SensitivitySensitivity ofof secondsecond harmonicharmonic ...
8
Sensitivity Sensitivity of of second second harmonic harmonic generation generation to to space space charge charge effects effects at at Si(111 Si(111 )/electrolyte )/electrolyte and and Si(111 Si(111 interfaces interfaces P. P. R. R. Fischer, Fischer, J. J. L. L. Daschbach, Daschbach, D. D. E. E. Gragson, Gragson, and and G. G. L. L. Richmond Richmond Department Department of of Chemistry, Chemistry, University University of of Oregon, Oregon, Eugene, Eugene, Oregon Oregon 97403 97403 (Received (Received 11 11 October October 1993; 1993; accepted accepted 15 15 June June 1994) 1994) The The potential potential dependence dependence in in the the surlace surlace second second hannonic hannonic response response from from hydrogen hydrogen terminated terminated n-Si(111) n-Si(l1l) and and oxidized oxidized n-Si(1l1) n-Si(1ll) surfaces surfaces has has been been examined examined in in aqueous aqueous NH 4 F and and H 2 2 S0 S0 4 4 NH4F solutions. solutions. The The relative relative phase phase of of the the nonlinear nonlinear response response as as measured measured by by rotational rotational anisotropy anisotropy experiments experiments is is found found to to be be highly highly sensitive sensitive to to the the presence presence of of the the oxide oxide and and the the field field applied applied across across Si(ll I)/mdde/electrolyte the the Si(lll)/oxide/electrolyte interlace. interlace. These These observations observations are are attributed attributed to to field field effects effects within within the the space-charge space-charge region region of of the the semiconductor semiconductor which which vary vary with with the the presence presence and and thickness thickness of of the the Si( 11 1) insulating insulating oxide oxide layer layer on on the the Si(lll) surlace. surlace. I. I. INTRODUCTION INTRODUCTION Understanding Understanding the the electrochemical electrochemical behavior behavior of of semicon- semicon- ductor ductor surlaces surlaces remains remains challenging challenging due due to to the the highly highly reac- reac- tive tive nature nature of of these these surlaces. surlaces. Nevertheless, Nevertheless, studies studies continue continue to to be be performed performed in in this this area area because because of of the the technological technological importance importance of of these these materials, materials, particularly particularly Si Si and and GaAs. GaAs. In In recent recent years, years, significant significant progress progress has has been been made made in in charac- charac- terizing terizing the the silicon silicon surface surface and and the the silicon/electrolyte silicon/electrolyte inter- inter- face face by by infrared infrared reflectance, reflectance, 1 1,2 ,2 ultrahigh ultrahigh vacuum vacuum (UHV) (UHV) transfer transfer experiments,3,4 experiments,3,4 scanning scanning tunneling tunneling microscope microscope (STM),5-10 (STM),5-10 and and atomic atomic force force microscope microscope (AFM)ll (AFM)ll imaging. imaging. Many Many of of these these studies studies have have taken taken advantage advantage of of the the relative relative stability stability to to oxidation oxidation of of silicon silicon surlaces surlaces prepared prepared in in a hydro- hydro- gen gen terminated terminated state. state. Much Much less less is is known known about about the the oxidized oxidized Si Si surface surface frequently frequently found found in in both both air air and and water water owing owing to to the the inapplicability inapplicability of of many many of of the the above above techniques techniques for for ex- ex- amining amining the the buried buried SilSi0 SilSi0 2 2 interfacial interfacial region. region. The The presence presence of of this this oxide oxide can can significantly significantly alter alter the the electrochemistry electrochemistry and and the the electronic electronic properties properties of of the the semiconductor, semiconductor, as as it it both both in- in- sulates sulates the the surlace surlace and and is is capable capable of of storing storing significant significant charge. charge. Second Second harmonic harmonic generation generation (SHG) (SHG) provides provides an an al- al- ternative ternative means means of of examining examining the the silicon silicon surface surface and and buried buried 14 14 SilSiO z - - SilSi0 2 interlace. interlace. 12 12 In In this this article, article, we we discuss discuss the the application application of of optical optical SHG SHG t6 studying studying Si(lll)/electrolyte and and Si(111)/electrolyte Si(l1l)/Si0 2 2 /electrolyte /electrolyte to Si(1U)/Si0 interfaces interfaces for for which which oxides oxides have have been been photoanodically photoanodically grown grown on on the the surlace. surlace. There There are are two two primary primary objectives objectives in in this this study. study. The The first first is is to to correlate correlate the the intensity intensity and and relative relative phase phase of of the the SH SH optical optical response response with with the the potential potential induced induced semiconductor/ variations variations in in the the static static field field applied applied to to the the semiconductorl electrolyte electrolyte interface, interface, and and related related to to this, this, how how the the presence presence of of oxide oxide layers layers of of varied varied thicknesses thicknesses on on the the Si(lll) Si(lll) surface surface al- al- ters ters this this potential potential dependence. dependence. Unlike Unlike metal metal electrodes, electrodes, where where SHG SHG is is highly highly sensitive sensitive to to potential potential variation variation due due to to the the screening screening of of the the charge charge at at the the surface,15,16 surface,15,16 the the field field for for semiconductors semiconductors can can extend extend several several hundred hundred angstroms angstroms into into the the material, material, the the depth depth being being dependent dependent upon upon the the doping doping den- den- sity sity of of the the semiconductor, semiconductor, possible possible Fermi-level Fermi-level pinning, pinning, and and the the strength strength of of the the applied applied field. field. The The second second focus focus of of this this study study has has been been of of a more more funda- funda- mental mental nature. nature. That That is is to to address address issues issues related related to to the the source source of of the the response response from from these these buried buried interfaces, interfaces, the the relative relative con- con- Si(111) tributions tributions from from the the depletion depletion layer, layer, the the Si(lll) surface surface adja- adja- cent cent to to the the electrolyte electrolyte or or oxide, oxide, and and the the Si0 Si0 2 2 layer, layer, as as well well as as determining determining the the coupling coupling of of the the applied applied static static field field to to various various susceptibility susceptibility tensors. tensors. For For semiconductors, semiconductors, the the bulk bulk response response can can be be significant significant and and thus thus any any study study which which attempts attempts to to un- un- derstand derstand the the source source of of the the response response must must be be concerned concerned with with the the higher higher order order terms. terms. In In the the present present study, study, this this issue issue has has been been addressed addressed by by potential potential dependent dependent measurements. measurements. The The H-Si -Si (111) (111) and and the the oxidized oxidized surface surface were were examined examined in in different different electrolyte electrolyte solutions solutions and and in in UHY. UHY. Much Much of of the the work work reported reported involves involves measuring measuring the the variation variation in in the the SH SH response response with with azimuthal azimuthal rotation rotation of of the the sample sample by by 360 360 0 0 about about its its normal. normal. The The simplest simplest system system is is a H-terminated H-terminated sample sample examined examined under under potential potential control control in in NH4F. NH 4 F. This This surface surface should should be be monohydrogen monohydrogen terminated, terminated, smooth, smooth, and and relatively relatively free free of of any any photogenerated photogenerated surface surface oxides oxides due due to to the the solubil- solubil- ity ity of of the the oxides oxides in in this this solution. solution. 17 17 The The results results obtained obtained at at the the flat flatband band potential potential are are compared compared with with similar similar rotational rotational an- an- isotropy isotropy measurements measurements conducted conducted in in UHV UHV for for a sputtered sputtered and and annealed annealed sample, sample, and and a hydrogen hydrogen terminated terminated sample sample transferred transferred to to UHV UHV without without further further surface surface cleaning. cleaning. These These studies studies are are followed followed by by potential potential dependent dependent measurements measurements of of the the SH SH rotational rotational anisotropy anisotropy from from initially initially H-terminated H-terminated sur- sur- NH4F face face immersed immersed in in a H 2 2 S0 S0 4 4 solution solution where, where, unlike unlike the the NH 4 F solution, solution, the the surface surface oxides oxides which which are are formed formed photoanodi- photoanodi- cally cally are are insoluble. insoluble. Similarly, Similarly, experiments experiments are are conducted conducted on on samples samples immersed immersed in in H 2 2 S0 S0 4 4 on on which which controlled controlled amounts amounts of of surlace surlace oxides oxides (0-40 (0-40 A) A) are are photoanodically photoanodically grown grown prior prior to to examination examination of of the the SH SH potential potential dependence. dependence. II. II. DESCRIPTION DESCRIPTION OF OF SHG SHG FROM FROM CUBIC CUBIC MEDIA MEDIA When When the the electromagnetic electromagnetic field field polarizes polarizes a medium, medium, this this interaction interaction is is governed governed by by Maxwell's Maxwell's equations equations and and the the con- con- stitutive stitutive relations. relations. The The second second order order polarization polarization that that gives gives rise rise to to SHG SHG can can be be expressed expressed by by a series series muItipole of of multipole terms, terms, 18,19 18,19 0734·2101/94112(5)12617/81$1.00 2617 2617 J. J. Vac. Vac. Sci. Sci. Technol. Technol. A 12(5), 12(5), Sep/Oct Sep/Oct 1994 1994 0734-2101/94112(5)/2617/81$1.00 ©1994 ©1994 American American Vacuum Vacuum Society Society 2617 2617
Transcript of SensitivitySensitivity ofof secondsecond harmonicharmonic ...
SensitivitySensitivity ofof secondsecond harmonicharmonic generationgeneration toto spacespace chargecharge effectseffects atat Si(111Si(111 )/electrolyte)/electrolyte andand Si(111Si(111 )/SiO~electrolyte)/SiO~electrolyte interfacesinterfaces
P.P. R.R. Fischer,Fischer, J.J. L.L. Daschbach,Daschbach, D.D. E.E. Gragson,Gragson, andand G.G. L.L. RichmondRichmond DepartmentDepartment ofof Chemistry,Chemistry, UniversityUniversity ofof Oregon,Oregon, Eugene,Eugene, OregonOregon 9740397403
(Received(Received 1111 OctoberOctober 1993;1993; acceptedaccepted 1515 JuneJune 1994)1994)
TheThe potentialpotential dependencedependence inin thethe surlacesurlace secondsecond hannonichannonic responseresponse fromfrom hydrogenhydrogen terminatedterminated n-Si(111)n-Si(l1l) andand oxidizedoxidized n-Si(1l1)n-Si(1ll) surfacessurfaces hashas beenbeen examinedexamined inin aqueousaqueous NH4F andand HH22S0S044NH4F solutions.solutions. TheThe relativerelative phasephase ofof thethe nonlinearnonlinear responseresponse asas measuredmeasured byby rotationalrotational anisotropyanisotropy experimentsexperiments isis foundfound toto bebe highlyhighly sensitivesensitive toto thethe presencepresence ofof thethe oxideoxide andand thethe fieldfield appliedapplied acrossacross
Si(ll I)/mdde/electrolytethethe Si(lll)/oxide/electrolyte interlace.interlace. TheseThese observationsobservations areare attributedattributed toto fieldfield effectseffects withinwithin thethe space-chargespace-charge regionregion ofof thethe semiconductorsemiconductor whichwhich varyvary withwith thethe presencepresence andand thicknessthickness ofof thethe
Si(11 1)insulatinginsulating oxideoxide layerlayer onon thethe Si(lll) surlace.surlace.
I.I. INTRODUCTIONINTRODUCTION
UnderstandingUnderstanding thethe electrochemicalelectrochemical behaviorbehavior ofof semiconsemiconductorductor surlacessurlaces remainsremains challengingchallenging duedue toto thethe highlyhighly reacreactivetive naturenature ofof thesethese surlaces.surlaces. Nevertheless,Nevertheless, studiesstudies continuecontinue toto bebe performedperformed inin thisthis areaarea becausebecause ofof thethe technologicaltechnological importanceimportance ofof thesethese materials,materials, particularlyparticularly SiSi andand GaAs.GaAs. InIn recentrecent years,years, significantsignificant progressprogress hashas beenbeen mademade inin characcharacterizingterizing thethe siliconsilicon surfacesurface andand thethe silicon/electrolytesilicon/electrolyte interinterfaceface byby infraredinfrared reflectance,reflectance,1
1,2,2 ultrahighultrahigh vacuumvacuum (UHV)(UHV) transfertransfer experiments,3,4experiments,3,4 scanningscanning tunnelingtunneling microscopemicroscope (STM),5-10(STM),5-10 andand atomicatomic forceforce microscopemicroscope (AFM)ll(AFM)ll imaging.imaging. ManyMany ofof thesethese studiesstudies havehave takentaken advantageadvantage ofof thethe relativerelative stabilitystability toto oxidationoxidation ofof siliconsilicon surlacessurlaces preparedprepared inin aa hydrohydrogengen terminatedterminated state.state. MuchMuch lessless isis knownknown aboutabout thethe oxidizedoxidized SiSi surfacesurface frequentlyfrequently foundfound inin bothboth airair andand waterwater owingowing toto thethe inapplicabilityinapplicability ofof manymany ofof thethe aboveabove techniquestechniques forfor exexaminingamining thethe buriedburied SilSi0SilSi022 interfacialinterfacial region.region. TheThe presencepresence ofof thisthis oxideoxide cancan significantlysignificantly alteralter thethe electrochemistryelectrochemistry andand thethe electronicelectronic propertiesproperties ofof thethe semiconductor,semiconductor, asas itit bothboth ininsulatessulates thethe surlacesurlace andand isis capablecapable ofof storingstoring significantsignificant charge.charge. SecondSecond harmonicharmonic generationgeneration (SHG)(SHG) providesprovides anan alalternativeternative meansmeans ofof examiningexamining thethe siliconsilicon surfacesurface andand buriedburied
1414SilSiOz -SilSi02 interlace.interlace. 1212
InIn thisthis article,article, wewe discussdiscuss thethe applicationapplication ofof opticaloptical SHGSHG t6 studyingstudying Si(lll)/electrolyte andandSi(111)/electrolyte Si(l1l)/Si022/electrolyte/electrolyteto Si(1U)/Si0 interfacesinterfaces forfor whichwhich oxidesoxides havehave beenbeen photoanodicallyphotoanodically growngrown onon thethe surlace.surlace. ThereThere areare twotwo primaryprimary objectivesobjectives inin thisthis study.study. TheThe firstfirst isis toto correlatecorrelate thethe intensityintensity andand relativerelative phasephase ofof thethe SHSH opticaloptical responseresponse withwith thethe potentialpotential inducedinduced
semiconductor/variationsvariations inin thethe staticstatic fieldfield appliedapplied toto thethe semiconductorl electrolyteelectrolyte interface,interface, andand relatedrelated toto this,this, howhow thethe presencepresence ofof oxideoxide layerslayers ofof variedvaried thicknessesthicknesses onon thethe Si(lll)Si(lll) surfacesurface alaltersters thisthis potentialpotential dependence.dependence. UnlikeUnlike metalmetal electrodes,electrodes, wherewhere SHGSHG isis highlyhighly sensitivesensitive toto potentialpotential variationvariation duedue toto thethe screeningscreening ofof thethe chargecharge atat thethe surface,15,16surface,15,16 thethe fieldfield forfor semiconductorssemiconductors cancan extendextend severalseveral hundredhundred angstromsangstroms intointo thethe material,material, thethe depthdepth beingbeing dependentdependent uponupon thethe dopingdoping dendensitysity ofof thethe semiconductor,semiconductor, possiblepossible Fermi-levelFermi-level pinning,pinning, andand thethe strengthstrength ofof thethe appliedapplied field.field.
TheThe secondsecond focusfocus ofof thisthis studystudy hashas beenbeen ofof aa moremore funda-funda
mentalmental nature.nature. ThatThat isis toto addressaddress issuesissues relatedrelated toto thethe sourcesource ofof thethe responseresponse fromfrom thesethese buriedburied interfaces,interfaces, thethe relativerelative concon
Si(111)tributionstributions fromfrom thethe depletiondepletion layer,layer, thethe Si(lll) surfacesurface adjaadjacentcent toto thethe electrolyteelectrolyte oror oxide,oxide, andand thethe Si0Si022 layer,layer, asas wellwell asas determiningdetermining thethe couplingcoupling ofof thethe appliedapplied staticstatic fieldfield toto variousvarious susceptibilitysusceptibility tensors.tensors. ForFor semiconductors,semiconductors, thethe bulkbulk responseresponse cancan bebe significantsignificant andand thusthus anyany studystudy whichwhich attemptsattempts toto ununderstandderstand thethe sourcesource ofof thethe responseresponse mustmust bebe concernedconcerned withwith thethe higherhigher orderorder terms.terms. InIn thethe presentpresent study,study, thisthis issueissue hashas beenbeen addressedaddressed byby potentialpotential dependentdependent measurements.measurements.
TheThe HH-Si-Si(111)(111) andand thethe oxidizedoxidized surfacesurface werewere examinedexamined inin differentdifferent electrolyteelectrolyte solutionssolutions andand inin UHY.UHY. MuchMuch ofof thethe workwork reportedreported involvesinvolves measuringmeasuring thethe variationvariation inin thethe SHSH responseresponse withwith azimuthalazimuthal rotationrotation ofof thethe samplesample byby 36036000 aboutabout itsits normal.normal. TheThe simplestsimplest systemsystem isis aa H-terminatedH-terminated samplesample examinedexamined underunder potentialpotential controlcontrol inin NH4F.NH4F. ThisThis surfacesurface shouldshould bebe monohydrogenmonohydrogen terminated,terminated, smooth,smooth, andand relativelyrelatively freefree ofof anyany photogeneratedphotogenerated surfacesurface oxidesoxides duedue toto thethe solubilsolubilityity ofof thethe oxidesoxides inin thisthis solution.solution.1717 TheThe resultsresults obtainedobtained atat thethe flatflatbandband potentialpotential areare comparedcompared withwith similarsimilar rotationalrotational ananisotropyisotropy measurementsmeasurements conductedconducted inin UHVUHV forfor aa sputteredsputtered andand annealedannealed sample,sample, andand aa hydrogenhydrogen terminatedterminated samplesample transferredtransferred toto UHVUHV withoutwithout furtherfurther surfacesurface cleaning.cleaning. TheseThese studiesstudies areare followedfollowed byby potentialpotential dependentdependent measurementsmeasurements ofof thethe SHSH rotationalrotational anisotropyanisotropy fromfrom initiallyinitially H-terminatedH-terminated sursur
NH4Ffaceface immersedimmersed inin aa HH22S0S044 solutionsolution where,where, unlikeunlike thethe NH4F solution,solution, thethe surfacesurface oxidesoxides whichwhich areare formedformed photoanodiphotoanodicallycally areare insoluble.insoluble. Similarly,Similarly, experimentsexperiments areare conductedconducted onon samplessamples immersedimmersed inin HH22S0S044 onon whichwhich controlledcontrolled amountsamounts ofof surlacesurlace oxidesoxides (0-40(0-40 A)A) areare photoanodicallyphotoanodically growngrown priorprior toto examinationexamination ofof thethe SHSH potentialpotential dependence.dependence.
II.II. DESCRIPTIONDESCRIPTION OFOF SHGSHG FROMFROM CUBICCUBIC MEDIAMEDIA
WhenWhen thethe electromagneticelectromagnetic fieldfield polarizespolarizes aa medium,medium, thisthis interactioninteraction isis governedgoverned byby Maxwell'sMaxwell's equationsequations andand thethe conconstitutivestitutive relations.relations. TheThe secondsecond orderorder polarizationpolarization thatthat givesgives riserise toto SHGSHG cancan bebe expressedexpressed byby aa seriesseries muItipoleofof multipole terms,terms, 18,1918,19
0734·2101/94112(5)12617/81$1.0026172617 J.J. Vac.Vac. Sci.Sci. Technol.Technol. AA 12(5),12(5), Sep/OctSep/Oct 19941994 0734-2101/94112(5)/2617/81$1.00 ©1994©1994 AmericanAmerican VacuumVacuum SocietySociety 26172617
26182618 FischerFischer etet 8/.:8/.: SensitivitySensitivity ofof SHGSHG toto spacespace chargecharge effectseffects 26182618
P~Y{2w)P~Y(2w) ==XD:E{ w)E{w) XP:E{w)V E{XD:E( w)E(w) ++XP:E(w)V E( w)w)
XQ-- VVXQ:E(XQ:E( ww )E()E(w)w) -- XQ:: VVE(E( ww )E()E(w)w)
CC MM++ -:--2-:--2 V[XV[X :E(w)E(w)],:E(w)E(w)], (1)(1) ll ww
wherewhere XX representsrepresents thethe susceptibilitysusceptibility tensortensor forfor eacheach termterm andand E(w)E(w) representsrepresents thethe incidentincident field.field. TheThe firstfirst twotwo termsterms areare electricelectric dipoledipole inin nature,nature, thethe thirdthird andand fourthfourth describedescribe thethe electricelectric quadrupolequadrupole contributioncontribution andand thethe lastlast termterm thethe magmagneticnetic dipoledipole contribution.contribution. UnderUnder thethe electricelectric dipoledipole approxiapproximation,mation, thethe secondsecond orderorder responseresponse vanishesvanishes inin thethe bulkbulk ofof centrocentrosymmetricsymmetric mediamedia andand isis allowedallowed onlyonly atat thethe interfaceinterface wherewhere thethe inversioninversion symmetrysymmetry isis broken.broken. WhenWhen describingdescribing thethe in-planein-plane response,response, onlyonly thethe firstfirst termterm shouldshould bebe considconsideredered sincesince thethe tangentialtangential componentscomponents ofof thethe incidentincident electricelectric fieldfield areare continuouscontinuous acrossacross thethe interface.interface. However,However, ifif eithereither thethe drivingdriving fieldsfields oror thethe SHSH polarizationpolarization underunder considerationconsideration containscontains aa fieldfield componentcomponent normalnormal toto thethe surface,surface, thenthen therethere willwill bebe contributionscontributions fromfrom thethe higherhigher orderorder termsterms toto thethe sursurfaceface response.response.
SHGSHG derivesderives itsits surfacesurface sensitivitysensitivity fromfrom notnot onlyonly thethe didipolepole allowedallowed surfacesurface terms,terms, butbut higherhigher orderorder (quadrupole)(quadrupole) termsterms whichwhich areare inseparableinseparable fromfrom thethe dipoledipole terms.terms. TheThe ininducedduced nonlinearnonlinear surfacesurface polarizationpolarization cancan bebe writtenwritten inin termsterms ofof anan effectiveeffective susceptibilitysusceptibility asas2020,21,21
P~Y(2w)P~Y(2w) ==x~Yx~y :E(w)E(w),:E(w)E(w), (2)(2)
wherewhere XmXm isis thethe secondsecond orderorder susceptibilitysusceptibility tensortensor reflectingreflecting thethe opticaloptical asas wellwell asas thethe symmetrysymmetry propertiesproperties ofof thethe surfacesurface layerlayer andand isis comprisedcomprised ofof contributionscontributions fromfrom thethe bulkbulk andand thethe surface.surface. IfIf thethe SHSH intensityintensity isis recordedrecorded asas aa functionfunction ofof aziazimuthalmuthal angleangle ofof rotation,rotation, thethe variationvariation inin intensityintensity reflectsreflects thethe overalloverall symmetrysymmetry ofof thethe surfacesurface andand allowsallows determinationdetermination ofof thethe tensorialtensorial propertiesproperties ofof Xeff'Xeff'
TheThe (111)(111) surface,surface, whichwhich isis thethe focusfocus ofof muchmuch ofof thethe workwork describeddescribed here,here, hashas CC3v3v symmetrysymmetry ifif thethe firstfirst bilayerbilayer ofof thethe SiSi surfacesurface isis considered.considered. WhenWhen thisthis surfacesurface isis rotatedrotated aboutabout itsits azimuth,azimuth, wherewhere 1>1> isis defineddefined asas thethe angleangle betweenbetween thethe [211][211] directiondirection onon thethe (111)(111) faceface andand thethe projectionprojection ofof thethe incidentincident wavewave vectorvector parallelparallel toto thethe surface,surface, thethe angularangular dependencedependence inin thethe SHSH responseresponse cancan bebe writtenwritten2222,23•
23
/1:;)(/1:;)( ¢)¢) == la(OO)la(OO) ++ c(3)c(3) cos(31))cos(31)) 1122 ,, (3)(3)
== Ih(3) sin{31> (4)(4)/~::)(/~::)( ¢)¢) Ib(3) sin(31)W,W, wherewhere thethe subscriptssubscripts referrefer toto thethe polarizationspolarizations ofof thethe fundafundamentalmental andand SHSH light,light, respectively.respectively. PP polarizationpolarization refersrefers toto thethe polarizationpolarization vectorvector ofof thethe lightlight residingresiding inin thethe planeplane ofof inciincidence,dence, whereaswhereas thethe polarizationpolarization vectorvector forfor ss polarizationpolarization isis normalnormal toto thethe planeplane ofof incidence.incidence. TheThe complexcomplex coefficientscoefficients aa (00),(00), h(3),b(3), andand c(3)c(3) containcontain thethe surfacesurface dipoledipole susceptibilitysusceptibility elementselements Xijk,Xijk, anyany contributioncontribution fromfrom higherhigher orderorder bulkbulk sussusceptibilityceptibility terms,terms, 'Yy andand ~,~, andand thethe appropriateappropriate FresnelFresnel facfac
alec)tors.tors. TheThe complexcomplex coefficientcoefficient a(cc) isis referredreferred toto asas thethe isotroisotropicpic coefficientcoefficient asas itit remainsremains invariantinvariant withwith rotationrotation andand
Xj2z ycontainscontains surfacesurface termsterms Xzxx'Xzxx' XxzxXxzx ,, Xrz '' andand bulkbulk termsterms 'Y andand ~.~. TheThe anisotropicanisotropic coefficients,coefficients, h(3b(3 andand c(3),c(3), includeinclude termsterms whichwhich varyvary withwith 1>1> andand containcontain bothboth XxxxXxxx andand ~.~.
IsotropicIsotropic ~~ ~~
f-----+-----I~~ ~-------+--------4
; L-____L-I____I~__~)5 !:-_-:-:~I-__=_I~---' oo 120120 240240 360360 o 120 240 360
CaseCase 11 CaseCase 22
iR0f\&\1iR\fY\&\1oo 120120 240240 360360
FIG.FIG. I.I. DemonstrationDemonstration ofof thethe interferenceinterference betweenbetween isotropicisotropic andand anisotropicanisotropic 1)contributionscontributions ofof thethe samesame magnitude.magnitude. (Case(Case I) IfIf thethe twotwo polarizationspolarizations areare inin
phase,phase, theirtheir interferenceinterference willwill resultresult inin threethree maximamaxima inin thethe rotationalrotational pattern.pattern. (Case(Case 2)2) IfIf thethe polarizationspolarizations areare outout ofof phasephase byby 180°,180°, thethe resultingresulting anisotanisotropyropy patternpattern willwill havehave sixsix maxima.maxima.
ForFor p-inputp-input andand p-outputp-output polarizationpolarization thethe observedobserved intenintensitysity modulationmodulation uponupon azimuthalazimuthal rotationrotation arisesarises fromfrom thethe ininterferenceterference betweeribetweeri thesethese anisotropicanisotropic andand isotropicisotropic terms.terms. TheThe SHSH patternspatterns cancan bebe understoodunderstood byby consideringconsidering howhow thesethese twotwo polarizationspolarizations interfere,interfere, asas illustratedillustrated inin Fig.Fig. 1.1. InIn thisthis simplesimple example,example, bothboth thethe isotropicisotropic andand anisotropicanisotropic polarizationspolarizations areare chosenchosen toto havehave thethe samesame magnitude.magnitude. IfIf thethe polarizationspolarizations areare in-phasein-phase (case(case 1),1), theythey willwill interfereinterfere andand resultresult inin threethree maximamaxima inin thethe rotationalrotational spectrum.spectrum. IfIf theythey areare out-of-phaseout-of-phase byby 180°,180°, sixsix maximamaxima areare observedobserved (case(case 2).2). TheseThese areare thethe twotwo extremesextremes sincesince inin mostmost casescases thethe contributionscontributions areare notnot comcom
yieldspletelypletely in-phasein-phase oror out-of-phase.out-of-phase. AA fitfit toto Eq.Eq. (3)(3) yidds thethe ratioratio c(3)/a(oo), whichwhich containscontains aa magnitudemagnitude andand relativerelativec(3)la(00),
phasephase angleangle reflectingreflecting thisthis interferenceinterference underunder thethe experimenexperimentaltal conditions.conditions. RotationalRotational anisotropyanisotropy measurementsmeasurements underunder specificspecific polarizationpolarization conditionsconditions areare veryvery sensitivesensitive toto relativerelative changeschanges inin bothboth magnitudemagnitude andand phasephase ofof eithereither thethe isotropicisotropic oror anisotropicanisotropic contributions.contributions.2424,25,25
TheThe applicationapplication ofof aa dcdc fieldfield toto thethe interfaceinterface asas inin thethe casecase ofof thesethese studiesstudies addsadds additionaladditional factorsfactors toto thethe polarizability,polarizability, wherewhere2626
P~Y(2w)=P~Y(2w)= x~y :E(w)E(w)+ x~y :E(w)E(w)E(dc).:E(w)E(w)E(dc). (5)(5)xiY :E{w)E{w)+ xiY
TheThe potentialpotential dependencedependence isis expressedexpressed inin thethe thirdthird orderorder Edcterm,term, wherewhere Edc isis thethe staticstatic electricelectric fieldfield orientedoriented normalnormal toto
thethe surface.surface. ThisThis fieldfield isis onon thethe orderorder ofof 1-51-5 VV andand dropsdrops acrossacross thethe spacespace chargecharge regionregion (SCR)(SCR) ofof thethe semiconductor.semiconductor. TheThe strengthstrength ofof thethe dcdc fieldfield cancan bebe onon thethe orderorder ofof 101066_10_1077
VV1m1m dependingdepending onon thethe depthdepth ofof thethe SCRSCR whichwhich isis governedgoverned largelylargely byby thethe dopingdoping densitydensity ofof thethe semiconductor.semiconductor. ThisThis cancan bebe viewedviewed asas aa mixingmixing ofof aa staticstatic fieldfield thatthat inducesinduces aa polarpolar
z directionizationization strictlystrictly orientedoriented inin thethe zdirection (normal(normal toto thethe elecelectrodetrode surface)surface) andand bothboth thethe surfacesurface andand bulkbulk polarizationpolarization ininducedduced byby lightlight waves.waves. ThisThis additionaladditional polarizationpolarization cancan bebe
following:27writtenwritten asas thethe following:27
J.J. Vac.Vac. Sci.Sci. Techno!.Techno!. A,A, Vol.Vol. 12,12, No.5,No.5, Sep/OctSep/Oct 19941994
26192619 FischerFischer etet al.:al.: SensitivitySensitivity ofof SHGSHG toto spacespace chargecharge effectseffects 26192619
P)2W)(E1 dc)=X:S/Ej(w)Ek(W)EldC' (6)(6)
SinceSince bothboth X(2)X(2) andand X'(3)X'(3) havehave thethe samesame symmetrysymmetry constraintsconstraints imposedimposed byby thethe electrodeelectrode sudace,sudace, thethe overalloverall symmetrysymmetry ofof thethe responseresponse remainsremains unchangedunchanged andand thethe netnet effecteffect isis thatthat mostmost tensortensor elementselements (including(including thethe bulkbulk termsterms becausebecause ofof thethe extensionextension ofof thethe staticstatic fieldfield intointo thethe bulkbulk semiconductor)semiconductor) willwill bebe affectedaffected2828,29,29 byby thethe intedacialintedacial charging.charging.
ForFor thethe generalgeneral discussiondiscussion ofof thethe observedobserved potentialpotential dedependence,pendence, Eq.Eq. (5)(5) cancan bebe writtenwritten inin aa simplersimpler formform asas
<xIIF(xi1) Xi~A<I»112,II sHsH<XIIF(xi1) ++X~2)+X~2)+ xi~A<I»1I2, (7)(7)
wherewhere FFcorrespondscorresponds toto thethe linearlinear FresnelFresnel factors,factors, X~2)X~2) andand X~~A<I>il)il) areare thethe sudacesudace andand bulkbulk susceptibilities,susceptibilities, Xi~A<I> isis thethe
effectiveeffective cubiccubic nonlinearitynonlinearity arisingarising fromfrom thethe staticstatic fieldfield whichwhich includesincludes allall termsterms thatthat varyvary linearlylinearly withwith fieldfield strength.strength.3030 A<I>A<I> isis thethe potentialpotential dropdrop acrossacross thethe semiconductorsemiconductor spacespace chargecharge regionregion andand isis proportionalproportional toto thethe differencedifference bebetweentween thethe appliedapplied (Eapp)(Eapp) andand thethe flatbandflatband (E(E tbtb)) potential.potential. WhenWhen thethe SHSH responseresponse isis dominateddominated byby thethe cubiccubic nonlinearnonlinearityity term,term, thethe observedobserved potentialpotential dependencedependence shouldshould bebe paraparabolicbolic withwith aa minimumminimum nearnear thethe flatbandflatband potential.potential. SuchSuch bebehaviorhavior hashas beenbeen observedobserved inin numerousnumerous studiesstudies ofof metaUmetaU electrolyteelectrolyte systems?systems? I1 UnderUnder conditions,conditions, wherewhere thethe sudacesudace and/orand/or bulkbulk quadraticquadratic nonlinearitiesnonlinearities dominate,dominate, parabolicparabolic popotentialtential dependencedependence withwith aa minimumminimum shiftedshifted fromfrom flatbandflatband
observed.wouldwould bebe observed,
III.III. EXPERIMENTALEXPERIMENTAL CONSIDERATIONSCONSIDERATIONS
TheThe opticaloptical measurementsmeasurements employemploy thethe fundamentalfundamental outputoutput fromfrom aa 1010 HzHz Nd:YAGNd:YAG laserlaser whichwhich illuminatesilluminates thethe sudacesudace atat aa 32°32° incidentincident angle.angle. 1010 nsns pulsespulses ofof =0.3=0.3 J/cmJ/cm22 werewere used,used, whichwhich isis belowbelow thethe damagedamage thresholdthreshold forfor silicon.silicon. TheThe elecelectrochemicaltrochemical cellcell hashas describeddescribed previously,24previously,24 andand consistsconsists ofof aa threethree electrodeelectrode systemsystem withwith thethe n-Si(lll)n-Si(lll) samplesample asas thethe workingworking electrode,electrode, aa platinumplatinum wirewire countercounter electrode,electrode, andand thethe saturatedsaturated calomelcalomel referencereference electrodeelectrode (SCE)(SCE) whichwhich isis isoisolatedlated fromfrom thethe workingworking electrodeelectrode byby aa ceramicceramic junction.junction. FlatFlatbandband potentialspotentials areare determineddetermined byby photocurrentphotocurrent transients.transients. AllAll potentialpotential scansscans areare restrictedrestricted toto thethe regionregion wherewhere stablestable photocurrentphotocurrent isis observed.observed.
ForFor thethe UHVUHV studies,studies, thethe samplesample waswas mountedmounted onon aa mamanipulatornipulator capablecapable ofof 360° azimuthalazimuthal rotationrotation inin3600 aa vacuumvacuum chamberchamber withwith aa basebase pressurepressure ofof 3.5X3.5X 10-10-99 Torr.Torr. TheThe chamchamberber isis equippedequipped withwith anan AugerAuger electronelectron spectrometerspectrometer (AES)(AES) andand lowlow energyenergy electronelectron diffractiondiffraction (LEED)(LEED) toto monitormonitor sursurfaceface cleanlinesscleanliness andand crystallinity.crystallinity. ForFor aa portionportion ofof thethe studies,studies, thethe SiSi sudacesudace waswas cleanedcleaned byby successivesuccessive cyclescycles ofof sputteringsputtering (Ne+,(Ne+, II kV,kV, 1515 pA)pA) andand annealingannealing (1000(1000 K).K).
TheThe Si(lll)Si(lll) waferswafers usedused inin thethe experimentsexperiments areare n-dopedn-doped (phosphorus)(phosphorus) withwith aa resistivityresistivity ofof 3.0-6.53.0-6.5 00 cm,cm, whichwhich reprepresentsresents aa dopingdoping densitydensity ofof =10=101515 cm-cm-33.. TheThe 0.66-0.710.66-0.71 mmmm thickthick samplessamples areare degreaseddegreased byby ultrasonificationultrasonification inin separateseparate bathsbaths ofof methylenemethylene chloride,chloride, acetone,acetone, andand methanolmethanol andand thenthen drieddried withwith nitrogen.nitrogen. TheThe backback ofof thethe waferwafer isis etchedetched forfor 11 minmin inin 48%48% HFHF toto removeremove thethe nativenative oxideoxide andand thenthen mountedmounted onon Ga-InGa-In eutecticeutectic thatthat hashas beenbeen placedplaced onon anan embeddedembedded coppercopper contactcontact inin aa Kel-FKel-F shaft.shaft. AA maskmask containingcontaining anan emembeddedbedded acidacid resistantresistant fluorocarbonfluorocarbon o-ringo-ring isis usedused toto sealseal thethe
backback sudacesudace fromfrom thethe electrolyte.electrolyte. TheThe polishedpolished sideside ofof thethe waferwafer isis etchedetched asas describeddescribed belowbelow andand loadedloaded intointo thethe elecelectrochemicaltrochemical cell.cell.
TheThe sudacesudace ofof thethe siliconsilicon waferwafer isis preparedprepared byby etchingetching inin bufferedbuffered NHNH4FIHF4FIHF solutionsolution (pH=8)(pH=8) whichwhich isis knownknown toto leaveleave thethe sudacesudace atomicallyatomically smoothsmooth andand inin aa mono-Hmono-Hterminatedterminated state.state. 1,32,33 InIn contrast,contrast, ifif thethe sudacesudace isis etchedetched withwith HF,HF, thethe resultingresulting sudacesudace isis mono-,mono-, di-,di-, andand tri-Htri-Hterminatedterminated andand consequentlyconsequently isis microscopicallymicroscopically rough.?rough.? IRIR studiesstudies havehave shownshown thatthat sudacesudace roughnessroughness graduallygradually dimindiminishesishes asas thethe pHpH ofof thethe bufferedbuffered solutionsolution isis increased,increased, leadingleading toto anan atomicallyatomically smooth,smooth, idealideal mono-H-terminatedmono-H-terminated Si(llSi(ll 1)1) sudacesudace atat aa pHpH ofof 8-9.8-9.11 Furthermore,Furthermore, thethe mono-Hmono-H
I•32,33
Si(1ll)terminatedterminated Si(Ill) sudacessudaces areare foundfound toto bebe resistantresistant toto oxioxidationdation inin electrolyteelectrolyte solutionsolution (until(until thethe sudacesudace isis subjectedsubjected toto anodicanodic potentialspotentials positivepositive ofof flatband).5,9,34flatband).5,9,34 AfterAfter thethe samplesample hashas beenbeen etched,etched, itit isis immersedimmersed inin 0.10.1 MM NH4FNH4F ifif thethe H-terminatedH-terminated surfacesurface isis toto bebe maintained,maintained, oror inin 0.10.1 MM HH22S0S044 ifif anan oxideoxide isis toto bebe grown.grown. OxidesOxides areare growngrown onon thethe SiSi samplessamples photoanodicallyphotoanodically byby thethe stepwisestepwise increaseincrease inin potenpotentialtial fromfrom thethe flatbandflatband toto aa finalfinal potentialpotential ofof +5.0+5.0 V,V, whilewhile
2 ,keepingkeeping thethe currentcurrent belowbelow 3030 j.LAlcmj.LAlcm2, andand illuminatingilluminating thethe
samplesample withwith aa diffusediffuse HeNeHeNe beam.beam. ThicknessesThicknesses ofof thethe oxidesoxides areare determineddetermined byby aa combinationcombination ofof ellipsometry,ellipsometry, photocurphotocurrentrent measurements,measurements, andand etchbacketchback times.times. FlatbandFlatband potentialspotentials havehave beenbeen determineddetermined byby photocurrentphotocurrent transienttransient measuremeasurements.ments.
IV.IV. RESULTSRESULTS FigureFigure 22 showsshows thethe rotationalrotational anisotropyanisotropy inin thethe SHSH rere
Si(Ill)sponsesponse withwith p-inputp-input andand p-outputp-output polarizationpolarization forfor aa Si(lll) NH4Fsudacesudace studiedstudied inin NH4F atat fourfour representativerepresentative potentials.potentials. AllAll
NH4Fsamplessamples werewere etchedetched forfor 33 minmin inin 2.02.0 MM NH4F atat pHpH 8.08.0 toto produceproduce aa H-terminatedH-terminated surface.surface. InIn situsitu electrochemicalelectrochemical etchetchinging ofof anodicallyanodically growngrown oxidesoxides showsshows thethe sudacesudace remainsremains H-terminatedH-terminated inin thethe darkdark atat potentialspotentials wellwell positivepositive ofof flatflatbandband inin 0.20.2 MM NHNH44F.F.3535 TheThe sudacesudace shouldshould bebe relativelyrelatively freefree ofof anyany photogeneratedphotogenerated sudacesudace oxidesoxides sincesince inin fluoridefluoride concontainingtaining electrolytes,electrolytes, thethe dissolutiondissolution ofof thisthis oxideoxide competescompetes withwith thethe photo-oxidationphoto-oxidation process.process. AtAt allall potentialspotentials examinedexamined
(111) sudacethethe threefoldthreefold symmetrysymmetry expectedexpected ofof thethe (Ill) surface ofof aa cubiccubic latticelattice isis observed.observed. FigureFigure 2(a)2(a) waswas takentaken atat -0.65-0.65 VV
c(3)/a(OO) o.o.andand aa fitfit toto Eq.Eq. (3)(3) yieldsyields aa valuevalue forfor c(3)la(OO) ofof 2.2e2.2e i21i21
FromFrom currentcurrent transienttransient measurements,measurements, thisthis isis determineddetermined toto bebe thethe flatbandflatband potential.potential. AsAs thethe potentialpotential isis drivendriven positivepositive intointo thethe depletiondepletion region,region, thethe rotationalrotational anisotropiesanisotropies showshow aa strongstrong andand progressiveprogressive variation.variation. TheThe magnitudemagnitude ofof thethe ratioratio ofof c(3)/a(OO)c(3)/a(OO) changeschanges byby approximatelyapproximately aa factorfactor ofof 22 overover thisthis 1.351.35 VV region.region. MoreMore strikingstriking isis thethe changechange inin thethe phasephase angleangle ofof thethe ratioratio fromfrom 22°22° toto 126°126° overover thisthis potentialpotential range.range. TheThe uncertaintyuncertainty inin thethe phasephase angleangle measurementmeasurement isis ::!:5°.::!:5°. AtAt popotentialstentials moremore negativenegative ofof thethe flatband,flatband, thethe magnitudemagnitude ofof thethe signalsignal intensityintensity continuescontinues toto increaseincrease althoughalthough thethe ratioratio ofof
sudace,c(3)/a(OO)c(3)/a(OO) decreasesdecreases onlyonly slightly.slightly. ForFor thethe H-Si(lll)H-Si(lll) surface, decreasingdecreasing thethe potentialpotential muchmuch beyondbeyond flatbandflatband ((-0.65-0.65 VV verversussus SCE)SCE) resultsresults inin significantsignificant hydrogenhydrogen evolutionevolution whichwhich inintederestederes withwith thethe opticaloptical processprocess andand preventsprevents examinationexamination ofof thethe samplesample underunder accumulationaccumulation conditions.conditions.
ToTo obtainobtain aa betterbetter understandingunderstanding ofof thethe potentialpotential dependepenaloe)dencedence ofof c(3)c(3) andand a(oo) independently,independently, thethe experimentsexperiments werewere
JVSTJVST AA •• Vacuum,Vacuum, Surfaces,Surfaces, andand FilmsFilms
2620 2620 26202620 FischerFischer etet al.:al.: SensitivitySensitivity ofof SHGSHG toto spacespace chargecharge effectseffects
22
~~
(fJ(fJ -+-'-+-' 00 Cc :::l:::l
22 LlLl Ll00
'-/'-/
>,>, -+-'-+-'
(fJ(fJ
CC Q)Q)
-+-'-+-' C 00c
Uu 22cc
00
EE LI
00 II
-0-0 CC 0a U 00u Q)Q)
b)
c)
b)
C)
(f)(f)
22 d)d)
RotationRotation AngleAngle
N14FFIG.FIG. 2.2. SHSH rotationalrotational anisotropyanisotropy fromfrom n-Si(lll)n-Si(lll) immersedimmersed inin 0.10.1 MM N~F atat ftatbandfiatband andand depletiondepletion conditions.conditions. AllAll scansscans werewere takentaken withwith p-inputp-input andand p-outputp-output polarizations.polarizations. TheThe solidsolid lineline isis thethe fitfit toto Eq.Eq. (3),(3), (a)(a) samplesample biasedbiased
ftatband c(3)/a(0»=2.2eatat thethe fiatband potential,potential, -0.65-0.65 VV (SeE),(SeE), c(3)/a(OO)=2.2e i21i21 ';'; (b)(b) 0.00.0 V,V, c(3)/a(0»=3.2ec(3)/a(OO)=3.2e i63i63 ';'; (c)(c) +0.2+0.2 V,V, c(3)/a(0»=3.4ec(3)/a(OO)=3.4e i86i86';'; (d)(d) +1.0+1.0 V,V, c(3)/a(OO)c(3)/a(OO) == 3.8e3.8e il26il26'.'.
conductedconducted atat l/J=l/J= 30°30° whichwhich allowsallows isolationisolation ofof thethe isotropicisotropic andand anisotropicanisotropic contributions.contributions. AsAs shownshown inin Fig.Fig. 3,3, bothboth thethe isotropicisotropic andand anisotropicanisotropic responseresponse areare foundfound toto havehave aa paraparabolicbolic potentialpotential dependencedependence withwith aa higherhigher overalloverall signalsignal levellevel fromfrom thethe latter.latter. TheThe isotropicisotropic responseresponse whichwhich wouldwould mostmost readilyreadily couplecouple toto thethe appliedapplied staticstatic fieldfield hashas aa minimumminimum nearnear
fJ.atband+0.26+0.26 V,V, nearlynearly 900900 mmVV fromfrom thethe flatband potential.potential. TheThe fieldfield cancan couplecouple toto thethe depthdepth ofof thethe spacespace chargecharge region,region, whichwhich forfor thisthis surfacesurface biasedbiased atat +0.3+0.3 VV isis onon thethe orderorder ofof 12001200 nm.nm. TheThe anisotropicanisotropic responseresponse hashas aa minimumminimum nearnear
fJ.atband.++ 1.61.6 V,V, eveneven furtherfurther awayaway fromfrom flatband. RotationalRotational anisotropyanisotropy experimentsexperiments werewere alsoalso conductedconducted inin
UHVDHV forfor comparison.comparison. AA samplesample waswas preparedprepared inin aa H-terminatedH-terminated statestate priorprior toto mountingmounting inin thethe chamberchamber whichwhich waswas thenthen evacuatedevacuated toto thethe basebase pressure.pressure. NoNo additionaladditional sursurfaceface cleaningcleaning waswas performed.performed. FigureFigure 4(a)4(a) showsshows thethe resultsresults withwith aa fitfit toto thethe datadata givinggiving aa phasephase angleangle ofof 61°.61°. TheThe resultsresults areare similarsimilar toto thatthat obtainedobtained inin solution.solution. TheThe bestbest corresponcorrespondencedence isis forfor thethe hydrogenhydrogen terminatedterminated samplesample inin solutionsolution
J.J. Vac.Vac. Sci.Sci. Technol.Technol. A,A, Vol.Vol. 12,12, No.5,No.5, Sep/OctSep/Oct 19941994
0.20.2 a)a) IsotropicIsotropic
/'0./'-..
~~ Ce ::J::l
.ci.ci \..\...
~~
.£.£ "'"' cc 22 0.00.0ce
.~.~ ce b)b) AnisotropicAnisotropic §§ 00 II
"0'0 cc 00 ()() (l)(l)
(/)(/)
00 rrss 0.00.0 0.50.5 1.01.0
PotentialPotential (V(V vs.vs. SeE)SeE)
FBFB
a("")FIG.FIG. 3.3. PotentialPotential dependencedependence inin thethe isotropicisotropic a(oo) andand anisotropicanisotropic responseresponse NH4F.10(3)10(3) forfor Si(llI)Si(llI) preparedprepared inin aa H-terminatedH-terminated statestate andand biasedbiased inin 0.10.1 MM NH4F.
TheThe isotropicisotropic responseresponse waswas monitoredmonitored withwith p-polarizedp-polarized incidentincident lightlight andand p-polarizedp-polarized SHGSHG withwith 4J=30°.4J=30°. TheThe anisotropicanisotropic responseresponse waswas monitoredmonitored withwith p-polarizedp-polarized incidentincident lightlight andand s-polarizeds-polarized outputoutput lightlight atat 4J=30°.4J=30°.
fJ.atwhichwhich isis biasedbiased aa fewfew hundredhundred millivoltsmillivolts positivepositive ofof thethe flatband.band. AfterAfter sputteringsputtering thethe samplesample withwith NeNe++ andand annealingannealing atat 900900 K,K, anan oxide-freeoxide-free surfacesurface withwith aa (7X7)(7X7) reconstructionreconstruction waswas obtained.obtained. TheThe SHSH responseresponse fromfrom thisthis samplesample isis shownshown inin Fig.Fig. 4(b).4(b). TheThe reconstructedreconstructed surfacesurface generatesgenerates aa muchmuch higherhigher signalsignal atat 10641064 nmnm thanthan thethe nonreconstructednonreconstructed surfacesurface withwith thisthis enhancementenhancement attributedattributed toto thethe increaseincrease ofof longlong rangerange orderorder
layer.4inging ofof thethe surfacesurface layer.4
FigureFigure 55 showsshows thethe resultsresults forfor thethe Si(lll)Si(lll) surfacesurface examexamfJ.atband 4.inedined atat flatband inin HH22S0S04 , TheThe surfacesurface waswas initiallyinitially preparedprepared
inin aa H-terminatedH-terminated state.state. ForFor thisthis systemsystem thethe flatbandflatband potentialpotential waswas measuredmeasured toto bebe -0.6-0.6 V.V. AsAs mentionedmentioned previously,previously, thethe oxidesoxides areare insolubleinsoluble inin thethe acidacid andand cancan formform onon thethe surfacesurface particularlyparticularly inin thethe presencepresence ofof thethe probeprobe laserlaser light.light. TheThe reresponsesponse fromfrom thethe surfacesurface studiedstudied inin HH22S0S044 atat thethe flatbandflatband
i44Dpotentialpotential givesgives aa fitfit toto c(3)/a(OO)c(3)/a(OO) ofof 1.7e1.7e i44 ° andand isis similarsimilar toto NH4Fthatthat obtainedobtained inin NH4F [Fig.[Fig. 2(a)].2(a)]. FromFrom oneone experimentexperiment toto
thethe next,next, aa variationvariation inin thethe phasephase (ranging(ranging fromfrom 30°30° toto 60°)60°) isis observed.observed. TheThe ftatbandftatband potentialpotential isis alsoalso foundfound toto varyvary slightlyslightly inin time,time, shiftingshifting anodicallyanodically withwith thethe formationformation ofof oxide.oxide. BothBoth variationsvariations areare attributedattributed toto photogeneratedphotogenerated surfacesurface speciesspecies whichwhich cancan storestore chargecharge atat thethe interfaceinterface inin anan unpredictableunpredictable manner.manner. ThisThis changeschanges thethe degreedegree ofof bandband bendingbending forfor aa givengiven appliedapplied potential,potential, thusthus leadingleading toto differentdifferent resultsresults fromfrom oneone experimentexperiment toto thethe next.next. TheThe photocurrentphotocurrent measurementsmeasurements givegive anan estimateestimate ofof thethe amountamount ofof photogeneratedphotogenerated speciesspecies toto bebe lessless thanthan 33 ML.ML.
InIn aa similarsimilar mannermanner thethe SHSH responseresponse hashas beenbeen studiedstudied Si(l1l)fromfrom Si(lll) surfacessurfaces onon whichwhich varyingvarying thicknessesthicknesses ofof Si0Si022
havehave beenbeen photoanodicallyphotoanodically grown.grown. FigureFigure 66 displaysdisplays thethe ro-ro
2621 2621 26212621 FischerFischer etet 81.:81.: SensitivitySensitivity ofof SHGSHG toto spacespace chargecharge effectseffects
4r--.--.,.-...--r-..---,.--r---,r--.--..---,.-.., a)a)
b)b) 2525
AzimuthalAzimuthal AngleAngle
FIG.FIG. 4.4. SHSH rotationalrotational anisotropyanisotropy fromfrom Si(Si(lll)Ill) initiallyinitially preparedprepared inin aa H-terminatedH-terminated statestate andand thenthen transferredtransferred toto UHV.UHV. (a)(a) ThisThis samplesample waswas loadedloaded intointo UHVUHV andand evacuatedevacuated toto basebase pressure.pressure. NoNo additionaladditional cleaningcleaning waswas perperformed.formed. TheThe resultresult isis SHSH rotationalrotational anisotropyanisotropy fromfrom aa H-terminatedH-terminated surfacesurface
e(3)/a(oo)=2.8ei61 °. (b)(b) ThisThis surfacesurface sputteredsputteredinin UHVUHV withwith aa fitfit toto c(3)/a(oo)=2.8ei6I waswaso.
theSi(lll)(Ne+)(Ne+) andand annealedannealed (900(900 K)K) toto removeremove anyany oxide.oxide. TheThe resultresult isis theSi(lll) 7X77X7 reconstructedreconstructed surfacesurface withwith aa fitfit toto c(3)/a(OO)e(3)/a(OO) == 1.8e1.8e i28i28".".
tationaltational anisotropyanisotropy fromfrom anan oxideoxide coveredcovered surfacesurface inin HH22S0S044 atat fourfour differentdifferent potentials,potentials, 0.00.0 VV [Fig.[Fig. 6(a)]6(a)] correspondingcorresponding toto thethe flatband potential,potential, andand threethreeHatband moremore positivepositive potentials,potentials, +1.5,+1.5, +2.5,+2.5, andand +3.5+3.5 VV inin Figs.Figs. 6(b)6(b) toto 6(d),6(d), respectively.respectively. ForFor allall oxidizedoxidized samples,samples, thethe flatbandflatband potentialpotential shiftsshifts anodi-anodi
2.52.5
2.02.0
,.......,-... 1.51.5 -<-'-<-'
cc :J:J 1.01.0
..ci..ci ~~ 0.50.5 00
'-./'-./
>..>.. 0.00.0 -<-'-<-' 2.52.5 (j)(j)
cc Q)(])
+.J 2.02.0 b)b)-<-' Cc
II 1.51.5 (/)(/)
1.01.0
0.50.5
6060 120120 180180 240240 300300 360360
RotationRotation angleangle FIG.FiG. 5.5. RotationalRotational anisotropyanisotropy fromfrom n-Si(lll)n-Si(lll) forfor p-inputp-input andand p-outputp-output popo
NH4Flarizations.larizations. (a)(a) Si(lll)Si(lll) inin 0.10.1 MM NH4F biasedbiased atat thethe flatbandflatband potentialpotential ofof -0.65-0.65 VV (SeE).(SeE). FittingFitting toto Eq.Eq. (3)(3) yieldsyields c(3)laoo=2.2ei21°,c(3)laoo=2.2ei21°, (b)(b) Si(lll)Si(lll) inin 0.10.1 MM HH2ZS0S044 biasedbiased atat thethe flatbandflatband potentialpotential ofof -0.60-0.60 V;V; c(3)la(OO) =e(3)la(OO)= t.7et.7e i44i44°.°.
JVSTJVST AA •• Vacuum,Vacuum, Surfaces,Surfaces, andand FilmsFilms
44 b)b)
>..>.. +-'+-' If)If)
CC (j)(j)
-+-'-+-' Cc
uu cc 44 c)c) oo EE I....I....
oo II
'0'D CC oo ()U (j)(j)
(f)(f)
44 d)d)
RotationRotation AngleAngle
FIG.FiG. 6.6. PotentialPotential dependentdependent rotationalrotational anisotropyanisotropy inin thethe p-inputp-input andand p-outputp-output SHSH responseresponse fromfrom aa 4040 AA photoanodicallyphotoanodically growngrown oxideoxide (SiO(Si02z)) inin 0.10.1 MM HH22S0S044 ,, (a)(a) atat fiatbandflatband (0.0(0.0 VV vsvs SeE);SeE); fitfit toto Eq.Eq. (3)(3) yieldsyields c(3)/a(X)= 2.Se I.S c(3)la(OO)=e(3)/a(X)= 2.5e i41i41 0,0, (b)(b) ++ 1.5 VV fromfrom flatband,flatband, e(3)/a(OO)= 33 AeAe i58i58°,°, c)c) +2.5+2.5 VV fromfrom flatband,flatband, c(3)/a(OO)""4.5ee(3)la(OO)=4.5e i92i92°,°, (d)(d) +3.5+3.5 VV fromfrom flatband,flatband,
""4.8ee(3)/a(OO)e(3)/a(OO) =4.8e il20il20°.°.
callycally asas determineddetermined byby thethe currentcurrent transienttransient measurementsmeasurements.. TheThe potentialpotential "window""window" forfor studyingstudying thethe oxide-coatedoxide-coated sursurfacesfaces widenswidens duedue toto thethe insulatinginsulating naturenature ofof thethe overlayer.overlayer. TheThe oxideoxide layerlayer isis estimatedestimated toto bebe ~40~40 AAbasedbased onon ellipsomellipsometryetry measurementsmeasurements andand etchbacketchback times.times. BecauseBecause thethe electroelectrochemicallychemically growngrown oxidesoxides havehave substantialsubstantial waterwater contentcontent inin thethe firstfirst monolayersmonolayers ofof oxide,oxide, withwith thisthis ratioratio decreasingdecreasing asas thethe thicknessthickness increases,17increases,17 thethe ellipsometricallyellipsometrically determineddetermined thicknessesthicknesses areare viewedviewed asas anan upperupper limit.limit.
Si(1l1)InIn comparingcomparing thethe responseresponse fromfrom thethe Si(lll) surfacessurfaces ofof Fig.Fig. 2(a)2(a) measuredmeasured atat flatbandflatband withwith thethe oxidizedoxidized samplesample inin Fig.Fig. 6(a),6(a), twotwo importantimportant observationsobservations cancan bebe made.made. TheThe rorotationaltational anisotropyanisotropy persistspersists inin thethe presencepresence ofof thethe oxideoxide andand thethe signalsignal fromfrom thethe oxideoxide coveredcovered surfacesurface isis enhancedenhanced relarelativetive toto thethe moremore oxideoxide freefree samplesample inin Fig.Fig. 1I (b).(b). TheThe formerformer observationobservation suggestssuggests thatthat inin thethe presencepresence ofof thethe oxide,oxide, thethe
Si(I 11)signalsignal fromfrom Si(lll) continuescontinues toto dominatedominate thethe responseresponse overover thethe moremore isotropicisotropic andand disordereddisordered oxideoxide overlayer.overlayer. TheThe signalsignal
2622 2622 26222622 FischerFischer etet al.:al.: SensitivitySensitivity ofof SHGSHG toto spacespace chargecharge effectseffects
enhancementenhancement observedobserved inin thethe presencepresence ofof oxideoxide indicatesindicates thatthat Si(1ll)thethe SHGSHG isis sensitivesensitive toto thethe Si(lll) surfacesurface adjacentadjacent toto thethe
oxideoxide andand thatthat itit isis notnot merelymerely aa bulkbulk response.response. However,However, thethe possibilitypossibility doesdoes existexist thatthat thethe signalsignal enhancementenhancement isis duedue toto aa changechange inin thethe FresnelFresnel factorsfactors whichwhich occursoccurs asas thethe oxideoxide growsgrows onon thethe surface,surface, thusthus alteringaltering thethe angleangle ofof incidenceincidence ofof thethe incomingincoming opticaloptical fieldfield andand thereforetherefore increasingincreasing thethe SHSH response.response.
AsAs withwith thethe Si(Si(111)111) surfaces,surfaces, therethere isis aa strongstrong potentialpotential dependencedependence inin thethe responseresponse fromfrom thethe Si(111)/Si0Si(111)/Si02 interfaceinterface asas manifestedmanifested inin bothboth thethe relativerelative magnitudemagnitude andand phasephase ofof thethe rotationalrotational anisotropy.anisotropy. ForFor anan appliedapplied biasbias ofof +3.5+3.5 VV beyondbeyond thethe flatbandflatband potentialpotential (0.0(0.0 VV forfor thisthis sample),sample), thethe relativerelative phasephase angleangle betweenbetween thethe anisotropicanisotropic andand isotropicisotropic responseresponse
80°. ForFor thethe H-terminatedH-terminated surface,surface, thethe changechange inin
2
changeschanges byby 80 0 •
phasephase angleangle occursoccurs overover aa muchmuch smallersmaller potentialpotential rangerange ofof 80°1.351.35 V.V. ForFor 1515 andand 2525 AA samples,samples, aa similarsimilar phasephase shiftshift ofof 800
occursoccurs overover aa potentialpotential rangerange ofof 2.02.0 andand 2.52.5 V,V, respectively.respectively. AtAt flatband,flatband, thethe fitsfits ofof thethe anisotropiesanisotropies forfor thethe differentdifferent oxideoxide coveredcovered samplessamples areare nearlynearly thethe same.same. TheThe observedobserved trendtrend isis thatthat thethe thickerthicker thethe oxideoxide overlayer,overlayer, thethe smallersmaller isis thethe obobservedserved changechange inin phasephase perper unitunit voltagevoltage droppeddropped acrossacross thethe interface.interface.
TheThe potentialpotential dependencedependence ofof thethe anisotropicanisotropic andand isotropicisotropic responseresponse waswas independentlyindependently examinedexamined forfor thesethese oxidizedoxidized samples.samples. BothBoth exhibitexhibit aa potentialpotential dependencedependence similarsimilar toto Fig.Fig. 3,3, withwith minimaminima whichwhich areare wellwell removedremoved fromfrom thethe flatbandflatband potential.potential. ForFor thethe 4040 AA thickthick oxideoxide sample,sample, thethe measuredmeasured minimaminima inin thethe anisotropicanisotropic andand isotropicisotropic responseresponse occursoccurs nearnear +3.3+3.3 andand +2.2+2.2 V,V, respectively;respectively; bothboth minimaminima occuroccur farfar fromfrom thethe flatbandflatband potentialpotential atat 0.00.0 V.V.
V.V. DISCUSSIONDISCUSSION
TheThe resultsresults ofof thethe electrochemicalelectrochemical studiesstudies describeddescribed aboveabove Si(lll)demonstratedemonstrate thethe sensitivitysensitivity ofof SHGSHG toto thethe Si(ll1) surfacesurface
whetherwhether itit isis adjacentadjacent toto anan electrolyteelectrolyte solutionsolution oror variousvarious thicknessesthicknesses ofof oxides.oxides. FitsFits toto thethe rotationalrotational anisotropyanisotropy datadata obtainedobtained asas aa functionfunction ofof potentialpotential showshow thatthat SHGSHG isis sensisensitivetive toto thethe appliedapplied fieldfield withinwithin thethe depletiondepletion layerlayer andand thatthat bothboth thethe amplitudeamplitude andand thethe relativerelative phasephase ofof thethe responseresponse varyvary withwith potential.potential. ForFor allall electrochemicalelectrochemical studies,studies, bothboth thethe anisotropicanisotropic andand isotropicisotropic responseresponse displaydisplay aa parabolicparabolic potenpotentialtial dependence.dependence. TheThe factfact thatthat thethe minimumminimum ofof thethe isotropicisotropic
fiatbandoror anisotropicanisotropic signalsignal doesdoes notnot occuroccur atat thethe flatband forfor eithereither
2 NH4Fthethe HH2S0S044 oror NH4F solutionssolutions suggestssuggests thatthat termsterms otherother thanthan x~~.:1<l>x~~.:1<I> inin Eq.Eq. (7)(7) mustmust bebe contributing.contributing. xl)xl) isis thethe mostmost likelylikely factorfactor consideringconsidering thatthat thethe penetrationpenetration depthdepth ofof thethe pumppump lightlight isis onon thethe orderorder ofof 11 cm,cm, considerablyconsiderably deeperdeeper thanthan thethe SCR.SCR. PreviousPrevious studiesstudies inin airair havehave determineddetermined thatthat thethe SHSH contributioncontribution fromfrom thethe surfacesurface andand thethe bulkbulk ofof Si(lll)Si(lll) areare ofof similarsimilar magnitude.magnitude.2323.,36
36 FurtherFurther evidenceevidence forfor thethe importanceimportance ofof thethe bulkbulk responseresponse comescomes fromfrom thethe observationobservation thatthat whenwhen Si(1ll)Si(lll) surfacessurfaces areare roughenedroughened byby etchingetching inin 48%48% HFHF priorprior toto introductionintroduction intointo thethe NHNH44F,F, thethe rotationalrotational anisotropiesanisotropies forfor thethe roughenedroughened andand unroughenedunroughened samplessamples areare quitequite similar.similar. ThereThere isis nono evidenceevidence ofof anan isotropicisotropic responseresponse fromfrom aa disordisordereddered siliconsilicon surfacesurface superimposedsuperimposed onon thethe responseresponse fromfrom thethe crystallinecrystalline lattice.lattice. TheThe intensityintensity ofof thethe overalloverall responseresponse isis howeverhowever slightlyslightly largerlarger inin thethe roughenedroughened case,case, suggestingsuggesting
J.J. Vac.Vac. Sci.Sci. Technol.Technol. A,A, Vol.Vol. 12,12, No.5,No.5, Sep/OctSep/Oct 19941994
thatthat X~2)X~2) isis notnot insignificant.insignificant. AnalysisAnalysis ofof thethe responseresponse inin Fig.Fig. 33 basedbased onon Eq.Eq. (7)(7) indicatesindicates thatthat thethe sumsum ofof (X~2)(X~2) ++ xkxk22»» isis atat leastleast ofof comparablecomparable magnitudemagnitude toto thethe contributioncontribution fromfrom X~~.:1<l>.X~~.:1<I>.
InIn anan earlierearlier work,26work,26 itit appearsappears thatthat thethe potentialpotential dropdrop inin thethe SCRSCR waswas assumedassumed toto bebe relativerelative toto thethe absoluteabsolute appliedapplied
fiatbandpotential,potential, notnot thethe flatband potential,potential, andand aa modelmodel ofof quadraticquadratic dependencedependence ofof thethe SHGSHG withwith appliedapplied potentialpotential waswas foundfound toto fitfit thethe data.data. AlthoughAlthough thethe flatbandflatband waswas notnot measuredmeasured inin thethe earlierearlier studies,studies, thethe current-voltagecurrent-voltage datadata presentedpresented suggestssuggests thatthat thethe flatbandflatband potentialpotential waswas wellwell negativenegative ofof 00 VV (versus(versus SCE).SCE). WithWith thisthis correction,correction, thethe datadata clearlyclearly dodo notnot fitfit aa quaquadraticdratic dependencedependence inin appliedapplied fieldfield strength,strength, butbut doesdoes reresemblesemble thethe datadata reportedreported here.here. Furthermore,Furthermore, thethe earlierearlier
37studiesstudies2626.,37 werewere performedperformed withoutwithout specifiedspecified polarizationpolarization andand samplesample orientationorientation informationinformation whichwhich makemake comparisonscomparisons difficult.difficult.
AnAn issueissue thatthat mustmust bebe addressedaddressed isis thethe effecteffect ofof thethe probeprobe lightlight onon thethe surfacesurface reactivity,reactivity, particularlyparticularly inin HH2S0S044 wherewhere thethe photogeneratedphotogenerated oxidesoxides areare insoluble.insoluble. TheThe influenceinfluence ofof thethe
2
Il-mintenseintense 1.0641.064 /l-m pumppump radiationradiation mustmust bebe consideredconsidered sincesince itit isis aboveabove thethe bandband gapgap forfor Si.Si. WithWith mostmost samples,samples, thethe signalsignalto-noiseto-noise ratioratio ofof thethe datadata improvesimproves markedlymarkedly afterafter thethe completioncompletion ofof oneone toto threethree rotationalrotational scansscans (4-12(4-12 min).min). PhoPhotocurrenttocurrent transientstransients areare observedobserved atat allall potentials,potentials, althoughalthough nearnear flatbandflatband theythey areare smallsmall andand bipolar.bipolar. TheseThese transientstransients reachreach aa steadysteady statestate atat anyany fixedfixed potential.potential. NearNear flatbandflatband thethe integratedintegrated chargecharge inin eacheach photocurrentphotocurrent transienttransient isis approxiapproximatelymately 10-10-77 CC (10-(10-66 ClcmClcm221laser1laser shot).shot). AssumingAssuming aa quantumquantum yieldyield ofof oneone forfor thethe oxidationoxidation ofof Si,Si, aa fourfour electronelectron process,process, thisthis indicatesindicates thatthat 30003000 laserlaser shotsshots willwill oxidizeoxidize aa monolayermonolayer ofof Si,Si, andand thusthus aa periodperiod ofof 1212 scansscans wouldwould bebe requiredrequired toto oxidizeoxidize aa monolayer.monolayer. ThisThis meansmeans thatthat thethe oxideoxide thicknessthickness inin HH22S0S044 isis lessless thanthan 33 ML.ML. InIn NHNH44F,F, ourour observationsobservations suggestsuggest thethe surfacesurface isis essentiallyessentially oxide-free.oxide-free. EarlierEarlier workwork3838 hashas shownshown aa slightlyslightly porousporous hydrogenhydrogen terminatedterminated surfacesurface growsgrows onon thethe siliconsilicon surfacesurface underunder continuouscontinuous illuminationillumination whenwhen biasedbiased positivepositive ofof flatband,flatband, butbut etchesetches awayaway whenwhen immersedimmersed
NH4Finin NH4F inin thethe dark.dark. TheThe lowlow dutyduty cyclecycle ofof thethe illuminationillumination usedused inin thesethese experimentsexperiments wouldwould produceproduce aa steadysteady statestate closeclose toto thethe nonporousnonporous surfacesurface condition.condition. Consequently,Consequently, nono changechange isis observedobserved inin thethe phasephase atat aa fixedfixed potentialpotential beforebefore andand afterafter allowingallowing aa samplesample toto sitsit inin thethe darkdark forfor aa periodperiod ofof 2020 min.min.
TheThe effecteffect ofof thethe oxideoxide isis clearlyclearly evidentevident inin thethe potentialpotential dependencedependence ofof thethe SHG.SHG. ComparedCompared toto thethe relativelyrelatively oxideoxidefreefree sample,sample, aa largerlarger biasbias mustmust bebe imposedimposed acrossacross thethe interinterfaceface toto obtainobtain aa relativerelative phasephase comparablecomparable toto thatthat ofof thethe H-terminatedH-terminated surface.surface. AtAt flatband,flatband, thethe fitsfits toto thethe anisotropiesanisotropies forfor thethe differentdifferent oxideoxide coveredcovered samplessamples showshow littlelittle variationvariation
X~) shouldshould bebe thethe similarsimilarasas oneone wouldwould expectexpect sincesince (X~2)(X~2) ++X~»
2 .forfor samplessamples withwith moremore thanthan aa fewfew monolayersmonolayers ofof Si0Si02. TheThe NH4Ffactfact thatthat thethe surfacesurface inin NH4F hashas aa slightlyslightly differentdifferent magnimagni
l(a)] 4tudetude andand phasephase [Fig.[Fig. l(a)J thanthan thethe HH22S0S04 casecase isis consistentconsistent withwith aa differentdifferent X~2)X~2) duedue toto thethe presencepresence ofof oxidizedoxidized surfacesurface species.species. AtAt potentialspotentials positivepositive ofof fiatband,flatband, aa progressivelyprogressively
Vapp isis necessarynecessary toto achieveachieve thethe magnitudemagnitude andand phasephaselargerlarger Vapp
angleangle thatthat isis obtainedobtained forfor surfacessurfaces withwith thinnerthinner oxideoxide coatcoatings.ings. ForFor example,example, anan additionaladditional ++ 1.01.0 VV mustmust bebe appliedapplied toto thethe 4040 AA samplesample toto obtainobtain aa phasephase angleangle comparablecomparable toto thatthat
26232623 al.:FischerFischer etet at.: SensitivitySensitivity ofof SHGSHG toto spacespace chargecharge effectseffects
obtainedobtained forfor thethe 2525 AAsamplesample biasedbiased atat thethe lowerlower voltage.voltage. WeWe attributeattribute thethe potentialpotential dependencedependence inin thethe relativerelative phasephase toto thethe variationvariation ofof thethe fieldfield atat thethe surfacesurface andand withinwithin thethe SCRSCR ofof thethe semiconductor.semiconductor. BecauseBecause ofof thethe insulatinginsulating naturenature ofof thethe oxide,oxide,
Si(11l)/Si0thethe fieldfield atat thethe Si(lIl)/Si022 interfaceinterface isis nono longer..1<1>longer..1<1> butbut isis reducedreduced byby thethe potentialpotential dropdrop acrossacross Vox,thethe oxideoxide layerlayer Vox'
ThisThis suggests,suggests, forfor example,example, thatthat thethe additionaladditional 1515 AAofof oxideoxide onon thethe surfacesurface screensscreens ~~ 11 VV ofof thethe overalloverall appliedapplied fieldfield relativerelative toto thethe 2525 AAsample.sample.
ForFor thethe UHVUHV studiesstudies thethe responseresponse fromfrom thethe H-terminatedH-terminated surfacesurface isis similarsimilar toto thatthat inin solutionsolution underunder conditions,conditions, wherewhere aa 100-100-200200 mmVV biasbias isis applied.applied. TheThe SHSH experimentsexperiments thusthus
fiatband.indicateindicate thatthat inin UHVUHV thethe surfacesurface isis notnot atat flatband. TwoTwo plauplausiblesible mechanismsmechanisms forfor thisthis includeinclude thethe existenceexistence ofof surfacesurface statesstates oror thethe chargingcharging ofof thethe surfacesurface (or(or contaminantcontaminant speciesspecies thereon)thereon) fromfrom straystray electronselectrons inin thethe chamber.chamber.
WeWe concludeconclude thatthat thethe potentialpotential dependencedependence predominantlypredominantly originatesoriginates inin thethe bulkbulk SCRSCR ofof thethe semiconductor.semiconductor. ItIt isis conconceivableceivable thatthat changeschanges inin thethe FresnelFresnel coefficientscoefficients ofof thethe oxide,oxide, especiallyespecially asas thethe oxideoxide thicknessthickness increases,increases, couldcould alsoalso acaccountcount forfor thisthis potentialpotential dependence.dependence. However,However, sincesince thethe comcomplexplex partpart ofof thethe dielectricdielectric forfor Si0Si022 variesvaries toto aa negligiblenegligible extentextent forfor thethe wavelengthwavelength regionregion employedemployed here,here, therethere isis veryvery littlelittle toto nono absorptionabsorption byby thethe oxide.oxide. Therefore,Therefore, thethe changeschanges inin thethe FresnelFresnel factorsfactors wouldwould bebe expectedexpected toto havehave onlyonly aa minorminor effect.effect. ClearlyClearly thethe bulkbulk bandband bendingbending inducedinduced byby thethe staticstatic electricelectric fieldfield isis thethe majormajor contributorcontributor toto thethe potentialpotential dependencedependence observedobserved here.here.
VI.VI. CONCLUSIONSCONCLUSIONS
TheThe studiesstudies presentedpresented herehere demonstratedemonstrate thethe potentialpotential deden-Si(Ul)pendencependence ofof n-Si(lll) samplessamples inin differentdifferent electrochemicalelectrochemical
environments.environments. TheThe mostmost strikingstriking resultresult isis thethe clearclear dependependencedence ofof thethe phasephase ofof thethe SHSH responseresponse onon thethe fieldfield presentpresent inin thethe SCR.SCR. TheThe abilityability toto opticallyoptically probe,probe, viavia SHSH phasephase meameasurements,surements, thethe staticstatic electricelectric fieldfield inin thethe presencepresence ofof oxideoxide onon thethe surfacesurface hashas notnot beenbeen demonstrateddemonstrated previously.previously. InIn contrastcontrast toto amplitudeamplitude measurements,measurements, phasephase measurementsmeasurements areare aa moremore accurateaccurate meansmeans ofof comparingcomparing surfacesurface changeschanges fromfrom oneone enenvironmentvironment toto anotheranother sincesince thethe phasephase willwill notnot bebe affectedaffected byby nonsignalnonsignal generatinggenerating elementselements inin thethe opticaloptical path.path.
TheThe amplitudesamplitudes ofof thethe isotropicisotropic andand anisotropicanisotropic responseresponse werewere shownshown toto bebe parabolicparabolic inin agreementagreement withwith earlierearlier work.work. However,However, thethe moremore detaileddetailed studiesstudies reportedreported herehere demondemonstratestrate thatthat therethere isis aa significantsignificant shiftshift inin thethe minimumminimum inin thethe responseresponse relativerelative toto thethe flatbandflatband potential.potential. ThisThis resultresult demdemonstratesonstrates thatthat therethere isis aa muchmuch largerlarger contributioncontribution fromfrom secondsecond orderorder termsterms thanthan assumedassumed previously.previously. TheThe parabolicparabolic naturenature ofof thethe responseresponse persistspersists withwith anan oxideoxide overlayeroverlayer but,but, extendsextends overover aa considerablyconsiderably widerwider potentialpotential rangerange duedue toto thethe insulatinsulatinging naturenature ofof thethe oxide.oxide. ThisThis isis consistentconsistent withwith thethe screeningscreening ofof thethe appliedapplied potentialpotential byby thethe oxideoxide whichwhich wouldwould leadlead toto reducedreduced bandband bendingbending inin thethe SCRSCR forfor anyany givengiven potential.potential. TheThe oxideoxide itselfitself doesdoes notnot appearappear toto opticallyoptically interfere.interfere.
TheThe mostmost significantsignificant resultresult isis thatthat thethe phasephase ofof thethe SHGSHG signalsignal forfor thethe buriedburied interfaceinterface isis directlydirectly correlatedcorrelated withwith thethe fieldfield inin thethe SCR.SCR. AwayAway fromfrom resonanceresonance thisthis providesprovides aa
26232623
uniqueunique probeprobe ofof thethe electronicelectronic environmentenvironment atat thethe surfacesurface applicableapplicable underunder aa varietyvariety ofof conditionsconditions includingincluding URV,URV, air,air, andand inin solution,solution, forfor bothboth clean,clean, H-terminated,H-terminated, andand oxidizedoxidized samples.samples. NoNo otherother singlesingle measurementmeasurement techniquetechnique isis appliapplicablecable inin allall thesethese environmentsenvironments forfor thethe determinationdetermination ofof suchsuch aa fundamentalfundamental propertyproperty ofof thethe surface.surface.
ACKNOWLEDGMENTACKNOWLEDGMENT
TheThe authorsauthors gratefullygratefully acknowledgeacknowledge fundingfunding fromfrom thethe DeDepartmentpartment ofof Energy,Energy, BasicBasic EnergyEnergy SciencesSciences (DE-FG06(DE-FG0686ER45273).86ER45273).
PresentedPresented atat thethe 40th40th NationalNational SymposiumSymposium ofof thethe AmericanAmerican VacuumVacuum Society,Society, Orlando,Orlando, FL,FL, 15-1915-19 NovemberNovember 1993.1993.
lp.lp. Jakob,Jakob, Y.Y. J.J. Chabal,Chabal, K.K. Raghavachari,Raghavachari, R.R. S.S. Becker,Becker, andand A.A. J.J. Becker,Becker, Surf.Surf. Sci.Sci. 275,275, 407407 (1992).(1992).
2G.2G. S.S. Higashi,Higashi, Y.Y. J.J. Chabal,Chabal, G.G. W.W. Trucks,Trucks, andand K.K. Raghavachari,Raghavachari, Appl.Appl. Phys.Phys. Lett.Lett. 56,56, 656656 (1990).(1990).
3T.3T. F.F. Heinz,Heinz, M.M. M.M. T.T. Loy,Loy, andand W.W. A.A. Thompson,Thompson, Phys.Phys. Rev.Rev. Lett.Lett. 54,54, 6363 (1985).(1985).
1. Technol.4T.4T. F.F. Heinz,Heinz, M.M. M.M. T.T. Loy,Loy, andand W.W. A.A. Thompson,Thompson, J. Vac.Vac. Sci.Sci. Techno!. BB 3,3, 14671467 (1985).(1985).
Appl.5K.5K. Itaya,Itaya, R.R. Sugawara,Sugawara, Y.Y. Morita,Morita, andand H.H. Tokumoto,Tokumoto, App!. Phys.Phys. Lett.Lett. 60,60, 25342534 (1992).(1992).
6p.6p. Allongue,Allongue, H.H. Brune,Brune, andand H.H. Gerischer,Gerischer, Surf.Surf. Sci.Sci. 275,414275,414 (1992).(1992). 1. Appl.7G.7G. S.S. Higashi,Higashi, R.R. S.S. Becker,Becker, Y.Y. 1.1. Chabal,Chabal, andand A.A. J. Becker,Becker, App!. Phys.Phys.
Lett.Lett. 58,58, 16561656 (1991).(1991). 8S._L.8S._L. Yau,Yau, F.-R.F.-R. F.F. Fan,Fan, andand A.A. 1.J. Bard,Bard, J.J. Electrochem.Electrochem. Soc.Soc. 139,139, 28252825 (1992).(1992).
Technol.9E.9E. Tomita,Tomita, N.N. Matsuda,Matsuda, andand K.K. Itaya,Itaya, J.J. Vac.Vac. Sci.Sci. Techno!. AA 8,8, 534534 (1990).(1990).
1. FestkoerperproblemeIOU.IOU. MemmertMemmert andand R.R. J. Behm,Behm, Festkoerperprob1eme 31,31, 189189 (1991).(1991). 1Iy'1Iy' KimKim andand C.C. M.M. Lieber,Lieber, J.J. Am.Am. Chern.Chern. Soc.Soc. 113,2333113,2333 (1991).(1991). l2p.l2p. R.R. Fischer,Fischer, J.J. L.L. Daschbach,Daschbach, andand G.G. L.L. Richmond,Richmond, Chern.Chern. Phys.Phys. Lett.Lett.
218,200218,200 (1994).(1994). 13N.13N. Sorg,Sorg, 1.J. Kruger,Kruger, W.W. Kautek,Kautek, andand 1.1. Reif,Reif, Ber.Ber. Bunsenges.Bunsenges. Phys.Phys. Chern.Chern.
97,40297,402 (1993).(1993). l4w. H.-J. Krause.14w. Daum,Daum, H.-I. Krause, U.U. Reichel,Reichel, andand H.H. Ibach,Ibach, Phys.Phys. Rev.Rev. Lett.Lett. 71,71,
12341234 (1993).(1993). l5G. J.15G. L.L. Richmond,Richmond, 1. M.M. Robinson,Robinson, andand V.V. L.L. Shannon,Shannon, Prog.Prog. Surf.Surf. Sci.Sci. 28,28,
II (1988).(1988). l6G.l6G. L.L. Richmond,Richmond, inin OpticalOptical SecondSecond HarmonicHarmonic GenerationGeneration asas anan InIn SituSitu
of Electrochemicallnteifaces,ProbeProbe oj Electrochemical Inteifaces, editededited byby A.A. J.J. BardBard (Marcel(Marcel Dekker,Dekker, NewNew York,York, 1991).1991).
l7H.l7H. 1.J. Lewerenz,Lewerenz, Electrochim.Electrochim. ActaActa 37,37, 847847 (1992).(1992). l8p. S. Pershan,18p. S.Pershan, Phys.Phys. Rev.Rev. 130,919130,919 (1963).(1963). 19p.19p. Guyot-SionnestGuyot-Sionnest andand Y.Y. R.R. Shen,Shen, Phys.Phys. Rev.Rev. BB 38,38, 79857985 (1988).(1988).
ofNonlinear20y.20y. R.R. Shen,Shen, TheThe PrinciplesPrinciples ojNonlinear OpticsOptics (Wiley,(Wiley, NewNew York,York, 1984).1984). 21y.21y. R.R. Shen,Shen, NatureNature 337,337, 519519 (1989).(1989). 221.221. E.E. Sipe,Sipe, D.D. J.J. Moss,Moss, andand H.H. M.M. VanVan Driel,Driel, Phys.Phys. Rev.Rev. BB 35,35, 11291129
(1987).(1987). 23H.23H. W.W. K.K. Tom,Tom, Ph.D.Ph.D. dissertation,dissertation, UniversityUniversity ofof California,California, Berkeley,Berkeley,
1984.1984. of24G.24G. L.L. Richmond,Richmond, inin SecondSecond HarmonicHarmonic GenerationGeneration asas anan In-situIn-situ ProbeProbe oj
SingleSingle CrystalCrystal ElectrodeElectrode Suifaces,Suifaces, editededited byby H.H. GerischerGerischer andand C.C. W.W. TobiasTobias (VCH,(VCH, 1992).1992).
Neff.25R.25R. Georgiadis,Georgiadis, G.G. A.A. Neff, andand G.G. L.L. Richmond,Richmond, 1.1. Chern.Chern. Phys.Phys. 92,92, 46234623 (1990).(1990).
26C.26C. Lee,Lee, R.R. K.K. Chang,Chang, andand N.N. Bloembergen,Bloembergen, Phys.Phys. Rev.Rev. Lett.Lett. 18,18, 167167 (1967).(1967).
27D.27D. A.A. Koos,Koos, V.V. L.L. Shannon,Shannon, andand G.G. L.L. Richmond,Richmond, 1.J. Phys.Phys. Chern.Chern. 94,94, 20912091 (1990).(1990).
28R.28R. A.A. Bradley,Bradley, R.R. Georgiadis,Georgiadis, S.S. D.D. Kevan,Kevan, andand G.G. L.L. Richmond,Richmond, J.J. Chern.Chern. Phys.Phys. 99,99, 55355535 (1993).(1993).
29J.291. Qi,Qi, M.M. S.S. Yeganeh,Yeganeh, 1.1. Koltover,Koltover, A.A. G.G. Yodh,Yodh, andand W.W. M.M. Theis,Theis, Phys.Phys. Rev.Rev. Lett.Lett. 71,71, 633633 (1993).(1993).
A. J.30p.30p. Guyot-SionnestGuyot-Sionnest andand A, Tadjeddine,Tadjeddine, 1. Chern.Chern. Phys.Phys. 92,92, 734734 (1990).(1990).
JVSTJVST AA -- Vacuum,Vacuum, Surfaces,Surfaces, andand FilmsFilms
81.: 2624262426242624 FischerFischer etet sl.: SensitivitySensitivity ofof SHGSHG toto spacespace chargecharge effectseffects
3'0. 35H.35H. J.J. LewerenzLewerenz andand T.T. Bitzer,Bitzer, J.J. Electrochern.Electrochern. Soc.Soc. 139,139, L21L21 (1992).(1992).3l0. L.L. Richmond,Richmond, inin Ref.Ref. 16.16. 32320.0. S.S. Higashi,Higashi, Y.Y. 1.1. Chahal,Chahal, O.O. W.W. Trucks,Trucks, andand K.K. Raghavachari,Raghavachari, Appl.Appl. 36H.36H. W.W. K.K. Tom,Tom, T.T. F.F. Heinz,Heinz, andand Y.Y. R.R. Shen,Shen, Phys.Phys. Rev.Rev. Lett.Lett. 51,51, 19831983�
Phys.Phys. Lett.Lett. 56,56, 656656 (1990).(1990). (1983).(1983).� 33p.33p. JakobJakob andand Y.Y. J.J. Chahal,Chahal, 1.1. Chern.Chern. Phys.Phys. 95,95, 28972897 (1991).(1991). 37370.0. A.A. AktsipetrovAktsipetrov andand E.E. D.D. Mishina,Mishina, Sov.Sov. Phys.Phys. Dokl.Dokl. 29,29, 3737 (1984).(1984).� 34R.34R. Houbertz,Houbertz, U.U. Mernmert,Mernmert, andand R.R. J.J. Behrn,Behrn, Appl.Appl. Phys.Phys. Lett.Lett. 58,58, 10271027 38L.38L. M.M. Peter,Peter, D.D. J.J. Blackwood,Blackwood, andand S.S. Pons,Pons, Phys.Phys. Rev.Rev. Lett.Lett. 62,62, 308308�
(1991).(1991). (1989).(1989).
J.J. Vac.Vac. Sci.Sci. Technol.Technol. A,A, Vol.Vol. 12,12, No.5,No.5, Sep/OctSep/Oct 19941994�
