Effect of Change in Angle between Microelectrode Surface and Jet

The influence of stream orientation versus surface of microelectrode detector was examined in the range be- tween vertical and parallel flow for vario...
0 downloads 0 Views 246KB Size
Download PDF
Anal. Chem. 2005, 77, 5174-5181

Effect of Change in Angle between Microelectrode Surface and Jet Direction in Flow System on Current Response in Solutions of Different Ionic Strength Anna Maria Nowicka, Mikolaj Donten, Marcin Palys, and Zbigniew Stojek*

Department of Chemistry, Warsaw University, ul. Pasteura 1, 02-093 Warszaw, Poland

The influence of stream orientation versus surface of microelectrode detector was examined in the range between vertical and parallel flow for various jet velocities and various levels of supporting electrolyte. The flow cell was equipped with a conical body Pt microdisk electrode, and the measurements involved voltammetry and chronoamperometry. Ferrocene, two its charged derivatives (sodium ferrocenylo sulfate and ferrocenylomethyltrimethylamino hexafluorophosphate) and sodium iodide were employed as the substrates in the experiments. The strongest convectional transport and the highest signal of the analytes was obtained for r ) 60° (r is the angle between the electrode surface and the stream direction). The measured current increased by up to 1.85 times versus the traditional setup, and therefore, this new geometry of the detector is analytically advantageous. The value of r corresponding to the highest signal tended to decrease to ∼45° in the absence of supporting electrolyte provided that either flow rate or analyte concentration was above a certain threshold value. The experiments indicated that the interplay of the convectional and migrational components in the analyte transport is different for the charge increase and the charge cancellation processes. These experimental facts were confirmed by digital simulation results. Microelectrodes have attracted considerable attention as electrochemical detectors in various fields of analytical chemistry, including flow systems.1-3 There are several reasons for applying them in flow systems.4 First, the spherical diffusive mass transport at microelectrodes is more effective than the linear one observed at regular-size electrodes. Due to this fact, at microelectrodes, the signal-to-noise ratio is enhanced and the measured current is less affected by the convection.5 Nevertheless, the problem of enhancing the rate of mass transport to a microelectrode by using * To whom correspondence should be addressed. E-mail: stojek@ chem.uw.edu.pl. (1) Ewing, A. G.; Mesaros, J. M.; Gawin, P. F. Anal. Chem. 1994, 66, 527A. (2) Brazill, S. A.; Kuhr, W. G. Anal. Chem. 2002, 74, 3421. (3) Brazill, S. A.; Kim, P. H.; Kuhr, W. G. Anal. Chem. 2001, 73, 4882. (4) To´th, K.; Sˇ tulı´k, K.; Kutner, W.; Fehe´r, Z.; Lindner, E. Pure Appl. Chem. 2004, 76, 1119. (5) Bond, A. M.; Luscombe, D. L.; Davey, D. E.; Bixler, J. W. Anal. Chem. 1990, 62, 27.

5174 Analytical Chemistry, Vol. 77, No. 16, August 15, 2005

convection in addition to spherical diffusion has been approached both numerically and experimentally.6 Second, the steady state is rapidly reached also when scan voltammetric techniques are used in analytical work in combination with flow techniques.7,8 The third advantage is, that the currents flowing at microelectrodes are very low. This significantly reduces the potential ohmic drop, making the microelectrodes applicable in highly resistive media.9-11 Finally, the small size of microelectrodes opens up the possibility of using them in microscale experiments, e.g., in the capillary electrophoresis12-14 or microanalytical systems. The electrochemical studies published so far and cited above have been generally carried out in high ionic strength solutions. The influence of the flow on the electrochemical response of microelectrodes in cells of different construction and under conditions of excess supporting electrolyte has been a subject of numerous experimental and theoretical studies. The responses of microarray channel electrode,15 wall-jet microelectrode,16 microdisk electrode,17,18 and channel microband electrodes19,20 were simulated using various numerical models. Some attention was already paid to the behavior of rotating microdisk electrodes in the absence of supporting electrolyte,21-23 and theoretical predictions for such systems were presented too.24 (6) Compton, R. G.; Dryfe, R. A.; Alden, J. A.; Rees, N. V.; Dobson, P. J.; Leigh, P. A. J. Chem. 1994, 98, 1270. (7) Rueda, M. E.; Sarabia, L. A.; Herrero, A.; Ortiz, M. C. Anal. Chim. Acta 2003, 479, 173. (8) Wallenborg, S. R.; Markides, K. E.; Nyholm, L. Anal. Chim. Acta 1997, 344, 77. (9) Bento, M. F.; Thouin, L.; Amatore, Ch.; Montenegro, M. I. J. Electroanal. Chem. 1998, 443, 137. (10) Bento, M. F.; Thouin, L.; Amatore, Ch. . Electroanal. Chem. 1998, 446, 91. (11) Amatore, Ch.; Thouin, L.; Bento, M. F. J. Electroanal. Chem. 1999, 463, 45. (12) Ertl, P.; Emrich, Ch. A.; Singhal, P.; Mathies, R. A. Anal. Chem. 2003 (13) Zeng, Y.; Chen, H.; Pang, D.-W.; Wang, Z.-L.; Cheng, J.-K. Anal. Chem. 2002, 74, 2441. (14) Wang, J.; Chen, G.; Chatrathi, M. P.; Musameh, M. Anal. Chem. 2004, 76, 298. (15) Moldoveanu, S.; Anderson, J. L. J. Electroanal. Chem. 1985, 185, 239. (16) Alden, J. A.; Hakoura, S.; Compton, R. G. Anal. Chem. 1999, 71, 827. (17) Amatore, C. A.; Fosset, B. J. Electroanal. Chem. 1992 , 328, 21. (18) Tait, R. J.; Bury, P. C.; Finnin, B. C.; Reed, B. L.; Bond, A. M. J. Electroanal. Chem. 1993, 356, 25. (19) Compton, R. G.; Fisher, A. C.; Wellington, R. G.; Dobson, P. J.; Leigh, P. A. J. Phys. Chem. 1993, 99, 10410. (20) Alden, J. A.; Booth, J.; Compton, R. G.; Dryfe, R. A. W.; Sanders, G. H. W. J. Electroanal. Chem. 1995, 389, 45. (21) Gao, X.; White, H. S. Anal. Chem. 1995, 67, 4926. 10.1021/ac050498t CCC: $30.25

© 2005 American Chemical Society Published on Web 07/06/2005

Figure 1. Wall-tube microelectrode flow cell with the capability of changing the angle between the electrode surface and the solution stream.

In this paper, we report on the changes in the response of a wall-tube microelectrode in a flow system when the angle, R, between the microelectrode surface and the flow direction is changed. For this purpose, an electrochemical wall-tube flow cell with a conical body Pt disk microelectrode has been constructed. In this cell, R can be freely changed from 0° to 90° with a precision of 1°. The influence of the absence of supporting electrolyte in such a measurement arrangement was also examined. Positively and negatively charged and uncharged ferrocene derivatives as well as iodide ion were used as the analytes. EXPERIMENTAL SECTION Electrochemical Measurements. Linear scan voltammetry and chronoamperometry were performed using an Autolab potentiostat controlled via a personal computer. Short pulse time chronoamperometry was used to shorten substantially the time of generation of the products at the electrode surface and, in this way, to limit the influence of possible follow-up precipitation reactions. Platinum conical body disk microelectrodes with active area of 5, 10, and 25 µm in radius (nLab, Warsaw, Poland) served as the working electrode. The diameter of the tip of the conical glass body was ∼60 µm. Two platinum wires were used as the counter and the quasi-reference electrode, respectively. The Pt quasireference electrode was used to avoid contamination of low ionic strength solutions by the ions from the reference electrode internal solution. Before each use, the working electrode was polished with 0.3-0.05-µm Al2O3 powders on a wet pad. After polishing, the electrode was rinsed with a direct stream of ultrapure water ((Milli-Q, Millipore) to remove completely the (22) Stevens, N. P. C.; Bond, A. M. J. Electroanal. Chem. 2002, 538-539, 25. (23) Hilmi, A.; Luong, J. H. Anal. Chem. 2001, 73, 2536. (24) Oldham, K. B. J. Phys Chem. B 2000, 104, 4703.

