210

Full Paper

Development of an Enzymeless/Mediatorless Glucose Sensor Using Ruthenium Oxide-Prussian Blue Combinative Analogue Annamalai Senthil Kumar,a Pei-Yen Chen,a Shu-Hua Chien,b Jyh-Myng Zena* a

Department of Chemistry, National Chung Hsing University, Taichung 402, Taiwan *e-mail: [email protected] b Institute of Chemistry, Academia Sinica, Taipei 11529, Taiwan Received: March 1, 2004 Final version: April 11, 2004 Abstract Presented in this work is the first step towards an enzymeless/mediatorless glucose sensor. We first observed remarkable electrocatalytic oxidation of glucose using combinative ruthenium oxide (RuOx)-Prussian blue (PB) analogues (designated as mvRuOx-RuCN, mv: mixed valent) at ca. 1.1 V (vs. Ag/AgCl) in acidic media (pH 2 Na2SO4/ H2SO4). Individual RuOx and PB analogs failed to give any such catalytic response. A high ruthenium oxidation state (i.e., oxy/hydroxy-RuVII, E8
1.4 V vs. RHE), normally occurring in strong alkaline conditions at RuOx-based electrodes, was electrogenerated and stabilized (without any conventional disproportionation reaction) in the mvRuOx-RuCN matrix for glucose catalysis. Detail X-ray photoelectron spectroscopic studies can fully support the observation. The catalyst was chemically modified onto a disposable screen-printed carbon electrode and employed for the amperometric detection of glucose via flow injection analysis (FIA). This system has a linear detection range of 0.3 – 20 mM with a detection limit and sensitivity of 40 mM (S/N ¼ 3) and 6.2 mA/(mM cm2), respectively, for glucose. Further steps towards the elimination of interference and the extendibility to neutral pHs were addressed. Keywords: Glucose, Prussian blue, Ruthenium oxide, Screen-printed electrode

1. Introduction With the rare exception of the Pb-Ir-O pyrochlore catalysis [1], most electrode materials are not amenable for the generation of significant glucose oxidation signals due to very high over-potentials in acidic and neutral media. Simple metal and metal oxide materials, i.e., ruthenium oxide (RuOx), are useful for glucose catalysis but only under strong alkaline, i.e., pH > 12, conditions [2 – 13]. Nowadays, glucose oxidase (GOx) in conjuction with a mediator, i.e., transducer, such as ferricyanide or ferrocene is the most common way of fabricating glucose biosensors. Other mediators, such as Prussian blue (PB, FeFeCN), a bimetallic zeolite type of macro-polymeric network [14 – 17], and various PB analogues, such as CrFeCN, CoFeCN, CuFeCN, NiFeCN, and FeRuCN, i.e., ruthenium purple, RP, [18 – 22], have also been reported. To develop a system that can directly catalyze glucose oxidation without the aid of an enzyme nor under strong alkaline conditions is certainly attractive. The main contribution of our study is thus to elucidate accomplishing these goals through the formation of a RuOx and PB combinative structure of RuMCN (M ¼ Fe, Ru). The work presented in this paper represents the first step towards the development of an enzymeless and mediatorless glucose sensor using a RuMCN (M ¼ Fe, Ru) catalyst. Surprisingly, the RuMCN-based system is not studied as much as other PB analogues [23 – 34]. To emphasize the combined structure and properties of RuOx and PB, Electroanalysis 2005, 17, No. 3

henceforth the RuMCN will be designated as mvRuOxMCN [34]. Earlier, Cox and co-workers reported a mixed valent ruthenium oxide-cyanometallate complex for analytical sensing of alcohols, aldehydes, and nitroamines except carbohydrates, e.g., glucose [23, 25, 27, 31]. The catalytic mechanism is also unknown. Recently, our preliminarily attempt on the redox sensitive organic redox probing (glucose, ethanol, and formaldehyde) analysis on RuOx and mvRuOx-FeCN indicated an unusual oxidation state of the RuOx-PB based combinative material [34]. Since both alcohols and aldehydes can be easily oxidized at RuOx electrodes in wide pH range (1 – 14) [35 – 38], we thus suspect that a specific form of RuOx present in the zeolitePB network may cause the desired electrocatalytic performance. To answer this question, X-ray photoelectron spectroscopy (XPS) was used to probe the Ru oxidation states and surface functional groups (especially O1s atomic energy levels so far ignored in the mvRuOx-MCN studies [32]) of this material. A detailed investigation about the unusually high Ru oxidation state in the mvRuOx-RuCN films was proposed to be essential for the uniquely observed behavior. The obtained results were then interpreted based on wellestablished RuOx and PB chemistry. Finally, a disposable type of mvRuOx-RuCN modified screen-printed carbon electrode (designated as mvRuOx-RuCN/SPE) is effectively demonstrated for sensitive glucose detection in acidic media by flow injection analysis (FIA).


2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

DOI: 10.1002/elan.200403086

211

Enzymeless/Mediatorless Glucose Sensor

2. Experimental

2.3. Electrode Preparation

2.1. Reagents and Materials

Prior to the modification, a bare SPE is electrochemically pretreated for 10 continuous cycles at a scan rate (v) of 50 mV/s in pH 2 KCl/HCl solution followed by washing with copious amount of distiller water. The mvRuOx-RuCN/ SPEs were prepared either by potential cycling method (from  0.5 to 1.25 Vat v ¼ 20 mV/s, n ¼ 20) or by potentiostatic method (1.1 V, t ¼ 900 s) using 2 mM RuCl3 þ 2 mM Ru(CN) 3 6 solution in pH 2 KCl/HCl solution. Even though both methods shows quantitatively and qualitatively similar electrochemical behavior, the film prepared by the potentiostatic method was found to be more stable (may be due to a more compact structure) in electroanalytical applications. The PB- and RuOx-modified SPEs, designated as PB/SPE and RuOx/SPE, respectively, were prepared by potential cycling method with 2 mM each of FeCl3 and Fe(CN) 3 6 and 2 mM RuCl3, respectively, for 20 cycles in pH 2 KCl/HCl solution. Note that the preparation condition for RuOx/SPE was very similar to the method for the preparation of hydrous RuOx electrodes reported by Hu and Huang [44]. In order to avoid the stability problems, freshly prepared electrodes were used for each run in the CV studies. The relative standard deviation (RSD) for the fresh mvRuOxRuCN/SPEs in CV was ca. 3% (n ¼ 5). Catalyst modified electrodes were pretreated in their respective base electrolytes by potential cycling between  0.5 to 1.25 V (vs. Ag/AgCl) at 50 mV/s in pH 2 Na2SO4/H2 SO4 solution. For FIA, the mvRuOx-RuCN/SPE was equilibrated in pH 2 Na2SO4/H2SO4 carrier solution at an applied potential (Eapp) of 1.1 V for 300 s with a flow rate (Hf) of 50 mL/min. To improve the stability and workability of this setup, 2 mM of RuCl3 and Ru(CN) 4 6 was added in the FIA carrier solution. Since both the detection and preparation potentials are set at 1.1 V, there was no complication in the current response of FIA. The quantification of glucose was achieved by measuring the anodic amperometric signals at room temperature. Freshly prepared mvRuOx-RuCN/ SPE strips were used for each set of FIA experiments. A simplified equation of ERHE ¼ [(EAg/AgCl þ 0.06 pH) þ o E Ag=AgCl (0.22)] was used for adjusting the potential between EAg/AgCl and ERHE, so that the pH-dependent ruthenium oxide redox potentials can be compared to the reported standard redox potentials of ruthenium in RHE (where o E RHE ¼ 0.0 V).

Ruthenium trichloride hydrate, RuO2, potassium hexacyanoferrate, potassium cyanocuprateI, potassium tetracyano nickellateII hydrate, potassium hexacyano chromateIII, potassium hexacyano platinateIV, and potassium hexacyano ruthenateII were from Aldrich. d(þ)Glucose was from Sigma. Other chemicals employed were of analytical grade and used without any purification. Aqueous solutions were prepared with doubly distilled deionized water. Stock solutions (0.1 M) were allowed to mutate at room temperature (25 8C) for 24 h before experiments. Caution! Ruthenium, chromium, copper, nickel, and platinum cyanometallate complexes are toxic, so care must be taken during their handling.

2.2. Apparatus Cyclic voltammetric (CV) experiments were carried out using a CHI 406 electrochemical workstation (Austin, TX). The three-electrode system consists of a working electrode either modified on SPE (geometric area ¼ 0.19 cm2) or glassy carbon electrode (geometric area ¼ 0.07 cm2), an Ag/ AgCl reference electrode (RE-5, BAS), and a platinum auxiliary electrode. The disposable SPEs were purchased from Zensor R&D (Taichung, Taiwan). Since dissolved oxygen did not interfere with the working system, no deaeration was performed in the present study. The FIA system consists of a BAS microprocessor pump drive, a Rheodyne 7125 sample injection valve (20 mL loop) with interconnecting Teflon tube, and a BAS thin wall-jet electrochemical cell system. The incorporation of the working SPEs into the wall-jet type cell was similar to our earlier reported procedure [39]. XPS analysis (Perkin-Elmer 1600 ESCA) was performed with either RuO2 powder (pellet) or RuOx-, mvRuOxRuCN-, and mvRuOx-FeCN-modified SPEs using an Mg Ka radiation source (1253.6 eV) with 0.1 eV resolution. The pressure inside the analyzer was maintained at about 109 Torr during the measurements. Prior to the experiments, the binding energy (BE) was standardized with Au 4f3/2 (84.0 eV, fwhm ¼ 1.2 eV). The high-resolution spectra were performed under ambient conditions and averaged from a number of scans to increase the signal-to-noise ratio. The Ru3d5/2 BE value of 280.9 eV for RuIVO2 and N1s BE value of 397.8 eV for the mvRuOx-MCN were uniformly taken as internal standard in the XPS analysis as per the established method [40 – 43]. Quantitative XPS analysis was performed by using the Origin 6.0 professional graphic program to pick up the intensity maximum and BE values. Freshly potentiostatically-prepared mvRuOx-RuCN(1.1 V for 900 s) and electrochemically-cycled mvRuOx-FeCN (  0.5 to 1.25 V for 20 cycles) films were subjected for XPS analysis. Some of the traces of higher Ru oxidation states electrogenerated and stabilized in the internal network could be identified from the high-resolution XPS measurements. Electroanalysis 2005, 17, No. 3

3. Results and Discussion 3.1. Cyclic Voltammetric Studies Figure 1A shows the CV responses using PB/SPE, RuOx/ SPE, SPE in 2 mM Ru(CN) 4 6 , and mvRuOx-RuCN/SPE in the presence and absence of glucose in pH 2 Na2SO4/H2SO4 (I ¼ 0.1 M) solution at v ¼ 10 mV/s. As can be seen, only the mvRuOx-RuCN/SPE shows a profound anodic catalytic glucose oxidation signal at E cat pa ¼ 1.06 Vaccompanied with a decrease in the cathodic peak current. The observation
2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

212

Fig. 1. A) CV responses of a) PB(FeFeCN)/SPE, b) RuOx/SPE, c) SPE with 2 mM Ru(CN) 4 6 , and d) mvRuOx-RuCN/SPE systems in pH 2 Na2SO4/H2SO4 (I ¼ 0.1 M) solution without (dotted lines) and with (solid lines) 0.1 M glucose at v ¼ 10 mV/s. B) Typical hydrodynamic FIA responses of a) SPE, b) PB(FeFeCN)/SPE, c) RuOx/SPE, and d) mvRuOx-RuCN/SPE with 10 mM glucose in pH 2 Na2SO4/H2SO4 (I ¼ 0.1 M) carrier solution. Eapp ¼ 1.10 V (vs. Ag/AgCl); Hf ¼ 50 mL/min; sample loop ¼ 20 mL.

A. S. Kumar et al.

indicates the powerful oxidation ability of the RuOx-PB combinative system over its individual analogues. Normally, the glucose oxidation requires powerful oxidants, such as RuVIIO 4 (1.40 V vs. RHE), Ce4þ (1.45 V vs. RHE), IO 4 (1.65 V vs. RHE), MnO 4 (1.68 V vs. NHE), etc [45, 46], and this is indeed the case for the well-established RuOx mediation in alkaline condition [11, 13]. Yet, the mediated oxidation of glucose in acidic media using the mvRuOxRuCN analogue is quite unique and thus it is suspected that different RuOx sites in the mvRuOx-RuCN network may play a role in catalysis. Previous oxy/hydroxy-RuOx studies indicate the existence of three pH-dependent Ru redox transitions, namely, RuIV/RuIII, RuVI/RuIV, and RuVII/RuVI, in aqueous solution [46 – 54]. Unfortunately, the RuVII oxidation state always leads disproportionation to the formation of RuIVO2 and oxy/hydroxy-RuVI species in pH < 12 solutions [46, 48, 49]. The oxy/hydroxy-RuOx individual system is thus not powerful enough to be applied for glucose oxidation at pH < 12 [53]. Scheme 1 compares the overall redox characteristic of classical RuOx systems with the unusual mvRuOx-MCN redox mediation effect in acidic media. As per Scheme 1, the redox selective organic probes like ethanol and formaldehyde can get oxidized in a wide pH range of 1 – 14 on the classical RuOx systems by the RuVI/ RuIV redox state; while, glucose can be oxidized only in strong alkaline conditions of pH > 12 by the higher RuVII/ RuVI redox state. The key question that needs to be clarified is can the unusual mediation effect in acidic media be related to the classical oxy/hydroxy-RuVII/RuVI catalyzed glucose oxidation on the RuOx electrode [11, 13, 34]? This can be true

Scheme 1. Conceptional representation for the dynamics of the Ru redox transitions against the solution pH and their specific oxidative functions towards organic redox probes at RuOx (A) and mvRuOx-MCN (B). The Ru(VII) is only stable in strong alkaline pHs, while at pH < 12, it gets disproportionated to lower redox states of (Ru(VI) and/or Ru(IV)). Electroanalysis 2005, 17, No. 3


2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Enzymeless/Mediatorless Glucose Sensor

213

only if the oxy/hydroxy-RuVII, presumably electrogenerated at 1.06 V on mvRuOx-RuCN (A2/C2), can be stabilized without any disproportionation reactions in acidic media. The first direct evidence to support this hypothesis is that the glucose oxidation potential observed at 1.06 V (i.e., 1.40 V vs. RHE) lies exactly inside the reported RuVII/RuVI redox transition potential region (Fig. 1A). The fact that the other two relatively weak redox transitions at 0.7 V (A2’/C2’) and 0.0 V (A1/C1) at the mvRuOx-RuCN did not participate in the glucose mediation was also in accordance with the electrochemical behavior of PB [55, 56]. Moreover, the mvRuOx-RuCN can also effectively oxidize ethanol and formaldehyde just like the RuOx system (data not included). Overall, these preliminary results all indicate the electrogeneration of unusual oxy/hydroxy-RuVII occurs in acidic media without any disproportionation reactions. Further evidence will be discussed in the following sections on XPS experiments, potential segments analysis, and kinetics measurements.

