Analytica Chimica Acta 576 (2006) 1722

Studies on electrochemical behaviors of acyclovir and its voltammetric determination with nano-structured lm electrode
Fang Wang a,b , Lin Chen a , Xiaoxia Chen a , Shengshui Hu a,b,
b

Department of Chemistry, Wuhan University, Wuhan 430072, PR China State Key Laboratory of Transducer Technology, Chinese Academy of Sciences, Beijing 100080, PR China Received 24 September 2005; received in revised form 7 December 2005; accepted 9 December 2005 Available online 31 January 2006

Abstract A multi-wall carbon nanotubes (MWNTs)-dihexadecyl hydrogen phosphate (DHP) lm-coated glassy carbon electrode (GCE) was fabricated, and the electrochemical behaviors of acyclovir on the MWNTs-DHP lm-coated GCE were investigated by using cyclic voltammetry (CV), linear sweep voltammetry (LSV), electrochemical impedance spectroscopy (EIS) and chronocoulometry (CC). The oxidation peak current of acyclovir increased signicantly and the peak potential shifted negatively at the MWNTs-DHP lm-modied GCE, compared with that at a bare GCE. The results showed that this nano-structured lm electrode exhibited excellent enhancement effects on the electrochemical oxidation of acyclovir. Consequently, a simple and sensitive electroanalytical method was developed for the determination of acyclovir. The oxidation peak current was proportional to the concentration of acyclovir from 8.0 108 to 1.0 105 mol/L. The detection limit was about 3.0 108 mol/L for 60 s accumulation at 0.00 V. The proposed method was demonstrated by using acyclovir tablets and the result was satisfying. 2005 Elsevier B.V. All rights reserved.
Keywords: Acyclovir; Multi-walled carbon nanotubes (MWNTs); Chemically modied electrodes (CMEs); Voltammetry; Glassy carbon electrode (GCE); Nanostructured lm

1. Introduction Acyclovir (9-2-hydroxyethoxymethyl guanine, ACV) plays a key role in the therapy of virus diseases. It is an acyclic nucleoside analogue antiviral drug used to treat infection with herpes simplex viruses (HSV), hepatitis B virus (HBV) and varicella zoster viruses (VZV) [1,2]. This drug has the best safety prole of all antivirals licensed so far [3]. Intravenous, oral and to a lesser extent topical formulations of acyclovir provide signicant therapeutic benet in therapy of virus diseases and no serious side effects of the drug are identied. Based on above description the quantitative determination of ACV become very important and has been widely studied. In recent years, high performance liquid chromatography (HPLC) [47] and radioimmunoassay (RIA) [8,9] have been mainly used for the determination of this antiviral drug. However, these methods are complicated and tedious, such as the optimization of chromatogram conditions and pretreatment of samples for HPLC

Corresponding author. Tel.: +86 27 8721 8904; fax: +86 27 6875 4067. E-mail address: sshu@whu.edu.cn (S. Hu).

system, and the radioactive pollution and expensive cost for RIA method. Compared with HPLC and RIA, the electrochemical methods based on chemically modied electrode (CME) look very simple, clear, convenient and inexpensive. To our knowledge, the studies on the electrochemical behaviors of acyclovir and its quantitative detection at carbon nanotubes (CNTs) lmcoated glassy carbon electrode have not been reported. CNTs modied electrodes, in particular, hold great promise for increasing the selectivity, sensitivity and reproducibility of voltammetric measurements. In the past decade CNT, including two distinct types of structures: the single-walled carbon nanotubes (SWNTs) and multi-walled carbon nanotubes (MWNTs) has become the focus of considerable interest on the basis of their large specic area, their unique architecture, remarkable mechanical and electrical properties [1012]. A lot of research ndings suggest that CNTs which could increase the electron transfer rate have full potential applications based on the fabrication of chemically modied electrodes and were mainly used to construct sensing lms of various biomolecules, including dopamine (DA), thyroxine and so on [1318]. The larger specic area could produce higher sensitivity because of its great adsorption capability to organic molecules. Nitric acid shortened

0003-2670/$ see front matter 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.aca.2005.12.023

18

F. Wang et al. / Analytica Chimica Acta 576 (2006) 1722

CNTs contains plenty of carboxyl or other oxygenic groups at their terminus. These probably so-called quinone-like groups as those at the glassy carbon electrode (GCE) surface could result in an increase in the effective electrochemically active surface area of CNTs towards electrochemical reaction. Combined with high electronic conductivity, better biocompatibility and useful mechanical properties all qualities of CNTs could promote electron transfer between reactant molecules and electrode surface. Based on the convenience of electrochemical method and characteristics of CNTs, and so one kind of MWNTs lm-coated GCE was developed in present work. The electrochemical behaviors of acyclovir had been investigated at this modied electrode. The results showed that the sensitivity for the determination of acyclovir increased markedly at MWNTs lm-coated GCE. Some experiment conditions were optimized and a voltammetric method for determination of acyclovir in acyclovir tablets was proposed and the result was satisfying. This method was convenient and available because of its higher sensitivity, lower detection limit and low costs. 2. Experimental 2.1. Apparatus and reagents Dihexadecyl hydrogen phosphate (DHP) was obtained from Fluka Chemical Reagent Co. and stored at 18 C. MWNTs were provided by Chengdu Organic Chemicals Co., Ltd., Chinese Academy of Sciences. Acyclovir (Wuhan Renfu Pharmaceutical Co., China) was dissolved in doubly distilled water to form 2.0 103 mol/L standard solution. Hundred milligrams of acyclovir tablets (Hubei Keyi Pharmaceutical Co., China) were dissolved in 20 mL doubly distilled water, and separated via centrifugal effect. The supernatant fractions were collected for determination. Other chemicals used were analytical reagents including all inorganic salts used to prepare buffer solution, interferences and all surfactants. All the chemicals were used without further purication and all the solutions were prepared with double distilled water. All the electrochemical measurements were performed on a CHI 830 electrochemical analyzer (Shanghai Chenhua Co., China) in a three-electrode system. The working electrode was a MWNTs lm-coated GCE. A saturated calomel electrode (SCE) and a Pt wire were used as the reference and counter electrodes, respectively. The electrochemical impedance spectroscopy (EIS) was carried out with the EG&G Model 273 electrochemical workstation and EG&G Model 5210 lock-in amplier (Princeton Applied Research, PAR, USA) powered by Echem Software. 2.2. Preparation of the MWNTs-DHP lm-coated GCE Five milligrams of MWNTs and 5 mg DHP were dispersed together in 5 mL water to form a uniform suspension with the aid of ultrasonic agitation for 20 min. Prior to modication, the GCE was mechanically polished to a mirror nish with 0.05 m Al2 O3 slurry and cleaned by ultra-sonication in water for several minutes, then air dried. At last, 9 L of the MWNTs-DHP

suspension was dropped on the GCE surface and dried under an infrared lamp to form a stable MWNTs-DHP lm. Similarly, DHP lm was prepared by casting DHP suspension on the surface of GCE. 2.3. Assay procedure At the beginning of the experiment, the MWNTs-DHP lmcoated GCE was scanned by successive cyclic voltammetric sweeps in 10 mL citratesodium hydrogen phosphate buffer solution (pH 7.36) between 0.00 and 1.10 V at 100 mV/s. When the current became steady, a certain volume of acyclovir standard solution was then added into the electrochemical cell. The mixture solution was stirred for 60 s and kept quiet for 2 s at 0.00 V. The linear sweep voltammograms were recorded while the potential initially sweeps from 0.00 to 1.10 V. The electrochemical oxidation peak current was measured at 0.93 V. After each measurement the modied electrode was refreshed by successive cyclic voltammetric sweeps in blank buffer solution to get a reproducible electrode surface. 3. Results and discussion 3.1. Characterization of the MWNTs-DHP lm-coated GCE by EIS Fig. 1 shows the Nyquist plots (Z versus Z ) for the electrochemical impedance spectroscopy (EIS) of K3 Fe(CN)6 / K4 Fe(CN)6 at MWNTs-DHP lm-coated GCE, DHP lmcoated GCE and bare GCE. Only a semicircle with a large diameter at full frequency range is observed at the DHP lmcoated GCE, indicating a slow electron transfer rate between K3 Fe(CN)6 /K4 Fe(CN)6 and the electrode surface. Because the DHP is an insulator and DHP lm on GCE blocks electron transfer between the solution and the electrode surface. The Nyquist

Fig. 1. The electrochemical impedance spectroscopy (EIS) of 0.5 mM K3 Fe(CN)6 /K4 Fe(CN)6 in 0.1 M KCl at MWNTs-DHP lm-coated GCE ( ), DHP lm-coated GCE ( ) and bare GCE ( ). Frequency range used, 100 kHz to 0.1 Hz with signal amplitude of 5 mV rms at ve steps per decade.

F. Wang et al. / Analytica Chimica Acta 576 (2006) 1722

19

Fig. 2. The impedance responses of 1.0 l05 mol/L acyclovir at the MWNTsDHP lm-coated GCE ( ), DHP lm-coated GCE ( ) and bare GCE ( ). Frequency range used, 100 kHz to 0.1 Hz with signal amplitude of 5 mV rms at ve steps per decade.

Fig. 3. Successive cyclic voltammograms of 1.0 l05 mol/L acyclovir in citratephosphate buffer solution (pH 7.36) at MWNTs-DHP lm-coated GCE. Scan rate, 100 mV/s.

plot of K3 Fe(CN)6 /K4 Fe(CN)6 at bare GCE is formed of a semicircle followed by a straight line at high frequency range. Compared with the resistive effects of DHP lm, the MWNTsDHP lm-coated GCE shows much different characteristics and only a line appears at high frequency range. It is clear that the MWNTs can greatly increase the electron transfer rate. The impedance responses of 1.0 l05 mol/L acyclovir were studied. Corresponding results are illustrated in Fig. 2. At DHP lm-coated GCE and bare GCE similar Nyquist plots are obtained. They both are semicircles with a large diameter at full frequency range. It is concluded that the electron transfer between DHP lm-coated GCE or bare GCE and acyclovir molecules is not easy to be achieved. On the contrary the Nyquist plot of an analogous line at the MWNTs-DHP lm-coated GCE testies that the MWNTS lm increases the electron transfer rate between electrode and acyclovir and the electrochemical oxidation of acyclovir would become easy. 3.2. Electrochemical behaviors of acyclovir at MWSTs-DHP lm electrode Fig. 3 shows successive cyclic voltammograms of 1.0 l05 mol/L acyclovir in 10 mL citratephosphate buffer solution (pH 7.36) at the MWNTs-DHP lm-coated GCE. A well-dened oxidation peak (O1 ) is observed at 0.93 V at the rst anodic sweep from 0.00 to 1.10 V. On the reversal scan, no corresponding reduction peak appears. However, during following successive cyclic sweeps the peak current of O1 decreases greatly and nally disappears. It is resulted from the fact that the electrode surface is blocked by the adsorption of the oxidation products which reduces the effective reaction sites at the modied electrode surface. Fig. 4 shows the linear sweep voltammograms of 1.0 l05 mol/L acyclovir at different working electrodes including bare GCE (curve d), DHP lm-coated GCE (curve c) and MWNTs-DHP lm-coated GCE (curve a). Curve b in Fig. 5

describes the background current of MWNTs-DHP lm-coated GCE in citratesodium hydrogen phosphate buffer solution (pH 7.36) in the absence of acyclovir. The signicantly increased oxidation currents of acyclovir at the MWNTs-DHP lm-coated GCE conrms that MWNTs show highly effective enhancement to acyclovir because of its large specic surface area and special electrical properties, which makes for easier adsorption of acyclovir and provides enough effective reaction sites. Furthermore, the trait of porous of the MWNTs-DHP lm on the GCE surface makes the adsorption of acyclovir to the electrode surface easy and the concentration of that at modied electrode surface increases. Under the same experiment conditions, the oxidation peak of acyclovir almost disappears at bare GCE and DHP lm-coated GCE.

Fig. 4. Linear sweep voltammograms of MWNTs-DHP lm-coated GCE in the absence (curve b) and presence of 1.0 l05 mol/L acyclovir at bare GCE (curve d), DHP lm-coated GCE (curve c) and MWNTs-DHP lm-coated GCE (curve a) in citratesodium hydrogen phosphate buffer solution (pH 7.36). Scan rate, 100 mV/s.

20

F. Wang et al. / Analytica Chimica Acta 576 (2006) 1722

Fig. 5. Effects of solution pH on the peak potential for 1.0 l05 mol/L acyclovir at the MWNTs-DHP lm-coated GCE in citratesodium hydrogen phosphate buffer. Inset: Effect of solution pH on peak current.

3.3. Optimizing of some experiment conditions 3.3.1. The optimum of the supporting electrolyte The electrochemical responses of 1.0 l05 mol/L acyclovir at the MWNTs-DHP lm-coated GCE in different supporting electrolytes such as 0.1 mol/L phosphate buffer solution (PBS, pH 1.08.0), 0.2 mol/L sodium acetateacetic acid buffer solution (NaAcHAc, pH 3.55.6), boraxhydrochloric acid buffer solution (pH 7.69.2), boraxsodium hydroxide buffer solution (pH 9.212.3), citratesodium hydrogen phosphate buffer solution (mixing 0.2 mol/L sodium hydrogen phosphate and 0.1 mol/L citrate, pH 2.28.0) and 0.05 mol/L potassium acid phthalate standard buffer solution (pH 4.0) were examined by linear sweep voltammetric (LSV). It is found that the oxidation peak current of 1.0 l05 mol/L acyclovir becomes the highest in citratesodium hydrogen phosphate buffer solution (pH 7.4) and the voltammogram shape is well dened, so citratesodium hydrogen phosphate buffer solution (pH 7.4) is the most suitable electrolyte for the electrochemical oxidation of acyclovir. Different pH of the supporting electrolyte was tested to research the oxidation current responses of 1.0 l05 mol/L acyclovir at the MWNTs-DHP lm-coated GCE. The results are shown in Fig. 5 suggesting that both the peak current and peak potential change with the change of solution pH. In acid buffer solution the voltammogram shape is well dened, but the oxidation peak potential is too positive. With the increase of solution pH the peak potential shifts negatively then the voltammogram shape becomes unshapely. It is found that the peak current becomes the highest in buffer solution (pH 7.36) and the voltammogram shape is well dened. Thus, citratesodium hydrogen phosphate buffer solution (pH 7.36) was chosen as electrolyte for the detection of acyclovir. 3.3.2. Optimization for the amount of MWNTs According to above discussion on the linear sweep voltammograms of 1.0 l05 mol/L acyclovir at different working electrodes in Section 3.2, MWNTs can facilitate the electron

Fig. 6. Dependence of the peak current for 1.0 l05 mol/L acyclovir on the amount of MWNTs-DHP dispersion on the GCE surface in citratesodium hydrogen phosphate buffer (pH 7.36).

transfer rate between acyclovir and electrode surface to enhance its electrochemical oxidation responses at the MWNTs-DHP lm-coated GCE. So the impact of lm thickness on oxidation peak current has been studied. Thus the MWNTs-DHP lm thickness was determined by the volumes of MWNTs-DHP solution used in making the membranes on the GCE surface. Then as displayed in Fig. 6, generally, the oxidation peak current of 1.0 105 mol/L acyclovir increases quickly with the increase of the volumes of MWNTs-DHP solution cast on the GCE surface in the range of 09 L. When 9 L of 1 mg/mL MWNTs-DHP suspensions are cast on the electrode surface, the oxidation peak current become very high. Subsequently, further improving the volumes of MWNTs-DHP solution, the oxidation peak current has a slow increase. However, when the volumes of MWNTs-DHP solution cast on the electrode surface is more than 13 L, the oxidation peak current decreases lightly. On one hand, excessive thick membrane is not helpful to the adsorption of acyclovir and goes against the electron transfer of acyclovir within the MWNTs-DHP lms based on the unimproved mass transfer. On the other hand, the increase of volumes of MWNTs-DHP solution used in making the membrane would result in increase of the amount of DHP and conductivity of the membrane would be reduced correspondingly. Because DHP is an insulator and it could block the electron transfer. As a result, 9 L of 1 mg/mL MWNTs-DHP suspension was therefore suitable for the fabrication of MWNTs-DHP modied electrode. 3.3.3. Inuences of accumulation potential and time The effects of the accumulation potential and the accumulation time are investigated, respectively. The inuences of the accumulation potential on the oxidation of 1.0 105 mol/L acyclovir were estimated rstly. At closed circuits, the peak potential shifts positively when the accumulation potential changes from 0.60 to 0.80 V and the oxidation peak current

F. Wang et al. / Analytica Chimica Acta 576 (2006) 1722

21

of acyclovir gets the maximum value at 0.00 V. It is the adsorption accumulation that increases the concentration of acyclovir at modied electrode surface, then the oxidation peak current of acyclovir increases accordingly. It comes to the conclusion that large specic surface area of MWNTs brings on the great accumulation efciency. Therefore, an accumulation potential of 0.00 V was chose. As for the accumulation time, it signicantly affects the oxidation peak current of 1.0 105 mol/L acyclovir. The peak current of 1.0 105 mol/L acyclovir increases greatly within the rst 1 min and then enhances slowly. Maybe this is attributed to the saturated adsorption of acyclovir on the MWNTs-DHP lm-coated GCE surface. 3.3.4. Effect of scan rate and the solution pH The effects of scan rate on the oxidation peak current were studied by the linear sweep voltammetry (LSV). The oxidation peak current of 1.0 105 mol/L acyclovir shows a linear relationship with scan rate in the range of 75500 mV/s. This result shows that the electrode process is controlled by the adsorption step. As far as the totally irreversible electrode process controlled by the adsorption step is concerned, the relationship between the oxidation peak potential and scan rate is described by the following equation [19]: Ep = E0 + RT na F ln RTk0 na F + RT na F ln

Fig. 7. Chronocoulometry of 2.0 l04 mol/L acyclovir at MWNTs-DHP lmmodied GCE (curve a). Curve b stands for that in the blank solution. The inset shows the linear relationship between the charges (Q) and the square roots of times (t1/2 ) for the oxidation reaction (background subtracted). Initial potential 0.8 V, nal potential 1.0 V, and pulse width 9 s.

where E0 is formal potential, T the temperature, the transfer coefcient, na the number of the electron transferred in the ratedetermining step, k0 the electrochemical rate constant and F is the Faraday constant. The slope indicates that the value of na is 0.97. In addition the number of the proton transfer in the electrochemical oxidation of acyclovir can be calculated from the data in Fig. 5. The oxidation peak potential negatively shifts with the increase of solution pH and obeys the following equation: Ep = 1.274 0.048 pH (R = 0.9987) the slope of 0.048 mV/pH suggests that the number of the electron transferred in the oxidation of acyclovir equals that of proton. 3.4. Chronocoulometry The method of chronocoulometry was used to characterize the oxidation of acyclovir on the MWNTs-DHP lm-coated GCE in citratesodium hydrogen phosphate buffer solution (pH 7.36) and determine the diffusion coefcient D and Qads containing 2.0 104 mol/L acyclovir, according to the formula given by Anson [20]. Q = 2nFAcD 1/2 1/2 t 1/2 + Qdl + Qads where n is the number of electron involved in the oxidation of acyclovir, F the Faraday constant, A the surface area of the working electrode, c the concentration of acyclovir, D the diffusion

coefcient of acyclovir, Qdl the double-layer charge and Qads is the Faradaic charge due to the oxidation of adsorbed acyclovir. Other symbols have their conventional signicance. The results are depicted in Fig. 7. The subtraction of the background charge can eliminate the effect of double-layer charge Qdl in our experiment. The inset shows the linearized plot for Q versus t1/2 and from the slope of the line one can calculate the diffusion coefcient of 2.0 104 mol/L acyclovir in citratesodium hydrogen phosphate buffer solution (pH 7.36) to be 7.07 106 cm2 s1 . Additionally, Qads can be obtained by the difference of the intercepts of the plot of Q versus t1/2 in the presence and absence of acyclovir. Here Qads is 38.71 C, according the equation: Qads = nFA the value of the surface concentration corresponding to a monolayer at the electrode surface, , can be obtained as 2.79 109 mol cm2 . 3.5. Calibration and interferences 3.5.1. Calibration graph Under the optimized experiment conditions, the calibration curve for acyclovir in citratesodium hydrogen phosphate buffer solution (pH 7.36) at MWNTs-DHP lm-coated GCE was characterized by LSV, and the linear dynamic range was comprised between 8.0 108 and l.0 l05 mol/L in terms of the relationship between acyclovir concentration and the oxidation peak current. That relationship can be described with the following linear regression equation in the mentioned concentration range: Ip = 4.848c + 5.824 106 (R = 0.9956) A detection limit of 3.0 108 mol/L acyclovir was obtained with an accumulation for 60 s. After each measurement the

22 Table 1 Determination of acyclovir in tablets Sample 1 2 3 4 Declared (mol/L) 5.0 107 5.0 106 2.0 106 3.0 106

F. Wang et al. / Analytica Chimica Acta 576 (2006) 1722

Detected by this method (mol/L) 5.1 107 1.05 106 2.1 106 3.09 106

Recovery (%) 102 105 105 103

modied electrode was refreshed by successive cyclic voltammetric sweeps in blank solution to get a reproducible electrode surface. The relative standard deviation (R.S.D.) of 2.86% for 10 times parallel detections of l.0 l05 mol/L acyclovir, suggesting good reproducibility of MWNTs-DHP lm-modied GCE. By casting 9 L of 1 mg/mL MWNTs-DHP suspensions on the electrode surface every time, seven different modied electrodes were fabricated and the relative standard deviation for the detections of l.0 l05 mol/L acyclovir is 1.46%, revealing excellent repeatability of MWNTs-DHP lm-modied GCE. 3.5.2. Interferences The effects of some small biomolecules on the electrochemical oxidation of l.0 l05 mol/L acyclovir have been evaluated. l.0 l03 mol/L ascorbic acid, glucose, dopamine, l-cystine, 5.0 l05 mol/L heteroauxing and isoniazide have almost no inuences on the current responses of l.0 l05 mol/L acyclovir (signal change below 5%) but the inuences by 2.0 l05 mol/L xanthine, hypoxanthine and adenine are serious (signal change is more than 10%). Because they have similar electroactive group with acyclovir, their oxidation potentials are close to that of acyclovir and then affect the oxidation peak current of acyclovir. 3.6. Determination of acyclovir in acyclovir tablets This medicine contains the active ingredient acyclovir. The proposed method was applied to the determination of acyclovir in acyclovir tablets. The results are illustrated in Table 1. In our experiments, the concentration of acyclovir was calculated using standard additions method. The relative standard deviation of each sample for three times parallel detections is less than 4.0%. In addition, the recovered ratio on the basis of this method was investigated and the value is between 102.0 and 105.0%. The recovered ratio indicates that the determination of acyclovir using MWNTs-DHP lm-coated GCE is effective and sensitive. 4. Conclusion The oxidation peak current of acyclovir was greatly enhanced at MWNTs-DHP lm modied electrode. The electrode pro-

cess was adsorption-controlled and totally irreversible. Because MWNTs had large specic surface area, high electronic conductivity, useful mechanical properties and strong adsorptive properties, more reaction sites could be provided for the electrochemical oxidation of acyclovir. Based on unique structure of MWNTs such as hollow geometry, MWNTs-DHP lm on the electrode surface was uniform and porous. Because of these characteristics the adsorption of acyclovir to electrode surface became easy and the concentration of acyclovir on the electrode surface increased. Combined with excellent conductivity of MWNTs, all show that the sensitivity for determination of acyclovir increased markedly. The modied electrode was easy to prepare and the detection process for acyclovir was simple, clear and convenient. This method was proposed to detect acyclovir in acyclovir tablets and the result was satisfying. The method was convenient and available because of its higher sensitivity, lower detection limit and low costs. If the electrodes system could achieve its micromation it would be better applied to vivo analysis of acyclovir. Acknowledgement This research is supported by the National Natural Science Foundation of China (nos. 30370397 and 60571042). References
[1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] E.D. Clercq, J. Clin. Virol. 22 (2001) 73. J.J. OBrien, D.M. Campoli-Richards, Drugs 37 (1989) 233. P. Wutzler, R. Thust, Antiviral Res. 49 (2001) 51. H. Mascher, C. Kikuta, R. Metz, et al., J. Chromatogr. 583 (1992) 122. R. Boulieu, C. Gallant, N. Silberstein, J. Chromatogr. B 693 (1997) 233. K.J. Swart, H.K.L. Hundt, A.M. Groenewald, J. Chromatogr. A 663 (1994) 65. R.L. Smith, D.D. Walker, J. Chromatogr. 343 (1985) 203. S.M. Tadepalli, R.P. Quinn, D.R. Averett, Antimicrob. Agents Chemother. 29 (1986) 93. K.M. Skubitz, R.P. Quinn, P.S. Lietman, Antimicrob. Agents Chemother. 21 (1982) 352. S. Iijima, Nature 354 (1991) 56. J.P. Salvetat, J.M. Bonard, N.H. Thomson, A.J. Kulik, L. Forro, W. Benoit, L. Zuppiroli, Appl. Phys. A 69 (1999) 255. D.T. Colbert, R.E. Smally, Trends Biotechnol. 17 (1999) 46. P.J. Britto, K.S.V. Santhanam, V. Alonso, A. Rubio, P.M. Ajayan, Bioelectrochem. Bioenerg. 41 (1996) 121. M. Musamech, J. Wang, A. Merkoci, Y.H. Lin, Electrochem. Commun. 4 (2002) 743. K.B. Wu, J.J. Fei, S.S. Hu, Anal. Biochem. 318 (2003) 100. K.B. Wu, X.B. Ji, J.J. Fei, S.S. Hu, Nanotechnology 15 (2004) 287. J.J. Davis, J.C. Richard, O. Allen, H. Hill, J. Electroanal. Chem. 440 (1997) 279. H. Luo, Z. Shi, N. Li, Z. Gu, Q. Zhang, Anal. Chem. 73 (2001) 915. E. Laviron, J. Electroanal. Chem. 52 (1974) 355. F.C. Anson, Anal. Chem. 36 (1964) 932.