemolecularspeciesand/orchargeredistribution[ 45 , 49 , 50 ].Thestrong interactionbetweentheextended ˇ electronsystemwiththesubstratecausestheorganic moleculestoadoptamolecularorientationregisteredtosplatticesitesonthe metalsurface.Therearesomeexceptionswheretheintermolecularinteractionsarestrong enoughtoovercomeeventherelativelylargemolecule-substrateinteraction,producingre- sultssimilartogrowthoninsulatingsurfacesasdiscussedin Section2.6.2 .However,the vastmajorityofmoleculargrowthobservationsonmetallicsurfacesdonotfollowthistrend [ 23 , 44 , 45 , 46 , 47 , 48 , 49 , 50 , 51 , 52 , 53 , 54 , 55 , 56 , 57 , 58 , 59 , 60 , 61 ]. Thesmallsurfacelatticeparametersofnoblemetalsurfacesaidinthemolecular processduetothereducedactivationbarrierfordThisallowsmoleculestofar enoughonthesurfacetoformsmallmoleculardomainswhereallthemoleculesmaintainthe molecularorientation,andtheweakintermolecularinteractionsstabilizetheformed domain.However,therelativelystrongmolecule-substrateinteractionstillhinderstheability forthemoleculestoelyproducelong-rangemolecularordering,andinsteadityieldsa completemonolayerconsistingofmanysmallmoleculardomains[ 60 , 61 , 123 , 124 ]. Itshouldbenotedthatthislyingmolecularmonolayerishighlydistortedfromthetypical bulkmolecularpacking[ 125 ].Themonolayeractsasapassivationlayerfor subsequentlayers,resultinginatreductioninthemolecule-substrateinteraction fortheselayers.Forthisreason,asthemolecularcoveragetransitionsfrommonolayerto multilayeredgrowth,moleculeswillgraduallystarttorelaxbacktothebulkpacking [ 46 , 51 , 52 , 54 , 126 , 127 ].Thisrelaxationprocesscanproduceahighdegree ofdisorderwithinthemolecularinthedirectionnormaltothesubstratesurface.This, 28 coupledwiththefactthatthein-planeorderingconsistsofsmallmoleculardomains,means thatorganicgrowthmorphologyonmetallicsurfaceshasatendencytoproducethin thatarenotideallyordered. 2.6.2InsulatingSurfaces Contrarytotheorganicmoleculargrowthonmetallicsurfaces,growthoninsulatingsurfaces resultsinaconsiderablereductioninthemolecule-substrateinteraction.Thisisindependent ofwhetherthesurfaceisamorphous,suchasSiO 2 ,orcrystalline,suchasNaCl.Thisisdueto theabsenceofavailablesurfaceelectronicdensityofstatesoninsulatingsurfaces,removing anychanceofstrongmolecularbondingtothesurface.However,sincethemolecule-substrate interactionhasbeenreduced,theintermolecularinteractionswillplayamoresigtrole inhowthegrowthproceeds.Oftenthemoleculeswillstandinanuprightorientationabove thesurface,maximizingthe ˇ - ˇ molecule-moleculeinteractionsratherthanthemolecule- substrateinteraction[ 27 , 28 , 29 , 30 , 31 , 32 , 33 ].Thisallowsthemoleculestomaintaina molecularpackingthatissimilartothebulkorganiccrystal[ 125 ]. Ifanamorphousinsulatingsurface,likeSiO 2 ,isused,thenthemoleculardomainswill formwithrandomrotationalorientationsrelativetootherdomainsonthesurface.Asthe coverageincreasestowardsacompletemonolayer,alargenumberofdomainboundaries willformwithtlylargeanglebetweenthepackinginthe domains[ 30 , 31 , 32 , 33 , 128 , 129 , 130 ].Iftheintermolecularinteractionsarestrongenough, organicmoleculardomainscanformonacrystallineinsulatingsurfacewithnoregistrationor restrictionsbasedonthecrystallographicorientationofthesubstrate;stillformingdomain boundarieswithcantlylargeangle[ 29 ].Overall,thistypeofgrowthis notidealfortransportrelateddevicesbecausethelargeangledomainboundariesthatare 29 formedcanscatterortrapchargecarrierswhenbeingtransportedthroughtheorganicthin [ 33 ]. Nevertheless,itispossibletomaintainmultilayeredthingrowthwiththesameover- allmolecularpackingsincelatticematchingtotheunderlyingsubstrateisno longerastrictrequirement.Layer-by-layergrowthbecomespossiblewhenthelateralinter- actionsdominateeitherthemolecule-substrateinteractionorinterlayerinteractionsbetween molecules.Thoughthisistheidealgrowthscenarioforthingrowth,thet numberofdomainboundariesformedduetothein-planemolecularorderingcouldbeprob- lematicforcertainapplications.Additionally,theremuststillbeenoughofadriving forceforthemoleculestowetthesurface.Otherwisethiscanquicklytransitiontothree- dimensionalislandgrowth,resultinginanevenhigherdegreeofdisorderwithinthe [ 30 , 31 , 32 , 33 , 128 , 129 ]. 2.6.3SemiconductingSurfaces Semiconductingsubstratesareconsideredthemostnotoriouslysurfacesforforming anorderedorganicmolecularthinusingthermalevaporationmethods.Thisislargely duetosurfacedanglingbondswhicharepresentatthesemiconductinginterface.Unsat- vacanciesthatarelocalizedonthesemiconductoradatomsarehighlyreactiveand willformstrongcovalentbondswithanyorganicspeciesthatisintroducedonthesurface. Thelargesurfacelatticeparametersassociatedwiththetypicalsurfacereconstructionsof semiconductorscombinedwiththestrongmolecule-substrateinteractiontlyhin- dersanymolecularonthesurface.Moleculeswillbecomelocalizedonthesites thattheyinitiallyadsorbon[ 36 ].Thislocalizationresultsinhighdisorderofthe fewmonolayers,becauseatacoverageapproachingacompletemonolayer,onlyunfavorable 30 adsorptionsiteswillbeavailableforincomingmolecules.Thisforcesthemoleculestoadopt adsorptionwhichcanintroduceroughnessinthenextlayer[ 131 ].However, aftertheserstfewlayersaregrown,thestrongmolecule-substrateinteractionwill bedrasticallyreduced,resultinginmoredominantintermolecularinteractionconsiderations inthegrowthasthecoverageincreases.Similartoobservationsonmetallicsubstrates,a rapidshiftwilloccurfromtheroughdistortedmolecularpackingtothebulk crystalpacking[ 46 , 51 , 52 , 54 , 125 ].Fromtheseobservations,theinterfaceorderingaswell astheout-of-planemolecularorderingisaconcernonsemiconductingsurfaces. Onestrategythatisoftenusedtoalleviatethisproblemistopassivatethesurfacedan- glingbonds[ 34 , 35 , 132 , 133 , 134 ].Aswillbediscussedinmoredetailin Section4.2 ,thisis commonlyachievedbychemicallybondingeitherhydrogenorbyusingahalogen,like rineorchlorine.Anotherpossiblestrategyistoformaself-assembledmonolayerconsisting ofmoleculeswhicharechemicallyboundateverydanglingbondsite.Byperformingeither ofthesemethods,theincomingmoleculeswillnolongerhaveastronginteractionwiththe substrate,allowingthemoleculargrowthtoproceed.However,itshouldbenotedthatthese surfacepassivationmethodscanoftenresultinalargereductionofthemolecule-substrate interaction,producingorganicmoleculargrowthsimilartoinsulatingsurfaceswiththesame orderingconcerns. 2.6.4Summary Takingintoaccountthepreviousstudiesoforganicmoleculargrowthonmetallic,semicon- ducting,andinsulatingsurfaces,itisapparentthatmoreexplorationintorouteswhich anintermediateinteractionregimeisrequired.Inthisregimethemolecule-substrateand intermolecularinteractionsarebalancedinawaythatallowsthemoleculestofreely 31 onthesurfacetoformstablegrowthbutstillmaintainenoughinteraction withthesubstratetoimposeorderingwithinthemolecularThisisthesubjectofa tportionofthisthesisandwillbediscussedindetailinthechaptersthatfollow. 32 Chapter3 ExperimentalandComputational Methodologies Inordertostudyorganicmolecularthingrowth,itisnecessarytohaveasolidunder- standingofthetmethodsthatcanbeutilizedforsurfacescienceinvestigations.This chapterincludesdetailsregardingtheexperimentalapparatusesandmethodsusedtoper- formthenecessarymeasurementsthroughoutthisdissertationaswellasthecomputational methodsutilizedtoaidintheanalysisandinterpretationoftheexperimentallyobserved phenomena. Section3.1 providesdetaileddescriptionsofthemainmeasurementequipment usedforexperimentationandincludesinformationaboutthecapabilitiesofeachset-up.In Section3.2 ,extensivediscussionisincludedrelatedtoeachoftheexperimentaltechniques thatareutilized.Thisincludesthetheoreticalworkingprincipleofeachtechniqueandwhat informationaboutthesamplecanbeextracted.Lastly, Section3.3 discussesthecomputa- tionalmethodsthathavebeenusedthroughoutthesestudiestoanalyzethedataobtained fromsamplemeasurements. 33 3.1ExperimentEquipment 3.1.1Ultra-HighVacuumSystem Theprimaryexperimentalapparatususedthroughoutthestudiesincludedinthisdisserta- tionisanultra-highvacuum(UHV)systemwhichiscommerciallyavailablethroughOmicron NanoTechnologyGmbH(OxfordInstruments).TheUHVsystemcanmaintainoperating Figure3.1: ImageoftheUHVsystemutilizedfororganicmolecularthin growthstudiesinvestigatedusingSTMmethodsandLEED . basepressuresbelow1 : 0 10 10 mbarthroughthecombinationofionpumpingandtita- niumsublimationpumpingtoremoveanyambientconditionsmighthaveonthe sampleduringprocessingandmeasurement[ 93 , 135 ].TheentireUHVsystemrestsontop ofvibrationdampenersinordertovibrationallyisolatetheentiresystemfromtheor. ThisparticularUHVset-upconsistsoftwoseparateUHVchambersandaload-lockportto controllablyintroducesamplesorothermaterialsintotheUHVconditionswithoutbreak- ingvacuum(see Figure3.1 ).Additionally,samplesthatconsistoflayeredmaterialscanbe 34 cleavedundervacuumconditionsintheload-lockportbeforeenteringtheUHVchambers. Thechamberclosesttotheload-lockportisthepreparationchamber.Asthename implies,thischamberisusedtopreparesamplesformeasurementsintheUHVsystem.This chamberhousesamanipulatorarmwithfull360 rotationwhichallowsforsamplestobe safelymovedthroughoutthepreparationchamberwiththeaidoftheattachedtransferarms. Themanipulatorcanbeusedtoannealsamplesifnecessarybyeitherindirectradiativere- sistiveheatingorbydirectcurrentheating,dependingonthetypeofsample.Radiative heatingistypicallyusedformetallicsurfaceswhereacurrentdriventhroughthesample cannotproduceattemperatureincrease.Thisheatingmechanismisrestrictedtoa maximumtemperatureof ˇ 900 C,althoughtypicallythetemperatureiskeptmuchlower thanthistoavoidexcessiveheatingoftheentirepreparationchamber.Forsemiconducting samples,suchasSi,directcurrentheatingcanbeappliedtoachievetheappropriatetemper- aturesrequiredforsurfacereconstructionwhichcanreachupto ˇ 1400 C.Thepreparation chambercurrentlyhousestwolow-temperaturecellswithKnudsenbasedboronni- tridecrucibledesignsfromCreatecFischerandCoGmbH,capableofevaporating organicmoleculestothesubstratesurface.Forsamplesthatrequirecontrolledamountsof Sitobedepositedonthesurface,aSievaporationset-upisavailablewithinthepreparation chamber.AniongunsourcefromSPECSSurfaceNanoAnalysisGmbHconnectedtoa leak-valvecontrolledargongaslineisalsoattachedtothepreparationchamberforcleaning metallicsubstrates,suchasAg(111).Besidestheequipmentforpreparingthesubstratesur- face,thepreparationchamberalsohousesamulti-channelplate(MCP)low-energyelectron (LEED)apparatusfromOmicronNanoTechnologyGmbH(OxfordInstruments) whichallowselectrondmeasurementstobemadeonavarietyofsamples.The LEEDopticsofthisset-uparecapableofproducinglowelectronbeamcurrentsdowntothe 35 pico-ampererange. Aftersampleshavebeenappropriatelypreparedinthepreparationchamber,theycan thenbetransferred insitu toasamplecarouselwhichcanholduptosixdtsamples orprobetips.Heretheyawaittransfertothescanningstageofthelow-temperaturescan- ningtunnelingmicroscope(LT-STM)andQ-Plusatomicforcemicroscope(Q-PlusAFM) chamberforsurfacemeasurements.Ahighqualityvibrationdecouplingschemeforthescan- ningstageisrequiredinordertoachievehigh-resolutionscanningprobemicroscopyimages, whichcanbeachievedbyhavingthescanningstagefreelysuspendedwithinthechamberby threesoftsprings(see Figure3.2 ).Additionally,thebaseofthescanningstageislinedwith Figure3.2: ImageoftheSTMscanningstagecontainedwithintheUHVsystem . metalplatesthatbetweenanarrayofpermanentmagnetsattachedtoacoppercontainer surroundingtheentirestagewhichisconnectedtothebaseoftheinnercryostat.Themetal platesandmagneticarrayproduceeddycurrentswhichfurtherdampenoutanymechanical 36 vibrationsthestagecouldencounter.Twocoppercontainerssurroundthescanningstage andareindirectcontactwiththetwo-cryostatarrangement.Theinnercryostatcanbe witheitherliquidheliumorliquidnitrogenwhiletheoutercryostatiswithliquid nitrogeninordertoprovideathermalshield.Thescanningstagecanmakethermalcontact withtheinnercryostat,andthecoppershieldssurroundingthestageallowforthescanning stagetomaintainlowtemperaturesformeasurements. Figure3.3: ImageofatypicalmechanicallycutPt/IrSTMtipconnectedtoatip carrierforuseintheUHVsystem . InthecenterofthescanningstageisthepiezoelectricscannerwhichtheSTMorQ- PlusAFMtipisattachedto.Thepiezoelectricmaterialisconnectedtoacoursemotorfor quickpositioningofthetiprelativetothesamplesurface.Theelectronicsthatcontrolthe 37 piezoelectrictubeareconnectedtoacontrolboxwhichisresponsibleforcommunicationto thecomputerandmanagingthefeedback,makingscanningprobemicroscopymeasurements possible.IfatipisdamagedorcannotberecoveredbySTMconditioningtechniques,itis possibletoremovethedamagedtipfromthepiezoelectrictubeandexchangeitforanew tip.Scanningprobetipsconsistofelectrochemicallyetchedtungstenwireormechanically cutplatinumiridiumwire.Thetipsarecuttotheappropriatelengthandmechanically tothetipcarrier,allowingforconnectiontothepiezoelectrictubeintheUHVsystem (see Figure3.3 ).Tipsusedthroughoutthedurationofthisdissertationconsistedofplat- inumiridium.Formoreinformationregardingthescanningprobemicroscopymeasurement processes,referto Section3.2.1 . 3.1.2AtomicForceMicroscopeandInvertedOpticalMicroscope Set-Up Figure3.4: ImageofthecombinedinvertedopticalmicroscopeandAFMset-up . 38 AFMmeasurementsoutsideoftheUHVsystemarecarriedoutonacommerciallyavailable AsylumResearch(OxfordInstruments)MFP-3DAFM.TheheadoftheAFMprovides precisemeasurementsoftheAFMcantileverpositionwhich,inturn,producesaccurateforce andtopographymeasurements.Thisisachievedbyusinganinternallasersourcewhichis Figure3.5: ImageofanAFMcantlieverinthecantileverholder . thebacksideofthecantilevertoadetector.Theresponsefromthedetector issyncedtopiezoelectricfeedbackelectronicstocontrolthecantilever.Theheadrestson topofan xy -scanningstagewhichcanmeasurethepositionofthecantileverandcorrectfor bothhysteresisissueswhenscanningand xy piezoelectriccreepintheimage,providing surfacescans.Thescanningstageisintegratedintoaninvertedopticalmicroscopeset-up, whichrestsontopofavibrationisolationopticaltable.Ifthesampleistransparent,this systemcanbeusedtoalignandfocusanexternallightsourcethroughthemicroscopeto thecantilevertip.Theentiresystemtogetherisstoredinasealedboxcontainerwhich restsontopofvibrationdampenersinordertoisolatetheentiresystemfromtheor(see Figure3.4 ). 39 Figure3.6: tviewsofthesealednitrogenenvironmentwhichcanbeat- tachedtotheAFMset-upallowingforAFMstudiesinnitrogen .Notethatthere isnosampleorAFMtipshowninthisimage. FromthedesignarchitecturefortheMFP-3DAFM,itispossibletosealthecantilever holder(see Figure3.5 )andthesampleofinterestwithinano-ringsealedcell.Thisallows formeasurementstobeobtainedwithinacontrolledorgasenvironment( Figure3.6 ). FortheAFMmeasurementscontainedherein,anitrogenatmosphereismaintainedtoavoid sampledegradationduetoexposuretoambientconditions.Thisisdonebyassemblingthe cantileverandsampleinsidethemodularcellinanitrogengloveboxenvironment(asshown in Figure3.7 ).Thecellisthensealedandtransportedoutofthenitrogengloveboxtobe connectedtotheAFMhead.Alowwrateofnitrogengasisintroducedintothecellso 40 Figure3.7: Imageofthenitrogengloveboxthatisusedtopreparethesealed nitrogenenvironmentattachmentforAFMstudies . astomaintainthenitrogenatmosphere,butnotdisrupttheAFMimagingprocess. ThisAFMscannerprovidesamultitudeofvariousmeasurements,conductiveorother- wise,andthereforethecantilevertipmaterialisimportant.Forthepurposesofthisdisser- tation,onlytopographyimagingandKelvinprobeforcemicroscopy(KPFM)measurements arenecessary.DiamondcoatedSitipswithspringconstant k =0 : 2N/mandresonance frequency f =13kHzandTi/IrcoatedSitipswithspringconstant k =2 : 0N/mandreso- nancefrequency f =70kHzareusedfortopographymeasurements,whileonlythelatterare usedforKPFM.FormoreinformationregardingtheAFMmeasurementtechniques,referto Section3.2.1.4 and Section3.2.1.5 . 41 3.1.3PhysicalPropertyMeasurementSystem TomeasuretransportpropertiesofSinanomembranedevicesaftercompletingfabrication, samplesareloadedintoaphysicalpropertymeasurementsystem(PPMS)fromQuantum Design(see Figure3.8 ).ThePPMSisavariabletemperatureandsystemwhichcan Figure3.8: ImageofthePPMSset-uputilizedforperformingtransportmeasure- mentsonSinanomembranedevices . performavarietyofautomatedmeasurementssuchasheatcapacityrelatedmeasurements, magnetometrymeasurements,andelectro-transportmeasurements.Externalmeasurement equipmentcanalsobeeasilyadaptedtorunautomatedmeasurementsinthePPMS.For example,aKeithley2400SSourceMetercanbeconnectedtothePPMStoactasthecurrent source,whileaKeithley6517BElectrometercanbeusedtomeasurethevoltageresponse. Samplescanbewiredtosamplepucks,providedbyQuantumDesign,andareinserted intothesamplechamberbaseforconnectionto12-pinwiredconnections.Theseareac- cessibletothePPMS'selectronics(seetheonlinebrochureforthePPMSwhichiscom- merciallyavailablefromQuantumDesign).Thesamplepuckcanbeattachedtot 42 measurement-dependentapplicationprobeinserts.Newapplicationprobeinsertscanbe designedandintegratedwiththesystemifthereisnosuitablealternativetoperformingthe necessarymeasurements.Formoreinformationregardingdesigninganewapplicationprobe insertsee Appendix A . ThePPMSsuperiortemperaturecontrolbetweenarange0.4Kto400K,depending ontheexperimentrequirements,whichisdonethroughtheuseofheliumexchangegas andsamplethermalcontact.Thegasiswarmedorcooledtotheappropriatetemperature withinthesamplechambertoreducethermalgradientsinthesystem.UsingthePPMS thermometers,thismethodallowsformeasurementstobetakenasthetemperatureisswept; therangeandrateofthesweepbytheuser.If,however,thesampleposition ischangedbyusinganewapplicationprobeinsert,thenitispossibleforuserstoeasily integratetheirownthermalcalibrationmeasurementclosertothesampletoensureproper temperaturemeasurement.Additionalfeaturesmaybenecessaryforspmeasurements, suchasHalltransportmeasurements(see Section3.2.4 ).TheparticularPPMSutilizedfor theseexperimentscanproduceeitheraDCorACmagneticof 9T. 3.2ExperimentalTechniques 3.2.1ScanningProbeMicroscopy 3.2.1.1ScanningTunnelingMicroscopy Scanningtunnelingmicroscopy(STM),pioneeredbyBinnigandRohrerinthe1980s,isa techniqueinvolvingmeasurementoftheelectrontunnelingsignalbetweentwoconductive electrodesthroughacontrollablevacuumbarrier.Afterdemonstratingtheobservation 43 oftheelectrontunnelingsignal,BinnigandRohrerproceededtodemonstratethatthis signalcouldbeusedtoproducetopographicimagesofsubstratesurfaceswithatomicscale resolution[ 136 , 137 , 138 ].Today,STMisnowroutinelyappliedtostudiesrequiringatomic scalesurfacetopographyimagesandtheelectronicstructureofthesubstrateonalocalscale. STMispossibleduetoaquantummechanicalcalledelectrontunneling.Whenthe STMtipisapproximatelyafewangstromsfromthesamplesurface,electronscantunnel throughthevacuumbarrierbetweenthetwo[ 139 , 140 , 141 ].Tounderstandthequantum tunnelingprinciple,considertwoconductingelectrodeswithseparationsmallenoughto allowtunnelingtooccur.Initiallythereisnovoltagebetweentheconductors,and themetallicelectrodeswillreachanelectronicequilibriumwheretheFermilevelsforboth materialsareinalignment.However,ifavoltage V isappliedbetweenthetwoelectrodes,the Fermilevelswillbeseparatedbyenergy eV ,where e isthechargeofanelectron.Thisgives Figure3.9: Diagramrepresentationofdirectedelectrontunnelingthroughavac- uumbarrierbetweenthesampleandthetip .TheFermilevelsofthesampleandthe tipareseparatedby eV duetotheappliedbiasvoltage V . 44 risetodirectedelectrontunneling,asshownin Figure3.9 .Inthisscenario,electronswhich occupystateswithin eV belowtheFermileveloftheconductingelectrodeexhibit ahigherFermilevelandtunnelintounoccupiedstateswithin eV abovetheFermilevel oftheconductingelectrodewithalowerFermilevel[ 139 , 140 , 141 ].Itshouldbenoted thatthistechniquecanonlybeappliedtosamplesdisplayingconductingorsemiconducting properties.Insulatingsurfacescannotbestudiedduetotheirlargebandgapwhich noavailabledensityofstatesforelectronstotunnelintoorbeextractedfrom. Amorecompleterepresentationofthetunnelingprocessforarbitrarygeometryinthree- dimensionscanbedonefollowingtheTmodel[ 142 , 143 ]whichmakesuseof thetunnelingmatrixputforthbyBardeen[ 144 ].Tosimplifythedescription,consider thevacuumbarrierbetweenthetwoelectrodes.Thisbarriercanbeapproximatedasa rectangularbarrieriftheappliedvoltagebetweenthetwoconductorsissmallincomparison totheworkfunctionsoftheSTMtipandthesample.Theworkfunction( ˚ )canbewritten as ˚ = V 0 E F (3.1) wherethebarrierpotentialis V 0 andtheFermienergyis E F .Forsimplicity,aone- dimensionaltunnelingbarriercanbeassumed.Duetothestandardorientationutilized forthevastmajorityofSTMs,theseparationbetweentheelectrodesisdenotedbythedis- placement z .TheelectronmotionisgovernedquantummechanicallybytheScodinger equation.Forthiscase,onlythetimeindependentScodingerequationisnecessary,andit isgivenbythefollowingexpression: 45 ~ 2 2 m d 2 dz 2 + V 0 = E (3.2) where ~ isthePlanckconstant, m isthemassoftheelectron, E istheenergyoftheelectron, and isthewavefunctionoftheelectron.Assumingthewavefunctionisintheplane-wave formandthevacuumbarrierspansthedistance d ,from Equation3.2 ,thegeneralizedsolution takesthefollowingform: = 8 > > > > > < > > > > > : A exp( {kz )+ B exp( {kz ) ;z< d 2 C exp( )+ D exp( ) ; d 2 d 2 9 > > > > > = > > > > > ; ; (3.3) where k and aregivenby k = p 2 mE ~ ; = p 2 m ( V 0 E ) ~ : (3.4) Inthisrepresentationtherectangularpotentialbarrieriscenteredattheorigin.Fromthe generalizedsolution,theelectronwavefunctionwithinthesampleissinusoidal.Whenthe electronentersthevacuumbarrierthewavefunctiondecaysexponentiallyuntilthevacuum- tipinterface.Atthispointthewavefunctiontransitionstobeingsinusoidalagain,though thereisadeceaseintheamplitudeduetothetransmissionprobability ˝ oftheelectrons throughthebarrier.Thisisdeterminedbytheamplituderatio j F j 2 = j A j 2 (see Figure3.10 ). Attheboundarylocations( z = d 2 and z = d 2 ), and d dz mustbecontinuouswhich constrains A , B , C , D ,and F .Bysolvingthesystemofequationsofthewavefunctionat thepotentialboundaries,thecot A canberepresentedintermsof F allowingfor ˝ tobederived,whichisgivenbythefollowing: 46 Figure3.10: Graphicalrepresentationoftheelectronwavefunctiontunneling throughavacuumbarrierofwidth d betweenthesampleandthetip .The Fermilevelsofthesampleandthetipareseparatedby eV duetotheappliedbiasvoltage V .Theelectronwavefunctionwithinthesampleissinusoidal.Astheelectronentersthe vacuumbarrierthewavefunctiondecaysexponentiallyuntilthevacuum-tipinterfacewhere thewavefunctiontransitionsbacktobeingsinusoidal. ˝ = j F j 2 j A j 2 = 1 1+ 1+ 2 k 2 2 k 2 sinh 2 ( ) (3.5) Ifthepotentialbarrierisconsideredhighincomparisontotheelectronenergy, V 0 >>E , thus >> 1,then ˝ canbetothefollowingrelation: ˝ = ) 16 k 2 2 2 k 2 2 exp( 2 )(3.6) Thetunnelingcurrent I inSTMisproportionalto ˝ ,resultinginthekeyrelationforthe atomicscaleresolutionbySTM[ 139 , 140 , 141 ]: 47 I / ˝ / exp( 2 )(3.7) From Equation3.7 ,achangeinthetip-sampleseparationwillcausethetunnelingcurrent tochangebythefactor: I final I initial =exp( 2 d )(3.8) SurfacestypicallyinvestigatedbySTMhaveworkfunctionsthatareapproximately4-5eV. Thus,forstatesbeingprobedattheFermilevel, ˇ 1 A 1 .Ifthetip-sampleseparation changesby1 A,thetunnelingcurrentwillcorrespondinglychangebyafactorof ˇ 0 : 135;an orderofmagnitude.IfoneatomattheendoftheSTMtipis1 Aclosertothesurfacethan theothers,thetunnelingcurrentwillprimarilywthroughthatsingleatomduetothe exponentialsensitivityinthetunnelingcurrent.Additionally,thetunnelingcurrentsignalis limitedtoasmallregionontheSTMtipsuchthattheatomiccorrugationsinthesubstrate ofinterestcanberoutinelyresolved[ 139 , 140 , 141 ]. Forminganatomic-scaletopographicimageofthesurfacerequiresextremelyprecise controlofthetippositionrelativetothesample.Suchprecisecontrolistypicallyachievedby utilizingthepiezoelectrictinherentinsomematerials.Whenavoltageisappliedtothese materials,itresultsinachangeintheirshapebyapreciseamountsuchthatcontrolonthe scaleofangstromsisachievable.Thepiezoelectricelementscommonlyutilizedinscanning probemicroscopymeasurementscanbecontrolledinallthreeorthogonaldirections( x , y , z ) (see Figure3.11 ).Movementofthepiezoelectricinthe xy -planeisconsideredmotionparallel tothesubstratesurfaceformostSTMsystemdesigns.Thepiezoelectrictubeimplemented insuchstudieshavefourelectrodes( x and y directions)surroundinganinnerelectrode( z 48 Figure3.11: Cross-sectionalschematicdrawingofthescanningpiezoelectricele- mentcommonlyutilizedinSPMexperiments .Thepiezoelectriciscapableofprecisely controllingthemotionofthetipinthe x -, y -,and z -directions.Thesegmenteddesignof theouterelectrodesresponsiblefor x and y motionallowsforthepiezoelectrictomoveby bending.The+ z -directionisintothepage. motionorthogonaltothesurface).Electronicsignalsofequalmagnitudebutoppositesign canbesenttotheouterelectrodestoinducebendingmotioninthe xy -plane.Thisallowsfor thepiezoelectrictorasteroverthesubstratesurfacewithveryprecisecontrol.TheSTMtip ispositionedattheendofthepiezoelectric,andthetip-sampleseparationcanbecontrolled bysendingasignaltotheinnerelectrodeonthepiezoelectrictodrivethe z motion. TheSTMistypicallyoperatedinconstant-currentmodewhenformingatopographymap ofthesurface.Inthisscenario,afeedbackloopisusedtomonitorthetunnelingcurrent signalandapplyadjustmentstothe z motioninordertomaintainatunnelingsignal. Thefeedbackloopworksinthefollowingway:anwithtlyhighgainis usedtoconvertthelowtunnelingcurrentsignaltoavoltage;thevoltageisthencompared toapresetvoltagevaluewhichcorrespondstothedesiredtunnelingcurrentset-point;the betweenthetwoissenttoafeedbackcircuit;afeedbackvoltageisthensentto 49 Figure3.12: SchematicdiagramdepictingtheoperationofSTMwhileinconstant currentmode . thepiezoelectricsuchthatthetippositionisadjustedtoyieldthedesiredtunnelingcurrent; thechangein z isplottedwithrespecttothepositioninthe xy -planeasthepiezoelectricis allowedtorasteroverthesurface.Theresultrepresentsthetopography[ 139 , 140 , 141 ].A detailedschematicoftheSTMoperationisincludedin Figure3.12 . 50 3.2.1.2ScanningTunnelingSpectroscopy Scanningtunnelingspectroscopy(STS)isanSTMmethodwherethetipismaintainedata positionrelativetothesubstratesurface.Thevoltage V betweenthetipand sampleisrampedthroughaspvoltagerangeandthetunnelingcurrent I ismeasured givingthe I ( V )characteristics.Atialconductancespectrum,commonlycalledan STSspectrum,canbeobtainedbytiatingthe I ( V )curve.Experimentallythisis typicallydonebyapplyinglock-indetectiontechniques[ 139 , 140 ].Inthismethod,asmall ACsignalisusedtomodulatethebiasvoltagesuchthat V isnowgivenby V = V 0 + V cos !t: (3.9) Thus,the I ( V )responsetoorderin V canberepresentedas I ( V )= I ( V 0 )+ dI dV V = V 0 V cos !t: (3.10) TheACportionofthetunnelingcurrent,whichisproportionaltothedtialconductance atagivenbiasvoltage,canbemeasuredusingalock-inThevoltage V 0 canalso berampedallowingforaspectrumtobesweptastheACtunnelingsignalismeasured. Theentialconductancemeasuredinthistunnelingschemehasfarmoreinteresting implicationsabouttheelectronicstructureofthesamplesurface.Ifitisassumedthat alltunnelingtransitionsbetweenthetipandsampleoccuratconstantenergy,thestates betweentherelativeFermilevelsofthetwoelectrodes,separatedby eV ,willcontribute tothetunnelingsignal.Inordertocalculatethetotaltunnelingcurrent,thetunneling contributionsfromelectronsatthevariousenergylevelswithin eV mustbeaccountedfor. 51 Thenumberofoccupiedstatesofthesample, n sample ,ataparticularenergy, E ,isgivenby thefollowingexpressionfortemperatures T> 0: n sample = D s ( E ) f ( E ) ; (3.11) where D s ( E )isthedensityofstatesofthesampleand f ( E )istheFermifunctionata spenergy E givenbytheexpression f ( E )= 1 exp E E F k B T +1 ; (3.12) where k B istheBoltzmannconstant.Similarly,thenumberofavailablestatesinthetip, n tip ,inwhichelectronscantunnelintoisgivenby n tip = D t ( E eV )(1 f ( E eV )) ; (3.13) where D t ( E eV )isthedensityofstatesoftheSTMtip.Fromtheseexpressions,the generalizedtunnelingcurrentfromsampletotip I s ! t canbewrittenas I s ! t / Z 1 ˝D s ( E ) f ( E ) D t ( E eV )(1 f ( E eV )) dE; (3.14) where ˝ isthetransmissioncotaspreviouslydescribedin Section3.2.1.1 .When T> 0,thermalexcitationscancausereversedirectiontunnelingcurrent I t ! s ,givenby I t ! s / Z 1 ˝D t ( E eV ) f ( E eV ) D s ( E )(1 f ( E )) dE: (3.15) Thetotaltunnelingcurrentisgivenbythebetween Equation3.14 and Equa- 52 tion3.15 whichyieldsthefollowingexpression: I / Z 1 D s ( E ) D t ( E eV )( f ( E ) f ( E eV )) dE: (3.16) Inthiscasethetransmissionprobability ˝ couldbeincorporatedintotheproportionality constantof I ifthetransmissionprobabilityisassumedtobeconstant,areasonableas- sumptionif E ˘ E F .Ifatipconsistingofanoblemetalisused,asdoneinthisstudywith platinum, Equation3.16 canbebyassumingtheSTMtiphasaconstantdensity ofstatesneartheFermilevel. D t isthenindependentof E andcanbeincorporatedintothe proportionalityconstantof I .Bytiatingthetunnelingcurrent I withrespectto V , arelationtotheACportionofthetunnelingcurrentdescribedinthelock-intechniquecan beexpressedas dI ( V ) dV / Z 1 D s ( E ) @f ( E eV ) @ ( eV ) dE: (3.17) Inthelimitthat T ! 0,thebehavioroftheFermifunctionresultsinthefollowingrelation: @f ( E eV ) @ ( eV ) = ) ( E eV )(3.18) which,whensubstitutedinto Equation3.17 ,givesthekeyresult: dI ( V ) dV / D s ( eV ) : (3.19) Hence,measuringthetiated I ( V )signalatlowtemperatures,i.e.thetial conductanceusingthelock-intechnique,providesadirectmeasureofthesampledensityof states[ 139 , 140 ]. 53 3.2.1.3tialConductanceMapping ThetialconductancemappingtechniquecombinestheaspectsofSTMtopography mappingwiththecorrelationbetweenthetiated I ( V )tunnelingsignalandthedensity ofstates.TheSTMisoperatedintheconstantcurrenttopographyimagingmode,while lock-inmeasurementtechniqueslikethosediscussedin Section3.2.1.2 arealsoutilizedto obtainthetialconductancesignalinordertoobtainSTSspectra.Aspbias voltageisselected,determinedbasedoninterestingfeaturesthatmaybepresentintheSTS spectrapreviouslyobtainedonthesurface.AstheSTMtiprastersalongthesurface,the feedbackloopisusedtomonitorthe z piezoelectricmotiontoproduceatopographicmap whilethetiated I ( V )tunnelingsignalissimultaneouslymonitoredfromthelock- in.Thelock-insignalissenttothecomputercontroltoproduceatial conductancemapataparticularbiasastheSTMtiprastersinthe xy -plane[ 139 , 140 ].This processprovidesawaytomakedirectcomparisonsbetweentheobservedsurfacetopography andthedensityofstatescontributionsataselectbias.Thespeedofacquisitionislimitedby thetimeconstantsettingofthelock-inandthepixelresolutionofthemap.Because ofthis,itcantakeconsiderablylongertoproduceatialconductancemapcompared totypicaltopographyimaging.Sincemoretimeisrequired,appropriateprecautionsmust betaken,suchastlyhighsample-noiseisolationandtlylowtemperatures toavoidthermaldriftwhileimaging. 3.2.1.4AtomicForceMicroscopy Becausesamplesmustbeconductive,asSTMreliesonquantumtunneling,thetypesof samplesthatcanbestudiedisrestricted.Realizingthisshortcoming,BinnigandRohrer continuedtoworkonscanningprobemicroscopymethodswhichledtotheinventionof 54 Figure3.13: SchematicdiagramdepictingtheoperationprincipleofAFM . AFMin1986[ 145 ].Theynotedthatthereisaiteinteractionforcebetweenthetipand surfacewhichcouldbeconsiderableincomparisonwiththeintermolecularinteractions.Such interactionscanresultinphysicaldisruptionsofthesurfacefeatures.Tosensethetip-sample interactionforce,thenatureoftheprobeconnectiontothepiezoelectricwasmosothat acantileverwithaforcesensingtipwasheldparalleltothesurfaceastheproberastersalong thesurface.Thetip-sampleinteractionwillvary,resultinginsorchangesinthe oscillatorybehaviorofthecantilever.Thesechangescanbemeasuredtoformatopographic image(see Figure3.13 )[ 141 , 145 , 146 , 147 , 148 , 149 ].Sincethismethodologyreliesonforce measurements,thesamplechoiceisnolongerrestrictedtoconductivesamples,allowingfor insulatingsurfacestobetopographicallyimaged. 55 Independentofthemodeused,itisnecessarytounderstandtheinteractionforcesinvolved betweenthetipandsamplesothattheappropriateinformationcanbeextracted.Theshort rangeforcesbetweenatomsofthetipandsamplecanbedescribedbytheLennard-Jones potential, U LJ ,givenby: U LJ ( r )=4 E ˙ r 12 ˙ r 6 (3.20) where r isthedistancebetweenthetipandsample, ˙ istheeseparationbetweenthe tipandsamplewhere U LJ ( r )=0,and E istheenergyofthepotentialwell.Aplotofthe Lennard-Jonespotentialisincludedin Figure3.14 .From Equation3.20 and Figure3.14 it Figure3.14: ExampleplotoftheLennard-Jonespotentialfortwointeracting atoms . isevidentthattheinteractionenergywillbeminimizedattheequilibriumseparation r 0 . 56 Whenthetip-sampledistanceislargerthan r 0 ,thepotentialisdominatedbylong-range attractiveinteractions.However,atshorterdistances,thePauliexclusionprinciplerestricts theelectroncloudsofthetipandsampleatomsfromoverlappingresultinginarepulsive interactionwhichwillbegintodominate[ 141 , 149 ]. Additionalcontributionstotheinteractionsbetweenthetipandsamplemustalsobe considered.Forexample,electrostaticallyinduceddipolesintheatomsresultinattractive interactions,i.e.vanderWaals(vdW)interactions.vdWinteractionsarealwayspresent betweenthetip-sampleatoms.InAFM,vdWinteractions F vdW canbeapproximatedby assumingtheendoftheAFMtipisasphereofradius R andthesamplesurfaceisrepresented byanplane: F vdW = HR 6 d 2 (3.21) where H istheHamakerconstantand d istheclosestdistancebetweenthesphereandthe plane,i.e.theclosesttip-sampledistance.Foraientlysmalltipradiuswhichmain- tainscontactwiththesurface,whentheseparationiscomparabletotheintermolecular distances F vdW becomesanadhesionforce, F adh bythefollowingexpression: F adh = 4 ˇR (3.22) where isthesurfaceenergyofthesample. Dependingonthematerialsbeingused,theelectrostaticforcemustalsobeconsideredin thetip-sampleinteractions.Thesystemcanbeapproximatedasacapacitor,thepotential energy U el ofwhichisgivenby: 57 U el = 1 2 CV 2 (3.23) where C isthecapacitancebetweenthetipandsampleand V isthetotalvoltage.From Equation3.23 theelectrostaticforce, F el ,canbewrittenasthefollowing: F el = U el = 1 2 @C @r V 2 CV @V @r (3.24) WhenperformingelectricalAFMstudies,thetipismetallic,andtheforcesperpendicular tothesamplesurfaceareconsideredthedominateinteractionsbetweenthetipandsam- ple.Becauseoftheseconsiderations, V isindependentof r andonlythe z -dependencyis t,simplifying Equation3.24 as F el = 1 2 @C @z V 2 : (3.25) Theelectrostaticconsiderationswillbeoftrelevancein Section3.2.1.5 [ 141 ].It shouldbenotedthatmagneticforcesmustbeconsideredifthetiporsampleconsistof magneticmaterials,however,thisisnotthecasefortheAFMstudiescontainedwithinthis dissertation. TheeasiestmodeofAFMoperationtounderstandiscontactmode.Inthismodethe cantilevertipremainsincontactwiththesamplesurface,andthetip-sampleinteraction forceismaintainedconstantbythefeedbackelectronics.Animportantpartofperforming contactmodemeasurementsistocalibratetheAFMtip,ensuringthattheappropriatetip- sampleforceisappliedwhenscanningthesurface.Thisiscommonlyperformedthroughtwo parameters:theresponseofthephotodetectorandthespringconstant, k ,ofthecantilever. 58 Figure3.15: PlotdepictingthechangeintheinteractionforcebetweentheAFM tipandthesampleasafunctionoftheseparationdistance z duringastandard force-distancecurve .Initially,thetipisnotincontactwiththesampleandthepiezo- electricisusedtomovethetiptowardsthesurface(1).Atacertainpointwhenthetipis veryclosetothesurface,mechanicalinstabilitycausesthetiptosnapintocontactwiththe surface(2).Asthepiezoelectriccontinuestoexpandinthe z -directionthetipispushedinto thesurfaceresultingintherepulsiveforceresponse(3).Thetipisthenretractedfromthe surface(4).Dependingonthemagnitudeof F adh thetipmayneedtobeliftedfurtherthan thepointatwhichitoriginallysnappedintocontactwiththesurface.Eventuallythetip snapsthesurfaceandisnolongerincontactwiththesurface(5)[ 141 ]. Tocalibratetheresponseofthephotodetector,aforcecurveistakenonthesurfaceby measuringtheonresponseasafunctionofthe z -displacementofthepiezoelectric,an exampleofwhichisshownin Figure3.15 .Theslopeoftheresponseintherepulsive-force regimecanbeusedtodeterminetheconversionfactorforthefromphotodetector voltageresponsetodistance, .Measuringthespringconstantofthecantileveristypically donebytakingaseriesoffrequencysweepswhilethetipisawayfromthesurface.This yieldsathermalnoisespectraofthecantileverbyplottingtheamplitudeversusfrequency ofeachsweep.Alargepeakwilloccuratthenaturalresonancefrequencyofthecantilever, 59 ! 0 : ! 0 =2 ˇf 0 = r k m (3.26) where m istheemasswhichaccountsforthetip-cantilevergeometry.Bythe peak,thespringconstantcanbedetermined. From and k ,theappropriatemagnitudeofthetip-sampleforceperpendiculartothesur- facecanbedeterminedfortopographymappingbyassumingthecantileverfollowsHooke's law: j F j = k z = k V (3.27) where V isthecantilevervoltageresponse,oftenusedtotheset-pointfor imaging. Becausethetipisalwaysincontactwiththesurface,itispossibletodamagesamples [ 150 ].OnemethodtoavoidtheseissuesistooperatetheAFMinnon-contactmode.In thismode,thetipisoscillatedatornearitsresonancefrequencyoverthesurface,andthe changesintheoscillatorybehaviorismonitoredtomapthesurfacetopography[ 141 ].The equationofmotionforthetipcanbedescribedbyadrivenoscillatorassumingthetipisa point-mass, m ,attachedtoaspring,givingthefollowingexpression: m z + m! 0 Q _ z + kz = F ts + F d cos( ! d t )(3.28) where ! 0 isgivenby Equation3.26 , Q isthequalityfactorofthecantilever, F ts isthe tip-sampleforceinteraction, F d isthedriveamplitude,and ! d isthedrivefrequency.When 60 thetipapproachesthesurface, F ts willcauseashiftinthecantileverresonance.With smallenoughoscillationamplitudes,thesystemcanberegardedasaperturbedharmonic oscillator.Inthiscase,anespringconstantcanbeassignedwhichaccountsforthis smallperturbation: k effective = k @F ts @z (3.29) Ifthetip-sampleforcegradient @F ts @z isconsideredsmall,then Equation3.26 canbeTaylor expandedtogivethefollowingapproximationforthefrequencyshift: f 0 = f 0 2 k @F ts @z (3.30) Innon-contactmode,thechangeintheresonancefrequency, f 0 ,isheldconstantbyfeed- backsignalssenttothepiezoelectricconnectedtothecantilever.Ashifttolowerfrequency correspondstoanattractiveforcewhileashifttohigherfrequencycorrespondstoarepulsive interaction.Typically,thenon-contactAFMmethodistoachieveandrequiresthe useofhighqualityfactorcantileversunderUHVconditions. Abalancebetweencontactmodeandnon-contactmodecanbeachievedbyoperating intappingmode.Inthismode,thetipisdrivenwithalargeamplitudeatitsresonance frequencyandmakesintermittentcontactwiththesamplesurface.Becausethetiptapson thesurface,itisonlyincontactforashorttime,thusresultinginatreduction inthelateralforceswhichtypicallyoccurincontactmode[ 151 ].Sincecontactisstill made,itisstillpossibletodamagethesamplesurface,butthiscanbereducedinthe feedbackprocess.Typicallyfortappingmode,tinteractionregimesaredetermined bycomparingthetiposcillationamplitudefarfromthesurfacetothatwhenitisintapping 61 range.Iftheratiobetweentheseamplitudesisclosetounity,thenitisconsidered\light tapping",reducingdamagetothesamplesurface.Dependingontheoscillationamplitude set-point,changesbetweenattractiveorrepulsiveimagingregimesarepossiblewhichcan resultinimprovedresolutionwhenimaging.However,itshouldbenotedthatthereisa betweenimprovedresolutionandincreasedpossibilitytodamagethesurface. WhiletheUHVsystemdescribedin Section3.1.1 iscapableofperformingnon-contact AFMtechniques,theworkinthisthesismadeuseofcontactandtappingmode. 3.2.1.5KelvinProbeForceMicroscopy In1898,LordKelvindevelopedamethodology,calledtheKelvinprobetechniqueformacro- scopicmeasurementofsurfacepotentialsofunknownmaterials[ 152 ].IntheKelvinprobe technique,thesampleofinterestactsasoneoftheplatesinformingaparallelplatecapaci- tor,theotherconsistsofaknownmetallicmaterial.Theseparationbetweenthesampleand knownplateisthenoscillatedatafrequency, ! ,toinducechangesinthecapacitance, C . Whenconnectedtoacircuit,thisresultsinanACresponsewhichcanbenbyapply- ingaDCvoltagetooneoftheplatesofthecapacitor.Thevoltagerequiredtocancelout theACresponsecorrespondstothesurfacepotentialbetweenthetwomaterials, asshownin Figure3.16 . Kelvinprobeforcemicroscopy(KPFM)makesuseofthesameworkingprinciplesasLord Kelvin'smethodology,howeveritcanbeintegratedwithAFMmeasurementsbysubstitut- ingtheplateforanAFMtip.RatherthanmonitoringtheACsignalgeneratedfromthe oscillationprocess,theelectrostaticforcebetweenthetipandsampleisused,andtherefore non-contactortappingmodesarenecessary.InKPFM,notonlyisaDCvoltage( V dc ) appliedbetweenthetipandsample,butanACvoltageisalsoapplied.Thisresultsinan 62 Figure3.16: SchematicsoftheenergylevelalignmentforthesampleandAFM tipduringtheKPFMmeasurementprocess .(a)Whenthesampleandthetipare separatedfarfromeachotherwithnoelectronic(b)Whenthetipandsampleare broughtclosetoeachotherallowingfortunnelingorelectricalcontact(tappingmodefor AFMstudies).ThisresultsinacurrentwtoaligntheFermilevelswhileinducingashift inthevacuumlevelequivalenttothesurfacepotentialduetothecharging (c)ADCbiasisappliedtonullifythesurfacepotentialwhichisequivalenttothe workfunctionbetweenthesampleandthetip. oscillatoryelectrostaticinteraction,oscillatingthecantileveratthesamefrequencyasthe ACvoltagesignal, ! ac [ 150 , 153 , 154 ].Theelectrostaticinteractionbetweenthetipand sample, F el ,isalreadyexpressedin Equation3.25 ,however V mustnowbemoto accountfortheappliedvoltagesandthesurfacepotentialbetweenthetipand sample.Thisresultsinthefollowing: 63 F el = 1 2 @C @z ( V dc V SPD + V ac sin( ! ac t )) 2 (3.31) where V SPD issurfacepotentialbetweenthetipandsample,i.e.thein theworkfunctionbetweenthetwo.Itisexpressedbythefollowing: V SPD = sample tip e (3.32) where sample and tip aretheworkfunctionsforthesampleandtip,respectively,while e istheelementarychargeofanelectron.Byexpandingtheequationfor F el ,thetip-sample electrostaticinteractioncanbebrokenupintothreeparts: F el = F dc + F ! ac + F 2 ! ac (3.33) where F dc , F ! ac ,and F 2 ! ac aregivenbythefollowingexpressions: F dc = @C @z 1 2 ( V dc V SPD ) 2 + V 2 ac 4 (3.34) F ! ac = @C @z ( V dc V SPD ) V ac sin( ! ac t )(3.35) F 2 ! ac = @C @z V 2 ac 4 cos(2 ! ac t )(3.36) F dc contributestothesignalusedfortopographymapping. F 2 ! ac isusedforcapacitance measurements.Moreimportantly,forKPFM, F ! ac isusedtomapthesurfacepotential ofthesample[ 150 , 153 , 154 ].Inordertodeterminetheworkfunctionofthesample 64 usingthismethod,acalibrationsamplewithawellworkfunctionisnecessary.By comparingthetip-samplesurfacepotentialbetweenthecalibratedsampleandthe sampleofinterest,theworkfunctionofthematerialcanbededuced.Atypicalcalibration sampleusedforKPFMmeasurementsthroughoutthisdissertationishighlyorderedpyrolytic graphite(HOPG). 3.2.2Low-EnergyElectron Inthelate1920s,oflow-energyelectronsofacrystallinematerialhadbeen showntobepossiblebyDavissonandGermer[ 157 ],thoughtheirinitialexperimentsfocused ontheprovidingevidenceofthewave-likepropertiesofelectrons.Furtherexplorationofthis techniquedidnotoccuruntil1960,whenthemodernlow-energyelectron(LEED) designhadbeendeveloped[ 158 ].LEEDisatechniquethatreliesontheuseoflow-energy electrons, E< 200eV,directedatnormalincidencerelativetoasurfacecrystalstructure. Elasticallybackscatteredelectronsarethendetected.Themeanfreepathlengthofthese low-energyelectronsisonly ˇ 10 A.Additionally,thewavelengthofanelectronbeam is ˇ 1 Awhichissmallerthantypicalinteratomicspacingsincrystalstructures[ 155 , 156 ]. TogetherbothofthesetraitsmakeLEEDsuitableforsurfacerelatedexperiments. TounderstandtheLEEDprocess,itisnecessarytodiscusssomebasic tiontheory.First,considertherelationshipbetweenthewavelengthoftheincidentbeam anditswavevectorgivenbythefollowing: k = 2 ˇ (3.37) Thewavevector k isameasureofthebeam'smomentum.Thechangeinthewavevectorwith 65 respecttotheincidentbeamultimatelydeterminesthedirectionofanyedbeams. Conversely,thepatternobtainediscorrelatedtochangesinthewavevector.For Figure3.17: SchematicdiagramdepictingtheEwaldsphereconstructionforan electronbeamwhichisincidentnormaltothesubstratesurface . thisreason,itiseasiertoworkinreciprocalspaceratherthanrealspace.Thisan inverselyproportionalinterpretationsimilartothewavevectorin Equation3.37 ,theresult ofwhichisthatthereciprocalspace(alsocalled k -space)patternbecomesa directinterpretationofthewavevector k .Amajoradvantageisthattheconditionsforcon- structiveinterference,i.e.whenabeamemergesfromthecrystal,isdetermined simplybythelawofconservationofmomentum,encapsulatedinanEwaldsphere[ 159 , 160 ]. Themagnitudeoftheincidentbeam'swavevector, j ~ k i j ,servesastheradiusoftheEwald spherepositionedattheoriginof ~ k i ,inreciprocalspace(see Figure3.17 ).Byconserva- 66 tionofmomentum,noisallowedoutsideofthissphere,leadingtothefollowing relationship: ~ k i = ~ k f + ~g (3.38) where ~g isthebetweenincidentandwavevectors. ForLEED,thereciprocalunitcellisconsideredtwo-dimensional.Bythedis- tancebetweenthereciprocallatticepointsisinverselyproportionaltotheseparationinreal space,andsincethesurfaceisconsideredtwo-dimensional,thereciprocallatticepointsnor- maltothesurfaceareseparated.Thisresultsinreciprocalspacerods.Neigh- boringrodsareseparatedbythereciprocalspacelatticeconstantforthetwo-dimensional crystalgivenby2 ˇ=a ,where a isthelatticeconstantinrealspace.Thecondition ( Equation3.38 )forLEEDisbyanybeamthatisdirectedatapointwhere thereciprocalrodsintersecttheEwaldsphere(see Figure3.17 ).Directionofthe beams,aswellasthenumberofedbeams,canbemobychangingtheenergy oftheincidentelectronbeam[ 159 , 160 ].Theenergyoftheincidentbeamisgivenby: E beam = ~ 2 2 m ( k i ) 2 ; where k i = 2 ˇ i (3.39) Thus, j k i j ismoedbychangingthebeamenergy. ThetypicalexperimentalarrangementforLEEDmeasurementisshownin Figure3.18 . Anelectrongunisusedtogenerateabeamofelectronsthataredirectedatnormalincidence tothegroundedsamplesurface.Amajorityoftheseelectronsareinelasticallybackscattered thesurface,butthosethatscatterelasticallyareresponsibleforformingtheobserved LEEDpattern.Theseelectronsgotoahemisphericalsystemofgridswhichsur- 67 Figure3.18: SchematicdiagramoftheLEEDopticsusedfordetecting electrons . roundtheelectrongun.Thegridsareusedtoacceleratetheelasticallyscatteredelectronsto atscreenwhilesimultaneouslyrejectinginelasticallyscatteredelectrons,reducing theirontheLEEDscreen.Thescreenissettoahighvoltage(near6kV)so thatelasticallyscatteredelectronswithenoughenergywillcauseanexcitationinthescreen, resultinginbrightLEEDspotstoshow.TorecordtheobservedLEEDpatternarear-view LEEDdesignisused,inwhichaCCDcameraisattachedtoarearwindowportbehind theLEEDscreen.ThepatternisthereforenotobscuredbyanyequipmentwithintheUHV chamber.ForthespLEEDset-upusedinthisstudy,theLEEDapparatuscontains multi-channelplates(MCP)intheLEEDopticswhichallowsforlowelectronbeamcurrents 68 inthepico-ampererange.Keepingthebeamcurrentlowreducesdamagetodelicatesam- plesorrapidsamplecharging,makingstudiesonorganicsystemsandinsulatingsurfaces possible. 3.2.3VanderPauwTransportMeasurement Inordertoperformsheetresistanceandresistivitymeasurementsofthinsamples,which arehighlysensitivetosurfaceprobeL.J.vanderPauwdevelopedwhatisnowcoined asthevanderPauwmethod[ 161 ].Thistechniqueallowsfortransportmeasurementson thinofarbitraryshapeandsize,aslongasthethinhasahomogeneousthickness, noholeswithintheregionbeingmeasured,andtheelectricalcontactsareconnectedto theouterlimitsofthethinsample.ThevanderPauwmethod,consideredafour-probe transportmeasurement,measuresthevoltagebetweentwocontactswhileacurrentispassed betweentheothertwo.Theofthistechniqueisthatavoltagemeasurement essentiallydrawsnocurrent,resultinginanegligiblecontactresistance,ahighlyimportant resultinordertoobtainaccuratesheetresistanceandresistivitymeasurementsofthin systems. In Figure3.19 ,thetvanderPauwmeasurementareshownfor aperfectsquaresamplegeometry.Asdemonstratedin Figure3.19 ,inordertodetermine thesheetresistanceandresistivity,measurementsmustbeperformedthroughtheeight of I and V .Inordertodeterminetheresistance,ageometricalfactor, f , mustbeintroduced,relatedtotheratioofthemeasuredvoltages, Q ,fromthetvan derPauwTheseratiosarebythefollowing: 69 Figure3.19: Circuitdiagramsofthetwiringrequiredfor vanderPauwtransportmeasurements .Startingfromthecocurrent I issentthroughcontactpoints1and2whilethevoltageismeasuredbetweencontacts3 and4toobtain V 1 .Maintainingthesamemeasurementcthepolarityof I is thenreversedand V 2 ismeasured.Asimilarschemeisusedfortheremaining toobtain V 3 through V 8 . Q A = V 1 V 2 V 3 V 4 (3.40) Q B = V 5 V 6 V 7 V 8 (3.41) withthe V 1 through V 8 shown Figure3.19 .Thevoltageratioandthegeometric factorarerelatedbythefollowingexpression: Q 1 Q +1 = f ln(2) arccos exp ln(2) f (3.42) Thisrelationshipcanbeplottedtogivethecorresponding Q - f curve,shownin Figure3.20 . Sincethethinsampleismaintainedasclosetoaperfectsquareaspossible,the Q ratios 70 Figure3.20: Graphicalrepresentationoftherelationshipbetweenthevoltageratio Q andthegeometricfactor f involvedinvanderPauwtransportmeasurements . willbecloseto1,making f ˇ 1.Thesheetresistancevalues R S;A and R S;B aredetermined bythefollowingequations: R S;A = ˇf A ln(2) V 1 V 2 + V 3 V 4 4 I (3.43) R S;B = ˇf B ln(2) V 5 V 6 + V 7 V 8 4 I (3.44) whichcanbecombinedfortotal R S as R S = R S;A + R S;B 2 (3.45) Theresistivity, ˆ ,canbedeterminedbyasimpleconversionifthethinthickness, t sample , isknown: ˆ = R S t sample (3.46) 71 Whenworkingwithresistivethinsamples,someprecautionsarenecessarytomin- imizepossiblesourcesoferrorinthevanderPauwmeasurements[ 162 , 163 ].Duetothe highresistanceofthelowcurrentsmustbeusedtominimizethermalheatingofthe sampleinordertopreventitfromburningout.Forthisreason,acurrentsourcethatis accuratedowntopAorlowerispreferred.Duetothehighresistanceofthesamplebeing measured,itisalsonecessaryfortheinputimpedanceofthevoltmetertobeatleasttwo ordersofmagnitudehigherthanthatofthesamplebeingmeasured,droppingmeasurement error < 1%. Tofurtherreducenoise,measurementcablesshouldberunascloseaspossibletothe sample,andextracableguardingisnecessarytopreventanyleakagecurrentinthesignal. Measurementsareperformedinastainlesssteelenclosuretoreducetheofexternal electromagneticnoise.Themeasurementenvironmentisalsokeptdarktoremoveanypos- siblephotoelectricPractically,thedimensionalityoftheelectriccontactconnections mayresultinplacementslightlyintothesampleareaasopposedtotheedge.Thiscanin- troducesomeexperimentalerrorinthesheetresistanceandresistivitymeasurementswhich canbeinthevaluesof Q A and Q B aswellastheircorresponding f values.Ifall ofthecontactsdonotextendoutfartherthan10%ofthetotalsamplesidelengththenthe errorintheresistivitycanbekept < 0 : 1%. Formeasurementsperformedinthisstudy,aKeithley2400SSourceMeterisusedas thecurrentsourceandaKeithley6517BElectrometerisusedtomeasurethevoltagere- sponsefromthethinlmmaterial.MeasurementsareperformedwithinthePPMSset-up. ForadditionalinformationregardinghowthesamplesareloadedintothePPMS,referto Appendix A . 72 3.2.4HallMeasurement HallmeasurementsmakeuseoftheHalldiscoveredbyEdwinHallin1879[ 164 ].Inthis scheme,avoltage,i.e.theHallvoltage, V H ,isproducedacrossaconductorwith bothacurrentpassingalongitandamagneticappliedperpendiculartothecurrent w.DuetotheLorentzforce,achargecarriermovingalongadirectionperpendicularto themagnetic ~ B ,experiencesaforce ~ F whichisperpendiculartoboththemagnetic andcarriervelocity,givenbythefollowingexpression: ~ F = q~v ~ B (3.47) where q istheelectricchargeofthecarrier.Ifanelectrondrivenbyanelectric ~ E , isdirectedperpendiculartothemagneticthentheforcetheelectronexperiencesis writtenas: ~ F = e ~ E + ~v ~ B (3.48) Assumingaconstantcurrent, I ,ismaintained,electronswhicharesubjectedtotheLorentz forcewilldriftawayfromthecurrentdirectionandcollectononesideofthesample.This buildupofchargeresultsinapotentialbetweenthetwosidesofthesample, V H , asshownin Figure3.21 . Hallmeasurementscanbeperformedonsamplesusingthesamegeometryusedforthe vanderPauwmeasurements,byapplyingamagneticdirectionperpendiculartothe sample.Inthisgeometry,thecurrentisdirectedbetweendiagonalelectricalcontactsand V H measuredbetweentheothertwocontacts,asshownin Figure3.22 .Itshouldbenoted 73 Figure3.21: SchematicdiagramdepictingtheHallandmeasuredHallpo- tential .ElectronsexperienceaLorentzforcewhichcausesthemtodriftaway fromthecurrentdirectionandcollectononesideofthesample.Thisbuildupofchargeis responsibleforthepotentialbetweenthetwosidesofthesample. that V H mustbemeasuredforbothcurrentdirectionsaswellasbothmagneticdirec- tions.Bydoingso,thevoltageoduetothealignmentoftheelectricalcontactsand magnetoresistancecanbecanceledout.Thisresultsineightmeasurementsof V H , whichcanbeusedtodeterminetheHallcot, R H ,givenbythefollowing: R H = ( V 42+ V 24+ + V 24 V 42 + V 31+ V 13+ + V 13 V 31 ) t sample 8 BI (3.49) where V xy isthevoltagemeasuredfortion xy ,asshownin Figure3.22 ,andthe + = subscriptreferstothemagneticdirection. MeasurementoftheHallcotcanyieldmaterialparametersthatarekeytounder- standingthenatureoftheelectricaltransport.FromtheHallcotalone,thecarrier 74 Figure3.22: Circuitdiagramsofthetwomainwiringtionsrequiredfor Halltransportmeasurements . concentration, n ,ofthesamplecanbedeterminedbythefollowingrelation: n = 1 eR H (3.50) ThesignoftheHallcotdeterminesthedominanttypeofcarrierinthematerial:a negativesignindicatesthatelectronsarethecarriertype,whileapositivesigncorresponds toholecarriers.IfvanderPauwmeasurementsarealsocarriedoutonthesamesampleto determineitssheetresistance R S ,asshownin Equation3.45 ,thenthecarriermobility of thesamplecanbedeterminedby: = j R H j R S 1 t sample (3.51) Whenworkingwithresistivethinsamples,similarprecautionsarenecessarywhen performingHallmeasurementsasinvanderPauwmeasurementstominimizepossible sourcesoferror[ 162 , 163 ].Themajorityofthesourcesoferrorarealreadydescribedin Section3.2.3 sotheywillnotbereiteratedhere.However,itshouldbenotedthatfor 75 Hallmeasurementsitisextremelyimportanttomaintainmacroscopicsamplehomogeneity. Failuretodosocanresultinincorrectassignmentofthecarriertype. Hallmeasurementsperformedinthisstudyusedthesamesourceandmeasurementset- upasthevanderPauwmeasurementwithsamplekeptinthePPMS.ThePPMSiscapable ofproducingthemagneticperpendiculartothesamplenecessaryforHallmeasurement. ForadditionalinformationregardinghowthesamplesareloadedintothePPMS,referto Appendix A . 3.3ComputationalTechniques 3.3.1EpitaxialRelationwiththeSubstrate Theroleofepitaxyininorganicgrowthhasalreadybeenwellestablishedwithinthe however,fororganicthingrowthspecialconsiderationsmustbemadewhenin- vestigatingtheinterfacebetweenthemolecularoverlayerandthesubstratesurface.Dueto thedelicatebalancebetweentheintermolecularinteractionsandthemolecularinteractions withthesubstrate,itistopredicteithertheorganizationofthemolecularor theepitaxialrelationship.Itisalsopossibleforlessobviousepitaxialmodestoarise.To addressthisissue,modelingmethodologieshavebeendevelopedtoenableanalysisofepitaxy occurringattheinterface.Thefocuswillparticularlybeonthegeometricphasecoherence model.However,inordertoappropriatelydescribetheinformationgainedfromthismod- elingmethod,itisnecessarytoprovideabriefdescriptionofthetepitaxialmodes thatcanbeencounteredinthecalculation. Theepitaxialrelationshipattheinterfacebetweenanoverlayerandthesubstratesurface canbeaccuratelydescribedbythefollowingkeyparameters:thesubstrateunitcellpara- 76 Figure3.23: Schematicrepresentationofthetpossibleepitaxialrelations betweenanatomicoverlayerandtheunderlyingsubstrate .Thesubstratesurface isrepresentedbytheblueatomsandtheoverlayerisrepresentedbytheredatoms.For simplicityintheinterpretation,theoverlayerlatticepointsareshowntobedirectlyontop ofsubstratelatticepoints,however,thisdoesnotnecessarilyhavetobethecase.(a)depicts thecaseofacommensurateepitaxialrelationwhereeveryatomoftheoverlayerregistersto aspsiteonthesubstrate.(b)depictsthecaseofaPOLcoincidentstructurewhere therearedistinctperiodiclinesintheoverlayerwhichregisteronsubstratesites.(c)depicts thecaseofacoicidentstructurewhichproducesatwo-dimensionalsuperstructuredueto theregistration. meters,denotedby a 1 , a 2 ,andtheangle between ~a 1 and ~a 2 ;theoverlayerlatticeparame- ters b 1 , b 2 ,andtheangle between ~ b 1 and ~ b 2 ;andtheazimuthalangle between ~a 1 and ~ b 1 whichtherotationbetweentheoverlayerandtheunderlyingsubstrate[ 165 ].From theseparametersthesubstratelatticevectorsandtheoverlayerlatticevectorscanberelated byastructuretransformationmatrix[ M ],givenby: 77 [ M ]= 2 6 4 M 11 M 12 M 21 M 22 3 7 5 = 2 6 4 b 1 sin ( ) a 1 sin ( ) b 1 sin ( ) a 2 sin ( ) b 2 sin ( ) a 1 sin ( ) b 2 sin ( + ) a 2 sin ( ) 3 7 5 (3.52) Frommatrixelementsof[ M ],themodeoftheepitaxycanbedetermined.Ifallmatrix elementsareintegernumbers,theneverylatticepointinthemolecularoverlayerwillregister withanunderlyingsubstratelatticepoint,asshownin Figure3.23 (a).Thisiscommonly referredtoasacommensurateepitaxyrelationship,whichisconsideredthemorefavorable conditionbecausetheoverlayerandsubstratesurfacehavecommonlatticeperiodicities, givingacoherentinterface.Ifinsteadthematrixelementsconsistofatleasttwointegers whicharetoasinglecolumnofthematrix,thenparticularlinesoftheoverlayer latticewillcoincidewithlinesoftheunderlyingsubstrate.Thisisreferredtoaspoint-on-line (POL)coincidenceepitaxyandresultsinaone-dimensional,registrationlinepattern,ascan beseenin Figure3.23 (b).Generally,POLcoincidenceislessfavorablethanacommensurate relation.Itisalsopossibletohavesplinesoftheoverlayerlatticecoincidewithlinesof theunderlyingsubstratesuchthatregistrationproducesatwo-dimensionalpatternasshown in Figure3.23 (c).Thisoccurswhenallmatrixelementsarerational,butnospcolumn consistsofintegers.Thelastepitaxyrelationtoiswhenoneofthematrixelements of[ M ]isanirrationalnumberandneithercolumnconsistsofintegers.Whenthisoccursthe overlayerdisplaysnospregistrationwiththesurface,alsocalledanincommensurate structure[ 165 ].Withtheseofpossibleepitaxyrelationsthatcanoccuratthe overlayer-substrateinterface,theepitaxialrelationshipbetweenamolecularoverlayerand thesubstratesurfacecanbemodeled. 78 3.3.1.1GeometricPhaseCoherenceModel Todeterminewhetherthereexistcommonlatticeperiodicitiesbetweentheoverlayerand substrate,theperiodicityofthesubstratelatticeandoverlayerlatticearerepresentedas plane-waves.Thisprocessresultsina\surfacepotential"landscapewhichispurelyageo- metricalconstruct.Theoverlayerlatticeisrotatedazimuthallywithrespecttothesubstrate, andthediscreteratioofthetwoplane-wavepotentialsarecomparedthroughouttherotation process,dbythefollowing: V V 0 = 1 2 N 2 2 N 2 sin( ˇNM 11 )sin( ˇNM 21 ) sin( ˇM 11 )sin( ˇM 21 ) sin( ˇNM 12 )sin( ˇNM 22 ) sin( ˇM 12 )sin( ˇM 22 ) (3.53) where N isrelatedtothesizeofthemolecularoverlayer.Therespectivelatticeparameters ofthesubstrateandtheoverlayercanbedeterminedfromavarietyofexperiments,such asSTMandLEED.Afterassigningtheappropriatesizeforthemolecularoverlayer,the overlayerisrotatedby tominimaintheplane-wavepotentialratio[ 165 ]. V=V 0 =0 correspondstothelowestamountoflatticemismatchattheinterface,resultinginacom- mensurateepitaxyrelationship,suchasthestructureshownin Figure3.23 (a).Alter- natively, V=V 0 =1correspondstothelargestlatticemismatchattheinterfacebetween theoverlayerandthesubstrate,indicatingtheoverlayerisincommensuratewithsubstrate. V=V 0 =0 : 5correspondstoacoincidentepitaxialrelationship,spPOLcoincidence. Thismethodologyprovidesawaytosearchforepitaxialcbysamplingthe geometryoftheoverlayer-substrateinterfaceandquicklyidentifywhichwill producecommensurateandcoincidentstructures. 79 3.3.1.2SimulatedMoirePattern Characterizationoftheepitaxyintermsofplane-wavesuperpositionsbetweentheoverlayer andthesubstratesurfacesharesmanyofthetraitsusedtodescribeamoirepattern.Amoire patternoccurswhentworepeatingpatternsaresuperimposedontopofeachotherwithone ofthepatternsslightlydisplacedorrotatedwithrespecttotheother.Thiscanproducea superstucturesuchastheoneshownin Figure3.24 .Mathematically,thedescriptionofthe Figure3.24: Exampleimageofamoirepatternproducedbytwoperiodicstruc- tureswhichhavebeenrotatedrelativetooneanother . moirepatterncorrespondswellwiththeplane-wavedescriptionoftheoverlayer-substrate interface,suchasinthecaseofPOLcoincidencewhereone-dimensionalregistrationlines canbeobserved.ThisisfurthercorroboratedbythefactthatSTMimagingofmolecular overlayersoncrystallinesurfaceshavedisplayedmoirepatterns.Whenimagingthesurface, theelectronicstatesoftheoverlayerandthesubstratearesuperimposed,whichcanresult inamoiresignaturetoappearwhenmappingthetopography. 80 Bycombiningtheuseofthegeometricphasecoherencemodelwithsoftwarewhichcan generateasnapshotoftheoverlayerrotatedwithrespecttothesubstratesurfacelattice,it ispossibletoproduceimagesofthepredictedoverlayerregistrationtothesubstrate,similar tothoseshownin Figure3.23 .Inordertoclearlydistinguishbetweensitesoftheoverlayer latticethatsharecommonsiteswiththeunderlyingsurface,andthosethatdonot,the substratelatticesitesarerepresentedbycoloreddotsoutlinedbyatcolor.When theoverlayerlatticesitesalignwithsubstratelatticesites,thisproducesanapparentcolor changewhencomparedtooverlayersiteswhicharenotaligned.Thus,latticesitesthatalign betweentheoverlayerandthesubstrate,formingacoincidentstructure,willproducean apparentsuperstructureindicativeofamoiresignaturewhichcanthenbecomparedwith experimentallyobtainedSTMimages. 3.3.2SimulatedLEEDPatterns Asdescribedin Section3.2.2 ,theLEEDpatternproducedexperimentallyisarepresenta- tionofthetwo-dimensionalsurfacelatticeinreciprocalspace.Ifthesubstratelatticeeither hasalreadybeendeterminedbypreviousexperimentsoristheoreticallyknown,thenitis possibletoformacorrespondingreciprocalspacerepresentationbysimplyusingtheknown parameterstomathematicallygenerateasimulatedpattern.However,themoreadvanta- geousroutefromanexperimentalperspectiveistobeabletosimulateaLEEDpattern forcomparisonwithanexperimentalobservedpattern.LEEDanalysiscaninthiswaybe comparedwiththeparametersdeterminedfromSTMtoensureproperlatticeassignment. ThereareseveralcommercialsoftwarepackagesavailableforLEEDsimulations.For LEEDstudiescontainedwithinthisdissertation,thepatternsaresimulatedusingeither LEEDLabincombinationwithLEEDCalsoftware,providedbyOmicronNanoTechnology 81 GmbH(OxfordInstruments),orfreeonline-availablesoftwareLEEDpat,providedbyK. HermannandM.A.VanHove.LEEDLaballowsforpatternswhichconsistofuptofour layerswithfullcontrolofthelatticeparametersofeachlayer,aswellasthero- tationbetweeneachlayer,tobesimulated.Additionally,LEEDLabcanaccountformirror planeandgliding-planesymmetryconsiderations.MoreimportantlyforsimulatingLEED patternsforlayeredstructureswithpossiblemoiresuperstructures,multiple-scatteringcan befactoredintothesimulatedpattern.ThoughLEEDpatdoesnothaveasmuchlayer control,ithastheaddedbofdisplayingwhichpeakscorrespondtoapar- ticulardomaintypeifthesymmetryofthelayerontopofthesubstrateallowsformultiple domains.Theseconsiderationsareallparticularlyimportantwhenperformingthin growthofmolecularspeciesonacrystallinesurface. 3.3.3DensityFunctionalTheory In1964and1965,twoseminalpapersbyHohenberg-KohnandKohn-Shamintroducedden- sityfunctionaltheory(DFT)[ 166 , 167 ].Inconventionaltheoreticalmethods,thewavefunc- tion,oftheelectronsystemisusedsinceitcontainsthefullsetofinformationregarding thesystem.However,becausecannotbeprobedexperimentallyanditcanbecomequite complicatedformany-bodysystems,adtquantitywhichcanstillgleantheimportant informationregardingthesystemisnecessary.DFTmakesuseoffunctionalsoftheelectron density.Thisisbecausetheexternalpotentialcanberepresentedtowithinaconstantby afunctionaloftheelectrondensity.Inprinciple,thepotentialrestrictstheHamiltonian, ^ H ,suchthatthegroundstateofthemany-bodyelectronsystemcanberepresentedbya functionaloftheelectrondensity.Thistheoryhasbeenbuiltuponinordertoaccountfor Coulombinteractions,exchange-correlation,etc.,whichrequirestheuseofframe- 82 worksandapproximationmethodstoensuretheappropriatepotentialconsiderationsare beingmadeforagivensystemofelectrons. Forthiswork,BingHuang,ChangwonPark,andMinaYoonatOakRidgeNational LaboratoryperformedDFTcalculationsfor M Pcmoleculesadsorbedonthedeactivated Si(111)-Bsurface.Assuch,onlydetailsregardingtheircalculationmethodsareincludedas detailedinReference[ 43 ]. Generalizedgradientapproximation(GGA)isusedandimplementedintheVienna ab initio simulationpackage(VASP)[ 168 ].ThePerdew-Burke-Ernzerhof(PBE)functionalis utilizedwiththeprojectoraugmentedwave(PAW)potentialsappliedtodescribethecore electrons.ToaccountforthevdWinteractions,aPBE+vdWschemewithaself-consistent electrondensityisutilized[ 169 ].Fortheplane-wavebasis,thekinetic-energyremained at400eV.TheSi(111)-B p 3 p 3surfaceisrepresentedbya4 4supercelltostudythe interactionsbetweenthe M PcmoleculesandtheSiadatoms(ad-Si)whereeverystructure isfullyrelaxeduntiltheforceoneachatomislessthan0.02eV/ A.Itshouldbenotedthat thereareissueswithPBEfunctionalinaccuratelydescribingtheenergylevellocationsand bandgaps[ 170 , 171 ].Becauseofthis,theabsolutecalculatedbindingenergyandenergy barriervaluesshouldbeviewedwithcaution.However,theobservedtrendscanbecorrectly reproduced.Potentiallandscapesareobtainedbyaconstrainedenergyminimizationalong thepathofamolecule.Atlythickslab(7atomiclayers)isusedfortheSisubstrate suchthatthepotentiallandscapeontheupperlayerisnotbytherelaxationofthe bottomlayer.Tothis,thenumberoflayerscanbeincreasedandstillshowsthatthe landscaperemainsessentiallyunchanged.ASiatominthesubstrateisasareference pointinordertopreventthewholesubstratefromdriftingtoitsglobalminimumenergy structure.Foratranslationalpotentiallandscape,the x and y coordinatesofthemetalatom 83 inthe M Pcmoleculearewhileforarotationallandscape,inadditiontothis,theforce componentresponsiblefortorqueisprojectedout.Allremainingatomsarefullyrelaxed. 84 Chapter4 BackgroundofMaterials Priortoorganicmolecularthingrowth,thoroughinvestigationofthesurfacesandin- terfaceswhichwillbeutilizedinthisstudyisrequired.Thischapterprovidesanextensive backgroundonthetmaterialsandsurfacesthatareusedthroughoutthisdissertation, aswellasthepreparationmethodsthatareutilizedtoensureanidealsampleispreparedfor measurement. Section4.1 providesabriefintroductiontosemiconductorphysics. Section4.2 detailsthetSisurfacesusedthroughoutthethingrowthevolutionstudieswith descriptionsofhowtopreparetheappropriateinterface.Thisisfollowedbydiscussionof silicon-on-insulatortechnology(SOI)in Section4.3 withspfocusesontheSinanomem- branematerial,includingfabricationmethodsandapplicationsofthismaterial.Lastly,in Section4.4 ,abriefbackgrounddescriptionregardingthe M Pcmoleculeisincludedalong withhowthesemoleculesarepreparedforuseinthinstudies. 