Anal. Chem. 2003, 75, 2703-2709

A Glucose Biosensor Employing a Stable Artificial Peroxidase Based on Ruthenium Purple Anchored Cinder Jyh-Myng Zen,* Annamalai Senthil Kumar, and Ching-Rue Chung

Department of Chemistry, National Chung Hsing University, Taichung 402, Taiwan

Iron-enriched industrial waste cinder (CFe*) has been recycled for efficient and stable anchoring of Ru(CN)64to the formation of a hybrid ruthenium purple complex. The cinder/ruthenium purple hybrid-modified carbon paste electrode (designated as CPE/CFe*-RP) was worked out for hydrodynamic analysis of H2O2 at a low detecting potential of 0.0 V versus Ag/AgCl in pH 7 ammonium buffer solution. The highly active, selective, and stable electrocatalytic system with a function similar to peroxidase enzyme shows a linear calibration curve up to 0.8 mM H2O2 at a rotation rate of 3600 rpm with slope and detection limit (S/N ) 3) of 0.11 µA/µM and 33 nM, respectively. Interference by direct electrochemical oxidation of easily oxidizable substances can be prevented as a result of the low detecting potential of the working system. A glucose biosensor was further constructed by coating with glucose oxidase and Tosflex on the CPE/ CFe*-RP (denoted as CPE/CFe*-RP/GOx/Ts). The proposed CPE/CFe*-RP/GOx/Ts with a two-layer configuration, that is, enzyme and protecting layers, exhibits good operational performance in terms of response time, linearity, detection limit, and lifetime. Tailoring of the electrode surface to construct a sensitive and cheap biosensor with low operating potentials is a challenging and continuous analytical research interest. Amperometric monitoring of H2O2 oxidation current signal in enzyme-coupled reactions often uses Pt, Pd, Rh, or some alloy electrodes.1-11 However, high detection potentials cause the interference of other electroactive substances rendering the biosensors unsuitable to further * To whom correspondence should be addressed. Fax: 886-4-22862547. E-mail: [email protected]. (1) Jung, S.-K.; Wilson, G. S. Anal. Chem. 1996, 68, 591-596. (2) Wang, J.; Zhang, X. Anal. Chem. 2001, 73, 844-847. (3) Wang, J. Anal. Chem. 1999, 71, 328R-332R and references therein. (4) Gunasingham, H.; Tan, C. B. Analyst 1989, 114, 695-98. (5) Sadik, O. A.; Brenda, S.; Joasil, P.; Lord, J. J. Chem. Ed. 1999, 76, 967970. (6) Xu, J.-J.; Chen, H.-Y. Anal. Biochem. 2000, 280, 221-226. (7) Bartlett, P. N.; Cooper, J. M. J. Electroanal. Chem. 1993, 362, 1-12 and references therein. (8) Yamato, H.; Koshiba, T.; Ohwa, M.; Wernet, N.; Matsumura, U. Synth. Met. 1997, 87, 231-236. (9) Chi, S.-M.; Oh, S.-M. Electrochim. Acta 1996, 15, 2433-2439. (10) Labat-Allietta, N.; Thevenot, D. R. Biosens. Bioelectron. 1998, 13, 19-29. (11) Hall, S. B.; Khudaish,E. A.; Hart, A. L. Electrochem. Acta 1999, 44, 24552462. 10.1021/ac020542u CCC: $25.00 Published on Web 04/30/2003

