ContentslistsavailableatScienceDirect
Catalysis Today
j o ur na l ho me p a g e :w w w . e l s e v i e r . c o m / l o c a t e / c a t t o d
Structure and reactivity of Au–Rh bimetallic clusters on titanate nanowires, nanotubes and TiO 2 (1 1 0)
J. Kiss
a,b, L. Óvári
b, A. Oszkó
a, G. Pótári
a, M. Tóth
a, K. Baán
a, A. Erdóhelyi
a,b,∗aDepartmentofPhysicalChemistryandMaterialScience,UniversityofSzeged,Aradivértanúkt.1,H-6720Szeged,Hungary
bReactionKineticsResearchLaboratory,InstituteofNanochemistryandCatalysis,ChemicalResearchCenterofHungarianAcademyofSciences, P.O.B.168,Aradivértanúkt.1,H-6701Szeged,Hungary
a r t i c l e i n f o
Articlehistory:
Received29March2011
Receivedinrevisedform25May2011 Accepted1June2011
Available online 30 June 2011
Keywords:
Core–shellstructure Titaniananowire Titaniananotube Gold
Rhodium
Ethanoldecomposition
a b s t r a c t
AuandRhclusters,aswellasAu–Rhbimetallicnanoparticleswerepreparedontitanatenanowires, nanotubesandonTiO2(110).TheywerecharacterizedbyX-rayphotoelectronspectroscopy(XPS),low energyionscatteringspectroscopy(LEIS)andFouriertransforminfraredspectroscopy(FTIR).Byper- formingcarefulLEISexperiments,itwasfoundthatforappropriateAuandRhcoverage,athinAulayer almostcompletelycoverstheRhnanoparticles,aRhcore–Aushellstructurewasdetected.Theforma- tionofthisstructurewasnotaffectedbyalkali(K)adatoms.LEISandFTIRmeasurementsdisclosedthat adsorbedCOat300KcausesthesegregationofRhatomstothesurfaceofmetalclustersinordertobind toCO.UponCOadsorptiononRh/titanatenanostructurestheIRstretchingfrequenciescharacteristicof thetwinformweredominant,whereasbimetallicnanosystemsfeaturedapronouncedlinearstretching vibrationaswell.InspiteofthisstructureadsorbedCOisdetectableduringtheethanoladsorptionon gold–rhodiumbimetallicclusterandtheethanoldecompositionrateistwicehigherthanonAu/TiO2.
© 2011 Elsevier B.V. All rights reserved.
1. Introduction
Highlydispersedgoldparticleshaveincreasinglygainedatten- tion in the catalytic community due totheir unusual catalytic propertiesinavarietyofoxidationandhydrogenationreactions [seee.g.Refs.[1–3]andreferencestherein].Supportedbimetallic catalystshave been shown toexhibit superior catalytic perfor- manceswhencomparedtotheirmonometalliccounterparts[4,5].
Dependingonthenatureofthesecondmetal,differentkindsof mechanismswereobservedintheinteractionswithbimetals.Both ligandeffects(i.e.anelectronicfactorduetothechangeinelectron density)andensembleeffects(i.e.surfacestructurefactordueto thepositioningofdifferentatomtypes)cancontributetosyner- geticeffectsoftenobservedforbimetallicsurfaceparticles[6–8].
Enhanceddispersionand stabilityofgoldnanoparticlesonstoi- chiometricandreducedTiO2(110)wereobservedinthepresence ofmolybdenum[9].IthasalsobeenobservedthatRhsignificantly changedthemorphologyandtopologyofAuonTiO2(110)sur- face[10,11].STMandlowenergyionscatteringspectroscopy(LEIS) experimentsrevealedthatatproperAuandRhcoveragethepost- depositedAucompletelyanduniformlycoverstheRhnanoparticles
∗Correspondingauthorat:DepartmentofPhysicalChemistryandMaterial Science,UniversityofSzeged,Aradivértanúkt.1,H-6720Szeged,Hungary.
Tel.:+3662343638;fax:+3662546482.
E-mailaddress:erdohely@chem.u-szeged.hu(A.Erdóhelyi).
(core–shellstructure).InthecaseoftheAu–Pdsystemtheresults showsegregationofgoldtothesurface.Densityfunctionaltheory calculationconfirmsthatAupreferstobeattheedgesofAuPdalloy particlesundervacuumconditions[12].ThepresenceofPtinthe bimetallicPt–Auclustersinhibitssintering,andtheaveragesizeof theclustersafterannealingdecreaseswithincreasingPtcontent.
BasedonLEISandSTMexperiments,performedonTiO2(110),it wasstatedthatthedepositionofAuonPtclustersresultsinthe formationofbimetallicclustersduetotheseedingofmobileAu atomsatexistingPtnuclei,butthedepositionofAuonPtdoes notproducecore–shellstructureswithAuontopatsmallcover- ages[13].LaterontheformationofPtcore–Aushellstructureswas demonstratedathighercoverages(0.25–0.5ML).AdsorptionofCO onthePtclusterscoveredbythegoldcappinglayerpromotedthe diffusionofPtatomstotheclustersurfaceinordertobindtoCO [14].
Inadditiontotheabovefindingsitturnedoutthatthenature ofthesupport(i.e.Fe2O3(111),MgO(100),andCeO2(111))often playsacriticalroleinreactivity[4,5,12].Theobservedvibrational frequencyofCOadsorbedonthesurfacecanbeusedtoidentify itsspecificbindingsitesandthustoelucidatethesurfacecomposi- tion.DesorptionofCOasafunctionoftemperatureprovidessimilar informationandsupportstheidentificationofbindingsitesonthe surface.Thisapproachhasbeensuccessfullyappliedinthepastto investigateAu–Pdalloysurfaces[15],Au–Pdparticlessupportedon SiO2[16]andAl2O3film[17],Au–Rhbilayersonsupportedpowder TiO2[18]andontitanatenanowireandnanotubes[19,20].
