Journal of Electroanalytical Chemistry Journal of Electroanalytical Chemistry 581 (2005) 294–302 www.elsevier.com/locate/jelechem Kinetic study of formic acid oxidation on carbon-supported platinum electrocatalyst J.D. Lović a, A.V. Tripković a,*, S.Lj. Gojković b, K.Dj. Popović a, D.V. Tripković a, P. Olszewski c, A. Kowal c b a ICTM-Institute of Electrochemistry, University of Belgrade, Njegoševa 12, P.O. Box 473, 11000 Belgrade, Serbia and Montenegro Faculty of Technology and Metallurgy, University of Belgrade, Karnegijeva 4, P.O. Box 3503, 11000 Belgrade, Serbia and Montenegro c Institute of Catalysis and Surface Chemistry, Polish Academy of Sciences, Krakow, Niezapominajek 8, 30-239 Poland Received 24 February 2005; received in revised form 20 April 2005; accepted 6 May 2005 Available online 15 June 2005 Abstract Oxidation of formic acid on the platinum catalyst supported on high area carbon was investigated by potentiodynamic and quasi-steady-state polarization measurements. It was found that the poisoning of the reaction occurred both in the hydrogen region and in the double-layer region, but poisons were formed faster at lower potentials. Kinetics of the reaction was consistent with the dual path mechanism. At lower potentials HCOOH was oxidized to CO2 at the Pt sites uncovered by COads. If high coverage by the poisoning species was attained, the reaction reached the limiting current plateau and further increase of the current densities started at the potential of COads oxidation. Kinetic parameters of the HCOOH oxidation suggested that the rate determining step was the transfer of the first electron from HCOOHads, which was adsorbed under the Temkin conditions. Oxidation of formic acid became pH-dependent reaction in the electrolytes of pH < 1 with the reaction order with respect to H+ ions of about
0.8. Ó 2005 Elsevier B.V. All rights reserved. Keywords: Formic acid; Electrochemical oxidation; Platinum supported catalyst; Fuel cells 1. Introduction Electrochemical oxidation of formic acid: HCOOH=CO2 + 2Hþ + 2e
E 0 =
0:25 V [1] ð1Þ has been investigated on platinum since early work of Breiter [2] and the results have been reviewed by Parsons and VanderNoot [3] and Jarvi and Stuve [4]. However, in the last several years this reaction has been attracting more attention [5–14] because a direct formic acid– oxygen fuel cell with polymer electrolyte membrane (PEM) has some advantages over a direct methanol fuel cell. Oxidation of formic acid commences at less positive * Corresponding author. Tel.: +381 11 347 0390; fax: +381 11 337 0389. E-mail address: amalija@tmf.bg.ac.yu (A.V. Tripković). 0022-0728/$ - see front matter Ó 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.jelechem.2005.05.002 potential than methanol oxidation [15] and crossover of formic acid through the polymer membrane is lower than that of methanol [16]. It has been widely accepted in the literature that HCOOH is oxidized to CO2 via a dual path mechanism [17,18] which involves a reactive intermediate (main path, dehydrogenation) and adsorbed CO as a poisoning species (parallel path, dehydration): ð2Þ Adsorbed formate (HCOO), rather than the formic acid fragment (COOH), was proposed as the reactive intermediate [6,19] and this assumption was recently confirmed by direct surface-enhanced infrared absorption J.D. Lović et al. / Journal of Electroanalytical Chemistry 581 (2005) 294–302 spectroscopy (SEIRAS) [10]. Adsorbed CO was assigned as the poisoning species and detected by IR spectroscopy [8,20]. However, oxidation of formic acid is not that simple reaction. COH/HCO species was detected by electrochemical thermal desorption mass spectroscopy (ECTDMS) [21] and assigned as a reactive intermediate, but there is an opinion that highly reactive intermediate may not be detected by the techniques currently in use [4] meaning that COH/HCO might be another poisoning particle. There is also a controversy about the formation of adsorbed CO. The following reaction sequence is possible [17,18]: (COOH)ads + 2Hads ! (COH)ads + H2 O þ (COH)ads ! COads + H + e
ð3Þ ð4Þ indicating that adsorbed hydrogen is necessary for the poison formation. Another rather complicate mechanism of the formation of COads was also proposed [22]: (HCOOH)ads + Hþ + e
! [CH(OH)2 ]ads ð5Þ [CH(OH)2 ]ads ! (CHO)ads + H2 O ð6Þ (CHO)ads ! COads + Hþ + e
ð7Þ assuming amphoteric nature of HCOOH. However, besides being the poisoning species, CO may act as the reactive intermediate while some fraction of COads can be oxidized with OHads to produce CO2 [22]. Platinum and platinum alloy nanoparticles supported on high area carbon are state of the art electrocatalysts in PEM fuel cells. Recently, Weaver and co-workers [8] investigated the oxidation of CH3OH, HCOOH, and HCHO on Pt/C electrocatalyst by cyclic voltammetry and IR spectroscopy and established a particle size effect in these reactions. Particles with d < l4 nm were found to be the most active for the oxidation of HCOOH. In the study presented here, we investigated the oxidation of formic acid on a Pt/C electrocatalyst with the average particle size of 3–4 nm. The aim was to explore the poisoning of the reaction, the influence of anions and to establish kinetic parameters under the steady-state conditions. in the constant loading of 20 lgPt cm
2. After drying in the stream of high-purity nitrogen at room temperature, the deposited catalyst layer was covered with 20 ll of a diluted aqueous NafionÒ solution (thickness of ca. 0.1 lm) and left to dry completely. Mass transfer resistance through the NafionÒ film covering the Pt/C layer was determined by recording the diffusion limiting currents of the hydrogen oxidation on the rotating disk electrode. Since Levich–Koutecky plots with the zero intercept were obtained, it was concluded that the mass transfer resistance through the NafionÒ film was negligible [24]. 2.2. Characterization of the catalyst The catalyst was characterized by the high resolution transmission electron microscopy (HRTEM) technique. The images of the electrode and the histogram of the particle size distribution [25] showed that Pt particles size ranged between 2 and 6 nm with an average mean particle diameter of 4 ± 0.3 nm. X-ray diffraction (XRD) measurements were carried out with a Siemens D5005 diffractometer using Cu Ka source operating at 40 mA and 40 kV and a graphite monochromator. The spectra were obtained in the 2h range from 10° to 110° at the rate of 0.04°/10 s. Commercial software (EVA) was used for subtracting the background and measuring of the integral breadth of the selected reflections. Apparatus peak broadening was determined in a separate experiment and taken into account in calculations. The diffraction pattern presented in Fig. 1 shows the characteristic peaks of Pt fcc structure. A broad peak at 2h of 25° originated from carbon support was also registered. The size of Pt crystallites was calculated from the broadening of (2 2 0) peak using Scherrer equation d¼ 0.9k ; B2h cos hmax 2. Experimental 2.1. Electrode preparation A platinum electrocatalyst supported on high area carbon (Pt/C) with 47.5 mass% Pt (Tanaka Precious Metal Group) was applied to a glassy carbon substrate in the form of a thin film [23]. A suspension of Pt/C in water was prepared in an ultrasonic bath and a drop of the suspension was placed onto the substrate resulting 295 Fig. 1. XRD pattern of 47.5 mass% Pt/C electrocatalyst. ð8Þ 296 J.D. Lović et al. / Journal of Electroanalytical Chemistry 581 (2005) 294–302 where d is average particle size in nm, k is the wavelength of X-ray (0.154056 nm), h is the angle at the peak maximum, and B2h is the breadth (in rad) of the peak at half height. The value of the particle size was calculated to be 3.1 nm. The Pt crystallite size was also determined using TOPAS program (Bruker-AXS, Germany), graphic based profile analysis program for structure determination and Rietveld refinement. Peak shape was modeled using fundamental parameter approach (FDA). Structures used in the refinement were Pt-fcc. This method gave the similar result as Scherrer equation, i.e., 2.9 nm, which is also in agreement with the HRTEM measurements. The total surface area of Pt particles was determined by using hydrogen adsorption/desorption charge from the steady-state cyclic voltammograms in the supporting electrolyte (see Fig. 2(a)) and a charge of 210 lF cm
2 for monolayer hydrogen adsorption. The specific surface area of Pt in the Pt/C catalyst was calculated to be 65 m2 g
1. Approximating cubo-octahedral Pt particles by ideal spheres [26], we calculated that the Pt particle diameter was 4.3 nm. This value is slightly higher than that obtained by XRD and HRTEM, which is to be expected because the total surface area of the supported catalyst particles in a thin film on the electrode is never accessible in the electrochemical experiments. The current densities for the oxidation of formic acid in this paper are given with respect to the surface area determined by cyclic voltammetry, because this surface area is relevant for the electrochemical studies. 2.3. Electrochemical measurements All electrochemical measurements were conducted in a standard electrochemical cell with a Pt wire spiral as the counter electrode and a saturated calomel electrode (SCE) as the reference electrode. All the potentials reported in the paper are expressed on the scale of SCE. The cell was thermostated at 22.0–60.0 °C, while the reference electrode was always at 22 °C. Most of the experiments were performed at 22.0 °C except those where an activation energy was determined. The electrolyte contained 10
2–1 M HClO4 or 0.1 M H2SO4 as a supporting electrolyte and 10
2–1 M HCOOH. All solutions were prepared with high purity water (Millipore, 18 MX cm resistivity) and p.a. grade chemicals (Merck). The electrolyte was deaerated with the bubbling of nitrogen. After having immersed a Pt/ C electrode in the supporting electrolyte, the potential was cycled between hydrogen and oxygen evolution regions at 50 mV s
1 until a steady-state voltammogram was obtained. Then, formic acid was added while the potential was held at
0.25 V for 3 min and the positive-going scan was initiated with the rate of 50 mV s
1 (potentiodynamic polarization curves) or 1 mV s
1 (quasi-steady-state polarization curves). 3. Results and discussion 3.1. Poisoning of the reaction Fig. 2. Steady-state cyclic voltammogram of Pt/C catalyst in 0.1 M HClO4 solution (a) and the first cyclic voltammogram after addition of 0.5 M HCOOH (b) recorded at a scan rate of 50 mV s
1. Steady-state voltammogram of Pt/C electrocatalyst in perchloric acid solution and the first forward and backward voltammogram for oxidation of formic acid at the same electrode are presented in Fig. 2. Voltammogram of Pt/C shows usual characteristics of Pt surface except that current peaks for adsorption/desorption of hydrogen are not as sharp and well resolved as on a smooth Pt electrode. The hydrogen region is followed by a double-layer charging current, adsorption of oxygen containing species and their reduction. Voltammogram of HCOOH oxidation shows the first scan after holding the electrode potential at
0.2 V for 3 min, while HCOOH was added into the electrolyte. It can be seen that the reaction commences in the hydrogen region and proceeds slowly in the positive scan direction reaching a plateau at 0.25 V. At the potentials more positive than 0.5 V, the reaction becomes significantly accelerated attaining a maximum rate at J.D. Lović et al. / Journal of Electroanalytical Chemistry 581 (2005) 294–302 297 0.8 V and finally the current density falls to zero within a narrow potential range. Immediately upon reversing the potential scan, a very steep increase of the reaction rate occurs. After reaching a maximum at 0.7 V, the current gradually decreases but the reaction remains much faster than in the forward scan. It should be noted that the voltammogram for oxidation of formic acid was found to be insensitive on the rotation of the electrode, indicating that the reaction is under activation control within the entire potential region. The spectra of voltammetric profiles recorded by reducing the negative potential limit are depicted in Fig. 3. When the scan was reversed at 0.1 V, the current densities were much higher than on the voltammogram started at
0.2 V (Fig. 2(b)) and a broad maximum instead of the plateau appeared preceding the peak at 0.8 V. Gradual shift of the negative potential limit incited increase in the reaction rate in the forward scan and overlapped the peaks. The results in Fig. 3 could suggest that more poisoning species are formed at more negative potentials. However, if the experiment was started by holding the electrode potential at 0.1 V for 3 min, the current densities in the first forward scan (curve a in Fig. 4) were low Fig. 4. The first cyclic voltammogram after holding the potential for 3 min at 0.1 V (a), the next cycle in the same limits but without holding the potential (b), the first cycle with the extended negative potential limit (c), and the subsequent scan after holding the potential 3 min at
0.2 V (d). Electrolyte: 0.1 M HClO4 + 0.5 M HCOOH, scan rate of 50 mV s
1. Fig. 3. Cyclic voltammograms of oxidation of 0.5 M HCOOH in 0.1 M HClO4 recorded with different negative potential limits. Scan rate of 50 mV s
1. and the plateau appeared. After oxidizing the poisoning species at more positive potentials, in the next scans with the same negative limit but without holding the potential, the current densities were much higher and stable during continuous cycling (curve b). Extending the negative potential limit to
0.2 V suppressed the reaction rate significantly (curve c). In the next scan performed after 3 min at
0.2 V (curve d), the current densities were even lower and the plateau expanded. These results clearly show that poisoning of the reaction is not determined only by the potential (curves b and c) but also by the time that electrode resides in a certain potential region (curves a and b). The reaction can be deactivated at the potentials outside of the hydrogen region which rules out Hads as the participant in the formation of the poison. The finding that the poisoning is faster and more expressed at more negative potentials does not contradict previous conclusion because the potential also influences the structure of the double layer and the orientation of the organic molecules. Deactivation of the formic acid oxidation was also investigated by the potential steps from
0.2 V to the 298 J.D. Lović et al. / Journal of Electroanalytical Chemistry 581 (2005) 294–302 Fig. 5. Current decays following the potential step from
0.2 V to the potentials indicated on the diagram. Electrolyte: 0.1 M HClO4 + 0.5 M HCOOH. Fig. 6. Cyclic voltammograms of oxidation of 0.5 M HCOOH in 0.1 M HClO4 recorded with different positive potential limits. Scan rate of 50 mV s
1. various potentials before the reaction reaches the plateau. Current–time transient curves presented in Fig. 5 show that the current drops rapidly upon potential steps to E 6 0 V, while at the potentials outside of the hydrogen region, E P 0.1 V, the quasi-stationary values of the current density are higher. A faster formation of the poisoning species at more negative potentials is in accord with the cyclic voltammetry given in Fig. 4. The influence of the positive potential limit on the reaction rate in the backward scan is presented in Fig. 6. Almost complete overlapping of the curves occurs up to 0.4 V, while an enhanced activity in the backward scan appears when the positive potential limit is more positive than 0.5 V. It can be seen that poisoning species cannot be removed substantially at the potentials less positive than 0.6 V. Oxidation of formic acid on the Pt/C electrocatalyst is in accord with the dual reaction path (Eq. (2)). At the potentials E < 0.4 V formic acid oxidizes through the main, dehydrogenation path. Dehydration reaction with COads as a final product occurs in the parallel path. Although substantial amount of COads is formed before main path commences (at E =
0.2 V, hCO  0.6 according to [8]), it seems that the coverage with COads continues to grow. This is implied by the decrease in the current densities during potential steps up to 0.2 V and by the shape of the voltammetric curve for the formic acid oxidation. The plateau on the curve could be caused by the slow diffusion, but it was found that the current densities were independent of the stirring of the electrolyte. If a slow chemical reaction preceding electrochemical step (CE mechanism) was responsible for the plateau, with the increase in the scan rate the plateau should be transformed in a peak [27], but in our experiments the shape of the voltammogram was independent of the scan rate. Since the plateau appeared only when the electrode surface was highly deactivated, it is probably caused by the increase in the coverage with the poisoning species and consequent decrease in the number of the available sites for the dehydrogenation reaction. Adsorbed CO is certainly the dominant poisoning species because the steep increase in the reaction rate at E  0.45 V coincides with the potential of the oxidative removal of COads on Pt/C electrocatalyst [28]. However, some other blocking species like COH/HCO or dimers might also be formed in a path parallel to the main reaction. J.D. Lović et al. / Journal of Electroanalytical Chemistry 581 (2005) 294–302 3.2. Anion effect Cyclic voltammograms of Pt/C electrocatalyst in perchloric and sulfuric acid solution as well as voltammograms of oxidation of formic acid in the same electrolytes are given in Fig. 7. The adsorption of anions originating from the supporting electrolytes affected cyclic voltammograms of Pt/C in both hydrogen region (E < 0.1 V) and at higher potentials where oxygen species are adsorbed at Pt surface. The adsorption of bisulfate anions at polycrystalline Pt commences deeply in the hydrogen region (at approximately
0.25 V) and extends to the more positive potentials reaching a maximum at 0.45 V [29]. In our experiments, the peaks for hydrogen adsorption/desorption at about
0.05 V were better expressed in the presence of bisulfate anions, but the total charge for upd of H was similar in both electrolytes. The adsorption of OH particles and the formation of Pt-oxide commence at lower potentials in perchloric acid solution than in sulfuric acid. The influence of bisulfate anions on the OH adsorption is usually attributed to the blocking of the Pt sites and/or to displacement of OHads by the bisulfate anions. The adsorption of perchlorate anions is questionable though recent mea- Fig. 7. Steady-state cyclic voltammogram of Pt/C catalyst in 0.1 M H2SO4 and 0.1 M HClO4 solution (a) and the first cyclic voltammograms after addition of 0.5 M HCOOH (b). H2SO4 solution (solid line), HClO4 solution (dashed line). Scan rate 50 mV s
1. 299 surement by electrochemical quartz crystal microbalance (EQCM) indicated adsorbed hydrated perchlorate anions on polycrystalline Pt [30]. Anyway, if perchlorate anions are present on the Pt surface, they should be bound more weakly than bisulfate anions and their influence on the adsorption of other species should be less pronounced. The presence of chloride anions, as an impurity in perchloric acid solution, may also contribute to the availability of the Pt active sites [31,32]. However, the voltammograms of oxidation of formic acid depicted in Fig. 7 show that there is no significant difference between the reaction rate in perchloric and sulfuric acid solution at the potentials up to the plateau, i.e., in the region where the main reaction path is operative. At higher potentials, oxidation of adsorbed CO commences at lower potentials in perchloric acid (which is in agreement with the Pt/C voltammogram), indicating that the adsorption of OH particles occurs at lower potentials in those media. Besides, the current densities for oxidation of COads in perchloric acid are higher than in sulfuric acid solution. The same behavior was observed under the steady-state conditions, which is illustrated by the Tafel plots in Fig. 8. In the main reaction path formic acid oxidizes via weakly chemisorbed species [4] at the Pt sites uncovered by the adsorbed particles, while in the parallel path COads is formed and oxidized at high potentials. According to the results in Figs. 7 and 8, adsorption of anions does not influence the number of free sites available for the main path but increases the amount of COads. This can be rationalized by two modes of action of bisulfate anions. Their adsorption on Pt surface decreases the number of free Pt sites. At the same time adsorbed bisulfate particles reduce the number of the ensembles of two adjacent Pt sites, which are necessary Fig. 8. Tafel plots for oxidation of 0.5 M HCOOH in 0.1 M HClO4 solution (open symbol) and in 0.1 M H2SO4 (bold symbol). Scan rate 1 mV s
1. 300 J.D. Lović et al. / Journal of Electroanalytical Chemistry 581 (2005) 294–302 for the formation of COads (ensemble effect) [33]. Consequently, the coverage by blocking adsorbates (bisulfates and COads in sulfuric acid and COads in perchloric acid) appears to be almost equal, giving similar current densities of the oxidation of formic acid in the main path. When oxidation of COads becomes predominant, higher coverage by COads in perchloric acid results in higher current densities. 3.3. Kinetics and mechanistic study When the polarization curves were recorded by the slow sweep of 1 mV s
1 (Fig. 8), the current densities were approximately the same as the values from the decay curves (Fig. 5) taken after 15 min. Therefore, the steady-state conditions can be approximated by the slow sweep polarization. Tafel plots presented in Fig. 8 show linear region between
0.05 and 0.2 V with a slope of about 150 mV dec
1. A similar slope was obtained in the experiments in 0.1 M HClO4 at temperatures up to 60 °C. From the corresponding Arrhenius plot (Fig. 9), the activation energy of about 20 kJ mol
1 was calculated. This value is close to 21 kJ mol
1 reported for bulk platinum [34]. The influence of the formic acid concentration was investigated in 0.1 M HClO4 solution with 0.01–1 M HCOOH. Reaction rate increased with the increasing HCOOH concentration up to 0.5 M. Further increase in the concentration of the reactant resulted in a decrease in the reaction rate. Fig. 10 shows the current densities at a constant potential in the Tafel region as a function of the HCOOH concentration. Line with a slope of about 0.5 implies that the reaction follows half order kinetics with respect to HCOOH in the electrolytes containing less than 0.5 M HCOOH. A similar concentration dependence was also observed at meso- Fig. 9. Arrhenius plot for oxidation of 0.5 M HCOOH in 0.1 M HClO4. Data taken from the Tafel lines at the potential indicated on the diagram. Fig. 10. Dependence of HCOOH oxidation current density on the HCOOH concentration in 0.1 M HClO4 solution. Data taken from the Tafel lines at the potential indicated on the diagram. porous Pt [9] as well as at Pt/C electrocatalyst [8]. In the electrolytes containing more than 0.5 M HCOOH, the surface coverage by COads is so high, i.e., the surface available for the active intermediate is so low, that the further increase in HCOOH concentration cannot produce increase in the current density. The extensive production of CO at high HCOOH concentration was postulated by Weaver at co-workers [8] and supported by IRAS measurements. The influence of the concentration of H+ ions was investigated in the electrolytes containing 0.5 M HCOOH and different concentrations of HClO4. In Fig. 11 the current densities at a constant potential in the linear Tafel region are plotted as a function of the concentration of HClO4. In the range of H+ ion concentration between 0.01 and 0.1 M the reaction rate does Fig. 11. Dependence of the HCOOH oxidation current density on the concentration of HClO4 in the solutions containing 0.5 M HCOOH. Data taken from the Tafel lines at the potential indicated on the diagram. J.D. Lović et al. / Journal of Electroanalytical Chemistry 581 (2005) 294–302 301 not depend on the H+ ion concentration, but in more acidic electrolytes a decrease in the reaction rate with the increase in H+ concentration was observed. The slope, i.e., reaction order with respect to H+ ion, was estimated to be
0.8. Data about pH dependence of the oxidation of formic acid are scarce in the literature. According to stationary voltammograms on polycrystalline Pt in [14], reaction rate decreased with increasing concentration of H2SO4 ion at c > 0.5 M with a reaction order of about
0.4. The Tafel slope of 120 mV dec
1 and the reaction order with respect to HCOOH of 0.5 can be rationalized by the following steps of the main path for oxidation of formic acid (Eq. (2)): HCOOH ¢ HCOOHads ð9Þ HCOOHads ! HCOOads + Hþ + e
ð10Þ HCOOads ¢ CO2 + Hþ + e
ð11Þ with the adsorption of HCOOH under the Temkin conditions [35] and the transfer of the first electron as the rate determining step. In that case, the rate of step (10) is given by the equation j ¼ Fk 2 ð1
hp ÞhHCOOH     brhHCOOH bFE  exp ; exp RT RT ð12Þ where hp is the coverage by the poisoning species and r is the rate of change of Gibbs energy of adsorption. According to the Temkin approximation of the Frumkin isotherm [35], coverage by HCOOHads is related to the concentration of HCOOH in the bulk of the solution by   rhHCOOH ð13Þ exp ¼ K 1 cHCOOH . RT Substituting (13) in (12) and taking b = 0.5, one obtains the following equation for the overall reaction rate:   0.5FE 0 0.5 j ¼ 2Fk 2 ð1
hp ÞcHCOOH exp . ð14Þ RT Reaction order with respect to HCOOH of 0.5 is consistent with the experimentally observed value at low HCOOH concentration. However, in our experiments the Tafel slope was 150 mV dec
1, i.e., higher than that predicted by Eq. (14). This can be attributed to the change in hp. If the coverage by COads at the beginning of the reaction is 0.6, as it was determined for the same type of the catalyst [8], and linearly grows with the potential, a simple simulation (Fig. 12) shows that the apparent Tafel slope will change to 150 mV dec
1 and that the Tafel-like behavior extends over 0.2 V (0.6 < hp < 0.8) which is followed by the increase in the slope. This model agrees with our experimental data in Fig. 12. Tafel plots simulated according to Eq. (14) with hp = 0.6 (bold symbols) and hp changing linearly from 0.6 to 0.95 (open symbols). the solutions containing less than 0.5 M HCOOH. In more concentrated solutions, the reaction rate does not increase further with the HCOOH concentration because of the high coverage by the poisoning species (extensive production of COads in concentrated solutions, vide supra). However, only results of the surface coverage by COads as a function of the potential and HCOOH concentration combined with the steady-state polarization measurements can prove the proposed mechanism. Reaction mechanism (9)–(11) does not predict the dependence of the reaction rate on the concentration of H+ ions. The change of the reaction order from zero to about
1 is an indication that the reactant participates in an acid–base equilibrium. Formic acid is a weak acid HCOOH ¢ HCOO
+ Hþ K = 1.8  10
4 ð15Þ
If HCOO anion was the only electroactive species, reaction order with respect to H+ ion would change from zero to
1, but at pH 2.4, which is more than an order of magnitude apart from pH 1 in our experiments (Fig. 11). Mechanism of formation of COads given by Eqs. (5)–(7) predicts first-order kinetics with respect to H+ ion, which should result in decrease of the number of the active sites for the main reaction with decreasing pH, but without discontinuity in the pH dependence of the reaction rate. Therefore, a spectroscopic analysis of the intermediates in electrolytes of different pH would be necessary to resolve the pH dependence of the formic acid oxidation. 4. Conclusions On the basis of the investigation of oxidation of formic acid on Pt/C electrocatalyst, the following can be concluded: 302 J.D. Lović et al. / Journal of Electroanalytical Chemistry 581 (2005) 294–302  Reaction follows the dual path mechanism comprising dehydrogenation of HCOOH as the main reaction and formation of poisoning species as the parallel reaction. At lower potentials dehydrogenation of HCOOH contributes to the current density, while poisons (probably COads) become reactive intermediate at E > 0.45 V.  Poisoning species are formed within the hydrogen region as well as in the double-layer region. Poisoning of the surface was found to be more rapid at lower potentials.  Oxidation of formic acid in the presence of the perchlorate and sulfate anions is close to each other in the potential region where the main path is predominant. Oxidation of COads commences at lower potentials and the current densities are higher in perchloric acid solution than in sulfuric acid solution.  Oxidation of formic acid increases with the increasing HCOOH concentration up to 0.5 M with the reaction order of 0.5, but at higher concentration a negative reaction order appeared.  Tafel slope of about 150 mV dec
1 was determined which is consistent with the transfer of the first electron as the rate determining step and the gradual increase in the coverage with the poisoning species.  Reaction rate is independent on concentration of H+ ions up to 0.1 M but in more acidic electrolytes the negative reaction order of about
0.8 was determined. Acknowledgment This work was supported financially by the Ministry of Science and Ecology, Republic of Serbia, Contract No. H-1796. References [1] T. Iwasita, in: W. Vielstich, H.A. Gasteiger, A. Lamm (Eds.), Handbook of Fuel Cells – Fundamentals, Technology and Applications, Electrocatalysis, Vol. 2, Wiley, New York, 2003, p. 604. [2] M.W. Breiter, Electrochim. Acta 8 (1963) 447. [3] R. Parsons, T. VanderNoot, J. Electroanal. Chem. 257 (1988) 9. [4] T.D. Jarvi, E.M. Stuve, in: J. Lipkovski, P.N. Ross (Eds.), Electrocatalysis, Wiley–VCH, Weinheim, 1998. [5] G.-Q. Lu, A. Crown, A. Wieckowski, J. Phys. Chem. B 103 (1999) 9700. [6] M.F. Mrozek, H. Luo, M.J. Weaver, Langmuir 16 (2000) 8463. [7] T.J. Schmidt, R.J. Behm, B.N. Grgur, N.M. Markovic, P.N. Ross Jr., Langmuir 16 (2000) 8159. [8] S. Park, Y. Xie, M.J. Weaver, Langmuir 18 (2002) 5792. [9] J. Jiang, A. Kucernak, J. Electroanal. Chem. 520 (2002) 64. [10] A. Miki, S. Ye, M. Osawa, Chem. Commun. (2002) 1500. [11] M.D. Maciá, E. Herrero, J.M. Feliu, J. Electroanal. Chem. 554– 555 (2003) 25. [12] C. Rice, S. Ha, R.I. Masel, A. Wieckowski, J. Power Sources 115 (2003) 229. [13] M. Zhao, C. Rice, R.I. Masel, P. Waszczuk, A. Wieckowski, J. Electrochem. Soc. 151 (2004) A131. [14] H. Okamoto, W. Kon, Y. Mukouyama, J. Phys. Chem. B 108 (2004) 4432. [15] J. Willsau, J. Heitbaum, Electrochim. Acta 31 (1986) 943. [16] X. Wang, J.-M. Hu, I.-M. Hsing, J. Electroanal. Chem. 562 (2004) 73. [17] A. Capon, R. Parsons, J. Electroanal. Chem. 44 (1973) 1. [18] A. Capon, R. Parsons, J. Electroanal. Chem. 45 (1973) 205. [19] C. Lamy, J.M. Leger, J. Chim. Phys. 88 (1991) 1649. [20] S.-C. Chang, L.-W.H. Leung, M.J. Weaver, J. Phys. Chem. B 94 (1990) 6013. [21] S. Wilhelm, W. Vielstich, H.W. Buschmann, T. Iwasita, J. Electroanal. Chem. 229 (1987) 377. [22] N.M. Markovic, P.N. Rorr Jr., Surf. Sci. Rep. 45 (2002) 117. [23] T.J. Schmidt, H.A. Gasteiger, R.J. Behm, Electrochem. Commun. 1 (1999) 1. [24] A.J. Bard, L.R. Faulkner, Electrochemical Methods: Fundamentals and Applications, Wiley, New York, 2001, p. 612. [25] A.V. Tripković, K.Dj. Popović, J.D. Lović, J. Serb. Chem. Soc. 68 (2003) 849. [26] F. Gloaguen, J.M. Léger, C. Lamy, A. Marmann, U. Stimming, R. Vogel, Electrochim. Acta 44 (1999) 1805. [27] A.J. Bard, L.R. Faulkner, Electrochemical Methods: Fundamentals and Applications, Wiley, New York, 2001, p. 495. [28] T. Kawaguchi, W. Sugimoto, Y. Murakami, Y. Takasu, Electrochem. Commun. 6 (2004) 480. [29] M.E. Gamboa-Aldeco, E. Herrero, P.S. Zelenay, A. Wieckowski, J. Electroanal. Chem. 348 (1993) 451. [30] M.C. Santos, D.W. Miwa, S.A.S. Machado, Electrochem. Commun. 2 (2000) 692. [31] N. Markovic, P.N. Ross, J. Electroanal. Chem. 330 (1992) 499. [32] N. Markovic, P.N. Ross, Langmuir 9 (1993) 580. [33] S.C. Chang, Y. Ho, M.J. Weaver, Surf. Sci. 265 (1992) 81. [34] R. Wang, D. Johnson, G.R. Luecke, J. Electrochem. Soc. 131 (1984) 2369. [35] E. Gileadi, Electrode Kinetics, VCH, New York, 1993. [36] P.W. Atkins, Physical Chemistry, fourth ed., Oxford University Press, Oxford, 1990, p. 951.