Oxygen evolution reaction on electrochemically generated thick cobalt oxide films

Mcctrochimica Aeta. Vol. 35, No. 2, pp. 437443, Printed in Great Bdam

0013~686/“XI 63.00 +O.oO Pergamon Press pk.

1990

OXYGEN EVOLUTION REACTION ON ELECTROCHEMICALLY GENERATED THICK COBALT OXIDE FILMS M. R. GENNERO Programa

DE

CHIALVOand A. C. CHIALVO

de Electroquimica Aplicada e Ingenieria Electroquimica (PRELINE), Facultad Quimica (UNL), Santiago del Ester0 2829, (3000) Santa Fe, Argentina (Received

17 January 1989; in revisedfirm

de Ingenieria

8 March 1989)

Abstract-The electrocatalytic characteristics for the oxygen evolution reaction in alkaline solutions of cobalt oxide films obtained by periodic potential treatments were studied. The influence of ageing processes and thermal treatments on the reaction kinetics were evaluated. In addition, the real exchange current was determined through the analysis of the dependence of the apparent electrocatalytic activity on the oxide charge.

INTRODUCTION A major goal in developing advanced electrolyser technology for high purity hydrogen production is to minimise the energy requirement of the electrolytic process. The oxygen evolution reaction (OER) is of special interest because of its high anodic overvoltage[lJ. To date, results using metal oxide electrocatalysts have generally shown more promise than those obtained with modifications of the structure of traditional nickel electrodes[2]. The spineltype Co,O, has shown promising efficiency and longterm performance[3]. However, published works of OER on this kind of electrode is rather scarce[4-81. They were prepared by a thermal decomposition method or by a freeze drying-vacuum decomposition technique and then incorporated or not into a TFEbonded coating. Other methods to obtain metal oxide electrocatalysts, mainly used for nickel, consist in the preparation of thick oxide films. They can be obtained through chemical or electrochemical precipitation of the metal hydroxide or by the direct anodic oxidation through the application of current or potential perturbations to the base metal[9-151. Recently, a procedure has been developed to produce the macroscopic growth of cobalt oxide films in alkaline solutions by the application of a repetitive square-wave potential signal (RSWPS) to a cobalt electrode. Through this method, reproducible cobalt oxide electrodes involving a relatively large amount of active material can be obtained. The present work deals with the study of electrocatalytic characteristics of the films obtained by means of the RSWPS technique and the influence of ageing processes on the reaction kinetics for the OER in alkaline solutions.

counter electrode was built with a platinum helicoid (10 cm2 geometric area) and concentrically mounted with the working electrode. Potentials were measured against a Hg/HgO reference electrode in the same electrolyte solution and are given with respect to the reversible oxygen electrode (roe) in the same solution. Special attention was paid to the construction and location of the Luggin-Haber tip, so as to ensure negligible ohmic drops in solution. Experiments were carried out in 1 M and 2 M solutions, which were prepared with triply distilled water. All measurements were made at 30 + O.l”C, under nitrogen atmosphere. The OER study consisted mainly in potentiostatic runs and slow potentiodynamic sweeps ( 10m4 V s-l). The values of the current density, apparent current density (Sp) and oxide charge (Q) were referred to the geometrical electrode area. Preparation

of cobalt oxide electrodes

Cobalt oxide films were obtained by the application of the RSWPS technique to the working electrode. This treatment was similar to that employed previously to develop thick oxide layers on the corresponding base metals[ 16,173. In the case of cobalt, the optimal values of the square-wave parameters were determined earlier[l8]. According to it, the following values were used: lower potential limit E, = - 1.8 V; upper potential limit E, = 0.5 V (both referred to Hg/HgO); frequency 100 Hz. Electrodes with different amounts of cobalt oxide were prepared, varying the duration of the RSWPS treatment (0.5 min < t < 6 min). The amount of cobalt oxide was determined through the charge involved in the cathodic halfcycle of the stabilized voltammogram obtained by a repetitive potentiodynamic sweep run at 0.1 V s- ’ between -0.6 and 0.4 V. The electrode charge follows a reasonably linear dependence with the time of application of the RSWPS.

EXPERIMENTAL Runs were made in a three electrode, two compartment Pyrex glass cell. Cobalt wires (0.5-0.8 cm2 geometric area) were used as working electrodes. The

Electrode

pretreatments

Three types of active materials were used in the OER study.

438

M.R. GENNERODECHIALVO

Fresh electrodes: the experiments were performed immediately after the attainment of the cobalt oxide film, without any previous voltammetric cycle. (ii) Aged electrodes: this type was obtained by subjecting the fresh electrode to long-term potentiostatic runs in the potential region where the OER takes place (charge passed 2 lo4 C cme2) or by leaving the fresh electrode in solution during a certain period of time. (iii) Thermally treated electrodes: this type was obtained by applying a thermal treatment to the fresh electrode. It was carried out in a quartz furnace at a selected temperature in the range 200-5OO”C, under air atmosphere. (i)

X-ray difiaction

analysis

The electrodes were characterised by X-ray analysis. It was carried out on a Rich-Seifert diffractometer, employing Ni filtered CuK, radiation (35 kV) at a scan rate of 0.6” min- ‘. Two kind of samples were used, cobalt oxide powders and cobalt sheets covered by an oxide overlayer. This analysis was complemented by SEM observations. RESULTS Tafel plots for the OER on fresh

films

Immediately after the preparation of the electrode through the application of the RSWPS technique, it was subjected to a potentiodynamic sweep at 10T4 V s- ’ between 0.2 and 0.6 V. The variation of the logarithm of the current density with the overpotential (log i-q) is depicted in Fig. 1 for electrodes covered with different amounts of cobalt oxide, in 1M NaOH solutions. Curve a shows the behaviour of a bright cobalt electrode. It can be observed that the log i-q relationship is approximately linear in a large range of potentials. The Tafel plots corresponding to

and A.C.

CHIALVO

the cobalt oxide electrodes (curves b-f) show a re-

markable increase in the current density as the oxide charge increases, particularly at low overpotentials. For 9 >0.4 V, the current density increases to a lower extent with increasing values of both the potential and the oxide charge. it can be observed that the Tafel slope (b = dq/dlog i) in the low overpotential region takes an approximately constant value for the different charges, which is a little smaller than that of the bright cobalt electrode. Besides, the relation between the current density corresponding to the highest oxide charge value and that of the bright cobalt is about 160. In the low overpotential region (tfeO.35 V), the current density varies linearly with the oxide charge at constant overpotential (Fig. 2). On the other hand, in the high overpotential region the current density varies slightly with the oxide charge. Similar results were obtained in 2 M NaOH solutions. Tafel plots for the OER on agedjlms

When the cobalt oxide electrodes were left in the electrolyte solution for a given period of time, the log i-q relationship obtained by subjecting the electrode to a slow potentiodynamic sweep in the OER potential region shows differences with regard to the response of a fresh electrode. Figure 3 shows the Tafel plots obtained in 2 M NaOH solution corresponding to a cobalt oxide film (Q = 137.5 mC cmM2) immediately after its preparation (curve a) and after six days (curve b). A marked decrease in the current density can be observed in the low over-potential region. On the other hand, the Tafel slope shows a lower value for the same overpotentials (Table 1). Due to this fact, the current density takes approximately the same values as that of the fresh electrode when the potential is higher than 0.4 V. This electrode was then subjected to several potentiostatic runs at different potentials corresponding to the OER potential region, each one having a duration of at least 50 h. After each potentiostatisation, a slow potentiodynamic sweep run at 10e4 V s- ’ was applied to the electrode, which shows differences as a function of the overpotential. After holding the electrode at q = 0.45 V, the Tafel plot shows a similar shape as the fresh electrode, but lower

1

15

Fig. 1. Apparent Tafel plots for the OER on fresh cobalt oxide electrodes from a potentiodynamic sweep run at 10e4 Vs-’ in 1 M NaOH, with different charges: (a) 1.8 (blank) (b) 36.2; (c) 46.5; (d) 153.6; (e) 237.3; (f) 629.7 mC cmWz.

Fig. 2. Dependence of the current density upon the oxide charge for different overpotential values (indicated in the

figure).

Oxygen evolution reaction on cobalt oxide films

439

20 77

IE 0

z ‘0 L;

H 00

3:[q

I

-10 I

i:

I

I

030

0.40

050 E,”

,

0.60

Fig. 3. Apparent Tafel plots for the OER on aged cobalt oxide electrodes from a potentiodynamic sweep run at lo-“ V s- 1 (full lines) and from potentiostatic runs of 50 h (dashed line) in 2 M NaOH. Charge = 137.5 mC cm-z. (a) Fresh electrode; (b) after 6 days in solution; (c) after polarization at 1 = 0.45 V during 50 h; (d) after polarization at q = 0.275 V for 50 h.

values of the current density (Fig. 3 curve c). After holding the electrode at q = 0.275 V, a surprisingly low value for the Tafel slope in the overpotential range from 0.25 to 0.32 V (b = 0.033 V dec- ‘, Table 1) is obtained, together with current densities higher than that of the fresh electrode (Fig. 3, curve d).

0

20

6t

0.03OV dec-’ and O.l20Vdec-’ overpotentials, respectively.

at high and low

Tafel plots for the OER on thermally treated electrodes

A cobalt oxide electrode was subjected to thermal treatment at 480°C during 2 h. Figure 5 (curve c)

r

A cobalt oxide electrode (Q = 137 mC cm-‘) was subjected to a constant potential in the OER potential region during approximately 50 h. This experience was repeated for several overpotential values, measuring the variation of the current with time (Fig. 4). For the higher overpotentials (0.45 and 0.50 V), the current decreases continuously with time, rapidly for the first 20 h and then more slowly. At 0.40 V the current decays very fast for the first 2 h and then takes an almost constant value. At 0.36 V the current diminishes during the first hour and then increases continuously. At 0.30 V, the current shows a continuous and small increase. The current values corresponding to 50 h of potentiostatisation were used to draw the stationary Tafel plot (Fig. 3, dashed line). Two overpotential regions clearly can be observed. The Tafel slopes are

0.0

-1.0

Fig. 5. Apparent Tafel plots for the OER on a cobalt oxide electrode from a potentiodynamic sweep run at 10m4V s-r in I M NaOH: (a) blank; (b) fresh film (Q = 311 mC cm-‘); (c) after thermal treatment at 480°C for 2 h.

Table 1.

Fresh electrode Six days in Solution After pot. at ‘I = 0.45 v After pot. at q = 0.275 V

t/h

Fig. 4. Current-time relationships at constant overpotential. 2 M NaOH; Q = 137.5 mC cm-‘.

Tafel plots for the OER from potentiostatic runs

Electrode Pretreatment

40

Range/V

b/Vdec-’

0.28-0.35 0.28-0.35

0.055 0.038

6.33 x 1O-9 6.53 x IO- I2

0.28-0.35

0.056

3.12 x 1O-9

0.25X1.32

0.033

3.31 x lo-r2

$‘/A cm - z

440

M. R. GENNERO DE CHIALVOand A. C. CHIALVO

shows the response of this electrode to a slow potentiodynamic sweep from 0.25 to 0.60 V. For comparison, the log i-q plots of the blank (Fig Scurve a) and of the oxide electrode immediately after its prep-

aration are also shown (Q = 311 mC cme2; Fig. 5, curve b). A marked decrease of the current density after the thermal treatment can be observed. While the Tafel slope remains approximately the same. Voltammetric characteristics of the cobalt oxide electrodes after the OER Immediately after the application of the slow potentiodynamic sweeps from 0.25 to 0.60 V to study the

OER, the electrodes were subjected to a voltammetric scan in the cathodic direction run at 0.1 V s - 1.in order to evaluate the charge and to reveal possible changes on the cobalt oxide film (Fig. 6). Two broad electroreduction current peaks can be distinguished. The following anodic scan exhibits the corresponding electrooxidation current peaks. In the case of the fresh electrode (Fig. 6, curve a), the next cathodic scan shows a marked decrease of the voltammetric charge, which is approximately a 65% of that of the first scan. Furthermore, the most positive electroreduction current peak shows a displacement towards more anodic potentials. The voltammetric shape remains practically unchanged during successive potential cycles. The voltammetric response of an electrode aged in solution for 6 days is similar to that of the fresh electrode, although the peaks are better defined, particularly the anodic peak located at ca 0.29 V near the OER threshold potential. Figure 6 (curve b) shows the voltammogram corresponding to the electrode subjected to long-term potentiostatiz-

ation at n = 0.45 V (high overpotential region) and then to a slow potentiodynamic sweep. Three differences can be distinguished in the voltammogram as referred to that of the fresh one: the more anodic value of the threshold potential for the OER, the increase of the anodic current peak located at ca -0.10 V, which is more well-defined and narrower and the appearance of a new cathodic peak located at ca -0.25 V. The voltammogram corresponding to the electrode potentiostatized 50 h at n = 0.27 V (low overpotential region), is similar both in shape and charge involved to that of the fresh electrode (Figure 6, curve c). Nevertheless, the threshold potential for the OER takes a more positive value and it remains constant for the following cycles. Besides, there is a very small difference between the charges of the first scan and the stabilized profile. It should be noticed that the electrode charge remains approximately unchanged through all these long-term treatments. Oxide electrodes subjected to a thermal treatment at 480°C during 2 h show a voltammetric shape which was quite different to that of the untreated electrodes (Fig. 7). Only one redox couple can be observed, which shows a highly reversible behaviour. Besides, the voltammetric charge is approximately 8% of that of the electrode before the thermal treatment. This fact should be related to the decrease in the current density in the Tafel plot (Fig. 5). On the other hand, practically no changes occurred in the potentiodynamic shape due to the ageing process of the oxide film. X-ray diffraction patterns

Figure 8 shows the X-ray diffraction patterns of three different samples of cobalt oxide obtained by the

30

0

-30 30

Fig. 7. Potentiodynamic profile of a thermally treated oxide electrode run at 0.1 V s-r in I M NaOH.

': :: 0

30

b

1

0 -06

-03

00

E/V O3

Fig. 6. Potentiodynamic profiles of cobalt oxide electrodes run at 0.1 V s- ’ in 2 M NaOH after a slow potentiodynamic sweep between 0.25 V and 0.60 V: (a) fresh film; (b) after potentiostatization at 0.45 V during 50 h; (c) after potentiostatization at 0.275 V for 50 h. Dashed line = first cycle; full line = stabilised profile.

I

I

20

1

I

4o

28

60

I

Fig. 8. X-ray diffractograms: (a) aged cobalt oxide powder; (b) cobalt oxide electrode after potentiostatization at 0.45 V for I7 h; (c) thermally treated cobalt oxide powder.

Oxygen evolution reaction on cobalt oxide films RSWPS treatment. Figure 8a corresponds to a dehydrated powder aged in 2 M NaOH solution for 6 days, without any electrochemical treatment. Figure 8b corresponds to a cobalt electrode covered by a thick oxide layer, after OER at 0.45 V during 17 h and Fig. 8c belongs to an oxide powder thermally treated at 480°C for 2 h.

DISCUSSION Cobalt electrodes covered by a cobalt oxide film obtained through the application of the RSWPS technique, were evaluated as electrocatalysts for the oxygen evolution reaction in alkaline solutions. The electrocatalytic activity of this kind of electrodes is characterized by the following experimental results.

