Electrochimica Acta 51 (2005) 133145

The reduction of l-cystine hydrochloride at stationary and rotating disc mercury electrodes
T.R. Ralph a,b,d , M.L. Hitchman a , J.P. Millington c , F.C. Walsh d,
Department of Pure and Applied Chemistry, University of Strathclyde, 295 Cathedral Street, Glasgow G1 1XL, UK b Johnson Matthey Fuel Cells, Lydiard Fields, Great Western Way, Swindon SN5 8AT, UK c School of Chemical Engineering and Analytical Science, University of Manchester, Sackville Street, Manchester M60 1QD, UK Electrochemical Engineering Group, School of Engineering Sciences, University of Southampton, Higheld, Southampton SO17 1BJ, UK Received 15 February 2005; received in revised form 11 April 2005; accepted 15 April 2005 Available online 23 May 2005
a

Abstract The kinetics of l-cystine hydrochloride reduction have been studied at a mercury-plated copper rotating disc electrode (RDE) and at a stationary mercury disc electrode (SMDE) in 0.1 mol dm3 HCl at 298 K. The reduction of the disulphide is irreversible and hydrogen evolution is the major side reaction. In contrast to steady state electrode kinetic studies at a mercury drop electrode (which shows a well-dened limiting current), the mercury-plated Cu RDE shows overlap between disulphide reduction and hydrogen evolution. These effects are attributable to strong reactant adsorption with a calculated surface coverage close to 100%. A Tafel slope of 185 mV per decade is found with a cathodic transfer coefcient of 0.32 and a formal rate constant of 6.7 109 m s1 . The relative merits of steady state voltammetry at a mercury-plated copper RDE and linear sweep voltammetry at the SMDE are discussed, as is the mechanism of l-cysteine hydrochloride formation. 2005 Elsevier Ltd. All rights reserved.
Keywords: l-Cysteine hydrochloride; l-Cystine hydrochloride; Mercury-plated copper; Rotating disc; Stationary mercury drop

1. Introduction There is a signicant body of literature on the polarography of l-cystine reduction in a variety of electrolytes [1]. In acid media, l-cystine reduction commences at 0.1 V versus SCE with a small pre-wave before the main wave, the latter being diffusion controlled and both chemically and electrochemically irreversible. A limiting current plateau starts at a potential of approximately 0.5 V versus SCE due to: RSSR + 2H+ + 2e = 2RSH (1)

Reaction (2) has been confused with mercury cysteinate formation in earlier studies in acid media, as noted in a review [1]. This situation has arisen due to the similar potentials for the pre-wave and the reversible oxidation of l-cysteine to mercury cysteinate: 2RSH + 2Hg 2e = (RS)2 Hg2(ads) + 2H+ 2RSH + 2Hg 2e = 2RSHg(ads) + 2H+ (3) (4)

where R = CH2 (NH2 HCl)COOH, the pre-wave in pH 7.4 buffer solution was attributed by Stankovich and Bard [2] to the reduction of adsorbed l-cystine to form l-cysteine: RSSR(ads) + 2H+ + 2e = 2RSH

(2)

Corresponding author. Tel.: +44 2380 598752; fax: +44 2380 598754. E-mail address: f.c.walsh@soton.ac.uk (F.C. Walsh).

Earlier studies have not, however, examined l-cystine hydrochloride reduction under conditions typically found in the electrosynthesis of l-cysteine hydrochloride. Electrode kinetic studies have involved reactant concentrations much less than 0.4 mol m3 to avoid problems with disulphide and thiol adsorption blocking the small electrode surface area available at a mercury drop electrode. This paper examines the reduction under conditions more closely resembling those in the electrosynthesis of l-cysteine hydrochloride and provides detailed kinetic information obtained from a variety of

0013-4686/$ see front matter 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.electacta.2005.04.012

134

T.R. Ralph et al. / Electrochimica Acta 51 (2005) 133145

complementary techniques. A large, stationary mercury disc electrode (SMDE) has been employed to allow higher reactant concentrations to be examined using cyclic voltammetry. This was combined with studies at a mercury-plated copper rotating disc electrode (RDE), which allowed mass transport to be controlled and kinetic parameters to be extracted using steady state voltammetry and controlled potential coulometry in addition to cyclic voltammetry.

cal electrode area of 0.80 cm2 . A platinum foil was placed around the rim to provide an approximately planar mercury surface. Electrical contact to the mercury was made via a platinum wire protruding into the cup and sealed within the syringe. 2.3. Electrochemical cells All of the cells used a saturated calomel electrode (SCE) as the reference electrode. Steady state, linear and cyclic voltammograms were recorded at the mercury-plated copper RDE. The Luggin capillary, manufactured from a syringe, communicated the SCE (Ingold) to within less than 1 mm of the RDE surface. Naon 324 cationic ion-exchange membrane divided the cell. The catholyte compartment accommodated 250 cm3 of electrolyte and was surrounded by a thermostated water jacket. The large volume ensured the cell walls did not modify the catholyte ow pattern at the disc. A Perspex top sealed the compartment mainly to prevent the dissolution of oxygen into purged catholytes, although it aided centralisation of the disc and facilitated attachment of a thermometer and purging line. A much larger platinum foil auxiliary electrode of geometrical area 4.7 cm2 ensured that oxygen evolution at the counter electrode was not rate controlling. Linear and cyclic voltammograms were recorded at the SMDE. The design was similar to the RDE cell. A two-piece Luggin capillary was used to allow easy removal of the capillary from the cell. Constant potential coulometry at the mercury-plated copper RDE was performed in a glass cell having a working electrode compartment of 25 cm3 . The SCE (Ingold) reference electrode was connected to within 1 mm from the RDE surface by a Luggin capillary formed from a length of silicone rubber tubing attached to a glass end section, which was manufactured as an integral part of the catholyte chamber. Naon 324 cationic ion-exchange membrane divided the cell. Efcient separation was particularly important because of the long duration of the coulometry experiments and the need to accurately determine the reactant and product concentrations after electrolysis. The geometrical area of the platinum foil counter electrode was 1.8 cm2 . Electrolysis times were minimised by using an RDE, rather than a stationary electrode, to provide efcient mass transport and to maximise the electrode area (0.50 cm2 ) to catholyte volume ratio (0.50 cm2 /25 cm3 = 0.02 cm1 ). To ensure that RDE laminar ow theory [4] was obeyed in the small catholyte compartment for the range of rotation rates investigated measurements were made at a platinum RDE with the ferricyanide/ferrocyanide ion redox couple. Well-dened limiting current plateaux were obtained for ferricyanide ion reduction and ferrocyanide ion oxidation. Application of the Levich equation produced diffusion coefcients of 6.3 1010 m2 s1 for ferricyanide and 7.1 1010 m2 s1 for the ferrocyanide ion in aqueous 1.0 mol dm3 potassium hydroxide, at 298 K, in close agreement with the literature [6].

