03.01.2014 Views

KV Rybalka

KV Rybalka

KV Rybalka

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

Journal of New Materials for Electrochemical Systems 10, 81-89 (2007)<br />

© J. New Mat. Electrochem. Systems<br />

Binary Pt– and Pd–based Electrocatalysts for Oxidation<br />

of Hydrogen with Co Admixtures<br />

* K. V. <strong>Rybalka</strong> 1 , M. R. Tarasevich 1 , B. M. Grafov 1 , V. A. Bogdanovskaya 1 , L. A. Beketaeva 1 ,<br />

E. N. Loubnin 1 and Yu. A. Kolbanovskii 2<br />

1 A.N. Frumkin Institute of Physical Chemistry and Electrochemistry, RAS, Moscow 119071, Leninsky pr.31, Russia<br />

2 A.I. Topchiev Institute of Petrochemical Synthesis, RAS, Moscow 119991 Leninsky pr.29, Russia<br />

Received: April 15, 2005, Accepted: December 20, 2006<br />

Abstract: The results are presented of studying the electrochemical and electrocatalytic properties of binary hydrogen electrooxidation<br />

catalysts including Pt or Pd and some other metal (Au, Mo, Ru) and tolerant towards CO admixtures. The methods of catalyst synthesis at<br />

the dispersed carbon carrier (XC72) are suggested, the binary system structure is studied. According to the results of the studies by the<br />

electrochemical impedance method, the quantitative estimation is performed for the developed catalysts of their tolerance towards CO<br />

admixtures. PdAu catalyst features the highest tolerance, which can result from the decoration of palladium surface by gold. The decoration<br />

of Pd nanoparticles by gold hinders CO adsorption while seemingly preserving enough accessible sites for hydrogen adsorption. In<br />

PdRu binary system, the tolerance is probably provided by the formation of an alloy, at which the oxidative desorption of CO is considerably<br />

facilitated in comparison with pure Pd.<br />

Keywords: : electrooxydation,catalyst,tolerance,impedance spectra<br />

1. INTRODUCTION<br />

Gas diffusion hydrogen electrodes with dispersed platinum content<br />

of 0.1–0.2 mg/cm 2 demonstrate excellent characteristics when<br />

pure hydrogen is used, but are poisoned already with traces of CO.<br />

The problem of employing gaseous hydrogen with CO admixtures<br />

in fuel cells (FC) was first studied in the 80-s [1,2] when it was<br />

shown that CO admixture of 10 ppm caused the voltage decrease<br />

of 0.2 – 0.3 V at the load of 0.8 A/cm 2 . At the load of 0.2 A/cm 2 at<br />

80 ºC the presence of CO in hydrogen, its amount being 250 ppm,<br />

leads to a drastic voltage drop of 0.4 V [3]. CO concentration in<br />

hydrogen obtained in the reforming process after its additional<br />

purification by means of the shift method usually constitutes from<br />

20 to 100 ppm. The inhibiting effect of CO on the hydrogen electrooxidation<br />

process results from the fact that the bonding strength<br />

of Pt–CO is higher than that of Pt–H [4], which complicates the<br />

dissociative hydrogen adsorption on platinum and its further ionization.<br />

At present, two possibilities of eliminating or decreasing the<br />

poisoning effect of CO presence in hydrogen are considered. The<br />

*To whom correspondence should be addressed:<br />

first of them is the increase of FC working temperature. It is<br />

shown experimentally in work [5] that at the temperature of hydrogen–oxygen<br />

FC being close to 200 ºC the specific parameters do<br />

not change even at CO content of 20000-30000 ppm. The second<br />

way is based on the development of new anodic catalysts tolerant<br />

towards CO admixtures in hydrogen, while retaining the construction<br />

and functioning conditions of FC.<br />

Platinum modification by adatoms [6] and, later, the decoration<br />

of its surface by metal submonolayers [7,8] seemed to be a prospective<br />

way of purposeful optimization of platinum properties in<br />

respect to CO oxidation. However, such systems are rather unstable<br />

in the conditions of FC functioning. Therefore, despite the<br />

importance of these studies for understanding the mechanism of<br />

electrocatalysis, they can hardly possess any practical value.<br />

At present it seems most prospective to use as electrocatalysts<br />

the binary systems based on platinum group metals which results<br />

in lower surface coverage by adsorbed CO. Such systems can be<br />

obtained by employing the methods of fusing, codeposition, thermochemical<br />

and mechanochemical synthesis, cathodic deposition<br />

etc. The most trustworthy and reproducible effects of increasing<br />

the tolerance towards CO admixtures in gaseous hydrogen were<br />

obtained by introducing Ru admixtures into Pt. As shown in [8, 9]<br />

81


82 K. V. <strong>Rybalka</strong> et al. / J. New Mat. Electrochem. Systems<br />

by means of STM and IR spectroscopy, the minimum oxidation<br />

overpotential of linearly adsorbed CO ads is achieved for the system<br />

of Pt 0.5 Ru 0.5 . According to [8], the promoting effect of Ru in the<br />

system of Pt–Ru also manifests itself in the absence of the alloy<br />

formation. The optimum ratio of Pt 0.5 Ru 0.5 was confirmed in a<br />

number of works [10–13], in particular, at the example of commercial<br />

E–TEK catalyst. The system of PtMo also proved to be of<br />

interest [7,12,14–16]. In papers [17,18] it was shown at well–<br />

characterized smooth PtMo surfaces that the presence of 20–25 at.<br />

% Mo accelerates the oxidative removal of CO and facilitates the<br />

adsorption of hydrogen. These data were confirmed by using<br />

PtMo system with the ratio of 4:1 on carbon black XC72 [14].<br />

Later in [15], the authors specified the optimum composition by<br />

increasing Mo content in PtMo system up to 3:1.<br />

Pd–based electrocatalytic systems are of real interest alongside<br />

with Pt catalysts. In [19], a method of PdAu black synthesis on<br />

carbon black XC72 is offered. In the case of these catalysts, hydrogen<br />

adsorption is observed at Pd atoms in the range of potentials<br />

of 0.05–0.3 V, same as the well–expressed limiting diffusion<br />

current of hydrogen oxidation at the rotating disc electrode [20].<br />

In respect to CO oxidation, the properties of PdAu are close to<br />

those of PtRu, but CO oxidation rate is somewhat lower. The advantages<br />

of PdAu system as compared to PtRu are better manifested<br />

at the conditions of 60 ºC and CO concentration in the<br />

range of 250–1000 ppm. It was demonstrated [21] that at the content<br />

of Au > 70% in PdAu alloys hydrogen adsorption is not<br />

observed. A conclusion was made as to its occurring only at Pd<br />

atoms. Herewith, CO adsorption on gold is also negligible [22].<br />

This system can probably turn out to be tolerant towards CO and<br />

at the same time to be active towards hydrogen electrooxidation.<br />

In the present work, the possibility is considered of developing<br />

Pt– and Pd–based binary catalysts of hydrogen oxidation that are<br />

tolerant towards CO admixtures. The electrocatalytic properties of<br />

such systems were compared both as to pure hydrogen oxidation<br />

and the oxidation of hydrogen with CO admixtures, their content<br />

being from 50 to 500 ppm. Herewith, Pt and PtMo supported on<br />

carbon black XC72 were used in this work for comparison with<br />

the properties of binary PdM system (М=Au, Mo,Ru).<br />

2. EXPERIMENTAL<br />

2.1. Method of catalyst preparation and performing<br />

electrochemical and structural studies<br />

All Pt– and Pd–based binary catalysts studied in this work were<br />

synthesized on carbon black XC72. The commercial E–TEK catalyst<br />

(XC72 + 20% Pt) that we used for comparison was heated<br />

prior to experiments in an inert atmosphere at 300 ºC [23].<br />

The catalyst synthesis was performed by simultaneous and successive<br />

deposition of the components onto carbon black. Various<br />

methods of reducing the initial compounds were employed, catalyst<br />

synthesis temperature and the duration of its reduction process<br />

were optimized. Several methods of catalyst preparation were<br />

developed.<br />

To obtain XC72 + 20 % PtMo (75:25) catalyst, successive application<br />

of Pt and Mo onto carbon black from the aqueous solutions<br />

of H 2 PtCl 6 and (NH 4 ) 6 Mo 7 O 24 was used. The reducing agent<br />

was formic acid or sodium borate. The mixture was evaporated,<br />

washed, and subjected to thermal treatment in an inert atmosphere<br />

and then in hydrogen.<br />

The most effective XC72 + PdAu catalysts for hydrogen oxidation<br />

are prepared as follows. In the process of catalyst synthesis,<br />

the diluted solutions of components (0.033 M) were used [24].<br />

The reducing agent was ethanol. The reduction was carried out in<br />

a synthesizer with vigorous stirring for four hours in the atmosphere<br />

of an inert gas at 100 ºC. The obtained catalyst was baked<br />

for three hours in the atmosphere of hydrogen at 530 ºC and then<br />

cooled in the atmosphere of an inert gas. After that, the catalyst<br />

was many times washed by water and dried in a vacuum compartment.<br />

The synthesis of binary catalysts of PdMo and PdRu was carried<br />

out on carbon black XC72 using metal salts (PdCl 2 ,<br />

RuOHCl 3 , (NH 4 ) 6 Mo 7 O 24 ). The total amount of promoting metals<br />

constituted 15–30 wt.%, the metal ratio was changed from 1:1 to<br />

4:1 atomic parts. The optimum method of PdMo catalyst preparation<br />

is the successive deposition of metals. PdRu catalyst features<br />

better characteristics at its preparation by means of the codeposition<br />

method.<br />

The catalyst activity was estimated from the run of the polarization<br />

curves of hydrogen electrooxidation reaction obtained at the<br />

rotating disc electrode and from the electrochemical impedance<br />

spectra. All the electrochemical measurements were performed at<br />

the disc pyrographite electrode sealed into Teflon with the diameter<br />

of 5 mm. 100 µg/cm 2 of the catalyst with 0.7 wt.% of Nafion<br />

were applied to the disc electrode according to the technique developed<br />

in [25]. 500 µl of ethanol and the necessary amount of 5%<br />

Nafion solution were added to the catalyst sample (2mg). The<br />

mixture was subjected to ultrasonic homogenization for 30 min. A<br />

sample of the obtained mixture was applied to the previously purified<br />

and polished electrode and dried in air for 12 hours.<br />

The potential sweep rate while registering the polarization<br />

curves at the rotating disc electrode (from the steady–state potential<br />

to more positive values) constituted 1 mV/s.<br />

The curves of hydrogen oxidation close to those obtained in<br />

[26] were observed at the same electrode with E–TEK catalyst of<br />

20wt.% Pt in 0.5 M H 2 SO 4 . Herewith, the value of 0.62 n Fc H2 D<br />

H2 2/3 ν -1/6 product in the equation of Levich [27] i lim = 0.62 n Fc H2<br />

D H2 2/3 ν -1/6 ω 0.5 constituted 7.65 10 –2 mA cm -2 rpm -0.5 , which<br />

somewhat differs from the value of 7.80 10 -2 mA cm -2 rpm -0.5<br />

specified by the authors of the above work. This evidences the<br />

satisfactory hydrodynamic conditions that are realized at the disc<br />

electrode with a thin catalyst layer.<br />

The measurements of the electrochemical impedance components<br />

were performed with the help of numeric impedance meter<br />

X-206 in the frequency range of 20Hz – 100 kHz. All the measurements<br />

were performed using a two–electrode scheme, i.e. the<br />

impedance included the impedance of both the studied and the<br />

auxiliary electrode, which was a platinated platinum disc with the<br />

diameter of 20 mm. The auxiliary electrode impedance is negligible<br />

as its surface considerably exceeded the studied electrode surface.<br />

All the impedance measurements were performed at the electrode<br />

steady–state potential.<br />

The tolerance of the catalysts towards CO was estimated from<br />

the variation of the impedance spectrum as dependent in the time<br />

of electrode conditioning in the electrolyte saturated by hydrogen<br />

containing CO admixtures. The ratio of the exchange current of<br />

the hydrogen reaction determined after a two–hour exposure of the<br />

electrode in 0.5 M H 2 SO 4 solution saturated by the gas mixture to


Binary Pt– and Pd–based Electrocatalysts for Oxidation of Hydrogen with Co Admixtures / J. New Mat. Electrochem. Systems<br />

83<br />

Table 1. Structural characteristics of the catalysts based on binary metal systems of Pd + M (M- Au, Ru, Mo).<br />

Catalyst Phases Particle size (nm) Lattice parameter (nm)<br />

XC72 +20% PdAu (70:30 at.%) Alloy, separate phases of gold and palladium particles 6 (±30%) Alloy - 0.3956<br />

XC72 +20% PdMo (70:30 at.%) Palladium 18-20 (±20%) 0.3897<br />

XC72 +20% PdRu (50:50 at.%) Alloy, separate phases of ruthenium and palladium particles 3-5 (±40%) Alloy - 0.3871<br />

Table 2. Electrochemical parameters of Pd–based binary catalysts.<br />

Catalyst: XC72+ S, m 2 /g met E st , V<br />

i, mА/см 2<br />

at E=0.05V<br />

I lim , mA/cm 2<br />

m=2000 rpm<br />

30 wt.% PdAu<br />

(50:50 at. %)<br />

10.7 0.001 1.79 2.33<br />

30 wt.% PdAu<br />

(70:30 at. %)<br />

- 0.000 0.94 1.45<br />

30 wt.% PdAu<br />

(80:20 at. %)<br />

- 0.000 0.70 1.30<br />

20 wt.%PdRu<br />

(50:50 at. %)<br />

97 0.010 1.88 2.53<br />

20 wt.%PdRu<br />

(70:30 at. %)<br />

40.5 0.000 2.17 3.32<br />

20 wt.%PdRu<br />

(80:20 at. %)<br />

146.7 0.003 0.94 2.42<br />

20 wt.%PdMo<br />

(50:50 at. %)<br />

37 0.005 1.15 2.29<br />

20 wt.%PdMo<br />

(70:30 at. %)<br />

110 0.004 1.78 1.96<br />

20 wt.%PdMo<br />

(80:20 at. %)<br />

111 0.001 2.10 3.28<br />

the exchange current value obtained in the absence of CO was chosen<br />

to be the quantitative tolerance parameter. Alongside with that,<br />

the electrocatalytic activity of different catalysts was compared in<br />

the reaction of CO oxidation from the solution bulk. The higher the<br />

rate of this reaction, the higher tolerance towards CO can be expected<br />

from the binary catalytic system.<br />

The specific surface area values for the metal component of the<br />

binary catalyst were determined from the amount of electricity<br />

consumed in hydrogen and CO desorption [28]. The cyclic potentiodynamic<br />

curves were registered at the potential sweep rate of 40<br />

mV/s, in the range of potentials of 0.03÷0.07 V, in order to reduce<br />

hydrogen evolution or its dissolution in palladium. Firstly, the solution<br />

was deaerated for 40 min by helium, after which a potentiodynamic<br />

curve was registered that allowed to make conclusions about<br />

the character of hydrogen adsorption at the studied catalyst. Then<br />

the solution was saturated by gaseous CO (30 min). After that, CO<br />

was removed from the solution by bubbling helium (40 min), and<br />

more potentiodynamic curves were registered (2 cycles). When<br />

calculating the surface area value, it was assumed that in the case of<br />

the monolayer coverage of the catalyst metal component by hydrogen<br />

its desorption consumes 210 mC/cm 2 , while in the case of CO<br />

the value is 420 mC/cm 2 .<br />

The catalyst compositions were analyzed by energy dispersive<br />

X-ray spectroscopy (EDX) and compared with the nominal ones.<br />

According to the EDX measurements atomic compositions were<br />

very close to the nominal composition of catalysts pointed in tables<br />

1-3.<br />

The physical and chemical characteristics of the catalysts were<br />

studied using the methods of X–ray photoelectron spectroscopy<br />

(XPS) at electron spectrometer VG ESCA 3МКII and X–ray diffraction<br />

analysis (XRD) at diffractometer JDX – 10 PA (JEOL,<br />

Table 3. Kinetic parameters of hydrogen oxidation catalysts in the<br />

mixture of H 2 + CO. 0.5 M H 2 SO 4 , the temperature is 60 ºC.<br />

Catalysts<br />

XC72 + 20wt.%<br />

PdMo(50:50 at.%)<br />

XC72 + 20wt.%<br />

PdMo(70:30at.%)<br />

XC72 + 20wt.%<br />

PdMo(80:20at.%)<br />

XC 72 + 30wt.%<br />

Pd<br />

XC 72 + 20wt.%<br />

PdRu(50:50at.%)<br />

XC 72 + 20wt.%<br />

PdRu(70:30at.%)<br />

XC 72 + 20wt.%<br />

PdRu(80:20at.%)<br />

XC 72 + 30wt.%<br />

PdAu(50:50at.%)<br />

XC72 + 20 wt.%<br />

Pt<br />

i 0 (0),<br />

A/g met<br />

i 0 (0),<br />

mA/cm 2 met<br />

i 0 (120)/i 0 (0)<br />

in H 2 +500 ppm CO<br />

1570 4.24 0.55<br />

700 0.64<br />

2800 2.52<br />

846 1.29<br />

476 0.49 0.50<br />

54.5 0.14<br />

1290 0.88<br />

641 6.0 0.60<br />

4400 8.55 0.24 (for 226 ppm CO)<br />

where i 0 (120) is the exchange current of the hydrogen reaction after the electrode was<br />

conditioned for 120 min in the solution saturated by H 2 +CO gas mixture; i 0 (0) is the<br />

exchange current in the absence of CO.<br />

Japan) using the filtered CuKα - radiation. The samples of dispersed<br />

catalysts for XPS analysis were stored without protecting<br />

the sample surface from the contact with the air. The studied catalysts<br />

were pressed onto a copper gauze with the cell size of 1 mm<br />

fixed on a gold support. The analyzed sample size constituted 5×10<br />

mm. The measurement of the spectra was performed in a chamber<br />

evacuated up to the vacuum of 3×10 -8 Pa. MgK α source was used<br />

for electron excitation. The treatment of the spectral lines was performed<br />

using a computer and service software package VG X<br />

DATA 1000. The nature of the chemical bonds was characterized<br />

basing on the variation of the core level bonding energies, i.e., the<br />

chemical shifts. All the bonding energies of the spectral lines of the<br />

core levels were calibrated vs. Au4f 7/2 level with the bonding energy<br />

of 83.8 eV. The accuracy of determining the chemical shifts in<br />

the spectra corresponded to the values of the bonding energy shifts<br />

from the average values measured in six spectra ±0.2 eV.<br />

The relative atomic abundances of the elements at catalyst surfaces<br />

were estimated by XPS. The atomic concentrations were<br />

calculated by integration of the signals recorded included the core<br />

level spectra and the atomic sensitivity factors. Similar procedures<br />

were used to calculate the surface compositions of the metallic<br />

nano-size particles in inorganic support [29-33]. In present investigation<br />

the errors of XPS method resulting from the effect of change<br />

of particle sizes and surface roughness were taken into account.<br />

The surface atomic ratios of catalysts obtained by XPS were calculated<br />

with the top error 20%. Thus, these XPS experiments could<br />

give indications about the surface atomic ratio within this error.


84 K. V. <strong>Rybalka</strong> et al. / J. New Mat. Electrochem. Systems<br />

Figure 2. . Cyclic potentiodynamic curves on PdRu catalysts, 60<br />

ºC, the potential sweep rate is 40 mV/s.<br />

The working electrode potential was measured vs. the saturated<br />

calomel reference electrode connected with the cell through an<br />

electrolytic bridge. All the values of the potentials given in the<br />

paper were re-calculated vs. the potential of the reversible hydrogen<br />

electrode in the same solution.<br />

3. RESULTS AND DISCUSSION<br />

Figure 1. Diffractogram fragments for the catalysts of (a) XC72 +<br />

20 wt.% PdAu (70:30 at.%) and (b) XC72 + 20 wt.% PdRu (50:50<br />

at.%).<br />

The diffraction spectra were registered at the room temperature<br />

using the mode of the continuous recording of spectra and the<br />

mode of scanning with the step of 0.1 0 with pulse accumulation per<br />

point for 20÷100 s. The appearance of alloys in the binary systems<br />

was observed from the angle shifts of the position of the diffraction<br />

reflex maximums in the range of high diffraction angles 2θ 0 of 64 0<br />

– 70 0 corresponding to reflexes from (220) planes of the cubic palladium<br />

phase. In the case of a very low intensity of this reflex, the<br />

structural data were obtained after analyzing the form of the maximums<br />

of lines (111) and (200) of the cubic palladium phase in the<br />

diffraction angle range of 2θ 0 of 39 0 – 48 0 . In this case, the line<br />

profiles for the expansion of complex lines were approximated by a<br />

combination of Voigt and Lorentz functions that provide the best fit<br />

for the line profiles of the dispersed palladium particles [34].<br />

All the electrochemical measurements were carried out in 0.5 M<br />

solution of twice distilled H 2 SO 4 in a thermostated cell with separate<br />

cathodic and anodic compartments. The compounds used for<br />

the synthesis of the catalysts produced by company “AURAT”<br />

were of chemical grade. The temperature was maintained by thermostat<br />

U-3. The polarization curves were registered with the help<br />

of numeric potentiostat IPC-20.<br />

3.1. Catalyst structure according to the data of<br />

XPS, XRD, and CO adsorption<br />

PdAu catalyst synthesized by the above method is characterized<br />

according to XRD method by a rather homogeneous particle size<br />

distribution, of which the average value constitutes 6 nm (table 1,<br />

fig. 1a). Herewith, the system of [XC72 + 20wt.% PdAu (70:30<br />

at.%)] is in fact heterophaseous and consists of the particles of<br />

metallic gold, palladium, and the alloy. The enrichment of the surface<br />

layer by gold is found from XPS spectra. At the nominal<br />

Au/Pd ratio of 0.43 the experimental value constituted 0.62.<br />

The system of PdRu synthesized on XC72 is the most highly<br />

dispersed one. For this sample, the intensity of reflex (220) proved<br />

to be extremely low due to the high dispersion degree of the metal<br />

particles. Therefore, all the data on the particle size, the particle<br />

atomic lattice parameters were obtained by analyzing the form of<br />

the complex line in the diffraction angle range of 2θ 0 of 39 0 – 48 0<br />

after the profile approximation and extension to the component<br />

lines corresponding to the reflexes of planes (111) and (200) of<br />

palladium phase. The corresponding result of the treatment of the<br />

profile of this line is presented at fig. 1b. This sample includes the<br />

phases of metallic ruthenium, metallic palladium, and their alloy<br />

containing up to 10% ruthenium in palladium. The alloy formation<br />

is evidenced by the decrease of Pd lattice parameter at the introduction<br />

of Ru, as compared to pure Pd. The concentration ratio at the<br />

surface is Pd:Ru=1:1. Two lines were observed in the spectra of the<br />

both metal core levels Pd3d5 and Ru3p that corresponded to the<br />

metals, ruthenium particles oxidized to RuO 2 and palladium particles<br />

oxidized to PdO. The concentration ratio of M:MO=4:1 was


Binary Pt– and Pd–based Electrocatalysts for Oxidation of Hydrogen with Co Admixtures / J. New Mat. Electrochem. Systems<br />

85<br />

Figure 3. Polarization dependences for the hydrogen oxidation reaction at the rotating disc electrode coated with Pt–based catalysts at<br />

the rotation rate, rpm: (1) 680; (2) 1080; (3) 1480; (4) 2040; 0.5 M H 2 SO 4 , 60 ºC.<br />

characteristic of both types of particles.<br />

At the diffractograms of the samples of {XC72 + 20 % PdMo},<br />

only the diffraction lines corresponding to the position of maximums<br />

of the cubic palladium lattice were expressed, which pointed<br />

to the absence of a separate metallic molybdenum phase. The lattice<br />

parameter determined for palladium particles proved to be<br />

somewhat higher than in the case of crystalline palladium<br />

(a=0.38902 nm). This means that the changes of the cubic dispersed<br />

palladium phase observed for this sample are negligible<br />

owing to the intercalation into the molybdenum atomic lattice.<br />

However, according to the data of the studies using XPS method,<br />

most of molybdenum in this sample was in the form of oxide. The<br />

structural characteristics of the synthesized catalysts is presented in<br />

table 1.<br />

The surface area of the metallic catalyst component was determined<br />

from the results of potentiodynamic measurements. These<br />

data, together with the other studies performed by means of XRD<br />

and XPS methods, also allow to estimate the effect of the binary<br />

system components on their catalytic properties. The comparison of<br />

the specific surface area of PdRu (50:50 at.%) determined from CO<br />

desorption (97 m 2 /g) and calculated from the particle size according<br />

to XRD data using relationship S=6 * 10 4 /ρ * D (160 m 2 /g, where ρ is<br />

the metal density, D is the particle diameter in Е) demonstrates that<br />

palladium particles are decorated by ruthenium oxides. This is evidenced<br />

by the decrease of the metallic component surface area<br />

accessible for CO adsorption.<br />

Fig. 2 contains, as an example, the cyclic potentiodynamic<br />

curves obtained at the catalysts of XC72+20 wt.%Pd, XC72+20<br />

wt.%PdRu (50:50 at.%) and XC72+20 wt.%PdRu(80:20 at.%).<br />

From the data presented at fig. 2, it is clearly seen that the potential<br />

of the maximum of oxidative CO desorption essentially depends on<br />

the catalyst composition. At the increase of Ru content in the catalyst,<br />

the potential of CO peak shifts towards the negative values. At<br />

the transition from Pd to PdRu (50:50 at.%), the shift of the peak<br />

potential constitutes more than 300 mV. The presented data unambiguously<br />

point to the fact that the introduction of Ru into the catalyst<br />

results in the decrease of the over–potential of the oxidative<br />

CO desorption adsorbed at the catalyst surface. This probably is the<br />

main cause of the decrease of the shown below inhibiting effect of<br />

CO on the process of hydrogen electrooxidation at the introduction<br />

of Ru into the catalyst.<br />

3.2. Electrochemical characteristics of PtM and<br />

PdM catalysts in the reaction of hydrogen oxidation<br />

and their tolerance towards CO<br />

At fig. 3, the polarization curves of hydrogen electrooxidation at<br />

the catalysts of XC72 + 20 % Pt and XC72 +20 % PtMo (75:25<br />

at.%) are presented for the different electrode rotation rates. In both<br />

cases, the process rate is limited by the step of hydrogen molecule<br />

diffusion towards the rotating electrode surface.<br />

The polarization curves of hydrogen electrooxidation at PdM<br />

catalysts at different ratios of the binary system components are<br />

presented at fig. 4. The limiting current value for these catalysts<br />

grows essentially at the increase of the electrode rotation rate,<br />

which evidences the considerable effect of the mass transport step<br />

on the hydrogen reaction rate. As seen from the data presented at<br />

fig. 4 and in table 2, PdM compositions synthesized on carbon<br />

black XC72 are sufficiently effective catalysts of hydrogen oxidation.<br />

However, one should note the more slow raise of the hydrogen<br />

oxidation current at the overpotential growth at these catalysts<br />

as compared to platinum (fig. 3).<br />

In table 2, some electrochemical parameters are presented that<br />

characterize the electrocatalytic properties of the obtained catalysts:<br />

catalyst specific surface areas s, electrode steady–state potential<br />

values in the hydrogen–saturated solution E st , and current densities<br />

at the electrode polarization being 0.05 V.


86 K. V. <strong>Rybalka</strong> et al. / J. New Mat. Electrochem. Systems<br />

Figure 4. Polarization dependences for the hydrogen oxidation reaction at the rotating disc electrode coated with Pd–based catalysts at the<br />

rotation rate, rpm: (1) 680; (2) 1080; (3) 1480; (4) 2040; 0.5 M H 2 SO 4 , 60 0 C.<br />

Figure 5. Dependence of the hydrogen oxidation limiting current<br />

on the square root of the electrode rotation rate: (1) calculated<br />

from the formula of Levich; (2) PdAu (50:50 at.%); (3) PdRu<br />

(50:50 at.%); (4) PdMo (50:50 at.%); 0.5 M H 2 SO 4 , 60 ºC.<br />

Figure 6. Polarization curves of CO oxidation at the catalysts of<br />

(1) XC72 + 30 wt.% Pd; (2) XC72 + 30 wt.% PdAu (50:50 at.%);<br />

(3) XC72 + 20 wt.% PdMo (50:50 at.%); (4) XC72 + 20 wt.%<br />

PdRu (50:50 at.%).<br />

At fig. 5, the dependences of the hydrogen electrooxidation limiting<br />

current on the square root of the electrode rotation rate are<br />

presented for three catalysts: PdAu, PdRu, PdMo at the component<br />

ratio being 50:50 at.%. These dependences are compared with the<br />

calculated data for the rotating disc electrode. For all the three catalysts,<br />

the limiting current changes linearly with the square root of<br />

the electrode rotation rate; however, unlike the calculated dependence,<br />

the experimental curves do not pass the origin of coordinates<br />

at the electrode rotation rate tending to zero. This evidences that the<br />

rate of the hydrogen electrooxidation reaction at these catalysts is<br />

limited not only by mass transport but also by kinetic limitations.<br />

As already noted above, negligible admixtures of CO in hydrogen<br />

already result in the poisoning of platinum catalysts used at<br />

present and to the drastic decrease of their activity in the hydrogen<br />

reaction. One of the methods that can be used to predict the catalyst<br />

tolerance towards CO is the comparison of the relative activity of<br />

different systems in CO oxidation reaction in the electrolyte solution<br />

saturated by this gas. At fig. 6, the dependences of CO electrooxidation<br />

current at various catalysts are presented. As CO<br />

molecules adsorbed at the catalyst surface block the hydrogen oxidation<br />

process, so the start of CO oxidation probably corresponds<br />

to the appearance of CO–free sites at the catalyst surface that can


Binary Pt– and Pd–based Electrocatalysts for Oxidation of Hydrogen with Co Admixtures / J. New Mat. Electrochem. Systems<br />

87<br />

Figure 7. Impedance plots of the hydrogen reaction at the equilibrium potential at various Pd–based catalysts in the frequency range of 20–<br />

100000 Hz. (1) pure H 2 ; (2–5) gas mixture of H 2 +500 ppm CO after conditioning the electrode in the solution (min): (2) 30, (3) 50, (4) 90,<br />

(5) 120. 0.5 M H 2 SO 4 , at the temperature of 60 ºC.<br />

be accessible for hydrogen adsorption and electrooxidation. As can<br />

be seen (fig. 6), the degree of the catalyst surface unblocking can<br />

grow in the series of Pd < PdMo < PdAu < PdRu. Herewith, the<br />

catalyst tolerance in the end is obviously characterized both by the<br />

potential of CO oxidation start and by the ratio of the rates of this<br />

reaction and that of oxygen electrooxidation.<br />

As was shown above, for all the considered catalysts the rate of<br />

hydrogen supply from the solution bulk produces a considerable<br />

effect on the hydrogen reaction flow. To eliminate the effect of the<br />

mass transport step and allow measuring directly the rate of the<br />

electrochemical step of the hydrogen reaction and the influence of<br />

CO admixtures on it, the electrochemical impedance method was<br />

used. At fig. 7, the impedance plots of the electrooxidation process<br />

are presented for hydrogen containing 500 ppm CO at three Pd–<br />

based catalysts. Similar dependences were obtained in H 2 SO 4 solution<br />

saturated by the gas mixture of H 2 + 150 ppm CO. However, in<br />

the latter case the growth of the impedance components in time is<br />

less expressed than in the case of the electrolyte saturation by the<br />

mixture of H 2 + 500 ppm CO.<br />

As follows from the above data, the radius of impedance plots<br />

grows with the increase of the time of electrode conditioning in the<br />

solution, which evidences the hydrogen reaction deceleration at the<br />

adsorption of CO at the catalyst surface. The comparison of the<br />

data presented at fig. 7 with the results of impedance measurements<br />

obtained in similar conditions at the catalyst of XC72 + 20 % Pt<br />

demonstrates already at lower CO content (fig. 8) that CO adsorption<br />

at the studied PdM catalysts produces a less inhibiting effect<br />

on the hydrogen reaction flow, as compared to platinum. In order to<br />

estimate quantitatively the effect of CO admixtures on the hydrogen<br />

reaction rate, the exchange current values were calculated for<br />

this reaction using the impedance spectra. The calculation was<br />

based on Randles model [35,36], which includes the double electric<br />

layer capacitance connected in parallel to the electrochemical reaction<br />

resistance. In [37], where the impedance method is used for<br />

treating the kinetics of hydrogen discharge/ionization on smooth<br />

palladium, a more complicated equivalent scheme offered in paper<br />

[38] was used; however, as shown in [37], in the range of high<br />

enough frequencies the impedance of Pd electrode is sufficiently<br />

well described by Randles model.<br />

At fig. 9, the dependences are presented of the ratios of the exchange<br />

current in pure hydrogen to that determined in the electrolyte<br />

saturated by H 2 + CO gas mixture on the time of electrode<br />

conditioning in the solution. As follows from the curves presented<br />

at the figure, the poisoning of the catalysts by CO molecules adsorbed<br />

at their surface results in the essential decrease of the hydrogen<br />

reaction rate. Simultaneously, these data illustrate the dynamics<br />

of the catalyst surface poisoning by CO traces in hydrogen. The<br />

data on the hydrogen reaction kinetics at various catalysts obtained<br />

by analyzing the impedance spectra are summarized in table 3. As<br />

can be seen, all the studied catalysts are much more tolerant towards<br />

CO admixtures in hydrogen than Pt. PdAu alloys feature the<br />

highest tolerance. This can be caused by the decoration of palladium<br />

surface by gold. As CO is much less adsorbed on gold, as<br />

compared to palladium (the maximum value of 0.2 vs. 1.0, correspondingly),<br />

thence palladium surface decoration with gold must<br />

be the cause of its tolerance towards CO. Although the activity of


88 K. V. <strong>Rybalka</strong> et al. / J. New Mat. Electrochem. Systems<br />

nanoparticles by gold hinders CO adsorption, while seemingly<br />

preserving enough sites that are accessible for hydrogen adsorption.<br />

In the binary system of PdRu, the main role is probably played by<br />

the alloy formation that facilitates greatly the oxidative desorption<br />

of CO, as compared to pure Pd.<br />

5. ACNOWLEGMENT<br />

The work was financially supported by the Presidium of the<br />

Academy of Sciences, Russian Academy of Sciences.<br />

Figure 8. Impedance plots of the hydrogen reaction at the equilibrium<br />

potential at Pt catalyst in the frequency range of 20–100000<br />

Hz. (1) pure H 2 ; (2–5) gas mixture of H 2 +500 ppm CO after conditioning<br />

the electrode in the solution (min): (2) 30, (3) 50, (4) 90,<br />

(5) 120. 0.5 M H 2 SO 4 , at the temperature of 60 ºC.<br />

Figure 9. Dependence of relative exchange current i 0 (t)/i 0 (0) on the<br />

electrode conditioning time in 0.5 M H 2 SO 4 solution saturated by<br />

the gas mixture of H 2 +500 ppm CO. 60 0 C. (1) XC72 + 30 wt.%<br />

PdAu (50:50 at.%); (2) XC72 + 20 wt.% PdMo (50:50 at.%); (3)<br />

XC72 + 20 wt.% PdRu (50:50 at.%); (4) XC72 + 20 wt.% Pt, CO<br />

content is 200 ppm.<br />

this catalyst in pure hydrogen is much lower than that of platinum,<br />

in time, due to the slower poisoning of the electrode by CO adsorption<br />

on its surface as compared to platinum, it proves to be more<br />

effective in the electrooxidation reaction of hydrogen with CO<br />

admixtures.<br />

4. CONCLUSION<br />

As follows from the comparison of the presented data, the binary<br />

systems of PdM are characterized by different mechanisms of tolerance<br />

towards CO. The decoration of Pd nanoparticles by Mo oxides,<br />

similarly to PtMo system, results in the earlier oxidation of<br />

adsorbed CO, which is probably due to the redox process with the<br />

participation of Mo 4+ /Mo 6+ pair [39]. The decoration of Pd<br />

REFERENCES<br />

[1] S. Gottesfeld, J. Pafford, J. Electrochem.Soc., 135, 2651 (1988).<br />

[2] H. F. Oetjen, V. M. Shmidt, U. Stimming, F. Trila, J. Electrochem.Soc.,<br />

143, 3838 (1996).<br />

[3] P.Costamagna, S.Srinivasan, J.Power Sources, 102, 242 (2001).<br />

[4] G. A. Camara, E. A. Ticianelli, S. Mukerjee, S. J. Lee, J. Mc<br />

Breen, J. Electrochem.Soc., 149, A748 (2002).<br />

[5] Li Qingfeng, H. A. Hjuler, С Hasiotis, J. K. Kallitsis, C. G.<br />

Kontoyannis, N.J. Bjerrum, Electrochem. and Solid-State Letters,<br />

5, A 125 (2002).<br />

[6] R. Parson, T. Vander Noot J. Electroanal Chem., 257, 9 (1988).<br />

[7] G. Samjeske, H. Wang, T. Loffler, H. Baltruschat, Electrochim.<br />

Akta, 47, 3681 (2002).<br />

[8] K. A. Fridrich, K. P. Geyzers, A. J. Dickinson, U. Stimming, J.<br />

Electroanal. Chem 524 - 525, 261 (2002).<br />

[9] A. Kabbabi, R. Faure, R. Durand, B. Beden, D. Hahn, J.-M.<br />

Leger, C. Lamy, J. Electroanal. Chem. 444, 41 (1998).<br />

[10]J. Divisek, H.-F. Oetjen, V. Peinecke, V.M. Schmidt, U. Stimming,<br />

Elektrochim.Acta 43, 3811 (1998).<br />

[11]A.M. Castro Luna, G.A. Camara, V.A. Paganin, E.A. Ticianelli,<br />

E.R.Gonzalez, Elektrochem. Commun. 2, 222 (2000).<br />

[12]T. Joroin, N. Fugiwara, Z. Siroma, K. Yasuda, Y. Miyazaki,<br />

Electrochem.Commun. 4, 442 (2002).<br />

[13]Z. Qi, A. Kaufman, J. of Power Sources 113, 115 (2003).<br />

[14]S. Mukerjee, S.J. Lee, E.A. Ticianelli, J. McBreen, B.N. Grgur,<br />

N.M.Markovic, P.N. Ross, J.R. Giallombardo, E.S. De Castro,<br />

Electrochem.and Solid-State Letters 2, 12 (1999).<br />

[15]S. Mukerjee, R.C. Urian, Electrochim. Akta 47, 3219 (2002).<br />

[16]E.M. Crabb, M.K. Ravikumar, Y.Qican, A.E. Russel, S.<br />

Maniquet, J.Yaa, D.Thompsett, M. Hurford, S.C. Ball., Electrochim.<br />

and Solid-State Letters 5, A5 (2002).<br />

[17]B.N. Grgur, G.Zhuang, N.M. Marcovic, P.N. Ross, J. Phys.<br />

Chem., 101, 3910 (1997).<br />

[18]B.N.Grgur, G. Zhuang, N.M. Marcovic, P.N. Ross, J. Phys.<br />

Chem. 102, 2494 (1998).<br />

[19]T.J.Schmidt, Z. Jusys, H.A.Gasteiger, R.J.Behm, U. Endruschat,<br />

H. Boennemann, J. Electroanal. Chem., 501, 132<br />

(2001).<br />

[20]T.J. Schmidt, H.A. Gasteiger, R.J. Behm, J. New. Mater. Electrochem<br />

Soc. 2, 27 (1999).<br />

[21]M. Lukaszewski, K. Kusmierczyk, J. Kotowski, H. Siwek,<br />

A.Czerwinski, J. Solid State Electrochem. 7, 769 (2003).<br />

[22]K. Gossner, E. Mizera, J. Elektroanal Chem 140, 35 (1982).


Binary Pt– and Pd–based Electrocatalysts for Oxidation of Hydrogen with Co Admixtures / J. New Mat. Electrochem. Systems<br />

89<br />

[23]T. J. Schmidt, H. A. Gasteiger, G. D. Stöb, P. M. Urban, D. M.<br />

Kolb, R. J. Behm, J. Electrochem. Soc. 145, 2354 (1998).<br />

[24]Toshima, Y. Yamazaki, K. Asakura, J. Phys. Chem. 96,<br />

9927(1992).<br />

[25]M. R. Tarasevich, L. A. Beketaeva, B. N. Efremov, N. M.<br />

Zagudaeva, L. N. Kuznetsova, K. V. <strong>Rybalka</strong>, and V. E. Sosenkin,<br />

Russian J. Electrochem., 40, 612 (2004).<br />

[26]T.J.Schmidt, H.A.Gesteiger, R.J.Behm, J.of Electrochem. Soc.,<br />

146, 1296 (1999).<br />

[27]G. Levich, Physicochemical Hydrodynamics, Fizmatgiz, Moscow,<br />

1959.<br />

[28]P. Delahay, “Double Layer and Electrode Kinetics” Interscience<br />

Publishers, John Wiley& sons, N.Y.-London-Sidney,<br />

1965.<br />

[29]Carter W.E., Schweitzer G.K., Carlson T. A., J. Electron. Spectrosc.<br />

, 5 , 827 (1974).<br />

[30]R. Bouwman, P. Biloen, J. Catal., 48, 209 (1977).<br />

[31]P. J. M. Kerkhof , J. A Moulin. , J. Phys.Chem. 83, 1612<br />

(1973).<br />

[32]C. Defosse , J. Electron. Spectrossc., 23, 157 (1981).<br />

[33]S. C. Fung, J. Catal. 56, 454 (1979).<br />

[34]M. R. Fitzsimons, J. A. Eastmen., M. Muller-Stach, G. Walner,<br />

Phys.Rev., 44, 2452 (1991).<br />

[35]J. E. Randles, Disc.Faraday Soc., 11, 1 (1947).<br />

[36]Ya. Kuta, E. Yeger, in the book “Measurement Methods in<br />

Electrochemistry”, Eds. E. Yeger, A. Zalkind, MIR, Moscow,<br />

1977.<br />

[37]Maximov Yu. M., Bruners R. U., Podlovchenko B. I., Elektrokhimiya,<br />

22, 1000 (1986).<br />

[38]M. W. Breiter, Z. Phys. Chem. N.F., 112, 183 (1978).<br />

[39]S. Makerjee, S. J. Lee, A. E. Ticianelli, J. McBreen, J. Electrochem.Solid-State<br />

Lett., 2, 12 (1999).<br />

[40]R. C. Urian, A. F. Gulla, S. Makerjee, J. Electroanal.Chem,<br />

554-555, 307 (2003).

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!