A fundamental in situ IR spectroelectrochemical study of the electrical polarization of nickel, copper and gold electrodes in the presence of the unstable tellurocyanate ion in DMSO and DMF solutions

Abstract

Subtractively normalized interfacial Fourier transform infrared spectroscopy (SNIFTIRS) studies have been performed for the first time on the interaction of the tellurocyanate ion (TeCN⁻) with electrically polarised nickel (Ni), copper (Cu) and gold (Au) electrodes in dimethyl sulfoxide (DMSO) and dimethyl formamide (DMF) containing tetrabutylammonium perchlorate (TBAP) as a supporting electrolyte. In general, it was conclusively shown that the TeCN⁻ ion decomposes at certain applied potentials on electrically polarized electrochemical cells to form elemental tellurium and cyanide (CN⁻) ion. The cause of the decomposition was reasoned to be due to the presence of electrodissolved metal ions from the electrode, which destabilise the TeCN⁻ ion through their polarising power as cations. The severity of the destabilisation was found to increase in the order Au⁺ > Cu⁺ > Ni²⁺, which is in general agreement with the relative polarising power of these metal ions. SNIFTIRS and IR transmission data from model solutions demonstrated that the metal complex ion speciation observed during the polarization experiments therefore arose directly from the interaction of the studied metal electrode with CN⁻ ions released by decomposition, rather than with the TeCN⁻ ion itself. The potential at which the electrochemical decomposition of the TeCN⁻ ion occurred and its rate of decomposition depended on the type of metal electrode used. In general, the instability of the TeCN⁻ ion in electrochemically polarized systems reflects its fundamental chemistry of decomposing in the presence of polarizing cations and that it may be of use as a sensor for these species in an electrochemical setting.

---

Introduction

In systems of electrochemical interest, elements such as Te and Se, for instance, have attracted much interest in their role as electrode surface modifiers and the effects they have on the electrochemical behaviour of adsorbates of relevance. For instance, Te-modified single crystal Pt(111) electrodes have been studied with respect to CO adsorption and oxidation. When present as an adatom on the surface, the Te was found not to catalyse CO oxidation because the adatom always existed in its reduced form on the electrode and hence could not transfer oxygen to effect the oxidation.¹ This was found to be of use in inhibiting poisoning and accumulation caused by adsorbed CO in systems where formic acid was being oxidised, particularly at intermediate coverages of the adatom.^2,3 In other studies, technologically valuable films, such as cadmium telluride thin films, have been prepared by electrodeposition of Te on a Au electrode followed by reduction with metal ions present.⁴ Broadening the knowledge base of electrochemical processes involving tellurium based species is hence valuable in understanding their electrochemical behaviour, especially if the traditional voltammetry techniques used can be combined with a spectroscopic method (e.g. infrared spectroscopy) to visualise actual molecular species. We have chosen the TeCN⁻ ion in this study, because it allows facile monitoring by IR spectroscopy due to the ν(CN) stretching vibration being in an area of the IR spectrum (2500–1800 cm⁻¹) that is uncrowded by fundamental IR vibrations. It is also a species that has not been studied previously by IR spectroelectrochemical methods but whose analogues, i.e. related pseudohalide ions of cyanate (NCO⁻), thiocyanate (NCS⁻) and selenocyanate (NCSe⁻), have been subjected to previous IR spectroelectrochemical investigations in aqueous and polar aprotic solvents.^5–7

The successful preparation of the TeCN⁻ ion, an inherently unstable ion, was reported in 1968 by Downs,⁸ who prepared it from the reaction of tetraethylammonium cyanide and elemental Te in DMF and isolated pale yellow crystals that were highly moisture and oxygen sensitive. A Raman spectrum of a solution of the ion showed a peak at 2080 cm⁻¹. “Feebly polarizing” cations were necessary in the isolation of the solid salt, because the TeCN⁻ ion cannot exist in a solid in the presence of more strongly polarizing cations like potassium or even cesium. Although the TeCN⁻ ion is relatively stable in DMF or acetone, the addition of water will instantly cause decomposition to elemental Te and a CN⁻ ion.⁸ Austad et al.⁹ isolated pure tetraphenylarsonium tellurocyanate, after the reaction of tetraphenylarsonium cyanide and tellurium powder in acetonitrile solution. It was demonstrated that the tetraphenylarsonium salt of tellurocyanate was reasonably stable during storage as long as no solvent traces were left on the pale yellow crystals. Austad et al. also published ultraviolet (UV) and infrared (IR) data for the TeCN⁻ ion in acetonitrile and reported the ν(CN) stretching frequency of the ion at 2081 cm⁻¹. The IR stretching frequencies for the –TeCN group in solid tetramethylammonium tellurocyanate ((CH₃)₄NTeCN) and tetraphenylarsoniumtellurocyanate ((C₆H₅)₄AsTeCN) were reported in 1972 by Ellestad et al.¹⁰ Spencer et al.¹¹ reported the first preparation of the TeCN⁻ ion in DMSO and described its use in organic reactions to produce stable benzyl tellurocyanates. This and other aspects of its use in inorganic chemistry have been discussed further by Al-Rubaie et al.¹² Klaeboe et al.¹³ stated that the Te–C bond is very weak, which explains why it can only be isolated as a salt of large non-polarising cations like the tetraethylammonium ion or tetraphenylarsonium ion.

As the TeCN⁻ ion is stable in only a limited range of non-aqueous solvents, the number of electrochemical studies reported that involve this ion as an electroactive species has been few. For instance, Cauquis et al.¹⁴ reported a voltammetric study in 1975, in which the electrochemical oxidation of TeCN⁻ in acetonitrile at a platinum microelectrode was discussed and compared with the electrochemistry of other structurally related pseudohalide ions, i.e. NCO, NCS⁻ and NCSe⁻ ions. Using fast sweep rates, they were able to detect the electrochemical oxidation of TeCN⁻ ions via a series of reactions, in which the formation of “tellurocyanogen”, (TeCN)₂, from the combination of two tellurocyanate free radical species was proposed. The (TeCN)₂ species so formed was then thought to react further with more TeCN⁻ to form the (TeCN)₃⁻ ion. This (TeCN)₃⁻ species (with a formal Te oxidation state of −1.33, assuming even distribution of the negative charge over the Te atoms) was then postulated to oxidise to reform (TeCN)₂ (with a formal Te oxidation state of −1). The (TeCN)₂ species was then said to subsequently decompose to form elemental Te and a cyanogen ((CN)₂) species. It was found that fast sweep rates minimized the importance of the decomposition of the ion to elemental Te and (CN)₂ species. Cauquis et al.¹⁴ had commented on how the instability of the TeCN⁻ ion can cause challenges in the electrochemical detection of such transitory species.

