The electrochemistry and performance of cobalt-based redox couples for thermoelectrochemical cells

Abstract

Thermoelectrochemical cells are a promising technology for sustainably generating electricity from waste heat. These electrochemical devices directly convert heat into electricity, with a performance governed by the properties of the redox couple, electrolyte and electrode. In this work the influence of the nature of the redox couple on fundamental properties such as the Seebeck coefficient, diffusion coefficient and charge transfer resistance was investigated. Four different cobalt complexes containing the ligands 2-(1H-pyrazol-1-yl)pyridine) (Co^2+/3+(py-pz)₃), 2-(1H-pyrazol-1-yl)-4-tert-butylpyridine (Co^2+/3+(bupy-pz)₃), 2,6-di(1H-pyrazol-1-yl)pyridine (Co^2+/3+(pz-py-pz)₂) and 1,10-phenanthroline (Co^2+/3+(phen)₃) were examined in a 3:1 dimethyl sulfoxide: 1-ethyl-3-methylimidazolium tris(pentafluoroethyl) trifluorophosphate mixture. The performance of each redox couple was governed by the ligand properties. The highest Seebeck coefficient was measured for Co^2+/3+(py-pz)₃ (2.36 mV K⁻¹) which was attributed to a combination of its small radius, bi-denticity and lower degree of aromaticity. This is higher than the previously reported Co(bpy)₃ couple. The highest power output was achieved with the Co^2+/3+(py-pz)₃ redox electrolyte, using platinum electrodes coated with a carbon layer, which gave 36 mW m⁻² from a ΔT of 30 ^○C. The power outputs achieved using the different redox couples was highest for those with a high Seebeck coefficient, good electrochemical reversibility and fast ion diffusion. The electrochemical reversibility depends significantly on the nature of the electrode substrate and it is demonstrated that carbon-coated platinum electrodes can be used to improve the electrochemical reversibility of these redox couples.

Introduction

The thermoelectrochemical cell (thermocell) is a device capable of converting waste heat to electricity with no carbon emissions [1]. Sources of sustainable waste heat include geothermal and industrial sources. Interest in the field of thermal energy harvesting has been increasing in recent years, driven by the urgent need to generate more energy from sustainable sources. Semiconductor-based thermoelectric devices are limited to high temperature operation (>200 °C) due to their decreased efficiency at lower temperatures [[2], [3], [4]]. This leaves the lower temperature range of waste heat (<200 °C) under-utilised, presenting an opportunity for the application of liquid-based thermoelectric devices. The increasing research into liquid-based thermocells is also driven by the higher open circuit voltages achievable for these systems, which is in the mV K⁻¹ range as opposed to the μV K⁻¹ range for a typical solid-state thermoelectric devices [2,[5], [6], [7], [8]].

The thermocell is a two electrode system that uses an electrolyte containing a redox couple. When a temperature gradient is applied to the cell, the temperature dependence of the redox reaction generates a potential difference. The cell potential achievable for a given temperature gradient is represented by the Seebeck coefficient (S_e) (Eqn (1)) [7]:

$$S_e=\frac{\Delta V}{\Delta T}=\frac{\Delta S_{rc}}{nF}$$

where, $\Delta V$ is the potential difference between the hot and cold electrodes, $\Delta T$ is the temperature gradient between the electrodes, $\Delta S_{rc}$ is the entropy change associated with the redox reaction, $n$ is the number of electrons involved in the reaction and $F$ is Faraday’s constant [7].

The entropy change of a redox reaction is influenced by factors such as the structural changes that take place in the redox species and the extent of rearrangement of the solvation shell. Until recently, most research into thermocells focussed on aqueous systems, particularly with the ferri/ferrocyanide redox couple that has a Seebeck coefficient of −1.4 mV K⁻¹ [1]. A number of other redox couples have been examined in both aqueous and non-aqueous media including iodide/tri-iodide (I⁻/I₃⁻) [[9], [10], [11]], Cu/Cu²⁺ [12,13], and mixed redox couple systems, such as ferrocene/ferrocenium (Fc/Fc⁺) and I⁻/I₃⁻ [14]. Recently, we reported the Seebeck coefficient of Co^2+/3+(bpy)₃ in 3-methoxypropionitrile, at 2.19 mV K⁻¹ [8]. The high Seebeck coefficient of this redox couple is attributed to the high-to-low electronic spin state transition for the Co^2+/3+ redox couple, which contributes an additional electronic component to the total entropy change [[15], [16], [17]]. There have also been reports in the literature of Seebeck coefficients as high as 7 mV K⁻¹ [18,19] in non-redox active electrolytes. In non-redox active electrolytes, the potential difference occurs via the formation of a concentration gradient due to the different mobility of ions at different temperatures, rather than the entropy change due to a redox reaction. As such, it is important to make the distinction between redox active and non-redox active electrolytes as they have different mechanisms and applications.

At present, the widespread application of thermocells is limited by their relatively low power and conversion efficiency. While recent research into device design and optimisation has led to some significant improvements in cell potential and power outputs [[20], [21], [22]], there is still significant scope for increased understanding and optimisation of the fundamental properties of thermocells and their materials. The investigation of new redox couples with high Seebeck coefficients is a particularly important avenue for improving the performance. Redox couples for thermocell applications should meet three main criteria: high Seebeck coefficients, electrochemical reversibility and electrochemical stability. Cobalt-based redox couples are a promising class of redox couples for this application as they can have higher Seebeck coefficients than other analogous transition metal complexes [16,23,24]. Cobalt complexes with pyridine-pyrazole chelating ligands have been investigated for use in dye-sensitised solar cells, demonstrating good electrochemical stability and reversibility [[24], [25], [26], [27]]. In this work, four cobalt-complexes with different chelating ligands have been investigated as redox couples for thermocells for the first time. The performance of these complexes is compared to that of Co^2+/3+(bpy)₃, which is the highest performing cobalt complex, tested in the same solvent systems. Thus, the effects of the different ligands on the fundamental properties of the redox couple, such as the Seebeck coefficient, and the corresponding effects of these properties on overall thermocell performance are discussed.