4.1IntroductiontoSemiconductorPhysics Forasemiconductingmaterial,thevalencebandandconductionbandareseparatedby whatisknownasthebandgapofthematerial, E gap ( Figure4.1 ).Considerasemiconductor maintainedat T =0K.Inthisscenariotheelectronswilloccupythevalencebandsuch thatitiscompletelyAsaconsequence,theconductionbandisempty.Whenthe temperatureofthesemiconductorisincreased,electronscanbethermallyexcitedfromthe 85 valencebandtotheconductionband,introducingfreeelectroncarriersatthebottomofthe conductionband.However,theprocessofexcitinganelectronoutofthevalencebandtothe conductionbandresultsinaholeatthetopofthevalenceband[ 172 , 173 ].Theelectrons intheconductionbandandholesinthevalencebandultimatelydeterminetheelectronic behaviorofthesemiconductor. Figure4.1: Energydiagramofdonorandacceptorlevelsthatcanbepresentin theSibandgapformedbetweenthevalenceandconductionbands . Themostcommonlyutilizedsemiconductingmaterialintheelectronicsindustry,aswell astheprimaryfocusofthisdissertation,isSi.SiisaGroupIVelementintheperiodic tablewhichpossessfourvalenceelectrons.Wheninitscrystalform,everySiatomisable toformfourcovalentbondswithneighboringatomsbysharingthesefourelectrons.Sialso displaysanindirectbandgap, E gap =1 : 12 e Vatroomtemperature.ConsideringapureSi crystal,theconductivityofthecrystalisrelatedtotheintrinsicnumberofcarriersinthe material.Inordertocalculatethenumberofchargecarriersatagiventemperature, T , theprobabilityforanelectrontooccupyastateofenergy, E ,relativetotheFermilevelof 86 thesemiconductor, E F ,mustbecalculated[ 172 , 173 ].ThisisgivensimplybytheFermi function: f ( E )= 1 exp E E F k B T +1 (4.1) where k B istheBoltzmannconstant.From Equation4.1 itispossibletowritetheequations forthedensityofstatesoftheconductionband, D CB ,andvalenceband, D VB ,ofthe materialbyconsideringfreeelectronsandholesinthree-dimensions.Theseexpressionsfor thedensityofstatesaregivenbythefollowing: D CB ( E )= V 2 ˇ 2 ~ 3 (2 m e ) 3 = 2 p E E gap (4.2) D VB ( E )= V 2 ˇ 2 ~ 3 (2 m h ) 3 = 2 p E (4.3) where V isthevolumeand m e and m h aretheelectronandholemass,respectively.Inorder todeterminethenumberofelectronsintheconductionbandforagivenvolume, n ,the productof f ( E )and D CB ( E )mustbeintegratedovertheenergyrangespanningfromthe bandgapenergy E gap throughallconductionbandenergies.Similarly,thenumberofholes inthevalencebandforagivenvolume, p ,canbecalculatedbyintegratingtheproductof (1 f ( E ))and D VB ( E )overtheenergyrangespanningallvalencebandstates.Thisgives thefollowingexpressions: n = 1 V Z 1 E gap D CB ( E ) f ( E ) dE = N CB exp E F E gap k B T (4.4) 87 p = 1 V Z 0 D VB ( E )(1 f ( E )) dE = N VB exp E F k B T (4.5) where N CB and N VB aregivenbythefollowing: N CB =2 2 ˇm e k B T h 2 3 = 2 (4.6) N VB =2 2 ˇm h k B T h 2 3 = 2 (4.7) ForapureSicrystal,theelectronandholeconcentrationsmustbeequalsincetheholesin thevalencebandaregeneratedbytheactofthermallyexcitingelectronstotheconduction band.Thus,theintrinsiccarrierconcentrationoftheholes, p i ,andelectrons, n i ,canbe writtenasthefollowing: n i = p i = p N CB N VB exp E gap 2 k B T (4.8) Itispossibletomodifycarrierconcentrationofthesemiconductoringmaterial,andconse- quentlyshifttheFermilevelbyintroducingwhatarecalleddopantsintothesemiconducting crystal.Thisiscommonlyachievedbyreplacingsomeoftheatomsinthepuresemiconductor crystalwithimpurityatoms.ForSi,thisinvolvesusingimpurityatomsfromeitherGroup III(typicallyboron,B)orGroupV(typicallyphosphorous,P,orarsenic,As)elements.If SiisreplacedwithB,thethreevalenceelectronswillcovalentlybondwiththreeofthefour Sielectrons,leavingavacancyinthecrystalwhichgeneratesaholecarrier.Inthecommon semiconductorvernacular,theBdopantisconsideredanacceptor.WhenSiisdopedwith Bitiscalledp-typedoping.IfinsteadofusingB,PdopantsareintroducedintotheSi 88 crystal,fouroutoftheeelectronsofthePatomwillcovalentlybondwiththefourSi electrons,leavinganexcessofoneelectroncarrier.Inthiscase,thePdopantisconsidered adonor.WhenSiisdopedwithP,itiscalledn-typedoping.DopingtheSiwillintroduce donorandacceptorlevelswithinthebandgapofSi( Figure4.1 ).Fromthere,electronsthat areintroducedbydonoratomscanbethermallyexcitedtotheconductionband(n-type) orelectronsinthevalencebandcanhoptowardstheacceptoratomsitesleavingbehind holes(p-type).Sincedopingaltersthecarrierconcentrationsofelectronsandholes,the formulationin Equation4.8 nolongerholds.Rather,theproductoftheelectronandhole concentrationisd[ 172 , 173 ].Thus,fordopedSi,therelationin Equation4.8 canbe rewrittenas: n 2 i = np = N CB N VB exp E gap k B T : (4.9) Asidefromtheconcentrationofthefreecarriersinthesemiconductor,theconductivityof thesemiconductorisalsodeterminedbywhatiscalledthemobility, ,ofthechargecarriers. Themobilityeshoweasilychargecarriersmovingthroughthecrystalareby anelectric[ 172 , 173 ].Supposeachargecarrierisplacedinanelectric ~ E .Asthe carriermovesthroughthecrystal,everycollisionwillresultinenergylossessuchthatthe averagevelocityofthecarrierreachesasteady-statevalueknownasthedriftvelocity, ~v d . Thedriftvelocityandelectricarerelatedbythefollowingrelation: ~v d = ~ E (4.10) Thedriftvelocitycanberelatedtotheemassofthecarrier, m ,andthescattering time, ˝ .Therefore,themobilitycanbewrittenasthefollowing: 89 = q˝ m (4.11) where q isthechargeofthecarrier.Thescatteringtime,andconsequentlythemobility,can bebychargecarrierscatteringeventsatcrystaldefects,dopantswithinthecrystal, andlatticephononsbasedonthetemperatureofthecrystal.Combiningthenitionsofthe carrierconcentrationandcarriermobilityresultsintheconductivityofthesemiconducting crystal: ˙ = q ( e n + h p )(4.12) 4.2SiliconSurfaces Siinbulkdisplaysadiamondcubiccrystalstructure.However,theatomsthatareonthe surfaceoftheSicrystalcanexhibitatstructure,knownasasurfacereconstruction. TheSidiamondcanbecleavedalongaspcrystalplanedirection,howeverindoingso, thesurfaceatomsoftenassumenewequilibriumpositions,creatinganewsurfacestructure. Additionally,thermaltreatmentofthesubstratecanresultinthesurfaceatomsadopting anewequilibriumsurfacereconstruction[ 95 , 96 ].OftheSisurfacesstudied throughoutthisthesis,Si(111)isthemostutilizedcrystalorientationforsmallorganic moleculargrowthevolutionstudies.Thisinterfacecandisplayvarioussurfacereconstructions bychangingthesubstrateannealingtemperatures,dopantconcentrationinthecrystal,or evenbyusingsurfacepassivationtechniques,thesubjectof Section4.2.1 , Section4.2.2 ,and Section4.2.3 .Thepropertiesofthevariousinterfacereconstructionsplayalargeroleinthe abilitytoformanorderedmolecularthinAsidefromtheSi(111)crystalorientation, 90 theSi(001)surfaceisalsoinvestigatedaspartofstudiesgearedtowardsdeterminingthin growthoftheorganicmolecules,sponhydrogenterminatedSisurfaces. 4.2.1TheSi(111) 7 7 SurfaceReconstruction InordertoprepareaSisubstrateforexperimentscommerciallyavailableSi(111)wafersare cuttotheappropriatesamplesizeforbeingintroducedintotheUHVsystem.Samples arethencleanedbyaseriesofsolventultrasonicationsintrichloroethylene,acetoneand2- propanol,followedbyavarietyofacidtreatmentsincludingpiranhasolution(1:3hydrogen peroxideandsulfuricacid,respectively),oxideetch(1:30),andstandardRCA1and RCA2waferpreparationprocedures.Attheendofthechemicaltreatmentprocess,athin surfaceoxideprotectionlayerisformedthatis1-2nmthick.Duringall exsitu preparation procedures,greatcaremustbetakenwhilehandlingtheSisamplessuchthattheyarenot contactedorcontaminatedbyanymetalimpurities,especiallyifthesampleisgoingtobe heatedafterwards.ThesamplescanthenbetransferredintotheUHVsystemandannealed insitu bydirectcurrentheatingat500 Covernighttoremovethesurfaceoxide.TheSi(111) 7 7surfacereconstructioniscompletedafterheatingandcoolingcyclesvaryingbetween roomtemperatureand1200 C. TheSi(111)7 7surfacereconstructionisthemostcommonlyobservedSi(111)surface reconstruction,oftenusedasacalibration/benchmarktestsampleforUHVstudies.Atypical LEEDcalibrationpatternoftheSi(111)7 7surfaceisshownin Figure4.2 (a)whichis inagreementwiththesimulatedpatternin Figure4.2 (b)forthissurfacereconstruction. Themeanfreepathoftheincidentelectronbeamis5-8 Aforthebeamenergiesusedin thisLEEDstudy,sothatboththebulkSi(111)andsurfacereconstructionfeaturescanbe observed[ 155 , 156 ].IftheSi(111)surfacemaintainedthesamestructureasthebulkcrystal, 91 Figure4.2: MeasuredLEEDpatternandsimulatedpatternonthe Si(111) 7 7 surface .(a)LEEDpattern( E beam =50eV)ofthereconstructedSi(111) 7 7surface(innerspots)andtheSi(111)1 1bulkstructure(outerhighintensityspots). (b)SimulatedLEEDpatternconsideringboththebulkSi(111)1 1structure(reciprocal spaceunitcelloutlinedinwhite)andtheSi(111)7 7surfacereconstruction(bluespots). thenthesurfaceisconsideredtohavea1 1surfacestructure.Thesurfacereconstruction isconsidered7 7becausetheunitvectorsofthenewreconstructedunitcellareseven timesthesizeofthe1 1unitvectors,asdisplayedin Figure4.2 .AfternumerousLEED, RHEEDandSTMstudies,the7 7surfacereconstructioncouldbeaccuratelymodeled asadimer-adatom-stackingfaultstructure.Thisparticularsurfacereconstructionextends throughthetopeatomiclayersoftheSisurface.ItconsistsoftwelveSiadatoms(ad- Si)thatareevenlydividedacrosstwotriangularsubunits,onefaultedandoneunfaulted. Alongthesidesofthesetriangularsubunits,dimersareformed,aswellasacornerhole thatextendsasdeepastheatomiclayer.AschematicmodeloftheSi7 7surface reconstructioncanbeeasilyfoundinthecurrentliterature[ 95 , 96 ],andisnotprovidedhere. Instead, Figure4.3 demonstratesatypicalsetofSTMtopographyimagesoftheSi(111) 7 7surfacereconstructionatandemptystateimagingconditions.Itshouldbe 92 Figure4.3: TypicalSTMtopographyimageofSi(111) 7 7.(a)STMtopography image( V s =-2.0V; I t =150pA)oftheSi(111)7 7surfaceed-state)obtainedat 77K.(b)STMtopographyimage( V s =+2.0V; I t =150pA)oftheSi(111)7 7surface (empty-state)obtainedat77K.Both(a)and(b)aretakeninthesamelocation.Thefaulted andunfaultedstructuresarehighlightedinboth(a)and(b). notedthattherearesomead-SivacanciesobservedintheSTMimages.Fromthe stateimagingcondition,thefaultedunitcellstructurecanbeclearlyidenbyt contrastforthetriangularsubunits,anditisalsoclearlymarkedintheimageaswell.Other dimer-adatom-stackingreconstructionsarepossibleontheSi(111)surface,howeverthe7 7 reconstructionistheenergeticallyfavorableinterfaceduetohowitbalancesthesurface stress,andthereforeitismorecommonlyobserved.Ofcourseitispossibletoobservesome oftheotherreconstructionsthroughthermaltreatmentmethodssuchasquenching,however thisisnotthefocusofthesestudies. OneofthemainchallengesinperforminganysortofthingrowthontheSi(111) 7 7reconstructionisduetothead-Siwhichterminatethesurface.Theseatomicsites leavesurfacedanglingbondswhicharehighlyreactiveandcaneasilyformcovalentbonds withmaterialthatisdepositedonthesurface.ThisisevidentintheSTScurveshownin Figure4.4 ,whichisanaverageofmultipleSTSspectratakenoverthecornerad-Siofthe 93 Figure4.4: NormalizedSTSspectratakenonthecornerad-Siintheunfaulted regionsoftheSi(111) 7 7 surface .Notethehighdensityofstatesavailableduetothe surfacereconstruction. unfaultedregionofthe7 7unitcell.IntheSTScurve,therearestrongenhancementsin thedensityofstateswithintheselectedbiasrangewhichindicatesaconsiderableamountof availablestatesforadsorbedatomsorspeciestointeractwithonceonthesurface.Forthis reason,atomsandmoleculesthataredepositedonthissurfaceinteractverystronglywith thead-Sisites,elycausingtheadsorbedspeciestobelocalizedontheadsorptionsite thatitlandson[ 36 , 131 ].Ultimately,thesedanglingbondsitesarethereasonthatthin growthonSisurfaces,aswellassemiconductorsingeneral,islimitedandconsiderably challenging. 94 4.2.2TheDeactivatedSi(111)-B p 3 p 3 R 30 SurfaceReconstruc- tion Inordertoavoidcomplicationsinthingrowthduetosurfacedanglingbonds,alter- nativemethodsforpreparingtheSisurfaceneededtobeexplored,withthegoalofeither deactivatingthesurfacedanglingbondsbydepletingthechargesfromthebondsitesorby passivatingthedanglingbondsitesdirectlywithasaatomicormolecularlayer.To date,inordertodirectlydeactivatethesurfacedanglingbonds,theonlyfeasibledirection thathasbeeninvestigatedthusfarinvolvestheuseofheavilyBdopedSi(111)substrates. TheSisamplesarepreparedfortransferintotheUHVsystemusingthesamechemical treatmentmethods.Similarly,thesampleisannealed insitu bydirectcurrentheatingat Figure4.5: SimulatedstructurediagramofthedeactivatedSi(111)-B p 3 p 3 R 30 surface .Theatomsareappropriatelycolorlabeledinthe. 95 500 Covernighttoremovethesurfaceoxide.Heatingcyclesat1200 Careusedtoestablish freshSisurfacefollowedbyextendedannealingat800 C.800 Cischosenbecause,atthis temperature,sub-surfaceBdopantsegregationcanoccurwhichallowsfortheBatomsto totheinterface.Atthiscondition,itisenergeticallyfavorableforBatomsubstitu- tionatSisitesneartheinterface.Thepresenceofsub-surfaceBatthesesitesallowsfor thetrivalentatomstobondwithfourneighboringSiatomsanddepleteelectronsfromthe surfacedanglingbondsofthead-Si.Whenthesub-surfaceBattheinterfaceis ˇ 1 = 3mono- layer,anatomicallysmoothanddeactivatedSi(111)-Bsurfaceisformed,whichdisplaysa p 3 p 3 R 30 surfacereconstruction[ 174 , 175 , 176 ]. AstructuralmodeloftheSi(111)-B p 3 p 3 R 30 surfacereconstructionisshownin Figure4.5 ,withtheatomsappropriatelycolorlabeled.Thissurfacereconstructionresults Figure4.6: STMtopographyimageofapristineregiononthedeactivatedSi(111)- B p 3 p 3 R 30 surface .STMtopographyimage( V s =+2.0V; I t =200pA)ofthe Si(111)-B p 3 p 3 R 30 surfacereconstructionobtainedat77K. inahexagonallatticestructurewhichisrmedbySTMtopographyimagingasshownin Figure4.6 .ThemeasuredandsimulatedLEEDpatternsin Figure4.7 arealsoinagreement 96 Figure4.7: MeasuredLEEDpatternandsimulatedpatternonthe deactivatedSisurface .(a)LEEDpattern( E beam =50eV)ofthereconstructedSi(111)- B p 3 p 3 R 30 surface(innerspots)andtheSi(111)1 1bulkstructure(outerspots). (b)SimulatedLEEDpatternconsideringboththebulk(reciprocalspaceunitcelloutlined inwhite)andsurfacereconstruction(reciprocalspaceunitcelloutlinedinblue). withtheexpectedsurfacereconstructionIn Figure4.6 ,thelocalregionbeing imagedispristine,displayingnoatomicdefects.Asaresult,allofthead-Siatomsappear thesamesincetheideal1 = 3monolayerofsub-surfaceBintheirpreferredatomicsiteshas beenachievedlocally.However,anumberofdefectscanoccuronthesurfaceduringsample processing.Ifasub-surfaceBatomisabsentattheinterface,anindividualdanglingbondsite willbeformed,asillustratedin Figure4.8 .Thedanglingbondsappearasbrightprotruding featuresintheSTMimagebecausetheunpairedelectroninducesalocalenhancementin availabledensityofstates.Inordertopreventadsorbedatomsormoleculesfrombeing localizedatthedanglingbonddefects,thedensityofthesedefectsiskepttoaminimum [ 36 ].Individualad-Sivacanciesaswellassub-surfaceBdefectscanalsooccur,howeverthey havealessdramaticonthemolecularadsorptionthandanglingbonddefects.These typesofdefectsoftenappearasdarkfeaturesintheSTMimages,asshownin Figure4.8 . 97 Figure4.8: STMimageofatomicdefectsthatarecommonlyobservedonthe Si(111)-Bterraces .STMtopographyimages( V s =+2.0V; I t =50pA)oftheSi(111)-B p 3 p 3 R 30 surfaceobtainedat77K.BrightfeaturescorrespondtoSidanglingbonds withtheabsenceofunderlyingboronatoms(greenarrow).Darkfeaturescorrespondto individualad-Sivacancies(redarrow)orsub-surfaceborondefects(bluearrow). ContrarytothosetakenontheSi(111)7 7surface,STScurvestakenonthedeac- tivatedad-SisitesintheSi(111)-B p 3 p 3 R 30 reconstruction(see Figure4.9 )display noavailablesurfacedensityofstateswithinthebandgap.Onlytheunoccupiedstate,at about1.5eVabovetheFermilevel,andanoccupiedback-bondstate,atabout1.5eVbelow theFermilevel,areobservedinthespectra,whichisconsistentwithpreviousstudies[ 174 ]. Theabsenceofdensityofstateswithinthebandgapindicatesthatthedanglingbondsites havebeendeactivatedandwillnolongerexhibitstrongcovalentbondingwithadsorbates, unlesssomeothermechanismisatplay.Thus,thedeactivatedSi(111)-B p 3 p 3 R 30 reconstructionprovidesanatomicallysmoothdeactivatedsemiconductingsurfacethatwill allowadsorbedmoleculestofullyexplorethesurfacepotentialenergylandscapeandinter- 98 Figure4.9: NormalizedSTSspectratakenonad-SisitesofthedeactivatedSi(111)- B p 3 p 3 R 30 surface .Notetheabsenceofanydensityofstateswithinthebadgap. ThenormalizedSTSspectratakenonthecornerad-SiintheunfaultedregionsoftheSi(111) 7 7surfaceisalsoincludedforreference. molecularinteractionsinordertoformorganizedstructures[ 174 , 175 ].Thisarobust templateforinvestigatingthemechanismsofmolecularthingrowth. 4.2.3TheHydrogenPassivatedSi(111)-HandSi(001)-HSurfaces Asalreadymentionedin Section4.2.2 ,analternativemethodtodeactivatingthesurface danglingbondsistopassivatethedanglingbondsitesdirectlywithaatomicor molecularlayer.ThismethodcanbebinscenarioswheredeactivatingSi(111)-B throughdopingisnotidealfortheapplicationsoftheSisurface.SurfacepassivationofSiis mostcommonlyachievedbyterminatingeveryad-Sionthesubstratewithhydrogenatoms, thusmakingeverydanglingbondsiteinert.Ofcourse,anotherterminationoptionisusing 99 Figure4.10: MorphologyofthehydrogenterminatedSi(111)surfaces .Atypical AFMtopographyimage(AdiamondcoatedSitipisusedwithspringconstant k =0 : 2N/m andresonancefrequency f =13kHz)oftheSi(111)-Hwhichhasbeenpreparedbythe roughetchingprocess(dashedredrectangle)isshownin(a).AtypicalAFMtopography image(AdiamondcoatedSitipisusedwithspringconstant k =0 : 2N/mandresonance frequency f =13kHz)oftheSi(111)-Hwhichhasbeenpreparedbythesmoothetching process(dashedbluerectangle)isshownin(b).Both(a)and(b)showtheAFMz-height responseaswellasthesimultaneousamplituderesponse. halidessuchaschlorineorHowever,thistendstomakethesurfacemorereactive. Therearetwowaysinwhichhydrogenterminationisperformed:crackinghydrogen gas(H 2 )byusingaheatedtinaUHVenvironmentanddirectingtheseparated hydrogenatomsattheSisurface[ 34 , 132 ];wetchemicalacidetchingtreatmentsoftheSi samples[ 35 , 133 , 134 , 177 ].WetchemicalprocessingoftheSisurfaceisutilizedinthis particularstudy.InitiallySisamplesarecleanedbyaseriesofsolventultrasonicationsin trichloroethylene,acetoneand2-propanol,similartotheotherSisurfaces.Thisisthen 100 Figure4.11: MeasuredLEEDpatternandsimulatedpatternonthe smoothSi(111)-Hsurface .(a)LEEDpattern( E beam =65eV)oftheSi(111)-H1 1surface.(b)SimulatedLEEDpatternconsideringthebulkSi(111)1 1interfaceat comparablebeamenergy. followedbyanextendeddipinpiranhasolution(1:3hydrogenperoxideandsulfuricacid, respectively)toremovecontaminatesandformauniformoxidelayeronthesurface.The stepinvolvesaseriesofacidetchinginvolvingoxideetch(1:30)orboth oxideetch(1:30)andammoniumdependingonthedesiredsurfacemorphology [ 177 ].Siacidetchingmustbeperformedinaninertatmosphere,suchasnitrogenorargon. Thisisbecausethepresenceofoxygenandwatercantheetchingprocess.The etchingprocessisalsoimprovediftheetchingsolutionispurgedofremnantoxygenand water,whichcanintothesolutionafterthebottleisopenedinambientconditions [ 177 ].AfteretchingtheSisampleusingpurgedsolutionsinaninertatmosphere,theSi surfacewillnowbehydrogenterminated.Dependingonthedesiredmorphology,theSi surfacewilleitherbecompletelyterminatedbyonlymonohydridesateveryad-Sisiteorby amixtureofmono-,di-,andeventrihydrides[ 177 ]. ThesurfacemorphologyofhydrogenterminatedSi(111)andSi(001)usingbothwet 101 Figure4.12: MorphologyofthehydrogenterminatedSi(001)surfaces .Atypical AFMtopographyimage(AdiamondcoatedSitipisusedwithspringconstant k =0 : 2N/m andresonancefrequency f =13kHz)oftheSi(001)-Hwhichhasbeenpreparedbythe roughetchingprocess(dashedredrectangle)isshownin(a).AtypicalAFMtopography image(AdiamondcoatedSitipisusedwithspringconstant k =0 : 2N/mandresonance frequency f =13kHz)oftheSi(001)-Hwhichhasbeenpreparedbythesmoothetching process(dashedbluerectangle)isshownin(b).Both(a)and(b)showtheAFMz-height responseaswellasthesimultaneousamplituderesponse. chemicaltreatmentmethodsneedstobeinvestigatedbyAFMimagingpriortotestingthin growthonthesurface.ItisnecessarytoperformtheAFMstudiesinasealednitrogen environmenttopreventthepossibilityofaircontaminationonthehydrogenterminated Sisurface,particularlyoxidationoftheinterface. Figure4.10 (a)displaysatypicalAFM topographyimageofSi(111)thathasbeenhydrogenpassivatedbyaseriesofdipsin oxideetch,while Figure4.10 (b)istakenonaSi(111)surfacethatishydrogenterminatedby aseriesofdipsinbothoxideetchandammoniumThereisastarkin 102 thesesurfacemorphologies,whichisstrictlyduetothedtetchingproceduresapplied toeachsample.Usingjusttheoxideetchsolutionresultsinnon-uniformetching, producingarough,small,granular-lookingsurfacemorphology.InthiscasetheSisurfaceis terminatedbyamixtureofmono-,di-,andtrihydrides.Itshouldbenotedthat,despitethis beingconsideredaroughsurface,theaveragecalculatedsurfaceroughnessfromtheAFM imageisonly2 : 0 0 : 2 A.Byetchingwithanadditionalammoniumdeprocess,a steady-statestepwetchingoftheSisurfaceresultsinsmoothSi-Hterraceswithevenly spacedsteps.Inthiscaseeveryad-Sisiteisterminatedbymonohydride[ 177 ].ALEED measurementtakenonthesmoothSi(111)-Halsothatduetotheclear peaksoftheSi(111)1 1surface,thesurfaceisatomicallysmooth,asshownin Figure4.11 . ThesamecomparisonhasbeencarriedoutonSi(001),asshownintheAFMtopography imagesin Figure4.12 .ForthisSicrystalorientation,whenetchingonlywithoxide etch,thesurfacedisplaysasimilarmorphologyasSi(111)ascanbeseenin Figure4.12 (a).However,whenammoniumisintroducedintotheetchingprocedureforSi(001) aroughsurfaceisproducedwithrelativelytallartifactsthatremainonthesurfaceafter etching,ascanbeseenin Figure4.12 (b).Itislikelythatbothetchingprocedureson theSi(001)surfacehydrogenterminatethead-Sisiteswithamixtureofmono-,di-,and trihydrides.ThesefourhydrogenterminatedSisurfacesareusedforthingrowthstudies inthisthesis. 4.3Silicon-on-InsulatorTechnology Astheelectronicsindustry,particularlythesemiconductorindustry,evolved,itbecamenec- essaryforthedevelopmentofimprovedelectronicswithincreasedfunctionality,complexity, 103 andperformance.Thisrequiredmorecomplexsubstratearchitecturesotherthanjustbulk wafermaterialsfordevicefabrication,whichleadtosilicon-on-insulator(SOI)technology. ASOIsubstrateconsistsofathin,single-crystallineSitemplate/devicelayerandathick, Sihandlewafer,bothofwhichareseparatedbyaburiedSiO 2 layer,oftenreferredtoas BOX.Thissubstratearchitecturebringswithitsomeadvantages,suchasfasterswitching Figure4.13: Schematicdiagramoftheseparationbyimplantationofoxygen methodforpreparingSOIwafers .TheBOXlayerisformedbyperforminghighdose oxygenimplantationfollowedbyhightemperatureannealingtoformtheSiO 2 layer.It shouldbenotedthatthismethoddoesnotproducehighqualitySOIwafers. duetoreductionsinparasiticcapacitancefromthepresenceoftheBOX,removalof whichlimitdevicesfrombeingincloseproximitytoeachother,andlimitingtheelectric betweencommonsource-draindevicesbyusingatlythintemplate/devicelayer [ 64 , 178 , 179 ].Itispossibletoetchthetemplate/devicelayeroftheSOIsubstratedown tothenanoscale,makingthistechnologyparticularlyusefulfortheeverdecreasingsize 104 Figure4.14: SchematicdiagramoftheSmart-CutmethodforpreparingSOI wafers .HydrogenimplantationisperformedonaSiwaferwithaSiO 2 surfacelayersuch thatahydrogenrichregionisformedinthewafer.ThiswaferisthenbondedtoanotherSi waferwhichactsasthehandlewafer.Thebondedwafersarethenannealedcausingthem toseparatealongthehydrogenrichregionleavingbehindanSOIwaferandabareSiwafer. demandsofthesemiconductorindustry[ 180 ]. SOIsubstratesaretypicallypreparedinoneoftwoways:SeparationbyImplantationof Oxygen(SIMOX);andtheSmart-Cutprocess[ 178 ].Asthenamesuggests,intheSIMOX method,theBOXlayerisformedbyperforminghighdoseoxygenimplantation.During theprocess,theSicrystalismaintainedatelevatedtemperaturestoreducethedamagethe oxygendosingprocesscaninduceonthecrystal.Thelaststeprequiresahightemperature annealingprocesstoformtheSiO 2 andrecovertheSicrystalstructure(see Figure4.13 ). Smart-Cut,ontheotherhand,astrategyforformingawaferofhigherqualitythan theSIMOXmethod.ThisisdonebytakingaSiwaferwhichhasaSiO 2 surfacelayerand 105 performinghydrogenimplantationsuchthatahydrogenrichzoneisformedinthewafer. ThiswaferisthenbondedtoanotherSiwaferwhichactsasthehandlewafer.Thebonded wafersarethenannealed,causingthetwowaferstoseparatealongthehydrogenrichregion leavingbehindanSOIwaferandabareSiwafer.TheSOIwafercanthenbepolishedusing standardmethodstoproduceawaferwithsimilarsurfacequalityasabulkSiwafer(see Figure4.14 ).Fromherethetemplate/devicelayercanbeetchedtothedesiredthicknessby aseriesofthermaloxideetchingtreatmentsandwetchemicaletching. 4.3.1SiNanomembranes TheSinanomembraneisasingle-crystalline,two-dimensionalnanostructurewithalarge surface-area-to-volumeratio.Thisservesasasystemtodiscusssurfaceandinterfaceef- fectsonthetransportpropertiesofasemiconductingnanostructure.Sinanomembrane isaversatilenanostructurewhichexhibitsahighdegreeofyandconformational properties.Becauseofthis,itawidevarietyofapplicationsinelectronics [ 78 , 79 , 80 , 81 , 82 , 83 , 84 , 85 , 86 ].TheSitemplate/devicelayerofSOIsubstratescanbe thinneddowntonanometerthicknesses,makingthetopSilayeraSinanomembranethat restsontopoftheBOX.Itispossibletoreleasenanomembranesofvarioussizesandshapes fromtheSOIsubstratebycombiningphotolithographystrategieswithcontrolledchemical etchingoftheBOX.Ifsuccessful,theresultisafree-standingmembranewhichcanbetrans- ferredtoothersurfaces[ 84 , 85 , 180 ].Thismembranetechnologyhasalsobeenextendedto othermaterialsforvarioustpurposes.Forexample,membranesconsistingofTiand Gehavealreadydisplayedtheabilitytostorelithiumionswithhighcapacity,whichcould bepotentiallyusedastheanodematerialinlithiumbatteries[ 86 ]. OneinterestingaspectoftheSinanomembraneisitselectricalconductivity.Consider 106 anSOIsubstratewherethetemplate/devicelayerhasbeennominallydopedto10 15 cm 3 . Ifthetemplate/devicelayerisnowthinneddowntobe ˇ 20nm(comparabletowhatis usedformembranestudiesinthisthesis),thenthenumberofdopantsin1cm 2 is ˇ 2 10 9 . Sincethenanomembraneisontopofanoxidelayer,thisSiO 2 canplayatrolein theconductivityofSinanomembranesbecauseofinterfacestatesthatformattheSi-SiO 2 interface.Theseinterfacestatescantrapcharges,causingthemtobedepletedfromthe Sinanomembrane[ 173 ].TypicallythenumberofinterfacestatesthatformattheSi-SiO 2 interfaceisontheorderof ˇ 10 11 .Thus,forthinenoughnanomembranes,allcarriersin themembranewillbetrappedattheSi-SiO 2 interface,elymakingtheminsulatorsat roomtemperature.However,inarecentstudyithadbeenshownthat,byreconstructingthe Figure4.15: Exampleschematicdiagramofthesurfacetransferdopingmechanism inducedbyclosely-lyingmolecularorbitalstotheSinanomembranebandedges . InthisspcaseamoleculehasbeenselectedwiththeLUMOclosetotheSivalence bandedge,allowingforthermalexcitationofmobilecarriersbetweenthemembraneand molecularspeciesonthesurface. topSinanomembraneinterfaceinanUHVenvironment,surfacebandscanbeintroducedat theinterfacewithoneofthembeingcloseinenergytothevalenceorconductionbandedges 107 Figure4.16: Schematicdiagramsofthetpossiblemethodsforintroducing aninterfacialdipolelayeronthesurfaceofSinanomembranes . oftheSinanomembrane.Thisallowsforthermalexcitationofonecarriertypefromthe membranetothesurfacebandsatroomtemperature,leavingbehindtheoppositecarrier intheclosely-lyingbandedgeoftheSinanomembrane[ 87 , 88 ].Thisisaprocessknown as\surfacetransferdoping,"amechanismwhichcanmodifytheecarrierdensity, makingitpossibletoinducemobilecarriersinthemembranetoenhancetheconductivity [ 89 ].The\surfacetransferdoping"mechanismisnotjustlimitedtousingreconstructed surfaces,butcouldalsobeachievedbyintroducingclosely-lyingmolecularorbitalstotheSi nanomembranebandedges[ 181 , 182 ].Selectingmoleculeswithahighestoccupiedmolecular orbital(HOMO)orlowestunoccupiedmolecularorbital(LUMO)closetotheSiconduction orvalencebandedgescouldalsoallowforthermalexcitationofmobilecarriersbetweenthe membraneandmolecularspeciesonthesurface(see Figure4.15 ).Organicmoleculescanbe easilyfunctionalizedthroughsyntheticchemistrymethodstotunetheHOMOandLUMO levelalignmentwiththebandedgesoftheSinanomembrane. AnothermethodforpossiblymodifyingtheelectricalconductivityofSinanomembranes 108 isthroughtheformationofaninterfacialdipolelayerbyusingorganicsmallmolecules. Thesedipolescanbeusedtoelytunetheelectrostaticpotentialatthesubstrate surface[ 183 ].SincetlythinSinanomembranesarehighlysensitivetothesurface condition,moofelectrostaticpotentialattheinterfacecanthecarrier distributionand,consequentlythetransportcharacteristicsofthemembrane.Aninterfacial dipolelayercanbeformedthroughthreetmethods:inducingchargetransferat theinterfacebyphysisorptionofamolecularspecies;physisorptionofamolecularspecies withaninherentdipolemoment;inducingachemicalinteractionbetweenthesurfaceanda molecularspecieswhichischemicallyboundtothesurface(see Figure4.16 ).Asdiscussed in Section2.5 ,smallorganicmoleculescanbeusedtoformclosedpackedmolecularthin onthesurface.Byforminganorderedmolecularthinutilizingmoleculesthatcan generateaninterfacialdipolelayerthroughoneofthesethreemethodstheopportunity topreciselycontroltheconductivityofSinanomembranesbysurfacechemistry[ 9 , 184 ]. Moreover,sincesmallorganicmoleculesareextendedobjectswithahighnumberinternal degreesoffreedom,theycaneitheradoptaorastanding-uprightorientation, whichcanbecontrolledbytemplatedmoleculargrowth[ 10 ].Manipulationofthemolecular orientationcouldprovideanothermethodfortuningtheinterfacialelectronicstructureand chargetransfer.Ultimately,thiscanthecarrierdensityandconductivityintheSi nanomembrane. 4.3.2FabricationofSiNanomembranes InordertofabricateanisolatedSinanomembranedeviceforelectricalstudies,amulti- stepprocessisrequired.Thiswillbedetailedhere.First,thetemplate/devicelayerof aSOIwaferisetchedto ˇ 20nmbyaseriesofthermaloxideetchingtreatmentsand 109 wetchemicaletching.AthickphotoresistprotectionlayerisspunontopoftheSOIand itiscuttotheappropriatesamplesize.Afterwards,theSOIsamplesareultrasonicated insolventstoremovetheoriginalphotoresistlayerandproduceacleandevicelayer.A thinuniformphotoresistlayeristhenspunontopofthedevicelayerandheattreatedfor photolithographyprocessing.AsquaremesaphotoresistpatternisontheSOIsample byexposingcertainregionstoultravioletlightandremovingtheexposedphotoresistusing adevelopercompound.Thequalityofthephotoresistischeckedbyanopticalmicroscope usingalightsourcetoensurethattherearenoholesinthephotoresistwhich thesquareSinanomembranedevice.Sampleswithawsquaremesaarethen placedinareactiveionetchplasmasystem,commonlyreferredtoasanRIEplasmaetcher, commerciallyavailablefromNordsonMARCH.UsingSF 6 intheplasmaallowsforSitobe etchedawaywhileleavingSiO 2 largelyThesquarephotoresistmesaprotects thedeviceregionfromtheplasmaetchingprocess.Siinthesurroundingareasisremoved leavinganisolatedSinanomembranedeviceontopoftheSiO 2 layer,asshownin Figure4.17 (a)and(b).Thedevicesampleisthenultrasonicatedinsolventstoremovethephotoresist layer,leavingaclean,wdevice.Atthispointitispossibletoperformfurther chemicalprocessinginordertohydrogenterminatethenanomembranesurface,asdescribed in Section4.2.3 ,anddepositamolecularthinAftercompletionofallnanomembrane preparationprocedures,transportmeasurementscanbeperformedbywiringthedevicesas shownin Figure4.17 (c).IndiumsolderisusedtoconnectthewirestotheSinanomembrane surface.Electrodes1through4areusedforapplyingthecurrentandmeasuringthevoltage responseofthenanomembrane.Electrode5isusedeitherasawaytoback-gatetheSi nanomembranefromthehandlewaferoritisgrounded. 110 Figure4.17: ImagesofafabricatedSinanomembranedevice .(a)showsacompletely fabricatedSinanomembranedevicebeforewiringconnectionshavebeenmade.Forclarity, asideschematicdiagramoftheSinanomembranedeviceisincludedin(b).(c) displaysthewiringnecessaryforperformingtransportmeasurementsonthe Sinanomembranedevices. 4.4MetalPhthalocyanineMolecules Thisthesisfocusesontheuseofmetalphthalocyanine( M Pc)moleculesfororganicthin studies. M Pcisawell-knownmetal-organiccoordinationcomplexwhichconsistsofacentral transition-metal(TM)ioncoordinatedviafourbenzene-pyrroleorganicligandgroups(Pc), asshownin Figure4.18 ,toformthecompletemolecule.ThediscoveryofpurePc moleculesoccurredintheearly1900sasabyproductofcolorreactions,producingablue- coloredpowder.TheeventualTMioncoordinationabilityhappenedtobediscoveredby accidentinadyefactory.Acolorreactionvesselbroke,whichexposedthepurePcmolecules tothepipinginthefactory,producingaslightcolorchangeinthepowder.Thisresulted inimmediateproductionof M Pcmoleculesforuseindyesonanindustrialscale[ 185 , 186 ]. 