Capacitive Detection in Ultrathin Chemosensors Prepared by

Anal. Chem. 2007, 79, 3220-3225

Correspondence

Capacitive Detection in Ultrathin Chemosensors Prepared by Molecularly Imprinted Grafting Photopolymerization Tetyana L. Delaney,*,†,‡ Denys Zimin,§ Michael Rahm,| Dieter Weiss,| Otto S. Wolfbeis,† and Vladimir M. Mirsky†

Institute of Analytical Chemistry, Chemo- and Biosensors, Institute of Experimental and Applied Physics, and Institute of Physical and Theoretical Chemistry, University of Regensburg, D-93040 Regensburg, Germany, and Department of Biology, St. Joseph’s College, Brooklyn, New York 11205

The usual applications of capacitive detection in chemoand biosensors are based on changes in effective thickness of insulating layers due to adsorption of analyte onto receptors. Ultrathin chemosensors based on molecularly imprinted polymerization enable a realization of another capacitive approach that exploits changes in electrical capacitance due to modification of the dielectric constant of the polymer. Such chemosensors were prepared by photografted molecularly imprinted polymerization on the surface of gold electrodes. An adsorbed layer of hydrophobic photoinitiator (benzophenone) provided grafted polymerization on the surface of the alkanethiol-modified gold electrode. The chemosensors were characterized by cyclic voltammetry, impedance spectroscopy, and scanning electron and atomic force microscopy. Binding of analyte was detected by measurements of electrical capacitance. The results indicate a decrease of the dielectric constant of the polymer layer due to analyte binding up to 20%. Capacitive detection is a routine technique to study adsorption in classical electrochemistry.1,2 During the last 10 years, it was extensively applied to detect binding in immunosensors,3-7 * To whom correspondence should be addressed. Fax: +1-603-908-7937. E-mail: [email protected]. † Institute of Analytical Chemistry, Chemo- and Biosensors, University of Regensburg. ‡ St. Joseph’s College. § Institute of Experimental and Applied Physics, University of Regensburg. | Institute of Physical and Theoretical Chemistry, University of Regensburg. (1) Bard, A.; Faulkner, L. Electrochemical Methods: Fundamentals and Applications, John Wiley & Sons: New York, 1980. (2) Damaskin, B.; Petrii, O.; Batrakov, V. Adsorption of Organic Compounds on Electrodes; Plenum: New York, 1971. (3) Mirsky, V. M.; Riepl, M.; Wolfbeis, O. S. Biosens. Bioelectron. 1997, 12, 977-989. (4) Sadik, O.; Xu, H. Anal. Chem. 2002, 74, 3142-3150. (5) Ameur, S.; Maupas, H.; Martelet, C.; Jaffrezic-Renault, N.; Ben Ouada, H.; Cosnier, S.; Labbe, P. Mater. Sci. Eng. 1997, C 5, 111-119. (6) Berggren, C.; Bjarnason, B.; Johansson, G. Biosens. Bioelectron. 1998, 13, 1061-1068. (7) Dijksma, M.; Kamp, B.; Hoogvliet, J.; van Bennekom, W. Anal. Chem. 2001, 73, 901-907.

3220 Analytical Chemistry, Vol. 79, No. 8, April 15, 2007

enzymatic biosensors with deposited substrate layers,8,9 DNA sensors,10,11 and biological12 and chemical sensors.13-16 Signal changes in such biosensors are typically in the range of a few percent, but the high stability of the signal allows one in many cases to obtain a relatively high signal-to-noise ratios. Higher sensitivity can be reached by using different signal amplification schemes. based on the deposition of second layers of antibodies,17 multilayers of avidin-biotin or biotinylated liposomes,11 or insoluble sediment formed by enzymatic labels of protein.18 In most cases, these techniques are based on changes of electrode capacity due to increase of thickness of the insulating layer (Figure 1a);1-15,18 in some cases, changes of double layer capacitance due to modification of surface or boundary potential are detected.19 In the past few years, molecularly imprinted polymerization was extensively used to form artificial receptors.20-25 The mechanism (8) Krause, C.; Mirsky, V. M.; Heckmann, K. Langmuir 1996, 12, 6059-6064. (9) Mirsky, V. M.; Mass, M.; Krause, C.; Wolfbeis, O. S. Anal. Chem. 1998, 70, 3674-3678. (10) Wrobel, N.; Riepl, M.; Schinkinger, M.; Mirsky, V. M.; Wolfbeis, O. S. Med. Biol. Eng. Comput. 1999, 37, suppl. 2, 362-363. (11) Patolsky, F.; Lichtenstein, A.; Willner, I. Angew. Chem. 2000, 112 (5), 970973. (12) Bajari, T.; Lindstedt, K.; Riepl, M.; Mirsky, V. M.; Nimpf, J.; Wolfbeis, O. S.; Dresel, H.; Bautz, E.; Schneider, W. Biol. Chem. 1998, 379, 10531062. (13) Mirsky, M. V.; Hirsch, T.; Piletsky, S.; Wolfbeis, O. S. Angew. Chem. 1999, 111, 1179-1181. (14) Hirsch, T.; Kettenberger, H.; Wolfbeis, O.; Mirsky, V. M. Chem Commun. 2003, 432-433. (15) Prodramidis, M.; Hirsch, T.; Wolfbeis, O. S.; Mirsky, V. M. Electroanalysis 2003, 15, 1-4. (16) Mirsky, V. M., Ed. Ultrathin Electrochemical Chemo- and Biosensors; Springer: Berlin, 2004. (17) Riepl, M.; Mirsky, V. M.; Wolfbeis, O. S. Anal. Chim. Acta 1999, 392, 77-84. (18) Katz, E.; Willner, I. In Ultrathin Electrochemical Chemo- and Biosensors; Mirsky, V. M., Ed.; Springer: Berlin, 2004. (19) Schweiss, R.; Mirsky, V. M.; Wolfbeis, O. S. Mater. Sci. Forum 1998, 287288, 427-430. (20) Panasyuk, T.; Mirsky, V. M.; Piletsky, S.; Wolfbeis, O. S. Anal. Chem. 1999, 71, 4609-4613. (21) Panasyuk-Delaney, T.; Mirsky, V. M.; Ulbricht, M.; Wolfbeis, O. S. Anal. Chim. Acta 2001, 435, 157-162. (22) Panasyuk-Delaney, T.; Mirsky, V. M.; Wolfbeis, O. S. Electroanalysis 2002, 14, 221-224. 10.1021/ac062143v CCC: $37.00

© 2007 American Chemical Society Published on Web 03/15/2007

Figure 1. Signals of capacitive sensors can be caused by changes of thickness of the total insulating layer (a) or by changes of dielectric constant of the insulating layer (b).

of capacitive effect in the impedometric sensors, based on the molecularly imprinted polymers,26-30 is different. In this case, binding occurs not only on the surface but also in the pores of a few tenths of a nanometer-thick porous polymer layer. This leads to the replacement of highly polar water molecules by lower polar analyte molecules, followed by essential increase of dielectric constant and corresponding changes of the electrode capacitance (Figure 1b). It is shown here that the signal changes in MIPbased ultrathin capacitive sensors are really caused by the changes of dielectric constant. The metabolite creatinine and the herbicide desmetryn were used as model analytes. In our earlier communication, we described detection of the clinically important analyte creatinine22 and the ecologically relevant herbicide desmetryn25 with the help of artificial receptors. The technique was based on the formation of an ultrathin (but not monomolecular) three-dimensional poly(23) Cheng, Z.; Wang, E.; Yang, X. Biosens. Bioelectron. 2001, 16, 179. (24) Piletsky, S., Turner, A., Eds. Molecular Imprinting of Polymers Landes Biosciences: Georgetown, TX, 2006. (25) Baggiani, C.; Anfossi; L.; Giovannoli, C. Curr. Pharm. Anal. 2006, 2 (3), 219-247. (26) Owens, P.; Karlsson, L.; Lutz, E.; Andersson, L. Trends Anal. Chem. 1999, 18, 146-154. (27) Dickert, F.; Hayden, O. Trends Anal. Chem. 1999, 18, 192-199. (28) Yano, K.; Karube, I. Trends Anal. Chem. 1999, 18, 199-204. (29) Haupt, K.; Mosbach, K. Chem. Rev. 2000, 100, 2495-2504. (30) Piletsky, S.; Panasyuk, T.; Piletskaya, E.; Nicholls, I.; Ulbricht, M. J. Membr. Sci. 1998, 4096, 1-16.

Figure 2. Impedance spectra of bare gold electrode, gold electrodes coated with hexadecanethiol (HD), and gold electrodes coated by MIP presented as Bode plots (A, B). Electrolyte: 5 mM K3[Fe(CN)6] in 100 mM NaCl, pH 7.5. Equivalent circuits (C) used to describe impedance spectra of coated electrodes: Rs, electrolyte resistance; Rct, reaction resistance; Cdl, capacitance.

mer layer directly on the transducer surface. This was first accomplished by electropolymerization,20,31 but low cross-linking of these polymers resulted in a low temporal stability. Recently, a grafting photopolymerization was suggested to form highly cross-linked molecularly imprinted polymers on the surface of a capacitive transducers.21,22 High selectivity, sensitivity, and stability were observed. In this paper, we characterize the chemosensors based on photografted molecularly imprinted polymers and discuss the mechanism of signal generation. The results suggest that capacity (31) Malitesta, C.; Losito, I.; Zambonin, P. Anal. Chem. 1999, 71, 1366.

Analytical Chemistry, Vol. 79, No. 8, April 15, 2007

3221

changes are caused by modification of the dielectric constant of the polymer layer. EXPERIMENTAL SECTION Preparation of Sensitive Layers on the Surface of Gold Electrodes. 2-Acrylamido-2-methyl-1-propanesulfonic acid and N,N-methylenediacrylamide were obtained from Aldrich, hexadecanethiol was from Fluka, and other reagents were from Merck. All chemicals were used as received. All measurements were performed at 22 °C. The gold electrodes consisted of a thin gold layer sputtered onto a glass or silicone support with an adhesive chromium sublayer. The electrodes had a square shape with a macroscopic area of 2.4 mm2 and connected to a contact pad by a contact line of 8-mm length and 10-µm width. The electrodes were cleaned with a hot mixture of piranha solution (a 1:3 mixture of 30% H2O2/ concentrated H2SO4), rinsed with water and dried. Caution: this solution reacts violently with most organic materials and must be handled with extreme care. Then, the electrodes were immersed into a 5 mM solution of hexadecanethiol in chloroform for at least 5 h and washed briefly with ethanol. A modification of the procedure previously established in our laboratory was used for photografting on gold surfaces covered with an alkanethiol monolayer: hexadecanethiol-covered gold surfaces were coated with a benzophenone layer by immersion for 15 min in a 150 mmol/L solution of benzophenone in acetone. Acetone was removed within 5 min in a drying cabinet at 40 °C. Then, 25 µL of the prepolymerization mixture was applied on the surface and covered with a quartz slide. Thus, a thin liquid film was formed between two surfaces. Irradiation was performed by a high-pressure mercury lamp (100 W, type HBO200/4 from Spindler & Hoyer, Go¨ttingen, Germany) for 15 min for a large number of electrodes simultaneously. The irradiated spot had a diameter of ∼15 cm. Characterization of a Modified Gold Surface. Impedance measurements were performed using Autolab PGSTAT-13 electrochemical workstation (EcoChemie, The Netherlands). A standard three-electrode electrochemical cell was used for the CV measurements. The electrode assembly consisted of a bare or modified Au electrode as the working electrode, a Pt wire as counter electrode, and a saturated calomel electrode (SCE) as reference electrode. Impedance measurements were perfomed at dc offset potential of +300 mV vs SCE, ac potential of 20 mV. The frequency range studied was from 10 mHz to 10 kHz. However, because of high capacitance and high resistance of the studied organic films, the low-frequency range (10 mHz to 100 Hz) was more informative and used in the most experiments. Cyclic voltammograms and impedance spectra were recorded in the electrolyte consisting of 5 mM K3Fe(CN)6, 10 mM phosphate buffer (pH 7.5), 100 mM NaCl. Precise capacitive measurements of sensor responses were performed in the 25 mM phosphate buffer, 100 mM NaCl, pH 7.5. The electrode capacitance was measured by registration of the 90° component of the capacity current by means of a lock-in amplifier (PAR, model 121) at 80 Hz in twoelectrode configuration with dc offset potential +300 mV versus SCE. Surface roughness and porosity of the modified gold electrodes were characterized by atomic force microscope (AFM) and scanning electron microscope (SEM). The SEM was a Topcon SM 510. The AFM Autoprobe CP (Park Scientfic Instruments, Sunnyvale, CA; presently Veeco) with multitask head was used. Imaging was performed in noncontact mode using the standard 3222 Analytical Chemistry, Vol. 79, No. 8, April 15, 2007

Figure 3. Changes in electrical capacitance of chemical sensors based on molecularly imprinted polymers with desmetryn (MIP-D) and creatinine (MIP-C) as template due to additions of corresponding analytes (A) and response kinetics of the sensor toward desmetrine on subsequent additions of creatinin and desmetryn (B).

NC-cantilevers from the same manufacturer. The postmeasurement processing of images and curves was performed using ProScan Version 1.6 Image Processing software of Autoprobe CP. RESULTS AND DISCUSSION Molecularly imprinted polymer (MIP) modification of the electrode led to major alterations of the impedance spectra. The Bode plot of bare electrodes and of electrodes coated by alkanethiol and alkanethiol with subsequently deposited polymer are shown in the Figure 2. As expected, impedance spectra of uncoated electrodes can be described by Randles equivalent circuit with essential contribution of Warburg’s impedance. Coating of the electrodes with alkanethiol leads to increase of the impedance and to switching the reaction limiting stage from the diffusion to the reaction one. At a frequency of more than 50 Hz, the coated electrodes have mainly capacitive properties (the phase angle becomes close to 90° and the log|Z| vs log f function is a straight line with a slope of -1). The log|Z| remains nearly unchanged at low frequencies (where the resistive components dominate) and decreases at high frequencies. The reaction resistance of the alkanethiol-coated electrode can be estimated as ∼150 kΩ/cm2 (Table 1). The specific capacitance of the alkanethiol-coated electrode corresponds to published data.8 Then, the alkanethiol-

Figure 4. Scanning electron microscopy of bare gold electrode (A), gold electrode coated by polymer without template (B), and gold electrode coated by molecularly imprinted polymer to creatinine (C). Acceleration voltage, 25 kV.

coated electrodes were deep into benzophenone solution in acetone and irradiated by UV light in the presence of polymerizable monomers and template molecules. It leads to an increase in capacitance and resistance (Figure 2). The increase in the capacitance can be explained by two reasons. The first reason could be some damage of the insulating layer of alkanethiols. The second reason may result from a strong increase in hydrophilicity of the surface. This effect was observed earlier21 in measurements of contact angles; an increase of surface hydrophilicity leads to an increase of electrical capacitance of thin layers. No decrease in the capacitance corresponded to the formation of the polymer layer; it indicates on high dielectric constant of the polymer. This

suggests the polymer consisted of small aqueous volumes separated from each other by a very thin insulating layer, probably as a result of micelle formation in the prepolymerization mixture. Cyclic voltammetry confirms an increase of insulating properties due to deposition of alkanethiol and polymer: typical peaks of oxidation and reduction of ferro-/ferricyanide disappeared after coverage of the gold electrode with a monolayer of hexadecanethiol. Finally, a complete blockage of the electrochemical activity was observed after MIP deposition. Some electrodes after polymerization demonstrated very low electrical capacitance (typically 2-8 nF) and extremely high resistance (100-300 MΩ). This could be explained by formation Analytical Chemistry, Vol. 79, No. 8, April 15, 2007

3223

of a nonporous polymer layer with very low water content. Such electrodes did not display any response on addition of analytes. Strongly prevalent capacitive properties of the MIP-coated electrode at 80 Hz allows us to relate changes in the capacitive current at this frequency to changes in the electrode capacitance. The electrodes have selective response on additions of analytes: no response of the polymer imprinted by desmetrine on addition of creatinine was observed (Figure 3). Binding of desmetrine to the desmetrine-imprinted electrode leads up to 20% decrease of the electrode capacitance; binding of creatinine to the creatinineimprinted electrode results in up to 10% decrease in the electrode capacitance. Detailed investigation of analytical properties of these electrodes was published elsewhere.21,22 The usual technology of molecularly imprinted polymerization required addition of some porogen (methanol or others) into the prepolymerized solution. In our experiments, it was not necessary; electron microscopy investigation (Figure 4) demonstrated high porosity of the polymers formed without additions of porogens. This may be caused by polymerization in aqueous phase: many monomers and cross-linkers as well as template molecules are surface active and can lead to formation of micelles in the prepolymerization solution. A comparison of Figure 4B and C demonstrates the effect of the template on the polymer porosity. Atomic force microscopy was performed on the strip part of the electrode close to the circle part. It allowed us to use the uncoated part of the silicon surface as a reference point for the height measurement. The results are shown in Figure 5. Bar gold electrodes have a smooth surface with a height of ∼280 nm (Figure 5A). This corresponds approximately to the total thickness of the gold layer and adhesive sublayer obtained by quartz microbalance monitoring during the electrode fabrication. Polymerization without template led to an increase in the height until ∼295 nm (Figure 5B). Therefore, a 15-nm polymer layer is formed. The surface is not as even as for a gold layer, but relatively smooth. The same polymerization procedure performed in the presence of template led to the formation of an ∼55-nm-thick polymer layer with a rough surface (Figure 5C, D). This confirms the suggestion on the formation of micelles in the prepolymerized solution. An estimation of the dielectric constant of sensing layers gives unrealistically high values. It can be caused by presence of electrically invisible large open pores on the rough surface. Comparison of Figure 5C and Figure 5D demonstrates that molecularly imprinted polymers after desorption of template and after its readsorption have the same thickness. Therefore, we can exclude a possible polymer swelling and its contribution to capacitive effects. Taking into account that the capacitance of a planar capacitor is completely determined by its geometry (area and thickness) and dielectric constant, any modification of its capacitance at constant geometry indicates definitely modification of the dielectric constant. It can be explained by replacement of water molecules in the polymer pores by template molecules. This explains the experimental result that the less polar analyte (desmetrine) caused higher capacitance changes then creatinine (Figure 2A). In contrast to effects on the polymer surface, these volume effects inside the capacitor lead to higher capacitive effects. Analyte binding led to 20% decrease of the polymer capacitance. 3224 Analytical Chemistry, Vol. 79, No. 8, April 15, 2007

Figure 5. Atomic-force microscopy of bare gold electrode (A, E), gold electrode coated by polymer without template (B), and gold electrode coated by molecularly imprinted polymer to creatinine with desorbed (C, F) and readsorbed (D) creatinine. The data are presented as profile (A-D) and 3-D image (E, F).

Let us assume that water-filled pores possess the same dielectric constant as water (∼80), and analyte-filled pores has the same dielectric constant as the polymer. A capacitor with randomly distributed pores can be roughly approximated by two serially connected capacitors; one of them has the volume and dielectric constant of the polymer without pores while the second one possesses the dielectric constant of water and its volume corre-

sponds to the total volume of pores. A substitution of the experimentally observed capacitance decrease in this model gives the total volume of the pores ∼30% of the total volume of the sensing layer if the dielectric constant of nonporous polymer material is between 5 and 10 or ∼40% for dielectric constant 20. Thus, we investigated changes of structural and electrochemical properties of gold electrodes during formation of a selective receptor layer by using the molecularly imprinted technology and analyzed the mechanism of the capacitive response. The results

demonstrate a principal difference in the mechanism of the capacity changes in comparison with other types of capacitive sensors: the thickness of the polymer layer was proved to stay constant; therefore, the capacity changes arise due to modification of dielectric constant of the polymer layer.

Received for review November 13, 2006. Accepted February 2, 2007. AC062143V

Analytical Chemistry, Vol. 79, No. 8, April 15, 2007

3225