Anal. Chem. 2003, 75, 2086-2091
Characteristic Electrochemical Responses of Polymer Microchannel-Microelectrode Chips Kosei Ueno, Haeng-Boo Kim,† and Noboru Kitamura*
Division of Chemistry, Graduate School of Science, Hokkaido University, Sapporo 060-0810, Japan
Polymer microchannel chips (dimension of the channel, 20 µm (depth) × 100 µm (width)) integrated with dual microband electrodes were fabricated by an imprinting method, and their characteristic electrochemical responses were elucidated in detail. A silicon micromachined template fabricated by photolithography and dry etching techniques was used for imprinting. An aqueous solution of a ferrocene derivative was brought into the microchannel by pressure-driven flow, and the electrode responses were studied on the basis of voltammetry and chronoamperometry. A linear sweep voltammetry of 1-hydroxyethylferrocene (FeCp-OH) in water demonstrated that the electrode responses in the microchannel chip were best characterized by one-dimensional diffusion along the channel length, reflecting the structural dimension of the channel. In generation-collection mode experiments, furthermore, a collection efficiency as high as ∼90% was attained in the microchannel owing to both restricted space and characteristics of solution flow in the channel. It was confirmed that diffusion and solution flow made molecular transport very efficient in the microchannel. The experimental results were also compared with those predicted by computer simulations. Microfluidic devices have been combined with various detection techniques, such as laser-induced fluorescence (LIF),1,2 mass spectrometry,3,4 thermal-lens microscopy,5 and so forth. Although these detection techniques are certainly very important to analyze a trace amount of a sample, a relatively large off-chip apparatus is required, with a typical example being a laser or microscope. In contrast to these methodologies as detection techniques, electroanalytical techniques are very advantageous, since an electrode as a detector can be miniaturized arbitrarily and incorporated into a microfluidic device. In practice, micromachining technologies have succeeded in fabricating microelectrodes, and the electrodes are employed as a sensitive detector in liquid chromatography,6,7 capillary electrophoresis,8,9 biosensors,10 and so on. Therefore, a * To whom correspondence should be addressed. E-mail: kitamura@ sci.hokudai.ac.jp. † Present address: Engineering Research Institute, The University of Tokyo, Yayoi, Bunkyo-ku, Tokyo, Japan. (1) Haab, B. B.; Mathies, R. A. Anal. Chem. 1995, 67, 3253-3260. (2) Chiem, N.; Harrison, D. J. Anal. Chem. 1997, 69, 373-378. (3) Xue, Q.; Foret, F.; Dunayevskiy, Y.; Zavracky, P. M.; McGruer, N. E.; Karger, B. L. Anal. Chem. 1997, 69, 426-430. (4) Ramsey, R. S.; Ramsey, J. M. Anal. Chem. 1997, 69, 1174-1178. (5) Hibara, A.; Sato, K.; Hisamoto, H.; Uchiyama, K.; Slyadnev, M. N.; Tokeshi, M.; Kitamori, T. Prog. Nat. Sci. 2001, 11 (Suppl.), S237-S241. (6) Armentrout, D. N.; McLean, J. D.; Long, M. W. Anal. Chem. 1979, 51, 1039-1045.
