Anal. Chem. 2004, 76, 5045-5050
Carbon Nanotube-Chitosan System for Electrochemical Sensing Based on Dehydrogenase Enzymes Maogen Zhang,†,‡ Audrey Smith,† and Waldemar Gorski*,†
Department of Chemistry, University of Texas at San Antonio, San Antonio, Texas 78249-0698, and Department of Chemistry, Nanjing Normal University, Nanjing 210097, P.R. China
Multiwalled carbon nanotubes (CNT) were solubilized in aqueous solutions of a biopolymer chitosan (CHIT). The CHIT-induced solubilization of CNT facilitated their manipulations, including the modification of electrode surfaces for sensor and biosensor development. The colloidal solutions of CNT-CHIT were placed on the surface of glassy carbon (GC) electrodes to form robust CNT-CHIT films, which facilitated the electrooxidation of NADH. The GC/CNT-CHIT sensor for NADH required ∼0.3 V less overpotential than the GC electrode. The susceptibility of CHIT to chemical modifications was explored in order to covalently immobilize glucose dehydrogenase (GDH) in the CNT-CHIT films using glutaric dialdehyde (GDI). The stability and sensitivity of the GC/ CNT-CHIT-GDI-GDH biosensor allowed for the interference-free determination of glucose in the physiological matrix (urine). In pH 7.40 phosphate buffer solutions, linear least-squares calibration plots over the range 5-300 µM glucose (10 points) had slopes 80 mA M-1 cm-2 and a correlation coefficient 0.996. The detection limit was 3 µM glucose (S/N ) 3). The CNT-CHIT system represents a simple and functional approach to the integration of dehydrogenases and electrodes, which can provide analytical access to a large group of enzymes for wide range of bioelectrochemical applications including biosensors and biofuel cells. Carbon nanotubes1 (CNT) are relatively new nanomaterials that display attractive structural, mechanical, and electronic properties, including improved electrochemical activity.2 Chitosan (CHIT) is a polysaccharide biopolymer (Figure 1), which displays excellent film-forming ability, high water permeability, good adhesion, and susceptibility to chemical modifications due to the presence of reactive amino and hydroxyl functional groups. It was for these reasons that we decided to explore the possibility of using the CNT-CHIT system as a platform for the development of electrochemical sensors and biosensors. * Corresponding author: (fax) 210-458-7428; (e-mail)
[email protected]. † University of Texas at San Antonio. ‡ Nanjing Normal University. (1) Iijima, S. Nature 1991, 354, 56-58. (2) Sherigara, B. S.; Kutner, W.; D’Souza, F. Electroanalysis 2003, 15, 753772. 10.1021/ac049519u CCC: $27.50 Published on Web 07/28/2004
© 2004 American Chemical Society
Figure 1. Chemical structure of chitosan.
The present paper discusses three aspects of the CNT-CHIT system. They relate to the solubilization of pristine CNT in aqueous solutions of CHIT, the development of new NADH electrodes based on CNT-CHIT films, and the integration of dehydrogenase enzymes within CNT-CHIT films for the development of electrochemical biosensors. In the first part of this paper, a biocompatible solubilization of nanotubes is presented, which is based on noncovalent association of CNT with chains of CHIT in aqueous solutions. The inherent insolubility of pristine CNT in solvents, in particular in water, has complicated their handling as chemical reagents and has limited their integration with biological systems. The challenge of solubilizing CNT in water has been addressed through their covalent modification with hydrophilic functionalities3,4 or noncovalent modification with surfactants5 or polymers.6 Of the three approaches, covalent modification has the disadvantage that it can impair the physical properties of CNT, while the use of surfactants has a potential to denature biological molecules. Such problems are minimized in the case of polymer-based solubilization. Thus far, however, the number of polymers that render CNT soluble in aqueous solutions, without the necessity for covalent modification, has been limited.6-9 (3) Sun, Y.-P.; Huang, W.; Lin, Y.; Fu, K.; Kitaygorodskiy, A.; Riddle, L. A.; Yu, Y. J.; Carroll, D. L. Chem. Mater. 2001, 13, 2864-2869. (4) Peng, H.; Alemany, L. B.; Margrave, J. L.; Khabashesku, V. N. J. Am. Chem. Soc. 2003, 125, 15174-15182. (5) Bandow, S.; Rao, A. M.; Williams, K. A.; Thess, A.; Smalley, R. E.; Eklund, P. C. J. Phys. Chem. B 1997, 101, 8839-8842. (6) O’Connell, M. J.; Boul, P.; Ericson, L. M.; Huffman, C.; Wang, Y.; Haroz, E.; Kuper, C.; Tour, J.; Ausman, K. D.; Smalley, R. E. Chem. Phys. Lett. 2001, 342, 265-271. (7) Bandyopadhaya, R.; Nativ-Roth, E.; Regev, O.; Yerushalmi-Rozen, R. Nano Lett. 2002, 2, 25-28. (8) Star, A.; Steuerman, D. W.; Heath, J. R.; Stoddart, J. F. Angew. Chem., Int. Ed. 2002, 41, 2508-2512.