111 Researchintothevarioussynthesisstrategiesof M Pc,thetmetalioncomplexesthat couldbeformed,andadditionalfunctionalizationmethodssuchascontinuedfor sometimeinantoextendtheapplicationsofthisdye. Figure4.18: Simulatedimageofafree-standing M Pcmoleculeshowingthecom- pleteatomicstructure .Atomsareappropriatelycolorlabeledinthe Today,amultitudeoftmetalionsthatcanbecoordinatedwithpurePcmolecules havebeendiscovered,makingitpossiblefor M Pcmoleculestodisplayawidevarietyofmag- neticandelectronicproperties.Additionally,thesemoleculesexhibitvarioustraitsthathave promotedtinterestintheirpossibletechnologicalapplications.Forexample, M Pc moleculesarethermallystableatrelativelyhightemperatures(500 600 Corhigherunder vacuumconditions[ 120 ]).Theyexhibitanextended ˇ -electronsystem,whichcanresult in ˇ - ˇ stackingbetweenmolecules,formingorganicstructureswithapreferredorientation. M Pcmoleculesintheirbulkformalsodisplaysemiconductingproperties[ 185 , 186 ]. Thevarietyoftechnologicalapplications,coupledwith M Pcmolecules'simplisticsquare- 112 planargeometry,havemade M Pconeofthemostfrequentlystudiedmoleculesinorganic molecularthingrowth,withvariousstudiesalreadyperformedoninsulating[ 27 , 28 , 29 , 30 , 31 , 32 , 33 ],semiconducting[ 34 , 35 , 36 , 37 , 38 , 39 , 40 , 41 , 42 , 43 ],andmorecommonly, metallicsurfaces[ 23 , 44 , 45 , 46 , 47 , 48 , 49 , 50 , 51 , 52 , 53 , 54 , 55 , 56 , 57 , 58 , 59 , 60 , 61 ]. Theuseofthesemoleculesinthinisalreadybeginningtoemergeintoday'selectronics market,applicationsinorganiclight-emittingdiodes(OLEDs),organic transistors(OFETs),andorganicphotovoltaics(OPVs)[ 17 , 18 , 19 , 20 ].Despitesomeearly successesoftheuseof M Pcmoleculesinthesedevicearchitectures,therearemanyfactors relatedtothemoleculargrowthandorderingof M Pcthinthatneedexplorationto advancetheuseofthesemoleculesintheelectronicsindustry.Therearefewstudiesof M Pc moleculesgrownonSisurfaces,whichissurprisingconsideringthefactthatSiisoneof themostwidelyusedmaterialsinthemicro-andnano-electronicsmarket.Asdiscussed in Section4.2 ,thisislikelybecausetheSisurfaceisconsiderednotoriouslyto formanorganicmolecularthinonduetothesurfacedanglingbonds[ 36 ].However, usingsurfacepassivationanddeactivationmethodsenablesstudiesofthegrowthdynamics, molecularordering,andmolecularorientationof M PcmoleculesonSisurfaces,whichisthe subjectofthisthesis. Forthepurposesofthestudiescontainedwithinthisthesisthevarious M Pcmolecules usedarepreparedinthefollowingwayformolecularthingrowth.Commerciallyavailable highpurity( ˇ 90%purityorhigher)ZnPc,CuPc,CoPc,FePc,F 16 ZnPc,andClAlPc powdersfromSigmaAldrichcanbepurchasedforuse,however,additionalstepsarerequired beforeuse.Afterreceivingthe M Pcpowder,thecontaineriswrappedinfoiltopreventlight exposureandarestoredinanitrogenorvacuumenvironmenttoavoidanypossibilityof thematerialaginginambientconditions.Whenreadytostartusingthemolecules,itis 113 necessarytoperformadditionalsublimationprocessestoensureonlypure M Pc areusedwithoutanyriskofcontaminates,especiallyifthepurityofthepowderiscloserto thelowerpuritylimitof ˇ 90%.Thisisdonebyplacingthemoleculesinaquartzevaporation boatwhichisplacedinsideaquartztubethatisvacuumsealedandpumpedonbyaturbo pump.Therearethreequartzsleevesinsidethetubewhichareplacedatthethreedt temperaturecontrolzonesalongtheentirequartztube.Theexitporttotheturbopumpis protectedbyquartzwooltoallowforreliablepumpingwhilepreventingmoleculesfromgoing directlytotheturbopumpduringthesublimationprocess.Thetemperatureofthethree theatingzonesareslowlyrampeduptoproduceatemperaturegradient.Typically, forthe M Pcmolecules,theheatingzonewiththemoleculeevaporationboatissetbetween 430 500 C.Thezonenexttothisissettoaslightlylowertemperature(350 C)sothatthe puresublimatedmoleculeswillcollectandcondensebacktopowderformonthequartzsleeve forthiszone,resultinginthehighlymoleculesthatareusedforexperimentation. Athirdheatingzoneissettoanevenlowertemperature(300 C)toensurethatthepure moleculescollectintheappropriateregionwithinthetube.Alongthisheatingzone,aswell astowardstheturbopump,contaminatesanddefectedmoleculesintheoriginalpowderare collected.Aftercompletionofthesublimationprocess,thehigh-puritymoleculesareeither collectedforadditionalsublimationcyclestoensuretheuseofhighqualitymaterialorthey aredirectlyutilizedinmolecularthinstudies,ifconsideredpureenough. Thehighpurity M Pcmolecularpowderisstoredinanitrogenorvacuumenvironment untilitisreadytobeinstalledintooneoftheevaporationsystemsusedthroughoutthis dissertation.Forgrowthevolutionstudiesof M PcmoleculesonthedeactivatedSi(111)-B surface,themolecularpowdersareplacedintoboronnitridecruiciblesthataretheninserted intorespective,low-temperaturecellsthatareconnectedtoaUHVsystem.The 114 molecularpowderisbakedat140 CforthreedaysduringtheUHVsystembakingprocess. Uponcompletionofthebakingprocess,theorganicevaporatorsarefurtherdegassedby cyclingtheevaporatortemperaturemultipletimesabovethemoleculardepositionoperating temperature.Afterthis,depositionofthe M Pcmoleculescanbeperformed.Thedeposition ratecanbeeasilycontrolledbyadjustingthetemperatureoftheorganicevaporatorandmon- itoringthepressurechangesintheUHVchamber,typicallymaintainedat ˇ 1 : 0 10 9 mbar duringthedepositionprocess.Whenusingverylowdepositionratesitisnecessarytocal- ibratetheexpectedmolecularcoverageusingscanningprobemicroscopytechniques.For studiesonhydrogenterminatedSisurfacesandSinanomembranedevices,thehighpurity M Pcpowdercanbedirectlyinsertedintoathermalevaporationboat,whichistheninstalled intoacommerciallyavailablephysicalvapordepositionsystemfromAngstromEngineering. Theboatisthenheatedtoevaporatethe M Pcmoleculesontothesamplesurface.Priorto evaporationontothesurfaceofinterest,depositioncalibrationtestsareperformedtoensure theappropriatethicknessisdepositedontothesamplesurface.Uponcompletionof thecalibrationteststheevaporationprocesscanbeautomatedusingthecontrollersonthe system. 115 Chapter5 AnisotropicStep-FlowGrowthofZinc PhthalocyanineontheDeactivated Si(111)-BSurface InthischaptertheinitialgrowthofZnPconthedeactivatedSi(111)-Bsurfaceissystem- aticallystudiedbySTMfromsub-monolayertotheformationofacompletemonolayer.In Section5.1 ,bycombiningSTM,STS,andtialconductancemapping,thepacking motifofZnPccanbeidenFurthermore,theobservedmoirepatternforZnPcdomains canbeusedasanindicatoroftheepitaxialrelationship.Thepackingdeter- minedfromtheSTMexperimentiscomparedtomolecularorbitalsimulations,geometric calculations,andsimulatedmoirepatternstoassessthemolecularassignment. Section5.2 discussesSTMinvestigationsofthegrowthevolutionofZnPc.TheZnPcmoleculesare abletorapidlyontheSisurfaceevenatroomtemperatureandnucleateatSistep edgesfollowedbytheformationofhighly-orderedanisotropicstructuresacrossSiterraces, i.e.anisotropicwgrowth.Thewgrowthmodereducestheallowedsymmetry ofthemoleculardomainssuchthatonlytwodominantmirror-rdomainsarepresent onthesurface.Thoughtheformationofahighly-orderedorganicmolecularmonolayeris possible,defectsonthedeactivatedSisurfaceandadditionalactivationbarrierscaninter- 116 rupttheanisotropicwgrowth,asdetailedin Section5.3 .Theseresultsareconcluded inthelastsection.PartsofthischapterareadaptedfromReferences[ 39 ]S.R.Wagner et al ., Phys.Rev.Lett. 110 ,086107(2013)and[ 42 ]S.R.Wagner etal ., Surf.Sci. 630 ,22 (2014). 5.1ZnPcMolecularPackingandEpi- taxialRelationonDeactivatedSi(111)-B AfterappropriatethermaltreatmenttoformtheSi(111)-B p 3 p 3 R 30 surfacerecon- struction(asdetailedin Section4.2.2 ),ZnPcweregrownbythermalevaporationata tlylowdepositionrateviathemethoddescribedin Section4.4 .Thisdeactivated andatomicallySisurfaceprovidesauniquesubstratetoformlarge-scalehighly-ordered organicthinasdemonstratedin Figure5.1 whereZnPcmoleculescrystallizeina highly-orderedclose-packedAsidefromthedrasticchangeintheperiodic structurefromthesurface p 3 p 3reconstruction,anadditionaltwo-dimensionalpattern canbeobservedasthebrightfeatures(greenlines)in Figure5.1 .Theobservationofthis superstructuresuggeststhatthemolecularoverlayerisnotcommensuratewiththesubstrate becauseitisonlywhenspmoleculescomeclosetoregisteringwithad-Sithatanen- hancedelectronicdensityofstates[ 165 ]occurs,givingrisetotheperiodicsuperstructure observedintheSTMimage.Theperiodicityofthepatternis8 : 3 0 : 2nmalongthe h 11 2 i directionoftheSisubstrateand5 : 0 0 : 2nmalongthedirectionofthesuperstructure,with anangleof104 2 betweenthetwo.Frompurelyageometricalperspective,thelattice mismatchbetweentheZnPcmolecularoverlayerandtheunderlyingsurfacecanresultina moirepattern(asdiscussedin Section3.3.1.2 )dependingontheepitaxialrelationship. 117 Figure5.1: ZnPcpackingandenhancedmoiresignaturefromSTM . AtypicalSTMtopographyimage( V s =+2.0V; I t =35pA)oftheZnPconthe deactivatedSi(111)-Bsurfaceobtainedat77K.Unitcellofthemolecularoverlayergivenby: b 1 =12 : 3 0 : 1 A, b 2 =6 : 7 0 : 1 A,and =92 1 .Greenlinesmarkbrightfringesofthe moirepattern.Theperiodicityofthepatternis8 : 3 0 : 2nmalongthe h 11 2 i directionof theSisubstrateand5 : 0 0 : 2nmalongthemoiredirectionwithangleof104 2 between thetwo,highlightedbyblackarrows. Inordertodeterminethenatureofthemolecularpackingandtheepitaxial relationbetweenthemolecularoverlayerandtheSisubstrate,STMexperimentsarecarried outonaSisurfacethatispartiallycoveredbyZnPcstructures.First,STSspectraare takenonboththeSisurfaceandtheZnPcmoleculestoidentifytheappropriatesample biastouseforimagingthemolecules. Figure5.2 showsthespectratakenonthead-Siand ontheorderedZnPcmolecules,respectively.Asalreadydiscussedin, Section4.2.2 ,theSi spectrumrevealsthatthesurfacehasbeendeactivatedwiththeunoccupiedstateatabout 1.5eVabovetheFermilevelandanoccupiedback-bondstateatabout1.5eVbelowthe Fermilevel,consistentwithpreviousstudies[ 174 ].SpectrumonZnPcshowstheavailable densityofstatesnear+2.5eV,correspondingtotheresonanttunnelingthroughtheLUMO 118 Figure5.2: DensityofstatesofZnPconthedeactivatedSi(111)-Bsurface .Normal- izedSTSspectratakenonthedeactivatedad-Sisites(red)andZnPcmolecules(blue)over thesameareaasshownin Figure5.3 .ThepeakintheZnPcspectranear+2.5V(resonant tunnelingtotheLUMO)isusedforthesimultaneousimagingoftopologyandtial conductancemapin Figure5.3 . [ 187 , 188 ]. DuetotheconvolutionbetweentopographyanddensityofstatesinherentinSTMimag- ing,additionalmeasurementsarenecessaryforappropriateassignmentofthemolecular Todistinguishbetweenthetwo,simultaneousdtialconductancemap- pingandtopographymeasurementsneededtobeperformed.In Figure5.3 (a)and(b),the comparisonbetweenthetopographyimageandtialconductancemapattheresonant tunnelingconditionshowsthatonlytwoofthebenzene-pyrroleringsoftheZnPcstructure areresolvedoneachmolecule,asrepresentedbytheblacklines.ThisimpliesthattheZnPc moleculesadopttiltedmolecularorientationwheretheapparentheightofthemoleculesis 119 Figure5.3: Simultaneoustopographyandtialconductancemappingof ZnPconSi(111)-B .(a)PartialimageshowingboththeZnPcoverlayerandtheSisurface ( V s =+2.5V; I t =50pA)obtainedat77K.(b)tialconductancemaptakensimulta- neouslyatthesameimagingconditions.(c)AtomicmodelofaZnPcdimerinobservedtilted orientationandLUMOoffreeZnPcmoleculescalculatedusingdensityfunctionaltheoryand thePerdew-Wanggeneralizedgradientapproximation.Incorrespondencewiththesimulated packing,blacklinesrepresentingindividualmoleculesareassigned.Theedge-to-edgedis- tancebetweenthetwobenzene-pyrroleringsismeasuredtobe13 : 0 0 : 2 A. representsthe anglebetween a 1 and b 1 ,theazimuthalrelationbetweenthemolecularoverlayer andthesubstrate. d =6 : 0 0 : 4 Adeterminedfromhistogramandlineanalysisofthez-piezoresponsein theSTMimages(see Figure5.4 foratypicallineAFMmeasurementsoftheZnPc arealsoinagreementwiththeapparentheightdeterminedfromSTM.Thus,fromthe apparentheightmeasurementstheZnPcmoleculesaretilted ˘ 30 abovetheSisurface. TobecertainthattheassessmentofthetialconductancemaptakenattheLUMO conditionisreasonable,thecalculatedLUMOoffree-standingZnPcmoleculesinthistilted orientationisdeterminedusingdensityfunctionaltheoryforcomparison( Figure5.3 (c)). Thecalculatedresultsdisplayagreementbetweentheelectrondensitydistributionofmolec- ularorbitalsobservedinthetialconductancemapandallowsfortheappropriate assignmentofindividualZnPcmolecules,whichareindicatedbytheblackbarsin Figure5.3 120 Figure5.4: TheapparentheightdeterminedfromSTMofZnPconthedeactivated Si(111)-Bsurface .(a)STMtopographyimage( V s =+2.0V; I t =20pA)showingaregion partiallycoveredwithZnPcontheSi(111)-Bsurface.Theredlinein(a)correspondstothe regionwherethelinepwastakenin(b).(b)TypicallinetakenontheZnPc TheaverageapparentheightfrommultiplelineandhistogramsfromSTMis d =6 : 0 0 : 4 A. (a)and(b).Thus,theunitcelloftheZnPcmolecularoverlayercanbeassignedbythe followingunitvectors: b 1 =12 : 3 0 : 1 A, b 2 =6 : 7 0 : 1 A,and =92 1 asshownin Figure5.1 and Figure5.3 (a). Withthecorrectmolecularassignment,theepitaxialrelationbetweentheZnPcoverlayer andtheSisubstratecanbedetermined.Asillustratedin Figure5.3 (a),theunitcellof themolecularoverlayerisrotatedby27 2 fromthe h 11 2 i directionofSi(111)-Bsurface. Furthermore,boththelateralunitcellandthe d -spacingofthedepositedZnPcoverlayers onthedeactivatedSisurfacearedistortedfromthebulkphase[ 125 ].Thiswell-d azimuthalorientationisthekeycharacteristicofthenon-commensurateepitaxialgrowth andthereasonfortheobservedmoirepattern. ThemoirepatternandtheepitaxialrelationshipoftheZnPcmolecularoverlayeron 121 Figure5.5: ZnPcepitaxialrelationshipandsimulatedmoirepattern .(a)Resultsof geometricanalysisasthemolecularoverlayerisrotatedazimuthallyontheSi(111)-Bsurface. Firstminimumin V=V 0 (=0.7)occursat28 0 : 5 withotherminimaspacedinincrements of60 duetothesubstrate C 6 symmetry.(b)Simulatedoverlayer-surfacegeometryatan azimuthalrotationangleof28 .Redcirclesoutlinedinbluerepresentad-Si.Blackdots representtheZnPcoverlayerlatticedeterminedfromSTM.Greenlinesmarkthebright fringesofthemoirepattern.Theperiodicityofthepatternis8 : 5 0 : 1nmalongthe h 11 2 i directionoftheSisubstrateand4 : 8 0 : 1nmalongthemoiredirectionwithangleof104 1 betweenthetwo,highlightedbywhitearrowsforcomparisonwith Figure5.1 . theSisubstratecanbequantitativelyinterpretedbycalculationsusingthegeometricphase coherencemodel(discussedin Section3.3.1.1 )[ 165 ].Asshownin Figure5.5 (a),anazimuthal rotationof28 0 : 5 givesaminimumvalueof V=V 0 ˘ 0 : 7,whichindicatesanepitaxial relationbetweenincommensurateandPOLcoincidence.Theinitialnucleationeventsare guidedbytheequilibriumminimumlandscapewhichappearsnearlyPOLcoincidentforsmall domainsbutbecomesmoreincommensurateforcompletelayers.Thisisfurthercorroborated bythecalculatedstructuretransformationmatrixusingthesubstratelatticeparameters, molecularoverlayerlatticeparameters,andtheazimuthalanglegivenbythefollowing: 122 2 6 4 b 1 sin ( ) a 1 sin ( ) b 1 sin ( ) a 2 sin ( ) b 2 sin ( ) a 1 sin ( ) b 2 sin ( + ) a 2 sin ( ) 3 7 5 = 2 6 4 1 : 13 0 : 021 : 00 0 : 02 1 : 01 0 : 031 : 01 0 : 03 3 7 5 (5.1) wherethematrixelementsintherightcolumnindicatethatPOLcoincidenceiswithinthe acceptableuncertaintyrange(referto Section3.3.1 ).Otherminimainthemodeloccurat 60 intervalsconsistentwiththesubstrate C 6 symmetry.Usingthecalculatedazimuthal angleandtheexperimentallydeterminedlatticeparameters,asimulatedimageoftheZnPc molecularoverlayeronthedeactivatedSisurface( Figure5.5 (b))canbegenerated.Analysis ofthesimulatedandSTMmoirepatterns(whicharebothhighlysensitivetothisazimuthal relationship)displayexcellentagreement,withtheperiodicityofthefringepattern(green linesin Figure5.1 and Figure5.5 (b))andtheangleofthemoirefringeswithrespectto the h 11 2 i directionoftheSi(111)-Bsurfacearebothwithinthemeasurementuncertainty (Thesimulatedperiodicityofthemoirepatternis8 : 5 0 : 1nmalongthe h 11 2 i directionof theSisubstrateand4 : 8 0 : 1nmalongthemoiredirectionwithangleof104 1 between thetwo).Theobservedmoirepatternsforeachdomainareuniquetothespazimuthal registrationofthemoleculardomain.Thus,themoirepatternhastheadditionalbof actingasanindicatorofthein-planeazimuthalorderingofeachdomainallowingfort ZnPcdomainstobeeasilydistinguished. 5.2ObservationsandDiscussionoftheAnisotropicStep- FlowGrowthMode Asdiscussedin Section2.4 ,wgrowthisgenerallyobservedininorganicepitaxial growthwhenthenucleationofthedepositedspeciesisenergeticallyfavoredatstepedgesand 123 thelengthislongenoughtoenablestepedgesitesamplingpriortoaccumulation ofthespeciesontheterrace[ 108 , 109 ].InthecaseofZnPc,themoleculesdemonstrate Figure5.6: AnisotropicwgrowthofZnPconSi(111)-B .STMtopography image( V s =+2.0V; I t =35pA)ofsub-monolayerZnPccoverageonthedeactivatedSi surfaceobtainedat77K.Duringthedepositionprocessthesubstratewasheldat25 C.The formationofparallelstripesatmultiplenucleationsitesalongthestepedgescanbeclearly observed. evidenceofthewgrowthmodestartingfromthemonolayeratroomtemperature ( Figure5.6 ).TheZnPcmoleculesrapidlyontheSi(111)-Bsurfaceandpreferentially nucleateattheSistepedges,aphenomenonuniformlyobservedacrossthesubstrate[ 39 , 40 , 42 ].ThelowdefectdensityoftheSi(111)-Bsurfaceelysuppressestheisland nucleationonterracesandallowsmoleculestofullyexplorethesurfaceatroomtemperature tothesesitesatthesteps.However,theanisotropyofthisgrowthdistinguishesit frominorganicwgrowth.Oncethestepedgesitesareoccupied,stripestructures expandacrosstheSiterracesuntiltheyreachthenextSistepedge(withoutthepresence ofadditionalnucleationontheSiterrace).Theanisotropicgrowthofstripestructurescan 124 beattributedto ˇ - ˇ stackingbetweenmolecules.AfterreachingthenextSistepedge,the ZnPcstripesproceedtogrowinthein-planeperpendiculardirectionandwidenuntilthe stripescoalescetoformacompletemonolayer.Itisimportanttonotethattheanisotropic wgrowthmodeisobservedatorneartheequilibriumgrowthcondition,i.e.theZnPc moleculesaredepositedataconsiderablylowdepositionratesuchthatthemoleculesare abletoenergeticminimumsitespriortoaccumulationonthesurface.Thishad beenthoroughlytestedbyaseriesofgrowthexperimentsattZnPcdepositionrates. Fromthe C 6 symmetryofthesubstrate,threetrotationaldomainsoftheZnPc stripestructuresareexpectedwitheachonehavingacorrespondingmirrodo- main,resultinginatotalofsixtZnPcstripesthatcanbeformed.However,from Figure5.6 ,thegrowthisentirelydominatedbyasingularrotationalorientationthatpro- videstheshortestpathbetweenstep-edges.Infact,thewgrowthmodereducesthe allowedsymmetryofthemoleculardomainssuchthatonlytwodominant domainsarepresentonthesurface.Thisphenomenonishighlydesirablesincetheformation oflow-angletwinboundariesoftheadjacentstripestructuresispreferredoverrandomlyori- entatedboundaries(oftenobservedduringtheislandgrowth)forachievingsuperiorin-plane electricaltransport[ 189 , 190 ]. ThoughSTMisalocalscanningtechniqueonthescaleofnanometerstomicrons,the observationoftheanisotropicwgrowthandthereductionintheallowedsymmetryof themoleculardomainsisalsoapparentonlargerscales.Asshownin Figure6.2 anddiscussed laterin Section6.1.1 ,LEEDmeasurementstakenonsub-monolayergrowthofZnPconthe deactivatedSi(111)-Bsurfacerevealsthatthepeaksareonlyduetothetwo dominantdomainspresentonthesurface.Thisistbecausethe spotsizefortheLEEDmeasurementis ˘ 1mmandthehighdegreeofmolecularordering 125 ismaintainedonthisscale.Furthermore,suchawgrowthoforganicmoleculeson theSi(111)-Bdeactivatedsurfacemayallowformacroscopicorderedstructurestoformon vicinalsubstrateswithanevenlowermiscut.Itispossiblethatthisgrowthmodecanbe extendedtootherorganicmoleculesonthedeactivatedSisurfaceprovidedthatadelicate balanceisachievedbetweenthemolecule-substrateandintermolecularinteractions. 5.3InterruptedAnisotropicStep-FlowGrowthMode 5.3.1DeactivatedSi(111)-BSurfaceDefects Itisknownthatthegrowthdynamicsofinorganicthinsdependsontheinitialdeposition conditionalongwiththeatomicprocessessuchasandnucleation.Thisalsoistrue fororganicmolecularthingrowth.Oneofthetfactorsindeterminingthe nucleationoftheZnPcmoleculesadsorbedonthedeactivatedSisurfaceistheinitialsurface condition.IfatnumberofdefectsarepresentontheSisurface,theycouldprovide additionalnucleationsiteswhichcouldresultindrasticchangesinthemoleculargrowth mode,suchasinducingatransitionfromwgrowthtoislandgrowth. Figure5.7 (a)and(b)presentatypicalSi(111)-Bsurfacethatisconsidered\pristine" attscales,wheretheatomicstructureofthesurfaceandthestraightstepedges areillustrated.Nevertheless,evenonsuchahighqualitysubstratesurfacevarioustypes ofdefectscouldbepresentandthegrowth.In Section4.2.2 , Figure4.8 has alreadyshownsomeoftherentatomicscaledefectsthatcanbepresentonthesurface. Additionally,Sidefectlinesareacommonobservationevenon\pristine"surfaces.These defectlinesareinducedbysmallatomicdisplacements,asshownin Figure5.7 (b),andoften extendacrossSiterraces,providinganactivationbarrierforadmoleculeInsome 126 Figure5.7: AtomicdefectlinesonthedeactivatedSi(111)-Bsurface .STMtopog- raphyimages( V s =+2.0V; I t =50pA)oftheSi(111)-B p 3 p 3 R 30 surfaceatt scaleswereobtainedat77K.(a)Alargescaleimageofa\pristine"Si(111)-Bsurface.Atomic Sidefectlinesarestillpresentontheterraces(highlightedbytheblackarrow)despitethe highsurfacequality.(b)ZoomedinimageofaSidefectline. instances,thesedefectlineshavecausedZnPcstripestoterminateinthemiddleofextending acrosstheSiterrace(highlightedbythedashedblackcirclein Figure5.8 (a)).This canactuallybedramaticenoughtoseparatetwodomainsofthesamepacking andregistrationwhichwouldnormallycoalesce,asseenin Figure5.8 (b). Inadditiontotheatomicdefectlines,Sistepedgedefectscanbeintroducedduringthe chemicaletchingorsampleannealingprocessespriortoZnPcdeposition.Thiscanresult indeformationsattheSistepedges,creating\bowl"-shapededgesandirregularsurface reconstructionnearthe\bowl"-shapededges.ThepresenceofthesedefectsontheSistep edgesincreasestheprobabilityofobservingallallowedmolecularrotationaldomainsby thesubstratesymmetryduetothechangeinthestepedgenucleationsites,thusremoving thebgainedbyaccessingtheanisotropicwgrowthmode.From Section5.1 , theazimuthalrelationshipbetweenthemolecularoverlayerandtheunderlyingsubstrate 127 Figure5.8: ofSidefectlinesonthegrowthofZnPc .(a)and(b)areSTM topographyimages( V s =+2.0V; I t =30pA)ofsub-monolayerZnPcmoleculesdeposited atroomtemperatureontheSi(111)-B p 3 p 3 R 30 surface,obtainedat77K.(a)The anisotropicstep-wgrowthofZnPconapristineSisurfaceinducesanereduction inthegrowthsymmetry,allowingonlyonestripegrowthdirectionwithtwomirror moleculardomains.Eachmirrordomainhasadistinctmoirepatternpointingalongthe directionindicatedbytheblackarrows.Stripesgrowalongtheshortestpathwayacrossthe SiterracesandterminateattheSistepedgesorattheSidefectlines(dashedblackcircle). AftertheZnPcstripeisterminatedatthestepedge,thegrowthextendslaterallyuntilthe ZnPcstripescoalesce(highlightedbythewhitearrows).(b)TwoZnPcdomainswiththe samemolecularpackingunabletocoalesceduetothepresenceofaSidefect line. surfaceresultsintheformationofdistinctmoirepatternsinSTM,whichallowsforiden- oftZnPcrotationaldomains.IntypicalSTMimagesoftheanisotropic wgrowthmode,despitetheexclusivein-planestripegrowthorientation,moirepat- ternspointingalongtwotdirectionsasmarkedbytheblackarrowsin Figure5.8 (a)arepresent.Thisindicatesthattherearetwomdomainsparalleltoeach other.Asthemoleculardomainsdisplayingtheidealwgrowthbegintocoalesce, twoscenariosmightbeexpected:ZnPcstripeswiththesamemolecularmerge togetherintoonemoleculardomain( Figure5.9 (b));ZnPcstripeswiththe molecularmeettoformatwinboundary( Figure5.9 (b)).However,iftheSi stepedgesbecomedeformedduringthesamplepreparationandintroducestepedgedefects, 128 Figure5.9: ofSistepedgedefectsonthegrowthofZnPc .(a)STM topographyimage( V s =+2.0V; I t =50pA)oftheSi(111)-B p 3 p 3 R 30 surface obtainedat77K.TheSi(111)-BsurfacedisplaysdeformationsoftheSistepedgesforming \bowl"shapedstructures(dashedbluecircle)andthedistinctsurfacereconstructionnear the\bowls"(dashedgreencircle).Sidefectlinesaremarkedbythewhitearrows.(b)STM topographyimage( V s =+2.0V; I t =30pA)ofsub-monolayerZnPcanisotropicw growthinterruptedbystepedgedefectsonthedeactivatedSisurface.EachZnPcdomain hasadistinctmoirepatternpointingalongthedirectionindicatedbytheblackarrows.The directionthatstripeswillcoalesceandmergeorformboundaries(dashedwhitelines)is highlightedbythewhitearrowswith.Thedashedredovalsindicateregionswhereother rotationaldomainshavebeenintroducedduetostepedgedefects.Dashedblackcircles highlightwherestripesterminateatSidefectlines. additionalrotationalmoleculardomainsareabletonucleateattheSistepedges,givingrise totheformationofrotationalmoleculardomainboundaries( Figure5.9 (b)).Byintroducing thesehigh-angledomainboundaries,thiswillultimatelyresultinhinderedin-planeelectrical transport[ 189 , 190 ]. 5.3.2Ehrlich-SchoebelBarrier AlthoughtheEhrlich-Schoebelbarrier(ESB)isoriginallyintermsofinorganic systemsusinganatomisticpicture,thesamenomenclaturehasbeenadoptedtodescribe 129 theorganicmolecularthingrowth.However,inthecaseoforganicmolecularthin [ 31 , 110 , 113 , 191 , 192 , 193 , 194 ],theanisotropicnatureandinternaldegreesoffreedom oftheorganicmoleculesoftencomplicatetheprocesses,thusdirectlythe magnitudesoftheESBspresent[ 31 , 114 , 115 , 116 , 117 , 118 , 119 , 191 , 192 ].Thusfar,the ESBhasbeendescribedasoneofthedominantfactorsindeterminingtheorganicmolecular morphology[ 31 , 110 , 113 , 115 , 116 , 117 , 118 , 119 , 191 , 192 , 193 , 194 , 195 , 196 , 197 ]. DuetothedramatictheESBcanhaveonthegrowthevolutionofmolecularthin itisnecessarytotakethisintoconsiderationforthecaseofZnPcgrowthonthedeactivated Si(111)-Bsurface. AlreadyfromthediscussionoftheanisotropicstepwgrowthmodeobservedforZnPc in Section5.2 itisapparentthataESBispresent.OncetheZnPcmolecularstripesextend acrosstheSiterraces,theyterminateattheupperSistepedges,suggestingthattheESB associatedwiththeSistepedgesislargeenoughtopreventanyZnPcdownwardmass transportascanbeseenin Figure5.6 andalsoshowndiagrammaticallyin Figure5.10 (a) [ 99 , 100 , 101 , 115 ].Infact,theESBassociatedwiththeSistepedgescontinuestorestrictthe ZnPcgrowthwithintheSiterracesevenasthecoverageincreases.Thismakesitpossible tohavevariationinthebuildupofZnPclayersoneachterraceduetothesteeringand shadowing(see Figure7.7 in Section7.2 )[ 198 , 199 , 200 ].Bydecreasingthemis-cutof theSisubstrate,themacroscopicroughnessacrosstheSiterracescouldbereduced. AsidefromtheESBassociatedwiththeSistepedges,astheZnPccoverageincreasesan additionalESBcanstarttoplayamoredominantroleinthemoleculargrowth;namelythe ESBassociatedwiththestepedgescreatedbyorderedZnPcdomains(see Figure5.10 (b)). AsthecoverageincreasesincomingZnPcmoleculeswillhaveanincreasedprobabilityto landontopofexistingZnPcordereddomainsandthedownwardmasstransportofthese 130 Figure5.10: Schematicdiagramsofthetactivationbarriersforad- moleculesduringthegrowthofZnPconthedeactivatedSisurface .(a)Depictsthe presenceofalargeESBattheSistepedges.(b)ThestepedgeoforderedZnPcmolecules alsoallowsforthepresenceofanadditionalESB.GrowthobservationsofZnPcindicate thatthisbarrierisnegligibleatroomtemperatureandenergeticallythemolecule-substrate interactionstronglyfavorsthewettingofthesubstrate.(c)AccumulationofZnPcmolecules atgrainboundariesintroducesanactivationbarrierforadmoleculescrossingbetweenthe moleculardomains.Theadmoleculeaccumulationcanalsoprovideadditionalnucleation sites. domainstotheSisurfacewilldetermineifacompletelayercanbeformedbeforeinitiating growthofasecondlayer,i.e.layer-by-layergrowth.FromtheSTMimagesshownin Fig- ure5.11 admoleculesareabletoeasilydescendtoclosethegapsbetweenmolecularstripes forZnPcandforCuPc.ThisdemonstratesthattheESBassociatedwithZnPcstepedgesis negligibleatroomtemperaturesincedownwardmasstransportisnothindered.ThisESBis expectedtoremainthesamewithincreasingthicknessaslongasthemolecularpacking motifismaintainedformultilayers[ 42 , 118 , 201 , 202 ].Thoroughdiscussionrelatedtomulti- layeredZnPcwillbeincludedin Chapter6 .ThenegligibleESBpreventstheformation 131 Figure5.11: CoalescingPcmoleculardomainsanddomainboundaryformations . STMtopographyimages( V s =+2.0V, I t =30pA)ofsub-monolayerPcmoleculesdeposited atroomtemperatureontheSi(111)-B p 3 p 3 R 30 surface.(a)Depictsthemergingof twoZnPcdomainsthatdisplaythesamemolecularpacking(b)Illustratesthe formationofatwinboundarybetweentwodZnPcdomains.(d)Evidenceof twotrotationaldomainscoalescingobservedforbothZnPcandCuPc(CuPcgrowth shownhere).Protrudingmoleculesareformedalongthegrainboundary. ofmoundedstripes[ 31 , 102 , 103 , 104 , 105 , 106 , 116 , 117 ],andisthusessentialforachieving asmoothmolecularthinItisworthnotingthatorganicmoleculesdepositedonother passivatedSisurfaces,mostnotablywithhydrogentermination,frequentlyadopttheisland growthmodewherethedegreeoftheorderingintheisnotnearlyascomparableas whathasbeenobservedonthedeactivatedSi(111)-Bsurface[ 34 , 35 , 132 , 133 , 134 ]. 5.3.3AdditionalActivationBarriersandNucleationatDomain Boundaries Asthegrowthevolves,grain(ordomain)boundarieswillbegintoformwhichcanalso actasactivationbarriersintheprocessandevenintroducenucleationsitesfor adatoms.However,therearefewstudiesdiscussingthetofthegrainboundarycrossing barrieronthemolecular[ 102 , 203 ]andnucleation[ 204 , 205 ].Thecombination oftheESBandthegrainboundarycrossingbarriercaninevitablyresultinarough 132 [ 102 ].Fromtheprevioussection,theESBassociatedwithdescendingtheZnPcstepedges canbeconsiderednegligible.Thus,itisexpectedthattheanisotropicwgrowthof ZnPcwillproceedinalayer-by-layerfashion.However,astheZnPccoverageapproachesone monolayerthenucleationofthesecondorthirdlayerareobservedpriortothecompletionof thelayerasshownin Figure5.12 .Thus,anadditionalactivationbarrierorpreferential nucleationsitesmustbepresentontopofthelayer,whichpreventsZnPcmolecules fromdowntocompletethemonolayer. Figure5.12: NucleationofZnPcdomainsatdomainboundariespriortomonolayer completion .STMtopographyimages( V s =+2.5V, I t =20pA)ofZnPc(0.9ML)deposited ontheSi(111)-B p 3 p 3 R 30 surface.(a)Depositionatroomtemperatureresultsinthe formationofamultilayeredstructurewheretheformationofthenextlayerbeginspriorto thecompletionofthepreviouslayer.(b)Zoomedinimageoftheregionoutlinedbytheblack rectanglein(a).ThemoirepatterndirectionidenthetZnPcdomains(black arrows).Thedashedblacklineshighlightthegrainboundariesoftheunderlyinglayerwhere nucleationofthetoplayerhasoccurred. Asshownin Figure5.11 (c),whentwotrotationalmoleculardomains(generated duetothepresenceofSistepedgedefects)coalesce,moleculesatthegrainboundaryare unabletoadoptathataccommodatesbothdomains.Thisleadstoabuildup ofmoleculeswhichcanactasgrainboundarycrossingbarriers,hinderingmolecular acrosstdomainsandthusterminatingmolecularstripesontopofthelayeras 133 showndiagrammaticallyin Figure5.10 (c).Ifthemoleculesatthegrainboundariesadopt apoorlypackedgurationandareorientedsuchthatalargeportionoftheirextended ˇ systemisexposed,newnucleationsitesforadmoleculescouldbeintroduced.Closeinspection ofthesecondlayerdomainswithintheregionhighlightedbytheblackboxin Figure5.12 (a)showsthatthenucleationofthesedomainsisinitiateinthemiddleoftheterraceinstead ofatthestepedge.TheexpandedSTMimageofthehighlightedareain Figure5.12 (a) furtherthatthenucleationoccursatselectlocationsalongthegrainboundaries oftheunderlyinglayerasindicatedbythedashedblacklinesin Figure5.12 (b).This suggeststhatwithinagivenSiterracethedomainboundariesareplayingamoredominant roleindeterminingthegrowthevolutionandtheroughnessoftheZnPcthantheESB associatedwithdescendingtheZnPcstepedges[ 102 ]. 5.4Conclusions ZnPcdisplaysanintermediatemolecule-substrateinteractionwiththedeactivatedSi(111)-B surfacewhichiswell-balancedwiththeintermolecularinteractionssuchthatthemolecules areabletoeasilyadoptaclose-packedhighly-orderedwhichistiltedabove theplaneparalleltotheunderlyingsubstrateevenatroomtemperature.Thisisfurther supportedbytheobservationofthemoiresuperstructureintheSTMimagesoftheor- ganicmolecularoverlayer,implyingthatthedomainsofthemoleculardisplayasp azimuthalregistrationwiththeunderlyingSisurface,butarenon-commensurate.Addition- ally,theZnPcmoleculescanserapidlyonthedeactivatedSisurfacesuchthatthey nucleateattheSistepedgesitesandproceedtogrowintheanisotropicwgrowth modeatroomtemperature.Thisistheobservationofstep-wgrowthforanorganic 134 molecularthinonasemiconductingsurfaceandhastheaddedbofrestricting thenumberofsymmetryallowedmoleculardomainssuchthatasingledominantin-plane molecularordering. DespiteallofthepositivegrowthevolutiontraitsofZnPconthedeactivatedSisurface, therestillremainsomechallengesinordertoensureaofthehighestqualitycanbe formed.TheinitialsurfaceconditionofthedeactivatedSi(111)-Bmustbecarefullycon- trolledsoasnottointroduceatnumberofdefectswhichcanalterthenatureof thewgrowth.ThoughtheESBforthestepedgesformedbytheZnPcmoleculesis considerednegligible,theESBduetotheSistepedgesandadditionalactivationbarriers and/ornucleationsitesassociatedwiththegraindomainboundariesofZnPctogethercan inducerougheningofthewithincreasingmolecularcoverage. Chapter6 and Chapter7 willincludemorediscussionregardingthegrowthevolutionofZnPconthedeactivatedSi surfaceasthegrowthtransitionstoamultilayeredlmandalsostrategiesforremovingsome oftherougheningthroughtemperaturecontrol. 135 Chapter6 Out-Of-PlaneGrowthStudyofZinc PhthalocyanineontheDeactivated Si(111)-BSurface InthischapterthegrowthofZnPconthedeactivatedSi(111)-BsurfaceisstudiedbyLEED fromsub-monolayertomultilayerthicknesses.In Section6.1 ,LEEDmeasurementsare takenonZnPcgrownattthicknesses.Appropriateassignmentoftheobserved peaksisaddressedandcomparisonwithLEEDpatternsimulationsareusedto trackthechangesintheZnPcpackingwithincreasingthickness. Section6.2 comparesobservationsfromLEEDwithSTMmeasurementsoftheZnPcgrowth.From thecombinedLEEDandSTMinvestigations,themolecularpackingofZnPc onlyexperiencesaverymildrelaxationafterthemonolayerandcanmaintainthis packingfortlythickTheseresultsareconcludedinthelastsection.Partsof thischapterareadaptedfromReference[ 41 ]S.R.Wagner etal ., J.Phys.Chem.C 118 , 2194(2014). 136 6.1Low-EnergyElectronMeasurements 6.1.1Sub-MonolayerGrowthofZincPhthalocyanineandSimu- latedLow-EnergyElectronPattern ThoughatamountofknowledgeandunderstandingofZnPcgrowthdynamicson thedeactivatedSi(111)-BsurfacecouldbeascertainedfromtheSTMtechniquesutilizedin Chapter5 ,astheorganicmolecularmthicknessincreasesitbecomeschallengingfora stabletunnelingjunctionfromthesubstratethroughmultipleorganiclayerstotheSTMtip tobeestablished.ThiscanresultinunstableSTMimagingandcomplicationsinobtaining measurementswhichadequatelyobservethephenomenaofinterest.Asanalternative,low currentLEEDtheabilitytoextractinformationregardingthemolecularpacking withoutdamagingtheasthethicknessincreasesbytrackingchangesintheelectron ComparisonbetweentheresultingLEEDpatternsandSTMimagesobtained atthicknesseswhichstillostableimagingensuresproperassignmentofraction peaks,whichisimportantwheninterpretingresultsonthickerorganics.Usingthis strategy,theout-of-planegrowthbehaviorofZnPconthedeactivatedSisurfacecanbe explored. InordertodistinguishtheelectronctionpeaksinLEEDmeasurementsbetween thedeactivatedSisurfaceandtheZnPcm,initialmeasurementsonthebareSisurface arerequired.In Section4.2.2 , Figure4.7 (a)showsatypicalLEEDpatterntakenonthe deactivatedSi(111)-B p 3 p 3 R 30 surface.Themeanfreepathoftheincidentelectron beamis5-8 AforthebeamenergiesusedinthisLEEDstudy[ 155 , 156 ].Therefore,both thebulkSi(111)andsurfacereconstructionfeaturescanbeobserved.Fromthe 137 pattern,theinnerspotsareassociatedwiththelargerunitcellparametersofthe reconstructedSi(111)-B p 3 p 3 R 30 surfacewhiletheouterspotsarefrom theSi(111)1 1bulkstructure.ToverifythattheobservedLEEDpatternisconsistent Figure6.1: TypicalLEEDpatternofsub-monolayerZnPconthedeactivatedSi surface .