Electrochemical degradation of Nafion ionomer to functionalize carbon support for methanol electro-oxidation

ContentslistsavailableatScienceDirect

Journal

of

Power

Sources

j ou rn a l h o m e pa g e :w w w . e l s e v i e r . c o m / l o c a t e / j p o w s o u r

Electrochemical

degradation

of

Nafion

ionomer

to

functionalize

carbon

support

for

methanol

electro-oxidation

Yu-Chi

Hsieh, Jing-Yu

Chen,

Pu-Wei

Wu

DepartmentofMaterialsScienceandEngineering,NationalChiaoTungUniversity,Hsin-chu300,Taiwan

a

r

t

i

c

l

e

i

n

f

o

Articlehistory:

Received8April2011

Receivedinrevisedform24May2011

Accepted25May2011

Available online 24 June 2011 Keywords: Nafionionomer Carbonfunctionalization Cyclicvoltammetry Methanolelectro-oxidation Pt

a

b

s

t

r

a

c

t

Aneffectiveelectrochemicalroutetoproducefunctionalgroupsoncarbonsurfaceisdemonstrated. Cyclicvoltammetric(CV)sweepsareperformedin0.5MH2SO4 electrolyteonelectrodescontaining carboncloth,VulcanXC72R,andNafionionomer.Withsupplyofambientoxygen,thegenerationof hydroxylradicalsfromtheoxygenreductionreactionduringCVcyclesinitiatesthedecompositionof Nafionionomerthatleadstoformationofoxygenatedfunctionalgroupsonthecarbonsurface.Ion chro-matographyconfirmsthedissolutionofsulfateanionsuponCVscans.Ramananalysissuggestsaminor alterationforthecarbonstructure.However,X-rayphotoelectronspectroscopyindicatesasignificant increaseofoxygenatedfunctionalgroupsinconjunctionwithnotablereductioninthefluorinecontent. TheamountoftheoxygenatedfunctionalgroupsisdeterminedbycurvefittingofC1sspectrawith knownconstituents.Thesefunctionalgroupscanalsobefoundbyimmersingtheas-preparedelectrode inasolutioncontainingconcentratedresiduesfromNafionionomerdecomposition.Thefunctionalized electrodeallowsa170%incrementofPtionadsorptionascomparedtothereferencesample.After elec-trochemicalreductions,thefunctionalizedelectroderevealssignificantimprovementsinelectrocatalytic abilitiesformethanoloxidation,whichisattributedtotheoxygenatedfunctionalgroupsthatfacilitates theoxidationofCOonPt.

© 2011 Elsevier B.V. All rights reserved.

1. Introduction

Carbonaceousmaterialshavebeenwidelyusedasthesubstrates

for catalystimpregnations inroom tempeaturare fuelcells like

polymermembranefuelcellsanddirectmethanolfuelcells[1–10].

Itisbecausewiththeselectionofcarbonsassupports,

nanopartic-ulatecatalystssuchasPtandPtRuareabletodistributeuniformly,

leadingtoreducedloadingandbettercatalystutilization.Todate,

carbonsinarichvarietyofformsincludingcarbonblacks,carbon

nanotubes(CNTs),mesoporouscarbons,carbonnanocapsules,

acti-vatedcarbons,andcarbonxerogelshavebeeninvestigatedasthe

catalystsupportswithimpressiveresults[11–16].Theinteractions

betweenthecarbonandcatalystarecriticalbecausealackof

suf-ficientbondingbetweenthem causespossibledetachmentsand

colaescence,which resultsin undesirableperformance

degrada-tion[17].Unfortunately,untreatedcarbonsareoftenhydrophobic

innaturethat allowpooradsorptionofcatalyst precursorsand

catalysts.Therefore,itisnecessarytocarryoutadditional

function-alizationtreatmentsonthecarbonstorenderahydrophilicsurface

instead.Afterpropersurfacefunctionalizations,thecarbonsare

∗ Correspondingauthor.Tel.:+88635131227;fax:+88635724727.

E-mailaddress:[email protected](P.-W.Wu).

expectedtoadsorbmorecatalystprecursorsforalargeramount

ofcatalystdeposition.

Earlierstudiesonthecarbonfunctionalizationsareconcerned

withcarboncorrosionsbecauseundertheoperationconditionsof

phosphoricacidfuelcells,thecarbonis pronetooxidation loss

bytheformationofsurfaceoxidizedgroups[18–22].Ingeneral,

thefunctionalizationofcarboninvolvesanodizationtreatmentsin

concentratedacidsatmoderatetemperature[23,24].Forexample,

Kangasniemietal.imposedpotentiostatictreatmentsonthe

Vul-canXC72(XC72) in1MH2SO4 solution,anddeterminedthatat

roomtermperature,asignficantoxidationwasoccurringforthe

anodizingvoltageof1.2Vfor16hbut0.8Vwassufficentat65◦C

toproducethesameeffect[25].Asimilaranodizationtreatment

of2VwasemployedbyStevanovicet al.tointroduceselective

functionalgroupsontheglassycarbons[26].Thedegreeof

sur-facefunctionalizationalsodependsonthetypeofcarbonmaterials

becausetheirsurfaceareaandmicrostructuredifferconsiderably.

For instance, theCNTsreveal notable oxidation resistanceover

theXC72 while theBP2000, witha larger specificsurface area

(m2_g−1_{),experiencesmoreoxidationlossasopposedtotheXC72}

withasmallerspecificsurfacearea[27].Sofar,after

functionaliza-tion,surfaceoxidizedgroupssuchasphenols,carbonyls,carboxylic

acids,ethers,quinones,andlactoneshavebeenidentified.Theexact

mechanismresponsiblefortheformationofselectivefunctional

groupsiscontingentontheprocessingstepsinvolvedandthetype

0378-7753/$–seefrontmatter © 2011 Elsevier B.V. All rights reserved.

ofcarbonmaterials.Sincethecarbonisrelativelyinertin

corro-siveelectrolytes,typicalsurfacefunctionalizationstepsarerather

time-consuming.Therefore,itisofparticularinteresttodevelopa

simpleandefficientprocessforfunctionalizationpurpose.

Analternativeapproachtofunctionalizethecarbonmaterials

isbychemicalalterationofpolymericbinders.Inelectrode

fabri-cations,Nafionionomerisoftenaddedinmixturewithcarbons,

servingsimultaneouslyasabinderandconductivepathforproton

transport.Therefore,itispossiblethattheNafionionomerwould

sufferfromstructuraldamageandlossofsulfonicacidsidechains

ifdeliberateelectrochemicaltreatmentsareapplied.Previously,

extensiveeffortshavebeendevotedtounderstandtheresponsible

mechanismforNafionmembranedegradationindifferent

environ-mentsandfactorsincludinghumidity,temperature,andoxygen

concentrationarefoundtoberelevant[28,29].Accordingto

litera-ture,hydroxyl(•OH)andperoxy(•OOH)radicalsformedduringfuel

celloperationsareabletoreactwithpolymerendgroupsthatstill

containresidualterminalH-groups[30–32].Furtherstudies

indi-catethatthesulfonicacidsidechainsarealsosusceptibletoradicals

[29,33,34].Inaddition,thedegradedspeciesofNafioncontainfree

radicalsthathavebeenreportedtoattackcarbonsandchemically

bondtotheirsurface[35–37].Moreover,itissuggestedthatthe

presenceoffunctionalizedgroupsonthecarbonsurfaceis

possi-bletoengenderadditionaloxidizedgroups[38–40].Inlightofthis

information,werealizethattheintentionaldegradationofNafion

ionomermightprovideaneffectiverouteforcarbon

functionaliza-tion.

In this work, we conducted multiple cyclic voltammetric

scans (CV) to introduce functional groups on the electrode

structure,followed byPtcationadsorptionand electrochemical

reduction to produce nanoparticulate Ptimpregrenated onthe

functionalizedsupport.Electrocatalyticanalysisonthemethnaol

electro-oxidationwasperformedtoelucidatetheeffectof

func-tionalizedsupportforcatalyticactions.

2. Experimental

2.1. Carbonsurfacefunctionalization

Theelectrodefor surfacefunctionalizationwasfabricatedby

depositingacarbon/Nafionmixtureontoacommerciallyavailable

carboncloth(E-TEK).First,10mgNafionionomersolution(5wt%)

and 8mg carbonpowders (VulcanXC72R) were mixedin 5mL

99.5wt%ethanolfor60minundersonicationtoformanink

dis-persion.Thedispersionwasdepositedrepeatedlyona4cm2_carbon

clothwhichwaskeptat80◦Catopahotplatetoevaporateresidual

solvent.TheloadingsfortheXC72RandNafionionomeronthe

car-bonclothwere2and0.125mgcm−2,respectively.Subsequently,

thesurfacefunctionalizationwasperformedbyimposingCVscans

ontheelectrode(activeareaof0.785cm2_{)for20cyclesbetween}

−0.2and1.1V(vs.Ag/AgCl)at50mVs−1inanaqueouselectrolyte

of0.5MH2SO4.APtfoilof10cm2wasusedasthecounterelectrode.

ThedurationfortheCVscanswas17min.Inordertointroduce

oxygenduringtheCVscans,thebacksidefortheelectrodewas

pressedagainstastainlesssteelfoilthatwaspartiallyexposedto

theambientoxygen.Aschematicforthecelldesignisillustratedin

Fig.1.Forcomparisonpurpose,wealsoimmersedtheas-prepared

electrodesin0.5MH2SO4or0.1MHClaqueoussolutionwith

con-centratedresiduesfromNafionionomerdecompositiontoanalyze

theirsurfacefunctionalgroups.

2.2. Electrochemicalanalysis

Thefunctionalizedelectrode wasimmersedin5mMH2PtCl6

aqueoussolution(pHadjustedto8)at40◦C.Theimmersionlasted

Fig.1.AschematicoftheelectrochemicalcellforCVscansin0.5MH2SO4aqueous

solution.Thecarbonclothispartiallyexposedtoambientoxygen.

for48htoensuresufficientadsorptionofPtCl62−.Toreducethe

adsorbedPtions,CVscanswerecarriedoutbetween−0.2and0.2V

in0.5MH2SO4aqueoussolutionat50mVs−1.Toevaluatethe

elec-trochemicalsurfacearea(ECSA)forthedepositedPt,weconducted

CVscansbetween−0.2and0.9Vin0.5MH2SO4at50mVs−1.The

ECSAwasestimatedbytheintegratedchargeinthehydrogen

des-orptionregion.Formethanolelectro-oxidation,multipleCVscans

wereperformedbetween−0.2and 0.9Vat50mVs−1 in500mL

aqueoussolution of0.5MH2SO4 and 1M CH3OH.The areafor

theworkingelectrodewas0.785cm2_{.Forlifetimedetermination,}

chronoamperogramswererecordedat0.5Vfor30minin500mL

of0.5MH2SO4and1MCH3OH.TheAg/AgClandPtfoil(10cm2)

wereusedasthereferenceand counterelectrodes,respectively.

Surfacefunctionalization,PtCl62−reduction,ESCAdetermination,

andmethanolelectro-oxidationwerecarriedoutat26◦Cina

three-electrodearrangementusingaEG&G263Apotentiostat.

2.3. Materialscharacterizations

X-rayPhotoelectronSpectroscopy(XPS;ThermoMicrolab350)

wasadoptedtoevaluatetheoxygenatedfunctionalgroupsonthe

functionalizedelectrodes.RamanSpectroscopy(LabRAMHR800)

wasconductedtodetectthemicrostructurevariationontheXC72R

afterCVscans.Ionchromatography(DionexDX120)wasutilized

toanalyzetheconcentrationofdissolvedSO42−andfluorocarbons

fromthedecompositionofNafionionomer.TransmissionElectron

Microscope(TEM;PhilipsTecnai-20)wasusedtoobservethe

mor-phologiesanddistributionsforthePtnanoparticles.TheaveragePt

sizewasobtainedbyTEMimageanalysis(Image-ProPlus6.0).The

amountofPtloadingswasdeterminedbyanInductivelyCoupled

PlasmaMassSpectrometry(ICP-MS;SCIEXELAN5000)wherethe

samplesweredissolvedinasolutioncontainingHCl,HNO3,andHF

ata2:2:1volumeratio.

3. Resultsanddiscussion

3.1. ElectrochemicaldegradationofNafionionomer

Fig.2 providestheCVprofilesatvariouscyclesfor the

elec-trodescontainingcarboncloth,Nafionionomer,andXC72Rwith

thesupplyofambientoxygen.Asshown,theCVprofilesexhibited

acharacteristicbehaviorforcapacitorswithsymmetricresponses

inwhichconsiderableanodicandcathodiccurrentswereobserved

above0.9Vandbelow_−0.1V,respectively.Notably,thecurrent

fromtheanodicscanforthefirstcyclewasnegligibleuntil0.9V

1.2 1.0 0.8 0.6 0.4 0.2 0.0 -0.2 -6 -4 -2 0 2 4 6 20th 10th I (mAcm -2 ) E (V) vs. Ag/AgCl 5th 1st

Fig.2.ProfilesfrommultipleCVscanswithambientoxygenforelectrodes

contain-ingcarboncloth,XC72R,andNafionionomer.

currentsforthecathodicscan,suggestingsurfaceactivationatan

oxidativepotentialabove0.9Vinthefirstcycle.Interestingly,both

theanodicandcathodiccurrentsdemonstratedsteadyincrements

withincreasingCVcycles.Weunderstoodthattherecorded

cur-rentsweremostlyfromtheXC72Rasthecarbonclothcontributed

aninsignificantamountwithitsrelativelyreducedsurfacearea.

However,inourobservation,samplesofXC72Rdepositedonthe

carbonclothrevealedCVcurvesthatwereinsensitivetoincreasing

cycles,agenericbehaviorforelectrochemicaldouble-layer

capac-itors.Therefore,werealizedthattherewaschemicaldegradation

ofNafionionomerthatledtotheincreasingcurrents.

InordertoobservetheeffectofNafionionomerdegradation

more clearly, we need to remove the capacitive currentsfrom

theXC72R.Therefore,wecarriedoutadditionalexperimentswith

theelectrodescontainingcarbonclothandNafionionomeronly.

Fig.3(a)exhibitstheCVprofilesforthesampleswiththesupply

ofambientoxygen.Asshown,thereappearedobviousoxidation

and reduction peaks centering around 0.55 and 0.34V,

respec-tively.Inaddition,thesesignalsincreasedsteadilywithincreasing

cycles. According to literature, these peaks were attributed to

hydroquinone-quinone redox couple on the carbon substrates,

suggestingtheformationofoxygenatedfunctionalgroupsonthe

surface[25–27].Alsoshownisthecarbonclothwithoutthe

addi-tionofNafionionomerbutwiththeoxygensuppliedfromambient.

Obviously,therewasnegligiblecurrentintheCVscans,indicating

thatwithoutNafionionomer,oxygenatedfunctionalgroupsonthe

carbonsurfacewerenotformedatnoticeableamount.

EarlierstudiesontheNafionmembranedegradationhave

iden-tifiedthehydroxyl (•OH)and peroxy(•OOH)radicalstobethe

activespeciestoattackthechemicalstructureofNafion.Itwas

sug-gestedthatthedissolvedoxygendiffusestotheanodesidereacting

withthehydrogentoformhydrogenperoxide[29,30].Inourcase,

withsufficientsupplyofambientoxygen,theCVscansinanacidic

environmentonthecarbonelectrodeswerelikelytoinitiatethe

oxygenreductionreactionbyatwo-electronroutewhichledto

theformationofhydrogenperoxide[41,42].Thishydrogen

perox-idesubsequentlyengenderedthedecompositionofNafionionomer

thatfurtheracceleratedtheoxidationofcarbon.Toverifythe

sig-nificanceofoxygeninthisprocess,werepeatedtheexperiments

withtheelectrodescontainingcarbonclothandNafionionomerbut

withoutthesupplyofambientoxygen.Theeliminationofoxygen

wasachievedbyimmersingtheworkingelectrodetotheelectrolyte

completelyinasealedthree-electrodecellinconjunctionwith

suf-ficientargonpurgingtoremoveanydissolvedoxygen.Theresulting

CVprofilesaredisplayedinFig.3(b).Interestingly,therewas

neg-1.2 1.0 0.8 0.6 0.4 0.2 0.0 -0.2 -3 -2 -1 0 1 2 3 I (mAcm -2 ) E (V) vs. Ag/AgCl

a

b

Fig.3.ProfilesfrommultipleCVscans(a)withambientoxygenand(b)without

ambientoxygenforelectrodescontainingcarbonclothandNafionionomer.Also

shownin(a)istheelectrodewithcarbonclothonly.

ligibledifferencefortheCVresponsesbetweenthefirstand20th

cycle,andthehydroquinone–quinoneredoxcouplewasrather

sub-dued.Thisminuteamountofredoxcouplewaspossiblypresentin

thesamplebeforetheCVscanswereimposed.AccordingtoFig.3,

withoutthesimultaneouspresenceofoxygenandNafionionomer,

theamountofnewlyformedoxygenatedfunctionalgroupsonthe

carbonsurfacewasinsignificant.Analternativeapproachto

pro-ducethehydroxylradical(•OH)isthedirectoxidationofwater

[30].ThisisascenariothatispossibleintheCVscanswithoutthe

supplyofambientoxygen.However,fromFig.3(b)weconcluded

thatthedirectoxidationofwaterwasunabletoproducesufficient

hydroxylradicals(•OH)forNafionionomerdegradation.Therefore,

theprincipalcausefortheformationofoxygenatedfunctionalized

groupsonthecarbonsurfacewastheoxygenreductionroutethat

engenderedthedecompositionofNafionionomer.

TofurthervalidatethecontributoryroleofNafionionomerand

oxygenforcarbonfunctionalizations,additionalexperimentson

thecarbonclothandNafionionomerwerecarriedout.We

per-formed the CVscans with and withoutthe supply of ambient

oxygen,andrecordedtheiranodiccurrentsatthe20thcycle.Fig.4

exhibitsthecomparisonfortheanodiccurrentat0.5Vforboth

sam-ples,aswellasdatafromFigs.2and3(a),respectively.Apparently,

withoutthesupplyofambientoxygen,theanodiccurrentbecame

relativelysubduedforeverysample.Ingeneral,thepresenceof

1E-5 1E-4 1E-3 0.01 0.1 1 10 without O₂ with O₂ CC/ XC72R Nafion ionomer CC/ Nafion ionomer CC I (mAcm -2 ) Nafion ionomer

Fig.4.Comparisoninthecurrentvalueobtainedat0.5Vfromthe20thCVcycle

forelectrodescontainingcarboncloth(CC),Nafionionomer,CC/Nafionionomer,

andCC/XC72R/Nafionionomer.TheseCVexperimentsareperformedwithambient

oxygenandwithoutambientoxygen,respectively.

oxidationcurrent.However,withtheadditionofNafionionomer,

theeffectofoxygenbecamemorepronounced.Itistherefore

con-cludedthatthedegradationofNafionionomer,promotedbythe

presenceofoxygen,ledtoacceleratedcarbonfunctionalization.

3.2. Carbonfunctionalization

Fig.5demonstratestheRamanspectrafortheelectrodesafter

CVscansandH2SO4immersion.Asshown,bothsamplesrevealed

characteristicpeakswhichweredefinedasD-band(1310cm−1)

andG-band(1596cm−1),respectively.TheD-bandrepresentedthe

presenceofdefectsanddisorderinthecarbonstructurewhilethe

G-bandreflectedthegraphiticin-planevibrationswithE2g

sym-metry[17].Hence,theratioofD/Gsignalssuggestedthedegree

ofcrystallinityinthecarbonstructure.Asmentionedearlier,the

electrodeundergoingtheH2SO4immersionwasselectedfor

com-parisonpurposeanditexhibitedaD/Gvalueof2.57.Incontrast,

thesampleafterCVscansrevealedaD/Gvalueof2.67.This

mod-eratevariationintheD/Gratioinferredthatthecarbonstructure

wasreasonablymaintainedafterCVtreatmentsandtheformation

2000 1800 1600 1400 1200 1000 G-band CV scans with O₂ Intensity (a.u.) Raman shift (cm-1) H 2SO4 immersion D-band

Fig.5.RamanspectraforelectrodesafterCVscanswithambientoxygenandH2SO4

immersiononly.Theseelectrodescontaincarboncloth,XC72R,andNafionionomer.

0 200 400 600 800 F1s O1s (b) (c) Intensity (a.u.)

Binding energy (eV)

(a) C1s

F

Fig.6.XPSsurveysfor(a)as-preparedelectrode,aswellaselectrodesafterCVscans

(b)withoutambientoxygenand(c)withambientoxygen.Theseelectrodescontain

carboncloth,XC72R,andNafionionomer.

ofoxygenatedfunctionalgroupswascausedbythedecomposition

ofNafionionomer.

TheXPSwasadoptedtoobtainvariationsonthesignalsfrom

carbon,oxygen,andfluorinefortheelectrodesunderCVscanswith

andwithoutthesupplyofambientoxygen.Wealsoperformedthe

XPSanalysisontheas-preparedelectrodewithoutCVscansfor

comparison.AsshowninFig.6,relevantpeaksontheXPSprofiles

(resolutionin1eV)werelabeledproperlyandtheywere

identi-fiedasF1s,FKLL,O1s,andC1s,respectively.Table1liststheir

respectiveatomicratios.Itcanbeseenthattherewasnegligible

differenceintheatomicratiosbetweensamplesintheas-prepared

stateandafterCVscanswithoutthesupplyofambientoxygen.

However,thesampleafterCVscanswiththesupplyofambient

oxygenrevealedasimilarcarbonamountbutitsatomicratioforthe

oxygenwasincreasedconsiderablyinconjunctionwithanotable

reductioninthefluorinecontent.Theseresultssuggestedthatthe

CVscanscoupledwiththesupplyofambientoxygenwereable

toproduceoxygen-richfunctionalgroupsontheelectrodesurface

whiletheNafionionomerwaspartiallydecomposed.

Fig.7(a)presentstheC1sXPSprofiles(resolutionin0.1eV)for

theas-preparedelectrodeandelectrodesafterCVscanswithand

withoutthesupplyofambientoxygen,respectively.Apparently,

theas-preparedsample andtheoneafterCVscanswithoutthe

supplyofambientoxygendisplayedsimilarpatternsasexpected.In

contrast,thesampleafterCVscanswiththesupplyofambient

oxy-gendemonstratedanotablepeakaround286–288eVinaddition

tothetypicalC1ssignalat284.5eV.Tounderstanditsnature,this

C1sprofilewassubjectedtocurvefittingwithknownfunctional

groupstodeterminetheirrelativeamounts.Fig.7(b)illustratesthe

curvefittingresultsandtheatomicratiosforindividualfunctional

groupsarelistedinTable2.Thesefunctionalgroupswereselected

fromearlierliteraturereportsandwerepresumedtobepresentin

thefunctionalizedelectrodes[23,25–27].FromTable2,the

sam-pleafterCVscanswiththesupplyofambientoxygenrevealeda

Table1

Theatomicratiosforcarbon,oxygen,andfluorinefromXPSprofilesforas-prepared

electrode,aswellaselectrodesafterCVscanswithandwithoutthesupplyof

ambi-entoxygen.

C(at%) O(at%) F(at%)

As-prepared 61 4.3 34.7

CVscanswithoutO2 62 3.5 34.5

Table2

TheatomicratiosfortheC–C,–OH,–C O,–COOH,andC–FfromXPScurvefittingforas-preparedelectrode,aswellaselectrodesafterCVscanswithambientoxygenand withoutambientoxygen.

C–C(at%)(backbone) –OH(at%)(a) –C O(at%)(b) –COOH(at%)(c) C–F(at%) (a+b+c)/C–C

As-prepared 63.7 14 6.3 5.7 10.3 40.8%

CVscanswithoutO2 63.7 14 6.3 5 11 39.7%

CVscanswithO2 48 14 20.6 9.1 8.1 91%

notablereductionintheamountforC–Fgroup.Inaddition,the

oxi-dized–C Oand–COOHgroupsweresubstantiallyincreasedalong

withconsiderablereductionintheC–Cbackbone.

Sofar,ourresultsindicatedthatthedecompositionofNafion

ionomerwasinitiated by theambientoxygenand this process

resultedintheformationofoxygenatedfunctionalgroupsonthe

carbonsurface.Tovalidateourpremise,weattemptedtoobtainthe

S2psignalbutthe0.5MH2SO4aqueoussolutionprovided

unnec-essarybackgroundnoises.Hence,wepreparedseveralelectrodes

(carboncloth/XC72R/Nafionionomer)andsubjectedthemtoCVs

in0.1MHClaqueoussolutioninstead.ThepurposefortheseCV

scanswastodecomposetheNafionionomersotheHClsolution

withconcentratedresidueswasformed.AccordingtoTeranishiet

al.,thedegradationofNafionproducedF−,SO32−,CO2,SO2,and

somefluorocarbons[43].Subsequently,weimmersedtheelectrode

madeofXC72RandcarbonclothintheHClsolutioncontaining

con-280 285

290 295

Intensity (a.u.)

Binding energy (eV) as-prepared CV scans without O₂ CV scans with O₂

a

Binding energy (eV)

b

Fig.7.(a)C1sXPSprofilesforas-preparedelectrode,aswellaselectrodesafterCV

scanswithoutambientoxygenandwithambientoxygen.(b)CurvefittingfortheC

1sXPSprofilefromelectrodeafterCVscanswithambientoxygen.Theseelectrodes

containcarboncloth,XC72R,andNafionionomer.

Fig.8.IonchromatogramforNafionionomerdegradationin0.1MHClaqueous

solution.

centratedNafionionomerresiduestoallowsufficientadsorptionof

thedecomposedspecies.AsshowninFig.8,signalsfromtheion

chromatographywereattributedtoSO42−andF−indifferent

inten-sities.SimilarconstituentswereobservedinearlierworkbyChen

andFullerforNafionmembranedegradation[44].Intheirwork,

aratherstrongCF3COO−peakwasidentifiedonthecathodeside

associatedwiththeoxygenreductionreaction.Unfortunately,in

ourcasetheamountofCF3COO− wasbelowthedetectionlimit.

ThisnotableabsenceofCF3COO− waspossiblyduetoits

imme-diatereadsorptionontothecarbonsurfaceafterdetachmentfrom

theNafionbackbone.

To monitor the extent of Nafion ionomer degradation, we

recordedthesignalfor SO42− uponCVcyclesandtheresulting

dataaredisplayedinFig.9.Thevalueforthe0cyclewasobtained

fromthesamplewithimmersioninthe0.1MHClaqueous

solu-tionfor17min,whichservedasthereferencebecausethesample

100 80 60 40 20 0 0 1 2 3 4 5 2rd cycle Sulfate concentration (ppm) CV cycle numbers Immersion in HCl 1st cycle

Fig.9.VariationofsulfateconcentrationasafunctionofCVscanswithambient

oxygen.Thedataat0thcycleisobtainedfromtheelectrodeimmersedin0.1MHCl

Table3

TheatomicratiosforC–C,–OH,–C O,–COOH,andC–FfromC1sXPScurvefittingforas-preparedelectrode,aswellaselectrodesmadeofXC72R/carbonclothwithand

withoutimmersioninHClsolutioncontainingconcentratedresiduesfromNafionionomerdecomposition.

C–C(at%)(backbone) –OH(at%)(a) –C O(at%)(b) –COOH(at%)(c) C–F(at%) (a+b+c)/C–C

As-prepared 63.7 14 6.3 5.7 10.3 40.8%

XC72R/carbonclothwithoutimmersion 63.7 17.2 8.3 7 3.8 51%

XC72R/carbonclothwithimmersion 46.5 11.6 8.2 27.9 5.8 102.5%

of20CVcyclesexperiencedthesameamountoftimeinthe0.5M

H2SO4aqueoussolution.Asshown,thereferencesamplerevealed

sulfateconcentrationof0.35ppm.Thisreducedamountwasnot

unexpectedastheNafionionomerlikelymaintainedreasonable

chemicalstabilityagainstthe0.1MHClaqueoussolutionat26◦C.

However,onceCVscanswereapplied,thesulfateanion

concentra-tionsbecamelargerconsiderablyreachingaplateauafter20cycles

at4.3ppm.Apparently,withinthefirst20cycles,thereappeared

alinearincreaseinthesulfateconcentrationwithcyclingnumber.

ThisindicatedthatadesirableamountofNafiondecompositionand

itssubsequentcarbonfunctionalizationwaspossiblebyselecting

appropriateCVscans.

Fig.10providestheC1sXPSprofiles(resolutionin0.1eV)for

theas-prepared electrode(carboncloth/XC72R/Nafion ionomer)

as well as electrodes (carbon cloth/XC72R) with and without

immersionintheHCl solutioncontainingconcentratedresidues

fromNafionionomerdecomposition.Apparently,theelectrodeof

XC72R/carbonclothdemonstrated a singleC 1speak at 284eV

whiletheas-preparedelectrodeexhibitedanadditionalC–Fpeak

around 291eV. However, the electrode of XC72R/carbon cloth

immersed in the HCl solution with concentrated decomposed

Nafionionomerresiduesrevealedastrongsignalaround289eV

whichwasattributedtotheoxygenatedgroupsonthecarbon

sur-280 284 288 292

(c)

(b)

Binding energy (eV)

(a)

Fig.10.C1sXPSprofilesfor(a)as-preparedelectrode(carboncloth/XC72R/Nafion

ionomer),aswellaselectrodes (carboncloth/XC72R)(b)before and(c)after

immersioninHClsolutioncontainingconcentratedresiduesfromNafionionomer

decomposition.

face.Table3liststheatomicratiosfortheindividualfunctional

groupsfromthecurvefittingresultsofFig.10.Apparently,the

sam-pleafterimmersingintheHClsolutionshowedalargeamount

of oxygenatedfunctional groups. We surmised that theNafion

ionomerresidueintheHClsolutionwaslikelypresentasCF3COO−.

Afterchemicaladsorption,theseresiduesweretransformedtothe

oxygenatedfunctionalgroupsonthecarbonsurface.

ThechemicaladsorptionofNafionionomerresiduescanalso

beconfirmedfromtheS2pXPSprofile(resolutionin0.1eV)

dis-playedinFig.11.TheelectrodeofXC72R/carbonclothrevealeda

characteristicS2psignalnear164eVwhichwasattributedtothe

impurityintrinsictothecarbonmaterial.However,theelectrode

ofXC72R/carboncloth/Nafionionomerdemonstratedanadditional

peakaround168eVwhichwascausedbytheHSO3fromtheNafion

ionomer.Interestingly,theXC72R/carbonclothsampleimmersed

intheHClsolutionwithconcentratedNafionionomerdecomposed

residuesalsoexhibitedtheHSO3signalinadditiontotheS2pfrom

impurity.Thisfurthersupportedourpremisethatthedecomposed

Nafionionomerresidueswereabletochemicallyadsorbontothe

carbonsurface. 150 155 160 165 170 175 180

Binding energy (eV)

(c)

(b)

(a)

Fig.11.S2pXPSprofilesfor(a)as-preparedelectrode(carboncloth/XC72R/Nafion

ionomer),aswellaselectrodes(carboncloth/XC72R)(b) beforeand(c) after

immersioninHClsolutioncontainingconcentratedresiduesfromNafionionomer

Table4

ElectrochemicalparametersobtainedfromCVprofilesonfunctionalizedandbaselineelectrodesformethanolelectro-oxidation.

Pta_{Loading(␮gcm}−2_{) Anodicscan Cathodicscan}

Vab iac iad iae Vcf icg ich ici

mV mAcm−2 mAPt−1mg−1 mAPt−1cm−2 mV mAcm−2 mAPt−1mg−1 mAPt−1cm−2

Functionalizedelectrode 511 729 38.9 76.1 0.47 502 43.7 85.5 0.53

Baselineelectrode 301 677 20.4 67.7 0.33 428 19.7 65.4 0.32

a_{totalweightofPtasdeterminedbyICP-MS.}

b_{peakpotentialinanodicscan.}

c _{peakapparentcurrentdensityinanodicscan.}

d _{peakmassactivityinanodicscan.}

e_{peakPtsurfaceactivityinanodicscan.}

f _{peakpotentialincathodicscan.}

g_{peakapparentcurrentdensityincathodicscan.}

h _{peakmassactivityincathodicscan.}

i _{peakPtsurfaceactivityincathodicscan.}

Fig.12.TEMimagesfordepositedPtnanoparticleson(a)functionalizedand(b)

baselineelectrodes.

3.3. Methanolelectro-oxidation

Fig. 12(a)demonstrates theTEMimage for Pt nanoparticles

deposited onthe functionalized electrode followed by

electro-chemicalreduction.Asshown,therewereplentyPtnanoparticles

uniformly distributed with notable aggregations. The primary

particlesizefromtheimageanalysissoftwarewas2.68_±1.62nm.

TheTEMimageonthebaselineelectrode(simpleH2SO4immersion

followedbyelectrochemicalreduction)ispresentedinFig.12(b).

Apparently, the amount of Pt nanoparticles was substantially

reduced, a fact consistent with earlier findings from ICP-MS.

In addition, their size was slightly smaller at 2.20±1.45nm.

These results confirmed that the functionalized electrode

enabled a larger amount of Pt deposits, albeit with moderate

coalescence.

Fig.13 presentsthe CVprofiles ofhydrogen desorptionand

adsorptionforthefunctionalizedandbaselineelectrodes,

respec-tively.Asexpected,thereappearedstrongerresponsesinhydrogen

desorptionandadsorptionforthefunctionalizedelectrodebecause

of its relativelylargeramount of Ptdeposit. Estimationon the

ECSA wasconductedby theintegral areafor hydrogen

desorp-tion in the anodic scan and the resulting ECSA values were

82.2 and 60.9cm2 _{for the functionalized and baseline}

elec-trodes, respectively. This ratio of 1.35 was smaller to that of

1.70for thePtloadingfromICP-MS.We attributedthereduced

ESCAratiototheobservedPtaggregationonthefunctionalized

electrode.

Fig. 14(a) provides the CV profiles for methanol

electro-oxidationinapparentcurrentdensityforthefunctionalizedand

1.0 0.8 0.6 0.4 0.2 0.0 -0.2 -15 -10 -5 0 5 10 functionalized baseline I / mAcm -2 E / V vs. Ag/AgCl

1.0 0.8 0.6 0.4 0.2 0.0 -0.2 -10 0 10 20 30 40 _{functionalized} baseline I / m A c m -2 E / V vs. Ag/AgCl

(a)

(b)

(c)

Fig.14.CVprofilesforfunctionalizedandbaselineelectrodesonmethanol

electro-oxidationin (a)apparent currentdensity, (b) mass activity, and (c) unit Pt

electrochemicalsurfacearea.

baselineelectrodes,respectively.Relevantelectrochemical

param-etersfromthesecurvesarelistedinTable4.Accordingtoliterature,

intheseprofilestheanodicpeak(ia)isattributedtotheoxidation

ofmethanolwhilethecathodicpeak(ic)correspondstothe

oxi-dationof carbonaceousspeciesproduced fromearliermethanol

oxidation[45–47].Inaddition,theratio(ia/ic)indicatesthe

elec-trocatalyticabilitytoremoveCO.Hence,anelectrodewithahigher

1800 1200 600 0 4 8 12 16 20 functionalized baseline I / mAmg -1 Time (sec)

Fig. 15.Chronoamperograms for functionalized and baseline electrodes on

methanolelectro-oxidationat0.5Vfor30mininmassactivity.

apparentcurrentandalargeria/icratioisalwaysdesirable.Notably,

thefunctionalizedelectrode demonstrateda substantialcurrent

incrementoverthatofbaselineelectrode.Thisnotable

improve-mentwaspartiallycausedbyalargerPtdepositswhichledtoa

highernominalcurrent.However,theia/icratioforthe

function-alizedelectrodewas0.89,whichwasslightlysmallerthan1.03of

baselineelectrode.Tocomparefairly,itisnecessarytoreplotthe

CVprofilesinmassactivityandunitPtECSA,asshowninFig.14(b)

and(c).Apparently,thereappearedaconsistenttrendinwhichthe

functionalizedelectroderevealedsignificantenhancementsover

thatofbaselineelectrode.Theseresultssuggestedthatthe

oxy-genatedfunctionalgroupswerelikelycontributingtothemethanol

electro-oxidation.Similarphenomenawerealsoreported

previ-ouslyinwhichtheoxidizedfunctional groupswerebelievedto

provideoxygen-richspeciestofacilitateCOoxidationonPtsurface

[26,48,49].

AfterconfirmingenhancementsinCVsforthefunctionalized

electrode, itis necessarytoevaluateitschronoamperogram for

lifetimedetermination.Fig.15providesthechronoamperograms

forthefunctionalizedandbaselineelectrodesat0.5Vfor30min

inmassactivity.Apparently,bothelectrodesdisplayedanotable

currentdecayinthefirst20min.However,theamountofcurrent

decaywasrelativelyconstantforbothelectrodesandthe

function-alizedelectrodewasconsistentlybetterthanthebaselineelectrode.

SincethePtwasusedinourstudy,theseperformance

degrada-tionswerenotunexpectedaspoisonousintermediateswereable

toadsorbonthePtsurfacecompromisingitscatalyticabilityfor

methanoloxidation.Itisnotedthatsimilarbehaviorswerealso

observedbyMaetal.intheirstudyofPt–Ru(OxHy)m

electrocata-lysts[50].

Sofar,ourworkdemonstratesafacileapproachto

functional-izecatalystsupportswithoutinvolvinghightemperatureandlarge

anodicpotentials.SincetheNafionionomeritselfisoftenusedas

abinderinelectrodefabrication,asimpleCVwithdissolved

oxy-gennearbycoulddecomposetheNafionionomerpartiallyresulting

intheformationofoxygenatedfunctionalgroups.Thesefunctional

groupsareactiveinassistingthePtformethanolelectro-oxidation.

It is noted that the enhancement effect observed in this work

is differentfrom conventional approaches in which the role of

NafionionomeristoextendtheinterfacialareabetweentheNafion

ionomerandelectrocatalyst[51].Moreover,thefunctionalgroups

couldpotentiallyenablealargerelectrocatalystimpregnation

lead-ingtoanimprovedstability[52].Furtherstudiesareunderwayto

exploretheanchoringeffectforthefunctionalizedgroupsandlife

4. Conclusions

WeconductedmultipleCVscansinanacidicelectrolyteonthe

electrodescontainingcarboncloth,XC72R,and Nafionionomer.

Withthesupplyofambientoxygen,theNafionionomer

experi-encedchemicalattacksleavingdecomposedresidueswhichwere

abletoadsorbontothecarbonsurfaceleadingtoanaccelerated

formationofoxygenatedfunctionalgroups.Thedecompositionof

Nafionionomerwasattributedtothehydrogenperoxideproduced

fromtheoxygenreductionreactionduringCVscans.Raman

analy-sisonthecarbonelectrodesrevealedminorstructuralmodification

afterCVscans.ResultsfromXPSsurveysindicated a significant

increaseoftheoxygenatedfunctionalgroupsonthecarbon

sur-faceinconjunctionwithanotablereductioninfluorinecontent.

The functionalized electrode was determined to allowa larger

amountofPtionadsorptionascomparedtothebaselineelectrode.

Afterelectrochemicalreduction,thePtnanoparticleswereevenly

formedonthecarbonsupports.Electrochemicalanalysisonthe

methanolelectro-oxidationwasperformedandweobserved

sig-nificantincrementsinapparentcurrentdensity,massactivity,as

wellasunitPtESCAforthefunctionalizedelectrode.

Acknowledgement

TheauthorsaregratefultoProfessorChuen-JinnTsaiandMiss

Yi-LinLiufromtheInstituteofEnvironmentalEngineeringfortheir

kindassistancewiththelaboratoryequipment.Financialsupports

fromNationalScienceCouncilofTaiwan(NSC-96-2221-E-009-110)

andNationalSynchrotronRadiationResearchCenter

(2009-2-063-1)areacknowledged.

References

[1] A.L.Dicks,J.PowerSources156(2006)128–141.

[2]M.S.Saha,A.Kundu,J.PowerSources195(2010)6255–6261.

[3]J.F.Drillet,H.Bueb,R.Dittmeyer,U.Dettlaff-Weglikowska,S.Roth,J.

Elec-trochem.Soc.156(2009)F137–F144.

[4]J.H.Kim,B.Fang,S.B.Yoon,J.S.Yu,Appl.Catal.B88(2009)368–375.

[5]J.F.Lin,V.Kamavaram,A.M.Kannan,J.PowerSources195(2010)466–470.

[6]A. Guha, T.A. Zawodzinski, D.A. Schiraldi, J. Power Sources 195 (2010)

5167–5175.

[7]N.Spataru,X.T.Zhang,T.Spataru,D.A.Tryk,A.Fujishima,J.Electrochem.Soc.

155(2008)B264–B269.

[8]V.Baglio,A.DiBlasi,C.D’Urso,V.Antonucci,A.S.Arico,R.Ornelas,D.

Morales-Acosta,J.Ledesma-Garcia,L.A.Godinez,L.G.Arriaga,L.Alvarez-Contreras,J.

Electrochem.Soc.155(2008)B829–B833.

[9]N.Y.Hsu,C.C.Chien,K.T.Jeng,Appl.Catal.B84(2008)196–203.

[10]S.Maass,F.Finsterwalder,G.Frank,R.Hartmann,C.Merten,J.PowerSources

176(2008)444–451.

[11]W.Phompan,N.Hansupalak,J.PowerSources196(2011)147–152.

[12] Y.F.Hsieh,Y.C.Hsieh,P.W.Wu,C.H.Liao,Y.M.Chang,J.Electrochem.Soc.157

(2010)B39–B44.

[13]J.R.C.Salgado,F.Alcaide,G.Alvarez,L.Calvillo,M.J.Lazaro,E.Pastor,J.Power

Sources195(2010)4022–4029.

[14]Y.M.Chang,Y.C.Hsieh,P.W.Wu,DiamondRelat.Mater.18(2009)501–504.

[15]R.I.Jafri,T.Arockiados,N.Rajalakshmi,S.Ramaprabhu,J.Electrochem.Soc.157

(2010)B874–B879.

[16]B.Liu,S.Creager,J.PowerSources195(2010)1812–1820.

[17]G.X.Zhang,S.H.Sun,D.Q.Yang,J.P.Dodelet,E.Sacher,Carbon46(2008)

196–205.

[18]P.L.Antonucci,F.Romeo,M.Minutoli,E.Alderucci,N.Giordano,Carbon26

(1988)197–203.

[19]N.Giordano,P.L.Antonucci,E.Passalacqua,L.Pino,A.S.Arico,K.Kinoshita,

Electrochim.Acta36(1991)1931–1935.

[20]E.Passalacqua,P.L.Antonucci,M.Vivaldi,A.Patti,V.Antonucci,N.Giordano,K.

Kinoshita,Electrochim.Acta37(1992)2725–2730.

[21]S.I.Pyun,E.J.Lee,T.Y.Kim,S.J.Lee,Y.G.Ryu,C.S.Kim,Carbon32(1994)

155–159.

[22] A.M.Puziy,O.I.Poddubnaya,V.N.Zaitsev,O.P.Konoplitska,Appl.Surf.Sci.221

(2004)421–429.

[23]Z.R.Yue,W.Jiang,L.Wang,S.D.Gardner,C.U.Pittman,Carbon37(1999)

1785–1796.

[24]R.Berenquer,J.P.Marco-Lozar,C.Quijada,D.Cazorla-Amoros,E.Morallon,

Car-bon47(2009)1018–1027.

[25]K.H.Kangasniemi,D.A. Condit,T.D.Jarvi, J.Electrochem.Soc.151 (2004)

E125–E132.

[26]S.Stevanovic,V.Panic,D.Tripkovic,V.M.Jovanovic,Electrochem.Commun.11

(2009)18–21.

[27]J.J.Wang,G.P.Yin,Y.Y.Shao,S.Zhang,Z.B.Wang,Y.Z.Gao,J.PowerSources171

(2007)331–339.

[28]C.Chen,T.F.Fuller,J.Electrochem.Soc.156(2009)B1218–B1224.

[29] C.Chen,G.Levitin,D.W.Hess,T.F.Fuller,J.PowerSources169(2007)288–295.

[30]F.A.deBruijn,V.A.T.Dam,G.J.M.Janssen,FuelCells8(2008)3–22.

[31]D.A.Schiraldi,Polym.Rev.46(2006)315–327.

[32]D.E.Curtin,R.D.Lousenberg,T.J.Henry,P.C.Tangeman,M.E.Tisack,J.Power

Sources131(2004)41–48.

[33]G.Hubner,E.Roduner,J.Mater.Chem.9(1999)409–418.

[34]T.Xie,C.A.Hayden,Polymer48(2007)5497–5506.

[35]H.P.Boehm,Carbon32(1994)759–769.

[36] A.Bravo,H.R.Bjorsvik,F.Fontana,L.Liguori,A.Mele,F.Minisci,J.Org.Chem.

62(1997)7128–7136.

[37]M.Sansotera,C.L.Bianchi,G.Lecardi,G.Marchionni,P.Metrangolo,G.Resnati,

W.Navarrini,Chem.Mater.21(2009)4498–4504.

[38] J.Xie,D.L.Wood,K.L.More,P.Atanassov,R.L.Borup,J.Electrochem.Soc.152

(2005)A1011–A1020.

[39]T.Morimoto,K.Hiratsuka,Y.Sanada,K.Kurihara,J.PowerSources60(1996)

239–247.

[40] C.T.Hsieh,H.Teng,Carbon40(2002)667–674.

[41] K.Kinoshita,Oxygenelectrochemistry,in:ElectrochemicalOxygenTechnology,

JohnWiley&Sons,NewYork,1992,pp.21–26.

[42]C.Chen,T.F.Fuller,Electrochim.Acta54(2009)3984–3995.

[43] K.Teranishi,K.Kawata,S.Tsushima,S.Hirai,Electrochem.Solid-StateLett.9

(2006)A475–A477.

[44]C.Chen,T.F.Fuller,Polym.Degrad.Stab.94(2009)1436–1447.

[45]T.C.Deivaraj,J.Y.Lee,J.PowerSources142(2005)43–49.

[46] Z.B.Wang,G.P.Yin,J.Zhang,Y.C.Sun,P.F.Shi,Electrochim.Acta51(2006)

5691–5697.

[47]Z.L.Liu,X.Y.Ling,B.Guo,L.Hong,J.Y.Lee,J.PowerSources167(2007)272–280.

[48]J.H.Chen,M.Y.Wang,B.Liu,Z.Fan,K.Z.Cui,Y.Kuang,J.Phys.Chem.B110

(2006)11775–11779.

[49]M.A.Scibioh,I.H.Oh,T.H.Lim,S.A.Hong,H.Y.Ha,Appl.Catal.B77(2008)

373–385.

[50]J.H.Ma,Y.Y.Feng,J.Yu, D.Zhao,A.J.Wang,B.Q.Xu,J. Catal.275(2010)

34–44.

[51]C.H.Park,M.A.Scibioh,H.J.Kim,I.H.Oh,S.A.Hong,H.Y.Ha,J.PowerSources

162(2006)1023–1028.

[52]W.M.Chen,Q.Xin,G.Q.Sun,Q.Wang,Q.Mao,H.D.Su,J.PowerSources180