1211

Full Paper

Organic Redox Probes for the Key Oxidation States in Mixed Valence Ruthenium Oxide/Cyanometallate (Ruthenium Prussian Blue Analogue) Catalysts Annamalai Senthil Kumar, Jyh-Myng Zen* Department of Chemistry, National Chung Hsing University, Taichung 402, Taiwan *e-mail: [email protected] Received: May 13, 2003 Final version: July 29, 2003 Abstract The electrochemical redox behavior of the polynuclear mixed valence ruthenium oxide cyanometallate complexes (mvRuOx MCN, M ˆ Fe, Cr, Ni, Cu, Ru and Pt) have been systematically studied in this report by using three redox sensitive organic probes of glucose, ethanol and formaldehyde. The results were interpreted by the well-established ruthenium oxide and Prussian blue chemistry. The mvRuOx MCN, under the category of Ru-based Prussian blue analogue, was found to possess superior electrocatalytic activity than either ruthenium oxide or Prussian blue in acidic mediums. The electrogenerated oxy/hydroxy-RuVII state (at ‡ 1.1 V vs. Ag/AgCl) was unusually stabilized in the mvRuOx MCN matrix without any disproportion reaction in acidic environments. In contrast to those of earlier studies, possible structure in terms of the RuIII/II NC M and RuIII/II O RuVII/VI sites was proposed here. Enzymeless analytical detection of glucose in acidic conditions was first time demonstrated with sensitivity comparable to that of ruthenium oxide-based electrodes in alkaline solutions. Keywords: Redox behavior, Mixed valence ruthenium oxide cyanametallate, Prussian blue analogues

1. Introduction Chemically modified electrodes prepared by the mixed valence ruthenium oxide/cyanometallates (mvRuOx MCN, M ˆ Fe and Ru) with the ™ M CN Ru ∫ type of 3D internal linkage are often applied in the oxidative determination of organic and biologically important molecules, as summarized in Table 1 [1 ± 17]. These catalysts have mutual properties of ruthenium oxide (RuOx) and Prussian blue (PB, KFeIIIFeIICN) analogues. Surprisingly, very little information was available regarding the basic redox characteristics of the mvRuOx MCN so far. Questions related to the key ruthenium species involved in the electrocatalytic process and the pH-dependent dynamics of each redox peaks are unclear. There is even inconsistency in the reported electrochemical behavior of the mvRuOx MCN films. For example, Cataldi et al. reported three redox peaks in the potential region from 0.2 to ‡ 1.2 V (vs. Ag/AgCl) for mvRuOx FeCN [14]; whereas, Fu et al. observed only two redox peaks in the same potential range [16]. Gorski and Cox reported oxy/hydroxy-RuIV/III and RuVI/IV redox transitions for mvRuOx RuCN at ‡ 0.8 and ‡ 1.0 V (vs. Ag/ AgCl), respectively, in pH 2 solution [13]. The redox assignment, however, is different from that of classical solid state and solution phase oxy/hydroxy-ruthenium chemistry, where the RuIV/III and RuVI/IV redox transitions should occur in the potential range from ‡ 0.06 to ‡ 0.14 V and ‡ 0.56 to ‡ 0.81 V (vs. Ag/AgCl), respectively, in pH 2 solution [18 ± 24]. Recently, Fu et al. selectivity determined Cs‡ ion by Electroanalysis 2004, 16, No. 15

using mvRuOx FeCN [16], which is against the property of size insensitive to alkali metals for the classical PB analogues [2, 14]. There are obvious qualitative changes in the redox and electrochemical features of the catalysts from time to time. We systematically investigate here the possible ruthenium oxidation states responsible for the electrocatalytic oxidations in the mvRuOx MCN (M ˆ Cr, Fe, Ni, Cu, Ru and Pt), especially for mvRuOx FeCN and mvRuOx RuCN. The information gained from previous studies on oxy/hydroxyRu, RuOx and PB-based electrodes were applied for the redox assignments [18 ± 40]. This is quite reasonable since the mvRuOx MCN is prepared by a similar procedure to that of PB and both mvRuOx MCN and PB are considered under the same category of zeolite type of inorganic polymeric materials [1, 2]. Furthermore, in some way, the mvRuOxMCN is close to RuOx in electrocatalytic behavior as well as in analytical applications [2, 3, 11, 13, 14]. It is noteworthy that, regardless of the physiochemical characterization of mvRuOx RuCN and mvRuOx FeCN by X-ray photoelectron (XPS), Auger, IR and UV-visible absorption spectroscopies (mainly intended to pursue the redox characteristic), the exact ruthenium oxidation state is not yet authentically concluded [3, 14]. The complications are due to the multiple redox centers and polymericcomplex nature of the mvRuOx MCN. For example, the C(1 s) energy levels at ca. 284 eV (native) and 288 eV (cyano group), respectively, are always mingled with Ru(3d5/2) and Ru(3d3/2) components in the binding energy window

¹ 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

DOI: 10.1002/elan.200302940

1212

A. S. Kumar, J.-M. Zen

Table 1. Literature survey about mixed valence (mv) ruthenium cyanometallate films. CE: Capillary electrophoresis. Year mv-Film

Purpose

Epa ( V ) Proposed higher redox state Ref.