2086 Analytical Chemistry, Vol. 75, No. 9, May 1, 2003
microfluidic device integrated with microelectrodes is very useful in µ-TAS and lab-on-a-chip systems.11,12 On the other hand, we reported recently that an imprinting method first proposed by Martynova et al.13 was a very powerful means to fabricate polystyrol microchannel chips and demonstrated that the fabricated polymer chips were applied successfully to an automated photochemical reaction14 and liquid/liquid extraction.15 Owing to its convenient fabrication method, we fabricated symmetrical and unsymmetrical zigzag side-walled microchannel chips and demonstrated channel-shape effects on a liquid/liquid extraction efficiency.16 In addition to such achievements, we recently found that a tough metal thin layer could be vacuumdeposited directly onto a polystyrol substrate. On the basis of these results, we succeeded in fabricating polystyrol microchannelmicroelectrode chips.17 It is worth emphasizing that metal microband electrodes so far reported are fabricated generally on a silicon or glass substrate with a chromium layer being used as an adhesive layer between the substrate and the electrode. Therefore, an electrode with a finite thickness sits on a solid substrate. It is inconvenient to fabricate a silicon- or glass-based microchannel-microelectrode chip, since the presence of an electrode, lead wire(s), and bonding pad(s) makes it very difficult to bond the substrate with a microchannel plate. In addition, a thick electrode will influence solution flow characteristics in the microchannel. In contrast to these chip materials, polymer substrates having elasticity are very advantageous, since an electrode and wire(s) can be buried into the polymer substrate, as shown by preliminary results reported by the present authors.17 As related works, furthermore, Girault and co-workers recently reported poly(ethylene terephthalate) microchannel-microelectrode chips fabricated by a laser ablation technique18 and demonstrated characteristic electrochemical (7) Takahashi, M.; Morita, M.; Niwa, O.; Tabei, H. J. Electroanal. Chem. 1992, 335, 253-263. (8) Wang, J.; Tian, B.; Sahlin, E. Anal. Chem. 1999, 71, 5436-5440. (9) Martin, R. S.; Gawron, A. J.; Lunte, S. M.; Henry, C. S. Anal. Chem. 2000, 72, 3196-3202. (10) Niwa, O.; Morita, M.; Tabei, H. Electroanalysis 1994, 6, 237-243. (11) Manz, A., Becker, H., Eds. Microsystem Technology in Chemistry and Life Sciences; Springer-Verlag: Berlin; 1999. (12) Harrison, D. J., Van Den Berg, A., Eds. Micro Total Analysis Systems ’98; 1998; Kluwer Academic Publishers: Dordrecht. (13) Martynova, L.; Locascio, L. E.; Gaitan, M.; Kramer, G. W.; Christensen, R. G.; MacCrehan, W. A. Anal. Chem. 1997, 69, 4783-4789. (14) Ueno, K.; Kitagawa, F.; Kitamura, N. Lab. Chip 2002, 2, 231-234. (15) Kim, H.-B.; Ueno, K.; Chiba, M.; Kogi, O.; Kitamura, N. Anal. Sci. 2000, 16, 871-876. (16) Ueno, K.; Kim, H.-B.; Kitamura, N. Anal. Sci. 2003, 19, 391-394. (17) Ueno, K.; Kitagawa, F.; Kim, H.-B.; Tokunaga, T.; Matsuo, S.; Misawa, H.; Kitamura, N. Chem. Lett. 2000, 858-859. 10.1021/ac0264675 CCC: $25.00
© 2003 American Chemical Society Published on Web 03/25/2003
Figure 1. Structural layouts of the microchannel chip integrated with electrodes: (a) the whole chip layout and (b) the structures of dual working electrodes. WE, working electrode; CE, counter electrode; and RE, reference electrode. The sizes in the figure are shown in millimeters (a) and micrometers (b).
responses in the microchannel. They reported that the microchannel-electrode chip was very useful as a detection system in an electrophoresis chip.19 Despite high potentials of a polymer microchannel-microelectrode chip, nonetheless, the number of reports on both a fabrication method and electrochemical characteristics is still limited, and work along that line is worth exploring in more detail. It is easily expected that, since the electrochemical responses of the electrode incorporated in a microchannel are governed by various factors, such as the dimensions of the channel and the electrode, a potential sweep rate, a solution flow rate, and so forth, systematic studies are absolutely necessary. In the present article, the fabrication method of a polystyrene microchannel-microelectrode chip and the characteristic responses of the microelectrode are reported. To discuss the electrode responses, the experimental results observed by linear sweep voltammetry and generationcollection mode experiments were compared with those predicted by numerical simulations. EXPERIMENTAL SECTION Sample Preparation. 1-Hydroxyethylferrocene (FeCp-OH, Tokyo Kasei Kogyo Co. Ltd., GR Grade) was purified by repeated recrystallizations from ethanol and used for electrochemical measurements throughout the study. Water was distilled and deionized by a Milli-Q Lab system. Potassium chloride (KCl, Wako Pure Chemical Industries Ltd., GR Grade) was used as a supporting electrolyte. The concentrations of FeCp-OH and KCl in water were set at 1.0 × 10-3 and 0.1 mol/dm3, respectively. Fabrication of a Polymer Channel-Electrode Chip. A fabricated polymer microchip was composed of channel and electrode substrates. Commercially available polystyrol substrates (Tamiya Co. Ltd., 15 × 30 mm) were used as the chip material. The structural layout of the microchannel chip integrated with electrodes is shown in Figure 1. In the present study, the width, depth, and total length of the microchannel were set at 20 µm, 100 µm, and 60 mm, respectively. A silicon template for imprinting was fabricated by photolithography and dry-etching techniques. The silicon template and a polystyrol substrate were fastened tightly between two glass plates and heated at 110 °C for 25 min to transfer the embossed structure to the polymer plate. The SEM images of the chip (see Supporting Information) indicate (18) Rossier, J. S.; Roberts, M. A.; Ferrigno, R.; Girault, H. H. Anal. Chem. 1999, 71, 4294-4299. (19) Rossier, J. S.; Ferrigno, R.; Girault, H. H. J. Electroanal. Chem. 2000, 492, 15-22.
that the depth of the channel is somewhat shallower (20 µm), as expected, than the height of the template (25 µm), and the bottom edges of the polymer channel are not necessarily sharp. Nonetheless, the embossed structure of the template was transferred fairly well to the polymer substrate. As discussed later in detail, the channel structures were enough for the present purpose of the study. On the other hand, an electrode substrate was fabricated as follows. First, a gold thin film (thickness of 100 nm) was vacuumdeposited directly onto a polystyrol substrate (5.0 × 10-6 Torr). The Au/polymer substrate was then annealed in an oven at 108 °C for 30 min, in which the annealing temperature was set just above the glass transition temperature of the polymer (∼108 °C). To fabricate microelectrodes, a photoresist (Tokyo Ohka Kogyo Co. Ltd., OFPR800: positive type) was spin-coated onto the Au/ polymer substrate and exposed to a 300-W tungsten lamp through a photomask whose structure was drawn on a transparency film by a computer software package and a 2400 dpi printer. The resist layer was developed by an alkaline solution (Tokyo Ohka Kogyo Co. Ltd., NMD-3) and rinsed with pure water, and the exposed Au layer was etched in an aqueous I2 (1.5 g)/NH4I (8 g) solution (100 mL) at 50 °C for 5 s. The fabricated Au band microelectrodes were employed as working (WE) and counter electrodes (CE). The Au band electrodes with the width of 47 µm were fabricated perpendicularly to the direction of the channel length, as shown in Figure 1. Thus, the electrode length is equal to the channel width (100 µm). To conduct generation-collection mode experiments, two Au working electrodes were set separated at 68 µm (edge-to-edge distance, Figure 1b). For integration of a reference electrode (RE, width ) 200 µm), a silver thin layer (∼100 nm) was vacuum-deposited directly onto the Au electrodes/polymer substrate through a plastic mask. The Au-Ag/polymer substrate was then clamped between two glass plates and heated at 108 °C for 25 min to bury the electrodes into the polymer substrate. For an electrode fabricated on a glass or silicon substrate, the finite thickness of the electrode is sometimes troublesome, as discussed above. Therefore, we think that elasticity of the polystyrol substrate, capable of burying the electrodes into the substrate, is one of the important properties as a chip material. Finally, the electrode substrate was covered and bonded with the channel substrate by pressing the two substrates between two glass plates at 108 °C for 18 min. Electrochemical Measurements. A solution flow system analogous to that reported previously was used in the present study.14-17 A syringe pump (Harvard Co., model 44) was used to control the flow velocity in the microchannel. A connection between the bonding pads on the channel-electrode chip and an electrochemical analyzer (ALS, model 701A) via lead wires was made by a silver paste as an electroconductive adhesive, and they were fixed with an epoxy resin. Numerical Simulations of Electrochemical Responses. To discuss both experimental results and the performances of the microchannel-microelectrode chip, numerical simulations of the electrochemical responses in the chip were performed on the basis of a finite element method (FEM) by using a computer software package (PDEase 2D, Macsyma Inc.). Although three-dimensional analyses of the concentration profiles at a microdisk electrode are necessary, we conducted two-dimensional analyses in the present Analytical Chemistry, Vol. 75, No. 9, May 1, 2003
2087
Figure 2. Linear sweep voltammograms of FeCp-OH observed by the channel chip (a) and the electrode substrate (without a channel cover, b) at ν of 5 (- - - -), 10 (‚‚‚‚‚‚), 20 (- - - -), or 50 (s) mV/s.
system, since the concentration profile along the direction of the electrode-length (equal to the channel width) is identical. Considering a simple redox reaction, as in the case for that of ferrocene without migration and convection, the concentration profile of a redox specimen can be calculated on the basis of the Fick’s second law, eq 1.18,20
∂CR ) DR∇2CR ∂t
(1)
where ∇2 is the Laplacian operator, and that for an inlaid band electrode is given by the following equation:
∇2)
∂2 ∂2 + 2 2 ∂x ∂z
For the calculations, the following initial and boundary conditions were assumed. Initial conditions: 0
0
at t ) 0, CR ) CR , CO ) 0
(2)
Boundary conditions:
at t > 0,
CR CR
) 0
0 θe-γ nF with θ ) enF/RT(Ei - E ) γ ) ν -γ RT 1+θ (3)
at t > 0, at the channel wall
[ ] ∂CR ∂N
wall
)0
(4)
where CR and CO are the concentrations of reduced (R, CR0 ) 1.0 × 10-6 mol dm-3) and oxidized species (O), respectively. DR and DO are the diffusion coefficients of R and O (DR ) DO ) 2.8 × 10-6 cm2 s-1), respectively. Ei and E0 are the initial and formal potentials, respectively, and t is a time. n, F, R, T, and ν are the number of an electron transferred (n ) 1 in the present case), the Faraday constant (C mol-1), the gas constant (J mol-1 K-1), a temperature (298 K), and a potential sweep rate (V s-1), respectively. N is the distance from the channel wall and taken as normal to the wall. A current at an electrode (i(t)) was then simulated as time-dependent flux of FeCp-OH at 2088
Analytical Chemistry, Vol. 75, No. 9, May 1, 2003
the electrode, as given by eq 5.
[
i(t) ) nFADR
]
∂CR(x, t) ∂x
x)0
(5)
where A is the cross-sectional area of a band electrode. Under solution flow conditions, on the other hand, the flow profile along the direction of the channel depth (z axis) is parabolic, while that along the channel width (y axis) is plug-like, since the aspect ratio of the present microchannel is low: z (20 µm)/y (100 µm). Therefore, the flow velocity along the y axis is assumed to be almost constant, while that along the z axis depends on the depth position in the microchannel. Consequently, the numerical simulation can be simplified and conducted by a twodimensional approximation, since the flow profile along the y axis is not a parameter.21,22 The flow profile along the z axis was then simulated by the Navier-Stokes equation, which was considered to simulate i(t) under solution flow conditions in the microchannel. RESULTS AND DISCUSSIONS Electrochemical Responses of the Microelectrode in the Channel Chip. Under non-solution-flow conditions, linear sweep voltammograms of FeCp-OH at several ν’s were measured by using both the channel-electrode chip and the electrode substrate (i.e., without a channel substrate). The results are shown in Figure 2. For the experiments with the channel chip, since complete deaeration of the sample solution was difficult, voltammetry was conducted under aerated conditions. Therefore, the anodic current by O2 reduction was observed in the potential region negative to 0 V. Since this is not a main issue of the present study, we neglected the anodic current in the following discussions. The oxidation potentials of FeCp-OH determined by the channel chip (Figure 2a) and the electrode substrate (Figure 2b) were 164 and 161 mV (vs Ag), respectively, which agreed very well with the value determined by a conventional disk electrode: 163 mV (vs Ag). It is well-known that the electrochemical response of a microband electrode is best characterized by hemicylindrical diffusion and, in practice, quasi-steady-state currents were ob(20) Bard, A. J., Faulkner, L. R., Eds. Electrochemical Methods: Fundamentals and Applications, 2nd ed.; John Willey & Sons: New York, 1980; 148-153, 168-190, (21) Kim, H.-B.; Kogi, O.; Kitamura, N. Anal. Chem. 1999, 71, 4338-4343. (22) Kamholz, A. E.; Weigl, B. H.; Finlayson, B. A.; Yager, P. Anal. Chem. 1999, 71, 5340-5347.
Figure 3. Simulated linear sweep voltammograms and the relevant concentration profiles (upper traces) of FeCp-OH predicted for the microchannel chip (a) and the electrode substrate (b). For the voltammograms, the sweep rate was set 5 (- - -), 10 (‚‚‚‚‚‚‚), 20 (- - - -), or 50 (s) mV/s. The concentration profiles are the results at 90 s after the start of the potential sweep (ν ) 5 mV/s) from 0 to 450 mV.
served in the voltammograms (Figure 2b). Furthermore, in the case of a band electrode, the shape of a voltammogram is governed by the electrode width. Namely, when the electrode width is small as compared to the thickness of a diffusion layer and ν is slow enough, the voltammogram becomes a sigmoidal shape. In the case of the electrode substrate (Figure 2b) at ν ) 5 mV/s, the sigmoidal voltammogram demonstrates that the thickness of the diffusion layer is large compared to the electrode width. On the other hand, even at the same sweep rate, the shape of the voltammogram observed by the channel chip (Figure 2a) was different from that in Figure 2b, and the voltammogram showed a clear peak irrespective of ν. These behaviors are very similar to the current-potential curve observed for one-dimensional linear diffusion. Clearly, the presence of the microchannel wall above the electrode influences the diffusion profile and, thus, the electrode responses. To understand these results in more detail, numerical simulations of the responses were conducted as described below. The simulated voltammograms and the relevant concentration profiles of the FeCp-OH cation generated at the electrode in the channel chip (a) or on the substrate (b) are shown in Figure 3. The concentration profiles shown by the contour lines or curves are the results at 90 s after the start of the potential sweep (ν ) 5 mV/s) from 0 to 450 mV. In the case of the electrode substrate (Figure 3b), hemicylindrical diffusion of the FeCp-OH cation takes place at the electrode, and this gives the voltammograms in the lower panel, which reproduces very well the observed ones in Figure 2b. For the microchannel chip, on the other hand, the simulations predict one-dimensional linear diffusion, since the diffusion space around the electrode is restricted by the presence of the microchannel (channel depth ) 20 µm) and the channel length is infinite (total length ) 6 cm), as compared to the thickness of the diffusion layer. The simulations also reproduced almost satisfactorily the observed voltammograms in Figure 2a. It is concluded, therefore, that the electrochemical responses in the microchannel chip are best characterized by one-dimensional
diffusion along the channel length and governed by the spatial geometry around the electrode. Generation-Collection Mode Experiments under NonSolution-Flow Conditions. To elucidate further the electrochemical characteristics of the channel-electrode chip, we explored generation-collection (GC) mode experiments. In the GC mode, a solute is electrolyzed at one electrode (generator, G) and the electrolyzed solute is oxidized or reduced at the adjacent electrode (collector, C).23-25 In the present experiments, WE1 and WE2 in Figure 1 were used as G and C, respectively, and the potential at G was swept from 0 to 450 mV while that at C was kept 0 V. The collection efficiency (η) defined as the ratio of the peak current at C (ic) to that at G (ig) was determined at several ν’s. The results on η determined with the microchannel chip (closed circles) and the electrode substrate (without a channel plate, closed triangles) under non-solution-flow conditions are summarized in Figure 4. It is worth emphasizing that the η value observed by the channel chip is almost two times larger than that by the electrode substrate, irrespective of ν. At ν ) 5 mV/s, as an example, the η value by the channel chip was as high as 45%, even for the edge-to-edge GC distance of 68 µm. Since the FeCpOH cation generated at G diffuses to both sides of the electrode along the channel-length direction, η should be 50% as the maximum value. The observed η value thus implies almost complete collection of the cation at C. In the case of dual-band microelectrodes without a cover substrate, such a high collection efficiency cannot be realized with an interelectrode distance larger than 10 µm. Therefore, the presence of the microchannel wall above G and C provides extraordinary effects on η. The open circles and triangles in Figure 4 represent the simulated collection efficiencies for the microchannel chip and the electrode substrate, respec(23) White, H. S.; Kittlesen, G. P.; Wrighton, M. S. J. Am. Chem. Soc. 1984, 106, 5375-5377. (24) Bard, A. J.; Crayston, J. A.; Kittlesen, G. P.; Sher, T. V.; Wrighton, M. S. Anal. Chem. 1986, 58, 2321-2331. (25) Niwa, O.; Morita, M.; Tabei, H. Anal. Chem. 1990, 62, 447-452.
Analytical Chemistry, Vol. 75, No. 9, May 1, 2003
2089
Figure 4. The potential sweep rate (ν) dependence of the collection efficiency (η). The closed circles and triangles are the data obtained by the microchannel chip and the electrode substrate, respectively. The open circles and triangles represent the results by numerical simulations for the channel chip and the electrode substrate, respectively.
tively. Although the observed collection efficiency is slightly larger than the relevant simulated value for the channel chip, the results almost agreed well with each other. One-dimensional linear diffusion along the channel length is the primary reason for the high collection efficiency in the chip. Furthermore, although the data are not shown here, we confirmed that the channel depth, shallower than 100 µm, gave results analogous to those in Figure 4. Generation-Collection Mode Experiments under Solution Flow Conditions. GC mode experiments were also conducted under solution flow conditions by using the microchannel chip, with WE1 (upstream side) and WE2 (downstream side) being used as G and C, respectively. A typical example of the voltammograms recorded at G and C (ν ) 20 mV/s) is shown in Figure 5a (linear flow velocity (u) ) 8.3 mm/s). It is worth noting that the voltammograms observed under the solution flow conditions are sigmoidal, which is in marked contrast to those in Figure 2. This is due to the fact that efficient mass transport of FeCp-OH or the cation to G or C, respectively, takes place by both diffusion and solution flow. Further information on this point is obtained by a ν dependence of the η value. As shown in Figure 6b (closed circles), the η value was almost constant at 28%, irrespective of ν (u ) 8.3 mm/s). It is interesting to note that without solution flow, the η value increased from 23 to 45% with a decrease in ν from 20 to 5 mV/s (Figure 4). The results indicate clearly that the electrode reaction is fast enough, and mass transport of FeCpOH or the cation to the relevant electrode is governed essentially by solution flow. Therefore, the collection efficiency does not depend on the sweep rate at such a fast flow rate. The η values obtained by the simulations shown by the open circles in Figure 5b agreed very well with the observed one, which supported the above discussion, as well. The above discussion indicates that the η value should depend on the solution flow rate (u). Furthermore, the absence of a ν dependence of η at ν ) 50-5 mV/s and u ) 8.3 mm/s in Figure 5 demonstrates that a slower flow rate and a much faster potential sweep rate are necessary to study a u dependence of η. Therefore, we explored chronoamperometry to study a u dependence of η. In the actual experiments, the time profiles of ig and ic were determined by stepping the potential at G from 0 to 350 mV, while C was set 0 V. A typical example of the results at u ) 1200 (a) and 120 µm/s (b) is shown in Figure 6. Since the cation of FeCpOH generated at G is transported compulsorily to C by solution 2090
Analytical Chemistry, Vol. 75, No. 9, May 1, 2003
flow, the anodic current observed at C increased with a time. On the basis of the results in Figure 6, the time necessary to reach the saturated ic value was determined to be 0.1 or 0.4 s at u ) 1200 or 120 µm/s, respectively. Experiments analogous to those in Figure 6 were conducted at several u’s, and the steady-state ig or ic value in each time profile was plotted against u, as shown in Figure 7a. The currents at G (closed circles) and C (open circles) increased to the positive and negative directions, respectively, with an increase in u and saturated at almost constant values above a certain u value. The results on ig indicate that FeCp-OH provided to G by solution flow is oxidized almost completely at a low u value (solid line in Figure 7a), while an increase in u resulted in over-feed of FeCp-OH to G, and therefore, the current is limited by the charge-transfer rate at G: leveling-off ig. On the other hand, since the flow velocity is very fast, most of the FeCp-OH cation produced at G is sped away to the downstream side of C without the electrode reaction. At a given u, therefore, ic is saturated at a slower flow velocity than u showing the leveling off of ig at G. Depending on the time scales of solution flow and self-diffusion of FeCp-OH/FeCp-OH cation, the collection efficiency shows a maximum value at a certain u. In the present case, the η value of 87% was attained at u ) 30 µm/s, as shown in Figure 7b. The solid curve in the figure represents the results by the numerical simulations, which reproduces very well the experimental observations. Therefore, the results in Figure 7a are reasonably explained along the context described above. It is worth emphasizing that the collection efficiency at u ) 30 µm/s was as high as 87%, even for the edge-to-edge GC distance of 68 µm. In the case of ordinary dual-band electrodes, such a high collection efficiency is realized only with an interelectrode distance of several micrometers. Therefore, solution flow in the microchannel provides extraordinary effects on η. Although the collection efficiency is almost constant at u > 500 µm/s, that is strongly dependent on u at u < 500 µm/s as the results of the relationship between the time scales of mass feed to G/C and the charge-transfer rate. The data in Figure 7 demonstrate that the current value increases in proportion to the cubic root of u, which is in accordance with the results predicted by hydrodynamic voltammetry with channel electrodes under convective diffusion rate-limiting conditions.26,27 CONCLUSION We fabricated the polymer microchannel chips integrated with dual-band microelectrodes on the basis of photolithography and an imprinting method. Linear sweep voltammetry and generationcollection mode experiments were explored to elucidate characteristic responses of the electrodes in the microchannel chip. The voltammogram obtained by the channel chip was different from that by the electrode substrate without a channel plate, owing to a restricted space around the electrode in the microchannel chip. Reflecting such characteristics, the collection efficiency observed in the microchannel chip was almost two times larger than that in the absence of a channel substrate. Furthermore, the solution flow-rate dependence of η in the channel chip demonstrated that the collection efficiency depended highly on the flow rate, and η (26) Matsuda, H. J. Electroanal. Chem. 1968, 16, 153. (27) Compton, R. G.; Fisher, A. C.; Wellington, R. G. J. Phys. Chem. 1993, 97, 10410-10415.
Figure 5. Cyclic voltammograms (ν ) 20 mV/s) of FeCp-OH observed at G (solid curve) and C (broken curve) in the microchannel (a). A potential sweep rate (ν) dependence of the collection efficiency (b). The closed and open circles are the data obtained by the experiments and the numerical simulation, respectively. The linear flow velocity (u) was set at 8.3 mm/s throughout the experiments.
Figure 6. Time profiles of the currents observed at G and C in the channel chip at u ) 1200 (a) and 120 µm/s (b).
Figure 7. (a) Flow rate dependencies of the steady-state currents observed at G and C. The closed and open circles exhibit the data at G and C, respectively. The solid line represents the predicted values when all FeCp-OH provided to G is oxidized completely at the electrode. (b) The flow rate dependence of the collection efficiency (closed circles) was calculated from the data in part a. The solid curve represents the results by the numerical simulations.
as high as 90% was realized at an optimum solution flow rate. This is another interesting characteristic of the microelectrodes integrated in a microchannel chip. Polymer microdevices integrated with electrodes could be, thus, very promising as a highly efficient detection device for µ-TAS, as well as an electrochemical sensor. ACKNOWLEDGMENT N.K. acknowledges a Grant-in-Aid from the Ministry of Education, Science, Sports, and Culture, Japan (nos. 13853004 and 14050001) for support in part of the research. K.U. also acknowledges J.S.P.S. for the fellowship. The authors also thank Prof. H.
Misawa and Dr. S. Matsuo at The University of Tokushima for kind help in the fabrication of a silicon template. SUPPORTING INFORMATION AVAILABLE Scanning electron microscope (SEM) images of the crosssectional views of the silicon template and the polymer channel chip. This material is available free of charge via the Internet at http://pubs.acs.org. Received for review December 25, 2002. Accepted February 28, 2003. AC0264675 Analytical Chemistry, Vol. 75, No. 9, May 1, 2003
2091