Recently, we have applied in situ IR methods, such as subtractively normalised interfacial Fourier transform infrared spectroscopy (SNIFTIRS), to the study of Ni and Au electrodes anodically polarized in the presence of DMSO and DMF electrolytes containing the structurally related pseudohalide ions of NCO⁻, NCS⁻ and NCSe⁻.^5,6 These studies indicated that complex ion species arose from the interaction of electrogenerated Ni(II) and Au(I) ions with the NCO⁻, NCS⁻ and NCSe⁻ ions in the electrolyte. As the TeCN⁻ ion is fortuitously stable when prepared in dry DMSO and DMF solvents, it was decided to extend the previously reported work ^8,9,11 and study by SNIFTIRS Ni, Cu and Au electrodes electrically polarized in the presence of the TeCN⁻ ion to (1) provide information in an area where there is a complete dearth of previous scientific investigation and also (2) to compare the electrochemistry of the TeCN⁻ ion with that of its structurally related, more stable, analogue ions. It was also of interest to investigate whether the electrochemical behaviour of the TeCN⁻ ion, as reported by Cauquis et al.,¹⁴ would also occur on these metal electrodes and whether intermediate species such as [TeCN]₃⁻ would be detectable by IR spectroscopy, as well as to compare the behaviour of TeCN⁻ with that of related pseudohalides already studied. No previous IR spectroelectrochemical studies involving the TeCN⁻ ion have ever been reported. Hence, this study reports SNIFTIRS work coupled with voltammetric methods on the electrochemical behaviour of the TeCN⁻ ion on electrically polarised electrodes of Ni, Cu and Au. Further work carried out as part of this study involved the preparation of model solutions (via an electrochemically independent method) consisting of mixtures of TeCN⁻ ions or CN⁻ ions with Ni(II) and Cu(II) salts to help verify the species observed in the SNIFTIRS studies.

Experimental section

The solvents used in the study (DMF, Ajax Finechem Pty, UNIVAR grade and DMSO, Scharlau, synthesis grade) were used without further distillation. However, due to the moisture sensitivity of the TeCN⁻ ion, both solvents were dried over activated type 3A molecular sieve pellets to remove traces of water. Before drying over molecular sieves, the solvents, DMF and DMSO, were quoted in their specifications as having 0.1% and 0.15% water impurity as measured by Karl Fischer titration. All glassware was cleaned thoroughly with doubly distilled water prior to its use in experiments.

The TeCN⁻ ion was prepared following the procedure of Spencer et al.¹¹ An excess amount (0.40 g) of preground elemental tellurium (Aldrich Chemical Co, USA, 99.9%) was mixed with 0.163 g of potassium cyanide (KCN, (BDH, Anala-R grade)) which had been pre-dried in a desiccator containing activated silica gel. 50 mL of dried DMSO or DMF were added to the mixture and heat was applied (with the mixture having a N₂ gas flow from an oxygen free N₂ (OFN) cylinder supplied by BOC gases pass over it) to a temperature of 80–90 °C for 2 h to facilitate reaction. Successful reaction was indicated by the production of a pale yellow solution, which was then filtered from the unreacted Te solid. An infrared (IR) spectrum of the pale yellow solution in DMF or DMSO featured a peak at 2079 cm⁻¹ due to the TeCN⁻ ion.⁹ The yellow solution was stored under OFN until required. The concentration of TeCN⁻ produced was assumed to be 0.05 mol L⁻¹ based on the amount of KCN added.

Electrolyte solutions for the IR spectroelectrochemical investigation were prepared by adding solid tetrabutylammonium perchlorate (TBAP) from Aldrich Chemical Co, >97.0%, directly to the prepared ∼0.05 mol L⁻¹ TeCN⁻ solution, so that a concentration of 0.1 mol L⁻¹ TBAP was produced. The TBAP functioned as an inert supporting electrolyte.

The thin layer spectroelectrochemical cell, windows, specular reflectance unit and electrodes (working, reference and secondary) used in this study were as described and illustrated in a recent publication.⁵ The IR window used in the cell was a 32 × 3 mm CaF₂ window. The angle of incidence at the electrode surface was effectively 30°, due to the use of the FT-30 Spectra-Tech Specular Reflectance Unit on which the cell was placed. Due to the configuration of the cell, the signals detected were due to solution species existing in the thin layer between the IR transparent window and the working electrode. The working electrode consisted of a 7 mm diameter polycrystalline Ni, Cu or Au foil electrode embedded in a glass syringe barrel using Araldite epoxy glue. The Ni electrode was made from a 0.15 mm, 99.9% foil obtained from BDH, while the Au (from a 0.1 mm 99.99% foil) and the Cu (from a 0.25 mm, 99.98% foil) electrodes were purchased from Sigma Aldrich Chemical Co. Ltd.

The electrochemical and spectroscopic instrumentation used to carry out the cyclic voltammetry (CV) and IR spectroscopy of the systems was all as described in a previous publication.⁵ The CV was performed under thin layer conditions from −800 mV (AgCl/Ag) to +2000 mV (AgCl/Ag). The SNIFTIRS spectral acquisition technique involved single potential alteration, as has also been described in a previous study.⁵ Generally, the background spectra were acquired as the average of 100 scans at −900 mV (AgCl/Ag), with all of the sample spectra (also the average of 100 scans) being acquired at applied potentials more anodic than this at 4 cm⁻¹ resolution.

When using SNIFTIRS as the method of subtractive normalisation in the present study, the spectral configuration used led to spectra that (when containing an increasing amount of an IR absorbing species and produced with an applied potential more anodic than the background potential) gave a positive-going signal. Accordingly, any species existing in greater concentrations at the background potential therefore appeared as a negative-going signal. SNIFTIRS spectra are hence displayed as +log₁₀(R₂(λ)/R₁(λ)) (R = reflectance, λ = IR wavelength, subscripts 1 and 2 refer to the reflectance of the electrode at the background potential and sampling potential respectively) rather than the opposite, i.e. as the negative log term, as has been the convention.¹⁵

Current–potential data from the systems were collected at the same time as the IR data by averaging over (1) the starting current in the cell at the beginning of an IR acquisition and (2) the final current (after several minutes of acquisition) at the end of the scan to generate effectively what could be regarded as a “slow”, “single sweep” voltammogram of the electrode system under study.

