Development of electrocatalysts for the oxygen evolution reaction in alkaline water electrolysis.
Thesis DisciplineChemical Engineering
Degree GrantorUniversity of Canterbury
Degree NameDoctor of Philosophy
With the ever depleting oil resources around the world, an alternative transportable and storable fuel is required for the future. Water electrolysis is one method of producing hydrogen (an alternative fuel) that does not require oil and gas. It has the advantage of producing extremely pure hydrogen without directly harming the environment through carbon emissions. This work focuses on improving the efficiency of the water electrolysis process through reducing the oxygen evolution overpotential by utilising electrocatalysis. The electrocatalysts need to exhibit good activity, good electrochemical stability and be economic to produce. The long-term goal is to utilise an electrocatalyst in a zero-gap, alkaline electrolyser developed by Callaghan Innovation. Based on the literature review, the following electrocatalysts were chosen for further investigation: LaₓSr₁₋ₓCoO₃, CoFe₂O₄, Co₃O₄, NiCo₂O₄, IrO₂ and NiOₓ. Due to differences in preparation procedures and testing conditions, it can be difficult to directly compare performance results from literature. The initial work enables the comparison of these catalysts through using a standard preparation procedure (thermal decomposition) and an identical testing procedure. The majority of the electrochemical studies in this research were performed using a 30 wt.% KOH electrolyte due to its high conductivity, with potentials referenced against Hg/HgO (30 wt.% KOH). The electrocatalyst with the lowest overpotential, and therefore best performance in this study is IrO₂, however due to its cost, Co₃O₄ is also considered as a promising option as it has the second best performance while only having a mid-range surface area when compared to the other electrocatalysts. This indicates high intrinsic activity. Another observation from the preliminary work is that the nickel substrate material itself has better electocatalytic performance than some of the other electrocatalysts tested. Thus, it was decided that methods of obtaining and maintaining good performance with an uncoated nickel electrode be investigated. To gain a better understanding of the nickel hydroxides which catalyse the oxygen evolution reaction (OER), the complex phase changes which occur on the electrode surface were investigated. It was found that a process in addition to the standard α-Ni(OH)₂/ γ-NiOOH and β-Ni(OH)₂/ β-NiOOH reactions occur in the more concentrated KOH electrolyte. It is also confirmed that the initial hydroxide layer formed anodically from metallic nickel is not α-Ni(OH)₂, but rather a layer which is more readily reducible. While in situ XAS suggested that γ-NiOOH is not transformed to any further phase up to 0.665 V vs Hg/HgO in 1 M KOH, at higher potentials, after extensive OER (at least 40 hrs) at 50 mA cm⁻² and in 30 wt.% KOH, an additional phase can be identified by cyclic voltammetry. During galvanostatic oxygen evolution, the nickel anodes follow an ageing behaviour characterised by a brief activation period, a short period of high activity (i.e., low overpotential) followed by deactivation and eventually stable but poor activity. After studying the structural changes which occur on the surface of the nickel electrode, as well as the ageing behaviour, methods of rejuvenation of nickel anodes are investigated. The deactivation caused by ageing can be mitigated by temporarily reducing the potential for brief periods. Continuous rejuvenation of nickel anodes is investigated and it is shown that rejuvenation at 0.5 V vs Hg/HgO for 10 min out of every 100 min can prevent ageing of the anode, thus maintaining a low overpotential during galvanostatic oxygen evolution at 50 mA cm⁻². It is suggested that the short potentiostatic rejuvenation periods at regular intervals prevents the ratio of Ni(IV) to Ni(III) from increasing, thereby maintaining the intrinsic activity of the material. Additionally it is found that the rejuvenation potential must be above 0.36 V vs Hg/HgO to ensure the rejuvenation is successful in improving performance, i.e., the material must not reduce to Ni(II). It is estimated that by using rejuvenation steps, an energy saving of 8% is possible in an alkaline water electrolyser using nickel anodes. While the performance decrease of the nickel anodes over time can be reduced by rejuvenation, addition of electrocatalytic coatings are still recommended where the goal is to further improve performance and stability without the need for rejuvenation steps. As cobalt oxide was identified as one of the most promising electrocatalysts earlier, its catalytic performance is further optimised by investigating deposition procedure options. Layers produced by both thermal decomposition and electrochemical deposition had similar electrochemical behaviour, provided that the layers were annealed at temperatures ≥ 350°C. This thermal treatment was required to mechanically stabilise the electrochemically deposited cobalt oxide layer. Due to this finding, the effect of annealing temperature was investigated for the electrochemically deposited layer, and it was found that the overpotential for oxygen evolution increased with annealing temperature. Using cyclic voltammetry and impedance spectroscopy, it is concluded that the decrease in performance with increasing annealing temperature is largely caused by the corresponding decrease in active surface area. However, for annealing temperatures ≥ 400°C, additional resistances are introduced which cause lowered performance. As iridium oxide is identified as another promising but costly electrocatalyst, deposition of thin iridium layers are investigated. Spontaneous deposition onto nickel substrates was investigated as a method of producing these thin electrocatalytic iridium layers. UV/Vis spectroscopy, cyclic voltammetry and other electrochemical methods are used to investigate the deposition process and the activity of the electrocatalytic coating towards the OER. From three solutions (IrCl₃ + HCl, H₂IrCl₆ + HCl, and H₂IrCl₆), H₂IrCl₆ is shown to give the most active and stable coating, with deposition times of 45 min at 60°C being enough to increase the activity of the nickel substrate for the OER. While the iridium oxide coatings are highly active at low current densities (less than 45 mA cm⁻², coatings with higher surface area would be needed to maintain a high performance at higher current densities. Finally, some preliminary work was carried out in order to scale up the cobalt oxide coating procedure to a 20 cm x 25 cm nickel foam electrode. Firstly it is shown that electrodeposition can be used to coat a 0.5 cm x 1 cm nickel foam electrode (most previous work had been done on nickel foil) and achieve an overpotential decrease of 150 mV at 50 mA cm⁻² relative to the uncoated nickel foam. Next a 20 cm x 25 cm cobalt oxide coated electrode was successfully produced via electrodeposition. Testing the performance in one cell of the zero-gap Callaghan Innovation electrolyser shows no increase in performance over the uncoated electrode. However, by testing a cobalt oxide coated electrode in a smaller scale zero-gap cell (nickel foam electrodes 5 cm x 4 cm in size) it is shown that a performance improvement can indeed be achieved in a zero-gap configuration. A number of reasons for the lower than expected performance at full-scale (20 cm x 25 cm) are discussed and recommendations for future work are made.