771

Full Paper

A Nonenzymatic Glucose Sensor Using Nanoporous Platinum Electrodes Prepared by Electrochemical Alloying/Dealloying in a Water-Insensitive Zinc Chloride-1-Ethyl-3-Methylimidazolium Chloride Ionic Liquid Chih-Hung Chou,a Jyh-Cheng Chen,a Chia-Cheng Tai,b I-Wen Sun,b Jyh-Myng Zena* a

Department of Chemistry, National Chung Hsing University, Taichung 402, Taiwan *e-mail: [email protected] b Department of Chemistry, National Cheng Kung University, Tainan 70101, Taiwan Received: November 21, 2007 Accepted: December 6, 2007 Abstract We report here a nonenzymatic sensor by using a nanoporous platinum electrode to detect glucose directly. The electrode was fabricated by electrochemical deposition and dissolution of PtZn alloy in zinc chloride-1-ethyl-3methylimidazolium chloride (ZnCl2-EMIC) ionic liquid. Both SEM and electrochemical studies showed the evidences for the nanoporous characteristics of the as-prepared Pt electrodes. Amperometric measurements allow observation of the electrochemical oxidation of glucose at 0.4 V (vs. Ag/AgCl) in pH 7.4 phosphate buffer solution. The sensor also demonstrates significant reproducibility in glucose detection; the higher the roughness factor of the Pt electrode, the lower the detection limit of glucose. The interfering species such as ascorbic acid and p-acetamidophenol can be avoided by using a Pt electrode with a high roughness factor of 151. Overall, the nanoporous Pt electrode is promising for enzymeless detection of glucose at physiological condition. Keywords: Glucose, Porous platinum, Nonenzymatic sensor, Ionic liquid DOI: 10.1002/elan.200704102

1. Introduction Numerous studies have been performed to overcome or to alleviate the drawbacks of enzymatic glucose sensors for their lack of stability due to the intrinsic nature of enzymes. To address this problem, amperometric measurements allow observation of the electrochemical oxidation of glucose on a bare platinum (Pt) surface [1 – 3]. Normal Pt electrodes, however, are low sensitivity and poor selectivity in glucose detection and sometimes are not free from poisoning by adsorbed intermediates [4]. Such disadvantages have been partly overcome by Pt-Pb alloy (Pt2Pb) electrodes [5]. Compared to pure Pt surfaces, Pt2Pb is relatively insensitive to interfering species, such as lascorbic acid (AA), uric acid, and p-acetamidophenol (AP). Moreover, Pt2Pb generates more stable and larger responses than that of pure Pt. Despite these valuable advantages, surface poisoning by chloride ion still remains a serious problem. Another successful approach for direct oxidation of glucose is through the application of either mesoporous or nanoporous Pt films [6 – 16]. Several studies have reported on discriminatively enhanced amperometric responses to glucose and on the performance of nonenzymatic glucose sensors based on these porous Pt structures. The key idea is based on the fact that the roughness of the porous electrodes is even smaller than the scale of the Electroanalysis 20, 2008, No. 7, 771 – 775

chronoamperometric diffusion field. As a result, the faradaic currents of rapidly oxidizable or reducible reactants are proportional to the apparent geometric area of the electrode, regardless of its porous roughness [17]. The faradaic currents associated with kinetic-controlled electrochemical events, on the other hand, are sensitive to the nanoscopic surface area of the electrode, rather than to its geometric area. The main goal of this work is to develop a nonenzymatic glucose sensor based on a nanoporous Pt electrode prepared by green chemistry. We recently reported a simple strategy for the fabrication of nanoporous Pt by electrochemical deposition and dissolution of PtZn alloy in zinc chloride-1ethyl-3-methylimidazolium chloride (ZnCl2-EMIC) ionic liquid [18]. Note that the Lewis acidic ZnCl2-EMIC ionic liquids (which contain more than 33 mol% ZnCl2) are potentially useful for the electrodeposition of zinc-containing alloys [19 – 22]. The advantages of the method include 1) the nano-channel size can be manipulated by varying the quantity of the PtZn surface alloy through the electrodeposition charge, 2) no corrosive acids or bases are used for the de-alloying and 3) extremely high working temperature is not required. In this study, the glucose oxidation on such Pt electrodes was thoroughly evaluated. Potential applications of these electrodes in direct sensing of glucose with high selectivity are discussed. The advantages over many nonH 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

772

C.-H. Chou et al.

enzymatic glucose sensors are the use of environmental friendly green solvent to fabricate the nanoporous platinum and the fabrication method is relatively simple.

2. Experimental 2.1. Reagents and Chemicals d(þ)-Glucose (dextrose anhydrous) (SHOWA), l(þ)-ascorbic acid (AA, Wako), p-acetamidophenol (AP, Sigma), Na2HPO4 (SHOWA), NaH2PO4 (SHOWA), sulfuric acid (Fisher), and anhydrous zinc chloride (Aldrich, 99.99%) were used without purification. The EMIC was prepared and purified according to the method described previously [21, 23, 24]. Preparation of the porous electrode was conducted inside a nitrogen-filled vacuum atmospheres glove box (Vacuum Atmospheres, HE-493-SB). The methods used to estimate the moisture and oxygen content of its atmosphere have also been described previously [25]. The oxygen and water levels of the atmosphere were kept below 1 ppm. The 40  60 mol% ZnCl2-EMIC melt was prepared by mixing 40% of ZnCl2 and 60% EMIC in a flask and then heated the mixture to 90 8C for 2 days. The as-prepared melt was a colorless liquid at temperatures near 40 8C.

2.2. Electrode Preparation Polished Pt wires (1 cm, r ¼ 0.1 cm) which have been immersed in 0.2 M NaOH solution and rinsed with deionized water with a specific resistivity of 18.2 MW/cm were cathodically treated in the 40 – 60 mol% ZnCl2-EMIC ionic liquid at 120 8C at a potential ( 0.22 V vs. Zn(II)/Zn) where bulk deposition of zinc occurs. During the electrodeposition of zinc, the color of the Pt electrode surface remained silver white. Subsequent anodic treatment was conducted at 1.2 V, where the PtZn alloy was stripped, until the anodic current dropped to zero. The PtZn-coated electrode surface turned black immediately after the start of the anodic treatment, indicating the formation of a porous surface. During anodic treatment, the PtZn alloy was selectively dissolved from the Pt electrode, resulting in the formation of nanochannels on the platinum surface. Note that, by inserting the as-prepared Pt wire into a PE tube, the geometric surface area of the Pt electrode can be manipulated for electroanalysis.

2.3. Electrochemical Experiments Electrochemical experiments of analyte testing were performed using a CHI 832a electrochemical analyzer (CH Instruments, USA). All electrochemical measurements were done in a three-electrode system with an Ag/AgCl (3 M KCl) and a Pt wire as reference electrode and counter electrode, respectively. The surface areas of the Pt electrodes were determined by measuring the areas under the hydrogen adsorption/desorption peaks of the cyclic voltamElectroanalysis 20, 2008, No. 7, 771 – 775

mograms at a scan rate of 100 mV/s in 2 M sulfuric acid solution. A conversion factor of 210 mC/cm2 was used to determine the electrode area [26]. The diffusion characteristics of the nanoporous Pt were compared to that of smooth Pt by checking the cyclic voltammetric peak currents and peak separations of 1.0 mM potassium ferricyanide solution. The nanoporous Pt electrode was evaluated as a glucose sensor in aerated 0.1 M phosphate buffered solution (PBS). Amperometric curves were obtained in a quiescent solution a few seconds after stopping the stirring that was required to mix the materials added (i.e., glucose, AA, and AP). Current changes 90 s after adding glucose to concentrations of 1, 2, 3, 4, and 5 mM were treated as specific responses to the glucose in the solution. The signals for AA and AP were measured by adding 0.1 mM of AA and AP sequentially and reading the current changes 90 s later. We confirmed that the AP signal was unaffected by the presence of AA.

3. Results and Discussion 3.1. Characterization of the Nanoporous Pt The effective surface area of the fabricated nanoporous Pt electrode was characterized by measuring the roughness factor (RF). Cyclic voltammetry was used to characterize the electrochemical properties of the nanoporous Pt electrodes. Figure 1A shows the cyclic voltammograms for the as-prepared nanoporous Pt electrodes with a RF of (a) 5 and (b) 67 in 2 M H2SO4 at a scan rate of 20 mV/s. As can be seen, the electrode with a RF of 67 shows a much higher current response compared to that of a RF of 5. It is well-known that the current response in the potential range from  0.1 to 0.2 V (vs. Ag/AgCl) originates from the hydrogen adsorption/desorption of hydrogen adatoms. Whereas, the anodic oxidation of the Pt film starting at ca. 0.6 V is due to the formation of platinum oxide that is subsequently reduced, as indicated by the appearance of a reduction peak at 0.65 V in the negative potential scan. The integrated intensity of the peaks represents the number of Pt sites available for hydrogen adsorption/desorption, and so the actual surface area. Obviously, the surface area effect (i.e., RF of electrode) can indeed reflect in the current response as indicated in Figure 1A. The microstructure of the as-prepared electrodes was also examined with SEM. As mentioned earlier, the merit of fabricating nanoporous Pt by electrochemical formation and dissolution of a PtZn surface alloy is that the channel size produced on the resulting porous Pt surface is easily tunable by the amounts of the electrochemically generated PtZn surface alloy. Figures 1B and 1C clearly illustrate the difference in the microstructure of the two dealloyed Pt electrodes. Overall, an increase in the PtZn surface alloy during the electrodeposition step produces a more porous surface with larger channels as well as channels of smaller size after de-alloying. The gorge-like structure of nanoporous Pt with an upper dimension of ca. 450 nm for both

www.electroanalysis.wiley-vch.de

H 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

773

Nonenzymatic Glucose Sensor Using Platinum Electrodes

Fig. 1. A) Cyclic voltammograms for the as-prepared porous Pt electrodes with RF of a) 5 and b) 67 in a solution of 2 M H2SO4 at a scan rate of 20 mV/s. SEM pictures for Pt electrode with RF of 5 (B) and 67 (C), respectively.

Fig. 2. Cyclic voltammograms of the nanoporous Pt electrode (RF ¼ 67) in a) blank, b) 5 mM glucose, and c) 10 mM glucose. The supporting electrolyte was 0.1 M, pH 7.4 PBS and the scan rate was 20 mV/s.

RF ¼ 67 and RF ¼ 5 electrodes and the bottom dimension of RF ¼ 67 electrode is smaller than that of RF ¼ 5 electrode. The crack as well as the depth of RF ¼ 67 electrode is also more than those of RF ¼ 5 electrode. Most importantly, diffusion-controlled behavior similar to those of a wellpolished flat Pt surface is observed when a redox couple like ferricynanide-ferocynaide is employed for voltammetric analysis. In general these features of the resulting Pt films are consistent with those previously reported for nanoporous Pt [27, 28].

discharging due to double layer capacitance occurs. Finally, at a potential of ca. 0.6 V, Pt-OH surface species start to form in the PBS. The Pt-OH species can oxidize the intermediate derived from the electrosorption of glucose, releasing free Pt active sites for the direct oxidation of glucose. The glucose signal being entirely suppressed [5]. The effect of chloride ion was thus further studied. Based on the cyclic voltammograms obtained, the peak at 0.6 V was virtually the same in the absence/presence of 0.15 M NaCl, indicating the nanoporous Pt electrode can retain sufficient sensitivity in the presence of chloride ion. In the present investigation, the performance of the nanoporous Pt electrode toward the oxidation of glucose has been optimized by recording the amperometric response. The effect of detection potential on the amperometric detection of glucose was first studied. The response current of the proposed electrode was found to abruptly increase to steady-state values upon addition of glucose with a very quick response time of ca. 4 s. As shown in Figure 3, the response sensitivity of glucose increases significantly with the potential increasing from 0.4 to 0.5 V. Note that, in cyclic voltammetry measurements using the nanoporous Pt electrode, the oxidation peaks of AA and AP appeared at around 0.1 V and 0.4 V vs. Ag/AgCl, respectively. Since amperometric measurement at a potential of > 0.1 V was positive enough to fully oxidize AA, the sensitivity was thus relatively consistent for AA than that of AP. On the other hand, because 0.4 V was close to the oxidation potential of AP, a sharp increase in current response was hence observed beyond this detection potential. Although a relatively higher current response was observed at 0.5 V vs. Ag/ AgCl, considering the fact that the amperometric measurement at 0.4 V vs. Ag/AgCl shows the most consistent in peak response for glucose, it was thus used in subsequent study. Note that the amperometric response at a detection potential of 0.4 V was also unaffected by chloride ion.

3.2. Electroanalysis of Glucose at the Nanoporous Pt Electrode The effective amperometric response to glucose is usually characterized in neutral media simulating physiological conditions. Therefore, the electrocatalytic activity of the nanoporous Pt electrode toward the oxidation of glucose in pH 7.4 PBS was first investigated by cyclic voltammetry. Figure 2 shows the cyclic voltammograms observed in the presence and absence of glucose at a scan rate of 20 mV/s. With glucose in solution, three enhanced anodic current peaks at around  0.4, 0.0, and 0.6 V vs. Ag/AgCl attributed to the hydrogen region, double layer region, and Pt oxide region, respectively, was detected. Note that the observation in good agreement with that reported earlier [28]. Most importantly, the three peaks increased accordingly as the glucose concentration increased from 5 to 10 mM verifying the electrocatalytic activity of the nanoporous Pt electrode toward glucose. Specifically, the first current peak at around  0.4 V should be due to the electrosorption of glucose to form as adsorbed intermediate, releasing one proton per glucose molecule. The second peak at ca. 0.0 V comes from the electrochemical oxidation of glucose on Pt surface in neutral media in the potential range where only charging or Electroanalysis 20, 2008, No. 7, 771 – 775

www.electroanalysis.wiley-vch.de

H 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

774

C.-H. Chou et al.

Fig. 3. Effect of detection potential on the amperometric responses of glucose (2 mM), ascorbic acid (100 mM), and 4acetamidophenol (100 mM) at the nanoporous Pt electrode (RF ¼ 151). The supporting electrolyte was 0.1 M, pH 7.4 PBS.

As mentioned earlier, the nanoporous Pt structures can be used to discriminatively enhance amperometric responses of glucose. To confirm this, the amperometric responses of the nanoporous Pt electrode (RF ¼ 151) and of a wellpolished flat Pt electrode (RF ¼ 2.6) to glucose, AA and AP in pH 7.4 PBS were compared. As can be seen in Figure 4A, the well-polished flat Pt electrode was very sensitive to AA and AP; but responded only negligibly to glucose. Since amperometric measurement at 0.4 V was positive enough to fully oxidize AA, the sensitivity was thus much higher for AA than that of AP. On the other hand, the nanoporous Pt electrode surface responded to glucose with a linear dependence. Most importantly the interference from AA

and AP is very effectively prevented by using the proposed nanoporous Pt electrode. The nanoporous Pt electrodes with various RF were then systematically studied to measure the signals for glucose, AA and AP. Figure 4B shows selective signal amplification by the porous surfaces. The results clearly indicated the dependence of the responses of the nanoporous Pt electrode to glucose on the roughness factors; whereas, the responses of the nanoporous Pt electrode to AA and AP were little affected. Consequently, the effective electrode area enlarged specifically enhanced the faradaic current so that selectivity to glucose over AA and AP was remarkably improved. Overall, high RF is essential for good sensitivity, but the selectivity is mainly originated from the morphology of Pt electrode. The signal of AA (or AP) is constant regardless of the RF of Pt electrode, indicating that it is the morphology not RF to increase the selectivity. Figure 5 shows the amperometric response of the nanoporous Pt electrode with a RF of 151 under experimental conditions of pH 7.4 PBS and detection potential at 0.4 V vs. Ag/AgCl. As can be seen, the calibration curve is linear in the range of 0 – 10 mM glucose with a correlation coefficient of 0.9976 and a sensitivity of 291.0 mA/cm2 mM (S/N ¼ 3). This value is much larger than the value of 31.3 mA/cm2 mM reported previously by using a three-dimensionally ordered, macroporous Pt electrode [29], indicating that our electrode can give higher sensitivity for glucose detection. Furthermore, this sensor also shows fast response, e.g. the response reaches 95% of the steady-state value within 4 s.

4. Conclusions In this study, we have demonstrated the attractive feature of using ambient temperature ionic liquid (i.e., green solvent) for fabricating of nanoporous Pt electrode. The surface of

Fig. 4. A) The amperometric responses for sensing glucose (1 – 20 mM), ascorbic acid (AA, 100 mM) and 4-acetamidophenol (AP, 100 mM) at the nanoporous Pt electrodes of RF ¼ 2 and 151, respectively. B) The effect of roughness factor on the signals for 1 mM glucose, 100 mM AA and 100 mM AP. The supporting electrolyte was 0.1 M, pH 7.4 PBS with a detection potential of 0.4 V. Electroanalysis 20, 2008, No. 7, 771 – 775

www.electroanalysis.wiley-vch.de

H 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

775

Nonenzymatic Glucose Sensor Using Platinum Electrodes

6. References

Fig. 5. Hydrodynamic responses (A) and the corresponding calibration curve (B) for various concentrations of glucose detected at the nanoporous Pt electrodes of RF ¼ 151 in pH 7.4 PBS. The linear range is 0 – 10 mM glucose at a detection potential of 0.4 V with a correlation coefficient of 0.9976.

the nanoporous Pt electrode is mechanically and chemically stable, and its surface can be easily regenerated by electrochemical cleaning. It provides nonenzymatic selectivity over representative interfering species at a potential that allows the oxidation of such interfering materials. The selectivity appears to be good enough for clinical application without enzyme or an additional outer membrane. The nanoporous Pt electrode also retains sufficient sensitivity in the presence of chloride ion and is thus useful for application in physiological fluids. In future, we hope to extend the application to detect the whole blood glucose concentration in diabetes patient.

5. Acknowledgements The authors gratefully acknowledge financial support from the National Science Council of Taiwan. This work is supported in part by the Ministry of Education, Taiwan under the ATU plan.

Electroanalysis 20, 2008, No. 7, 771 – 775

[1] Y. B. Vassilyev, O. A. Khazova, N. N. Nikolaeva, J. Electroanal. Chem. 1985, 196, 105. [2] B. Beden, F. Largeaud, K. B. Kokoh, C. Lamy, Electrochim. Acta 1996, 41, 701. [3] I. T. Bae, E. Yeager, X. Xing, C. C. Liu, J. Electroanal. Chem. 1991, 309, 131. [4] S. Ernst, J. Heitbaum, C. H. Hamann, J. Electroanal. Chem. 1979, 100, 173. [5] Y. Sun, H. Buck, T. E. Mallouk, Anal. Chem. 2001, 73, 1599. [6] B. Gollas, J. M. Elliott, P. N. Bartlett, Electrochim. Acta 2000, 45, 3711. [7] G. S. Attard, J. G. Glyde, C. G. Goltner, Nature 1995, 378, 366. [8] G. S. Attard, C. G. Goltner, J. M. Corker, S. Henke, R. H. Templer, Angew. Chem. 1997, 36, 1315. [9] J. M. Elliott, G. S. Attard, P. N. Bartlett, N. R. B. Coleman, D. A. S. Merckel, J. R. Owen, Chem. Mater. 1999, 11, 3602. [10] J. M. Elliott, P. R. Birkin, P. N. Bartlett, G. S. Attard, Langmuir 1999, 15, 7411. [11] S. A. G. Evans, J. M. Elliott, L. M. Andrews, P. N. Bartlett, P. J. Doyle, G. Denuault, Anal. Chem. 2002, 74, 1322. [12] A. H. Whitehead, J. M. Elliott, J. R. Owen, G. S. Attard, Chem. Commun. 1999, 331. [13] G. S. Attard, M. Edgar, C. G. Goltner, Acta Mater. 1998, 46, 751. [14] G. S. Attard, P. N. Bartlett, N. R. B. Coleman, J. M. Elliott, J. R. Owen, Langmuir 1998, 14, 7340. [15] P. R. Birkin, J. M. Elliott, Y. E. Watson, Chem. Commun. 2000, 1693. [16] J. M. Elliott, J. R. Owen, Phys. Chem. Chem. Phys. 2000, 2, 5653. [17] A. J. Bard, L. R. Faulkner, Electrochemical Methods: Fundamentals and Applications, Wiley, New York 2001. [18] J.-F. Huang, I-W. Sun, Chem. Mater. 2004, 16, 1829. [19] J.-F. Huang, I-W. Sun, J. Electrochem. Soc. 2002, 149, E348. [20] J.-F. Huang, I-W. Sun, J. Electrochem. Soc. 2003, 150, E299. [21] M. C. Lin, P. Y. Chen, I-W. Sun, J. Electrochem. Soc. 2001, 148, C653. [22] P. Y. Chen, I-W. Sun, Electrochim. Acta 2001, 46, 1169. [23] J. Robinson, R. A. Osteryoung, J. Electrochem. Soc. 1980, 127, 122. [24] J. S. Wilkes, J. A. Levisky, R. A. Wilson, C. L. Hussey, Inorg. Chem. 1982, 21, 1263. [25] J. S. Y. Liu, I-W. Sun, J. Electrochem. Soc. 1997, 144, 140. [26] S. Trasatti, O. A. Petrii, Pure Appl. Chem. 1991, 63, 711. [27] S. Park, T. D. Chung, H. C. Kim, Anal. Chem. 2003, 75, 3046. [28] S. Park, H. Boo, T. D. Chung, Anal. Chim. Acta 2006, 556, 46. [29] Y.-Y. Song, D. Zhang, W. Gao, X.-H. Xia, Chem. Eur. J. 2005, 11, 2177.

www.electroanalysis.wiley-vch.de

H 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim