Mechanism of the Chlorine Evolution on a Ruthenium Oxide/Titanium Oxide Electrode and on a Ruthenium Electrode

Home , Exchange current density, Overpotential

Mechanism of the chlorine evolution on a ruthenium oxide/titanium oxide electrode and on a ruthenium electrode

Citation for published version (APA): Janssen, L. J. J., Starmans, L. M. C., Visser, J. G., & Barendrecht, E. (1977). Mechanism of the chlorine evolution on a ruthenium oxide/titanium oxide electrode and on a ruthenium electrode. Electrochimica Acta, 22(10), 1093-1100. https://doi.org/10.1016/0013-4686(77)80045-5

DOI: 10.1016/0013-4686(77)80045-5

Document status and date: Published: 01/01/1977

Document Version: Publisher’s PDF, also known as Version of Record (includes final page, issue and volume numbers)

Please check the document version of this publication:

• A submitted manuscript is the version of the article upon submission and before peer-review. There can be important differences between the submitted version and the official published version of record. People interested in the research are advised to contact the author for the final version of the publication, or visit the DOI to the publisher's website. • The final author version and the galley proof are versions of the publication after peer review. • The final published version features the final layout of the paper including the volume, issue and page numbers. Link to publication

General rights Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.

• Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain • You may freely distribute the URL identifying the publication in the public portal.

If the publication is distributed under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license above, please follow below link for the End User Agreement: www.tue.nl/taverne

Take down policy If you believe that this document breaches copyright please contact us at: [email protected] providing details and we will investigate your claim.

Download date: 27. Sep. 2021 MECHANISM OF THE CHLORINE EVOLUTION ON A RUTHENIUM OXIDE/TITANIUM OXIDE ELECTRODE AND ON A RUTHENIUM ELECTRODE

L. J. J. JAN=,* L. M. C. ST.&S&W.,* J. G. Vrssmt and E. BARENDRECHT* * Department of Electrochemistry, Eindhoven University of Technology and t Computing Centre, Eindhoven University of Technology, Eindhoven, The Netherlands

(Receioed 28 September 1976, in final form 7 December 1976)

Abstract-The mechanism of the chlorine evolution on titanium electrodes coated with a layer of ruthenium oxide and titanium oxide under different experimental conditions, and on a ruthenium electrode, both in acidic chloride solution, has been investigated Potentiodynatnic current density- potential curves were recorded as a function of the time of anodic pre-polarisation, the composition of the solution and the temperature. Moreover, potential decay curves were determined. Theoretical potential decay curves were deduced for both the Tafel reaction (2C&,-rCls) and the Heyrovsky reaction (Cl- + Cl,,+ Cl, + e-) as a rate determining step in the formation of molecular chlorine. They were compared with those found experimentally. The intluence of possible diffusion of atomic chlorine out of the electrode was also taken into consideration. It was found that for all the electrodes investigated, molecular chlorine is formed both at anodic polarisation and on open circuit according to the Volmer-Heyrovsky mechanism, where the Heyrovsky reaction is the rate-determining step. The transfer coefficient is 0.5 for the chlorine evolution at an “ideal” ruthenium oxide titanium oxide electrode and at a ruthenium electrode.

