On-Chip Microfluidic Transport and Mixing Using Electrowetting and

Sep 29, 2005 - crofluidic transport, mixing, and sensing on a single chip. The operation ... in a network of flow channels by switching on necessary e...
0 downloads 0 Views 523KB Size
Download PDF
Anal. Chem. 2005, 77, 6857-6863

On-Chip Microfluidic Transport and Mixing Using Electrowetting and Incorporation of Sensing Functions Wataru Satoh, Hiroki Hosono, and Hiroaki Suzuki*

Graduate School of Pure and Applied Sciences, University of Tsukuba, 1-1-1 Tennodai, Tsukuba, Ibaraki 305-8573, Japan

An integrated system was developed that performs microfluidic transport, mixing, and sensing on a single chip. The operation principle for the microfluidic transport was based on electrowetting. A solution to be transported was confined in a space between a row of gold working electrodes and a protruding poly(dimethylsiloxane) (PDMS) structure. When a negative potential was applied to one of the gold working electrodes, it became hydrophilic, and the solution was transported through the flow channel. The solution could be transported in any desired direction in a network of flow channels by switching on necessary electrodes one by one. Furthermore, two solutions transported through two flow channels could be mixed using a mixing electrode based on the same principle. To demonstrate the applicability of a lab-on-a-chip, an air gap ammonia electrode was integrated by taking advantage of the open structure of the flow channel. Gaseous ammonia that was produced after pH adjustment and diffused through an air gap caused a pH change in the electrolyte layer, which was measured with an iridium oxide pH indicator electrode. The 90% response time was less than 1 min for the millimolar order of ammonia. The calibration curve was linear down to 10 µM. The ammonia-sensing system was also applied to construct biosensing systems for urea and creatinine. A linear relationship was observed between the potential and the logarithm of the concentration of the analytes down to 50 µM for both urea and creatinine. The developed microfluidic system can be a basic building block for future systems. The recent progress of micro total analysis systems (µTAS) or labs-on-a-chip has already resulted in a high level of integration of microfluidic and sensing functionalities,1-3 and more sophisticated and automated devices are expected. For the development of future generation µTAS, some critical problems still need to be solved. One of them is the development of an integrated on* To whom correspondence should be addressed. E-mail: hsuzuki@ ims.tsukuba.ac.jp. Tel: +81-29-853-5598. Fax: +81-29-855-7440. (1) Burns, M. A.; Johnson, B. N.; Brahmasandra, S. N.; Handique, K.; Webster, J. R.; Krishnan, M.; Sammarco, T. S.; Man, P. M.; Jones, D.; Heldsinger, D.; Mastrangelo, C. H.; Burke, D. A. Science 1998, 282, 484-487. (2) Duffy, D. C.; Gillis, H. L.; Lin, J.; Norman F. Sheppard, J.; Kellogg, G. J. Anal. Chem. 1999, 71, 4669-4678. (3) Liu, J.; Hansen, C.; Quake, S. R. Anal. Chem. 2003, 75, 4718-4723. 10.1021/ac050821s CCC: $30.25 Published on Web 09/29/2005

© 2005 American Chemical Society

chip microfluidic transport system.4 Although many micropumps have been proposed over the last two decades,5,6 the development of highly integrated systems with these pumps has not progressed as expected. As a result, many researchers have been forced to rely on external conventional pumps to inject, transport, and mix solutions in micro flow channels. Although the integration of an on-chip microfluidic system may not be required, depending on the case, it will certainly play a critical role in a group of devices, such as those designed for micro health care systems. As a very simple system that can be used to perform on-chip microfluidic transport, systems based on the electroosmotic principle have been used for the separation of molecules, such as DNA, and for the transport of solutions.7 Solution can be mobilized even in a complicated microfluidic network by applying a voltage through an intended path. However, this technique usually requires a very high driving voltage, which restricts the use of these devices outside a laboratory. In addition, the power consumption or Joule heating is significant, and the presence of a bubble can have a serious effect, which prevents the injection of solutions into empty flow channels or reaction chambers, as often required in clinical analyses. To mobilize a solution freely with a low driving voltage and power consumption, the use of interfacial tension has recently attracted attention.8-15 The influence of interfacial tension is negligible in conventional systems of macroscopic scales. However, it becomes dominant as the dimensions of the flow system are reduced. In light of the current trend toward the reduction of (4) Walt, D. R. Science 2005, 308, 217-219. (5) Gravesen, P.; Branebjerg, J.; Jensen, O. S. J. Micromech. Microeng. 1993, 3, 163-182. (6) Kovacs, G. T. A. In Micromachined transducers sourcebook; WCB McGrawHill: Boston, 1998; pp 823-855. (7) Manz, A.; Effenhauser, C. S.; Burggraf, N.; Harrison, D. J.; Seiler, K.; Fluri, K. J. Micromech. Microeng. 1994, 4, 257-265. (8) Pollack, M. G.; Fair, R. B.; Shenderov, A. D. Appl. Phys. Lett. 2000, 77, 1725-1726. (9) Geng, X.; Yuan, H.; Oguz, H. N.; Prosperetti, A. J. Micromech. Microeng. 2001, 11, 270-276. (10) Yun, K. S.; Cho, I. J.; Bu, J. U.; Kim, C. J.; Yoon, E. J. Microelectromech. Syst. 2002, 11, 454-461. (11) Moon, H.; Cho, S. K.; Garrell, R. L.; Kim, C.-J. J. Appl. Phys. 2002, 92, 4080-4087. (12) Ren, H.; Fair, R. B.; Pollack, M. G.; Shaughnessy, E. J. Sens. Actuators, B 2002, 87, 201-206. (13) Lee, J.; Moon, H.; Fowler, J.; Schoellhammer, T.; Kim, C.-J. Sens. Actuators, A 2002, 95, 259-268. (14) Yoon, J. Y.; Garrell, R. L. Anal. Chem. 2003, 75, 5097-5102. (15) Satoh, W.; Loughran, M.; Suzuki, H. J. Appl. Phys. 2004, 96, 835-841.

Analytical Chemistry, Vol. 77, No. 21, November 1, 2005 6857

sample and reagent volumes, the operation principle is very attractive and promising. In applying this principle to microfluidic transport, a method often used is to mobilize a droplet on an electrode array.8,11-14 However, such systems are incompatible with developed microanalysis systems that require the transport of continuous liquid. To cope with this problem, we have developed an integrated on-chip microfluidic system that transports a solution in the form of a continuous liquid column.15 Using this principle and the unique structure of these flow channels, a solution can be transported in the flow channels and introduced into selected flow channels without using any valves. The principle can also be used to mix two solutions. The microfluidic system developed in this study can be used in many bio/chemical analyses that require the handling of solutions. In this study, we demonstrated its applicability by integrating a Severinghaus-type ammonia electrode in a microfluidic system. To achieve a fast response, the separation of the electrolyte solution from the sample solution with an air gap is a very effective strategy.16 Although the construction of the air gap structure in a discrete miniaturized sensor has been very difficult, the open structure of the flow channel used for microfluidic transport can be used advantageously to realize it. Furthermore, the reduction of the length of the air gap is easily achieved using planar two-dimensional patterns on a single substrate. We demonstrate here that our system is effective for on-chip microfluidic transport, mixing of solutions, and bio/chemical sensing. EXPERIMENTAL SECTION Reagents. The materials used for fabrication and performance characterization were obtained from the following commercial sources: negative photoresist (OMR-83) from Tokyo Ohka Kogyo (Kawasaki, Japan), thick-film photoresist (SU-8) from MicroChem (Newton, MA), precursor solution of PDMS (KE-1300T) from Shin-Etsu Chemical (Tokyo, Japan), fluorescein, urease (EC 3.5.1.5, from jack bean, 116 units/mg) and horseradish peroxidase (HRP; EC 1.11.1.7, 100 units/mg) from Wako Pure Chemical Industries (Osaka, Japan), creatine deiminase (EC 3.5.4.21, from Bacillus sp., 12.7 units/mg) from Toyobo (Osaka, Japan), the precursor solution of PVA-SbQ from Toyo Gosei (Chiba, Japan), and Amplex Red (10-acetyl-3.7-dihydroxyphenoxazine) from AnaSpec (San Jose, CA). L-Glutamate oxidase was a gift from Yamasa (Chiba, Japan). Standard solutions were prepared with distilled deionized water. Structure and Fabrication of the System. The basic structure of the flow channel is shown in Figure 1. The system consists of a glass substrate with thin-film electrodes and a PDMS substrate that forms flow channels with the electrodes. An essential element to generate the driving force is an array of elongated gold working electrodes. In our preliminary study, we found that the strip of the gold working electrode and a simple flat PDMS substrate could not control the flow because the solution spread as the solution went forward. Therefore, we used a protruding structure on the PDMS substrate formed along the patterns of the gold electrodes. The transported solution was confined in a space between the protruding structure and the gold working electrode by interfacial tension (Figure 1). To confine the solution in the flow channel (16) Ross, J. W.; Riseman, J. H.; Krueger, J. A. Pure Appl. Chem. 1973, 36, 473487.

6858 Analytical Chemistry, Vol. 77, No. 21, November 1, 2005

Figure 1. Microfluidic transport of a solution in a flow channel based on electrowetting. (a)-(c) show the cross section along the flow channel, which shows the basic structure and arrangement in the flow channel and the procedure for mixing. (d) shows the cross section viewed from the right of (a)-(c). W.E., working electrode; R.E., reference electrode; A.E., auxiliary electrode.

more effectively, a hydrophobic negative photoresist layer was also formed in areas other than the flow channel. A Ag/AgCl reference electrode and a platinum auxiliary electrode were formed in the inlet to complete a three-electrode system. A unique feature of this microfluidic system is that the direction of the flow can be controlled without using any valves. To achieve this, the elongated working electrode was separated into fractions of working electrodes with an intervening glass gap area. The driving potential was applied independently to the respective working electrodes. To demonstrate the flow control in a relatively complicated network of flow channels, a tree-shaped network and a star-shaped network of flow channels were formed (see Figure 3). Our interest was not limited to the flow control. To demonstrate the applicability of this device to a lab-on-a-chip, three flow channels for the sample solution, a NaOH solution to change the solution pH, and an electrolyte solution for the ammonia electrode were formed using the basic structure described above (Figure 2a-c). An additional gold working electrode was formed between the flow channels to mix the former two solutions (Figure 2c). The width of the working electrode for the sample solution was 300 µm, whereas that of the electrode for the NaOH solution and the mixing electrode was 200 µm. The gap between the respective electrodes was 100 µm. The entire area consisting of the three working electrodes formed a mixing area. A flow channel for the ammonia electrode (200 µm wide) was formed in the vicinity of the mixing electrode. An iridium oxide pH indicator electrode and a Ag/AgCl reference electrode were formed as detecting elements

Figure 2. Integration of an air gap ammonia electrode with the microfluidic system. (a) Arrangement of the glass substrate with the electrodes and the PDMS substrate with the protruding structure. (b) Planar layout of the entire system. The leads for the electrodes are omitted. (c) Magnified view of the mixing and sensing areas. The shaded area attaches to the glass substrate, and the areas other than the electrode areas form a recess (see Figure 1). The electrolyte solution wets the area including the reference electrode and the pH indicator electrode to the right edge of the latter. (d)-(f) Cross section of (c) along the direction X-X′ and Y-Y′. The position of the immobilized enzyme is also indicated. (d)-(f) also show the procedure for the mixing of the solutions and the detection of ammonia, urea, or creatinine. W.E., working electrode; R.E., reference electrode; A.E., auxiliary electrode.