© 2003 American Chemical Society

real sample assays. Horseradish peroxidase-based enzyme systems, on the other hand, provide a specific detection of H2O2 but always lack practical stability. The addition of mediators (such as metal hexacyanoferrates,12-15 methyl viologen,16 tetrathiafulvalene tetracyanoquinodimethane,17 luminol,18 etc.) that can operate at relatively lower potentials is a good alternative for the metal-based biosensors. Recently, a Prussian blue (PB)-based “artificial peroxidase” was reported as a transducer for H2O2.12-14 As a result of high activity and selectivity in H2O2 reduction, the specially deposited PB was denoted as “artificial enzyme peroxidase”. The main problem for the classical method of PB preparation on conducting substrate, however, is the low stability upon strong mechanical stress during measurements, especially in neutral pHs.19 A highly stable mediator system that can operate at a low detection potential under hydrodynamic conditions for H2O2 detection would thus be interesting in analytical chemistry. Our group recently disclosed a novel and stable electrocatalytic system in which PB and its related analogues are electrochemically prepared directly on the interlayer resume of cinder-modified carbon paste (CPE) or screen-printed electrodes.20-24 The ironenriched cinder (CFe*) can help in the anchoring of hybridized bimetallic hexacyanometalate electrodes, such as CFe*-[M(CN)6], on the working electrode surface (where M ) Fe, Co, Ni, etc). This modification procedure gave extraordinary stability to the mediator system, even under strong hydrodynamic stress.21 The cinder is considered to be industrial waste and can be obtained anywhere in the world where there is the steel industry. Recycling (12) Karayakin, A. A.; Gitelmacher, O. V.; Karyakina, E. E. Anal. Chem. 1995, 67, 2419-2423. (13) Karayakin, A. A.; Karyakina, E. E.; Gorton, L. Anal. Chem. 2000, 72, 17201723. (14) Karayakin, A. A.; Kotel’nikova, E. A.; Lukachova, L. V.; Karyakina, E. E.; Wang, J. Anal. Chem. 2002, 74, 1597-1603. (15) Shyu, S.-C.; Wang, C.-M. J. Electrochem. Soc. 1998, 145, 154-158. (16) Zen, J.-M.; Lo, C.-W. Anal. Chem. 1996, 68, 2635-2640. (17) Li, Q.-S.; Ye, B.-C.; Liu, B.-X.; Zhong, J.-J. Biosens. Bioelectron. 1999, 14, 327-334. (18) Ouyang, C.-S.; Wang, C.-M. J. Electroanal. Chem. 1999, 474, 82-88. (19) Scharf, U.; Grabner, E. W. Electrochim. Acta 1996, 41, 233-239. (20) Zen, J.-M.; Senthil Kumar, A.; Chen, H.-W. Electroanalysis 2000, 12, 542545. (21) Zen, J.-M.; Senthil Kumar, A.; Chen, H.-W. Electroanalysis 2001, 13, 11711178. (22) Zen, J.-M.; Chen,-P. Y.; Senthil Kumar, A. J. Chin. Chem. Soc. 2002, 49, 915-920. (23) Senthil Kumar, A.; Zen, J.-M. Proceeding of the Seventh International Symposium on Advances in Electrochemical Science and Technology, SAEST, India, 2002, D95-98. (24) Senthil Kumar, A.; Zen, J.-M. Electroanalysis 2003, in press.

Analytical Chemistry, Vol. 75, No. 11, June 1, 2003 2703

Figure 1. Schematic diagram for the tailoring and structure of the CPE/CFe*-RP/GOx/Ts system.

of such an industrial waste to an efficient catalytic system is economically attractive. In this work, we extended our hybridization route to Ru(CN)64- to the formation of ruthenium purple (RP, CFe*-Ru(CN)6). A hydrodynamic detecting system was further used to increase the sensitivity. This approach provides the feasibility for the sensitive determination of H2O2 at 0.0 V vs Ag/ AgCl. An amperometric biosensor for glucose was then constructed by successive coating of an enzyme and a protecting layer of Tosflex (Ts, a perfluorinated polycationic polymer)25,26 on the working system. Note that, similar to Nafion,14,27 Tosflex can work as an oxygen resource and helps to improve the activity of enzyme, even in an oxygen-free solution. Herein, we describe the optimal compositions of two polymeric composites, that is, an enzyme and protecting layer, on the glucose biosensor performance (e.g., sensitivity, stability, and selectivity). The characteristic of the glucose sensor illustrates the advantages of hybrid RP-based “artificial peroxidase” as a transducer for H2O2 detection. Finally, the practical applicability of the proposed glucose sensor is demonstrated by determining the glucose content in some real samples. EXPERMENTAL SECTION Chemicals and Reagents. Glucose oxidase (GOx, EC 1.1.3.4, 181,600 units/g, from Aspergillus niger), β-D-glucose, bovine serum albumin (BSA), Ru(CN)64-, L-ascorbic acid, uric acid, and acetaminophen were bought from Sigma (St. Louis, MO). Glutaraldehyde (GA) was obtained from E-Merck. Graphite powder (Aldrich) with diameter of 1-2 µm and mineral oil (Mallinckrodt) of viscosity 25 cs were used for making the carbon paste electrode (CPE). The Tosflex membrane (IE-SA 48) was obtained from Tosoh Soda, Japan, and dissolved per our recent report.25 All of the other compounds (ACS-certified reagent grade) in this work were used without any further purification. Deionized double-distilled water was used for making the solutions. Apparatus. All electrochemical experiments were performed on a BAS CV-100W electrochemical analyzer (West Lafayette, IN). (25) Zen, J.-M.; Tsai, D.-M.; Senthil Kumar, A.; Dharuman, V. Electrochem. Comm. 2000, 2, 782-785. (26) Zen, J.-M.; Tsai, D.-M.; Senthil Kumar, A. Electroanalysis 2003, in press. (27) Wang, J.; Fang. L. J. Am. Chem. Soc. 1998, 120, 1048-1050.

2704

Analytical Chemistry, Vol. 75, No. 11, June 1, 2003

A BAS model VC-2 electrochemical cell was employed in these experiments. For the rotating disk electrode (RDE) studies, a BAS rotator system with 3-mm diameter BAS cavity electrode was used. The three-electrode system consisted of cinder-modified working electrodes, a Ag/AgCl reference electrode (model RE-5, BAS), and a platinum wire auxiliary electrode. The flow injection system consisted of a Cole-Parmer (Vernon Hills, IN) microprocessor pump drive and a Rheodyne (Cotati, CA) model 7125 sample injection valve (20-µL loop) with an interconnecting Teflon tube. A supporting electrolyte of 0.1 M, pH 7 ammonium buffer was used in most cases. Other buffer solutions (I ) 0.1 M) of phosphate (PBS), tris(hydroxymethyl)-aminomethane/HCl (Tris), and KCl/KOH were also prepared for the studies. UV-visible spectra were recorded with a Hitachi U-3000 spectrophotometer. In real sample application, a portable glucose meter (Precision Plus) was used for direct estimation of glucose and for the comparison. Electrode Preparation. The cinder material was obtained from the industrial waste of a local steel mill. The processing of the cinder and the preparation of the cinder-modified carbon paste electrode (designated as CPE/CFe*) generally followed our earlier reports.20,21 Figure 1 shows the systematic scheme for the tailoring procedure of the cinder-modified electrodes. For the preparation of the cinder/ruthenium purple hybrid-modified carbon paste electrode (designated as CPE/CFe*-RP), 2 mM of Ru(CN)64in 0.1 M, pH 2 ammonium buffer solution was continuously cycled for 60 scans at a scan rate of 50 mV/s. The formation of RP in the cinder matrix was confirmed by the absorption spectra (with λmax ) 550 nm).28 A layer of 5 µL of enzyme followed by another layer of Tosflex was coated onto the CPE/CFe*-RP to prepare the glucose biosensor. A typical enzyme casting solution was prepared by dissolving 25.0 mg of GOx, 21.3 mg of BSA, and 25 µL of 25% GA in sequence in 1 mL of water. The Tosflex coating solution used in the biosensor preparation was 0.9 wt % in a 1:1:1 mixture of ethanol, 2-propanol, and water. General Procedure. The glucose responses of the electrode were measured as baseline-corrected peak current values either from dynamic response of rotating disk electrode or from flow (28) Rajan, K. P.; Neff, V. D. J. Phys. Chem. 1982, 86, 4361-4368.

Figure 2. Cyclic voltammograms of the CPE/CFe*-RP before (a) and after (b) the addition of 0.1 mM H2O2 in pH 7 ammonium buffer solution (I ) 0.1 M) at v ) 50 mV/s. Insert figure is the comparison graph for H2O2 reduction reaction at various chemically modified electrodes.

injection analysis (FIA). The enzyme electrode was kept in pH 7 ammonium buffer at 4 °C when not in use. Three commercially available real samples of glucose tablet (Becton Dickinson), 5% dextrose monohydrate (as an intravenous drip solution, Gitose, Taiwan), and oral electrolyte maintenance solution (Pedialyte, Ross Production) were used in the practical analysis. The samples were properly diluted with pH 7 ammonium buffer before being used for detection under E ) 0.0 V at room temperature. RESULTS AND DISCUSSION Electrochemical Behavior and H2O2 Mediated Reduction on the CPE/CFe*-RP. Figure 2a shows the cyclic voltammetric response of the CPE/CFe*-RP in pH 7 ammonium buffer (I ) 0.1 M) solution. A well-defined redox peak was noticed at a formal potential (Ef ) (Epa + Epc)/2) of 250 mV versus Ag/AgCl. Similar redox behavior was also observed on the electrodeposited RP analogue at the Pt electrode (Ef ) 245 mV versus SCE in 1 N K2SO4) and clay modified electrode (Ef ) 190 mV versus SCE in pH 5.1 solution).15,28 The electrochemical behavior confirms the anchoring of the hybridized RP by the fixed CFe* in the cinder matrix corresponding to the following electrochemical electrontransfer reaction with the ammonium ion.28

NH4CFeIII*RuII(CN)6 + e- + NH4+ T (NH4)2CFeII*RuII(CN)6 (1) In the above transformation, the nude high spin CFeIII*/CFeII* couple derived from cinder could directly participate in the redox process. The unit cell of RP is a face-centered cubic structure with the low-spin Ru ions coordinated to C terminals and high-spin Fe ions linked to N groups, such as FeIII-NC-RuII, in the threedimensional framework similar to PB analogues.29 A linear

Figure 3. Dynamic response of H2O2 at the CPE/CFe*-RP rotating disk in pH 7 ammonium buffer under various potentials (A) and rotation rates (B). Insert figure is the typical Koutecky-Levich plot for the data of part B. [H2O2] ) 10 µM.

dependence of the cathodic peak current (ipc) versus square route of scan rate (v1/2) plot and a slope (∂ log(ipc)/∂ log(v)) ∼0.5 from the double logarithmic plot of ipc versus v in blank solution indicates the diffusion-controlled charge-transfer nature of the working system.20,21,30 The surface excess (ΓRP) was calculated as 8.5 × 10-9 moles/cm2, which is ∼2 orders higher than the value of 5.0 × 10-11 moles/cm2 reported from clay-modified RP electrodes.15 The reductive current signals are effectively mediated at the redox potential of RP upon the addition of H2O2 (Figure 2b). Under identical experimental conditions, the electrochemical signal observed at the CPE/CFe*-RP is much higher than that of GCE, classical GCE/RP, and CPE/CFe* systems. Meanwhile, interference by direct electrochemical oxidation of easily oxidizable substances can also be prevented because of the low detecting potential of the working system. Figure 3A shows a typical hydrodynamic i-t curve for the CPE/CFe*-RP at various applied potentials (400 to -400 mV) with 10 µM H2O2 at a rotation rate (ω) of 3600 rpm. As can be seen, the reduction limiting current (iL) starts to level off at ∼0.0 V. Similarly, the rotation rates also affected the iL values, as shown in Figure 3B. The Levich plot (iL versus ω1/2) was linear up to 3600 rpm, and after that, it reached the kinetic limitation. A typical (29) Itaya, K.; Uchida, I.; Neff, V. D. Acc. Chem. Res. 1986, 19, 162-168 and references therein. (30) Bard, A. J.; Faulkner, L. R. Electrochemical Methods, Fundamentals and Applications; John Wiley & Sons: New York, 2001.

Analytical Chemistry, Vol. 75, No. 11, June 1, 2003

2705

Scheme 1. (A) Schematic Representation of the Enzyme Catalyzed Glucose Oxidation and (B) Reaction Pathway in Terms of the Michaelis-Menten Kineticsa

Figure 4. Typical plot of current versus [H2O2] on the CPE/CFe*RP rotating disk electrode in pH 7 ammonium buffer solution at E ) 0.0 V and ω ) 3600 rpm.

linear Koutecky-Levich plot (iL-1 versus ω-1/2) can be used to evaluate the heterogeneous electron-transfer rate constant (k°) using the following linear expressions.30

1/iL ) 1/ik + 1/0.62nFAD2/3ν-1/6ω1/2[H2O2]

(2)

ik ) nFAk°[H2O2]

(3)

Here, ik and ν correspond to the kinetic current measured at ω-1/2 f 0 and kinematic viscosity, respectively, and other symbols have their own significance. The k° value obtained was 0.012 cm/s with ΓRP ) 8.5 × 10-9 moles/cm2. Note that the reported k° values for artificial peroxidase using PB-modified electrodes (ΓPB ) 6-10 × 10-9 moles/cm2) are in the range of 0.015-0.020 cm/s.31,32 In other words, the present system is comparable to earlier reported PB-based systems regarding the heterogeneous electron-transfer rate constant. A possible electrocatalytic mechanism is proposed through redox mediation of a lower oxidation state in the CPE/ CFe*-RP (i.e., (NH4)2CFeII*RuII(CN)6). The calibration plot shows a systematic increase in current signals up to 800 µM H2O2, with a slope and regression coefficient of 0.11 µA/µM and 0.994, respectively (Figure 4). The detection limit (DL, S/N ) 3) was 33 nM. Successive addition of 10 µM H2O2 (n ) 13) showed a relative standard deviation (RSD) of 4.16% verifying the good reproducibility of the system under strong hydrodynamic condition. Meanwhile, as can be seen from the i-t RDE curves (Figure 3), another advantage of the present system was the very fast response time (t95% < 15 s). Major interference from ascorbic acid, acetaminophen, and dopamine with equal concentrations was also negligible, as shown in Figure 5A. Finally, critical parameters, such as dissolved O2 and stability of the loaded mediator, that can affect the application in real samples were especially studied. As can be seen in Figure 5B, no appreciable change in the redox behavior was observed, regardless the (31) Karyakin, A. A.; Karyakina, E. E.; Gorton, L. J. Electroanal. Chem. 1998, 456, 97-104. (32) Zhang, Y.; Wilson, G. S. J. Electroanal. Chem. 1993, 345, 253-271.

2706

Analytical Chemistry, Vol. 75, No. 11, June 1, 2003

a

rds ) rate determining step.

existence of O2. Continuous CV scans at ω ) 3000 rpm on the GCE/RP prepared by the classical method showed complete leaching of the catalyst under this hydrodynamic stress, whereas ∼90% of the peak response was retained at the CPE/CFe*-RP. This result clearly demonstrates the superior stability of the present system. Very recently, our group has made detailed X-ray photoelectron spectroscopic (XPS) studies of the bare cinder material. The results showed that the cinder is a complex composite of calcium and iron silicates similar to a glass matrix.24 The iron ions were found to be tagged in bridging (BO) and nonbridging oxygen (NBO) sites with a relative percentage of 85%, 15% in the silicate matrix. Moreover, the free and labile NBO-tagged Fe was found to help for the derivatization of the M(CN)6n- complexes as hybrid PB analogues. Hence, it is expected that the rigid backbone of the free iron ions in the cinder can lead to a hybrid RP system more stable than the classical RP film. Furthermore, experiments regarding the glucose biosensor from this “artificial peroxidase” system through the detection of H2O2 (formed as a dissolved O2 reduced product in the intermediate step during the specific enzymatic reaction with glucose, as shown in Scheme 1A) are studied subsequently. Evaluation of Biosensor Performance. The importance of the overlayer coating on top of the GOx layer (i.e. CPE/CFe*RP/GOx) was first investigated in order to improve the stability and durability of the working system. Both Nafion and Tosflex were taken into consideration under identical experimental conditions. Interestingly, the Tosflex coating (i.e. CPE/CFe*-RP/GOx/ Ts) showed a better performance, ∼10 times higher in current response than that of the Nafion coating. It is likely that the positively charged Tosflex attracts the enzyme having a negative charge at physiological pH and further enriches the concentration of enzyme near the electrode/electrolyte interface and, thus, for the increased detecting current signals. Further optimized experimental factors include solution pH and rotation speed. The optimum values of pH ) 7 and ω ) 3600 rpm are similar to the optimized conditions obtained in H2O2 detection, further confirming the proper function of the CPE/ CFe*-RP/GOx/Ts for glucose sensing. The effect of enzyme loading on the electrode performance was evaluated next, and

Figure 5. Dynamic responses of the CPE/CFe*-RP rotating disk electrode with 10 µM H2O2 (first three spikes) and with additional 10 µM each of ascorbic acid (AA), acetaminophen (AP), and dopamine (DA) (last three spikes) (A). CV response of the CPE/CFe*-RP in the absence (i) and presence (ii) of 0.1 mM H2O2 with N2-saturated (a) and O2-saturated (b) pH 7 ammonium buffer solution at v ) 50 mV/s (B). Other conditions are as in Figure 4.