aluminum oxide powder from the surface. The presence of remains of aluminum oxide on the electrode surface was found to lead to a decrease in the measured current. The electrode surface was inspected optically with an Olympus, model PME 3, inverted metallurgical microscope. In all experiments, the electrochemical cell was kept in a Faraday cage to minimize the electric noise. The experimental flow setup consisted of a pumping system based on the syringe principle, a laboratory-designed flow cell, and an optical stereoscopic microscope to control the position of the working electrode tip with respect to the microtube outlet. The construction of the flow cell (see Figure 1) is based on two independently movable Teflon cylinders: one holding the working electrode and the other one, the tube outlet. This mechanical solution allowed the change of R with the precision of 1°. An automatic piston buret (715 Dosimat, Metrohm) of very uniform flow and very low electrical noise was used as the injection system. The capillary tubing used was made of PEEK and had an internal diameter of 500 µm. The distance between the working electrode and the tube outlet was between 0.9 and 1.2 mm. The calculated Reynolds number, Re, was significantly smaller than the critical value of 2100 over which the flow becomes turbulent. This number was calculated using the following formula: Re ) FVl/η, where V is flow velocity, l is diameter of the capillary tubing, F is density, and η is dynamic viscosity. In our experiments, for the acetonitrile solution containing 0.1 M tetrabutylammonium perchlorate (TBAP), the Reynolds number equaled 1325 (η ) 0.4491 cP, V ) 1.5 m s-1, and F ) 0.7937 g cm-3), and for the O.1 M NaClO4 aqueous solution, Re amounted to 743 (η ) 1.036 cP, V ) 1.5 m s-1, and F ) 1.027 g cm-3). Therefore, the solution flow was assumed to be laminar. During microscopic examinations, no evidence of the formation of turbulence in the stream around the conical microAnalytical Chemistry, Vol. 77, No. 16, August 15, 2005

5175

5176

Analytical Chemistry, Vol. 77, No. 16, August 15, 2005

4.12 ( 0.15 4.26 ( 0.25 4.39 ( 0.51 7.48 ( 0.24 6.30 ( 0.19 7.62 ( 0.24 8.33 ( 0.19 6.71 ( 0.17 8.02 ( 0.14 8.27 ( 0.17 6.81 ( 0.15 8.27 ( 0.21 8.02 ( 0.20 6.34 ( 0.16 7.98 ( 0.17 6.91 ( 0.15 5.36 ( 0.13 6.69 ( 0.14 4.41 ( 0.09 4.49 ( 0.08 4.51 ( 0.10 7.87 ( 0.18 7.52 ( 0.11 7.54 ( 0.08 8.16 ( 0.11 8.14 ( 0.16 8.36 ( 0.21 Normalization factor: wave height in quiet solution.

7.96 ( 0.14 8.09 ( 0.12 8.54 ( 0.26 1.00 0.85 2.00 1.01 0.85 1.82

a

90° 0°

7.05 ( 0.11 7.08 ( 0.12 7.45 ( 0.10

8.46 ( 0.09 8.59 ( 0.11 8.94 ( 0.27

0° 30° 45°

IR,V)1.5/IV)0

60° 72°

γ)0

30° 45°

IR,V)1.5/IV)0

60° 72° 90° Theor

Fc Fc-TMA+ Fc-SO3-

(25) Lindsay, J. K.; Hauser, Ch. J. Org. Chem. 1957, 22, 355. (26) Myland, J. C.; Oldham, K. B. J. Electroanal. Chem. 1993, 347, 49. (27) Amatore, C. A.; Fosset, B.; Bartlet, J.; Deakin, M. R.; Wightman, R. M. J. Electroanal. Chem. 1988, 256, 255.

Exp

Preliminary Experiments in Quiet Solutions. In quiet solutions, the height of the voltammetric wave of ferrocene was generally independent of the support ratio (γ, the ratio of concentrations of supporting electrolyte and analyte). This is in agreement with the theory.26,27 The observed small departures for various support ratios resulted from the change in diffusion coefficients caused by the change in the viscosity of the solutions. For the charged ferrocene derivatives, the migrational contributions to the transport of the electroactive species in solutions with no deliberately added supporting electrolyte were in agreement with the theory. The experimental data obtained and the corresponding theoretical predictions are presented in the second and third columns of Table 1. Under chronoamperometric conditions and for all electroactive substrates, the oxidation limiting current was achieved in a short time and did not change within tens of seconds. Flow Direction Perpendicular to the Electrode Surface. As can be expected, for both low and high ionic strength, the solution flow directed perpendicularly toward the microelectrode surface caused significant increase in the current response in comparison to the quiet solutions. The shape of the chronoamperometric curves obtained under the hydrodynamic conditions was influenced as illustrated in Figures 2-4. Each figure in this set also contains two insets. The bottom insets present the corresponding voltammograms with the potentials selected for

system

(charge neutralization process, solvent: water) (III)

γ ) 50

FcSO3- f FcSO3° + e

(IV)0,γ)0/ IV)0,γ)50)

FcTMA+ f FcTMA2+ + e (charge increase process, solvent: acetonitrile) (II)

hydrodynamic conditions

Fc f Fc+ + e (charge production process, solvent: acetonitrile) (I)

quiet solution

RESULTS AND DISCUSSION We have employed ferrocene and its two derivatives as the voltammetric model systems. The corresponding electrode processes are

Table 1. Influence of Migration, Convection, and r on Normalized Height of Voltammograms of Selected Ferrocene Derivativesa

electrode was observed. This observation was further supported by digital simulations of the flow conditions (see section on simulation results). Temperature was maintained at 22 °C in all experiments. Reagents. Ferrocene (Fc, 98%) was purchased from Aldrich. Reagent grade sodium iodide was purchased from POCh, Poland. Sodium ferrocenesulfonate (Fc-SO3-Na+) and ferrocenylmethyltrimethylammonium hexafluorophosphate (FcTMA+PF6-) were synthesized in our laboratory according to the procedures available in the literature.25 The above substrates were dissolved either in deionized water (Milli-Q, Millipore, conductivity of 0.056 mS cm-1) or in acetonitrile (p.a., Fluka). Lithium perchlorate (99%) and TBAP (99%), both purchased from Fluka, were used as supporting electrolytes.

Figure 2. Experimental chronoamperograms of Fc obtained in the absence (b, d, γ ) 0) and the presence of excess (a, c, γ ) 50) supporting electrolyte. Quiet solution (a, b); hydrodynamic system (c, d). Top inset: limiting current as a function of flow rate for R ) 90°. Bottom inset: corresponding voltammograms with arrows indicating the potentials selected for the amperometric experiments. Conditions: acetonitrile, applied potential 4.3 V for γ ) 0 and 0.7 V for γ ) 50, conical body Pt disk microelectrode of 10 µm in radius, Fc concentration, 2 mM; flow rate, 0.63 m/s; supporting electrolyte, TBAP; potential scan rate, 10 mV/s.

Figure 3. Experimental chronoamperograms of ferrocenylmethyltrimethylammonium hexafluorophosphate (Fc-TMA+PF6-) obtained in the absence (a, c, γ ) 0) and in the presence of excess supporting electrolyte (b, d, γ ) 50). Quiet solution (a, b); hydrodynamic system (c, d). Top inset: limiting current as a function of flow rate for R ) 90°. Bottom inset: corresponding voltammograms with arrows indicating the potentials selected for the amperometric experiments. Conditions: acetonitrile, applied potential 0.7 V, conical body Pt disk microelectrode of 10 µm in radius, Fc-TMA+PF6- concentration, 2 mM; flow rate, 0.63 m/s; supporting electrolyte, TBAP; potential scan rate, 10 mV/s.

