Influence of Mixed Diffusional, Migrational, and Convective Mass

The current response of a 10-µm wall-tube microelectrode in a flow injection system under the conditions of low ionic strength has been examined ...
0 downloads 0 Views 146KB Size
Download PDF
Anal. Chem. 1999, 71, 4926-4931

Influence of Mixed Diffusional, Migrational, and Convective Mass Transport on the Response of a Wall-Tube Microelectrode in a Flow Injection System Fredrik Bjo 1 refors,† Joanna Gadomska,‡ Mikołaj Donten,‡ Leif Nyholm,†,§ and Zbigniew Stojek*,‡,|

Department of Chemistry, University of Warsaw, ul. Pasteura 1, PL 02-093 Warsaw, Poland, and Department of Analytical Chemistry, Uppsala University, P.O. Box 531, SE-751 21 Uppsala, Sweden.

The current response of a 10-µm wall-tube microelectrode in a flow injection system under the conditions of low ionic strength has been examined experimentally for several redox systems such as ferrocene in methanol, undiluted methanol, and water in acetone. The examination involved the dependence of the current on the positioning of the microelectrode relative to a 760-µm-i.d. capillary outlet, flow rate, potential pulse time, and ratio between the concentrations of the supporting electrolyte and electroactive species (Cel/Credox). For Cel/Credox ratios smaller than ∼0.001 and a flow rate of 200 µL/min, the dependencies of the current on the flow rate and the positioning of the microelectrode versus the capillary tip were reversed compared to the presence of excess supporting electrolyte. The current was thus found to decrease with increasing flow rate while a local current maximum could be seen when the microelectrode was center-aligned with the capillary tip. The changes in the current are the results of local differences in convective transport. These differences alter the rate of migrational accumulation of counterions at the electrode surface. It is shown that results similar to those obtained for the excess supporting electrolyte case can be obtained for a low value of Cel/ Credox and a given flow rate if the electrode potential and time scale of the experiments are chosen appropriately. Microelectrodes, due to their small dimensions, do not generate large ohmic drops and can therefore be used for electrochemical experiments in very resistive media.1,2 For stationary solutions, several successful applications have been demonstrated including determinations in the absence of added supporting electrolyte, voltammetry of undiluted solvents, and electrochemical determinations in supercritical media.1-11 In these cases, the influence of migration on the mass transport is usually substantial and much †

Uppsala University. University of Warsaw. § E-mail: [email protected]. | E-mail: [email protected]. (1) Wightman, R. M.; Wipf, D. O. In Electroanalytical Chemistry; Bard, A. J., Ed.; Marcel Dekker: New York, 1988; Vol. 15. (2) Bond, A. M. Analyst 1994, 119, R1. (3) Bond, A. M.; Fleischmann, M.; Robinson, J. J. Electroanal. Chem. 1984, 168, 299. ‡

4926 Analytical Chemistry, Vol. 71, No. 21, November 1, 1999

theoretical work has hence been undertaken to investigate the influence of migration on the electrochemical response of different compounds in stationary solutions.12-23 It has been shown that an electrochemical reaction involving a neutral compound leads to the formation of a kind of supporting electrolyte layer in the vicinity of the electrode as a result of the drawing of counterions from the bulk of the solution to maintain electroneutrality in the depletion layer. The effects of such migrational processes on voltammetric and chronoamperometric curves have been described for a number of electrochemical systems in stationary solutions7,12,18,19,21 and for a rotating microdisk electrode.24 Electrochemical detection in flow systems represents an analytical application of microelectrodes that is becoming increasingly popular in capillary flow injection analysis, capillary chromatography, and capillary electrophoresis.2,25 In this context, much effort has been made to study the effects of combined diffusional and convective mass transport to microelectrodes.2,26-30 In particular, the influence of the flow rate on the electrochemical (4) Ciszkowska, M.; Stojek, Z. J. Electroanal. Chem. 1986, 213, 189. (5) Wipf, D. O.; Wightman, R. M. Anal. Chem. 1990, 62, 98. (6) Olsen, S. A.; Tallman, D. E. Anal. Chem. 1996, 68, 2054. (7) Amatore, C.; Deakin, M. R.; Wightman, R. M. J. Electroanal. Chem. 1987, 225, 49. (8) Crooks, R. M.; Bard, A. J. J. Electroanal. Chem. 1988, 243, 117. (9) Wallenborg, S. R.; Markides, K. E.; Nyholm, L. Anal. Chem. 1997, 69, 439. (10) Malmsten, R. A.; White, H. S. J. Electrochem. Soc. 1986, 133, 1067. (11) Ciszkowska, M.; Stojek, Z. Analyst 1994, 119, 239. (12) Oldham, K. B. J. Electroanal. Chem. 1988, 250, 1. (13) Baker, D. R.; Verbrugge, M. W.; Newman, J. J. Electroanal. Chem. 1991, 314, 23. (14) Bond, A. M.; Fleischman, M.; Robinson, J. J. Electroanal. Chem. 1984, 172, 11. (15) Norton, J. D.; White, H. S.; Feldberg, S. W. J. Phys. Chem. 1990, 94, 6772. (16) Amatore, C.; Fosset, B.; Bartelt, J.; Deakin, M. R.; Wightman, R. M. J. Electroanal. Chem. 1988, 256, 255. (17) Cooper, J. B.; Bond, A. M. J. Electroanal. Chem. 1991, 315, 143. (18) Palys, M.; Stojek, Z.; Bos, M.; van der Linden, W. Anal. Chim. Acta 1997, 337, 5. (19) Hyk, W.; Stojek, Z. J. Electroanal. Chem. 1997, 422, 179. (20) Cooper, J. B.; Bond, A. M.; Oldham, K. B. J. Electroanal. Chem. 1992, 331, 877. (21) Myland, J. C.; Oldham, K. B. Electroanal. Chem. 1993, 347, 49. (22) Bento, M. F.; Thouin, L.; Amatore, C. J. Electroanal. Chem. 1998, 446, 91. (23) Bento, M. F.; Thouin, L.; Amatore, C.; Montenegro, M. I. J. Electroanal. Chem. 1998, 443, 137. (24) Gao, X.; White, H. S. Anal. Chem., 1995, 67, 4057. (25) Ewing, A. G.; Mesaros, J. M.; Gavin, P. F. Anal. Chem. 1994, 66, 527A536A. 10.1021/ac990430b CCC: $18.00

© 1999 American Chemical Society Published on Web 09/25/1999

response of microelectrodes in flow cells of different construction and for excess supporting electrolyte conditions has been a topic of several experimental and theoretical studies.26,29,31-33 It has also been shown that the presence of spherical diffusion in microelectrode-based thin layer27,32,34 and wall-tube flow cells33 modifies the flow rate dependence of the current. We are, however, not aware of any similar studies that include the effects of migrational mass transport on the voltammetric and chronoamperometric responses in a flow system. Recently, we became interested in employing the steady-state oxidation current of undiluted methanol in a wall-tube microelectrode-based flow injection system for analytical purposes.35 In the latter study, which was carried out in the absence of excess supporting electrolyte, it was found that an increase in the flow rate led to a decrease in the current, in contrast to the behavior usually found for microelectrodes in flow systems in the presence of excess supporting electrolyte.26,29,31-33 These findings, which are in good agreement with the findings of Gao and White24 obtained with a rotating microdisk electrode in the absence of an excess of supporting electrolyte, can be explained by the negative effect of convective mass transport on the buildup of a supporting electrolyte or ionic layer at the electrode surface. The aim of the present work was to investigate the influence of convective mass transport on the current for a wall-tube microelectrode-based flow injection system for different ratios between the concentrations of the supporting electrolyte and electroactive species (Cel/Credox). The experiments included the oxidation of water in acetone, ferrocene in methanol, and undiluted methanol. The buildup of the layer enriched in ions at the walltube microelectrode was also studied by analyzing current-time transients obtained after the application of potential pulses. EXPERIMENTAL SECTION Apparatus. The experimental setup consisted of a PU-980 HPLC pump (Jasco, Tokyo, Japan), a Rheodyne injection valve (Rheodyne, Cotati, CA) with a 80-µL loop, and a laboratorydesigned wall-tube microelectrode flow cell in Kel-F, which is shown in Figure 1. The distance between the working electrode surface and the capillary end, Z, was usually between 0.5 and 1.5 mm. The capillary tubing used was made of PEEK and had an internal diameter of 760 µm and an outer diameter of 1.6 mm. The working electrode was a 10-µm platinum disk manufactured by Project (Warsaw, Poland). The platinum fiber was sealed into a glass capillary with an outer diameter of 6 mm. The reference electrode was a platinum coil, and all potentials reported in the paper are given versus this quasi-reference electrode. The counter electrode was made of stainless steel. Before use, the working electrode was polished to mirror finish with 1- and 0.1-µm alumina (Buehler Ltd., Lake Bluff, IL) on a wet pad. Subsequent renewals (26) Tait, R. J.; Burry, P. C.; Finnin, B. C.; Reed, B. L. J. Electroanal. Chem. 1993, 356, 25. (27) Alden, J. A.; Compton, R. G. J. Electroanal. Chem. 1996, 404, 27. (28) Moldoveanu, S.; Anderson, J. L. J. Electroanal. Chem. 1985, 185, 239. (29) Cope, D. K.; Tallman, D. E. J. Electroanal. Chem. 1986, 205, 101. (30) Anderson, J. L.; Ou, T. Y.; Moldoveanu, S. J. Electroanal. Chem. 1985, 196, 213. (31) Corti, H. R.; Goldfarb, D. L.; Ortiz, M. S.; Magallanes, J. F. Electroanalysis 1995, 7, 569. (32) Niwa, O.; Morita, M. Anal. Chem. 1996, 68, 355. (33) Wallenborg, S. R.; Markides, K. E.; Nyholm, L. Anal. Chim. Acta 1997, 344, 77.

Figure 1. The flow cell used: (1-3) working microelectrode, reference electrode, and counter electrode, respectively; (4-6) working, reference, and counter electrode leads; (7-9) micrometer screws for X, Y, and Z settings, respectively; (10) capillary tip of internal diameter φ; cell volume 25 mL.

of the electrode surface prior to the amperometric measurements were accomplished by brief polishing with 0.1-µm alumina. A PARC model 283 potentiostat together with the ECHEM software was used to control the potential and measure the current. Reagents. Methanol and acetone were of analytical reagent grade and purchased from POCh (Gliwice, Poland). According to the manufacturer, the acetone contained not more than 0.03% (w/w) water. Lithium perchlorate of analytical reagent purity was obtained from Aldrich (Milwaukee, WI) and was dried under vacuum before use. Ferrocene was obtained from Merck (Munich, Vien, Germany). The water used was of Milli-Q (Millipore, Austria) quality. RESULTS AND DISCUSSION Influence of Electrode Positioning. It is well-known that, in the absence of excess supporting electrolyte, the oxidation or reduction of a neutral compound leads to an accumulation of counterions in the vicinity of the electrode due to the necessity of preserving electroneutrality.7,12,18,19,21 In a flow system, it can be expected that such a buildup of an ionic layer near the electrode will be impeded by the convection caused by the flow, in analogy with the situation at a rotating microdisk electrode.24 This should Analytical Chemistry, Vol. 71, No. 21, November 1, 1999

4927

lead to increased ohmic drop across the depletion layer compared to stationary solutions. In a wall-tube microelectrode-based flow cell, the magnitude of the convection at the electrode surface, for a given flow rate, depends critically on the position of the electrode with respect to the capillary end. To study the influence of the position of the wall-tube working electrode (with respect to the capillary end) on the current signal under low Cel/Credox conditions, two model systems were used. The first consisted of 5.0 × 10-3 M ferrocene dissolved in methanol. Although supporting electrolyte was not added, the Cel/ Credox ratio was estimated to be ∼1 × 10-3 due to the presence of ∼5 µM ionic impurities in the methanol (as determined by conductometry). In the experiments performed, the distance, Z, between the electrode surface and the plane of the capillary end was first kept constant while the electrode was moved in the X-Y plane. The X coordinate was changed while the Y setting was kept constant. In the following discussion, the X and Y coordinates have been normalized with respect to the inner diameter, φ, of the flow capillary. The X/φ and Y/φ values are equal to zero when the electrode is perfectly aligned with respect to the center of the capillary. As there was no ferrocene in the cell at the start of the experiment, the shapes of the peaked curves in Figure 2 are mainly determined by the dimensions of the jet hitting the electrode (and hence the dispersion in the jet between the capillary end and the electrode) for the different electrode positions. The currents in Figure 2 have been normalized with respect to the steady-state limiting current of ferrocene in the quiescent solution. For sufficiently large X/φ values, the current was hence found to be equal to the background current found in a quiescent solution. In contrast to the behavior seen in the presence of excess supporting electrolyte36 (see curve a in Figure 2B), the oxidation current in Figure 2A (curve a) exhibits a much wider minimum at the top of the peak, and the highest current ratio obtained is smaller than unity. A precise examination of curve a in Figure 2A reveals the presence of a small current maximum at around X/φ ) Y/φ ) 0. This small maximum is in contrast to the larger minimum also seen in Figure 2A and the minimum previously observed36 under excess supporting electrolyte conditions when the current depends only on the mass-transfer rate of the analyte. The width of the maximum is ∼0.38, which is similar to that of the minimum of curve a in Figure 2B. It is therefore reasonable to assume that both effects are caused by the same phenomenon. The appearance of a narrow maximum (or narrow minimum in the presence of excess supporting electrolyte) for X/φ ) Y/φ ) 0 stems from the presence of a stagnation point at the electrode surface when the electrode and the jet are perfectly aligned. No such maximum within a larger minimum is therefore seen at curves b and c in Figure 2A, and no minimum is seen at curves b and c in Figure 3B, as the center of the jet does not hit the electrode surface in these cases; such conditions exist for either Y/φ > 0.66 or Y/φ < -0.66. Next experiments were carried out with undiluted methanol containing 4.0 × 10-3 M lithium perchlorate and present both in (34) Takahashi, M.; Morita, M.; Niwa, O.; Tabei, H. J. Electroanal. Chem. 1992, 335, 253. (35) Gadomska, J.; Donten, M.; Stojek, Z.; Nyholm, L. Analyst 1996, 121, 1869. (36) Macpherson, J. V.; Beeston, M. A.; Unwin, P. R. J. Chem. Soc., Faraday Trans. 1995, 91, 899.

4928 Analytical Chemistry, Vol. 71, No. 21, November 1, 1999

Figure 2. Normalized current, due to the oxidation of 5.0 × 10-3 M ferrocene dissolved in methanol (the ferrocene was only present in the flow) containing no deliberately added supporting electrolyte (A) and 0.1 M LiClO4 (B), plotted as a function of the dimensionless X/φ coordinate: Y/φ ) 0 (a), 1.3 (b), and 2 (c). E ) +0.65 V; flow rate 200 µL/min; Z/φ ) 0.66. The currents have been normalized with respect to the steady-state current in a quiescent solution.

the cell and in the jet. The electrooxidation of undiluted methanol at sufficiently positive potentials in the presence of 4.0 × 10-3 M LiClO4 is known to be a well-defined process: the steady-state limiting current is reached within a reasonable time in quiet solutions and the magnitude of the current is controlled by diffusion.37 In this system, the concentration of ionic impurities was considerably smaller than the concentration of LiClO4 added, and the value of Cel/Credox was therefore well determined. In the present experiments, this ratio equaled 1.8 × 10-4, which is ∼6 times smaller than in the ferrocene experiments. For the purpose of better demonstration of the effect of the convective mass transport on the current, the methanol oxidation experiments were undertaken at +3.9 V, a potential on the rising portion of the methanol oxidation wave, where the steady-state current is not attained rapidly.37

Figure 3. Normalized currents due to the oxidation of methanol in the presence of 4.0 × 10-3 M LiClO4 (the same solution was present in both the flow and the cell) as a function of the dimensionless X/φ coordinate: Y/φ ) 0 (a), 2 (b), and 4 (c); Z/φ ) 2, flow rate 200 µL/ min; E ) +3.9 V. The currents have been normalized with respect to the steady-state current in a quiescent solution.

As is seen in Figure 3, the methanol oxidation current first exhibits its steady-state value and then decreases when the electrode is moved into the jet along the X/φ axis. The decrease, which is similar to the wide minimum in Figure 2A, and which can be as large as 28% for Y/φ ) 0, clearly reflects the increasing influence of convective mass transport as the electrode is moved toward the center of the jet. As in the ferrocene case, the increased mass transport makes it more difficult to assemble a layer with increased concentrations of ions at the electrode surface. Similarly to curve a in Figure 2A, a current maximum is seen when the electrode is positioned close to the point X/φ ) Y/φ ) 0 due to the presence of a stagnation point. In Figure 3, curve a, this increase in the current is ∼29% of the total minimum value. No local maximum within the minimum is seen for curve b in Figure 3, since the electrode is misaligned with respect to the center of the jet; such conditions exist, as in the ferrocene case, for either Y/φ > 0.66 or Y/φ < -0.66. The appearance of the maximum for X/φ ) Y/φ ) 0 in the absence of excess supporting electrolyte (or a minimum in the presence of excess supporting electrolyte) can therefore be used to check the alignment of the electrode with respect to the capillary end in electrochemical detection using wall-tube microelectrodes. As the distance, Z/φ, between the capillary end and the electrode and the flow rate is known to affect the center line velocity of the jet for wall-tube microelectrode-based flow cells,33,36 some experiments were also carried out to examine the influence of these factors on the current under Cel/Credox < 10-3 conditions. In the presence of excess supporting electrolyte, the current decreases with increasing Z/φ values, as is seen for curve b in Figure 4 for the oxidation of ferrocene in the presence of 0.1 M LiClO4. An opposite dependence is seen in the absence of excess supporting electrolyte (curve a in Figure 4), which reflects the (37) Ciszkowska, M.; Stojek, Z. J. Electroanal. Chem. 1993, 344, 135.

Figure 4. Normalized currents for the oxidation of 5.0 × 10-3 M ferrocene dissolved in methanol (ferrocene present only in the flow) as a function of Z/φ for 0.0 (a) and 0.1 (b) M LiClO4. X/φ ) Y/φ ) 0; E ) +0.65 V. The currents have been normalized with respect to the current obtained for Z/φ ) 2.5 in the absence of added supporting electrolyte.

decreased convective removal of ions from the electrode vicinity with increasing Z/φ values. For sufficiently large Z/φ values, the current is, however, expected to start to decrease because the ferrocene concentration reaching the electrode decreases as a result of increased mixing of the jet with the surrounding solution. The fact that curves a and b in Figure 4 do not reach the same value at high values of Z/φ is due to a lower viscosity of the solution in the absence of 0.1 M LiClO4. Our results also show that variations in the Z/φ value result in changes in the dependence of the current on the positioning of the electrode in the X-Y plane. For example, for the oxidation of ferrocene in the absence of added supporting electrolyte, at Z/φ values of 1.3 and 2.0, the obtained curves lacked the large minimum seen in curve a of Figure 2A as a result of the reduced center line velocity of the jet. As the current for X/φ ) Y/φ ) 0 was practically the same for Z/φ ) 1.3 and 2.0, it is clear that the concentration of ferrocene reaching the electrode was essentially the same in both cases. This indicates that the dispersion in the center of the jet is small under these conditions. In a flow injection system, this could be used to minimize the influence of convective mass transport on the buildup of an ionic layer in the vicinity of the electrode under the conditions of low Cel/Credox, while still maintaining high sensitivities and sample throughput. Flow Injection Experiments. Injection of water into the mobile phase composed of acetone and a low concentration of lithium perchlorate was chosen for flow injection experiments. This system was recently studied by us in conjunction with the determination of the water content in fat.38 Before injection experiments, cyclic voltammetric waves of water were obtained under flow conditions. The waves were well-defined and their heights were previously found11 to be mass transport controlled. As is shown in Figure 5, the height of the water oxidation wave was found to be practically independent of the flow rate for flow rates between 100 and 400 µL/min, using a Z/φ value of 1.3 and Analytical Chemistry, Vol. 71, No. 21, November 1, 1999

4929

Figure 5. Linear scan voltammograms for the oxidation of water, present as an impurity in acetone, for flow rates of 100 (a), 200 (b), and 400 µL/min (c). 1.0 × 10-4 M LiClO4 in both flow and cell. Z/φ ) 1.3.

1.0 × 10-4 M LiClO4 in the mobile phase. An increase in the flow rate, however, leads to a shift in the half-wave potential to more positive values. This shift from about +2.75 to +3.08 V seen for an increase in the flow rate from 100 to 400 µL/min is similar to that seen for increasing rotation rates for a rotating microdisk electrode24 and can be interpreted in terms of flow increasing iR drop, which is a result of an increased rate of removal of ions from the depletion layer at higher flow rates. The steady-state current can also be obtained in an appropriate constant-potential amperometric experiment. The background current (0.6 µA) due to the oxidation of water impurities in acetone in the presence of 1.0 × 10-4 M LiClO4 should hence correspond to the limiting current in Figure 5, provided that a sufficiently positive potential is imposed. As is seen in Figure 6, injections of the same acetone (with the same water content), but with different concentrations of lithium perchlorate compared to the mobile phase, lead to significant modifications of the current. This is a result of the change in the Cel/Credox ratio and hence in the ohmic drop across the depletion layer. While an injection of acetone containing no LiClO4 leads to a ∼70% decrease in the current compared to the background level, the injection of acetone containing more than 1 × 10-4 M lithium perchlorate results in an increase in the current by a factor of up to 7. This increase may be surprising, as the baseline in Figure 6 can be regarded as the limiting steady-state current for the concentration of water impurities present in the acetone. What happens is that the increase in the LiClO4 concentration leads to a decrease in the ohmic drop which renders the electrode potential so positive that the oxidation of acetone starts to contribute to the current measured. On the other hand, an addition of water to the injected sample, decreases the Cel/Credox ratio and thus increases the ohmic drop. Increasingly positive detection potentials are therefore needed for increasing water concentrations. Flow Injection Pulse Amperometry. The shape of the current-time transient, for an electroactive species in a flow injection system with low Cel/Credox, will, most likely, depend on 4930 Analytical Chemistry, Vol. 71, No. 21, November 1, 1999

Figure 6. Amperometric signals due to two sets of injections of acetone containing: 0 (a), 1.0 × 10-3 (b), 5.0 × 10-3 (c), 1.0 × 10-2 (d), and 2.0 × 10-2 M (e) LiClO4. The acetone in the mobile phase contained 1.0 × 10-4 M LiClO4. E ) +3.4 V; Z/φ ) 1.3; flow rate 100 µL/min; injection volume 80 µL.

the rate by which the ions are accumulated in the vicinity of the electrode. For a given Cel/Credox ratio, the contact time between the injected sample and the electrode may, for instance, be too short to ensure that the desired potential is reached. To estimate the time needed to collect a sufficient amount of ions in the vicinity of the wall-tube electrode in the presence of a flow, water injection experiments were performed in which the potential was repeatedly stepped from +2.0 to +3.4 V with a pulse time of 1.0 s and with a 2.0-s rest at +2.0 V between the pulses. In these experiments, the current was sampled at different times during the pulses. As presented in Figure 7, the water oxidation current corresponding to the injection of acetone containing 0.05% (v/v) of added water (yielding a Cel/Credox ratio of ∼4 × 10-3) was found to increase with increasing sampling time for sampling times ranging from 10 to 500 ms. This behavior is in contrast to that normally seen in the presence of excess supporting electrolyte, but is in a good agreement with theoretical39,40 and experimental19 findings for potential pulse experiments in quiescent solutions and for low Cel/ Credox. As is seen in the inset of Figure 7, a nearly constant current is obtained for sampling times longer than ∼500 ms, which suggests that less than 1 s is needed to accumulate a sufficient amount of ions in the vicinity of the electrode under these conditions. A very similar dependence was also seen when the magnitude of the background current was plotted as a function of the sampling time. The dependence of current on time is more clearly seen in Figure 8, which presents current transients recorded at different advancements of one of the peaks in Figure 7. It is seen that all three current-time transients practically overlap at short time; then the current is mainly controlled by the ohmic drop across the depletion layer and not by the concentration of the analyte.19 For a sufficiently long time, when the accumulation of the ions is well advanced and the ohmic drop across the depletion layer therefore is relatively small, the current finally becomes related to the water concentration. A slightly longer time is needed to reach the mass-transfer-controlled

increase is restricted by the convection, which tends to flush the ions away from the vicinity of the electrode. For a given water concentration, the time required to reach the steady-state current also depends on the concentration of lithium perchlorate in the mobile phase. In analogy with the situation in stationary solutions,40 the required time increases with decreasing Cel/Credox ratios. It is, however, clear that a lithium perchlorate concentration of 1.0 × 10-4 M (i.e., a Cel/Credox ratio of ∼4 × 10-3) is sufficient to ensure that the flow injection peak currents for the injection of 80 µL of acetone containing 0.05% water are mass transfer controlled in the flow injection system used.

Figure 7. Normalized pulse amperometric currents for the injection of acetone containing 0.05% (v/v) added water and 1.0 × 10-4 M LiClO4. Z/φ ) 1.3; flow rate 100 µL/min. Sampling time: 10 (a), 25 (b), 50 (c), 100 (d), 250 (e), and 500 ms (f). The currents were normalized with respect to the steady-state current (see the text for details). Inset: normalized flow injection peak currents plotted as a function of the sampling time.

Figure 8. Normalized current-time transients recorded prior to (a), on the rising portion (b), and at the maximum of the peak in Figure 7 (c).

conditions when the transient is recorded at the maximum of the flow injection peak (curve c in Figure 8) and the water concentration is the highest. The current, nevertheless becomes mass transfer controlled within less than 1 s also in this case. The results depicted in Figure 8 show that the time needed to reach masstransfer-controlled conditions increases with decreasing Cel/Credox ratios, which is in qualitative agreement with previous findings for stationary solutions.19 In the present case, the rate of current (38) Bjo ¨refors, F.; Nyholm, L.; Donten, M.; Stojek, Z. Analyst 1999, 124, 143. (39) Jaworski, A.; Donten, M.; Stojek, Z. J. Electroanal. Chem. 1996, 407, 75. (40) Hyk, W.; Palys, M.; Stojek, Z. J. Electroanal. Chem. 1996, 415, 13.

CONCLUSIONS The results of this study clearly show that the convective removal of ions (charged oxidation products and counterions) from the vicinity of the electrode decreases the oxidation current for neutral compounds in a wall-tube microelectrode-based flow injection system under low Cel/Credox conditions. Similarly to the effect of increase in the rotation rate at a rotating microdisk electrode,24 an increase in flow rate and/or a decrease in the electrode-capillary distance leads to decreased currents, which finally is a result of the increased potential drop across the depletion layer. We have shown that the dependences of the current on the position of the wall-tube microelectrode and the flow rate obtained for Cel/Credox ratios smaller that ∼1.0 × 10-3 are reversed compared to those for the presence of excess supporting electrolyte. A local current maximum is hence seen when the walltube microelectrode is perfectly aligned with the capillary. This feature could be used to facilitate the alignment of end-column electrochemical detectors in conjunction with, for example, capillary liquid chromatography and capillary electrophoresis. If one cannot center align the working microelectrode with the capillary tip, then to maintain quality of analytical measurements, he should rigorously keep the position of the microelectrode versus the tip unchanged. The present results also indicate that the buildup of the ionic layer at the wall-tube microelectrode is fast enough to ensure mass-transfer-controlled currents within seconds for Cel/Credox ratios larger that ∼1.0 × 10-3. This means that it is possible to obtain results similar to those found in the presence of an excess of supporting electrolyte also under low ionic strength conditions in a wall-tube microelectrode-based flow system. To further minimize the influence of convective mass transport of ions away from the electrode, it is possible to increase the distance between the electrode and the capillary. In this way, it would still be possible to utilize the high sample throughput inherent to a flow injection system operated at high flow rates. ACKNOWLEDGMENT This work was supported in part by a Polish grant from the State Committee for Scientific Research (BST-592/5/98) and by Grant K-AA/KU 09368-320 from the Swedish Natural Science Research Council. Received for review April 23, 1999. Accepted August 11, 1999. AC990430B Analytical Chemistry, Vol. 71, No. 21, November 1, 1999

4931