Model solutions in the case of the Ni and Cu systems were also prepared to provide experimentally independent confirmation of the species detected and assigned in spectra acquired of the electrodes under SNIFTIRS conditions. These involved mixing partially pre-dried¹⁶ metal salts (e.g. CuCl₂·2H₂O and Ni(NO₃)₂·6H₂O) in DMSO along with TeCN⁻ or CN⁻ ions in metal ion : TeCN⁻ or metal ion : CN⁻ molar ratios ranging from 1 : 1 to 1 : 4. These model solutions were sampled by FTIR in conventional transmission mode using a Press-Lok™ cell via procedures described previously.⁵

Results and discussion

1. Cyclic voltammograms and single sweep voltammograms

1.1. Ni/TeCN⁻/TBAP/DMSO and Ni/TeCN⁻/TBAP/DMF electrochemical systems. Fig. 1(a) and (b) show the cyclic voltammograms of the nickel electrode in TeCN⁻/TBAP/DMSO and TeCN⁻/TBAP/DMF based electrolytes, respectively. Both voltammograms share the common trait of indicating a sharp increase in current in the anodic region above 0.0 mV (AgCl/Ag) with the formation of a broad peak, albeit at different values of the applied potential. The cyclic voltammograms were recorded under thin-layer conditions. Some of these (e.g. Fig 1(c) and (d)) were affected by resistance occurring in the thin layer resulting in a sharp “resistive tilt” in the appearance of the voltammogram.^17,18

Fig. 1 Cyclic voltammograms of the nickel, copper and gold electrodes in DMSO and DMF solvents containing tellurocyanate ions and 0.1 mol L⁻¹ TBAP (sweep rate = 20 mV s⁻¹): (a) Ni electrode and ∼0.05 mol L⁻¹ TeCN⁻ in DMSO, (b) Ni electrode and ∼0.05 mol L⁻¹ TeCN⁻ in DMF, (c) Cu electrode and ∼0.05 mol L⁻¹ TeCN⁻ in DMSO, (d) Cu electrode and ∼0.05 mol L⁻¹ TeCN⁻ in DMF, (e) Au electrode and ∼0.05 mol L⁻¹ TeCN⁻ in DMSO, and (f) Au electrode and ∼0.05 mol L⁻¹ TeCN⁻ in DMF. Arrows show the path traced upon conducting the sweep of potentials.

Fig. 2(a) and (b) show the “single sweep voltammograms”, which are effectively “slow-scan” voltammograms for the systems studied. These are the current–voltage plots for the systems obtained during the SNIFTIRS spectral acquisitions. In common with the cyclic voltammograms, there is a broad similarity in appearance for the single sweep voltammograms, although the peaks observed occur at different applied voltages for the different solvents. Differences in appearance between the cyclic voltammograms recorded under relatively faster scanning conditions and the single sweep voltammograms can be ascribed to mass transport issues in the thin layer. The plots in Fig. 2(a) and (b) (and others) do not show the resistive tilt effect, as the electrode has more time to respond to changes in the applied potential under the conditions of the SNIFTIRS experiment.

Fig. 2 Single sweep voltammograms acquired of the nickel, copper and gold electrodes during a SNIFTIRS experiment in DMSO and DMF solvents containing tellurocyanate ions and 0.1 mol L⁻¹ TBAP: data were obtained from calculating the average current at the beginning and end of the spectral acquisition period for each applied potential at which SNIFTIRS spectra were acquired. (a) Ni electrode and ∼0.05 mol L⁻¹ TeCN⁻ in DMSO, (b) Ni electrode and ∼0.05 mol L⁻¹ TeCN⁻ in DMF, (c) Cu electrode and ∼0.05 mol L⁻¹ TeCN⁻ in DMSO, (d) Cu electrode and ∼0.05 mol L⁻¹ TeCN⁻ in DMF, (e) Au electrode and ∼0.05 mol L⁻¹ TeCN⁻ in DMSO and (f) Au electrode and ∼0.05 mol L⁻¹ TeCN⁻ in DMF.

The suspicion of decomposition occurring in both of the studied systems was raised by the observation that the Ni electrodes turned black when either the CV experiment was initiated or a SNIFTIRS experiment was begun. In further discussion on the SNIFTIRS spectra, more detail, such as when the decomposition begins to occur on the electrode surface, will be provided.

1.2. Cu/TeCN⁻/TBAP/DMSO and Cu/TeCN⁻/TBAP/DMF electrochemical systems. Fig. 1(c) and (d) show the cyclic voltammograms of the copper electrode in TeCN⁻/TBAP/DMSO and TeCN⁻/TBAP/DMF based electrolytes, respectively. In contrast to the Ni electrode systems, there is a greater similarity in the appearance of the voltammograms of the Cu electrode systems recorded in DMSO and in DMF. In addition, it is clear that the Cu electrode systems appear more reactive, as their maximum anodic currents (8–10 mA) are 3–5 times the maximum currents observed in the Ni electrode systems (1.8–3 mA). Compared to the Ni electrode cyclic voltammograms, the Cu electrode systems are generally featureless and show merely an increase in current starting from negative currents at cathodic potentials to positive currents at anodic potentials. As stated earlier, the resistive tilt observed in some of the cyclic voltammograms would be caused by mass transport issues, however, the single sweep voltammograms do not exhibit this effect. In common with the Ni electrode systems, the Cu electrodes showed evidence of the decomposition of the TeCN⁻ ion occurring in the electrochemical systems during CV and SNIFTIRS experiments.

1.3. Au/TeCN⁻/TBAP/DMSO and Au/TeCN⁻/TBAP/DMF electrochemical systems. Fig. 1(e) and (f) show the cyclic voltammograms of the gold electrode in TeCN⁻/TBAP/DMSO and TeCN⁻/TBAP/DMF based electrolytes, respectively. The single sweep voltammograms are illustrated in Fig. 2(e) and (f). In Fig. 1(e) and (f), the Au electrode cyclic voltammograms look similar in appearance below 0 mV (AgCl/Ag), but have different appearances in the anodic region above 0 mV (AgCl/Ag). In common with the other two electrode systems, a blackening of the electrode was observed, suggesting decomposition of the TeCN⁻ ion had occurred at the electrode surface.