1984 1987 1989 1990

Ru Ru Ru Ru

0.80 1.01 0.92 0.90

[a] [a] [b] [a]

RuIII/II RuIII/II ± ±

[2] [3] [5] [6]

1993 1994 1995 1995 1995 1995 1995 1996

Ru Ru Ru Ru Ru Ru Ru Ru/Nafion Ru Ru Ru Ru Ru Ru Ru Ru, Ru Fe and Os Ru Ru Ru Ru Fe WO3/Ru Ru/Pt Ru M ( M ˆ Fe, Cr, Cu, Ni and Ru)

Oxidative detection of arsenic Methanol oxidation and detection Insulin oxidative detection Oxidative detection of cysteine, cystine, methionine and thiocyanate Insulin detection from pancreatic b-cells Oxidative detection of N-nitrosamines Oxidative detection of aliphatic and furanic aldehydes Oxidative determination of neutral organic compounds Sensor for aliphatic alcohols Amperometric detection of ethanol Oxidation of N-nitrosamines Voltammetric and XPS investigation

0.85 0.95 1.00 1.00 1.10 1.05 1.00

[a] [a] [a] [b] [a] [a] [a]

± RuVI/IV RuIV/III RuVI/IV RuIV/III RuVI/IV RuVI/IV RuVI/IV or RuVI/V

[7] [8] [9] [10] [11] [12] [13] [14]

CE-coupled ethanol detection CE detection of Cs‡ Oxidation of methanol Oxidation of carbohydrate (glucose), ethanol and formaldehyde

1.05 0.80 1.10 1.10

[a] [b] [a] [b]

± Low spin FeIII/II RuIV/III RuVII/VI

[15] [16] [17] This work

1996 1999 2001 2004

Ru Ru Ru Ru

Reference electrodes: [a] SCE and [b] Ag/AgCl.

of ca. 280 ± 288 eV in the XPS studies of mvRuOx MCN. The quantitative deconvolution of individual elements is thus very complicated, not to mention the difficulty in studying the electrogenerated intermediates with XPS. Xray single crystal and powder diffraction studies are also not easy because of the very slow solution phase formation rate of the mvRuOx MCN (M ˆ Fe, Cr, Ni, Cu, Ru and Pt). Overall, there is no proper procedure for the mvRuOx MCN formation apart from electrochemical deposition methods.

We first adopt redox sensitive organic probes to investigate the key ruthenium oxidation states in the mvRuOx MCN. Scheme 1 sketches the established dynamics of oxy-ruthenium redox states against solution pH and their specific organic molecules. It is understandable that oxy-RuVI (ruthenate) is available in pH 1 ± 14; while the higher oxy/hydroxy-RuVII (perruthenate) can exist only in strong alkaline conditions of pH > 12 [18 ± 20]. As a result of the disproportion reaction, the oxy/hydroxy-RuVII always transforms to lower oxidation states when pH < 12 [18 ± 20, 29]. Regarding to the specific redox oxidations, oxy/

Scheme 1. A) Redox transitions and standard potentials of the oxy-Ru oxidation states. B) Conceptional representation for the dynamics of the oxy-Ru redox transitions against the solution pH and their specific oxidative functions towards organic redox probes. The RuVIIO 4 is only stable in strong alkaline pHs, while at pH < 12, it gets disproponated to lower redox states. Electroanalysis 2004, 16, No. 15

¹ 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

1213

Mixed Valence Ruthenium Oxide/Cyanometallate Catalysts

hydroxy-RuVI can oxidize aliphatic alcohols (e.g., ethanol) and formaldehyde in a wide pH range of 1 ± 14 [25, 28, 41 ± 43]; while the oxy-RuVII can even mediate the carbohydrates (e.g., glucose) and aromatic alcohols (e.g., benzyl alcohol) oxidations under strong alkaline conditions [20, 26, 27, 29, 43]. The carbohydrate oxidation was chosen in his study to specific account for the presence of oxy-RuVII since benzyl alcohol has solubility problem in aqueous solution [43]. By choosing suitable organic probes at desired pH and potential, one can then acquire easily about the existing ruthenium redox states. Indeed, similar redox probe assignment has already been successfully addressed for the lead ruthenate pyrochlore chemically modified electrode in our recent studies [43, 44]. Glucose (for probing oxy-RuVII/VI), ethanol and formadehyde (for probing oxy-RuVI/IV) are used to specifically assign the ruthenium redox transitions in the mvRuOx MCN. Normally, glucose oxidation reaction requires a powerful oxidant (with E8  ‡ 1.40 V vs. RHE) like RuO 4 (‡ 1.40 V vs. RHE), Ce4‡ (‡ 1.45 V vs. RHE), IO 4 (‡ 1.65 V vs. RHE) and MnO 4 (‡ 1.68 V vs. NHE) to make the reaction from happening [45 ± 48]. Interestingly, superior electrocatalytic responses were noticed for glucose as well as the other two organic probes for the mvRuOx MCN in acidic pHs at ca. ‡ 1.10 V (vs. Ag/AgCl). This is unique since such a catalytic oxidation for glucose has never been reported on either RuOx (only in pH > 12) or any PB analogues in acidic environments. The unusual entrapment and generation of oxy/hydroxy-RuVII species in acidic pH is proposed as a key step for the superior electrocatalysis of the mvRuOx MCN. The stable polynuclear zeolite type of macro-structure with internal CN bridges can somehow protect the oxy/ hydroxy-RuVII oxidation state electrogenerated in the mvRuOx MCN from the disproportion reaction. To understand the phenomenon, the investigation is first focus mainly on the redox function of the catalytic ruthenium oxidation states of the mvRuOx FeCN. Various mvRuOx MCN (M ˆ Fe, Cr, Ni, Cu, Ru and Pt) films were then further studied for comparison. Finally, a preliminary result for the enzymeless detection of glucose was demonstrated by using the most sensitive mvRuOx RuCN.

2. Experimental 2.1. Reagents and Materials Ruthenium trichloride hydrate, potassium hexacyanoferrate, potassium cyanocuprate, potassium tetracyanonickelate hydrate, potassium hexacyanochromate, potassium hexacyanoplatinate, potassium hexacyanoruthenate and formaldehyde were purchased from Aldrich. Ethanol (RDH), d(‡)glucose (Sigma) and other chemicals employed were of analytical grade and used without any purification. Aqueous solutions are prepared with doubly distilled deionized water. The cyanometallate complexes from Ru, Cu and Ni are toxic, so proper care must be taken before handling. Electroanalysis 2004, 16, No. 15

2.2. Apparatus Cyclic voltammetric experiments were carried out using a CHI 660 electrochemical workstation (Austin, TX, USA). The three-electrode system consists of a 3 mm diameter modified/unmodified GCE working electrode, an Ag/AgCl reference electrode (RE-5, BAS) and a 3 mm diameter platinum auxiliary electrode. For nonaqueous electrochemical experiments, a silver wire was used as the semireference electrode with 0.4 V shift (positive) to the Ag/ AgCl reference electrode. All of the solutions were deaerated by argon gas for at least 20 min before the electrochemical experiments.

2.3. Preparation of Modified Electrodes Prior to the modification, the GCE surface was mirror polished with a BAS polishing kit followed by the sonication. The type I mvRuOx MCN and PB were prepared with 1 mM of RuCl3 ¥ xH2O ‡ 1 mM of M(CN) n6=4 (M ˆ Cr, Ni, Fe, Cu, Pt and Ru) and 1 mM of FeCl3 ‡ 1 mM of Fe(CN) 36 , respectively, by potential cycling method in the window of 0.5 to ‡ 1.2 V (vs. Ag/AgCl) in 10 mL solution of 0.1 M, pH 2 KCl/HCl at a scan rate (v) of 50 mV/s for 20 cycles. As to the ruthenium oxide electrode (designated as EC/RuOx), a cleaned GCE is electrochemically cycled with 1 mM RuCl3 in the above said experimental condition. The way of RuOx preparation is close to the method reported by Hu et al. [31]. A type II mvRuOx FeCN was also prepared in pure 10 mM HClO4 with 1 mM each of RuCl3 ¥ xH2O ‡ Fe(CN) 36 solution. Unless otherwise mentioned, all electrodes were prepared under type I condition. The freshly prepared mvRuOx FeCN showed a reddish brown color with a shinny crystalline appearance. All electrodes were stabilized in respective base electrolyte by potential cycling in the window of 0.50 to ‡ 1.25 V (vs. Ag/AgCl) (v ˆ 50 mV/s, n ˆ 10) before experiments. Comparative redox probe experiments with glucose, ethanol and formaldehyde were consistently performed with 0.5 M in pH 2 KC/HCl at a scan rate of 50 mV/s. Since the oxy-ruthenium redox potentials are pH-dependent, in order to compare to the reported standard redox potentials of ruthenium metal ions in the reversible hydrogen scale (RHE, where E8RHE ˆ 0.0 V), a simplified equation of ERHE ˆ [(EAg/AgCl ‡ 0.06 pH) ‡ E8Ag/AgCl (0.22)] is used for converting the potential between EAg/AgCl and ERHE.

3. Results and Discussion 3.1. Electrochemical Redox Behavior of the mvRuOx FeCN Figure 1A shows a typical cyclic voltammetric response for the electrochemical growth of the type I mvRuOx FeCN in pH 2 KCl/HCl solution. Three well-defined redox couples at 0.00 V (A1/C1), ‡ 0.80 V (A2/C2) and ‡ 1.10 V (A3/C3) ¹ 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

1214

A. S. Kumar, J.-M. Zen

Fig. 1. Typical CV response for the growth (n ˆ 20) of mvRuOx FeCN films on GCE with 1 mM each of Fe(CN) 36 ‡ RuCl3 in A) pH 2 KCl/HCl (type I) and B) 0.01 M HClO4 (type II) solutions. Potential segment analysis of mvRuOx FeCN film in the scan window of 0.5 to ‡ 1.2 V (C) and ‡ 1.2 to 0.5 V (D). E) Hydrodynamic effect of mvRuOx FeCN (BAS cell stand stir mode 3). F) CV response of EC/RuOx and GCE. Fresh electrodes were used for each set of experiments. Except (B), for all other cases, pH 2 KCl/HCl base electrolyte was uniformly used at a scan rate of 50 mV/s.

together with a redox reaction at ‡ 0.25 V related to the solution phase Fe(CN) 63 =4 were observed. As to the classical PB, only a couple of well-defined redox peaks at ‡ 0.20 V (A1/C1) and ‡ 0.90 V (A2/C2) corresponding to the high-spin FeIII/II (i.e., PB/ES, ES: Everitts salt) and lowspin Fe(CN) 36 =4 redox transitions (PB/PY, PY: Prussian yellow), respectively, were detected [32 ± 36]. By comparing the above results, the A1/C1 and A2/C2 redox couples in the mvRuOx FeCN can be regarded as high-spin RuIII/II and low-spin Fe(CN) 36 =4 redox transitions, respectively, as the same assignment was indeed reported earlier [14]. The shift in redox potentials between PB and mvRuOx FeCN is related to the formation of oxy- and hydroxy-sites as will be discussed in a later section. The A1/C1 redox couple, however, does not participate in most catalytic oxidations of organic molecules. As to the A3/C3 redox process, which is different from the reported redox transitions of ruthenium as listed in Table 1, and therefore the exact oxidation state of ruthenium is unclear. We first attempt to rationalize the difference between the mvRuOx FeCN and classical PB and RuOx films with special attention to the unknown A3/C3 redox characterElectroanalysis 2004, 16, No. 15

