Insights into the influence of the Ag loading on Al2O3 in the H2-assisted C3H6-SCR of NOx

(1)

HAL Id: hal-00979536

https://hal.sorbonne-universite.fr/hal-00979536

Submitted on 16 Feb 2016

HAL is a multi-disciplinary open access

archive for the deposit and dissemination of

sci-entific research documents, whether they are

pub-lished or not. The documents may come from

teaching and research institutions in France or

abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, est

destinée au dépôt et à la diffusion de documents

scientifiques de niveau recherche, publiés ou non,

émanant des établissements d’enseignement et de

recherche français ou étrangers, des laboratoires

publics ou privés.

Distributed under a Creative Commons Attribution| 4.0 International License

Insights into the influence of the Ag loading on Al2O3 in

the H2-assisted C3H6-SCR of NOx

Tesnim Chaieb, Laurent Delannoy, Guylène Costentin, Catherine Louis,

Sandra Casale, Ruth L. Chantry, Z.Y. Li, Cyril Thomas

To cite this version:

(2)

ContentslistsavailableatScienceDirect

Applied

Catalysis

B:

Environmental

jo u r n al ho m e p ag e :w w w . e l s e v i e r . c o m / l o c a t e / a p c a t b

Insights

into

the

inﬂuence

of

the

Ag

loading

on

Al

2 O

3 in

the

H

2 -assisted

C

3 H

6 -SCR

of

NO

x

Tesnim

Chaieb

a,b

_,

_Laurent

_Delannoy

a,b

_,

_Guylène

_Costentin

a,b

_,

_Catherine

_Louis

a,b

_,

Sandra

Casale

a,b

_,

_Ruth

_L.

_Chantry

c

_,

_Z.Y.

_Li

c

_,

_Cyril

_Thomas

a,b,∗

a_Sorbonne_{Universités,}_UPMC_Univ_Paris_06,_UMR_7197,_Laboratoire_de_Réactivité_de_Surface,₄_Place_Jussieu,_F-75005_Paris,_France

b_CNRS,_UMR_7197,_Laboratoire_de_Réactivité_de_Surface,₄_Place_Jussieu,_F-75005_Paris,_France

c_Nanoscale_Physics_Research_Laboratory,_School_of_Physics_and_Astronomy,_University_of_Birmingham,_Edgbaston,_Birmingham_B15_2TT,_United_Kingdom

a

r

t

i

c

l

e

i

n

f

o

Articlehistory:

Received8January2014

Receivedinrevisedform7March2014

Accepted11March2014

Availableonline20March2014

Keywords:

Ag/Al2O3

H2effect

Selectivecatalyticreduction

Kinetics

Compensationphenomenon

a

b

s

t

r

a

c

t

TheadditionofH2hasbeenreportedtopromotedrasticallytheselectivecatalyticreductionofNOx byhydrocarbons(HC-SCR).Yet,theinﬂuenceoftheAgloadingontheH2-promotedHC-SCRhasbeen thesubjectofaverylimitednumberofinvestigations.TheH2-HC-SCRearlierstudiesreportedmostly onAg/Al2O3samplescontainingabout2wt%Ag,sincethisparticularloadinghasbeenshowntoprovide optimumcatalyticperformancesintheHC-SCRreactionintheabsenceofH2.Thepresentstudyhighlights fortheﬁrsttimethattheH2-C3H6-SCRcatalyticperformancesofAg/Al2O3samplesimprovedinthe 150–550◦_C_temperature_domain_as_the_Ag_loading_(Ag_surface_density:_x_(Ag/nm2

Al2O3))decreasedwell

below2wt%.AdetailedkineticstudyofH2-C3H6-SCRwasperformedinwhichthereactionordersin NO,C3H6andH2,andtheapparentactivationenergiesweredeterminedforthereductionofNOxtoN2 onaAg(x)/Al2O3catalystsseries,forwhichAgwasfoundtobeinahighlydispersedstatebyTEMand HAADF-STEM.Remarkably,changesinthesekineticparameterswerefoundtooccuratanAgsurface densitycloseto0.7Ag/nm2

Al2O3(Agloadingof2.2wt%)coincidingwiththechangesobservedearlierin

theNOxuptakesoftheAl2O3supportingoxide[18].Interpretationoftheactivityandkineticdataled ustoconcludethattheH2-C3H6-SCRreactionproceedsviatheactivationofH2andC3H6onAgspecies andtheirfurtherreactionwithNOxadspeciesactivatedontheAl2O3support.Theunexpectedhigher catalyticperformancesoftheAgsampleswiththelowerAgsurfacedensitieswasattributedtothehigher concentrationofactivesitesontheAl2O3supportingoxideabletochemisorbNOxspecies,inagreement withtheNOxuptakedata.ThekineticdataobtainedforAgsurfacedensitieslowerthan0.7Ag/nm2Al2O3

alsosuggestthattheinteractionbetweenNOxandC3H6 adspecieswouldberatedetermininginthe C3H6-SCRprocess.

1. Introduction

Sincethe late 1970s, thepreservation of theair qualityhas becomeone of themajorconcerns for theOECD (Organization forEconomicCo-operationandDevelopment)countriesinorder to minimize the impact of air pollutants on environment and health.InEuropeinparticular,theEuro6standards,expectedto enterintoforceinJanuary2014,willmainlyfurtherreducethe NOx(NO+NO2)emissionsfromdieselautomotives,from180to

∗ Correspondingauthorat:SorbonneUniversités,UPMCUnivParis06,UMR7197,

LaboratoiredeRéactivitédeSurface,4PlaceJussieu,Case178,F-75005Paris,France.

Tel.:+33144273630;fax:+33144276033.

E-mailaddress:cyril.thomas@upmc.fr(C.Thomas).

80mg/km[1].Tomeettheseevermorestringentstandards, fur-therimprovementsintheefﬁciencyofthecatalyticconvertersare required.

Asemphasizedearlier[2],thecatalyticreductionofNOxtoN2

inastronglyoxidizingmediumisnottrivial.Theselectivecatalytic reductionofNOxbythehydrocarbons(HC-SCR)wouldbean

ele-gantalternativetotheselectivecatalyticreductionofNOxbyNH3

(NH3-SCR)andtheleanNOxtrap(LNT)technologies,whichboth

exhibitintrinsicshortcomingssuchastheNH3slipandCO2

penal-ties,respectively.Amongthecatalyticformulationsevaluatedto date,Ag/Al2O3hasbeenreportedtobethemostpromisingcatalyst

inthepioneeringworkbyMiyadera[3].Inthisstudy,itwasshown thatNOxcouldbeefﬁcientlyandselectivelyreducedtoN2by

var-ioushydrocarbonsinthepresenceofwaterandinalargeexcess ofO2.YettheuseofsuchAg/Al2O3catalystsislimitedbecauseof

http://dx.doi.org/10.1016/j.apcatb.2014.03.025

(3)

therestrictedoperatingtemperaturewindow(300–500◦C)within whichthesematerialsefﬁcientlycatalyzetheHC-SCRreaction[2,3], asexhauststemperaturesaslowas150◦C aretypically encoun-teredinthecatalyticconvertersofdieselcars[4].ImprovedHC-SCR efﬁcienciesatthelowertemperatureshavebeenobtainedwith higherhydrocarbons[5,6]orethanol[7],but,despitetheobserved broadeningoftheoperatingtemperaturewindowwiththehigher hydrocarbons,theAg/Al2O3catalystsdonotmeettheSCR

efﬁcien-ciesneededatthelowertemperaturesyet[8].

Thediscoveryofalow-temperaturepromotingeffectofH2on

theHC-SCRofNOxbySatokawaetal.forC1–C4hydrocarbons[9,10],

whichwaslaterconﬁrmedforhigherhydrocarbons[2,11,12],is undoubtedlyamajorbreakthroughforalumina-supportedsilver catalysts.Interestingly,theH2 effecthasbeenshowntobe

spe-ciﬁctoAg/ZSM-5[13]andAg/Al2O3[10],asAg/SiO2,Ag/TiO2and

Ag/ZrO2werefoundtobeinactiveinthepresenceofH2intheC3H8

-SCRfeed[10].Thispromptedtheimportanceofaluminumspecies andthepotentialcontributionofAl2O3intheHC-SCRprocessas

outlinedbyseveralgroups[14,15].

Recently,thecharacterizationoftheAl2O3supportingoxideof

aseriesofAg/Al2O3samples,viaanewly-developed

characteriza-tionmethod:namelythetemperature-programmeddesorptionof NOx[16,17],allowedustoprovidefurtherinsightsintothe

ori-ginoftheoptimumloadingofAgonAl2O3 fortheC3H6-SCRof

NOx[18],reportedtobeabout2wt%inmoststudies[3,19–24].

WecametotheconclusionthatthisparticularAgloadingresulted fromthemaximumloadingofsilverperunitsurfaceareaofAl2O3

(Agsurfacedensityconcept)forwhichAg2Oclustersremainhighly

dispersedonfreshlycalcinedsamples[18].Toourknowledge,the inﬂuenceoftheAgloadingontheH2-promotedHC-SCRofNOxhas

beenthesubjectofaverylimitednumberofinvestigations[24,25], mostoftheworksinthisﬁeldreportingonAg/Al2O3sampleswith

anominalAgloadingcloseto2wt%[2,6].Sadokhinaetal.[24]and Shimizuetal.[25]concludedtoanoptimumAgloadingof2wt% intheH2-C6H14-SCRandH2-C3H8-SCRreactionsinthepresence

of0.1%and0.5%H2,respectively.Intheseworksthealiquotsof

catalystsevaluatedintheH2-HC-SCRreactionswereeitherkept

constant[24]orvariedtomaintainconversions below30%[25] sothattheamountsofAgchangedintheexperimentsperformed. Thisbringsaboutadditionalcomplexityintheinterpretationofthe catalyticdata,especiallywhenthoseareexpressedasNOx

conver-sions[24]orratesofNOreduction[25]pergofAg/Al2O3sample.

Additionally,Shimizuetal.reportedontheNOreductionratesasa functionoftheAgloadingoftheAg/Al2O3catalystsatagiven

tem-perature(300◦C)[25],whereasSadokhinaetal.showedthatthe catalyticactivityoftheinvestigatedAg/Al2O3sampleswas

inﬂu-encedsigniﬁcantlybythereactiontemperatureandpreferredto providecomparisonofthecatalystsoverawiderangeofreaction temperatures(100–550◦C)[24].

Theaimofthepresentworkistogainfurtherunderstanding ontheinﬂuenceoftheAgloadingofAg/Al2O3samplesintheH2

-assistedC3H6-SCRofNOx(H2-C3H6-SCR).Forthispurpose,andas

doneinourearlierstudyontheC3H6-SCRofNOx[18],theamount

ofAginthealiquotsofsamplesevaluatedinH2-C3H6-SCRwaskept

constantviadilutionoftheAg/Al2O3catalystswiththebareAl2O3

oxide.Thisstudyalsoreportsonoriginalﬁndingsintheinﬂuenceof AgloadingonthekineticsofH2-C3H6-SCRthathelpexplainingthe

observedcatalytictrendintheH2-promotedC3H6-SCRreaction.

2. Experimental

2.1. Catalystsynthesisandcharacterization

The seriesof Ag/Al2O3 samples investigated in thecatalysis

studycorrespondstothat usedand characterizedearlierinthe

0

1

2

3

0.0

0.4

0.

8

1.2 Ag surface density (Ag/nm

2Al2O3

)

NO

x

uptake

(µmol/m

2 Al2O3

)

4

3

2

1

0 Ag loading (wt%)

Fig.1.NOxuptakesasafunctionoftheAgsurfacedensity/Agloading:(䊉)newly

synthesizedAg/Al2O3samplesand()dataextractedfrom[18].

C3H6-SCRstudy[18].The␥-Al2O3support(Procatalyse,180m2/g)

was ground and sieved, and the fraction between 0.200 and 0.315mm was used toprepare the Ag-promoted samples.The depositionofAgwasperformedbyincipientwetnessimpregnation ofthebareAl2O3 support(0.7cm3/gporousvolume)with

aque-oussolutionsofAgNO3(Aldrich,>99%)toachievesilverloadings

varying from0.5 to 4.3wt%,which wereascertained by induc-tively coupled plasma atomic emission spectroscopy (ICP-AES, CNRS–Solaize).Afterimpregnation,theAg-loadedsampleswere agedfor4hunderambientconditionsandsubsequentlydriedat 100◦Covernight.Finally,theAg-loadedsampleswerecalcinedat 600◦C(3◦C/min)for4hin amufﬂe furnace.Fromhereon, the sampleswillbedenotedasAg(x)/Al2O3,wherexrepresentsthe

AgsurfacedensityexpressedasthenumberofAgatomspernm2

ofsupport(Ag/nm2

Al2O3)[18].TheAgloadingsofAg(0.0)/,Ag(0.1)/,

Ag(0.3)/,Ag(0.4)/,Ag(0.6)/,Ag(0.7)/,Ag(0.8)/,Ag(0.9)/,Ag(1.1)/and Ag(1.3)/Al2O3amountedto0.0,0.5,0.9,1.3,1.8,2.2,2.6,3.1,3.5and

4.3wt%,respectively[18].

Ag(0.6)/, Ag(0.7)/,Ag(0.8)/Ag(0.9)/and Ag(1.1)/Al2O3 samples

werenewly preparedaccordingtotheabove-mentioned proce-duretobeinvestigatedbyelectronmicroscopytechniquesafter being characterizedby theNOx-TPD methoddescribed in Refs.

[16–18].TheAgloadingandtheBETsurfaceareaofthe newly-prepared samples were measured using an X-ray ﬂuorescence (XRF)spectrometerXEPOSHE(AMETEK)andaBelsorpmax(Bell Japan) equipments, respectively. The characterization of these newly-preparedsamplesbytheNOx-TPDmethodwasinexcellent

agreementwiththatreportedearlier[18]withinthelimitsof accu-racyofthetechnique(Fig.1),thusattestingforthereproducibility ofthepreparationmethod.Theseparticularsilverloadingswere selectedonthebasisoftheearlierNOx-TPDresultswhichsuggested

thatAgremainedinanoptimumdispersedstateonfreshlycalcined samplesuptoanAgsurfacedensityofabout0.7Ag/nm2_[18]_.

BrightﬁeldTEMandenergydispersiveX-rayspectroscopy(EDS) (PGTdetector)characterizationofAg(0.6)/andAg(1.1)/Al2O3was

performed using a JEOL 2010 microscope operating at 200kV equippedwithanOriusCCDcamera(Gatan).Aberration-corrected scanningtransmissionelectronmicroscopy(STEM)imaging was carriedoutattheUniversityofBirminghamusinga200kVJEOL 2100F microscope, ﬁtted with a high angle annular dark ﬁeld (HAADF)detectorandaBrukerEDSdetector.Forthepurposesof imagingthesamplesweredropped,indrypowderform,onto amor-phouscarbon-coatedcopperTEMgrids.Forcomparisonpurposes, theAg(0.7)/Al2O3samplewasalsodepositedonaTEMgridafter

beingdispersedinethanol.

2.2. H2-C3H6-SCRruns

ThesteadystatecatalyticH2-C3H6-SCRexperimentswere

(4)

tonotethatunlessspeciﬁedotherwiseandincontrastwithmost studiespublishedtodate,theamountofsilverintroducedinthe catalyst beds remainedessentially constant. The samples were heldonplugsofquartzwoolandconsistedin0.38gof mechan-icalmixturesofAg(x)/Al2O3andAl2O3ofthesamegrainsizesin

whichtheamountofAgwasequalto30.9±1.2␮mol.The tem-perature of the tubular furnace was set by a Eurotherm 2408 temperaturecontrollerusingaKtypethermocouple.Priortothe H2-assisted C3H6-SCR experiments, the samples were calcined

insituinO2(20%)–Heat550◦C(3◦C/min)for2hwithaﬂowrate

of100mLNTP/min.Aftercoolingdownto150◦C,thesampleswere

submittedtoaC3H6-SCRexperimentfrom150to550◦C[18].The

samplesweresubsequentlyexposedtotheH2-C3H6-SCRfeedat

150◦C. H2 (2%/He),NO(4000ppm/He),C3H6 (2000ppm/He),O2

(100%)andHe(100%)werefedfromindependentcylinders(Air Liquide)withoutanyfurtherpuriﬁcationviamassﬂowcontrollers (Brooks5850TR).Typically,thecompositionoftheH2-NOx-C3H6

-O2-Hefeedwas:0.21%H2,385ppmNOx(∼96%NO),400ppmC3H6

and8%O2inHe,andthetotalﬂowratewas230mLNTP/min.The

temperaturewasthenincreasedstepwisefrom150to550◦Cwith 25◦Cincrementsandleftforabout1hateachtemperaturestep. Thereactoroutflowwasanalyzedusinga_␮-GC(Agilent Technolo-gies,CP4900)equippedwithtwochannels.Thefirstchannel,a5A molecularsievecolumn(80◦C,150kPaHe,200msinjectiontime, 30sbackflushtime),wasusedtoseparateH2,N2,O2andCO.The

secondchannel,equippedwithaporaplotQcolumn(60◦C,150kPa He,200msinjectiontime),wasusedtoseparateCO2,N2O,C3H6and

H2O.AchemiluminescenceNOxanalyzer(ThermoEnvironmental

Instruments42C-HT)allowedthesimultaneousdetectionofboth NOandNO2.NOxconversionstoN2andN2Owerecalculatedas

follows:

XNOxtoN2(%)=(2×[N2])/[NOx]inlet)×100 (1)

XNOxtoN2O(%)=(2×[N2O])/[NOx]inlet)×100 (2)

where[NOx]inlet,[N2]and[N2O]weretheconcentrationsinNOx

measuredattheinletofthereactorandinN2andN2Oattheoutlet

ofthereactor.C3H6conversionswerecalculatedonthebasisofthe

COx(CO+CO2)productsformed:

XC3H6(%)=([CO]+[CO2])/([C3H6]inlet×3)×100 (3)

where[CO],[CO2]and[C3H6]inletweretheconcentrationsofCO

andCO2 measuredattheoutletofthereactorandthatofC3H6

measuredattheinletofthereactor,respectively.

Thecomparisonofthecatalytic performancesofthe materi-alsinvestigatedinthepresentstudywasalsomadeonthebasis ofanefﬁciencycriterion(%)inthereductionofNOxtoN2inthe

150–550◦C range of temperatures. This criterion comparesthe catalyticperformancesoftheinvestigatedsamplestothoseofa catalystthatwouldallowforthefullreductionofNOxtoN2from

150to550◦C(100%efﬁciency).

2.3. Kineticmeasurements

Priortothekineticmeasurements,thesampleswerecalcined insituinO2(20%)–Heat550◦C(3◦C/min)for2hwithaﬂowrateof

100mLNTP/minandthetemperaturewascooleddownto325◦C.In

theH2-C3H6-SCRkineticinvestigations,theamountofAg(x)/Al2O3

catalystsdilutedinAl2O3,toobtaincatalystbedsof0.38gofthe

correspondingmechanicalmixtures,variedfrom1to5mgto main-taintheconversionsofNOx,C3H6andH2below13%,16%and30%,

respectively.Forcomparisonpurposes,C3H6-SCRkineticswasalso

studiedonamechanicalmixtureof0.15gofAg(0.7)/Al2O3(2.2wt%

Ag)and0.23gofAl2O3 at325and375◦C. Inthiscase,the

con-versionsofbothNOxandC3H6variedfrom5toabout30%when

thetemperaturewasincreasedfrom325to375◦C.At325◦C,the

conversionsofNOxandC3H6on0.38gofthebareAl2O3supportin

theC3H6-SCRandH2-C3H6-SCRreactionswerefoundtobeabout

3%and2%,respectively.Afterbeingpretreated,thesampleswere contactedwiththestandardreactingmixture(0.21%H2,385ppm

NOx, 400ppmC3H6 and 8%O2 in He) usedin theH2-C3H6-SCR

runs(Section2.2)for1–2hat325◦C.Afterreachingsteady-state, thedeterminationofthekineticparameters(˛,ˇ, andEa)was

performedonthebasisofthefollowingpowerratelawequation:

rN2=k[NO]

˛_[C

3H6]ˇ[H2]=Aexp(−Ea/RT)[NO]˛[C3H6]ˇ[H2](4)

where rN2, k, [NO],[C3H6],[H2], ˛, ˇ,, A, Ea, Rand T arethe

rateofreductionofNOxtoN2(mol/sg),thekineticconstant,the

inletconcentrationsinNO,C3H6andH2,thereactionorderswith

respecttoNO,C3H6andH2,thepre-exponentialfactor,the

appar-entactivationenergy(J/mol),thegasconstant(8.314J/molK)and thereactiontemperature(K),respectively.

ThedeterminationofthereactionorderswithrespecttoNO, C3H6andH2wasperformedat325◦Cbyvaryingtheconcentration

ofoneofthesereactantswhiletheconcentrationsoftheothers werekeptconstant.Thefollowingrangesofconcentrationswere used:200–700ppmNO,200–700ppmC3H6and1400–2400ppm

H2. Note that theinﬂuence of O2 wasnot investigated and its

concentrationwas maintainedto 8%.The reaction orderswere determinedfromtheslopeofthestraightlinesobtainedbyplotting thelogarithmoftherateofN2productionasafunctionofthe

loga-rithmoftheconcentrationoftheinvestigatedreactant.Therateof N2production(Eq.(5))wasproportionaltotheconversionofNOx

toN2(Eq.(1))andcalculatedasfollows:

rN2=FNOxXNOxtoN2/W (5)

whereFNOx,XNOxtoN2 andWaretheNOxmolarﬂowrate(mol/s),

theconversionofNOxtoN2(Eq.(1))andthecatalystloading(g),

respectively.ItwasalsoveriﬁedthattheN2reactionratesobtained

initiallyunderthestandardconditions(0.21%H2,385ppmNOx,

400ppmC3H6and8%O2inHe)werenotaffectedbythechangesin

theconcentrationsinthereactingfeedsusedforthedetermination ofthedifferentreactionorders.

Apparentactivationenergieswereestimatedfromtheslopeof thestraightlinesobtainedintheArrhenius-typeplotsunderthe standardfeed(0.21%H2,385ppmNOx,400ppmC3H6and8%O2in

He)atreactiontemperaturesinthe325–375◦Crangewith incre-mentsof10◦C.

The presence of external and internal diffusion limitations wasveriﬁed accordingto thecriteria deﬁned by Koros–Nowak [26]. External diffusion limitations were checked by changing the amount of Ag(0.7)/Al2O3 introduced in the reactor while

keeping a constant ﬂow rateto catalyst loadingratio. The H2

-C3H6-SCR data obtained under the standard conditions (0.15g

Ag(0.7)/Al2O3+0.23gAl2O3andatotalﬂowrateof230mLNTP/min)

were compared to those obtained for an H2-C3H6-SCR

reac-tion in which the amounts Ag(0.7)/Al2O3 and Al2O3, and the

ﬂow rate werehalf of those of thestandard conditions (0.08g Ag(0.7)/Al2O3+0.11gAl2O3andatotalﬂowrateof115mLNTP/min)

(Fig.2a).Internaldiffusionlimitationswereinvestigatedby chang-ing the grain size of a Ag(0.8)/Al2O3 catalyst from 50–125 to

200–315_␮m(Fig.2b),whilekeepingaconstantﬂowrateto cat-alystloadingratio(0.12gAg(0.8)/Al2O3+0.26gAl2O3andatotal

ﬂowrateof230mLNTP/min).TheNOxandC3H6conversions

plot-tedinFig.2 indicatetheabsenceof both externaland internal diffusionlimitationsunderthepresentexperimentalconditions. ComparableexperimentscarriedoutfortheC3H6-SCRreactionalso

suggestedtheabsenceofdiffusionlimitationsintheabsenceofH2

(5)

0 50 100 200 0 400 600 Temperature (°C) Conversions (%) (a) C3H6 NOx 0 50 100 200 0 400 600 Temperature (°C) Conversions (%) (b) 200-315 125-200 50-125

Fig. 2.Examination ofthe external (a)and internal (b) diffusionlimitations

in the reduction of NOx to N2 (gray curves) and the oxidation of C3H6

to COx (blackcurves) inthe H2-C3H6-SCR reaction. Feed composition:0.21%

H2, 385ppm NOx, 400ppmC3H6, 8% O2 and He balance. Mechanical

mix-tures of (a) 0.15g Ag(0.7)/Al2O3+0.23g Al2O3 and 230mLNTP/min (—), and

0.08g Ag(0.7)/Al2O3+0.11g Al2O3 and 115mLNTP/min (---), and (b) 0.12g

Ag(0.8)/Al2O3+0.26gAl2O3and230mLNTP/minwith50–125and125–200(—),and

200–315␮m(---)particlesizes.

3. Resultsanddiscussion

3.1. TEMandHAADF-STEMcharacterizationofthepost-NOx-TPD

samples

A thorough TEM examination of Ag(0.6)/Al2O3 and

Ag(1.1)/Al2O3, previously characterized by NOx-TPD, could

notrevealunambiguouslythepresenceofAgand/orAg2Oclusters

onthese samples, although the presence of Ag wasconﬁrmed invariousareasonbothsamplesbymeansofEDSanalyses(not shown).AsthepoorcontrastofAgand/orAg2OclustersonAl2O3

mayaccountforthedifﬁcultyintheirobservation,further charac-terizationofthesamplesexhibitingAgsurfacedensitiesranging from 0.6 to1.1Ag/nm2 _was _performed_by _HAADF-STEM. _As _in

thecase oftheTEMinvestigation,itwasfoundtobeextremely difﬁculttoobserveanyAgand/orAg2O clustersonallsamples

withHAADF-STEM.Agand/orAg2Oclustersofabout1nmcould

thereby beseen on only one areaof theAg(0.6)/Al2O3 sample

(examplesofclustersarehighlightedinFig.3a).Interestinglywhen theAg(0.7)/Al2O3samplewasdepositedontotheTEMgridfrom

anethanolsuspension,largeAgparticles(>10nm)couldbeclearly seenonallareasinvestigated(Fig.3band c),whereas no com-parablelargerparticleswereobservedonthesamplesdeposited inadrypowderform.Thisobservationisconsistentwithearlier studiesofSayahetal.[27]inwhichitwasconcludedthattheuse ofethanoltodisperseAg/Al2O3 samplesonTEMgridsshouldbe

avoided,asdoingsuchleadstothereductionofhighlydispersed oxidizedAgspeciesbyethanolandthecorrespondingformation oflargeAgparticleswhicharefoundtobenotrepresentativeof thestateofAgintheinvestigatedsamples.Thelargersizeofthese particlesallowedEDScharacterization,atypicalexampleofwhich isshowninFig.3c.

Tosummarize,thecharacterizationof ourAg/Al2O3 samples

byelectronicmicroscopytechniquesshowsthatAgispresentin ahighlydispersedstateonAl2O3.Suchaconclusionisconsistent

withearlierXAS(X-rayabsorptionspectroscopy)studiesinwhich

Fig.3. HAADF-STEMimagesofpost-NOx-TPDsamples:(a)Ag(0.6)/Al2O3deposited

ontheTEMgridindrypowderform(withsomeexampleclustersringed)and(b)

Ag(0.7)/Al2O3depositedontheTEMgridwithanethanolsuspension.(c)EDS

ele-mentalmapoverlaidonto(b,dottedsquare)inwhichtheAlandAgsignalsare

shownrespectivelyinredandgreen.(Forinterpretationofreferencestocolorin

thisﬁgurelegend,thereaderisreferredtothewebversionofthearticle).

itwasshownthatAgwasalmostatomicallydispersedafter calci-nationwhensupportedonAl2O3[25,28].

