Composite Carbon Paste Biosensor for Phenolic Derivatives Based on

Anal. Chem. 2003, 75, 5422-5428

Composite Carbon Paste Biosensor for Phenolic Derivatives Based on in Situ Electrogenerated Polypyrrole Binder P. Mailley,*,† E. A. Cummings,‡ S. C. Mailley,‡ B. R. Eggins,‡ E. McAdams,‡ and S. Cosnier*,§

Laboratoire d’Electrochimie Mole´ culaire et Structure des Interfaces, DRFMC/SI3M/EMSI, CEA Grenoble, 17 Avenue des Martyrs, 38054 Grenoble Cedex 9, France, Northern Ireland BioEngineering Centre (NIBEC), University of Ulster at Jordanstown, Newtownabbey Co. Antrim BT37 OQB Northern Ireland, U.K., and Laboratoire d’Electrochimie Organique et de Photochimie Redox, UMR CNRS 5630, Institut de Chimie Mole´ culaire de Grenoble, FR CNRS 2607, Universite´ Joseph Fourier, BP 53, 38041 Grenoble Cedex 9, France

Amperometric biosensors based on new composite carbon paste (CPE) electrodes have been designed for the determination of phenolic compounds. The composite CPEs were prepared by in situ generation of polypyrrole (PPy) within a paste containing the enzyme polyphenol oxidase (PPO). The best paste composition (enzyme/ pyrrole monomer/carbon particles/Nujol) was determined for a model enzyme, glucose oxidase, according to the enzymatic activity of the resulting electrodes and to the enzyme leakage from the paste during storage in phosphate buffer. The in situ electrogenerated PPy enables improvement in enzyme immobilization within the paste since practically no enzyme was lost in solution after 72 h of immersion. Moreover, the enzyme activity remains particularly stable under storage since the biocomposite structure maintains 80% of its activity after 1-month storage. Following the optimization of the paste composition, PPO-based carbon paste biosensors were prepared and presented excellent analytical properties toward catechol detection with a sensitivity of 4.7 A M-1 cm-2 and a response time lower than 20 s. The resulting biosensors were finally applied to the determination of epicatechin and ferulic acid as flavonol and polyphenol model, respectively. Since the pioneering works of Adams in 1958,1 carbon paste electrodes (CPE) have been extensively used for sensor design in electroanalysis.2,3 Effectively, these CPEs, which are based on a mixture of graphite powder and water-immiscible pasting liquid, offer low background current, wide operating potential window, renewability, and low cost.4,5 Moreover carbon pastes, as they are * To whom correspondence should be addressed. E mail: [email protected]; [email protected].. † DRFMC/SI3M/EMSI, CEA Grenoble. ‡ University of Ulster at Jordanstown. § UMR CNRS 5630. (1) Adams, R. N. Anal. Chem. 1958, 30, 1576-1573. (2) Chi, Q.; Gopel, W.; Ruzgas, T.; Gorton, L.; Heiduschka, K. Electroanalysis 1997, 9, 357-365. (3) Kalcher, K.; Kauffmann, J.-M.; Wang, J.; Svancara, I.; Vytras, K.; Neuhold, C.; Yang, Z. Electroanalysis 1995, 7, 5-22.