3.2. Surface Analysis by XPS Figure 2 shows XPS survey scan of SPE, mvRuOx-RuCN/ SPE, mvRuOx-FeCN/SPE, and RuO2 powder (pellet) in the BE window of 1000 to 0 eV. An interesting thing noted here is the existence of an O1s peak in both the mvRuOx-RuCN and mvRuOx-FeCN films, as confirmed by the very poor O1s intensity-peak observed in a control experiment using an underlying SPE (Fig. 2). The importance of the O1s peak in the mvRuOx-MCN was not mentioned in an earlier structural characterization of the mvRuOx-MCN film by Cataldi et al. [32]. Since classical PB films should not contain internal network oxygen species, it is likely the RuOx-PB network leads to the observed characteristics. Figures 3 show high-resolution spectra of the N1s, O1s, Ru3d, and Ru3p core energy levels for the mvRuOx-MCN material. Usually quantitative deconvolution of Ru peaks is extremely difficult in the presence of carbon species, since the major Ru3d3/2 BE level at ca. 285.0 eV is mingled with the C1s level at ca. 284.6 eV. As various carbon species are expected in the case of mvRuOx-MCN, the study here is therefore restricted to qualitative analysis. The internal standard assignment of the N1s level at 397.8 eV with the mvRuOx-FeCN and mvRuOx-RuCN films is in agreement with those reported in the literature [32, 43]. The RuO2 standard sample subjected to high resolution XPS was then used for comparison except neglecting Ru3d3/2 due to the complication of co-existing C1s at the same energy level. The BE values of 280.9 and 282.1 eV observed at the Ru3d5/2 level for the RuO2 standard powder samples corresponds to RuIVO2 and RuVIO3, respectively, as they are close to the reported values of either 280.7 and 282.5 eV or 280.9 and 282.5 eV [40, 41]. Note that the O1s speciation analysis from XPS also indicates the presence of various oxidation states of Ru in RuO2 and will be discussed later. The mvRuOxMCN shows a significant positive BE shift, especially with Ru and O1s core energy levels, over those of the RuO2. The Electroanalysis 2005, 17, No. 3

Fig. 2. XPS survey scans of SPE and its modified electrodes and RuO2 powders (pellet) samples. cps: counts per second.

shift again has something to do with the high valence ruthenium oxidation species in the RuOx-PB combinatory analogues. Two BE maximum values of 281.3 and 282.3 eV at Ru3d5/2 shoulder were noticed for the mvRuOx-FeCN; while, for the mvRuOx-RuCN, three BE values of 281.6, 282.3 and 283.0 eV were obtained. In comparison with those of the RuO2, the second BE value at 282.3 eV can be assigned as the RuVI specie. The highest BE value of 283.0 eV should therefore correspond to a highly valent RuOx species higher than the RuVI state. Previously Kim and Winograd reported a Ru3d5/2 BE value at 283.3 eV for a RuVIIIO4 compound [40]. With the slightly lower BE value in this study, it is likely the RuVIIOx species reflects efficient catalytic current signals over other mvRuOx-MCN analogues. Finally, the lowest BE of 281.6 eV for Ru3d5/2 can be assigned as RuIIINClinkage in the mvRuOx-MCN network based on the following reasons. Firstly, a specific BE value at 708.2 eV corresponding to the Fe2p3/2 core energy level of the FeCNwas identified for the mvRuOx-FeCN analogue (Figure 2). Secondly, Meyers group observed a Ru III 3d5=2 BE value of 281.3 eV for a [(bpy)2(H2O)RuIIIORuIII(H2O)(bpy)2]4þ complex [57]. Thirdly, the mvRuOx-RuCN film possesses a PB type basic 3D-network having a high spin bare metal-ion complexed with the “N” terminal of the cyano group in its internal cubic structure [55]. The high-resolution XPS of O1s energy levels can also provide information about the RunþO species. Both the mvRuOx-FeCN and mvRuOx-RuCN films show qualitatively similar O1s responses with two diffused BE maximum
2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

214

A. S. Kumar et al.

bonding. This strong hydrogen bonding effect further helps the catalyst to stabilize in the internal structures (Scheme 2). In fact, for the case of ruthenium hydroxides (RuO(OH)2 (H2O)2, (Scheme 2), Belova et al. noticed a strong hydrogen bonding effect by XPS via the O1s energy level at ca. 534.9 eV [60]. Table 1 provides detailed information on BE values and possible components. Overall, three major components were identified in the mvRuOx-MCN network using classical literature and PB models: (i) RuIIINCM, (ii) RuIIIORu, and (iii) RuIIIORuVII/VI(¼O)(H2 O). Scheme 2 shows the proposed structure of the mvRuOx-MCN analogues having oxo, hydroxy, aqua ruthenium species in addition to RuORubridges based on the XPS studies. The A2/C2 peak with unusual catalytic behavior to glucose should thus originate from the RuIIIORuVII/VI(¼O)(H2O)species. A recent investigation on the corrosion protection of aluminum by chromate conversion coating described the formation of similar oxy/ hydroxy-CN bridges in the network [61 – 64]. In that process, complex was used as an accelerator to the Fe(CN) 3 6 increase chromate coating formation, and CrVIIO 4 acted as a self-healing agent for corrosion protection. The oxy-Cr group was found to form a covalent inter-linkage with FeCNresulting in its stabilization [63]. In addition, Os and V based cyano-metallates were also reported for imperfect PB structures having oxy/hydroxy functional groups [65, 66]. This information supports the possibility of the existence of an unusually high valent ruthenium oxidation in the RuOx-PB combinatory analogue. Presumably, strong cyano-bridges together with inter-molecular hydrogen bonding in the internal network of the mvRuOxMCN can stabilize the high-valent ruthenium species without any classical disproportionation reaction occurring. Fig. 3. High-resolution XPS responses of N1s, O1s, Ru3d, and Ru3p core energy levels for various samples.