2. Experimental details 2.1. Mercury-plated copper electrode Construction of the three electrode, three compartment RDE cell has been considered in detail elsewhere [3]. Briey, it consisted of a brass shaft in an epoxy resin shell with a detachable bottom section containing the copper disc, which was machined to give a tight, press t into a PTFE sheath. A brass spring provided electrical contact between the disc and the shaft. The sheath diameter (2 cm) was close to 2.5 times larger than that of the disc, to provide well-dened laminar ow conditions [4]. The active electrode had a geometrical electrode area of 0.50 cm2 . The copper (Goodfellow) was 99.99 wt% with 70 ppm silver; 2 ppm iron, lead, nickel and silicon; 1 ppm aluminium, bismuth, calcium, magnesium and tin; <1 ppm chromium, manganese and sodium. A method due to Daly et al. [5] was used to plate the copper RDE with mercury. The copper electrode was wet polished to a mirror nish, initially on ne silicon carbide paper followed by 1, 0.1 and 0.05 m alumina, in turn, on polishing cloths. After rinsing with absolute alcohol and doubly distilled water, the electrode was ultrasonicated in pure water to remove any remaining alumina particles. Electroplating was performed from aqueous 0.15 mol dm3 mercuric nitrate and 0.12 mol dm3 potassium cyanide (250 cm3 ) previously deoxygenated with nitrogen for an hour. The copper electrode was rotated at 30 Hz and a constant current density of 50 mA cm2 was passed for 5 min using a platinum foil anode, while nitrogen blanketed the cell to prevent air ingress. The mercury surface was rinsed with doubly distilled water then wiped with a paper tissue, resulting in a mirror nish. The electroplating solution was prepared from BDH Analar grade chemicals and doubly distilled water. A constant current density was supplied by an E G & G PARC Model 371 potentiostat/galvanostat and the electrode was rotated with an Oxford Electrodes assembly. The calibrated rotor was capable of providing rotation frequencies of between 0 and 50 Hz to an accuracy of 0.1 Hz. 2.2. Stationary mercury disc electrode The three electrode, three compartment glass cell, which provided an upward facing mercury working electrode, was based on a vertical, glass syringe altered to give a cup at its upper end. The cup diameter of 1.01 cm gave a geometri-

T.R. Ralph et al. / Electrochimica Acta 51 (2005) 133145

135

2.4. Chemicals and solutions BDH Analar grade l-cystine and l-cysteine were used as supplied. Most experiments were performed in aqueous 0.1 or 2.0 mol dm3 hydrochloric acid. The effect of pH on the system was investigated using aqueous solutions of constant ionic strength and pH [7]. 2.5. Voltammograms of l-cystine hydrochloride reduction The mercury-plated copper RDE and the SMDE were maintained at a potential at least 50 mV negative of that required for calomel formation and corrosion of copper. The initial potential chosen was approximately 0.2 V versus SCE in aqueous 0.1 mol dm3 hydrochloric acid and 0.25 V versus SCE in aqueous 2.0 mol dm3 hydrochloric acid. Steady state voltammograms were recorded at the mercury-plated copper RDE by sweeping the electrode potential from the initial to the nal potential limit then back to the initial potential (to check for hysteresis) at a rate of 2 mV s1 . The current output was monitored with a Houston 2000 XY chart recorder or with a Keithly 178 digital multimeter connected across the current output terminals of the potentiostat. Linear sweep and cyclic voltammograms were recorded at the mercury-plated copper RDE and the SMDE by sweeping the electrode potential linearly at sweep rates of between 25 and 500 mV s1 , and recording the current response on a Houston 2000 XY chart recorder. All voltammograms were recorded at 298 K, unless indicated otherwise, and at least in triplicate. The cells were thermostated using a Townson and Mercer water bath linked to the cell with silicone tubing. Electrolytes were deoxygenated with nitrogen for an hour prior to measurements and the cell blanketed with the gas during measurements to prevent air ingress. A PARC Model 371 potentiostat/galvanostat (capable of delivering 1 A at 30 V maximum) or a Model 363 potentiostat/galvanostat (7 A at 20 V maximum) was used. Potential sweep rates were controlled using a Thompson DRG 16 sweep generator. Current interrupt studies showed that the maximum uncompensated ohmic drop between the working and reference electrodes was approximately 10 mV at a current density of 150 A m2 , the majority of measurements being made at much lower current densities. 2.6. Coulometry of l-cystine hydrochloride reduction at the mercury-plated copper RDE l-Cystine hydrochloride reduction was examined using coulometry. Immediately upon contacting the electrolyte the electrode was set at the desired potential for the electrolysis and the electrode rotation rate was xed. During electrolysis, the charge passed and electrolysis time was recorded. Electrolytes were deoxygenated with nitrogen prior to elec-

trolysis and a steady stream of the gas was passed over the catholyte during measurements to prevent air ingress. At the end of an experiment, voltammograms were recorded and the catholyte was analysed for l-cystine hydrochloride and l-cysteine hydrochloride concentrations by high performance liquid chromatography (HPLC). Further details of the HPLC measurements are available elsewhere [3,8]. An E G & G PARC Model 173 potentiostat/galvanostat (1 A at 100 V maximum) with an E G & G PARC Model 179 digital coulometer as a plug in accessory was used. The HPLC employed a Hewlett Packard HP 1090 Series M Liquid Chromatograph system with autoinjector, 1040A diode array UV detector and computer control. The diode array, with 211 photodiodes reading every 10 ms, allowed simultaneous monitoring of different detection wavelengths and simultaneous recording of chromatographs and UV spectra. Low signal to noise ratios were possible by signal averaging which enabled low absorbances to be used. Peak purity was readily determined by comparing UV spectra with those for standard samples. 2.7. Measurement of electrolyte properties Solution viscosity was determined by standard U-tubes in a Gallenkamp viscometer bath. Solution density was determined using density bottles. A Corning 150 pH/ion meter and Ingold pH electrode were used to measure solution pH.

3. Results and discussion 3.1. Mercury-plated copper RDE There is always some concern over the purity of mercuryplated electrode surfaces. Both Daly et al. [5], whose method was employed in this study and Yoshida [9] have suggested that the mercury-plated electrode surface consists of a mercury lm, containing not more than 3 ppm copper, on top of a copper amalgam. After approximately 24 h, however, both authors observed dissolution of the mercury lm in the underlying amalgam. Three experiments were, therefore, performed to determine whether the surface behaved essentially as a mercury lm, or an amalgam, on the timescale of the voltammetry (<30 min) and the coulometry (8 h) experiments used in this study. The electrochemical reduction of uorescein was examined at a fresh mercury-plated copper RDE surface. The E1/2 value for the reversible, one electron reduction of uorescein to semi-uorescein in aqueous 0.1 mol dm3 sodium hydroxide was measured as 1.175 V versus SCE. This is in close accord with an E1/2 value of 1.168 V versus SCE measured at a dropping mercury electrode (DME) [5]. The Levich equation: jL = 0.062zFD2/3 1/6 cb 1/2 (5)

136

T.R. Ralph et al. / Electrochimica Acta 51 (2005) 133145

was used to measure the diffusion coefcient of uorescein at the RDE. Here, jL is the limiting current density (A m2 ), z the number of electrons transferred to the uorescein molecule (z = 1), D the diffusion coefcient of uorescein (m2 s1 ), v the kinematic viscosity of the electrolyte (1.08 106 m2 s1 [5]), cb the bulk concentration of uorescein (5 mol m3 ) and is the rotation rate of the electrode (rad s1 ). A linear plot of jL versus 1/2 enabled the diffusion coefcient of uorescein to be measured as 3.0 0.2 1010 m2 s1 in close agreement with the value of 3.2 1010 m2 s1 found using a channel electrode [10]. While the uorescein studies provided condence in the quality of the mercury-plated surface, a key feature of mercury in the studies of l-cystine hydrochloride reduction is a high hydrogen overpotential. Consequently, the cathodic potential limit for hydrogen evolution from aqueous 0.1 mol dm3 hydrochloric acid was measured. At a fresh mercury-plated copper surface, a current density of 0.010 A m2 at 1.0 V versus SCE was recorded. This is close to the value of 0.008 A m2 measured at the same potential with a hanging mercury drop electrode (HMDE) [11]. It is likely that the mercury-plated electrode represents the optimum achievable by this method, in terms of a high hydrogen overpotential for a moderate mercury loading. Yoshida [9] noted that above 0.1 mg cm2 of deposited mercury, the hydrogen overpotential of the electrode is not improved. Around 150 mg cm2 of mercury is deposited on the copper base metal in this study; although some of this is wiped off during the preparation, a value in excess of 0.1 mg cm2 is estimated to have remained on the surface. The l-cystine/l-cysteine system differs from hydrogen evolution and uorescein reduction in one important respect. There is strong interaction between the amino acids and mercury [11]. Consequently, the disulphide reduction was examined at an SMDE of similar area to the RDE, to check that comparable performance for the reduction of l-cystine hydrochloride was observed in the cyclic voltammograms. Fig. 1 shows the main wave in cyclic voltammograms for the

reduction of dissolved l-cystine hydrochloride to l-cysteine hydrochloride: RSSR.2HCl + 2H+ + 2e = 2RSH.HCl (6)

in aqueous 0.1 mol dm3 hydrochloric acid, at both the mercury-plated copper RDE and the SMDE. Both the peak potential and the peak current are very similar at the two electrodes. In terms of the longer time-scale of the coulometry studies, the current density for hydrogen evolution is only slightly increased from 0.010 to 0.013 A m2 at a potential of 1.0 V versus SCE. Also, the main wave in the cyclic voltammograms for reduction of dissolved l-cystine hydrochloride did not change after 8 h. The evidence suggests that, on the timescale of both voltammetry and coulometry experiments, the plated surface is a mercury lm rather than an amalgam. After a period of 8 h, there may be some increase in the copper concentration of the lm but this effect does not greatly alter the hydrogen overpotential or the electrocatalytic activity of the electrode for disulphide reduction. 3.2. Controlled potential coulometry at the mercury-plated copper RDE 3.2.1. Aqueous 2.0 mol dm3 hydrochloric acid Initially, l-cystine hydrochloride concentrations appreciably above 10 mol m3 in aqueous 2.0 mol dm3 hydrochloric acid were investigated at the mercury-plated copper RDE. Potentials more negative than 1.2 V versus SCE were required to achieve reasonable rates of reduction. At such negative potentials, however, rates of reduction were variable and current efciencies no better than 80%. This is most probably an effect of adsorption of the amino acids at the mercury surface. Previous studies [11] at a much smaller area DME showed evidence of the disulphide and thiol adsorption blocking the mercury electrode surface and moving the reduction to more negative electrode potentials. At disulphide concentration of 10 mol m3 and below, current densities moved to less negative potentials and adsorption effects were not observed. It was not possible, however, to separate l-cystine reduction from hydrogen evolution using constant potential coulometry. Current efciencies were below 50% for l-cystine hydrochloride reduction and neither z nor the diffusion coefcient of l-cystine hydrochloride could be determined in aqueous 2.0 mol dm3 hydrochloric acid due to the competitive hydrogen evolution reaction. 3.3. Determination of z and the reaction order for l-cystine hydrochloride Using 10 mol m3 l-cystine hydrochloride in aqueous 0.1 mol dm3 hydrochloric acid at the mercury-plated copper RDE, a large, electrochemically irreversible wave was evident for the electrosynthesis, corresponding to electrode reaction (6). This disulphide concentration represents

Fig. 1. Cyclic voltammograms of 10 mol m3 l-cystine hydrochloride in aqueous 0.1 mol dm3 HCl at a mercury-plated copper disc electrode and an SMDE. Potential sweep rate: 100 mV s1 . Temperature: 298 K. Electrolyte deoxygenated with N2 .

T.R. Ralph et al. / Electrochimica Acta 51 (2005) 133145

137

rotation rate of 50 Hz to maximise the rate of reduction. Under these conditions, in a constant volume system, Faradays Laws can be stated as: q (7) z= c(t) FV where z is the number of electrons transferred in the overall electrode reaction, q the charge passed, c(t) is the change in the bulk reactant concentration during electrolysis and V is the electrolyte volume. Table 1 shows values for q and c(t) in addition to the concentration of l-cysteine hydrochloride produced for several disulphide electrolyses. All electrolyses were consistent with a two electron change, in agreement with electrode reaction (6). In addition, the concentrations of reactant and product conrm that 1 mol of disulphide was converted into 2 mol of thiol, within experimental error, in agreement with reaction (6).
Fig. 2. Steady state voltammograms of 10 mol m3 l-cystine hydrochloride in aqueous 0.1 mol dm3 HCl at a mercury-plated copper RDE. Electrolyte deoxygenated with N2 . The background current from the HCl electrolyte is also shown. Temperature: 298 K. Electrode rotation frequency: (1) 5 Hz, (2) 10 Hz, (3) 20 Hz, (4) 30 Hz, (5) 40 Hz and (6) 50 Hz.

the experimentally determined solubility limit in aqueous 0.1 mol dm3 hydrochloric acid. Fig. 2 shows the current density versus electrode potential plots at various electrode rotation rates. At electrode potentials, which are more positive than 0.7 V versus SCE the rate of the reduction reaction is controlled completely by the kinetics of electron transfer. The current is therefore independent of the electrode rotation rate. At more negative electrode potentials, there is an extensive region of mixed kinetic-mass transport control. At values more negative than 1.1 V versus SCE, an apparent convective-diffusion limited current is hidden by simultaneous hydrogen evolution from the background electrolyte. The plateau denition is increasingly lost at higher electrode rotation rates due to an extension of the mixed control region to progressively more negative potentials. This is a well-established effect at an RDE for electrochemically irreversible systems [12]. Constant potential coulometry at the mercury-plated copper RDE was performed with 10 mol m3 l-cystine hydrochloride in aqueous 0.1 mol dm3 hydrochloric acid at a cathode potential of 0.950 V versus SCE and an electrode