3.2. Catalyticperformances

TheinﬂuenceoftheadditionofH2 ontheC3H6-SCRcatalytic

performancesoftheAg(x)/Al2O3 catalystsisillustratedinFig.4.

(6)

0 50 100 0 200 400 600 Temperature (°C) Conversions (%) _0.0 0.1 0.2 H2 (%) 0 50 100 0 200 400 600 Temperature (°C) Conversions (%) _0.0 0.1 0.2 H2 (%) 0 50 100 0 200 400 600 Temperature (°C) Conversions (%) _0.0 0.1 0.2 H2 (%) 0 50 100 0 200 400 600 Temperature (°C) Conversions (%) _0.0 0.1 0.2 H2 (%) C3H6 NOx H2 0 50 100 0 200 400 600 Temperature (°C) Conversions (% ) 0.0 0.1 0.2 H2 (% ) 0 50 100 0 200 400 600 Temperature (°C) Conversions (% ) 0.0 0.1 0.2 H2 (% )

(a)

(c)

(b)

(d)

(f)

(e)

Fig.4. Inﬂuenceoftheadditionof0.21%H2ontheconversionsofNOxtoN2(—)andC3H6toCOx(---)(C3H6-SCR(,)andH2-C3H6-SCR(䊉,)reactions)on(a)Ag(0.3)/Al2O3,

(b)Ag(0.4)/Al2O3,(c)Ag(0.6)/Al2O3,(d)Ag(0.7)/Al2O3,(e)Ag(0.8)/Al2O3and(f)Ag(1.1)/Al2O3formechanicalmixturesofAg(x)/Al2O3catalystsandAl2O3forwhichtheamount

ofAgwaskeptessentiallyconstantto30.9±1.2␮mol.Feedcompositions:0%or0.21%H2,385ppmNOx,400ppmC3H6,8%O2andHebalancewitha230mLNTP/minﬂow

rate.TheconcentrationofH2inthecourseoftheH2-C3H6-SCRreactionisalsoshown( ).

H2drasticallypromotedtheC3H6-SCRreactionfortemperatures

lowerthan400◦C.Overall,theadditionof0.21%H2 in thefeed

shiftedtheNOxtoN2andC3H6toCOxconversionstotemperatures

approximately150◦ClowerthanthoseobtainedintheC3H6-SCR

reaction.OnAg(0.7)/Al2O3 (Fig.4d),theconversions ofNOxand

C3H6increasedsteeplyfrom150to225◦CinthepresenceofH2in

thefeed,whilethissamplehardlycatalyzedtheC3H6-SCRreaction

from150to300◦C.Fortemperaturesincreasingfrom225to400◦C, theconversionsofNOxandC3H6alsoincreasedinthepresenceof

H2buttoamuchmorelimitedextentthanfrom150to225◦C.The

pseudoplateauobservedintheconversionsofNOxandC3H6 in

the225–400◦Cregioncorrespondedtoatemperaturedomainfor whichH2wasfullyconsumed(Fig.4d).Fortemperatureshigher

than400◦C,theconversionsofNOxandC3H6werenolonger

sig-niﬁcantlyinﬂuencedbythepresenceofH2inthefeed.Comparable

commentscanbemadeontheotherinvestigatedsamples(Fig.4). OnAg(1.1)/Al2O3,however,thecatalyticperformanceswerefound

tobehigherinC3H6-SCRcomparedtothoseinH2-C3H6-SCRfrom

400to500◦C(Fig.4f).

ThecatalyticperformancesoftheAg(x)/Al2O3 samplesinthe

H2-C3H6-SCR of NOx are compared in Fig. 5. It can be seen

thattheNOxreductiontemperaturewindowbroadenedtolower

temperaturesas the Ag surface density (Ag loading) increased (Fig. 5a). In parallel, themaximum in N2Oconversion and the

conversionofC3H6 shiftedtolowertemperaturesastheAg

sur-face density increased (Fig. 5b and c). The broadening of the NOxreductionactivitytolowertemperaturesoccurred,however,

at theexpense of the NOx conversions at the higher

tempera-tures(Fig.5a).Thisillustrateshowcomplexisthecomparisonof

thecatalyticperformancesofAg/Al2O3 catalystswithvariousAg

surfacedensities(Agloadings)overabroadtemperaturedomain (150–550◦C).

Fig.6ashowsthatplottingtheNOxconversionsasafunction

oftheAgsurfacedensityatagivenreactiontemperaturedoesnot allowforaneasycomparisonofthesamples.Indeed,itcanbeseen inthisﬁgurethattheAgsurfacedensityprovidingoptimum con-versionofNOxtoN2intheH2-C3H6-SCRreactiondecreasesfrom

0.8to0.3Ag/nm2

Al2O3 asthetemperatureincreasesfrom200to

300◦C.Thismaybeattributedtodifferencesintheconversionsof C3H6andH2,inparticular,andtothefactthattheseconversions

variedsigniﬁcantlyonthesamplesinvestigated(Fig.4).The com-parisonoftheSCRperformancesofaseriesofcatalystsatagiven reactiontemperatureshouldbemadeideallywiththesamelevel ofconversionsforallreactants.Notethatthisisparticularly chal-lengingwhentworeactantsofverydifferentreactivityareused, suchasH2andhydrocarbonsintheH2-HC-SCRreaction.InaH2

-C3H8-SCRstudyperformedat300◦C,Shimizuetal.[25]suggested

theexistenceofanoptimumloadingofAgatabout2wt%.Despite thefactthattheseauthorsreportedthatthecorrespondingH2

-C3H8-SCRreactionratesweremeasuredunderconditionswhere

conversionswerebelow30%,itisuncertainthattheterm “con-versions” alsoincludedthat of H2.It mustbe emphasizedthat

mostoftheliteraturedataintheH2-HC-SCRﬁeldusuallydonot

reportontheconversionofH2.Inagreementwiththedatashown

inFig.4,Richteretal.,whowereamongtheveryfewauthorsto providetheH2conversions,alsoobservedanincreaseintheH2

con-sumptionintheH2-C3H8-SCRreactionastheAgloadingincreased

(7)

0

50

100

0

200

400

600 Temperature (°C)

NO

x

to N

2

(%

)

Ag(0.3)/Al2O3 _(a) Ag(1.1) Al2O3

0

5

10

15

20

0

200

400

600 Temperature (°C)

NO

x

to

N

2

O (

%)

(b)

0

50

100

0

200

400

600 Tempe

rature

(°C)

C

3

H

6

t

o

C

O

x

(%

)

(c)

Fig.5.InﬂuenceofAgsurfacedensityofAg(x)/Al2O3catalystsontheconversions

of(a)NOxtoN2,(b)NOxtoN2Oand(c)C3H6toCOxintheH2-C3H6-SCRofNOxon

Al2O3( ),andAg(0.3)/Al2O3(black),Ag(0.4)/Al2O3(gray),Ag(0.6)/Al2O3(blue),

Ag(0.7)/Al2O3(red),Ag(0.8)/Al2O3(purple)andAg(1.1)/Al2O3(green)for

mechani-calmixturesofAg(x)/Al2O3catalystsandAl2O3forwhichtheamountofAgwaskept

essentiallyconstantto30.9±1.2␮mol.Feedcomposition:0.21%H2,385ppmNOx,

400ppmC3H6,8%O2andHebalancewitha230mLNTP/minﬂowrate.(For

interpre-tationofreferencestocolorinthisﬁgurelegend,thereaderisreferredtotheweb

versionofthearticle).

Thecomparisonofthecatalyticperformancesofthematerials investigatedinthepresentstudywasalsomadeonthebasisofan efﬁciencycriterioninthereductionofNOxtoN2inthe150–550◦C

range oftemperatures. AsillustratedinFig.6b,this criterionis deﬁned,foragivencatalyst,astheratiooftheareaundertheNOxto

N2conversioncurvebetween150and550◦Ctotheareaunderthe

samecurveassuming100%conversion,forthesamerangeof tem-perature(i.e.theareaofthedottedsquareinFig.6b).Thiscriterion isexpressedasapercentage.Fig.6cﬁrstlyshowsthattheuseofsuch acriterionallowsconcludingtoanoptimumAgsurfacedensityof 0.7Ag/nm2

Al2O3 (Agloadingof2.2wt%)intheC3H6-SCRreaction

(opensymbols).Thisisinagreementwithearlierliteraturereports inwhichitwasclaimedthatoptimumNOxreductionactivitywas

obtainedforAg/Al2O3sampleswithanAgloadingcloseto2wt%

[3,18–24].Fig.6calsoclearlyillustratesthatthecatalytic perform-ancesintheC3H6-SCRoftheAg(x)/Al2O3samplesweredrastically

promotedbytheadditionofH2,astheNOxreductiontoN2

efﬁ-cienciesintheH2-C3H6-SCRreaction(fullsymbols:55–32%)were

foundtobemuchhigherthanthoseintheC3H6-SCRreaction(open

symbols:20–27%). IncontrasttotheC3H6-SCRreaction,Fig.6c

allowsconcludingthattheconceptofoptimumAgsurface den-sity(Agloading)doesnotapplytotheH2-C3H6-SCRreactionover

abroadtemperaturedomain(150–550◦C),sinceincreasingNOx

reductionefﬁciencieswereobtainedwithdecreasingAgsurface densities.Thisconclusiondifferssubstantiallyfromthosereported bySadhokinaetal.[24]andShimizuetal.[25]forwhoman opti-mumAgloadingcloseto2wt%wasidentiﬁedintheH2-promoted

C6H14-andC3H8-SCRreactions,respectively.Itcannotbeexcluded

thatthenatureofthereducing hydrocarbons[2]maybeatthe originofthediscrepancywithintheresultsofthepresentwork.It mustalsoberecalledthatcontrarytotheprocedurefollowedinthe presentwork(constantsilvercontentsinthealiquotsofsamples tested),thedatareportedbySadhokinaetal.[24]andShimizuetal. [25]wereobtainedwithAg/Al2O3aliquotsinwhichtheamounts

ofAgvaried,thusbringingadditionalcomplexityinthe interpre-tationofthedata.Tomimictheexperimentalconditionsusedby Sadhokinaetal.[24]andShimizuetal.[25],thecatalytic perform-ancesofaliquotsofsamplesinwhichtheamountsofAg(x)/Al2O3

catalysts(withx=0.3and0.6Ag/nm2

Al2O3)dilutedwithAl2O3were

keptconstant(0.19gAg(x)/Al2O3+0.19gAl2O3)wereinvestigated

0 20 40 60 80 0.0 0.4 0.8 1.2

Ag surface density (Ag/nm2Al2O3)

Conversion of NO x to N 2 (%) 200 °C 225 °C 250 °C 275 °C 300 °C _(a) 0 50 100 Temperature (°C) Co nv ers io n of NO x to N 2 (% ) (b) 150 550 30 40 50 60 0.0 0.4 0.8 1.2

NO x reduction to N 2 efficiency (%) 19 24 (c)

Fig.6. (a)conversionofNOxtoN2intheH2-C3H6-SCRreactionforgivenreactiontemperatures:200(––),225(–䊉–),250(––),275(––)and300◦C(----),(b)description

oftheNOxreductionefﬁciencycriterionintheH2-C3H6-SCRreactiononAg(0.6)/Al2O3and(c)NOxreductiontoN2efﬁcienciesintheC3H6-SCR(----)andH2-C3H6-SCR

(–䊉–)reactionsinthe150–550◦_C_range_of_temperatures_as_a_function_of_the_Ag_surface_density_for_0.38_g_of_mechanical_mixtures_of_Ag(x)/Al₂_O₃_and_Al₂_O₃_for_which_the

amountofAgwaskeptessentiallyconstantto30.9±1.2␮mol.Feedcompositions:0%or0.21%H2,385ppmNOx,400ppmC3H6,8%O2andHebalancewitha230mLNTP/min

(8)

0

50

100

200

0

40

0

600 Temperature (°C

)

NO

x

to N

2

(%)

(a) Ag(0.3)/Al2O3 + Al2O3 Ag(0.6)/Al2O3 + Al2O3

0

50

100

200

0

40

0

600 Temperature (°C

)

C

3

H

6

to CO

x

(%)

(b)

Fig.7. InﬂuenceoftheamountsofAgintroducedontheconversionsof(a)NOxtoN2

and(b)C3H6toCOxintheH2-C3H6-SCRofNOxon(—)0.19gAg(0.3)/Al2O3+0.19g

Al2O3(15.4␮molAg)and(---)0.19gAg(0.6)/Al2O3+0.19gAl2O3(31.9␮molAg).

Feedcomposition:0.21%H2,385ppmNOx,400ppmC3H6,8%O2andHebalance

witha230mLNTP/minﬂowrate.

(Fig.7).Asalreadynoticedonaliquotsofsamplesforwhichthe amountsofAgwaskeptessentiallyconstant(Fig.5aandc),Fig.7 showsabroadeningintheNOxreductiontemperaturewindowto

lowertemperaturesattheexpenseoftheNOxconversionsatthe

highertemperaturesandashiftintheconversionofC3H6tolower

temperaturesastheAgsurfacedensity(Agloading)increased.As a consequence,thehighly-loadedsample (Ag(0.6)/Al2O3)is the

moreactive intheNOx reduction toN2 at temperaturesbelow

290◦C,whereasitbecomeslessactivethanthelowly-loaded sam-ple(Ag(0.3)/Al2O3)at higher temperatures (Fig.7a). Moreover,

despitethefactthattheamountofAginAg(0.3)/Al2O3(15.4␮mol)

wastwicelowerthanthatinAg(0.6)/Al2O3(31.9␮mol),theNOx

reductiontoN2 efﬁciencyestimatedforAg(0.3)/Al2O3 (52%)was

foundtobeslightlyhigherthanthatofAg(0.6)/Al2O3(48%).Thus,it

appearsthattheAgsurfacedensity(Agloading)forwhichoptimum catalyticperformancesinH2-C3H6-SCRwereachievedwasstrongly

dependentontheselectedreactiontemperature(Figs.6aand7a).

3.3. Kineticinvestigations

HC-SCRprocesseshavebeenthesubjectofaverylimited num-berof kinetic studies[5,20,30–41]. Among thesestudies, those thathaveaimedatinvestigatingthekineticsoftheH2-promoted

HC-SCRareevenscarcer[30,34–36,40]althoughkineticshasbeen showntobeextremelyproﬁtableinprovidinguniqueinformation abouttheunderstandingofthecatalyticreactionsatamolecular level[26].

Kineticmeasurementswereperformedtogainfurtherinsights intotheoriginoftheimprovedH2-C3H6-SCRcatalytic

perform-ancesobtainedwithAg(x)/Al2O3samplesofdecreasingAgsurface

density(Figs.5aand6c(–䊉–)). Fig.8ashows thattheNO reac-tion order increased from0.1 to about0.5 up toa Ag surface densityof0.7Ag/nm2

Al2O3 andthenlevelsoffathigherAgsurface

densities.Fig.8bindicatesthattheC3H6reactionorderremained

essentiallyconstant (0.4)for Agsurfacedensitieslower thanor equalto0.7Ag/nm2

Al2O3andthenincreasedbeforelevelingoff(0.8)

forAgsurfacedensitieshigherthanorequal to0.9Ag/nm2 Al2O3.

Fig.8cshows that theH2 reactionorder (0.5) remained

essen-tially constant for Ag surface densities lower than or equal to

0 1 2 3 0 0.2 0.4 0.6 NO reaction orde r (a) Ag loading (wt%) 0 1 2 3 4 NO x uptak e (µ mol/n m 2 Al2O 3 ) 0 1 2 3 0 0.4 0.8 1.2 C3 H6 reaction orde r (b) 0 1 2 3 Al2O3 0.2 0.6 1.0 H2 reaction orde r (c) 0 1 2 3 0.0 0.4 0.8 1.2

0 30 60 90 Ea (kJ/mol ) (d)

Fig.8.Inﬂuence ofAgsurfacedensityofAg(x)/Al2O3 catalystsonthekinetic

parameters(a)NO(--♦--),(b)C3H6(----)and(c)H2(----)reactionorders,

and(d)apparentactivationenergies(Ea,--♦--)oftheH2-C6H6-SCRreaction.The

NOxuptakesdeterminedpreviouslyonthefreshlycalcinedsamples[18]arealso

reported(–䊉–).Forcomparisonpurposes,thereactionorderswithrespecttoNOand

C3H6andtheactivationenergymeasuredonAg(0.7)/Al2O3intheC3H6-SCRreaction

(absenceofH2)arealsoreported(circledsymbols)in(a),(b)and(c),respectively.

0.8Ag/nm2

Al2O3 and then increased upto 0.9 for anAg surface

densityof1.1Ag/nm2

Al2O3.Finally,Fig.8dshowsadecreaseinthe

apparentactivationenergyfrom61toabout26kJ/molwhenthe Agsurface densityincreasedfrom 0.3to0.7Ag/nm2

Al2O3 before

remaining essentially constant up to a Ag surface density of 1.1Ag/nm2

Al2O3.

Itisremarkablethatchangesinthereactionorderswithrespect toNO,C3H6andH2,andintheapparentactivationenergy(Ea)for

thereductionofNOxtoN2intheH2-C3H6-SCRreactionoccurredat

anAgsurfacedensity(Fig.8:∼0.7Ag/nm2

Al2O3=2.2wt%Ag)

coin-cidingwiththechangesobservedintheNOxuptakesoftheAl2O3

supportoftheAg(x)/Al2O3catalysts[18].

Thekineticparametersreportedin earlierHC-SCR investiga-tionsperformedonAg/Al2O3catalystsaresummarizedinTable1

intheorderofincreasingcarbonnumberintheHCreductant.For comparisonpurposes,theAgloadingshavealsobeenexpressed asAgsurfacedensitiesaccordingtoRef.[18].Itcanbeseenfrom thistablethatthesestudieshavemainlyfocusedonthe determina-tionoftheapparentactivationenergies(Ea)oftheHC-SCRprocess

(9)

Table1

ComparisonofthekineticdatareportedearlierintheHC-SCRandH2-HC-SCRofNOxonAl2O3andAg/Al2O3catalystswiththoseobtainedinthepresentstudy.