(a)SimulatedLEEDpatternofZnPc(sixcolordomains)ontheSi(111)-B p 3 p 3 R 30 surface(whitecircles)usingtheunitcellparametersandtheazimuthalrelation betweenthemolecularoverlayerandthesubstratedeterminedfromSTMmeasurements in Section5.1 .ThefullsymmetryofthedeactivatedSi(111)-Bsurfaceisconsidered.(b) LEEDpattern( E beam =21.8eV)of0.6monolayersZnPccoveragedepositedonSi(111)-B withsomesurfacedefectspresent.TheblackcrossesindicatethelocationofexpectedZnPc peakswhilethegreencirclesindicatethelocationofexpectedsatellitepeaksthat werenotintenseenoughtobeobserved.Thefeatureshighlightedbytheblue,red,purple, andblackrectanglesaretoaccuratelyassignthepeaksthatareassociatedwith theZnPcdomainswhilesatellitepeaksarehighlightedbythedashedgreencircles. withwhatisexpectedtheoreticallyfromtheperiodicityofthetwostructures,asimulated LEEDpatternconsideringboththebulkSi(111)1 1bulkstructure(reciprocalspaceunit 138 celloutlinedinwhite)andthe p 3 p 3 R 30 surfacereconstruction(reciprocalspaceunit celloutlinedinblue)isshownin Figure4.7 (b).SimulationandtheobservedLEEDpattern displaygoodcorrespondenceatabeamenergyof E beam =50eV.ForLEEDmeasurements onZnPcsincetheunitcellparametersdeterminedin Section5.1 arelargerthanor closetothatoftheSi(111)-B p 3 p 3 R 30 surfacereconstruction,typicallyalowerbeam energyof E beam =21-23eVhadtobeusedtoobservetherelevantfeatures.At thisenergythe p 3 p 3 R 30 surfacereconstructioncanbeobservedaswellasanyZnPc features. Afterestablishingacleansurface,ZnPcmoleculesaredepositedwiththesubstrateheldat 110 Ctoinvestigatethefactorsthatgovernthequalityofthemultilayeredgrowth.The elevatedsubstratetemperatureisselectedforreasonswhichwillbecomeclearin Chapter7 . Oncedepositioniscompleted,samplesarethentransferred insitu totheSTMand/or transferred insitu totheLEEDfordataacquisition.Thoughtheanisotropicw growthobservedforZnPccanrestrictthestripegrowthtoonedominantorientation,as discussedin Section5.3.1 ,additionalrotationaldomainsmaystillbeobservedifirregularSi stepedgesareformedduringthesamplepreparation.Thisresultsinanincreasedprobability toobserveallsixallowedmoleculardomains,threefromtherotationalsymmetryofthe substrateandtwoforeachrotationaldomainfromthemirroronsymmetry,whichwill contributetotheobservedLEEDpattern.Theadditionalrotationaldomainswillcontinue topropagatethroughthemultilayeredandtheLEEDpatterninasimilar fashion.Additionally,itshouldbenotedthatthebeamspotsizeforLEEDmeasurements is ˘ 1mm,thus,theobservedpatternisanaveragesamplingoftheZnPc growthonalargerscalethancanbeobservedinSTMorAFMmeasurements.Thiscan resultinanincreasedprobabilitytoencounterdefectsintheSisurfacewhichultimately 139 determinetheZnPcgrowthandtheobservedpattern.In Figure6.1 (a)theLEEDpatternforZnPcissimulated,takingintoaccounttheSiandZnPcunit cells,thefullsurfacesymmetryofthesubstrate,andtheazimuthalregistrationbetweenthe molecularoverlayerandtheSisurface,asdetailedin Section5.1 [ 39 , 40 ].Itisworthnoting thatthemultiscatteringbetweentheZnPcmoleculesandtheSisurfaceandthe fromtheenhancedelectronicdensityofstatesofthemoirepatternarenotconsideredinthe simulation.Thesixcolordomains,withreciprocalspaceunitcellincluded,correspondto thesixallowedZnPcdomainswhiletheSi(111)-B p 3 p 3 R 30 surfacereconstructionis outlinedinwhite.ThecommonlyobservedLEEDpatterntakenonasub-monolayerZnPc shownin Figure6.1 (b)isingoodagreementwiththesimulationresult.However,extra peaksareobserved.ThekeyfeaturesoftheLEEDpatternarehighlightedbythe blue,red,purple,andblackrectanglesandareedtoaccuratelyassignthepeaksthat areassociatedwiththeZnPcdomains.Thesatellitepeaksarehighlightedbythedashed greencircles.Peakswithlowerintensitythanexpectedaremarkedwithgreenspots.The explanationoftheobservedsatellitepeaksisdiscussedindetailin Section6.2 . ThoughalloftheZnPcrotationaldomainsarecommonlyobservedintheLEEDpatterns ofsub-monolayerandmultilayerlms,ifapristineSi(111)-Bsurfaceisprepared,asshown in Figure5.7 (a),itispossibletoachievesuperiorin-planemolecularorderingduetothe anisotropicwgrowthmodeonalargescale,comparabletothebeamspotsizefor LEEDmeasurements.Suchacaseisshownin Figure6.2 .FromtheLEEDpattern,it isobviousthatpeakshavedisappearedduetothereductioninthenumberof allowedZnPcdomainstoonlytwomirrordomainsexhibitingthesamein-plane growthdirection.Thesixmostintensedarkspotswiththesamehexagonalorientationas theinnerspotsobservedin Figure4.7 arethepeaksassociatedwiththeSi(111)- 140 Figure6.2: LEEDpatternofZnPcexhibitingtheidealanisotropicw growth .LEEDpattern( E beam =22.8eV)ofsub-monolayerZnPcdepositedonaSi(111)-B surfacewithdefect-freeterracesandregularstepedges.Thereciprocalspaceunitcellsof thetwomirrordomainsaredepictedinredandblue.Theblackcrossesindicatethelocation ofexpectedZnPcpeakswhilethegreencirclesindicatethelocationofexpected satellitepeaksthatwerenotintenseenoughtobeobserved. B p 3 p 3 R 30 reconstructionwithfourofthemalsoincludingcontributionsfromthe ZnPcmolecularstructures.Thetionpeaksforthetwomolecularmirrordomainsare highlightedbythecornersofthereciprocalspaceunitcellsoutlinedinredandblue.From thereciprocalspaceunitcellsdepictedintheLEEDimage,anangleof4 : 0 0 : 5 canbe observedbetweenthetwounitvectors.Thisanglearisesfromtheazimuthalregistrationof theZnPcmolecularoverlayerandtheunderlyingSisubstrate[ 39 , 40 ].Somekeyfeatures intheLEEDpatternareobserved,similarto Figure6.1 (b),a\bow-shaped"patternis highlightedintheblackrectangleanda\linear"featureishighlightedintheredrectangle. 141 6.1.2MultilayeredGrowthofZincPhthalocyanineandSimulated Low-EnergyElectronPattern NextthegrowthphenomenaofmultilayeredZnPcthinisaddressed.TheLEEDpat- ternforZnPccoveragerangingfrom0.6monolayersto40monolayersismonitoredtotrack changesinthein-planemolecularordering.Achangeinthein-planecrystallizationcorre- spondstoachangeinthemolecularwhichwilllikelyresultinachangein theinterlayerspacing[ 54 , 206 , 207 , 208 , 209 ].LEEDpatternsatthecoverageof0.6mono- layersand15monolayersareshownin Figure6.3 (a)and(b),respectively.Comparison Figure6.3: LEEDpatterncomparisonbetweensub-monolayerandmultilayer ZnPconSi(111)-B .(a)and(b)LEEDpatterns( E beam =21.8eV)ofZnPcdeposited onaSi(111)-Bsurfaceatcoveragesof0.6monolayersand15monolayers.Arotationinthe oftheZnPcpeakswithincreasingcoverageishighlightedbythe redhexagon.SlightspotsplittingoccursfortheinnerZnPcactionpeakshighlightedby thepurplecircle. betweenthesetwopatternsdisplaysthreedistinctfeatureswithincreasingcoverage:(i)a rotationofthepeaksneartheouteredgeoftheLEEDpattern(redhexagon);(ii) 142 Figure6.4: Schematicreal-spaceunitcellrepresentationsofthelayerand multilayeredZnPc .(a)TheZnPcreal-spaceunitcellregistrationwiththeSi(111)-B surfaceforthemonolayerusingtheepitaxialrelationshipdeterminedin Section5.1 .(b) Thereal-spacerelaxedmolecularunitcellformultilayeredZnPcgrowth.(c)Simulated LEEDpatternofZnPc(sixcolordomains)consideringthefullsymmetryoftheSi(111)-B surfaceusingtheunitcellthatisslightlyrelaxed( b 1 =12 : 4 A, b 2 =6 : 6 A,and =90 )incomparisonwiththemolecularpackingin Section5.1 ,whilestill maintainingthesameazimuthalregistrationwiththesubstrate. slightspotsplittingfortheinnerpeaks(purplecircle);(iii)disappearanceofthe satellitepeaksinthepattern.Despitethesesubtlechanges,thegeneralshapeoftheLEED patternismaintainedsuggestingonlyaminorvariationinthein-planemolecularordering withincreasingcoverage.Thisminorvariationisassociatedwithaslightlyrelaxedunitcell ( Figure6.4 (b))inthemultilayeredZnPcwithunitcellparameters b 1 =12 : 4 0 : 1 A, b 2 =6 : 6 0 : 1 A,and =90 1 incomparisonwiththemonolayerunitcell( Figure6.4 (a))determinedfromSTMmeasurementsin Section5.1 ( b 1 =12 : 3 0 : 1 A, b 2 =6 : 7 0 : 1 A, and =92 1 ). Figure6.4 (c)showstheLEEDsimulationcarriedoutusingthisrelaxed ZnPcunitcellandmaintainingthesameazimuthalregistrationangleof28 whichcaptures bothfeatures(i)and(ii).Asdenotedin Figure6.4 (a)and(b),maintainingthesameaz- imuthalregistrationforthisrelaxedZnPcunitcellagaintheanglebetween ~ b 1 and ~ b 1 0 to4 resultinginthesamespotsplittingformirrordomainsaspreviouslyobservedwithin 143 Figure6.5: LEEDpatternsofmultilayeredZnPcanddecreasingsatellitepeak intensity .(a)-(f)LEEDpatterns( E beam =21.8eV)ofZnPcdepositedontheSi(111)-B surfaceatcoveragesof3,5,and7monolayerswhere(a)-(c)aretakenatnormalincidence fromthesamplewhile(d)-(f)aretakenat7 normalincidencetoobservethe(00) peak.Theintensityofthesatellitepeaksareobservedtograduallydecreaseswithincreasing coverageuntiltheLEEDpatternresemblesthesimulationinshownin Figure6.4 . theblackrectanglesin Figure6.1 (b)and Figure6.2 ).However,thechangeintheunit cellangle causes ~ b 2 and ~ b 2 0 tonolongerliealongaSisymmetryaxis,yieldingadditional spotsplittingfortheinnerspotsintheLEEDpattern,ashighlightedbythepurplecircle in Figure6.3 (b)and Figure6.4 (c)).Notethatfeatures(i)and(ii)areobservedaslong asthecoverageincreasesbeyondtherstmonolayerandaremaintaineduptoacoverage of40monolayers.Relaxationofthemolecularpackingbeyondthemonolayerhasbeen observedonsubstrateslikeSiO 2 [ 128 , 129 ],suggestingthatevenontheweaklyinteracting 144 surfacesthemolecule-substrateinteractioncanleadtoaslightlystrainedlayerfromthe toplayers.However,theseLEEDobservationssuggestthatotherthantheslightunitcell relaxationfromthesecondmonolayerthein-planeorderingoftheZnPcdoesnotchange withincreasingcoverage.Infact,depositionuptoacoverageof40monolayersresultsin thesameLEEDpatternthatisobservedforacoverageof15monolayers,suggestingthat ahighlyorderedmolecularstructureismaintainedinmultilayeredZnPcwhichisdif- ferentfromany( hkl )sliceofthereportedbulkstructureforcrystalsgrownbythevacuum deposition[ 125 ]. ThedisappearanceofthesatellitepeakswithintheLEEDpatterniselaboratedonin Figure6.5 .LEEDpatternsaresystematicallyobtainedforZnPcdepositedonthedeactivated Sisurfaceatcoveragesof3,5,and7monolayerstotrackthechangeinthesesatellite peaks.CloseinspectionoftheLEEDpatternin Figure6.5 (c)showsthatall thesignaturesofsatellitepeakshavedisappeared,resultinginaLEEDpatternthatclosely resemblesthesimulatedpatternin Figure6.4 (c).Adistinctexampleofthesatellitepeak intensitydecreasecanbeobservedintheLEEDpatternstakenat7 normalincidence. Figure6.5 (d)showsthatforacoverageof3monolayerstherearesixsecondarypeaks surrounding(00),whilethesepeaksarenolongerpresentatacoverageof7monolayersas canbeseenin Figure6.5 (f).FromtheseriesofLEEDmeasurementsin Figure6.5 ,the intensityofthesatellitepeaksdecreaseswithincreasinglmthickness.Inorder todeterminetheoriginofthesesatellitepeaksandtheirimplicationsontheZnPcmolecular growth,STMmeasurementsatthesameZnPccoveragesarenecessarytoidentify anystructuralorelectronicchangeswiththickness,whichisthesubjectofthenext section. 145 6.2ComparisonwithScanningTunnelingMicroscopy Measurements ToidentifytheoriginoftheadditionalpeaksobservedinLEED,Fast-Fourier- Transform(FFT)ofSTMimagestakenonthetwoZnPcmirrordomainsisperformed,as shownin Figure6.6 (a)-(c)[ 210 , 211 , 212 , 213 , 214 ].Thereciprocalspaceunitcellsare Figure6.6: LEEDpatterncomparisonwithFFTsobtainedfromSTMimages ofZnPc .(a)and(b)FFTsobtainedfromSTMtopographyimagestakenontwomirror domainsofZnPconthedeactivatedSisurface.Thereciprocalspaceunitcellsofthe twodomainsarehighlightedbythedashedredandbluelines.ThemoiredirectioninFFT isindicatedbytheblackarrowwithsatellitepeaksinducedateachmajorpeakalongthis direction.(c)STMtopographyimage( V s =+2.5V, I t =30pA)obtainedat77K,takenatthe boundarybetweentwomirrordomains.Thereal-spaceunitcellsforeachdomain areoutlinedinredandbluewhichcorrespondtothereciprocalspaceunitcellsshowninthe FFTs.(d)ResultingFFTwhentheFFTsfromthetwomirrordomainsin(a)and (b)areoverlayedtogether.Theredandblackrectangleshighlightfeaturesforcomparison with Figure6.2 .Thereciprocalspaceunitcellvectorsforbothdomainsareshowninred andbluewiththeanglebetweenthetwoishighlighted. 146 depictedinbothFFTimages.AhighresolutionSTMimageofthedomainboundaryis alsopresentedin Figure6.6 (c),wherethemolecularpackingforthetwodomainsismirror withrespecttoeachotherasindicatedbytheunitcells( b 1 0 = b 1 =12 : 3 0 : 1 A, b 2 0 = b 2 =6 : 7 0 : 1 A,and 0 = = 92 1 )[ 39 ].Closeinspectionofthe(00)points oftheFFTsshowsthattherearetwodistinctpeaksneartheoriginwhichcorrespondtothe periodicityofthemoirepatternforeachdomain.MeasurementsfrombothSTMandFFT analysisgiveamoirelineperiodicityof7 : 9 0 : 1nm.Thisslightlyfromoneofthe moiresupercellvectorsdiscussedin Section5.1 becauseinthissectionthemoirepatternis discussedasatwo-dimensionalsuperstructure.However,fromFFTsofZnPcdomains,the moiresignaturesappeartogiveone-dimensionalsatellitepeaks,mostnotablybecauseonly twodistinctpeaksarefoundnearthe(00)pointofeachFFT.Furthermore,satellitepeaks correspondingtothemoireperiodicityarealsoobservedaroundeachpeakassociatedwith themolecularoverlayer,formingafringedlinepatternwhichisalongtdirections (highlightedbytheblackarrows)forthetwomirrordomains. IntheFFTs,thereal-spaceperiodicityfromtheSTMimagesisconvertedtoreciprocal spacewhichcanbecomparedwithLEEDmeasurements.Thus,theFFTsofthetwomirror domainscanbecomparedwiththeLEEDpatternshownin Figure6.2 byoverlayingthe twoFFTs.Asdepictedin Figure6.6 theoverlayedFFTencompassingbothdomainsshows thatthecombinationofthepeaksassociatedwiththemolecularoverlayerandthemoire patternsproducesa\bow-shaped"patternhighlightedintheblackrectangleanda\linear" featurehighlightedintheredrectangle.ThesefeaturesarealsoobservedintheLEED patternshownin Figure6.2 ,wherethesatellitesignaturesareattributedtothe contributionsofmultiscatteringbetweentheSi(111)-B p 3 p 3 R 30 surfaceandtheZnPc molecularoverlayer,i.e.thesuperpositionofthecorresponding ~ k vectors[ 215 , 216 , 217 , 218 ], 147 Figure6.7: STMimagesofmultilayeredZnPconthedeactivatedSisurface .(a)-(c) ZoomedinSTMtopographyimages( V s =+2.5V, I t =30pA)takenonthree(a),e(b), andseven(c)ZnPclayers,obtainedat77K.Notethattheintensityofthemoiresignature intheSTMimagesdecreaseswithincreasingthickness. and/ortheenhancedelectronicdensityofstatesfromthemoirepatternwhichmightcause aphaseshiftintheelectronbeamtoproduceadditionalscatteringpeaks[ 219 ]. CombiningLEEDobservationandanalysissoftware,thevalidityofthisassumptionwas testedbyinvestigatingthesixallowedsatellitepeaksthatoccuraroundthe(00)point. Fromthesecalculationsthemoirelineseparationispredictedtobe8 : 0 0 : 1nm,consistent withSTM/FFTmeasurements.Itisworthnotingthattheincommensuratenatureofthe ZnPcoverlayer[ 39 ]leadstotheobservationofmoiresatellitepeaksthatdonotcompletely thebrillouinzone[ 220 ]. CombinedwiththeFFTanalysis,theoriginoftheintensitydecreaseofthesatellite peaksintheLEEDmeasurementswithincreasingZnPcthicknesscanbeexplainedin connectiontothemoiresignatureobservedintheSTMimages.In Figure6.7 ,theSTM imagesillustratethatevenaftertheslightrelaxationintheunitcelltheZnPc moleculescontinuetoshownearlythesamein-planemolecularorderingandazimuthalreg- istrationwiththesubstrateasindicatedbytheobservedmoirepattern.However,asthe thicknessincreasesadrasticdropinthesignatureofthemoirepatternisobservedmuch 148 likethecaseoftheLEEDmeasurement.Itisimportanttonotethatitcouldbechallenging toimagemultilayersoforganicmoleculesduetothelowelectricalconductanceoforganic [ 221 ].However,throughoutthisstudyalowtunnelingcurrentisusedwhichallows forthemolecularorderingtostillbediscernedinupto7monolayers[ 222 ].Thegrad- ualdecreaseofthemoirepatternwithincreasingthicknessisindicativeofthedecreasing molecule-substrateinteraction.Thus,theappearanceofthemoirepatternsinLEEDimages providesameanstotrackthemolecule-substrateinteractionwithincreasingthickness. Asthemolecule-substrateinteractiondecreases,themolecule-moleculeinteractionwillbegin toplayamoredominantroleinthegrowthwhichcouldleadtoachangeinthemolecular fromthethinphasetothebulkphase.However,thisisnotthecasefor ZnPcgrowth.SincetheZnPcmoleculesstartingfromtheinitialmonolayeraretiltedabove Figure6.8: Side-viewschematicoftheZnPcpackingonthedeacti- vatedSisurface .ZnPcmoleculesaretiltedby ˘ 30 abovetheSisurfaceasdescribedin Section5.1 .Thisallowsfortinterlayer ˇ - ˇ stackingtostabilizemultilayered theSi(111)-Bsurfaceby ˇ 30 (see Figure6.8 )[ 39 ],italarge ˇ - ˇ interactionbetween adjacentlayers.Thus,thisstronginterlayermolecularinteractiondrivesthehighdegree ofout-of-planeorderinginthemolecularthinphasewhichcanbemaintainedwith- outrelaxingtothebulk-likestructuredespitethegraduallydecreasingmolecule-substrate interaction[ 223 ].Thecombinedin-planeorderingfromaccessingtheanisotropicw growthmodeandout-of-planeorderingallowedbytheinterlayer ˇ - ˇ interactionthe 149 possibilityofformingamolecularorganicthatdisplaysbothin-planeandout-of-plane molecularordering( Figure6.9 ). Figure6.9: SummaryschematicofZnPcgrowthonthedeactivatedSi(111)-B surface .Theanisotropicwgrowthmodeallowsforsuperiorin-planemolecular orderingwhilethetiltedmolecularallowsfortinterlayer ˇ - ˇ stacking tostabilizemultilayeredleadingtoout-of-planemolecularordering. 6.3Conclusions AccesstotheanisotropicwgrowthmodeofZnPconthedeactivatedSi(111)-B p 3 p 3 R 30 surfaceprovidesameanstoachievelong-rangein-planemolecularordering. However,adeepunderstandingoftheout-of-planemolecularorderingforgrowthbeyond thelayerisnecessary.ToexplorethemultilayeredgrowthofZnPconthedeactivated Si(111)-B p 3 p 3 R 30 surface,combinedSTMandLEEDmeasurementsarerequired. Beyondthemonolayer,amildrelaxationintheZnPcin-planemolecularorderingoccurs whichisnottenoughtochangetheazimuthalorderingordisrupttheout-of-plane ordering.WithincreasingthicknesstheZnPcgrowthisabletomaintainthisnewre- laxedmolecularpackinguptoacoverageof40monolayers,despiteagradualdecreaseinthe 150 molecule-substrateinteraction.ThetiltedorientationoftheZnPcmoleculesallowsfor ˇ - ˇ stackingbothin-planeandout-of-plane,resultinginastablemolecularstructure. 151 Chapter7 TemperatureDependentGrowthof MetalPhthalocyaninesonthe DeactivatedSi(111)-BSurface PreviouschaptershaveprimarilyfocusedoncharacterizingthegrowthofZnPcbothin- planeandout-of-planeatasubstratetemperatureduringmoleculardepositionon thedeactivatedSi(111)-Bsurface.Thesubjectofthischapterfocusesonchangesinthe observedgrowthmodesof M PcmoleculesdepositedontheSisurfaceatvaryingsubstrate temperatures,studiedbySTMtopographyimages.In Section7.1 ,STMimagingisused toexploremoinsub-monolayergrowthofZnPcandCuPcatvaryingsubstrate temperaturesduringmoleculardeposition.Theseresultsarecomparedwiththetypical observationsfrominorganicgrowthmodechangeswithvaryingsubstratetemperature.Then in Section7.2 ,STMmeasurementsaretakenonZnPcvaryingfromsub-monolayerto multilayercoveragesaterentsubstratetemperaturesduringmoleculedeposition.All resultsareconcludedinthelastsection.PartsofthischapterareadaptedfromReferences [ 39 , 40 , 42 ]:S.R.Wagner etal ., Phys.Rev.Lett. 110 ,086107(2013);S.R.Wagner etal ., MaterRes.Soc.Symp.Proc. 1550 ,609(2013);S.R.Wagner etal ., Surf.Sci. 630 ,22 (2014). 152 7.1GrowthModeTransitionsofZincandCopperPh- thalocyanine Asalreadydiscussedin Chapter5 and Chapter6 ,evenwiththesubstratetemperatureheld atroomtemperatureduringthedepositionprocess,itispossibletoachievehighlyordered ZnPcdomainsdisplayingtheanisotropicwgrowthmode.Thisisduetothelargedif- fusionlengthofZnPconthedeactivatedSi(111)-Bsurfaceallowingtheadsorbedmoleculesto thestepedgesitesfornucleation.Fromthediscussion Section2.4 ,iftheorganicgrowth isanalogoustoobservationsofatomisticinorganicgrowththenasthesubstratetemperature increases,itisexpectedthatlargerZnPcstripeswillformwithalowernucleationdensityon thesurface. Figure7.1 displayscharacteristicSTMimagesoftheZnPcgrowthwiththesub- strateheldattemperaturesof25 C,110 C,180 C,and270 Cusingadepositionrate andcoverage.Itshouldbenotedthatfororganicgrowththesubstratetemperaturecannot beelevatedashighastypicallyperformedforinorganicgrowth.Thisisbecausetoohigh ofasubstratetemperaturecancausetheorganicspeciestodissociate.For M Pcmolecules inaninertgasenvironmentthedissociationtemperatureis ˇ 450 C[ 186 ],though,under vacuumconditionsasperformedinthesemeasurementsthedecompositiontemperaturecan beevenhigherthanthis(500 600 Corhigher[ 120 ]).Tobecertainthatdecompositionof the M Pcmoleculesisnotaconcern,thesubstratetemperatureismaintainedwellbelowthe decompositiontemperature[ 31 , 112 , 121 , 122 ]. Figure7.1 (a)displaysthetypicalw growthobservedatroomtemperature.Whengrowthisperformedat110 C( Figure7.1 (b)) thewgrowthismaintained,however,tlywiderstripesareformedconsisting ofZnPcmoleculesinthehighly-orderedtiltedtion,allwiththesameazimuthal orientationwithrespecttotheunderlyingsubstrate.TheZnPcdomainalsomaintainsthe 153 samemolecularunitcellaswhatisobservedforroomtemperaturedeposition.Additionally, asexpectedfromthediscussioninboth Section2.3 and Section2.4 ,thenumberofstripes decreaseswithincreasingsubstratetemperature.However,asthetemperatureisincreased furtherthegrowthbeginstodeviatefromwhatistypicallyobservedforinorganicgrowth. Figure7.1: STMimagesofZnPcgrowthmodechangeswithincreasingtemper- ature .(a)-(d)STMtopographyimages( V s =+2.0V, I t =35pA)ofZnPcdepositedon theSi(111)-B p 3 p 3 R 30 surfacewithvaryingsubstratetemperatureduringdeposition. Thecoverageanddepositionratearekeptthroughouttheseexperiments.Transitions inthegrowthmodefromwgrowthtotwo-dimensionalislandnucleationtorandom roughgrowthareobserved. At180 C,thegrowthnolongermaintainstheanisotropicwgrowthmodeas evidencedbytheformationofislandsthatnolongernucleateatthestepedges,butinstead 154 growontheterraces,asshownin Figure7.1 (c).Theseislandsconsistofverylargesingle domainswithallmoleculeshavingthesameazimuthalregistrationtothesubstrateand maintainingthetiltedmolecularpacking.Consistentwithclassicalnucleation theory,theislandsthatnucleateonagiventerraceareverylargeinsizeincomparison withlowertemperatures.Consequently,thenucleationdensityoftheseorderedislandshas alsodecreasedsuchthattherearetlylargeregionsonagiventerracewhereno orderedZnPcislandsareobservedandbetweenthelargedomainsonlymoleculesadsorbed atterracedefectsitesoccurs.Thoughthemoleculargrowthcontinuestofollowwhatis expectedfromclassicalnucleationtheory,thereversalinthegrowthmodefromstw growthtotwo-dimensionalislandgrowthwithincreasingtemperatureincomparisonwith typicalinorganicgrowthobservationsisunexpected,andcontinuestobethetrendasthe temperatureisincreasedfurther.At270 CtheZnPcgrowthdisplaysthecompletelossof in-planeazimuthalorderingandonlyrandomroughgrowthisobservedacrosstheentire substratesurface.Thistrendseemscounterintuitiveasonewouldexpectalongermolecule length,however,thegrowthmodetransitionsarereversedfromthetypicaltrend observedininorganicsystems. Onepotentialconcernisthatthesecouldbeduetotheofmolecular re-evaporation.However,post-annealingthesamplesat270 CforZnPcdoesnotchange themolecularcoverageonthesurface,excludingthepossibilityofre-evaporation.Toverify thattheobservedgrowthcorrespondstothermalequilibriumstructurestwotap- proachescanbeused:furtherreductioninthedepositionrate;post-annealthesamplesafter deposition.However,nochangesinthestructuresformedateachcorrespondingtempera- turecouldbeobserved.Anotherpossiblescenarioisthatparameters,including thepath,barrier,andattemptfrequency,mightbedisturbedtly 155 athightemperaturesduetothevibrationalinternalenergyofmoleculesduring thecourseofIftheprocessisimpeded,onewouldexpectthegrowth modetotransitfromwtotwo-dimensionalislandnucleationandeventuallytothe roughgrowth.Inordertoevaluatethescenario,futuremoleculardynamicscalcu- lationsarewarranted.Nevertheless,theresultstakenat180 C,inparticularthelowdensity andlargedomainsizeofislandsnucleatedontheterraces,donotseemtosuggestthatthe molecularsionisimpeded.Thisleadstothehypothesisthatthecompetitionbetween andnucleationratesasdiscussedin Section2.3 arecontributingtothechangesin thegrowthmodesobservedatelevatedtemperatures. From Equation2.18 in Section2.3 ,theprobabilityforstablenucleationisdetermined bythecombinedprobabilityforformingacriticalnucleus, N p ,andtheprobabilityforan admoleculetotothecriticalnucleustomakeitstable, D p [ 91 , 92 ].Inthiscasethe dominantfactorsthatcontributetothenucleationarethebarrier E ds ,theaverage edgeenergyassociatedwithnucleationofatwo-dimensionalislandthesupersaturation level ˙ ,andthesubstratetemperature T .Forappropriatecomparisonwithobservations fromtheZnPcexperiments, ˙ issettoareasonableconstantvalue,i.e.deposition rateusedforeachtemperature.Ifthetwoexponentialexpressionsfrom Equation2.18 are plottedwithrespecttotemperaturethereexistsacross-pointduetothettemperature dependencesfor D p and N p .Thecross-pointdescribesthetransitionofthenucleation ratefrombeingusiondominanttonucleationdominant.Sincethelengthof molecules,asevidencedinthewgrowth,is mlongevenatroomtemperature, moleculesexploretheenergylandscapeofthesurfaceandpreferentiallynucleateatthe stepsfollowedbystackingwithothermoleculestoformorderedstructureswhenthegrowth temperatureisbelowthecross-point,i.e.dominant(RegionIin Figure7.2 ).At 156 Figure7.2: Plotoftheandnucleationcontributionstotheoverallnu- cleationprobabilityoftwo-dimensionalclustersonthesurfacewithrespectto thesubstratetemperature .Thebluecurverepresentsthecontributiontotheoverall nucleationprobabilityduetoadmolecule(inthiscase E ds =55meV).Theredand blackcurvesconsiderthetcontributionstotheoverallnucleationprobabilityfrom theenergeticcostsforformingaclusterduetotheedgeenergyassociatedwithnucleating atthestep=56meV)orinthemiddleoftheterrace=62meV).Thesupersaturation level ˙ =100isheldconstantinbothcurves. temperaturesbelowthecrossingpointtheprobabilityforformingacriticalnucleusis low.However,asthegrowthtemperatureincreases,thenucleationtermstartstohavea moretimpactontheoverallprobabilityforstablenucleation,whichleadstoa transitioninthegrowthfrombeingwgrowth,totwo-dimensionalislandgrowth,and eventuallytorandomroughnucleation. Inordertounderstandthetransitionfromanisotropicwgrowthtotwo-dimensional islandgrowth(RegionIIin Figure7.2 ),itisnecessarytoconsiderhowthemoleculesare nucleatingonthesurface.Asillustratedin Figure7.3 ,theenergyforformingacritical 157 Figure7.3: Pcnucleationatthesurfacestepedgesorinthemiddleoftheterrace . (a)and(b)depicttwonucleationscenariosfortheobservedPcmoleculegrowth:onthe terraceandalongtheSistepedge.Theaverageedgeenergyassociatedwiththesetwo scenariosistduetotheamountofexposedsurfacearea.Nucleationatthestepedge wouldgivealoweraverageedgeenergyduetothelowerexposedsurfacearea. nucleusisconsideredheterogeneousonthesurfacewhereitissmallerfornucleiformedat thestepedgesandlargerfornucleiformedinthemiddleoftheterrace.Thisisduetothe edgeenergyofthemolecularstructuresassociatedwiththenucleation.Toillustrate,two nucleationprobabilitycurvesareincludedin Figure7.2 usingempiricalvaluesfor E ds , and ˙ .Whenthesubstratetemperatureislow,nucleationwilloccuratthestepedges(black curve).However,oncethesubstratetemperatureisincreasedhighenoughitispossibleto startformingnucleiinthemiddleoftheterraces(redcurve).Despitetheincreased ityofthemoleculeswithtemperature,whennucleistartforminginthemiddleoftheterrace, moleculeswillattachtoexistingnucleibeforethepreferentialnucleationsitesatthe stepedges,transitioningthegrowthawayfromanisotropicwgrowth.Sincenucle- ationatthestepedgescannolongeroccurattemperaturesbeyondthecrossingpointin Figure7.2 ,theblackcurvecannolongerbeconsidered.Insteadtheprobabilityforforming acriticalnucleusinthemiddleoftheterraceisused(redcurvein Figure7.2 ).InRegionII theprobabilityformoleculartoanucleusstilloutweighstheprobabilityforforming 158 Figure7.4: Graphicalrepresentationoftherelationshipbetweenthebar- rierandvdWbindingenergyformolecularnucleationformation .(a)Depictsthe relationbetweenthebarrierandbindingenergyforthegrowthoforderedmolec- ularstructures.ThebluecurverepresentsthevdWpotentialwiththepotentialminimum relatedtotheintermolecularbindingenergy.WhenthevdWpotentialminimumissig- tlylargerthantheusionbarrier(redcurve),anobviousenergyminimumisstill accessibleinthetotalenergy(blackcurve)allowingadmoleculestoandnucleateinto orderedstructures.(b)DepictsthecasewheretheenergyminimuminthevdWpotential iscomparableinmagnitudetothebarrieronthesurface.Adistinctglobalmini- muminthetotalenergyisnolongerpresent,resultinginrandomnucleationofdisordered molecularclustersontheterraces. criticalnucleiinthemiddleoftheterraces,makingthegrowthinthistemperaturerange dominant.Thisallowsfortheformationoflarge,ordered,two-dimensionalislands ontheterraces. Asthegrowthtemperatureincreasesbeyondthesecondcrossingpoint,theoverallproba- bilityforstablenucleationtransitionsfrombeingndominanttonucleationdominant (RegionIIIin Figure7.2 ),resultinginrandomroughgrowthontheterraces.Toexplain thenucleationofsmalldisorderedmolecularclustersconsiderthefollowingenergeticargu- mentsinvolvingthebindingenergyofanucleusandthemolecularbarrieronthe surface.Asthedisorderinnucleiincreaseswithentropy,thiswillresultinareductioninthe 159 Figure7.5: STMimagesofCuPcgrowthmodechangeswithincreasingtemper- ature .(a)-(c)STMtopographyimages( V s =+2.0V, I t =25pA)ofCuPcdepositedon theSi(111)-B p 3 p 3 R 30 surfacewithvaryingsubstratetemperatureduringdeposition. Thecoverageanddepositionratearekeptthroughouttheseexperiments.Transitions inthegrowthmodefromwgrowthtotwo-dimensionalislandnucleationtorandom roughgrowthareobserved. bindingenergyofadmoleculestothenucleus,i.e.thedepthofthevdWpotentialminimum, whencomparedtoahighlyorderednucleuswherethemolecularpackingmaximum ˇ - ˇ interaction.WhenthedepthofthevdWpotentialminimumislargerthanthe barrier,admoleculeswillbeabletotheglobalenergyminimumandformastableor- deredstructure( Figure7.4 (a)).However,whenthevdWpotentialminimumiscomparable orsmallerthanthebarrier,admoleculescannolongerthevdWpotentialmin- imumandinstead\see"manylocalminimainthetotalsurfacepotentiallandscape.This willresultinadmoleculesrandomlyadsorbingonthesurfaceforminglocalizeddisordered structures( Figure7.4 (b))[ 224 ]. Toidentifyother M Pcmoleculesthatfollowthesamegrowthtransitionswithsubstrate temperatureasZnPc,STMtopographyimagingwascarriedoutonsampleswithCuPc depositedonthedeactivatedSi(111)-Bsurfaceasshownin Figure7.5 .Observationsof CuPcgrowthdisplaythesametemperaturedependenceasdemonstratedwithZnPc.Thekey isthatthetemperaturesatwhichthegrowthmodetransitionsareobservedhave beenshiftedfromthecaseofZnPc.ForCuPcdepositionatroomtemperature,molecules 160 adoptthetiltedorderedpackingwheredomainsdisplaythecoexistenceof wgrowthandtwo-dimensionalislandgrowth( Figure7.5 (a)).Thisiscorroborated bythenucleationprobabilitycurvesin Figure7.2 whichdisplaythepossibilityofobserving thecoexistenceoftheanisotropicwgrowthandthetwo-dimensionalislandgrowth intheintermediatetemperatureregion.Increasingthetemperatureto110 C,onlylarge two-dimensionalislandsareformedwhicharerestrictedtotheterracethegrowthisinitiated in.BetweenthelargeCuPcislandssomeindividualmoleculeshaverandomlynucleatedon theterrace( Figure7.5 (b)).Fortemperaturesof180 Candabove,onlyrandomrough growthofCuPcisobserved( Figure7.5 (c)). Theseresultssuggestthatthebarrier,nucleationenergy,andbindingenergy needtobedelicatelybalancedtoachievetheobservedwgrowth.ForZnPconSi(111)- B,thisoptimalgrowthmodeisobservedbetween25 Cand170 C,whereasforCuPcthe optimalgrowthmodemustberealizedatatemperatureslightlylessthan25 Cbecauseat roomtemperaturetheanisotropicwgrowthandthetwo-dimensionalislandgrowth coexist.ThebetweentheZnPcandCuPcgrowthcouldbeattributedtothe tmolecule-substrateandmolecule-moleculeinteractionscausedbysubstitutingthe ZnmetalionforCu,whichimpactthenucleationandgrowthbehavior.Despitetheshiftin temperatureforobservingtheoptimumwgrowthmode,CuPcgrowthdisplaysthe sametemperaturedependenttrendasZnPcdepositedondeactivatedSi(111)-B. 161 7.2SmoothingofZincPhthalocyanineFilmsatEle- vatedTemperatures In Section5.3.3 ,defectsonthedeactivatedSisurfacecancausetheformationofrotational domainboundariesduringthegrowthofZnPc,resultingingrainboundarycrossingbarriers whichactasactivationbarriersintheprocessandevenintroducenucleationsites foradmolecules.Thiscanbedetrimentalforformingahighlyorderedmolecularthin becauseitcaninduceatransitionfromtheideallayer-by-layergrowthtodomainsnucleating priortothecompletionofthepreviouslayer,asisthecaseforZnPc.However,itispossible topreventthisrougheningbyborrowingfromsomeoftheclassicalnucleationtheory discussionin Section2.3 .Despitesomeoftheobservedbetweeninorganic growthandthetemperaturedependentgrowthfor M PconthedeactivatedSi(111)-Bsurface, onecommontrendthatcanbetakenadvantageofistheincreaseinthesizeofdomainsformed withacorrespondingreductioninthenucleationdensityasthesubstratetemperatureis increasedduringmaterialdeposition[ 91 , 92 ].Meanwhile,admoleculesaremorelikelyto overcomethegrainboundarycrossingbarrieratelevatedsubstratetemperatures.Both ofthesewillresultinatreductioninthedensityofrotationaldomain boundariesformedand,thus,provideasmootherthinmorphology. Asshownin Figure7.6 ,monolayercoverageofZnPconthedeactivatedSisurfacetransi- tionsfromroughinterruptedwgrowth(roomtemperature)toasmoothobtained atagrowthtemperatureof110 C,wherethemajorityoftheZnPcdomainsareabletoex- tendlaterallyuntilbeingterminatedattheSidefectlines(highlightedbythedashedblack ovalsin Figure7.6 (b)).However,asthethicknessisincreasedwiththesubstrate temperatureheldat110 C,again,roughnessintheZnPccontinuestopersistasseen 162 Figure7.6: EvidenceofZnPcsub-monolayersmoothingbysubstratetemper- aturecontrol .STMtopographyimages( V s =+2.5V, I t =20pA)ofZnPcdepositedonthe Si(111)-B p 3 p 3 R 30 surfaceatacoverageof0.9monolayers.(a)Depositiononthesur- facewiththesubstrateheldatroomtemperatureresultsintheformationofamultilayered structurewherethenextlayerbeginstoformpriortothecompletionofthepreviouslayer. (b)Increasingthesubstratetemperatureduringdepositionto110 Creducesthenumber ofgrainboundariesandenablesgrainboundarycrossing,allowingfortheformationofa completemonolayer.DefectsfromtheSisurface,e.g.theatomicSidefectlines,propagate intotheZnPclayerashighlightedbythedashedblackovals. in Figure7.7 (a).Eventhoughelevatingthesubstratetemperatureto110 Cntly reducesthenumberofrotationaldomainboundariesformed,butenoughstillremaininthe whichcontinuetopropagatethroughtheasthethicknessincreases,resultingin theroughin Figure7.7 (a).Ifthesubstratetemperatureisincreasedhighenough suchthatnucleationalongthegrainboundariescanbeeliminatedthroughouttheonly contributionsfromtheSidefectlinesthatseparatethemoleculardomainswillbepresent. 163 Figure7.7: EvidenceofZnPcmultilayeredsmoothingbysubstratetempera- turecontrol .STMtopographyimages( V s =+2.5V, I t =20pA)ofZnPcdepositedonthe Si(111)-B p 3 p 3 R 30 surfaceatacoverageof5monolayers.Theapparentheightofeach layerisillustratedinthehistogramplotsforthecorrespondingSTMimages.Peaksinthe histogramplotsarenumberedwiththecorrespondinglayerassignmentinthetopography imageswherethe\+"indicatesZnPclayergrowthontheupperSiterraces.Similartothe sub-monolayergrowth,increasingthesubstratetemperatureduringdepositionresultsinthe formationofsmootherascanbeseenbycomparingtheSTMimagesin(a)and(b). Duetothegrowthtransitionsdiscussedin Section7.1 ,selectionofthesubstratetemper- aturemustbedonecarefullysuchthatthePcmoleculescontinuetomaintainthehighly orderedtiltedmolecularpackingion.In Figure7.7 (b),maintainingthesubstrate temperatureat200 CresultsinZnPcgrowththatisonlylimitedbytheSistepedgesand Sidefectlines.ThoughonagivenSiterracethemoleculargrowthisverysmooth,anad- ditionalrougheningmechanismbeginstotakenamelysteeringandshadowing previouslymentionedin Section5.3.2 .Inthiscase,asmoleculesaredeposited,theybecome 164 restrictedtotheterracetheylandinduetothestrongSistepedgeESBwhichcanyield varyingbuildupofmthicknessoneachterrace,ascanbeseenin Figure7.7 (b).Someof thisshadowingcouldberemovedbyloweringthemis-cutangleofthedeactivatedSi surfacetoproducelargeseparationbetweentheSistepedges.Nevertheless,theseresults showthatelevatingthesubstratetemperatureduringgrowthreducesthenucleationdensity andconsequentlythenumberofgrainboundaries,yieldingasmoother 7.3Conclusions Thesetemperaturedependentgrowthstudiesof M PcmoleculesonthedeactivatedSi(111)- Bsurfacedemonstratethedegreeofcontroloverthemolecularpackingandstructures thatcanbeachievedforthissystem.Throughsubstratetemperaturecontrol,the M Pc growthmodecanbemoFurthermore,thesizeanddensityoforderedmoleculardo- mainscanbealteredwithinalimitedtemperaturerangesuchthatthenumberofdefects anddomainboundariesformedistlyreduced.Ultimately,thisallowsforthefor- mationofasmooth,highly-orderedmolecularthinTheseresultsgiveanoverviewof strategiesforinterpretationandimprovingthemoleculargrowthwhichcouldbeapplied tootherorganicmolecularsystem.Thoughclassicalnucleationtheorycouldbeappliedto formulateanunderstandingofthe M Pcmoleculargrowth,thereversetrendinthegrowth modetransitionsfortheorganicmolecularsystemscomparedwithtypicalinorganicgrowth observationswarrantfurtherinvestigationthroughtheoreticalcomputationalmethods. ThetransitionsinthegrowthmodesforZnPcandCuPcmappedinthisstudyprovide temperaturerangesthatcanbeusedtoformhighlyorderedorganicmolecularthinon alargescale.Forexample,theanisotropicwgrowthmodeexhibitedbyZnPccanbe 165 extendedtotemperaturesnearing170 C,allowingfortheformationofverylargemolecular domainswithanexclusivein-planemolecularordering.Additionally,growthatthiselevated temperaturewillallowforthemultilayered M Pcgrowthtohaveasmoothmorphologywhere themolecularpackingismaintainedevenforcientlythickByreducingthedensity ofSiatomicdefectlinesandthemis-cutangleofthesubstrate,thegrowthcouldbe evenfurtherimprovedbeyondthis,highly-orderedorganicmolecularthatcan beapplicableonscalesrelevantforindustrialuses. 166 Chapter8 MetalPhthalocyanineGrowthon DeactivatedSi(111)-BMediatedby SelectiveOrbitalCoupling Chapter5 through Chapter7 exploredthegrowthevolutionof M Pconthedeactivated Si(111)-Bsurface,particularlyZnPc,fromsub-monolayercoveragestomultilayeredmolec- ularandalsoatvaryingsubstratetemperaturesduringmoleculardeposition.From thesechaptersitisapparentthatthemolecularqualitycanbemobytemper- ature.However,anotherroutetowardsmodifyingthegrowthevolutionof M Pcis tochangethemolecule-substrateinteraction.Thisisthesubjectofthefollowingchapter. In Section8.1 ,thegrowthoft M PcmoleculesisexploredusingSTMtopography imaging. Section8.2 providesanexplanationofthemolecularbindingmechanismfor M Pc molecules,whichultimatelygovernstheirgrowthevolutiononthedeactivatedSisurface. Howthismechanismmothegrowthevolutionisexploredin Section8.3 throughDFT calculationsandanalysis.Finally,theresultsareconcludedinthelastsection.Partsofthis chapterareadaptedfromReference[ 43 ]:S.R.Wagner etal ., Phys.Rev.Lett. 115 ,096101 (2015). 167 8.1GrowthComparisonbetweenZinc,Copper,and CobaltPhthalocyanine Fordevicearchitecturesmakinguseofsmallorganicmolecules,thereisintenseinterestin understandingandmanipulatingtheinterfacialelectronicstructureaswellasthemolecule- substrateinteractionthatgovernstheorganicmoleculargrowth[ 39 , 40 , 41 , 42 , 46 , 44 , 47 , 45 , 48 , 225 , 226 , 227 ].Ofallsubstratetypesthat M Pcmoleculeshavebeendepositedon, metallicsurfaceshavebeenmostextensivelystudied,duetothecommonuseofmetallic contactsindevicearchitectures.Noblemetalsurfacesprovideyintuningthe M Pc- substrateinteractionthroughorbitalhybridizationand/orchargeredistributionmediated throughthecentralTMionofthePcmolecule[ 45 , 49 , 50 ].However,despitethevariation inallowedorbitalhybridizationforthesesystems,thereislittleontheorientation aswellastheorderingof M Pcmoleculesonthesurface. M Pcmoleculesonmetallicsurfaces commonlyadoptaclose-packedmolecularstructureforthemonolayerwhich thengraduallyrelaxesbacktothebulk M Pcpackingformultilayergrowth [ 46 , 51 , 52 , 54 ]. Fromthestudiesdetailedthusfarin Chapter5 through Chapter7 ,thedeactivated Si(111)-B p 3 p 3 R 30 surfacefacilitateslong-rangeorderedgrowthofZnPcinboth thein-planeandout-of-planedirections[ 39 , 41 ].However,itisstillunknownwhatthe underlyingmolecule-substratebindingmechanismisandwhatrolethismechanismplaysin thelong-rangemolecularordering.Additionally,ifthismechanismcanbeidenitis unknownwhetheritcanberationallytunedsuchthatitimpactsthethinmorphology. Inlinewithpreviousstudiesonmetallicsurfaces,here,varioustypesof M Pcmolecules ( M =Zn,Cu,andCo)aredepositedonthedeactivatedSi(111)-Bsurface. Figure8.1 168 Figure8.1: Largescalecomparisonofmoleculardomainformationconsistingof ttypesof M Pcmolecules .(a)-(c)LargescaleSTMtopographyimagesofsub- monolayercoverageof(a)ZnPc,(b)CuPc,and(c)CoPcdepositedonthedeactivatedSi surfacewiththesubstratemaintainedatroomtemperatureduringthedepositionprocess. ZnPcexhibitsthehighlyorderedtiltedmolecularasdiscussedin Section5.1 . CuPcdomainsmaintainthetiltedmoleculartionbutdisplaydefectedregions. CoPcshowsbothtiltedandmolecularorientations.Scanningconditionsfor(a)- (c): V s =+2.0-2.5V, I t =4-30pA. presentsthedistinctmoleculargrowthbehavioronalargerscaleofeach M Pcmolecule typewhile Figure8.2 displaysthegrowthonalocalscale.Ascanbeseenin Figure8.1 (a)and Figure8.2 (a),ZnPcmoleculesadoptthetiltedmolecular asdiscussedextensivelyin Chapter5 .Toguidetheview,greenbarmarkersareincluded intheSTMimagein Figure8.2 (a)whichrepresentindividualZnPcmoleculeswithinthe tiltedmoleculardomain.Ontheotherhand,forCuPcthereisanotabledegradationinthe qualitywhereatamountofgapdefectsandshiftsinthemolecularpacking occurwithinthemoleculardomainsascanbeseenin Figure8.1 (b).Despitethedefectsin themoleculardomains,theyareabletostillmaintainthesamemolecularorientationand packingmotifastheZnPccase[ 39 , 40 , 41 ]. Figure8.2 (b)clearlyillustratesthattheseshifts aretheresultofzig-zagpatternedgrowthofCuPcmolecules,i.e.thegreenandbluebars aswellastheregionoutlinedbydashedwhitelineshighlightedintheSTMimage.These shiftsinthemolecularpackingcanleadtothedevelopmentofgapdefectsasoutlinedby 169 Figure8.2: Comparisonofmolecularpackinganddefectswithindomainsconsist- ingoferenttypesof M Pcmolecules .(a)-(c)ExpandedSTMtopographyimages of(a)ZnPc,(b)CuPc,and(c)CoPcshowingthepackingof M Pcmoleculeswithinatyp- icaldomain.ZnPcexhibitsthehighlyorderedtiltedmoleculardiscussedin Section5.1 .CuPcdisplaysshiftsinthemolecularpackingwhichareoutlinedbythedashed whitelinesandgapdefectsashighlightedbythedashedredovals.CoPcshowsbothtilted andgmolecularorientations,withthetilteddomainoutlinedbythedashedgreen lines.Thegreen/bluebarsin(a)-(c)representindividualPcmoleculeswithtiltedorienta- tion.Molecularassignmentin(c)isfurthercorroboratedin Figure8.3 . thedashedredovalsin Figure8.2 (b).GrowthofCoPc,however,isdrasticallyt fromthatofZnPcandCuPc.Moleculespredominantlylieonthesurface. Figure8.3 (a)includeslinetakenonZnPcandCoPcstripesintheregionsshownintheSTM topographyimagesin(b)and(c).Thelinedistinguishtheeintheapparent heightbetween M Pcmoleculeshavingat-lyingorientationandthoseinthetiltedpacking Itisnotuntilthecoverageisincreasedaboveacertainthresholdthatsmall disorderedstripesformexhibitingthetiltedasillustratedin Figure8.1 (c) and Figure8.2 (c).Thebluebars Figure8.2 (c)indicatethelocationofindividualtilted moleculessimilarto(a)and(b)ofthesameHowever,inthecaseofCoPcitisa littlemorechallengingtoassignthemoleculesduetothedisorderwithinthesesmallstripe domains. Toaddressthisissueandaccuratelyassignthetiltedmoleculesin Figure8.2 (c),com- parisonwithSTMimagesofZnPcandCuPcstripedomainsiscarriedout,spfor 170 Figure8.3: Shiftsin M Pcmolecularpackingandapparentheightofmolecules inaortiltedorientation .(a)LineptakenonZnPc(red)andCoPc (blue)tiltedmoleculardomainsonthedeactivatedSi(111)-Bsurface.Thelinefor CoPcalsodisplaystheapparentheightofCoPcmoleculesintheration. (b)STMtopographyimage( V s =+2.0V, I t =20pA)ofatiltedZnPcmoleculardomain ontheSi(111)-Bsurface.Theredlinecorrespondstothelineeregionin(a).(c) STMtopographyimage( V s =+2.5V, I t =20pA)ofsub-monolayercoverageofCoPc onSi(111)-B.Thebluelinecorrespondstothelineregionin(a).Thegreen,blue, andredbarsindicatetheassignmentofindividualmoleculeswithinmoleculardomains.(d) STMtopographyimage( V s =+2.5V, I t =20pA)ofsub-monolayercoverageofCoPcon Si(111)-B.Thisisazoomed-outimageinaregionnearbytheoneshownin(c).Thedashed bluerectangleoutlinesaregionofCoPcmoleculeswhichshowsasimilarpackinggeometry totheorderedregionsofZnPcandCuPc. instanceswheretheshiftsinthemolecularpackinggivetheappearanceofa\hexagonal" geometry.AZnPcstripeisshownin Figure8.3 (b),wherethebluebarsrepresentindi- vidualmoleculeswithinthedomainthatarerotatedwithrespecttothemolecularordering (greenbars)oftheprimarygrowthdirection(greenarrows).Thisresultsina\kink"for- mationinthemolecularstripewitha60 angleinthegrowthdirection(blue arrow),producingtheobserved\hexagonal"geometrynearthechangeinthepacking,as highlightedbythepurplehexagon.NotethatwhentheZnPcdomainsbecomelarge,this 171 typeofmolecularshiftobservedintheearlystagesoftheZnPcgrowthisself-healedwith increasingmolecularcoverage.InthecaseofCoPc,thestripesaretypicallyverynarrowwith atamountofmolecularshiftswithinthestripes,asindicatedbythegreenandred bars(individualtiltedCoPcmolecules)in Figure8.3 (c).Again,thegreenbarsrepresent moleculespackedalongtheassignedprimarygrowthdirection(greenarrows)whileinthis casetheredbarsareindividualCoPcmoleculesthatarelaterallyfromtheprimary packingSimilartotheZnPccase,theselateralmolecularshiftsalsoinduce \kinks"inthedomainwitha60 anglenceinthegrowthdirectionandgivesthe appearanceofa\hexagonal"structure.DuetothehighincidenceofCoPcmolecularshifts thetilteddomainscancloselyresembletheunderlyingSi(111)-Bsubstrate.However,itis stillpossibletoobservesomeCoPctiltedmoleculardomainswhichdisplayregionswithout thesemolecularshifts,thusremovingthe\hexagonal"geometry,ascanbeseenintheregion highlightedbythebluedashedrectanglein Figure8.3 (d).Nevertheless,thehighdegreeof disorderintheCoPcmoleculargrowthonthescalesshownin Figure8.1 and Figure8.2 continuetopersistonevenlargerscales. Thedrasticallytmolecularassemblyandgrowthbehaviorobservedforthevarious M PcdepositedonthedeactivatedSisurfaceisinsharpcontrasttothegrowthofthese moleculesonmetallicsurfaces[ 46 , 51 , 52 ].Theseresultssuggestthatthecriticalinteractions governingthemoleculargrowthcouldbemediatedbythecentralTMion.Thisisfurther corroboratedbytheintheinitialadmoleculeadsorption.Asdetailedinprevious chapters,theZnPcareabletorapidlyacrosstheSiterracesandnucleateatthe energeticallyfavorableSistepedgesitespreventingtheobservationofsingle,ZnPc moleculesunlesstheyadsorbatdefectsitesintheSiterraces[ 39 , 40 , 41 , 42 ].CuPcmolecules continuetodisplayalargeityonthedeactivatedSisurface,however,thereisanotable 172 Figure8.4: Individual M Pcmoleculesintheorientation .(a)and(b)are STMtopographyimagesof(a)CuPcand(b)CoPcmoleculestakeninregions betweentiltedmoleculardomains.ThethreestableCoPcorientationsallowed bythesubstratesymmetryarehighlightedbythered,blueandgreendashedrectangles, exhibitingregistrationwiththeunderlyingsurface(solidrectangles).Thedashedwhite circlehighlightsasinglemoleculerotatingbetweenall3orientations.(c)and(d)areSTM topographyimageofasingle(c)CuPcmoleculeand(d)CoPcmoleculeonthe deactivatedSi(111)-Bsurface.TheSi(111)-B p 3 p 3 R 30 unitcellisoverlaidinthe imagesin(c)and(d).Scanningconditionsfor(a)-(d): V s =+1.3-2.0V( V s =-2.0Vin(c)), I t =3-25pA. increaseinthenucleationoforientedmoleculescomparedtoZnPc,whichexceeds theexpecteddefectdensity.IsolatedCuPcmoleculesareshownin Figure8.4 (a),withahighresolutionimageofanindividualmoleculeshownin(c).Thefourbenzene- pyrroleringsoftheindividualCuPcmoleculecanbeidenwhilethecentralCumetalion appearsasadepressionintheSTMimage[ 51 ].FromacollectionofSTMimagesofindividual CuPcmolecules,itispossibletodeducethatthesemoleculesdonotdisplayspesites forregistrationwiththeunderlyingdeactivatedSisurface,suggestingthattheinteraction 173 betweentheCuPcmoleculeandthead-Siisonlymildlystrongerthanthatexperiencedby ZnPc.Incontrast,CoPcmoleculesarecommonlyobserved,displayingastrong registrationwiththeunderlyingad-Siatoms,ascanbeseenin Figure8.4 (b)and (d).BoththeCoionsandthefourbenzene-pyrroleringsoftheindividualCoPcmolecule preferentiallyregisteronad-Siatoms,wheretheCoionappearsasabrightprotrusioninthe STMimagesduetoorbital-mediatedtunneling[ 51 , 228 ].ComparingtheindividualCuPc moleculein Figure8.4 (c)withtheCoPcmoleculein(d)showsthatthesymmetryofthe CoPcmoleculeisreducedfrom C 4 to C 2 ,givingtheappearanceofanapparentrectangular shapeintheSTMimaging.FromthehighdegreeofCoPcmolecularregistrationalong thecommonsymmetryelementswiththeunderlyingsurface,whichdisplays C 6 symmetry, onlythreestableorientationsareobservedasindicatedbythedashed/solidred, blue,andgreenrectanglesin Figure8.4 (b).Often,asinglemoleculecanbeobservedtobe rotatingbetweenallthreetriggeredbyfasttunnelingfromtheSTMtip[ 229 ]. Suchacaseisindicatedbythedashedwhitecirclein Figure8.4 (b).Thecombinedsingle moleculeandlarge-scalemoleculargrowthobservationssuggestastrongmolecule-substrate interactionbetweentheCoPcmoleculesandthedeactivatedSi(111)-Bsurface. 8.2MolecularBindingMechanismtotheSi(111)-B From Section8.1 itisapparentthattheinteractionsinvolvedinthemoleculargrowthprocess couldbemediatedbythecentralTMion.Tounderstandwhy,considerthecentralTM ionwithinafree-standing M Pcmolecule.TheinteractionbetweentheTMionandligands formedbythefourinwardlyprojectingnitrogencentersresultsinanattractionbetweenthe TMionandtheligandgroupstoformaTMcoordinationcomplexinthecenterofthePc 174 molecule.Consideringthe d -orbitalsoftheTMion,thereareedegenerate d -orbitals. WhenagroupofligandssurroundtheTMion,theelectronsofligandgroupswillbecloser tosomeofthe d -orbitalorientationsandfartherawayfromothers.Thiswillmanifesta variationintherepulsionbetween d -orbitalelectronsandtheelectronsassociatedwiththe ligandgroupswhichwillultimatelyresultinenergysplittingofthe d -orbitals.Fromthe D 4 h symmetryofthe M Pcmoleculethiswillinduce d -orbitalsplittingintoadoublydegenerate state( d xz , d yz )andthreesinglydegeneratestates( d xy , d x 2 y 2 ,and d z 2 )where d xz , d yz , and d z 2 protrudefromthemolecularplane[ 230 , 231 ].AstheTMionischangedfromZnto Co,thechargedistributionwithineachorbitalisalteredasshownin Figure8.5 (a).Note thatwhentheTMionisCo,the d z 2 orbitalbecomesdepopulated[ 231 , 232 ]. Afterconsideringthe d -orbitalsplittingofafree-standing M Pcmolecule,tounderstand howtheseTM d -orbitalsthemolecularbindingtothesubstrate,andconsequently, thegrowthbehaviorof M PcmoleculesonthedeactivatedSi(111)-Bsurface,DFTcalcula- tionsarenecessary.Forinformationregardingthecomputationprocessandconsiderations usedforthisstudypleasereferto Section3.3.3 .FromtheSTMimageanalysisin Section8.1 , theinteractionbetweentheCuPcmoleculeandthead-Siisonlymildlystrongerthanfor ZnPc,yetbothZnPcandCuPcarefoundtoformordered,tiltedmoleculardomains.Since thegrowthisdistinctlytforCoPcfromtheothertwo M Pcmoleculesaswellas indicatingsignsofaverystrongmolecule-substrateinteraction,thefocusofthemolecular bindingmechanismwillinitiallyrevolvearoundCoPcsinceitistheonlymoleculewithan unpopulated d z 2 orbital.DFTcalculationsindicatethatthe d z 2 orbitalisstronglylocalized ontheCoion,butmoreimportantlythesingly-occupied d z 2 orbitalisthemajorsymmetry- allowedorbitaldisplayingstrongoverlapwiththeempty p z stateofad-Si.Thus,astrong bondisanticipatedtoformresultinginatlyhigherenergygaincomparedtoZnPc 175 andCuPcadsorption. Figure8.5 (b)illustratestheorbitalhybridizationschemebetween theCoPcmoleculeandthedeactivatedSi(111)-Bsurface.Fromthediagram,theenergy Figure8.5: Schematicsofthe M Pcorbitalhybridizationmechanism .(a)showsa schematicdiagramofthe d -orbitalforZnPc,CuPcandCoPc.(b)illustratestheorbital hybridizationmechanismbetweenaCoPcmoleculeandthedeactivatedSi(111)-Bsurface. Notethattheback-bondsurfacestate,thediscreteenergylevelinpink,isinresonancewith thebulkSivalenceband. leveloftheCo d z 2 orbitalislocatedwithinthebulkSibandgapandthead-Si p z orbitalis locatednearthebulkSiconductionbandminimum[ 233 ].However,whenCoPcisadsorbed onthesurface,the p - d orbitalcouplingmechanismresultsinahybridized p z - d z 2 bonding statewithalowerenergythanthebulkSivalencebandmaximum.Thelowerenergyofthe hybridizedstateallowsforoneelectrontobetransferredfromthesubstratetothebonding statemakingitfullyoccupied. Theemergenceofthe p z - d z 2 hybridizationisapparentincalculatedchargedensity enceplotsofthet M PcmoleculesadsorbedonthedeactivatedSisurface.ForZnPc andCuPc( Figure8.6 (a)and(b))onlyanegligibleorminuteamountofchargeaccumula- tionisobserved,respectively,asshownbytheblueandredregionsinthecalculatedplots. However,forthecaseofCoPc( Figure8.6 (c)),thestrongorbitalhybridizationintroducesa tchargeaccumulationonthe p z orbitalsofad-Si.Itshouldbenotedthatforthe calculatedchargedensityplots,allowedlocalenergeticminimamolecular rationsforZnPcandCuPcareselectedsuchthattheyarecomparabletoCoPc.Thisisdone 176 Figure8.6: Chargedensitymapsof M Pcmoleculesadsorbedonthe deactivatedSi(111)-Bsurface .(a)-(c)Calculatedchargedensitymapsof molecularadsorptionof(a)ZnPc,(b)CuPc,and(c)CoPconthedeactivatedSi(111)-B p 3 p 3 R 30 surface.Thesameisosurfacechargedensityvalue( 0.001 e= Bohr 3 )is selectedforalloftheseplots.Redregionsdenoteelectronaccumulationandblueregions indicateelectrondepletion.Thegreenatomsrepresentad-Siwhilethepinkatomsdenote sub-surfaceboron. inordertotiatetheinteractionschemesbetweenthecentralTMionandthead-Si. Themagnitudeofthechargeredistributionissetbytheenergylevelbetweenthe p z orbitalofad-Siandthe d z 2 orbitalof M Pcwhichultimatelydeterminesthedegreeof mixingbetweenthetwoorbitals.Thisallowsforthenetchargetransferfromthesubstrate tothe M Pcmoleculestobedetermined.Fromtheanalysis,thenetchargetransferforCoPc andZnPcadsorptiononthedeactivatedSisurfaceiszero.Interestingly,CuPcexperiencesa smallamountofnetchargetransfer( < 0 : 1 e ),wherechargedonationtotheempty p z state isbalancedbypartialchargeback-donationfromthesubstratetothesingly-occupiedCuPc d x 2 y 2 orbital( Figure8.5 (a)). 177 8.3Transition-MetalIonInducedModulationinthe PotentialEnergyLandscape Withtheunderstandingthatthe p - d orbitalhybridizationbetweenthe M PccentralTM ionsandthead-Sideterminesthemolecularbindingtothesubstrate,thefocuscannow beshiftedtowardshowthisbindingmechanismcancethemoleculargrowth.This requirescalculationsofthepotentialenergylandscapeofthesurfaceforthet M Pc moleculesexaminedalongselectivedirectionswheretheinternalmolecularcoordinationand rotationarefullyoptimizedateachpointalongthepathway.Initially,theCuPcmolecule isconsideredonthedeactivatedSisurface.Notethatsimilarpotentialenergylandscape trendsareexpectedforZnPc. Figure8.7 (a)illustratesonepossiblepathwayfor themoleculetomoveonesubstrateunitcellwithagiveninitialmolecularorientation, whileitsorientationremainsunchanged.Thepotentiallandscapealongthispathwayis representedin Figure8.7 (b).Thedotsinthegraphofthepotentiallandscapecorrespond tothesnapshotsofthemolecularshownin Figure8.7 (c).From Figure8.7 (b) itcanbeseenthatthevdWenergyminimum(redline)doesnotcorrespondwiththetotal energyminimum(blueline).Thisisbecausethereisaweakattractiveinteractionbetween theNatomsofthePcmoleculeandthesurfacead-Si.Thoughthisinteractionisweakit canstillhaveanimpactonthemolecularascanbeseenbycomparingthe tpanelsin Figure8.7 (c).Anobviousdistortionofthemolecularplanewheread-Si arenearthe M PcNsitescanbeobservedwhencomparingtheparticularly thethirdpanelandpanel.ThesetwopanelscorrespondtothevdWenergyminima andthetotalenergyminima,respectively.TheCuPcmoleculedonthedeactivated Sisurfaceexperiencesmanylocaladsorptionminimawithsimilarbindingenergyvalueson 178 Figure8.7: M PcvdWinteractionwiththedeactivatedSisurfaceandrotational potentialenergylandscapecomparison .(a)displaysthepathwayofaCuPc moleculeonthedeactivatedSisurface.Thepotentiallandscapealongthepathin(a)is representedbythesolidbluelinein(b).Thesolidredlinein(b)correspondstothevdW contributiontothepotentiallandscape.SelectatomicgeometriesoftheCuPcmolecule alongthepotentiallandscape(dots)in(b)areshownin(c).In(d)and(e)thebluelines displaytheDFTtotalenergyoftherotationalpotentiallandscapefor(d)CuPcand(e) CoPc,withthevdWenergycontributionsplottedinred.(f)Theenergyminimum( =0 ) ofCoPc.(g)Theenergymaximum( =15 )ofCoPc.The colorcodingin(c),(f),and(g)representstheatomicheights(in A)abovead-Si. thesubstratesurface.However,themolecularatmostoftheselocalminima donotshowanyregistrywiththesubstrate,similartotheSTMobservationspreviously discussed. Duringtheprocessthe M Pcmoleculesareabletorotate,thus,itisnecessary toinvestigatetherotationalpotentialenergylandscapesofthe M Pcmoleculesonthead- Sisurface.First,consideraCuPcmoleculeadsorbedonthedeactivatedSisurfaceinone ofthelocaltotalenergyminimashowninthepanelof Figure8.7 (c). 179 Figure8.8: inthepotentialenergylandscapesfort M Pc moleculesadsorbedonthedeactivatedSisurface .(a)showstheenergylandscapes ofCuPcalongthetpathways.Linesconnectingthedatapointsareshown toguidetheview.Inset:Depictsthelinearlyindependentminimumenergypathwaysfor molecularThead-Sipositionsaredenotedbytheorangecircles.Dependingon theinitialmolecularorientation,therolesofthetwopathwayscanbeexchanged.(b)the energylandscapeofCoPcalongthesolidpathshownintheinsetin(a). ThisisselectedtobecomparabletothatofCoPcinordertoexcludeother contributionsandconcentrateonidentifyingtheoriginoftherotationalbarrier. Figure8.7 (d)displaysboththecalculatedtotalenergyandthevdWcontributiontotherotational potentiallandscapefortheCuPcmoelcule.Fromthecalculationtherotationalpotential barrieris ˘ 0 : 3eVandappearstobeexclusivelyfromthecorrugationinthevdWenergy (redline).Thisrotationalbarrierisapparentinthepotentialenergylandscape(dashed blueline)calculatedalongthedashedpathshownin Figure8.8 (a).Inthiscase,theCuPc moleculeundergoesarotationchangetominimizeitsenergybarrieralongthis pathway.Thisproducesarotationalbarrier(peakatthebeginningofthedashedpotential landscape)andonceitisovercome,theCuPcmoleculefollowsthesamepathasthesolid bluelinein Figure8.8 (a)(redrawnfrom Figure8.7 (b)),wherethemolecularorientation remainsunchanged.DespitethesesmallenergybarriersmainlyduetothevdW 180 energycorrugationfromtheadaptiverelaxationoftheCuPc/ZnPcmoleculeandtheweak interactionfromNatomswiththesurface,unhinderedmolecularofCuPc/ZnPc moleculesonthedeactivatedSi(111)surfacecanstillresultduetointermolecularattraction. NowconsidertheadsorptionofaCoPcmoleculeonthedeactivatedSisurface.Dueto thestrong p - d orbitalhybridizationandregistrationwiththead-Si,thereisaciently deeppotentialwellwithatenergybarrier,asevidentintherotational potentialenergylandscapeaswellasthetranslationalpotentialenergylandscapealongthe easiestpath(solidbluearrowin Figure8.8 (a))shownin Figure8.7 (e)and Figure8.8 (b) (bluelines).Fromthesecalculationstheenergeticminimais ˘ 1 : 2eVlowerthanthevdW energycontribution(redlines).Thelargeenergybarrieroriginatesfromthebreakingofthe chemicalbondbetweentheCoandthead-Siatom.Theconsequenceofthislargebarrierisa totalpotentialenergylandscapewhichwillstronglylocalize M Pcmoleculesonthesurfaceas observedin Figure8.4 (b)and(d).Fromtherotationalpotentialenergylandscape,though itisevidentthatthereishigherCoPcadsorptionenergythanCuPc,thecorrugatedsignal inthetotalenergyasthemoleculerotatesremains.Thisisbecausethemoleculeadaptively relaxesitselftomaximizetheadsorptionenergy.Sincead-Siisconsidereddeactivated,they onlyexhibitweakvdWinteractionwithnearbycarbonandhydrogeninthe M Pcmolecule hinderingmolecularapproachtothesurface.Moleculardistortionsandchargeoverlapcan arisebycompromisingthevdWenergygainfromthesurfacetomaximizetheadsorption energy.In Figure8.7 (f),thefourpyrrole-benzeneringsoftheCoPcmoleculeevadead-Si sitesbybendinginwardallowingformaximumapproachtothesubstrate.Thisresultsin anenergyminimumat =0 forCoPc.Themolecularrelaxationlowersthe symmetryofCoPcfrom C 4 to C 2 ,causingthemoleculetodeviateconsiderablyfromthe squareshapeasobservedintheSTMimagesin Figure8.4 (b)and(d).Similarobservations 181 havebeenreportedfor M PcmoleculesdepositedonCu(111)[ 234 , 235 ],however,inthese cases,themolecularbendingwasatleasttwotimessmallerduetothesmallmetalliclattice parameter.WhentheCoPcmoleculeisrotatedto =15 ,thead-Siatomspreventthe moleculefrombendingtoallowformaximummolecularapproachtothesurface,ascanbe seenin Figure8.7 (g).ThedegreeofmolecularrelaxationastheCoPcmoleculerotatesis whatgivesrisetothepronouncedvdWpotentialcorrugation. Table8.1: M Pc- M Pcand M Pc-Sibindingenergy(BE). MPctype MPc-SiBE phase[ 125 ]BE phase[ 125 ]BE ZnPc 3.18eV 3.92eV 4.00eV CuPc 3.04eV 3.85eV 3.88eV CoPc 4.37eV 3.90eV 4.00eV Themajorbetween M Pcgrowthonmetallicsurfacesandthedeactivated Si(111)-Bsurfaceutilizedinthisstudyisthepotentialenergylandscape,spallyits corrugationandhomogeneity.Typicallythecalculatedcorrugationinthepotentialenergy landscapeformetallicsurfacesisanorderofmagnitudesmallerthanthedeactivatedSi surfaceforevenweaklyinteracting M Pc.Forexample,ifthepotentialenergylandscapefor aCuPcadsorbedonAu(111)andthedeactivatedSisurfacearecalculatedandcompared, itisquiteobviousthatthemetallicsurfaceprovidesanotablysmootherpotentialenergy landscape,asshownin Figure8.9 .Theinthesepotentialenergylandscapeswill haveadramaticonhowthemoleculargrowthwillproceed.OntheAu(111)surface,as themolecularcoverageincreasestheintermolecularinteractionsresultinanenergywhichcan befullyachievedduetothepotentialenergylandscape.Thisisultimatelywhatdrives theformationofclose-packed, M Pcmolecularstructureswhicharecommonly observedonsurfaceslikeAu(111)[ 46 , 48 , 51 ].OnthedeactivatedSi(111)-B,the M Pc moleculesmustinitiallyadsorbatlocalenergyminimumsitesduetothecorrugationinthe 182 Figure8.9: Comparisonbetween M Pcgrowthonametallicsurfaceandthedeac- tivatedSisurface .Schematicpotentialenergylandscapes(leftpanel)ofaCuPcmolecule ontheAu(111)(orangecolor)andSi(111)-B(lightbluecolor)surfaces,andtheresulting morphologies(rightpanels)ofCuPcaggregates(bluestars)withincreasingmolecularcov- erage. potentialenergylandscape.Asthemolecularcoverageincreases,thelocalenergyminimum siteswillbecomeoccupied,leavingonlyunfavorableadsorptionsitesforincoming M Pc molecules.Atthispointthedegreeofcorrugationinthepotentialenergylandscape,i.e. therespectivemolecule-substrateinteractionforthet M Pcmolecules,willdetermine howthemoleculargrowthproceeds(see Table8.1 fortherelativeenergycomparisonsforthe t M Pcmoleculesusedinthisstudy).Ifthemolecule-substrateinteractionisweak, asforthecaseofZnPcandCuPc,thenitispossibleforthemolecule-moleculeinteractionto overcomethepotentialenergylandscapecorrugationstomaximizethe ˇ - ˇ intermolecular attraction.Bymaximizingthisintermolecularattraction,aportionofthesurfaceadsorption energyishowever,thisresultsinastructuraltransitiontothetiltedmolecular 183 packingwhichisenergeticallyfavorable[ 39 , 40 , 41 , 42 ].IfCoPcisconsidered, themolecule-substratebindingenergyistlylarger,resultinginahighlycorrugated surfacepotential.Inthiscase,theintermolecularinteractionsareunabletoovercomethe energybarrierssuchthattheincomingmoleculesarelocalizedonthesurfaceandadopta molecularorientation,preventingtheformationofanylong-rangeordereddomains asdemonstratedin Figure8.1 (c). 8.4Conclusions FromSTMandDFTanalysisof M PcgrowthonthedeactivatedSi(111)-Bsurface,the molecularorderingandorientationcanbedrasticallyalteredbyexchangingthecentral TMion.Thead-Siatomsfacilitatehighlyselectiveorbitalcouplingwiththe d -orbitals oftheTMionthroughtheirlocalized p z orbitals.Thisyieldsamechanismfortuningthe molecule-substrateinteractionstrengthwhich,consequentlywillmodifythepotentialenergy landscapeformolecularDependingonhowtlythemolecule-substrate interactionintroducescorrugationinthepotentialenergylandscape,thiscanresultingrowth transitionsandfurtherimpactthemolecularpackingandorientation.Thispavesthewayfor understandinghowtheassemblyandgrowthphenomenaofTM-incorporatedorganic moleculescanbemotoyieldoptimumgrowthforincorporationintosilicon-based devices. The p - d orbitalhybridizationmechanismcanalsomanifestforotherTMbasedmolec- ularsystemssuchasporphyrin.Porphyrinisamoleculewhichnotonlysharesthesame technologicalapplicationsas M Pc,butisalsohighlyrelevantinbiologicalsystems.Studies ofporphyrinadsorptiononmetallicsubstratessuchasAg(111)havealreadyshownstriking 184 similaritiestothatof M Pcinhowthe d -orbitalsoftheTMionshybridizewiththeunder- lyingsurface[ 236 , 237 , 238 , 239 , 240 , 241 , 242 ].Consequently,themolecularadsorptionof porphyrinontheSi(111)-Bsurfacecouldexhibitthesameselective p - d orbitalhybridization phenomenon.AsidefrommoleculargrowthonthedeactivatedSisurface,the p - d orbitalhy- bridizationmechanismfor M Pcmoleculesisalsorelevanceontopologicalinsulator systems[ 63 , 62 ].Ultimately,thegeneralapplicationsofthisobservedorbitalhybridization phenomenonwidensthescopeofthestudydescribedinthischapter. 185 Chapter9 TuningSiNanomembraneTransport PropertiesbyanInterfacialMetal PhthalocyanineThinFilm Usingtheknowledgegainedfrom Chapter5 through Chapter8 regardingthethingrowth of M PconthedeactivatedSi(111)-Bsurface,thegoalistonowintegratetheorganicthin withanovelsilicon-baseddeviceliketheSinanomembranedescribedin Section4.3.1 . SincetheSinanomembranedeviceishighlysensitivetotheinterfacecondition,theorganic- inorganichetero-interfaceformedbytheorganicgrowthprocesscanbeusedtomodifythe transportproperties.Unfortunately,byperformingtypicalbulknon-uniformdopingmethods ontheSinanomembranetothelevelsrequiredtoformthedeactivatedSi(111)-B p 3 p 3 R 30 surfacereconstruction,transportbehaviorthatisdominatedbythehighdoping levelsinthenanomaterialwillbecomeprevalent.Forthisreason,hydrogentermination neededtobeusedinordertopassivatetheSidanglingbondsitesonthenanomembrane. Section9.1 detailsthe M Pcthingrowthtestscarriedoutonhydrogenterminated Si(111)andSi(001)totheoptimumgrowthconditionforformingasmoothmolecular AfterproducingtheidealhydrogenterminatedSisurfaceonthenanomembranes, consistingofrent M Pcmolecules(F 16 ZnPc,CuPc,andClAlPc)atvaryingthicknesses 186 areintroducedonthesurfacetoseehowchangesintheorganic-inorganichetero-interface changetheelectricalpropertiesofthenanomembrane.Initialmeasurementsinvolvedtheuse ofKPFMtomonitorchangesintheworkfunctionoftheorganic-inorganichetero-interfaceon theSinanomembraneforaparticularmoleculeasthethicknessvaries.Theseresultsare detailedin Section9.2 .Changesintheworkfunctionduetothepresenceofonlyafewlayers of M Pccanbeindicativeoftheformationofaninterfacialdipolelayerformedbycharge transferattheinterface.Whethertheworkfunctionincreasesordecreasesdeterminesifan electronwillbedonatedtothenanomembraneorremovedfromthenanomembraneinthe chargetransferprocess.ThisproducesanenhancementorreductionintheSinanomembrane conductivity.Tomonitortheconductivityofthesedevices,vanderPauwandHalltransport measurementshavebeencarriedoutonhydrogenterminatedSinanomembraneswithand withouta M Pcmolecularasdiscussedin Section9.3 .Thesepreliminarymeasurements illustratethepossibilityofutilizingorganicsmallmoleculestomodifytheelectricalproperties oflow-dimensionalnanomaterialsforusefulapplicationsintheelectronicsindustry. 