Anal. Chem. 2003, 75, 350-354
Chemicofunctional Membrane for Integrated Chemical Processes on a Microchip Hideaki Hisamoto,† Yuki Shimizu,† Kenji Uchiyama,‡ Manabu Tokeshi,‡ Yoshikuni Kikutani,‡ Akihide Hibara,† and Takehiko Kitamori*,†,‡
Department of Applied Chemistry, Graduate School of Engineering, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8656, Japan, and Integrated Chemistry Project, Kanagawa Academy of Science and Technology, 3-2-1 Sakado, Takatsu-ku, Kawasaki, Kanagawa 213-0012, Japan
Here we report a design and synthesis of a chemically functional polymer membrane by an interfacial polycondensation reaction and multilayer flow inside a microchannel. Single and parallel dual-membrane structures are successfully prepared by using organic/aqueous twolayer flow and organic/aqueous/organic three-layer flow inside the microchannel followed by an interfacial polycondensation reaction. By using the inner-channel membrane, permeation of ammonia species through the innerchannel membrane is successfully achieved. Furthermore, horseradish peroxidase is immobilized on one side of the membrane surface to integrate the chemical transform function onto the inner-channel membrane. Here substrate permeation through the membrane and subsequent chemical transformation at the membrane surface are realized. The polymer membrane prepared inside the microchannel has an important role in ensuring stable contact of different phases such as gas/liquid or liquid/ liquid and the permeation of chemical species through the membrane. Furthermore, membrane surface modification chemistry allows chemical transformation of permeated chemical species. These methods are expected to lead to development of complicated and sophisticated chemical systems involving membrane permeation and chemical reactions. Microchip technology or microreaction technology using microfabricated devices has been the focus of much attention in recent years.1 Various kinds of effective chemical processes, such as solution mixing or electrophoretic separation, have been integrated onto microchips in order to realize the micro total analysis system (µ-TAS) or microreactors.2-4 We have been focusing on the integration of various kinds of microunit operations (MUOs), and we successfully developed * Corresponding author. E-mail:
[email protected]. † The University of Tokyo. ‡ Kanagawa Academy of Science and Technology. (1) Freemantle, M. Chem. Eng. News 1999, (February 22), 27-36. (2) Proceedings of the µTAS’2001 Symposium 2001; Ramsey, J. M., van den Berg, A., Eds.; Kluwer Academic Publishers: Dordrecht, The Netherlands, 2001. (3) Proceedings of the Third International Conference on Microreaction Technology; Ehrfeld, W., Ed.; Springer: Berlin, 1999. (4) Ehrfeld, W.; Hessel, V.; Lowe, H. Microreactors: New Technology for Modern Chemistry; Wiley-VHC: Weiheim, 2000.
350 Analytical Chemistry, Vol. 75, No. 2, January 15, 2003
continuous-flow chemical processing (CFCP).5 During its development, we observed that the formation of a stable organic/aqueous interface along the longitudinal direction of the microchannel played an important role for integration of various kinds of MUOs.5-17 Although exploiting multiphase flow using the same solvents (aqueous/aqueous or organic/organic) for analytical or fabrication purposes has been the focus of much attention in this research area,18-20 flow using different solvents (organic/aqueous) also provides attractive applications. By using aqueous/organic multilayer flow, we successfully demonstrated highly efficient molecular transport between two phases for analytical and synthetic applications, which exploited the characteristics provided by a liquid microspace such as short molecular diffusion distance and large specific interfacial area.5-17 We briefly reported on the preparation of a polyamide membrane at the organic/aqueous interface along the longitudinal direction, which was accomplished by a well-known interfacial polymerization reaction and linker group modification of the inner (5) Tokeshi, M.; Minagawa, T.; Uchiyama, K.; Hibara, A.; Sato, K.; Hisamoto, H.; Kitamori, T. Anal. Chem. 2002, 74, 1565-1571. (6) Tokeshi, M.; Minagawa, T.; Kitamori, T. Anal. Chem. 2000, 72, 17111714. (7) Sato, K.; Tokeshi, M.; Sawada, T.; Kitamori, T. Anal. Sci. 2000, 16, 455456. (8) Tokeshi, M.; Minagawa, T.; Kitamori, T. J. Chromatogr.. A 2000, 894, 1923. (9) Minagawa, T.; Tokeshi, M.; Kitamori, T. Lab Chip 2001, 1, 72-75. (10) Hisamoto, H.; Horiuchi, T.; Tokeshi, M.; Hibara, A.; Kitamori, T. Anal. Chem. 2001, 73, 1382-1386. (11) Hisamoto, H.; Horiuchi, T.; Uchiyama, K.; Tokeshi, M.; Hibara, A.; Kitamori, T. Anal. Chem. 2001, 73, 5551-5556. (12) Surmeian, M.; Hibara, A.; Slyadnev, M. N.; Uchiyama, K.; Hisamoto, H.; Kitamori, T. Anal. Lett. 2001, 34, 1421-1429. (13) Hibara, A.; Tokeshi, M.; Uchiyama, K.; Hisamoto, H.; Kitamori, T. Anal. Sci. 2001, 17, 89-93. (14) Hisamoto, H.; Saito, T.; Tokeshi, M.; Hibara, A.; Kitamori, T. Chem. Commun. 2001, 2662-2663. (15) Hibara, A.; Nonaka, M.; Hisamoto, H.; Uchiyama, K.; Kikutani, Y.; Tokeshi, M.; Kitamori, T. Anal. Chem. 2002, 74, 1724-1728. (16) Surmeian, M.; Slyadnev, M. N.; Hisamoto, H.; Hibara, A.; Uchiyama, K.; Kitamori, T. Anal. Chem. 2002, 74, 2014-2020. (17) Shimizu, Y.; Hisamoto, H.; Hibara, A.; Tokeshi, M.; Kitamori, T. Digest of Papers, 2001 International Microprocesses and Nanotechnology Conference, Japan Society of Applied Physics: Tokyo, Japan, 2001; pp 20-21. (18) Kamholz, A. E.; Weigl, B. H.; Finlayson, B. A.; Yager, P. Anal. Chem. 1999, 71, 5340-5347. (19) Kenis, P. J. A.; Ismagilov, R. F.; Takayama, S.; Whitesides, G. M.; Li, S.; White, H. S. Acc. Chem. Res. 2000, 33, 841-847. (20) Kenis, P. J. A.; Ismagilov, R. F.; Whitesides, G. M. Science 1999, 285, 8385. 10.1021/ac025794+ CCC: $25.00
© 2003 American Chemical Society Published on Web 12/13/2002
transformation of permeated species by membrane surface modification inside the microchannel.
Figure 1. General idea of polymer membrane formation under organic/aqueous two-phase flow in an X-shaped microchannel.
microchannel surface (Figure 1).17 Very recently, Zhao et al.21 also reported a similar membrane formation inside a millimeter-sized channel by countercurrent flow, as an application of immiscible liquid control by inner-surface modification. Peterson et al. took a different approach to preparing a polymer monolith structure by using a similar surface modification.22 In our opinion, active addition of chemical functions to the inner-channel membrane is one promising way to open up new directions for microchip technologies in chemistry. During the past decade, photolithography technologies have provided excellent, complicated, structures for using physical functions of a microspace such as damming bead material or increasing the surface-to-volume ratio.23,24 In contrast, the synthetic approach proposed in this work provides organic membrane structure allowing chemical functions such as a separation of chemical species based on chemical properties of the membrane itself and an active addition of chemical functions by fully established surface modification chemistry. In this paper, we present a full account of the membrane preparation, and integration of chemical functions of the membrane such as permeation of chemical species, and chemical (21) Zhao, B.; Viernes, N. O. L.; Moore, J. S.; Beebe, D. J. J. Am. Chem. Soc. 2002, 124, 5284-5285. (22) Peterson, D. S.; Rohr, T.; Svec, F.; Frechet, J. M. J., Anal. Chem. 2002, 74, 4081-4088. (23) Andersson, H.; van der Wijngaart, W.; Stemme, G. Electrophoresis 2001, 22, 249-25. (24) He, B.; Tait, N.; Regnier, F. Anal. Chem. 1998, 70, 3790-3797.