NOMENCLATURE overpotential on open circuit; q. = E, - E, 11. at t, = 0 Tafel slope of the E/log J curve degree of coverage with atomic chlorine concentration of atomic chlorine present in 0 at t. = 0; this being equal to the degree an electrode at a distance x at time r. after of coverage at the potential of the electrode switching off the polar&ion current on current flow c(x, 0) at x 2 0 or 403, t,) at t, > 0 co 6 0 at E, D diffusion coefficient of atomic chlorine E electrode potential us see . on open E” electrode potential circuit 1. lNTRODUCTlON J% reversible electrode potential as see F Faraday constant A few years ago the titanium electrode coated with k,; kr, the slope of the linear part of the a layer of mixed oxides of titanium, ruthenium and/or (q,(O) - q&log t. curve for respectively the other noble metals was introduced into the chlorine- Tafel and Heyrovsky reactions alkali industry. J geometry cd during anodic polar&ion JO geometric exchange cd Several investigators[lkI] have already examined j real cd; j = J/r the bebaviour of this anode material in concentrated In real exchange cd acidic chloride solutions. Moreover, the mechanism j0.T; &,A j, tor respectively the Tafel and the Hey- of the chlorine evolution has also been stud- rovsky reactions ied[2, 5,7]. reaction constant of respectively the Tafel The experimental results are often contradictory. and the Heyrovsky reactions For instance, the following Tafel slopes were found: k+ i32mV for a coating of oxides of Ti, Ir and Ru, 6 108 mV[4] and 40 mV[7] for a coating of RuO, and P constant factor; p = ejc R gas constant 35 mV for a coating of RuO, and TiO,[6]. The differ- i- roughness factor ences in Tafel slopes cannot be explained only in T temperature in “C or K terms of chemical composition of the oxide layer. TR the firing temperature Possible important factors may be the conditions tA time of anodic polarisation used in the preparation and the electrochemical pre- CR the firing time treatment of the electrode. The influence of these fac- t. time after switching off the polarization cur- tors are presented in this paper for two different types rent of electrodes, namely for a titanium electrode coated rate of chlorine evolution for respectively the Tafel and the Heyrovsky reaction with an oxide mixture composed of 30mole% RuO, distance in the electrode measured from the and 7Omole% TiO, and for a pure ruthenium eleo boundary electrode surface/electrolyte trode. The mechanisms proposed in the literature are transfer coefficient mainly based on Ejlog J curves. In addition, we also a for the Heyrovsky reaction examined potential decay curves for elucidating the overpotential during anodic polar&ion mechanism of the chlorine evolution. 1093 1094 L. J. J. JANSSEN, L. M. C. STARMAN& J. G. VISER AND E. BARENDRECHT

t.EXPERlMENTAL cedure was applied to minimise the change of the composition of the electrolyte during the long con- Preparation of the Ru0,fli02-electrode and the 2.1 tinued electrolysis. Ru-electrode The RuO,/TiO, electrodes were prepared by coating titanium sheets (40 x 5 x 0.5 mm, purity of Ti 3. RESULTS 9849%) with a layer of ruthenium oxide and titanium 3.1 Roughness of the electrodes oxide. The RuO, content of the coating was 30% The capacitance determined by the potential-pulse The titanium sheets were cleaned with acetone and method[8] was used as a measure for the roughness subsequently etched in concentrated hydrochloric of the RuOfliOL electrode. acid at 80” for IO min. It appeared that in the investigated range from 800 The coating was achieved by painting the titanium to IO00 mV, both the direction and the amplitude of sheet with an aqueous solution consisting of 1.4ml the potential-pulse had no influence upon the capaci- TiCl,, 3 ml H20, 0.5 ml 30% H,O, and 120mg tance. Assuming that the capacitance measured is RuCl, .3 Hz0 and then by drying in air at about SO”. equal to the double layer capacitance (viz 17.4 pF/cm2 Thereafter the coated sheet was fired at 500” for 1 h, determined at a mercurcy electrode in a 4M except when otherwise stated. In applying multilayer NaCL + 1 M HCL solution), the roughness factor of coatings the procedure was repeated. Unless other- the different Ru02fli0, electrodes varied between 20 wise stated RuO,/IIO, electrodes with two coating and 40. No systematic increase of the roughness of layers were used. the electrodes was found with increasing numbers of For comparison, a pure ruthenium rod was used; coating layers. a disc having a geometric area of l.OOcm’ was the _- exposed electrode surface; the cylinder part of the rod 3.2 Influence of the sweeprate and of the scanning was completely isolated with a Perspex cylinder. potential range on the E/J curves To investigate the effect of the sweeprate on the 2.2 Experimental techniques and apparatus morphology of the E/J curve, this curve was Two different electrolytic cells were used. The measured at a sweeprate varying between 0.5 and measuring cell, a normal H-shaped diaphragm cell 50 mV/s. A typical E/J curve for a Ru03fliOz elec- (volume of both compartments was lOOmI), was trode is represented in Fig. 1. For these electrodes applied for the determination of the potential-current hysteresis occurs; however, at sweeprates lower than curves, the potential-time curves and of the capaci- 2 mV/s the hysteresis disappeared practically. The E/J tance of the electrodes. This cell was thermostatted curves were determined in the potential range from (unless otherwise stated at 2Y). The other cell, a ves- 1050 to 1500mV. This potential range was chosen sel divided into an anodic compartment of 5000 ml to prevent a change of the nature of the &&rode and a cathodic compartment of 2COOml was used for surface during the sweeping of the potential. The the anodic polarization of the electrodes for long minimum potential E,i” was varied between 1050 and periods. In both cases, platinum electrodes served as 1250mV and the maximum potential E,,,,, between counterelectrodes. The voltammograms were mea- 1200 and 1500 mV. It appeared that the shape of the sured with the usual setup. E/J curve remained the same and that only some hys- The potential of the electrode was continuously scanned between two fixed potentiah, Emin and E,,,. Usually, the rate of the potential sweep was 1.25 mV/s. At arriving the steady state (mostly after sweeping for 10 min) the E/J curve was recorded The relation between the potential and the time passed after switching off the polarisation current, t., was determined in the usual way. The capacitance of the electrode was determined by the potential-pulse method, described by Berndt[8]. The magnitude of the potential-pulse was 50 mV. As a reference electrode a see was used. The reversible potential of the CI,/Cl- redox system was determined with a platinum electrode, activated by a chlorine evolution for a short period. It appeared that the activity of the RuO,/TiO, electrodes depends on the polarisation time. To deter- mine this effect the electrode was polarised with a constant current density, viz 25 mA/cm”. After a fixed polarisation time the electrode was removed from the great electrolytic cell and was put into the small H-shape cell containing a 4 M NaCl + 1 M HCI solu- 0 I 10 I5 tion, saturated with chlorine. The voltammogram was J. flbA/C& recorded after a sweep period of about 10 min. There- after the electrode was again polarised anodically in Fig. 1. E/J plot for a RuO,fliO,electrode. Sweeprate, the great electrolytic cell for a fixed period. This pro- 20 mV/s. Chlorine evolution on titanium electrodes 1095

IO-

OS- I I I I Fig. 2. The current density J at an overpotential of 90 mV 350 ‘Ko 4cu xx) 550 6M) T and the Tafel slope b for a RuO,fliO,-electrode as a func- R. *C tion of the total electrolysis time r,. The RuO,/TiO,-elec- Fig. 4. The current density J at an overpotential of 90 mV trade was formed by firing at 500°C for 2 h. and the Tafcl slope b for a Ru02/Ti02-electrode as a function of the firing temperature tR for 2 h. teresis occurred Its effect depended on the difference between E,,,inand Em,. The electrodes were polarised with 25 mA/cmZ for Owing to the preceding results, the following about 150 h. In Fig. 4 the cd and the Tafel slope are potentiodynamic E/J curves were determined at a plotted us the temperature of the firing process. The sweeprate of 1.25 mV/s and at a potential range of results of Fig. 4 were supported by a second series 1050-l 300 mV. of experiments. In this series, the minimum value however, was 50 mV, but was also obtained for a fir- 3.3 The nature of the electrode surface ing temperature of 500”. Figure 4 shows that the cur- 3.3.1 Time of adic polarisation. The influence of rent density, which is a measure for the activity of the anodic polar&ion on the behaviour of the Ru- the electrode evolving chlorine, decreases with in- electrode and the RuO,/TiO,-electrode evolving creasing temperature of the firing process. The results chlorine were investigated by determining potentio- agree with that of Kalinovskii et al[9] found for firing dynamic E/J curves after the electrodes have been temperature above 400”. polarised for several times in succession. In Fig. 2, 3.3.3 Time of the firing process. The effect of the for a RuOz/TiO, electrode the current density at a firing time was investigated for RuO,/TiO, electrodes constant overpotential and the Tafel slope b of the formed at 500” with times varying between 0.25 and E/log J curve are plotted versus the total time ta 6h. of anodic polarisation of the electrode. The Ru02/TiOz layer on the electrode fired for 6 h The shapes of both curves of Fig. 2 are cbaracter- started to scale off after electrolysis of 100h and had istic. The decrease of both the current density and practically disappeared after a 150 h electrolysis. It the Tafel slope occurred mainly during the first 20 h. appeared that the current density and the Tafel slope This phenomenon occurred most markedly for elec- are practically independent of the firing time. trodes formed by firing at 350” for 15 min. For a Ru- Kuhn and Mortimeti4] found also no systematic electrode the J/t,, and the bitA relation are repre- variation in the activity of the electrodes fired at 400 sented in Fig. 3. for l-50 h. In the following only the current density and the 3.3.4 Thickness of the coating layer. The thickness Tafel slope are given for electrodes of which the of the coating layer of the RuO,/TiO, electrode was nature of the surface has become constant. varied by covering the electrodes 1,2, 5 and 10 times 3.3.2 Temperature of the firing process. The effect with the RuCl,/I’iCl, solution. After a pre-electrolysis of the temperature of the firing process on the chlor- for about 150 h, both the E-J curve and the capaci- ine evolution at a RuO,/Tio, electrode was investi- tance were determined From these experiments it fol- gated for the temperature range from 350 to 600”. lowed that the thickness of the electrode had no sig- nificant influence upon the Tafel slope and upon the ratio between the current density and the capacitance of the electrode. The exchanged current density J, was determined by extrapolation of the linear section of the E/log J curve. The electrochemical surface area was calculated from the capacitance of the electrode and the capacitance of the double layer. The obtained value varied between 8 x 10m5 and 20 x 10-5m4/cm’, the average value of Jo was , , 13 x 10-5mA/cmZ. 3.4 Electrolytic conditions 3.4.1 Composition ofthe solution. For solutions with a chloride concentration of 5 M and a pH varying tnv h between 0 and 3, it appeared that the E/J curve does Fig. 3. The current density J at an overpotential of 90 mV not depend on the PH. The influence of the chloride and the Tafel slope b for a Ru-electrode as a function concentration on the E/J curve was investigated for of the total electrolysis time t,. NaCl/HCl solutions of which the ratio between the 1096 L. J. J. HANSEN, L. M. C. STARMANS, J. G. VMER AND E. BARENDRECHT