istics. Figure 1B shows the typical CV obtained for the type II mvRuOx FeCN in pure acid solution of 10 mM HClO4. Interestingly, in the absence of alkali metal ion, a similar well-defined three-redox characteristic to that of Figure 1A was also noticed. It is well known that classical PB analogues always need alkali metals or ammonium ions for charge compensation and network stabilization with the following order: NH ‡4 > K‡ > Cs‡ > Na‡ >> Li‡ > Ba2‡ > Ca2‡ > H‡ [32 ± 34, 37, 49]. The behavior observed for the mvRuOx FeCN is thus very different from that of the simple PB case. In this case, it is expected that H‡ ion should play a key role for the electrochemical stabilization and redox behavior of the mvRuOx FeCN film. Potential segment analysis was next carried out to study the additional A3/C3 redox couple on the mvRuOx FeCN. If the A3/C3 peaks are coupled with other redox sites, the CV behavior should be strongly influenced in the potential segment analysis. As shown in Figures 1C and D, by holding the starting potential either at 0.5 V or ‡ 1.2 V and then scanned to different potential region, very little change in CV behavior was observed. Only the potential scan experiment from 0.5 to ‡ 1.2 V shows a small increase in the C3 ¹ 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Mixed Valence Ruthenium Oxide/Cyanometallate Catalysts

1215

Fig. 2. CV responses of the EC/RuOx (A), PB (B) and mvRuOx FeCN (C) in absence (a) and presence (b) of 0.5 M HCHO (i), C2H5 OH (ii) and glucose (iii) at v ˆ 50 mV/s in pH 2 KCl/HCl.

cathodic response (Fig. 1C). This is presumably caused by electrogeneration of the oxy/hydroxy-RuVII from its oxyRuVI site at high positive potentials ( ‡ 1.1 V). To check whether the electroactive species is loosely trapped in the network with an ideal-diffusional characteristic, the influence of the mass transfer effect within the films was then done under stirring. Since no considerable alteration in the mass transfer current was noticed under the hydrodynamic working conditions, the possibility is thus ruled out. Overall, the results indicate the existence of three different and individual redox sites in the mvRuOx FeCN polymeric network. This is further confirmed by comparing the typical response of the EC/RuOx with a clear redox peak of RuIV/III (most probably of RuO2) in pH 2 KCl/HCl solution [31]. The investigation with organic redox probes of formaldehyde, ethanol and glucose at PB and mvRuOx FeCN was found to provide very important information regarding the assignment of the A3/C3 ruthenium species. As can be seen in Figure 2A, the EC/RuOx shows an obvious catalytic response to formaldehyde and ethanol at potentials >‡ 0.7 V, but no further catalysis to glucose. Note that similar results were also obtained in our recent study on a lead ruthenate pyrochlore chemically modified electrode [43, 44]. The behavior on the EC/RuOx in pH 2 solutions is consistent with that of RuOx and oxy/hydroxy-Ru chemistry

[19, 20, 23]. The PB electrode, on the other hand, shows no catalytic responses to all three probes (Fig. 2B). Most important of all, the mvRuOx FeCN shows catalystic response to ethanol and formaldehyde in addition to unusual glucose oxidation by the A3/C3 redox couple at ‡ 1.10 V (vs. Ag/AgCl) (Fig. 2C). The catalytic potential is situated exactly on the oxy/hydroxy-RuVII/VI standard redox potential window of ‡ 1.35 to ‡ 1.45 V (vs. RHE) in pH 2 solution (Scheme 1). The existence of such an active and powerful redox site is the key for the efficient and enhanced catalysis of the mvRuOx FeCN. This is quite unique because glucose oxidation at low pHs has never been reported on any RuOx electrodes. Scheme 2 illustrates the mediation route for the oxidation of organic substrate (glucose) through the electrogenerated RuVII oxidation state. The unit cell of PB is a face-centered cubic structure (unit cell length ca. 10.2 ä) with the low spin iron coordinated to carbon terminals and high spin iron linked to nitrogen as FeII CN FeIII [32, 40]. Scheme 3 represents PB like structure of the mvRuOx FeCN with cyano-bridged Fe and oxy/hydroxy-Ru species in the 3D network. Such a complex structure was reported earlier for the case of osmium tetraoxide-modified hexacyanoruthenate films [50]. Nevertheless, this simple configuration cannot explain the extra

Scheme 2. Mediated oxidation of glucose using electrogenerated RuVII in the mvRuOx MCN film in acidic condition at 1.10 V (vs. Ag/AgCl). Electroanalysis 2004, 16, No. 15

¹ 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

1216

A. S. Kumar, J.-M. Zen

Scheme 3. Basic structural representation of Prussian blue (PB) and mvRuOx FeCN networks. The mvRuOx FeCN is a tentative structure.

based structure is the most probable arrangement, which can somehow stabilize the electrogenerated oxy/hydroxyRuVII without any further disproportion to lower oxidation states [19, 20]. The strong internal hydrogen bonding between oxy- and hydroxy- functional groups provides an extra force in stabilization. So far, only Os- and V-based cyanometallates were reported for the imperfect PB structures with oxy/hydroxy functional groups [40, 49, 50]. This report includes the Ru-PB analogue as a third compound in that category with quite unusual features. Of course, carefully designed XPS, XRD and some in situ electron spin resonance (ESR) spectroscopic experiments are necessary to confirm the above structure.

Scheme 4. Conceptional representation for the mvRuOx MCN catalytic film (most probable model) at electrogenerated conditions. The iRuVII(ˆO)hin the scheme is to illustrate the possible structure of higher oxidation state in the network.

redox transition of A3/C3 in the network. One possibility is that the oxy/hydroxy-RuVII species (similar to perruthenate: RuVIIO4 ) might be interlinked or trapped (presumably interlinked) with another oxy/hydroxy-RuIII/II in the PBtype 3D network as Fe CN RuIII/II O RuVII/VI . Similar to the presence case, a recent investigation on the corrosion protection of aluminum by chromate conversion coating (CCC) described the formation of such oxy and CN bridges in the network [51 ± 54]. In that process, the Fe(CN) 36 complex was used as an accelerator to increase the chromate coating formation, and the CrVIIO 4 acted as a self-healing agent for the corrosion protection. The oxy-Cr group was found to form a covalent inter-linkage with Fe CN and thereby to its stabilization [54]. Scheme 4 shows a possible configuration for the mvRuOx FeCN. It is similar to the structure of PB except with the oxy-RuVII/VI in the interstitial position. Two types of Ru species, namely RuIII/II NC Fe (A1/C1) and RuVII/VI O Ru N (A3/ C3) together with a low spin FeIII/II CN (A2/C2) were proposed here. The iRuVII(ˆO)h(oxo functional group) &

Electroanalysis 2004, 16, No. 15

3.2. Role of Base Electrolytes The alkali metal ions were found to play a major role in stabilization and to the redox properties of the PB analogues through insertion mechanism [40]. On the other hand, the H‡ (in acidic solution) and/or OH (in basic solution) insertion reaction was essential for the case of RuOx [20 ± 22]. To study the alkali metal and H‡ ions effects for the mvRuOx FeCN, two separate experiments in non-aqueous medium (CH3CN) with 0.01 M NaClO4 (case I) and pure 0.1 M, pH 5 alkali metal ion (Li‡ to Cs‡) solutions (case II) were carried out. As can be seen in Figure 3, under case I, the redox transitions in the mvRuOx FeCN completely disappear with very weak responses to HCHO and C2H5OH and no response to glucose. The shinny crystalline appearance is also vanished. As to the case II, the observed characteristics are similar to those of case I except that an additional irreversible peak appears at ca. ‡ 0.80 V. The film turns dull-brown in color with a catalytic behavior similar to the classical RuOx (Fig. 2A). It can thus be concluded that (i) the mvRuOx FeCN redox behavior depends more on the proton intercalation, (ii) there is no appreciable penetration effect of alkali metal ions in the network and (iii) the mvRuOx FeCN becomes destabilized around neutral conditions due to the formation of RuOx-like secondary structures. ¹ 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

1217

Mixed Valence Ruthenium Oxide/Cyanometallate Catalysts

Fig. 3. I) CV responses of the mvRuOx FeCN in A) 0.01 M HClO4 before (a) and after (b) continuous (n ˆ 5) cycling in CH3CN ‡ 0.01 M NaClO4 solution and B) CH3CN ‡ 0.01 M NaClO4. II) CV response of the mvRuOx FeCN in A) 0.01 M HClO4 before (a) and after (b) continuous (n ˆ 6) cycling in 0.1 M Cs‡, B) 0.1 M Li‡ and C) 0.1 M Cs‡. v ˆ 50 mV/s.

Jayalakshmi et al. recently investigated the electrochromic effect of PB (prepared in 0.1 M KCl) in nonaqueous medium (CH3CN ‡ alkali metal ions like K‡, Na‡ and Li‡) and reached the conclusion that Na‡ and Li‡ ions can increase the reversibility of the PB/ES redox peaks and the peak currents [37]. Since no such behavior was observed at the mvRuOx FeCN, the channel size of the film is therefore incompatible with the alkali metal ion insertion. It is further recognized that the basic cubic structure of the film, presumably occupied by bulky oxy/hydroxy groups and/or by strong internal hydrogen bonding in the framework, is not as open as the classical PB film. Disturbing of such a complicated systems may result in some secondary structure with predominant RuOx characteristics. The selective determination of Cs‡ in the presence of other alkali metal cations like Li‡ and Na‡ by Fu et al. should therefore not be expected at the mvRuOx FeCN [16]. In their studies the mvRuOx FeCN was prepared in 0.1 M KCl/HCl of pH 2 bath followed by scanning the electrode in the potential window of 0.5 to ‡ 1.1 V (vs. Ag/AgCl) in 0.1 M KCl (pH 5) until a steady profile was obtained. Such a film showed a total absence of the A3/C3 redox couple with only two redox transitions at ‡ 0.20 and ‡ 0.80 V (vs. Ag/AgCl). The selective determination of Cs‡ is believed to be an effect of the transformation of the mvRuOx FeCN into secondary structure discussed above. However, the exact behavior and characteristic of this secondary structure has not been addressed in detail.

3.3. Effect of pH As shown in Figure 4, the peak potentials are almost constant up to pH 3; whereas, sharp decrease in the A3/C3 peak currents were observed as the solution pH increases. Electroanalysis 2004, 16, No. 15

A new peak starts to form at the expense of the A3/C3 and A2/C2 redox processes in the region of pH 4 ± 8. This observation is different to the effect of pH on both the RuOx [20, 22, 23] and PB electrodes [36, 38, 39]. The abrupt change in the response at pH 4 is reasonable since the CN linkage in PB structure is highly unstable under neutral and alkaline pHs. Again, at pH > 4, the mvRuOx FeCN tends to convert into the secondary structures discussed earlier. The effect of ionic strength of [K‡] on the mvRuOx FeCN under pH 2 condition is also studied. As shown in Figure 5, unlike the pH effect, no obvious change in the redox peaks was observed. Both the A3/C3 and A2/C2 couples were almost unaltered by over two decades of difference in [K‡]. Only the A1/C1 redox peak showed a variation close to the Nernstian case (ca. 60 mV/decade) with an equal number of K‡/e . The variation of A1/C1 was considered to intercalate with alkali metal ion similar to the PB redox characteristic reported earlier [38, 39].

3.4. Other mvRuOx MCN Complexes Cyanometallate complexes from Cr, Ni, Cu, Pt and Ru were reacted with RuCl3 for the preparation of different mixed valence films in pH 2 solution. Table 2 summarizes the quantitative experimental results obtained from the organic redox probe studies. The basic feature about the higher oxidation state, which is responsible for the organic redox probes oxidations, is again compared. Interestingly, the net analyte catalytic current (i cat i bulk pa ˆ i pa pa † is in the order of mvRuOx CrCN < mvRuOx NiCN < mvRuOx CuCN < mvRuOx FeCN < mvRuOx PtCN < mvRuOx RuCN. This observation indicates the existence of the electrogenerated oxy-RuVII redox transition in all of the ¹ 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

1218

A. S. Kumar, J.-M. Zen

Fig. 4. Typical CV responses of the mvRuOx FeCN in 0.1 M KCl solution of different pHs at v ˆ 50 mV/s. The insert figures are typical plots of ipc and Epc against solution pH for various redox processes.

mvRuOx MCN films at ca. ‡ 1.1 V (vs. Ag/AgCl). Figure 6 shows the typical CV responses of the highly catalytic mvRuOx RuCN with various redox probes at a scan rate of 50 mV/s. Taking the catalytic oxidation of glucose at ‡ 1.1 V as a scale, the mvRuOx RuCN obviously possesses the highest capability for stabilizing the oxy/hydroxy RuVII species. The internal interaction-effect between the redox

Fig. 5. Typical CV responses of the mvRuOx FeCN in different [KCl] at pH 2. Insert figure is the typical Ep against log[KCl] plot for various redox processes at v ˆ 50 mV/s.

groups and oxy-/hydroxy- hydrogen bonding is likely the main reason for such an alteration in the i cat pa . It is important to notice that the A3/C3 redox peak at the mvRuOx RuCN is not as reversible as that at the mvRuOx FeCN. The highly diffused in nature of the A3/C3 redox couple at the mvRuOx RuCN in pH 2 solution is because that the A2/ A2 peak is buried in the huge A3/C3 peak. The i cat pa for glucose oxidation on the mvRuOx RuCN depends on the acidity of solution. As shown in Figure 7, a boosted catalytic response with a more reversible character-

Table 2. Electrode responses with various redox probes (0.5 M ) in pH 2 KCl/HCl ( I ˆ 0.1 M ) [a]. Electrode

C2H5OH

HCHO Epa (mV ) i

GCE PB EC/Ru [b] EC/RuOx mvRuOx FeCN mvRuOx FeCN[b] mvRuOx CrCN mvRuOx NiCN mvRuOx CuCN mvRuOx PtCN mvRuOx RuCN

± ± ± 800 [c] 1 099 1 087 1 049 1 039 1 046 1 166 [c] 1 149 [c]

cat pa

Glucose

(mA ) Epa (mV ) i

± ± ± 5.7 77.3 79.7 13.9 18.8 31.5 193.3 177.8

± ± 900 [c] 1 027 1 091 1 093 1 038 985 1 012 1187 [c] 1 140 [c]

cat pa

Remarks

(mA ) Epa (mV ) i

± ± 4.1 15.0 111.7 121.2 7.7 11.9 19.3 160.5 146.1

± ± ± ± 1 064 1 056 903 997 998 1 166 [c] 1 258

cat pa

(mA )

± ± ± ± 22.7 23.3 1.3 2.7 3.8 65.5 100.5

Not electroactive No electrocatalytic activity Catalysis almost nil Brown color film Stable and active More stable and active Peaks are very feeble Similar to mvRuOx-CrCN -doNo A3/C3 peak and predominant of Pt feature Very active and peaks are diffused

[a] CV at a scan rate of 50 mV/s. [b] Prepared in 0.01 M HClO4. [c] Measured as E1/2. analyte i cat i blank pa ˆ i pa pa

Electroanalysis 2004, 16, No. 15

¹ 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

1219

Mixed Valence Ruthenium Oxide/Cyanometallate Catalysts

Fig. 6. Typical CV responses of the mvRuOx RuCN with 0.5 M of different organic redox probes in 0.1 M, pH 2 KCl/HCl at v ˆ 50 mV/s.

istic at the A3/C3 redox couple was noticed when [HClO4] > 2 M. Both the A2/C2 and A1/C1 couples (which do not participate in catalysis), on the other hand, almost disappeared at such a high acidity. The exact reason, however, is not clear up to this point. Most importantly, the same observation for the A3/C3 couple was also found on HCl, H2 SO4 and H3PO4 suggesting the critical role of [H‡] on the catalysis. It is presumably because of the protonation of the oxy- and hydroxy- sites and induces a positive interaction of internal hydrogen bonding. Note that the concept of hydrogen bonding was reported previously for the oxy/hydroxy groups in classical RuOx [30, 55, 56]. However, such an unusual characteristics like this study has never been reported. Linear response was observed with increasing concentration of glucose at the mvRuOx RuCN in acid solutions. The film stability at high acidic conditions is much appreciable. The calibration plots show slopes of 0.26 mA/mM (high concentration range) and 1.1 mA/mM in 0.5 M and 5 M HClO4, respectively. The sensitivity is comparable to a recent report on cyclometalated Ru(II) complex/glucose oxidase enzyme in neutral pH (0.042 mA/mM) and RuOxmodified carbon paste electrode in strong alkaline solution (0.63 mA/mM) [26, 57]. Repetitive formation of the mvRuOx RuCN and detection of 0.1 M glucose in 0.5 M and 5 M HClO4 showed a relative standard deviation of < 3% (n ˆ 7). All these results indicate the superior performance of the mvRuOx MCN over the classical RuOx and PB systems because of the existence of unusual oxy/hydroxy-RuVII in acidic pHs.

2)

4. Conclusions

7)

Fig. 7. Typical CV response of the mvRuOx RuCN in A) 0.5 M and B) 5 M [HClO4] with/without 0.1 M glucose at v ˆ 10 mV/s. a) i cat pa vs. log[HClO4] plot for 0.1 M glucose (v ˆ 50 mV/s). b) The calibration plots.

3)

4)

5)

6)

A series of important conclusions can be derived from the above discussion. 8) 1) The electrochemical behavior of the mvRuOx MCN shows both the PB and oxy/hydroxy-RuOx character-

Electroanalysis 2004, 16, No. 15

istics. Unlike PB, the mvRuOx MCN depends more on [H‡] than alkali metal ions. Organic redox probes studies with formaldehyde, ethanol and glucose indicate the superior electrocatalytic property of the mvRuOx MCN over classical oxy/ hydroxy-RuOx in acidic pHs. Major redox couples of the mvRuOx FeCN are high spin CN RuIII/II (A1/C1), low spin FeIII/II CN (A2/ C2) and CN Ru O RuVII/VI (A3/C3). The unusual existence of the electrogenerated oxy/ hydroxy-RuVII in the mvRuOx MCN in acidic pHs at ‡ 1.1 V (vs. Ag/AgCl) plays a key role for the meditative electrocatalytic behavior of organic probes. The mvRuOx FeCN is stable only when pH < 3 and may transform into secondary structures with the predominant characteristic of classical RuOx as pH > 3. Among various cyanometallates, the mvRuOx RuCN was found to possess the highest ability of stabilizing the oxy/hydroxy-RuVII in the network. The mvRuOx RuCN offers a direct and enzymeless glucose sensing method in acidic pHs with a sensitivity comparable to that of RuOx and some enzyme-based systems. The superior electrocatalytic property of the mvRuOx MCN can be applied in glucose detection (chemical sensor), fuel cell (pace makers) and electroorganic synthesis. ¹ 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

1220 Further work is in process to further understand the unusual characteristic and for the practical chemical sensor applications.

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

6. References [1] J. A. Cox, L. Cheng, Electrocatalytic Determination of Biochemical Compounds. in Electroanalytical Methods for Biological Materials, (Eds: A. Brajter-Toth, J. Q. Chambers), Marcel Dekker, New York 2002, pp. 417 ± 438. [2] J. A. Cox, P. J. Kulesza, Anal. Chem. 1984, 56, 1021. [3] P. J. Kulesza, J. Electroanal. Chem. 1987, 220, 295. [4] J. Wang, Z. Lu, Electroanalysis 1989, 1, 517. [5] J. A. Cox, T. J. Gray, Anal. Chem. 1989, 61, 2462. [6] J. A. Cox, T. J. Gray, Anal. Chem. 1990, 62 2742. [7] R. T. Kennedy, L. Huang, M. A. Atkinson, P. Dush, Anal. Chem. 1993, 65, 1882. [8] W. Gorski, J. A. Cox, Anal. Chem. 1994, 66, 2771. [9] T. R. I. Cataldi, C. Campa, D. Centonze, Anal. Chem. 1995, 67, 3740. [10] J. A. Cox, K. S. Alber, C. A. Brockway, M. E. Tess, W. Gorski, Anal. Chem. 1995, 67, 993. [11] T. R. I. Cataldi, D. Centonze, A. Guerrieri, Anal. Chem. 1995, 67, 101. [12] T. R. I. Cataldi, D. Centonze, E. Desimoni, V. Forastiero, Anal. Chim. Acta 1995, 310, 257. [13] W. Gorski, J. A. Cox, J. Electroanal. Chem. 1995, 389, 123. [14] T. R. I. Cataldi, A. M. Salvi, D. Centonze, L. Sabbatini, J. Electroanal. Chem. 1996, 406, 91. [15] T. R. I. Cataldi, D. Centonze, E. Desimoni, Food Chem. 1996, 55, 17. [16] C. Fu, L. Wang, Y. Fang, Anal. Chim. Acta 1999, 391, 29. [17] P. J. Kulesza, B. Grzybowska, M. A. Malik, M. Chojak, K. Miecznikowski, J. Electroanal. Chem. 2001, 512, 110. [18] A. J. Bard, R. Parsons, J. Jordan, Standard Potentials in Aqueous Solutions, IUPAC, Marcel Dekker, New York 1983, p. 413. [19] K. W. Lam, K. E. Johnson, D. G. Lee, J. Electrochem. Soc. 1978, 125, 1069. [20] L. D. Burke, J. F. Healy. J. Electroanal. Chem. 1981, 124, 327. [21] M. E. G. Lyons, L. D. Burke, J. Chem. Soc. Faraday Trans I 1987, 83, 299. [22] S. Trasatti, Transition Metal Oxides: Versatile materials for Electrocatalysis. In Electrochemistry of Novel Materials (Eds.: J. Lipkowski, P. N. Ross), VCH, New York 1994, ch. 5; pp 207 ± 295 and references there in. [23] A. Senthil Kumar, K. C. Pillai, J. Solid State Electrochem. 2000, 4, 408 and references there in. [24] A. Senthil Kumar, Ph. D. Dissertation, A. C. College of Technology, University of Madras, India 1998.

Electroanalysis 2004, 16, No. 15

A. S. Kumar, J.-M. Zen [25] E. J. M. O×Sullivan, J. R. White, J. Electrochem. Soc. 1989, 136, 2576. [26] J. Wang, Z. Taha, Anal. Chem. 1990, 62, 1413. [27] M. E. G. Lyons, C. A. Fitzgerald, M. R. Smyth, Analyst 1994, 119, 855. [28] J.-W. Kim, S.-M. Park, J. Electrochem. Soc. 1999, 146, 1075. [29] V. Dharuman, K. C. Pillai. Bull. Electrochem. 1999, 15, 476 and references there in. [30] I. D. Belova, R. R. Shifrina, Y. E. Roginskaya, A. V. Popov, T. V. Varlamova, Soviet Electrochem. 1987, 23, 1138. [31] C.-C. Hu, Y.-H. Huang, J. Electrochem. Soc. 1999, 146, 2465. [32] K. Itaya, I. Uchida, V. D. Neff, Acc. Chem. Res. 1986, 19, 162. [33] A. Dostal, B. Meyer, F. Scholz, U. Schrˆder, A. M. Bond, F. Marken, S. J. Shaw, J. Phys. Chem. B. 1995, 99, 2906. [34] N. F. Zakharchuk, B. Meyer, H. Hennig, F. Scholz, A. Jaworksi, Z. Stojek, J. Electroanal. Chem. 1995, 398, 23. [35] J.-M. Zen, A. Senthil Kumar, H.-W. Chen, Electroanalysis 2000, 12, 542. [36] J.-M. Zen, A. Senthil Kumar, H.-W. Chen, Electroanalysis 2001, 13, 1171 and references there in. [37] M. Jayalakshmi, H. Gomathi, G. Prabhakara Rao, Solar Energy Mat. & Solar Cells 1997, 45, 201. [38] U. Scharf, E. W. Grabner, Electrochim. Acta 1996, 41, 233. [39] V. Krishnan, X. L. Xidis, V. D. Neff, Anal. Chim. Acta 1990, 239, 7. [40] N. R. de Tacconi, K. Rajeshwar, R. O. Lezna, Chem. Mater. 2003, 15, 3046. [41] De Andrade, R. Adalgisa R, M. P. Donate, P. P. D. Alves, C. H. V. Fidellis, J. F. C. Boodts, J. Electrochem. Soc. 1998, 145, 3839. [42] D.-T. Shieh, B.-J. Hwang, J. Electrochem. Soc. 1995, 142, 816. [43] J.-M. Zen, A. Senthil Kumar, J.-C. Chen, J. Mol. Catalysis A. Chemical 2000, 160, 277. [44] J.-M. Zen, A. Senthil Kumar, Acc. Chem. Res. 2001, 34, 772. [45] H. S. El Khadem, Carbohydrate Chemistry, Academic Press, New York, 1988, p. 144. [46] K. K. Gupta, S. S. Gupta, A. Mahapatra, J. Carbohydr. Chem. 1989, 8, 713. [47] S. Honda, K. Adachi, K. Kakehi, H. Yuki, K. Takiura, Anal. Chim. Acta 1975, 78, 492. [48] T. Saeki, J. Pharm. Soc. Japan 1950, 70, 690; B.-H. Perlmutter, A. Persky, J. Am. Chem. Soc. 1960, 82, 276. [49] D.Shaojun, L.Fengbin, J. Electroanal. Chem. 1986, 210, 31. [50] J. A. Cox, B. K. Das, J. Electroanal. Chem. 1987, 233, 87. [51] L. Xia, R. L. McCreery, J. Electrochem. Soc. 1999, 146, 3696. [52] L. Xia, E. Akiyama, G. Frankel, R. L. McCreery, J. Electrochem. Soc. 2000, 147, 2556. [53] G. S. Frankel, R. L. McCreery, J. Electrochem. Soc. Interface 2002, 10, 34. [54] P. Campestri, E. P. M. van Westing, A. Hovestad, J. H. W. de Wit, Electrochim. Acta 2002, 47, 1097. [55] J. Wang, C. Y. Fan, Q. Sun, K. Reuter, K. Jacobi, M. Scheffler, G. Ertl, Angew. Chem. Int. Ed. 2003, 42, 2151. [56] D. A. McKeown, P. L. Hagans, L. P. L. Carette, A. E. Russell, K. E. Swider, D. R. Rolison, J. Phys. Chem. 1999, 103, 4825. [57] A. D. Ryabov, V. S. Sukharev, L. Alexandrova, R. Le Lagadec, M. Pfeffer, Inorg. Chem. 2001, 40, 6529.

¹ 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim