Anal. Chem. 2004, 76, 6802-6807
Potentiometric Immunosensor Using Artificial Antibody Based on Molecularly Imprinted Polymers Tatsuya Kitade,* Keisuke Kitamura, Tomoyuki Konishi, Shigehiko Takegami, Takayoshi Okuno, Michie Ishikawa, Manabu Wakabayashi, Kayoko Nishikawa, and Yoko Muramatsu
Kyoto Pharmaceutical University, 5 Nakauchi-cho, Misasagi, Yamashina-ku, Kyoto 607-8414, Japan
A potentiometric artificial immunosensor based on a molecularly imprinted polymer was prepared as a detecting element in micro total analysis systems with the intent of providing easy clinical analysis. As the structure and transducing mechanism of this sensor are very simple, construction of a single microsensor should be quite easy. Multimicrosensor arrays applicable to several kinds of analytes will be attainable by both changing the template molecule to be imprinted and reducing the sensor size. The response characteristics of this sensor were evaluated by measuring the response potential to serotonin, which was used as a model material. The obtained sensor was highly responsive to serotonin in water but not to tryptamine, acetaminophen, or procainamide. This phenomenon confirms that the sensor recognizes serotonin and that it functions as a specific artificial immunosensor. Quick measurement is possible because the response time, defined as the time required to achieve 95% of the magnitude of the equilibrated signal, correspond to ∼12 s. The sensor’s determination and detection limits were found to be 1 µmol/L and 100 pmol/L, respectively. These results suggest that our strategy can be applied to construction of a potentiometric artificial immunosensor. Some kinds of immunosensors are currently under investigation in numerous applications due to their high specificity for organic substances.1-4 However, these immunosensors have some defects; i.e., they have instable performance, they deteriorate easily, they are not usable in harsh environments because of their poor chemical and physical stability, their manufacture and handling are complicated, and they are expensive due to complicated biosynthesis. These defect points are caused primarily by biomolecules of an antibody or antigen being used for a recognition element. On the other hand, molecularly imprinted polymers (MIPs) have been widely studied for numerous applications. The binding sites of the MIPs have affinities and selectivities approaching those * To whom correspondence should be addressed. E-mail: kitade@ mb.kyoto-phu.ac.jp. Fax: +81-75-595-4660. (1) Mosiello, l.; Laconi, C.; Gallo, M. D.; Ercole, C.; Lepidi, A. Sens. Actuators, B 2003, 95, 315-320. (2) Elcole, C.; Gallo, M. D.; Mosiello, L.; Baccella, S.; Lepidi, A. Sens. Actuators, B 2003, 91, 163-168. (3) Purvis, D.; Leonardova, O.; Farmakovsky, D.; Cherkasov, V. Biosens. Bioelectron. 2003, 18, 1385-1390. (4) Kojima, K.; Hiratsuka, A.; Suzuki, H.; Yano, K.; Ikebukuro, K.; Karube, I. Anal. Chem. 2003, 75, 1116-1122.
6802 Analytical Chemistry, Vol. 76, No. 22, November 15, 2004
of antigen-antibody systems, and molecularly imprinted materials have therefore been dubbed “artificial antibodies”, “antibody mimics”, or “antibody analogues”. MIPs have several advantages over real antibodies. For example, they are mechanically and chemically stable and physically robust, they can easily be synthesized in similar manner even if the kind of template molecule is different, and they are applicable to almost all organic compounds with inexpensive manufacturing. MIPs have recently been utilized as a selective molecular recognition membrane or layer in chemical sensing systems.5-8 For transducers of the MIP-based sensors, voltammetry,9,10 quartz crystal microbalances,11-13 amperometry,14,15 surface plasmon resonance,16 field-effect devices,17,18 and potentiometry19 have been applied. In these transducers, the structure and transducing mechanism of the potentiometric sensing system are relatively simple, and the system can easily be scaled down and applied to both the single microsensor or multimicrosensor arrays. However, there have been only a few reports dealing with potentiometric immunosensors utilizing artificial antibodies based on MIPs.19 The reason that the potentiometric artificial immunosensor has hardly been developed despite its many advantages may be partially related to the difficulty of improvement in the sensitivity, accuracy, precision, and stability. If MIPs could be applied to the sensing element of a potentiometric immunosensor, it would be possible to make a potentiometric artificial immunosensor that could be used even in harsh environments, be repetitively measured, and be miniaturized. (5) Haupt, K.; Mosbach, K. Chem. Rev. 2000, 100, 2495-2504. (6) Marx, S.; Zaltsman, A.; Turyan, I.; Mandler, D. Anal. Chem. 2004, 76, 120126. (7) Shoji, R.; Takeuchi, T.; Kubo, I. Anal. Chem. 2003, 75, 4882-4886. (8) Malitesta, C.; Losito, I.; Zambonin, P. G. Anal. Chem. 1999, 71, 13661370. (9) Blanco-Lopez, M. C..; Fernandez-Llano, L.; Lobo-Castanon, M. J.; MirandaOrdieres, A. L.; Tunon-Blanco, P. Anal. Lett. 2004, 37, 915-927. (10) Huang, H. C.; Lin, C. I.; Joseph, A. K.; Lee, Y. D. J. Chromatogr., A 2004, 1027, 263-268. (11) Das, K.; Penelle, J.; Rotello, V. M. Langmuir 2003, 19, 3921-3925. (12) Fu, Y.; Finklea, H. O. Anal. Chem. 2003, 75, 5387-5393. (13) Cao, L.; Li, S. F. Y.; Zhou, X. C. Analyst 2001, 126, 184-188. (14) Kriz, D.; Mosbach, K. Anal. Chim. Acta 1995, 300, 71-75. (15) Shoji, R.; Takeuchi, T.; Suzuki, H.; Kubo, I. Chem. Sens. 2002, 18 (B), 7-9. (16) Kugimiya, A.; Takeuchi, T. Biosens. Bioelectron. 2001, 16, 1059-1062. (17) Pogorelova, S. P.; Zayats, M.; Bourenko, T.; Kharitonov, A. B.; Lioubashevski, O.; Katz, E.; Willner, I. Anal. Chem. 2003, 75, 509-517. (18) Hedborg, E.; Winquist, F.; Lundstrom, I.; Andersson, L. I.; Mosbach, K. Sens. Actuators, A 1993, 37-38, 796-799. (19) Zhou, Y.; Yu, B.; Shiu, E.; Levon, K. Anal. Chem. 2004, 76, 2689-2693. 10.1021/ac040098q CCC: $27.50
© 2004 American Chemical Society Published on Web 10/19/2004
Figure 1. Construction of the potentiometric artificial immunosensor based on molecularly imprinted polymers.
In the present study, we developed a novel potentiometric artificial immunosensor with a sensing element based on MIPs instead of a real antibody. We also present a method that can be applied to preparation of MIPs from a hydrophilic template molecule, which is ordinarily relatively difficult compared with using a hydrophobic template. It is expected that the sensor would have not only advantages that are caused by using MIPs but also superior performance as follows. Because the very thin plasmapolymer layer functioning as an interface is put between the sensing element and transducer, adhesion between the transducer and sensing element is improved. Consequently, repetitive measurement becomes possible by improving the mechanical durability of the device. Moreover, because the film thickness of the plasma-polymer layer used as an interface material is controllable, controlling the thickness of a very thin sensing element can be achieved by absorbing and immobilizing the MIPs into the plasma-polymer layer. Therefore, very small change in potential could be measured, and the response speed would be increased. As the structure and transducing mechanism of this sensor are very simple, construction of a single microsensor will be easily developed and multimicrosensor arrays applicable to several kinds of analytes will possibly be attainable by changing the template molecule to be imprinted and by reducing the sensor size. The final purpose of this study is to apply this sensor to a detecting element in a micro total analysis system, thereby facilitating easy clinical analysis. EXPERIMENTAL SECTION Materials and Reagents. Ethylbenzene (Kanto Chemical Inc.) and styrene (Wako Pure Chemical Inc.) were used as the monomers of the plasma-polymer for interface material; dibutyl phthalate (Wako Pure Chemical Inc.) was used as the activating solvent; sodium dodecyl sulfate (SDS) (Wako Pure Chemical Inc.)
was used as the surfactant; 2,4-dimethylvaleronitrile (V-65; Wako Pure Chemical Inc.) was used as the radical initiator; serotonin hydrochloride (Tokyo Kasei Kogyo Inc.) was used as the template; methacrylic acid (Wako Pure Chemical Inc.) was used as the functional monomer; ethylene glycol dimethacrylate (Wako Pure Chemical Inc.) was used as the cross-linking monomer; toluene (Wako Pure Chemical Inc.) was used as the porogenic solvent; poly(vinyl alcohol), degree of polymerization 1000 and degree of hydrolysis 86-90 mol % (Wako Pure Chemical Inc.), was used as the dispersion stabilizer; methanol (Kanto Chemical Inc.), tetrahydrofuran (Wako Pure Chemical Inc.), acetone (Wako Pure Chemical Inc.), tryptamine hydrochloride (Tokyo Kasei Kogyo Inc.), procainamide (Nacalai Tesque Inc.), and acetaminophen (Tokyo Kasei Kogyo Inc.) were used without further purification, except for ethylene glycol dimethacrylate, which was purified to remove the polymerization inhibitor. Distilled water from deionized water was used as the solvent for the stock and test solutions. Construction and Structure of Sensor. The construction method and structure of the potentiometric artificial immunosensor are shown in Figure 1. In this study, the sensing element of serotonin-imprinted polymer was absorbed and immobilized in a plasma-polymer layer by a method of swelling and polymerization.20 First, thin-layer platinum was sputter-coated by an ion coater (IB-3, Eiko Engineering Inc.) on a surface of a completely cleaned glass plate (76 × 26 mm) to make a Pt electrode functioning as a transducer. Then, the very thin (not more than 1 µm) plasmapolymer functioning as an interface was deposited on the platinum layer. Plasma coating was carried out with a model BP-1 plasma deposition system (Samco International Lab., Kyoto, Japan) composed of a glass bell jar, a stainless steel base, and a pair of (20) Ugelstad, J.; Kaggerud, K. H.; Hansen, F. K.; Berge, A. Macromol. Chem. 1979, 180, 737-744.
Analytical Chemistry, Vol. 76, No. 22, November 15, 2004
6803
Figure 2. Response curves of the serotonin-imprinted polymer-modified sensor (SIPS) and the nonimprinted polymer-modified sensor (NIPS). Interface, plasma-polymerized ethylbenzene (0.5 µm). Concentration of the substances, 1 × 10-3 M.
parallel electrodes 25, 28, and 10 cm in diameter, respectively. The upper electrode was connected to a high-frequency (13.56 MHz) power supplier via a matching network. The reactor was evacuated by a rotary pump. The gas pressure in the reactor was monitored by a thermocouple vacuum gauge. Monomer gas, styrene or ethylbenzene, was transported into the reactor by evaporation from the reservoir. A very thin plasma-polymer was deposited on the surface of the platinum layer of the Pt electrode by placing it on the lower electrode. The thickness of the plasmapolymer layer was determined by a calibration curve relating deposition time and film thickness on a glass slide with the time and thickness being determined using a reflex interferometer (Mizojiri Optical Instruments, model 2, Tokyo, Japan). First-step swelling was performed by immersing the plasma-coated device in an emulsified mixture of dibutyl phthalate (3.8 mL) functioning as the activating solvent, V-65 (0.34 g) functioning as the radical initiator, SDS (0.16 g) functioning as the surfactant, and distilled water (40 mL), all of which were emulsified by sonication. This first-step swelling was carried out at room temperature with stirring for 24 h. Second-step swelling was continuously performed by immersing the swollen device in a dispersed mixture of serotonin hydrochloride (2.3 mmol) functioning as the template, methacrylic acid (3.0 mmol) functioning as the functional monomer, ethylene glycol dimethacrylate (25.0 mmol) functioning as the cross-linking monomer, toluene (47 mmol) functioning as the porogenic solvent, and distilled water (45 mL) containing poly(vinyl alcohol) (0.96 g) functioning as the dispersion stabilizer, all of which were uniformly dispersed by sonication. The dissolution of the serotonin to the hydrophobic phase is promoted by combining the serotonin in the water phase with methacrylic acid. This second-step swelling was carried out at room temperature with stirring for 24 h. After the second-step swelling was completed, the medium was fully deaerated by helium gas for 10 min. The polymerization procedure was then carried out at 7075 °C under helium atmosphere with stirring for 24 h. The device was washed successively in distilled water, methanol, tetrahydrofuran, and acetone for 30 min or more to remove the template. After the sensor was completely dried under the atmosphere, it was used for the measurement. 6804 Analytical Chemistry, Vol. 76, No. 22, November 15, 2004
Measurement of Sensor Response. The response potential of the obtained sensor to serotonin or other chemical substances in water versus the potential of the Pt electrode serving as the reference was monitored by a potentiometer (pH meter F-23, Horiba Inc.). The Ag/AgCl electrode is generally used for a reference electrode. However, we used the Pt electrode as a reference electrode in this study, because miniaturization is easier by using a Pt electrode than an Ag/AgCl electrode. The potentiometer was controlled by a personal computer through an RS232C interface to obtain real-time data acquisition and recording. All the measurements were carried out in a glass beaker containing 100 mL of distilled water. When the potential change of the sensor was measured, a suitable amount of stock solution (0.1 mol/L) was injected into the beaker with a microsyringe. After the measurement, the sensor was washed with distilled water to remove serotonin in the MIP for the next analysis. RESULTS AND DISCUSSION Selectivity. Potential response curves of the serotoninimprinted polymer-modified sensor (SIPS) to serotonin, tryptamine, acetaminophen, and procainamide were compared. Tryptamine was tested due to its structural similarity to serotonin. Acetaminophen has a Log P value almost equal to that of serotonin. Procainamide has a chemical structure different from serotonin. To confirm the efficiency of the MIP, the response curves of the nonimprinted polymer-modified sensor (NIPS) to the same substances were also examined. These results are shown in Figure 2. We find that response time, defined as the time required to achieve 95% of the magnitude of the equilibrated signal, corresponds to ∼12 s. On the other hand, stabilization time of this sensor is ∼10 min. In the response curve of serotonin measured by NIPS, a pronounced initial increase followed by a decrease in potential was observed immediately after injection. We cannot clearly explain this phenomenon. The selectivities for SIPS and NIPS were compared. A specific response to the serotonin could be observed for the SIPS but not for the NIPS, suggesting that the molecular imprinting is effective in giving serotonin selectivity to this sensing mechanism and the serotonin-imprinted polymer has a complementary cavity of
Table 1. Selectivity Coefficients for the Serotonin Sensor Determined by the Separate Solutions Method interfering material
concn (× 103 M)
selectivity coefficient pot (Log K 5-HT,j )
tryptamine procainamide acetaminophen
1 1 1
-3.88 -4.39 -3.43
serotonin molecule. The potential change of the NIPS may be caused by nonspecific binding to the carboxyl group on the surface of the nonimprinted polymer, adsorption and distribution to the sensing element, or both. Selectivity coefficients for the SIPS are pot summarized in Table 1. The selectivity coefficients Log K 5-HT,j of the sensor toward different species were determined by a separate solutions method, in which the following equation was used: pot Log K 5-HT,j ) a5-HT/aj
where a5-HT and aj are the serotonin and interfering material concentrations, which show the even electrode potential. SIPS of course had high selectivity for procainamide, which has a chemical structure different from that of serotonin. SIPS also shows high selectivity for tryptamine, which is structurally similar to serotonin, and acetaminophen, whose Log P value almost equals that of serotonin. A superior feature of this method is that the shape of the complementary cavity on the MIP that forms at an interface between MIP and the water phase reflects the molecular conformation of the serotonin in the water phase. This relationship may explain the strong selectivity of this sensor. Based on the above results, it was found that there is a strong possibility of preparing the potentiometric artificial immunosensor by our suggested strategy. Interface Material. An analyte can be specificically determined by measuring a potential change that is caused when the analyte is fit into a complementary cavity of a MIP at the surface of a sensing element. To transmit the surface potential of the sensing element to a transducer of a Pt electrode, appropriate conductivity of the sensing element is necessary. However, the MIP does not have appropriate conductivity to efficiently transmit a surface potential. Hence, it is difficult to measure the surface potential by using only MIP as a sensing element. Our proposed sensing element acquires appropriate conductivity by absorbing the MIP into the plasma-polymer, which has appropriate conductivity as will be described next, and by making it ultrathin layer. By forming this sensing element onto the Pt electrode functioning as a transducer, surface potential of the sensing element can be measured. On the other hand, MIP functioning as the sensing element of commonly investigated artificial immunosensors, which was made from acrylic-based MIPs, is directly formed on the metal surfaces of transducers such as Au and Pt; hence the sensor is not particularly durable because the MIP shows poor adhesiveness to the metal and easily peels off the metal surface. If the MIP does not closely adhere to the metal surface, the signal of electrochemical change occurring on the surface of the MIP is
Figure 3. Response curve of the sensor, which was constructed by plasma-polymerized styrene as the interface. Thickness of the plasma-polymer, 0.5 µm. Concentration of the substances, 1 × 10-3 M.
