RuO₂ electrode surface effects in electrocatalytic oxidation of glucose

Abstract

We have studied the electrocatalytic activity of RuO₂-PVC film electrodes, fabricated using RuO₂ powders prepared at five different temperatures, viz., 300, 400, 500, 600 and 700°C, for the oxidation of glucose in high alkaline media, 1 to 3 M NaOH. The RuO₂-PVC film electrodes have been first characterized in 1 to 3 M NaOH solution by cyclic voltammetry (CV) and rotating disc electrode (RDE) techniques in a wide potential range −1,100 to 450 mV (SCE), and three redox pairs representing Ru(IV)/Ru(III), Ru(VI)/Ru(IV) and Ru(VII)/Ru(VI) transitions have been identified. The voltammetric peaks at low sweep rates have been analyzed using surface activity theory formulated for interacting electroactive adsorption sites, and interaction terms have been evaluated. The total voltammetric surface charges have been analyzed as per Trassatti’s formalism with respect to their dependence on potential sweep rate, and charges associated with less accessible and more accessible surface sites have been calculated. For glucose oxidation, the results have indicated that RuO₂ (700°C)-PVC electrode shows two oxidation peaks in contrast to RuO₂ (300°C)-PVC electrode. Also, RuO₂ (700°C)-PVC electrode exhibits higher intrinsic electrocatalytic activity than the 300°C electrode, although the former possesses lower electrochemically active surface area. Additionally, kinetic analyses made from RDE results with reference to Michealis–Menten (MM) enzyme catalysis has shown that RuO₂ (700°C) electrode possesses extended glucose-sensing range in terms of MM kinetic constant, K _M , compared to other electrodes. Possible reasons for such differences in the behavior of the electrodes of different temperatures towards glucose oxidation are identified from studies on oxidation of glucose in solutions of different pH, oxidation of different glucose derivatives, and also from physicochemical results from BET, XRD, SEM, DTGA, XPS analysis of RuO₂ powder samples.

Introduction

The literature on the electrooxidation of carbohydrates, especially, glucose, shows that it can be oxidized directly on various metal electrodes, such as, Pt [1–3], Cu [4, 5], Au [6, 7], Ag [8], Ni [9, 10], Co [11] and on cobalt phthalocyanine-modified electrode [12]. In recent times a few conducting metal oxides like tungsten oxide [13], ruthenium oxide [14–17], and iridium oxide [18] have been studied for glucose oxidation revealing many useful and unique properties, like, reduced operation over potential and enhanced efficiency of the glucose oxidation reaction. In particular ruthenium dioxide has received special attention due to its high metallic conductivity, unusual stability, presence of Ru in several oxidation states (Ru(III), Ru(IV), Ru(VI), Ru(VII)) within the potential region in which solvent (water) is stable, and excellent activity towards electrocatalytic oxidation of carbohydrates and several other organic compounds [19–30].

Activity of pure metal electrodes on glucose [1–10] reveals the following aspects: glucose oxidation is a potential-dependent and multi-step multi-electron transfer reaction; the rate of oxidation reaction and the type of reaction intermediates formed depend on the nature and structure of the electrode material. For example, glucose was found to react differently on different single crystals of Pt [2] and Au [7], and in this respect still research is being conducted with many advanced in-situ spectroscopic techniques to identify the reaction intermediates to obtain intricate details of the oxidation mechanism. We expect a similar electrode structure-sensitive glucose oxidation reaction with RuO₂ electrode as well. With RuO₂-based electrode material it is significantly convenient to impart structural and morphological changes by varying the electrode preparation temperature [31–33].

Ruthenium-dioxide-coated electrodes were widely investigated, and RuO₂ electrodes are more traditionally prepared from its chloride salt by high temperature pyrolysis technique, wherein the oxide is directly formed on suitable substrates like Ti, Ta, Pt or silica glass, Ti being the most predominantly used [31, 32]. Literature reports [31] indicate that RuO₂ formation starts from 285°C, with complete formation at 450°C, and continuing up to 850°C. But note that in preparing RuO₂ coatings by direct thermal decomposition at temperatures higher than 500°C, problems arise at the contact between the support and the active layer, such as, interphase swelling, formation of fragile overlayers, interdiffusion etc. [31, 32]. Recently, we overcame this problem by initially preparing the RuO₂ powders at high temperatures even up to 700°C, and then coating a Pt electrode surface with a mixture of RuO₂ powders and PVC (80:20 w/w%) in THF solvent. This method appears to be more convenient, leading to a well-bonded and mechanically stable layer on Pt surface. RuO₂-PVC film electrodes were prepared previously by the above method in the temperature range 300 to 700°C and characterized in solutions of various pH, 1 to 14 [33].

In this paper, we report electrocatalytic oxidation of glucose in NaOH solutions of 1 to 3 M concentration at RuO₂-PVC film electrodes, fabricated using oxides prepared at five different temperatures viz., 300, 400, 500, 600 and 700°C, and illustrate that the reaction mechanism and the extent of oxidation of glucose, indeed, strongly depend on the morphological and structural features of these electrodes. Interestingly, higher intrinsic electrocatalytic activity and extended linearity for glucose sensing are observed for RuO₂ (700°C)-PVC electrode and not for RuO₂ (300°C)-PVC electrode, although the 700°C electrode possesses smaller surface charge and lower electrochemically active surface area. The possible reasons are exploited with supporting results from other non-electrochemical techniques like BET, X-ray diffractometry (XRD), scanning electron microscopy (SEM), differential thermogravimetry (DTGA), and X-ray photoelectron spectroscopy (XPS). In the first part of the work RuO₂-PVC film electrodes are characterized in 1 to 3 M high alkaline solutions, followed by electrooxidation of glucose using both CV and RDE techniques.

Experimental details

Chemicals

Glucose, NaOH and tetrahydrofuran (THF) of HPLC grade were obtained from Sd fine chemicals, India. RuCl₃.xH₂O was from Arora Matthey, India, and poly(vinylchloride) (PVC) was from Fluka. All the experimental solutions were prepared using doubly distilled water.