0920-5861/$–seefrontmatter© 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.cattod.2011.06.002
Lowenergyion scatteringspectroscopy(LEIS),besidesother surfacesensitivetechniques,wassuccessfullyappliedinthechar- acterizationofbimetallicnanoclustersonTiO2(110)[9–11,13,14], becauseithasveryhighsensitivityinthetopmostlayer.In this paperweusethistechniquewithXPSandFTIRmethodstoinvesti- gatetheeffectofpotassiuminthearrangementofAu–Rhclusters onTiO2(110)surfaceandontitanatenanowireandnanotubessup- ports.
One-dimensionalnanostructureshavebeeninthefocusofthe material sciencecommunity for well more thana decade now [21].Althoughcarbonnanotubesmaysolvemanychallenges of materials engineering in the long run, at present their practi- calapplicabilityappearstobelimitedbythelackofadequately selective synthesis technologies. On the other hand, inorganic nanostructures(e.g.metallicnanowires,oxides,sulfides,selenides, etc.)canbepreparedinarathercontrolledmanner,andtherefore theirindustrial-scaleapplicationisveryclosenow.
Titanatenanotubesarealsoofgreatinterestforcatalyticappli- cations,since their highcation exchange capacity providesthe possibilityofachievingahighmetaldispersion[22].Inthisregard, thereareseveraldemonstratedexamplesofsuccessfulutilization oftitanatenanotubesasmesoporouscatalystsupportsfordiffer- entnanoparticles.High-aspect-ratioTiO2andtitanatenanoobjects areintensivelystudiedatpresentbecauseoftheirpromisingphoto- electrical[23,24],biomedical[25],andhydrogenstorageproperties [26–28]. Gold-containing titania nanotubes were found to dis- playhigheractivityinthephoto-oxidationofacetaldehyde[29], thewater–gasshiftreaction[30],andCOoxidation[31]thanthe DegussaP-25catalyst.
In this paper, we shall use the term “nanotube” for high- aspect-ratioobjectswithahollowinnerchannelbelow50nmand
“nanowire”forhigh-aspect-ratioobjectswithoutaninnervoidand adiameterbelow200nm.
The catalytic activity of titanate structures was tested in theethanol decomposition. Catalytic conversion of bio-ethanol receivedconsiderableattentioninthelastdecade[32–37]andthe searchforthemostactive,stableandselectivecatalysthasresulted inmanyheterogeneouscatalyticsystems.
SupportedRh isoneof thebestsamplesfor thesereactions, butonlyafewpapers dealwiththeefficiencyofsupportedAu.
Idrissetal.[38]studiedtheoxidationofethanolonAu/CeO2,and theadsorptionand desorption of it onAu/TiO2 [39]. Guan and Hensen[40]recentlyexaminedthedehydrogenationofethanol onAunanoparticlesdepositedonvariousSiO2supports.Astrong influenceoftheAuparticlesizewasobserved.Itisnoteworthy that,inthepresenceof oxygen,theintrinsicactivityofAu/SiO2 increasedconsiderably.LippitsandNieuwenhuys[41]investigated theconversionofethanolintoethyleneoxideongold-basedcata- lysts.ItwasfoundthatwhengoldwasaddedtotheAl2O3support thecatalystproducedethyleneoxidewhichwasnotobservedon thesupportalone.Thepresenceofgoldnanoparticlesisnecessary fortheformationofethyleneoxide.Solymosiandco-workers[37]
studiedtheethanoldecompositiononAunanoparticlesdeposited onvariousoxides.Theformationofhydrogenandtheproductdis- tributiondependedsensitivelyonthenatureofthesupport.These samplescatalyzemainlythedehydrogenationofethanol,topro- ducehydrogenand acetaldehyde. Thehighest rateofhydrogen evolutionwasobservedonAu/CeO2.
2. Experimental
Titanatenanowiresandnanotubeswerepreparedbyhydrother- malconversionofanataseTiO2 as describedelsewhere[42,43].
Briefly,thenanostructureswerepreparedbymixing2gofanatase into140cm310MaqueousNaOHsolutionuntilawhitesuspen- sionwasobtained,agingthesuspensioninaclosed,cylindrical,
Teflon-lined autoclave at 400K for 1–72h while rotating the wholeautoclaveintensivelyat60rpmarounditsshortaxis,and finally washingthe productwith deionized water and neutral- izingwith0.1MHClacidsolutiontoreachpH=7;atthispoint, theslurrywasfilteredandtheresiduewasdriedinairat353K.
Au,Rhand theircoadsorbedlayers withdifferentcompositions wereproducedbyimpregnatingtitaniananowiresandnanotubes withthemixturesof calculated volumesof HAuCl4 (Fluka) and RhCl3·3H2O(JohnsonMatthey)solutionstoyield1wt%metalcon- tent.InthebimetalliccatalysttheAu/Rhatomicratiowas1.The impregnatedpowdersweredriedinairat383Kfor3h,thanoxi- dizedat573K.Thefinalpre-treatmentwasat573Kinhydrogen atmosphere[18].
Titanatenanotubesareopenendedhollowtubularobjectsmea- suring7–10nminouterdiameterand50–170nminlength.They featureacharacteristicspiralcrosssectioncomposedof4–6wall layers.Thetypicaldiameteroftheirinnerchannelis5nm.Titanate nanowirerepresentsthethermodynamicallymoststableformof sodiumtrititanateundertheappliedalkalinehydrothermalcon- ditions(notethatthepost-syntheticneutralizationstepconverts theoriginal Na2Ti3O7 into its hydrogenform withoutaffecting thenanowiremorphology).Theirdiameteris45–110nmandtheir lengthisbetween1.8and5m.Thespecificsurfaceareaoftitanate nanotubesisratherlarge(∼185m2g−1)duetotheirreadilyaccessi- bleinnerchannelsurface,whereasthatofsolidtitanatenanowires is∼20m2g−1.Theimpuritylevelofproducednanocompositeswas lessthan1%.TheforeignelementsdeterminedbyXPSwereK,Ca, MgandNaremainedfromthepreparationprocess.
Theultrahighvacuumexperimentsweredonein twosepa- ratechambers.Oneofthem(1)wasusedtoanalyzethehigharea titanatesamples,whilemeasurementsrelatedtoTiO2(110)were conductedintheother(2).