not efficiently transmitted to the transducer. By a plasma polymerization method, the gradient layer, which was a mixture of Pt, served as the transducer material, and plasma-polymer was formed between the transducer and the plasma-polymer layer. Therefore, the durability and transmissibility of the surface potential change may be improved by using a plasma-polymer as the interface material. Furthermore, a uniform, pinhole-free very thin film can be obtained by plasma polymerization, and the thickness of the plasma-polymer film is strictly controllable to below 1 µm. By absorbing and immobilizing the MIP into this very thin plasma-polymer layer, a very thin sensing element, for which the thickness is controlled, can be made. Consequently, a very small change in potential can be measured, and the response speed is increased. Because the chemical structure and properties of the plasmapolymer change greatly according to the monomer species, response curves measured by sensors that were constructed with different plasma-polymers were compared. The monomers investigated were ethylbenzene and styrene. The results of the plasma-polymerized ethylbenzene and styrene as the interface are shown in the left part of Figure 2 and Figure 3, respectively. In the case of using plasma-polymerized ethylbenzene, a clearly intense potential change in response to the serotonin, compared to that observed with the other substances, was observed; hence it can be concluded that the sensor recognized the serotonin. On the other hand, in the case of using plasma-polymerized styrene, the sensor had less recognition ability. We cannot clearly explain the reason that different results were obtained by plasmapolymerized ethylbenzene and styrene. Potential responses of the ethylbenzene and styrene plasmapolymer layer without MIP, i.e., bare plasma-polymer, to serotonin, tryptamine, and procainamide were investigated. Both plasma-polymers almost evenly responded to these substances, and the potential changes were 10-20 mV. This phenomenon may have originated from nonspecific adsorption, distribution to the plasma-polymer, or both. The effects of the thickness of the plasma-polymer made from ethylbenzene, i.e., the thickness of the sensing element, on the potential response of the sensor were investigated. However, it must be noticed that the final thickness of the sensing element is not same as the plasma-polymer layer because the MIP is created inside the plasma-polymer layer after swelling. The results are Analytical Chemistry, Vol. 76, No. 22, November 15, 2004
6805
Figure 4. Effects of the plasma-polymer thickness on the potential response of the sensor. Interface, plasma-polymerized ethylbenzene. Serotonin concentration, 1 × 10-3 M.
Figure 5. Effects of the heating time for polymerization on the potential response of the sensor. Interface, plasma-polymerized ethylbenzene (0.5 µm). Serotonin concentration, 1 × 10-3 M.
shown in Figure 4. The sensor with a 0.5-µm plasma-polymer layer gave a higher and quicker potential response than the 0.25µm layer. Therefore, the sensor with a 0.5-µm plasma-polymer layer had a superior response property than that with a 0.25-µm layer. Obtaining a smaller potential change with the thinner plasma-polymer layer may have been caused by the short effect, by which the electrical conductivity of a sensing element is increased when it is excessively thin or if a pinhole forms in the element due to its weak mechanical strength. Effect of Heating Time for MIP Polymerization on Potential Response. To improve the MIP selectivity, the shape of the complementary cavity of the MIPs must be completely retained after removing the template molecule. A high degree of crosslinking and a rigid chemical structure is necessary to maintain the shape of the complementary cavity. In this study, we adopted a heating polymerization method, and an extremely short heating time may cause insufficient rigidity of the MIPs. Hence, the effects of heating time on the potential response were investigated to determine an appropriate heating time. Figure 5 shows the response curves measured by the sensors prepared with various heating times. It was confirmed that the potential changes increased and stabilized in a short time with increases in the heating time. These results indicate that a sufficient heating time (24 h) to achieve polymerization is necessary to form a rigid complementary cavity and to improve sensor performance. Contact Area of the Sensing Element with Test Solution. To develop a single microsensor or multimicrosensor arrays, the sensor must maintain a constant potential under various contact 6806 Analytical Chemistry, Vol. 76, No. 22, November 15, 2004
Figure 6. Response curves measured under different contact areas with test solution. The length is the distance from the lower end of the sensor. The values in parentheses correspond to the contact area with test solution. Interface, plasma-polymerized ethylbenzene (0.5 µm). Serotonin concentration, 1 × 10-3 M.
areas of the sensing element with a test solution. Hence, the effects of changing the contact area with test solution on the response potential were investigated. The results are shown in Figure 6. The lengths shown in the figure correspond to the distance from the lower end of the sensor, and these values are proportional to the contact area with the test solution. The values in parentheses correspond to the contact areas with test solution. The response curves labeled 0.1, 1, and 2 cm were measured by dipping the part of the sensor consisting of the MIP-absorbing plasmapolymer into the test solution. On the other hand, the response curves labeled 3 and 4 cm were measured by dipping the part of the sensor consisting of the plasma-polymer with and without MIP into the test solution. In the case of 0.1, 1, and 2 cm, all corresponding to a normal sensor, the change in the contact area in response to the test solution had almost no effect on the response potential. The slight decrease in the response potential observed on the 0.1 cm response curve may have been caused by a short effect occurring due to a very small crack on the bottom edge of the sensor. The contribution ratio of this short effect seems to be high due to its narrow contact area. The performance of the sensor severely deteriorated in the case of 3 and 4 cm. The potential response of the 3 and 4 cm condition primarily originates from nonspecific adsorption, distribution, or both to the plasma-polymer. These results indicate that miniaturization of the sensor is possible by reducing the size of sensing element. Calibration Curve. A dynamic response curve for the sensor representing step changes in the serotonin concentrations is shown in Figure 7. Test solutions in the range from 1 pmol/L to 10 mmol/L were analyzed. As the response potential change appeared at a concentration of 100 pmol/L, the detection limit of this sensor could be considered to be 100 pmol/L. The calibration curve is shown in Figure 8, and it showed a linear range from 1 µmol/L to 10 mmol/L with a correlation coefficient of 0.996 for 15 data points. The relative standard deviation of the five successive determinations was 2.4% at a level of 100 µmol/L. The slope of the linear dynamic range was 24.7 mV/decade. The calibration curve did not show the Nernstian response slope. This phenomenon may have been caused by incompleteness of the optimization of conditions that was concerned with electric
Figure 7. Dynamic response of the serotonin sensor for step changes in the serotonin concentrations. Interface, plasma-polymerized ethylbenzene (0.5 µm). Serotonin concentration, 1 × 10A mol/ L.
Figure 8. Calibration curve of serotonin in water. Interface, plasmapolymerized ethylbenzene (0.5 µm). Each point shows the average of three determinations.
transmittance. Investigation of the conditions for showing the Nernstian response is now being carried out. Therefore, we cannot clearly explain this phenomenon now. These results indicate that highly sensitive, accurate, and precise measurements will be possible with this sensor.
Repeatability. After the determination, serotonin could be removed from the MIP by immersing the sensor in water. With this washing treatment, the response potential was easily returned within 5 mV of the primary value, and the sensor was recovered within 5 min. It was therefore confirmed that repetitive use is possible with this sensor. After the measurement, the sensor can be washed, dried, and then stored under atmosphere until the next determination. Moreover, the sensor has superior durability because the MIP is absorbed and immobilized in the highly crosslinked plasma-polymerized ethylbenzene with a very rigid chemical structure. After the measurements were carried out more than 50 times for 200 days, the response still remained at 96% of its initial magnitude. Hence, the sensor performance was found to be sufficiently stable. CONCLUSION A potentiometric artificial immunosensor was developed by using artificial antibody based on molecularly imprinted polymer using serotonin as a model template. The sensor consists of a Pt electrode functioning as a transducer and MIP absorbed in the plasma-polymer functioning as a sensing element. The sensor was found to have superior selectivity and specificity to serotonin and to proportionally respond to serotonin concentrations. MIP preparation is easy and inexpensive and applicable to various chemical substances. Application of MIP to the sensing element of a potentiometric artificial immunosensor will provide a simpler, more specific, and less expensive sensing system than an ordinal immunosensor using real antibodies. As the structure and transducing mechanism of the proposed sensor are very simple, construction of a single microsensor should be quite easy, and multimicrosensor arrays applicable to several kinds of analytes will be attainable by changing the template molecule to be imprinted and by reducing the sensor size. The effects of interfering substances are now under investigation. Received for review May 18, 2004. Accepted August 18, 2004. AC040098Q
Analytical Chemistry, Vol. 76, No. 22, November 15, 2004
6807