Experimental

Cyclic voltammograms were recorded using Wenking potentiostat (ST 72) coupled with a Wenking signal generator (VSG 83) and a Graptech X-Y-t recorder (WX 2300). RDE experiments were carried out using Analytical rotator, Pine Instruments USA. A home-made three-compartment glassy cell consisting of a main body (200 ml) and a smaller compartment, isolated from the main body by coarse glass frit containing a large surface area (4 cm²) Pt foil employed as auxiliary electrode, was used. The working electrode was placed in the center of the main body which also contained the reference electrode (SCE) mounted in a Luggin capillary to minimize the iR drop.

RuO₂-PVC film electrode preparation

RuO₂-PVC film electrodes were prepared as described elsewhere [33]. Briefly, RuCl₃.xH₂O was heated in a muffle furnace at preset preparation temperature in presence of a stream of O₂ for 12 h with intermittent cooling and grinding. Finally the oxide samples were water-washed thoroughly for the removal of trace quantities of Cl⁻, dried in air oven at 110°C for 12 h and stored in desiccator. RuO₂ powder was mixed with PVC in 4:1 (wt%) ratio in dry THF, and the paste like slurry was coated on to a pretreated Pt electrode of 5 mm diameter. It was then dried at room temperature for 2 days. The effective loading was around 40 mg cm⁻². Prior to coating, Pt was pretreated by polishing it with an emery paper (No. 800, Carborandum Universal, India) followed by washing in 20% HCl and water.

Results and discussion

Characterization of electrodes

The electrochemical characterization of conductive metallic oxide electrodes is generally accomplished through cyclic voltammetry. This technique is very powerful for in situ characterization of electrode surfaces, its quantitative and qualitative behavior being intimately related to the surface conditions of the electrode. RuO₂-PVC film electrodes, fabricated using oxides prepared at different temperatures ca. 300 to 700° C, have already been characterized in solutions of different pH, 0 to 14 [33]. The following is the detailed analysis in different NaOH solutions. Initially, RuO₂ (700°C)-PVC film electrode is studied in detail in 3 M NaOH followed by other NaOH solutions. Subsequent to this, other temperature RuO₂ electrodes are investigated.

Cyclic voltammetric behavior of RuO₂ (700°C)-PVC film electrode in NaOH solutions of different concentrations

Figure 1 illustrates the cyclic voltammograms, CV, recorded for RuO₂ (700°C)-PVC film electrode in 3 M NaOH at different scan rates (v) in the potential range −1,100 to 450 mV. The voltammetric curve within the potential window for oxygen and hydrogen evolution reactions shows three pairs of reversible peaks localized at −400 to −350 mV, −100 to 100 mV, and 300 to 400 mV, corresponding, respectively, to Ru(IV)/Ru(III), Ru(VI)/Ru(IV), and Ru(VII)/Ru(VI) solid-state redox transitions, in broad agreement with the one described previously in NaOH solution of similar concentration for Ti/RuO₂ electrode prepared by direct thermal decomposition of RuCl₃ on Ti at 425°C [34]. The peaks are designated as A1/C1, A2/C2, and A3/C3 for further discussion.

Fig. 1

Cyclic voltammograms of RuO₂ (700°C)-PVC film electrode in 3 M NaOH at various sweep rates (mV s⁻¹) indicated in the figure. Electrode geometric area=0.0707 cm²

The peak currents of anodic and cathodic peaks increased with scan rate. The couple A1/C1 is less well-defined than the other two. A linear relation is observed for log(peak current density) with log(scan rate); however, with two slopes ca. 1 in the lower scan rate region (10–150 mV s⁻¹ for A3/C3 and 20–300 mV s⁻¹ for A2/C2), and 0.5 at higher scan rates (150–800 mV s⁻¹ for A3/C3 and 300–800 mV s⁻¹ for A2/C2). The peak potential, E _p, is invariant with scan rate at slower scan rates, but is displaced anodically at high scan rates, by 40 and 50 mV per decade increase in sweep rate for A3 and A2 peaks, respectively. All these results support the surface redox-state behavior of the system at low scan rates, and diffusional behavior at high scan rates; the rate determining diffusional process in the latter case occurring either in solution bulk or in RuO₂ material. The RDE technique is suitable to delineate the solution mass transfer from a similar process in electrode solid material [35]. The peak position and peak currents for all the three pairs of peaks are not dependent on electrode rotation, signifying that the surface electron transfer processes of the three pairs of peaks is coupled to the diffusion of electrolyte ion occurring within the solid matrix [35].

Fig. 2

E _1/2 versus log(c _OH) plots for A1/C1, A2/C2, and A3/C3 redox couples of RuO₂ (700°C)-PVC film electrode

In aqueous solution of NaOH of varying concentrations ca. 2, 1.5, and 1 M, the back ground current and the peak currents decreased systematically and the E _1/2, obtained as (E _pa + E _pc)/2, is shifted anodically with decrease in NaOH concentration. This is true for all the three redox couples. The shift of E _1/2 with log(c _OH) is shown in Fig. 2. The coefficient [∂E _1/2/∂log(c _OH)] is −85, −82 and −95 mV decade⁻¹ for A1/C1, A2/C2 and A3/C3, respectively. These values are close to −3/2 (2.303RT/F) observed by Lyons and Burke for 450°C pyrolyzed Ti/RuO₂ electrode in NaOH solutions of concentrations 1 to 6 M [34]. Thus, the solid-state redox reactions of RuO₂(700°C)-PVC film electrode in 1 to 3 M NaOH solutions can be described as follows:

A1/C1

$$({\text{--}}\hbox{O}{\text{--}})_{2}\hbox{Ru}^{\text{III}} \hbox{O}^{-} (\hbox{H}_{2}\hbox{O})_{3}+1.25 \hbox{OH}^{-} \rightleftarrows ({\text{--}}\hbox{O}{\text{--}})_{2}\hbox{Ru}^{\text{IV}} \hbox{O} ^{-} [(\hbox{OH})_{1.25}]^{-0.25} (\hbox{H}_{2}\hbox{O})_{1.75}+1.25 \hbox{H}_{2}\hbox{O}+e$$

(1)

A2/C2

$$({\text{--}}\hbox{O}{\text{--}})_{2}\hbox{Ru}^{\hbox{IV}} \hbox{O} ^{-} [(\hbox{OH})_{1.25}]^{-0.25} (\hbox{H}_{2}\hbox{O})_{1.75} + 2.5 \hbox{OH}^{-} \rightleftarrows ({\text{--}}\hbox{O}{\text{--}})_{2}\hbox{Ru}^{\text{VI}} [\hbox{O}_{1.75}]^{-1.75} (\hbox{OH})_{2.25} + 2.5\hbox{H}_{2}\hbox{O} + 2e$$

(2)

A3/C3:

$$({\text{--}}\hbox{O}{\text{--}})_{2}\hbox{Ru}^{\text{VI}} [\hbox{O}_{1.75}]^{-1.75} (\hbox{OH})_{2.25} + 1.25 \hbox{OH}^{-} \rightleftarrows ({\text{--}}\hbox{O}{\text{--}})_{2}\hbox{Ru}^{\text{VII}} [\hbox{O}_{3}]^{-2} (\hbox{OH}) + 1.25 \hbox{H}_{2}\hbox{O} + e$$

(3)

Considering the voltammetric behavior at low sweep rates, the results, although, indicate reversible surface reactions (symmetrical peak shapes, nearly equal anodic and cathodic peak heights, linear increase of i _pa versus v, scan-rate-independent E _pa, and electrode-rotation-independent CV traces) for all the three redox pairs, nevertheless peak-to-peak separation (E _pa − E _pc) and the potential of full width at half maximum (E _FWHM) are not equal to those predicted for ideal systems [36–38], namely, zero and (90/n) mV (where n is the number of electrons transferred per electroactive site), respectively. The deviation from ideal behavior may arise due to the site-to-site interactions (thermodynamic effect), slow charge transfer (kinetic effect), or large uncompensated electrolyte resistance effects [37–39]. Following this, the interaction term r of each of the redox pairs ca. A2/C2 and A3/C3, associated with the RuO₂ (700°C)-PVC film electrode is estimated from the relationship

$$i_{\rm pa}= \frac{nFq_{\rm s}v}{RT(4-2r \Gamma _{\rm s})} $$

(4)

formulated for interacting electroactive adsorption sites [37–39]. Here q _s (in coulombs) is the charge due to solid-state redox transition and Γ_s (in mol cm⁻²) is the surface concentration of the redox sites, related to each other, as

$$q_{\rm s} = nFA_{\rm geo} \Gamma _{\rm s}$$

(5)

A _geo (in cm²) is the geometric area of the electrode, and R, T and F have their usual significance. The ratio of i _pa–v slope to the charge q _s gives, from Eq. 4, nF/[RT(4−2r Γ_s)], from which r is evaluated. q _s for these calculations was obtained integrating the i–E curve over the potential interval between the foot of the oxidation process and the peak ending for low v, typically 5 mV s⁻¹.

For A3 (Ru(VII)/Ru(VI) transition), with n=1 and Γ_s3=3.69×10⁻⁷ mol cm⁻² (estimated from the charge q _s3 under A3 peak), r _A3 is calculated to be −2.16×10⁶ cm² mol⁻¹. For A2 (Ru(VI)/Ru(IV) transition), r _A2=−4.40×10⁷ cm² mol⁻¹ with n=2 and Γ_s2=1.53×10⁻⁷ mol cm⁻². The negative r for A2 and A3 indicates that repulsive interaction exists amongst the redox sites of A3/C3 and A2/C2 in the surface layer. The greater negative value for A2/C2, by one order, indicates that repulsion is more for A2/C2 than the A3/C3 case. The observation that peak A2 is quite broad and charge possessed by this peak (1.05 mC) is less than double the charge of peak A3 (2×2.5 mC), in spite of the fact that A2 involves two electrons while A3 involves only one electron, is probably due to the greater negative interaction associated with A2.

Note that the r values −2.16×10⁶ cm² mol⁻¹ and −4.40×10⁷ cm² mol⁻¹ for A3 and A2 in 3 M NaOH can be compared with their respective values ca. 7.24×10⁶ cm² mol⁻¹ and −1.26×10⁷ cm² mol⁻¹ in 1 M NaOH [33]. It can be noticed that with increase in NaOH concentration from 1 to 3 M, the interaction in the adsorption layer becomes increasingly repulsive for both A3 and A2. This behavior can be associated with oxide solvation. In aqueous medium the oxide surface is covered by OH⁻ groups forming hydroxylated species [34, 40] which are responsible for solid-state redox transitions (Eqs. 1, 2, 3) and associated surface charge. An increase in surface hydroxylation with an increase in NaOH concentration, consequently, promoting repulsion in the surface layer, is reasonable. This is confirmed by total voltammetric surface charge for different NaOH solutions. Figure 3 shows anodic and cathodic charge densities, q _a and q _c respectively, calculated based on the electrode geometric area as a function of NaOH concentration for 700°C electrode (and also for other electrodes, see later). q _a and q _c were obtained integrating the i–E curve of the respective NaOH solution over the complete potential interval −1,100 to 450 mV for v=5 mV s⁻¹. Figure 3 shows that for 700°C electrode, there is a slight, but definite, increase with NaOH concentration, in q _a and q _c.

Fig. 3

Dependence of total voltammetric surface charge densities normalized over electrode geometric area, q _a and q _c, on NaOH concentration for RuO₂-PVC film electrodes of different oxide preparation temperatures (°C) indicated in the figure. q _a and q _c were obtained integrating the CV curve over the complete potential interval −1,100 to 450 mV (SCE) for 5 mV s⁻¹ sweep rate

Note that the potential region −1,100 to 450 mV comprises solid-state redox reactions involving the various oxidation states of Ru (Eqs. 1, 2, 3) superimposed on the double-layer-charging [31, 32]. It appears, therefore, reasonable to consider that q _a and q _c are composed of greater contribution by the redox transitions and only marginally by double-layer-charging [27, 28]. Thus, the charge values can be taken as a relative measure of active surface sites and can be used to follow the electrochemically active surface area of the electrodes [27, 28, 41].

