JPhysChemB_103_5787_1999

StructureandDensityofMoandAcidSitesinMo-ExchangedH-ZSM5CatalystsforNonoxidativeMethaneConversion

RichardW.BorryIII,YoungHoKim,AnneHuffsmith,JeffreyA.Reimer,andEnriqueIglesia*

MaterialsSciencesDiVision,E.O.LawrenceBerkeleyNationalLaboratory,andDepartmentofChemicalEngineering,UniVersityofCalifornia,Berkeley,California94720ReceiVed:March12,1999;InFinalForm:May7,1999

Mo/H-ZSM5(1.0-6.3wt%Mo;Mo/Al)0.11-0.68)catalystsforCH4aromatizationwerepreparedfromphysicalmixturesofMoO3andH-ZSM5(Si/Al)14.3).X-raydiffractionandelementalanalysisofphysicalmixturestreatedinairindicatethatMoOxspeciesmigrateontotheexternalZSM5surfaceatabout623K.Between773and973K,MoOxspeciesmigrateinsidezeolitechannelsviasurfaceandgasphasetransport,exchangeatacidsites,andreacttoformH2O.TheamountofH2OevolvedduringexchangeandtheamountofresidualOHgroupsdetectedbyisotopicequilibrationwithD2showedthateachMoatomreplacesoneH+duringexchange.Thisstoichiometryandtherequirementforchargecompensationsuggestthatexchangedspeciesconsistof(Mo2O5)2+ditetrahedralstructuresinteractingwithtwocationexchangesites.TheproposedmechanismmayprovideageneralframeworktodescribetheexchangeofmultivalentcationsontoAlsitesinzeolites.AstheMoconcentrationexceedsthatrequiredtoformaMoOxmonolayerontheexternalzeolitesurface(∼4wt%MofortheH-ZSM5used),Mospeciessublimeas(MoO3)noligomersorextractAlfromthezeoliteframeworktoforminactiveAl2(MoO4)3domainsdetectableby27AlNMR.These(Mo2O5)2+speciesreducetoformtheactiveMoCxspeciesduringtheinitialstagesofCH4conversionreactions.OptimumCH4aromatizationrateswereobtainedoncatalystswithintermediateMocontents(∼0.4Mo/Al),becausebothMoCxandacidsitesarerequiredtoactivateCH4andtoconverttheinitialC2H4productsintoC6+aromaticsfavoredbythermodynamics.

Introduction

Theselectiveconversionofnaturalgastohigherhydrocar-bonsandaromaticsremainsanimportantindustrialchallenge.ThediscoveryofcatalyticCH4aromatizationonMo/H-ZSM51wasfollowedbyseveralreportsofnear-equilibriumCH4conversionsat973Kwithhighselectivitytobenzene.2-4CH4reactionratesincreasedwithtimeonstreamasdispersedMoOxformedtheactiveMoCxspecies.3EthyleneandethaneareformedonMoCxsitesasprimaryproductsandtheyconverttoC6+aromaticsviaoligomerization,cracking,andcyclizationreactions.Thesereactionsrequirechaingrowthanddehydro-genationstepsthatoccuronBronstedacidsitesaidedbyhydrogendesorptionsitesprovidedbyMoCxspecies.ThenetreactionrateislimitedbytherateofCH4activationandbytheapproachtothermodynamicequilibriumoftheoverallmethanearomatizationreaction.3

Inpreviousstudies,Mo/H-ZSM5catalystswerepreparedbyslurryorincipientwetnessimpregnationofH-ZSM5withaqueousammoniumheptamolybdate(AHM;(NH4)6Mo7O24)solutions,followedbytreatmentinairat723-973K.1-4LargeaqueousmolybdateionsdonotexchangedirectlyontoH-ZSM5cationexchangesitesduringimpregnation.5Infraredspectros-copyanddifferentialthermalanalysis6showedthatAHMdecomposedinairbetween500-650KtoformMoO3crys-tallitesontheoutersurfaceofzeolitecrystals.7,8At773K,MoO3infraredbandsdisappearedasMoO3crystallitesdispersedontheexternalsurfaceofZSM5crystalsandthenmigratedas

*Correspondingauthor.E-mail:iglesia@cchem.berkeley.edu.

(MoO3)noligomersintozeolitechannels.6Theseauthorscon-cludedthatMoOxinteractswithframeworkoxygensinH-ZSM5at973Ktoformisolatedmonomolybdatespecies(MoO42-)basedontheappearanceofinfraredbandsforModOvibrationsintetrahedralMo6+.

SurfacemigrationofMoOxontoH-ZSM5canoccurathightemperaturesbecauselatticemobilitywithinMoO3becomespossibleaboveitsTammanntemperature(534K).9SublimationofMoO3becomesdetectableabove623-673K,10andMoO3reachesavaporpressureof56Paat973K11as(MoO3)noligomers(n)2-5).12H2OformedduringdecompositionofAHMprecursorscanformMoO2(OH)2,12,13whichhasavaporpressureof4.9Paat973K.11

AsisolatedMoOxspeciesmigrateintozeolitechannelsviagasphaseorsurfacediffusion,theyreactwithH+atomsatexchangesitestoform(MoO2(OH))+species,whichcancondensewithanotheronetoforma(Mo2O5)2+dimerandH2O:

Throughoutthisreport,werefertothesedimersas(Mo2O5)2+inspiteofthefactthatsuchcationicspeciesarenotknownin

10.1021/jp990866vCCC:$18.00©1999AmericanChemicalSociety

PublishedonWeb06/25/1999

5788J.Phys.Chem.B,Vol.103,No.28,1999

solution,becauseitiscustomarytothinkofexchangedspeciesinzeolitesascationic.Thestablestructuralanalogistheknown(Mo2O7)2-anion,consistingoftwoMocenterswithtetrahedralsymmetry,withtwooftheoxygenatomsresidingatframeworkpositionsinZSM5.(MoO2(OH))+speciescanalsoreactwithazeoliteOHgrouptoforma(MoO2)2+cationbridgingtwoacidsitesandwater,astructurepreviouslyproposedfortheexchangeofMoO2Cl2ontoH-Yzeolite.14

SimilarcondensationreactionsofOHgroupsleadtotheextractionofAlionsfromtheframework,withtheformationofwaterandthedisappearanceoftwoBronstedacidsites.InMo/H-ZSM5,extraframeworkAlatomsformsmalldomainsofAl2O3orAl2(MoO4)3;thelatterspeciesweredetectedby27AlNMRinMo/H-ZSM5sampleswithhighMocontent(>10wt%).15

ImpregnationorionexchangeofH-ZSM5withAHMsolutionsleadsinitiallytoexternalMoO3crystalsduringtreatmentinair.Thus,itseemedpossibletoprepareMo/H-ZSM5viaexchangefromintimatemixturesofMoO3andH-ZSM5powders.ThisapproachavoidsAHMdecompositionproducts(N2,NH3,andH2O)6,whichinterferewithmeasure-mentsofthekineticsandtheextentofexchangefromtheamountofwaterevolvedduringsynthesis.Here,wereportmechanisticdetailsofthesynthesisofMoOx/H-ZSM5frommixturesofMoO3andH-ZSM5powders,thestructureoftheMoOxspeciesformedduringexchange,andtheroleanddensityofMoOxspeciesandofBronstedacidsitesinalkanereactions.ExperimentalSection

CatalystSynthesisandCharacterization.H-ZSM5waspreparedbyaqueousexchange(repeatedfourtimes)ofNa-ZSM5(Zeochem,Si:Al)14.3)with1.0MNH4NO3(Fisher)(∼10gNa-ZSM5/L).Thesamplesweredriedat400Kfor24handtreatedindryairat773Kfor24h(10-15gofcat.,50cm3/min).Mo/H-ZSM5waspreparedfromphysicalmixturesofMoO3(JohnsonMatthey,99.5%purity)andH-ZSM5;thesemixturesweregroundtogetherforabout0.1hinanaluminamortarandpestle.PowderX-raydiffractionpatternswereobtainedinaSiemensDiffractometerD5000usingCuKRradiation(λ)1.5406Å).SurfaceareasandporevolumeswereobtainedfromN2physisorptionatitsboilingpoint(Autosorb6;Quantachrome,Inc.),aftertreatingsamplesat773Kfor1hinflowingdryair(Medicalgrade,Praxair).