Analytical Chemistry, Vol. 76, No. 17, September 1, 2004 5045
In the second part, a reliable determination of NADH at the CNT-CHIT film electrodes is presented. The NADH is a reduced form of the nicotinamide adenine dinucleotide (NAD+), which is a cofactor for a large number of dehydrogenase enzymes (>300). Thus, the significance of NADH detection at CNT-CHIT films is that such films can provide a signal transduction in dehydrogenase-based electrochemical biosensors for a number of substrates (analytes). The oxidation of NADH at the CNT-based electrodes is a recent development.10-13 The third part of this paper demonstrates the capability of the CNT-CHIT films for immobilization of dehydrogenase enzymes for electrochemical biosensing. The enzyme immobilization relied on tethering the amino groups of both CHIT and enzyme with glutaric dialdehyde. A model enzyme glucose dehydrogenase was attached to the CNT-CHIT film and used for the interferencefree determination of glucose in urine matrix. To the best of our knowledge, this is the first demonstration of an electrochemical biosensor based on the integration of a dehydrogenase enzyme with carbon nanotubes using a biopolymeric modifiable scaffold. Recently, a biosensor utilizing alcohol dehydrogenase, carbon nanotubes, and Teflon as a binder has been reported.12 Such a composite electrode material has allowed for a relatively fast (∼60 s) detection of ethanol at low potentials (0.20 V) without the need for redox mediators. Materials and methods for the integration of dehydrogenases and electrodes are of particular interest because they can provide access to a large group of enzymes for biosensor and biofuel cell development. EXPERIMENTAL SECTION Reagents. Multiwalled CNTs (∼95% nominal purity) were purchased from Nanolab (Brighton, MA). CHIT (MW ∼ 1 × 106; ∼80% deacetylation), NADH, NAD+, glucose dehydrogenase (GDH, from Pseudomonas sp., EC 1.1.1.47), and glutaric dialdehyde (GDI) were purchased from Sigma-Aldrich. Other chemicals, NaH2PO4‚H2O, Na2HPO4, HCl, and NaOH, were from Fisher. All solutions were prepared using deionized water that was purified with a Barnstead NANOpure cartridge system. Electrochemical Measurements. A CHI 832 workstation (CH Instruments, Inc.) was used to collect electrochemical data. Experiments were performed at room temperature (20 ( 1 °C) in a conventional three-electrode system with a 3.0-mm-diameter glassy carbon disk working electrode (Bioanalytical Systems, Inc.), a platinum wire as the auxiliary electrode, and a Ag/AgCl/3MNaCl (BAS) reference electrode. Prior to use, the glassy carbon electrodes were wet polished on an Alpha A polishing cloth (Mark V Lab) with successively smaller particles (0.3- and 0.05-mm diameter) of alumina. The slurry that accumulated on the electrode surface was removed by ultrasonication for 30 s in deionized water and methanol. The pH 7.40 phosphate buffer solution (0.05 M) served as a background electrolyte in all experiments. The experiments were repeated at least three times, and the means of measurements (9) Wang, J.; Musameh, M.; Lin, Y. J. Am. Chem. Soc. 2003, 125, 2408-2409. (10) Musameh, M.; Wang, J.; Merkoci, A.; Lin, Y. Electrochem. Commun. 2002, 4, 743-746. (11) Wang, J.; Deo, R. P.; Poulin, P.; Mangey, M. J. Am. Chem. Soc. 2003, 125, 14706-14707. (12) Wang, J.; Musameh, M. Anal. Chem. 2003, 75, 2075-2079. (13) Liu, P.; Yuan, Z.; Hu, J.; Lu, J. Proc.-Electrochem. Soc. 2003, 13, 346-356.
5046
Analytical Chemistry, Vol. 76, No. 17, September 1, 2004
are presented with the standard deviations or relative standard deviations. Solutions of Chitosan and Carbon Nanotubes. Chitosan is soluble in acidic aqueous solutions in which it behaves as a cationic polyelectrolyte. At pH >6, dissolved chitosan flocculated due to the deprotonation of its amine groups. A 0.50 wt % CHIT stock solution was prepared by dissolving chitosan flakes in hot (80-90 °C) aqueous solution of 0.05 M HCl. The solution was cooled to room temperature, and its pH was adjusted to 3.5-5.0 using a concentrated NaOH solution. The chitosan solutions were filtered using a 0.45-µm Millex-HA syringe filter unit (Millipore) and stored in a refrigerator (4 °C) when not in use. All chitosan solutions were colorless. The carbon nanotubes were solubilized in chitosan solutions (0.50-3.0 mg of CNT mL-1) using a short 15-min sonication. Film Electrodes. The CNT-CHIT film electrodes were prepared by casting 30.0 µL of CNT-CHIT solution (0.50 wt % CHIT, 0.50 mg of CNT mL-1, pH 3.5) on the surface of a glassy carbon (GC) electrode. Evaporation of water from such a solution (1-2 h, room temperature) yielded robust surface films that contained ∼10 and 90 wt % CNT and CHIT, respectively. The enzyme films were synthesized by tethering the GDH to CHIT with GDI following our recent protocol.14 Briefly, the GDI and CHIT solutions were reacted first. A high molar ratio of GDI to CHIT glucosamine units (200:1) was used in order to react only one aldehyde group of GDI with amino groups of CHIT. The unreacted excess of GDI was removed from the mixture by multiple extractions with ethyl ether. The enzyme immobilization was accomplished by reacting free aldehyde groups of CHITGDI product with amino groups of GDH. The optimization studies showed that the most reliable CNT-CHIT-GDI-GDH enzyme films for glucose determination were formed by mixing 30.0 µL of 0.10 wt % CHIT-GDI solution (pH 5.0) that contained CNT (0.50 mg mL-1) with 20.0 µL of GDH solution (1.0 unit µL-1) on the surface of the glassy carbon electrode and evaporating water for 2 h. All film electrodes were stored in air at room temperature when not used, unless otherwise stated. Thermogravimetric and Spectroscopic Analysis. The thermogravimetric analysis (TGA) was performed with a TGA-50 thermogravimetric analyzer (Shimadzu). The experiments were carried out in an atmosphere of flowing air (7 mL min-1) at a heating rate of 10 °C min-1. The solid samples of CHIT and CNTCHIT were prepared by evaporating water from solutions of CHIT and CNT-CHIT (3.0 mg of CNT mL-1), respectively. The solids were washed with water in order to remove NaCl salt that was formed in CHIT solutions during the pH adjustment. The infrared spectra were collected with a Bruker Equinox 55 spectrometer. Chitosan-containing films were cast on polyethylene cards type 61 (3M), dried at room temperature, washed with water, and redried before analysis. RESULTS AND DISCUSSION Characterization of the CNT-CHIT System. The multiwalled CNT precipitated in water, and in acidic aqueous solutions (pH 3.5), due to the aggregation caused by the hydrophobic interactions and van der Waals attractive forces between the tubes. In an aqueous solution of CHIT (pH 3.5), the CNT were solubilized (14) Wei, X.; Cruz, J.; Gorski, W. Anal. Chem. 2002, 74, 5039-5046.
Figure 2. (A) TGA and (B) DTG curves for (a) CHIT, (b) CNTCHIT, (c) and CNT samples. In panel B, curves a and c were shifted along the y-axis for clarity.
as indicated by the black color of the solution (see Supporting Information). Such solutions retained their black color for several months. Apparently, stabilization of the CNT-CHIT solutions was caused by the interactions between the CHIT chains and CNT. The thermodynamic rationale for such a stabilization has been recently discussed in terms of wrapping of CNT in polymer chains,6 which led to the disruption of both the hydrophobic interface with water and intertubular attractions in aggregates. The addition of a concentrated salt (e.g., 1.0 M NaCl) to a CNT-CHIT solution caused the formation of a compact black precipitate. This salting-out effect was due to the shrinkage of the particles’ electrical double layer at a high ionic strength and, thus, identified the CNT-CHIT system as a colloidal suspension of charged particles. The CNT-CHIT particles were charged because of the protonation of amino groups of CHIT (pKa ) 6.3) in the acidic solution. The deprotonation of CHIT at pH > pKa led to the formation of a bulky black gel of CNT-CHIT. Such behavior of the CNT-CHIT system is significant because it provides simple chemical means (salting-out, pH precipitation) for the manipulation of pristine carbon nanotubes. Figure 2 presents the comparative thermogravimetric studies of the solid-state samples of CHIT, CNT-CHIT, and CNT. The TGA graph of CHIT sample (Figure 2A, curve a) showed that, after the initial evaporation of residual water, the thermal decomposition of chitosan took place in two steps. One can speculate15 that these steps involved depolymerization and de(15) Peniche-Covas, C.; Arguelles-Monal.; W.; San Roman, J. Polym. Degrad. Stab. 1993, 39, 21-28.
composition of glucosamine units of chitosan at ∼200-400 °C, which was followed by the oxidative decomposition of the residues in the temperature range ∼400-600 °C. The two decomposition steps yielded a distinct narrow peak at 300 °C and a broad peak at 550 °C, respectively, on the differential TGA graph (Figure 2B, curve a). The TGA of CNT-CHIT composite sample (Figure 2A, curve b) showed that it decomposed within the temperature range ∼200-580 °C in two steps, which generally resembled the thermal decomposition of CHIT sample (Figure 2A, curve a). However, the differential TGA of the two samples revealed a major difference in their decomposition patterns. Instead of a distinct peak at 300 °C, which was observed for the CHIT sample (Figure 2B, curve a), a small ill-defined peak at 220 °C was recorded for the CNTCHIT composite sample (Figure 2B, curve b). One can hypothesize that such a shift in the decomposition peak was due to the adsorption of CHIT on CNT. The TGA of CNT sample (Figure 2A, curve c) demonstrated that CNT were more stable than CHIT. The mass loss of a CNT sample took place in the temperature range ∼430-660 °C. The remaining material at higher temperatures (∼6%) corresponded to transition metals that were introduced during the synthesis of CNT.1 The differential TGA graph of the CNT sample (Figure 2B, curve c) displayed a narrow peak at 470 °C, which could be ascribed to amorphous carbon impurities, and a wide plateau afterward due to oxidation of CNT. The larger integrated area below the plateau indicated that the sample was composed predominantly of CNT. Interestingly, the 470 °C peak was absent in the differential TGA graph of the CNT-CHIT sample (Figure 2B, curve b). This suggested the absence of amorphous carbon in the CNT-CHIT solution that was used to prepare the CNT-CHIT solid-state sample. Apparently, the CHIT-assisted solubilization of CNT was selective enough to allow for the separation of CNT from carbonaceous impurities. Thus, the solubilization of CNT in chitosan solutions has a potential for the development of an inexpensive and nondestructive purification method for CNT. A purification of carbon nanotubes based on their solubilization in starch solutions has been recently reported.8 CNT-CHIT-Based Sensor for NADH. The CNT-CHIT films were cast on the surface of GC electrodes in order to prepare NADH sensors. Preliminary studies showed that a minimum of 5 wt % CNT in CNT-CHIT films was necessary in order to detect the oxidation current of NADH. In the following experiments, the CNT-CHIT film electrodes with ∼10 wt % CNT were used without any preceding thermal or electrochemical pretreatment. Figure 3 shows that the oxidation of NADH at a GC/CHIT control electrode (curve a) and at a GC/CNT-CHIT electrode (curve b) yielded a current peak at 0.60 and 0.34 V, respectively. Such a large decrease in the overpotential for the oxidation of NADH at a CNT-based electrode could be ascribed to the high local density of electronic states in CNT, which has been attributed to their helicity and possible topological defects.16 However, a simpler explanation can be that a large surface area of CNT-based electrode, as indicated by large background currents (Figure 3, curve b1), resulted in lower current density, which led to a lower (16) Nugent, J. M.; Santhanam, S. V.; Rubio, A.; Ajayan, P. M. Nano Lett. 2001, 1, 87-91.
Analytical Chemistry, Vol. 76, No. 17, September 1, 2004
5047
Figure 3. Linear scan voltammograms recorded at (a) GC/CHIT and (b) GC/CNT-CHIT film electrodes in 0.10 mM NADH solution. Traces a1 and b1 were recorded in a background electrolyte solution, pH 7.40 phosphate buffer. For the clarity of presentation, traces b and b1 were shifted downward by 6 µA along the y-axis. Scan rate, 50 mV s-1.
oxidation overpotential. The analytical significance of the CNTbased electrodes was that they allowed for the oxidation of NADH at relatively low potentials without the need for redox mediators to shuttle the electrons from the NADH to electrode surface.17,18 Figure 4A presents an amperometric trace recorded at the GC/ CNT-CHIT electrode (E ) 0.40 V) during the spiking of NADH aliquots into a stirred buffer solution. The trace illustrates a fast response time (t90% < 5 s) of the film electrode. The inset in Figure 4A shows a calibration curve for the NADH based on the current steps of the amperometric trace. The dynamic range covered 3 orders of magnitude of NADH concentrations from 5 µM to 10 mM (not shown). A linear least-squares calibration plot over the range 5-300 µM (8 points) had a slope of 130 ( 6 mA M-1 cm-2 (disk geometrical area) with a correlation coefficient R2 ) 0.996. The detection limit was 3 µM NADH (S/N ) 3). The above analytical performance was comparable to or better than that of the other NADH detectors based on carbon nanotubes,12,13 carbon cloth,19 ceramic carbon,20 or oxidized carbon fiber electrodes.21 Recently, the boron-doped diamond films for NADH determination have been prepared using the microwave plasma chemical vapor deposition method.22 Such film electrodes have displayed very low detection limit (0.01 µM NADH) at 0.6 V, although the linear dynamic range of the determination was rather limited (0.01-0.50 µM). A variety of electrochemical NADH sensors have been developed using electrodes based on redox mediators, which facilitated oxidation of NADH at low potentials.23,24 In general, such sensors have displayed detection limits and linear ranges similar to those of CNT-based sensors; however, (17) Katakis, I.; Dominguez, E. Mikrochim. Acta 1997, 126, 11-32. (18) Armstrong, F. A.; Wilson, G. S. Electrochim. Acta 2000, 45, 2623-2645. (19) Campbell, C. E.; Rishpon, J. Electroanalysis 2001, 13, 17-20. (20) Sampath, S.; Lev, O. J. Electroanal. Chem. 1998, 446, 57-65. (21) Hayes, M. A.; Kuhr, W. G. Anal. Chem. 1999, 71, 1720-1727. (22) Rao, T. N.; Yagi, I.; Miwa, T.; Tryk, D. A.; Fujishima, A. Anal. Chem. 1999, 71, 2506-2511. (23) Lobo, M. J.; Miranda, A. J.; Tunon, P. Electroanalysis 1997, 9, 191-202. (24) Gorton, L.; Dominguez, E. Rev. Mol. Biotechnol. 2002, 82, 371-392.
5048 Analytical Chemistry, Vol. 76, No. 17, September 1, 2004
Figure 4. (A) Amperometric response of the GC/CNT-CHIT film electrode to additions of NADH aliquots (5.0 µM-1.0 mM) into a stirred solution of pH 7.40 phosphate buffer. (B) Amperometric response of the GC/CNT-CHIT-GDI-GDH film electrode to additions of glucose aliquots (10 µM-5.0 mM) into a stirred solution of pH 7.40 phosphate buffer that contained 0.10 mM NAD+. Insets: corresponding calibration plots. Potential, 0.400 V.
their stability could be severely limited due to deactivation or leaching of the mediators. Nonelectrochemical assays of NADH have been recently developed using fluorescence detection coupled to capillary electrophoresis,25-27 which yielded detection limits as low as 0.02-0.1 µM with the linearity range of e10 µM.27 The above data show that combined contemporary detection methods cover a wide range of NADH concentrations (nM-mM). CNT-CHIT-Dehydrogenase-Based Biosensor for Glucose. The biosensor was prepared by incorporating a model enzyme GDH into the CNT-CHIT film. CHIT is a convenient polymeric scaffold for enzyme immobilization and has been used previously for the immobilization of other dehydrogenases in enzymatic reactors.28-31 To accomplish the covalent immobilization of GDH, the CHIT was first chemically modified with bifunctional tether molecules of GDI.14 The modification reaction was con(25) Wise, D. D.; Shear, J. B. Anal. Biochem. 2004, 326, 225-233. (26) Wang, Z.; Yeung, E. S. J. Chromatogr., B 1997, 695, 59-65. (27) Markuszewski, M. J.; Britz-McKibbin, P.; Shigeru, T.; Matsuda, K.; Nishioka, T. J. Chromatogr., A 2003, 989, 293-301. (28) Soni, S.; Desai, J. D.; Devi, S J. Appl. Polym. Sci. 2001, 82, 1299-1305. (29) Shin, S. J.; Yamanaka, H.; Endo, H.; Watanabe, E. Enzyme Microb. Technol. 1998, 23, 10-13. (30) Cochrane, F. C.; Petach, H. H.; Henderson, W. Enzyme Microb. Technol. 1996, 18, 373-378. (31) Petach, H. H.; Driscoll, J. Biotechnol. Bioeng. 1994, 44, 1018-1022.
Figure 5. Infrared spectra of (a) CHIT, (b) CHIT-GDI, and (c) CHITGDI-GDH films.
firmed by FT-IR spectroscopy, which showed that, in contrast to the CHIT sample (Figure 5, curve a), the CHIT-GDI product displayed a characteristic absorption band at 1720 cm-1 due to the free aldehyde groups (curve b). The enzyme immobilization relied on the reaction between the free aldehyde groups of the CHIT-GDI product and amino groups of GDH to form Schiff bases. That the aldehyde groups were used up in this reaction was supported by the disappearance of the band at 1720 cm-1 in the spectrum of the CHIT-GDI-GDH film (curve c). The presence of enzyme in the film gave rise to characteristic bands at 1656 and 1540 cm-1, which could be assigned to the amide CdO and N-H deformation of GDH, respectively. Interestingly, these glucose dehydrogenase bands were similar to the ones observed previously for glucose oxidase.14 The combination of CNT with the CHIT-GDI-GDH product yielded robust biocomposite films that responded to glucose in the presence of NAD+. Figure 4B shows a steady-state response of the GC/CNT-CHIT-GDI-GDH biosensor to additions of glucose aliquots to a stirred solution of NAD+. The response was a current due to the electrooxidation of NADH that was produced in the enzymatic reaction GDH
glucose + NAD+ 98 gluconolactone + NADH (1) The biosensor displayed a fast response time (t90% < 5 s), wide linear dynamic range (5-300 µM glucose), high sensitivity (80 ( 4 mA M-1 cm-2, R2 ) 0.996), and low detection limit (3 µM glucose). That the sensitivity toward glucose (80 mA M-1 cm-2) was comparable to the sensitivity to NADH (130 mA M-1 cm-2) indicated an efficient signal transduction in the GC/CNT-CHITGDI-GDH biosensor. In addition, the relative standard deviation of glucose signal measured at five independently prepared biosensors was below 5%, which documented good reproducibility of the biosensor preparation. Perhaps the most attractive feature of the GC/CNT-CHITGDI-GDH biosensor, when compared to other dehydrogenasebased electrodes,18 was its operational stability. Figure 6 (curve a) shows that the biosensor maintained practically constant current under extended polarization (24 h at 0.400 V) in a stirred solution of 0.50 mM glucose and 0.10 mM NAD+. This demonstrated a remarkable resistance of the GC/CNT-CHIT-GDI-GDH elec-
Figure 6. Amperometric trace recorded at (a) GC/CNT-CHITGDI-GDH, E ) 0.400 V, and (b) GC/CHIT-GDI-GDH, E ) 0.700 V, film electrodes in a stirred solution of 0.50 mM glucose that contained 0.10 mM NAD+. Background electrolyte, pH 7.40 phosphate buffer.
trode to fouling, which is common to oxidation of NADH at other solid electrodes. The latter point is illustrated by the decaying current that was recorded at a CNT-free film electrode (Figure 6, curve b). One can hypothesize that improved operational stability of the CNT-based electrode was due to faster charge transfer and lower overpotentials for the oxidation of NADH at CNT. The improved kinetics of electron transfer limited the amount of radical intermediates and their dimerization, which typically cause fouling of the electrode surface during the oxidation of NADH via two successive one-electron-transfer steps.17 In addition, the fast diffusion of species away from nanotubes and the protection of enzyme by the chitosan matrix contributed to the improved stability of the GC/CNT-CHIT-GDI-GDH electrode. The store stability of the GC/CNT-CHIT-GDI-GDH electrode was studied over the period of 8 days by recording daily the current (E ) 0.40 V) in a stirred buffer solution that contained 0.50 mM glucose and 0.10 mM NAD+. The electrode was stored at 4 °C in the pH 7.40 phosphate buffer solution when not in use. On day 4, the current decreased to ∼20% of the initial response and stayed practically constant afterward. The common problem in the electrochemical determination of glucose is the interference from redox-active species such as ascorbic acid, uric acid, and acetaminophen that are usually present in physiological samples of glucose. The detection of glucose at the CNT-based film electrode also suffered from such interferences because all three species yielded the oxidation currents at potentials used for the oxidation of NADH. However, good operational stability and sensitivity of the GC/ CNT-CHIT-GDI-GDH film electrode allowed for the design of an interference-free glucose assay based on the standard addition method. This was demonstrated using a real physiological matrix (urine) that was spiked with known amounts of glucose. The assay relied on the current measurement at the GC/CNT-CHIT-GDIGDH electrode (E ) 0.400 V), which was immersed in a stirred buffer solution to which relevant aliquots were sequentially added (Figure 7). First, an aliquot of the sample (glucose-containing urine) was added to a buffer solution. This generated a current iinterf (step a) due to the electrooxidation of redox-active interfering species (ascorbic acid, uric acid, etc.) that were present in urine. Second, an aliquot of NAD+ was added to the solution in order to trigger the enzymatic reaction 1 and, thus, to produce the NADH. Analytical Chemistry, Vol. 76, No. 17, September 1, 2004
5049
interferences from changing levels of oxygen in the samples. Work is in progress on the development of reagentless biosensors based on coimmobilization of dehydrogenases, NAD+, and interference rejecting agents in CNT-CHIT films.
Figure 7. Amperometric trace (E ) 0.400 V) recorded at GC/CNTCHIT-GDI-GDH film electrode in a stirred buffer solution (10.0 mL) during the additions of (a) 10.0 µL of urine containing 10 mM glucose, (b) 40.0 µL of 25.0 mM NAD+, and (c) 10.0 µL of 10 mM glucose standard solution. Inset: standard addition plot. Background electrolyte, pH 7.40 phosphate buffer.
Since the amount of NADH produced was proportional to the glucose content of the sample, the electrooxidation of NADH yielded a current (step b) that was proportional to glucose concentration in the sample. To calibrate this current, the aliquots of a standard glucose solution were added to the assay solution (steps c). The standard addition method yielded a good recovery of glucose (116 ( 3%, N ) 7). This illustrated the merit of the biosensor system based on the GC/CNT-CHIT-GDI-GDH electrode for the interference-free determination of glucose in the matrix of urine. Another advantage of such an electrode system was that the oxygen did not influence its analytical performance. This is in contrast to oxidase-based electrodes that are prone to
5050 Analytical Chemistry, Vol. 76, No. 17, September 1, 2004
CONCLUSIONS The nondestructive solubilization of multiwalled CNT in an aqueous solution of poly(aminosaccharide) CHIT offers straightforward means for their manipulation, purification, and modification. The CNT-CHIT system represents a new biocomposite platform for the development of dehydrogenase-based electrochemical biosensors. In such a system, CNT provide a signal transduction based on the electrooxidation of dehydrogenase cofactor NADH, while CHIT serves as a biocompatible and chemically modifiable scaffold for enzyme immobilization. The CNT-CHIT biocomposites have a potential to provide operational access to a large group of dehydrogenase enzymes, which encompass hundreds of members, for designing of a variety of bioelectrochemical devices (e.g., sensors, biosensors, biofuel cells). ACKNOWLEDGMENT The NIH/MBRS/SCORE Grant GM 08194 supported this work. The authors gratefully acknowledge Dr. David M. Johnson for help with TGA experiments. SUPPORTING INFORMATION AVAILABLE Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org.
Received for review March 26, 2004. Accepted June 14, 2004. AC049519U