(9 A remarkable dependence on the film pretreatment, electrochemical or physical, which produces marked differences in the apparent Tafel slopes in the low overpotential region (0.03 < b < 0.06 V dec- ‘). Besides, in the high overpotential region, the electrode does not follow a Tafel behaviour when it is analysed by slow potentiodynamic sweeps. (ii) Two well-defined Tafel regions with the following slopes: 0.03 V dec-’ for q < 0.3 V and 0.12 V dec-’ for q > 0.3 V were obtained when potentiostatic runs of 50 h were applied to the oxide electrode. (iii) The linear dependence of the current density at constant overpotential upon the amount of electrochemically detected oxide present in the film. These facts show the existence of two phenomena originated by different causes. On the one hand, there is a slow ageing process of the hydrated cobalt oxide film. This is related to the film growth process, which consists of the periodic precipitation of hydroxylated species of cobalt in the metalLoxide interface. During the growth process, periodic variations of the species flow, charges, pH, elc are produced, generating an amorphous film with a high water content. This film cannot be considered in an internal equilibrium state. Consequently, when the application of the RSWPS is interrupted, a slow process of microscopic restructuring of the hydrated oxide starts. These phenomena can be appreciated when the electrochemical response of a fresh electrode is compared with that of an electrode left in the electrolyte solution during six days. It can be observed that the stabilised electrode exhibits better defined peaks in the voltammetric response. A similar behaviour was observed in electrodes stabilized by long-term potentiostatizations in the low overpotential region (TV Q 0.3 V). Nevertheless, the most noticeable change was observed in the Tafel slopes corresponding to the low overpotential region, which varies from approximately 0.055 V dec-’ for fresh electrodes to approximately 0.030 V dec- * for stabilised electrodes. The response is almost the same for the different stabilisation processes, open-circuit during several days or long-term potentiostatisations in the low overpotential region. Besides, it can be considered that a slow potentiodynamic sweep in the

441

OER potential region (0.25 < rl < 0.60 V) allows evaluation of the present electrocatalytic activity, since the sweep duration is much more lesser than the duration of the stabilization phenomena. The variation of the Tafel slopes should be attributed to changes in the adsorption energy of the reaction intermediate (OH, 0, etc), due to structural modifications (increase in the size of the microcrystals, etc.) of the oxide during the stabilisation process. These statements are related with the fact that a stabilized oxide shows an incipient state of crystallinity. Rather diffuse lines are observed at 20= 19.1, 20.3, 38.0, 38.9 and 65.6” in the X-ray diffraction pattern (Figs 8a and b). They can be assigned to the cobaltous hydroxide (Co(OH),) and to the cobaltic hydroxide (CoH,O)[19]. Other transformation phenomena of the oxide film are observed when the electrode is polarised at overpotentials greater than 0.30 V during long periods of time (more than 50 h). The potentiostatic runs define in this overpotential value a break in the Tafel slope, which takes the value 0.12 V dec- ‘. This sharp change in the electrochemical response of the cobalt electrodes covered by a hydrated oxide film should be caused by the slow transformation of the outer layers of this film into a hydroxylated form of Co(IV). This compound may produce a change in the reaction mechanism. In order to clarify this concepts, the following reaction can be written: CoOOH+OH-eCoO,+H,O+e-. From the Pourbaix diagram[20], for the experimental conditions, an equilibrium potential of approximately 0.24 V is obtained, which is slightly smaller than the experimentally obtained value (q = 0.30 V). It must be pointed out that this reaction involves compounds with a defined stoichiometry, which is not in agreement with the actual behaviour of the system. Other phenomena that give evidences in favour of a change in the chemical nature of the oxide surface are as follows. (i) The behaviour of an electrode stabilized at q < 0.30 V. When it is subjected to a slow potentiodynamic sweep in the OER potential region starting from this overpotential, it reproduces the response of the potentiostatic runs up to approximately 0.30 V. At higher overpotential values, the Tafel slope increases continuously, although the electrocatalytic activity is greater than that of the steady-state (Fig. 3, curve a and dashed line). carried out with Mossbauer (ii) Studies spectroscopy[21, 221 have demonstrated the existence of Co(IV) species, together with Co(II1) species, when a cobalt electrode is potentiostatised at high overpotentials, the relationship Co(IV)/Co(HI) becoming greater as the polarisation potential increased. On the other hand, the hydrated oxide film obtained by the RSWPS technique is constituted by non-stoichiometric complex, compounds which possibly included OH- ligands and other ions present in the electrolyte solution. Taking into account these facts, it should be considered that the oxidation of this film gives a tetravalent cobalt oxide with similar morphological and

M. R. GENNERO

442

DE CHIALVO

structural characteristics. Therefore, it seems unlikely this compound is related to COO,, but to a hydroxylated form which should have a structure similar to a cobaltite IV (CoO,Na,). This statement is supported by the result obtained when the Co(OH), is plriced in alkaline media in the presence of oxygen or other oxidant such as hydrogen peroxide, which gives a cobaltite IV compound[23]. (iii) When an electrode is subjected to a long-term polarisation at q 2 0.30 V and then to a potentiodynamic sweep towards cathodic potentials, its voltammetric response exhibits a new cathodic peak located at ca -0.25 V. Besides, the next positive potential scan shows a marked increase of the anodic current peak located at ca -0.10 V. The charge involved in this couple increases slowly with the polarisation time. It is likely that this couple is related to the phenomenon which takes place at q 2 0.3 V, although the experimental evidence is not enough to explain convincingly the processes carried out at such potentials. Finally, the other experimental fact that takes advantage of the important role of the hydroxylated species present in the film is the electrochemical behaviour of the cobalt spinel. It has been obtained by thermal treatment X-ray diffraction

of the hydrated oxide film and its spectrum (Fig. 8c) coincides with

those described in the literature for the cobalt spine1 structure. Its voltammetric response exhibits only one cathodic current peak located at ca 0.21 V and the corresponding anodic current peak at ca 0.24 V (Fig. 7). As it has an anhydrous volumetric structure, the only process that take place on its surface are the electrosorption and electrodesorption of OH-[24]. Consequently, it can be immediately inferred that in the hydrated oxide film, the current peaks located between -0.3 and 0.0 V are related to oxidationreduction processes of hydroxylated cobalt species. Current density vs charge dependence of fresh jilms

Figure 1 shows that, at constant q values corresponding to the low overpotential region, the Tafel lines move towards greater current densities as the oxide charge increases. This behaviour should be related to an increase in the roughness factor of the oxide-solution interface, where the OER takes place. The displacement magnitude of the Tafel lines can be related to the oxide charge [see Appendix, equation

W)l:

log(Pp/Q) = q/b +logi,k

The validity of this expression can be verified through the experimental data. Figure 9a shows the log i”P/Q us q plot, for the values obtained from Fig. 2. The slope of

this straight line is equal to 0.054 V dec- ‘. On the other hand, the equation (A4) must be verified: I:~= io+ki,Q. Figure 9b shows the itp us Q plot. The slope this straight

line gives the value of the product

of (ki,).

This value was equal to that obtained from the Fig. 9a, for q = 0. On the other hand, the real i,

and A. C. CHIALVO

0.28 _

032

v0.36

39 -

Fig. 9. (a) log iaP/Qvs q plot, values obtained from Fig. 2; (b) itp us Q plot.

value can be obtained (i,=3x10-7mAcm-2).

from Fig. 9b, for Q =0

Acknowledgements-This work was supported by the Consejo National de Investigaciones Cientificas y T&enieas (CONICET-Argentina) and the Stiftung Volkswagenwerk, Federal Republic of Germany.

REFERENCES 1. D. E. Hall, J. electrochem. Sot. 132,41 (1985). 2. S. Trasatti and G. Lodi, in Electrodes of Conductive Metallic Oxides. Parr B (Edited by S. Trasatti), pp. 521-626, Elsevier, Amsterdam (1980). 3. M. R. Tarasevich and B. N. Efremov, in Electrodes of Conductive Metallic Oxides. Part A (Edited by S. Trasatti), pp. 221-259, Elsevier, Amsterdam (1980). 4. H. Vandenborre, R. Leysen, H. Nackarts and Ph. Van Asbroeck, Advances Hydrogen Energy 3, Hydrogen Energy Progress IV, p. 107, Pergamon Press, Oxford (1982). 5. H. Willems, A. G. C. Kobussen, J. H. W. de Wit and G. H. J. Broers, J. electroanal. Gem. 170, 227 (1984). 6. H. Willems, A. G. C. Kobussen, J. C. Vinke, J. H. W. de Wit and G. H. J. Broers, J. electroanal. Chem. 194, 287 (1985). 7. S. M. Jasem and A. C. C. Tseung, J. electrochem. Sot. 126, 1353 (1979). 8. C. Iwakura, A. Honji and H. Tamura, Electrochim. Acta 26, 1319 (1981). 9. P. Benson, G. W. D. Briggs and W. F. K. WynneJones, Electrochim. Acta 9, 275 (1964).

10. H. Gomez Meier, J. R. Vilche and A. J. Arvia, J. electroanal. Chem..138, 367 (1982). 11. S. I. C6rdoba. R. E. Carbonio. M. Lboez Teiielo and V. A. Macagnb, Electrochim. Acta 31, lj21 (1986). 12. M. Paszkiewicz, J. appl. Electrochem. 11,443 (1981). 13. L. D. Burke and T. A. M. Twomey, J. electroanal. Chem. 162, 101 (1984). 14. C. K. Dyer, J. electrochem. Sot. 132, 13 (1985). 15. A. Visintin, A. C. Chialvo, W. E. Triaea and A. J. Arvia, J. electroanal. Chem. 225, 227 (1987). 16. A. C. Chialvo, W. E. Triaea and A. J. Arvia, J. electroanal. Chem. 146, 93 (1983). 17. A. Visintin, A. C. Chialvo, W. E. Triaca and A. J. Arvia, J. electroanal. Chem. 225, 227 (1987). 18. T. Kessler, A. Visintin, M. R. Gennero de Chialvo, W. E. Triaca, and A. J. Arvia, J. e[ecrroanal. Chem., in press.

Oxygen evolution reaction on cobalt oxide films 19. Kondrashev and Fedorova, Dokl. Akad. Nauk. 94, 229 (1954). 20. M. Pourbaix, Atlas of Electrochemical Equilibria in Aqueous Solutions, p. 322, Pergamon Press, Oxford (1966). 21. G. W. Simmons, E. Kellerman and H. Geidheiser, J. e/ectrochem. Sot. 123, 1276 (1976). 22. G. W. Simmons, A. Vertes, M. L. Varsanyi and H. Leidheiser, J. electrochem. Sot. 126, 187 (1979). 23. J. Amiel in Nouveau Traite de Chimie Minerale, Vol. XVIII, (Edited by P. Pascal), p. 319 Masson, Paris (1963). 24. B. N. Efremov, G. I. Zakhaarkin, S. R. Zhukov and M. R. Tarasevich, Sov. Electrochem. 14, 805 (1978).

443

The roughness factor can be written as follows: A (geom)

i,

i

642)

A linear relationship between f and the oxide charge Q is proposed: f=l+k(Q-Q”)-l+kQ;

Q%Q’,

(A3)

where Q” is the charge of the monolayer. Substituting in equation (A2) and rearranging: itp = i, + ki,Q,

(A4)

and: izp= ki,Q;

APPENDIX

kQ $ 1.

(AS)

Substituting equation (AS) in (Al) and operating: The Tafel equation is: q + b log itp = b log i’p.

log (P/Q) = q/b + log i, k.

641)

(A6)