EXPERIMENTAL SECTION Reagents. Reagents of the highest grade commercially available were used for preparation of the aqueous test electrolytes. Adipoyl chloride, hexamethylenediamine, and glutaraldehyde were purchased from Tokyo Chemical Industry Co., Ltd. (Tokyo, Japan). Biochemistry grade 4-aminoantipyrine (4-AAP) and N-ethylN-(2-hydroxy-3-sulfopropyl)-m-toluidine (TOOS) were purchased from Dojindo Laboratories (Kumamoto, Japan). Peroxidase from horseradish (HRP, EC 1.11.1.7, specific activity 200 units/mg) and 35% hydrogen peroxide solution (electronic industry grade) were obtained from Wako Pure Chemical Industries (Osaka, Japan). All reagents were used without further purification. Distilled and deionized water used had resistivity values of more than 1.7 × 107 Ω cm at 25 °C. Microfabrication of the Glass Microchip. The microfabrication procedure has been described elsewhere.13 Briefly, the microchannel was fabricated on a Pyrex glass plate (30 mm × 70 mm) by the photolithographic wet etching method. A cover plate with holes (500-µm diameter) was bonded on the etched plate by thermal fusion bonding to complete the glass microchip fabrication. The glass microchips had a channel pattern of 240-µm width and 115-µm depth or 210-µm width and 40-µm depth. The connection between the fabricated chip and a syringe via capillary tubing has also been described in the literature.11 Interfacial Polycondensation Reaction. Before carrying out the interfacial polycondensation reaction, the entire inside surface of the microchannel was chemically modified by (3-aminopropyl)triethoxysilane as a linker layer. Surface modification was done by introducing an ethanol solution of (3-aminopropyl)triethoxysilane (1 wt %) into the microchannel under a constant-flow condition for 17 h. Volume flow rate was fixed at 1 µL/min. After surface treatment, pure ethanol was introduced to remove unreacted species. After the surface modification procedure, two types of nylon membrane structures were prepared inside the microchannel. The first was a single membrane structure formed by using organic/aqueous two-layer flow, and the second was a parallel dual-membrane structure formed by using organic/aqueous/ organic three-layer flow. For each type, adipoyl chloride in 1,2dichloroethane (0.01 M) and hexamethylenediamine in 0.1 M NaOH solution (0.01 M) were used as organic and aqueous solutions. Volume flow rates for preparing the single and parallel dual-membrane structures were 20 µL/min (org)/20 µL/min (aq) and 10 µL/min (org)/20 µL/min (aq)/10 µL/min (org), respectively. Preparation of the Microchip Sample for Taking CrossSectional SEM Photographs. Scanning electron microscope (SEM) photographs were taken in order to observe a crosssectional view of the inner-channel membrane. Since the nylon membrane prepared inside the microchannel was elastic, the membrane-prepared microchip was cut according to the following procedure. The cutting position was marked by a straight line using a dicer before forming the membrane. After the interfacial polymerization, pure water was introduced into both sides of the microchannels separated by the membrane, and all the liquid inlet holes were sealed with adhesive tape. Then the microchip was Analytical Chemistry, Vol. 75, No. 2, January 15, 2003
351
slowly immersed into liquid nitrogen to freeze the water in it. The chip was removed and immediately cut along the marked line manually. Chemical Modification of the Membrane Surface by an Enzyme. Surface modification of the nylon membrane by an enzyme was done using a modified procedure from a previous study.25 One side of the microchannel separated by the membrane was treated with 3.5% hydrochloric acid for 1 h under continuous flow (5 µL/min), to remove residual reagents at the channel wall and activate the amino group at the membrane surface. After this, the microchannel was washed with pure water for 30 min under continuous flow (5 µL/min). In the next step, 2.5% glutaraldehyde solution was introduced into the channel to form a Schiff base on the membrane surface (5 µL/min, 1 h), followed by washing with pure water (5 µL/min, 30 min). Horseradish peroxidase solution (0.5 mg/mL, 0.1 M phosphate buffer pH 7.4) was introduced to immobilize the enzyme on the surface by Schiff base formation (5 µL/min, 1 h). After washing with pure water (5 µL/min, 30 min), 0.1% sodium borohydride solution was introduced to reduce the Schiff base (5 µL/min, 5 min). Washing with pure water to remove unreacted sodium borohydride completed the enzyme immobilization on the nylon membrane surface (5 µL/min, 30 min). Permeation Studies. All the permeation experiments were performed by using a single membrane prepared inside the microchannel under a stopped-flow condition. (1) Permeation of Ammonia Species. Permeation of ammonia species through the inner-channel membrane was detected as a color change of phenolphthalein solution by using the thermal lens microscope (TLM) developed by our group. TLM allows ultrasensitive detection of nonfluorescent molecules with high spatial resolution. The TLM principle has been described in detail elsewhere.26-28 One side of the channel separated by the membrane was filled with a neutral, aqueous phenolphthalein solution (10-4 M), while on the other side of the channel, 10-1 M sodium hydroxide solution and 10-2 M ammonium chloride solution were introduced to form dissolved ammonia gas in the channel. The solutions on both sides of the membrane were introduced under continuous flow (5 µL/min). When the measurements of membrane permeation were started, the syringe pumps were stopped and the thermal lens microscope detection was begun at the center of the microchannel (∼60 µm from the channel wall and membrane). Then the time course of the TLM signal was recorded. During the measurement, a high-pH solution was used. It is well known that the silanized surface is not stable in highly acidic or basic solutions. Therefore, during the measurement shown in Figure 4, solution inside the channel was fully washed with pure water before each measurement. (2) Substrate Permeation and Subsequent Enzyme Reaction at the Membrane Surface. Substrate permeation through the inner-channel membrane and the subsequent enzyme reaction were also detected by TLM. In this case, horseradish peroxidase (25) Ishimori, Y.; Karube, I.; Suzuki, S. Eur. J. Appl. Microbiol. Biotech. 1981, 13, 19-23. (26) Uchiyama, K.; Hibara, A.; Kimura, H.; Sawada, T.; Kitamori, T. Jpn. J. Appl. Phys. 2000, 39, 5316-5322. (27) Tokeshi, M.; Uchida, M.; Uchiyama, K.; Sawada, T.; Kitamori, T J. Luminesc. 1999, 83-84, 261-264. (28) Tokeshi, M.; Uchida, M.; Hibara, A.; Sawada, T.; Kitamori, T Anal. Chem. 2001, 73, 2112-2116.
352 Analytical Chemistry, Vol. 75, No. 2, January 15, 2003
Figure 2. Channel patterns and top and cross-sectional views of the nylon membrane prepared inside the microchannel. (a) Single membrane formed under organic/aqueous two-layer flow: (b) parallel dual membranes formed under organic/aqueous/organic three-layer flow.
was used as a model enzyme to be immobilized on one side of the membrane, and 4-AAP, TOOS, and hydrogen peroxide were used as substrates, respectively. The enzyme-immobilized side of the microchannel was filled with 4-AAP and TOOS (10-3 M, in phosphate buffer each), and the other side was filled with hydrogen peroxide (10-4 M, in phosphate buffer). Hydrogen peroxide can permeate through the membrane; thus, detection of enzyme reaction product at the enzyme-immobilized side of the channel allowed detection of substrate permeation and the subsequent enzyme reaction. To start the measurements, flow and stop of solutions were done in the same manner as the previous experiment. The detection point was also the same as described there. To carry out the control experiments, introduction of substrates without hydrogen peroxide (control 1) and introduction of all the substrates but no membrane-immobilized enzyme (control 2) were also performed. RESULTS AND DISCUSSION Preparation of Nylon Membrane inside the Microchannel by Interfacial Polymerization and Multilayer Flow. We have been focusing on the organic/aqueous interface formed inside the microchannel to realize the various types of analytical or synthetic applications. On the other hand, interfacial polymerization is a well-known method for synthesizing a polyamide membrane at the organic/aqueous interface. Thus, combination of these technologies is expected to enable membrane formation inside the microchannel. Here we investigated the nylon membrane formation inside the microchannel under organic/aqueous two-layer flow or organic/aqueous/organic three-layer flow. Figure 2 shows the channel patterns and top and crosssectional views of the nylon membranes prepared inside the microchannels. After introducing organic and aqueous phases containing membrane monomers, a membrane-type 3D structure immediately formed at the liquid/liquid interface position. As clearly shown in these photographs, the nylon membrane was successfully formed in the microchannel by interfacial polycondensation. The membrane thickness was determined to be ∼10 µm for both types. In the case of the 6,6-nylon membrane
Figure 3. Sequential photographs showing an air bubble passing through one side of microchannel separated by the nylon membrane 3D structure.
preparation by interfacial polycondensation, the prepared membrane avoids interdiffusion of starting matrixes across the liquid/ liquid interface. Therefore, the thickness of the membrane at the liquid/liquid interface can be determined by the concentration and reactivities of the starting matrixes. In our case, the same organic and aqueous solutions containing starting matrixes were used for preparing single or parallel dual membranes. Although linear flow velocities for these two conditions were not exactly the same, monomer concentration governed membrane thickness. From the cross-sectional photographs, no significant thickness distribution of the prepared membrane from top to bottom was observed. This fact also supports the explanation given above. Figure 3 shows sequential photographs when an air bubble was introduced into one side of the microchannel separated by the nylon membrane. From these photographs, considerable leaking of water into the air phase (air bubble) was not observed during a few-hour experiment, indicating that the membrane successfully worked as a diaphragm separating the single channel into gas and liquid channels. Here, polymer membrane 3D structures were successfully prepared by interfacial polycondensation reaction and immiscible multilayer flow formation inside the microchannel. An innerchannel membrane can be applied for integrating permeationbased chemical processes into the microchannel. Some applications are described below. Permeation Study 1: Permeation of Ammonia Species. To confirm the molecular transport function of the membrane, we carried out the simple permeation experiment (Figure 4a) using the nylon membrane prepared inside the microchannel. In this case, the generated ammonia species (NH3 gas, NH4OH) is expected to permeate the nylon membrane and change the pH of the phenolphthalein phase. Therefore, permeation of ammonia species through the 6,6-nylon membrane can be detected as the color change of phenolphthalein phase by using TLM. Figure 4b shows the time course of the TLM signal generated at the phenolphthalein phase when ammonium chloride and
Figure 4. Ammonia species permeation experiment. (a) Schematic illustration of experimental system. (b) Time course of the TLM signal just after stopping the syringe pump (for details, see the Experimental Section).
sodium hydroxide solutions were introduced into the other side of the channel separated by the membrane. To confirm leaking of sodium hydroxide solution to the phenolphthalein phase through the membrane, introduction of sodium hydroxide solution without ammonium chloride was also investigated. In both cases, TLM measurements were done at the phenolphthalein phase under the stopped-flow condition. As can be seen, the TLM signal at the phenolphthalein phase did not change when sodium hydroxide solution was introduced without ammonium chloride; on the other hand, the signal rapidly increased when ammonia species was generated at the other side of the membrane. Zhao et al.21 reported water permeation through a nylon membrane under a pressure-applied condition. However, in our case, permeation of hydrophilic sodium hydroxide species was not observed. Generally, nylon membranes are a hydrophobic polyamide membrane. Thus, highly hydrophilic sodium hydroxide could not permeate through the membrane under a normalpressure condition. These results clearly suggested that the nylon membrane prepared inside the microchannel successfully worked as a permeation membrane of ammonia species. In Figure 4b, the TLM signal reached steady state at ∼1 min. In our experiment, response time is mainly governed by diffusion of ionic species (NH4+, OH-) in liquid phase rather than that of ammonia species in the nylon membrane, due to the large difference in diffusion coefficients. It is well known that the diffusion time of small chemical species in liquid is much slower (generally 10-5 cm2/s in diffusion coefficient) compared to that of gas species (generally 10-1 cm2/s in diffusion coefficient.). Based on Fick’s law of diffusion, the diffusion time of small molecules for a 100-µm-sized channel is approximately several tens of seconds. In our case, we prepared a highly permeable nylon membrane, and the color change (deprotonation) of phenolphthalein is considered as a diffusion-controlled reaction, so that our Analytical Chemistry, Vol. 75, No. 2, January 15, 2003
353
TLM signal shown in Figure 5b shows that the hydrogen peroxide permeation and subsequent enzyme reaction at the membrane surface were successfully completed. Since one of the substrates, TOOS, is ionic for its sulfonate group, the produced dye molecule should also have an ionic group.29 As mentioned in the ammonia species permeation experiment, the nylon membrane prepared here did not easily transport hydrophilic species under a normal-pressure condition. When the detection point of TLM was shifted to the other side separated by the membrane ((B) in Figure 5b), no signal was observed, supporting the finding that the colored product did not permeate through the membrane. These results suggested that surface modification chemistry was successfully implemented, and complicated and sophisticated chemical processes involving substrate permeation and subsequent chemical reaction could be realized.
Figure 5. Substrate permeation and subsequent enzyme reaction experiment. (a) Schematic illustration of experiment system. (b) Time course of the TLM signal just after stopping the syringe pump (for details, see the Experimental Section). TLM detection points A and B indicate detection points in the microchannel separated by the innerchannel membrane, shown in (a). (A) Enzyme immobilized side. (B) Hydrogen peroxide supplying side. Controls 1 and 2 were obtained at detection point A.
experimental result corresponded well with diffusion theory. For membrane permeation-based chemical processes, a micrometersized channel is quite advantageous for obtaining rapid equilibrium. Permeation Study 2: Substrate Permeation and Subsequent Enzyme Reaction at the Membrane Surface. Surface modification of the inner-channel membrane by catalyst molecules is one promising way to integrate complicated and sophisticated chemical processes involving substrate permeation and chemical reaction. Here we employed horseradish peroxidase as a biological catalyst molecule and hydrogen peroxide as a nylon membranepermeable substrate. Horseradish peroxidase catalyses organic dye formation by using hydrogen peroxide, TOOS, and 4-AAP as substrates.29 Figure 5a illustrates the experiment. Horseradish peroxidase was immobilized on one side of the membrane surface, and hydrogen peroxide was supplied from the other side. The enzyme-immobilized side of the channel was filled with TOOS and 4-AAP. When the flow stopped, the TLM signal increased rapidly and reached a steady state at ∼250 s ((A) in Figure 5b). In this case, no signal was obtained by introduction of TOOS and 4-AAP without hydrogen peroxide (control 1 in Figure 5b). Since the TLM signal obtained when all the substrates were mixed inside the bare channel (without membrane-immobilized enzyme) was rather slow (control 2 in Figure 5b), the rapid increase of (29) Tamaoku, K.; Murao, Y.; Akiura, K.; Ohkura, Y. Anal. Chim. Acta 1982, 136, 121-127.
354
Analytical Chemistry, Vol. 75, No. 2, January 15, 2003
CONCLUSIONS We have demonstrated the preparation of nylon membrane structures by an interfacial polycondensation reaction and multilayer flow inside a microchannel. Multilayer flow such as organic/ aqueous two-layer flow or organic/aqueous/organic three-layer flow allowed single or parallel dual-membrane formations. The prepared membrane provided stable liquid/gas or liquid/ liquid contact, and the membrane also served as a permeation membrane of chemical species inside the microchannel. Surface modification by an enzyme allowed active addition of a molecular transform function to the inner-channel membrane. By using the enzyme-modified membrane, substrate permeation and subsequent molecular transformation were realized. The proposed method of preparing the polymer membrane can be applied for other starting matrixes possessing two reactive sites. Surface modification chemistry may also be applicable. Furthermore, parallel dual membranes are expected to be applied for the multiple-analyte determination based on different enzyme modification to each membrane or the efficient multistep synthesis required for general organic chemistry. These applications are currently under investigation. Thus, the methods proposed here are expected to allow integration of complicated and sophisticated chemical processes involving molecular transport between liquid/ liquid or liquid/gas phases and chemical transformation of transported species in a liquid microspace. We expect the chemicofunctional membrane proposed here will lead to a novel 3D structure construction methodology in future microchip technology. ACKNOWLEDGMENT This work was partially supported by Grant-in-Aid for Creative Scientific Research (13GS0024), from the Ministry of Education, Science, and Culture, Japan. Received for review May 23, 2002. Accepted November 12, 2002. AC025794+