1.XPspectraweretakenwithaSPECSinstrumentequippedwith aPHOIBOS150MCD9hemisphericalanalyzer.Theanalyzerwas operatedintheFATmodewith20eVpassenergy.TheAlK␣radi- ation(h=1486.6eV)ofadualanodeX-raygunwasusedasan excitationsource.Thegunwasoperatedatthepowerof150W (12.5kV,12mA).Theenergystepwas25meV,electronswerecol- lectedfor100msinonechannel.Typicallyfivescansweresummed togetasinglehighresolutionspectrum.Forbindingenergyrefer- encetheTi2p3/2 maximum(458.9eV)wasused.Thesamedata wereobtainedwhenC1s(adventitiouscarbonat285.1eV),orO1s latticeoxygen(530.4eV)wasusedasreferences.Asampleprepa- rationchamberwasdirectlyconnectedtothemeasuringchamber toavoidthecontaminationofsamplesbetweeneachstep.Forspec- trumacquisitionandevaluationbothmanufacturer’s(SpecsLab2) andcommercial(CasaXPS,Origin)softwarepackageswereused.
ASPECSIQE12/38ionsourcewasusedforgenerationoflow- energyscattering(LEIS)spectra.He+ionsof800eVkineticenergy wereappliedatalowionfluxequalto0.03A/cm2,whichwasnec- essarytoavoidthesputteringofsurface.Theincidentangleswas 55◦(withrespecttosurfacenormal),ionsejectedalongthesurface normalweredetected.Theionenergies(LEIS)weremeasuredby thesamehemisphericalenergyanalyzerasmentionedabove.
2.Experiments related totheTiO2(110)single crystalwere conductedina separateUHVchamber.Thecrystal wasaprod- uctofPI-KEM.Itstemperaturecouldbechangedbetween150and 1000K.Thesamplewascleanedat1000K,applyingAr+ionsputter- ing(2.5A,1.5keV,10min)andvacuumannealing(5min)cycles.
ThefinaltreatmentwasAr+ionsputteringat300K(30min).Met- alsweredepositedfromane-beamevaporatorofOxfordApplied Research. The chamber had facilities for Auger electron spec- troscopy(AES),XPSandLEIS.ElectronsaswellasHe+ionswere detectedbyaLeyboldEA10/100hemisphericalanalyzer.During LEISmeasurements,theincidentanddetectionangleswere50◦ (withrespecttosurfacenormal),andthescatteringanglewas95◦.
AnAlK␣anodewasusedasanX-raysource,andthebindingenergy scalewasreferencedtothe4f7/2peakofa thickAulayersetto 83.8eV.
IRstudieswereperformedinahighvacuumsystem.Thesam- ples werepressed onto a Ta-mesh.The mesh wasfixedtothe bottomofaconventionalUHVsamplemanipulator.Itwasresis- tivelyheatedandthetemperatureofthesamplewasmeasured withaNiCr–Nithermocouplespotweldeddirectlytothemesh.IR spectrawererecordedwithaGenesis(Mattson)FTIRspectrometer.
ThewholeopticalpathwaspurgedbyaBalston75-62FTIRpurge generator.
DRIFTspectrawereregisteredduringtheethanoladsorption withBio-RadFTIRspectrometerwithwavenumber accuracyof
±4cm−1.Theinstrumentwasequippedwithadiffusereflectance attachment(Spectra-Tech)withBaF2windows.Typically32scans wereregistered.Thecatalystswerepretreatedasmentionedabove and the sample was cooled down to room temperature. Then ethanolwasintroducedintothecellbybubblingArgasthrough ethanolat273Kwhilethesamplewasheatedlinearlywithaheat- ingrateof10K/minupto573K.TheIRspectrawereregistered continuously.
Thecatalyticexperimentswerecarriedoutinafixedbedcon- tinuousflowreactor(8mmo.d.quartztube),which washeated externally.Thedeadvolumeofthereactorwasfilledwithquartz beads.Forcatalyticstudiessmallfragments(1–3mm)ofslightly compressedpelletswereused.Typically50mgof catalystswas used.Ethanolwasintroducedintothereactorbybubblingthecar- rier gas(Ar 80ml/min) throughliquid ethanol cooledto 273K.
Analysis of the products and reactants was performed with a Chrompack9001gaschromatographusingPorapackQScolumn.
ThegasesweredetectedsimultaneouslybyTCandFIdetectors.
3. Resultsanddiscussion
3.1. LEISexperimentsobtainedonapotassiumcontaining TiO2(110)surface
InourpreviouspapersitwasdemonstratedthatRhcore–Au shellclusterscanbepreparedonTiO2(110)ifAuispostdeposited byphysicalvapordeposition(PVD)ontheoxidesurfacecontaining Rhclusters[10,11].NoseparateAuclustersformedintheprocess atasubstratetemperatureof500K.TheexistingRhclustersact asnucleationcentersforgoldatomsdepositedsubsequently.The highdiffusionrateofAuatomsontheoxideduringevaporation probablyplaysanimportantroleinthisprocess.Earlier[11]we demonstratedthattheformationofcore–shellRh–Auparticlesdid notdependonthestateofreductionofthesurface:itproceeded bothonunreconstructedTiO2(110)andalsoonthemorereduced TiO2(110)−(1×2).
ThethermodynamicdrivingforcethatAuatomstendtobeout- sidewithinbimetallicclusterscomesfromthesignificantlylower surfacefreeenergyofAucomparedtoRh[44].Whenthereverse depositionsequence was applied,i.e. Rh wasdeposited onthe Au/TiO2(110)surface,averyefficientplaceexchangebetweenRh andAuatomswasobservedbyLEISevenatroomtemperature, movingRhatomsintosubsurfaceregionsofbimetallicclustersand theclustersurfaceremainedcoveredbygold[10].Thisindicates thatdiffusionprocesseswithinsmallmetalclustersareratherfacile comparedtobulkmaterials.
Consideringtheimportantroleofalkalipromotersincatalysis andthatalkalineatoms/ionsaretypicalconstituents/impuritiesin titanatenanotubesandnanowiresitseemedworthinvestigating whethertheformationofcore–shellRh–Auclustersproceedsona potassiumcontainingTiO2(110)surfaceaswell.
The potassium overlayer was prepared by thermal segrega- tion:annealingthecleanedsputteredsurfaceatT≥800K(5min)
Fig.1.AugerspectraobtainedafterannealingtheTiO2(110)surface–previously Ar+sputteredat300K– athighertemperaturesinUHVfor5min.
resultedintheappearanceofpotassiumonthesurfacedetected byAESandXPS.Generally,thepotassiumcontentofaTiO2crystal isexhaustedbysputtering-annealingcycleswithinafewweeks, butwiththissampletheKcoverageobtainedbythe5minanneal- ingat1000Kwasstableduringtheperiodoftheseexperiments (∼2month),characterizedbyaK/TiAESratioof0.4monitoring theKLMM(252eV)andTiLMM(387eV)peaks(Fig.1).Othertypical impurities(Ca,Na)werenotpresentonthissurface.
ThechemicalstateofthepotassiumcontainingTiO2(110)sur- faceobtainedafter5minannealingat1000KwasanalyzedbyXPS.
TheTi2p3/2 peakwasobservedat458.7eV(notshown),which istypicalforTi4+ionsin thetitaniasurface.Apparentlynodra- maticshiftsorincreasedreductionwasfoundduetothepresence ofpotassium.TheTi2pfeature,collectedatadetectionangleof16◦ (withrespecttosurfacenormal)containedasmallcontributionof Ti3+sites,withanareaofca.6.5%comparedtotheareaoftheTi4+
doublet.ThisisrathertypicalforausedTiO2(110)crystal,already coloreddarkblue.Thepeakpositionswereidentical,whentheTi2p regionwasmonitoredatagrazing(71◦)detectionangle,resulting inasmallerinformationdepth.However,theextentofreduction wassomewhatlarger,resultinginaTi3+componentwithanareaof ca10%oftheareacorrespondingtotheTi4+component.Evidently thisdifferenceisaclearsignthatmoredefectsitesexistinthetop surfaceregion.TheK2p3/2peak(notshown)wasfoundat294.0eV indicatingthatKatomsatthesurfacearepartiallyionized[45].
The possible formation of core–shell Rh–Au clusters on the potassiumcoveredTiO2(110)surfacewasinvestigatedbyLEIS.The scatteringangleinthatchamberwasrelativelysmall(95◦),which isadvantageousforthesensitivity,butresultsinalowerresolu- tion.Consequently,theLEIScontributionsofKandTiaremerged intoonepeak,becauseofthesimilarmassesoftheseelements.
First,0.5MLofRhwasevaporatedonK/TiO2(110)atasubstrate temperatureof330K.Annealingto500Kdidnotcauseanyappre- ciablechangemonitoredbyLEIS.Goldwassubsequentlydeposited at500KandthechangesintheLEISintensitiesasafunctionofAu coverageareshowninFig.2.
TheRhpeakdecreasedsteeplywithincreasingamountsofgold, whiletheintensitiesoftheOpeakandofthepeakcorresponding toKandTiwereonlymildlyattenuated.EvidentlyAuatomsevap- oratedonthesamplenucleatedonexistingRhclusters.IfonlyAu atomslandedontopofRhclustersduringevaporationhadbeen stabilizedontopofRh,thanthecompletedisappearanceofRhsig- nalwouldhaveoccurredapproximatelyatthecoverageof1ML orevenabove,consideringthecurvatureoftheclusters.However, theRhsignalinourmeasurementsdecreasedtozeroatamuch
Fig.2.ChangesinLEISintensitiesduetoAudepositionat500KontheK/TiO2(110) surfaceprecoveredwith0.5MLofRh.
smallercoverage(0.35ML).Thisindicatesthatamajorfractionof goldatoms–impingedonthefreeoxidesurfaceduringevaporation – migratedtoRhclustersandwasstabilizedthere,similarlyasit wasdemonstratedforthepotassium-freeTiO2(110)surfaceinour previouspapers[10,11].Atasubstratetemperatureof450Kana- logueresultswereobtained,buttheprocesswasnotinvestigated atevenlowerTvaluesinordertoavoidtheeffectofbackground COonLEISintensities.
3.2. Resultsobtainedontitanatenanowiresandnanotubes
ThesurfacecompositionofAu–Rhclustersontitanatenanocom- positewasalsoinvestigatedbyLEIS.Aswasobservedinthecase ofTiO2(110)substrate,withincreasinggoldcontenttheRhLEIS intensitydecreaseddramatically.Themostpregnantfeaturewas observedinthe0.5%Au+0.5%Rhcase.Onthemonometallicsys- temsthegoldandrhodiumHe+scatteringsignalsappearedat753 and707eV,respectively.Onbimetallicnanocomposite,however, onlythegoldsignalshowedup(Fig.3)[20].Therhodiumpeak wasjustabithigherthanthenoiselevel.Itisveryinterestingthat whilethegoldcontentisless(0.5%)inbimetallicsystem,theLEIS intensityofAuishigherthaninthemonometallic(1%)case.
OnthemonometallicRh/TiO2nanowirethedominantXPpeak forRh 3d5/2 appeared at307.1eVafter reductionat 573K(not shown).Thetailingatthehigherbindingenergysideofthisemis- sionmayconsistoftheoriginalasymmetryandtraceofforeign element(MgKLL Augerpeak)thatremainedfromtheprepara- tionprocess.Acarefuldeconvolutionrevealedsomeemissionat 309.3eV,presumablyduetomoredispersednanoparticles.Very similarrhodiumXPspectrawererecordedontitanatenanotubes.
Inthecaseofthe1%Au/TiO2nanowiretwopeakswereobserved onthereducedsampleforAu4f7/2at83.7eV(metallicstate)and 85.6eV.Theemissionat85.6eVcannotbeattributedtoakindof higheroxidationstate,becauseitdevelopedafterhydrogentreat- mentat 573K.We may attribute this featuretogold atoms in verysmallsizednanoparticles(“final-state”effect)[19,20].Under drasticconditionsofanoxygenplasmagoldoxideformationwas observed;theAu4f7/2wasobservedabove85eV[46].Wefound thatthefeatureat85.6eVappearedonlyafterreductionandnot afteroxidationsoweexcludetheoxidizedAuformation.
TheXPspectraofthebimetallicAu+Rhlayersupportedontita- niananowiresareshowninFig.4.Surprisingly,theemissionforthe higherenergypeakofAu4f7/2at85.6eVcorrespondingtotheatom- icallydispersedstateisverysmallinthepresenceofRh(Fig.4A), andatthesametimetheemissionforRh3d5/2ataround309.3eV
780 760
740 720
700 680
Rh
Rh Au
Kinetic energy [eV]
100 a.u.
a
b
c LEIS spe ctra t aken on titanate na nowire
Au
Fig.3. LEISspectraof1%Au/TiO2(b),1%Rh/TiO2nanowire(c),0.5%Au+0.5%
Rh/TiO2nanowire(a)[20].
alsodiminished.TheadsorptionofCOat300Kdidnotalterthe positionofgoldemissions.TheRhXPSfeatureat308.9–309.3eV appearedorintensifiedalittlebitafter60minCOadsorptionmay beduetotheappearanceofmoredispersedparticles(Fig.4B).The COpressurewas1.3mbarinthepreparationchamber.
The morphology of Au, Rh and Au+Rh supported on TiO2 nanowires and nanotubes wasinvestigated by FTIR,employing adsorbedCO asa probe moleculesensitivetothe localsurface structure. AdsorbedCO exhibitsatleastthree differentstretch- ing frequencies belonging tocertain adsorption sites of Rh on oxidesupports[46].Thebandat2070–2030cm−1 isdue toCO adsorbedlinearlytoRh0(depending onthecoverage),theband at∼1855cm−1representsthebridgebondedCO(Rh2–CO),andthe featureat∼2100cm−1andat∼2020cm−1correspondstothesym- metricandasymmetricstretchingsofRh+(CO)2(twinCO).These latterIRsignalsweredetectedwhenthecrystallitesizewasvery small[47].COadsorptionongoldsupportedontitaniananostruc- turewasnotobserved.OnmonometallicRhsupportedonnanowire practicallythetwinformwaspresent(2027and2097cm−1),the linearformbetweenthesetwopeakswasmuchsmallerandthe bridgeformwashardlyobservable(Fig.5A).Onnanotubesthelin- earlyadsorbedCOfeaturesshowedupat2075cm−1betweenthe peaksat2100and2036cm−1(twinform)(notshown).Fromthese IRstudieswemayconcludethatasignificantpartofRhexistsin smallparticlesizesonbothnanowiresandtubes.
Itisplausibleifthegoldcompletelyanduniformlycoversthe Rhnanoparticles(core–shellstructure)adsorbedCOcouldnotbe detectedbyFTIR.However,thesituationisthatrelativelystrongCO bandsappearedat300Katapressureof1.3mbar.Inthepresence
80 82 84 86 88 90 92
Binding energy [eV]
200 cps
Binding energy [eV]
B A
83.7
85.6
H
2306 308 310 312 314
307.0
H
2200 cps
80 82 84 86 88 90 92
85.6
83.7
CO CO
200 cps
306 308 310 312 314
Binding energy [eV]
Binding energy [eV]
200 cps
307.0
308.9
Fig.4. XPspectraofAu4f(A)andRh3d(B)of0.5%Au+0.5%Rh/TiO2nanowirebeforeandafterCOtreatment.TheCOpressurewas1.3mbar.
ofgold,thepeakcorrespondingtothelinearformbecamestronger at 2070–2071cm−1 and the twin CO stretching frequencies decreased(Fig.5B).Seeminglythere isa contradictionbetween theresultsofLEISandCOadsorptioninfraredexperiments.Onthe topmostlayertherearenoRhatoms(Fig.3),butadsorbedCOwas detectedbyFTIRonthissurface(Fig.5B).Thisdiscrepancycanbe explainedbyaCOinducedsurfacereconstruction.Theadsorption ofCOonAu–RhclustersmaypromotethediffusionofRhtothe surfaceofthecluster.Similarphenomenawasobservedrecentlyin thecaseofbimetallicPt–AuclustersonTiO2(110)[14].Thesame effect was also observed on Pd–Au bimetallic model catalysts, synthesizedeitherasthinfilmsonMo(110)orasnanoparticleson TiO2thinfilminCOoxidationatelevated(8–16Torr)COpressures [48].Pdpreferentiallysegregatestothesurfacetoformcontiguous PdsitesandCO oxidationreactivityisregained.Thedifferences obtainedonFTIRspectrabetweenmonometallicRhandbimetallic Au–Rhnanoclustersontitanatenanowirearenotsurprising.The presence of the core–shell composites and theseparate highly dispersedgold clustersmay significantlyaffectthemorphology
andelectronicstructureofRhnanoparticlesthereforethebonding modesofCOadsorbedonRhsites.TheobservedFTIRfeaturemay be related to theXPS result detected afterCO adsorption.The morphologyandcrystallitesizeofsegregatedRhcouldbedifferent in surrounding of Au–Rh core–shell and in the inside of the core–shellstructure.