RuO₂-PVC film electrodes of different temperatures in NaOH solutions of different concentrations

Typical CV curves measured in 3 M NaOH for RuO₂ electrodes of five different temperatures 700, 600, 500, 400 and 300°C in the potential region −1,100 to 450 mV are shown in Fig. 4. It can be seen that when oxide preparation temperature is decreased from 700 to 300°C, the overall voltammetric current increases gradually and the surface redox peaks also show an increase in intensity. However, the peaks loose their sharpness and become broader for low temperature samples. The electrochemistry of different temperature electrodes can be now conveniently understood using total voltammetric surface charge, as a criterion, and analyzing the potential sweep rate dependence of q _a and q _c, as per Trasatti formalism [41].

Fig. 4

Cyclic voltammograms in 3 M NaOH for RuO₂-PVC film electrodes of different oxide preparation temperatures (°C) indicated in the figure. Sweep rate 10 mV s⁻¹. Electrode geometric area=0.0707 cm²

For this purpose, CV curves over the entire potential range −1,100 to 450 mV were integrated at various scan rates for all the five temperature electrodes and for all the four NaOH solutions. Typical plots of q _a and q _c for the five electrodes in 3 M NaOH are given in Fig. 5 as a function of scan rate. Note that at any particular scan rate, there is a systematic increase in q _a and q _c with decrease in oxide preparation temperature from 700 to 300°C. This can be understood as: with decrease in oxide preparation temperature, high surface area material results with the existence of countless fissures, cracks etc. [31, 32]. Due to this characteristic morphology of thermal oxides normally two regions can be distinguished: (1) the external oxide/solution interface and (2) the internal oxide/solution interface located inside the fissures and cracks. It is the less accessibility of these more internal, difficult-to-reach surface sites in the pores and cracks for the diffusing OH⁻ ion for the electron transfer reactions (Eqs. 1, 2, 3) that is responsible for the steady decrease of q _a (q _c) with potential sweep rate, as observed in Fig. 5, for different temperature electrodes. Interestingly, for high temperature sample ca. 700°C, q remains nearly independent of sweep rate signifying that the “screened” surface is relatively smaller for this electrode.

Fig. 5

Variation of total voltammetric surface charge densities normalized over electrode geometric area, q _a and q _c, with potential sweep rate in 3 M NaOH for RuO₂-PVC film electrodes of different oxide preparation temperatures (°C) indicated in the figure

The charges associated with external oxide/solution interface ($q^{{\text{*}}}_{{{\text{out}}}}$) and internal oxide/solution interface ($q^{{\text{*}}}_{{{\text{in}}}}$) are calculated from experimental q _a values. According to Ardizzone et al. [41], for an oxide material, the measured voltammetric charge, q _a, at various scan rates can be linearized if 1/q _a is plotted against v ^1/2. The linear extrapolation to v→0 gives $q^{{\text{*}}}_{{{\text{tot}}}}$, the surface charge arising due to infinitely slow OH⁻ ion exchange, and thus denoting the charge related to whole active surface. That is

$$\frac{1} {{q_{{\text{a}}} }}= \frac{1}{q^{*} _{\rm tot}} + \hbox{constant}1\; (v^{1/2})$$

(6)

On the other hand, a plot of q _a against v ^−1/2, and extrapolation to v→ ∞ gives an intercept of $q^{{\text{*}}}_{{{\text{out}}}}$, the amount of charge related to outer surface of the oxide that is straight forwardly accessible to OH⁻ ion exchange.

$$q_{\rm a} = q^{*} _{\rm out} + \hbox{constant}2\; \left(\frac{1}{v^{1/2}}\right) $$

(7)

Once $q^{{\text{*}}}_{{{\text{tot}}}}$ and $q^{{\text{*}}}_{{{\text{out}}}}$ are calculated, as above, $q^{{\text{*}}}_{{{\text{in}}}}$ is got remembering that

$$q^{*} _{\rm tot} = q^{*} _{\rm out}+ q^{*} _{\rm in} $$

(8)

Figure 6 illustrates a plot of $q^{{\text{*}}}_{{{\text{out}}}}$ versus $q^{{\text{*}}}_{{{\text{tot}}}}$ for all the five electrodes in the four NaOH solutions. It is interesting to note that the points gather around a single line of slope ∼0.55 for all low temperature electrodes viz., 300 to 600°C (curve (a)), irrespective of NaOH concentration, indicating that the whole surface is almost equally divided between outer and inner regions. On the other hand, for 700°C electrode the ratio $q^{{\text{*}}}_{{{\text{out}}}} /q^{{\text{*}}}_{{{\text{tot}}}}$ is closer to unity (0.78) as clearly shown in the expanded line (curve (b)), implying that this electrode possesses negligibly small inner area ($q^{{\text{*}}}_{{{\text{in}}}}$ ∼ 25%), as already inferred from constant q-v data in Fig. 5. The physicochemical data for the RuO₂ samples collected from Ref. [33, 42] are summarized in Table 1. The pore-specific volume data from BET measurements are sufficient for the discussion at this stage. The pore-specific volume being larger for low temperature samples, 300 to 600°C, and smaller for the 700°C sample amply support the $q^{{\text{*}}}_{{{\text{out}}}} /q^{{\text{*}}}_{{{\text{tot}}}}$ results that 300 to 600°C materials are more porous, but the 700°C sample is not. As one can see later, this factor has a decisive effect on the electrocatalytic behavior of 700°C electrode towards glucose oxidation in possessing extended linear calibration range, compared to other low temperature electrodes.

Fig. 6

$q^{{\text{*}}}_{{{\text{out}}}}$ versus $q^{{\text{*}}}_{{{\text{tot}}}}$ plots for RuO₂-PVC film electrodes of different oxide preparation temperatures in NaOH solutions of different concentrations (M): (open circle) 1; (filled circle) 1.5; (open square) 2; (filled square) 3. Curve (a): for 300 to 600°C electrodes and curve (b): for 700°C electrode

Table 1 Physicochemical parameters of RuO₂ samples prepared at different temperatures^a

Note that Ardizzone et al. [41] calculated $q^{{\text{*}}}_{{{\text{out}}}}$ and $q^{{\text{*}}}_{{{\text{tot}}}}$ for 300 to 500°C thermally prepared Ti/RuO₂ electrodes in 1 M HClO₄ and 1 M KOH. The correlation between $q^{{\text{*}}}_{{{\text{out}}}}$ and $q^{{\text{*}}}_{{{\text{tot}}}}$ was found to be linear only in HClO₄ solution with a slope ∼ 0.55; whereas in KOH solution it was nonlinear with $q^{{\text{*}}}_{{{\text{out}}}}$ decreasing more in KOH than in HClO₄ as the active site quantity increases. Such a behavior was explained in terms of rate of surface diffusion of proton-donating species, being faster in acid and slower in alkali solution. Our results in Fig. 6 indicate that $q^{{\text{*}}}_{{{\text{out}}}}$ versus $q^{{\text{*}}}_{{{\text{out}}}}$ is linear in alkaline solutions, 1 to 3 M NaOH, for 300 to 600°C RuO₂-PVC film electrodes. In other words, the presence of PVC constitutes a favorable microenvironment for faster surface diffusion of proton-donating species even in alkaline medium, the more the higher the active site concentration. The exact nature of this enhancement is not known but interaction between PVC and OH⁻ may be suspected.

Catalytic glucose oxidation

Oxidation of glucose at RuO₂ (700°C)-PVC film electrode

The curve (a) of Fig. 7 shows the CV for the oxidation of 15 mM glucose at a RuO₂ (700°C)-PVC film electrode in 1 M NaOH. In Fig. 7, also included is the response of RuO₂ (700°C)-PVC film electrode in pure base electrolyte devoid of glucose (curve (b)). In presence of glucose, the anodic currents corresponding to both Ru(IV) oxidation to Ru(VI) (A2 peak) and Ru(VI) oxidation to Ru(VII) (A3 peak) increased, while the cathodic peak current of C3 decreased and the C2 peak totally disappeared. It was observed that the peak currents of A2 and A3 increased steadily with glucose concentration. These are indications that the glucose oxidation at RuO₂ (700°C) electrode in NaOH may occur by mediation by surface-confined ruthenium redox species. There are several possibilities for the appearance of two oxidation peaks for the glucose oxidation at RuO₂ (700°C) electrode in 1 M NaOH solution. In the first place, note that the peaks are centered in the potential region of Ru(IV) to Ru(VI) oxidation and Ru(VI) to Ru(VII) oxidation. Apparently, both ruthenate Ru(VI) (at peak A2) and perruthenate Ru(VII) (at peak A3) are electroactive towards glucose oxidation. However, there are conflicting reports regarding the catalytically active surface species of RuO₂ electrode [15, 19, 20, 25–27], implying that the nature of the catalytic species is dependent on the type of the substrate molecule. For example, the perruthenate (Ru(VII)) surface group is regarded as active towards the oxidation of difficultly oxidizable compounds like benzyl alcohol [19, 25] and formic acid [20]. Both Ru(VII) and Ru(VI) have been proposed to be active for benzaldehyde [19], formaldehyde [20], and glucose [15]. In the case of ethanol, however, while Shieh and Hwang [26] have suggested that only Ru(VII) is active; Ru(VI) species has been identified to be the only active species by de Andrade et al. [27].

Fig. 7

Cyclic voltammograms of glucose at RuO₂(700°)-PVC film electrode in 1 M NaOH: (curve (a)) 15 mM glucose; (curve (b)) supporting electrolyte without glucose. Sweep rate 10 mVs⁻¹. Electrode geometric area=0.0707 cm²

It is well established that in solutions of pH<13 the highest surface-oxidized form of ruthenium species that is generated during RuO₂ oxidation is only Ru(VI) within the potential range in which the supporting electrolyte is stable [19, 20, 25–27, 33]. Thus, the oxidation of glucose was carried out at RuO₂(700°C)-PVC electrode in pH 10.6 (borax+NaOH) buffer solution, to examine whether Ru(VI) species is catalytically active for glucose oxidation, and the results are presented in Fig. 8. RuO₂(700°C)-PVC electrode in pure base electrolyte shows two redox pairs of peaks (curve (a)): A1/C1 corresponds to Ru(IV)/Ru(III) at E _1/2≈ 0 mV and A2/C2 corresponds to Ru(VI)/Ru(IV) at E _1/2≈ 340 mV. When glucose (10 mM) is added, the voltammogram is virtually unaffected (curve (b)), suggesting that Ru(VI) species is not catalytically active towards glucose oxidation. Therefore, the glucose oxidation peak observed for RuO₂(700°C)-PVC electrode in 1 M NaOH at lower potential 100–200 mV in Fig. 7 is not due to mediation by ruthenate Ru(VI), but it has a different origin. Note that the necessity of higher alkaline condition for the electrocatalytic oxidation of glucose was reported with RuO₂ carbon-paste electrode [14], and also with Ni [9, 10] and Co [11]. Thus, the glucose oxidation in alkaline solution occurs via redox mediation by Ru(VII) at higher potentials in the range 300 to 400 mV as follows:

$$\hbox{Ru}(\hbox{VI})_{\rm Sur} \rightleftharpoons \hbox{Ru}(\hbox{VII})_{\rm Sur} + e$$

(9)

$$2\hbox{Ru}(\hbox{VII})_{\rm Sur} + \hbox{G}_{\rm Sol} \to 2\hbox{Ru}(\hbox{VI})_{\rm Sur} + \hbox{GL}_{\rm sol}$$

where G denotes glucose and GL represents gluconolactone.

Fig. 8

Cyclic voltammograms of glucose at RuO₂(700°)-PVC film electrode in borax + NaOH solution of pH 10.6: (curve (a)) supporting electrolyte without glucose; (curve (b)) 15 mM glucose. Sweep rate 10 mV s⁻¹. Electrode geometric area=0.1964 cm²

We may consider that the two oxidation peaks for glucose on RuO₂(700°C)-PVC electrode in 1 M NaOH, Fig. 7, correspond to successive oxidation of glucose substrate—the mechanism similar to that noticed with metal electrodes [1–3, 7]. On metal electrodes glucose is oxidized step-wise with the first peak at low potentials corresponding to the abstraction of α-hydrogen from glucose molecule followed by further oxidation at higher potential region. The abstraction of α-hydrogen in the case of glucose on metal electrodes was confirmed from oxidation studies with α-hydrogen lacking carbohydrates, like, sodium gluconate, fructose etc., which showed one-stage oxidation at higher potentials [2, 7]. But, on RuO₂ (700°C)-PVC electrode, we have observed experimentally, as shown in Fig. 9, that all these α-hydrogen-lacking carbohydrates are oxidized with two oxidation peaks similar to glucose oxidation. Besides, the intensity of both the anodic peaks does increase with increase in bulk concentration of these derivatives. This observation suggests that the first glucose oxidation peak on RuO₂ (700°C) electrode in 1 M NaOH (Fig. 7, curve (a)) is not connected with α-hydrogen abstraction, signifying that the glucose oxidation on RuO₂ does not proceed following the same mechanism as with metal electrodes.

Fig. 9

Cyclic voltammograms of glucose and its derivatives at RuO₂(700°)-PVC film electrode in 1 M NaOH: (curve (a)) glucose; (curve (b)) fructose; (curve (c)) sodium gluconate. Substarte concentration 10 mM. Sweep rate 10 mV s⁻¹. Electrode geometric area=0.1964 cm²

We believe that the glucose oxidation peak at lower potentials on RuO₂, Fig. 7, may be an adsorption pre-peak as a result of the strong adsorption of the oxidized product, as well-documented in the polarography and CV literature [36]. The effect of potential sweep rate on the oxidation of glucose at the RuO₂ electrode confirms this proposal. With increase in scan rate, peak current increases for both the peaks (Fig. 10); however, the pre-peak increases at a faster rate than the main peak for several glucose concentrations (Fig. 10, inset (A)). Moreover, note that the peak current of the pre-peak increases linearly with v, i.e., ∂i _pa/∂(log(v)) ≈ 1 (Fig. 10, inset (B)), typical of a surface electron transfer process. But, for the main peak, the peak current increases linearly with v ^1/2, i.e., ∂i _pa/∂(log(v)) ≈ 0.5, typical of diffusion-controlled electron transfer process.

Fig. 10

Cyclic voltammograms of glucose at RuO₂(700°)-PVC film electrode in 1 M NaOH at various sweep rates (mV s⁻¹) indicated in the figure. Glucose concentration 15 mM. Electrode geometric area=0.1964 cm² Inset (A): dependence of peak current ratio (ρ) of pre-peak to main peak on sweep rate for various glucose concentrations (mM) indicated in the figure Inset (B): Log(i _pa) versus log(v) plots for 15 mM glucose: (filled circle) pre-peak; (open circle) main peak

Thus, in effect, the glucose oxidation at RuO₂(700°C)-PVC film electrode in alkaline solution occurs via redox mediation by electrogenerated Ru(VII) in two ways: one involving soluble forms of both the reactant and the product at the main peak in the potential region 300 to 400 mV (reaction (Eq. 10)), and the other involving strong product adsorption at the pre-peak in the potential region 100 to 200 mV (reaction (Eq. 11)).

$$2 \hbox{Ru}(\hbox{VII})_{\rm Sur} + \hbox{G}_{\rm Sol} \to 2 \hbox{Ru}(\hbox{VI})_{\rm Sur} + \hbox{GL}_{\rm Sol}$$

(10)

$$2 \hbox{Ru}(\hbox{VII})_{\rm Sur} + \hbox{G}_{\rm Sol} \to 2 \hbox{Ru}(\hbox{VI})_{\rm Sur} + \hbox{GL}_{\rm Ads}$$

(11)

Oxide preparation temperature effect on the catalytic oxidation of glucose

The oxide preparation temperature effect was investigated for 15 mM glucose oxidation in 1 M NaOH for all the five RuO₂-PVC film electrodes. Figure 11 shows comparative CVs recorded at v=10 mV s⁻¹ for the five electrodes. It is quite interesting to note that the pre-peak, which is present prominently with high temperature electrodes, decreases gradually with decrease in oxide preparation temperature and it is totally absent with 300°C electrode. The peak potential of the main catalytic peak is displaced towards more positive potentials as the preparation temperature decreases. This behavior suggests that the overall catalytic activity of the electrodes for glucose oxidation is higher for high temperature electrodes.

Fig. 11

Cyclic voltammograms of glucose in 1 M NaOH at RuO₂-PVC film electrodes of different oxide preparation temperatures (°C) indicated in the figure. Sweep rate 10 mV s⁻¹. Electrode geometric area=0.0707 cm²

The efficiency of the oxide electrodes for glucose oxidation was characterized using the glucose oxidation current (i _cat) normalized over electrode geometric area (A _geo), i.e., the apparent electrochemical activity [25, 27, 28], j _cat=i _cat/A _geo. Figure 12a shows a plot of j _cat as a function of oxide preparation temperature. The j _cat data presented in Fig. 12(A) correspond to oxidation current measured at 350 mV from cyclic voltammogram recorded at v=1 mV s⁻¹ at a rotating electrode with 1,000 rpm for the oxidation of 15 mM glucose in 1 M NaOH at RuO₂-PVC film electrodes of different oxide preparation temperatures. It is clear that j _cat decreases with increase in oxide preparation temperature indicating that the low temperature RuO₂ electrodes (300–500°C) possess higher activity than the 600 and 700°C samples.

Fig. 12

a Catalytic current density normalized over electrode geometric area, j_cat, as a function of oxide preparation temperature in 1 M NaOH. b Catalytic current density normalized over electrochemically active surface area (EASA), j _cat’, as a function of oxide preparation temperature in 1 M NaOH. For details see text

The j _cat parameter reflects changes in both morphologic (area) and intrinsic electrocatalytic effects. To compare the real catalytic activity of the different electrodes, one must separate the individual contributions. It is possible to eliminate morphological effects from the electrochemical response by dividing the oxidation current by the electrochemically active surface area (EASA). The normalized oxidation current, free of area effects, is given by j _cat’= i _cat/EASA, and j _cat’ as a function of preparation temperature is illustrated in Fig. 12b. EASA for each of the five electrodes was estimated from double-layer-charging curves in 1N NaOH using the small amplitude CV technique, as described in detail elsewhere [33], and the EASA for all the electrodes are listed in Table 1.

Figure 12b shows a trend reverse to that observed in Fig. 12a clearly evidencing that high-temperature-prepared RuO₂-PVC film electrodes have better intrinsic catalytic properties than the low temperature samples.

Now referring to Table 1, it shows that the high temperature sample, 700°C RuO₂, possesses smaller amount of Cl⁻ (XPS), higher Ru(VI) content (XPS), and smaller water content (DTGA), compared to low temperature sample, 300°C RuO₂. Basing on these data, the behavior of higher temperature RuO₂ electrodes showing better intrinsic catalytic activity can be associated with the following possibilities.

• 1. The pK_A of glucose being 12.28, the reactant molecule in alkaline solution is anionic [43]. Thus, its oxidation at a 300°C electrode may proceed with a higher activation energy due to the presence of submonolayer quantities of anionic Cl⁻ in the network structure of water in the surface layer of this electrode. Consequently, with the complete absence of Cl⁻ (Table 1), the 700° RuO₂ electrode can allow glucose oxidation reaction to occur at enhanced rates resulting in high electrocatalytic activity.

• 2. RuO₂ particles in solution are known to present with extensive hydration of surface oxyhydroxide groups [34, 40]. The lesser quantities of water for 700°C sample (Table 1) reflect that its surface is hydrated to a lesser extent (less hydrophilic). Near the surface of such an arrangement, the glucose molecule with its six –OH groups may form strong water-like network structure as a result of extensive interaction between –OH groups and the Ru metal sites at the oxide surface. Thus a stronger adsorption interaction, a prerequisite for a more facile electron transfer, can be envisaged between glucose and 700°C RuO₂ surface (than 300°C electrode).

• 3. Other explanations for high reactivity of the 500–700°C electrodes consider higher quantity of the pre-catalyst species, Ru(VI), identified with these samples (Table 1). These Ru(VI) species can support preferential adsorption of glucose oxidation products producing additional path for glucose oxidation in the form of adsorption pre-peak along with the main peak (Fig. 7), with overall higher activity of these electrodes.

Oxide preparation temperature effect on the linear calibration range

Figure 13 shows the plot between j _cat (measured at 350 mV) and glucose concentration (c ⁰_G ) for various RuO₂ electrodes in 1 M NaOH. The data presented in Fig. 13 were obtained at a RDE rotated at 1,000 rpm and v=1 mV s⁻¹. It may be noted that for all the electrodes the catalytic oxidation current increases with glucose concentration, but levels-off at higher concentration. It can, therefore, be inferred that glucose oxidation at all the RuO₂-PVC film electrodes follows saturation kinetics similar to that described by Michaelis–Menten (MM) reaction scheme involving the formation of a complex between the oxyruthenium surface species and dissolved glucose, which subsequently decomposes to generate the pre-catalyst and the product glucanolactone.

$$ {\text{2Ru(VII)}}_{{{\text{Sur}}}} + {\text{G}}_{{{\text{Sol}}}} \overset {K_{{\mathbf{M}}} } \leftrightarrows [2{\text{Ru}}({\text{VII}}) - - - {\text{G}}]_{{{\text{Sur}}}} \xrightarrow{{k_{{\text{c}}} }}2{\text{Ru}}({\text{VI}})_{{{\text{Sur}}}} + {\text{GL}}_{{{\text{Sol}}}} $$

(12)

Fig. 13

j _cat variation with glucose concentration (c ⁰_G ) in 1 M NaOH for RuO₂-PVC film electrodes of different oxide preparation temperatures (°C) indicated in the figure. Inset: variation of j _cat in the lower region of c ⁰_G

One may write the corresponding equation as [15, 23]

$$ \frac{1} {{j_{{{\text{cat}}}} }} = \frac{1} {{nFk_{{\text{c}}} \Gamma _{{\text{S}}} }} + \frac{{K_{{\mathbf{M}}} }} {{nFk_{{\text{c}}} \Gamma _{{\text{S}}} c^{{\text{0}}}_{{\text{G}}} }} $$

(13)

where k _c (cm s⁻¹) and K _M (in mol dm⁻³) are the MM kinetic constants, n the number of electrons transferred for oxidation of one molecule of glucose (here n=2), Γ_s surface concentration of redox sites and F the Faraday constant. The behavior that glucose follows MM kinetics can be further confirmed by reasonably linear Lineweaver-Burke plots (j ⁻¹_cat versus (c ⁰_G )⁻¹) for all the RuO₂ electrodes, as shown in Fig. 14, with

$$ {\text{Slope = }}\frac{{K_{{\mathbf{M}}} }} {{nFk_{{\text{c}}} \Gamma _{{\text{S}}} }} $$

(14)

$$ {\text{Intercept = }}\frac{1} {{nFk_{{\text{c}}} \Gamma _{{\text{S}}} }} $$

(15)

The constants k _c and K _M were calculated from intercept and slope using n=2 and Γ_s for each individual electrode, as listed in Table 1. Figure 14-inset shows that the calculated K _M , which signifies the maximum concentration up to which linearity between j _cat and c ⁰_G is maintained, increases with oxide preparation temperature (curve (a)).

Fig. 14

Lineweaver-Burk plot (j ⁻¹_cat versus (c ⁰_G )⁻¹) for the data in Fig. 13. Inset: Dependence of maximum c ⁰_G of linearity range on oxide preparation temperature (T/°C): (Curve a) predicted from Michaelis–Menten kinetic equation; (Curve b) experimental

The above result is quite significant and important, since it predicts that the linear calibration range is more wider and extended to higher glucose concentrations for high temperature RuO₂ electrodes than for electrodes prepared at low temperatures. It is indeed true experimentally, as illustrated in Fig. 13-inset, which shows linear calibration plots between j _cat and c ⁰_G over lower range of glucose concentrations. Note that the linear range for 700°C electrode is extended over 0 to 20 mM, 0 to 15 mM for 600°C electrode, but it is restricted to 0 to 5 mM for 300°C electrode. The experimentally observed maximum c ⁰_G of linearity range as a function of oxide preparation temperature is included in Fig. 14-inset as curve (b).

The wider linearity region extended to higher glucose concentrations associated with 700°C RuO₂ electrode may be understood as due to its pores of smaller size (Table 1), which can cause restricted diffusion of the analyte towards active Ru(VII) sites resulting in active sites getting saturated only when glucose concentration level approaches higher values, a phenomena already noticed with enzyme electrodes using microporous and homogeneous membranes of different pore diameters employed as outer membranes [44]. Thus, as far as the application of RuO₂ as amperometric electrode for the estimation of glucose is concerned, the morphological effects rather than the intrinsic catalytic properties determine its performance.

Now referring to two sets of maximum c ⁰_G values of linearity region, one experimentally observed and the other predicted from Eq. 13, Fig. 14-inset shows that experimental values (curve (b)) are always lower than the predicted values from MM Eq. 13 (curve (a)), for all the five electrodes. In other words, experiments indicate surface site saturation at relatively lower glucose concentration, whereas higher glucose concentrations are predicted to saturate as per MM Eq. 13. This is probably because MM Eq. 13 concerns only with uncomplicated idealized reaction (Eq. 12), which contains soluble forms of glucose and glucanolactone; but in reality, we know now that the glucose oxidation on RuO₂ electrodes is fairly complex occurring by two paths one involving soluble forms of glucose and glucanolactone (reaction (Eq. 12)) and the other involving strong product adsorption with the following MM reaction scheme (reaction (Eq. 16))

$$ {\text{2Ru(VII)}}_{{{\text{Sur}}}} + {\text{G}}_{{{\text{Sol}}}} \overset {K^{\prime }_{{\mathbf{M}}} } \leftrightarrows [2{\text{Ru}}({\text{VII}}) - - - {\text{G}}]_{{{\text{Sur}}}} \xrightarrow{{k^{\prime }_{{\text{c}}} }}2{\text{Ru}}({\text{VI}})_{{{\text{Sur}}}} + {\text{GL}}_{{{\text{Ads}}}} $$

(16)

with K _M ’ and k _c’ as the corresponding MM kinetic constants. This product adsorption reaction can additionally contribute, and higher oxidation current can be registered in experiments leading to surface site saturation at relatively low glucose concentrations. This explanation is supported by the observation in Fig. 14-inset that the difference in c ⁰_G between the two sets of values is larger for high temperature electrodes, because only at these electrodes the strong product adsorption (pre-peak) is more pronounced.

Conclusions

• 1. RuO₂-PVC film electrodes, fabricated using RuO₂ powders prepared at five different temperatures, viz., 300, 400, 500, 600 and 700°C, showed in high alkaline media 1 to 3 M NaOH three redox pairs Ru(IV)/Ru(III), Ru(VI)/Ru(IV) and Ru(VII)/Ru(VI) exhibiting non-Nerntian behavior with [∂E _1/2/∂log(c _OH)] shift of −85, −82, and −95 mV, respectively, similar to traditionally prepared RuO₂/Ti electrodes.

• 2. Both the redox pairs Ru(VI)/Ru(IV) and Ru(VII)/Ru(VI) are associated with repulsive interaction among the redox sites in the surface layer with corresponding interaction terms r _A2=−4.40×10⁷ cm² mol⁻¹ and r _A3= −2.16×10⁶ cm² mol⁻¹, respectively, in 3 M NaOH.

• 3. Estimation of $q^{{\text{*}}}_{{{\text{tot}}}}$, $q^{{\text{*}}}_{{{\text{out}}}}$, $q^{{\text{*}}}_{{{\text{in}}}}$ for all the five electrodes shows that the low temperature electrodes viz., 300 to 600°C, are associated with almost equally distributed outer and inner active sites, whereas the 700°C electrode possesses fewer inner active surface sites, in total agreement with pore-specific volume data from BET measurements.

• 4. The glucose oxidation on RuO₂ does not follow the same mechanism as on metal electrodes. Here glucose undergoes oxidation via redox mediation by electrogenerated Ru(VII) in two ways: one involving soluble forms of both the reactant and the product, and the other involving strong product adsorption, with the latter process more pronounced and persistent for high temperature RuO₂-PVC film electrodes.

• 5. The effect of oxide pyrolysis temperature of RuCl₃.nH₂O revealed that high-temperature-prepared RuO₂-PVC film electrodes have poorer apparent electrochemical activity, but better intrinsic electrocatalytic activity compared to the low temperature samples. It appears that smaller amount of Cl⁻ and higher Ru(VI) content (from XPS measurements), and smaller water content (from DTGA measurements) associated with high temperature RuO₂ samples are responsible.

• 6. The high-temperature-prepared RuO₂ electrodes exhibit more wider linear calibration range extended to higher glucose concentrations, and this is intimately connected to their lesser porosity and pores of smaller size (from BET measurements), which favor restricted-diffusion of the analyte and delayed-saturation of their active sites only at higher glucose concentrations.