HC-SCRofNOx H2-HC-SCRofNOx Aga_(wt%) _ıb_(Ag/nm2 Al2O3) HC Ea c_(kJ/mol) _NOd _HCd _E ac(kJ/mol) NOd HCd Ref. 0.0 0.0 CH4 124 – – – – – [31] 0.8–21.5 0.2–4.8 CH4 95–124 – – – – [31] 2.0 0.7 C2H5OH 57±3 −0.28 0.34 – – – [32] 0.0 0.0 C2H5OH 47 – – – – – [33] 2.1 0.5 C2H5OH 39 – – – – – [33] 3.5 0.8 C2H5OH 37 – – – – – [33] 8.0 1.9 C2H5OH 39 – – – – – [33] 2.0 0.5

C2H5OH −C12H26− m−xylene 166 – – 116 – – [34] 3.8 1.1 123 – – 89 – – [34] 6.2 1.9 83 – – 65 – – [34] 0.0 0.0 C3H6 85 0.3 0.3 [35] 1.0 0.2 C3H8 147 – – 33 – – [30] 5.0 1.1 C3H8 250 – – 30 – – [30] 2.0 0.4 C3H8 224 –2.53 1.91 61 0.49 0.76 [36] 2.0 0.6 C6H14 67 0.8 0.9 – – – [20] 2.0 0.6 C8H18 – ≥0 ≥0 – – – [5] 2.0 0.6 C8H18 – 0 >1.0 – – – [37] 2.0 0.6 C8H18 – 0 1.0 – – – [38] 1.5 1.2 C8H18 – 0.5–0.7 0–1.7 – – – [39] 1.9 0.6 C8H18 – −0.39 – – 0.14–0.40 0.26–0.40 [40] 1.9 0.6 C16H34 – 0.5–0.6 ≥0 – – – [41] 2.2 0.7 C3H6 105 0.0 1.0 26 0.4 0.4 This 0.5–4.3 0.1–1.4 C3H6 – – – 61–23 0.1–0.5 0.4–0.8 study

a_Ag_loading_in_the_Ag/Al

2O3catalysts.

b_Ag_surface_density_corrected_for_the_content_of_Ag_as_Ag

2OfollowingtheproceduredescribedinRef.[18].

c _Apparent_activation_energies.

d _Reaction_orders_with_respect_to_NO_and_HC_for_the_production_of_N

2.

H2-promotedHC-SCR.Thekineticdatareportedfortheﬁrsttimein

Fig.8fortheH2-C3H6-SCRreactionsoveranextendedrangeofAg

surfacedensitiesare,therefore,oftheutmostinterest.Itcanalso bededucedfromTable1thatthenatureofthehydrocarbonshasa signiﬁcantinﬂuenceonthekineticparametersoftheHC-SCR reac-tion.Nevertheless,somecommontrendscanbedrawnfromthese setsofdata.

Inparticular, Richteret al.[30], Kimet al.[34] andShimizu etal.[36]reportedonasigniﬁcantdecreaseinEawiththe

pro-motionoftheHC-SCRreactionbyH2 (Table1).Thisﬁndingcan

alsobeobserved inFig.8dwhere theEa oftheC3H6-SCR

reac-tiondecreasestremendouslyfrom105(circledfulldiamond)to 26kJ/mol(opendiamond)whenpromotedbyH2onAg(0.7)/Al2O3.

Regarding theevolution of Ea with theAg loading (Agsurface

density),conﬂictingresultswerereportedbyKimetal.inC2H5

OH-C12H26-m-xylene-SCR[34]andRichteretal.inC3H8-SCR[30].Kim

etal.[34]concludedtoadecreaseinEa,whereasRichteretal.[30]

reportedoneitheranincrease inEa inC3H8-SCRorconstantEa

inH2-C3H8-SCRwithincreasingloadingsofAg(Table1).Thedata

plottedinFig.8dindicateadecreaseinEaoftheH2-C3H6-SCR

reac-tionwithincreasingAgloadings,butinamorerestrictedrangeof Agsurfacedensities(0.3–0.7Ag/nm2

Al2O3)thanthatreportedby

Kimetal.(0.5–1.9Ag/nm2

Al2O3)[34].

Regardingthereactionorders,earlierstudiesconcludedthatthe promotionoftheHC-SCRreactionbyH2ledtoasigniﬁcantincrease

intheNOreactionorder[36,40]andadecreaseinthereactionorder inC3H8[36]foragivenAgloading(0.4–0.6Ag/nm2_Al₂_O₃)(Table1).

ThedatashowninFig.8aandbforAg(0.7)/Al2O3areconsistent

withtheseﬁndings.TheNOreactionorder(Fig.8a)increasedfrom 0(C3H6-SCR,circledfulldiamond)to0.4(H2-C3H6-SCR,open

dia-mond),whereastheC3H6reactionorder(Fig.8b)decreasedfrom

1.0(C3H6-SCR,circledfulltriangle)to0.4(H2-C3H6-SCR,open

trian-gle).Toourknowledge,thetrendsintheNO,C3H6andH2reaction

ordersshowninFig.8a–cforanH2-HC-SCRreactionhavenotbeen

reportedtodate.

Beforeprovidinginterpretationfortheevolutionsofthekinetic parameterswithincreasingAgsurfacedensities(Fig.8),themain conclusions of thecharacterization of theAg(x)/Al2O3 catalysts

by the NOx-TPD method, and the associated evolution of the

NOxuptakesofthesupportingAl2O3 oxide(NOxspeciesdonot

chemisorbonAgspecies)[18],shallberecalled.Thedecreasein theNOx uptakesobserved withincreasing Agsurface densities

uptoanAgdensityofabout0.7Ag/nm2

Al2O3([18]andFig.8),for

whichoptimumC3H6-SCRactivitywasobtained(Fig.6c,----),

wasinterpretedasthemaximumloadingofsilverperunitsurface areaofAl2O3forwhichoptimalAgdispersionwaspreserved[18].

TheobservedlineardecreaseintheNOxuptakebelowanAgsurface

densityofabout0.7Ag/nm2

Al2O3 wasalsoattributedtothe

forma-tionofhomogeneouslydistributedAgspeciesofincreasingdensity [18].ForAgsurfacedensitieshigherthan0.7Ag/nm2

Al2O3,theNOx

uptakesoftheAl2O3supportingoxideremainedessentially

con-stant(Fig.8).Thiswasassignedtoanincreaseinthesizeofthe Agclusters,theAl2O3 surfacesitesontowhichAgwasanchored

beingsaturatedforanAgsurfacedensityofabout0.7Ag/nm2 Al2O3

[18].DespitetheNOx-TPDcharacterizationoftheAg(x)/Al2O3

cat-alystswasdoneoncalcinedsamples[18],thecorrelationsobtained inthepresentworkbetweentheNOxuptakesandthechangesin

thekineticparametersoftheH2-HC-SCRreaction(Fig.8)suggest

thattheAgspeciesinthecalcinedsamplesmaybetakenas repre-sentativeofthoseexistingunderthereactionconditionsfroman Agdispersionpointofview.Thisassumptionissupportedbythe EXAFSstudiesperformedbyShimizuetal.[25]andBurchetal. [28]onAg/Al2O3sampleswithAgloadingsofabout2wt%(Ag

sur-facedensityof0_.4Ag_/nm2

Al2O3 in[25])whichconcludedtoavery

limitedsinteringoftheAgphaseunderHC-SCRreactionconditions, promotedornotbyH2,andtothepreservationofAginahighly

dispersedstatewiththeformationofAgclustersmadeof3–4Ag atomsforthisparticularAgloading.Theseearlierstudies,together withthecharacterizationofthenewly-preparedsamplesbyNOx

(10)

forwhich Ag wasshown toremainin a highlydispersed state whateverthesilverloadinginthedrypowdersamples,provide supportfortheuseoftheAgsurfacedensityconceptinthepresent study.

Fig.5cclearlydemonstratesthattheactivationofC3H6intheH2

-C3H6-SCRreactionoccurredonAgsites,asthebareAl2O3support

doesnotcatalyzeC3H6oxidationattemperaturesaslowasthose

observedfortheAg(x)/Al2O3samples.Themuchlowerconversion

ofH2onAl2O3(<5%at500◦C,notshown)thanthosemeasuredon

theAg(x)/Al2O3catalysts(100%above325◦C,Fig.4)alsoindicates

thattheactivationofH2occurredonAgspecies.Theseconclusions

areconsistentwiththerecentlypublishedelegantstudyofKim etal.inwhichthepromotionaleffectofH2inHC-SCRwasattributed

tomorphologicalandchemicalchangesoftheAgphases[34].ForAg surfacedensitieslowerthanorequalto0.7Ag/nm2

Al2O3,the

reac-tionordersof0.4inC3H6(Fig.8b)and0.5inH2(Fig.8c)remained

essentiallyconstantandpositive.WithinthesamerangeofAg sur-facedensities(<0.7Ag/nm2

Al2O3),theNOreactionorderincreased

from0.1to0.4(Fig.8a).ThisindicatesthattheactivatedformofNO (NOxadsorbedspecies:NOxadspecies)doesnotcompetefortheAg

sitesresponsiblefortheactivationofC3H6andH2,assucha

compe-titionshouldhaveresultedinchangesinthereactionordersinC3H6

andH2,whichwasnotobserved(Fig.8bandc).Asaconsequence,

thesitesresponsiblefortheadsorptionofNOxareratherrelatedto

theAl2O3 supportingoxide.Thisproposalisconsistentwith

ear-lierFTIRstudiesinwhichitwasshownthatsigniﬁcantamounts ofNOxwerestoredontheAl2O3supportingoxide[12,34,36,42,43]

anditwassuggestedthatnitrateswouldbereactionintermediates ofHC-SCR[20].

TheincreaseintheNOreactionorderwiththeadditionofH2

onAg(0.7)/Al2O3(Fig.8a),from0fortheC3H6-SCRto0.4forthe

H2-C3H6-SCR(Fig.8a),couldbeassignedtoadepletioninthe

cov-erageoftheNOxadspecies[2,36] duetothedrasticincrease in

theH2-C3H6-SCRreactionratecomparedtothatofC3H6-SCRat

325◦C (Fig.4d).Themuchlower C3H6-SCRreactionratewould

thusallowforsaturationcoverageoftheAl2O3sitesbytheNOx

adspecieswhichwouldresultintheobserved0threactionorder withrespecttoNO.Thisexplanationwouldbeconsistentwiththe decreaseinnitratecoveragewiththeintroductionofH2observed

byShimizuetal.[36].Likewise,theincreaseintheNOreaction orderas theAgsurface density increases(Fig.8a)is attributed toadecreaseinthecoverageofthecatalystsurfacebytheNOx

adspecieswhichiscoherentwiththeobserveddecreaseintheNOx

uptakesoftheAg(x)/Al2O3samplesforAgsurfacedensitieslower

than0.7Ag/nm2

Al2O3.TheobservedlevelingoffoftheNOreaction

orderforAgsurfacedensitieshigherthan0.7Ag/nm2

Al2O3 canbe

associatedtotheconstantcoverageofthecatalystsurfacebythe NOxadspecies,inagreementwiththesteadinessoftheNOxuptakes

observedinthisAgsurfacedensitydomain(Fig.8a).Itwouldhave beenoftheutmostinteresttohavebeenabletoprovideadditional quantitativedatasupportingthesuggesteddifferencesinNOx

cov-erageaccountingforthechangesobservedintheNOreactionorder. OnemayhavethoughtabouttheuseoftheSSITKA(Steady-State Isotopic Transient Kinetic Analysis) techniquesto providesuch information[44].Recently,Burchandco-workers[45,46]put par-ticularemphasisonthefactthatconventionalSSITKAshouldbe usedwithanextremecautioninidentifyingtruereaction inter-mediatesin theH2-C8H18-SCRreaction. Thesedifﬁcultiescould

beovercome for isocyanate intermediates by using short time onstreamSSITKA(STOS-SSITKA)insteadofconventionalSSITKA [45,46].Whiletheinvolvementofnitrate-typespeciesadsorbed onorclosetotheactiveAgsitescouldberevealedbySTOS-SSITKA intheH2-NH3-SCRreaction [47],nosuchinformationcouldbe

obtainedinthemorecomplexH2-C8H18-SCRreaction[46].These

argumentsthereforeclearlypreventtheuseofSTOS-SSITKAinthe

Fig.9. SchematicrepresentationoftheH2-C3H6-SCRreactiononAg(x)/Al2O3

cata-lystswithincreasingAgsurfacedensities(x).

estimationoftheNOxad-speciesinvolvedintheH2-C3H6-SCR

pro-cessofthepresentstudy.Finally,exhaustsgenerallycontainhigh concentrationsofCO2andH2O(5–12%).AstheNOxspecieswere

deducedtobechemisorbedonAl2O3inthepresentwork,itcannot

beexcludedthatthepresenceofCO2andH2Ointhefeedwould

leadtoadecreaseintheperformancesoftheAg/Al2O3catalysts

intheSCRreactionsduetotheircompetitiveadsorptionwiththe NOxspecies.Inagreementwiththis,theinhibitingeffectofH2O

hasbeenhighlightedpreviouslyontheC3H6-SCR[48–50]andH2

-C3H8-SCR[30]performances.

TheincreaseintheC3H6andH2reactionordersforAgsurface

densitieshigherthan0.7Ag/nm2

Al2O3(Fig.8bandc)canbeassigned

tothefactthatthecombustionofthesemoleculesonthelargerAg nanoparticles,presentatsuchhighAgloadingsandwhich oxida-tioncapabilitieshavebeenclearlyillustrated[51],prevailsover theirefﬁcientuseintheSCRreactionfortheproductionofN2[3].

Thedecreaseintheapparentactivationenergy(Ea)with

increas-ing Ag surface densities up to 0.7Ag/nm2

Al2O3, also reported

recentlybyKimetal.[34],ismoredifﬁculttoexplainasthehigher Ea(Fig.8d)werefoundonthemoreactivesamplesinH2-C3H6-SCR

(Fig.5a).AsrecalledbyRichteretal.[30]andBondetal.[52],the rateconstantk(Eq.(4))notonlyvarieswithEabutalsowiththe

pre-exponentialfactor(A),whichitselfdependsontheconcentrationof activesites.Thispeculiarityisdeﬁnedastheso-called compensa-tionphenomenonincatalysis[52].Theunexpectedhigheractivity intheH2-C3H6-SCRreactionoftheAg(x)/Al2O3sampleswiththe

lowerAgsurfacedensities,associatedtotheirhigherEa,mightbe

relatedtothehighernumberofactivesitesontheAl2O3

suppor-tingoxideabletochemisorbNOxspeciesinagreementwiththe

NOxuptakedata(Fig.8),hencetotheirhigherpre-exponential

fac-tors.YettheobservedchangesintheNO,C3H6 andH2 reaction

orders(Fig. 8a–c) prevented the estimation of the correspond-ingpre-exponentialfactors.ForAgsurfacedensitieshigherthan orequal to0.7Ag/nm2

Al2O3,Ea remainedessentiallyconstantas

weretheNOx uptakes(Fig.8d),and therefore alsothenumber

ofactivesitesontheAl2O3 supportingoxideabletochemisorb

NOxspecies andthecorrespondingpre-exponentialfactors.The

lower catalytic performancesof thesesamples (Fig. 4b)is thus ratherassignedtothecombustionofC3H6onlargerAg

nanopar-ticles[3] formedattheseparticularlyhighAgsurface densities [18].

Insummary,theinterpretationoftheactivityandkineticdata ledus toconcludethat theH2-C3H6-SCR reactionproceedsvia

theactivationofH2 andC3H6onAgspecieswhichfurtherreact

withNOxadspeciesactivatedontheAl2O3support,as

schemat-icallyillustratedinFig.9.Suchaproposalisconsistentwiththe conclusionsdrawnearlierbySheandFlytzani–Stephanopouloson CH4-SCRstudies[31].TheunexpecteddecreaseintheH2-C3H6

(11)

activesitesontheAl2O3supportingoxideabletochemisorbNOx

species[18] for Agsurfacedensities lowerthan0.7Ag/nm2 Al2O3

andthentothecombustionofC3H6 onthelargerAg

nanoparti-clesforAgsurfacedensitieshigherthan0.7Ag/nm2

Al2O3 (Fig.9).

Finally,thekineticdataobtainedforAgsurface densitieslower than 0.7Ag/nm2

Al2O3, in particular, suggest that theinteraction

between theNOx and C3H6 would be rate determining in the

C3H6-SCRprocessinagreementwiththeearlierproposalofBurch

etal.[49].

4. Conclusion

The promotional effect of H2 in the C3H6-SCR reaction was

conﬁrmedonaseriesofAg(x)/Al2O3samplesexhibitingvarious

Agsurfacedensities(0.3<x<1.1Ag/nm2

Al2O3).TEMand

HAADF-STEManalysesoftheAg(x)/Al2O3 catalystsindicatedthat silver

wasinahighlydispersedstate,whateverthemetalloading.The introductionofH2 resultedin(i)abroadeningintheNOx

reduc-tiontemperaturewindowtolowertemperaturesattheexpenseof theNOxconversionsattemperatureshigherthan300◦Cand(ii)

a shiftintheconversion ofC3H6 tolowertemperaturesasthe

Agsurface densityincreased.In contrasttotheC3H6-SCR

reac-tion,theconceptofoptimumAgsurfacedensityat0.7Ag/nm2 Al2O3

(Agloadingof 2.2wt%) [18] didnot applyto theH2-promoted

C3H6-SCR reaction over a broad temperature domain. Indeed,

the catalytic performances in the H2-C3H6-SCR reaction in the

150–550◦C range of temperaturesimproved as the Ag surface densityoftheAg(x)/Al2O3 samplesdecreased.Adetailedkinetic

studywasperformedinwhich thereactionordersinNO,C3H6

and H2, and theapparentactivation energies weredetermined

ontheAg(x)/Al2O3 series.Remarkably,changes inthesekinetic

parameters werefoundtooccuratan Agsurface densityclose to 0.7Ag/nm2

Al2O3 (Ag loading of 2.2wt%) coinciding with the

changes observedearlier in theNOx uptakesof theAl2O3

sup-portingoxide[18].Interpretationoftheactivityandkineticdata ledus toconcludethat theH2-C3H6-SCR reactionproceedsvia

theactivationofH2 andC3H6onAgspecieswhichfurtherreact

withNOx adspecies activatedonthe Al2O3 support.The

unex-pecteddecreaseintheH2-C3H6-SCRperformancesobservedwith

increasing Ag surface densities was assigned to a decrease in number of active sites on the Al2O3 supporting oxide able to

chemisorbNOx species[18] forAgsurfacedensities lowerthan

0.7Ag/nm2

Al2O3 andtothecombustionofC3H6 onthelargerAg

nanoparticlesforAgsurfacedensitieshigherthan0.7Ag/nm2 Al2O3

(Fig. 9).Finally, thekinetic dataobtainedfor Ag surface densi-tieslowerthan0.7Ag/nm2

Al2O3 alsosuggestthattheinteraction

betweentheNOxandC3H6wouldberatedeterminingintheC3H6

-SCRprocess.

Acknowledgments

TCgratefully acknowledgesUPMCforfinancialsupport(PhD GrantNo.:322/2012).TheauthorswouldliketothanktheIMPC (InstitutdesMatériauxdeParisCentre)throughthe“plateforme demicroscopieélectronique”.Theauthorsalsoacknowledge finan-cialsupportfromCOSTActionMP0903Nanoalloys.Theaberration correctedSTEMinstrumentusedinthisworkwasfundedbythe BirminghamScienceCityproject. RLCandZYLacknowledgethe EngineeringandPhysicalSciencesResearchCouncilU.K.for finan-cialsupport(GrantNo.:EP/G070326/1).J.Olek(Engineer,Agilent

Technologies)isacknowledgedforhispromptandusefulassistance intheuseofthe␮-GCsystem.

References

[1]http://ec.europa.eu/environment/air/transport/road.htm

[2]R.Burch,Catal.Rev.Sci.Eng.46(2004)271.

[3]T.Miyadera,Appl.Catal.B:Environ.2(1993)199.

[4]F.Klingstedt,K.Eränen,L.-E.Lindfors,S.Andersson,L.Cider,C.Landberg,E. Jobson,L.Eriksson,T.Ilkenhans,D.Webster,Top.Catal.30/31(2004)27.

[5]K.-I.Shimizu,A.Satsuma,T.Hattori,Appl.Catal.B:Environ.25(2000)239.

[6]K.-I.Shimizu,J.Shibata,A.Satsuma,T.Hattori,Phys.Chem.Chem.Phys.3(2001) 880.

[7]S.Kameoka,T.Chaﬁk,Y.Ukisu,T.Miyadera,Catal.Lett.51(1998)11.

[8]C.Petitto,H.P.Mutin,G.Delahay,Appl.Catal.B:Environ.134–135(2013)258.

[9]S.Satokawa,Chem.Lett.29(2000)294.

[10]S.Satokawa,J.Shibata,K.-I.Shimizu,A.Satsuma,T.Hattori,Appl.Catal.B: Environ.42(2003)179.

[11]K.-I.Shimizu,A.Satsuma,Phys.Chem.Chem.Phys.8(2006)2677.

[12]R.Burch,J.P.Breen,C.J.Hill,B.Krutzch,B.Konrad,E.Jobson,L.Cider,K.Eränen, F.Klingstedt,L.-E.Lindfors,Top.Catal.30/31(2004)19.

[13]J.Shibata,Y.Takada,A.Shishi,S.Satokawa,A.Satsuma,T.Hattori,J.Catal.222 (2004)368.

[14]K.A.Bethke,H.H.Kung,J.Catal.172(1997)93.

[15]J.H.Lee,S.J.Schmieg,S.H.Oh,Appl.Catal.A:Gen.342(2008)78.

[16]H.Y.Law,J.Blanchard,X.Carrier,C.Thomas,J.Phys.Chem.C114(2010)9731.

[17]C.Thomas,J.Phys.Chem.C115(2011)2253.

[18]T.Chaieb,L.Delannoy,C.Louis,C.Thomas,Appl.Catal.B:Environ.142–143 (2013)780.

[19]T.E.Hoost,R.J.Kulda,K.M.Collins,M.S.Chattha,Appl.Catal.B:Environ.13 (1997)59.

[20]K.-I.Shimizu,J.Shibata,H.Yoshida,A.Satsuma,T.Hattori,Appl.Catal.B: Envi-ron.30(2001)151.

[21]L.-E.Lindfors,K.Eränen,F.Klingstedt,D.YuMurzin,Top.Catal.28(2004)185.

[22]K.Arve,L. ˇCapek,F.Klingstedt,K.Eränen,L.-E.Lindfors,D.Y.Murzin,J.Dˇedeˇcek, Z.Sobalik,B.Wichterlová,Top.Catal.30/31(2004)91.

[23]R.Zhang,S.Kaliaguine,Appl.Catal.B:Environ.78(2008)275.

[24]N.A.Sadokhina,A.F.Prokhorova,R.I.Kvon,I.S.Mashkovskii,G.O.Bragina,G.N. Baeva,V.I.Bukhtyarov,A.YuStakheev,Kinet.Catal.53(2012)107.

[25]K.-I.Shimizu,M.Tsuzuki,K.Kato,S.Yokota,K.Okumara,A.Satsuma,J.Phys. Chem.C111(2007)950.

[26]M.Boudart,G.Djéga-Mariadassou,KineticsofHeterogeneousCatalytic Reac-tions,PrincetonUniversityPress,Princeton,NJ,1984.

[27]E.Sayah,D.Brouri,Y.Wu,A.Musi,P.DaCosta,P.Massiani,Appl.Catal.A:Gen. 406(2011)94–101.

[28]J.P.Breen,R.Burch,C.Hardacre,C.J.Hill,J.Phys.Chem.B109(2005)4805.

[29]X.Zhang,Y.Yu,H.He,Appl.Catal.B:Environ.76(2007)241.

[30]M.Richter,U.Bentrup,R.Eckelt,M.Schneider,M.-M.Polh,R.Fricke,Appl.Catal. B:Environ.51(2004)261.

[31]X.She,M.Flytzani-Stephanopoulos,J.Catal.237(2006)79.

[32]W.L.Johnson,G.B.Fisher,T.J.Toops,Catal.Today184(2012)166.

[33]Y.Yan,Y.Yu,H.He,J.Zhao,J.Catal.293(2012)13.

[34]P.S.Kim,M.K.Kim,B.K.Cho,I.-S.Nam,S.H.Oh,J.Catal.301(2013)65.

[35]M.Haneda,Y.Kintaichi,H.Shimada,H.Hamada,J.Catal.192(2000)137.

[36]K.-I.Shimizu,J.Shibata,A.Satsuma,J.Catal.239(2006)402.

[37]K.Arve,F.Klingstedt,K.Eränen,J.Wärna,L.-E.Lindfors,D.Yu.Murzin,Chem. Eng.J.107(2005)215.

[38]K.Eränen,L.-E.Lindfors,F.Klingstedt,D.Yu.Murzin,J.Catal.219(2003)25.

[39]J.R.HernàndezCarucci,A.Kurman,H.Karhu,K.Arve,K.Eränen,J.Wärnå,T. Salmi,D.Yu.Murzin,Chem.Eng.J.154(2009)34.

[40]K.Arve,H.Backman,F.Klingstedt,K.Eränen,D.Yu.Murzin,Appl.Catal.A:Gen. 303(2006)96.

[41]K.Arve,J.R.HernàndezCarucci,K.Eränen,A.Aho,D.Yu.Murzin,Appl.Catal.B: Environ.90(2009)603.

[42]P.Sazama,L. ˇCapek,H.Drobná,Z.Sobalík,J.Dˇedeˇcek,K.Arve,B.Wichterlová, J.Catal.232(2005)302.

[43]S.Chansai,R.Burch,C.Hardacre,J.P.Breen,F.Meunier,J.Catal.276(2010)49.

[44]S.L.Shannon,J.G.Goodwin,Chem.Rev.95(1995)677.

[45]S.Chansai,R.Burch,C.Hardacre,J.Breen,F.Meunier,J.Catal.276(2010)49.

[46]S.Chansai,R.Burch,C.Hardacre,J.Breen,F.Meunier,J.Catal.281(2011)98.

[47]S.Chansai,R.Burch,C.Hardacre,J.Catal.295(2012)223.

[48]F.C.Meunier,R.Ukropec,C.Stapelton,J.R.H.Ross,Appl.Catal.B:Environ.30 (2001)163.

[49]R.Burch,J.P.Breen,F.C.Meunier,Appl.Catal.B:Environ.39(2002)283.

[50]S.T.Korhonen,A.M.Beale,M.A.Newton,B.M.Weckhuysen,J.Phys.Chem.C115 (2011)885.

[51]N.Bogdanchikova,F.C.Meunier,M.Avalos-Borja,J.P.Breen,A.Pestryakov,Appl. Catal.B:Environ.36(2002)287.