InordertogetclosertotheunderstandingofCO-inducedmor- phologicalchangessomelowenergyionscatteringmeasurements (LEIS)werecarriedoutbeforeandafterCOadsorptiononAu–Rh bimetallicclustersupportedontitanatenanotubes.Intheseexper- imentstwocompositionswereused:0.25%Au+0.75%Rh/TiO2and 0.5%Au+0.5%Rh%/TiO2nanotubes.Theresultsarerepresentedin Fig.6.InordertoavoidorminimizethesputteringeffectoftheHe ionsweappliedrelativelylowionflux(0.03A/cm2).Atthelower goldcontentrhodiumscatteringappearedat707eVafterhydro- gentreatmentat573K.AfterCOadsorptionat300Kitsintensity increasedbyabout20%(Fig.6A).TheeffectofadsorbedCOwas morepronouncedinthe0.5%Au+0.5%Rhcase.Practicallynoor verysmallRhscatteringsignalwasdetectedindicating thatthe
1800 1950 2100 2250
2097 2027
1845
Wavenumber (cm-1) 0.1
1800 1950 2100 2250
0.05
Wavenumber (cm-1) 1843 2033 2070 2098
B A
Fig.5. FTIRspectraofadsorbedCOat300K:(A)1%Rh/TiO2nanowireand(B)0.5%Au+0.5%Rh/TiO2nanowire.TheCOpressurewas1.3mbar.
goldcoveredtherhodiumparticles(Fig.6B),similarlythenanowire case(Fig.3).WhenCOwasaddedtothebimetalliccomposite,the rhodiumionscatteringsignalshowedup.Theseexperimentaldata stronglysupporttheabovementionedphenomena,namelythatCO inducesareconstructionofthesurface.Rhpreferentiallysegregates tothesurfacetoformRh–CObond.
ThisprocessmayimplythemigrationofRhorAuatomswithin theclustersoronthesurfaceoftheclusters.ConsideringthatAu–Au andAu–Rhbondsareweaker thanRh–Rhbonds,itmayappear moreprobablethatthemovementofAuatomsismoreimportant
780 760 740 720 700 680
R h
K inetic energy [eV ] K inetic energy [eV ]
B A
200 a.u.
C O
H2
C O
H2 A u
780 760 740 720 700 680
R h
A u
200 a u.
Fig.6. LEISspectraof0.25%Au+0.75%Rh/TiO2nanotube(A),0.5%Au+0.5%Rh/TiO2
nanotube(B)beforeandafterCOadsorptionat300K.TheCOpressurewas1.3mbar.
(theshell“opens”), whileRhatoms moveless. Theinfluenceof COgasinthispossibleprocessisanimportantquestion.Though theAu–COinteractionisratherweakcomparedtoRh–CObond,in therelativelyhighCOpressureapplied,theCOcoverageonthe Aucappinglayerisprobablysignificant.TheadsorbedCO layer presumablydecreasesthemetal-metalbondsbothforAuandRh, leadingtoanincreaseddiffusionrate.ItisalsopossiblethatCO, adsorbedinsomehollowsiteoftheAulayercanformsomekindof bondwithsubsurfaceRhatomsandthisinteractionpromotesthe displacementofAuatomsandthesegregationofRh.
Anotherpossiblescenarioisthatthereisacontinuousthermal fluctuationofmetalatomswithinthebimetallicclustersindepen- dentofthepresenceofCO,andfor shortperiodsRhatomscan befoundontheclustersurface,whicharesubsequentlytrapped byCO.Theexistenceofthisfluctuationmaybepossible,sincesig- nificantmovementofmetalatomswithinsmallclustersatroom temperaturewasobservedinseveralcases[10,49,50]characterized bydiffusioncoefficientsmanyordersofmagnitudehigherthanthe correspondingbulkvalues.Anyhow,furtherstudiesareneededto understandindetailtheCOinducedrestructuringofRh–Auclus- ters.
3.3. Ethanoladsorptionontitanatenanowiresupportedcatalysts
Duringethanoladsorptionatroomtemperaturepredominantly thecarbon–carbonbonddissociatesandthecarbon–oxygenbond ispreservedleadingtoadsorbedCObutnotoadsorbedatomicOon Rh(111)[51]surfaces.InadditiontoCO,alsosignificantamounts ofmethylidyneandethylidynespeciesareformedonRh(111),the latterpresumablyviaareformingreaction.GongandMullins[52]
reportedthatonAu(111)ethanoladsorbsonlyweaklyanddesorbs molecularly.EthanoladsorptionhasbeenextensivelystudiedbyIR spectroscopyoverTiO2andTiO2supportedmetals[53,54]among othersonRh/TiO2[55]andAu/TiO2[39],too.
Fig. 7 shows the DRIFT spectra of ethanol adsorbed on 1%
Au/TiO2,0.5%Au+0.5%Rh/TiO2and1%Rh/TiO2catalystatdifferent temperatures.Similarspectrawereregisteredinallcasesexcepting theCOregion.Theseresultssupportthatthesespeciesadsorbed onthe nanowire. At 300K absorption bands were observed at 2974,2929and2875cm−1 in theC–Hstretchingregion.Inthe lowfrequencyrangeabsorptionbandsweredetectedat1448–1450
1500 2000
3000 0.0
0.5 1.0 1.5 2.0
2036
145014481448 114211401144
138313831384
16331633
2929 2875
2974
2038
2028
Absorbance (a.u.)
Wavenumber (cm-1) 1
2 3 4 5 6 7 8 9 10 11 12
2028
2974 2929 2875
2974 2929 2875 1633
1847
Fig.7.Infraredspectraofethanoladsorbedatroomtemperature(1,5,9),at373(2, 6,10),423(3,7,11),473K(4,8,12)on1%Au/TiO2nanowire(1–4),0.5%Au+0.5%
Rh/TiO2nanowire(5–8),and1%Rh/TiO2nanowire(9–12).
and 1383–1384cm−1, both could be assigned as ␦CH3 vibra- tionsofethanol.Thebandsobservedat1140–1144,1124–1121, 1069–1074, 1045–1047cm−1 could be attributed to C–O and C–C vibrations of monodentate and bidentate ethoxide species [39,55].Theintensitiesofthesebandsdecreasedasthetemperature increasedbutdramaticchangeswerenotdetected.
HerewefocusontheCOregionwheresignificantdifferences wereobserved.OnAu/TiO2wedidnotobserveanyabsorbance between2100and1800cm−1 duringethanoladsorptionaswas foundafterCO adsorption [20]. WhentheRh/TiO2 wastreated with ethanol above 373K new peaks were detected at 2028 and 1847cm−1. The intensities of these bands increased with increasingthetemperatureandthepeakobservedat2028cm−1 at room temperature shifted to higher wave numbers. The absorbanceat 1847cm−1 could beattributedto bridgebonded COonRhsites[47,55].ThefeatureoftheadsorbedCO detected at2028–2038cm−1issignificantlydifferingfromthatdepictedin Fig.5. Thisdiscrepancy couldbe explained bythe Rhcarbonyl hydrideformation(H–Rh–CO)[55,56].
When0.5%Au+0.5%Rh/TiO2 wasthecatalysttooursurprise nearlythesamespectrawereobservedasinthecaseof1%Rh/TiO2 onlytheintensitiesofthebandswereweaker;at373Kasmallpeak wasdetectedatabout2028cm−1attributedtoadsorbedCO.
AccordingtotheXPSresultsbindingenergiesofRhandAudid notalterduringtheethanoladsorptionevenathighertempera- ture.TheC1sspectrashowedformationofdifferentcarbonoxygen containingspecies(∼289and286.3eV).
120 100
80 60
40 20
0 0 5 10 15 35 40 45 50
Conversion [%]
Time [min]
Fig.8. Theconversionofethanolinthedecompositionreactionat603KonTiO2
nanowire(),1%Au/TiO2nanowire(),0.5%Au+0.5%Rh/TiO2nanowire()and1%
Rh/TiO2nanowire(䊉).
3.4. Ethanoldecompositionontitanatenanowiresupported catalysts
Theethanol decompositionwasinvestigatedat603KonAu, Rhand Au–Rhsupportedontitanatenanowire.Themainprod- ucts were in all cases carbon monoxide, methane, hydrogen, acetaldehyde, but ethylene, diethyl ether and acetic acid were alsodetectable.Theethanolconversionontitanatenanowirewas slightlyhigher(7.1%)thanonAu/TiO2(6%)(Fig.8)butinthelat- tercasetheacetaldehydeselectivitywasmorethantwicehigher (37%) than ontheclean support.Earlierit wasalsofoundthat thesupportedAuparticlescatalyzemainlythedehydrogenation ofethanol,toproducehydrogenandacetaldehyde[37].Whenthe catalystcontainsRhtheconversionsignificantlyincreased;on0.5%
Au-0.5%Rh/TiO2itwas12.2%andonRh/TiO2theconversionwas aboutfivetimeshigher(36.9%)thanonAu/TiO2(Fig.8).Theproduct distributionsweresimilaronbothsamples,mainlyCOwasformed theacetaldehydeselectivitywasonly4–6%.Itshouldbenotedthat inthesecasestheethanolconsumptiondecreasedintimewhilein othercasesitwasstable.
FromDRIFTandcatalyticmeasurementsweconcludethatgold doesnotblockentirelytheadsorptionandtheactivesitesofRhina core–shellstructure.TheethanolortheCOproducedintheethanol decompositionmayalsoinducethesegregationofRhandthe0.5%
Au+0.5%Rhsupportedontitanatenanowireactsasaneffective catalystinethanoldecomposition.
4. Conclusions
1.Au and Rh clusters, as well as Au–Rh bimetallic nanoparti- cleswerepreparedontitanatenanowires,nanotubesand on TiO2(110).ThesampleswerecharacterizedbyX-rayphotoelec- tronspectroscopy(XPS),lowenergyionscatteringspectroscopy (LEIS)and Fourier transforminfraredspectroscopy (FTIR).By performing careful LEIS experiments, it was found that, for appropriate Au and Rh coverage, the Au almost completely covers theRh nanoparticles. Rhcore–Au shell structure was detected.Theformationofthisstructure wasnot affectedby alkali(K)adatoms.LEISandFTIRshowedthatadsorbedCOat 300KcausesthesegregationofRhtothesurfaceinordertobind toCO.
2.Rhsupportedontitanatenanowireexhibitahighcatalyticactiv- ityinethanoldecomposition.Inbimetallicformitpreservesits activitybecausetheethanolsimilarlytoCOmayalsoinducethe segregationofRhfromtheAu-shellRh-corestructure.
Acknowledgements
The financial supports of the Hungarian Scientific Research Fund (OTKA) through projects K69200, K76489 and TÁMOP- 4.2.1/B-09/1/KONV-2010-0005are acknowledged.Wethankthe preparation of titanate nanocomposites to Ákos Kukovecz and AndrásSápi.
References
[1]G.C.Bond,C.Louis,D.T.Thompson,CatalysisbyGold,ImperialCollegePress, London,2006.
[2]M.Haruta,GoldBull.37(2004)27.
[3]G.Hutchings,J.Chem.Commun.(2008)1148.
[4]J.A.Rodriguez,Surf.Sci.Rep.24(1996)23(andreferencestherein).
[5] J.G.Chen,C.A.Menning,M.B.Zeller,Surf.Sci.Rep.63(2008)201(andreferences therein).
[6]A.Gross,Top.Catal.37(2006)29.
[7]L.Guczi,Catal.Today101(2005)53(andreferencestherein).
[8] L.Guczi,G.Boskovic,E.Kiss,Catal.Rev.Sci.Eng.52(2010)133(andreferences therein).
[9]L.Bugyi,A.Berkó,L.Óvári,A.M.Kiss,J.Kiss,Surf.Sci.602(2008)1650.
[10]L.Óvári,L.Bugyi,Zs.Majzik,A.Berkó,J.Kiss,J.Phys.Chem.C112(2008)1811.
[11]L.Óvári,A.Berkó,N.Balázs,Zs.Majzik,J.Kiss,Langmuir26(2010)2167.
[12] H.L.Abbott,A.Aumer,Y.Lei,C.Asokan,R.J.Meyer,M.Sterrer,S.Shaikhutdinov, H.-J.Freund,J.Phys.Chem.114(2010)17099.
[13]J.B.Park,S.F.Conner,D.A.Chen,J.Phys.Chem.C63(2008)5490.
[14]S.A.Tenny,J.S.Ratliff,C.C.Roberts,W.He,S.C.Ammal,A.Heyden,D.A.Chen,J.
Phys.Chem.C114(2010)21652.
[15]Z.Li,F.Gao,O.Furlong,W.T.Tyose,Surf.Sci.64(2010)136.
[16]K.Luo,T.Wei,C.W.Yi,S.Axnada,D.W.Goodman,J.Phys.Chem.B109(2005) 23517.
