Electrooxidation of thiocyanate on the copper-modified gold electrode

Analyst, June 1998, Vol. 123 (1359–1363)

1359

Electrooxidation of thiocyanate on the copper-modified gold electrode and its amperometric determination by ion chromatography Innocenzo G. Casella* , Maria R. Guascito and Giuseppe E. De Benedetto Dipartimento di Chimica, Universita’ degli Studi della Basilicata, Via N. Sauro 85, 85100 Potenza, Italy,

Cyclic voltammetry was used to investigate the electrochemical behavior of an Au/Cu electrode towards the electrooxidation of thiocyanate ion in alkaline medium. The effects of pH, copper loading, scan rate and applied potential on the electrocatalytic oxidation of thiocyanate have been investigated. Flow injection experiments and ion-chromatography (IC) were performed to characterise the electrode as an amperometric sensor for the thiocyanate determination. The effects of carbonate concentration and common interferents on the retention time were also estimated. The electrode stability, precision, limit of detection and linear range were evaluated at a constant applied potential of 0.7 V vs. Ag/AgCl. Calibration plots, obtained in IC, were linear from 1.0 to 195 mm (correlation coefficient of 0.9984). The detection limit (LOD) was 0.5 mm (29 ppb) in a 50 ml injection. An example of analytical application, which includes the IC separation and detection of thiocyanate ion present in human urine, is given. Keywords: Copper-modified gold electrode; thiocyanate ion determination; alkaline media; ion-exchange chromatography; urine

The thiocyanate ion is present in humans as a result of the digestion of some vegetables and as a metabolic product of compounds in tobacco smoke containing cyanide.1,2 In this respect, the concentration level of thiocyanate is considered to be a good probe to distinguish between smokers and nonsmokers. Thiocyanate is also known to block iodine uptake by the thyroid gland. Determinations of thiocyanate in serum have been advocated to monitor therapy with sodium nitroprusside. The concentration level of thiocyanate is also considered to be important to assure the quality of vegetables: in fact antinutritional effects of glucosinolates, processing effects on oilseed rape and on cruciferous vegetables rich in indol-3-ylmethylglucosinolates could be revealed by its presence.3 A large number of methods have been developed for the determination of this ion in several matrices. Spectrophotometric techniques based on the formation of a red complex with FeIII, molecular fluorescence based on the modified Köning method,4 generation of carbonyl sulfide by the SCN2 ion and its detection by gas-phase molecular absorption spectrometry5 have been developed. Ion chromatography (IC) with a conductivity detector is widely used as methodology for the separation and detection of the common anions in aqueous samples.6–8 Although ion exchange techniques can easily separate thiocyanate from a host of common anions, a long time per sample is required (retention time of thiocyanate is about 20–25 min). In this respect, separations of inorganic anions were achieved using stationary phases of zwitterionic micelles on a reversed-phase octadecyl silica (ODS)-packed column.9

Recently, interesting methods for the separation of anions by high-performance capillary electrophoresis (HPCE) with indirect or direct photometric detection have been described.10–12 In other reports, membrane-coated carbon rod ion-selective electrodes or anion-selective membranes of the anti-Hofmeister type [i.e., Cu–poly(dithiooxamide) films] have been used as detectors in potentiometric flow analysis for thiocyanate determination.13,14 The direct monitoring of activities of the selected species without any sample pre-treatment was feasible, but sometimes significant interferences by other anions present in high concentration were found. Ion chromatography combined with electrochemical detection (EC) appears a very attractive analytical approach to the determination of electroactive anions owing to its high sensitivity, selectivity and wide dynamic range.15–18 Generally, the determination of thiocyanate at bare metal electrodes has been severely limited by the poisoning caused by its oxidation products, so both cleaning and regeneration steps are needed to obtain stable amperometric responses.19 For this reason other transition metal-based electrodes have been proposed as electrode materials which were used as amperometric sensors at fixed applied potential. In addition, chemically modified electrodes (CMEs) containing surface-confined chemical functionalities have shown some advantages over conventional bare electrode in terms of catalytic activity, selectivity and protection from fouling effects. For example, platinum electrodes modified by adsorption of iodine and coated with cellulose acetate or glassy carbon electrodes modified by anodization in a RuCl3/ K4Ru(CN)6 mixture are used as stable amperometric sensors for thiocyanate in flow systems.20 A good physical dispersion of the catalytic centres on an inert electrode surface leads to a highly catalytic activity suitable for efficient electrocatalysis. In this respect, we have recently proposed a stable modified bimetallic electrode used as an amperometric sensor for the detection of carbohydrates,21 prepared by electrodeposition of copper onto a gold surface. It demonstrated that the favourable combination of gold and copper leads to a sensing electrode with strong catalytic activity over a wide range of potentials. Indeed, in alkaline medium a synergic catalytic contribution of gold and copper towards the electrooxidation of glucose was observed. Moreover, good mechanical stability under flow forced hydrodynamic conditions was found also upon changing different detection modes and considering that gold and copper electrode materials have been successfully employed as sensing probes for the determination of many polar aliphatic compounds in ion-exchange chromatography,22–24 their combination represents a very attractive CME design for analytical applications. In this article a bimetallic gold/copper electrode (Au/Cu) and its use as an electrochemical sensor for the determination of thiocyanate in alkaline medium both in flow injection analyses and in ionic chromatography are described. The amperometric sensor has been applied to the determination of thiocyanate in human urine.

1360

Analyst, June 1998, Vol. 123

Experimental Reagents All solutions were prepared from analytical-reagent grade chemicals without further purification and using deionized and doubly distilled water. Individual 0.200 m sodium thiocyanate (Aldrich, Steinheim, Germany) solutions were prepared in distilled water. The relevant solution was standardised by the Volhard method and stored in the dark. Appropriate dilutions were performed with 50 mm NaOH solutions. Unless otherwise specified, experiments in cyclic voltammetry were performed using 50 mM NaOH as background electrolyte. In voltammetry the solutions were protected from oxygen by purging with highpurity nitrogen. All experiments were carried out at ambient temperature. Apparatus A potentiostat/galvanostat Model 273 Princeton Applied Research (PAR EG&G, Princeton, NJ, USA) was used for electrochemical measurements. Cyclic voltammetry (CV) was done in a three-electrode cell using the Au/Cu bimetallic electrode, a SCE (4 m KCl) reference electrode and a platinum foil counter-electrode. The gold electrode used in CV (geometric area, 0.125 cm2) was purchased from PAR. All current densities in this paper are quoted in terms of mA cm22 of apparent geometric area. Amperometric measurements in flowing streams were performed using a PAR Model 400 electrochemical detector and a flow-through thin-layer electrochemical cell consisting of Au/ Cu as working electrode, an Ag/AgCl (4 m KCl) reference electrode, and a stainless steel counter electrode. A Servogor 120 BBC was used to record the output signal. Flow injection experiments were carried out with a Varian 2510 pump (Palo Alto, CA, USA) equipped with a Model 7125 Rheodyne injector (Cotati, CA, USA) using a 50 ml sample loop. The mobile phase was purged from oxygen with an on-line degassing system (Hewlett Packard Series 1050). Chromatographic separations were performed with a IonPac AS12A (Dionex, Sunnyvale, CA, USA) anion-exchange column (200 3 4 mm ID) coupled with the IonPac AS12A (50 3 4 mm) guard column, using 50 mm NaOH plus 0.1 m Na2CO3 as the mobile phase. Electrode preparation Before each electrode modification, copper particles were removed from the gold surface by electrode sonication for 5 min in concentrated nitric acid and washed with distilled water. The electrodeposition of copper on gold was accomplished at 20.3 V vs. SCE for 90 s in a 50 mm CuSO4 solution. Subsequently, the modified electrode was conditioned by continuous CV cycling between the scan limits 20.3 and +0.8 V for 10 min at a scan rate of 100 mV s21 in 50 mm NaOH electrolyte. The amount of copper, calculated from the total charge passed during the deposition process, was estimated by assuming that the reduction of Cu2+ to Cu0 is 100% efficient.

with the CuIII/CuII transition, while the cathodic peak (IIIc) at 20.18 V is due to the oxygen reduction. As it can be seen, the voltammogram of the Au/Cu electrode is the weighted sum of the electrochemical properties of the bare gold and copper electrodes in alkaline medium.25–28 In the presence of 3 mm thiocyanate (solid curve), a new oxidation peak (IIa) at  0.6 V appeared and between 0.5 V and 0.8 V the electrooxidation process is always operative in positive and negative potential directions. It is interesting to observe that at an unmodified gold electrode no oxidation currents were observed in the presence of thiocyanate ions at potentials greater than 0.6 V. Thus, the oxidation currents between 0.6 V and 0.8 V are related to the copper loading. The currents of peaks Ia, Ic and IIIc were lowered by increasing the thiocyanate concentration. This last behavior is possibly explained assuming an adsorption process of thiocyanate on the gold substrate before its oxidation. It is likely that the thiocyanate ions are strongly adsorbed on the catalytic sites and a partial replacement of oxygen and/or hydroxyl species follows. The peak current IIa increases linearly on increasing the thiocyanate concentration in the range 0.5–10 mm (correlation coefficient > 0.998). Moreover, the anodic peak potential (IIa), in the presence of increasing concentrations of thiocyanate, shifts to higher potentials by about 20 mV mm21. The reaction order with respect to thiocyanate concentration has been obtained from the slope of log(concentration of thiocyanate) vs. log(current density) in the range of potentials between 0.55 and 0.65 V. In the concentration range of thiocyanate between 0.2 and 25 mm, the slope is 0.76. The fractional reaction order obtained indicates that adsorption of analyte and/or inhibition by a side product can play a role in the reaction mechanism. Anodic currents were evaluated in the range of potentials between 0.4 V and 0.8 V in the presence of 1.6 mm thiocyanate; an increase with increasing scan rate (v) was observed. In particular, plot of the log(peak current IIa) vs. log(scan rate) gave a straight line whose slope was 0.71 (correlation coefficient > 0.995). Furthermore, the slope obtained from the linear relationship between the peak potential (IIa) and log (v) was about 118 mV per decade. These findings suggested that the electrochemical process was not an ideal diffusion-limited step, but electrocatalysis involved a relatively slow adsorption of the thiocyanate species on the Au/Cu electrode. In order to elucidate some aspects of the thiocyanate oxidation at Au/Cu electrode, the effects of changing the lower potential limit on cyclic voltammograms were investigated. In Fig. 2 the relevant results are presented; it can be seen that as the lower potential limit was increased, the anodic peak IIa disappeared completely (curve B). This behavior is similar to

Results and discussion Cyclic voltammetry measurements A representative cyclic voltammogram in 50 mm NaOH of a bimetallic Au/Cu electrode, having about 25 mg cm22 copper loading is reported in Fig. 1 (dashed curve). The voltammetric profile of the Au/Cu electrode reveals a cathodic peak (Ic) at 0.05 V vs. SCE during the cathodic scan, corresponding to the formation of Au0; the re-formation of gold oxide is associated at the small and broad anodic wave (Ia) at  0.3 V. During the cathodic scan a small cathodic wave (IIc) at 0.6 V is associated

Fig. 1 Steady-state cyclic voltammogram (10th cycle) at an Au/Cu electrode in the presence of 3 mm thiocyanate (solid line). The dashed line corresponds to a voltammogram run in blank 50 mm NaOH. Scan rate, 100 mV s21.


that observed on the platinum electrode in acidic medium,29 where the electrooxidation of thiocyanate ions proceeds via a preliminary dissociation/adsorption step: SCN2 ) Sads + CN2

(1)

and adsorbed sulfur undergoes a subsequent concerted oxidation: Sads + 4H2O ) SO422 + 8H+ + 6e2

(2)

On the contrary, considering that in this case between 0.65 V and 0.8 V the oxidation currents are practically independent of the lower potential limit, a probable direct adsorption/oxidation of thiocyanate on the active catalytic sites (i.e., CuOOH) is operative: 4CuOOH + SCN2 ) 4CuO + SO422 + CN2 + 4H+ + 2e2 (3) However, details of the electrooxidation mechanism and relevant kinetics of this process are subject to further specific investigations. Effect of hydroxide concentration It is well known that an alkaline medium is required to enhance the electroactivity of gold and copper electrode for the electrooxidation of several organic compounds.24–28 Therefore, the effect of hydroxide concentration on the electrode signal was investigated by cyclic voltammetry. Solutions of 3 mm thiocyanate maintained at a constant ionic strength (1 m NaClO4) were used and their pH was varied between 11.6 and 13.6. In Fig. 3 some representative cyclic voltammograms of a Au/Cu electrode obtained at different pH values are shown. As

1361

it can be seen, the peak potential IIa shifts to more negative potentials on increasing pH, the plot Ep vs. pH is linear and the relevant slope is 2188 mV pH21 (correlation coefficient > 0.997). Moreover the peak current IIa decreases with increasing pH (about 120 mA pH21). On the contrary, the oxidation currents in the range of potentials between 0.6 V and 0.8 V showed a significant increase with increasing pH (about 180 mA pH21). Although the maximum sensitivity was obtained using a background electrolyte containing high hydroxide concentrations (in the range of potentials between 0.6 V and 0.8 V), a solution containing 50 mm NaOH (i.e., pH 12.7) was usually chosen both to maintain a low background current and not to compromise analytical sensitivity. FIA measurements The Au/Cu electrode was tested as an amperometric sensor in flowing streams using a conventional thin-layer cell configuration. In accordance with the CV behavior, an increase in the applied potential results in an amperometric signal increase with a good catalytic activity at potentials higher than 0.7 V. The background current shows a nearly constant value in the potential range between 0.5 V and 0.7 V, while at higher potentials the residual current increases drastically. Consistent with the last observations, an applied potential of 0.7 V was chosen for detection of thiocyanate in flow injection analysis and chromatographic separations. In fact it represented the best compromise between maximum analytical response and background current noise. The effect of flow rate on detector response was tested by flow injection analysis. Consecutive injections of 15 mm thiocyanate solution showed that peak current is strongly dependent upon flow rate with a sharp decrease in current on increasing the flow rate from 0.3 to 3.0 ml min21. The observed trend may be due to the slow rate of the catalytic reaction between the modified electrode and analyte. This result agrees with those reported in the previous section (voltammetry), where the sorption/desorption of species rather that mass transport processes are the rate determining steps. Effect of copper loading

Fig. 2 Influence of the lower potential limit on the oxidation currents. Potential limit: A, 20.3 V vs. SCE; B, 0.2 V vs. SCE. Thiocyanate concentration, 3 mm. Experimental conditions as in Figure 1.

Fig. 3 Effects of pH on current density–potential curves in a solution of constant ionic strength (1 M NaClO4). Experimental conditions as in Fig. 1.

The influence of copper loading on catalytic activity was examined in flow injection conditions. The flow rate of the mobile phase, 50 mm NaOH, was 1.0 ml min21 and the Au/Cu electrode was maintained at 0.7 V (vs. Ag/AgCl). Fig. 4 shows the amperometric response of sensors with different copper loading by injections of a 25 mm thiocyanate solution. As can be seen, the current increases with copper loading up to a

Fig. 4 Amperometric response on loading copper catalyst for 25 mm thiocyanate. The peak currents were determined in FIA at 0.7 V vs. Ag/ AgCl. Mobile phase, 50 mm NaOH; flow rate, 1.0 ml min21; injection volume, 50 ml.

1362


maximum value of 15–20 mg cm22, then it remains practically constant up to 50 mg cm22. As scanning electron microscopy has shown,18 at low copper loading the electrode surface was characterised by a uniform dispersion of globular structures of copper particles with an average 1.3 mm size, while higher copper loadings (above 40 mg cm22) resulted in non-uniform dispersion of the catalyst which formed large clusters on the gold substrate. It is interesting to observe that variation of the copper loading between 15 and 50 mg cm22 produced very low variation in the analytical signal (see Fig. 4). The low effect of copper loading can be a key to the appreciable reproducibility of electrode performances notwithstanding the possible corrosion of copper particles under amperometric flow-through conditions. Performances of the Au/Cu electrode The Au/Cu electrode maintained its catalytic activity upon repeated thiocyanate injections. Representative responses of 2.5 mm (A) and 5 mm (B) thiocyanate solutions, obtained with an Au/Cu electrode (about 25 mg cm22) using 50 mm NaOH as mobile phase at 1.0 ml min21 flow rate and a constant potential of 0.7 V vs. Ag/AgCl are shown in Fig. 5. The calibration graph, obtained for thiocyanate ion, had a range linear up to 80 mm (correlation coefficient = 0.998) and the detection limit evaluated at a signal-to-noise ratio of 3 from the lowest injected concentration was 0.1 mm ( 6 ppb). A linear least-squares analysis of the points (n = 3) in the range 1.0–50 mm thiocyanate yielded a slope of 60 nA mm21 ± 3. Further, the RSD of 50 consecutive injection of 15 mm thiocyanate solution was 3.3%. The catalytic response is very enduring; after more than 10 h of flow injection experiments, corresponding to about 100 injections of 25 mm thiocyanate, a variation of 4–6% peak height was observed. Moreover, injection measurements periodically performed over a 2-month period were sufficiently reproducible. In fact only an average 20% decrease in electrode response was observed. However, if the 20% error cannot be tolerated, interspersing standards with the samples is possible to correct the change in sensitivity. Ion chromatography with electrochemical detection of thiocyanate The rate of ion migration through the column is directly dependent upon the type and concentration of ions in the eluent. Using 50 mm NaOH as mobile phase the thiocyanate ions tended to be strongly retained on the column (they were eluted after 25 min at a flow rate of 1.5 ml min21). In this respect, the use of mobile phase with high eluent capacity, such as mixtures of carbonate, caused the retention times of thiocyanate ion to be significantly reduced. Table 1 reports the effect of carbonate

concentration on the retention time of thiocyanate. As it can be seen, the retention time of the chromatographic peak decreases while the peak area remains practically independent with increasing carbonate concentration. Considering the above experimental results, a mobile phase containing 50 mm NaOH plus 0.1 m Na2CO3 was employed to evaluate the analytical performance of the Au/Cu electrode in IC. A calibration graph (n = 3) was obtained for the thiocyanate ion with a linear range between 1.0 and 195 mm; the correlation coefficient was 0.9984. The detection limit (S/N = 3) was 0.5 mm (29 ppb). Further, the precision expressed as RSD of seven chromatographic experiments (about 2 h of operation time) was 2.9% for 25 mm thiocyanate. Several potential interferents such as ascorbic, uric, tartaric and gluconic acids, glucose, chloride, sulfate, nitrite and nitrate were studied by IC using a solution containing 110 mm thiocyanate at 0.7 V and 1.5 ml min21 flow rate. All electroactive interferents considered here (i.e., ascorbic, tartaric, uric and gluconic acid, glucose and nitrite; 0.4 mm of each compound) were eluted after 2–5 min, while the retention time of the thiocyanate (13.3 min) and its peak current, were practically independent of the presence of the above interferents. In addition, the presence of electroinactive ions as sulfate and nitrate ( 0.3 mm) did not produce any variation in the chromatographic profile. In general, the variations in the chromatographic peak of thiocyanate in the presence of each interferent were less than a few percent, which are well within analytical uncertainty (i.e., 2.9% as RSD). Analytical applications To demonstrate the usefulness of the proposed modified electrode and to show the simplicity of the method, a rapid determination of thiocyanate ions in human urine (obtained by a smoker) was performed. The original sample was diluted 1:250 with the mobile phase (50 mm NaOH plus 0.1 m Na2CO3) and then injected in the column. A relevant chromatogram of a human urine sample obtained in IC with the Au/Cu electrode as amperometric sensor is shown in Fig. 6. From the peak currents before and after sequential standard additions, the thiocyanate concentration was determined. The estimated concentration of thiocyanate in the real sample was 0.43 mm ± 0.02 (n = 3). Further work is in progress to compare this proposed method and the spectrophotometric detection of thiocyanate in urine and other biological samples. Conclusion Copper particles finely dispersed on a gold substrate were investigated by cyclic voltammetry in an alkaline medium

Table 1 Retention times and peak area of the thiocyanate chromatographic signal at various concentrations of carbonate eluent. Peak area Retention time/min (arbitrary units) 0.000 25.6 1.40 0.015 21.2 1.35 0.032 18.5 1.41 0.055 15.9 1.60 0.080 14.5 1.50 0.100 13.3 1.50 0.150 13.1 1.55 * Conditions: anion-exchange column (Dionex) AS12A (200 3 4 mm ID) plus precolumn; mobile phase, 50 mm NaOH plus various concentrations of carbonate; flow rate, 1.5 ml min21; applied potential, 0.7 V vs. Ag/AgCl; injections of 62 mm thiocyanate. [CO322]/m

Fig. 5 Multiple flow injection peaks of 2.5 mm and 5.0 mm; peaks A and B respectively. Other experimental conditions as in Fig. 4.


1363

References 1 2 3 4 5 6 7 8 9 10 11

Fig. 6 Liquid chromatogram of human urine diluted 1 : 250 with mobile phase. Column, IonPac AS12A (Dionex) anion-exchange (200 3 5 mm id) coupled with IonPac AS12A (50 3 4 mm id) guard column; isocratic elution with 50 mm NaOH plus 0.1 m Na2CO3 as mobile phase; flow rate, 1.5 ml min21; peak 1, thiocyanate ion; applied potential, 0.7 V vs. Ag/AgCl.

12 13 14 15 16 17 18 19

towards the electrooxidation of thiocyanate. The oxidation process takes place through a preliminary adsorption of the analyte on the AuOH and CuOOH sites. Good catalytic activity in the range of potentials between 0.5 V and 0.8 V (SCE) was observed. The suitability of the Au/Cu electrode as an amperometric detector for the thiocyanate determinations in flowing streams, including IC has been explored. A sensitive and simple method for thiocyanate determination in a biological matrix is proposed. The only required preliminary treatment of the real sample was an adequate dilution with the mobile phase. The ease of preparation of the modified electrode, the good temporal stability and sensitivity, also in presence of high concentration of common interferents, confirmed the suitability of the Au/Cu electrode as an amperometric sensor in IC for the determination of thiocyanate ions in biological samples.

20 21 22 23 24 25 26 27 28 29

This work was carried out with the financial assistance of Ministero dell’Universita’ e della Ricerca Scientifica (M.U.R.S.T., Rome).

Poutts, W. C., Kuehneman, M., and Widdowson, C. M., Clin. Chem., 1974, 20, 1344. Bendtsen, A. B., and Hansen, E. H., 1991, Analyst, 116, 647. Montgomery, R. D., in Toxic Constituents of Plant Foodstuffs, ed. Liener, I. E., Academic, New York, 1969, p. 143. Tanaka, A., Deguchi, K., and Deguchi, T., Anal. Chim. Acta, 1992, 261, 281. Pinillos, S. C., Vincente, I. S., Bernal, J. G., and Asensio, J. S., Anal. Chim. Acta, 1996, 318, 377. Small, H., Stevens, T. S., and Bauman, W. C., Anal. Chem., 1975, 47, 1801. Romano, J. P., and Krol, J., J. Chromatogr., 1992, 602, 205. Jackson, P. E., Romano, J. P., and Wildman, B. J., J. Chromatogr., 1995, 706, 3. Hu, W., Takeuchi, T., and Haraguchi, H., Anal. Chem., 1993, 65, 2204. Harrold, M. P., Wojtusik, M. J., Riviello, J., and Henson, P., J. Chromatogr., 1993, 640, 463. Bjergegaard, C., Moller, P., and Sorensen, H., J. Chromatogr. A, 1995, 717, 409. Röder, A., and Bächmann, K., J. Chromatogr. A, 1995, 689, 305. Wang, E., and Kamata, S., Anal. Chim. Acta, 1992, 261, 399. Gong, Y.-T., Won, M.-S., Shim, Y.-B., and Park, S.-M., Electroanalysis, 1996, 8(4), 356. Rocklin, R. D., and Johnson, E. L., Anal. Chem., 1983, 55, 4. Casella, I. G., and Marchese, R., Anal. Chim. Acta, 1995, 311, 199. Liu, A., Xu, L., Li, T., Dong, S., and Wang, E., J. Chromatogr., 1995, 699, 39 Casella, I. G., and Salvi, A. M., Electroanalysis, 1997, 9(8), 596. Austin, D. S., Polta, J. A., Polta, T. Z., Tang, A. P. C., Cabelka, T. D., and Johnson, D. C., J. Electroanal. Chem., 1984, 168, 227. Cox, J. A., Gray, T., and Kulkarni, K. R., Anal. Chem., 1988, 60, 1710. Casella, I. G., Gatta, M., Guascito, M. R., and Cataldi, T. R. I., Anal. Chim. Acta, 1997, 357, 63. Johnson, D. C., Dobberpuhl, D., Roberts, R., and Vandeberg, P., J. Chromatogr. 1993, 640, 79. Luo, P., Zhang, F., and Baldwin, R. P., Anal. Chem., 1991, 63, 1702. Ueda, T., Mitchell, R., Kitamura, F., and Nakamoto, A., J. Chromatogr. 1992, 592, 229. Bruckenstein, S., and Shay, M., J. Electroanal. Chem., 1985, 188, 131. Angerstein-Kozlowska, H., Conway, B. E., Hamelin, A., and Stoicoviciu, L., J. Electroanal. Chem., 1987, 228, 429. Miller, B., J. Electrochem. Soc., 1969, 116, 1675. Marioli, J. M., and Kuwana, T., Electrochim. Acta, 1992, 32(2), 1187. Loucka, T., and Janos, P., Electrochim. Acta, 1996, 41(3), 405.

Paper 7/09041B Received December 16, 1997 Accepted March 4, 1998