values of 531.8 and 533.3 eV together with a third BE value at 533.1 and 534.4 eV, respectively. In comparison with that of RuO2, a substantial shift in the BE values was noticed in the present case. The 531.8 eV O1s response can be assigned as a bridging oxygen, BO, i.e., O, in the network of the RuORulinkage, as reported earlier [31, 32]. Meanwhile, the SiOSitype BO in silicate glass ((SiO2)0.7-x (Na2O)0.3(Fe2O3)x) and the industrial waste cinder having a complex CaOSiO2-Fe2O3 framework were also reported with BE values of ca. 532 eV and 531.2 eV, respectively [58, 59]. The second O1s energy level at 533.3 eV can be assigned as oxo-ruthenium species connected with RuVI, i.e., > RuVI¼O, with a presumably strong inter-molecular hydrogen bonding effect. The highest O1s energy level at 534.1 eV represents the fraction of oxygen species derived from oxy, hydroxy and aqua groups (chemisorbed); while, the energy level at 534.4 eV, is due to the fraction of oxygen species coupled with the high valent Ru VII oxidation state, i.e., > RuVII¼O, and in turn to the strong inter-molecular hydrogen Electroanalysis 2005, 17, No. 3

3.3. Catalyst Evaluation Since the working electrode was unstable at pH > 2, experiments were restricted to pH 2 only. We recently investigated the pH effect (pH 2 – 8) on a mvRuOx-FeCN/GCE system [34], and the A2/C2 catalytic peak current at ca. 1.05 V was found to start to sharply decrease from pH 2 ! pH 8 and at neutral pHs the mvRuOx-FeCN tends to convert into a less catalytic RuOx based secondary structures. Hence, further detailed pH study with the mvRuOx-RuCN is another subject as well as a separate work. Various ruthenium PB analogues with a general formula of mvRuOx-MCN (M ¼ Cr, Cu, Ni, Fe, Pt, and Ru) were investigated by cyclic voltammetry for glucose mediation in a pH 2 KC/HCl solution. Percentage of relative catalytic glu bulk current (i cat pa ¼ i pa  i pa Þ with respect to M ¼ Ru is in the order of Cr (1.2%) < Ni (2.7%) < Cu (3.7%) < Fe (22.7%) < Pt (65.5%) < Ru (100%). Note that no catalytic response in the pH 2 solution was observed for RuOx and PB and neither for PB in couple with GOx enzyme. Hence, it is the ability to form the combinative surface oxide structure in the PB framework that controls the glucose catalytic
2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

215

Enzymeless/Mediatorless Glucose Sensor

Scheme 2. Representative structures of various RuOx compounds. Cases (a) and (b) are from [51] and [60], respectively, and (c) & (d) are proposed structures for the RuOx-PB combinatory analogues. Table 1. XPS binding energy ( BE) data for different ruthenium oxide based systems. Bold letters corresponding to the proposed species. System

BE values (eV )

Component

Ru3d5/2

Ru3p3/2

O1s

RuO2 powder

280.9 282.1

462.9 465.7

RuO2 powde [a]

280.7, 280.9 282.5

462.7 466.1

mvRuOx – FeCN [b]

281.3 282.3

463.5 465.9

mvRuOx-RuCN [b]

281.6 282.3 283.0

463.7 465.9 467.1

529.6 530.7 533.0 529.4 530.7 533.2 531.8 533.3 534.1 531.8 533.3 534.4

RuIVO2 RuVIO3 OH, H2O [c] RuIVO2 RuVIO3 OH, H2O [d] RuIIIN¼CFe, RuIIIORu > RuVI¼O > RuVI¼O · ( H2O ) [e] > RuIIINCRuIII, RuIIIORu > RuORuVI¼O > RuORuVII¼O ( H2O ) [e]

[a] References [40, 41]. [b] BE data with a reference to 397. 8 eV of N1S standard level. [c] Physisorbed H2O with hydrogen bonding. [d] Reference [42]. [e] Chemisorbed H2O with strong hydrogen bonding.

activity with the RuOx-PB combinative analogue. As per the XPS studies, it is obvious that the mvRuOx-RuCN yielded relatively higher intensity count values for the surface O1s species over mvRuOx-FeCN (Fig. 3) and can thus partly validate this expectation. Since the mvRuOxRuCN film showed the highest sensitivity for glucose oxidation, it was chosen for the subsequent catalytic studies [34]. The effect of supporting electrolyte was next optimized under identical experimental conditions at the mvRuOxRuCN/SPE in various pH 2 (I ¼ 0.1 M) solutions of NaH2 PO4, NaClO4, KCl, NaCl, K2SO4, and Na2SO4. As summarized in Table 2, the i cat pa of the catalytic peak at ca. 1.1 V was found in the order of NaH2PO4 < K2SO4 < NaClO4 < KCl < Electroanalysis 2005, 17, No. 3

NaCl < Na2SO4. The surface interactive and diffusion effects of the cation and its counter anion are expected to cause such an alteration and in turn affect stability. It is wellknown that alkali metal ions can penetrate into the PB structures for charge neutralization and network stabilization [55]. Instead of alkali metal dependence, strong proton dependence results were observed for mvRuOx-RuCN [23]. Hence, it is expected that the combined insertion and intercalation properties of the proton and alkaline metal together with anionic counter ions may influence the variation in the stability values as shown in Table 2. As mentioned earlier, the effect of CV scan rate on the glucose mediation behavior can provide information on the unusual existence of a higher oxidation state. Figure 4
2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

216

A. S. Kumar et al.

Table 2. Effect of various buffer solutions (pH 2, I ¼ 0.1 M ) on the catalytic response of 0.1 M glucose at v ¼ 50 mV/s. No. 1 – 6 are prepared by potential cycling at  0.5 to 1.25 V (n ¼ 20) in KCl/HCl ( I ¼ 0.3 M ) pH 2 with 2 mM of RuCl3 þ Ru( CN ) 4 6 . No. 7 was obtained by potentiostatic method. [a] After ca. 15 min of continues cycles in corresponding solutions. No.

Buffer

E cat pa ( V )

i cat pa (mA )

% of loss during initial stabilization [a]

1. 2. 3. 4. 5. 6. 7.

NaH2PO4/H3PO4 NaClO4/HClO4 KCl/HCl NaCl/HCl K2SO4/H2SO4 Na2SO4/H2SO4 Na2SO4/H2SO4 ( E ¼ 1.10 V, t ¼ 900 s)

1.10 1.10 1.10 1.09 1.07 1.08 1.08

71.1 87.0 105.8 112.4 75.7 122.3 153.1

17.54 12.00 16.40 19.50 14.01 20.11 < 8.00

Fig. 4. Effect of scan rate on the oxidation of 0.1 M glucose at the mvRuOx-RuCN/SPE in pH 2 Na2SO4/H2SO4 (I ¼ 0.1 M) solution. cat cat vs. v1/2. The insert figures are a) log(i cat pa Þ vs. log(v), b) E pa vs. log(v), and c) i f

indicates that both ipa and E cat pa regularly increased upon glucose increasing v. Double logarithmic plots of i bulk vs. v pa and i pa resulted in slopes of 0.92 and 0.61 in the absence and presence of glucose, respectively, on mvRuOx-RuCN/SPE. This observation indicates that the charge-transfer mechanism switches from pure adsorption to diffusion controlled in the presence of glucose. A Tafel slope (ba) of 122 mV was indirectly measured using the E cat pa vs. log(v) plot (slope ¼ 0.061 V/decade in the low v region) as per the irreversible kinetics of E cat pa ¼ (ba/2) log(v) þ constant [67 – 69]. Based on the Tafel equation of ba ¼ 2.303RT/aanaF, where na is the Electroanalysis 2005, 17, No. 3

number of electrons involved in the rate-determining step (rds). Assuming that na ¼ 1 in the rds, the value of aa was thus found to be ca. 0.5 indicating a common symmetric energy profile for a mediating system [56]. cat 1/2 The catalytic current function (i cat f ¼ i pa /v ) for glucose oxidation was calculated and plotted against v1/2. As can be seen in Figure 4c, the i cat f shows a maximum at an initially low v of 5 mV/s and parabolic behavior at high v. In other words, the accelerated cross-exchange electron transfer reaction is operated only at low v, indicating a typical example of an EC type catalytic mechanism, in which the charge transfer is
2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

217

Enzymeless/Mediatorless Glucose Sensor

Scheme 3. EC type of general catalytic mechanism

Fig. 5. Typical potential segment analysis response of the mvRuOx-RuCN film in 0.1 M Na2SO4/H2SO4 of pH 2 solution. Potentials were scanned either at anodic (I) or at cathodic directions (II and III). III) Enlarged view of IIb at v ¼ 10 mV/s (a) and a response at v ¼ 50 mV/s (b).

coupled with an irreversible chemical reaction to regenerate the starting electroactive species [67, 68, 70]. In fact, solution phase studies of perruthenate (RuVIIO 4 Þ in alkaline conditions resulted in a similar current profile and mechanism [48]. The electrogeneration of oxy/hydroxy-Ru(VII) acts as an intermediate in the proposed general mechanism for the catalytic oxidation of glucose as shown in Scheme 3: In the above scheme, the first step involving the mvRuOVII/VIx-RuCN redox system and the second step with the interaction of glucose and electrogenerated mvRuOVIIx-RuCN were considered electrochemical (E) and chemical (C) reactions, respectively. To validate Scheme 3, the oxidized product was further analyzed using fast atom bombardment mass spectra in Na2SO4/H2SO4. Regardless of the unstable molecular ion peaks, complicated rearrangements, and bond breakage steps, specific m/z values of 220 and 236 with ca. 20% of relative intensity were noticed for D-glucose and the oxidized product, respectively [45]. The calculated m/z difference of 16 in the above case Electroanalysis 2005, 17, No. 3

presumably corresponds to D-gluconic acid as an oxygenated compound.

3.4. Potential Segment Analysis In order to get information about the nature of the electrogenerated oxy- and/or hydroxy-RuVII/RuVI redox state, a potential segment analysis was performed at a low scan rate as depicted in Figure 5. As observed, the nature of the applied potential window influences the catalytic peak at 1.1 V but not the low potential redox peak at 0 V. Anodic forward sweep experiments in the potential window of  0.5 ! 1.2 V resulted in a positive peak potential shift on the A2/C2 redox peak; while the A1/C1 system is almost unaltered. It is interesting to observe that upon the negative sweep experiments starting from 1.2 to  0.5 V; obvious qualitative and quantitative changes were observed in A2/ C2 response. The cathodic sweep experiments from 1.2 !
2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

218 1.1 V revealed no faradic responses. For the case of the 1.2 ! 0.8 V scan experiment, an interesting crossover current response with the A2/C2 was observed (Fig. 5IIIa). Experiments with scans 1.2 ! 0.4 V, 1.2 ! 0.2 V, and 1.2 !  0.5 V showed no such obvious changes. Hence, the particular crossover experiments at ca. 1.0 V attenuate the time dependent electro-generation of some intermediate species. Note that a faster scan rate potential analysis showed no such interesting results (insert Fig. 5IIIb). It is most likely that the electrogenerated oxy/hydroxy-RuVII species at 1.06 V is involved with a slow-water oxidation reaction in the absence of any organic analyte, resulting in the formation of lower Ru oxidation states (such as RuIV and/or RuVI). The situation is similar to the case of perruthenate ions in alkaline conditions [48]. Moreover, the cathodic sweep experiments at starting potentials more than 1.06 V may easily deactivate the film to lower oxidation states. Ultimately, the nature and direction of the sweep rate is very critical to achieving the unusual ruthenium oxidation state and thus catalysis. It is noteworthy that in our recent investigation with the mvRuOx-FeCN film did not show any such crossover reaction and the potential dependent redox current values [34]. It is obvious from the XPS that the mvRuOx-FeCN material is oriented with less amount of oxy and/or oxyhydroxy surface functional groups (O1s response) and in turn to the poor RuVII stabilization. Hence, the mvRuOx-FeCNs feature is different from that of mvRuOxRuCN.

A. S. Kumar et al.

Fig. 6. CV response with increasing concentration of glucose at the mvRuOx-RuCN/SPE in pH 2 Na2SO4/H2SO4 (I ¼ 0.1 M) solution at v ¼ 10 mV/s. The insert figures are plots of a) i cat pa vs. [glucose] and b) log(i cat pa Þ vs. log[glucose].

3.5. Concentration Effect of Glucose The effect of glucose concentration on the mvRuOx-RuCN was investigated by CV at a quasi-steady state condition (v ¼ 10 mV/s). As shown in Figure 6, the i cat pa values regularly increased against the concentration of glucose in the range of 0 – 100 mM with a slope of 1.12 mA/mM (R ¼ 0.997). Five repeated experiments with freshly prepared mvRuOxRuCN/SPE strips yielded similar catalytic current profiles with a RSD value of 3.5% indicating reliability of the presented system. The reaction order can also be calculated from the slope of the log(i cat pa Þ vs. log[glucose] plot. A slope of 0.97 indicates a first order reaction of the glucose oxidation on the mvRuOx-RuCN/SPE. This observation suggests an EC mechanism with Michaleis-Menten (MM) reaction pathway as illustrated in Scheme 4 [54, 71, 72].

Scheme 4. Michaleis Menten reaction pathway. G ¼ Glucose; GO ¼ gluconic acid Electroanalysis 2005, 17, No. 3

3.6. Kinetic Analysis Initial experiments were carried out with a glassy carbon rotating disk electrode (GCE-RDE) modified with a mvRuOx-RuCN film in a pH 2 Na2SO4/H2SO4 solution scanning at a rate of 5 mV/s from  0.5 ! 1.2 V. Unfortunately, there is no rpm effect in the presented case. It was expected that there was no kinetic limitation for the masstransfer of the analyte from bulk to the electrode surface under hydrodynamic conditions and most probably, the mass-transfer assisted chemical reaction is relatively faster than the electron-transfer reaction at mvRuOx-RuCN with glucose (Scheme 2). At this stage, the RDE analysis is not useful for further mechanistic estimation. Alternatively, we used a quasi-steady state CV (QSCV) approach at a scan rate of 10 mV/s for further kinetic analysis. Note that same approach was used for the MM analysis [11], and in our own results with lead ruthenate pyrochlore, we demonstrated the QSCV for many biological substances [54]. Recently, Lyon derived a theoretical model for a microheterogeneous catalytic system with detailed information about individual rate constants of the reaction pathways in MM analysis [71]. Later, they extended their model for glucose oxidation studies under alkaline conditions using RuO2-modified graphite-epoxy electrodes [11, 71]. Here we adopt the same model for glucose catalysis at mvRuOx
2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

219

Enzymeless/Mediatorless Glucose Sensor Table 3. Michaleis-Menten ( MM ) kinetic data on ruthenium oxide based electrodes. Modified electrode

