Journal of Electroanalytical Chemistry 644 (2010) 144149

Contents lists available at ScienceDirect

Journal of Electroanalytical Chemistry

journal homepage: www.elsevier.com/locate/jelechem

Calculations of the exchange current density for hydrogen electrode reactions: A short review and a new equation
Daoping Tang, Juntao Lu *, Lin Zhuang *, Peifang Liu
College of Chemistry and Molecular Sciences, Hubei Key Lab of Electrochemical Power Sources, Wuhan University, Wuhan 430072, China

a r t i c l e

i n f o

a b s t r a c t
After reviewing relevant equations for the calculation of exchange current density, a new equation is derived for hydrogen electrode reactions to correct for the inuences of the hydrogen concentration change in the vicinity to the electrode surface. This equation is able to describe the polarization curve shape in the small polarization region as well as to calculate the exchange current (density). The abilities of this equation are demonstrated by the data obtained with a Pt rotating disk electrode in 0.1 mol l1 KOH solution. The exchange current density at 298 K under 1 atmosphere hydrogen pressure is found to be 0.103 mA cm2 with an apparent activation energy of 33.5 kJ mol1. At a constant temperature, the exchange current is found to be proportional to the square root of the hydrogen partial pressure in the solution. 2009 Published by Elsevier B.V.

Article history: Received 15 October 2009 Received in revised form 19 November 2009 Accepted 30 November 2009 Available online 16 December 2009 Keywords: Exchange current density Hydrogen electrode reactions Platinum Concentration polarization Fuel cells

1. Introduction The hydrogen electrode reactions, including hydrogen oxidation reaction (HOR) and hydrogen evolution reaction (HER), are among the most important reactions in electrochemistry and the exchange current (density), i0, has been widely used as a parameter to evaluate the relevant catalysts in technical applications, such as fuel cells and water electrolyses. As well-known, the solubility of dihydrogen in aqueous solutions is rather low and this makes experimental determination of i0 difcult for catalysts highly active to the hydrogen reactions. In order to obtain i0 as a characteristic for a given electrochemical system, either the Tafel plot in large overpotential region or the linear equation in small overpotential region may be used, in principle. However, for hydrogen reaction catalysts, it is better to deduce i0 from low overpotential measurements. This is because the catalyst surface may change with potential and the parameter extrapolated from the large overpotential region may not reect the real situation round the reversible hydrogen electrode potential (RHE). On the other hand, the real overpotential of a technical hydrogen electrode is usually not high (no more than one-tenth of a volt), the i0 obtained from measurements round the RHE is certainly more applicable than those obtained by extrapolating from large overpotential mea-

* Corresponding authors. Tel.: +86 27 87262592; fax: +86 27 87162672. E-mail addresses: jtlu@whu.edu.cn (J. Lu), lzhuang@whu.edu.cn (L. Zhuang). 1572-6657/$ - see front matter 2009 Published by Elsevier B.V. doi:10.1016/j.jelechem.2009.11.031

surements. Therefore, we shall concentrate on the equations for small overpotential in this paper. Pt has been the best catalyst for the hydrogen reactions. In the early stage of electrochemistry, the reported i0 values for the hydrogen reactions at Pt electrodes were obtained without the correction for the concentration polarization [112]. Since 1970s, it has been realized that the concentration overpotential is not negligible even at low overpotentials for Pt. For catalysts other than Pt, the same problem appears as long as the catalytic activity becomes high. Therefore, it is necessary to work out an equation which is able to correct properly for the inuences of the hydrogen concentration changes in the vicinity of the electrode surface. There have been a few such equations found in the literature, but they were somewhat different from each other. In this paper, we shall rst review these equations and compare them in two aspects, i.e., the deduction of i0 and the description of the polarization curve shape in low overpotential region; and then we shall derivate a new equation, which is of higher accuracy in polarization curve description and yet simple in form, and verify it using experimental data obtained with a Pt rotating disk electrode (RDE) in KOH solutions. The reason for choosing KOH solution, rather than H2SO4 solution, in the present work was that the hydrogen electrode reactions in alkaline media are kinetically slower than in acidic media, so that we can readily verify the ability of the equations for i0 calculation. For a Pt RDE in acidic solutions, the exchange current is too large to be determined by conventionally RDE method but is accessible for carbon-diluted Pt/C using the equations derived in this paper [13].

D. Tang et al. / Journal of Electroanalytical Chemistry 644 (2010) 144149

145

2. Experimental The data of hydrogen reactions at Pt electrodes in alkaline solutions were obtained by carrying out electrochemical measurements with a single compartment cell which was equipped with a water jacket for temperature control. The working electrode was a Pt RDE (Autolab, Mode ECO CHEMIE BV), and the counter electrode was a Pt wire. The reference electrode was an RHE [14] in the same solution as that in the cell and was connected to the cell through a Luggin capillary. All solutions were prepared using carefully puried water, which was obtained by distillation of the 18 MX cm deionized water with KMnO4 to remove traces of organics. Measurements were conducted on an electrochemical analyzer (CHI 634, Shanghai). Before electrochemical measurements, the Pt electrode was polished with alumina powers successively from 0.3 to 0.05 lm and then carefully washed with acetone and deionized water. The polished Pt electrode was pretreated and characterized by cyclic voltammetry (CV) in 0.5 mol l1 H2SO4 purged with high purity N2 gas (99.999%) till the standard CV was obtained. Then the Pt electrode was washed with water thoroughly and transferred into the cell containing 0.1 mol l1 KOH solution that was saturated by high purity hydrogen. The electrode was conditioned again by CV over 0.051.3 V (versus RHE), till reproducible CV curves were obtained. Immediately after surface conditioning, a slow potential scan was performed from 0.05 to 0.45 V to obtained a steady state polarization curve. Fig. 1 shows a set of typical polarization curves acquired at different rotation rates. On the cathodic branch (negative to RHE), the hydrogen evolution current increases quickly with potential moving to the negative. On the anodic branch (positive to RHE), with the increase of potential, the hydrogen oxidation current increases rst and then levels off. A careful inspection into Fig. 1 would nd that the plateaus are not really at. When the pseudo-plateau currents taken at 0.3 V were plotted against the square root of rotating rates, a slightly curved line resulted as shown by the solid line in Fig. 2. It can be seen that the pseudo-plateau currents deviated from the expected proportionality (the dashed line in Fig. 2) for high rotating rates, indicating an inuence of kinetic limitation. The kinetic limitation would be seen more clearly at higher rotating rates. For example, at 6400 rpm, pseudo-plateau current was found to be about 20% than the extrapolated (not shown in Fig. 2). In order to reduce the inuence of the kinetic limitation, we concentrated on relatively low rotating rates and made corrections by extrapolating from the initial slope as shown in Fig. 2 where the corrected mass-transport limiting currents are marked by the numbers close to the open diamonds.

Fig. 2. Pseudo-plateau currents (lled squares) and extrapolated mass-transport limiting currents (open diamonds) for different rotation rates.

Fig. 3 shows the plot of normalized current, i/ia,d, versus polarization in the small overpotential region; the deviation of the polarization curves from straight lines indicates the effects of concentration polarization. 3. Review of literature equations In this section, we shall rst present the general equations correcting for the inuences of surface concentration changes, followed by a few literature equations specically for the hydrogen reactions. In all the equations below, the anodic current (ia) is dened as positive current; and the difference between the working potential (E) and the equilibrium potential (Eeq) as the overpotential (g), i.e., g = E Eeq. According to these denitions, hydrogen oxidation and hydrogen evolution are accompanied by positive and negative overpotentials, respectively. 3.1. General equations 3.1.1. Concentration polarization correction for single-step oneelectron reactions [15] Concentration polarization is usually referred to the overpotential caused by the change of the concentrations in the vicinity of the electrode surface (surface concentration changes). In fact, the surface concentration change causes not only concentration

Fig. 1. Typical polarization curves of Pt RDE in hydrogen saturated 0.1 mol l1 KOH solution (100, 400, 600, 900, and 1200 rpm for curves from a to e, respectively, 298 K).

Fig. 3. Plots of normalized currents versus overpotential (data originated from Figs. 2 and 3).

146

D. Tang et al. / Journal of Electroanalytical Chemistry 644 (2010) 144149

overpotential but also the change of exchange current. However, for the sake of simplicity, we shall use the term concentration polarization for both effects in this paper. For a single-step one-electron reaction R = O + e, the wellknown ButlerVolmer equation applies

cated. However, for the small overpotential region, the following equation applies in general [15]:

i ni0 gF =RT

11

i FkfC R exp1 af E E0 C O expaf E E0 g

where E0 and k are the standard equilibrium potential and standard rate constant; CO and CR represent the concentrations of oxidized and reduced states, respectively; a is the transfer coefcient for the cathodic component reaction; and f = F/RT. At equilibrium potential Eeq, the absolute value of the anodic and cathodic component currents (the internal currents) are both equal to the exchange current and the net current is zero:

where n is the number of electrons involved in the total reaction. Eq. (11) is the basis for the determination of exchange current by small polarization measurements. For hydrogen oxidation, n = 2,

i 2i0 gF =RT

12

Unfortunately, there has been no general equation with correction for the change of surface concentration for multi-step multielectron reactions.

i0 ia;eq FkC R exp1 af Eeq E0 i0 ic;eq FkC O expaf Eeq E0

2a 2b
3.2. Concentration polarization correction for hydrogen electrode reactions: Harrison equation Harrison and Kham [16] treated the hydrogen electrode reaction as a 2-electron single-step reaction. Recognizing that only the concentration change of dissolved hydrogen needed correction, they omitted the term of i/ic,d and introduced a factor of 2 into the exponential term in Eq. (8). After rearrangement, they reached

Substituting the Nernst equation Eeq = E0 + (RT/F) ln(CO/CR) into either Eqs. 2a or 2b results in

i0 FkC R C O

a 1a

Eq. (3) reects the effects of concentrations on the exchange current. It can be recognized from the deduction that concentration change inuences the exchange current via two means, i.e., changing the concentrations and changing the reaction energy barrier due to the change of equilibrium potential. Substituting Eq. (3) into Eq. (1), we have

logfi=1 i=id exp2f gg log i0 21 af g=2:303

13

i i0 fexp1 af g expaf gg

Expanding the exponential function using the approximation exp(x) = 1 + x, Eq. (4) is reduced to the well-known linear equation for small overpotentials:

i i0 F g=RT

In the presence of concentration polarization, it can be deduced that

i i0 fC R;s =C R;b exp1 af g C O;s =C O;b expaf gg

where i0 is the exchange current corresponding to the concentrations in the bulk; Cb and Cs stand for the concentrations in the bulk and in the vicinity to the electrode surface, respectively. The surface concentration can be calculated from the ratio of the net current to the mass-transport limiting current:

C O;s =C O;b 1 i=ic;d C R;s =C R;b 1 i=ia;d

After substituting Eqs. (7a) and (7b), Eq. (6) becomes

7a 7b

According to Eq. (13), plotting logfi=1 i=i d exp2f gg against g will result in a straight line from which i0 and a can be deduced. Fig. 4 shows such plots obtained from the original data in Figs. 1 and 2. In small overpotential region, say 20 mV round RHE, the plots appear to be straight lines to a good approximation. The kinetic parameters deduced from Fig. 5 are given in Table 1. The data processing using Eq. (13) seems successful that the deduced i0 values are essentially independent of electrode rotation rates and the transfer coefcient is close to 0.5, which is a popular value in electrode kinetics. Although the assumption of 2-electron single-step reaction is certainly wrong, Eq. (13) turned out to work well in the determination of i0 and a and in the description of the polarization curve shape. A shortcoming of Eq. (13) is the noise amplication effect. The noise level round RHE in Fig. 4 is much higher than that in the original data (Fig. 3). This effect seems to stem from the fact that the denominator of the logarithm term in Eq. (13) is a small difference between two values very close to unity.