TheeffectofairtreatmentwasexaminedusingMoO3/H-ZSM5mixtures(0.3g,0-8wt%Mo)driedat623Kfor24hin20%O2/Ar(100cm3/min,Praxair,>99.999%).Sampleswereheatedat10K/minto973KandH2Oevolutionrateswere

Borryetal.

measuredbymassspectrometry(LeyboldInficon,THP-TS200)usingheatedtransferlines(373K)andArasaninternalstandard.Sampleswereheldat973Kfor0.5handthencooledto300K.ThenumberofresidualOHgroupsinMo/H-ZSM5wasmeasuredfromtheevolutionofHDandH2bymassspectrometryassampleswereheatedfrom300to873K(10K/min,2minholdtime)in5%D2/Ar(100cm3/min,Praxair,>99.999%).16,17Sampleswerethencooledto300Kandtheexperimentwasrepeatedusing5%H2/Ar(100cm3/min,Praxair,>99.999%)whilemeasuringHDandD2evolutionrates.TheMocontentwasmeasuredbyatomicabsorption(GalbraithLaboratories)aftertreatmentinairat973K.

27AlNMRspectrawerecollectedusingahome-built400MHzspectrometer(9.1Tmagneticfield)at104.2MHzwhilespinningat4kHz.18Sampleswerehydratedatroomtemperaturebyplacingtheminanemptydesiccatorcontainingliquidwater.HydrationofAlsitesweakensquadrupoleinteractionsthatbroaden27AlNMRlinesinH-ZSM5.19,20

Steady-StateCH4AromatizationonMo/H-ZSM5.CH4reactionswerecarriedoutat950Kinatubularreactor(4mmi.d.)withplug-flowhydrodynamics(25cm3/min,1:1CH4/Ar(Praxair,>99.995%)).Mo/H-ZSM5catalysts(1.0g)wereloadedontoaporousquartzdisk(1.0cmi.d.)locatedwithinthisreactor.TemperaturesweremeasuredwithatypeKthermocouplelocatedinsideaquartzsheathincontactwiththecatalystbed.Thereactoreffluentwassampledusingtransferlinesheldat400K,21andthereactantandproductconcentrationsweremeasuredbygaschromatography.22ReportedselectivitiesaredefinedasthepercentageoftheCH4convertedappearingasagivenproduct.CH4wasusedasthereferencepeakbetweentheflameionizationandthermalconductivitydetectorsandArwasusedasaninternalstandard.Theamountofcarbonmissingwithinthemeasuredproducts(1-10%)isreportedas“carbon”intheresults,butitincludescarbonconsumedtoformMoCxandcondensableproductsremaininginzeolitechannelsorintransferlines.ThereportedvaluesforCH4conversion,hydro-carbonyields,andcarbonbalanceswerereproduciblewithin(0.5%(absolute),andthedetectionlimitswere50and1ppmforthermalconductivityandflameionizationdetectors,respec-tively.Catalystsweretreatedin20%O2/He(100cm3/min,Praxair,>99.999%)at950Kfor2hbeforecatalyticreactions.23Results

CatalyticCH4AromatizationonMo/H-ZSM5.CH4reac-tionratesandselectivitiesmeasuredonMo/H-ZSM5(4.0wt%)preparedfromMoO3/H-ZSM5mixtures(Figure1)weresimilartothosereportedonsamplespreparedbyimpregnationofH-ZSM5withaqueousAHM(Table1).3Duringtheinitialactivation(∼1-2h),CH4conversionreachedabout20%,butcarbon,H2,CO2,H2O,andCOwerethemostabundantinitialproducts.After2h,COandCO2werenolongerdetectableandthecarbonselectivitywasbelow7%.Benzene(∼70%selectivity),toluene(∼5%),ethylene(4-8%),ethane(2-4%),andnaphthalene(10-20%)wereformedatCH4conversionsof7-10%.Propeneando-xylenewerealsodetectedintraceamounts(<2%selectivity).Bycomparison,aMo/H-ZSM5(2wt%)samplepreparedbyimpregnationofH-ZSM5withaqueousAHM3gave10%CH4conversionand70%selectivitytobenzeneat973K(Table1).

MoO3MigrationandSublimationduringCatalystSyn-thesis.OurinitialmeasurementsoftherateofMoOxmigrationfromtherateofH2OevolutionduringtreatmentinairwereinfluencedbyMoO3sublimationfromthephysicalmixtures.HeatingMoO3/H-ZSM5mixtures(4.0wt%Mo)from300to

Mo-ExchangedH-ZSM5CatalystsJ.Phys.Chem.B,Vol.103,No.28,19995789

TABLE1:ComparisonofCH4AromatizationonMo/H-ZSM5PreparedfromAqueousImpregnationofMolybdateSolutiononH-ZSM5andPhysicalMixtureofMoO3/H-ZSM5

Mocontent(wt%)

2.0a4.0b4.3e

a

Mo/Alatomicratio

0.370.420.45

reactiontemp(K)973950950

GHSV(h-1)800750d750d

CH4

conversion(%)

10.010.21.7

selectivity(%)

C2H43217

C6H6706627

C10H820184

solidcarbon51252

IncipientwetnessimpregnationofAHM/H-ZSM5.3bPhysicalmixtureofMoO3/H-ZSM5(thisstudy).c“Vapor-exchanged”preparationmethod(thisstudy,Figure6).dSpacevelocitywascalculatedassumingcatalystpackeddensityof0.5gcat./cm3reactorvolume.

Figure1.CH4pyrolysison4wt%Mo/H-ZSM5,(a,top)CH4conversion,C2-C10hydrocarbonproductyield,andsolidcarbonyieldvs.timeonstream,(b,bottom)carbonselectivityofgas-phaseproductsvstimeonstream(1.0g,25cm3/min,50%CH4/Ar,105kPa,950K).

TABLEOxidation2:MoContentafterTemperature-Programmed0.3-1.0g,of100MoOcm33/min,/H-ZSM520%MixturesO(∼4wt%Mo,2inHe)

temperatureprogramMoaspreparedaMoafterTPOb

(wt%)(wt%)

I300to973K(10K/min)4.03.47II

300to623K(hold24h)4.03.60623to973K(10K/min)III300to623K(hold24h)

3.94

3.95

623to773K(10K/min,48h)773to950K(10K/min,2h)

a

BasedonamountofMoaddedtophysicalmixture.bBasedonelementalanalysisofoxidizedsample.

973K(at10K/min)ledtothesublimationof13%oftheMoas(MoO3)norMoO2(OH)2(Table2,programI).ThesublimedMoOxappearedassmallMoO3crystallitesonthewallsaftertheheatedsample.WhenMoO3/H-ZSM5mixtureswereheldat623Kfor24hinordertoremovephysisorbedH2OandtodisperseMoOxspeciesontotheexternalzeolitesurface,

Figure2.H2OdesorptionfromH-ZSM5duringfirst(a)andsecond(b)TPO.Thesamplewasheatedat10K/minfrom323to623K,heldfor24h,thenheatedat10K/minto973K(0.3g,100cm3/min,20%O2/Ar,101kPa).

subsequentheatinginairto973KledtoMolossesoflessthan10%(Table2,programII).TheH2Oevolvedduringexchangeappearedasawell-definedpeakbetween623and973K(asshownlaterinFigure3).Holdingat773Kfor48hduringtreatmentinair(Table2,programIII)eliminatedMoloss(<0.1%),apparentlybecauseMoOxcrystallitesspreadasstronglyinteractinglayerswithmuchlowervaporpressurethanbulkMoO3.ThismethodwasusedtopreparetheMo/H-ZSM5samplesusedinthereactionstudies(Figure1,Table1).

H2ODesorptionduringOxidativeTreatmentofMoO3/H-ZSM5Mixtures.Figure2showstheH2Oevolutionrateduringtemperature-programmedoxidation(TPO)ofH-ZSM5(usingprogramIIinTable2).Thepeakat300-500KcorrespondstothedesorptionofwateradsorbedonH-ZSM5fromambientair(2.62molH2O/Al).Thesmallerpeakbetween673and973Kcorrespondstozeolitedealumination(eq3)bylossofOHgroupsandextractionofAlatomsfromtetrahedralframeworkpositions.ApreviousthermogravimetricstudyofH-ZSM5reachedsimilarconclusions.24Figure2bshowsH2OevolutionratesduringasecondTPOaftercoolingthesampleindry20%O2/Ar(100cm3/min)toroomtemperature.Thephysisorbedwaterpeakisnolongerdetectedandthedealumi-nationpeakwassmalleranditappearedathighertemperature,indicatingthatthezeolitestructurewasstabilizedagainstfurtherdealuminationbytheinitialtreatment.ThenumberofOHgroupsdesorbedasH2Oduringthefirstandsecondtreatmentscorrespondsto0.23and0.06H/Al,respectively.

H2OevolutionduringTPOofaMoO3/H-ZSM5physicalmixture(4.5wt%Mo)25showsthattheamountofphysisorbedH2OisslightlysmallerthanonH-ZSM5(1.80molH2O/Al),butthesecondpeakismuchlarger(Figure3).ThisreflectstheanchoringofMoatcationexchangesitestoformMospeciescontainingOHgroupsandthesubsequentcondensationofthese

5790J.Phys.Chem.B,Vol.103,No.28,1999Figure3.H2OdesorptionfromMoO3/H-ZSM5mixture(4wt%Mo)duringTPO.Thesamplewasheatedat10K/minfrom323to623K,heldfor24h,thenheatedat10K/minto973K(0.3g,100cm3/min,20%O2/Ar,101kPa).

Figure4.H2Odesorptionbetween623and973KfromMoO3/H-ZSM5mixtureswithvaryingMoinitialconcentration(0.3g,100cm3/min,20%O2/Ar,101kPa,heatat10K/min).

OHgroupswiththoseinH-ZSM5orinneighboringMocations(eqs1or2).ThesmalldifferenceinH2Odesorptionratesbelow623KbetweenH-ZSM5(Figure2)andMoO3/H-ZSM5mixtures(Figure3)iscausedbyOHgroupsboundtoMoO3,andnotbyMoexchange.ThisconclusionisconfirmedbytheobservationthatH2Oevolutionratesbetween623and773KdidnotdependonMoloading(Figures4and14),suggestingthatMocationsmigrateintochannelsandexchangeonlyabove773K,asalsoproposedbyothers.26Figure4showsthattheamountofH2Oformedbetween773and973KincreasedwithincreasingMocontent.SomeMosublimesfrommixtureswithmorethan2wt%Mo(Table3)whentreatedbyprogramII(Table2);thiswasconfirmedbyelementalanalysesbeforeandafterairtreatmentandbytheMoO3residueformedattheexitofthetreatmentcell.

ThenumberofH2Omolecules(perAl)formedduringairtreatmentisshowninTable4.IfthelossofOHgroupsviarecombinationandAlextractionfromframeworkpositionsoccurredtothesameextentinMo/H-ZSM5andH-ZSM5samples,H2OdesorptionvaluesonMo/H-ZSM5canbecorrectedbysubtractingtheamountofH2OdesorbedfrompureH-ZSM5(0.23H/Al,Table4).ThesecorrectedvaluesarereportedasH/AlF(AlF)frameworkaluminum).Thisassump-

Borryetal.

TABLESpectrometry3:MoPhysicalbeforeContentandMeasuredafterOxidationbyAtomicofMoOAbsorption3/H-ZSM5(TemperatureMixturesProgramofVaryingIIinTableMoConcentration

2,0.3g,100cm320%O/min,2inHe)

targetMoanalyzedMoanalyzedMocontentcontentinphysicalMo/AlcontentafterMo/Al(wt%)mixture(wt%)beforeTPOTPO(wt%)afterTPO

11.10.111.00.1122.20.222.00.2044.00.423.60.378

7.9

0.87

6.3

0.68

TABLEcm3100/min,4:H2ODesorption(TPOProgramII,0.3g,100Mixtures

cm3/min,20%5%O2(DinHe)andIsotopicExchange(10K/min,2orH2)inAr)ofMoO3/H-ZSM5MoaftercontentTPO%)Mo/AlH2OdesorptionresidualHcontentratioH/AlaH/AlbalanceHf(wtFbH/MocH/AldD/AleH/Al

000.22900.6300.5640.919g1.00.110.3470.1181.110.4000.4560.8032.00.200.4740.2451.220.2540.3550.8293.60.370.6860.4571.230.0990.1700.8566.30.680.6650.436

0.64

0.0350.1070.772

4.3

0.45

0.0190.074

aFromintegratedareaofH2Odesorptionabove623K(Figure4).b

Subtractvalueincolumn3fromthatforH-ZSM5(0wt%Mo).cDividecolumn4byMo/Alratio.dFromD2

(g)/O(s)-Hisotopicexchange(Figure9).eFromH2(g)/O(s)-Disotopicexchange(Figure10).fSumofcolumns3and7.gIncludessecondTPOandD2(g)/O(s)-Hvalues(seetext).

Figure5.NumberofHdesorbedasH2OperAlatomduringtreatmentofMoO3/H-ZSM5mixturesinairto973K.

tionwasconfirmedusing27AlNMRtomeasurethechangeinintensityofthepeakcorrespondingtotetrahedral(framework)Alcenters(asdiscussedlater).Thesedatashowthatforsampleswithlessthan4wt%Mo,MoOxspeciesexchangewithH+withastoichiometryofaboutoneMoperH(Figure5,slope)1.24H/Mo).ForH-ZSM5withaSi/Alratioof14.3,thenumberofcationexchangesitesisnotsufficienttoaccommodateallMoatomsinthe8wt%Mosample,andtheexcessMoOxsublimesabove773Kduringairtreatment.

Vapor-ExchangeSynthesisfromSeparatedMoO3andH-ZSM5Powders.Inaseparateexperiment,Mo/H-ZSM5sampleswerepreparedbyflowing20%O2/He(100cm3/min)overbulkMoO3(0.185g),andcontactingthe(MoO3)nvaporwithabedofH-ZSM5(1.509g;Figure6).TheMocontentwouldbe7.3wt%ifalltheMoO3remainedintheexchanged

Mo-ExchangedH-ZSM5CatalystsFigure6.Preparationof“vapor-exchanged”Mo/H-ZSM5sample(1.5g,50cm3/min,20%O2/He,101kPa,950K,96h).

TABLEMixtures5:N2(D973afterPhysisorptionK

TPOandTPRatTreatment77KofMoOinAir3/H-ZSM5andH22)toMoconcnBETsurfaceareamicroporevolume(wt%)(m2/gzeolite)

(cm3/gzeolite)

02720.1281.12780.1192.02660.1203.62180.1076.31320.0594.3a

201

0.084

a

PreparedviaMoO3vaporexchangeat950K,96h(seetext).

samplesaftertreatmentinair.AlltheMoO3powdersublimedafter96hat950K,contactedtheH-ZSM5sample,andeitherexchangedatZSM5sitesordepositedaslightgreencrystals(bulkMoO3)oncolderreactorwallsabovethezeolitebed.TheresultingMo/H-ZSM5contained4.3wt%Mo.TheseresultsshowedthatsurfacediffusionisnotrequiredforMoOxmigrationduringtreatmentinairathightemperature(>873K)andthatvapor-phasetransportof(MoO3)noligomerscanoccurduringsynthesisofMo/H-ZSM5fromphysicalmixtures.CH4aroma-tizationreactionrates1wereverylowonthissample.Ataspacevelocityof750h-,CH4conversionwasonly1.7%(cf.4.3wt%Mosample;Table1).

ZeoliteStructuralChangesduringThermalTreatmentofMoO3/H-ZSM5Mixtures.TheunusualshapeoftheH2Oevolutionratesonthe8wt%Mo/H-ZSM527(Figure4)suggestedsomestructuraldifferencesbetweenthissampleandtheothers,whichweexploredusingpowderX-raydiffraction(XRD),N2physisorption,and27AlNMR.N2physisorptionresultswereusedtoestimatesurfaceareasandmicroporevolumes(Table5).NeitherdiffractionnorN2physisorptiondatashowedanystructuraldegradationofthezeoliteinsampleswith1.0or2.0wt%MorelativetopureH-ZSM5,butbothmethodsdetectedamodestlossofcrystallinityinthe3.6and4.3wt%Mosamples,andsignificantdestructionofthezeoliteframeworkinthe6.3wt%sample.27AlossofsurfaceareawithincreasingMocontentwaspreviouslyreportedalsoforMoOx/ZSM5preparedbyAHMslurrymethods.28

27AlNMRspectroscopycanprobethelocalstructureofAlcationsinMo/H-ZSM5samples.TetrahedralAlcentersinthezeoliteframeworkbondedtoterminalO-Hgroupshaveachemicalshiftof56ppm;extraframeworkoctahedralAlcenters(inZSM5andinAl2O3)shownoshift(δ)0ppm),whilecrystallineAl2(MoO4)3(octahedralAl,tetrahedralMo)showsalineat-13ppm.1527AlNMRspectraforMoOx/H-ZSM5samples(Figure7)showedthatthedensityofunperturbedframeworkAlcenters(correspondingtoBronstedacidsites)

J.Phys.Chem.B,Vol.103,No.28,19995791

Figure7.27AlNMRspectraofMoO3/H-ZSM5physicalmixturesafterTPOandTPRtreatmentsto973K(*)4.3wt%MosamplepreparedviaMoO3vaporexchangeat950K,96h.

decreasedwithincreasingMocontent.27AlNMRlinescorre-spondingtocrystallineAl2(MoO4)3appearedinthe6.3wt%Mo/H-ZSM5sample.27TheAl2(MoO4)3crystallitesaretoosmalltobedetectedbyX-raydiffractioninthissample,buttheybecomedetectableinMo/H-ZSM5sampleswithhigherMocontent(15wt%).15

Thus,itappearsthattheunusualshapeoftheH2Odesorptioncurvefor6.3wt%Mo/H-ZSM5reflectstheextractionofAlfromtheframeworktoformAl2(MoO4)3andtheconcomitantcollapseofthezeoliteporestructure.H2Odesorptionratesfrommixturesinitiallycontaining4and8wt%Mowereidenticaluptoabout850K(Figure4),indicatingthatMoOx-inducedextractionofframeworkAldidnotoccurbelow850K.Thisagreeswith27AlNMRdatafor15wt%Mo/H-ZSM5,whichshowedAl2(MoO4)3insamplestreatedinairat873or973K,butnotinthosetreatedat773K.15TheratioofOHgroupsremovedasH2OtoMoatomsinthe6.3wt%Mosampleisonly0.64;thisvalueisverysimilartothatexpectedforAl2-(MoO4)3(0.67),butitissmallerthaninMo/H-ZSM5sampleswithlowerMocontent.

IsotopicEquilibrationofD2withOHGroupsinMo/H-ZSM5Samples.TheisotopicequilibrationofsurfaceO-HgroupswithD2wascarriedoutaftertreatingMo/H-ZSM5inairat973K(procedureII,Table2).Thismethodwasused

previouslyinordertodeterminethenumberofOHgroupsinH-ZSM5andcation-exchangedH-ZSM5.16

D2/O-HexchangedataareshowninFigure8for3.6wt%Mo/H-ZSM5(preparedbyprogramII;Table2).ThetotalnumberofHatomsremovedasHDandH2correspondsto0.1HperAl.Figure9showsHDdesorptionratesduringD2/O-Hexchangeforallsamples(0-6.3wt%Mo;preparedbyexchangeinairat973Kfor0.5h).ThenumberofO-HgroupsdecreasedwithincreasingMocontent,asO-HgroupswereincreasinglyreplacedbyModuringexchange(eq1).ThedatainFigure9alsoshowthatMospeciescatalyzeD2dissociation,therate-determiningstepinD2/O-Hexchange,andthus

5792J.Phys.Chem.B,Vol.103,No.28,1999Figure8.ProductionrateofH2,HD,andD2Oduringtemperature-programmedD2(g)/O-H(s)isotopicexchangeon3.6wt%Mo/H-ZSM5(0.3g,100cm3/min5%D2/Ar,101kPa).

Figure9.RateofHDdesorptionduringD2(g)/O-H(s)isotopicexchangeonMo/H-ZSM5(0.3g,100cm3/min5%D2/Ar,101kPa).

decreasethetemperaturerequiredforD2-OHexchangebyabout200K.

TheformationofD2OduringD2/O-Hexchangeabove500K(Figure8)reflectssomereductionofMo6+species.Thisprocessinfluencesthecalculatedexchangevalues,becausebothH2OandHDOformviaD2OexchangewithOHgroups.H2OandHDOcontainHatomsfromsurfaceO-Hgroups,buttheirconcentrationsandisotopiccontentsaredifficulttomeasurebecauseofwateradsorptionandexchangeontheheatedwallsofthecellandthetransferlines.Theseeffectswereshowntobesmallbycoolingthesamplesto300KafterD2(g)/O(s)-Hexchange,andthenheatingto973KinflowingH2;themeasuredamountsofHDandD2formed(Figure10,denotedasH2(g)/O(s)-Dexchange)gavedensitiesofresidualODgroupsthatwereonlyslightlydifferentfromthosemeasuredfromtheinitialD2-OHexchange.

D2/O-HandH2/O-DexchangedatashowthateachMoaddedtoH-ZSM5replaces1.0(0.2OHgroupsduringexchangefromphysicalmixturesat973K(Figure11).29TetrahedralAlatomscontaininganOHgroup(AlF)givea27AlNMRlinethatdecreasesinintensityasMocontentincreases.ThesedataallowustocalculatethefractionoftheAlcenters

Borryetal.

Figure10.RateofHDdesorptionduringH2(g)/O-D(s)isotopicexchangeonMo/H-ZSM5(0.3g,100cm3/min5%D2/Ar,101kPa).

Figure11.Numberofacidsites(H)orframeworkAl(AlF)remainingpertotalAlforseriesofMo/H-ZSM5samplesafterTPOandTPRtreatmentto973K(0.5h)measuredby27AlNMR,D2/O-H,andH2/O-D(0.45Mo/AlsamplepreparedviaMoO3vaporexchangeat950K,96h).

thatinteractwithMoOxspeciesinsteadofOHgroups(Figure7),usingavalueof0.92fortheAlF/AlratioinaH-ZSM5samplepretreatedinairat773Kfor1h(Table4).30TheresultsfromD2-OH,H2-OD,and27AlNMRareinexcellentagreement(Figure11);theyshowthateachMoreplacesoneOHgroup,asexpectedfor(Mo2O5)2+dimersbridgingtwocationexchangesites.

ThesumoftheOHgroupsremovedasH2Oduringexchangeandthoseremaining(fromD2/O-Hdata)shouldcorrespondtothetotalnumberofOHgroupsinthestartingH-ZSM5samples(Table4).OnH-ZSM5,thetwoTPOexperimentsandtheD2-OHexchangeaddtoavalueof0.92OH/Al(Table4).30AllMosampleshadvaluesof0.77-0.86totalOH/Al,inreasonableagreementwiththevalueof0.92measuredonH-ZSM5(Table4).Discussion

X-raydiffraction,N2physisorption,and27AlNMRmethodsshowthataftertreatmentinairat973K,MoOx/H-ZSM5

Mo-ExchangedH-ZSM5CatalystsFigure12.FormationofMo/H-ZSM5activesitesfromMoO3/H-ZSM5physicalmixtures.

Figure13.StructureofbulkMoO3andsolid-statemigrationonH-ZSM5surface.

samplespreparedfromMoO3-H-ZSM5physicalmixturesleadtoexchangedMospeciessimilartothoseobtainedfromaqueousexchangemethods.LowpHconditionsmayalterthedensityandnatureofexternalOHgroupsonzeolitecrystals,31whichcanleadtoMoOxbindingandmayaccountforthehighexternalMoconcentrationsdetectedbyX-rayphotoelectronspectroscopyonsamplespreparedbyaqueousexchange.26,31Methaneconver-sionratesandselectivitiesaresimilaronsamplespreparedbythetwomethods,suggestingthatactivesitedensitiesandaromatizationpathwaysaresimilarinMoOx/H-ZSM5samplespreparedfromphysicalmixtures(thisstudy)andinthosepreparedusingaqueousexchange.1,3,6,28,32,33

MigrationofMoOxSpeciesduringTreatmentinAir.OurresultsareconsistentwiththeprocessesdepictedschematicallyinFigure12.Below623K,X-raydiffractionshowsthatMoO3andH-ZSM5crystallitesremainintact.Between623Kand773K,MoO3crystallitesdisappearasMoOxmigratesontoexternalzeolitesurfaces.Kno¨zinger,etal.34haveproposedthatMoO3spreadsonAl2O3surfacesviaan“unrollingcarpet”mechanism.Above600K,MoO3layersmigrateuntiltheyreachAlOxsites,wheretheyanchorandbecomeimmobile(Figure13).ThisprocessdecreasessurfaceenergiesasMoO3formsstrongbondswithAl2O3;similarprocesseswerenotdetectedonSiO2.35ThesedataandthestabilityofAl2(MoO4)3suggestthatAl-O-MobondsarestrongerthanSi-O-Mobonds.36

MoO3consistsofMoO6octahedrawithdisplacedcentralMoatoms.37MoO6octahedraformbilayerswithvanderWaalsinterlayerbinding(Figure13).Between623and773K,

J.Phys.Chem.B,Vol.103,No.28,19995793

Figure14.H2OdesorptionrateduringTPOofMoO3/H-ZSM5mixturesat740and840K(fromFigure4).(*)Mo/Alratiosareasprepared.MoconcentrationsafterTPOareshowninTable3.

vibrationsovercomethisbindingandbilayersslideovereachotheruntiltheyanchoratAlsitesonexternalZSM5surfaces.38MoOxspeciesmayalsomigrateintozeolitechannelsatthesetemperatures,buttheydonotreactwithO-Hgroups,asshownbythelowrateofH2Oformationbelow∼750KandbythelackofdependenceofthisrateontheMocontent(Figure14).AftertheexternalZSM5surfaceiscoveredcompletelywithaMoO3layerabove773K,furtherspreadingcannotdecreasethesurfaceenergy.Inourstudy,sampleswithmorethan4wt%MocontainmoreMothanthatrequiredtoformaMoO3monolayerontheexternalsurfaceareaoftheZSM5crystalsused.TheexcessMoOxcansublimeas(MoO3)n,asexpectedfromtheMoO3vaporpressureandfromtheobservedlossofMoinsampleswithhighMocontent(Table3).ExcessMocanalsomigrateintoZSM5channelsasisolatedMoOxspecies(asshownbythedisappearanceofXRDpeaksandbytheobservedincreaseofH2Odesorptionrateabove773K;Figure14)orextractAlfromtheZSM5frameworktoformAl2(MoO4)3(detectedinthe8wt%Mosampleby27AlNMRandconfirmedbytheunusualshapeoftheH2Odesorptioncurveforthissampleabove850K;Figure4).

ThestrongAl-O-MointeractionsthatleadtothespreadingofMoO3onexternalzeolitesurfacesalsopreventMoO3sublimation.Thus,physicalmixtureswithlessMothanthatrequiredtoformasingleMoO3layer(1-2wt%Mo)donotloseMobysublimation.Themixturecontaining4wt%Morequiredtreatmentinairat773Kfor24hinordertoattainfullspreadingofaMoOxlayerandtoavoidsublimationduringexchange(Table2,programIII).Evenafterthistreatment,the8wt%Mo/H-ZSM5samplelostabout20%ofitsMoO3bysublimation(Table3).ThisshowsthatthemaximumMoO3coveragefortheexternal(mesoporous)surfaceareaoftheZSM5used(24m2/g)correspondstoabout4wt%Mo.39

H2Oevolutionrates(perAlorg-catalyst)duringairtreatmentofMoO3/H-ZSM5mixturesathightemperatures(Figure4)reflecttherateofreplacementofH+withcationicMooxospecies.ThisrateofexchangeinturndependsonthemobilityofmigratingMospeciesandonthenumberofZSM5channelsavailablefortransport.Below740K,H2OevolutionratesareindependentofMocontent(Figure14),indicatingthatMoOxspeciesdonotundergothecondensationreactionsthatanchorthematexchangesites.Between750and850K,H2OevolutionratesbecomeproportionaltotheMocontentforsampleswith

5794J.Phys.Chem.B,Vol.103,No.28,19991.0-3.6wt%Mo(Mo/Al)0.11-0.37;Figure14).ThesedatasuggestthatzeolitechannelsbecomeaccessibletotheexternalMoO3surfacelayerinproportiontotheamountofMospreadasalayerontheexternalzeolitesurface(atleastupto3.6wt%Mo).AthigherMocontents,H2Odesorptionratesremainconstantastemperaturesincreasefrom750and850K,suggestingthattheexternalMoO3exceedsonemonolayerandthatallchannelopeningsbecomeaccessibletoMoO3;then,condensationreactionsbecomelimitedbyrateofMomigrationwithinZSM5channels.Thesedataareconsistentwiththeschemeineq5andwithexpectedformationofasingleMoO3layerinthe3.6wt%MosamplesfortheexternalareaavailableintheH-ZSM5crystalsusedinthephysicalmixtures.

OnthesamplewiththehighestMocontent(7.9wt%Moinmixture;6.3wt%afterexchange),multipleMoO3layerscanformonexternalzeolitesurfaces.Above843K,AlOxspeciescanbeextractedfromthealuminosilicateanddissolvedintoliquidlikeMoO3layers,withinwhichtheycanreacttoformstableAl2(MoO4)3domains.Al2(MoO4)3wasdetectedby27AlNMRandX-raydiffraction15insampleswithhighMocontent.Inourstudy,Al2(MoO4)3wasnotdetectedinsampleswithlessthan4wt%Mo,becauseMoO3monolayersapparentlylackthelocalthree-dimensionalstructureandthethermodynamicincentiverequiredtostabilizeAlinAl2(MoO4)3structures.TheextractionofAlOxspeciesfromthealuminosilicateframeworktoformAl2(MoO4)3causesthestructuralcollapsedetectedbyX-raydiffractionandN2physisorptionmeasure-mentsintheMo/H-ZSM5samplewith6.3wt%Mo.27Hightemperaturesalonedonotcausetheobservedstructuralcollapse,because“vapor-exchanged”(4.3wt%Mo)samplesweretreatedat950Kfor96hwithoutsignificantlossofcrystallinityand27AlNMRdidnotdetectanyAl2(MoO4)3

.ItappearsthathighlocalMoOxconcentrationsarerequiredinordertoextractAlfromthealuminosilicateframework.Thesamplewith6.3wt%Mo(7.9wt%Moinstartingmixture)andthevapor-exchangedsamplewereexposedtosimilarMoOxconcentrations(7.9wt%vs7.3wt%Mo),butvapor-exchangedsampleswerecontactedby(MoO3)n(n∼3-4)oligomersatapressureofabout25Pa(250ppm),11whilethe8wt%MosamplecontainedsignificantinterfacialareabetweenliquidlikebulkMoO3andzeolitesurfaces.Thus,localMoO3concentrationsarelowerduringvaporexchangesynthesis(Figure6),because(MoO3)noligomersadsorbontotheZSM5surface,migratewithinzeolitechannels,andbindirreversiblyatsiteswheretheyformMo-O-AlbondsbeforehighlocalMoO3concentrationsarereached.ExchangeofCationicMoOxoSpeciesviaCondensationReactions.TheamountofwaterevolvedduringexchangeandthenumberofO-HgroupsremainingafterexchangeshowthateachmigratingMoOxreplacesoneH+inH-ZSM5(Figures5and11).Thesedataandtherequiredchargebalancesuggestthat(Mo2O5)2+dimersinteractingwithtwoexchangesitesformduringexchange(eq1),whichacquiretheditetrahedralstructureof(Mo2O7)2-dimerswhentwoframeworkoxygenatomsprovidetheanchoringsitesforsuchdimers.InfraredbandsBorryetal.

correspondingtotetrahedralMo6+centersweredetectedprevi-ouslybutincorrectlyassignedto(MoO4)2-monomersbridgingtwoexchangesites,6astoichiometrythatisnotconsistentwithourresults.Thesizeof(Mo2O7)2-ditetrahedraintetra-n-butylammoniumdimolybdate(∼5.7Å)40andsimplegeometricalconsiderationsbasedonaZSM5crystalstructurewithrandomlydistributedAlionsshowedthatabout60%oftheAlatomsinH-ZSM5(Si/Al)14)canresidesufficientlyclosetoanotherAltoallowModimerstobridgeAlpairs.41Theexclusivepresenceof(Mo2O5)2+dimersisthenpossibleatallMo/Alratiosbelow0.6(i.e.,<5wt%MoforourSi/Alzeoliteratio).27AlNMRshowsthatallAlcentersinvapor-exchangedMo/H-ZSM5havebeendisplacedfromframeworkcrystallographicpositions(Figure7).Thus,eitherAlcentersareinteractingwith(Mo2O5)2+dimersortheyhavebeenextractedfromtheframeworkviacondensationoftwoO-HgroupsatnextnearestneighborAlatoms.AllAl-OHspeciesinH-ZSM5ultimatelyfindthenextnearestneighborrequiredtoanchoraModimerortodehydrateduringtheextendedtreatmentinair(96hat950K)requiredforvapor-phaseexchange.TheapparenthighdensityofAlsitepairsinthissamplemayreflectthemigrationofAlcentersbyexchangewithSiatomsabove773K,42aprocessthatstopswhenthermodynamicallystableAl-O-AlorMo-O-Albondsformviacondensationreactions.Therequiredmigrationcanoccurvia“hopping”intolatticevacanciesordefects,viathermally-inducedexchangeofSiwithAl,orbyH2O-assisteddiffusionpathways.WhenAlatomsresidemomentarilynearanotherone(asanextnearestneighbororacrossachannel),theirOHgroupsreacttoformH2OandanO2-vacancy,orexchangewithtwo(MoO2OH)+,whichcanthencondensetoforma(Mo2O5)2+dimercontainingabridgingoxygenatom.EithereventstabilizesAl-O-AlorAl-O-Mositepairsslightlydetachedfromframeworkpositions,distortedfromtetrahedralsymmetry,andinvisibleby27AlNMR.

27AlNMRdata(Figure7)confirmedthenatureoftheexchangedMospeciesandthedetailsofthemigrationandexchangeprocesses.Theintensityofthelineat56ppm,correspondingtotetrahedralAlcenterintheZSM5framework,decreasedastheMocontentincreasedforsampleswithlessthan4wt%Mo;thisdecreaseinintensitywasnotcompensatedbythedetectionofnewormoreintenseNMRlines.AthigherMocontents,theframeworkAllinewasnotdetected,butstronglinescorrespondingtoAl2(MoO4)3appear.NoNMRlinesweredetected(δ)-416and543ppm)invapor-exchanged(4.3wt%)Mo/H-ZSM5samples.ThisshowsthatvaporexchangecreatesneitherAl2O3norAl2(MoO4)3insampleswiththisMocontent.Theanchoringof(Mo2O5)2+dimersappearstoformMo-O-AlbondsthatdistortthetetrahedralstructureofframeworkAlcenters,leadingtothetypicalbroadeningoflinesforquadrupolarAlnucleiwithdecreasingsymmetry.43TheAlsymmetryatexchangesitesbondedto(Mo2O5)2+dimersdiffersmarkedlyfromthatobservedinknownAlcompounds.Al3+centershavetetrahedralcoordinationinaluminosilicateframe-worksandoctahedralsymmetryinγ-Al2O3.Al2(MoO4)3has

Mo-ExchangedH-ZSM5CatalystsFigure15.LiteraturereportsofCH4conversion,normalizedbythemaximumconversionreportedineachstudy,asafunctionofMo/Alratio.Thesolidlinesconnectdatafromasinglestudy.

octahedralAlcentersandtetrahedralMocenters,while(Mo2O5)2+dimerslocatedattwocationexchangesiteswouldhavedistortedtetrahedralAlcentersboundtodistortedtetrahedralMo6+;itisthisrequireddistortionthatleadstothebroadeningofNMRlinesafterexchangeordehydroxylationofzeolites.

FormationofActiveSitesduringCH4Aromatization.ThepredominantformationofCOx,H2O,andH2duringinitialcontactofMoOx/H-ZSM5withCH4at950K(Figure1)andtheincreaseinhydrocarbonformationrateswithincreasingtimeonstreamsuggestthatactivesitesarecreatedbythereductionandcarburizationof(Mo2O5)2+speciesthatarenotactiveformethaneconversiontohydrocarbons.Sitesin(Mo2O5)2+dimersreactwithC-HbondsinCH4,buttheycannotdesorbtheresultingfragments,whichforminsteadMoCxspeciesviasubsequentdehydrogenationsteps.CatalyticCH4activationbeginsonlyafterreducedMospeciesarepassivatedbytheformationofthesecarbidespecies,whicharethenabletodesorbreactionproductsandcompleteaturnover.ThehighmeltingpointandthelowvolatilityofMosuboxidesandcarbides11suggestthatmigrationisunlikelytooccurafterreduction-carburizationofexchanged(Mo2O5)2+.Theseconclusionsareconsistentwithdetailedcharacterizationstudiesreportedelse-where.44,47

Exchanged(Mo2O5)2+speciesareprecursorstoCH4activa-tionsitesinMo/H-ZSM5catalysts.Therefore,CH4conversionratesincreasewithincreasingextentofMoexchangeduringsynthesis.Reactionpathways,however,alsorequireBronstedacidsites,providedbytheremainingH+speciesinZSM5.TheseacidsitescatalyzechaingrowthandcyclizationreactionsoftheinitialethyleneproductsformedinCH4activationsteps;thus,acidsitesshifttheproductdistributionstowardaromatics,whicharefavoredoveralkenesbythermodynamics.Asaresult,thefractionoftheAlsitesexchangedbyMo(Mo/Al),andnottheMocontent,becomestherelevantparameterdeterminingcatalyticrates.OurH-ZSM5samples(Si/Al)14.3)loseabout40%oftheframeworkAlsitesduringairtreatmentat973K;forthesesamples,theoptimumMoconcentrationforCH4aromatizationisabout0.4Mo/Al(Figure15)orabout0.7Mo/AlFifonlyframeworkAlareassumedtostabilizeH+.Inthesesamples,theremainingBronstedacidsitesaresufficienttocatalyzeC2H4aromatizationreactionstonear-equilibriumlevels.Figure15showsourdataandthosefrompreviousstud-ies,1,6,28,45,46asCH4reactionratesvsMo/Alratio.Allstudiesdetectmaximumrelativeratesoncatalystswith0.3-0.5Mo/

J.Phys.Chem.B,Vol.103,No.28,19995795

Alratios,eventhoughthezeoliteSi/Alratiovariedoverawiderange(12.5-50)inthesesamples.Conclusion

CH4conversionratesoncatalystspreparedfrommixturesofMoO3andH-ZSM5powdersareverysimilartothoseobtainedonsamplespreparedviaaqueousimpregnation-exchangemethods.Thesimplersynthesisfromphysicalmixturesallowsmeasurementsofthekineticsofformationandofthestoichi-ometryofexchangedMoOxspecies.H2OevolvedduringheatingreflectsthekineticsofcondensationpathwaysthatanchorMoOxatzeoliteexchangesites.IsotopicequilibrationbetweenD2andresidualOHgroupsafterexchangegivesthefractionofH+speciesreplacedbyexchangedMooxocations.ThemechanismofexchangeinvolvestheinitialformationofanexternalMoO3monolayeronzeolitecrystalsviasurfacemigrationat623-773K.WhentheMocontentexceedsthatrequiredtoformamonolayer(∼4-5wt%MointheH-ZSM5used),MoOxspeciesarelostas(MoO3)noligomersviasublimationorasunreducibleandinactiveAl2(MoO4)3domainsviareactionswithframeworkAlatoms.Between773and973K,surfaceandgas-phasetransportpathsleadtomigrationofMoOxspeciesintozeolite+channelsandtoreactionswithOHgroupstoformMoO2-(OH)speciesthatcondensequicklytoformH2OandastrongMo-O-Alanchoringbond.TheamountofH2OevolvedduringexchangeandthenumberofH+remainingafterexchangeareconsistentwiththereplacementofoneH+byeachexchangedMo.ThisstoichiometryandtherequirementforchargeneutralityduringexchangeleadtotheproposalthatMospeciesexistas(Mo2O5)2+ditetrahedrainteractingwithtwozeoliteexchangesites.These(Mo2O5)2+dimersreduceandcarburizeduringmethanereactionstoformtheactiveMoCxsitesrequiredforcatalyticC-Hbondactivation.MaximumratesareobservedoncatalystswithintermediateMo/Alratios(∼0.4),becausebothexchangedcationsandresidualBronstedacidsitesarerequiredfortheconversionofmethanetoC6+aromatics.

Acknowledgment.R.B.wassupportedbyaNationalSci-enceFoundationFellowship.Y.-H.K.wassupportedbytheKoreanScienceandEngineeringFoundation(KOSEF).TheprojectwasfundedbytheFederalEnergyTechnologyCenteroftheU.S.DepartmentofEnergy(DE-AC03-76SF00098)underthetechnicalsupervisionofDr.DanielDriscoll.ReferencesandNotes

(1)Wang,L.;Tao,L.;Xie,M.;Xu,G.Catal.Lett.1993,21,35.(2)Solymosi,F.;Erdohelyi,A.;Szoke,A.Catal.Lett.1995,32,43.(3)Wang,D.;Lunsford,J.H.;Rosynek,M.P.Top.Catal.1996,3,289.

(4)Liu,S.;Dong,Q.;Ohnishi,R.;Ichikawa,M.J.Chem.Soc.,Chem.Commun.1997,1455.

(5)Stakheev,A.Y.;Khodakov,A.Y.;Kustov,L.M.;Kazansky,V.B.;Minachev,Kh.M.Zeolites1992,12,866.

(6)Xu,Y.;Shu,Y.;Liu,S.;Huang,J.;Guo,X.Catal.Lett.1995,35,233.

(7)Dong,Y.;Liu,S.;Zhang,Q.;Liu,J.;Yang,K.ActaPet.Sin.(Pet.Process.Sect.)1992,8,66.

(8)Xu,Y.;Liu,S.;Wang,L.;Xie,M.;Guo,X.Catal.Lett.1995,135,135.

(9)Satterfield,C.N.HeterogeneousCatalysisinIndustrialPractice,2nded.;McGraw-Hill,Inc.:NewYork,1991.

(10)Brewer,L.Molybdenum:Physico-chemicalpropertiesofitscompoundsandalloys;InternationalAtomicEnergyAgency:Vienna,1980;Vol.7.

(11)Knacke,O.,Kubaschewski,O.,Hesselmann,K.,Eds.Thermo-chemicalPropertiesofInorganicSubstancesII,2nded.,;Springer-Verlag:Berlin,1991;Vol.2.

(12)Vahel,A.;Hubener,S.;Eichler,B.Radiochim.Acta1995,69,233.

5796J.Phys.Chem.B,Vol.103,No.28,1999

(13)Vahel,A.;Hubener,S.;Dressler,R.;Eichler,B.;Turler,A.Radiochim.Acta1997,78,53.

(14)Minming,H.;Howe,R.F.J.Catal.1987,108,283.

(15)Liu,W.;Xu,Y.;Wong,S.-T.;Wang,L.;Qiu,J.;Yang,N.J.Mol.Catal.A1997,120,257.

(16)Biscardi,J.A.;Meitzner,G.D.;Iglesia,E.J.Catal.1998,179,192.

(17)H-ZSM5samples(0%Mo)wereheatedtohighertemperatures(973K)inordertoensurecompleteisotopicexchange.

(18)Therelaxationdelaywas1.0s,thesweepwidthwas50kHz,andallchemicalshiftswerereferredtoadiluteAl(NO3)3solution.A90°/180°echopulsesequencewasusedanddatawerecollectedfor0.35h.Increasingtherelaxationdelaybetweenacquisitionsdidnotincreasethe27AlNMRsignal,indicatingthat1.0swassufficienttoallowfullrelaxationofallAlspinstates.Inthiscase,theNMRsignalpeakarea,integratedusingPeakfit(JandelScientific),isproportionaltothenumberofAlcentersrepresentedinthatpeak.

(19)Kentgens,A.P.M.;Scholle,K.F.M.G.J.;Veeman,W.S.J.Phys.Chem.1983,87,4357.

(20)Deng,F.;Du,Y.;Ye,C.;Wang,J.;Ding,T.;Li,H.J.Phys.Chem.1995,99,15208.

(21)Experimentsusingunheatedlinesoverestimatedtheamountofcarbondepositedonthecatalystbecauseofcondensationofnaphthaleneandheavierproductsinthegastransferlines.

(22)HP6890gaschromatograph(Hewlett-Packard)equippedwithaCarboxen1000packedcolumn(3.2mm×2m,Supelco)connectedtoathermalconductivitydetectorandwithaHP-1capillarycolumn(methylsilicone,0.32mm×50m×1.05mm,Hewlett-Packard)connectedtoaflameionizationdetector.

(23)Allgaseswere>99.99%purity(fromPraxairandMatheson)andwerepassedthroughmolecularsieve(13Xand5A)trapsbeforeuse.Flowratesweremeteredusingelectronicmassflowcontrollers(PorterInstru-ments).

(24)Wolthuizen,J.P.;vandenBerg,J.P.;canHoof,J.H.C.InCatalysisbyZeolites,Imelik,B.,Ed.;ElsevierScientificPublishingCo.:Amsterdam,1980.

(25)TheinitialconcentrationofMoforthissamplewas4.0wt%,butaftertemperature-programmedtreatmentinairupto973Kthesamplecontained3.6wt%Mo(Table3).

(26)Wang,D.;Lunsford,J.H.;Rosynek,M.P.J.Catal.1997,169,347.

(27)TheinitialMoconcentrationforthissamplewas8wt%,butaftertemperature-programmedtreatmentinairthesamplecontained6.3wt%Mo(Table3).

(28)Chen,L.;Lin,L.;Xu,Z.;Li,X.;Zhang,T.J.Catal.1995,157,190.

(29)Inprinciple,thenumberofDatomscountedduringH2(g)/O(s)-DexchangeshouldequalthenumberofHatomsdetectedduringD2(g)/O(s)-Hexchange.Thelatter,however,waslargerforallsamplescontainingMo,becausemoreD2wasformedduringH2(g)/O(s)-DexchangethanH2

Borryetal.

wasduringtheinitialD2(g)/O(s)-Hexchange.OneexplanationforthisdiscrepancyisthatthemassfragmentationpatternoftheD2OreductionpeakduringthefirstD2(g)/O(s)-HexchangecontainssomeH2O.

(30)ThisvalueislessthanunitybecausesomeacidsiteswerelostduringtheconversionofNH4-ZSM5toH-ZSM5at773K.ThissamplealsolostadditionalO-HgroupsduringtheinitialD2(g)/O(s)-Hexchange,becauseitrequiredheatingto973Kinordertocompletetheisotopicexchange.TheH/AlvaluesfromthesecondH2(g)/O(s)-DexchangewereusedforMo-containingsamplesinthecalculationsshowninTable4,becausethisledtoamoreaccuratehydrogenatombalance.

(31)Zhang,J.-Z.;Long,M.A.;Howe,R.F.Catal.Today1998,44,293.

(32)Lunsford,J.H.;Rosynek,M.P.;Wang,D.Presentedatthe4thInternationalNaturalGasSymposium,KrugerNationalPark,SouthAfrica,1995.

(33)Solymosi,F.;Szoke,A.SymposiumonHeterogeneousHydrocarbonOxidation,211thNationalMeeting,AmericanChemicalSociety(NewOrleans,LA),1996;p147.(34)Mestl,G.;Kno¨zinger,H.Langmuir1998,14,3964.(35)Regalbuto,J.R.;Ha,J.-W.Catal.Lett.1994,29,189.

(36)ThefactthatMoOxspreadingoccursonexternalZSM5surfacessuggeststhattheAlcontentinH-ZSM5hasahighenoughsurfacedensityofAlsitestoform(MoOx)noligomersthatbridgeAlspeciesontheexternalsurface.Si-O-Mobondsdonotappeartobestrongenoughtosurviveairtreatmentat950K,sinceMoO3vaporexchangepreparation(Figure6)resultsinonly4.3wt%Mo,whichismuchlessthanexpectediftheexternalZSM5surfaceareawascoveredandframeworkOHgroupswereexchangedwithMo.

(37)Andersson,G.;Magneli,A.ActaChem.Scand.1950,4,793.(38)MoO3mayeventuallyspreadintoamonolayer,butbelow773KtheweakbondsbetweenbilayersaremorelikelytobreakthanthestrongMo-O-Mobondswithinbilayers.

(39)ThisMocontent(4wt%)correspondstoabout19m2/gforaMoO3bilayeror38m2/gforamonolayer.

(40)Pope,M.T.HeteropolyandIsopolyOxometalates;InorganicChemistryConcepts;Jorgensen,C.K.,Ed.;Springer-Verlag:Berlin,1983.(41)Feng,X.B.;Hall,W.K.Catal.Lett.1997,46,11.

(42)TheTammanntemperatureofquartzSiO2is940K,somobilityofionsintheO-Si-OlatticeofZSM5shouldoccurbetween673and973K.

(43)Engelhardt,G.;Michel,D.High-resolutionsolid-stateNMRofsilicatesandzeolites;Wiley,NewYork,1987.

(44)Borry,R.W.;Li,W.;Meitzner,G.;Iglesia,E.J.Catal.submittedforpublication.

(45)Xu,Y.;Liu,W.;Wong,S.-T.;Wang,L.;Guo,X.Catal.Lett.1996,40,207.

(46)Pinglian,T.;Zhusheng,X.;Tao,Z.;Laiyuan,C.;Liwu,L.React.Kinet.Catal.Lett.1997,61,391.

(47)Borry,R.W.;Lu,E.C.;Kim,Y.H.;Iglesia,E.Stud.Surf.Sci.Catal.1998,119,403.