In conclusion, observations from the preliminary electrochemical investigations using cyclic voltammetry have shown evidence of the decomposition of the tellurocyanate ion during the electrical polarisation of the metal electrodes (Ni, Cu and Au). The cause of this decomposition is not due to water traces, but due to the destabilisation of the tellurocyanate ion when coming into contact with highly polarizing cations.^8,9 This inherent chemical instability has precluded in the past⁸ the isolation of such stable salts as potassium tellurocyanate, for instance. In the case of the electrochemical systems being studied, the “ions” regarded as the fundamental cause for this decomposition are most likely the Ni²⁺, Cu⁺ and Au⁺ ions that have formed during the dissolution of the electrode during polarisation. SNIFTIRS and average current measurements have provided some insight into the point at which decomposition of the TeCN⁻ ion occurred, as well as the rate of decomposition, and the consequences in terms of the species produced as a result of the decomposition in the three studied metal electrode systems.

2. SNIFTIRS spectra of the Ni, Cu and Au electrodes in the presence of tellurocyanate ions

2.1. Ni/TeCN⁻/TBAP/DMSO and Ni/TeCN⁻/TBAP/DMF electrochemical systems. Fig. 3(a) and (b) present SNIFTIRS spectra of the nickel electrode systems in DMSO and DMF, respectively, when electrically polarised in the presence of the tellurocyanate ion. Table 1 summarises the FTIR data from the systems studied.

Fig. 3 Series of SNIFTIRS spectra of the nickel, copper and gold electrodes at different applied potentials in DMSO and DMF solvents containing tellurocyanate ions and 0.1 mol L⁻¹ TBAP. (a) Ni electrode and ∼0.05 mol L⁻¹ TeCN⁻ in DMSO, (b) Ni electrode and ∼0.05 mol L⁻¹ TeCN⁻ in DMF, (c) Cu electrode and ∼0.05 mol L⁻¹ TeCN⁻ in DMSO, (d) Cu electrode and ∼0.05 mol L⁻¹ TeCN⁻ in DMF, (e) Au electrode and ∼0.05 mol L⁻¹ TeCN⁻ in DMSO, and (f) Au electrode and ∼0.05 mol L⁻¹ TeCN⁻ in DMF.

Table 1 FTIR data from in situ IR spectroelectrochemical studies of the Ni/TeCN⁻, Cu/TeCN⁻ and Au/TeCN⁻ systems electrically polarized in 0.1 mol L⁻¹ TBAP in DMSO or DMF solvents

System studied	ν(CN) of free TeCN^− ion (cm^−1)	ν(CN)^a of Ni(II)/CN^− complex ion (cm^−1)	ν(CN)^a of Cu(I)/CN^− complex ion (cm^−1)	ν(CN) of solid NCO^− (cyanate) species (cm^−1)	ν(CN)^a of Au(I)/CN^− complex ion (cm^−1)	ν(CO) of CO2 dissolved in solvent (cm^−1)	Colour of cell solution after SNIFTIRS experiment
a Note that the TeCN^− ion decomposed in the experiment, so the ν(CN) stretching frequencies refer to complexes of the respective metal ions Ni(II), Cu(I) and Au(I) ions with the CN^− ion, rather than with TeCN^−. nd = not detected.
Ni/DMF/TeCN^−	2078	2112		2158		2338	Colourless
Cu/DMF/TeCN^−	2079		2127			2337	Gold yellow
Au/DMF/TeCN^−	2079			2156		2337	Colourless
Ni/DMSO/TeCN^−	2079	2112		2165		2337	Colourless
Cu/DMSO/TeCN^−	2079		2125			2337	Gold yellow
Au/DMSO/TeCN^−	2077				2140	nd	Colourless

---

The most characteristic feature in the SNIFTIRS spectra of the Ni/TeCN⁻/TBAP/DMSO and Ni/TeCN⁻/TBAP/DMF systems is a negative-going peak at 2078–2079 cm⁻¹ appearing at all potentials, which can be clearly attributed to the ν(CN) stretching vibration of the TeCN⁻ ion. This is always present as a negative-going peak of varying intensity in the spectra, indicating that it is present in greater amounts in the background spectrum. This negative feature becomes very noticeable in the spectra at applied potentials >0 mV (AgCl/Ag), where the decomposition of the TeCN⁻ ion has proceeded rapidly.

As stated earlier, the observation of a negative-going peak in all of the spectra indicates that there were more TeCN⁻ ions in the original background spectrum acquired at −900 mV (AgCl/Ag) than at the potentials anodic to this background potential, although the (negative) intensity of the peak does not increase in magnitude until the Ni electrode is adjusted to more positive potentials (see Fig. 4(a) and (b)). This and the observation of blackened electrodes (see earlier) in each system are proof that the tellurocyanate ion is decomposing in the thin layer cell when subjected to electrode polarisation, especially in the regions where the nickel electrode is adjusted to potentials where Ni(II) ion species are obviously being produced.

Fig. 4 Plots of the intensity changes representing the quantities of various molecular species generated in the thin layer during electrochemical polarisation and observed in the SNIFTIRS spectra at different applied potentials in the nickel, copper and gold electrodes in DMSO and DMF solvents containing tellurocyanate ions and 0.1 mol L⁻¹ TBAP. (a) Ni electrode and ∼0.05 mol L⁻¹ TeCN⁻ in DMSO, (b) Ni electrode and ∼0.05 mol L⁻¹ TeCN⁻ in DMF, (c) Cu electrode and ∼0.05 mol L⁻¹ TeCN⁻ in DMSO, (d) Cu electrode and ∼0.05 mol L⁻¹ TeCN⁻ in DMF, (e) Au electrode and ∼0.05 mol L⁻¹ TeCN⁻ in DMSO, and (f) Au electrode and ∼0.05 mol L⁻¹ TeCN⁻ in DMF.

The plots in Fig. 4(a) and (b), as discussed, constitute the electrochemical evidence that the TeCN⁻ ion is decomposing, and it was observed that the potential at which this began to occur on the Ni electrode surface was −400 mV (AgCl/Ag). At potentials more cathodic than this value, the intensity of the ν(CN) stretching frequency peak of the TeCN⁻ ion changes negligibly, but after −400 mV (AgCl/Ag) it slowly becomes more negative and undergoes a very sharp decrease to even more negative intensities after 0 mV (AgCl/Ag). The reason that the decomposition started to occur on the electrode is that at −400 mV (AgCl/Ag) a small amount of anodic dissolution (i.e. oxidation of Ni to the Ni(II) ion) began to occur on the Ni electrode. Although the current increase in this region (i.e. between −400 mV (AgCl/Ag) is low initially, it increases significantly in the single sweep voltammograms (see Fig. 2(a) and (b)) as the applied potential is adjusted to more anodic values. The spectral evidence for Ni being oxidised to Ni(II) in the SNIFITRS spectra is provided by the fact that the intensity of the peak at 2112 cm⁻¹, attributed to the species [Ni(CN)₄]²⁻ (see later), also begins to increase in intensity from −400 mV (AgCl/Ag) and is maximised at +300 mV (AgCl/Ag). Hence, Ni²⁺ ions are being produced at the electrode, which are causing the decomposition of the tellurocyanate ion, with the Ni²⁺ ions resultantly becoming complexed with the released CN⁻ ions. The TeCN⁻ ion is hence sensitive to these ions when they are generated at the electrode surface.

