Precapacitive processes in single potential-step chronoamperometry

Tahra,

0039-9140/91$3.00+ 0.00 Perpmon Press plc

Vol. 38, No. 2, pp. 189494, 1991

Printed in Great Britain

PRECAPACITIVE POTENTIAL-STEP

PROCESSES IN SINGLE CHRONOAMPEROMETRY

XIANG YUAN and RAY VONWANDRUSZKA* Department of Chemistry, University of Idaho, Moscow, ID 83843, U.S.A. (Received 28 February 1990. Revised 26 June 1990. Accepted 5 July 1990)

Summary-Extremely short-lived anodic currents were observed in the early parts of the transient response following the application of a cathodic potential step to a mercury working electrode. It is proposed that this phenomenon is due to the existence of a brief precapacitive period, which precedes full development of the double-layer charging current, and which allows momentary reaction (reduction) of species present at the electrode surface. The observed anodic currents are explained in terms of a re-oxidation of such “resident” species that were reduced during this precapacitive period. The subsequent capacitive surge produced by the charging of the electrical double layer leads to an anodic shift of the electrode potential that can be suihcient for the re-oxidation of the precapacitive amalgam. The anodic peak is linearly related to depolarizer concentration and varies with supporting electrolyte concentration, ion mobility and potential step size. Cathodic preconcentration of the depolarizer enhances the effect.

The behaviour of transient faradaic currents produced by a single potential step has been described theoretically by Miaw and Perone,’ and Miaw.2 They used the classical treatment of Feldberg3 to develop algorithms that allowed them to generate digitally simulated chronoamperometric curves. In subsequent work in this laboratory, high-speed instrumentation was used to produce experimental transient currents in simple reversible electrochemical systems.4 It was shown that the faradaic peak current varied with depolarizer concentration, yielding a calibration curve with two distinct linear regions. The temperature effect was found to be significant, giving a concentration-dependent “switch temperature”. The application of a potential step to an electrode in an electrolyte solution gives rise to a relatively high initial current, due to the charging of the electrical double layer at the electrode-electrolyte interface. This capacitive current is given by AE

/ --t \

where AE is the size of the potential step, Ru the uncompensated cell resistance, t the time, and C,, the double-layer capacitance. During the charging of the double layer, as the capacitive current decays the potential of

*Authorfor correspondence.

the working electrode “scans” toward its final value, as indicated by equation (l).’ Faradaic processes can only occur when the potential reaches the decomposition potential of the active species in the course of this “pseudocharging currents are scan”. Double-layer normally regarded as interferences in electroanalytical work and various steps are taken to minimize them. In the present investigation, their property of effectively generating a fast scan of the electrode potential is utilized as such. We have shown in previous work4sS that the similarity of this process to ultra-fast linearsweep voltammetry gives it the potential for interesting analytical applications. However, the fundamental phenomena are not well understood and further research is needed before a feasible technique can be developed. Algorithms developed by Miaw’ show that a chronoamperometric plot of the faradaic current component flowing during the first 10 time constants of the cell is peak-shaped (Fig. l), not unlike a conventional linear-sweep voltamperogram. Laitinen and Koltho@ have described the transient faradaic response represented by the area beyond the peak, which decays with time, t, as a function of t-‘12. In this region, the current is under diffusion control and is described by the well-known Cottrell equation: i = nFA(D/xt)1’2C

(2)

where the symbols have their usual electrochemical significance. During the period 189

190

XLMG 1200

WAN

and

RAY VON WANDRUSZKA

EXPERIMENTAL

r

Electrodes and control/data system

Y 0

I

I

I

4

8

12

Time (RC)

Fig. 1. Theoretical shape of a transient faradaic chronoamperogram produced by a single potential step.

preceding the faradaic peak, the rising current is caused by the potential pseudo-scan; to explain the peak shape of the chronoamperometric curve, it is useful to invoke depolarizer depletion effects, commonly discussed with reference to linear-sweep voltammetry.4~7 The previous theoretical treatment’ indicates that linearity of the current-concentration curve should not be expected for relatively high-level and wide-ranging analyte concentrations. This leaves the relationships at lower concentrations and within narrower concentration ranges to be investigated. Digital simulation also predicts that current-concentration behaviour will vary significantly at different measurement times during the existence of the transient. In our previous investigation,4 we chose to obtain our measurements at the peak of the faradaic transient, which occurs after about one cell time constant. This point is easily identifiable and does not vary significantly among solutions giving similar time constants. In the present work, unexpected faradaic behaviour manifested itself at the very beginning of the cathodic potential step, at times less than approximately 0.3 times the cell time constant. A reproducible “dip” appeared in the faradaic current curve, indicating an oxidative process that briefly precedes the reductive one. The magnitude of this reverse current was found to be proportional to the depolarizer concentration in low concentration ranges and to increase with supporting electrolyte concentration. An explanation for the phenomenon is proposed.

A three-electrode cell was used, with a hanging mercury drop working electrode (Metrohm Herisau AG CH-9100), a calomel reference electrode, and a platinum counter-electrode. The control/data system has been discussed in detail elsewhere* and will be described only briefly here. It consists of three interactive parts: a potential-step generator and a high-speed data-aquisition module, both linked to an Apple IIe microcomputer. The instrument is capable of taking current readings (1Zbit resolution) at a rate of 500 kHz after application of the potential step. A block diagram is shown in Fig. 2. The digital portion of the potential-step generator is built around a 1Zbit digital-toanalogue converter (DAC) and data are provided by the computer through a custom-built interface.s The data-acquisition circuit comprises a highspeed 1Zbit analogue-to-digital converter (Date1 ADC-817) with a maximum conversion time of 2 psec. The digital data generated by the ADC are stored in 768 on-board memory locations. After application of the potential step, the data-acquisition system cycles at maximum speed, the ADC restarting itself after each conversion. When 768 current readings have been gathered in this way, the system shuts itself off and control is returned to the computer for data reduction and storage. Chemicals Distilled, demineralized water treated with a Millipore purification system to give lOI

control

c

1

digital data

\

data "Cqt@$ .

2

Fig. 2. Diagram of instrumentation used for measurement of faradaic transients.

Precapacitive processes

in sinalepotential-stepchronoamperometry

MR. cm resistivity was used for all solutions. All electrolytes were of analytical grade from EM Science and were used without further purification. The mercury used in the working electrode was high quality electronic grade (Aldrich). Triton X-405 was obtained from Sigma as a 70% aqueous solution. Procedure

All solutions were deaerated by passage of nitrogen for at least 10 min. A single cathodic potential step, from 0.0 to - 1.0 V us. SCE, was applied to the working electrode. The length of time the electrode was held at the starting potential generally varied from 5 to 10 sec. The total current readings (capacitive + faradaic) for a sample were first recorded in the data module memory, then transferred to the computer memory and committed to disk. The reported data are averages of triplicate determinations and a fresh mercury drop electrode was used for each experiment. The same procedure was followed for a blank solution, containing only supporting electrolyte. The latter data, consisting of capacitive current readings only, were digitally subtracted from the sample readings, yielding the faradaic current produced by the depolarizer. Induced charging-current effects may be operative during the period of changing faradaic current, but the focus of this investigation is on the precapacitive section of the transient, which precedes the flow of substantive charging current of any type.

191

4.00 r

3.25 2 3 E z ZY -I

3.00

2.75

2.50

:7LLLL-

-6

-5

-4

-3

LoQ Cd concentration)

Fig. 4. Variation of anodic peak current with Cd2+ concentration (0.5M KCI supporting electrolyte). RESULT!3 AND DISCUSSION

In the present study it was noted that at the very beginning of the pseudo-scan resulting from the cathodic potential step, at times less than 0.3 times the cell time constants, the current-time curve showed a brief negative excursion (Fig. 3). In practical terms, this means that the blank (capacitive only) cathodic current during that period was greater than the sample (capacitive + faradaic) current. This suggests that this initial faradaic portion of the current was anodic in nature, leading to a smaller total current, which was otherwise entirely cathodic. The magnitude of the anodic faradaic peak was found to be linearly related to depolarizer concentration over a limited concentration range (Fig. 4). Moreover, the current depended on the concentration of the supporting electrolyte, increasing in a non-linear fashion (Fig. 5). The data in Table 1 show that the size of the anodic peak for Cd2+ in different supporting electrolytes varied almost linearly with cation mobility. While the data were mostly obtained with chlorides, additional work has shown that potassium nitrate as electrolyte gave similar results. The size and position of the cathodic potential step significantly influenced the anodic peak: it decreased in magnitude with decreasing step size. This was the case when the initial I I I -7501 potential was kept constant and the final poten1.0 1.5 0 0.5 tial was changed in the anodic direction and also Time (msec) when the initial potential was shifted cathodiFig. 3. Actual shape of a faradaic chronoamperogram tally and the final potential was left at - 1.0 V produced by a single potential step from 0.0 to - 1.OV on a mercury electrode in 0.5M KC1 with 2.0 x 10m6M (Fig. 6). The addition of 0.0035% Triton X-405 Cd2+. Inset: total component currents for; a, blank and b, sample. detergent obliterated the entire transient effect,

192 1000

?hANG -

YUAN

and

bY

VON WANDRUSZ~CA

part of the faradaic transient and does not recognize the possibility that the conditions created by the double-layer charging current require a finite time to develop fully. We propose that the slowness of movement of hydrated ions in aqueous solution, relative to the movement of electrons in a metallic conductor, results in a brief “precapacitive” period immediately following the application of the potential step. During this time, the electrode potential is close (or even equal) to the applied potential, since the potential drop due to the charging current has not yet developed. This allows for the momentary reaction of the electroactive species 0.6 0.4 0.2 0 resident at the electrode surface at that instant, [KC11 (MI provided that the charge transfer is sufficiently Fig. 5. Variation of anodic peak current of 5.0 x 10m6M fast. This contention is supported by our obserCdz+ with KC1 concentration. vation that the capacitative (blank) current initially rises to a maximum, before it decays according to equation (1). The rise typically both anodic and cathodic, in the chronoamperogram. Only at high concentrations of Cd2+ lasts 6-8 psec in our system with OSM potass(1 .O x 10w3M and above) were both peaks ium chloride solution and is resolved into 3 or 4 data points. The faradaic processes that occur discernible again. As pointed out in the procedure section, the during the precapacitive period, i.e., before the capacitive current has risen sufficiently to cause electrode was normally left at the initial potenan electrode potential drop below the decompotial of 0.0 V for 5-10 set before the potential step was applied. The length of this period was sition potential, are resolved into 1 or 2 data found to have no significant effect-the final points. The bulk of the precapacitive reduction chronoamperogram showed no changes even probably takes place during the first l-2 psec when the initial potential was applied for several following the potential step and therefore falls outside the time resolution capability of our minutes. To find further evidence for the origins of instrument. The fast reduction and concurrent amalgathe anodic peak, the influence of cathodic preconcentration was investigated. A potential of mation of cadmium and lead ions renders them amenable to the precapacitive surface - 1.0 V us. SCE was applied to the working electrode for a duration of 100 set in Cd’+ reduction described above. The beginning of solutions ranging from 5.0 x lo-’ to 1.0 x 10-5M. This was followed by the customary 0.0 1000 to - 1.0 V potential step. In all cases it was h . found that the size of the anodic peak increased 600 about fourfold relative to the corresponding situation without preconcentration. The propagation delay and current settling time of the potentiostatic control circuit were less than 300 nsec, while the relevant electrochemical responses typically extended to approximately 100 msec. This precludes electronic artifacts as the causes of the observed effects and indicates the presence of electrochemical transients, even in the earliest measurements. The -9.1 I -0.9 -0.7 -0.5 I -0.3 -0.11 most plausible explanation of the phenomenon Potentlot range (V vs. SCE) involves the time-average resident (not necessarily adsorbed) depolarizer species at the elec- Fig. 6. Variation of the anodic peak current of 10 x 10e5M trode surface.4 It should be noted that existing Cd2+ in OSM KC1 with range of potential step sizes; A, from theory does not specifically address the early 0.0 V to final potential and B, from initial potential to - 1.OV.

Precapacitive processes in single potential-step chronoamperometry

the capacitive process therefore sees a quantity of amalgam, formed under the influence of the precapacitive full applied potential, in the very surface layer of the mercury drop. The subsequent flow of double-layer charging current shifts the electrode potential in an anodic direction and causes the re-oxidation of this amalgam. The resulting anodic current is effectively subtracted from the capacitive current, which is overwhelmingly “cathodic” in direction. The initially measured cathodic current in a solution containing a reversible depolarizer is, therefore, smaller than the corresponding current in a solution with supporting electrolyte only. The various currents involved in the precapacitive and early capacitive processes are shown in Fig. 7. The explanation offered here accounts for the observed dependence of the anodic peak on depolarizer concentration. The time-average presence of depolarizer at the electrode surface is essentially a bulk effect and the amount of precapacitive amalgam formed therefore depends directly on the depolarizer concentration. The loss of linearity observed at higher concentrations (Fig. 4) results from surface saturation of the electrode by resident species. The number density of depolarizer ions suthciently close to the surface to be available for precapacitive reduction has an upper limit dictated by the hydrated radius of the ion. Further increases in concentration beyond this saturation point cannot produce more amalgam for subsequent re-oxidation and the anodic peak current therefore reaches a maximum.

Anodac rsdlssolutvm current

Fig. 7. Magnified view of earliest measurable portions of transient currents. A, double-layer charging current; B, faradaic current. First data point approx. 2 psec after potential step.

193

The enhancement of the anodic peak by an increase in supporting electrolyte concentration and by greater ion mobility provides further evidence for the proposed mechanism. Both result in increased capacitive surges upon double-layer formation and hence lead to increased anodic shifts during the process. This, in turn, gives rise to a higher re-oxidation current. As mentioned above, a smaller potential step produces a smaller anodic peak, whether the initial potential is more cathodic or the tlnal potential is more anodic. This is also in agreement with the precapacitive mechanism; the smaller step decreases the capacitive surge and hence the anodic potential drop, leading to decreased re-oxidation current. If the size of the potential step is reduced by shifting the final potential anodically, an additional effect is operative: the precapacitive electrode potential is now less cathodic, so the extent of initial amalgamation is decreased. This leads to a yet smaller re-oxidation current. The observation that the addition of detergent to the solution eliminates both anodic and cathodic transient peaks is expected in view of electrode surface coverage by the detergent and disturbances in the extent and speed of double-layer formation. Strong evidence for the proposed mechanism is provided by the cathodic preconcentration experiment, especially in contrast with the case where non-electrolytic initial potentials are used. A separate and prolonged reductive accumulation of cadmium and lead in the electrode leads to a greatly enhanced anodic transient, which is no longer dependent on the presence of resident species at the electrode surface. Since the reduction in this case is not limited to the brief precapacitive period and the accumulation is therefore significantly greater, the amount of amalgam available for re-oxidation is increased. The anodic potential drop during the capacitive surge, however, is still effective, and the anodic peak increases in accordance with the amount of preconcentrated depolarizer. For all bulk concentrations considered at the chosen preconcentration time, the anodic peak increased linearly, suggesting that the mechanism can be likened to an ultra-fast anodic stripping procedure. It is worth noting that depolarizer adsorption on the electrode does not appear to play a major role under the conditions of this study, since leaving the electrode at the starting potential (0.0 V us. SCE)

?CL%NO YUAN and RAY VONWANDRUSZKA

194

for relatively long periods did not change the eventual anodic peak current.

Acknowledgemenr-The authors gratefully acknowledge support of this work by the University of Idaho Rem&h Council.

CONCLUSIONS

The proposed process, involving a precapacitive period during which species resident at the electrode surface are reduced and re-oxidized during the capacitive surge, accounts for the observed anodic component in an otherwise entirely cathodic process. The mechanism is suggested as a plausible expl~ation of the experimental findings. The present investigation has dealt with a mercury-drop electrode and amalgam-forming depolarizers and further studies are needed to determine whether similar processes are operative with non-amalg~ating depolarizers and at noble metal and carbon electrodes. Interesting possibilities exist for analytical applications of such fast anodic stripping of pre~n~ntrat~ analytes, with speed of analysis as the prime benefit.

REFERENCES 1. L.-H. Lai Miaw and S, P. Perone, Anal. Gem., 1979, Sl, 1645. 2. L.-H. Lai Miaw, Ph.D. Thesis,Furdue University, West Lafayette, IN, 1978. 3. S. W. Foldberg, in Electroanalytical Chemistry, A. J. Bard (ed.), Vol. 3, p. 199. Dekker, New York, 1966. 4. S. 0. Arhunmwunde and R. von Wandruszka, Alsal. Chim. Acta, 1989, 226, 337. 5. S. 0. Arhunmwunde, M.S. 7?hesr$University of Idaho, Moscow, ID, 1989. 6. H. A. Laitinen and I. M. Kolthoff, J. Am. Chem. Sot.,

1939, 61, 3344. 7. A. J. Bard and L. R. Faulkner, ~~e~tr~~~al Me&&: ~~~~ta~s rmdApplications,p. 214. Wiley, New York, 1980. 8. R. von Wandruszka, Anal. Instrum., 1984, 13, 193. 9. R. von Wandruszka and M. Maraschin, Anal. Len., 1981, 14,463. 10. S. S. Fratoni, Jr. and S. P. Perone, Anal. Gem.. 1976,

48, 287.