IO' T-I. K-’

Fig. 5. Plot of the current density J at an overpotential of 4OmV us the reciprocal absolute temperature for a RuO,fliO,-electrode formed at a firing temperature of 500” for 2 h.

NaCl and the HCI concentration was equal to 4, and the total chloride concentration varied between 0.5 Fig. 7. Potential decay curves for a Ru-electrode which and 5 M. For a RuO,/TiO,-electrode it appeared that had been polarised anodically with various cds, Before the chloride concentration has no influence upon tbe these curves were determined, the Ru-electrode was polar- current density at E = 1250mV; this potential is ised anodically with 2.5 mA/cm2 for one day. By this treat- minimal 5OmV higher than the reversible potential ment the electrode surface became black due to small ruthenium oxide particles. of the Cl-/CI, coupte obtained the results were in agreement with those of Kuhn and Mortimer[4] and those of Erenburg et aI[7]. 3.5 Potential-decay 3.4.2 Temperature. The temperature was varied The electrodes were polarised at a constant anodic between 2 and 70” to establish the effect of the tem- current to a steady-state overvoltage. Thereafter the perature of the electrolysis on the E/J relation for polarisation current was switched off and the poten- a RuO,-electrode. From the T&l slope b the transfer tial Q, on open circuit/log t. curve is represented in coefficient a = (2.3 RT/bF) - 1 was calculated. This Fig. 6 for a RuO,/TiO,-electrode which had been equation is used when the Heyrovsky reaction is the polarised anodically with different cd. The q,,/log tn- rate-determining step[lO]. It appeared that ccis inde- curve is practically linear at t, r 0.1 s. The slope of ‘pendent of the temperature. The average value of the the linear section of this curve is independent of the transfer-coefficient was 0.45 f 0.05. current density during the anodic polarisation, viz In Fig 5 the current density at an overpotential 20 mV. For J = 0.75 mA/cm’ qn at t, = 0 approached of 60 mV is plotted us the reciprocal absolute tem- the overpotential during the anodic polarisation. A perature. From the slope of the straight line of this slope of 20mV for the Qog t,-curve was obtained figure it was calculated that the activation energy is also in the time range from 2@--200s. Also for a Ru- 6.5 kcal/mol at an overpotential of 6OmV. electrode and for different current densities the over- voltige on open circuit is plotted vs the logarithm of time t,, see Fig. 7. The curves are straight lines with nearly tbe same slope, viz 22mV. The Ru-electrode used had been polarised anodically with 25 mA/cm’ for 1 day. After this pre-treatment the Ru- electrode was black.

d. THEORY :m2 During the electrolysis of an acidic chloride solution, chlorine is formed by anodic oxidation of Cl- ions Tbe chlorine formation can occur according to the Volmer-Tafel mechanism or the Volmer-Hey- rovsky mechanism Both mechanisms are already ex- tensively discussed in the literature for graphite and platinum electrodes. The theoretical relation between the cd, J, the overpotential, q, and the degree of coverage, 0, are given by Janssen and Hoogland[lO]. For 0 I I I 1 t I a graphite anode it has been observed that after 001 002 0.05 01 0.2 a5 1.0 switching off &hepolarisation current, the rate of the t,* 5 chlorine evolution and the potential of the electrode Fig. 6. Potential decay curves for a RuOfliOzelectrode decreases slowly[:113. which had been polarised anodically with various current A slow overpotential decay is observed also for a densities. The RuO,fliO+&ctrode was formed at a firing Ru02/II02 electrode at which chlorine had been temperature of 500” for 2 h. evolved. Although on open circuit the overpotential Chlorine evolution on titanium electrodes 1097 in this case decreased more strongly and no gas evo- If 0, < 1, then for 0 at low overpotentials, lution could be observed, it is obvious that otherwise the behaviour of the RuOJlIO,-electrode is compar- (5) able to that of the graphite electrode. The chlorine, evolved at the electrode on open cir- From (2) and (5) it follows that cuit is formed from atomic chlorine. Chlorine can be formed according to two mechanisms.

The Tare1 mechanism refers to the reaction If 2 kT 0, t. @ f3,/&,, equation (6) shows that the slope 2 Cl,, - Cl,. The potential on open circuit is deter- of the ~,/ln t, curve is -RT/F. mined by the equilibrium between the cathodic and In extension we now suppose that atomic chlorine the anodic Volmer reaction is also present in the bulk of the anode material. This means that atomic chlorine diffuses out of the bulk cl- F! Cl,, + e-. of the anode material and reacts to molecular chlor- First we discuss the case where atomic chlorine is ine according to the Tafel reaction. present only on the outer surface of the electrode. We suppose that the diffusion of chlorine atoms Moreover, we neglect the dissociation reaction of can be described as a one-dimensional diffusion prob- molecular chlorine. For this case it can be deduced lem. Consequently, the diffusion satisfies to the fol- that[ IO] lowing second equation of Fick, 64x, tn) D 62c(x,L) - ; = 2k,t3=.

” St, -’ 6x2 Integration of equation (1) gives : Moreover, it is supposed that at t, = 0 the concentration of atomic chlbrine is independent of the distance (g= O” x and that the total thickness is so great that this 1 + 2k,f+,t,’ thickness can be considered as infinity. The boundary The rate of the chlorine evolution, I+-, according to conditions are then: the Tafel reaction is equal to[lO] c&,0) = c0 for x > 0 c(cc, t.) = c0 for t, k 0 (3)

/.) t.1 wx, = kFc2(0, t.) Neglecting the current due to the discharge of the 6x capacitance of the double layer, the rate of the catho- where : dic Volmer reaction is then equal to the rate of anodic Volmer reaction. So, it follows that the degree of k*T _ 2$T coverage with atomic chlorine, 0, is given by[ 1l] I 1 and the degree of coverage 0 = pc B= (4) i+l” i + {(I - 0,)/&}exp -- where p is a constant factor. RT The relation between ~(0, t,)/cO and t, was calculated for different values of D, k; and c@ The calcula- tion was done by means of a Crank-Nicolson difference scheme with nonlinear boundary conditions[ 151 on a Burroughs B6700 computer. The time step size we used was At = 0.01 s and for the x-coordinate 640 gridpoints from O-cc, were taken. All these three factors have a great influence upon the log [c (0, tJco]/log t,, curves. Since the concentration of atomic chlorine at the electrode surface is proportional to exp (Fq/RT), qn (0) - r~, instead of log c (0, t&/co is represented graphically. The results of the calculations are shown in Figs. 8 and 12. In Fig. 8 q, (0) - 7. is plotted us log to for c,, = 10T5 mole/cm3, D = 10e4 cm’/s and various values of k+; in Fig. 9 for c0 = 10-l mole/cm3, kg = 106cm4/ moles and various values of D; and in Fig. 10 for D = IO-’ ad/s, k; = lo4 cm4/mole s and various values of co From the theoretical (II. (0) - Qlog t,-relations, Potential decay curves for the chlorine formation (5) and from the proportionality between c and 0 on open circuit according to the Tafel mechanism at it follows that in the range 20mV < q. (0) - c0 = 10-5mole/cm3. D = 10~4cm2/s and various values q. < 12OmV at 0.1 s -=zt, -=z10s the (q, (0) - q&log of kf. t,-curve is practically linear, its slope hr is generally 1098 L. J. J. JAN=. L. M. C. %L+RMANS, J. G. vlS&t ANDE. hRENDRECHT

I,_ a 2

Fig. 9. Potential decay curves for the chlorine formation on open circuit according to the Tafel mechanism at c0 = IO-’ mole/cm3, kf. = IO6cm+/moIe s and various Fig. 11. Potential decay curves for the chlorine formation values of D. on open circuit according to the Vohner-Heyrovsky mechanism at c0 = 10-5mole/s3, D = 10-4cm2/s and practically independent of c,, and k$, whereas D can various values of kg. vary from 10m4 to 10m5 c&/s. The slope hr is then equal to about 14rnV at 25”. At smrdler values of is present on the electrode surface. The rate of the D however, the slope of the (tl. (0) - ~&log Qcurve decrease of the degreeof coverage is given by changes strongly with increasing log t.. 4.2 Volmer-Heyrousky mechanism The other mechanism according to which molecular chlorine is formed on open circuit, is the Volmer- If 0, 4 1, then 8 at low overpotentials can be calcu- Heyrovsky mechanism, consisting of lated with equation (5). Substitution of 0 from (5) into Cl, + Cl- -Zlz + e- (8), followed by integration and rearrangements, deliver anodic Heyrovsky reaction ; us the following expression @U,(O) Q,= -Eln 2kHaf,+exp-- (9) cathodic Volmer rz&$.“- ““- . aAF RT 1 As mentioned before, the rate of the chlorine evolu- In extension, we discuss in the following also the case tion is determined by the anodic Heyrovsky reaction, when atomic chlorine diiuscs out of the bulk of the while neglecting the cathodic Heyrovsky reaction. electrode material. We apply also the one-dimensional First we discuss the case where only atomic chlorine second equation of Fick with the following boundary

1_1m 001 0.1 IO + “l 3 Fig. 10. Potential decay curves for the chIorinc formation Fig. 12. Potential decay curves for the chlorine formation on open circuit according to the Tafel mechanism at on crpen circuit according to the Volme-Heyrovsky mech- D = 10e5 cm’/& k$ = 10’ cm*/mole s and various values anism at c0 = 10-6mole/cm3*k~ = 10zcm3’2/mole* s of cg. and various values of D. Chlorine evolution on titanium electrodes 1099

