Effect of chlorine on performance of Pd catalysts prepared via colloidal immobilization

ContentslistsavailableatScienceDirect

Catalysis

Today

jou rn al h om ep a g e :w w w . e l s e v i e r . c o m / l o c a t e / c a t t o d

Effect

of

chlorine

on

performance

of

Pd

catalysts

prepared

via

colloidal

immobilization

Yingnan

Zhao

_,

_Wanwei

_Liang

_,

_Yongdan

_Li

_,

_Leon

_Lefferts

a_{CatalyticProcessesandMaterials,FacultyofScienceandTechnology,MESA+InstituteforNanotechnology,UniversityofTwente,Enschede,7500AE,}

TheNetherlands

b_{SchoolofChemicalEngineering,TianjinUniversity,Tianjin,300072,China}

c_{DepartmentofBiotechnologyandChemicalTechnology,SchoolofChemicalTechnology,AaltoUniversity,P.O.Box16100,00076Aalto,Finland}

a

r

t

i

c

l

e

i

n

f

o

Articlehistory:

Received8November2016 Receivedinrevisedform 28December2016 Accepted17January2017 Availableonline22January2017 Keywords: Pdcatalysts Nitritehydrogenation Chlorine Colloid Selectivity

a

b

s

t

r

a

c

t

ThiscontributionshowstheeffectofresidualchlorineonthecatalyticperformanceofaPd-basedcatalyst

inthehydrogenationofnitriteforcleaningofdrinkingwater.Thecatalystwaspreparedvia

immobiliza-tionacolloidalPdnanoparticlesusingactivatedcarbonassupport.Differentamountofhydrochloric

acid(HCl)wasaddedtoimmobilizethePdcolloidonthecarbonsupport,facilitatingtheremovalof

theresidualstabilizer,polyvinylalcohol(PVA),fromthesurfaceofthePdnanoparticles(NPs).The

cat-alystswerecharacterizedbyTEM,CO-chemisorption,XRF,N2physisorption,UV–visspectroscopy,and

XPS.Theactivityandselectivityofthecatalystsweremeasuredfornitritehydrogenationinsemi-batch

operation.TheresultsshowthatPVAcanberemovedcompletelyatpHbelow2.Theresidualchlorine

onthecatalystscanberemovedbyreductioninH2/N2atamildtemperature,i.e.200◦C,regardlessthe

amountofHClused.Nevertheless,highconcentrationofHClduringimmobilization(pH1)causes

par-tialPdre-dissolutionaccordingtoUV–visspectroscopy,resultinginformationofhighlydispersedPd

clustersthatcouldnotbedetectedwithTEM.Reductionofthiscatalystwithhighchlorinecontentin

H2at200◦CisresultinginformationofrelativelylargePdparticlesviasintering.Withoutpre-reduction

at200◦C,residualchlorinecanalsoberemovedalmostcompletelyduringthehydrogenationreaction

atroomtemperature.TheactivityofthePdcatalystisinsensitivetothechlorineconcentrationbelow

30␮molL−1intheaqueousreactionmixture.Interestingly,theselectivitytoN2isimprovedbyadding

chlorinetothereactionmixture,independentofthewaychlorineisadded,i.e.viathecatalystoradded

directlytothereactionsolution.

©2017TheAuthors.PublishedbyElsevierB.V.ThisisanopenaccessarticleundertheCCBYlicense

(http://creativecommons.org/licenses/by/4.0/).

1. Introduction

Nitrateandnitritecontaminationingroundwaterhasbecome arisingriskforsupplyingofdrinkingwater,especiallyin agricul-turalareaswheresyntheticnitrogenfertilizersareextensivelyused [1–3].Catalytichydrogenationofnitrate/nitrite(Eqs.(1)–(3))using Pd-basedcatalystshasbeendevelopedsincethelate1980s,asa highlyefficientmethodoperatedundermildconditions(typically around25◦Candambientpressure)[4,5].Thismethodprevents formationofanycontaminatingbrinesasinion-exchange

proce-∗ Correspondingauthor.

E-mailaddress:l.lefferts@utwente.nl(L.Lefferts).

dures,andisabletoconvertnitrateandnitriteinwaterlackingany organiccontamination,asisrequiredforbiologicaltreatment[6,7]. 2NO3-+2H2 Pd-Cu −−−−→ 2NO2+2H2O (1) 2NO2-+3H2+2H+ Pd −→ N2+4H2O (2) NO2-+3H2+2H+ Pd −→ NH4++2H2O (3)

Colloidalmethodhasbeendevelopedextensivelyfor prepara-tionofmetalnanoparticles(NPs)forcatalytic applicationinthe lastfewdecades[8,9].Theadvantageofthemethodiswell-known: thesizesoftheNPscanbewellcontrolledandmanipulated, facil-itatingstudiesonstructure-performancerelationships.Advanced methodsallowformationofNPswithwell-definedshapes, offer-inginterestingopportunitiesonevenmoredetailedstudiesonthe influenceofsurfacestructureoncatalysis[10,11].

http://dx.doi.org/10.1016/j.cattod.2017.01.028

Pdbasedcatalystsfornitrate/nitritehydrogenationhavebeen studiedusingcolloidpreparationmethods[11–16],reportingthat therateofnitritehydrogenationisindependentonPdparticlesize [14,17,18].Nevertheless,itisalsowellreportedthatresidual sta-bilizers,suchaspolyvinylalcohol(PVA)andpolyvinylpyrrolidone (PVP),can blockpartoftheactivesitesby coveringthe major-ityofthemetalsurface[14,19–22].Stabilizersandcappingagents canalsomanipulateadsorbedreactivespeciesonmetalsurface, influencingtheactivityaswellastheselectivityofPdcatalysts [14,22–25].Asaresult,itisgenerallypreferredtoremovethe resid-ualstabilizersonmetalsurface.

Removal of the residual stabilizers is challenging, normally includingoxidationandthermaltreatment,makingitdifficultto maintainparticlesizeandcrystalstructureoftheNPs,andstill,the removalcanbeincomplete[26–30].Inourpreviouswork[13],the Pdcolloidwasimmobilizedonactivated-carbon(AC)inaqueous HClsolutionatpH2.ItwasobservedthatPVAcouldbecompletely removedfromthePdsurface,accordingtoTEMandCO chemisorp-tionresults.ItwasproposedthatextensivechemisorptionofClon thePdsurface,promotedbythepresenceofO2accordingto

reac-tionsinEqs.(4)and(5),isweakeningtheinteractionbetweenPVA andtheNPs[13].

2Pd₊O2→2PdO

G=−186.82kJ (4)

PdO+2Cl−₊2H+_→PdCl2+H2O(l)

G=−41.38kJ (5)

AfterthePVA removal,thecatalystswerereduced mildlyin H2/N2 atmosphere at 200◦C, in order to reduce PdCl2 to Pd0,

removingchlorinefromthePdsurface.Ontheotherhand, chlo-rineremovalisnotcompleteandasmallbutsignificantamountof chlorineremainsoncatalystsupporti.e.activatedcarbon.

ThiscontributionaimsatoptimizationoftheHClconcentration duringimmobilizationofPd-PVAcolloidonactivatedcarbon.The effectofresidualchlorineinfluencingtheactivityandtheselectivity innitritehydrogenationwillalsobeshown.

2. Experimental

2.1. Chemicals

Sodium tetrachloropalladate(II) (Na2PdCl4≥99.995% (metal

basis)), polyvinyl alcohol (PVA, average MW=13000–23000, 87%–89% hydrolyzed), sodium borohydride (NaBH4,≥96%

(gas-volumetric)),andformic acid(98%–100%)werepurchasedfrom Sigma-Aldrich.Sodiumnitrite(>99%)waspurchasedfromMerck. Activatedcarbon(AC,SBET=1000m2g−1)wassuppliedbyNorit.

ACwassievedintherangeof38–45_␮mindiameterbeforeused ascatalystsupport.Alltheaqueoussolutionswerepreparedusing ultrapurifiedwaterobtained(Millipore,Synergy).

2.2. Pdcolloidpreparation

Thepreparationofpalladiumnanoparticlesviacolloidalmethod hasbeendescribedpreviously,whichcanbesummarizedasfollows [13].PVAwasdissolvedinwaterat70◦Cwithstirringforatleast2h. Thesolution(2wt%)wasthencooleddowntoroomtemperature. AqueoussolutionofNa2PdCl4(20mL,containing0.086mmolPd)

and1.76mLoffreshlypreparedPVAsolutionwereaddedto240mL water,obtainingayellow-brownsolution.After3min,NaBH4

solu-tion(1.72mL,0.172mmol)wasaddedundervigorousstirring.The

brownPdcolloidsolutionwasimmediatelyformed.ThefinalpH wastypically8–8.5.

2.3. Colloidimmobilization

Typically,0.75gACorgraphitewasaddedtothePdcolloid solu-tion (260mL,3.3×10−4molL−1)immediatelyafterpreparation. Hydrochloricacid(HCl)wasaddedtoadjustpHtoeither1,2or 3.Theslurrywasstirredinairwithamechanical6-blade-stirrer (␾44mm,1000rpm)withthepropellerpositionedatthecentreof liquidfor2hatroomtemperature,filteredandthoroughlywashed withwater.Afterthat,thecatalystsweredriedinvacuumat40◦C overnight.

2.4. Catalystreduction

Catalystspreparedasdescribedabovewerecarefullytreatedin atubefurnace.Inatypicalprocedure,thetemperaturewasraised to200◦Catarateof5◦Cmin−1,thenkeptfor1hat200◦C,in10 vol%H2/90vol%N2.ThenthesamplewasflushedinN2for30min

at200◦C,andcooleddownatarateof20◦Cmin−1toroom temper-atureinthesameatmosphere.ThecatalystswereflushedinN2for

24hbeforeexposuretoair.Inthefollowing,thesamplenotation willbeusedasshowninTable1.

2.5. Characterization

PdparticlesizedistributionwasdeterminedusingTEM(Philips CM300ST-FEG)allowingreliabledetectionofmetalnanoparticles of1nmandlageronAC.TheACsupportedcatalystswerefirstly groundintosub-micronfragmentsanddispersedinethanol.Then thesuspensionwasdroppedonacoppergridcoveredwithhollow carbonforTEMimagetaking.Atleastfiveofthesefragmentswere randomlyselectedfordeterminationofPdparticlesizes,and typi-cally300Pdparticlesweremeasured.Notethatinformationonthe spatialdistributionofnanoparticlesthroughthesupportcannotbe obtainedasthesampleswereground.Themetalloadingonthe sup-portswereanalyzedbyXRF.Thetotalsurfaceareaofsampleswere calculatedbasedonN2physisorptiondata,usingtheBETmethod

forp/p0valuesbetween0.03and0.13forcatalystspreparedwith

ACfollowingtherecommendationsofRouqueroletal.[31],witha typicalerrormarginof5%.

Re-dissolutionofPdbyHClwasmeasuredwithUV–vis spec-troscopyofthecolloidalsuspensioninaUV-spectrometer(Perkin Elmer Lambda 850, wavelength from 200 to 800nm, scanning speed266.75nmmin−1)atroomtemperature.ThepHoffreshly prepared unsupportedPd-PVA colloidsuspension wasadjusted to 1, 2 and 3 by adding HCl solution,followed by stirring the suspensioninairatmospherefor2h.Then500␮Lofthetreated suspensionwasintroducedinaquartzcell(QS1000)forperforming themeasurement.

COchemisorptionatroomtemperaturewasusedtodetermine themetalsurfaceareathatis accessibleingasphase.Typically, thesample waspre-reducedat roomtemperature in hydrogen and thenflushed inHeat thesametemperature.ThenCO was Table1

Samplenotationsanddetailsofcorrespondingpreparationprocedure.

Sample PreparationProcedure

PdAC1A Pd-PVAcolloidimmobilizedonACusingHCltoadjustpHto1 PdAC2A Pd-PVAcolloidimmobilizedonACusingHCltoadjustpHto2 PdAC3A Pd-PVAcolloidimmobilizedonACusingHCltoadjustpHto3 PdAC1R PdAC1AreducedinH2/N2at200◦C

PdAC2R PdAC2AreducedinH2/N2at200◦C

introducedaspulsesandtheresponsewasrecordedusingaTCD detector.Weassumedthatthestoichiometricratioofnumberof adsorbedCOmoleculesandnumberofaccessiblePdsurfaceatoms is1:1.ThePddispersion(Pddisp.)wasdefinedas

Pddisp.=number_numberofPdatoms_ofPdatomsinthesurface_intotalofNPs (6) The surface composition of the catalysts was analyzed by X-ray photoelectron spectroscopy (XPS, Quantera SXM, Al K␣ (1486.6eV)). The powder samples were stored in air without anyfurtherpretreatmentbeforeanalysis.Typically,afew micro-gram sample was pressed into an indium foil, and four spots (600×300␮m2_{)onthesamplewererandomlyselectedfor}

mea-surementstoaverageoutanyinhomogeneityinthecatalysts.The accuracyof theresultingpeak positionswaswithin 0.2eV.The spectrawerefittedusingthesoftware“Multipakv.9.4.0.7”. Typi-cally,thebindingenergyinallspectrawasfirstcalibratedusing thecarbon1speakat284.8eVasaninternalreference.The spec-tradetectedatfourspotsofonesamplewereaveragedinorder toimprovethesignal-to-noiseratio,followedwithShirley back-groundsubtraction.ThePdpeakswerefittedusinganasymmetric model,necessarybecauseofinteractionofthephotoelectronswith thevalencebandelectrons[32],whereastheSandClpeakswere fittedusing mixed Gaussian-Lorentzianmodel, assuggested by HandbookofX-rayPhotoelectronSpectroscopy[33].Thepeaksfor eachsample(Pd5d,Cl2pandS2p)werefittedwithsetsofdoublets withidenticalFHWM.Bothwidthandpeakpositionwereallowed tooptimize.Thedistancewithinthedoubletswasfixed,according tothedatasuggestedinthehandbook[33].

2.6. Nitritehydrogenation

Thereactionwasperformedinahome-buildapparatus includ-ingaglasstankreactor(␾98mmwithfour5mmbaffles),equipped withamechanical6-blade-stirrer(␾44mm,1000rpm)withthe propellerpositionedatthecentreofliquid.Typically,50mg cat-alystwasaddedto300mLH2O.Themixedsuspensionwasthen

stirredvigorously,whileaH2/He/CO2gasmixturewasintroduced

viaadippedpipe(H2/He/CO2=6/3/1byvolumeflowrate,totalflow

rate=100mLmin−1,totalpressure=1bar)foratleast1h.CO2was

usedasabufferaccordingtoreactionshowninEqs.(7)and(8)to partlycompensatefortheprotonsconsumedbynitrite hydrogena-tion.

CO2+H2O→HCO3−+H+ (7)

Table2

SummaryofXRFelementalanalysis,TEMandCOchemisorption.

Sample CCl(wt%)a dTEM(nm) Pddisp.(%)

CO chemisorptionb TEMc PdAC1A 1.1 3.0±1.2 12±3 30 Pd AC 2 A 0.8 2.8±0.9 12±2 32 Pd AC3A 0.6 3.1±1.2 8±2 29 PdAC1R n.d.d _{4.1±1.1 32±5 22} PdAC2R n.d. 3.0±1.0 30±4 30 PdAC3R n.d. 3.2±1.2 18±4 28

a_{ChlorinecontentinthecatalystsaccordingtoXRF.}

b_{COchemisorptioningasphase.ThesamplesreducedinH}

2flowfor1hat

21±1◦_C.

c _{EstimatedassumingcleansphericalPdNPs.}

d _{ThecontentisunderXRFdetectivelimitation(<0.05wt%).}

HCO3−+H2O→CO32−+H+ (8)

Then3mL4.2mmolL−1NaNO2solutionwasintroduced,

start-ingthereaction.Samplesof1mLweretakenwithasyringeevery 5min.Catalystswereremovedusingasyringefilter(PTFE,0.2␮m) beforeinjecting intoion chromatograph(DIONEX,ICS 1000)to determinethecontentofnitriteandammonium.

3. Results

3.1. Elementalanalysisandphysicalproperties

XRFresultsin Table2shows thatthechlorineconcentration decreased with increasing pH (less HCl used) for as-prepared catalysts. Chlorine content was below the XRF detection limit (<0.05wt%)afterreductioninH2/N2at200◦C.ThePdcontentof

allcatalystswas1.2±0.1wt%.

TEMresultsinTable2andFig.1andS-1showPdparticlesizes ofabout3nmonaverageinallcatalysts,exceptforPdAC1Rwith 4nmaveragedparticlesize,preparedusingthehighestHCl con-centrationandreducedat200◦C.

COchemisorptionwasusedtodeterminetheaccessibilityof Pdsurfaceingasphase.AsshowninTable2,theapparent disper-sionswereaslowas10%forallas-preparedcatalysts.Notethatthe sampleswerefirstreducedinH2flowatroomtemperaturebefore

CO waschemisorbed fromgas phase. Afterreduction in H2/N2

at 200◦C, the apparent dispersion was significantly increased, agreeingwithourpreviousobservations[13].Nevertheless,the

200

250

300

350

400

450

500

278

236

222

Absorbance (a. u.)

Wavelength (nm)

pH =

1

pH =

2

pH =

3

As-prepared (pH = 8.3)

207

Fig.2.UV–visspectraofunsupportedPd-PVAcolloidsuspensionafterstirringinair for2hatdifferentpHadjustedwithHCl.Theabsorptionpeaksat207nmand236nm canbeattributedto[PdCl3(H2O)]−andpeaksat222nmand278nmto[PdCl4]2−

[34,35].Notethereisnoactivatedcarbonsupportaddedinthecolloidsuspension.

Table3

XPSresultsofPdandCloxidation-statesandsurfaceconcentration.

Sample Pd2+_{/Pd Cl(Pd)/Cl}a _Cl(Pd)/Pd2+ PdAC1A 0.38 0.75±0.04 1.7 PdAC2A 0.38 0.79±0.05 1.8 PdAC3A 0.07 0.72±0.04 3.0 Pd AC 1R 0.12 0.37±0.11 2.1 PdAC2R 0.03 0.43±0.12 3.1 PdAC3R 0.03 0.36±0.11 1.7

a_{Cl(Pd)standsforchlorinebondedonPdsurface.}

persionofthecatalystpreparedwithpH3wasonly18%afterthe reduction,muchlowerascomparedwithcatalystspreparedwith pH1and2(about30%).

FigureS-2showsthatPVAisblockingmicro-poresexclusively inallcatalysts,withoutinfluencingmeso-poreswherethePdNPs aremostprobablylocated,agreeingwiththepreviousstudy[13]. 3.2. UV–visspectroscopy

Fig.2showspartialre-dissolutionofPdforunsupportedPd-PVA colloidstirredinairatdifferentpHcontrolledbyHClconcentration. NoPd-Clcomplexwasdetectedinas-preparedcolloidsuspension, indicatingnoPd-containingionsexistrightaftercolloid prepara-tion.Afterstirringinairfor2h,[PdCl3(H2O)]−(207nmand236nm)

wastheonlydetectedPd-Clcomplexinthecolloidsuspensionat lowHClconcentration(pH3).Incontrast,[PdCl4]2−(278nm)is

detectedinthesuspensionatpH2,becomingthemajorityPd-Cl complex(222nmand278nm)whenthepHwasfurtherdecreased to1byaddingmoreHCl.

3.3. XPS

XPSresultsaresummarizedinTable3,andtypicalspectraare shown in Fig.3 and Figure S-3 in SupportingInformation.The resultsagreewithpreviousobservationsingeneral:thesurfaceof PdNPswasoxidizedbyformationofPdCl2; reductioninH2/N2

at200◦C reducesPd2+_toPd0_{,while removingchlorine[13].In}

addition,theresultsshowthattheseeffectdependontheHCl con-centration:whileonly7%Pdisoxidizedintheas-preparedcatalyst preparedwithlowHClconcentration(pH3),muchhighervalues areobserved(upto38%)forcatalystspreparedwithlowerpH.Itis

alsoshownthatashighas12%Pdisoxidizedincatalystprepared withpH1afterreductioninH2/N2at200◦C.

Ontheotherhand,twotypesofchlorineweredetectedwith formalchargeCl−,asshowninFig.3andTable3,whichcanbe attributedtoClbondedtoPd(ca.198eV)andClinorganic com-poundspresumablypresentonAC(ca.200eV),respectively[36]. Theamountsofbothtypesofchlorinedecreasedsignificantlyafter reduction at 200◦C in H2/N2, whereasthe signalofchlorine in

organiccompoundsdecreasedtolesserextentascomparedto chlo-rineonPd.Furthermore,themolarratioofCl(Pd)/Pd2+_wasinthe

rangeof2–3forallsamples,bothbeforeandafterreduction.

3.4. Nitritehydrogenation

Fig.4(a)and(b)presentconcentrationsofnitriteand ammo-nium as function of time. The initial rate per total Pd can be estimatedusingthePdloadingasdeterminedbyXRFinTable1,as showninFig.4(c).Alternatively,Fig.4(d)showstheinitialrateper surfacePd,wheretheamountofsurfacePdwasdeterminedbyXRF resultstogetherwithCOchemisorptionresultsinTable2.Inboth cases,thereactionrateshowednosignificantchangewithvariation ofthepHduringcolloidimmobilizationforas-preparedPd-PVA/AC catalystsinFig.4(d).Forthereducedcatalysts,anincreaseof reac-tionratewasobservedwithincreasingthepH-value.Additionally, ammoniumformationcontinuedafternitritewasconverted com-pletely.Thiswasexplainedinpreviousworkbythepresenceof residual N-containing species on Pd surface, probably nitrogen atoms,reactingveryslowlywithhydrogenatclose-to-complete conversionlevel[15].Consequently,theselectivityofthecatalysts toammoniumcanonlybecomparedatconversionlevelsbelow 100%.

Fig. 5 shows that selectivity to ammonium decreased with increasingpHduringcolloidimmobilization,forbothas-prepared andreducedcatalysts.Thereductiontreatmentat200◦CinH2/N2

resultedinhigherselectivitytoammonium,regardlesstheamount ofHClusedforimmobilization.

The amountof Cl introducedby adding thecatalyststo the batchreactor(MCl,cata)canbecalculatedusingtheXRFresultsin

Table2,asshowninTable4.Cl−isreleasedfromthecatalysttothe solutionduringthe1hpre-reductiontreatment,beforethe reac-tionisinitiatedbyinjectionofthenitritesolution.Theamounts ofCl−inaqueousphasedetectedbyIC(MCl,solu)arealsoshownin

Table4.Inallcases,thevaluesofMCl,cataandMCl,soluareingood

agreement,indicatingthatchlorineonthecatalystindeeddissolves completely.InordertorevealtheinfluenceofthefreeCl−inthe reactionsolutiononcatalystperformance,NaClwasaddedtothe reactionslurrywithPdAC3R,asshowninTable4.Thecatalyst activityremainedconstantwithinexperimentalerrorasaresultof addingNaCl;surprisingly,theselectivitytoammoniumdecreased significantly.

4. Discussion

4.1. InfluenceofHClontheaccessibilityofPdsurface

COchemisorptionresultsinTable2showsignificantincrease ofaccessiblePdsurfaceafterreductioninH2/N2at200◦Cforthe

catalystpreparedatpH1and2,resultinginPddispersions sim-ilartotheobservationswithTEM.ThisconfirmsthatPVAcanbe removedcompletelyusingHClatpH2inthepresenceofair,as reportedpreviously[13].AfterPVAremovalviawashingand sub-sequentdrying,H2reductionat200◦Cremovedchlorinefromthe

PdsurfaceviaformationofgaseousHCl,asconfirmedbyXRFresults inTable2.

210 208 206 204 202 200 198 196 194

(eV)

Banding energy

Inte

nsity (a. u.)

Pd_

AC

_2_A

350

348

346

344

342

340

338

336

334

332

330

Banding energy (eV)

Intensity (a. u.)

Pd_

AC

_2_A

Pd_

AC

_2_R

Pd_

AC

_2_R

Pd

Pd

Cl (C) Cl (Pd)

(a)

(b)

Fig.3. XPSspectraofactivatedcarbonsupportedPd-PVAcolloidspreparedatpH2:(a)Pd3dspectra,(b)Cl2pspectra.Originaldata(hollowdots)wassubtractedwithShirley background(blackline)andfittedusingmethoddescribedinSection2.5.ThefittedPd3d5/2peaks,andCl2p3/2peaksarehighlighted(blueandorange)forcomparison.The fittedPd3d3/2peaksandCl2p1/2peaksarekeptingray.Thesumofallfittedpeaksshowedasredlinewitherrorshowedasdashline.Cl(C)andCl(Pd)standforchlorine inorganiccompoundsandchlorinebondedonPdsurface,respectively.(Forinterpretationofthereferencestocolourinthisfigurelegend,thereaderisreferredtotheweb versionofthisarticle.)

Table4

Summaryofreactionrateandselectivitytoammoniumrelatedtochlorineconcentration.

Sample MCl,cata

(␮mol)a

MCl,solu

(␮mol)b

Initial ratepertotalPd

(molnitritemolPd−1L−1min−1)

Initialrateper surfacePd

(molnitritemolPd−1L−1min−1)

NH4+selectivityat80% conversion(%) Pd AC1A 16 16 7.7±1.5 62.6±12.5 9.1±0.5 PdAC2A 11 11 6.0±1.5 51.3±12.8 3.6±0.4 PdAC3A 8.4 8.0 4.2±1.5 50.7±17.8 0.8±0.2 PdAC3R – 0.3 12.6±1.5 71.9±8.5 2.3±0.3

PdAC3RwithextraNaCl 9.2 13.3±1.5 76.1±8.5 1.1±0.2

a_{AmountofClintroducedbyaddingthecatalystinthereactor,accordingtoXRFresultsinTable2.} b _{AmountofClinaqueousphaseduringreactiondetectedbyIC.}

Ontheotherhand,COchemisorptiondatainTable2alsoshow arelativelylowPddispersionof18%onthereducedcatalyst pre-paredwithlowHClconcentration(pH3).Thisapparentdispersion islowerthanwhatwouldbeexpectedbasedonTEMandwe sug-gestthatthisiscausedbythepresenceofPVA.XRFresultsinTable2 showarelativelylowchlorineconcentrationintheas-prepared cat-alyst,inagreementwithXPSresultsinTable3.Inthisas-prepared catalystsonly 7%ofthePdis oxidizedascanbeseen inFigure S-3(a);furthermore,thePd2+_{:Cl(Pd)molarratiois1:2,}

indicat-ingthatthechlorinecoverageiswellbelowonemonolayer(about 30%),muchlowerascomparedtocatalystpreparedusinghigher HClconcentration(pH1and2).Obviously,boththelowchlorine andPd2+_{contentarecausedbythelowHClconcentrationused}

dur-ingthepreparation.WesuggestthatthesmallamountofCl−isnot sufficienttoinducedesorptionofPVAtothesameextentathigh Cl−concentration.

Interestingly,thedatainTable4revealsthatalmostall chlo-rinepresent onthecatalyst desorbedand dissolvedduringthe reaction,basedonthesimilaritybetweentheamountofCl− intro-duced,according XRF,and Cl− contentinaqueousphase inthe

batchreactorasdetectedbyIC.Apparently,themajorityof chlo-rinecanberemovedfromthecatalystsurfaceduringreductionin H2inaqueousphase,beforenitritewasintroduced,byreducingthe

PdClxspeciestoPdmetalandproducingHCl.However,thisdoes

notruleoutsmallamountofchlorineonPdcanstillinfluencethe catalyticreactions,asdiscussedbelowinSection4.3and4.4. Addi-tionally,theproducedHClisinsuchalowconcentration(lower than54␮molL−1ascalculatedbasedondatainTable4),and can-notinfluencethepHinthereactorsignificantly,especiallyinthe presenceofCO2.

In short, a critical amount of chlorine is necessary for PVA removalfromthePdsurface.Thechlorineconcentrationinthe cat-alystisdeterminedbytheHClconcentrationintheaqueousphase duringimmobilizationofPd-colloidonthecarbonsupport. 4.2. InfluenceofHClonPdoxidationstateandparticlesize

UV-visspectrainFig.2showthepresenceofCl-containingPd complexanionsintheunsupportedcolloidsuspensionafter stir-ringin airwithdifferentHCl concentration,and thecomplexes

Fig.4.ActivityofPd-PVAsupportedonACbeforeandafterreductionat200◦_{CinH2/N2:concentrationofnitrite(solidsymbols)andammonium(opensymbols)with(a)as}

preparedand(b)reducedcatalysts;initialreactionrate(c)pertotalPdor(d)persurfacePd(inunitofmolnitritemolPd−1L−1min−1)withcatalystspreparedatdifferentpH.

Theerrorbarsrepresentstandarddeviation.

converted from[PdCl3(H2O)]− to[PdCl4]2− withincreasing HCl

concentration(i.e.,decreasingpH).Apparently,thesePdcomplex anionsoriginatefrompartiallyre-dissolutionofthePdNPs, accord-ingtoEq.(9). Pd+Cl→− O2H+ PdCl2 +Cl− → +H2O [PdCl3(H2O)]− +Cl − → -H2O [PdCl4]2− (9)

XRFresultsinTable2clearlyshowverysimilarPdloadingsfor allthecatalystsupportedonAC,independentoftheHCl concen-tration.ThisindicatesthatthemajorityofdissolvedPdClx·(H2O)4-x

anions(2<x≤4)re-adsorbonAC.Inthisstudy,thePddispersion determinedbyCOchemisorption,afterreductionat200◦Cfor2h inordertoremovechlorinefromthePdnanoparticles,is32%forthe catalystpreparedwiththehighestHClconcentration(pH1), signif-icantlyhigherthanthedispersionestimatedbasedonTEM(22%). ThisindicatestheexistenceofextremelysmallPdNPs,which can-notbedetectedwiththeTEMusedinthisstudy.Simonovetal. reportedthat[PdCl4]−caneitherform␲-complexesofPdCl2with

C C fragments of thecarbonmatrix, or bereduced tometallic Pdparticlesbyspontaneousreductiononcarboninthepresence ofHCl[36–38].Probably,extremelysmallPdandPdCl2particles

areformedinthesamewayhere.Ontheotherhand,TEMshows evenlargerPdparticleswithanaveragesizeof4nminthesample PdAC1R,indicatingsomesinteringduringreductioninH2/N2at

200◦C.

Surprisingly,ashighas12%ofPdatomsremainedoxidizedafter reductioninH2/N2at200◦Cfor2hwithcatalystpreparedwith

thehighestHClconcentration(pH1),accordingtoXPSresultsin Table3andFigureS-3.Ascomparison,only3%ofPdwasoxidized inreducedcatalystspreparedatpH2and3.Allcatalystswerekept

inambientovernightbeforeXPSperformed.Thehighamountof oxidizedPdinthereducedcatalystpreparedwithpH1is support-ingthesuggestionthatverysmallPdparticlesarepresent,asthese aremoresusceptibleforoxidationinambient.

Insummary,excessamountofHClcausespartialre-dissolution ofPdNPs,causingincreasingPdparticlesizesaswellasresidual PdCl2afterreductionbyH2.ThusamoderateamountofHClshould

beusedtoremovePVAfromPdsurfacecompletely,atthesame timeminimizingside-effectscausedbyextraHCl.

4.3. InfluenceofHClonactivity

4.3.1. Influenceofchlorineontheactivityofas-preparedcatalysts AsshowninFig.4(c)and(d),allas-preparedcatalystsshowed similaractivitywithinexperimentalerror,regardlessthedifferent chlorineconcentrations.Accordingly,XRFresultsinTable2and ICresultsinTable4 showthemajorityof chlorinere-dissolved inaqueousphaseduringthereductiontreatmentinthereactor, immediatelybeforethecatalyticexperiment,cleaningthePd sur-face and making thesurface Pd atoms accessible for reactants. Furthermore,nosignificantchangeinactivityforPdAC3Rwas foundwhenaddingchlorineupto9.2␮mol(30␮molL−1)asshown inTable4.Apparently,chlorinedissolvedinthereactionmixture attheconcentrationsinthisstudyhasnosignificantinfluenceon theapparentactivity.

Atfirstsightthisseemstodisagreewithresultsreportedby Pintar,etal.onnitratehydrogenationonPd/Al2O3[39]andChaplin

etal.onnitritehydrogenation[40].Inbothstudiesreactionrates decreasewithchlorineaddition;itshouldbenotedthoughthat thechlorineconcentrationinthisstudyis atleasttwoorderof

0

20

40

60

80

100

0

10

20

30

40

(a)

As-prepared catalysts

Pd_AC_1_A

Pd_AC_2_A

Pd_AC_3_A

NH

Selectivity (%

)

Conversion (%)

0

20

40

60

80

100

0

10

20

30

40

(b) Reduced catalysts

Pd_AC_1_R

Pd_AC_2_R

Pd_AC_3_R

NH

Selectivity (%

)

Conversion (%)

1

2

3

NH4+ selectivity

at 80% conversion (

%

)

pH

(c)

Fig.5.SelectivitytoammoniumfornitritehydrogenationwithPd-PVAsupportedonAC:(a)selectivitytoammoniumasfunctionofnitriteconversionwithas-prepared catalysts;(b)selectivitytoammoniumasfunctionofnitriteconversionwithcatalystsreducedat200◦_{CinH2/N2;(c)comparisonofammoniumselectivityat80%nitrite}

conversionwithcatalystpreparedatdifferentpH.

magnitudelower,explainingwhytheeffectoncatalystactivityis negligible.

4.3.2. Activityafterreduction

TheactivityofthecatalystsafterreductioninH2/N2 at200◦C

iscomparedinFig.4(c)and(d).Theinitialratesofthereaction increasewithincreasingpH,usinglessHClduringimmobilization ofthecolloid.Itisclearfromthediscussionabovethatdissolution andre-depositionofPdislikely tooccurduringimmobilization ofthecolloid,probablyinfluencingtheparticlesizedistributionof PdNPsonthesupport,especiallyinthecaseofhighHCl concentra-tion(lowpH).Possibly,extremelysmallPdparticlesinducealower activityperactivesite,asparticlesizeeffecthavebeen experimen-tallyexcludedexclusivelyformetalparticleslargerthan 2.5nm [14,17,18].Thishypothesisneedstobefurthertestedusingcatalyst withnarrowparticlesizedistributionbelow2nmtoevenatomic level.

Insummary,catalystpreparedwithlowestHClconcentration (pH 3) resultsin the highest activity afterreduction in H2/N2,

despiteincompleteremovalofPVAfromthePdsurface.Instead, there-dissolutionofPdinpresenceofhighHClconcentrationinair maycausethedecreaseoftheactivitydespitecompleteremovalof thepolymer.

4.4. EffectofHClontheselectivitytoammonium

ReductioninH2/N2at200◦Cincreasestheselectivityto

ammo-nium as shown in Fig. 5(c), together with removal of chlorine contentasdetectedbyXRFinTable2andXPSinTable3.Onthe otherhand,Table4showstheselectivitydecreaseswithadding extraNaClinPdAC3Rcatalyst.Clearly,Cl−decreasesthe selec-tivitytoammonium,nomatterwhetherCl−isreleasedfromthe catalystduringthereactionorisaddedtotheaqueousphase

reac-tionmixture.Areliablecomparisontosimilareffectsreportedin literatureisnotpossible,againbecausetheconcentrationof chlo-rineinthisstudyismuchlowerthaninliterature[39,40].

Surprisingly, the selectivity to ammonium increased with increasingHCl concentrationduring colloidimmobilization, for bothas-preparedandreducedcatalysts,asshowninFig.5.Table4 confirms thatcatalysts preparedwithhigher HClconcentration inducedhigherchlorineconcentrationduringthecatalytic reac-tion.Clearly,thisdisagreeswiththeeffectoftheCl−concentration inthereactionmixtureasdiscussedaboveandthereforeanother effectonselectivityapparentlycontributes.

Itwasfoundinourpreviouswork[14]thatPVAisnot influenc-ingtheselectivitywhereastheselectivitytoammoniumincreases withincreasingPd particlesize.Therefore, wesuggest thatthe selectivityisinfluencedviathemetalparticlesizeastheparticle size distributions might bedifferent,as discussed above, influ-encedbythedissolution andre-depositionofPd inpresenceof HCl.Clearly,moreworkwouldbeneededtotestthishypothesis.

Inshort,theselectivitytoammoniumdecreaseswithadding lowconcentrationofchlorineinthereactionmixture,whereasthe selectivitytoammoniumincreaseswithincreasingconcentration ofHClduringcolloidimmobilization.

5. Conclusion

Pd-PVAcolloidshavebeenimmobilizedonACwithdifferent HClconcentrationinthepresenceofair.HighconcentrationofHCl (pH1)usedduringcolloidimmobilizationcausespartialdissolution ofPd,decreasingactivityfornitritehydrogenationandincreasing selectivitytoammonium.Incontrast,highactivityandlow selec-tivitytoammoniumcanbeachievedbyusinglowconcentrationof HCl(pH3)duringcolloidimmobilization,despitethefactthatPVA canonlybepartlyremovedfromthePdparticles.Finally,low

con-centrationoffreeCl−inthereactionmixtureinducesadecreasing selectivitytoammonium,withoutinfluencingthereactionrate.

AppendixA. Supplementarydata

Supplementarydataassociatedwiththisarticlecanbefound, intheonlineversion,athttp://dx.doi.org/10.1016/j.cattod.2017.01. 028.

References

[1]EuropeanEnvironmentAgency(EEA), http://www.eea.europa.eu/data-and-maps/figures/nitrate-concentration-in-groundwater(Lastupdated01Jan 2002).

[2]B.Gu,Y.Ge,S.X.Chang,W.Luo,J.Chang,Glob.Environ.Change23(2013) 1112–1121.

[3]A.Scheidleder,J.Grath,G.Winkler,U.Stärk,C.Koreimann,C.Gmeiner,S. Nixon,J.Casillas,P.Gravesen,J.Leonard,M.Elvira,Eur.Environ.Agency(EEA) (1999).

[4]K.-D.Vorlop,T.Tacke,Chem.Ing.Tech.61(1989)836–837.

[5]T.Tacke,Dissertation,TechnischenUniversitätCarolo-Wilhelmina,1991.

[6]A.Kapoor,T.Viraraghavan,J.Environ.Eng.123(1997)371–380.

[7]N.Barrabés,J.Sá,Appl.Catal.B104(2011)1–5.

[8]Y.Xia,Y.Xiong,B.Lim,S.E.Skrabalak,Angew.Chem.Int.Ed.48(2009)60–103.

[9]G.Somorjai,J.Park,Top.Catal.49(2008)126–135.

[10]B.RoldanCuenya,Acc.Chem.Res.46(2013)1682–1691.

[11]D.Shuai,D.C.McCalman,J.K.Choe,J.R.Shapley,W.F.Schneider,C.J.Werth, ACSCatal.3(2013)453–463.

[12]K.A.Guy,H.Xu,J.C.Yang,C.J.Werth,J.R.Shapley,J.Phys.Chem.C113(2009) 8177–8185.

[13]Y.Zhao,L.Jia,J.A.Medrano,J.R.H.Ross,L.Lefferts,ACSCatal.3(2013) 2341–2352.

[14]Y.Zhao,J.A.Baeza,N.KoteswaraRao,L.Calvo,M.A.Gilarranz,Y.D.Li,L. Lefferts,J.Catal.318(2014)162–169.

[15]Y.Zhao,N.KoteswaraRao,L.Lefferts,J.Catal.337(2016)102–110.

[16]J.K.Chinthaginjala,A.Villa,D.S.Su,B.L.Mojet,L.Lefferts,Catal.Today183 (2012)119–123.

[17]D.Shuai,J.K.Choe,J.R.Shapley,C.J.Werth,Environ.Sci.Technol.46(2012) 2847–2855.

[18]J.K.Chinthaginjala,J.H.Bitter,L.Lefferts,Appl.Catal.A383(2010)24–32.

[19]A.Villa,D.Wang,D.S.Su,L.Prati,Chemcatchem1(2009)510–514.

[20]A.Quintanilla,V.C.L.Butselaar-Orthlieb,C.Kwakernaak,W.G.Sloof,M.T. Kreutzer,F.Kapteijn,J.Catal.271(2010)104–114.

[21]X.Wang,P.Sonström,D.Arndt,J.Stöver,V.Zielasek,H.Borchert,K.Thiel,K. Al-Shamery,M.Bäumer,J.Catal.278(2011)143–152.

[22]J.N.Kuhn,C.-K.Tsung,W.Huang,G.A.Somorjai,J.Catal.265(2009)209–215.

[23]B.Feng,Z.Hou,H.Yang,X.Wang,Y.Hu,H.Li,Y.Qiao,X.Zhao,Q.Huang, Langmuir26(2010)2505–2513.

[24]Z.Niu,Y.Li,Chem.Mater.26(2014)72–83.

[25]S.Campisi,D.Ferri,A.Villa,W.Wang,D.Wang,O.Kröcher,L.Prati,J.Phys. Chem.C120(2016)14027–14033.

[26]J.A.Lopez-Sanchez,N.Dimitratos,C.Hammond,G.L.Brett,L.Kesavan,S. White,P.Miedziak,R.Tiruvalam,R.L.Jenkins,A.F.Carley,D.Knight,C.J.Kiely, G.J.Hutchings,Nat.Chem.3(2011)551–556.

[27]M.Crespo-Quesada,J.-M.Andanson,A.Yarulin,B.Lim,Y.Xia,L.Kiwi-Minsker, Langmuir27(2011)7909–7916.

[28]Y.Borodko,H.S.Lee,S.H.Joo,Y.Zhang,G.Somorjai,J.Phys.Chem.C114 (2009)1117–1126.

[29]C.Aliaga,J.Y.Park,Y.Yamada,H.S.Lee,C.-K.Tsung,P.Yang,G.A.Somorjai,J. Phys.Chem.C113(2009)6150–6155.

[30]A.M.Venezia,V.LaParola,G.Deganello,B.Pawelec,J.L.G.Fierro,J.Catal.215 (2003)317–325.

[31]F.Rouquerol,J.Rouquerol,K.Sing,AdsorptionbyPowdersandPorousSolids, AcademicPressLondon,1999,pp.237–285.

[32]DavidBriggs,J.T.Grant,SurfaceAnalysisbyAugerandX-RayPhotoelectron Spectroscopy,IMPublications,Chichester,2003.

[33]J.F.Moulder,W.F.Stickle,P.E.Sobol,K.D.Bomben,HandbookofX-ray PhotoelectronSpectroscopy,Perkin-ElmerCorporation,EdenPrairie,1992.

[34]L.I.Elding,Inorg.Chim.Acta6(1972)647–651.

[35]C.DrewTait,D.R.Janecky,P.S.Z.Rogers,GeochimicaetCosmochimicaActa55 (1991)1253–1264.

[36]P.A.Simonov,A.V.Romanenko,I.P.Prosvirin,E.M.Moroz,A.I.Boronin,A.L. Chuvilin,V.A.Likholobov,Carbon35(1997)73–82.

[37]P.Simonov,S.Troitskii,V.Likholobov,Kinet.Catal.41(2000)255–269.

[38]H.Chu,J.Wang,L.Ding,D.Yuan,Y.Zhang,J.Liu,Y.Li,J.Am.Chem.Soc.131 (2009)14310–14316.

[39]A.Pintar,M.vetinc,J.Levec,J.Catal.174(1998)72–87.

[40]B.P.Chaplin,E.Roundy,K.A.Guy,J.R.Shapley,C.J.Werth,Environ.Sci. Technol.40(2006)3075–3081.