(Figure 2c) following the procedure mentioned elsewhere.17 The distance between the edge of the mixing area and the flow channel for the electrolyte solution was 500 or 300 µm. The area worked as the air gap. To achieve highly sensitive ammonia detection, (17) Suzuki, H.; Shiroishi, H.; Sasaki, S.; Karube, I. Anal. Chem. 1999, 71, 50695075.

unwanted dissipation of ammonia should be minimized. Therefore, the flow channels and the mixing area were surrounded by the walls of PDMS (dashed lines in Figure 2c). The separation between the flow channels and the walls was 100 µm. For the construction of the biosensing systems for urea and creatinine, an enzyme (urease or creatinine deiminase) was immobilized on Analytical Chemistry, Vol. 77, No. 21, November 1, 2005

6859

the protruding structure near the end of the flow channel (Figure 2d). To this end, 2 mg of either of the enzymes was dissolved in 20 µL of distilled deionized water and mixed well with 20 µL of the precursor solution of PVA-SbQ. A 2-µL aliquot of the solution was spotted at the end of the PDMS flow channel for the sample solution using a micropipet. After water was evaporated from the solution, solution was cross-linked under UV light, and the immobilized enzyme was rinsed in distilled deionized water. Principle of Microfluidic Transport and Sensing. Electrowetting refers to a phenomenon or a technique that changes the wettability of a metal electrode by controlling its potential.18-20 Our device was constructed by taking advantage of this phenomenon. Figure 1a-c shows the procedure for microfluidic transport. First, the potential is set at -0.1 V (vs on-chip Ag/AgCl). In this state, the surface of the gold electrode is not sufficiently hydrophilic. Therefore, the meniscus of the liquid column remains at the edge of the first working electrode when the solution is filled in the injection port (Figure 1a). The potential of the electrode is then changed to -0.9 V. Following this change, the electrode becomes more hydrophilic,15 and the liquid column is mobilized along the flow channel to the end of the first working electrode by capillary action. When the liquid column reaches the other end of the working electrode, the column exudes through the hydrophilic glass gap area and stops at the edge of the neighboring working electrode (Figure 1b). Again, the column is mobilized along the second working electrode when the driving potential is applied (Figure 1c). By repeating the procedure, the direction of the extending solution can be controlled without using any valves. The same principle is applicable to the mixing of two solutions (Figure 2d-f). First, two solutions are transported to the mixing area (Figure 2e). As in the transport, the solutions exude through the glass gap areas and wet the longitudinal edges of the mixing electrode. When the potential is applied to the mixing electrode, the two solutions wet the mixing area and are mixed (Figure 2f). To operate the integrated sensing system, three solutions (a sample solution, a 50 mM NaOH solution, and the electrolyte solution) were transported through the respective flow channels, and the former two solutions were mixed as mentioned earlier. Gaseous ammonia produced by the accompanying pH change diffused into the air gap and then into the electrolyte solution for the ammonia electrode. This increased the pH of the electrolyte solution, which was measured using the pH indicator electrode. The relationship between the potential and the ammonia concentration becomes Nernstian except for the concentration region near the detection limit. The ammonia-sensing system can also be used for biosensing related to the production of ammonia. Here, the determination of urea and creatinine was tested. When a sample solution containing the analytes is transported and comes in contact with either of the enzymes (urease or creatinine deiminase), ammonium ions are generated enzymatically as follows. urease

urea + 3H2O 98 HCO3- + 2NH4+ + OHcreatinine deiminase

creatinine + 2H2O 98 N-methylhydantoin + NH4+ + OHAfter the enzymatic reaction is allowed to proceed for 15 s, the 6860

Analytical Chemistry, Vol. 77, No. 21, November 1, 2005

produced ammonium ions are converted into ammonia by changing the solution pH to the alkaline side. Again, the amount of produced ammonia is measured using the ammonia electrode. In conventional discrete biosensors, the optimum operational condition for the sensors is determined as a compromise between those for the receptor and the transducer. With the ammonia electrode used as a transducer, the optimum conditions should be met with respect to the production of gaseous ammonia and the enzymatic reaction of urease or creatinine deiminase. Note that the conditions can be optimized separately in our microsystem and larger responses should be obtained compared with the discrete biosensors. Other Procedures. The potential of the working electrode was controlled by two sets of Hokuto-Denko HA-151 potentiostat/ galvanostats. A 1.0 M KCl solution was used as a test solution. Fluorescein (1 mM) was added to the solution when recording the movement of the liquid column along the flow channels. Fluorescence from the liquid column was recorded using a Keyence VB-6000 fluorescence microscope. For the excitation and the detection of fluorescence, Omega Optical XF1067 and XF3081 band-pass filters (450-700 nm, 500-800 nm) were used, respectively. The fluorescence intensity was analyzed using Keyence VHH1A5 software. The electrolyte solution for the ammonia electrode was 0.1 M KCl containing 0.1 M NH4Cl. Urea or creatinine standard solutions were prepared with a 0.1 M KH2PO4-NaOH buffer solution (pH 6.0 and pH 8.0) and a 0.1 M H3BO3-NaOH buffer solution (pH 10.0), all containing 1.0 M KCl. All experiments were conducted at room temperature. RESULTS AND DISCUSSION Controlling the Direction of Flow. To examine the function of the system, the KCl solution was filled in the injection port, and the edge of the first working electrode was wetted. Once the potential was applied to the electrode, the motive force generated as a result of the change in the interfacial tension was sufficient to mobilize the liquid column from one end of the electrode to the other end. The end of the liquid column then exuded out of the first electrode, passed the hydrophilic gap area, and wetted the next electrode. Even in a complicated network of flow channels, the direction of the flow could be controlled by switching on necessary electrodes one by one. To demonstrate this point, a star-shaped network and a tree-shaped network of flow channels were fabricated, and the fluorescein solution was transported there (Figure 3). In both cases, the solution was first filled in the injection port and transported on the first working electrode. After the liquid column arrived at the end of the electrode and exuded to the neighboring electrodes, the working electrode placed in a desired direction was switched on. This caused the extension of the liquid column only on the electrode. By repeating the procedure, the solution could be transported and change its direction as desired. As already examined in our previous study, the flow velocity increased as the potential became more negative and the flow channel was more restricted.15 A limitation with this microfluidic system is that the direction of flow cannot be reversed even if the polarity of the potential is changed. At present, we are (18) Beni, G.; Tenan, M. A. J. Appl. Phys. 1981, 52, 6011-6015. (19) Bockris, J. O. M.; Reddy, A. K. N.; Gamboa-Aldeco, M. Modern Electrochemistry; Kluwer Academic/Plenum Publishers: New York, 2000. (20) Quilliet, C.; Berge, B. Curr. Opin. Colloid Interface Sci. 2001, 6, 34-39.

Figure 4. Change of the fluorescence intensity (in arbitrary unit) after mixing of solutions A and B. The pictures show a magnified view of the mixing area at 30 s, 3 min, and 9 min after mixing.

whether the solutions could be mixed as expected, the change in the intensity of fluorescence based on the following enzymatic reactions was measured.21,22 L-glutamate

L-glutamate

oxidase

+ H2O + O2 98 R-ketoglutarate + NH3 + H2O2 horseradish peroxidase

Amplex red + H2O2 98 resorufin + O2

Figure 3. Transport of the fluorescein solution in the networks of flow channels. (a)-(d) show the transport in the tree-shaped network, and (e)-(h) show the transport in the star-shaped network.

not sure whether the reverse movement can be made possible or not. Hydrophobic modification along with an additional mechanism might make it possible, although the excellent features such as low operating voltage and power consumption might be lost. The mixing of solutions transported through two different flow channels was also successfully completed. In the areas near the end of the flow channels, the solutions exuded through the hydrophilic gap area into the edge of the mixing electrode. If the width of the electrode was appropriate, the solutions were separated on the electrode under the “off” state. When the potential was applied to the mixing electrode, the two solutions wetted the electrode surface and mixed (Figure 2). To determine

Here, a 50 mM Tris-HCl buffer solution (pH 7.4) containing 4 g/L HRP and 1 mg/L L-glutamate oxidase (solution A) and a 50 mM Tris-HCl buffer (pH 7.4) containing 1 mM L-glutamate and 10 mM Amplex Red (solution B) were mixed. Resorufin is highly fluorescent, and the fluorescence intensity is an indicator of L-glutamate concentration. Figure 4 shows the change in the fluorescence intensity. The fluorescence was observed immediately after mixing, and the intensity increased as time elapsed. This demonstrated that the two solutions could be mixed and reacted in the very small space on the chip. Response of the Ammonia Electrode. The original configuration of the Severinghaus-type electrode has a gas-permeable membrane.23,24 However, to achieve a faster response, the air gap structure is effective, which particularly applies to the microfabricated ammonia electrode.25 The planar layout of the components (21) Kusakabe, H.; Midorikawa, Y.; Kuninaka, A.; Yoshino, H. Agric. Biol. Chem. 1983, 47, 179-182. (22) Seong, G. H.; Heo, J.; Crooks, R. M. Anal. Chem. 2003, 75, 3161-3167. (23) Stow, R. W.; Baer, R. F.; Randall, B. F. Arch. Phys. Med. Rehabil. 1957, 38, 646-650. (24) Severinghaus, J. W.; Bradley, A. F. J. Appl. Physiol. 1958, 13, 515-520. (25) Suzuki, H.; Matsugi, Y. Sens. Actuators, B 2004, 98, 101-111.

Analytical Chemistry, Vol. 77, No. 21, November 1, 2005

6861

Figure 5. Calibration plot for the ammonia electrode with 500-µm air gap (0) and 300-µm air gap (O). The lines were obtained by leastsquares fitting of the data points. The inset shows the dependence of the 90% response time on the concentration of ammonia with the respective air gaps. The time “0” was set at the point at which the pH of the solution was changed.

and the open structure of the flow channel facilitated the formation of a narrow air gap because it was easily realized by just placing two flow channels in proximity (Figure 2). The effect of using the air gap was checked by examining the performance of the electrode with two different gap lengths (300 and 500 µm). After a sample solution containing NH4Cl and the NaOH solution were mixed following the procedure mentioned earlier, a potential change was immediately observed in the ammonia electrode. The inset to Figure 5 shows the dependence of the 90% response time on the ammonia concentration. With this kind of electrode, which involves the diffusion of the analyte, the response depends on the analyte concentration.16 The 90% response time to ammonia of mM order was 2 min with a gap of 500 µm and less than 1 min with a gap of 300 µm. The response time increased as the concentration of ammonia decreased. Overall, the response was faster with the narrower air gap. Figure 5 shows the dependence of the potential of the pH indicator electrode on the ammonia concentration. A linear relationship was observed between the potential and the logarithm of the ammonia concentration down to 10 µM. The slope of the linear region was 81.8 mV/decade with a gap of 500 µm but was increased to 85.0 mV/decade with a gap of 300 µm. The so-called super-Nernstian behavior is ascribed to the iridium oxide used for the pH indicator electrode.26 It is evident that the produced ammonia was trapped more effectively in the electrolyte solution and caused a larger response in the ammonia electrode with the smaller air gap. In relation to this, the existence of the PDMS walls around the sensitive area played a critical role to minimize the unwanted dissipation of ammonia. Without the walls, the response was much smaller. The result indicates that the efficiency of the ammonia sensing is improved by further miniaturizing the device, which agrees with the tendency provided by the microfluidic system itself. (26) Hitchman, M. L.; Ramanathan, S. Analyst 1988, 113, 35-39.

6862 Analytical Chemistry, Vol. 77, No. 21, November 1, 2005

Figure 6. Calibration plot for the urea-sensing system. Urea solutions of different pHs were transported to the enzyme-immobilized area, and the NaOH solution was mixed to liberate ammonia. O, pH 6.0; 0, pH 8.0; 4, pH 10.0. The lines were obtained by least-squares fitting of the data points. The inset shows the dependence of the 90% response time on the concentration of urea in the respective solutions. The time “0” was set at the point at which the pH of the solution was changed.

Performance of the Urea- and Creatinine-Sensing Systems. To demonstrate the applicability to biosensing, the ammonia-sensing system was used to construct the urea- and creatinine-sensing systems. When a sample solution containing urea was transported to an enzyme-immobilized site, ammonium ions were generated enzymatically at a neutral pH. The ammonia was then liberated and diffused into the ammonia electrode after the NaOH solution was mixed. Figure 6 shows the dependence of the potential of the pH indicator electrode on the urea concentration. Here, solutions of different pHs were prepared, and changes in the response were examined using one of the solutions. Since the enzymatic reaction proceeds before changing the solution pH, a higher response should be obtained as the pH of the solution approaches a neutral value. The obtained result was exactly as anticipated. The slope was -69.4 mV/decade with a sample solution of pH 10.0, but it was -81.4 mV/decade with a sample solution of pH 6.0, reflecting the higher activity of the enzyme in the latter. The result suggests that the sensitivity can be improved when the enzymatic reaction and the diffusion of ammonia is conducted separately under the respective optimum conditions. The inset to Figure 6 shows the dependence of the 90% response time on the urea concentration. The response time was less than 5 min for urea of millimolar order with a sample solution of pH 8.0. On the other hand, the response time was longer at pH 10.0, reflecting the lower enzyme activity at that pH. Note that the response time depends on the amount of accumulated ammonium ions or the length of the waiting time as well as the structural parameters. Therefore, by balancing the waiting time with the length of the response time, the measurements will take much less time. The ammonia-sensing system was also used for the determination of creatinine. The activity of the used creatinine deiminase is the highest at pH 9.5. Therefore, standard solutions were prepared with a borate buffer solution of pH 10.0. Since the sample

solution was basic from the beginning, the produced gaseous ammonia diffused immediately through the air gap to the ammonia electrode. A linear calibration plot was obtained between the potential of the pH indicator electrode and the logarithm of creatinine concentration down to 50 µM. The slope was -78.2 mV/decade. The obtained results demonstrated the possibility of applying the microfluidic system to various types of bio/chemical sensing. CONCLUSIONS Electrowetting provides a promising methodology to construct highly sophisticated micro analysis systems. Two solutions injected from different injection ports can be transported on a row of working electrodes through flow channels. No valves are necessary. The direction of flow can be changed freely by applying a potential to the corresponding working electrode. Furthermore, two solutions can be mixed in the mixing flow channel based on the same principle. The open flow channel structure is suitable

for the integration of chemical sensors that require the separation of specific gases from a sample solution. The sensing system can also be used to construct a biosensing system by coupling it with an enzymatic reaction. This promising technique has potential in constructing highly sophisticated systems that may be called a chemical integrated circuit. ACKNOWLEDGMENT This study was supported by Grants-in-Aid for Scientific Research in Priority Areas and by the 21st Century COE Program, both of which are under the Ministry of Education, Culture, Sports, Science, and Technology, and Grants-in-Aid for Scientific Research (B), which is under the Japan Society for the Promotion of Science.

Received for review May 11, 2005. Accepted August 19, 2005. AC050821S

Analytical Chemistry, Vol. 77, No. 21, November 1, 2005

6863