5422 Analytical Chemistry, Vol. 75, No. 20, October 15, 2003

issued from a blending process, enable bulk modification of the entire electrode material with catalysts, electron mediators, or biomolecules. This property was widely exploited for the design of amperometric biosensors based on immobilized enzymes. A wide variety of additive materials or molecules can also be incorporated within the paste during the fabrication process to mediate the enzyme reaction, to stabilize the biomolecule activity, or to improve the sensor selectivity. However, CPEs exhibits severe limitations due mainly to their weak mechanical stability and their fabrication reproducibility.6 In the case of enzyme-based biosensors, the erosion of the electrode material results in the leakage of the enzyme or, eventually, of additional artificial redox mediators in the solution. Many efforts have been made to design new carbon pastes allowing an efficient retention of the biocatalyst and of the associated electron mediators. A number of these works involve the coating of CPE by an additional membrane such as a Nafion gel,7 a poly(o-phenylene diamine) (PPD) film,8 a PPD-polypyrrole (PPy) copolymer,9 or a poly(4-vinyl-pyridine) film.10 Such a methodology of preparation allows extension of the dynamic range of the sensor and can be employed to decrease the amperometric response of electroactive interfering species. However, this kind of configuration eliminates the possible renewing of the carbon paste surface and lowers dramatically the sensitivity of the resulting biosensor. To circumvent these drawbacks, bulk modification of the carbon paste by polymeric structures or organic stabilizers has been largely employed to improve its mechanical stability or the biomolecule microenvironment.6,11 In particular, the use of an in situ-generated polymeric matrix within the paste has led to the (4) Kulys, F.; Gorton, L.; Dominguez, E.; Enneus, J.; Jarskog, H. J. Electroanal. Chem. 1994, 372, 49-55. (5) Boujita, M.; Boitard, M.; El Murr, N. Biosens. Bioelectron. 1999, 14 545553. (6) Svorc, J.; Miertu, S.; Katrlik, J.; Stred’ansk, M. Anal. Chem. 1997, 69, 20862090. (7) Andrieux, C. P.; Audebert, P.; Divisia-Blohorn; B.; Linquette-Mailley, S. J. Electroanal. Chem. 1993, 353, 289-296. (8) Labo-Castanon, M. J.; Mirada Oderies, A. T.; Tunon-Blanco, P. Anal. Chim. Acta 1996, 346, 165-174. (9) Vidal, J.-C.; Garcia, E.; Mendez, S.; Yarnoz, P.; Castillo, J. R. Analyst 1999, 124, 311-324. (10) Zahia, M. A.; Abtain, S. Anal. Chim. Acta 1997, 354, 351-352. 10.1021/ac034177y CCC: $25.00

© 2003 American Chemical Society Published on Web 09/09/2003

development of rigid biocomposites. Then, the polymer material can be activated by the evaporation of the volatile fraction of a polymeric binder or can be chemically generated via a catalyst to form a rigid binder.12 Currently several binders have been used including epoxy resin13,14 sol-gel-derived ceramics,15 methacrylate,16 silicone,17 polyester or polyurethane,18 and Teflon.19 These composite biomaterials present an improved mechanical stability of the paste as well as the intrinsic advantage of surface regeneration of classical carbon pastes.19 However, the fabrication of these materials involves chemical reactions or heating processes that can deactivate the biological activity of the enzyme or denature the protein. In this context, we report here a new approach for the soft fabrication of rigid biocomposites based on the in situ electrochemical generation of a polymer network. The rigidifying process results from the electrochemical polymerization of pyrrole monomer previously incorporated within the paste. This new carbon paste hardening concept has been applied to the entrapment of glucose oxidase (GOx) as an enzyme model within the biocomposite structure to assess the efficiency of the immobilization process and to investigate the influence of an immobilized protein on the physicochemical properties of the composite. Then, tyrosinase, a polyphenol oxidase enzyme (PPO), was incorporated within the paste for the design of a phenolic biosensor focused on the determination of phenolic derivatives. EXPERIMENTAL SECTION Chemicals. All chemicals were of the highest analytical grade and were used as received. Aqueous solutions were prepared in deionized water prepared by passing distilled water through an Elga Prima/Maxima purification system. Carbon pastes were made with graphite powder (BDH) and Nujol (BDH). Pyrrole monomer and ferrocenecarboxylic acid were purchased from Aldrich. Glucose oxidase (EC 1.1.3.4, typeVII, from Aspergillus niger, 168 Sigma units mg-1), polyphenol oxidase (EC1.14.18.1, from mushroom, 123 units mg-1), phenol, catechol, ferulic acid, and (-)-epicatechin were purchased from Sigma. All buffers were prepared from potassium monobasic and dibasic phosphate salts (Aldrich). All chemicals were of analytical grade and were used without further purification. Apparatus. Electrochemical characterization and analytical testing were carried out using a carbon paste electrode as the working electrode and a large platinum foil electrode (Metrohm) as the counter electrode. All potentials were measured and reported with respect to a saturated calomel electrode (SCE). All electrochemical measurements were made in a one-compartment electrochemical cell. A water jacket, thermostatically controlled (11) Pravda, M.; Petit, C.; Michotte, Y.; Kauffmann, J.-M.; Vytras, K. Belgium J. Chromatogr., A 1996, 727, 47-54. (12) Cespedes, F.; Alegret, S. Food Technol. Biotechnol. 1996, 34, 143-146. (13) Falat, L.; Chen, H. Y. Anal. Chem. 1982, 54, 2108-2111. (14) Wang, J.; Fang, L.; Lopez, D. Analyst 1994, 119, 455-458 and reference therein. (15) Tsionsky, M.; Gun, G.; Glezer, V.; Lev, O. Anal. Chem. 1994, 66, 17471753. (16) Santandreu, M.; Cespedes, F.; Alegret, S.; Martines-Fabregas, E. Anal. Chem. 1997, 69 2080-2085. (17) Connor, M. P.; Sanchez, J.; Wang, J.; Smith, M. R.; Mannino, S. Analyst 1989, 114, 1427-1429. (18) Alegret, S.; Alonso, J.; Bartroli, J.; Cespedes, F.; Martines-Fabregas, E.; Del Valle, M. Sens. Mater. 1996, 8, 147-153. (19) Serra, B.; Riveiro, A. J.; Parrado, C.; Pingarron, J. M. Biosens. Bioelectron. 1999, 14, 503-513.

(thermostat liquitherm F) at 25 ( 0.2 °C, was used for the characterization and analytical evaluation of the biosensor. All other electrochemical measurements were carried out at room temperature. An Autolab PGStat 20 Ecochemie was used for ac impedance and other electrochemical measurements. Impedance spectra were fitted using the software EQUIVCRT based on the work of Boukamp.20 Electrochemical equipment consisting of a PAR model 173 potentiostat equipped with a model 179 digital coulometer and a model 175 programmer in conjunction with a Sefram TRP recorder was also used for the in situ electrochemical polymerization of pyrrole. The electropolymerization processes were performed in a 5-mL three-electrode cell. Methods. Preparation of the Carbon Paste Electrodes. The carbon pastes employed for the fabrication of CPE, polypyrrole CPE, and biocomposite CPE were prepared as follow: (i) Carbon paste was prepared by mixing graphite powder with Nujol in a mortar until it was uniformly wetted using a graphite/ Nujol ratio of 4/1 w/w. This ratio was employed as it provides convenient analytical properties.21 (ii) Pyrrole monomer was incorporated in the carbon paste using a pyrrole ratio of 10% w/w. (iii) The enzymes (GOx or PPO) were mixed with the pure and pyrrole-modified carbon pastes with ratios of protein ranging from 1 to 6% w/w. (iv) After blending, the pastes were packed into a Teflon electrode holder (geometric surface area 0.07 cm2) with electrical contact made via a vitreous carbon rod. (v) The electrode surface was smoothed on a paper to produce a reproducible working surface. Electropolymerization Process. Polypyrrole CPE electrodes were prepared by electropolymerization of a polymer membrane on the aforementioned pyrrole-free carbon paste enzyme mixture (iii). The electrosynthesis was performed in a solution containing LiClO4 (0.1 M) and pyrrole (25 mM) by cycling the potential seven times between -0.3 and 0.9 V/SCE at a scan rate of 50 mV s-1. Biocomposite electrodes were fabricated by in situ electropolymerization of the pyrrole monomer mixed within the paste (iii) in a LiClO4 (0.1 M) solution. The electropolymerization process was performed either by cycling the potential between -0.3 and 0.9 V/SCE at a scan rate of 50 mV s-1 or by applying a fixed potential of 0.77 V/SCE to zero current. ac Impedance. Impedance characterization were carried out in a 1 mM ferrocenecarboxylic acid potassium phosphate buffer solution (0.05 M, pH 7.4). A frequency range of 65 000-0.1 Hz was utilized with a potential amplitude of 5 mV rms. A bias of 0.3 V/SCE was used in order to study the electrode reaction at the half-wave potential of the oxidation of ferrocenecarboxylic acid. Scanning Electron Microscopy (SEM). Microstructure characterization of the carbon paste electrode surface was performed using a Hitachi S-3200N low-vacuum scanning electron microscope. A working distance of 8 mm was used with 20.0 kV at a magnification of 1200×. Determination of the Amount of Enzyme Released under Storage. The three enzyme electrodes (CPE, PPy-CPE, and biocomposite (20) Boukamp, B. A. EQUIVCRT, University of Twente, Department of Chemical Technology, P.O. Box 217, 7500AE Enschede, The Netherlands. (21) Cummings, E. A.; Mailley, P.; Linquette-Mailley, S.; Eggins, B. R.; McAdams, E. T.; McFadden, S. Analyst 1998, 123, 1975-1980.

Analytical Chemistry, Vol. 75, No. 20, October 15, 2003

5423

Figure 1. In situ electropolymerization of the pyrrole monomer mixed with carbon paste. The electropolymerization process is carried out by cycling the potential between -0.3 and 0.8 V/SCE in 0.1 M LClO4 aqueous solution (sweep rate 50 mV s-1).