In the SNIFTIRS spectra illustrated for both of the studied Ni/TeCN⁻ systems (Fig. 3(a) and (b)), other weak spectral features are evident. For example, in the Ni/DMSO/TeCN⁻ and Ni/DMF/TeCN⁻ systems, a weak peak at 2112 cm⁻¹ is observed. This peak, as discussed earlier, can be confidently assigned to the ν(CN) stretching vibration of the species [Ni(CN)₄]²⁻ in the DMSO or DMF solvent. Although the literature on the preparation of Ni–cyano complex ions in polar aprotic solvents is quite limited, a study by Semenishin¹⁹ has reported an IR ν(CN) stretching frequency for [Ni(CN)₄]²⁻ prepared in DMF and DMSO of 2117 cm⁻¹. In the same study and others,²⁰ it was also reported that there was a bathochromic shift of the IR stretching frequency when going from aqueous media to polar aprotic solvent media. In earlier reported Ni/CN⁻ SNIFTIRS studies²¹ in aqueous systems, the corresponding ν(CN) stretching frequency for this species was reported to be 2124 cm⁻¹. Hence the species at 2112 cm⁻¹ in the present study can be assigned to [Ni(CN)₄]²⁻ and not to species such as [Ni(CN)₅]³⁻, which are known not to form in these polar aprotic solvents due to the formation of a stronger bond between the metal ion and the DMSO or DMF solvent.¹⁹ The assignment of the 2112 cm⁻¹ peak to [Ni(CN)₄]²⁻ was also experimentally verified by model solution studies (in DMSO). In these studies, a solution of [Ni(CN)₄]²⁻ was prepared independently via the addition of Ni²⁺ salts and KCN, and Ni(II) salts were also added to TeCN⁻ ions directly, and their ν(CN) stretching frequency was confirmed using transmission IR spectroscopy (see later). In terms of the SNIFTIRS experiment, the complex ion [Ni(CN)₄]²⁻ would have formed through the release of CN⁻ ions in the solution from the decomposition of the TeCN⁻ ions (see eqn (1) for the suggested half equation) and also because of the high thermodynamic stability of this ion (which in water, for instance, has a log(β₄/(mol L⁻¹))⁴ value of 30.5 ± 0.3).²²

4TeCN⁻ + Ni_(s) → [Ni(CN)₄]²⁻ + 4Te_(s) + 2e⁻ (1)

The species, though stable, does not constitute a very dominant feature in the SNIFTIRS spectra. It is suspected that the electrode surface has become modified by the deposited Te layer, such that only a small amount of the Ni/cyanide complex can form due to a blocking effect.

Other peaks observed in Fig. 3(a) and (b) were weak and there were broad features from 2158–2165 cm⁻¹. It is believed that these are due to the ν(CN) stretching frequencies of solid cyanate species, i.e. KOCN(s), that have formed on the electrode during the in situ IR experiment. These will have formed as a result of CN⁻ ion oxidation at the electrode surface. The source of the CN⁻ ion may be from the [Ni(CN)₄]²⁻ complex, especially in the Ni/TeCN⁻/TBAP/DMSO electrochemical system, where the spectral intensity of the peak assigned to [Ni(CN)₄]²⁻ is seen to decrease to zero when the peak due to KOCN(s) starts to appear in the spectra (see Fig. 4(a)). The reason for observing KOCN(s) in the Ni/TeCN⁻ system is due to the low solubility of the NCO⁻ ion in DMSO, which would be exacerbated by the fact that thin layer conditions were being used, which are prone to supersaturation if a low solubility species (in the solvent being used) was being electrogenerated. Earlier SNIFTIRS studies involving cyanate ions in DMSO or DMF have used concentrations of half the value of 0.05 mol L⁻¹ (the expected value of [CN⁻] if generated from the decomposition of 0.05 mol L⁻¹ TeCN⁻ ion, assuming a 1 : 1 stoichiometry of ions), at which the cyanate ion is known to be soluble from experimental experience. Hence, the reason for only seeing a peak attributable to solid KOCN (as opposed to a solution species) is that it is being produced in the thin layer at a concentration that exceeds the solubility of the cyanate ion in DMSO. Had soluble cyanate species been detected, a peak at the lower wavenumber value of 2137 cm⁻¹, as has been observed in previous SNIFTIRS studies in this solvent, would have been detected.⁵

In the Ni/TeCN⁻/TBAP/DMF electrochemical system, the spectral intensity trend for the [Ni(CN)₄]²⁻ ion is not so clear-cut but does show a reduction in intensity at very anodic potentials, where the KOCN(s) formation is observed to occur. An additional feature observed in SNIFTIRS spectra is the appearance of a weak peak at 2337–2338 cm⁻¹, which can be assigned on the basis of earlier studies⁵ to dissolved CO₂ in the DMSO or DMF solvent. This peak was at its most intense only in the Ni/DMSO/TeCN⁻ system (the Cu and Au electrode systems either showed minor CO₂ content or none). In contrast, CO₂-associated peaks were weak or non-existent in the DMF solvent-based system of Ni. This CO₂ was electrogenerated during the in situ IR studies and arose from the oxidation of either the DMSO solvent or, more likely, from the liberated cyanide via the decomposition of TeCN⁻ at the electrode.

2.2. Cu/TeCN⁻/TBAP/DMSO and Cu/TeCN⁻/TBAP/DMF electrochemical systems. Fig. 3(c) and (d) illustrate the in situ IR (SNIFTIRS) spectra for the Cu/TeCN⁻/TBAP/DMSO or DMF electrochemical systems. Fig. 4(c) and (d) are graphs showing the intensity trends of species detected in the in situ IR experiments. Table 1 summarises the IR data from this system. As observed in the studied Ni electrode systems, the (negative-going) peak due to the TeCN⁻ ion at 2079 cm⁻¹ was an important feature in the spectrum, with the negative intensity indicating decomposition over part of the range of the applied potentials studied. The spectral intensity graphs of detected species as a function of applied potential (Fig. 4(c) and (d)) confirm that the decomposition of the TeCN⁻ ion is accelerated as the electrode is increasingly polarized in the anodic direction of potential, with a starting point of approximately −500 mV (AgCl/Ag). On Ni, the decomposition of the TeCN⁻ ion starts to occur at approximately −400 mV (AgCl/Ag). The earlier decomposition (though small) of the TeCN⁻ ion on the Cu electrode (starting at −500 mV (AgCl/Ag)) is due to the greater reactivity of the Cu electrode when it is immersed in the DMSO or DMF solvent and would indicate that Cu(I) species are being released from the electrode at these more cathodic applied potentials relative to Ni. As in the Ni electrode case, these released Cu(I) ions are likely to be destabilising the TeCN⁻ ion, and so leading to its decomposition into Te and CN⁻.

At the same potential that the decomposition of the TeCN⁻ ion is observed to occur on the Cu electrode (from −500 mV (AgCl/Ag)), weak, broad peaks at 2125–2127 cm⁻¹ are observed to occur in the SNIFTIRS spectra of the Cu/DMSO/TeCN⁻ and Cu/DMF/TeCN⁻ systems (see Fig. 3(c) and (d)). These are presumed, on the basis of earlier in situ IR studies of Cu/CN⁻ systems,²³ to be due to [Cu(CN)₂]⁻ ions. As these form at the time that the TeCN⁻ ion is beginning to decompose, we propose that these electrogenerated Cu⁺ ions are causing the decomposition and reacting with the released CN⁻ ions with a suggested formation half equation, shown in eqn (2), of:

2TeCN⁻ + Cu_(s) → [Cu(CN)₂]⁻ + 2Te + e⁻ (2)

In support of this, model solution IR data from solutions prepared by mixing Cu(II) salts and CN⁻ ions in DMSO have convincingly shown that this peak is in fact due to the Cu(I) species [Cu(CN)₂]⁻ (see later).

Additional features observed in the spectral intensity vs. applied potential plots (Fig. 4(c) and (d)) that also suggest the inherently greater reactivity of the Cu electrode systems are fluctuations in the observed intensity trends visible only at very high anodic applied potentials. This behaviour might be associated with reactive electrochemical systems showing a degree of turbulence or disturbance in the thin layer due to chemical changes caused through the application of the very positive potentials to the electrode surface. This would affect the intensities of species referenced to a background potential, where the activity in the thin layer is more quiescent.

2.3. Au/TeCN⁻/TBAP/DMSO and Au/TeCN⁻/TBAP/DMF electrochemical systems. Fig. 3(e) and (f) illustrate the SNIFTIRS spectra for the Au/TeCN⁻/TBAP/DMSO or DMF electrochemical systems. Fig. 4(e) and (f) are graphs showing the intensity trends of species detected in the SNIFTIRS experiments. Table 1 summarises the IR data from these systems. As observed in the studied Cu and Ni electrode systems, the (negative-going) peak due to the TeCN⁻ ion at 2079 cm⁻¹ was an important feature in the spectrum, with the negative intensity also suggesting that the TeCN⁻ ion is decomposing on the Au electrode systems over the range of the applied potentials studied. The spectral intensity graphs of detected species at different applied potentials (Fig. 4(e) and (f)) confirm that the decomposition of the TeCN⁻ ion accelerates as the electrode is increasingly polarized in the anodic potential direction. On Au, however, the potential at which this decomposition begins to occur is approximately −200 mV (AgCl/Ag). This event is paralleled in the Au/DMSO/TeCN⁻ SNIFTIRS spectra by an increase in the intensity of a peak at 2140 cm⁻¹ (see Fig. 4(e) and (f)), which has been assigned to the ν(CN) stretching frequency of the complex ion [Au(CN)₂]⁻ (see later). This indicates that the Au electrode is beginning to dissolve to form Au⁺ ions around this potential, which is in turn destabilising the TeCN⁻ ion. Compared to the Ni and Cu system intensity data for this species, the rate of decline of the TeCN⁻ species on the Au electrode, as indicated by the intensity curve for the peak at 2079 cm⁻¹ as a function of applied potential, is less acute. This suggests that decomposition proceeds to a lesser extent on Au, which agrees with the fact that Au⁺ as a cation has lower polarising power relative to the Cu⁺ and Ni²⁺ ions.

In the Au/DMF/TeCN⁻ SNIFTIRS spectrum, even though there is no characteristic peak due to [Au(CN)₂]⁻, but weak peaks due to solid KOCN (2156 cm⁻¹) and CO₂ (2337 cm⁻¹),^5,6 it is still obvious from the intensity–potential plots that TeCN⁻ decomposes in this system. It is thus posited that Au is still dissolving in some form in this system to form Au⁺ (which in turn is causing the decomposition of the TeCN⁻ ion), even though the presence of Au⁺ is not able to be verified in the SNIFTIRS spectra. The reason for there being no observation of [Au(CN)₂]⁻ in the DMF-based system is most likely due to the lower solubility of KOCN (an oxidation product of CN⁻) in the DMF solvent that was used. Indeed, this lack of solubility of KOCN in DMF was confirmed when making different model solutions of various metal salt : potassium cyanate molar ratios in DMF and DMSO (see later).

As mentioned earlier, SNIFTIRS spectra for the Au/DMSO/TeCN⁻ system (Fig. 4(e)) featured a weak peak at 2140 cm⁻¹, which can be assigned to the [Au(CN)₂]⁻ ion.^6,24 In DMSO, this would have formed from an interaction between the electrogenerated Au(I) ion and the tellurocyanate ion, thereby causing its decomposition with the release of a cyanide ion that would have subsequently complexed the Au⁺ ion to form [Au(CN)₂]⁻ (see eqn (3) for the suggested half equation demonstrating its formation).

2TeCN⁻ + Au_(s) → [Au(CN)₂]⁻ + 2Te + e⁻ (3)

The formation of this ion is not surprising, as the [Au(CN)₂]⁻ complex ion has one of the highest thermodynamic stabilities known.²⁵ Its formation would be far more highly favoured over a complex ion of Au(I) with the TeCN⁻ ion, for instance, the formation of which would be precluded anyway by the mild polarising power of Au(I) ions.

3. IR data from the model solutions prepared by mixing Ni(II) and Cu(II) salts with tellurocyanate solutions prepared in DMSO

Model solutions were prepared to verify the assignments made for peaks observed in the Ni and Cu SNIFTIRS spectra, as discussed above. Gold/tellurocyanate model solutions were not generated in this study. The model solution methodology has been used in earlier studies for confirming⁵ electrochemically generated species via an electrochemically independent generation method. These solutions were prepared by the direct addition of solid Ni(II) and Cu(II) salts to freshly generated solutions of TeCN⁻ ions in DMSO. Fig. 5 illustrates the model solution spectra for Ni(II) and Cu(II) ions interacting with the tellurocyanate solution generated in DMSO. Fig. 5(a) shows the IR spectrum of the free tellurocyanate ion when it is prepared from the reaction of elemental tellurium with potassium cyanide. The intense peak at ca. 2078 cm⁻¹ is due to the ν(CN) stretching frequency of the TeCN⁻ ion.⁹ Fig. 5(b) and (c) present the spectra obtained from mixtures of the TeCN⁻ ion with Cu(II) and Ni(II) salts, respectively, in the metal ion (Cu(II) or Ni(II)) : TeCN⁻ ion molar ratio of 1 : 2.

Fig. 5 IR transmission spectra for the (a) pale yellow solution of tellurocyanate ions generated in DMSO from the reaction of the 1 : 1 molar ratio mixture of finely ground elemental Te and KCN, (b) 1 : 2 molar ratio model solution prepared by mixing partially dehydrated CuCl₂ (0.025 mol L⁻¹) and TeCN⁻ (∼0.05 mol L⁻¹) in DMSO (solution observed to deposit black solid), and (c) 1 : 2 molar ratio model solution prepared by mixing partially dehydrated Ni(NO₃)₂ (0.025 mol L⁻¹) and TeCN⁻ (∼0.05 mol L⁻¹) in DMSO (solution observed to deposit black solid).

However, when adding Ni(II) or Cu(II) metal salts to solutions of free TeCN⁻ ions in DMSO, an immediate reaction was observed to occur, which resulted in the formation of a black solid that deposited in the reaction container. This was a firm indication that the TeCN⁻ ion was very unstable in the presence of Ni(II) and Cu(II) ions, with the decomposition of the ion into black finely divided elemental Te and a free cyanide ion⁹ being a likely consequence. This result is not only expected given the work of Austad et al.⁹ in proving that the TeCN⁻ ion is unstable in the presence of highly polarizing cations, but also strengthens the above interpretations of the SNIFTIRS spectra, regarding the instability of the TeCN⁻ ion when metal ions begin to be released during the polarisation-induced dissolution of the electrodes in the studied systems. This property of the TeCN⁻ ion may be of utility as a highly sensitive and reactive “marker” for the presence or release of oxidised, soluble metal species on electrodes. Although the TeCN⁻ ion has decomposed in the model solutions, the IR spectra of these solutions (Fig. 5(b) and (c)) clearly feature two weak peaks at 2127 cm⁻¹ and 2112.8 cm⁻¹, respectively, for the Cu(II)/TeCN⁻ and Ni(II)/TeCN⁻ model solutions. These can be assigned with reasonable confidence, on the basis of similar observations (for the Cu/CN⁻ system from earlier work²³), to the [Cu(CN)₂]⁻ ion (ca. 2127 cm⁻¹) and the [Ni(CN)₄]²⁻ ion (ca. 2112 cm⁻¹).¹⁹ These assignments were further experimentally proven by recording FTIR transmission spectra of other model solutions prepared by mixing Ni(II) and Cu(II) salts with KCN salt in DMSO in various Ni(II) : KCN and Cu(II) : KCN molar ratios varying from 1 : 1 to 1 : 4 (see Table 2 for FTIR data). In all of the Ni/KCN model solutions examined by transmission FTIR, a single peak at ca. 2112 cm⁻¹ was observed, which can be assigned to the [Ni(CN)₄]²⁻ complex ion in this solvent. This therefore proves the identity of the peak observed at this value in the SNIFTIRS spectra discussed earlier for Ni electrodes electrically polarized in the presence of TeCN⁻ ions in DMF and DMSO-based electrolytes.

Table 2 FTIR data from IR studies of DMSO model solutions of Ni(NO₃)₂ or CuCl₂ with potassium cyanide salts prepared with different molar ratios

Model solution studied and molar ratio of Ni(NO3)2 or (CuCl2) : KCN prepared in DMSO	ν(CN) of free NCO^− ion or cyanate-containing complex ion species (cm^−1)	ν(CN) of Ni(II) and Cu(I)/CN^− complex ion species (cm^−1)	ν(CO) of CO2 dissolved in solvent (cm^−1)	Observed colour of model solution preparations
a [Ni(CN)4]^2− (ref. 19).b [Cu(NCO)2]^− (ref. 26).c [Cu(NCO)4]^2− (ref. 27).d [Cu(CN)2]^− (ref. 28), nd = not detected, s = strong, sh = shoulder.
DMSO
Ni(NO3)2/KCN 1 : 1		2112^a	nd	Colourless
Ni(NO3)2/KCN 1 : 2		2112	nd	Colourless
Ni(NO3)2/KCN 1 : 4		2112	nd	Colourless
CuCl2/KCN 1 : 1	2224^b (s), 2200^c (sh)	2127^d	2337	Gold yellow
CuCl2/KCN 1 : 2	2224 (s)	2127	2337	Gold yellow

---

In the case of the model solutions prepared by mixing TeCN⁻ or CN⁻ with Cu(II) salts, some fortuitous redox chemistry has occurred between (effectively) the CN⁻ ion and Cu(II) ions to produce the observed [Cu(I)(CN)₂]⁻ species (which also featured in the Cu/TeCN⁻ SNIFTIRS spectra).

In Fig. 5(b), the peak at 2127 cm⁻¹ arose from the destabilisation of TeCN⁻ caused by the presence of the polarising Cu²⁺ ions (from the Cu(II) salt), which led to the release of a free CN⁻ ion. This then reacted with Cu(II) ions in a fortuitous redox side-reaction²⁹ (known to occur between Cu(II) and pseudohalide ions) that led to Cu(I) species being present in the mixture, which then became complexed with CN⁻ ions to form the [Cu(CN)₂]⁻ species that gives rise to the peak at 2127 cm⁻¹.

In the IR spectra (see Fig. 6) of model solutions prepared by mixing Cu(II) salts with KCN salt in different CuCl₂ : KCN molar ratios in DMSO the fortuitous redox chemistry was also observed, which led to the appearance of the 2127 cm⁻¹ peak due to [Cu(CN)₂]⁻, as well as additional peaks at 2224 cm⁻¹ (shoulder at 2200 cm⁻¹) and at 2337 cm⁻¹. These have been assigned to [Cu(I)(NCO)₂]⁻,²⁶ [Cu(II)(NCO)₄]^2− 27 and CO₂ dissolved in the DMSO solvent.⁵ These species are all obviously due to the products of the redox reaction. The shoulder peak at 2200 cm⁻¹ is ostensibly due to Cu(II)–isocyanato complex ion species disappearing when more cyanide ions were supplied in the solution. This served to reduce more Cu(II) to Cu(I). These species were not relevant to the SNIFTIRS spectra, but merely observed as side products in the transmission IR spectra of these Cu-based model solutions.

Fig. 6 IR transmission spectra for the (a) 1 : 1 molar ratio model solution for CuCl₂ : KCN in DMSO prepared by mixing partially dehydrated CuCl₂ (0.025 mol L⁻¹) and KCN (0.025 mol L⁻¹) in DMSO (solution) and (b) 1 : 2 molar ratio model solution for CuCl₂ : KCN in DMSO.

Conclusions

In summary, IR spectroelectrochemical studies of the anodic dissolution of Ni, Cu and Au in DMSO and DMF solutions of the unstable TeCN⁻ ion have been presented for the first time. These studies have conclusively shown, via the observation of by-products formed at the electrode, that the TeCN⁻ ion decomposes in polarized electrochemical cells at certain potentials to form CN⁻ ion and a black substance, which would be elemental tellurium. The cause of the decomposition is due to the generation of dissolved metal ions from the electrode, which destabilise the TeCN⁻ ion due to their polarising power, with the destabilising effect increasing in the order Au⁺ > Cu⁺ > Ni²⁺. This is in agreement with the known trends with regard to polarising power for these cations.³⁰ The CN⁻ ion from the decomposition then goes on in these systems to become incorporated in metal complex ions generated from the electrochemically induced dissolution of the metal electrodes. This was shown through a comparison of observed IR data from SNIFTIRS experiments with transmission IR data of model solution experiments (in DMSO), in which Ni(II) and Cu(II) salts were codissolved in solution with TeCN⁻ and CN⁻. In general, the tendency of the TeCN⁻ ion to decompose in electrochemically polarized systems where metal ions are being formed is reflective of its well-known chemical instability in the presence of highly polarizing cations. The effect may be of use in developing sensors to detect when soluble ions from electrodissolved metals are generated at electrodes.

Acknowledgements

The authors are grateful to the Faculty of Science and Engineering at the University of Waikato for the provision of specialized chemicals to make the tellurocyanate salts.

References

1. E. Herrero, A. Rodes, J. M. Pérez, J. M. Feliu and A. Aldaz, J. Electroanal. Chem., 1996, 412, 165–174.
2. E. Herrero, M. J. Llorca, J. M. Feliu and A. Aldaz, J. Electroanal. Chem., 1995, 383, 145–154.
3. E. Herrero, M. J. Llorca, J. M. Feliu and A. Aldaz, J. Electroanal. Chem., 1995, 394, 161–167.
4. S. Ham, B. Choi, N. Myung, N. R. de Tacconi, C. R. Chenthamarakshan, K. Rajeshwar and Y. Son, J. Electroanal. Chem., 2007, 601, 77–82.
5. L. K. H. K. Alwis, M. R. Mucalo and B. Ingham, J. Electrochem. Soc., 2013, 160, H803–H812.
6. L. K. H. K. Alwis and M. Mucalo, J. Electrochem. Soc., 2014, 161, H738–H750.
7. K. H. K. L. Alwis, B. Ingham, M. R. Mucalo, P. Kappen and C. Glover, RSC Adv., 2015, 5, 15709–15718.
8. A. W. Downs, Chem. Commun., 1968, 1290–1291.
9. T. Austad and J. Songstad, Acta Chem. Scand., 1971, 25, 331–333.
10. O. H. Ellestad, P. Klaeboe and J. Songstad, Acta Chem. Scand., 1972, 26, 1724–1726.
11. H. K. Spencer, M. V. Lakshmikantham and M. P. Cava, J. Am. Chem. Soc., 1977, 99, 1470–1473.
12. A. Z. Al-Rubaie, in New Developments in Organometallic Chemistry Research, ed. M. A. Cato, Nova Science, New York, 2004, pp. 191–213.
13. P. Klaeboe, C. J. Nielsen and J. Songstad, Acta Chem. Scand., Ser. A, 1977, 31, 884–886.
14. G. Cauquis and G. Pierre, Bull. Soc. Chim. Fr., 1976, 736–738.
15. C. Korzeniewski, in Handbook of Vibrational Spectroscopy, ed. J. M. Chalmers and P. R. Griffiths, John Wiley, New York, 2002, p. 2699.
16. W. M. Keely and H. W. Maynor, J. Chem. Eng. Data, 1963, 8, 297–300.
17. J. Wang, J. Harris, M. Li, D. Plana, M. E. Barbour, D. J. Fermin and S. Mann, CrystEngComm, 2013, 15, 7152–7156.
18. A. Heather, in Nanostructured Ceramic Oxides for Supercapacitor Applications, CRC Press, 2014, pp. 71–118.
19. D. I. Semenishin, Ukr. Khim. Zh. (Russ. Ed.), 1998, 64, 85–87.
20. N. M. Levinson, S. D. Fried and S. G. Boxer, J. Phys. Chem. B, 2012, 116, 10470–10476.
21. M. R. Mucalo, R. P. Cooney and G. A. Wright, J. Chem. Soc., Faraday Trans., 1990, 86, 1083–1086.
22. W. Beck, Pure Appl. Chem., 1987, 59, 1703–1720.
23. K. A. B. Lee, K. Kunimatsu, J. G. Gordon, W. G. Golden and H. Seki, J. Electrochem. Soc., 1987, 134, 1376–1678.
24. K. Kunimatsu, H. Seki, W. G. Golden, J. G. Gordon and M. R. Philpott, Langmuir, 1988, 4, 337–341.
25. J. Emsley, The Elements, Clarendon press, Oxford, 2nd edn, 1991.
26. A. Ray, G. M. Rosair, R. Rajeev, R. B. Sunoj, E. Rentschler and S. Mitra, Dalton Trans., 2009, 9510–9519.
27. J. L. Burmeister and T. P. O’Sullivan, Inorg. Chim. Acta, 1969, 3, 479–486.
28. K. A. B. Lee, K. Kunimatsu, J. G. Gordon, W. G. Golden and H. Seki, J. Electrochem. Soc., 1987, 134, 1676–1678.
29. S. K. Tobia, E. R. Souaya and W. G. Hanna, Polyhedron, 1985, 4, 425–428.
30. S. Prakash, Modern Inorganic Chemistry, S Chand & Co. Ltd, India, May, 1987.

---