3.3.1. Mass transport control and determination of the diffusion coefcient A second series of constant potential electrolyses were performed at 1.150 V versus SCE with an electrode rotation frequency of 5 Hz. Fig. 2 shows under these conditions, lcystine hydrochloride reduction may be under complete mass transport control. Assuming all current losses are due to hydrogen evolution and the current for this secondary reaction is independent of the l-cystine hydrochloride and l-cysteine hydrochloride concentrations during electrolysis, then from Equation (7): dc(t) = q(RSSR2HCl) jL(t) dt = zFV zFV (8)

where jL(t) is limiting current density at time t. For the RDE, the time dependent limiting current density, due to convective-diffusion is given by: jL(t) = zFkm cb (9)

where km is the average mass transport coefcient. The decay of reactant concentration is expected to be rstorder, according to: km A t (10) V If l-cystine hydrochloride reduction is under complete mass transport control, a plot of ln c(t) against t ln c(t) = ln c(o)

Table 1 Constant potential coulometry at 0.950 V vs. SCE at a mercury-plated copper RDE at 298 K. t (min) 0 63 129 186 288 381 480 q (kC) 0 3.412 6.862 9.313 13.741 17.011 20.662 cRSSR2HCl (mol m3 ) 10.0 9.3 8.6 8.1 7.2 6.5 5.8 cRSHHCl (mol m3 ) 0 1.4 2.7 3.9 5.7 6.9 8.4 ct 0 0.7 1.4 1.9 2.8 3.5 4.2
(RSSR2HCl)

(mol m3 )

z 2.02 2.03 2.01 2.04 2.02 2.04

138

T.R. Ralph et al. / Electrochimica Acta 51 (2005) 133145

should be a straight line with a slope of km A/V. For an RDE, the mass transport coefcient may be expressed as [13]: km = 0.62D2/3 v1/6 1/2 (11)

3.4. Analysis of the steady state and linear sweep voltammograms for l-cystine hydrochloride reduction at the mercury-plated copper RDE In Fig. 2, the electrochemically irreversible waves at the mercury-plated copper RDE due to l-cystine hydrochloride reduction can be analysed to obtain the important electrochemical kinetic parameters associated with electron transfer. The assumption is made that the reaction is rst order at the potentials where the reaction is under mixed kinetic-mass transport control and the behaviour ts the KouteckyLevich equation: 1 1 1 = + j jk jL (12)

where D is the diffusion coefcient of the reactant, v the kinematic viscosity of the electrolyte and is the electrode rotation rate calculated using the formula = 2f, where f is the rotation rate in Hz. Using Equation (11), it is possible to determine the diffusion coefcient of the reactant from the slope of the ln c(t) against t plot. This method has the particular advantage that neither z nor cb is required in contrast with voltammetric techniques, as noted by Hitchman and Albery [14]. A plot of ln c(t) against t for l-cystine hydrochloride reduction at 1.150 V versus SCE is linear with a slope of 4.6 105 s1 . The determination of the diffusion coefcient of l-cystine hydrochloride from such a plot requires an accurate value for the kinematic viscosity (dynamic viscosity/density) of the electrolyte as shown by Equation (11). This was determined experimentally by measuring the solution density, using density bottles and the dynamic viscosity in a viscometer bath. Table 2 lists the measured physical properties of several electrolytes of importance to the kinetic studies. The kinematic viscosity of aqueous 0.1 mol dm3 hydrochloric acid was compared with available literature [15,16] to check the accuracy of the measured data. A plot of uid density versus hydrochloric acid concentration (110 wt%), constructed from reference [15], gave a solution density of 998.8 kg m3 , following a short extrapolation. From reference [16] the dynamic viscosity of aqueous 0.1 mol dm3 hydrochloric acid is 1.007 0.003, which from the listed dynamic viscosity of water (i.e., 8.949 104 kg m1 s1 ) gives the viscosity of the acid as 9.012 104 kg m1 s1 at 298 K. The literature prediction of kinematic viscosity is 9.0 107 m2 s1 in good agreement with Table 2. Using the kinematic viscosity of 10 mol m3 l-cystine hydrochloride in aqueous 0.1 mol dm3 hydrochloric acid from Table 2 the diffusion coefcient of the disulphide is (4.85 0.05) 1010 m2 s1 at 298 K. Reasonable agreement of this diffusion coefcient with the available literature values of 4.8 1010 [17] and 5.3 1010 m2 s1 [18] conrms pure mass transport control of the reduction due to convective-diffusion of the disulphide at very negative potentials. The limiting current is hidden by simultaneous hydrogen evolution from the background electrolyte.

and a mixed control equation can be written for the RDE as: 1 1 1 = + j zFkf cb 0.62z F D2/3 v1/6 cb 1/2 (13)

At a constant potential, a KouteckyLevich plot of 1/j against 1/1/2 is a straight line. The intercept (extrapolation of 1/2 to yield 1/jk ) gives the kinetic current and, therefore, kf at the selected potential. The slope allows determination of the diffusion coefcient of the reactant, although this requires knowledge of z and cb . Fig. 4 shows that at a range of electrode potentials, in the mixed control region of the irreversible waves in Fig. 2, plots of 1/j against 1/1/2 give a series of parallel straight lines. From the intercepts, kf is calculated to increase from 2.4 105 m s1 at 0.850 V versus SCE to 1.5 104 m s1 at 1.000 V versus SCE. For z = 2 and a kinematic viscosity of 9.1 107 m2 s1 (Table 2) the diffusion coefcient of l-cystine hydrochloride is calculated as (5.01 0.12) 1010 m2 s1 at 298 K. The error represents the spread of values obtained from all the straight lines in Fig. 3, including the error in the slopes determined from a least squares analysis. From the denition of the heterogeneous rate constant:
0 kj = kf exp

c F E RT c F E 2.303RT

(14)

or, in logarithmic form:

0 log kf = log kf

(15)

A plot of log kf against E is a straight line the gradient of 0 which yields the Tafel slope (2.303RT/c F) and c , while kf

Table 2 Measured physical properties of electrolytes at 298 K Electrolyte 0.1 mol dm3 HCl 10 mol m3 RSSR.2HCl in 0.1 mol dm3 HCl 2.0 mol dm3 HCl 10 mol m3 RSSR.2HCl in 2.0 mol dm3 HCl (103 kg m3 ) 1.000 1.010 1.033 1.037 (104 kg m1 s1 ) 9.0 9.2 9.5 9.75 (106 m2 s1 ) 0.90 0.91 0.92 0.94

T.R. Ralph et al. / Electrochimica Acta 51 (2005) 133145

139

against E for the voltammograms in Fig. 2 recorded at an electrode rotation frequency of 10 Hz is linear with a Tafel 0 slope of 186 2 mV decade1 ; c is 0.32 0.01 and kf is (6.7 0.5) 1010 m s1 . These values are in close accord with those obtained from the plot of log kf against E. The voltammograms for the other rotation rates in Fig. 2 gave very similar kinetic parameters. 3.5. Analysis of the linear sweep voltammetry for l-cystine hydrochloride reduction at the SMDE 3.5.1. Aqueous 0.1 mol dm3 hydrochloric acid As seen in Fig. 1, linear sweep voltammograms at the SMDE and the mercury-plated copper RDE show an identical large diffusion controlled peak for the irreversible reduction of l-cystine hydrochloride to l-cysteine hydrochloride according to reaction (6). It is possible to analyse linear sweep voltammograms to obtain the kinetic parameters for l-cystine hydrochloride reduction. In order to provide comparative kinetic data to the steady state values at the mercury-plated copper RDE, the linear sweep voltammograms were analysed at the SMDE. Nicholson and Shain [19] have shown that, for a rst-order, electrochemically irreversible reaction, assuming semi-innite linear diffusion j = zFcb D1/2 v1/2 c F RT
1/2

Fig. 3. KouteckyLevich plot of 1/j against 1/1/2 for l-cystine hydrochloride reduction at a mercury-plated copper RDE. Ten moles per cubic metre of l-cystine hydrochloride in aqueous 0.1 mol dm3 HCl deoxygenated with N2 . Temperature: 298 K. Electrode potential: (1) 0.850 V, (2) 0.875 V, (3) 0.900 V, (4) 0.925 V, (5) 0.950 V, (6) 0.975 V and (7) 1.000 V vs. SCE. The dotted lines show extrapolation of to .

1/2 (bt)

(17)

is obtained from the intercept. The values of kf from the intercepts in Fig. 3 were plotted in the form of Equation (15). The Tafel slope is 189 1 mV decade1 , c is 0.32 0.01 and 0 kf is (7.0 0.5) 1010 m s1 for l-cystine hydrochloride 0 reduction at 298 K. The relatively large error in kf mainly reects the long extrapolation to 0 V versus SCE. Alternatively, manipulation of Equations (13) and (15) followed by taking logarithms produces: log 1 c F 1 0 = E + log zFkf cb j jL 2.303RT (16)

where cb is the bulk reactant concentration, D the diffusion coefcient of the reactant, v the potential sweep rate and (bt) is a current function which is dependent upon the electrode potential [19]. Indeed, (bt) goes through a maximum at 1/2 (bt) = 0.4958 and introduction of this value into Equation (17) gives the peak current density as: jp = 0.4958zFcb D1/2 c F RT
1/2

v1/2

(18)

This value occurs when c (Ep E0 ) + RT ln F Dc Fv RT

1/2

A plot of log[1/j 1/jL ] against E is a straight line which yields from the gradient the Tafel slope and c , while the in0 tercept can be used to nd kf . The use of log[1/j 1/jL ] corrects the normal Tafel plot for the degree of mass transport control evident in the mixed control region of the voltammograms. This plot should be more accurate than the plot of log kf against E, which is derived from KouteckyLevich plots, since it uses the currents directly measured at the electrode. An accurate measurement of jL is required, however. As shown in Fig. 2, this is not available for l-cystine hydrochloride reduction at mercury-plated copper because of simultaneous hydrogen evolution. The problem may be solved by calculating the theoretical jL using Equation (5) since the kinematic viscosity of the electrolyte and diffusion coefcient of the disulphide are known. A plot of log[1/j 1/jL ]

1 k0

= 5.34 mV (19)

or Ep = E0 RT 0.780 + 2.303 log c F c Fv RT

1/2

D1/2 k0 (20)

+2.303 log

Another equation sometimes employed in kinetic analyses. From Equation (20): exp c F [Ep E0 ] RT = 2.183 Dc Fv RT
1/2

1 k0

(21)

140

T.R. Ralph et al. / Electrochimica Acta 51 (2005) 133145

Multiplying both sides by jp , using Equation (19), and rearranging gives jp = 0.227zFcb k0 exp Taking logarithms: c F [Ep E0 ] (23) 2.303RT From Equation (20), a plot of Ep against log v is a straight line, which yields the Tafel slope and c from the gradient and k0 from the intercept (provided that the diffusion coefcient of the reactant is known). Alternatively, Equation (23) indicates that a plot of log jp against [Ep E0 ] for different potential sweep rates is a straight line which yields the Tafel slope from the gradient together with c and k0 from the intercept (provided z is known). The plot of log jp against [Ep E0 ] might be expected to be the less accurate of the two plots, since it requires the measurement of both jp and 0 Ep from the voltammogram rather than just Ep . To obtain kf rather than k0 from Equations (20) and (23) merely requires a value of 0 V to be assumed for E0 . A plot of the negative shift in Ep with increasing log v in the linear sweep voltammograms for reduction of 10 mol m3 l-cystine hydrochloride in aqueous 0.1 mol dm3 hydrochloric acid at 298 K was linear. From the slope of the straight line, the Tafel slope is 185 3 mV decade1 and c is 0.32 0.01; from the 0 intercept, kf is (6.7 0.8) 1010 m s1 (assuming D 10 m2 s1 ). The corresponding plot of log j is 5.0 10 p against [Ep E0 ] is also linear with a Tafel slope of 185 5 mV decade1 and an c value of 0.32 0.02 and, 0 from the intercept, a kf of (0.8 1.4) 1010 m s1 (assuming z is 2). Table 3 compares the kinetic parameters at the SMDE and the mercury-plated copper RDE. The close comparison amongst the measured kinetic parameters provides strong evidence for the reliability of the measurements and for the integrity of the mercury-plated copper RDE surface. The diffusion coefcient of l-cystine hydrochloride can also be obtained from the linear sweep voltammetry by using Equation (18). A plot of jp against v0.5 is a straight line through the origin, indicating diffusion control. From the slope of this plot, assuming z is 2 and c is 0.32, the diffusion coefcient of the disulphide is (5.1 0.4) 1010 m2 s1 at 298 K. This is in close agreement with measurements at a log jp = log(0.227zFcb k0 )
Table 3 Kinetic parameters from RDE studies and linear sweep voltammetry Kinetic parameter

mercury-plated copper RDE in this paper and with those available in the literature [17,18]. 3.5.2. Aqueous 2.0 mol dm3 hydrochloric acid electrolyte Linear sweep voltammograms at the SMDE offered a key advantage in kinetic studies. It allowed determination of the kinetic parameters in more concentrated aqueous 2.0 mol dm3 hydrochloric acid. In contrast to the lack of an electrochemically irreversible wave in steady state voltammetry at the mercury-plated copper RDE, linear sweep voltammetry at the SMDE showed a diffusion controlled peak. This difference is probably the result of two effects. The peak potential for the reduction in linear sweep voltammograms is unchanged in 0.1 and 2.0 mol dm3 hydrochloric acid and occurs between 0.86 and 0.98 V versus SCE. This is positive of the hidden diffusion limited current in steady state voltammograms at the mercury-plated copper RDE. Also, the current increases over 10 s to a steady value in the steady state measurements. Consequently, on the time scale of linear sweep voltammetry experiments, the hydrogen evolution rate is probably lower at a given electrode potential. Analysis of the diffusion controlled peak in linear sweep voltammograms using plots of Ep against log v and log jp against Ep for reduction of 10 mol m3 l-cystine hydrochloride in aqueous 2.0 mol dm3 hydrochloric acid indicates the kinetic parameters are very similar to those measured in aqueous 0.1 mol dm3 hydrochloric acid. The diffusion coefcient of l-cystine hydrochloride is, however, lower in the more concentrated acid, as evidenced by a lower peak current at a given potential sweep rate. In aqueous 2.0 mol dm3 hydrochloric acid from Equation (18), assuming z is 2 and c is 0.32, the diffusion coefcient of l-cystine hydrochloride is (4.0 0.3) 1010 m2 s1 at 298 K. 3.6. Comparison of diffusion coefcients The diffusion coefcients of l-cystine hydrochloride in aqueous hydrochloric acid determined from the voltammetry studies are listed in Table 5. In aqueous 0.1 mol dm3 hydrochloric acid at 25 C, the diffusion coefcients are in accord with the available literature values of 4.8 1010 [17] and 5.3 1010 m2 s1 [18]. The literature values were measured by polarography using the Ilkovic equation. Of the techniques used in this study, it is generally accepted [20] that the RDE is capable of determining diffusion coefcients with

c F [Ep E0 ] RT

(22)

RDE voltammetry at a mercury-plated copper RDE log kf vs. E log[1/j 1/jL ] vs. E 186 2 0.32 0.01 6.7 0.5

Linear sweep voltammetry at an SMDE Ep vs. log v 185 3 0.32 0.01 6.8 1.4 log jp vs. [Ep E0 ] 185 5 0.32 0.02 6.8 1.4

Tafel slope c kf (1010 m s1 )

(mV decade1 )

189 2 0.32 0.01 7.0 0.5

Ten moles per cubic metre l-cystine hydrochloride in aqueous 0.1 mol dm3 HCl at 298 K.

T.R. Ralph et al. / Electrochimica Acta 51 (2005) 133145 Table 4 0 Heterogeneous rate constants (kf ) for the reduction of l-cystine hydrochloride to l-cysteine hydrochloride at 298 K Electrode DME [11] Electrolyte 0.1 mol m3 RSSR in 0.1 mol dm3 HCl Technique Sampled dc polarography log kf vs. E log[1/j 1/jL ] vs. E Linear sweep voltammetry Ep vs. log v Steady state voltammetry log kf vs. E log[1/j 1/jL ] vs. E Linear sweep voltammetry Ep vs. log v Steady state voltammetry log kf vs. E log [1/j 1/jL ] vs. E Linear sweep voltammetry Ep vs. log v
0 kf (m s1 )

141

2.3 0.3 107 2.3 0.2 107 1.9 0.2 107 7.0 0.5 1010 6.7 0.5 1010 6.7 0.8 1010 2.2 0.3 1011 2.6 0.3 1011 1.7 0.2 1011

Mercury-plated copper RDE SMDE Lead RDE

10 mol m3 RSSR in 0.1 mol dm3 HCl 10 mol m3 RSSR in 0.1 mol dm3 HCl or 2.0 mol dm3 HCl 10 mol m3 RSSR in 0.1 mol dm3 HCl

a precision of better than 1%, while linear sweep voltammetry often produces values no better than 20%, in practice. The poor performance of linear sweep voltammetry has been attributed not only to the difculty of accurately measuring the peak current, but to the uncertainty in the theoretical equation for this current. Linear sweep voltammetry gives an error of 8%, while the RDE gives an error of 2.5% from KouteckyLevich plots, and 1% from constant potential coulometry studies. The diffusion coefcients in both 0.1 and 2.0 mol dm3 hydrochloric acid obtained from sampled dc polarography by us [11] are slightly higher than those from the above techniques, as shown in Table 5. The latter utilised, however, a 100-fold lower concentration of l-cystine hydrochloride. The increased value of diffusion coefcient probably reects the effect of this reduced concentration of disulphide. 3.7. Signicance of the kinetic parameters The Tafel slope and c value are in very close accord with the values measured in our laboratories at the DME [11]. The high Tafel slope and low c are attributable to the effect of adsorption of the disulphide at the mercury surface, which has an increasingly negative impact on the rate of reduction at progressively more negative potentials. Similar kinetic parameters were also measured at lead [3]. In contrast to mercury, there was no direct evidence of adsorption of the disulphide at lead. The heterogeneous rate constant for electron transfer is, however, much smaller at the SMDE than the value recorded at the DME and much closer to the corresponding values mea0 sured at the lead RDE [3]. Table 4 shows the comparative kf values. The main difference between the SMDE and the DME studies is the much higher disulphide concentrations used at the SMDE. Consequently, the effect of changing the disulphide concentration on the linear sweep voltammograms at the SMDE and the steady state measurements at the mercuryplated copper RDE was considered. 3.8. Concentration dependence of the kinetic parameters Fig. 4 shows the mass transfer corrected Tafel plots recorded at the mercury-plated copper RDE for l-cystine hy-

drochloride concentrations between 0.4 and 10 mol m3 in aqueous 0.1 mol dm3 hydrochloric acid. At all disulphide concentrations, the Tafel slope is 186 3 mV decade1 and c is 0.32 0.01. From the Tafel plots, it is possible to obtain the reaction order for l-cystine hydrochloride, by plotting log[1/j 1/jL ] against log(l-cystine hydrochloride concentration) at a constant cathode potential. Such plots, at a series

Fig. 4. Mass transport corrected Tafel plots for l-cystine hydrochloride reduction at a mercury-plated copper RDE at 10 Hz. The background electrolyte is aqueous 0.1 mol dm3 HCl deoxygenated with N2 at 298 K. lCystine hydrochloride concentration: (+) 0.4 mol m3 , ( ) 0.6 mol m3 , ( ) 0.8 mol m3 , ( ) 1 mol m3 , ( ) 2 mol m3 , ( ) 4 mol m3 , ( ) 6 mol m3 , ( ) 8 mol m3 and ( ) 10 mol m3 . The background current has been subtracted.

142

T.R. Ralph et al. / Electrochimica Acta 51 (2005) 133145

1.0 and 0.6 mol m3 l-cystine hydrochloride give from 0 the intercept a kf value of (8.4 0.7) 1010 and 10 m s1 , respectively (assuming a diffu(10.1 0.8) 10 sion coefcient of 5.0 1010 m2 s1 ) in accordance with Table 5. This shows the importance of measuring the reaction order by varying the reactant concentration, which is the prescribed method [21], rather than relying upon methods, which utilise only a single reactant concentration. Over a wide range of disulphide concentration, the KouteckyLevich plot, mass transport corrected Tafel plot and linear sweep voltammo0 gram suggest the reaction is rst-order. This increase in kf at lower disulphide concentrations also explains the discrepancy shown in Table 4 in the heterogeneous rate constants measured at the DME and the SMDE or the mercury-plated copper RDE. 3.9. Effect of adsorption of l-cystine hydrochloride at the mercury surface on the electrode kinetics The kinetic measurements at the mercury-plated copper RDE and the SMDE are consistent with the more detailed studies at the DME [1] and at a lead RDE [3] in terms of the mechanism of l-cystine hydrochloride reduction. The rst electron transfer to the thiol controls the rate of the reduction with proton transfer either preceding or occurring simultaneously with the rst electron transfer:
Fig. 5. Plot of log[1/j 1/jL ] against log(RSSR2HCl concentration); 0.4 mol m3 < cRSSR2HCl < 10 mol m3 . Potential: (1) 0.950 V, (2) 0.925 V, (3) 0.900 V, (4) 0.875 V, (5) 0.850 V and (6) 0.825 V vs. SCE. The dotted line shows a unity slope (reaction order of one) for the results at 0.825 V vs. SCE.

RSSR.2HCl + H+ + e = RSH.HCl + RS HCl

rds

(24) (25)

RS HCl + H+ + e = RSH.HCl

of cathode potentials, are shown in Fig. 5. Above a concentration of 4 mol m3 l-cystine hydrochloride, the reaction order is +1 but at lower disulphide levels the kinetic currents are progressively higher than predicted for a rst-order reaction. 0 This is consistent with the progressively higher kf values for l-cystine hydrochloride reduction as the reactant concentra0 tion decreases. kf values determined from the intercepts of the Tafel plots in Fig. 4 range from 6.6 1010 to 14.8 1010 over the disulphide concentration range 10 to 0.4 mol m3 . 0 Conrmation of the kf values recorded at the mercuryplated copper RDE is provided by linear sweep voltammetry at the SMDE. Plots of Ep against log v for

The role of disulphide adsorption on the electrochemical kinetics of the rate-determining step (rds) has been interrogated using cyclic voltammetry at the HMDE [11] and the SMDE. In agreement with the ndings of Stankovich and Bard [2] in a buffer solution of pH 7.4, there is a clear peak for the reduction of adsorbed l-cystine hydrochloride at approximately 0.25 V versus SCE in aqueous 0.1 mol dm3 hydrochloric acid according to: RSSR.2HCl(ads) + 2H+ + 2e = 2RSH.HCl (26)

At the DME, at l-cystine hydrochloride concentrations of 0.0250.1 mol m3 , the adsorption was shown to t the Langmuir adsorption isotherm and from a linear plot of 1/ versus 1/cb (where is the surface excess of adsorbent) the saturation

Table 5 Diffusion coefcients of l-cystine hydrochloride in aqueous hydrochloric acid at 298 K obtained from kinetic studies Technique Constant potential coulometry mercury-plated copper RDE KouteckyLevich plot mercury-plated copper RDE Linear sweep voltammetry SMDE Sampled dc polarography DME [11] Assumptions required None z=2 z = 2 and c = 0.32 z = 2 and c = 0.32 z=2 z=2 Electrolyte (mol dm3 HCl) 0.1 0.1 0.1 2.0 0.1 2.0 D (1010 m2 s1 ) 4.85 0.05 5.02 0.12 5.1 0.4 4.0 0.3 5.3 0.3 4.2 0.2

T.R. Ralph et al. / Electrochimica Acta 51 (2005) 133145

143

Fig. 6. Cyclic voltammetry of 10 mol m3 l-cystine hydrochloride in aqueous, 0.1 mol dm3 HCl deoxygeneated with N2 at 298 K, showing disulphide adsorption peaks at a SMDE. Potential sweep rate: 100 mV s1 .

coverage, s , was estimated as (2.7 0.8) 107 mol m2 . Based on this saturation coverage over the concentration range of 0.0250.1 mol m3 the surface coverage increases from 0.01 to 0.68. Examination of much higher disulphide concentrations at the larger area SMDE is interesting. Between 4 and 10 mol m3 l-cystine hydrochloride, a large current spike is present on the peak for reduction of the adsorbed disulphide as shown in Fig. 6. This is most likely due to rearrangement of l-cystine hydrochloride on the electrode surface after saturation coverage. This behaviour has been observed by Stankovich and Bard [2] for the reversible oxidation of the thiol to mercury cysteinate, at pH 7.4, as shown in electrode reactions (3) and (4). Determination of the charge passed under the adsorption peak (ignoring the current spike) yields a saturation coverage of 3.1 107 mol m2 in accord with the value derived at the DME [11]. At disulphide concentrations from 0.4 to 0.6 mol m3 the current spike is removed from the adsorption peak and the surface coverage is calculated to vary from 0.90 to 1. At these high coverages, departures from the Langmuir adsorption isotherm are observed because of the interaction between adjacent disulphide molecules on different sites.

At the DME, in the presence of very low concentrations of disulphide (0.0250.1 mol m3 ) and a range of surface coverages (0.01 < < 0.68) the reduction of solution phase l-cystine hydrochloride is rst-order. At intermediate disulphide concentrations (0.40.6 mol m3 ) and high surface coverages (0.9 < < 1.0) the reaction is not rst-order and becomes progressively more sluggish as the disulphide concentration increases. At much higher disulphide concentrations (0.610 mol m3 ) equivalent to saturation coverage ( = 1.0) of the SMDE, the reduction of solution phase l-cystine hydrochloride is again rst-order. This effect might not be expected since the adsorbed disulphide is reduced prior to the reduction wave for the electrosynthesis reaction. The most likely explanation is further adsorption of the disulphide at electrode potentials on the main irreversible wave, with the adsorption strength sufcient to prevent reduction until potentials negative of the main irreversible wave are reached. Any adsorption post-peak may be hidden by copious hydrogen evolution from the background electrolyte. Strongly adsorbed reactants are known to give a post-peak in cyclic voltammograms [22]. There are a number of observations that support this argument. The constant potential coulometry studies show that the little reduction of adsorbed disulphide at potentials close to the peak potential in the linear sweep voltammograms for the diffusion controlled reduction of dissolved l-cystine hydrochloride. The disulphide molecule has a double positive charge in acid electrolyte, which, purely on an electrostatic basis, suggests adsorption will be stronger at more negative potentials. At the DME [11] using high disulphide concentrations, there is evidence of strong adsorption phenomena in voltammograms and the irreversible wave for reduction of dissolved l-cystine hydrochloride is shifted to more negative 0 potentials. If this explanation is correct, the unchanged kf for saturation coverage of the electrode with disulphide indicates reduction of dissolved l-cystine hydrochloride occurs through the adsorbed disulphide lm, which reduces the rate of electron transfer.

4. Conclusions 1. The reduction of dissolved l-cystine hydrochloride to lcysteine hydrochloride is chemically and electrochemically irreversible at mercury. In aqueous 0.1 mol dm3 hydrochloric acid, there is no evidence of alternative products and the only current losses are due to hydrogen evolution from the background electrolyte. In contrast, to steady state measurements at a DME, which show a diffusion limited current starting at approximately 0.5 V versus SCE, which is well separated from hydrogen evolution, with higher disulphide concentrations at a mercury-plated copper RDE, the reduction wave shifts cathodically and the limiting current at above ca. 1.1 V versus SCE is hidden by hydrogen evolution.

144

T.R. Ralph et al. / Electrochimica Acta 51 (2005) 133145

2. The principal reason for the large shift in the potentials required for reduction of solution phase l-cystine hydrochloride is strong adsorption of l-cystine hydrochloride at the electrode surface, through which the reduction reaction must occur. At surface coverages of 0.92 < < 1.0 the kinetics of the rst electron transfer to the disulphide molecule is hindered although the symmetry factor is unaltered. For = 1, the adsorbed disulphide layer has no additional effect upon the kinetics of the electrosynthesis reaction. 3. For the conditions most closely resembling those used in electrosynthesis, the Tafel slope is 185 mV decade1 , c 0 is 0.32 and kf is 6.7 1010 m s1 . Linear sweep voltammetry at the SMDE shows the kinetic parameters are very similar in moving from 0.1 to 2.0 mol dm3 hydrochloric acid. Corresponding steady state measurements at the mercury-plated copper RDE could not separate l-cystine hydrochloride reduction from hydrogen evolution in the more concentrated acid, highlighting the advantage of linear sweep measurements in this case. 4. The mechanism of the electrosynthesis reaction proposed at lead electrodes [3] and at the DME [11] is similar to that found in the studies at the SMDE and the mercuryplated copper RDE. The rst electron transfer to l-cystine hydrochloride is the RDS with the Tafel slope and c modied by disulphide adsorption at mercury. 5. The diffusion coefcient for l-cystine hydrochloride in aqueous hydrochloric acid has been obtained from constant potential coulometry and from a KouteckyLevich analysis of the steady state voltammetry at the mercuryplated copper RDE. The coulometry studies have the advantage of not requiring either the number of electrons or the bulk reactant concentration. Linear sweep voltammetry at the SMDE of HMDE provided less precise diffusion coefcients.

jp 0 kf k0 kf km q r R t T v V z

peak current density (A m2 ) heterogeneous rate constant (at 0 V versus SCE) (m s1 ) heterogeneous rate constant (at E0 ) (m s1 ) heterogeneous rate constant for forward (cathodic) process (m s1 ) mass transport coefcient (m s1 ) electrical charge (C) radius of mercury drop (m) molar gas constant (8.314 J K1 mol1 ) time (s) temperature (K) potential sweep rate (V s1 ) electrolyte volume (m3 ) number of electrons in the electrode process

Greek letters c cathodic transfer coefcient adsorption coefcient dynamic viscosity (kg m1 s1 ) uid density (kg m3 ) kinematic viscosity (m2 s1 ) (bt) function dened by Equation (17) surface coverage surface excess of adsorbate saturation coverage value of electrode rotation rate (rad s1 )

References
[1] T.R. Ralph, M.L. Hitchman, J.P. Millington, F.C. Walsh, J. Electroanal. Chem. 375 (1994) 1. [2] M.T. Stankovich, A.J. Bard, J. Electroanal. Chem. 75 (1977) 487. [3] T.R. Ralph, M.L. Hitchman, J.P. Millington, F.C. Walsh, J. Electroanal. Chem., in press. [4] Southampton Electrochemistry Group, Instrumental Methods in Electrochemistry, Ellis Horwood Ltd., Chichester, 1985 (Chapter 4). [5] P.J. Daly, D.J. Page, R.G. Compton, Anal. Chem. 55 (1983) 1191. [6] A. Arvia, S. Machiano, J.J. Podesta, Electrochim. Acta 12 (1967) 259. [7] D.D. Perrin, B. Dempsey, Buffers For pH and Metal Ion Control, Chapman and Hall Ltd., London, 1974. [8] T.R. Ralph, M.L. Hitchman, J.P. Millington, F.C. Walsh, J. Electrochem. Soc. 52 (2005) D54. [9] S. Yoshida, Bull. Chem. Soc. Jpn. 54 (1981) 562. [10] B.A. Coles, R.G. Compton, J. Electroanal. Chem. 144 (1983) 87. [11] T.R. Ralph, M.L. Hitchman, J.P. Millington, F.C. Walsh, J. Electroanal. Chem., submitted for publication. [12] C.M.A. Brett, in: C.H. Bamford, R.G. Compton (Eds.), A.M.C.F.O. Brett in Comprehensive Chemical Kinetics, vol. 26, Elsevier, Amsterdam, 1986 (Chapter 5). [13] V.G. Levich, Physiochemical Hydrodynamics, Prentice Hall, New Jersey, 1962. [14] M.L. Hitchman, W.J. Albery, Electrochim. Acta 17 (1972) 787. [15] International Critical Tables of Numerical Data Physics, in: E.W. Washburn, C.J. West, N.E. Dorsey, F. Bichowsky, M.D. Ring (Eds.), Chemistry and Technology, vol. 3, McGraw-Hill, New York, 1929, p. 54.

Appendix A. Nomenclature electrode area (m2 ) cathodic Tafel slope (V decade1 ) bulk reactant concentration (mol m3 ) initial reactant concentration (mol m3 ) reactant concentration at time t (mol m3 ) diffusion coefcient of reactant (m2 s1 ) electrode potential (V) half-wave potential, (V) peak potential (V) standard formal potential (V) rotation frequency of RDE (s1 ) Faraday constant (96,485 C mol1 ) current density (A m2 ) limiting current density (A m2 ) limiting current density at time t (A m2 ) charge transfer current density (A m2 )

A bc cb c(0) c(t) D E E1/2 Ep E0 f F j jL jL(t) jk

T.R. Ralph et al. / Electrochimica Acta 51 (2005) 133145 [16] International Critical Tables of Numerical Data Physics, in: E.W. Washburn, C.J. West, N.E. Dorsey, F. Bichowsky, M.D. Ring (Eds.), Chemistry and Technology, vol. 5, McGraw-Hill, New York, 1929, p. 10 and 12. [17] I.M. Issa, A.A. Samahy, R.M. El Issa, Y.M. Temerik, Electrochim. Acta 17 (1972) 1615. [18] I.M. Kolthoff, C. Barnum, J. Am. Chem. Soc. 63 (1941) 520.

145

[19] R.S. Nicholson, I. Shain, Anal. Chem. 36 (1964) 706. [20] R.N. Adams, Electrochemistry at Solid Electrodes, Marcel Dekker, New York, 1969 (Chapter 8). [21] J.OM. Bockris, A.K.N. Reddy, Modern Electrochemistry, vol. 2, Plenum, 1973, p. 1008. [22] R.J. Wopschall, I. Shain, Anal. Chem. 39 (1967) 1514.