[17]D.R.Rainer,C.Xu,P.M.Holmblad,D.W.Goodman,J.Vac.Sci.Technol.15(1997) 1653.
[18]J.Kiss,R.Németh,Á.Koós,J.Raskó,J.Nanosci.Nanotechnol.9(2009)3828.
[19]A.Oszkó,G.Pótari,A.Erd ˝ohelyi,A.Kukovecz,Z.Kónya,I.Kiricsi,J.Kiss,Vacuum 85(2011)114.
[20] A.Kukovecz,G.Pótari,A.Oszkó,Z.Kónya,A.Erd ˝ohelyi,J.Kiss,Surf.Sci.605 (2011)1048.
[21]Y.N.Xia,P.D.Yang,Y.G.Sun,Y.Y.Wu,B.Mayers,B.Gates,Y.D.Yin,F.Kim,H.Q.
Yao,Adv.Mater.15(2003)353.
[22]F.Cesano,S.Bertarione,M.J.Uddin,G.Agostini,D.Scarano,A.Zeccina,J.Phys.
Chem.C114(2010)169.
[23]M.Hodos,E.Horváth,H.Haspel,A.Kukovecz,Z.Kónya,I.Kiricsi,Chem.Phys.
Lett.399(2004)512.
[24]G.K.Mor,K.Shankar,M.Paulose,O.K.Varghese,C.A.Grimes,NanoLett.5(2005) 191.
[25]S.Kubota,K.Johkura,K.Asanuma,Y.Okouchi,N.Ogiwara,K.Sasaki,T.Kasuga, J.Mater.Sci.:Mater.Med.15(2004)1031.
[26]L.Kavan,M.Kalbac,M.Zukalova,I.Exnar,V.Lorenzen,R.Nesper,M.Graetzel, Chem.Mater.16(2004)477.
[27]D.V.Baviykin,A.A.Lapkin,P.K.Plucinski,J.M.Friderich,F.C.Walsh,J.Phys.Chem.
B109(2005)19422.
[28]R.Huang,F.Chung,E.M.Kelder,J.Electrochem.Soc.153(2006)459.
[29] S.S.Malwadkar,R.S.Gholap,S.V.Awante,P.V.Korake,M.G.Chaskar,N.M.Gupta, Photochem.Photobiol.A:Chem.203(2009)24.
[30] V.Idakiev,Z.Yuan,T.Tabakova,B.-L.Su,Appl.Catal.A:Gen.281(2005)149.
[31]T.Akita,M.Okumura,K.Tanaka,K.Ohkuma,M.Kohyama,T.Koyanagi,M.Date, S.TsubotaS,M.Haruta,Surf.InterfaceAnal.37(2005)265.
[32]A.Erd ˝ohelyi,J.Raskó,T.Kecskés,M.Tóth,M.Dömök,K.Baán,Catal.Today116 (2006)367.
[33] J.Raskó,M.Dömök,K.Baán,A.Erd ˝ohelyi,Appl.Catal.A:Gen.299(2006)202.
[34]M.Scott,M.Goeffroy,W.Chiu,M.A.Blackford,Top.Catal.51(2008)13.
[35]M.Dömök,A.Oszkó,K.Baán,I.Sarusi,A.Erd ˝ohelyi,Appl.Catal.A:Gen.383 (2010)33.
[36]R.M.Navarro,M.A.Pe ˜na,J.L.G.Fierro,Chem.Rev.107(2007)3952.
[37]A.Gazsi,A.Koós,T.Bánsági,F.Solymosi,Catal.Today160(2011)70.
[38]P.-Y.Sheng,G.A.Bowmaker,H.Idriss,Appl.Catal.A:Gen.261(2004)171.
[39]M.A.Nadeem,M.Murdoch,G.I.N.Waterhouse,J.B.Metson,M.A.Keane,J.Llorca, H.Idriss,J.Photochem.Photobiol.A:Chem.216(2010)250.
[40]Y.Guan,E.J.M.Hensen,Appl.Catal.A:Gen.361(2009)49.
[41]M.J.Lippits,B.E.Nieuwenhuys,J.Catal.274(2010)142.
[42] E.Horváth,Á.Kukovecz,Z.Kónya,I.Kiricsi,Chem.Mater.19(2007)927.
[43]Á.Kukovecz,M.Hodos,E.Horváth,Gy.Radnóczi,Z.Kónya,I.Kiricsi,J.Phys.
Chem.B109(2005)17781.
[44]L.Z.Mezey,J.Giber,Jpn.J.Appl.Phys.21(1982)1569.
[45] J.Kiss,K.Révész,F.Solymosi,Surf.Sci.207(1988)36.
[46]L.K.Ono,B.R.Cuenya,J.Phys.Chem.C112(2008)4676.
[47]F.Solymosi,M.Pásztor,J.Phys.Chem.89(1985)4789.
[48]F.Gao,Y.Wang,D.W.Goodman,J.Phys.Chem.C114(2010)4036.
[49] O.Ozturk,J.B.Park,S.Ma,J.S.Ratliff,J.Zhou,D.R.Mullins,D.A.Chen,Surf.Sci.
601(2007)3099.
[50]R.Ferrando,J.Jellinek,R.L.Johnston,Chem.Rev.108(2008)845.
[51]A.Resta,J.Gustafson,R.Westerström,A.Mikkelsen,E.Lundgren,J.N.Andersen, M.M.Yang,X.F.MabXF,X.H.Bao,W.X.Li,Surf.Sci.602(2008)3057.
[52] J.Gong,B.Mullins,J.Am.Chem.Soc.130(2008)16458.
[53]G.A.M.Hussein,N.Sheppard,M.I.Zaki,R.B.Fahim,J.Chem.Soc.FaradayTrans.
87(1991)2661.
[54] L.Óvári,F.Solymosi,Langmuir18(2002)8829.
[55] J.Raskó,A.Hancz,A.Erd ˝ohelyi,Appl.Catal.A:Gen.269(2004)13.
[56]F.Solymosi,A.Erd ˝ohelyi,M.Kocsis,J.Catal.65(1980)428.