9.1GrowthofMetalPhthalocyanineFilmsonHydro- genPassivatedSiSurfaces 9.1.1GrowthontheSi(111)-HSurface TheSi(111)substratesurfacecouldbepassivatedbytwotwetchemicaletching methods,asalreadydetailedin Section4.2.3 .Thoughbothofthemethodsrequiredetching theSisubstrateinanitrogenatmosphere,thelastchemicalusedtoetchthesurface(either oxideetchorammoniumhadadramaticonthesurfacemor- 187 Figure9.1: MorphologyofCuPcdepositedonthesmoothSi(111)-Hsurface .A typicalAFMtopographyimage(AdiamondcoatedSitipisusedwithspringconstant k =0 : 2N/mandresonancefrequency f =13kHz)ofa10nmthickCuPconSi(111)-H whichhasbeenpreparedbythesmoothetchingprocessareshown.TheCuPcmolecules formtalltriangularthree-dimensionalislandsratherthanacontinuous(a)showsthe AFMz-heightresponseand(b)displaysthesimultaneousamplituderesponse. phologyobtained[ 177 ].Fromthemoleculargrowthstudiesdiscussedin Chapter5 through Chapter8 ,anatomicallysmoothSisurfaceisrequiredtoobservethehighlyorderedmolec- ulargrowth.Followingthistrend,theinitialgrowthexperimentsinvolvedtheuseofthe smoothhydrogenterminatedSi(111)surfaceasshownin Figure4.10 .Uponthe etchingprocessthissurfaceissealedinthenitrogenenvironmentusingano-ringsealedKF samplecarrierandtransportedintoannitrogengloveboxenvironmentinorderto minimizesampleexposuretoair.Thegloveboxisattachedtoanautomatedthermalevapo- 188 Figure9.2: MorphologyofCuPcdepositedontheroughSi(111)-Hsurface .A typicalAFMtopographyimage(AdiamondcoatedSitipisusedwithspringconstant k =0 : 2N/mandresonancefrequency f =13kHz)ofa10nmthickCuPconSi(111)- Hwhichhasbeenpreparedbytheroughetchingprocessareshown.TheCuPcmolecules formtallthree-dimensionalislandswhichhaveadisorderedappearance.(a)showstheAFM z-heightresponseand(b)displaysthesimultaneousamplituderesponse. rationset-upsosamplescanbepreparedinanitrogen/vacuumenvironmentandtransported backtothelaboratoryforAFMmeasurementswithouttheriskofcontamination. Figure9.1 showsanAFMimageofthesurfacetopographyforCuPcdepositedonthe smoothhydrogenterminatedSi(111)surface.MoleculargrowthstudiesofF 16 ZnPcand ClAlPchavealsobeencarriedoutonthissurface.However,duetothestrikingsimilarityin themorphologyforallmolecularspeciesused,onlyAFMimagesoftheCuPcgrowth areshown.Fromthisimageitisclearthatthemoleculesadoptathree-dimensionalisland 189 growthmode,producingtallislandsonthesurface.InlookingattheAFMimage,itcouldbe possiblethattheCuPcmoleculesformacompletemonolayerorevenafewlayersbefore transitioningtothethree-dimensionalislandgrowth.However,experimentsatthicknesses aslowastwomonolayers(3nmthick)stillresultedinthree-dimensionalislandsastall as25nm,suggestingthatthemoleculesdonotprefertowetthesurfaceandinsteadform three-dimensionalmolecularislands.Thiscouldbeindicativeofgrowththatisdominatedby theintermolecularinteractions.Itshouldbenotedthatthethree-dimensionalislandsthat areformedhaveatendencytoadoptatriangularshape.Oneconcernisthattheobserved growthisduetoakineticgrowthlimitationsuchasdepositingtheorganicmaterialat toohighofanevaporationrateormaintainingtoolowofasubstratetemperatureduring thegrowth[ 91 , 92 , 93 , 94 , 95 , 96 , 97 , 98 ].Toverifythatthisisnotthecase,aseriesof experimentshavebeencarriedoutattlylowermoleculardepositionratesaswell assometestsatelevatedsubstratetemperatures.Fromtheseexperiments,theobserved growthexperiencednochange,implyingthatthegrowthcannotbeelytunedwithin thelimitsexplored.TheresultsonthesmoothSi(111)-Hsurfaceindicatethatthissurface isnotidealforstudiesontheSinanomembraneduetothehighdegreeofnon-uniformityin themolecularwhichwillresultinpoorKPFMmeasurementsaswellasthe transportmeasurements. M PcmoleculargrowthisalsotestedontheSi(111)-Hsurfacepreparedbytheetching methodwhichproducesaroughsurfacemorphology(see Figure4.10 (a)).In Figure9.2 ,the surfacetopographyfromtheAFMimageofCuPcdepositedontheroughhydrogentermi- natedSi(111)surfaceshowsthattheCuPcmoleculesstilladoptathree-dimensionalisland growthevenonthisroughsurface.However,thereisanoticeabledegradationthatoccurs inthethree-dimensionalislandsthatareformed.Thoughtheislandsarestillconsiderably 190 tallasinthegrowthonthesmoothsurface,theynolongermaintainthetriangularshape andinsteadadoptarandomstructurewhichisduetotheroughsurfacemorphologyprior tomoleculardeposition.ItshouldbenotedthatmoleculargrowthstudiesofF 16 ZnPcand ClAlPchavealsobeencarriedoutonthissurface,andsimilartogrowthonthesmooth Si(111)-Hsurface,thesamemorphologyisobservedforallmolecularspeciesused.The inabilitytoformauniformthinoneithertheroughorsmoothetchedhydrogenter- minatedSi(111)surfacesindicatesthat,ingeneral,Si(111)isnotsuitableforstudyingthe of M PcthinonSinanomembranes.Toalleviatethisproblem,thereareacouple oftroutesthatcouldbeexplored.HydrogenpassivationcanbeperformedonSiwith atcrystalinterfaceinthehopesthatthechangeinthesubstrateunitcellwillresult inmorefavorablegrowthdynamicsforformingasmoothmolecularAnotheroptionis thatitcouldbepossibleothermoleculesgrownontheSi(111)hydrogenpassivatedsurfaces mightbeabletoavoidthisnon-uniformissue.However,sincethe M Pcmoleculeshave alreadybeenthoroughlystudiedinthisthesis,switchingtoanewSicrystalstructurepro- videdamoresimplisticapproachtomodifyingthemoleculargrowthwhilestillbeingable tomaintainthesamesamplepreparationproceduresandborrowfromtheknowledgegained in M PcthegrowthstudiesonthedeactivatedSisurface. 9.1.2GrowthontheSi(001)-HSurface SimilartoSi(111),theSi(001)substratesurfacecouldalsobepassivatedbythemethods describedin Section4.2.3 .Asalreadyshownin Figure4.12 (b),theetchingprocessused toproduceanatomicallysmoothsurfaceonSi(111)insteadcreatesaratherroughsurface withsomeetchingfeaturesdistributedacrosstheentiresample.Whentheetchingprocess usedforformingaroughsurfaceontheSi(111)isappliedtotheSi(001)surface,itresults 191 Figure9.3: MorphologyofCuPcdepositedontheroughSi(001)-Hsurface .Typical AFMtopographyimages(AdiamondcoatedSitipisusedwithspringconstant k =0 : 2N/m andresonancefrequency f =13kHz)ofthebareSi(001)-Hsurfacewhichhasbeenprepared bytheroughetchingprocessandofa10nmthickCuPconthesamesurfaceareshown. TheCuPcmoleculesformsmallgrainswhichcomformtotheSisurfacegivingasmooth (a)showsthebareroughetchedSi(001)-Hsurface(dashedredrectangle)while(b)displays theCuPconthissurface(dashedbluerectangle).TheAFMz-heightandamplitude responseisshownforboth. inthesamesurfacemorphologyascanbeseenin Figure4.12 (a).Thoughthistypeof surfacehadbeenpreviouslyastheroughsurfaceforSi(111),inthecaseofSi(001), thissurfaceappearstohaveamoreuniformmorphologythanthesurfaceresultingfromthe smoothetchingprocess.InordertoseeifchangingtheSicrystalorientationwouldhave anyonthemolecularthatisformed,thesurfacethatmostcloselyresembles asurfacemorphologyofSi(111)isselected,i.e.theSi(001)surfacehydrogenpassivatedby theroughetchingprocedure. 192 Figure9.3 showsdirectcomparisonbetweenthetopographyofthebareSi(001)-Hsurface andathinlayerofCuPcmoleculesdepositedontopoftheSi(001)-Hsurface.From Fig- ure9.3 ,asmooththinisformed,consistingofsmallmoleculargrainswhichisdrastically tfromthegrowthobservedonSi(111)-Hatthesamecondition.AsidefromCuPc, moleculargrowthstudiesofF 16 ZnPcandClAlPconcarriedoutontheSi(001)-Hsurface presentedthesamelowroughness,morphology.Thoughahighdegreeoforderingin themolecularcannotbeobtainedasonthedeactivatedSi(111)-Bsurface,thelowaver- agesurfaceroughnessofthe M PcgrownonSi(001)-Hmakesthissurfacesuitablefor performingKPFMandtransportmeasurementsonSinanomembranes. 9.2WorkFunctionMeasurementsofPhthalocyanine onPassivatedSiNanomembranes SimilartotheAFMmeasurementsin Section9.1.1 and Section9.1.2 ,allKPFMmeasure- mentsareperformedinasealednitrogenenvironmenttopreventanyairexpo- suremayhaveonthemeasurement.InordertodeterminetheworkfunctionoftheSi nanomembraneswithandwithoutmolecules,aswellasensurethattheAFMtipdoesnot changebetweenmeasurements,itisnecessarytomeasureacalibrationsamplewithawell workfunctionbeforeandafterameasurementofaSinanomembranedevice.For thepurposesofthisstudy,HOPGthathasbeenfreshlycleavedandmeasuredinanitrogen environmentisusedasthecalibrationsample.TheworkfunctionoftheSinanomembrane interfaceisdeducedbycomparisonbetweenthetip-samplesurfacepotentialbe- tweentheHOPGsurfaceandtheSinanomembrane. Figure9.4 displaysatypicalsurface potentialmapandsimultaneouslyobtainedtopographymapofthefreshlycleavedHOPG 193 Figure9.4: Simultaneoustopographyandsurfacepotentialmappingofthe Si(001)-Hnanomembranesurface .(a)AFMtopographyandKPFMsurfacepoten- tialimages(ATi/IrcoatedSitipisusedwithspringconstant k =2 : 0N/mandresonance frequency f =70kHz)takenonthebareSi(001)-Hnanomembranesurfacewhichhasbeen preparedbytheroughetchingprocess(dashedredrectangle).(b)IncludestypicalAFM topographyandsurfacepotetialimagesontheHOPGcalibrationsample(dashedbluerect- angle).TheHOPGsampleisusedbeforeandaftermeasurementsonthenanomembrane surfacetoensurenotipchangesoccurredandtoassigntheappropriateworkfunctionforthe nanomembranesurface. surface.Notethatthesurfacepotentialmapin Figure9.4 displayslittletonofrom thesurfacemorphology.TheHOPGsampleisthenswappedoutforaSinanomembrane deviceforKPFMmeasurement.Toensureappropriatesurfacepotentialmeasurementofthe Sinanomembranesurface,thenanomembraneandtheSihandlewafermustbegroundedto preventanysub-surfacechargingfromthehandlewaferandoxideonthesurface potentialmap.AtypicalKPFMmeasurementtakenonabareSi(001)-Hnanomembrane 194 Figure9.5: Simultaneoustopographyandsurfacepotentialmappingof1.5nm ontheSi(001)-Hnanomembranesurface .(a)AFMtopographyandKPFMsurface potentialimages(ATi/IrcoatedSitipisusedwithspringconstant k =2 : 0N/mand resonancefrequency f =70kHz)takenontheSi(001)-Hnanomembranesurfacewhichhas beenpreparedbytheroughetchingprocess,witha1.5nmthickofCuPcdepositedonthe surface(dashedredrectangle).(b)IncludestypicalAFMtopographyandsurfacepotetial imagesontheHOPGcalibrationsample(dashedbluerectangle).TheHOPGsampleis usedbeforeandaftermeasurementsonthenanomembranesurfacetoensurenotipchanges occurredandtoassigntheappropriateworkfunctionforthenanomembranesurface. isshownin Figure9.4 whichistakenbetweenHOPGmeasurements.Bysettingthework functionofHOPGtotheexpectedvalueof ˘ 4 : 6eV[ 243 ]andcomparingthetwosur- facepotentialmaps,theworkfunctionofthebareSi(001)-Hnanomembraneisdetermined tobe4 : 40 0 : 05eV,comparabletopreviousUPSmeasurementsofthehydrogentermi- natedSisurfaceforbulksubstrates[ 244 , 245 , 246 ].Workfunctionmeasurementofthebare nanomembranesurfaceisperformedformultipleiterationstoensurereproducibility. Threetypesof M Pcmolecules(F 16 ZnPc,CuPc,andClAlPc)havebeenbeengrownon theSi(001)-Hnanomembranesurfaceatthicknessesof1.5nm,3.0nm,and6.0nm, 195 Figure9.6: Simultaneoustopographyandsurfacepotentialmappingof3.0nm ontheSi(001)-Hnanomembranesurface .(a)AFMtopographyandKPFMsurface potentialimages(ATi/IrcoatedSitipisusedwithspringconstant k =2 : 0N/mand resonancefrequency f =70kHz)takenontheSi(001)-Hnanomembranesurfacewhichhas beenpreparedbytheroughetchingprocess,witha3.0nmthickofCuPconthesurface (dashedredrectangle).(b)IncludestypicalAFMtopographyandsurfacepotetialimages ontheHOPGcalibrationsample(dashedbluerectangle).TheHOPGsampleisusedbefore andaftermeasurementsonthenanomembranesurfacetoensurenotipchangesoccurred andtoassigntheappropriateworkfunctionforthenanomembranesurface. whichcanalsobedas1,2,and4monolayersbyassumingthemoleculesarestand- inguprightonthesurface(consistentwithpreviousgrowthexperimentsonroughetched hydrogenterminatedSi[ 35 ]). Figure9.5 , Figure9.6 ,and Figure9.7 displaystypicalsimul- taneoustopographyandsurfacepotentialmappingmeasurementstakenonCuPconthe Si(001)-HnanomembranewithcomparisontoHOPGmeasurementstakenprior.Forsim- plicity,onlythecaseofCuPcmoleculargrowthisshownhere.ImagestakenonClAlPcand F 16 ZnPcareomitted,however,theresultingworkfunctionmeasurementsaresummarized 196 Figure9.7: Simultaneoustopographyandsurfacepotentialmappingof6.0nm ontheSi(001)-Hnanomembranesurface .(a)AFMtopographyandKPFMsurface potentialimages(ATi/IrcoatedSitipisusedwithspringconstant k =2 : 0N/mand resonancefrequency f =70kHz)takenontheSi(001)-Hnanomembranesurfacewhichhas beenpreparedbytheroughetchingprocess,witha6.0nmthickofCuPconthesurface (dashedredrectangle).(b)IncludestypicalAFMtopographyandsurfacepotetialimages ontheHOPGcalibrationsample(dashedbluerectangle).TheHOPGsampleisusedbefore andaftermeasurementsonthenanomembranesurfacetoensurenotipchangesoccurred andtoassigntheappropriateworkfunctionforthenanomembranesurface. forallmoleculesandlmthicknessesinthegraphshownin Figure9.8 .Withafewlayersof CuPcorClAlPcdepositedontheSi(001)-Hnanomembranesurface,bothmoleculesdisplay ashifttolowerworkfunction,whileforthecaseofF 16 ZnPcdepositionanincreaseinthe workfunctionisinduced.Noticeableshiftsintheworkfunctionoccurevenwithjustthe applicationofasingle M Pcmonolayer.Changesintheworkfunctionduetothepresence ofonlyafewlayersof M Pccanbeindicativeoftheformationofaninterfacialdipolelayer formedbychargetransferattheinterface(see Figure9.9 )[ 183 ].Fromprevioustransport measurementsofhydrogenpassivatedSi(001)-Hnanomembranesthatarenominalp-type 197 Figure9.8: SummaryplotofthechangeintheworkfunctionofSinanomembranes forallmoleculesandthicknessesmeasuredbyKPFM .Theerrorbarsforeach measurementareincludedintheplot. doped,asthenanomembraneisthinneddownto ˇ 20nmandhydrogenpassivatedcarrier inversionisobservedinthetransportton-type[ 90 ].Thus,fromtheworkfunctionmea- surementsitisexpectedthatjustafewmonolayersoftheelectrondonormaterialsCuPc orClAlPcwillenhancetheelectricalconductivityoftheSinanomembranebyintroducing additionaln-typecarrierstothemembraneduringthechargetransferprocessforforming theinterfacialdipole.Consequently,adecreaseintheelectricalconductivityforF 16 ZnPcis expectedsincethisisanacceptormaterialwhichremoveselectronsfromtheSinanomem- braneinthechargetransferprocess.Thisiswhyanincreaseintheworkfunctionofthe organic-inorganichetero-interfaceisobserved.Thesepreliminarymeasurementsillustrate thepossibilityofutilizingthinconsistingofsmallorganicmoleculestomodifytheelec- tricalpropertiesoflow-dimensionalnanomaterialsforusefulapplicationsintheelectronics 198 Figure9.9: Exampleschematicrepresentationofworkfunctionshiftinducedby dipolelayerattheinterfacebetweenametallicsurfaceandaoforganic smallmolecules .In(a)themetallicsurfaceandtheorganicthinarespearatefrom eachotherandacommmonvirtualvacuumlevelisassumed.In(b),whentheorganic isinterfacedwiththemetallicsurfaceitresultsinashiftinthevacuumlevel,duetothe formationofaninterfacialdipolelayer.Thisshiftinthevacuumlevelyieldsachangeinthe workfunctionforthecombinedsystem. industry.However,inordertobecertainthatthisisindeedthecase,transportmeasure- mentsarerequiredtoverifyifanincreaseordecreaseinthenanomembraneconductivityis observedforthecorrespondingexpected M Pcmolecule. 9.3TransportMeasurementsofSiNanomembranes PreliminaryHallandvanderPauwtransportmeasurementstakenonhydrogenterminated Si(001)-Hnanomembraneswithandwithouta M Pcmolecularhavebeenperformed, 199 however,severalissuesoccurredduringtheinitialtransportexperiments.Inordertoalle- viatemanyofthecomplicationsintroducedbytheoriginaltransportprobe anewapplicationinsertprobeforthePPMS,thesystemusedtoperformthetransport measurementson,hadtobedesigned(see Appendix A )andnewsoftwarehadtobewritten beforereliabletransportmeasurementscouldbeestablished.Forthisreason,thetransport measurementshadtobestoppedinordertoaddressthevariousissues.Asaresult,reliable transportmeasurementscouldnotbeperformedpriortothecompletionofthisthesis.How- ever,asdiscussedin Section10.2 ,theapplicationinsertprobehassincebeencompletedand transportmeasurementsontheSinanomembranedevicescannowcontinue. 9.4Conclusions HydrogenterminationofSinanomembranesisnecessarytopassivatethesurfacedangling bondstoallowfororganicmolecularthinlmgrowth.Thisisbecausedopinglevelsrequired toformthedeactivatedSi(111)-B p 3 p 3 R 30 surfacereconstructionwillresultintrans- portbehaviorthatisdominatedbythehighdopinglevelsofthenanomaterial.Growthtests of M PcmoleculesonhydrogenterminatedSi(111)andSi(001)showedthatasmooth,gran- ularmolecularcanbeformedontheSi(001)-Hsurfaceetchedtoproducearoughsurface morphology,ratherthanonaregularlysteppedsurface.HydrogenterminatedSisurfaces whichareformedusingetchingmethodstoproduceasmoothterracedsurfaceresultedin tallnon-uniformthree-dimensionalislandgrowthnotsuitableforthesestudies. IsolatedSi(001)-Hnanomembranesdevicesthatarepassivatedusingtheroughetching procedureareusedtoseeiftheelectricalpropertiesattheorganic-inorganichetero-interface ofthenanomembranecanbemobytheformationofaninterfacialdipolelayer.This 200 isdonebyintroducingaof M Pcmoleculesonthesurface.Thedipolelayermanifests duetothechargetransferatthehetero-interfaceinducedbythefew M Pcmolecular layers.Thechangeintheworkfunctionoftheorganic-inorganichetero-interfaceonthe Sinanomembraneistrackedfort M Pcmoleculesasthethicknessvaries.This measurementdetermineswhetherornotaninterfacialdipoleisformedandultimatelyhow theinterfacialdipolelayerformedbyaparticularmoleculewillpossiblychangethecon- ductivityofthenanomembrane.IntroducingaofCuPcorClAlPcontheSi(001)-H nanomembranesurfacebothdisplayashifttolowerworkfunction,whileaofF 16 ZnPc inducesanincreaseintheworkfunction.Fromthesemeasurementsitisexpectedthatthe CuPcorClAlPcwillenhancetheelectricalconductivityofSinanomembranesanda ofF 16 ZnPcwilldecreaseintheelectricalconductivity.Tobecertainthatthisisindeed thecase,transportmeasurementsarerequiredtoverifyifanincreaseordecreaseinthe nanomembraneconductivityisobservedfortherespective M Pcthinlm.Duetoseveral issuesthatoccurredduringtheinitialtransportexperiments,anewapplicationinsertprobe neededtobedesignedbeforemovingforwardwiththenecessarytransportmeasurements. ThishassincebeencompletedandtransportmeasurementsofSinanomembranedevicescan beexplored.Nevertheless,thepreliminaryKPFMmeasurementsillustratethepossibilityof utilizingthinconsistingofsmallorganicmoleculestomodifytheelectricalproperties oflow-dimensionalnanomaterials. 201 Chapter10 ConclusionandFutureProspects 10.1SummaryofResults Toformhighlyorderedorganicthinwithmolecularprecision,requiresanunderstanding ofhowthesemoleculesinteractwiththesubstratesurfaceaswellastheirintermolecular interactions.Theresultspresentedinthisthesisaimtoprovideacomprehensiveanalysis ofthegrowthevolutionoforganicsmallmoleculesontechnologicallyrelevantSisurfaces andhowtheobservedgrowthcanbemotoproducehighly-orderedmolecularthin Theprototypicalmetalphthalocyanine( M Pc)molecule( M =Zn,Cu,Co)isutilized throughoutthesegrowthstudieswhichprimarilyoccuronthedeactivatedSi(111)-Bsurface, withsomepreliminaryresultsshownonhydrogenpassivatedSinanomembranes. ThegrowthstudiescenteringaroundtheuseoftheZnPcmoleculeonthedeactivedSi surfacedisplaygreatpromisetowardstheformationofanorganizedmolecularthinwith ahighdegreeofmolecularordering.TheZnPcmoleculeshowsanintermediatemolecule- substrateinteractionwiththedeactivatedSi(111)-Bsurfacethatisbalancedbytheinter- molecularinteractions.ThisallowstheZnPcmoleculestoeasilyformclose-packedhighly- orderedorganicnanostructureswherethemoleculesareinatiltedabovethe substrate.TheSTMimagesoftheorganicmolecularoverlayershowthepresenceofasu- perstructureknownasamoirepattern,implyingthatthemoleculardomainsmaintaina spazimuthalregistrationwiththeunderlyingSisurfacethatisnotcommensurate. 202 Someconsiderationsregardingthemolecularnbarriersarenecessaryduetothe drasticimpacttheycanhaveontheformationofacontinuousTheinitialsurface conditionofthedeactivatedSi(111)-Bmustbecarefullycontrolledsoasnottointroducea tnumberofdefects,asthesemayalterthemoleculargrowth.Fromsub-monolayer ZnPcgrowth,itisevidentthattheEhrlich-Schoebelbarrier(ESB)duetotheSistepedges isconsiderableandcanpreventadsorbed M PcmoleculesfrombetweenSiterraces. However,onapristinedeactivatedSisurface,theZnPcmoleculescanrapidlyacross theSiterracessuchthattheynucleateattheSistepedgesitesandproceedtogrowinthe anisotropicwgrowthmode.Thisgrowthmodecanrestrictthenumberofsymmetry allowedmoleculardomainsinwhichasingledominantin-planemolecularorderingispossible. Asthemolecularcoverageincreasestowardacompletemonolayer,thereisconcernasto whetherornotadmoleculesthatlandontopofpreexistingdomainswillbehinderedfrom formingacompletemolecularlayer.StepedgesformedbyZnPcmoleculeswithinexist- ingdomainscanexhibitanadditionaltypeofESB,preventingdownwardmasstransport. However,thesestudiesindicatethatthisESBisconsiderednegligible.Somechallengesstill remaininordertoensureamolecularofthehighestqualitycanbeformed.Thisislargely duetoadditionalactivationbarriersand/ornucleationsiteswhicharisefromgraindomain boundariesformedastheZnPcmolecularcoverageapproachesamonolayer.Togetherthese caninduceroughness.AccesstotheanisotropicwgrowthmodeofZnPc onthedeactivatedSisurfaceprovidesameanstoachievelong-rangein-planemolecularor- dering.Thisreducestheformationofundesirablegraindomainboundarieswhichintroduce theseactivationbarriers.Additionally,bycontrollingthesubstratetemperatureduringthe depositionprocess,thenumberofdefectsanddomainboundariesformedcanbetly reduced,promotingtheformationofalong-range,highly-ordered,smoothZnPcmolecular 203 Theobservedgrowthmodesarehighlydependentonthetemperature,socaremustbe takenwhenmodifyingthesubstratetemperatureduringthemoleculargrowth.Thegrowth modetransitionsfor M Pcmoleculesaremappedinthisstudy,providingtemperatureranges thatcanbeusedtoformahighlyorderedorganicmolecularthinonalargescale.For example,theanisotropicwgrowthmodeexhibitedbyZnPccanbeextendedto temperaturesnearing170 C,allowingfortheformationofverylargemoleculardomains withanexclusivein-planemolecularordering. Asthetransitionsfrommonolayertomultilayeredgrowth,amildrelaxation intheZnPcin-planemolecularorderingoccurs.However,therelaxationisnott enoughtochangetheazimuthalordering.Withincreasingthickness,theZnPcgrowth isabletomaintainthisnewrelaxedmolecularpackingfor40monolayersormore.The tiltedorientationoftheZnPcmoleculesenoughinterlayer ˇ - ˇ interactiontostabilize theasthethicknessincreases,despiteagradualdecreaseinthemolecule-substrate interaction.Thisresultsinahighdegreeoforderingintheout-of-planedirection.Coupling thisobservationwiththesuperiorin-planemolecularorderingbytheanisotropic wgrowthopensupthepossibilityofformingasmooth,organicmolecularthin whichdisplaysbothin-planeandout-of-planemolecularordering. Toamethodformodifyingmolecularorderingandorientationof M Pconthedeac- tivatedSisurface,thenatureofthemolecule-substrateinteractioncanbeinvestigated.Due totheabilityof M Pcmoleculestoeasilyexchangetheircentraltransitionmetal(TM)ion throughsyntheticchemistrymethods,itispossibletoexploretheroletheTMionplaysin theinteractionwiththesubstrate.InthecaseofthedeactivatedSisurface,thelocalized p z orbitalsofthead-Siatomsfacilitatehighlyselectiveorbitalcouplingwiththe d -orbitalsof 204 theTMion.Thisyieldsamechanismfortuningthemolecule-substrateinteractionstrength that,consequently,willmodifythemolecularprocessas M isswappedoutforZn, Cu,andCo.Thiscanresultingrowthtransitionsandfurtherimpactthemolecularpacking andorientation.Thechangeinthisinteractionisresponsibleforthebetweenthe highdegreeofmolecularorderingthatcanbeachievedbyZnPctoformatiltedmolecular andthehighlydisorderedstructuresformedbyCoPc. Thesestrategiesforformingahighlyorderedorganicthincanbedirectlyappliedto anovelsilicon-baseddevice.Sinanomembraneisalow-dimensionalmaterialwithelectronic propertiesthatarehighlysensitivetotheinterfacecondition.Thus,thegoalistobeableto modifythetransportpropertiesoftheSinanomembranebychangingtheconditionsofthe organic-inorganichetero-interfaceformedbythethingrowthprocess.Thesurfacesand interfacesplayacriticalroleindeterminingpropertiesandfunctionsofthenanomembrane, oftendominatinganyofthebulkSiproperties.Consequently,iftypicalbulknon-uniform dopingmethodsareperformedontheSinanomembranetohighdopinglevels,thetransport behaviorwillbedominatedbythedopingofthenanomaterial.Thisisproblematicfor formingthedeactivatedSi(111)-B p 3 p 3 R 30 surfacereconstruction.Forthisreasonan alternateSisurfacepassivation,i.e.hydrogentermination,isutilized.Growthtestsof M Pc thingrowthonhydrogenterminatedSi(111)andSi(001)showsthatasmoothgranular molecularcanbeformedontheSi(001)-Hsurfacewhichhasbeenetchedtoproduce aroughsurfacemorphologyratherthanaregularlysteppedsurface.Hydrogenterminated Sisurfaces,whichareformedusingetchingmethodstoproduceasmooth,terracedsurface, resultintallnon-uniformthree-dimensionalislandgrowthnotsuitableforthesestudies. AfterproducingtheroughetchedSi(001)-HsurfaceonSinanomembranes,aof M Pc moleculescanbeintroducedonthesurfacethatcanmodifytheelectricalpropertiesatthe 205 organic-inorganichetero-interfacebytheformationofaninterfacialdipolelayer.Thisdipole layermanifestsduetothechargetransferatthehetero-interfaceinducedbythefew M Pcmolecularlayers.Bytrackingthechangeintheworkfunctionoftheorganic-inorganic hetero-interfaceforaparticularmoleculeasthethicknessincreasesontheSinanomem- brane,itcanbedeterminedwhetherornotaninterfacialdipoleisformed.Thechange intheworkfunctiondetermineswhetheranelectronwillbedonatedorremovedfromthe Sinanomembraneinterfaceinthechargetransferprocess,whichultimatelydeterminesif thepresenceofthe M PcmolecularproducesanenhancementorreductionintheSi nanomembraneconductivity.Threetypesof M Pcmoleculeshavebeentested:F 16 ZnPc, CuPc,andClAlPc.FewlayersofCuPcorClAlPcontheSi(001)-Hnanomembranesur- facebothdisplayashifttolowerworkfunction,whilefewmonolayersofF 16 ZnPcinduces anincreaseintheworkfunction.ThereforeitisexpectedthatCuPcandClAlPcwillen- hancetheelectricalconductivityofSinanomembranes,whichbecomen-typeafterhydrogen termination,andadecreaseintheelectricalconductivityforF 16 ZnPc.Thesepreliminary measurementsillustratethepossibilityofutilizingorganicsmallmoleculesorganizedinto athintomodifytheelectricalpropertiesoflow-dimensionalnanomaterials,usefulfor applicationsintheelectronicsindustry. PreliminaryHallandvanderPauwtransportmeasurementstakenonhydrogentermi- natedSi(001)-Hnanomembraneswithandwithouta M Pcmolecularhavebeenper- formed,howeverseveralissuesoccurredduringtheinitialtransportexperiments.Toelim- inatetheseissuesforfuturemeasurements,anewapplicationinsertprobeforthePPMS systemforuseintransportmeasurementshasbeendesigned(see Appendix A )andnew softwarehasbeenwrittentoensurereliabletransportmeasurementscanbeestablished.As consequence,transportmeasurementscametoahaltwhilemanufacturingthenewprobe 206 andcouldnotbeperformedpriortothecompletionofthisthesis. 10.2FutureWork Thepurposeofthisthesisistolaythegroundworkformoreextensiveinvestigationinto thevariouspossibilitiesorganicsmallmoleculescanerinthesemiconductingindustryby utilizingthingrowthprocessesaswellasopenotheravenuesofexploration.Thiswork directlyrelatestoanumberofpotentialapplications,someofwhichwillbedetailedhere. Fromthegrowthevolutionstudiesof M PconthedeactivatedSi(111)-B p 3 p 3 R 30 surface,itisevidentthatZnPcdisplaysthehighestpossibilityforforminghighlyordered organicmolecularthatcanbeapplicableonscalesrelevantforindustrialuses.The abilitytomaintainsuperiorin-planeandout-of-planemolecularorderinginmultilayer awellinterfacewhichisofkeyimportancefororganicphotovoltaics,organic transistors,andorganiclightemittingdiodes.Thusfar,thegrowthstudiescon- tainedwithinthisthesishaveidenfactorswhichcantheabilityofZnPcto formasmoothmultilayeredwhilestillmaintainingthebenoftheanisotropicstep- wgrowthmode.However,adistinctZnPcdepositionandthermaltreatmentprocedure thatiscapableofformingamacroscopicscaleZnPccrystalmaintainingthehighdegree ofmolecularorderingremainstobeidenThisisonepossibledirectionwhichcould warrantfutureinvestigations.Somelimitsonthemolecularhavealreadybeen idenwhichcanbeattributedtothelargeESBassociatedwiththeSistepedgesaswell astheSidefectlinespresentevenonpristinedeactivatedSisurfaces.However,byreducing thedensityoftheSiatomicdefectlinesandloweringthemis-cutangleofthesubstrate, thegrowthcouldbefurtherimproved,pushingthelimitsofhowfarZnPcmolecular 207 orderingcanbemaintained. Fromamoleculargrowthperspective,asidefromexploringthegrowthofotherstandard M Pcmolecules,thereisanothersmallorganicmoleculewithinthePcfamilywhichcould bebtoexplore,callednaphthalocyanine.Thenaphthalocyaninemoleculestillco- ordinatesaTMioninthecenterofthemoleculelike M Pc,howeverithasanadditional benzeneringattachedtoeachofthebenzene-pyrroleligands,makingtheentiremolecule larger.Byslightlychangingthesizeofthemolecule,theintermolecularinteractions,aswell asthemolecule-substrateinteractions,willbemoThiscouldhaveadramaticimpact onthecorrugationandhomogeneityofthepotentialenergylandscape,possiblymodifying thelandscapetobesmoothsincethemoleculeisnowevenlargerthanthesubstratelattice parameters(similartothediscussionin Section8.3 ).Indoingso,thiswilldictatehow molecularwillproceedforthesemoleculesand,ultimately,whattypeofmolecular packingandorientationwillbeobserved. Inrecentyearstherehasbeenaconsiderableresearchinthedirectionoforganic photovoltaicstomatchtheofconventionalSi-basedphotovoltaicsusingmorecost ematerials.Smallorganicmoleculesareamongsomeofthecommonlystudiedma- terialsfororganicphotovoltaics,with M Pcmoleculesbeingamongthem.Duetothedevice architecturefororganicphotovoltaics,adonor-acceptorheterojunctionisrequiredtosepa- ratethechargecarriersonceanexcitonisgeneratedinthephoto-activematerial.Partof thereasonorganicvoltaicshavebeenlimitedintheirachievableisdueto thatoccuratthehetero-interface.Inthetypicalorganicphotovoltaicdevicearchitecture, thehetero-interfaceiseitherablendedoraverticallylayeredstructuremakingitto characterizetheinterfaceely.Tothisend,furtherinvestigationofhetereo-interface onamolecularscaleisnecessarytoimprovethecurrentunderstanding.Onepossibleroute 208 istotakeadvantageofthehighlyorderedZnPcstripestructuregrownonthedeactivatedSi surfaceasdescribedinthisthesis.AZnPcstripeprovidesatwo-dimensionaldonordomainin whichasecondtwo-dimensionaldomainconsistingofanacceptormaterialsuchasF 16 ZnPc canattach.Providedthegrowthoftheacceptormaterialdoesnotdisruptthe ZnPcstripegrowth,thisthepossibilityofformingahetero-interfacethat isorderedonamolecularlevel,one-dimensional,andcanbeinvestigatedusingSTM/AFM techniquesincombinationwithamethodforphoto-excitation.Duetotheanisotropyofthe ZnPcstripe,ahetero-interfaceformedalongthe ˇ - ˇ stackingdirectioncouldgiveat responsecomparedtogrowthonthesides.Ofcoursesomeelectronictscouldbecome convolutedwiththedeactivatedSisurface.However,ifsimilargrowthcanbeachievedon ametallicsurfacewithathininsulatorgrownontop,thiscoulddecouplethenanoscale heterojunctionfromthesubstrate. The p - d orbitalhybridizationmechanismfor M Pcdiscussedin Chapter8 canalsoman- ifestforotherTMbasedmolecularsystems,suchasporphyrin.Studiesofporphyrinad- sorptiononmetallicsubstrates,suchasAg(111),havealreadyshownstrikingsimilarities tothatof M Pcinhowthe d -orbitalsoftheTMionshybridizewiththeunderlyingsurface [ 236 , 237 , 238 , 239 , 240 , 241 , 242 ].Consequently,themolecularadsorptionofporphyrin, aswellasotherTMbasedmolecules,onthedeactivatedSi(111)-Bsurfacecouldexhibit thesameselective p - d orbitalhybridizationphenomenon.Thoughporphyrinsharessomeof thesametechnologicalapplicationsas M Pc,italsoishighlyrelevantinbiologicalsystems. Ifthe p - d orbitalhybridizationcanalsomodifythemolecularorderingandorientationof porphyrins,similartothe M Pcmoleculesdiscussedinthisthesis,thenthiscouldexpand thefunctionalityofSibasedelectronics.Particularly,combiningthingrowthofpor- phyrinswithamaterial,suchasSinanomembranes,couldhaveaprofoundimpact 209 onbiologicalsensorsandmedicalapplications. Sinceonlypreliminaryresultshavebeenshowninthisthesis,moreinvestigationsre- gardingmooftransportpropertiesofSinanomembranesthroughtheuseofor- deredorganicthinisrequired.Theinitialmeasurementshaveshownenoughevidence towarrantcontinuedinvestigationsofphysisorbed M PconSinanomembranes.Also,the applicationinsertprobehasbeencompleted,asshownin Appendix A ,enablingreliable transportmeasurementstobetakenontheSinanomembranes.Othertypesofmolecules thatdisplaychargetransferbehaviorwhendepositedontheSinanomembrane,thuscreat- inganinterfacialdipolelayer,couldbefurtherexplored.However,thereareothermethods forintroducingadipolelayerattheorganic-inorganichetero-interfacethatcouldalsobe studied.Theformationofathinconsistingofamoleculewithaninherentdipoleori- entationcanbeusedtomodifythecarriermobilityinthenanomembrane.Additionally,a directchemicalinteractionbetweentheorganicmolecularspeciesandtheSinanomembrane surfacecouldproduceaninterfacialdipole.Aninterestingdirectionwouldbetochemically anchoranorganicmonolayertotheSinanomembranesurfaceviawet-chemicalmethods, wherethemonolayerconsistsofamoleculewithapermanentdipolewhoseorientationcan beswitchedthroughaphoto-isomerizationprocess.Thechangeinthedipoleorientation couldresultinamointhetransportpropertiesoftheSinanomembraneandcan beeasilycontrolledbyphoto-excitationmethods. 210 APPENDICES 211 AppendixA TransportApplicationInsertProbe Design InordertoperformthenecessaryvanderPauwandHalltransportmeasurementsonSi nanomembranedevices,itbecamenecessarytodesignanewapplicationinsertprobefor useinthePPMSdescribedin Section3.1.3 .Thisisisduetoseveraltexperimental factors.First,theexistingsamplepuckforHallmeasurementsonlyallowsfortwoofthe eightnecessarywiringunlessthesampleisremovedfromthePPMSand exposedtoambientconditionforre-wiring.Thisisnotanidealscenariosincethehydrogen terminatedSinanomembraneishighlysensitivetoexposuretoairwhichcanalterthe electronicpropertiesovertime.Moreimportantlythispuckdoesnotallowforswapping oftheelectricalconnectionstotestwhetherornotacurrentleakagepathwayiscreated throughtheoxidelayertothehandleSiwaferbeforeoraftertheHallmeasurement.Thisis becausethepuckrequiresthatthehandlewaferbegroundedtothesystemratherthanwired suchthatitcanbeeithergroundedorusedinanelectricaltodetermineifa leakagepathwayhasbeencreated.Itisextremelyimportanttohavethisoptioninorderto ensuredeviceintegritybetweenmeasurements.Atexistingtransportprobeoptionis availablethatmoreyinwiringconnections.Thisallowsforall inthevanderPauwandHallmeasurementstobewiredformeasurement,however,Hall 212 measurementsarenotpossibleusingthisprobeduetothesampleorientation.Similarto theissueswiththepuckforHallmeasurements,thisprobealsoonlyallowsforfourwired electricalconnectionsmeaningtheSihandlewafermustgroundedtothesystemratherthan wiredforleakagepathwaychecksorback-gating.Ofcourse,withthisset-upitispossible tore-wirethedeviceinordertocheckifacurrentleakagepathwaythroughtheoxidehas formedduringthroughoutmeasurements,however,thisrequiresthesampletobeexposedto ambientmultipletimes.Forthesereasons,anewdesignwasnecessarywhichwouldaddress allthewsoftheexistingprobeswhilealsoaddingmorefunctionalityandadaptabilityin thedesigntomultipletypesofmeasurementstobeperformedwithouteverhavingto removethesamplefromthePPMS. Thenewtransportapplicationinsertprobewasdesignedbymewithhelpfromthe PhysicsandAstronomyMachineShopandDr.TaoZouatMichiganStateUniversity.The designconsistsofacopperframewhichisconnectedtoatypicalpuckforconnectiontothe baseofthePPMS.Aslitinthebaseofthecopperframeallowsforawireconnectionfrom thebasepucktoathermalsensorwhichismountedonthecopperframenearwherethe sampleislocated.Thisthermalsensorisprogrammedtooverwritethetypicaltemperature measurementschemeneartheendofthepuckinordertoprovideamoreaccuratetemper- aturemeasurementofthesample.Thepuckandadapterwiththeelectricalconnectionslit areconnectedtothecopperframebyasetofscrewssothatifnecessarythethermalsenor canbeeasilyadjustedorotherelectricalsensorscanbeintroduced.Thesamplerestsona copperblockinthecentralpartofthecopperframe.Ifthesampleneedstobeelectrically isolatedfromthisblockaninsulatingsapphirepiececanbeeasilymountedontheblock. Thesamplemountingblockcanbesettobeinoneoftwoorientationsviasetscrewswhich connecttotheframe;eitherperpendiculartothedirectionofthemagneticinthePPMS 213 orparalleltothemagnetic.Settingthemountingblockperpendiculartothemagnetic isparticularlyadvantageousforthisstudybecauseitallowsforHallmeasurementstobe performed.Theparallelorientationcanbeusedforotherstudiesifnecessary.Additionally, havingthisoptionforthesampleblockorientationthesamplemountingprocess. Abovethesamplemountingblockisabreadboardforwiringconnectionstothesample.This boardcanbesubstitutedoutforacopperblockwithsapphireregionsattached.Shielded wiringrunsthroughastainlesssteeltop,whichisconnectedtothetopofthecopperframe aswellasthebreadboardorcopperblockbyscrews.Thesewiresarethensolderedto thebreadboardorcopperblockforformingconnectionstothesample.Uptoeightwiring connectionsareavailable,althoughwithsomeminormoionsmorecanbeaddedif necessary.Thestainlesssteelcapisweldedtoastainlesssteeltubewhichalsohasevenly spacedstainlesssteeldisksweldedtoit.Thesedisksactasthermalshieldstohelpmaintain thesampleatcoldtemperaturebyreducingspreadinthetemperaturegradientalongthe probe.Eachofthesediskshasaslitinthemsothatthecablingcanbespiraledfromthe topoftheprobedowntotheprobeendwherethesampleislocated.Eachdiskisrotated by90 withrespecttoeachothertopreventattemperaturegradientfromform- ingalongtheprobe.Ventholeshavebeendrilledinatthescrewholelocationsaswellas inthestainlesssteeltubetoallowfortvacuumpumpingoftheprobeinsert.The stainlesssteeltubeisconnectedtoaKFbyanweldedonadapterwhichallowsfor minorheightadjustmentstobemade.TheeisconnectedtoKFcrossangeconnec- torandsealedtightusingano-ring.ThetwosidehavebeenmowithSMA cablefeedthroughconnectorswhichareappropriatelysolderedtothecablingthatspirals downtheprobeforelectricalconnectionstothesample.Thesecanbemoto includemoreelectricalconnectionstothesampleorevenanopticalerifnecessary.The 214 remainingKFangeconnectorisusedtothesealtheprobeinsertonthePPMS.Referto FigureA.1 , FigureA.2 , FigureA.3 , FigureA.4 , FigureA.5 ,and FigureA.6 forpictures oftheprobedesign.Usingthisdesign,allelectricalconnectionstothesamplecan beexchangedasneededwithoutremovingthesamplefromthesystem,thesamplecanbe orientedineithertheperpendicularorparalleldirection,anditispossibletocoolthe sampledownto2.5K.Ifalowertemperatureisrequired,thecablingcanbeeailyswapped outforthinnercablesmoresuitableformeasurementsatthesetemperatures.Ingeneralthis probeallthecriterianecssaryforperformingvanderPauwandHallmeasurements onthesamedevicewithoutexposuretoambientconditions,whilealsocheckingforcurrent leakagepathwaysthroughtheinsulatingoxidelayerbetweenmeasurements. FigureA.1: Three-dimensionaldesignoftheapplicationprobeinsertendusedfor transportmeasurements . 215 FigureA.2: TopoftheapplicationprobeinsertwithSMAcablefeedthrough connectorsandaKFconnectorusedtosealtheprobeonthePPMS . FigureA.3: Imageofthecablingwhichspiralsdownfromtheonthetop oftheprobeforelectricalconnectionstothesample .Thermalshieldsareusedto maintainthesampleatcoldtemperaturebyreducingspreadinthetemperaturegradient alongtheprobe.Additionally,thethermalshieldsareusedasawaytoholdanddirectthe cablingdowntothebottomoftheprobe. 216 FigureA.4: Aside-viewimageofthebottomoftheapplicationprobeinsertwhere thesampleconnectionismade . FigureA.5: Otherside-viewimageofthebottomoftheapplicationprobeinsert wherethesampleconnectionismade . 217 FigureA.6: Imageoftheadjustmentschemeforchangingtheorientationofthe samplemountingplatform . 218 AppendixB OtherContributedWorks Besidesthestudiescontainedwithinthisthesis,Iplayedaroleinsomeotherworksthatare notnecessarilyin-linewiththegeneralfocusofthingrowthoforganicsmallmolecules onSisurfaces.However,Idothinkitisworthprovidingabriefdescriptionofeachstudy andthecontributionsImadetoeach. Conductivepolymermaterialsareanotheravenueofinterestfororganicelectronicsdueto theirlowcost,behavior,andeaseoffabricationusingspin-coatingtechniques.Among thevariouspolymersinvestigated,poly(3-hexylthiophene),commonlyreferredtoasP3HT, hasbeenextensivelystudiedandshowngreatpromisefororganicelectronicapplications, particularlyforOPVs,OLEDs,andOFETs.Theelectricalpropertiesofaconsisting ofP3HTarehighlydependentonhowcrystallinetheP3HTstructureis,witht packingorientationsresultinginvariationsintheelectricaltransport.Inthisstudy,thermal treatmentmethodsareusedtoproduceaP3HTconsistingoflongP3HTerson topofanITOsubstratewithanelectronblockinglayer.ConductiveAFMmeasurements performedinanitrogenenvironment(asalreadydescribedin Section3.1.2 )onthermally treatedandnon-thermallytreatedP3HTshowthatthethermallytreatedyields regionswithhighcurrentalongtheers.Thisobservationiscorroboratedbyafullthree- dimensionaldevicemodelthataccountsfortheanisotropicmobilityintheP3HTers, currentspreading,andtrapMycontributiontothisstudyprimarilyinvolvedthe experimentalapparatuses,particularly,theinstallationandinitialtestsusedforfabricationof 219 thesamplesaswellaspreparingtheAFMset-upforperformingairsensitivemeasurements inacontrollednitrogenenvironment.Formoreinformationregardingthisstudyreferto Reference[ 247 ]:J.Sun,K.Pimcharoen,S.R.Wagner,P.M.Duxbury,andP.P.Zhang, Org.Electron. 15 ,441(2014). AkeycomponentoftheOPVdevicearchitectureiswhatiscalledawindowlayer,which consistsofawidebandgapmaterialsuchthatitisopticallytransparent.Typicallythe windowlayerislocatedonthecathodesideoftheOPV.Thewindowlayerisusedtoblock excitonquenchingatthecathode,serveasaprotectionlayerfortheorganicactivelayers whilethecathodeisdeposited,increasethedevicelifetime,andimprovechargecollection. OnepossiblewindowlayermaterialworthexploringisZnSduetoitswidebandgapand theabilitytothermallydepositthematerialinvacuum,similartowhatisdonefordepo- sitionofactivelayersconsistingofsmallorganicmolecules.However,duetotherelatively highresistivityofthermallydepositedZnSthedeviceperformanceshavebeenlowin comparisontootherwindowlayermaterials.Onesolutiontothisproblemistodopethe ZnSwithAl 2 S 3 inaco-depositionprocess.Atoptimaldoping,theZnSbasedwin- dowlayerdisplaysenhancedperformanceandcanactasanacceptablereplacementwindow layermaterial.MycontributionstothisstudyinvolvedAFMtopographymappingofthe ZnSatvaryingAl 2 S 3 dopinglevels.TheseresultsshowedthatastheAl 2 S 3 doping levelincreasedsmallZnSgrainsbegantodiminishandarethenreplacedbylargeAl 2 S 3 domains,whichisalsoconsistentwiththedecreaseinZnScrystallinityobservedinthex-ray measurements.Thisisindicativeofwhythedeviceperformancewouldd athigherdoping.FormoreinformationregardingthisstudyrefertoReference[ 248 ]:C.J. Traverse,M.Young,S.R.Wagner,P.P.Zhang,P.Askeland,M.C.Barr,andR.R.Lunt, J.Appl.Phys. 115 ,194505(2014). 220 Inrecentyearstherehasbeenagrowinginterestinexploringtwo-dimensionallayerma- terialswhichexhibituniqueelectronicproperties.Amongthemostfamousisgraphene, whichexhibitsmasslessDiracfermions,resultinginsuperiorcarriermobility.Siliceneisa two-dimensionalmonolayermaterial,likegraphene,wheretheSiatomsarearrangedina honeycombstructure.Silicenehasbeenconsideredanalternativetographenebecausewhen itisfreestandingitcanalsoexhibitmasslessDiracfermions,itcanbeeasilyintegrated intoSi-basedelectronics,anditalsopossessesstrongerspin-orbitcouplingthangraphene. However,freestandingsilicenehasyettobeproduced,andSistructuresgrownonasup- portingsubstratetendtodisplaystronginteractionswhichcausesthepropertiestodeviate drasticallyfromwhatisexpectedinthefreestandingcase.Inthisstudy,thegrowthofmul- tilayeredSi p 3 p 3ontopofAg(111)isstudiedusingSTMandSTS.Frommeasurements ofthelayerandsecondlayer,itcouldbedeterminedthatthe p 3 p 3phaseisgrown ontopofaninterfacial p 7 p 7layer.Furthermore,astrongsubstrateonthe Sithinelectronicpropertiesisevidentbythedecreaseinthefree-electron-likesurface stateofthe p 3 p 3phaseneartheAg(111)surface.Myinvolvementinthisstudyfocused onstructuralcharacterizationofthe p 3 p 3phasegrownontopofaninterfacial p 7 p 7 layer,whichresultsina p 21 p 21moirepattern.LEEDsimulations,overlayersimulations, andFFTanalysis,similartothemethodspreviouslyutilizedin Chapter6 ,werenecessaryto accuratelydeterminethethinstructureonAg.Additionally,Ialsoperformedroutine UHVoperationsandprovidedsystemsupportperiodicallythroughoutthisstudy.Formore informationregardingthisstudyrefertoReference[ 249 ]:J.Feng,S.R.Wagner,andP.P. Zhang, NatureSci.Rep. 5 ,10310(2015). 221 BIBLIOGRAPHY 222 BIBLIOGRAPHY [1] B.Cui, RecentAdvancesinNanofabricationTechniquesandApplications , (InTech,2011). [2] A.E.GrigorescuandC.W.Hagen, Nanotechnol. 20 ,292001(2009). [3] T.ItoandS.Okazaki, Nature 406 ,1027(2000). [4] C.Vieu,F.Carcenac,A.Pepin,Y.Chen,M.Mejias,A.Lebib,L.Manin- Ferlazzo,L.Couraud,H.Launois, Appl.Surf.Sci. 164 ,111(2000). [5] J.V.Barth,G.Costantini,andK.Kern, Nature 437 ,671(2005). [6] C.A.Mack, Proc.SPIEAdv.ResistMater.Process.Technol.XXVII 7639 , 763931(2010). [7] G.WitteandC.oll, J.Mater.Res. 19 ,1889(2004). [8] J.Hwang,A.Wan,andA.Kahn, Mat.Sci.Eng.R 64 ,1(2009). [9] J.V.Barth, Annu.Rev.Phys.Chem. 55 ,375(2007). [10] S.Duhm,G.Heimel,I.Salzmann,H.Glowatzki,R.L.Johnson,A.Vollmer, J.P.Rabe,andN.Koch, Nat.Mater. 7 ,326(2008). [11] J.Lewis, Mater.Today 9 ,38(2006). [12] S.R.Forrest, Nature 428 ,911(2004). [13] S.R.Forrest, Chem.Rev. 97 ,1793(1997). [14] C.G.Claessens,U.Hahn,andT.Torres, Chem.Rec. 8 ,75(2008). [15] K.K.McGrath,K.Jang,K.A.Robins,andD.-C.Lee, Chem.Eur.J. 15 , 4070(2009). 223 [16] J.A.A.W.Elemans,R.vanHameren,R.J.M.Nolte,andA.E.Rowan, Adv.Mater. 18 ,1251(2006). [17] D.Y.Kim,F.So,andY.Gao, SolarEnergyMater.Sol.Cells 93 ,1688(2009). [18] C.D.DimitrakopoulosandP.R.L.Malenfant, Adv.Mater. 14 ,99(2002). [19] L.S.HungandC.H.Chen, Mater.Sci.Eng.,R 39 ,143(2002). [20] R.R.Lunt,J.B.Benziger,andS.R.Forrest, Adv.Mater. 22 ,1233(2010). [21] G.E.Thayer,J.T.Sadowski,F.MeyerzuHeringdorf,T.Sakurai,andR.M. Tromp, Phys.Rev.Lett. 95 ,256106(2005). [22] S.Kowarik,A.Gerlach,andF.Schreiber, J.Phys.:Condens.Matter 20 , 184005(2008). [23] H.Huang,S.L.Wong,W.Chen,andA.T.S.Wee, J.Phys.D:Appl.Phys. 44 ,464005(2011). [24] D.Yokoyama, J.Mater.Chem. 21 ,19187(2011). [25] Z.-X.Xu,V.A.L.Roy,Z.-T.Liu,andC.S.Lee, Appl.Phys.Lett. 97 ,163301 (2010). [26] Y.-Y.Noh,J.-J.Kim,Y.Yoshida,andK.Yase, Adv.Mater. 15 ,699(2003). [27] L.Ramoino,M.vonArx,S.Schintke,A.BaratoH.-J.Guntherodt,andT. A.Jung, Chem.Phys.Lett. 417 ,22(2006). [28] A.Scarfato,S.-H.Chang,S.Kuck,J.Brede,G.andR.Wiesen- danger, Surf.Sci. 602 ,677(2008). [29] Y.Wang,J.oger,R.Berndt,andH.Tang, J.Am.Chem.Soc. 132 ,12546 (2010). [30] A.C.Brieva,T.E.Jenkins,D.G.Jones,F.ossner,D.A.Evans,andG. F.Clark, J.Appl.Phys. 99 ,073504(2006). 224 [31] A.C.Mayer,R.Ruiz,H.Zhou,R.L.Headrick,A.Kazimirov,andG.G. Malliaras, Phys.Rev.B 73 ,205307(2006). [32] A.R.Woll,T.V.Desai,andJ.R.Engstrom, Phys.Rev.B 84 ,075479(2011). [33] S.E.Fritz,S.M.Martin,C.D.Frisbie,M.D.Ward,andM.F.Toney, J. Am.Chem.Soc. 126 ,4084(2004). [34] M.Nakamura,T.Matsunobe,andH.Tokumoto, J.Appl.Phys. 89 ,2021 (1992). [35] M.Nakamura,Y.Morita,Y.Mori,A.Ishitani,andH.Tokumoto, J.Vac. Sci.Technol.B 14 ,1109(1996). [36] N.Nicoara, O.Paz,J.Mendez,A.M.o,J.M.Soler,andJ.M.omez- Rodr Phys.Rev.B 82 ,075402(2010). [37] Y.Makoudi,F.Palmino,M.Arab,E.Duverger,andF.Cherioux, J.Am. Chem.Soc. 130 ,6670(2008). [38] B.Baris,V.Luzet,E.Duverger,P.Sonnet,F.Palmino,andF.Cherioux, Angew.Chem.Int.Ed. 50 ,4094(2011). [39] S.R.Wagner,R.R.Lunt,andP.P.Zhang, Phys.Rev.Lett. 110 ,086107 (2013). [40] S.R.WagnerandP.P.Zhang, MaterRes.Soc.Symp.Proc. 1550 ,609(2013). [41] S.R.WagnerandP.P.Zhang, J.Phys.Chem.C 118 ,2194(2014). [42] S.R.WagnerandP.P.Zhang, Surf.Sci. 630 ,22(2014). [43] S.R.Wagner,B.Huang,C.Park,J.Feng,M.Yoon,andP.P.Zhang, Phys. Rev.Lett. 115 ,096101(2015). [44] K.Xiao,W.Deng,J.K.Keum,M.Yoon,I.V.Vlassiouk,K.W.Clark,A.-P. Li,I.I.Kravchenko,G.Gu,E.A.Payzant,B.G.Sumpter,S.C.Smith,J. F.Browning,andD.B.Geohegan, J.Am.Chem.Soc. 135 ,3680(2013). [45] A.Mugarza,R.Robles,C.Krull,R.ar,N.Lorente,andP.Gambardella, Phys.Rev.B 85 ,155437(2012). 225 [46] S.C.Bobaru,E.Salomon,J.-M.Layet,andT.Angot, J.Phys.Chem.C 115 ,5875(2011). [47] W.Dou,S.Huang,R.Q.Zhang,andC.S.Lee, J.Chem.Phys. 134 ,094705 (2011). [48] C.Stadler,S.Hansen,I.oger,C.Kumpf,andE.Umbach, Nat.Phys. 5 , 153(2009). [49] Y.Y.Zhang,S.X.Du,andH.-J.Gao, Phys.Rev.B 84 ,125446(2011). [50] Y.L.Huang,E.Wruss,D.A.Egger,S.Kera,N.Ueno,W.A.Saidi,T.Bucko, A.T.S.Wee,andE.Zojer, Molecules 19 ,2969(2014). [51] X.Lu,K.W.Hipps,X.D.Wang,andU.Mazur, J.Am.Chem.Soc. 118 , 7197(1996). [52] Q.Guo,Z.Qin,K.Zang,C.Liu,Y.Yu,andG.Cao, Langmuir 26 ,11804 (2010). [53] J.Uihlein,H.Peisert,M.Glaser,M.Polek,H.Adler,F.Petraki,R.Ovsyan- nikov,M.Bauer,andT.Chasse, J.Chem.Phys. 138 ,081101(2013). [54] S.ohnchen,S.Lukas,andG.Witte, J.Chem.Phys. 121 ,525(2004). [55] K.Yang,W.D.Xiao,Y.H.Jiang,H.G.Zhang,L.W.Liu,J.H.Mao,H.T. Zhou,S.X.Du,andH.-J.Gao, J.Phys.Chem.C 116 ,14052(2012). [56] S.K.ainen,M.Stepanova,R.Drost,P.Liljeroth,J.Lahtinen,andJ. Sainio, J.Phys.Chem.C 116 ,20433(2012). [57] P.arvinen,S.K.ainen,M.as,A.Harju,andP.Liljeroth, J.Phys. Chem.C 118 ,13320(2014). [58] M.Scardamaglia,C.Struzzi,S.Lizzit,M.Dalmiglio,P.Lacovig,A.Baraldi, C.Mariani,andM.G.Betti, Langmuir 29 ,10440(2013). [59] Y.-L.Wang,J.Ren,C.-L.Song,Y.-P.Jiang,L.-L.Wang,K.He,X.Chen, J.-F.Jia,S.Meng,E.Kaxiras,Q.-K.Xue,andX.-C.Ma, Phys.Rev.B 82 , 245420(2010). 226 [60] Y.Wakayama, J.Phys.Chem.C 11 ,2675(2007). [61] R.A.Rehman,W.Dou,H.Qian,H.Mao,F.Floether,H.Zhang,H.Li,P. He,andS.Bao, Surf.Sci. 606 ,1749(2012). [62] T.Bathon,P.Sessi,K.A.Kokh,O.E.Tereshchenko,andM.Bode, Nano Lett. 15 ,2442(2015). [63] P.Sessi,T.Bathon,K.A.Kokh,O.E.Tereshchenko,andM.Bode, Nano Lett. 14 ,5092(2014). [64] Y.C.YeoandT.-J.King, IEEEElectronDeviceLett. 21 ,161(2000). [65] Y.Xia,P.Yang,Y.Sun,Y.Wu,B.Mayers,B.Gates,Y.Yin,F.Kim,and H.Yan, Adv.Mater. 15 ,353(2003). [66] F.LeonardandA.A.Talin, Nat.Nanotechnol. 6 ,773(2011). [67] J.Stangl,V.Holy,andG.Bauer, Rev.Mod.Phys. 76 ,725(2004). [68] H.Y.Li,O.Wunnicke,M.T.om,W.G.G.Immink,M.H.M.van Weert,M.A.Verheijen,andE.P.A.M.Bakkers, NanoLett. 7 ,1144(2007). [69] M.T.ork,H.Schmid,J.Knoch,H.Riel,andW.Riess, Nat.Nanotechnol. 4 ,103(2009). [70] J.C.Ho,R.Yerushalmi,Z.A.Jacobson,Z.Fan,R.L.Alley,andA.Javey, Nat.Mater. 7 ,62(2008). [71] M.Li,G.Xing,L.F.N.A.Qune,G.Xing,T.Wu,C.H.A.Huan,X.Zhang, andT.C.Sum, Phys.Chem.Chem.Phys. 14 ,3075(2012). [72] A.D.Schricker,F.M.DavidsonIII,R.J.Wiacek,andB.A.Korgel, Nan- otechnology 17 ,2681(2006). [73] J.Jie,W.Zhang,K.Peng,G.Yuan,C.S.Lee,andS.-T.Lee, Adv.Funct. Mater. 18 ,3251(2008). [74] D.JenaandA.Konar, Phys.Rev.Lett. 98 ,136805(2007). 227 [75] J.Tang,K.W.Kemp,S.Hoogland,K.S.Jeong,H.Liu,L.Levina,M.Fu- rukawa,X.Wang,R.Debnath,D.Cha,K.W.Chou,A.Fischer,A.Amassian, J.B.Asbury,andE.H.Sargent, Nat.Mater. 10 ,765(2011). [76] M.V.Wolkin,J.Jorne,andP.M.Fauchet, Phys.Rev.Lett. 82 ,197(1999). [77] M.Soreni-Harari,N.Yaacobi-Gross,D.Steiner,A.Aharoni,U.Banin,O. Millo,andN.Tessler, NanoLett. 8 ,678(2008). [78] D.-H.Kim,R.N.Lu,andJ.A.Rogers, Annu.Rev.Biomed.Eng. 14 ,113(2012). [79] M.Ying,A.P.Bonifas,N.Lu,Y.Su,R.Li,H.Cheng,A.Ameen,Y.Huang, andJ.A.Rogers, Nanotechnology 23 ,344004(2012). [80] S.-W.Hwang,S.-K.Kang,X.Huang,M.A.Brenckle,F.G.Omenetto,and J.A.Rogers, Adv.Mater. 27 ,47(2015). [81] S.-K.Kang,S.-W.Hwang,S.Yu,J.-H.Seo,E.A.Corbin,J.Shin,D.S.Wie, R.Bashir,Z.Ma,andJ.A.Rogers, Adv.Funct.Mater. 25 ,1789(2015). [82] C.C.B.Bufon,J.D.A.Espinoza,D.J.Thurmer,M.Bauer,C.Deneke,U. Zschieschang,H.Klauk,andO.G.Schmidt, NanoLett. 11 ,3727(2011). [83] J.Yoon,A.J.Baca,S.-I.Park,P.Elvikis,J.B.GeddesIII,L.Li,R.H.Kim, J.Xiao,S.Wang,T.-H.Kim,M.J.Motala,B.Y.Ahn,E.B.Duoss,J.A. Lewis,R.G.Nuzzo,P.M.Ferreira,Y.Huang,A.Rockett,andJ.A.Rogers, Nat.Mater. 7 ,907(2008). [84] Y.Yang,Y.Hwang,H.A.Cho,J.-H.Song,S.-J.Park,J.A.Rogers,andH. C.Ko, Small 7 ,484(2011). [85] F.CavalloandM.G.Lagally, SoftMatter 6 ,439(2010). [86] C.Yan,W.Xi,W.Si,J.Deng,andO.G.Schmidt, Adv.Mater. 25 ,539 (2013). [87] P.P.Zhang,E.Tevaarwerk,B.-N.Park,D.E.Savage,G.K.Celler,I.Kneze- vic,P.G.Evans,M.A.Eriksson,andM.G.Lagally, Nature 439 ,703(2006). 228 [88] P.P.Zhang,E.P.Nordberg,B.-N.Park,G.K.Celler,I.Knezevic,P.G. Evans,M.A.Eriksson,andM.G.Lagally, NewJ.Phys. 8 ,200(2006). [89] W.Peng,Z.Aksamija,S.A.Scott,J.J.Endres,D.E.Savage,I.Knezevic, M.A.Eriksson,andM.G.Lagally, Nat.Commun. 4 ,1339(2013). [90] S.A.Scott,W.Peng,A.M.Kiefer,H.Jiang,I.Knezevic,D.E.Savage,M. A.Eriksson,andM.G.Lagally, ACSNano 3 ,1683(2009). [91] K.Byrappa,T.Ohachi,W.Michaeli,H.Warlimont,andE.Weber, Crystal GrowthTechnology ,(WilliamAndrewInc.,2003),Ch.2:\TheoryofCrystal GrowthfromVaporandSolution",25-54. [92] I.V.Markov, CrystalGrowthforBeginners:FundamentalsofNucleation, CrystalGrowth,andEpitaxy , 2Ed. ,(WorldScienPublishingCo.Pte. Ltd.,2003). [93] H.Luth, SolidSurfaces,InterfacesandThinFilms , 5Ed. ,(Springer,2010). [94] H.-C.JeongandE.D.Williams, Surf.Sci.Rep. 34 ,171(1999). [95] D.E.Savage,F.Liu,V.Zielasek,andM.G.Lagally, Semiconductorsand Semimetals 56 ,(AcademicPress,1999),Ch.2:\FundamentalMechanisms ofFilmGrowth",49-100. [96] F.Liu,M.Hohage,andM.G.Lagally, EncyclopediaofAppliedPhysics , Up- date2 ,(Wiley-VCH,1999),\SurfacesandInterfacesofSolids,Structureof", 321-352. [97] J.A.Venables,G.D.T.Spiller,andM.Hanbucken, Rep.Prog.Phys. 47 ,399 (1984). [98] Z.ZhangandM.G.Lagally, Science 276 ,377(1997). [99] R.L.SchoebelandE.J.Shipsey, J.Appl.Phys. 37 ,3682(1966). [100] S.-C.Li,Y.Han,J.-F.Jia,Q.-K.Xue,andF.Liu, Phys.Rev.B 74 ,195428 (2006). [101] H.Zheng,M.H.Xie,H.S.Wu,andQ.K.Xue, Phys.Rev.B 77 ,045303 (2008). 229 [102] M.J.Rost, Phys.Rev.Lett. 99 ,266101(2007). [103] J.W.Evans,P.A.Thiel,andM.C.Bartelt, Surf.Sci.Rep. 61 ,1(2006). [104] J.Krug,P.Politi,andT.Michely, Phys.Rev.B 61 ,14037(2000). [105] F.F.Leal,T.J.Oliveira,andS.C.Ferreira, J.Stat.Mech. ,P09018(2011). [106] M.KalP. Smilauer,G.Comsa,andT.Michely, Surf.Sci. 426 ,L447(1999). [107] I.ElkinaniandJ.Villain, SolidStateCommun. 87 ,105(1993). [108] M.H.Xie,S.M.Seutter,W.K.Zhu,L.X.Zheng,H.Wu,andS.Y.Tong, Phys.Rev.Lett. 82 ,2749(1999). [109] L.Bai,J.TandF.Liu, Phys.Rev.Lett. 92 ,225503(2004). [110] F.-J.MeyerzuHeringdorf,M.C.Reuter,andR.M.Tromp, Nature 412 ,517 (2001). [111] M.B.Casu,S.-A.Savu,B.-E.Schuster,I.Biswas,C.Raisch,H.Marchetto, Th.Schmidt,andT.Chasse, Chem.Commun. 48 ,6957(2012). [112] S.Verlaak,S.Steudel,P.Heremans,D.Janssen,andM.S.Deleuze, Phys. Rev.B 68 ,195409(2003). [113] R.Ruiz,D.Choudhary,B.Nickel,T.Toccoli,K.-C.Chang,A.C.Mayer, P.Clancy,J.M.Blakely,R.L.Headrick,S.Iannotta,andG.G.Malliaras, Chem.Mater. 16 ,4497(2004). [114] T.V.Desai,A.R.Woll,F.Schreiber,andJ.R.Engstrom, J.Phys.Chem. C 114 ,20120(2010). [115] J.E.Goose,E.L.First,andP.Clancy, Phys.Rev.B 81 ,205310(2010). [116] G.HlawacekandC.Teichert, J.Phys.:Condens.Matter 25 ,143202(2013). [117] G.Hlawacek,P.Puschnig,P.Frank,A.Winkler,C.Ambrosch-Draxl,andC. Teichert, Science 321 ,108(2008). 230 [118] X.N.Zhang,E.Barrena,D.Goswami,D.G.deOteyza,C.Weis,andH. Dosch, Phys.Rev.Lett. 103 ,136101(2009). [119] B.Krause,F.Schreiber,H.Dosch,A.Pimpinelli,andO.H.Seeck, Europhys. Lett. 65 ,372(2004). [120] E.Kol'tsov,T.Basova,andP.Semyannikov, Mater.Chem.Phys. 86 ,222 (2004). [121] M.Breeman,G.T.Barkema,andD.O.Boerma, Surf.Sci. 323 ,71(1995). [122] M.GiesenandH.Ibach, Surf.Sci. 431 ,109(1999). [123] D.G.deOteyza,E.Barrena,H.Dosch,andY.Wakayama, Phys.Chem. Chem.Phys. 11 ,8741(2009). [124] D.Hao,C.Song,Y.Ning,Y.Wang,L.Wang,X.-C.Ma,X.Chen,andQ.-K. Xue, J.Chem.Phys. 134 ,154703(2011). [125] T.Kobayashi,Y.Fujiyoshi,F.Iwatsu,andN.Uyeda, ActaCryst. A37 ,692 (1981). [126] J.otzen,S.Lukas,A.Birkner,andG.Witte, Surf.Sci. 605 ,577(2011). [127] M.ohr,M.Gabriel,R.oller, Surf.Sci. 507-510 ,330(2002). [128] X.N.Zhang,E.Barrena,D.G.deOteyza,andH.Dosch, Surf.Sci. 601 , 2420(2007). [129] X.N.Zhang,E.Barrena,D.Goswami,D.G.deOteyza,C.Weis,andH. Dosch, Phys.Rev.Lett. 103 ,136101(2009). [130] V.Kalihari,E.B.Tadmor,G.Haugstad,andC.D.Frisbie, Adv.Mater. 20 , 4033(2008). [131] P.Kury,K.R.Roos,D.Thien,S.ollenbeck,D.Wall,M.Horn-vonHoegen, andF.-J.MeyerzuHeringdorf, Org.Electron. 9 ,461(2008). [132] Y.Suzuki,M.Hietschold,andD.R.T.Zahn, Appl.Surf.Sci. 252 ,5449 (2006). 231 [133] S.Nishikata,G.Sazaki,J.T.Sadowski,A.Al-Mahboob,T.Nishihara,Y. Fujikawa,S.Suto,T.Sakurai,andK.Nakajima, Phys.Rev.B 76 ,165424 (2007). [134] T.Shimada,H.Nogawa,T.Hasegawa,R.Okada,H.Ichikawa,K.Ueno,and K.Saiki, Appl.Phys.Lett. 87 ,061917(2005). [135] R.Wilson, SurfaceAnalysis-ThePrincipalTechniques , 2Ed. ,(JohnWiley andSons,Ltd,2009),Appendix1:\VacuumTechnologyforAppliedSurface Science",613-647. [136] H.BinnigandG.Rohrer, Helv.Phys.Acta. 55 ,726(1982). [137] G.Binnig,H.Rohrer,C.Gerber,andE.Weibel, Appl.Phys.Lett. 40 ,178 (1982). [138] G.Binnig,H.Rohrer,C.Gerber,andE.Weibel, Phys.Rev.Lett. 49 ,57 (1982). [139] J.StroscioandW.Kaiser, MethodsofExperimentalPhysics 27 ,(Academic Press,1993),\ScanningtunnelingMicroscopy". [140] C.J.Chen, MonographsonthePhysicsandChemistryofMaterials 64 , 2Ed. , (OxfordSciencePublications,2008),\IntroductiontoScanningTunneling Microscopy". [141] G.J.Leggett, SurfaceAnalysis-ThePrincipalTechniques , 2Ed. ,(John WileyandSons,Ltd,2009),Ch.9:\ScanningProbeMicroscopy",479-562. [142] J.TandD.R.Hamann, Phys.Rev.Lett. 50 ,1998(1983). [143] J.TandD.R.Hamann, Phys.Rev.B 31 ,805(1985). [144] J.Bardeen, Phys.Rev.Lett. 6 ,57(1961). [145] G.Binnig,C.F.Quate,andC.Gerber, Phys.Rev.Lett. 56 ,930(1986). [146] R.Proksch,T.E.ScJ.P.Cleveland,R.C.Callahan,andM.B.Viani, Nanotechnology 15 ,1344(2004). 232 [147] Y.Martin,C.C.Williams,andH.K.Wickramasinghe, J.Appl.Phys. 61 ,4723 (1987). [148] H.-J.Butt,B.Cappella,andM.Kappl, Surf.Sci.Rep. 59 ,1(2005). [149] N.JaliliandK.Laxminarayana, Mechatronics 14 ,907(2004). [150] S.SadewasserandT.Glatzel, KelvinProbeForceMicroscopy:Measuringand CompensatingElectrostaticForces ,(Springer,2012). [151] Q.Zhong,D.Innis,K.Kjoller,andV.B.Elings, Surf.Sci.Lett. 290 ,L688 (1993). [152] L.Kelvin, Phil.Mag. 46 ,82(1898). [153] M.Nonnenmacher,M.P.O'Boyle,andH.K.Wickramasinghe, Appl.Phys. Lett. 58 ,2921(1991). [154] W.Melitz,J.Shen,A.C.Kummel,andS.Lee, Surf.Sci.Rep. 66 ,1(2011). [155] M.P.SeahandW.A.Dench, Surf.InterfaceAnal. 1 ,2(1979). [156] M.Vos,S.G.Anderson,andJ.H.Weaver, Phys.Rev.B 39 ,3274(1989). [157] C.J.DavissonandL.H.Germer, Phys.Rev. 30 ,705(1927). [158] E.J.Scheiber,L.H.Germer,andC.D.Hartman, Rev.Sci.Inst. 31 ,112 (1960). [159] C.A.Lucas, SurfaceAnalysis-ThePrincipalTechniques , 2Ed. ,(JohnWiley andSons,Ltd,2009),Ch.8:\SurfaceStructureDeterminationbyInterference Techniques",391-478. [160] M.A.VanHove,W.H.Weinberg,andC.-M.Chan, Low-EnergyElectron action:Experiment,Theory,andSurfaceStructureDetermination 6 , (Springer,1986). [161] L.J.vanderPauw, PhilipsRes.Rep. 13 ,13(1958). [162] D.W.Koon, Rev.Sci.Inst. 60 ,271(1989). 233 [163] O.Bierwagen,T.Ive,C.G.VandeWalle,andJ.S.Speck, Appl.Phys.Lett. 93 ,242108(2008). [164] E.H.Hall, Am.J.Math. 2 ,287(1879). [165] D.E.Hooks,T.Fritz,andM.D.Ward, Adv.Mater. 13 ,227(2001). [166] P.HohenbergandW.Kohn, Phys.Rev. 136 ,B864(1964). [167] W.KohnandL.J.Sham, Phys.Rev. 140 ,A1133(1965). [168] G.KresseandJ.Furthmuller, Comput.Mater.Sci. 6 ,15(1996). [169] A.TkatchenkoandM.Sc Phys.Rev.Lett. 102 ,073005(2009). [170] N.Marom,O.Hod,G.E.Scuseria,andL.Kronik, J.Chem.Phys. 128 , 164107(2008). [171] N.MaromandL.Kronik, Appl.Phys.A 95 ,159(2009). [172] S.M.SzeandK.K.Ng, PhysicsofSemiconductorDevices , 3Ed. ,(John WileyandSons,Inc.,2007). [173] D.K.Schroder, SemiconductorMaterialandDeviceCharacterization , 3ed. , (Wiley-IEEEPress,2015). [174] I.-W.Lyo,E.Kaxiras,andPh.Avouris, Phys.Rev.Lett. 63 ,1261(1989). [175] A.V.Zotov,M.A.Kulakov,S.V.Ryzhkov,A.A.Saranin,V.G.Lifshits, B.Bullemer,andI.Eisele, Surf.Sci. 345 ,313(1996). [176] E.J.Spadafora,J.Berger,P.Mutombo,M.Telychko,M. Svec,Z.Majzik,A. B.McLean,andP.Jel J.Phys.Chem.C 118 ,15744(2014). [177] M.A.Hines, Annu.Rev.Phys.Chem. 54 ,29(2003). [178] G.K.CellerandS.Cristoloveanu, J.Appl.Phys. 93 ,4955(2003). [179] Y.K.Choi,K.Asano,N.Lindert,V.Subramanian,T.-J.King,J.Bokor,and H.Chenming, IEEEElectronDeviceLett. 21 ,254(2000). 234 [180] M.M.Roberts,L.J.Klein,D.E.Savage,K.A.Slinker,M.Friesen,G.Celler, M.A.Eriksson,andM.G.Lagally, Nat.Mater. 5 ,388(2006). [181] P.Strobel,M.Riedel,J.Ristein,andL.Ley, Nature 430 ,439(2004). [182] JurgenRistein, Science 313 ,1057(2006). [183] H.Ishii,K.Sugiyama,E.Ito,andK.Seki, Adv.Mater. 11 ,605(1999). [184] G.Ashkenasy,D.Cahen,R.Cohen,A.Shanzer,andA.Vilan, Acc.Chem. Res. 35 ,121(2002). [185] N.B.McKeown, PhthalocyanineMaterials:Synthesis,Structure,andFunc- tion ,(CambridgeUniversityPress,1998). [186] K.M.Kadish,K.M.Smith,andR.Guilard, ThePorphyrinHandbook:Ap- plicationsofPhthalocyanines 19 ,(ElsevierInc.,2003). [187] K.W.Hipps,D.E.Barlow,andU.Mazur, J.Phys.Chem.B 104 ,2444 (2000). [188] M.Lackinger,T.Muller,T.G.Gopakumar,F.Muller,M.Hietschold,and G.W.Flynn, J.Phys.Chem.B 108 ,2279(2004). [189] J.Rivnay,L.H.Jimison,J.E.Northrup,M.F.Toney,R.Noriega,S.Lu,T. J.Marks,A.Facchetti,andA.Salleo, Nat.Mater. 8 ,952(2009). [190] A.B.ChwangandC.D.Frisbie, J.Appl.Phys. 90 ,1342(2001). [191] M.FendrichandJ.Krug, Phys.Rev.B 76 ,121302(R)(2007). [192] S.Yim,K.-ilKim,andT.S.Jones, J.Phys.Chem.C 111 ,10993(2007). [193] S.Zorba,Y.Shapir,andY.Gao, Phys.Rev.B 74 ,245410(2006). [194] R.Ruiz,B.Nickel,N.Koch,L.C.Feldman,R.F.Haglund,A.Kahn,andG. Scoles, Phys.Rev.B 67 ,125406(2003). [195] P.Clancy, Chem.Mater. 23 ,522(2011). 235 [196] F.-C.Chen,Y.-P.Chen,Y.-J.Huang,andS.-C.Chien, J.Phys.D:Appl. Phys. 43 ,405103(2010). [197] B.R.Conrad,J.Tosado,G.Dutton,D.B.Dougherty,W.Jin,T.Bonnen, A.Schuldenfrei,W.G.Cullen,E.D.Williams,J.E.Reutt-Robey,andS.W. Robey, Appl.Phys.Lett. 95 ,213302(2009). [198] S.vanDijken,L.C.Jorritsma,andB.Poelsema, Phys.Rev.Lett. 82 ,4038 (1999). [199] J.Seo,H.-Y.Kim,andJ.-S.Kim, Phys.Rev.B 71 ,075414(2005). [200] A.Toma,D.Chiappe,B. SetinaBatic,M.Godec,M.Jenko,andF.Buatier deMongeot, Phys.Rev.B 78 ,153406(2008). [201] F.F.Leal,S.C.Ferreira,andS.O.Ferreira, J.Phys.:Condens.Matter 23 , 292201(2011). [202] R.ZhangandH.Huang, Appl.Phys.Lett. 98 ,221903(2011). [203] A.Toma,B. SetinaBatic,D.Chiappe,C.Boragno,U.Valbusa,M.Godec, M.Jenko,andF.BuatierdeMongeot, J.Appl.Phys. 104 ,104313(2008). [204] J.H.KangandX.-Y.Zhu, Chem.Mater. 18 ,1318(2006). [205] G.Witte,K.anel,S.ohnchen,andCh.oll, Appl.Phys.A 82 ,447(2006). [206] D.V.Klyachko,J.-M.Lopez-Castillo,andJ.-P.Jay-Gerin, Phys.Rev.B 60 , 9026(1999). [207] Y.L.Wang,W.Ji,D.X.Shi,S.X.Du,C.Seidel,Y.G.Ma,H.-J.Gao,L. F.Chi,andH.Fuchs, Phys.Rev.B 69 ,075408(2004). [208] A.H.Scafer,C.Seidel,andH.Fuchs, Adv.Funct.Mater. 11 ,193(2001). [209] J.otzen,D.afer,C.oll,andG.Witte, Phys.Rev.B 81 ,085440(2010). [210] L.Meng,R.Wu,L.Zhang,L.Li,S.Du,Y.Wang,H.-J.Gao, J.Phys.: Condens.Matter 24 ,314214(2012). 236 [211] L.Kilian,U.Stahl,I.Kossev,M.Sokolowski,R.Fink,andE.Umbach, Surf. Sci. 602 ,2427(2008). [212] M.C.Cottin,J.ScA.Sonntag,H.Karacuban,R.oller,andC.A. Bobisch, Appl.Surf.Sci. 258 ,2196(2012). [213] A.Ueno,T.Fujita,M.Matsue,H.Yanagisawa,C.Oshima,F.Patthey,H.-C. Ploigt,W.-D.Schneider,andS.Otani, Surf.Sci. 600 ,3518(2006). [214] A.Crepaldi,S.Pons,E.Frantzeskakis,F.Calleja,M.Etzkorn,A.P.Seitso- nen,K.Kern,H.Brune,andM.Grioni, Phys.Rev.B 87 ,115138(2013). [215] N.J.Taylor, Surf.Sci. 4 ,161(1966). [216] E.Bauer, Surf.Sci. 7 ,351(1967). [217] E.G.McRae, Surf.Sci. 11 ,492(1968). [218] M.Meissner,M.Gruenewald,F.Sojka,C.Udhardt,R.Forker,andT.Fritz, Surf.Sci. 606 ,1709(2012). [219] J.M.Carpinelli,H.H.Weitering,E.W.Plummer,andR.Stumpf, Nature 381 ,398(1996). [220] A.T.N'Diaye,J.Coraux,T.N.Plasa,C.Busse,andT.Michely, NewJ. Phys. 10 ,043033(2008). [221] G.Beernink,T.Strunskus,G.Witte,andCh.oll, Appl.Phys.Lett. 85 , 398(2004). [222] D.afer,L.Ruppel,andG.Witte, Phys.Rev.B 75 ,085309(2007). [223] W.Dou,Y.Tang,C.S.Lee,S.N.Bao,S.T.Lee, J.Chem.Phys. 133 ,144704 (2010). [224] M.Nakamura,Y.Morita,andH.Tokumoto, Appl.Surf.Sci. 113/114 ,316 (1997). [225] L.Bartels, Nat.Chem. 2 ,87(2010). 237 [226] T.-C.Tseng,C.Urban,Y.Wang,R.Otero,S.L.Tait,M.AlcamD. Ecija, M.Trelka,J.M.Gallego,N.Lin,M.Konuma,U.Starke,A.Nefedov,A. Langner,C.oll,M. A.Herranz,F.MartN.MartK.Kern,andR. Miranda, Nat.Chem. 2 ,374(2010). [227] G.Heimel,S.Duhm,I.Salzmann,A.Gerlach,A.Strozecka,J.Niederhausen, C.Burker,T.Hosokai,I.Fernandez-Torrente,G.Schulze,S.Winkler,A. Wilke,R.Schlesinger,J.Frisch,B.Broker,A.Vollmer,B.Detlefs,J. S.Kera,K.J.Franke,N.Ueno,J.I.Pascual,F.Schreiber,andN.Koch, Nat. Chem. 5 ,187(2013). [228] L.Scudiero,D.E.Barlow,U.Mazur,andK.W.Hipps, J.Am.Chem.Soc. 123 ,4073(2001). [229] B.C.Stipe,M.A.Rezaei,andW.Ho, Science 279 ,1907(1998). [230] Y.NishidaandS.Kida, Coord.Chem.Rev. 27 ,275(1979). [231] M.-S.LiaoandS.Scheiner, J.Chem.Phys. 114 ,9780(2001). [232] W.A.Harrison, ElectronicStructureandthePropertiesofSolids ,(Dover Publications,1989). [233] E.Kaxiras,K.C.Pandey,F.J.Himpsel,andR.M.Tromp, Phys.Rev.B 41 , 1262(1990). [234] R.Cuadrado,J.I.Cera,Y.Wang,G.Xin,R.Berndt,andH.Tang, J. Chem.Phys. 133 ,154701(2010). [235] S.-H.Chang,S.Kuck,J.Brede,L.Lichtenstein,G.andR.Wiesen- danger, Phys.Rev.B 78 ,233409(2008). [236] J.D.Baran,J.A.Larsson,R.A.J.Woolley,Y.Cong,P.J.Moriarty,A.A. Cafolla,K.Schulte,andV.R.Dhanak, Phys.Rev.B 81 ,075413(2010). [237] J.D.BaranandJ.A.Larsson, J.Phys.Chem.C 117 ,23887(2013). [238] T.G.Gopakumar,T.Brumme,J.oger,C.Toher,G.Cuniberti,andR. Berndt, J.Phys.Chem.C 115 ,12173(2011). 238 [239] G.DiSanto,C.Castellarin-Cudia,M.Fanetti,B.Taleatu,P.Borghetti,L. Sangaletti,L.Floreano,E.Magnano,F.Bondino,andA.Goldoni, J.Phys. Chem.C 115 ,4155(2011). [240] T.Lukasczyk,K.Flechtner,L.R.Merte,N.Jux,F.Maier,J.M.Gottfried, andH.-P.Steinruck, J.Phys.Chem.C 111 ,3090(2007). [241] F.Buchner,K.-G.Warnick,T.A.orling,H.-P.Steinruck,W.Hi- eringer,andH.Marbach, J.Phys.Chem.C 113 ,16450(2009). [242] W.Auwarter,K.Seufert,F.Klappenberger,J.Reichert,A.Weber-Bargioni, A.Verdini,D.Cvetko,M.Dell'Angela,L.Floreano,A.Cossaro,G.Bavdek, A.Morgante,A.P.Seitsonen,andJ.V.Barth, Phys.Rev.B 81 ,245403 (2010). [243] K.Berke,S.Tongay,M.A.McCarthy,A.G.Rinzler,B.R.Appleton,andA. F.Hebard, J.Phys.:Condens.Matter 24 ,255802(2012). [244] I.Arbi,B.B.Hamada,A.Souissi,S.Menzli,C.B.Azzouz,A.Laribi,A. Akremi,andC. Appl.Surf.Sci. 305 ,396(2014). [245] B.B.Hamada,A.Souissi,S.Menzli,I.Arbi,A.Akremi,C.andM. Derivaz, ThinSolidFilms 567 ,96(2014). [246] M.GorgoiandD.R.T.Zahn, Org.Electron. 6 ,168(2005). [247] J.Sun,K.Pimcharoen,S.R.Wagner,P.M.Duxbury,andP.P.Zhang, Org. Electron. 15 ,441(2014). [248] C.J.Traverse,M.Young,S.R.Wagner,P.P.Zhang,P.Askeland,M.C. Barr,andR.R.Lunt, J.Appl.Phys. 115 ,194505(2014). [249] J.Feng,S.R.Wagner,andP.P.Zhang, NatureSci.Rep. 5 ,10310(2015). 239