the chronoamperometric experiments, and the top insets illustrate how the limiting currents depend on the flow rate. The changes in the shape of the chronoamperograms were different for different substrates. For the oxidation of Fc and Fc-TMA+ (Figures 2 and 3), the limiting hydrodynamics currents were achieved in short time, and the initial decrease in the current could not be observed. Alternatively, in the case of Fc-SO3-Na+ (Figure 4), the initial current decrease is very well visible in the no-supportingelectrolyte chronoamperogram. In addition to the differences in the initial parts of the chronoamperograms, the changes in the steady-state limiting currents due to the flow, obtained in the absence of supporting electrolyte, were significantly different for all reagents examined, while these differences were within 5% for

Figure 4. Experimental chronoamperograms of sodium ferrocenesulfonate (Fc-SO3-Na+) obtained in the absence (b, d, γ ) 0) and in the presence of excess supporting electrolyte (a, c, γ ) 50). Quiet solution (a, b); hydrodynamic system (c, d). Top inset: limiting current as a function of flow rate for R ) 90°. Bottom inset: corresponding voltammograms with arrows indicating the potentials selected for the amperometric experiments. Conditions: water, applied potential 0.8 (flow conditions) and 0.55 V (quiet solution), conical body Pt disk microelectrode of 10 µm in radius; Fc-SO3-Na+ concentration, 2 mM; flow rate, 0.63 m/s; supporting electrolyte, NaClO4; potential scan rate, 10 mV/s.

the excess supporting electrolyte. This phenomenon will be addressed in other parts of the paper. Change of the Angle. We have found that for any applied flow rate the magnitude of the limiting voltammetric current depends strongly on the angle between the electrode surface and the stream direction. To avoid the changes in the current related to the departure from the coaxial positioning of the electrode versus the capillary-tube end,28 the microelectrode tip was placed exactly in the center of the flow stream using the microscope. Similarly, the distance between the microelelectrode tip and the capillary end was rigorously controlled. The obtained angular dependencies reflect the changes in the convectional transport rate and are presented in Figures 5-7. Ferrocene is an uncharged species; therefore, its transport rate should not be influenced by the presence or the absence of supporting electrolyte. As a result, the convectional contributions to the total transport for any R are identical and the two curves presented in Figure 5 overlap. Figure 5 also demonstrates that the intensity of convection is a function of R, and the largest contribution of convection to the transport occurs for the angle of 60°. The minimum current is reached at 0° (flow parallel to the electrode surface) and amounts to ∼52% of the maximum value. Fc-TMA+ is a monovalent cation, and its oxidation product is a divalent cation (charge increase process). Thus, both species migrate under conditions of low ionic support. While the γ ) 50 plot presented in Figure 6 is identical to the ferrocene plots in Figure 5, the plot obtained for γ ) 0 deviates from it substantially and this deviation (decrease in current) is a function of R. For R located between 60° and 90°, where the convection has the strongest effect, this decrease is the largest and amounts to ∼34%, which is significantly more than the drop of 15% predicted (28) Bjorefors, F.; Gadomska, J.; Donten, M.; Nyholm, L.; Stojek, Z. Anal. Chem. 1999, 71, 4926.

Analytical Chemistry, Vol. 77, No. 16, August 15, 2005

5177

Figure 5. Limiting voltammetric current of oxidation of Fc (normalized with respect to the value obtained in the quiet solution with excess supporting electrolyte) as a function of the angle between the electrode surface and the stream direction. No supporting electrolyte (solid line); excess supporting electrolyte (dashed line). Conditions: acetonitrile, conical body Pt disk microelectrode of 10 µm in radius; concentration, 2 mM; flow rate, 1.5 m/s.

Figure 6. Limiting voltammetric current of oxidation of ferrocenylmethyltrimethylammonium (Fc-TMA+) (normalized versus the value obtained in the quiet solution with excess supporting electrolyte) as a function of the angle between the electrode surface and the stream direction. No supporting electrolyte (solid line); excess supporting electrolyte (dashed line). Conditions: acetonitrile, conical body Pt disk microelectrode of 10 µm in radius; concentration, 2 mM; flow rate, 1.5 m/s.

theoretically for a quiet solution without supporting electrolyte versus excess supporting electrolyte.26 As the angle is set to 0°, where the convectional transport is the weakest, the difference between the currents measured in the presence and the absence of the supporting electrolyte is the smallest: it reaches 19%, which is only a little more than in the quiet solution. Apparently, for the Fc-TMA+/Fc-TMA2+ system, the presence of convection enhances the negative migrational effect in the measured current. Fc-SO3- is a singly charged anion, and its oxidation product is neutral. Thus, in the absence of excess supporting electrolyte, the substrate migrates and the measured currents are appropriately increased. In the quiet solution, this increase is by a factor of 2. Contrary to the “unsymmetrical” behavior of the previous system illustrated in Figure 6, now both IV/ISS plots obtained for the presence and the absence of excess supporting electrolyte are of similar dependence on R and the proportionality factor for the currents is close to 2. This is illustrated in Figure 7. According to the theory for a quiet solution,26 the migrational 5178 Analytical Chemistry, Vol. 77, No. 16, August 15, 2005

Figure 7. Limiting voltammetric current of oxidation of ferrocenesulfonate (Fc-SO3-) (normalized with respect to the value obtained in the quiet solution with excess supporting electrolyte) as a function of the angle between the electrode surface and the stream direction. No supporting electrolyte (solid line); excess supporting electrolyte (dashed line). Conditions: water, conical body Pt disk microelectrode of 10 µm in radius; concentration, 2 mM; flow rate, 1.5 m/s.

contribution to the limiting current is 100% of the diffusion current. Under the hydrodynamic conditions, the contribution of migration is slightly depressed in the range of high R and not at all at low R. Quantitatively, the influence of migration and convection on the voltammetric wave heights of ferrocene and its two derivatives, as shown in Figures 5-7, is presented in Table 1. In that table, the ratios of all limiting currents to the corresponding currents obtained in the quiet solution are given for selected values of R. The difference between Table 1 and Figures 5-7 is that that, in Table 1, in the γ ) 0 section, the currents are normalized versus the limiting current obtained in the quiet solution in the absence of supporting electrolyte. The data presented in the γ ) 50 section of Table 1 indicate that in the presence of excess supporting electrolyte the convective enhancements of the oxidation currents of all three ferrocene compounds are similar and do not differ from each other by more than 7%. In the absence of supporting electrolyte (section γ ) 0) and for the angle of the maximum convection (R ) 60°), the convectional current enhancements for Fc and Fc-SO3- again do not differ much. However, the oxidation current for Fc-TMA+ is substantially lower (by ∼21%). This decrease is substantially limited for R ) 0°, where the convection is much weaker. To examine the influence of the electrode radius and high concentration of the analyte, we have done additional voltammetric measurements for the oxidation of another anion. The anion selected was I-. Since the corresponding electrode product is uncharged (I2), one may expect for the conditions of no supporting electrolyte and high analyte concentrations lower current enhancements (IV/Iv)0) compared to those of lower analyte concentrations. First, we have found that with an increase in the electrode radius the IV/Iv)0 ratio increases. However, this increase was much smaller for R ) 90° than for R ) 0°. This effect is illustrated in Figure 8. The sensitivity to R has some consequences: the maximum in the top plot of Figure 8 (r ) 25 µm) shifts from 60° to 45°. A similar effect was observed for the oxidation of I- to I2 when the concentration of the analyte for a given electrode radius

Table 2. Values of Parameter b in Eq 2 for V in the Range 0-1.5 M/s parameter b experimental values γ)0

a

R (deg)

uncharged substrate Fc

90 72 60 45 30 0

0.536 ( 0.017 0.509 ( 0.008 0.493 ( 0.003 0.476 ( 0.007 0.434 ( 0.006 0.304 ( 0.018

charged substrates FcTMA+ FcSO30.488 ( 0.041 0.416 ( 0.038 0.404 ( 0.037 0.354 ( 0.040 0.336 ( 0.035 0.321 ( 0.029

0.505 ( 0.031 0.442 ( 0.028 0.421 ( 0.025 0.387 ( 0.022 0.377 ( 0.016 0.313 ( 0.028

γ ) 50b 0.504 ( 0.003 0.462 ( 0.002 0.450 ( 0.004 0.424 ( 0.002 0.406 ( 0.002 0.364 ( 0.006

lit. valuesa 0.488 ÷ 0.532

0.321 ÷ 0.391

Reference 29. b Mean for all substrates.

Figure 8. Limiting voltammetric current of oxidation of iodide (normalized with respect to the value in the quiet solution) as a function of the angle between the electrode surface and the stream direction. Conditions: no supporting electrolyte, water, conical body Pt disk microelectrode of 5 (a), 10 (b), and 25 µm in radius (c). Concentration, 2 mM; flow rate, 1.5 m/s.

was increased. This is illustrated in Figure 9. We attribute the shift of the maximum in the angle plots to the increased potential ohmic drop in the solution. This drop is apparently related to angle R and reflects varying abilities of the flow to hinder the accumulation of the counterions in the diffusion layer. Algebraic Formulas Fitting the Experimental Data. The angular dependencies of the currents obtained for all the studied compounds in the presence of excess supporting electrolyte and under hydrodynamics conditions are presented in Table 2 (γ ) 50). It can seen that for all examined analytes the numbers illustrating the current increase by convection for a given R are very similar. Also, by examining all γ ) 50 data obtained, we have found that for each analyte the current is a simple function of R and can be represented by

IR,V ) cos(R)I0°,v + sin(R)I90°,v

(1)

where I0°,v and I90°,v are the convective currents obtained at a given flow rate, v, and for the parallel and perpendicular orientations, respectively. The deviations of the experimental results from eq 1 are smaller than 4% for flow rates up to 1 m/s.

Figure 9. Limiting voltammetric current of oxidation of iodide (normalized with respect to the value in the quiet solution) as a function of the angle between the electrode surface and the stream direction. Conditions: water, conical body Pt disk microelectrode of 10 µm in radius; concentration, 0.5 (a), 2 (b), 4 mM (c); flow rate, 1.5 m/s.

The currents obtained for each angle were also examined versus the flow rate. The following equation was fitted, using the Sigma Plot v. 5.0 software, to the data obtained:

IR,v ) IR,V ) 0 + aVb

(2)

where IR,V)0 is the limiting current obtained in the quiet solution for angle R, and a and b are the coefficients that depend on the flow rate (a), bulk concentration (a), viscosity (a), and geometry of the electrode (a and b),29 and finally, as we present in this paper, on the angle between the electrode surface and the stream direction (a and b). Coefficient b changes from 0.5 to 0.33 with the change in the angle from 90° to 0°. The value of 0.5 for angle 90° suggests that in this case the convectional contributions correspond well to the case of rotating disk electrode. For position R ) 0°, coefficient b equals ∼0.33 and this appears to be a good analogy to the behavior of a channel electrode. Interestingly, while the function of the microelectrode current response versus R is peak shaped, the values of coefficient b diminish continuously with the decrease in R. This is illustrated in Table 2. (29) Kubiak, W. W.; Strozik, M. M. J. Electroanal. Chem. 1996, 417, 95.

Analytical Chemistry, Vol. 77, No. 16, August 15, 2005

5179

Digital Simulations. To facilitate the interpretation of the experimental results, we have carried out digital simulations and made appropriate comparisons. The simulations have been done employing software package MIOTRAS that uses the finite volume method combined with the multidimensional upwinding method.30 The chosen software is capable of modeling mixed diffusional, migrational, and convectional transport. Because of symmetry requirements imposed by the package, only the coaxial position of the working electrode (i.e., R ) 90°, wall-jet microelectrode) could be modeled. The simulation of the flow field at the electrode surface has shown that if the jet direction is perpendicular to the electrode surface the solution stream is dispersed by the electrode body and a zone of relatively quiet solution is formed. This zone is of conical shape with the cone tip pointing toward the jet outlet. The cone base rests on the disk, and its size exceeds the size of the active area of the electrode body; thus, the flow of the solution near the electrode has laminar character. The obtained simulated voltammograms for the flow rate of 0.63 m/s are compared with the experimental ones in Figure 10a and b. The current scale for the set of simulated voltammograms was defined in such a way that the experimental and calculated waves have got equal heights in the case of excess supporting electrolyte and zero flow conditions. The agreement between theory and the experiment in terms of the wave heights is very good. Especially important here is the fact that, for the case of the charge increase (FcTMA+ f FcTMA2+ + e, Figure 10 a), the simulations confirm the enhancement of the negative effect of migration on the limiting currents. In other words, the simulations confirm that the experimentally observed effects result entirely from the transport phenomena and not from chemical complications that could take place in the studied systems. Some differences appear between the voltammogram slopes and positions. This concerns the case of γ ) 0 only. Most likely the differences arise from unequal ohmic potential drops in the simulations and experiments. In the experiments, the γ ) 0 condition cannot be obtained, since some unwanted ions are always present in the solution and their concentration cannot be determined easily. The use of quasi-reference electrodes is also partly responsible for the difference in the half-wave potentials. CONCLUSIONS Voltammetric and chronoamperometric experiments show that the orientation of the microelectrode versus stream direction significantly influences the current response in both the diffusional/convectional and diffusional/migrational/convectional conditions. For all studied analytes and most experimental conditions, the largest enhancement of the current by the convection was observed for angle 60°. This means that for this angle the convectional transport is the strongest. The flow parallel to the electrode surface, R ) 0°, results in the weakest convection and therefore was characterized by the smallest enhancement of the current versus the quiet solution. Quantitatively, for R ) 60°, the measured current increased by up to 1.85 times versus the traditional setup and therefore this new geometry of the detector is apparently analytically advantageous. It is important that that (30) Bortels, L.; Deconinck, J.; Van Den Bossche, B. J. Electroanal. Chem. 1996, 404, 15.

5180 Analytical Chemistry, Vol. 77, No. 16, August 15, 2005

Figure 10. Experimental voltammograms (symbols) compared with simulation results (solid lines). The current scale for the theoretical voltammograms was fixed by assuming that for the case of the presence of excess supporting electrolyte the experimental and theoretical voltammograms are identical. Analytes, Fc-TMA+ (a), Fc-SO3- (b); potential scan rate, 10 mV/s.

signal increase is not affected by deficiency of supporting electrolyte. In the absence of deliberately added supporting electrolyte, the convection influenced the measured current in different ways for different types of electrode processes. For uncharged analytes (ferrocene, charge production process), the voltammetric wave plateaus were practically identical in the absence and presence of supporting electrolyte. For the oxidation of the ferrocenesulfonate anion (charge cancellation process), the convection slightly decreased the migrational effect seen in quiet solutions. And finally, for the oxidation of the ferrocenylmethyltrimethylammonium, the convection enhanced substantially the migrational effect. While in a quiet solution, in the absence of supporting electrolyte, the voltammetric wave is lower only by 15% compared to excess supporting electrolyte; in the solution with the flow, it is lower by 34% at the angle of strongest convection. The latter number becomes much closer to 15% for the R ) 0 orientation, where the convection is much weaker. The strikingly different influence of convection on the voltammetric wave heights of the charged species was clearly confirmed by the digital simulations. However, some comments are needed here. The oxidation of the ferrocenesulfonate anion leads to the uncharged product that formally does not need a counterion (for the preservation of electroneutrality). The case of the oxidation of ferrocenylmethyltrimethylammonium cation is different. The electrode product is doubly charged (2+) and needs

an extra counterion. In the convective environment, to preserve the electroneutrality, it is apparently easier to repel the substrate of the same charge sign than to attract the negative counterion.

4 T09A 052 24 from KBN, the Polish Committee for Scientific Research.

ACKNOWLEDGMENT

Received for review March 24, 2005. Accepted June 7, 2005.

We thank Wojciech Ochmanski for his mechanical artwork in the cell construction process. This work was supported by Grant

AC050498T

Analytical Chemistry, Vol. 77, No. 16, August 15, 2005

5181