i=i0 1 i=ia;d exp1 af g 1 i=ic;d expaf g

For afg 1, Eq. (8) can be simplied by the Taylor expansion,

g 1=i0 1=ic;d 1=ia;d iRT =F

Note that ic,d is minus according to denition. Therefore, the measured overpotential g is a summation of electrochemical overpotential iRT/Fi0 and two concentration overpotentials (1/ic,d + 1/ ia,d) iRT/F. In usual acidic or alkaline solutions, the concentration polarization of the oxidized states are negligible, i.e., the absolute value of ic,d can be considered as innite. As a result, for the hydrogen electrode reactions, the reduced form applies:

g 1=i0 1=ia;d iRT =F

10

3.1.2. Multi-electron multi-step electrode reactions When the electrode reaction involves more than a single-step and more than single-electron, the kinetic equation will be compli-

Fig. 4. Plots of log{i/[(1 i/id) exp(2fg)]} versus g (data originated from Figs. 2 and 3).

D. Tang et al. / Journal of Electroanalytical Chemistry 644 (2010) 144149 Table 3 i0 values deduced from Eq. (15).

147

x (rpm)
i0 (mA)

100 0.0352

400 0.0337

600 0.0343

900 0.0342

1200 0.0347

This equation is identical to Eq. (10) except for the factor 2 added for 2-electron reactions. This equation is linear, not descriptive for the real polarization curve shape. However, from the slope of g versus i round RHE, reasonable i0 values can still be obtained (Table 3). 4. New equation for the concentration polarization correction for hydrogen electrode reactions 4.1. Derivation of the new equation
Fig. 5. Plots of g(1 i/ia,d) versus (i/ia,d) (data originated from Figs. 2 and 3).

3.3. Concentration polarization correction for hydrogen electrode reactions: Vogel equation Vogel and coworkers [17] assumed a TafelVolmer mechanism, i.e., the dihydrogen molecule is rst dissociated into two adsorbed hydrogen atoms and then oxidized to two protons. According to this mechanism, they proposed the following equation for the small overpotential region:

i=i0 C s =C b exp2F g=RT 1

14

Using approximation exp(x) = 1 + x (for x  1) and noting Cs/ Cb = 1 i/ia,d, this equation can be simplied to

2F gRT i=i0 i=ia;d =1 i=ia;d or g1 i=ia;d RT =2F ia;d =i0 1i=ia;d

14a 14b

Our purpose was to derive an equation which is not only applicable to the determination of i0 but also able to describe the polarization curve shape in small overpotential region for the hydrogen electrode reactions. The basic idea of our approach is to correct for both the change of effective equilibrium potential and the change of effective exchange current (density). We simply adopted Eq. (12) as our starting point and noted that the g and i0 in that equation are the effective electrochemical overpotential gec and the effective exchange current i0 , respectively. When the surface concentrations differ from the bulk concentrations, concentration overpotential gcn occurs and the measured overpotential is the sum of electrochemical and concentration overpotentials, g = gec + gcn. In the meantime, the effective exchange current dif fers from that for the bulk concentrations, i.e., i0 i0 . In order to emphasize the effective nature of the two variables, Eq. (12) is modied as Eq. (12a):

According to Eq. (14b), the plot of g(1 i/ia,d) versus (i/ia,d) should be a straight line passing through the origin and i0 can be calculated from the slope of the line. However, such plots for the original data of Figs. 1 and 2 did not appear to be straight lines as shown in Fig. 5. Nevertheless, the slopes round RHE were taken to calculate i0. As shown by Table 2, the i0 values obtained using Vogels approach are essentially the same as those in Table 1. 3.4. Concentration polarization correction for hydrogen electrode reactions: ChenKucernak equation Chen and Kucernak [18] studied the mechanism of hydrogen electrode reactions using Pt ultramicroelectrodes in sulfuric acid solutions. According to the mechanism proposed, they reported an equation for the small overpotential region:

i 2i0 gec F =RT

i0

12a

In the text below, we shall explain how to replace and gec with experimentally accessible variables to obtain the equation we wanted. The correction for the concentration overpotential can be performed by using the Nernst equation,

gcn RT =2F lnia;d =ia;d i

16

Substituting gec = g gcn into Eq. (12a) and using ln(1 + x) = x approximation result in Eq. (17):

g RTi=2F 1=i 0 1=ia;d

i0

17

g RTi=2F 1=i0 1=ia;d

15

Table 1 i0 and a deduced from Fig. 5.

x (rpm)
i0 (mA)

100 0.0327 0.42

400 0.0301 0.48

600 0.0304 0.49

900 0.0303 0.50

1200 0.0303 0.51

If the difference between i0 and is ignored, Eq. (17) is exactly identical to Eq. (15). It means that Eq. (15) corrects only the concentration overpotential and ignores the change in the effective exchange current. For single-step single-electron reactions, Eq. (3) describes the dependence of i0 on the concentrations. As the hydrogen electrode reactions are 2-electron multi-step reactions, Eq. (3) is not directly applicable. However, it can be seen from the derivation of Eq. (3) that i0 should still be proportional to the concentrations raised to certain powers. Because the concentration changes of the oxidized states can be omitted, we need only to consider the concentration change for reduced form (dihydrogen):

i0 1C c R
Table 2 i0 values deduced from Eq. (14a).

18

x (rpm)
ia,d (mA) i0 (mA)

100 0.0522 0.0348

400 0.105 0.0333

600 0.128 0.0335

900 0.157 0.0340

1200 0.181 0.0317

where c is a constant depending on the reaction mechanism which is beyond the scope of this work. According to Eq. (18), the effective exchange current can be expressed as:

i0 i0 1 i=ia;d c

19

148

D. Tang et al. / Journal of Electroanalytical Chemistry 644 (2010) 144149

Substituting Eq. (19) into Eq. (17) results in the equation we wanted:

g iRT =2F f1=i0 1 i=ia;d c 1=ia;d g

c

20

In comparison to Eq. (15), the term (1 i/ia,d) is introduced to correct for the change of effective exchange current due to surface concentration changes. The unknown parameter c can be determined by tting Eq. (20) with experimental data (ig curves) using known i0 values, or directly obtained from experimentally measured i0CH2 relationship (Eq. (18)). The relevant results will be given in the text below.

4.2. Demonstration and tests of the new equation As seen in Tables 13, the i0 values calculated by the three literature equations are essentially the same; one can take these values to determine c by tting Eq. (20) to the polarization curves. It was found that the best result of curve tting was obtained with c = 0.5 (Fig. 6). Fig. 6 also proves the ability of the new equation to describe the shape of polarization curve. Therefore, Eq. (20) can be implemented as:

Fig. 7. Plot of ln i0 versus ln pH for determination of c.

2F g=RT i=i0 1 i=ia;d 0:5 i=ia;d

21

We also measured c in a direct way, i.e., measuring the i0CH2 relationship (Eq. (18)). The polarization measurements were carried out under different hydrogen partial pressures and then the i0 and partial pressure were plotted in logarithmlogarithm plane (Fig. 7). From the slope c was found to be 0.497, in good agreement with the result of curve tting (0.5). Eq. (21) can be directly used for the calculation of i0. By plotting (2Fg /RT i/ia,d) against i/(1 i/ia,d)0.5, the i0 value can be obtained from the slope. Table 4 gives the data for different rotating rates. We have also determined i0 at different temperatures, and the results showed a good Arrhenius plot (Fig. 8) with an apparent activation energy of 33.5 kJ mol1. By normalizing to the electrochemical surface area, the exchange current density of hydrogen electrode reactions at Pt in 0.1 mol l1 KOH at 298 K can be obtained, which turns out to be 0.103 mA cm2 for the set of experimental data described above. According to the results obtained in repeated experiments, we report here the exchange current density to be 0.100.11 mA cm2. No data were found in literature for exactly the same experimental conditions. However, this result seems to be compatible with the literature data for similar conditions (Table 5).

Fig. 8. The Arrhenius plot of i0 for hydrogen electrode reactions at Pt in 0.1 mol l1 KOH solution.

5. Summary A new equation was derived for the hydrogen electrode reactions to correct for both the change in effective equilibrium potential and the change in effective exchange current (density). The

Table 4 i0 values deduced from Eq. (21).

x (rpm)
i0 (mA)

100 0.0353

400 0.0336

600 0.0346

900 0.0345

1200 0.0337

Table 5 Comparison between the reported i0 values and that of this work. Author Bockris [2] Harrison [16] Ernst [3] Tilak [5] [9] Markovic This work Fig. 6. Comparison of calculated (dots) and experimental (solid lines) polarization curves for different rotation rates with c = 0.5.
*

Medium (temp/K) 0.5 M NaOH (rt ) 1 M NaOH (rt*) 1 M KOH (298 K) 0.5 M NaOH (298 K) 0.1 M KOH (293 K) 0.1 M KOH (298 K)
*

i0 (mA cm2) 0.0871 0.233 1.05 0.31 (111): 0.035 (110): 0.3 0.100.11

rt = room temperature.

D. Tang et al. / Journal of Electroanalytical Chemistry 644 (2010) 144149

149

equation has demonstrated to be able to describe the polarization curve shape in the small overpotential region as well as to serve for the deduction of exchange current. From RDE measurements in 0.1 mol l1 KOH at 298 K, the exchange current density of hydrogen reactions at Pt was found to be 0.100.11 mA cm2 with an activation energy 33.5 kJ mol1. Moreover, the exchange current was found to be proportional to the square root of the partial pressure of dissolved hydrogen. Acknowledgements This work was nancially supported under the National Science Foundation of China ( 20933004, 20773096, 50632050, 20433060 and J0730426) and the National Hi-Tech R&D Program (2007AA05Z142). References
[1] J. OM. Bockris, I.A. Ammar, A.K.M.S. Huq, J. Phys. Chem. 61 (1957) 879886.

[2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18]

J. OM. Bockris, S. Srinivasan, Electrochim. Acta 9 (1964) 3144. S. Ernst, C.H. Hamann, J. Electroanal. Chem. 60 (1975) 97100. H. Kita, S. Ye, Y. Gao, J. Electroanal. Chem. 334 (1992) 351357. B.V. Tilak, C.P. Chen, J. Appl. Electrochem. 23 (1993) 631640. , S.T. Sarraf, H.A. Gasteiger, P.N. Ross, J. Chem. Soc. Faraday N.M. Markovic Trans. 92 (1996) 37193725. , B.N. Grgur, P.N. Ross, J. Phys. Chem. B 101 (1997) 54055413. N.M. Markovic J. Zhou, Y. Zu, A.J. Bard, J. Electroanal. Chem. 491 (2000) 2229. , J. Electroanal. Chem. 524525 (2002) T.J. Schmidt, P.N. Ross, N.M. Markovic 252260. C.G. Zoski, J. Phys. Chem. B 107 (2003) 64016405. F. Alcaide, E. Brillas, P.L. Cabot, J. Electrochem. Soc. 152 (2005) E319E327. A.F. Innocente, A.C.D. Angelo, J. Power Sources 162 (2006) 151159. Yubao Sun, Juntao Lu, Lin Zhuang, Electrochim. Acta doi:10.1016/ j.electacta.2009.09.047. S. Gong, J. Lu, H. Yan, J. Electroanal. Chem. 436 (1997) 291293. A.J. Bard, L.R. Faulkner, Electrochemical Methods, second ed., John Wiley & Sons, New York, 2001 (Chapter 3). J.A. Harrison, Z.A. Kham, J. Electroanal. Chem. 30 (1971) 327330. W. Vogel, J. Lundquist, P. Ross, P. Stonehart, Electrochim. Acta 20 (1975) 79 93. S. Chen, A. Kucernak, J. Phys. Chem. B 108 (2004) 1398413994.