ContentslistsavailableatScienceDirect
Applied Surface Science
j o ur na l ho me pa g e :w w w . e l s e v i e r . c o m / l o c a t e / a p s u s c
XPS for probing the dynamics of surface voltage and photovoltage in GaN
Hikmet Sezen
a, Ekmel Ozbay
b, Sefik Suzer
a,∗aDepartmentofChemistry,BilkentUniversity,06800Ankara,Turkey
bDepartmentsofPhysicsandElectricalandElectronicsEngineering,BilkentUniversity,06800Ankara,Turkey
a r t i c l e i n f o
Articlehistory:
Received10February2014
Receivedinrevisedform14June2014 Accepted14June2014
Availableonline20June2014
Keywords:
Chargingdynamics Surfacephotovoltage XPS
GaN Doping
a b s t r a c t
WedescribeapplicationoftwodifferentdatagatheringtechniquesofXPSforprobingthedynamicsof surfacevoltageandsurfacephotovoltage(SPV)developedinmicrosecondstosecondstime-domain,in additiontotheconventionalsteady-statemeasurements.Forthelonger(secondstomilliseconds)regime, capturingthedatainthesnapshotfashionisused,butforthefasterone(downtomicroseconds),square wave(SQW)electricalpulsesatdifferentfrequenciesareutilizedtoinduceandprobethedynamicsof variousprocessescausingthesurfacevoltage,includingtheSPV,viathechangesinthepeakpositions.
Thefrequencyrangecoversanywherefrom10−3to105Hzforprobingchangesduetocharging(slow), dipolar(intermediate),andelectronic(fast)processesassociatedwiththeexternalstressesimposed.We demonstrateitspowerbyapplicationton-andp-GaN,anddiscussthechemical/physicalinformation derivedthereof.Inaddition,themethodallowsustodecomposeandidentifythepeakswithrespectto theirchargingnatureforacompositesamplecontainingbothn-andp-GaNmoieties.
©2014ElsevierB.V.Allrightsreserved.
1. Introduction
GaNisawideband-gapcompoundsemiconductorwithsupe- rioroptoelectronicpropertiesutilizedinnumerousdevicesforhigh powerlightemittingdiodeandsensorapplications[1].Surfaceand defectstructuresofthematerialsusedcompletelydeterminemany ofthedeviceperformance,hencecontroloverthesepropertiesare ofutmostimportance.Surfacephotovoltage(SPV)[2,3],isoneofthe propertiesmeasuredtoevaluatethequalityandtheperformanceof suchmaterialsanddevices,andhasrecentlybeencombinedwith microscopictechniquessuchastheKelvinProbe(KP)[4,5].SPV measurementsrelateelectricalinformationtobandstructure,as wellastothenatureofsurfacemoieties, andmostimportantly tothenature of impurities and defects.However,even though asuperb (sub-micron)lateralresolutionisachievable,theKPis basedonelectricalmeasurements,hencedoesnothavechemical specificity.Ontheotherhand,X-rayPhotoelectronSpectroscopy (XPS),whichhasalsobeenusedforprobingSPV,providesexcel- lentchemicalinformation,inadditiontotheelectricalpotentials developedonsurfacestructures,althoughtheprecisionformea- suringthelatterismuchlower(∼20meV)whencomparedtothat ofKPmeasurements(∼1meV)[2].
∗ Correspondingauthor.Tel.:+903122901476;fax:+903122664068.
E-mailaddress:suzer@fen.bilkent.edu.tr(S.Suzer).
EarlierSPVmeasurementsweremostlysynchrotronbaseddue totheneededbrightnessofthesource[6–10],butcombinedwithX- raymicrofocusingandfastaccumulationtools,thenewgeneration commercialXPspectrometerscanalsoprovidecriticalinformation aboutsuchmaterialsanddevices[11,12].Inarecentpublication wereportedoncapturingofthetransientsurfacephotovoltagein n-andp-GaNbyXPSwherenewinformation aboutthemecha- nismofthetransientsformedbyilluminationwitha∼50mWviolet (405nm)laserwasbroughtoutanddiscussed[13].However,faster measurementsarenotpossibleinthismode,therefore,wehadto introduceothermodulationtechniques[14–19].
XPSisawidelyusedchemicalanalysistechniqueforprobing chemicalandphysicalcompositionofsurfacestructures.Incon- ventionaldatagatheringmodethesampleis grounded,and the positionaswellastheintensityofthepeaksarerecordedinastatic fashion[20].Althoughseveralreportshaveappearedintheliter- aturerelatedtodynamicalXPSmeasurementsinthetwoextreme ends,veryfast(nanosecondstoattoseconds)[21–24],andveryslow (minutestohours)[25–27].Theintermediatedomain(microsec- ondstoseconds)hasbeenoverlooked.Theveryfastmeasurements require either synchrotron facilities and/or specialized instru- mentationwhich arenot easilyaccessibletomany researchers.
Theintermediatedomainisimportantespeciallyfortestingand improvingperformanceofvariousmaterialsanddevicesusedfor charge storage, photovoltaics, ferroelectrics, chemical and bio- chemicalsensing,etc.Recently,ourgroupandothershavereported http://dx.doi.org/10.1016/j.apsusc.2014.06.089
0169-4332/©2014ElsevierB.V.Allrightsreserved.
throughvariouspublications,thebasicprinciplesandtechniques forimplementingsuchmeasurementsemployingverysimplemod- ificationtoconventionalinstruments[28–39].
Herein,we reportsuchXPSmeasurementsonn-andp-GaN samples,coveringtimeintervalsfromkilo-seconds(10−3Hz)down tomicrosecond(106Hz),anddiscusstheresultsobtainedpertain- ingtoelectricalpropertiesofthesampleasreflectedbythebinding energyshifts of theGa core-levels. Someof themeasurements includedinthepresentpaperhavealreadybeenpublishedinour previouspapers,buttheyarereproducedhereforthesakeofclar- ityandtoprovideacontinuousflowofthepresentation[13,40].
Thenextsectiondescribestheexperimentaldetailsandisfollowed byourresultsanddiscussions.Wefinishthepresentationbyour conclusions,tobefollowedbyacknowledgementsandreferences.
2. Experimental
GaNsamplesweregrownondoublepolishedc-planesapphire bylow-pressureMOCVD(AIX200/4RF-S).TheMgdopedp-andSi- dopedn-GaNsampleshaveconductivitiesof0.8and58S−1cm−1, respectively.AThermoFisherK-Alphaelectronspectrometerwith monochromatic AlK␣X-rays is used for XPS analysis,which is slightlymodifiedforimposingexternalvoltagestress(D.C.orA.C.) tothesampleduringdataacquisition. Thespectrometerisalso equippedwithalowenergyflood-gunfacilityforchargeneutraliza- tion,utilizingonlyelectronsorbothelectronsandAr+ions.Sample surfaceswerecleanedbysputtering withalowenergy(200eV) Ar+ionbeamforavoidingdamagebytheAr+ions,untiltheC1s peakfellbelowdetectionlimits.Noannealingofthesampleswas performedaftercleaning.Toinducesurfacephotovoltagea50mW 405nm(violet)CrystaLaserlaserisemployedeitherintheC.W.
mode,orthroughachopperwithafrequencyrangeof10−1–103. Forprobingthedynamicsofcharging/dischargingproperties,the dataisgatheredeitherinthefastsnapshotmodewith0.1stime resolutionorthesampleissubjectedtosquarewavepulses(SQW) of±10Vamplitudewithvaryingfrequenciesinthe10−3–105Hz rangeusingaStanfordResearchSystemDS345pulsegenerator, whilethedataisgatheredinnormalscanningmode.Ingeneralthe accuracyofourmeasurementsarebetterthan20meVinthepeak positions.However,sinceweallowthesamplestochargeanddis- charge,throughtheapplicationoftheSQWmodulation,thepeaks arebroader.Eveninsuchcase,wecanstillclaimanaccuracyof 20meV,sincethistimewemeasurethedifferencebetweentwo peakspositions,whichisinherentlymoreaccuratewhencompared toabsolutepeakpositionmeasurements.
3. Resultsanddiscussions
3.1. Photoresponseofn-andp-typeGaN
3.1.1. Steady-statemeasurements
Asurveyspectrumofthep-typeGaNhasanintenseGa2p3/2 peakat 1118.49eVand3s,3p,3dpeaksat161.41, 106.48,and 20.15eV,respectivelyasshowninFig.1.Inaddition,Gahasanum- berofLMMAugerlinesspanningtheintervalfrom380to630eV.
Unfortunately,abroadLMMAugerlineofGaat397eVisoverlap- pingwiththeN1speak,soitisnotpossibletofollowindividual N1sregionwithourXPSinstrumentusingAlK␣X-raysource.We generallyfollowtheGa2pifweneedhighintensity(inthesnapshot mode),buttheGa3dpeakisrecordedforbetterenergyresolution.
Whenexposedtoelectromagneticradiation,materialsrespond inavarietyofwaysdependingonthewavelength,intensity,and durationof theexposure. TheGa2p3/2 spectraof then-andp- GaNsamples,recordedwithandwithoutvioletlaserillumination, areshowninFig.2.Thehigheststeady-stateSPVforthen-and
Fig.1.AsurveyXPSspectrumofthep-GaNsampletogetherwiththeexperimental set-up.
p-GaN,aremeasuredas+0.15eVand−0.39eV,respectively,while thefloodgunisnotemployed.Whenthefloodgunisfunctioning, theresponseofespeciallythep-typeGaNissignificantlyaltered andprovidesadditionalinformationabouttheelectricalproperties ofthesemiconductor,aswasdiscussedinourpreviouspaper[13].
Whenthesamplesareintroducedintothespectrometer,elec- tronsflowfromthen-typeandholesfromthep-typeGaNsothat theyarechargeddifferently,duetotheirsignandextentofdoping.
But,thechargesonthesurfacesarescreenedbybandbendingand byaccumulationofopposite-signchargesnearthesurface.When lightwithsufficientenergyisincidentonthesampleadditional chargecarriersareintroducedwhichnormallyreturnthebandsto theirflat-bandcondition[5,10,13,40–45].GaNhasaband-gapof 3.4eV,thereforethevioletlaser,whichhasonlyanenergyof3.1eV isnotexpectedtocauseanyband–bandexcitation.Presenceofsur- facestates,justabovethevalanceband,aswellasdefects,aidedby thermalenergyatroomtemperaturemustbethereasonforthis band-flattening.Asdepictedbytheschematicsshowninthelower panelofFig.2,theseprocessesmustinvolveanexcitationofanelec- tronandaholefromsurfacestatesintobulkforthen-andp-GaN, respectively,resultinginpartialflatteningofthebands.
Fig.2. Ga2p3/2regionofbothsamplesrecordedwithoutandunderillumination withaC.W.VioletLaser(405nm)of∼50mWpower.Thelower panelsshow schematicallyflatteningoftheband-bendingunderillumination.
Fig.3.VariationofthepositionofGa2p3/2peakofp-GaNwithrespecttothepower oftheilluminationsource,controlledbyNeutralDensityOpticalFilters(NDF).
Wehavealsocarriedoutmeasurementsbyvaryingtheintensity oftheexcitationsourcetoobtainarelationshipbetweenthemea- suredvalueoftheSPVandthelogarithmoftheintensityoflightas giveninFig.3.Themostimportantfindingofthesemeasurements, isthatitrevealsasaturationbehaviortowardthehighintensity regionofthephotoexcitement.Thispowerdependencyofthep- GaNindicatesthattheresponseswemeasuredaremostlydueto surface-statesanddefects’relatedprocesses,whichisalsoingood agreementwithourmechanisticexplanationofthepartialband flatteningphenomenoninvolvingthesub-bandgapexcitations.
3.1.2. Transientmeasurementsusingthesnapshotmode
TheK-Alphaspectrometerallowsustorecord,withreasonably highsignal-to-noiseratio,anarrowspectralregionwithlessthan 0.1sintervals.InFig.4(a)and(b),thetime-resolvedGa2p3/2spectra
areshownforthetwosampleswhilethelaseristurnedonand offevery50s,whereweobservethattheSPVdevelopsinamuch shortertimeinterval(<0.1s)thanweareabletomeasure,butitgets severelyscreenedbytheflood-gunelectronsforthep-GaN.This indicatesthatwhenthefloodgunisoperativeanadditionalmech- anismistriggeredjustafterthechangeofthestateofthelaser,asa resultofinteractionbetweentheelectronsofthefloodgunandthe surfaceofthep-GaNsample.Thiscanbeexplainedasfollows.The p-GaNdevelopsacertainmagnitudeofdownwardband-bending atitssurface.ThisdownwardbandbendingatthelaserOFFstate providesaproperwelltocapturetheflood-gunelectronsatornear theconductionband,asschematicallyshowninFig.4(c)tocause anoppositeshifttolowerbindingenergyasexpectedandmea- sured.WhenthestateofthelaserischangedtoON,thedownward band-bendingis immediatelyflattenedand thewelldisappears.
Hence,theaccumulatedelectronsduring thepreviousstateare quicklysweptawaytothebulk,becausenowtheconductionband ofthebulkhasmanyfavorableenergylevelsfortheelectronsas alsoindicatedinFig.4(c).Thiseventisobservedasafurthershiftto ahigherbindingenergy.Moreover,thetimeconstantsofelectron accumulationandelectronsweepingawayfromthesurfaceofthe p-GaNsamplearedifferentandare6.3and0.85s,respectively.This differencearisesfromthedifferentnatureofthetwomechanisms.
Theimpingingelectronstakelongertime tobeaccommodated, sinceeachaccumulatedelectroncontributestothebuild-upofthe negativesheetofchargeatthesurface,whichinturnsuppresses the rate of accumulation of further negative charge. However, whenthelaseristurnedon,thesweepingofaccumulatedelectrons tothebulkisveryeffectiveandquickduetoitslargevolumeof thebulkcomparedtothesurface.Noteinpassingthatwhenthe flood-gunisoperativetheSPVaresmearedout,andundercertain conditions, might be completely unobservable by the normal scanning mode of the spectrometer due to the time-averaged dataacquisition.Only,throughthecarefulimplementationofthe
Fig.4.VariationsofthecenteroftheGa2p3/2peaksrecordedwith0.1sintervalsusingthesnapshotmode,asthelaseristurnONandOFF,withtheflood-gunturned-offand on;for(a)n-GaN,and(b)p-GaN.(c)SchematicsofformationoftheSPVtransientsforthep-GaN,andtheexponentialfitsaregivenfortheLaserONandOFFperiods.
Fig.5.Ga3dpeakofthen-andp-GaNsamplerecorded;(i)grounded,andunder theSQWmodulationat(ii)2MHz,and(iii)10kHzfrequencies,withoutandunder photoillumination.
snapshotmodewithrespecttothetimewindowsweusedwere weableustorecoverthefullmagnitudeoftheSPV[13].
3.1.3. ApplicationofSQWpulses
Sinceneitherthesnap-shotmodenorutilizationof anopti- calchoppercanbeusedintheshortermicrosecondsregime,we makeuseofapplicationofelectricalvoltagestressesintheform ofSQWpulses.InFig.5,wedisplaytheGa3dpeakofthen-and thep-GaNsamplerecordedwhilethesamplesaregrounded,as wellasunder2MHzand10kHzSQWmodulations,andwithout andwithvioletlaserillumination.SQWmodulationofthesample resultsintwinningofallthepeaks,sincethesamplesexperience either+10Vor−10Velectricalstressduringtheup(positive)and thedown(negative)cycles,respectively.Ifthesampleisconduct- ingtheresultistrivialshiftsofthepeakpositionsto+10.00and
Fig.6. Measuredpeakpositionsat+10Vand−10VcyclesoftheGa3dpeakasa functionofthefrequencyoftheappliedSQWexcitation,withoutandunderphoto- illumination.Thelowerpartplotsthedifferencebetweenthetwinnedpeaks.
−10.00eV,respectively,andwithabindingenergydifferencemea- surementofexactly20.00eVbetweenthem.However,ifthesample is poorly conducting,a difference of less than 20.00eV results, sincechargeneutralizationisdifferentunder+10and−10Vbiasing cycles[14–19].Thismeasurednegativedeviationfromthe20.00eV islargeratlow frequenciessincemoretimeis allocatedforthe sampletochargeordischarge,andapproacheszeroastheSQW modulationfrequencyincreases,dependingonthemechanismof theoperatingprocesses. Hence,aplotof themeasuredbinding energydifferenceasafunctionoftheappliedSQWfrequenciesis highlyinformative(seebelowFig.6)[14–19].
Accordingly,asshowninFig.5,thebindingenergydifference iscloserto20.00eV,yetslightlylowerat2MHz,fromwhichwe
Fig.7. Ga3dpeakofacompositesampleformedbyplacingthen-andp-GaNsamplesside-by-side,recorded;(i)grounded,andundertheSQWmodulationat(ii)2MHz, and(iii)10kHzfrequencies,withoutandunderphotoillumination.Theinsetdepictstheexperimentalset-up.
cannowdeducethatn-GaNismuchmoreconductingcomparedto thep-GaNsample,inagreementwiththemeasuredvaluesgiven intheexperimentalsection.Thissimpleagreementisactuallyan importantproofofthevalidityofourmethodology,whichcould leadtonumerousanalyticalapplicationsforharvestingdielectric propertiesofbothsimpleandcompositesurfacestructures[28,46].
Inaddition,themeasureddifferenceof19.81eVatthelowerfre- quencyof2MHzislowerthan19.98eVmeasuredat10kHz,while thelatteriswithintheexperimentaluncertaintyofthetheoreti- calvalueof20.00eV.Thesituationisverydifferentforthep-GaN, whichisdrasticallycharged,andthemeasuredvalueis16.74eV at2MHz, and approaches(19.93eV)to thetheoreticalvalueat 10kHz,butdoesnotquitereachit.Whatonelearnsfromthesemea- surementsisthat,especiallyforthep-GaN,charging/discharging propertiesarestronglyaffectedbyamultitudeofprocessesopera- tiveinthemicrosecondsallthewayupsecondsrange.
Inthesamefigure,datarecordedunderlightilluminationare alsogiven.For then-GaN,themagnitude ofthephotoresponse issmall,andisthesameirrespectiveofthemodeofdatagath- ering,whichrevealsthat photoresponseof thesampleis much faster(<10s)thanwecanmeasureevenifweemploytheSQW modulationuptothefrequencyof105Hz.Themagnitudeofthe photoresponseofthep-GaNismuchlarger,and,similartothecase oftheelectricalcharging,hasmultiple(slow,intermediateandfast) components,ascanbededucedfromthefactthatevenat10kHz modulation,themeasuredbindingenergydifferenceof19.97eV, underilluminationislargerthanthat(19.93eV)withoutit.
Acomprehensivesetof measurementsfor thep-GaNisdis- played in Fig.6, where themeasuredpositions of thetwinned peaksoftheGa3dpeak,aswellasthedifferencebetweenthem areplottedasafunctionoffrequencyoftheSQWpulses,bothwith andwithoutC.W.violetlaserillumination.Themeasuredcharging shiftsaresignificantlyalteredintwodomains;(i)0.2–20Hz,and (ii)0.2–10kHz,andcontinuetopersistevenathigherfrequencies.
Wemustemphasizethatallthechargingshiftswehavereportedso farinourpreviousstudiesonvariousdielectricmaterialslikeSiO2, CdSandpolymershavealwaysbeeninthelowerfrequencyrangeof 10−1–102Hz,andthisisthefirsttimeweareabletorecordthemin thekHzrange,whicharetoofastfortheusualchargecompensation processesinvolvingseveralcascadingprocesses,andmustsome- howberelatedwithelectronicprocesses.Accordinglywepropose thatatleastforthep-GaN,wehavesuccessfullydemonstratedthat withacombinationofthesnap-shotmeasurementsandutilization ofSQWpulses,XPSisabletoprobeanddistinguishamongtheslow (ionic),intermediate(dipolar),andthefast(mostlikelyduetoelec- tronic)chargingprocessesofmaterialsandtheirphotoresponses.
Asimple analyticalextensionofourmethodisdemonstrated inFig.7,whereweanalyzeacompositeregionconsistingofboth n-andp-GaNsamplesplacedside-by-sideinanad-hocp–njunc- tionfashion.TheGa3dpeakofthecompositesampleappearsas a singleone,when recorded withthesampleis grounded, and thephotoresponseisnotstrongenoughtodistinguishbetween thedifferentlydopeddomains.However,whentheSQWpulses with10kHzmodulationareapplied,thepeaksdecomposetotwo distinctcomponents,assignabletothen-orthep-domains,respec- tively.Atthelowerfrequencythedecompositionhappensonlyfor the+10Vcycle,whilestillonlyoneoverlappingcompositepeakis observableinthe−10Vcycle.Thephotoresponsesofthen-and thep-componentsare notasstrong,but neverthelessarealso indicativeofthenatureofthedopingascanalsobeseenfromfigure.
4. Conclusions
Byutilizingdifferentopticaland/orelectricalmodulationtech- niques, XPS spectroscopy, which is conventionally used in the
staticscanningfashion,canalsobeusedforprobingthedynam- ics of surface structures. A combination of time-resolved XPS with0.1sintervals,incorporationofSQWmodulation,anduseof photoilluminationprovidesusnewwaysofinvestigatingsurface electronicstructureandothersurfacepropertiesofsemiconduct- ing and dielectric materials in a chemically specificfashion, in themicrosecondstosecondstimedomains,which hasnotbeen reportedbefore.Oneparticularadvantageofthemethodisitsabil- itytodistinguishandidentifythepeaksrepresentingtheirdoping natureforcompositesurfacestructures,usingtheirresponsesto electrical,electromagneticstresses,aswellastoacombinationof both.
Acknowledgements
ThisworkwassupportedbytheScientificandTechnological ResearchCouncilofTurkey(TUBITAK)GrantNo:212M051.
References
[1]M.Kocan,A.Rizzi,H.Luth,S.Keller,U.K.Mishra,Phys.StatusSolidiB234(2002) 773.
[2]L.Kronik,Y.Shapira,Surf.Sci.Rep.37(1999)1.
[3]D.K.Schroder,Meas.Sci.Technol.12(2001)R16.
[4]M.Foussekis,J.D.Ferguson,A.A.Baski,H.Morkoc,M.A.Reshchikov,Phys.B404 (2009)4892.
[5]M.A.Reshchikov,M.Foussekis,A.A.Baski,J.Appl.Phys.107(2010)113535.
[6]J.E.Demuth,B.N.J. Persson,A.J. Schell-Sorokin,Phys.Rev.Lett.51(1983) 2214.
[7]J.E.Demuth,W.J.Thompson,N.J.DiNardo,R.Imbihl,Phys.Rev.Lett.56(1986) 1408.
[8]A.Schellenberger,R.Schlaf,C.Pettenkofer,W.Jaegermann,Phys.Rev.B45 (1992)3538.
[9]R.Schlaf,A.Klein,C.Pettenkofer,W.Jaegermann,Phys.Rev.B48(1993) 14242.
[10]M.Moreno,H.Yang,M.Horicke,M.Alonso,J.A.Martin-Gago,R.Hey,K.Horn, J.L.Sacedon,K.H.Ploog,Phys.Rev.B57(1998)12314.
[11]H.Shpaisman,E.Salomon,G.Neher,A.Vilan,H.Cohen,A.Kahn,D.Cahen,J.
Phys.Chem.C113(2009)3313.
[12]S.Schafer,S.A.Wyrzgol,R.Caterina,A.Jentsy,S.J.Schoell,M.Havecker,A.Knop- Gericke,J.A.Lercher,I.D.Sharp,M.Stutzmann,J.Am.Chem.Soc.136(2012) 12528.
[13]H.Sezen,E.Ozbay,O.Aktas,S.Suzer,Appl.Phys.Lett.98(2011)111901.
[14]H.Sezen,G.Ertas,A.Dana,S.Suzer,Macromolecules40(2007)4109.
[15]S.Suzer,H.Sezen,A.Dana,Anal.Chem.80(2008)3931.
[16]S.Suzer,H.Sezen,G.Ertas,A.Dana,J.ElectronSpectrosc.Relat.Phenom.176 (2010)52.
[17]H.Sezen,G.Ertas,S.Suzer,J.ElectronSpectrosc.Relat.Phenom.178-179(2010) 373.
[18]H.Sezen,S.Suzer,J.Vac.Sci.Technol.A28(2010)639.
[19]H.Sezen,S.Suzer,Surf.Sci.Lett.604(2010)L159.
[20]D.Briggs,M.P.Seah,PracticalSurfaceAnalysis,2nded.,Wiley,Chichester,1999.
[21]M.Marsi,R.Belkhou,C.Grupp,G.Panaccione,A.Taleb-Ibrahimi,L.Nahon,D.
Garzella,D.Nutarelli,E.Renault,R.Roux,M.E.Couprie,M.Billardon,Phys.Rev.
B61(2000)R5070.
[22]M.Marsi,M.E.Couprie,L.Nahon,D. Garzella,R. Bakker,A.Delboulbe,D.
Nutarelli,R.Roux,B.Visentin,C.Grupp,G.Indlekofer,G.Panaccione,A.Taleb- Ibrahimi,M.Billardon,Nucl.Instrum.MethodsPhys.Res.A393(1997)548.
[23]M.Marsi,M.E.Couprie,L.Nahon,D.Garzella,T.Hara,R.Bakker,M.Billar- don,A.Delboulbe,G.Indlekofer,A.Taleb-Ibrahimi,Appl.Phys.Lett.70(1997) 895.
[24]M.Marsi,L.Nahon,M.E.Couprie,D.Garzella,T.Hara,R.Bakker,M.Billar- don,A.Delboulbe,G.Indlekofer,A.Taleb-Ibrahimi,J.ElectronSpectrosc.Relat.
Phenom.94(1998)149.
[25]K.Hirose,M.Yamawaki,K.Torii,T.Kawahara,S.Kawashiri,T.Hattori,Appl.
Surf.Sci.237(2004)411.
[26]K.Hirose,H.Nohira,K.Azuma,T.Hattori,Prog.Surf.Sci.82(2007)3.
[27]K.Hirose,J.ElectronSpectrosc.Relat.Phenom.176(2010)46.
[28]S.Suzer,Anal.Chem.75(2003)7026.
[29]F.Karadas,G.Ertas,S.Suzer,J.Phys.Chem.B108(2004)1515.
[30]S.Suzer,A.Dana,J.Phys.Chem.B110(2006)19112.
[31]G.Ertas,U.K.Demirok,A.Atalar,S.Suzer,Appl.Phys.Lett.86(2005)183110.
[32]G.Ertas,U.K.Demirok,S.Suzer,Appl.Surf.Sci.249(2005)12.
[33]S.Suzer,A.Dana,G.Ertas,Anal.Chem.79(2007)183.
[34]C.P.Conger,S.Suzer,Langmuir25(2009)1757.
[35]H.Cohen,R.Maoz,J.Sagiv,NanoLett.6(2006)2462.
[36]H.Cohen,S.K.Sarkar,G.Hodes,J.Phys.Chem.B110(2006)25508.
[37]A.Samokhvalov,R.W.Gurney,M.Lahav,S.Cohen,H.Cohen,R.Naaman,J.Phys.
Chem.B107(2003)4245.
[38]S.K.Sarkar,G.Hodes,L.Kronik,H.Cohen,J.Phys.Chem.C112(2008)6564.
[39]H.Cohen,J.ElectronSpectrosc.Relat.Phenom.176(2010)24.
[40]H.Sezen,S.Suzer,J.Spectrosc.Dyn.2(2012)3.
[41]N.J.Halas,J.Bokor,Phys.Rev.Lett.62(1989)1679.
[42]J.P.Long,H.R.Sadeghi,J.C.Rife,M.N.Kabler,Phys.Rev.Lett.64(1990)1158.
[43]J.P.Long,V.M.Bermudez,Phys.Rev.B66(2002)121308.
[44]E.A.Kraut,R.W.Grant,J.R.Waldrop,S.P.Kowalczyk,Phys.Rev.Lett.44(1980) 1620.
[45]S.A.Ding,S.R.Barman,K.Horn,H.Yang,B.Yang,O.Brandt,K.Ploog,Appl.Phys.
Lett.70(1997)2407.
[46]M.T.Camci,S.Suzer,J.Vac.Sci.Technol.A32(2014)021510.
