Journal of Electroanalytical Chemistry 671 (2012) 24–32

Contents lists available at SciVerse ScienceDirect

Journal of Electroanalytical Chemistry

journal homepage: www.elsevier .com/locate / je lechem

The kinetics of the hydrogen oxidation reaction on WC/Pt catalyst with lowcontent of Pt nano-particles

M.D. Obradovic a, S.Lj. Gojkovic b, N.R. Elezovic c, P. Ercius d, V.R. Radmilovic b, Lj.D. Vracar b,N.V. Krstajic b,⇑a Institute of Chemistry, Technology and Metallurgy – ICTM, University of Belgrade, 11000 Belgrade, Njegoseva 12, Serbiab Faculty of Technology and Metallurgy, University of Belgrade, 11000 Belgrade, Karnegieva 4, Serbiac Institute for Multidisciplinary Researches, University of Belgrade, 11030 Belgrade, Kneza Vislesava 1, Serbiad National Center for Electron Microscopy, LBLN University of California, Berkeley, USA

a r t i c l e i n f o a b s t r a c t

Article history:Received 14 October 2011Received in revised form 30 December 2011Accepted 27 January 2012Available online 24 February 2012

Keywords:Hydrogen oxidationMechanismTungsten carbidePlatinum catalyst

1572-6657/$ - see front matter � 2012 Elsevier B.V. Adoi:10.1016/j.jelechem.2012.01.026

⇑ Corresponding author. Tel.: +381 113303682.E-mail address: nedeljko@tmf.bg.ac.rs (N.V. Krstaj

The catalytic activity of WC/Pt electrocatalysts towards hydrogen oxidation reaction (HOR) in acid solu-tion was studied. Tungsten carbide (WC) prepared by polycondensation of resorcinol and formaldehydein the presence of ammonium metatungstate salt and CTABr surfactant was used as the support of a Ptelectrocatalyst (WC/Pt). The obtained WC/Pt electrodes were characterized by XRD, HRTEM, EDS, EELSand electrochemical measurements. HRTEM analysis showed that the WC particles possess a core–shellstructure with a metallic tungsten core and a shell composed of a mixture of tungsten carbides shell (WCand W2C). The WC/Pt catalyst is composed of well-dispersed sub-nanometer Pt clusters which consist of afew to several tens of Pt atoms. EELS measurements indicate that the WC particles function as nucleationsites for Pt nanoparticles. Based on the Tafel–Heyrovsky–Volmer mechanism the corresponding kineticequations were derived to describe the HOR current–potential behavior over the entire potential regionon RDE. The fitting showed that in the lower potential region HOR on Pt proceeds most likely via theTafel–Volmer (TV) pathway. The kinetic results also showed that the WC/Pt(1%) when compared to thestandard C/Pt(1%) electrode led to a remarkable enhancement of the hydrogen oxidation in an acidicmedium, which was explained by H-spill-over between platinum and tungsten carbide.

� 2012 Elsevier B.V. All rights reserved.

1. Introduction

Tungsten carbide (WC) has been considered as an anode mate-rial for hydrogen [1] or methanol [2] polymer electrolyte mem-brane fuel cells (PEMFCs) since Leavy and Boundort [3] firstshowed that WC materials possess catalytic properties similar tothose of platinum group metals, due to their isoelectronic structureto platinum. Unfortunately, WC is not an inert material. When ex-posed to water, WC undergoes continuous oxidation and dissolu-tion [4,5]. The exact nature of the formed tungsten oxides isdifficult to characterize and is strongly dependent on the appliedpotential, electrolyte composition and surface pretreatment [6].The relatively undefined composition of the surface oxide layersis probably the major reason for the irreproducible hydrogenadsorption potentials at tungsten carbide [7]. The activity forhydrogen oxidation reaction (HOR) and stability of high area tung-sten carbides in acidic electrolytes depend on their preparationmethod. The activity of WC was related to the carbon deficiencyand oxygen replacement in carbon layers of the WC lattice [8,9].

ll rights reserved.

ic).

Most significant for the electrocatalysis at tungsten carbide is thepresence of oxygen species (tungsten oxides) at the catalyst surface.

Ross and Stonehart [10] discussed the effect of the surface com-position of tungsten carbides on the activity for HOR. The HORactivity of the carbon deficient, oxygen containing carbides wassignificantly higher than that of the stoichiometric carbide. Theincreased activity of the oxygen substituted carbide was due to areduced interaction of the surface with the electrolyte, resultingfrom the covalent tungsten-oxygen bonding. However, when usedas an anodic material under PEFC conditions, WC alone exhibitspoor electrocatalytic activity although it showed tolerance to COpoisoning [11,12]. The HOR activity of WC was four orders ofmagnitude lower than that of Pt [10]. One of the reasons is itslow specific area, due to a high preparation temperature whichlead to formation of rather large particle sizes.

Lately, a great effort has been directed towards studies on thefundamental surface properties and new methods for preparationof WC materials that could be used as an electrocatalyst [13,14].It has been reported that WC possesses three different crystallinephases (b-W2C, a-WC, and b-WC1�x) depending on the syntheticroute and reaction conditions [15]. In order to improve the catalyticactivity of tungsten carbide towards HOR the effect of the addition

M.D. Obradovic et al. / Journal of Electroanalytical Chemistry 671 (2012) 24–32 25

of a second metal, such as Ni, Co, Fe, Mn or Mo was also investigated[16–18]. However, addition of Ni, Fe or Mn resulted in a lower activ-ity for HOR, owing to the reduced surface area of the catalysts. Atthe same time, Co–WC catalyst carburized at 600 �C exhibited bet-ter catalytic activity as compared to WC catalyst [18].

It has been found that combination of Pt with botha-WC [19–21],b-W2C [22] and b-WC [23] resulted in a high catalytic activity forhydrogen oxidation. While the W2C/Pt catalyst showed similarHOR activity and kinetics as C/Pt catalyst [22], WC/Pt catalystsprepared with high surface area WC exhibited higher specific activ-ity then C/Pt catalyst [24,25]. It has been postulated that Pt wasaccelerating the dissociative adsorption of H2, which was the rate-determining step for the HOR [25] and WC could take over the reststeps in the mechanism of the hydrogen oxidation reaction.

In addition, it has been found that the stability of WC wasextended to higher positive potentials in the presence of a smallamount of Pt [26]. The enhanced stability of the Pt–WC surfacewas attributed to the strong bonding between Pt and WC, mostlikely at defect regions of the WC surface, which prevents thesurface oxidation. Some reports found that the charge for Hadsorption/desorption on Pt supported on WC was higher com-pared with that on C supports [27,28], which was explained byH+-spill-over between Pt and WC.

According to our best knowledge, in the published reports, theHOR activity of various tungsten carbides or Pt/WC catalysts wasmainly treated by comparing RDE polarization curves, or the ki-netic currents determined using Koutecky–Levich linear relation-ship, at the selected potential.

The aim of the present paper is to demonstrate that the kineticsof the HOR at WC/Pt catalysts can be generally treated through theTafel–Heyrovsky–Volmer pathway, and it is possible to determinekinetic parameters of those elemental steps by analyzing RDEpolarization curves. The kinetics of the HOR was investigated athome-made WC/Pt and commercial C/Pt catalysts in order toexamine the specific role of WC on enhanced activity of Pt towardsthe HOR. The Pt catalyst content was very low (1 mass%) in bothelectrodes, because that is the only way to see any noticeable dif-ference in their RDE polarization curves, taking into account thefact that the HOR is an extremely fast reaction on Pt.

2. Experimental

2.1. Preparation of WC

Mesoporous WC was prepared by polycondensation of resorcinol(99% purity E. Merck) and formaldehyde (Fluka Chemie) in the pres-ence of cetyltrimethylammonium bromide (CTABr) surfactant,using the modified method proposed by Ganesan et al. [28]. In a typ-ical synthesis, 6.14 g of CTABr was dissolved in 20 ml of distilledwater and added to the solution containing 4.073 g of ammoniummetatungstate (AMT), 1.13 g of resorcinol and 1.7 ml of formalde-hyde in 10 ml H2O. Then the solution was decanted in a glass tube,sealed and placed for 3 days at 25 �C, 1 day at 50 �C and 3 days at85 �C. During this procedure the solution was transformed to gelfrom which cryogel was prepared by the freeze-drying method[29]. The gel was immersed in a tenfold higher volume of t-butanol(p.a. quality, Centrohem, Beograd) for one day and rinsed twice withnew t-butanol to displace the liquid contained in the gel. The samplewas pre-frozen at�30 �C for 24 h. After that, it was dryed frozen for20 h at the pressure of 4 mbar. The red colored cryogel was calcinat-ed at 1173 K for 1 h in Ar flow and 2 h in H2 flow (2 cm3 s�1) [30].

2.2. Physicochemical characterization

The BET surface area and pore size distribution of the WCsample were calculated from nitrogen adsorption/desorption

isotherms at �196 �C, using the gravimetric McBain method. Poresize distribution was estimated by applying the BJH method [31]to the desorption branch of isotherms, and mesoporous surfaceand micropore volume were estimated using the high resolutionas plot method [32].

The phase structure, crystallinity and size of the synthesizedWC and WC/Pt catalysts were studied with an X-ray diffractometer(XRD, JEOL 6300F microscope) with Cu Ka radiation(k = 0.154056 nm).

Transmission electron microscopy (TEM), measurements wereperformed using the FEI (Fillips Electronic Instruments)-CM200-FEG super-twin and TEAM0.5 ultra-twin transmission electronmicroscopes, operating at 200 kV and equipped with the Gatan1k � 1k and 2k � 2k CCD cameras, respectively. Specimens wereprepared for transmission electron microscopy by making suspen-sion of the catalyst powder in ethanol, in an ultrasonic bath. Thesuspension was dropped onto clean holey carbon grids and thendried in air. The chemical composition of W–WC core–shell parti-cles were characterized by energy dispersive X-ray spectroscopy(EDS) and electron energy loss spectroscopy (EELS).

2.3. Catalyst preparation

The Pt catalyst was deposited on a WC support by a conven-tional borohydride reduction method. The preparation processcan be described as follows: 40 mg of the WC powder was dis-persed in 20 ml of high-purity water (Millipore, 18 MX cm) inultrasonic bath, and then mixed with appropriate amount ofH2PtCl6 aqueous solution (10 mg ml�1). The mixture of metal saltand support was reduced by using an excess of sodium borohy-dride solution. The precipitate was washed with high-purity waterand then dried at 80 �C. The Pt loading of the samples was 1 mass%and 10 mass%.

2.4. Electrode preparation

Four milligrams of WC/Pt catalysts was ultrasonically sus-pended in 1.0 ml of 2-propanol and 50 ll of Nafion solution(5 mass%, Aldrich) to prepare catalyst inks. Then, 10.0 ll of inkwas transferred with an injector to the clean gold disk electrode(5 mm diameter, with geometric surface area of 0.196 cm2). Aftervolatilization of alcohol, the electrode was heated at 80 �C for10 min. The Pt loading was 0.4 lg (1 mass%). In order to comparethe catalytic activity of WC/Pt against a conventional C supportedPt catalyst, a commercial catalyst XC-72R/Pt (20 mass%, E-Tek)was thoroughly mixed and homogenized with appropriate amountof XC-72R powder to attain 1 mass% of Pt loading.

2.5. Electrochemical characterization

A conventional three-compartment cell was used for electro-chemical characterization. The working electrode compartmentwas separated by fritted glass discs from other two compartments.A reversible hydrogen electrode (RHE) in the same solution as thatof the working electrode was used as the reference electrode. Alarge-area platinum sheet of 5 cm2 geometric area was used asthe counter electrode. The electrochemical measurements wereperformed in 0.5 mol dm�3 HClO4 solution (Spectrograde, Merck),prepared with high-purity water, at the temperature of 25 �C.

The experiments were performed by potentiodynamic method.A PAR Model 273 Potentiostat/Galvanostat was used for all electro-chemical experiments. Polarization curves for the HOR were re-corded at the scan rate of 2 mV s�1.

The cyclic voltammetry (CV) experiments were carried out inthe potential range between hydrogen and oxygen evolution in

26 M.D. Obradovic et al. / Journal of Electroanalytical Chemistry 671 (2012) 24–32

N2 saturated 0.5 mol dm�3 HClO4 with various scan rates at a rotat-ing speed of 2500 rpm.

The electrochemically active surface area of Pt was determinedby the stripping of underpotentially deposited (upd) copper [33].This method is applicable to WC + Pt system since the anodic peakof Cu stripping is more positive than the peak for hydrogen dein-tercalation from the hydrous tungsten oxide. In addition, WC alonewas found [34] to be inactive for Cu upd. Copper was underpoten-tially deposited from the unstirred solution of 0.10 mol dm�3

H2SO4 and 2.0 � 10�3 mol dm�3 CuSO4 at the potential of 0.330 Vvs. RHE, which is about 15 mV more positive than the equilibriumpotential of Cu electrodeposition in the electrolyte applied. After2 min of deposition, which is enough to form a for full monolayer,the electrode potential was swept anodically and a stripping vol-tammogram was recorded.

3. Results and discussion

3.1. Characterization of WC and WC–Pt

An XRD pattern of the as-prepared WC is presented in Fig. 1. Thediffraction peaks at 31.77�, 35.98�, and 48.26� correspond to (001),(100), and (101) facets of WC, respectively, while the 2H of34.52�, 38.03�, 39.57�, 52.3�, 61.8�, and 69.77� correspond to(100), (002), (101), (102), (110), and (103) facets of W2C. Inaddition, the sharp peaks at 40.26�, 58.27�, and 73.19� correspondto (110), (200), and (211) facets of metallic W. Hence, the XRDpattern consists of well defined diffraction peaks of W, W2C, andWC phases and does not indicate the presence of tungsten oxidesafter preparation procedure, which means that during carburiza-tion process all of the starting amount of tungsten oxide was con-verted into a mixture of WC, W2C and metallic W.

An XRD pattern of WC/Pt (not shown) does not show peaks aris-ing from the Pt nanoparticles due to very low Pt content.

Figs. 2a and 3a depict a typical high-angle angular dark field(HAADF) scanning transmission electron microscope (STEM)images of the as prepared WC and WC/Pt with 1 wt.% Pt, respec-tively. The tungsten carbides particles are nanocrystals with a typ-ical bimodal particle size distribution: larger above 5 nm indiameter and smaller ones below 2 nm in diameter (Fig. 2a). Thecorresponding HRTEM image of WC particle (Fig. 2b) shows a com-plex core–shell structure with a core of metallic tungsten and a shellmade of a mixture of tungsten carbides (WC and W2C). Elemental

Fig. 1. XRD pattern of WC support.

analysis of the shell by EDS is shown in the inset of the Fig. 2b.The O element detected in the shell is probably associated withthe reaction of the highly active surface of tungsten carbides withthe oxygen from the air and corresponding tungsten oxides areprobably amorphous in nature. Taking into account the fact thatpreparation procedure was conducted at 900 �C the origin of O ele-ment in the shell is not non-converted tungsten oxides, because thecorresponding diffraction peaks of oxide phases must be present inXRD pattern. It has been well documented that under normal condi-tions, on contact with ambient atmosphere the fresh tungsten car-bide chemisorbed oxygen resulting in the formation of WOx

species [35,36]. It can be concluded that after 3 h of carburizationat 900 �C, carbon reduced the tungsten oxide in the precursor to amixture of W, WC, and W2C. Also, there is carbon in excess andthe supporting material consists of un-reacted amorphous carbonwith embedded W particles covered with WC and W2C phases.

The WC/Pt(1 wt.%) sample (Fig. 3a) is composed of well-dis-persed sub-nanometer Pt clusters which mostly consist of a fewto several tens of Pt atoms. Some of the individual atoms areclearly resolved in this HRSTEM image. FFT taken from the HAADFimage shown in Fig. 3b shows clearly that this Pt particle is imagedclose to 110 zone axis. Even subnanometer sized cluster marked inFig. 3b shows similar, although slightly distorted, fcc pattern, typ-ical for Pt cluster. The EELS spectrum in the inset of Fig. 3b, showsthe presence of W and Pt by their M4,5 energy loss edges, andthereby indicates that WC particles are functioning as nucleationsites for Pt nanoparticles.

Nitrogen adsorption isotherms of WC, as the amount of N2 ad-sorbed as a function of relative pressure at �196 �C, are shown inFig. 4a. The isotherms are identified as type IV which is character-istics of mesoporous materials. The specific surface area calculatedby the BET equation, SBET, was 85 m2 g�1. The pore size distribution(PSD) is shown in Fig. 4b shows that the WC is mesoporous withthe most of the pore radius lower than 2 nm.

3.2. Cyclic voltammetry results

Cyclic voltammetry (CV) data for the WC and WC/Pt catalystswith different Pt loadings are presented in Fig. 5. CV was per-formed to determine the electrochemically active surface areaand to elucidate the adsorption properties of the catalysts. TheCV curve of the WC (Fig. 5a, curve 3) shows a very clear adsorp-tion/desorption peak at 0.1 V vs. RHE. Its reversible nature suggeststhat WC nanoparticles are active for H(upd) adsorption/desorptionwithout the presence of Pt species. The presence of WC nanoparti-cles in the WC/Pt catalysts results in a few changes in the CV of thePt electrode. In the presence of very low Pt content, the H(upd)adsorption/desorption reaction (WC/Pt(1%) catalyst) is enhancedand takes place in a narrow potential range (curve 2 in Fig. 5a).The WC/Pt catalyst with higher Pt content has a CV with twoH(upd) adsorption/desorption peaks (curve 1) whose shape is sim-ilar to the CV of Pt but its first H(upd) peak is suppressed until thesecond desorption peak is shifted to a more positive potential(�0.3 V) and that indicates the H-spill-over.

In order to check the possibility that these WC particles to exhi-bit a H-spill-over effect when they are in contact with Pt, the elec-trochemically active surface area of the catalysts, Seasa, wascalculated from the charge associated with the anodic desorptionpeak of (upd) hydrogen with a reference to 210 lC cm�2 for poly-crystalline Pt, and from Cu stripping measurements (Fig. 5b). Thecorresponding results are presented in Table 1.

The electrochemically active surface area of the catalysts calcu-lated from the H(upd) region is about 30% higher than the value ob-tained from Cu stripping measurements. These results prove thepresence of H-spill-over effect of WC particles, which could be re-verse or direct.

Fig. 2. (a) A HAADF-STEM micrograph of WC with the typical bimodal particle size distribution of WC nanocrystals; (b) TEM image of the core–shell structure of a WCparticles. Black dot in the shell indicates e-beam size (1.2 nm) and its position during EDS analysis.

M.D. Obradovic et al. / Journal of Electroanalytical Chemistry 671 (2012) 24–32 27

Hydrogen that is adsorbed and dissociated on the Pt surface,(Eq. (1a)) could spill over onto the tungsten oxide species, WOy

(Eq. (1b)), which are present at WC surface. Then, desorption ofspiltover H atoms could take place from HxWOy by migration backto Pt species (reverse spill-over).

Hþ þ eþ Pt$ Pt—Hads ð1aÞ

xPt—Hads þWOy $ HxWOy þ xPt ð1bÞ

However, having in mind the CV response of pure WC, H(upd)could be directly adsorbed on WOy species and then spill over ontothe Pt surface.

Due to the presence of a H-spill-over effect of WC particles, thedetermination of electrochemically active surface area of Pt nano-particles from the stripping of underpotentially deposited Cu(UPD) copper is relevant.

3.3. Kinetics of the hydrogen oxidation at WC/Pt catalysts

3.3.1. Theoretical considerationsThe hydrogen oxidation reaction (HOR) in acid media can be

written as:

H2 ! 2Hþ þ 2e ð2Þ

The corresponding elementary steps for the Tafel–Heyrovsky–Volmer mechanism on Pt catalysts are:

H2 þ 2Pt ()kT

k�T

2H� Pt Tafel reaction ð3Þ

H2 þ Pt ()kH

k�H

H� PtþHþ þ e Heyrovsky reaction ð4Þ

H� Pt ()kV

k�V

PtþHþ þ e Volmer reaction ð5Þ

In the Tafel–Volmer pathway the dissociative adsorption of ahydrogen molecule is followed by two separate one-electron oxi-dations of adsorbed H atoms. However, in the Heyrovsky–Volmerpathway, a one-electrooxidation occurs simultaneously withchemisorption, followed by another one-electron oxidation of theadsorbed H atom.

In a steady-state, the kinetics of the HOR for the simultaneousoccurrence of the Tafel–Volmer and Heyrovsky–Volmer routescan be described in terms of the reaction rate of the elementarysteps by the following equations:

Fig. 3. (a) A high resolution HAADF-STEM image of WC/Pt with sub-nanometer Pt clusters attached to the WC particles. (b) EELS spectrum of Pt cluster (inset of Fig. 3b) showsthe presence of W and Pt by their M4,5 energy loss edges.

28 M.D. Obradovic et al. / Journal of Electroanalytical Chemistry 671 (2012) 24–32

v � F ¼ I ¼ FðvH þ vVÞ ¼ 2FðvT þ vHÞ ð6Þ

where v is the overall reaction rate and vs with different subscriptsare the reaction rates of the elementary steps (3–5).

It has been recently shown [37,38] that the overall reaction rateof the HOR when a Langmuir-type adsorption is considered for theadsorbed hydrogen, in the presence of mass-transfer limitationsand assuming that the electron transfer coefficient is ½, can bewritten as:

I ¼ I0T ð1�HHÞ2 1� IIL

� �� ð1�H0

HÞ2 HH

H0H

� �2� �

þI0H ð1�HHÞ 1� IIL

� �exp FE

2RT

� �� ð1�H0

HÞHH

H0H

� �exp � FE

2RT

� �h i ð7Þ

where HH and H0H are the coverages of the reaction H intermediate

at the potential E, and at the reversible potential, E0, respectively; IL

is the H2 diffusion limiting current at high potential where cH2 ap-proaches zero; I0T and I0H are the exchange current densities of Tafeland Heyrovsky steps, respectively.

Dependence of HH on g (or E vs. RHE) can be obtained by solv-ing the equation dH=dt ¼ 2vT þ vH � vV ¼ 0, following the approx-imation that the concentration of the reaction intermediate doesnot change in time. Wang et al. [38] found that simple expressioncan be used under the assumption that the Volmer reaction rate issufficiently larger than the Tafel and Heyrovsky reaction rates. Forsufficiently small H0

H and E P 0 follows:

HH

H0H

¼ exp � FEcRT

� ð8Þ

where c is also a function of H0H. Previously, it was shown by simu-

lations [39] that H0H ranges from 0.09 to 10�7. Incorporating Eq. (8)

into Eq. (7) and letting (1 �HH) and (1 �H0H) be unity, Eq. (7) can

be presented in more simplified form:

I ¼ I0T 1� IIL

� � exp � 2F E

cRT

� � �

þ I0H 1� IIL

� exp

F E2RT

� � exp � F E

cRT

� exp � F E

2RT

� � �ð9Þ

This can be rearranged as:

I ¼I0Tð1� exp �2F E

c RT

� �h iþ I0H exp F E

2RT

� �� exp �F E

c RT

� �� exp �F E

2RT

� �h i1þ I0T

ILþ I0H

ILexp F E

2RT

� �

¼ Ik

1þ IfIL

ð10Þ

The numerator in the above equation corresponds to the kineticcurrent, Ik, when the limiting current, IL ?1. The sum of the twopositive terms in the denominator represents the kinetic current ofthe forwards reactions, If.

Fig. 4. (a) Nitrogen adsorption isotherms for WC support. Solid symbols-adsorp-tion, open symbols-desorption. (b) BJH pore size distribution.

Fig. 5. (a) Cyclic voltammograms for WC (curve 3) and WC/Pt electrodes withdifferent amount of Pt, recorded at a sweep rate of 100 mV s�1 in N2 saturated0.5 mol dm�3 HClO4 solution a 25�C. (b) Cu stripping voltammetry for WC/Pt(1 wt.%)electrode in 0.1 mol dm�3 H2SO4 solution at sweep rate of 20 mV s�1.

M.D. Obradovic et al. / Journal of Electroanalytical Chemistry 671 (2012) 24–32 29

Here, I0T, I0H and c are the three essential kinetic parameters.Wang et al. [38] showed that the HOR takes place dominantly

through the Tafel–Volmer pathway on Pt for E 6 50 mV, wherethe current rises rapidly in an inverse exponential function. In thiscase, Eq. (9) can be presented in a more simplified form:

I ¼ I0T 1� IIL

� � exp � 2FE

cRT

� � �ð11Þ

and rearranged as:

I ¼ I0T � IL

I0T þ IL

� 1� exp � 2FE

cRT

� � �ð12Þ

or:

exp � 2FEcRT

� ¼ 1� I

I0T� I

ILð13Þ

If the I0T� IL condition is fulfilled (this condition can be easilyfulfilled even in RDE measurements), then the second term onthe right side of Eq. (14) can be neglected and Eq. (14) becomes:

E ¼ � cRT2F

lnIL � I

IL

� ð14Þ

Eq. (15) is the similar to the Nernstian equation (E ¼� RT

2F lnðIL�IILÞ) for a pure diffusion controlled reaction and which is

frequently used to prove that the HOR takes place as the reversiblereaction.

3.3.2. Determination of the kinetic parametersFig. 6a presents the hydrogen oxidation polarization curves for

several rotation speeds, obtained at a scan rate of 2 mV s�1 for theWC/Pt(1 wt.%) catalyst. The current increases rapidly with poten-tial on each curve and reaches a limiting value at ca. 75 mV(RHE). Fig. 6b displays the Levich–Koutecky plot for the HOR,where the near-zero intercept and the linear behavior indicate thatthe current at the positive potential limit is essentially diffusionlimited for this catalyst.

The HOR polarization curves presented in Fig. 6a were first ana-lyzed in low potential region using Eq. (15) in order to calculate thekinetic parameter, c. The solid lines in Fig. 7 show that agreementwith the data obtained for the value of c = 1.42, which does not de-pend on the rotation rate.

Assuming also that HOR takes place dominantly through the Ta-fel–Volmer pathway in the low potential region, and if, for instanceE 6 10 mV the overall reaction rate can be presented by Eq. (11) inwhich the exponential may be expanded and higher terms ne-glected, so that one obtains:

Table 1Calculated kinetic parameters for the HOR at WC/Pt(1%) and C/Pt(1%) catalysts.

Electrode (0.4 lg Pt) Electrochem.active surface area (cm2) I0T I0H c

HUPD desorpt. CuUPD desorpt. mA mA cm�2 A mg�1Pt mA mA cm�2 A mg�1

Pt

C/WC–Pt(1%) 0.53 0.40 16 40 40 1 2.5 2.5 1.42Vulcan/Pt(1%) 0.23 0.23 4 17 10 0.3 1.3 0.8 2.2

The exchange current densities are calculated using the value of electrochemically active Pt surface area estimated by Cu stripping voltammetry.

Fig. 6. (a) Polarization curves obtained with RDE at 2 mV s�1 for H2 oxidation in0.5 mol dm�3 HClO4 solution at C/WC–Pt(1%) catalyst, for several rotating speeds.(b) Corresponding Levich–Koutecky plot for the HOR at 0.3 V vs. RHE.

Fig. 7. Measured (symbols) and fitted (line) polarization curve for the HOR on WC/Pt(1%) at different rotation rates of RDE in hydrogen saturated 0.5 mol dm�3 HClO4

solution at 25 �C. Fitted curve was obtained using Eq. (15) in order to determine thekinetic parameter c.

Fig. 8. Experimental data (symbols) and fitted (lines) linear polarization curves forthe HOR on WC/Pt(1%) RDE in hydrogen saturated 0.5 mol dm�3 HClO4 solution at25 �C. Fitted curves were obtained using Eq. (15) and the kinetic parameter c = 1.42.

30 M.D. Obradovic et al. / Journal of Electroanalytical Chemistry 671 (2012) 24–32

IE¼ 2F

cRTIL � I0T

I0T þ IL

� ð15Þ

Fig. 8 presents the corresponding polarization data recorded atdifferent rotation rates in the low potential region and the corre-sponding values of the exchange current for Tafel step, I0T were cal-culated from the slope of linear I–E (RHE) response (Eq. (15)),(mean value is presented in Table 1.

Now, it is possible to estimate the last third kinetic parameter,(I0H) by fitting the RDE polarizations curves in the whole potentialrange with Eq. (9). The solid line in Fig. 9 shows that agreementwith the polarization data (symbols) is obtained by the best fitswith one variable kinetic parameter. The values of the calculated

Fig. 9. Measured (symbols) and fitted (line) polarization curves for the HOR on WC/Pt(1%) RDE in hydrogen saturated 0.5 mol dm�3HClO4 solution at 25 �C. Fitted curvewas obtained using Eq. (9) and previously determined kinetic parameters c = 1.42and I0T = 16 mA.

Fig. 10. Kinetic current for the HOR on WC/Pt(1%) electrode at 25 �C calculatedusing the numerator in Eq. (10) with I0T = 16 mA and I0H = 1.0 mA and c = 1.42. Dashand dotted lines represent the contributions from TV and HV pathways,respectively.

Fig. 11. Polarization curves obtained with RDE at 2 mV s�1 for the HOR in0.5 mol dm�3 HClO4 solution at C/Pt(1%) catalyst, for several rotation speeds.

Fig. 12. Measured (symbols) and fitted (line) polarization curve for the HOR on C/Pt(1%) RDE at different rotating rates in hydrogen saturated 0.5 mol dm�3 HClO4

solution at 25 �C. Fitted curve was obtained using Eq. (15) in order to determine thekinetic parameter c.

M.D. Obradovic et al. / Journal of Electroanalytical Chemistry 671 (2012) 24–32 31

kinetic parameters of the HOR at WC/Pt(1 wt.%) electrode that fitthe polarization RDE curves are presented in Table 1.

It can be concluded that the Tafel–Volmer pathway is responsi-ble for the high HOR activity on WC–Pt(1 wt.%) catalyst atE 6 50 mV, where the current rises rapidly in an inverse exponen-tial fashion to a value close to I0T. Further increase is realizedmainly through Heyrovsky–Volmer pathway, which is importantat the more positive potentials, where ITV levels off.

It is interesting to note that in the literature, the E vs. ln ((IL � I)/IL) relationship is most often identified with the Tafel plot (assum-ing reversible kinetics of the HOR) That is wrong, because when

Tafel step controls the overall reaction rate, the current, I, dependsas an inverse exponential fashion on potential, E. This is in agree-ment with the theoretical prediction that if a preceding chemicaladsorption is followed with fast electron transfer reaction, as it isproposed in the kinetic analysis for HOR, then it is not correct touse the Butler–Volmer equation in performing the kinetic of thereaction.

In addition, by comparing the dependence of the kinetic cur-rent, Ik on the potential E, for the HOR, (Ik is calculated by usingthe numerator in Eq. (10) and the determined kinetic parameters),and anodic polarization curve for the HOR, derived from a half-cellunder actual PEMFC operating condition [40,41], one can see verygood conformity (Fig. 10).

Fig. 13. Polarization curves obtained with rotating disk electrode at 2 mV s�1 for H2

oxidation in 0.5 mol dm�3 HClO4 solution at C/Pt(1%) and WC/Pt catalysts, atrotating speed of 600 rpm.

32 M.D. Obradovic et al. / Journal of Electroanalytical Chemistry 671 (2012) 24–32

In order to find the influence of WC on the activity of Pt catalystfor the HOR, the polarization measurements were also conducted ona commercial C/Pt electrode, with the same Pt content (mass.1%).Fig. 11 presents the HOR polarization curves for several rotationspeeds, recorded at a scan rate of 2 mV s�1 for a C/Pt(1 wt.%) elec-trode. For this electrode, the linearity of E � ln((IL � I)/IL), irrespec-tive of rotation rate is also achieved, indicating that the Tafel–Volmer (TV) pathway for the HOR is operative in the lower potentialrange (Fig. 12).

The polarization curves for the HOR on WC/Pt(1 wt.%) and C/Pt(1%) RDE electrodes are presented together in Fig. 13, for com-parison. It is evident that WC/Pt(1 wt.%) catalyst exhibits highercatalytic activity. The kinetic parameters of the HOR at C/Pt(1 wt.%)electrode were determined in an identical way as in the case ofWC/Pt(1 wt.%) catalyst and presented in Table 1, as well. In orderto compare the activities of these two catalysts, specific and massspecific activities are also presented.

On the basis of the determined kinetic parameters for the HORit can be concluded that the WC/Pt(1 wt.%) electrode exhibits fourtimes higher mass specific activity and about 2.5 times higherspecific activity compared to C/Pt(1 wt.%) catalyst. The higheractivity of the WC/Pt catalyst could be a result of the presence ofthe H-spill-over effect which is probably also operative also inthe potential range where the HOR takes place. WC plays the rolein accelerating the dissociative adsorption of H2 at Pt, which is therate-determining step for the HOR.

4. Conclusions

The kinetics and mechanism of the HOR on a home-made WC/Pt(1%) electrode in a perchloric acid solution has been studied.Based on the Tafel–Heyrovsky–Volmer mechanism the corre-sponding kinetic equations were derived to describe the HOR cur-rent behavior on RDE over the entire potential range, Applyingthese equations in the analysis of the polarization curves measuredfor WC/Pt(1%) at RDE, the exchange currents of Tafel, (I0T), andHeyrovsky, (I0H), elemental steps and adsorption parameter, (c),

were determined. The fitting procedure shows that the HOR onWC/Pt and C/Pt catalysts most likely proceeds most likely via theTafel–Volmer pathway in the lower potential region, while theHeyrovsky–Volmer pathway is operative in the higher potential re-gion. The WC/Pt(1 wt.%) catalyst exhibits four times higher massspecific activity for the HOR compared to the conventional carbonsupported Pt catalyst. The enhanced catalytic activity of the WC/Ptcatalyst is probably cause by the presence of a H-spill-over effect ofWC nanoparticles.

Acknowledgments

This work is financially supported by the Ministry of Scienceand Technological Development, Republic of Serbia, under ContractNo. 172054. All TEM characterizations have been performed at Na-tional Center of Electron Microscopy, LBLN, University of California,Berkeley.

References

[1] H. Böhm, F.A. Pohl, Wiss, Ber. AEG-Telefunken 41 (1968) 46.[2] D. Baresel, W. Gellert, J. Heidemeyer, P. Scharner, Angew. Chem. Int. Ed. 10

(1971) 194.[3] R.B. Levy, M. Boudart, Science 181 (1973) 547.[4] G.A. Tsirlina, O.A. Petrii, Electrochim. Acta 32 (1987) 637.[5] B. Bozzini, G.P. De Gaudenzi, A. Fanigliulo, C. Mele, Mater. Corrosion 54 (2003)

295.[6] R. Fleischmann, H. Böhm, Electrochim. Acta 22 (1977) 1123.[7] J.W. Johnson, C.I. Wu, J. Electrochem. Soc. 118 (1971) 1909.[8] P.N. Ross, P. Stonehart, J. Catal. 39 (1975) 298.[9] R.D. Armstrong, J. Electrochem. Soc. 118 (1971) 568.

[10] P.N. Ross, P. Stonehart, J. Catal. 48 (1977) 42.[11] V.S. Palanker, R.A. Gaiyev, D. Sokolsky, Electrochim. Acta 22 (1977) 133.[12] H. Okamoto, G. Kawamura, A. Ishikawa, T. Kudo, J. Electrochem. Soc. 134 (7)

(1987) 1645.[13] H. Meng, P.K. Shen, J. Phys. Chem. B 109 (2005) 22705.[14] W. Liu, Y. Soneda, H. Hatori, Chem. Lett. 35 (10) (2006) 1148.[15] M.K. Neylon, S. Choi, H. Kwon, K.E. Cury, L.T. Thompson, Appl. Catal. A 183

(1999) 253.[16] E.J. Rees, K. Essaki, C.D.A. Brady, G.T. Burstein, J. Power Sources 188 (2009) 75.[17] M. Nagai, M. Yoshida, H. Tominaga, Electochim. Acta 52 (2007) 5430–5436.[18] S. Ishar, M. Yoshida, M. Nagai, Electrochim. Acta 54 (2009) 1255.[19] N. Liu, K. Kourtakis, J.C. Figueroa, J.G. Chen, J. Catal. 215 (2003) 254.[20] G. Lu, J.S. Cooper, P.J. McGinn, J. Power Sources 161 (2006) 106.[21] Z.J. Mellinger, E.C. Weigert, A.L. Stottlemyer, J.G. Chen, Electrochem. Solid State

Lett. 11 (2008) B63.[22] L.G.R.A. Santos, K.S. Freitas, E.A. Ticianelli, J. Solid State Electrochem. 11 (2007)

1541.[23] J. Kim, Ji-H. Jang, Y.-H. Lee, Y.-Uk. Kwon, J. Power Sources 193 (2009) 441.[24] Y. Hara, N. Minami, H. Itagi, Appl. Catal. A 323 (2007) 86.[25] Y. Hara, N. Minami, H. Matsumoto, H. Itagi, Appl. Catal. A 332 (2007) 289.[26] E.C. Weigert, D.V. Esposito, J.G. Chen, J. Power Sources 193 (2009) 501.[27] R. Ganesan, J.S. Lee, Angew. Chem. Int. Ed. 44 (2005) 6557.[28] R. Ganesan, D.J. Ham, J.S. Lee, Electrochem. Commun. 9 (2007) 2576.[29] H. Tamon, H. Ishizaka, T. Yamamoto, T. Suzuki, Carbon 37 (1999) 2049.[30] D.J. Ham, Y.K. Kim, S.H. Han, J.S. Lee, Catal. Today 132 (2008) 117.[31] E.P. Barret, L.G. Joyner, P.P. Halenda, J. Am. Chem. Soc. 73 (1951) 373.[32] K. Kaneko, C. Ishii, H. Kanoh, Y. Hanzawa, N. Setoyama, T. Suzuki, Adv. Colloid

Interf. Sci. (76–77) (1998) 295.[33] C.L. Green, A. Kucernak, J. Phys. Chem. B 106 (2002) 1036.[34] M.D. Obradovic, B.M. Babic, A. Kowal, V.V. Panic, S.Lj. Gojkovic, J. Serb. Chem.

Soc. 73 (2008) 1197.[35] V. Glibin, L. Svirko, I. Bashtan-Kandybovich, D. Karamanev, Surf. Sci. 604 (2010)

500.[36] K. Page, J. Li, R. Savinelli, H.N. Szumila, J. Zhang, J.K. Stalick, Th. Proffen, S.L.

Scott, R. Seshadri, Solid State Sci. 10 (2008) 1499.[37] B. Babic, J. Gulicovski, Lj. Gajic-Krstajic, N. Elezovic, V.R. Radmilovic, N.V.

Krstajic, Lj.M. Vracar, J. Power Sources 193 (2009) 99.[38] J.X. Wang, T.E. Springer, R.R. Adzic, J. Electrochem. Soc. 153 (2006) A1732.[39] M.R. Gennero de Chialvo, A.C. Chialvo, J. Electrochem. Soc. 147 (2000) 1619.[40] G.A. Camara, E.A. Ticianelli, S. Mukerjee, S.J. Lee, J. McBreen, J. Electrochem.

Soc. 149 (2002) A748–A753.[41] R.C. Jiang, H.R. Kunz, J.M. Fenton, J. Electrochem. Soc. 152 (2005) A1329–

A1340.