this electrode (Fig 4) and by the temperature of the electrolysis (3.4.2) At the firing temperature of 500” the slope b has a minimum value. For one series of experiments, this Mlue is about 41 mV and for an other series about SOmV. Mostly, a minimum value of about 4OmV at 25’ was found for RuO,/TiO,- electrode fired at 500”. It is already mentioned that the Tafel slope increases linearly with the absolute temperature (3.4.2). Taking into account that GL= 0.45 f 0.05 for the series of experiments, mentioned in section 3.4.2., and that the Tafel slope at 25” for an ‘ideal’ RuO,/TiO~-electrode. is about 4OmV, it, follows that a = 0.5 for the chlorine evolution at an ‘ideal’ RuO,/TiO,-electrode. For a Ru-electrode pre-electrolysed with an anodic current density of 25mA/cm2 for 1 day, the Tafel slope at 25” was also about 40mV. Comparing a slope of 40 mV with the theoretical Fig. 13. Potential decay curves for the chlorine formation ones for the possible reaction mechanism[lO], it fol- on open circuit according to the Volmer-Heyrovsky mech- lows that chlorine is formed according to the Volmer- anism at D = 10-5cm’/s, ki = 103’2cm”2/molef s and Heyrovsky mechanism and that the Heyrovsky reac- various cw tion is the rate-determining step. A slope higher than 40mV may be caused by a conditions bad electrical contact between the RuOJTiO, coating layer and the titanium sheet of the de&ode. It is c(x,O) = c0 and x z 0, possible that the experimental value of 40mV is not c(co, t.) = co for t, 3 0, the true minimum value of the Tafel slope. In the literature smaller values of the Tafel slope are given. D L) wx,~ = k;c(O, tn)3’2 Faita and Fiori[2] found a slope of about 30 mV for 6X electrodes consisting of oxides of Ru, Ti and Ir. Bon- The latter condition can be deduced from the rate of dar et al[3] obtained a slope of 34 mV for a TifRuO, the chlorine evolution on open circuit, vH, according to electrode with 90% Ti and 10% Ru in the active layer. the Volmer-Heyrovsky mechanism. The rate is given A slope of 30mV pertains to the Tafel reaction as the rate-determining step[lO]. The difference in ex- by perimental values of the Tafel slope may be explained jO,H tP+an by a different composition of the electrode. “H=F8:+““’ where j,, is the exchange current density of the 5.1.2 The relation between the overpotential on open Heyrovsky reaction. Since 0 = pc and assuming circuit and the time. Section 4 shows that the relation c(* = 0.5, we get uH = kBc(O,t,J312 where between the overpotential on open circuit and the k* = p3’zj0,rr time can be used in order to elucidate the reaction li- mechanism. Owing to the conclusions in 5.1.1, we dis- Fe;‘2 cuss the Volmer-Tafel mechanism only for the case For this case also the ratio between ~(0, Q/c0 and that the Tafel reaction is the rate-determining step t,, was calculated for different D, kf, and co values. and the Tafel slope is 30mV at 25”. If on open circuit In analogy with the Tafel treatment, q, (0) -q. is no diffusion of atomic chlorine from the bulk of the plotted OSlog t, for c,, = 10-smole/cm3, D = lOA electrode material occurs the theoretical slope of the cm’/.s and various ki (see Fig. 11); for c0 = low6 q,Jlog t. curve is 60mV for a sufficiently long time. mole/cm3, kg = IO* cm3~2/mol’~*s and various D (see This value is much greater than the experimental Fig. 12) and for D = 10-scm2/s, kg = 103’2cm3/2/ value of 22mV. In the case that diffusion plays a mole”‘s and various c0 (see Fig. 13) The slope h, role the (q,(O)-q&logf, relation is given for different of the (II. (0) - q,,)/log bcurve at 20mV < q. (0) - values of k$, c,, and D in the Figs. 8-10. A diffusion qn < 120 mV and at 0.1 s < tn < 10 s is also practi- coefficient in the order of 10-s cm’/s is a reasonable cally independent of cO, kg and of D varying from supposition,> since the diffusion of atomic chlorine 1OL4 to 10~5cm2/s. The slope h, is then equal to occurs on the surface of the crystallites of the active about 20 mV. layer of the RuO,,TiO, and of the Ru-electrode. The degree of coverage B0 at the moment of switch-

5. DISCUSSION ing-off the current increases with the overvoltage 4 during the electrolysis according to 8 + 6, exp 5.1 The mechanism of the chlorine formntion (Fq/KI). At sufficient great values of qn(0)-qrl, the 5.1.1 The T&l slope of the E/log J c&e. The Tafel theoretical (qn(0)-qJlog t” curve is linear and has a slope, b, generally used to elucidate the mechanism slope of about 14 mV (4.1). Unfortunately, from the of the chlorine formation. It appears that the slope experimental VJlog t. curves for J > 0.75 mA/cm2, ie b for a Ru0,/Ti02-electrode is influenced by the tem- for 9 > 58 mV (Fig. 6), the determination of q,(O) is perature of the firing process at the preparation of impossible. Especially for high overpotenfials the 1100 L. J. J. JANSSEN,L. M. C. STARMANS, J. G. VISSER AND E. BARENDRECHT double layer will have a great effect on the potential v > 60 mV, the cathodic reaction can be neglected decay curve t, smaller than 0.1 s. However, for the with respect to the anodic reaction. The current den- results given in Fig. 7 and in Fig. 6 with exception sity j for the Volmer-Heyrovsky reaction where the of those for J = 0.75 mA/cm’, it is likely that q,,(Oh,, Heyrovsky reaction is the rate-determining step is will be sufficiently great in order to obtain a linear given by[ IO] g,,/log t, relation. The experimental slope of the q,/log t, curve, viz 20mV (Fig. 6) and 22 mV (Fig. j = 2jo.R~exp!!!E!?. RT 7) does not agree with the theoretical one for the r Tafel mechanism Since the Tafel slope is about 40 mV, the litera- In the following we discuss the Volmer-Heyrovsky ture[lO] shows that Q/t?, is proportional to exp mechanism if the Heyrovsky reaction is the rate- W//R T). determining step and the Tafel slope is 40 mV at 25”. Taking this factor into account, it follows from de Analogous to the discussion of the Tafel mechanism activation energy at q = 6OmV, viz 6.5 k&/mole we distinguish two cases uiz without and with diffu- (3.4.2), that the activation energy at the reversible sion of atomic chlorine. When the diffusion is not potential is equal to 8.5 kcal/mole for the RuOz/ taken into consideration, a comparison of the experi- TiO,-electrode. mental slope of the q,/log t. curve with the theoretical It appeared that the Tafel slope b depends on the one, ie 120mV, shows that both slopes do not coin- firing temperature. For an ‘ideal’ RuOz/TiOz-elec- cide. However, when the difision is taken into con- trade we obtain a transfer coefficient of 0.5. This value sideration, the theoretical slope of the q,/log r,, at a is also found for the Ru-electrode. A smaller transfer- sufficiently long time is equal to 20 mV (Figs 11-13): coefficient may be caused by an oxide layer. The real So, there is a very good agreement between the theo- experiment exchange current density is equal to that retical slope and the experimental slope. The experi- of the Heyrovsky reaction. For RuO,/Tio,-electrodes, mental slope at a sufficiently long time is not in- formed by firing at 500” for 2 h in a 4 M NaCl solu- fluenced by the current density which was applied tion saturated with chlorine, the real exchange current during the polarization. This agrees also with the density j,, of the Heyrovsky reaction is equal to theoretical model, at which the diffusion is con- 13 x lo- mA/cm* at 25” (3.3.4.). sidered. For low current densities viz J < 0.75 mA/cm2 it REFERENCES appeared that v_(O) was practically equal to 21during the current flow. The q,/log tn curve at low current 1. S. Trasatti and G. Buzzanea, J. elecrraunol. Interfociol densities can also be described with the theoretical Chem 29, 1 (1971). model. 2. G. Faita and G. Fiori, J. appl. Electrochem. 2, 31 (1972). From the preceding discussion and from 5.1.1. it 3. R. U. Bondar, A. A. Borisova and E. A. Kalinovskii, follows that chlorine is formed according to the Vol- Soviet Electrochem. 10, 37 (1974). mer-Heyrovsky mechanism and that the Heyrovsky 4. A. T. Kuhn and C. J. Moitimer, J. electrochem Sot. reaction is the rate-determining step both on current 120, 231 (1973). flow and on open circuit. 5. R. G. Ehrenburg, L. J. Krishtalik and V. J. Bystrov, In the literature the potential decay is mostly plot- Soviet Electrochem. 8, 1690 (1972). ted us log (t,, + to), where t, is a positive quantity[13]. 6. R. G. Ehrenburg, L. J. Krishtalik and J. P. Varoshovs- This treatment of the potential decay is based on the kova. Soviet Ekctrochem 11. 898 (1975). 7. R. b: Ehrenburg, L. I. Krishtalik and J.? P. Yaroshevs- assumption of a constant electrode capacity. Under, koya, Soviet Electrochem 11, 993 (1975). this condition the decay slope and the Tafel slope/ D. Berndt, Elrctrochim. Acta 10, 1067 (1965). are identical[lKJ. For oxygen evolution at a nickel :: E. A. Kalinovskii, R. N. Bon&r and N. N. Meshkova, oxide electrode in 0.15 M KOH at 25” Conway and Soviet Electrochem. 8. 1430 (1972). Bourgault[14] found different slopes. In this case the 10. L. J. J. Janssen and J. G. H&J&d, EIectrochim. Acta ionic double layer capacitance is negligible compared 15, 941 (1970). with the pseudo-capacitance of adsorbed OH radicals. 11. L. J. J. Janssen and J. G. Hoogland, Electrochim Acra For a hydrogen electrode at which ,the logarithm of 15, 1667 (1970). the degree of coverage of hydrogen atoms is propor- 12. D. U. van Rosenberg Method,for the Numeric& SOB- tional to the overpotential, Conway and Bourgault rion of Partial Differential Eauations. D. 75. American Else&r, New Yolk (1969). _ _ expect the same behaviour[14]. 13. H. B. Morley and F. E. W. Wetmore, Can. J. Chem 34, 359 (1956). 5.2 Acrivativn energy and kinetic parameters 14. B. E. Conway and P. L. BourgauJt, Trans. Faraday At sufficient high overpotentials, for instance Sot. 58, 593 (J962)<