CPE) were stored in a 0.1 M phosphate buffer (pH 7) solution for 72 h. The quantity of enzyme released from the paste in the storage solution was estimated periodically through the enzymatic activity of this solution measured by spectrophotometry as previously described.22 The release experiments were carried out with GOx as an enzyme model because of its robustness. Amperometric Analysis. The amperometric measurements of PPO substrates were performed in stirred air-saturated phosphate buffer, 25 cm3 (0.05 M, pH 7.4), by holding the carbon paste electrodes at -0.2 V. Various phenolic compounds from 10-3 to 1 M stock solutions were added to this stirred solution, and the current corresponding to the reduction of the enzymatically generated o-quinones was recorded. RESULTS AND DISCUSSION In Situ Electropolymerization of PPy within Carbon Paste. The generation of polymeric structures at CPE interfaces was already carried out for the design of membranes enabling substrate diffusion control23 or electrooxidizable interference rejection.9,24 However, to our knowledge, in situ electropolymerization of PPy, or another electropolymerizable monomer, within CPE structures was not yet performed. Since pyrrole is soluble in the hydrophobic pasting liquid, the electropolymerization properties of pyrrole molecules previously mixed with the carbon paste was examined by cyclic voltammetry. Electropolymerization of incorporated pyrrole was attempted by repeatedly scanning the potential between -0.2 and 0.85 V/SCE (Figure 1). The appearance of an oxidation wave due to monomer oxidation is followed by the appearance and growth of a quasi-reversible peak system corresponding to the polypyrrole oxidation and reduction at 0.4 and -0.1 V, respectively. This clearly indicates the formation of an electropolymerized polymeric film within the carbon paste structure. The PPy redox system exhibits a lower reversibility than those obtained for films electropolymerized at a platinum or (22) Coche-Gue´rente, L.; Cosnier, S.; Innocent, C.; Mailley, P. Anal. Chim. Acta 1995, 311, 23-30. (23) Zahir, M. D.; Ab Ghani, S. Anal. Chim. Acta 1997, 354, 351-358. (24) Lobo-Castanon, M. J.; Miranda-Orderies, A. J.; Tunon-Blanco, P. Anal. Chim. Acta 1997, 346, 165-174.

5424 Analytical Chemistry, Vol. 75, No. 20, October 15, 2003

glassy carbon electrode. Such behavior could be related to the hydrophobic nature of the pasting liquid, which acts as a diffusion barrier for the doping cations. Electropolymerization of incorporated pyrrole was also carried out by controlled-potential oxidation. After optimization of the polymerization process, the applied potential was fixed at 0.77 V, allowing the generation of the polypyrrole network with an electrical yield of 15%. The current efficiency of the electropolymerization process was determined by dividing the charge recorded under the oxidation wave of the polymerized polypyrrole by the charge passed during the controlled-potential electrolysis and multiplying by 7. Indeed, the electropolymerization process requieres 2.33 electrons per pyrrole monomer. The integration of the charge under the polypyrrole activity was estimated equal to 3.8 mC. This charge corresponds to the oxidation of a polymeric structure containing 120 nmol of pyrrole units (for a doping factor of 0.33). Different pyrrole/carbon paste ratios from 2 to 25% in weight of monomer were tested. The 10% composition corresponded to the optimum ratio in terms of reproducibility of the electropolymerization process and quality of the resulting composite paste and was thereby utilized in the further experiments. Electrochemical, Structural, and Biochemical Characterization of Enzyme-Containing Carbon Pastes. With the aim to fabricate amperometric biosensors, the possibility of immobilizing an enzyme in composite carbon paste was investigated. For this purpose, GOx was chosen as the enzyme model. Moreover, various carbon paste configurations were tested to highlight the role of the in situ-electrogenerated polypyrrole on the properties of the resulting biocomposite. Three different carbon pastes were tested to compare their electrochemical, structural, and biochemical behavior: polypyrrole-free carbon paste containing GOx (socalled CPE), carbon paste containing GOx coated with an ex situgenerated PPy membrane obtained by cyclic voltammetry in a monomer aqueous solution (so-called PPy-CPE), and carbon paste containing GOx and in situ-generated PPy (so-called biocomposite CPE). We were able to verify that the in situ-electrogenerated polypyrrole improved enzyme retention within the paste and that it was electropolymerized in the volume of the CPE and not at the top of the CPE surface. To assess the structure and the electrochemical behavior of each enzyme electrode, their characterization was realized by ac impedance spectroscopy and SEM. Finally, the enzyme leaching out of the biosensor was estimated for each paste in order to determine the retention capabilities of the different structures. Analysis of the impedance data was performed using a commercial program from Boukamp20 based on the equivalent circuit model where the total series resistance of the system (Rs) is in series with a parallel combination of a resistance (RCT) and a constant-phase angle impedance (ZCPA). The resistance, RCT, is thought to represent charge-transfer resistance across the interface. Due to the irregular structure and surface roughness of carbon paste electrodes, an empirical constant-phase element (ZCPA) K(Jω)-β) was used to represent the double-layer capacitance,2,25 where K is a measure of the magnitude of the nonfaradic impedance and the exponent, β, reflects the extent of deviation (25) Zhou, D. M.; Jones, J. G.; McAdams, E. T.; Lackeermeir, A. Innovation Technol. Biol. Me´ d. 1995, 16, 77-86.

Figure 2. ac impedance spectra of GOx-based CPEs: (O) CPE, (9) PPy-CPE, and (2) biocomposite CPE obtained for CPEs containing 6% w/w enzyme. The impedance spectra were recorded within the frequency range 65 kHz-0.1 Hz at a potential bias of 0.3 V/SCE in 10-3 M ferrocene buffered solution. Table 1. Electrical Components of CPE, PPy-CPE, and Biocomposite CPE Interfaces Extracted from the Described Equivalent Circuit Model (RS in Series with a Parallel Combination of RCT and ZCPAa) electrode

RS, Ω

RCT, kΩ

K, µF

β

CPE PPy-CPE biocomposite CPE

395 245 314

14.4 11.5 6.7

2.7 31 22

0.69 0.73 0.62

a

ZCPA) K(jω)-β.

from an ideal capacitive behavior. Figure 2 shows the impedance results for the three-enzyme electrode configurations. For all fabricated electrodes, the impedance plots had the form of a “depressed” semicircle over a wide frequency range. At higher frequencies, however, a deviation from this behavior was observed. Such a “distortion” of the impedance locus at the highfrequency domain, which was observed for all the configurations, has been attributed to surface roughness effects.26 For the CPE, a clearly defined semicircle was evident at the high-frequency domain. However, in the case of PPy-CPE and biocomposite CPE, the semicircle was less defined due to the decrease in time constants between the two equivalent circuits, thus indicating a “second” process at the electrode surface. Nevertheless, the results were extracted from the impedance data at the low frequency arc using the equivalent circuit model described above. The results indicate a decrease in charge-transfer resistance from CPE (14.4 kΩ) to biocomposite CPE (6.65 kΩ) (Table 1). Furthermore, an increase in K was observed from bare to biocomposite CPEs, 2.7 and 22 µF, respectively, with β indicating a mixture of Warburg-like diffusion and capacitive behavior. In addition, compared to a CPE, the phase angle increased and decreased for PPy-CPE and biocomposite CPE, respectively. The impedance results would suggest that the presence of the conducting polymer has improved the electrical properties of the fabricated electrodes, since a decrease in charge-transfer resist(26) De Levie, R. In Advancements in Electrochemistry and Electrochemical Engineering; Delahay, P., Tobias, C. W., Eds.; Interscience: New York, 1967; Vol. 6, p 329.

ance was noted. Such a decrease in charge-transfer resistance can be related to an improvement in electrical connection between the graphite grains and an enhancement in electroactive surface area due to the presence of the polymer. Furthermore, an increase in phase angle was observed for PPy-CPE. Such behavior has been interpreted by De Levie26 in terms of a decreased surface roughness. It seemed thus that the presence of the polymer on the electrode surface has decreased the surface roughness compared to a CPE. For biocomposite CPE, however, the phase angle decreased. This decrease in phase angle can be related to the presence of the polymer throughout the paste, which provides a conducting pathway between the graphite grains. However, unlike a PPy-CPE where the polymer is solely present at the electrode/electrolyte interface, the in situ-formed polymer does not influence significantly the electrode surface roughness compared to a CPE. Consequently, a slight change in phase angle was observed (Figure 2). This hypothesis is corroborated by the comparison of SEM morphologies of the different enzyme electrode configurations (Figure 3). The surface topography for a CPE (Figure 3A) shows a granular surface with areas of high charge-up. This charge-up decreases for PPy-CPE (Figure 3B) and is virtually absent for the biocomposite CPE (Figure 3C), which exhibits a homogeneous surface. As charge-up can be related to the insulating properties of the sample under observation,27 the resistance of each sample can be qualified and compared if the experimental conditions are consistent between each sample viewed. Thus, a CPE (Figure 3A) is more resistive compared to a biocomposite CPE (Figure 3C), which is in agreement with impedance measurements As demonstrated by ac impedance and SEM images, PPy was electropolymerized within the paste in biocomposite CPE structures. To characterize the impact of the in situ-electrogenerated polypyrrole on the retention properties of the biocomposite CPE toward entrapped proteins, the phenomenon of enzyme release was investigated for the three configurations with GOx. The enzyme leakage in solution was estimated using a spectrometric method22 for four different GOx loadings, namely, 1, 2, 4, and 6% w/w. Since the addition of a high amount of biomolecules within the paste may alter the mechanical cohesion of the resulting biomaterial due to the hydrophilicity of the biomolecule,21 the enzyme ratio was limited to 6% w/w. For instance, Figure 4 shows the evolution of the loss of GOx molecules with time for the three configurations (enzyme ratio 6% w/w) soaked in a phosphate buffer solution. It clearly appears that the quantities of GOx released in solution for the unmodified carbon paste are markedly higher than those determined for the carbon pastes modified by polypyrrole. Effectively, the amount of released enzyme reached 45 µg after 2 h whereas 3.5 µg and no leakage were recorded for PPy-CPE and biocomposite CPE, respectively. Due to the large enzyme release recorded for CPE, longer release experiments were restricted to PPy-CPE and biocomposite CPE. For longer soaking duration, the amount of GOx lost in solution by the biocomposite CPE remained extremely low, reaching only 0.1 µg after 72 h soaking, whereas the PPy-CPE configuration leads to 150 µg. Such result highlights the improvement in the enzyme retention brought by the in situ-electropolymerized polypyrrole (27) Newbury, D. E.; Joy, D. C.; Echlin, P.; Fiori, C. E.; Goldstein, J. I. Advanced Scanning Electron Microscopy and X-Ray Microanalysis, 2nd ed.; Plenum Press: New York, 1986; p 47.

Analytical Chemistry, Vol. 75, No. 20, October 15, 2003

5425

Figure 3. SEM micrographies of CPE (A), PPy-CPE (B), and biocomposite CPE (C) containing 6% w/w GOx.

Figure 4. Estimated quantities of GOx released in phosphate buffer versus time for the CPE (A), PPy-CPE (B), and biocomposite CPE (C). The experiments were performed at room temperature in 3 mL of 0.1 M phosphate buffer solution (pH 7).

compared to PPy film generated at the electrode surface. This correlates the electrochemical and structural characterizations that demonstrated the presence of electrogenerated PPy in the bulk of the carbon paste material. In situ-electropolymerized polypyrrole seems to act as a polymeric binder that reinforces the stability of the carbon paste material. Since the biocomposite CPE configuration provides the most efficient enzyme retention, the enzymatic activity of such biomaterial, which reflects its ability for biosensing applications, has been examined for different enzyme ratios of 1, 2, 4, and 6% w/w. Since the immobilized GOx molecules catalyze, in the presence of molecular oxygen, the oxidation of glucose with the production of H2O2, the enzymatic activity of the modified electrodes was evaluated by soaking these electrodes in phosphate buffer containing 50 mM glucose and by following spectrometrically the time-dependent increase in H2O2 concentration. As expected, the enzymatic activity increases with the enzyme loading from 0.14 to 1.82 units. In addition, the storage stability of the biocomposite CPE based on an enzyme ratio of 6% and stored dry at -4 °C was investigated. The enzymatic activity of the biocomposite remained stable for 3 days without any loss. An activity lower than 20% was recorded after 30 days of storage. Finally, the aforementioned characterizations have demonstrated the effectiveness of this new concept of carbon paste 5426

Analytical Chemistry, Vol. 75, No. 20, October 15, 2003

modification. The obtained biomaterials exhibited improved electrical and mechanical properties with the highest enzymatic activity being obtained for an enzyme loading of 6% w/w. This enzyme ratio was employed for the fabrication of PPO-based carbon paste sensors dedicated to the detection of flavonols in beers. Design of a PPO-Based Sensor for Flavonols. PPO catalyses, in the presence of oxygen, the oxidation of phenol, catechol, and their derivatives such as polyphenols and flavonols. To fabricate a biosensor for the detection of flavonols, we previously designed plant tissue carbon paste electrodes based on apple tissues containing PPO.21 Due to the weak enzyme activity of apple tissues, these biosensors exhibited a rather low sensitivity and a poor selectivity toward flavonol. To increase the sensitivity of the biosensor, we have used here the isolated PPO enzyme. We have built up the CPE, PPy-CPE, and biocomposite CPE containing PPO in the ratio of 6% w/w. The analytical properties of these different biosensor configurations were compared for a model substrate of PPO. Then, the steady-state responses of the different biosensors to catechol were recorded in stirred air-saturated 0.1 M phosphate buffer (pH 7) by holding the sensor at a potential of -0.2 V/SCE in order to reduce the enzymatically generated quinones. Figure 5 presents the calibration curves for catechol obtained at the three CPEs as well as the steady-state currenttime response of the biocomposite CPE to successive injection of catechol (Figure 5D). Table 2 summarizes the different analytical characteristics extracted from the aforementioned curves (sensitivity, maximum current density Jmax, and apparent MichaelisMenten constant Kapp M ) for the three bioelectrodes. The sensitivity of the biocomposite CPE to catechol was 1.6 and 5.2 times higher than at the CPE and PPy-CPE configurations, respectively. Such behavior may be related to a better accessibility to the enzymatic active sites within the biocomposite structure or a higher density of these sites at the electrode surface. On the other hand, Jmax increased markedly from PPy-CPE (31 µA cm-2) to biocomposite CPE (103 µA cm-2) and CPE (129 µA cm-2). It should be noted that Jmax is directly related to the available quantity of immobilized enzymes and corresponds to the saturation of the active sites. These results seem to indicate that the ex situgenerated polypyrrole coating acted here as a membrane, contrary to the in situ-electrogenerated polymer, and blocked partially the

Figure 5. Calibration curves for catechol obtained at CPE (A), PPyCPE (B), and Biocomposite CPE (C). Steady-state current-time response of the biocomposite CPE (D) for increasing catechol concentrations in 5 × 10-7 M steps. Amperometric responses recorded at -0.2 V/SCE in stirred air-saturated 0.1 M phosphate buffer at 25 °C with CPEs containing 6% w/w PPO.

Figure 6. Calibration curve for catechol (A) and phenol (B) obtained at the biocomposite CPE containing 6% w/w PPO. Experimental conditions as in Figure 5. Table 3. Sensitivitya of the Biocomposite CPE Containing 6 % w/w PPO to Catechol, Phenol, (-)-Epicatechin, and Ferulic Acid

Characteristicsa

Table 2. Analytical toward Catechol Detection of CPE, PPy-CPE, and Biocomposite CPE Containing 6% w/w PPO electrode

sensitivity, A M-1 cm-2

Jmax, µA cm-2

b Kapp M , µM

CPE PPy-CPE biocomposite CPE

2.91 0.92 4.75

129 31 103

26 18 11

a Experimental conditions as in Figure 5. b Estimated by considering the catechol concentration at which the biosensor response is half of the saturation value Jmax.

catechol diffusion. Such behavior correlated with ac impedance characterization, proving that in situ PPy was generated evenly in the carbon paste and acted as a polymeric binder. The slight increase in Jmax values from biocomposite CPE to CPE may be attributed to a more eroded surface due to the gradual degradation of the unmodified paste and hence to a more important active surface. Finally, the recorded Kapp M were markedly lower than such obtained for the free enzyme.28 Such behavior is classically observed in the case of PPO-based biosensors and is related to an amplification phenomenon due to the electrochemical regeneration of the enzyme substrate.29-31 It should be noted that the lower value determined for the biocomposite CPE (11 µM) reflects a more efficient recycling phenomenon that may be due to a higher enzyme density at the electrode surface.31 It appears, in Figure 5D, that the amperometric biosensor response of the obtained biosensor was rapid since 95% of the steady-state deviation was reached within 20 s. These increments in catechol concentration were within the concentration range corresponding to the linear part of the calibration curve and hence illustrate the repeatability of the recorded signal (RSD ) 2.1%). (28) Barman, T. E., Ed. Enzyme handbook; Springer-Verlag: New York, 1985; Vol. 1, p 226. (29) Cosnier, S. Innocent, C. Bioelectrochem. Bioenerg. 1993, 31, 147-160. (30) Cosnier, S.; Gondran, C.; Watelet, J.-M.; De Giovani, W.; Furriel, R. P. M.; Leone, F. A. Anal. Chem. 1998, 70, 3952-3956. (31) Wang, B.; Dong, S. J. Electroanal. Chem. 2000, 487, 45-50.

a

substrate

sensitivity, A M-1 cm-2

catechol phenol (-)-epicatechin ferulic acid

4.75 3.13 0.63 0.07

Experimental conditions as in Figure 5.

The biocomposite CPE exhibited a good operational stability since its catechol sensitivity remained constant after five consecutive assays (including stabilization, sensing, and washing process). In contrast, the catechol sensitivity at the CPE decreased strongly to less than half of its initial value, highlighting the protecting effect of the in situ generated PPy. The reproducibility of the biosensor preparation was also determined. For a same batch of paste and for five different electrodes, the average sensitivity was 4.7 A M-1 cm-2 (RSD 7.6%). Thereby, although the handmade preparation does not allow perfect control of the surface roughness of the resulting carbon paste, the biocomposite electrodes were easy to fabricate and rather reproducible for a same batch of paste. As previously mentioned, PPO oxidizes specifically phenols and other o-phenol derivatives with catalytic rates depending on its substrate specificity. For instance, the oxidation rate of phenol was ∼40% lower than that of catechol for the free enzyme. Figure 6 shows the amperometric response of a biocomposite CPE as a function of catechol and phenol concentrations. The resulting calibration curves exhibit the same shape for phenol analysis as for catechol detection. The analytical characteristics of the biosensor, in terms of Jmax and sensitivity, are lowered for phenol compared to catechol from approximately 38 and 34%, respectively. Such behavior shows that the immobilized enzyme retained its substrate selectivity when entrapped within a biocomposite CPE. Finally, the response of the biocomposite CPE to various phenolic compounds was investigated. More particularly, flavonols are of interest for the brewing industry since they are responsible for cloudiness in beers32 and are good indicators of beer quality. (32) Gardner, R. J.; McGuiness, J. D. Technol. Q. Master Brew. Assoc. Am. 1977, 14, 250-261.

Analytical Chemistry, Vol. 75, No. 20, October 15, 2003

5427

Flavonols originate, as other phenolics, from the raw material of brewing. It becomes important to evaluate the level of flavonol with the best accuracy as possible regarding other phenolics. We have tested the response of the biocomposite sensor to (-)epicatechin, as the flavonol model, and to ferulic acid as the polyphenol model. Table 3 displays sensitivity to these two substrates as well as those previously recorded for catechol and phenol. As reported in the literature,33 the substitution on the phenolic ring markedly reduced the enzyme activity and hence the biosensor performances for ferulic acid and epicatechin determination. The biosensor, nevertheless, exhibited low detections limit and a short response time for all substrates with, in particular, 15 and 200 nM as detection limits for epicatechin and ferulic acid, respectively. It clearly appears that the biocomposite CPE exhibits higher sensitivity toward (-)-epicatechin than toward ferulic acid with a 9 times lower answer for the polyphenol model than for the flavonol one. Further work is currently underway to assess the real selectivity of the biocomposite sensor toward flavonols in real beer samples. (33) Wilcox, D. E.; Porras, A. G.; Huang, Y. T.; Lerch, K.; Wilker, M. E.; Solomon, E. L. J. Am. Chem. Soc. 1985, 107, 4015-4027.

5428

Analytical Chemistry, Vol. 75, No. 20, October 15, 2003

CONCLUSION In this report, we have described glucose oxidase and polyphenol oxidase carbon paste electrodes prepared via a new strategy of carbon paste modification based on the in situ electropolymerizaton of pyrrole monomer previously mixed within the paste. Such modification induced a better electrical percolation of the carbon structure and improved to a large extent the enzyme entrapment within the electrode material. We have, in particular, illustrated the attractive potentialities offered by a biocomposite electrode based on PPO for the detection of flavonols. Works dedicated to the control of beer samples are now under investigation. ACKNOWLEDGMENT The authors kindly acknowledge the financial assistance from Guinness Ireland Group, Dublin, and the Department of Higher and Further Education, Training and Employment (DHFETE) Northern Ireland. Received for review February 21, 2003. Accepted August 6, 2003. AC034177Y