Figure 6. Effect of GOx-enzyme loading (A) and applied potential (B) on the determination of glucose at CPE/CFe*-RP/GOx/Ts rotating disk system. Other experimental conditions are as in Figure 4.

the results are shown in Figure 6A. As can be seen, the sensitivity (µA/mM) increased upon increasing the enzyme loading and reached a maximum at ∼25 mg/mL of GOx. The optimum applied potential of 0.0 mV (Figure 6B) is more negative than the redox potential of RP (Ef ) 250 mV) and lies between there and the redox potential of the H2O2 reduction reaction.30 It is a typical

example of mixed-potential mechanism,33 at which both of the half reactions (RP and H2O2) are operating simultaneously. Overall, the advantage for the construction of a low potential biosensor is clear. (33) Power, G. P.; Ritche, I. M. J. Chem. Ed. 1983, 60, 1022-1026.

Analytical Chemistry, Vol. 75, No. 11, June 1, 2003

2707

Table 1. Determination of Glucose in Various Samples Using the CPE/CFe*-RP/GOx/Ts

a

parameters

intravenous drip

oral electrolyte

tablet

linear equation R original value, µM spiked value, µM after spike, µM recovery, % exptl value exptl valuea labeled value

y ) 1.8748 + 0.0064x 0.9998 292.93 300.00 609.15 102.74 0.293 M 0.288 M 0.278 M

y ) 1.295 + 0.00505x 0.9999 256.43 300.00 563.44 101.26 23.08 g/L 23.70 g/L 20.00 g/L

y ) 1.5833 + 0.00653x 0.9999 242.46 300 556.60 102.60 4.37 g/tablet 4.82 g/tablet 5.00 g/tablet

Sample analysis using a commercial glucose meter.

Figure 7. Variation of base electrolyte (A) and Tosflex amount (B) on the determination of 0.1 mM glucose using the CPE/CFe*-RP/ GOx/Ts rotating disk system. Other experimental conditions are as in Figure 4.

The effect of different buffer solutions on the performance of the glucose sensor was also taken into consideration, as shown in Figure 7A. The obtained current signals were in the order of ammonium > KCl > Tris > PBS. The higher efficiency of the ammonium buffer over the rest may be presumably due to the positive interaction with GOx and, in turn, to the current amplification. The amount of Tosflex to the biosensor activity was also studied, as shown in Figure 7B. Thick film (>1.8 wt %) showed considerable inhibitory effect to biosensor activity. Thin films with a predominance of monolayer characteristics lead to effective signal communication from the internal electrode surface to the electrolyte interface, and so the time of access and the energy barrier are lower. A similar physical characteristic was also noticed on the Nafion-RuO2-Ru(bpy)32+ multicomposite 2708 Analytical Chemistry, Vol. 75, No. 11, June 1, 2003

Figure 8. Typical FIA response for the CPE/CFe*-RP/GOx/Ts with increasing concentration of glucose in pH 7 ammonium buffer solution (A) and its calibration plot (B). E ) 0.0 V and flow rate ) 2.4 mL/min. Curve-fitting was performed using nonlinear hyperbolic function.

electrodes for water oxidation reaction.34 An optimum value of 0.9 wt % Tosflex was thus derived for further experiments. In addition to the oxygen resource, the Tosflex-modified biosensor showed a remarkable stability, even under strong hydrodynamic stress, at ω ) 8100 rpm in neutral pHs. Note that this level of stability is impossible to reach with classical PB- and RP-modified systems. The calibration plot using a simple hydrodynamic i-t curve (by RDE) at ω ) 3600 rpm with E ) 0.0 V showed a linear range (34) Chandrasekara Pillai, K.; Senthil Kumar, A.; Zen, J.-M. J. Mol. Catal. A 2000, 160, 277-285.

up to 2 mM, with a slope of 5.4 mA/M and regression coefficient of 0.998. Reproducibility of the biosensor system was evaluated under the optimized conditions, and a relative standard deviation (RSD) of 3.6% was obtained for 12 successive analyses of 50 µM glucose. The DL (S/N ) 3) of 0.35 µM is more sensitive than a similar system based on the Clay/RP/GOx electrode.15 Finally, considering the convenience of FIA in an analytical application, an FIA-based system was also constructed using a low working volume at low operating potential. We also hope that working with FIA can eliminate the diffusion restriction of glucose at high concentrations. The effect of flow rate (Rf) to the performance was studied at E ) 0.0 V with 2 mM glucose, and the obtained peak height (µA) and corresponding area (µC) were found to decrease sharply from 0.2 to 0.5 mL/min and then decreased slowly up to 2.8 mL/min. The optimized Rf was then evaluated as 2.4 mL/min by calculating the FIA peak resolution (Rp) (i.e., peak height (µA)/peak area (µC)). Under this optimum Rf value, typical FIA responses for various concentrations of glucose are shown in Figure 8. The calibration plot was linear up to 20 mM glucose, with a sensitivity of 182 µA/M and a regression coefficient of 0.9998. The slope value is ∼30 times lower than hydrodynamic i-t analysis. However, the upper detection value is 10 times increased by FIA over the earlier method, and it further offers experiments with low working volumes (20 µL). Reproducibility of the biosensor system was again evaluated by successive injection (n ) 16) of 1 mM glucose, and an acceptable RSD value of 4.6% was observed. Michaelis-Menten Kinetics. Scheme 1 shows the possible reaction mechanism based on general (A) and Michaelis-Menten (MM) (B) pathways for the glucose oxidation, where GOxox and GOxred represent the oxidized and reduced forms of the glucose oxidase enzyme, respectively, Kmapp is the MM’s apparent rate constant, and kcat is the catalytic rate constant for the overall oxidation reaction (rds). The intermediate [GOxox-G] is considered a high-energy complex with a key-lock type of enzymesubstrate interaction. The kinetic parameters were deduced using simple nonlinear curve-fitting analysis based on the MarquardtLavenberg algorithm using the following expressions.35-37 (35) Senthil Kumar, A.; Zen, J.-M. Electroanalysis, 2002, 14, 671-678. (36) Zen, J.-M.; Lai, Y.-Y.; Ilangovan, G.; Senthil Kumar, A. Electroanalysis 2000, 12, 280-286. (37) Lyons, M. E. G. In Advances in Chem. Phys. Polymeric Systems; Pyrigogine, I., Rice, S. A., Eds.; Wiley: New York, 1996; p 297. (38) Rogers, M. J.; Brandt, K. G. Biochemistry 1971, 10, 4624-4630. (39) Swoboda, B.; Massey, V. J. Biol. Chem. 1965, 240, 2209-2215. (40) Cho, J.-H.; Shin, M.-C.; Kim, H.-S. Sens. Actuators, B 1996, 30, 137-141.

ipc ) nFAkcatΓE[glucose]/(Kmapp + [glucose])

(4)

imax ) nFAkcatΓE

(5)

In the above equations, ΓE and imax correspond to the surface enzyme concentration and maximum oxidation current, respectively. Figure 8B shows the nonlinear fitted curve for the data obtained from FIA in glucose determination (Note that glucose detection was operated indirectly from the H2O2 assays, as shown in Scheme 1A). The calculated Kmapp and imax were 28.7 mM and 8.4 µA, respectively. The Kmapp is reasonable, as compared to an earlier reported value of 26-33 mM for the soluble enzyme systems38,39 and 33 mM for a Pt/polypyrrole/GOx biosensing electrode.40 Overall, the CPE/CFe*-RP/GOx/Ts fulfills most of the criteria for the development of an ideal glucose sensor, especially the operation under low detecting potential. Real Sample Analysis. Analytical application to the determination of glucose content in real samples by the standard addition method is summarized in Table 1. For comparison, a commercial glucose meter was used to measure the glucose content simultaneously. As can be seen, the values measured by the present system and a commercial glucose meter are very close. These results indicate the suitability of the present system to practical applications. CONCLUSIONS A hybrid RP-anchored cinder modified carbon paste electrode was successfully demonstrated for effective H2O2 reduction at low operative potentials. Basic physical and mechanistic aspects of the CPE/CFe*-RP toward the H2O2 reduction reaction indicate an action similar to horseradish peroxidase enzyme. In other words, the present system can also be denoted as “artificial enzyme peroxidase”. Further modification with GOx enzyme and Tosflex on the external surface showed excellent glucose biosensing activity and marked stability, even under strong hydrodynamic stress. The derived apparent Michaelis-Menten rate constant (Kmapp) is comparable to solution-phase enzyme and indicates the suitability of the present system. Real sample analyses of some glucose-containing drugs gave satisfactory results with good recoveries. This system has the potential for further extension to single-use disposable sensors, and the work is in progress. Received for review August 21, 2002. Accepted March 13, 2003. ACKNOWLEDGMENT The authors gratefully acknowledge financial support from the National Science Council of the Republic of China. AC020542U

Analytical Chemistry, Vol. 75, No. 11, June 1, 2003

2709