Graphite-epoxy RuO2 (450 8C ) [f] PVC-RuO2 (300 8C) [g] mvRuOx-RuCN [e] mvRuOx-RuCN [h]

Analyte

pH

E cat pa ( V vs. RHE )

GRu (mol/cm2)

MM rate constants ( LB plot) Km (mmol/cm3)

kc (s1)

Glucose

14

0.35 [b] (1.41)

1.65  106

0.13

0.06

Glucose Glucose Ethanol

14 2 2

0.35 [a] (1.43) 1.06 [b] (1.40) 1.00 [b] (1.34)

0.73  106 2.31  109 5.00  109

55.00 37.08 2.80

0.79 0.60 32.0

k’ME (102 cm/s) 83.00 1.80 0.04 0.006

Linearity (mM ) [c]

0–1 0 – 20 0 – 100 0 – 270 [d]

[a] vs. SCE; [b] vs. Ag/AgCl; [c] by CV analysis; [d] by amperometric i-t curve analysis; [e] this work; [f] reference [11]; [g] reference [13]; [h] reference [30].

RuCN and compare the heterogeneous kinetic rate constant with that of classical RuO2. As per the MM pathway, the electrogenerated catalyst, mvRuVIIOx-RuCN, first equilibrates with glucose (G) to form a high-energy intermediate complex [(-mvRuVIIOx-RuCN)-G] with a rate constant of KM. In the second step, the intermediate complex decomposes into gluconic acid and mvRuVIOx-RuCN at a first order rate constant of kc (considered as rds for the overall reaction). Finally, the catalyst was regenerated back to the active form using an applied potential of 1.06 V (vs. Ag/ AgCl), as shown in Scheme 4, at an electrochemical rate constant of k’ME. Similar kinetic studies have already demonstrated for the ethanol oxidation at mvRuOxRuCN (in acidic solution) [30], glucose oxidation at RuO2 (in alkaline condition) [11, 13], and many biological molecules using lead ruthenate pyrochlore (under wide pHs) [54]. Linewear-Burk (LB) equation, 1/i cat pa ¼ Km/nFAkcGRu[glucose] þ 1/nFAkcGRu, is further used for the calculation of the rate constants [71, 72]. In the above equation n ¼ total number of electrons involved in the reaction, i.e., 2, GRu ¼ surface coverage of the active ruthenium sites (calculated as 2.31  109 mol/cm2 for the anodic peak at 1.06 V (vs. Ag/AgCl) by CV at v ¼ 10 mV/s in blank solution), and other symbols have their own significance. Based on the LB expression, a plot of 1/i cat pa vs. 1/[glucose] yielded a linear response of 1/ipa ¼ 0.0178 mA1 þ 0.660 mA1/mM1[glucose] with a regression coefficient of 0.997. The calculated kc, Km, and k’ME (¼ kcGRu/Km) are listed in Table 3 and earlier MM kinetics results at other RuOx-based systems are also shown for comparison. The catalytic rate constant kc is very close to earlier glucose oxidation results at graphite-epoxy RuO2 (450 8C) composite and PVC-RuO2 (300 8C) electrodes at strong alkaline conditions [11, 13]. However, the kc for ethanol oxidation [30] is much higher than that for glucose oxidation at the mvRuOx-RuCN. The relatively fast kc for ethanol is acceptable, since the alcohol oxidation can be mediated even by a lower RuVI oxidation state [49, 53]. In classical synthetic organic chemistry, the alkaline medium is highly unsuitable for glucose oxidation because of epimer and fragmentation formation in addition with mutarotation [45]. Moreover, H2O2 in aerobic condition (which is the key detecting step in Ist generation glucose biosensor) can also oxidize glucose in the homolytic (with radical step) pathways [45]. Hence, a very wide linear detection range in the Electroanalysis 2005, 17, No. 3

Fig. 7. Hydrodynamic FIA responses at various Eapp (a) and Hf (b) for the detection of 10 mM glucose at the mvRuOx-RuCN/ SPE in pH 2 Na2SO4/H2SO4 (I ¼ 0.1 M) carrier solution.

presence case may be due to the absence of such complications. All the voltammetric investigation validates the useful behavior of the mvRuOx-RuCN towards enzymeless glucose mediation. The catalytic behavior is further utilized as a sensitive detection scheme for glucose by FIA.

3.7. Analytical Applications The advantage of the mvRuOx-RuCN/SPE over classical SPE, SPE/PB, and SPE/RuOx for glucose detection is clearly demonstrated in Figure 1B. In accordance with that of CV studies, only the mvRuOx-RuCN/SPE shows response to glucose in FIA. The hydrodynamic factors like applied potential (Eapp) and flow rate (Hf) were systematically optimized. As shown in Figure 7, the FIA signals reach
2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

220

A. S. Kumar et al.

Fig. 8. A) FIA responses for different concentration of glucose, (B) FIA responses for 5 mM each of glucose (G), ascorbic acid (AA), dopamine (DA), and uric acid (UA), and (C) 0.3 mM glucose after 24 h of running time. Carrier solution also contains 2 mM RuCl3 and Ru(CN) 4 6 . All the other conditions are the same as in Figure 1B.

a maximum at an Eapp of 1.10 V. The observation is in agreement with the CV results for the electrogeneration of ruthenium higher oxidation states at 1.06 V. While choosing an Eapp of 1.10 V, the effect of Hf was studied in the range of 50 – 1000 mL/min. Again, because of the electrogeneration rate of oxy/hydroxy RuVII, a lower Hf resulted in a sensitive response to glucose. Hence, Eapp ¼ 1.1 Vand Hf ¼ 50 mL/min were chosen as optimized conditions for further glucose sensing experiments. To increase the stability, the strategy of adding reactants for ethanol detection at similar type of electrode was adopted here [30]. Figure 8 shows the typical FIA response for increasing concentration of glucose by adding diluted concentration of 2 mM each of RuCl3 and Ru(CN) 3 6 in the carrier buffer solution. The calibration plot was linear in the window of 0.3 – 20 mM with slope and regression coefficient of 1.2 mA/mM (6.12 mA/(mM/cm2)) and 0.9952, respectively. Continuous injection of 0.3 mM glucose (n ¼ 10) resulted in a RSD value of 4.3% indicating a detection limit (S/N ¼ 3) of 40 mM. Even though the slope is not as good as the PB/GOx based electrode (50 mA/(mM/cm2) [17], it is comparable to some cyclometalated RuII complex/GOx (ca. 0.042 mA/ mM) electrodes in neutral pHs [73, 74] and a RuO2 modified carbon paste electrode (8.9 mA/(mM/cm2)) in strong alkaline solution [12]. It is interesting to note that the hydrodynamic rotating disk electrode (RDE) failed to give any response to glucose in an earlier section, while the hydrodynamic FIA was more successful here. Nature of the hydrodynamic flow is critical for this observation. In the FIA experiments, a stationary mvRuOx-RuCN/SPE system was coupled with very slow hydrodynamic flow rate (50 mL/min), where as at RDE, both the bulk solution and the electrode systems were under dynamic state. Hence, the slow hydrodynamic flow of the carrier solution in FIA can allow proper crossover kinetics Electroanalysis 2005, 17, No. 3

near the electrode/electrolyte interface and further leads to the catalytic oxidation current signals. Interference effect was studied with 5 mM each of glucose, ascorbic acid (AA), dopamine (DA), and uric acid (UA) in FIA under optimized conditions. As expected, all compounds show considerable interference to glucose determination (Fig. 8). It is no doubt that all these substances are simply oxidized at the electrode due to the existence of powerful oxidant of RuVII. If such a detection system is combined with separation techniques like HPLC or CE, one can easily get complete sequencing of many biological compounds. Finally, the stability of the working electrode was checked for 24 h run and a very small deviation of < 3% was noticed. Overall, the mvRuOxMCN offers a new catalytic system for glucose oxidation and detection. Further step toward the elimination of interference and the extendibility to neutral pHs are in progress in our laboratory. It is noteworthy that in our recent publication on ruthenium purple (FeRuCN, RP) anchored cinder, we had demonstrated selective H2O2 detection and in turn as a glucose biosensor (in couple with GOx enzyme) in pH 7 buffer [75]. The silicate type of cinder-base matrix helps the formation of new hybrid PB analogues and thus stabilizes it in neutral pHs [59]. Moreover, Tosflex over-layer coating was found to be effective to prevent the common interferences. Hence, in future work, we planned to anchor the mvRuOx-RuCN catalyst into silicate type of macronetwork and protect with permselective membrane to mimics the GOx enzyme. Meanwhile, regarding the practical sensing application of the mvRuOx-RuCN, it is quite suitable for the analysis (in couple with separation techniques) of glucose in gastric systems in stomach, where strong HCl (0.1 M) is produced, and classical enzyme and other simple electrochemical systems are in vain for the assays.
2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

221

Enzymeless/Mediatorless Glucose Sensor

Recently, Loetanantawong et al reported electrocatalytic oxidation of tetracyclines on mvRuOx-RuCN/GCE at ca. 1.00 V (vs. Ag/AgCl) in pH 1 H2SO4/(0.5 M)K2SO4 solution [76]. The important thing to be noted is the proposed oxidation state of RuIV/III in series with earlier paper [28], which is in agreement with our studies based on the recent organic-redox probes [34] and the XPS investigations. This also indicates the lack of information about the Ru redox states with the RuOx-PB based combinative analogues.

4. Conclusions A combinative material of ruthenium oxide and Prussian blue was for the first time found to mediate glucose oxidation, without the aid of glucose oxidase, in acidic media. Classical PB and RuOx electrodes alone did not show such catalytic current signals. The catalytic behavior of mvRuOx-RuCN in acid media is comparable to the classic RuO2 electrode in strong alkaline conditions due to the mediation of the high oxy/hydroxy-RuVII/RuVI redox state (ca. 1.4 V vs. RHE). Applied kinetic parameter for the glucose oxidation on the mvRuOx-RuCN also supports the electrogeneration of a high valent oxy/hydroxy-ruthenium oxidation state and in turn to the cross-exchange electron transfer to glucose. Comparative XPS analysis of the RuO2 with the mvRuOx-MCN further supports this conclusion. Termination of strong CNlinkages with internal hydrogen bonding effects may be the likely reason for the stabilization of high valent ruthenium in the PB matrix. The catalytic oxidation was further utilized for the sensitive analytical application by flow injection analysis (FIA) at high-dead volumes. The FIA analytical results are comparable to those of classical RuO2 (at alkaline conditions) and some enzyme-based biosensing (neutral media) systems. Since the preparation route of the film is easy, simple and in the form of disposable screen-printed electrodes, the procedure can be used for real practical applications. Most importantly, the direct and irreversible mediation of glucose on mvRuOx-RuCN/SPE should also be extendable to both analytical, synthetic organic, fuel cell (especially to bio-fuel cell with glucose as a fuel source) and heterogeneous catalysis applications in acidic media. Numerous applications are readily being imagined and further work is in progress. Enzymatic sensing of glucose is a highly selective analysis, while the artificial inorganic-enzyme analogue based mimicking methods can always offer acceptable selectivity if the electrode, solution, and instrumental parameters are properly tuned. Since the present work is the first investigation for the development of suitable catalyst system with nonalkaline condition operation, no attention was paid to the elimination of interference and particular for neutral pH operation. Further work in couple with suitable hybrid matrix and permselective membranes is possible to shape the catalyst to mimics the GOx enzyme function. Further steps are all in progress in this direction as well as to the practical sensing applications. Electroanalysis 2005, 17, No. 3

5. Acknowledgement The authors gratefully acknowledge financial support from the National Science Council of the Republic of China.

6. References [1] C. Mullens, M. Pikulski, S. Agachan, W. Gorski, J. Am. Chem. Soc. 2003, 125, 13602. [2] S. V. Prabhu, R. P. Baldwin, Anal. Chem. 1989, 61, 852. [3] Y. Xie, C. O. Huber, Anal. Chem. 1991, 63, 1714. [4] G. Wittstock, A. Strubing, R. U. Szargan, G. Werner, J. Electroanal. Chem. 1998, 444, 61. [5] T. R. I. Cataldi, E. Desimoni, G. Ricciardi, F. Lelj, Electroanalysis 1995, 7, 435. [6] S. Berchmans, H. Gomathi, P. Rao, Sens. Actuators B 1998, 50, 156. [7] I. G. Casella, M. R. Guascito, Electrochim. Acta 1999, 45, 1113. [8] I. G. Casella, E. Desimoni, A. M. Salvi, Anal. Chim. Acta 1991, 243, 61. [9] N. Xonoglou, I. Moumtzis, G. Kokkinidis, J. Electroanal. Chem. Interfacial Electrochem. 1987, 237, 93. [10] T. R. I. Cataldi, C. Campa, G. E. De Benedetto, Fresenius J Anal.Chem. 2000, 368, 739. [11] M. E. G. Lyons, C. A. Fitzgerald, M. R. Smyth, Analyst 1994, 119, 855. [12] J. Wang, Z. Taha, Anal. Chem. 1990, 62, 1413. [13] V. Dharuman, K. C. Pillai, Bull. Electrochem. 1999, 11, 476. [14] A. A. Karyakin, O. V. Gitelmacher, E. E. Karyakina, Anal. Chem. 1995, 67, 2419. [15] A. A. Karyakin, Electroanalysis 2001, 13, 813. [16] M. P. OHalloran, M. Pravda, G. G. Guilbault, Talanta 2001, 55, 605. [17] A. A. Karyakin, E. A. Kotelnikova, L. V. Lukachova, E. Karyakina, J. Wang, Anal. Chem. 2002, 74, 1597. [18] M.-S. Lin, W.-C. Shih, Anal. Chim. Acta 1999, 381, 183. [19] J. Z. Zhang, S. J. Dong, Anal. Lett. 1999, 32, 2925. [20] I. L. Mattos, L. Gorton, T. Laurell, A. Malinauskas, A. A. Karyakin, Talanta 2000, 52,791. [21] S. Milardovoc, I. Kruhak,. D. Ivekovic, V. Rumenjak, M. Tkalcec, B. S. Grabaric, Anal. Chim. Acta 1997, 350, 91. [22] C.-F. Chen, C.-M. Wang, J. Electroanal. Chem. 1999, 466, 82. [23] J. A. Cox, P. J. Kulesza, Anal. Chem. 1984, 56, 1021. [24] P. J. Kulesza, J. Electroanal. Chem. 1987, 220, 295. [25] J. A. Cox, T. J. Gray, Anal. Chem. 1990, 62, 2742. [26] R. T. Kennedy, L. Huang, A. Atkinson, P. Dush, Anal. Chem. 1993, 65, 1882. [27] W. Gorski, J. A. Cox, Anal. Chem. 1994, 66, 2771. [28] T. R. I. Cataldi, C. Campa, D. Centonze, Anal. Chem. 1995, 67, 3740. [29] J. A. Cox, K. S. Alber, C. A. Brockway, M. E. Tess, W. Gorski, Anal. Chem. 1995, 67, 993. [30] T. R. I. Cataldi, D. Centonze, E. Desimoni, V. Forastiero, Anal. Chim. Acta 1995, 310, 257. [31] W. Gorski, J. A. Cox, J. Electroanal. Chem. 1995, 389, 123. [32] T. R. I. Cataldi, A. M. Salvi, D. Centonze, L. Sabbatini, J. Electroanal. Chem. 1996, 406, 91. [33] P. J. Kulesza, B. Grzybowska, M. A. Malik, M. Chojak, K. Miecznikowski, J. Electroanal. Chem. 2001, 512, 110. [34] A. Senthil Kumar, J.-M. Zen, Electroanalysis 2004, 16, 1211; A. Senthilkumar, J.-M. Zen, ChemPhysChem. 2004, 5, 1227. [35] E. J. M. OSullivan, J. R. White, J. Electrochem. Soc. 1989, 136, 2576.
2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

222 [36] J.-W. Kim, S.-M. Park, J. Electrochem. Soc. 1999, 146, 1075. [37] R. De Andrade, R. Adalgisa, M. P. Donate, P. P. D. Alves, C. H. V. Fidellis, J. F. C. Boodts, J. Electrochem. Soc. 1998, 145, 3839. [38] D.-T. Shieh, B.-J. Hwang, J. Electrochem. Soc. 1995, 142, 816. [39] J.-M. Zen, H.-P. Chen, A. Senthil Kumar, Anal. Chim. Acta 2001, 449, 95. [40] K. S. Kim, N. Winograd, J. Catal. 1974, 35, 66. [41] J. Augustynski, L. Balsenc, J. Hinden, J. Electrochem. Soc. 1978, 125, 1093. [42] I. M. Kodinstev, S. Trasatti, M. Rubel, A. Wieckowski, N. Kaufher, Langmuir 1992, 8, 283. [43] M. Oku, J. Chem. Soc. Faraday Trans. 1993, 89, 743. [44] C.-C. Hu, Y.-H. Huang, J. Electrochem. Soc. 1999, 146, 2465. [45] H. S. El Khadem. Carbohydrate Chemistry. Monosaccharides and Their Oligomers, Academic Press, New York 1988, p. 144. [46] A. J. Bard, R. Parsons, J. Jordan. Standard Potentials in Aqueous Solutions, IUPAC, Marcel Dekker, New York 1983, p. 413. [47] S. Trasatti, in Electrochemistry of Novel Materials (Eds: J. Lipkowski, P. N. Ross), VCH, New York 1994, ch. 5 and reference therein. [48] K. W. Lam, K. E. Johnson, D. G. Lee, J. Electrochem. Soc. 1978, 125, 1069. [49] L. D. Burke, J. F. Healy, J. Electroanal. Chem. 1981, 124, 327. [50] L. D. Burke, D. P. Whelan, J. Electroanal. Soc. 1979, 103, 179. [51] A. Senthil Kumar, K. C. Pillai, J. Solid State Electrochem. 2000, 4, 408. [52] A. Senthil Kumar, Ph. D. Dissertation, A. C. College of Technology, University of Madras, India 1998. [53] J.-M. Zen, A. Senthil Kumar, J.-C. Chen, J. Mol. Catal. A: Chemical 2001, 126, 177. [54] J.-M. Zen, A. Senthil Kumar, Acc. Chem. Res. 2001, 34, 772. [55] K. Itaya, I. Uchida, V. D. Neff, Acc. Chem. Res. 1986, 19, 162. [56] J.-M. Zen, A. Senthil Kumar, H.-W. Chen, Electroanalysis 2001, 13, 1171.

Electroanalysis 2005, 17, No. 3

A. S. Kumar et al. [57] D. A. Geselowitz, W. Kutner, T. J. Meyer, Inorg. Chem. 1986, 25, 2015. [58] A. Mekki, D. Holland, C. F. McConville, M. Salim, J. Noncryst. Solids 1996, 208, 267. [59] A. Senthil Kumar, P.-Y. Chen, J.-M. Zen, Electroanalysis, 2004, 16, 242. [60] I. D. Belova, R. R. Shifrina, Yu. E. Roginskaya, A. V. Popov, T. V. Varlamova, Soviet Electrochem. 1987, 23, 1138. [61] L. Xia, R. L. McCreery, J. Electrochem. Soc. 1999, 146, 3696. [62] L. Xia, E. Akiyama, G. Frankel, R. L. McCreery, J. Electrochem. Soc. 2000, 147, 2556. [63] G. S.Frankel, R. L. McCreery, J. Electrochem. Soc. Interface 2002, 10, 34. [64] P. Campestri, E. P. M. Van Westing, A. Hovestad, J. H. W. De Wit, Electrochim. Acta 2002, 47, 1097. [65] J. A. Cox, B. K. Das, J. Electroanal. Chem. 1987, 233, 87. [66] N. R. de Tacconi, K. Rajeshwar, R. O. Lezna, Chem. Mat. 2003, 15, 3046. [67] A. J. Bard, L. R. Faulkner, Electrochemical Methods. Fundamentals and Applications, 2nd ed., Wiley, New York 2001. [68] Electroanalytical Methods. Guide to Experiments and Applications (Ed: F. Scholz), Springer, Berlin, 2002. [69] J.-M. Zen, A. Senthil Kumar, J.-C. Chen, Anal. Chem. 2001, 73, 1169. [70] M. H. Pournaghi-Azar, H. Razmi-Nerbin, Electroanalysis 2001, 13, 465. [71] M. E. G. Lyons, in Advances in Chemical Physics Polymeric Systems (Eds: I. Pyrigogine, S. A. Rice), Wiley, NewYork 1996, p. 297. [72] A. Senthil Kumar, J.-M. Zen, Electroanalysis 2002, 14, 671. [73] A. D. Ryabov, V. S. Sukharev, L. Alexandrova, R. Le Lagadec, M. Pfeffer, Inorg. Chem. 2001 40, 6529. [74] E. S. Ryabov, V. N. Goral, E. Csçregi, B. Mattiasson, A. D. Ryabov, Angew. Chem. Int. Ed. 1999, 38, 804. [75] J.-M. Zen, A. Senthil Kumar, C.-R. Chung, Anal. Chem. 2003, 75, 2703. [76] B. Loetanantawong, C. Suracheep, M. Somasundrum, W. Surareungchai, Anal. Chem. 2004, 76, 2266.


2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim