Catalysis to produce solar fuels: From the production of hydrogen via water splitting, to hydrogen conversion to methanol by its reaction with CO2

Increasing CO2-anthropogenic concentration in the atmosphere is surpassing sustainable levels and unambiguously jeopardising the global climate. The main share of CO2 emissions corresponds to the production of energy, currently 17 TW per year. By analysing historical trends on population growth and energy consumption, it is expected that the global energy demand will reach 30 TW by 2050, which in turn will aggravate the stress on the environment. In order to alleviate adverse climate-change-related consequences, it was agreed to reduce global CO2 emissions. This could be achieved by shifting the current energy system towards a carbon-free energy vector, hydrogen. Hydrogen can be obtained from the hydrolysis of water using surplus electricity generated from renewables, without contributing to CO2 emissions. The main drawback to switching from a carbon to a hydrogen-based energy system is that H2 is a gas and current technology has evolved around liquid fuels. In order to circumvent this energy transition, hydrogen can be further transformed to liquid fuels via its reaction with CO2. In a polymer electrolyte membrane (PEM) electrolyser, water is oxidised at the anode to oxygen and protons, protons migrate through the membrane to the cathode where they recombine with electrons to form hydrogen. Due to the slow kinetics and multiple reaction steps on the anode compared to the cathode, the oxidation of water to oxygen (or oxygen evolution reaction, OER) is responsible for high overpotentials. Additionally, the anode needs to be made of iridium based catalysts, and because of its low natural abundance, it is essential to develop materials with an efficient Ir usage and to optimise its catalytic activity and stability. This research is divided in two defined themes. The first (chapter 3 to chapter 6) focuses on the optimisation of IrO2 catalysts for the oxygen evolution reaction, necessary half reaction in a PEM water electrolyser for the production of hydrogen. The second (chapter 7) comprises of the optimisation of PdZn catalysts supported on TiO2 for the further transformation of H2, by its reaction with CO2, to solar fuels. In chapters 3 and 4, the effect of the base on the hydrothermal synthesis of unsupported and supported amorphous iridium oxo-hydroxides is studied. The hydrothermal synthesis was chosen because it allowed the synthesis of amorphous IrOx materials without the need for heat treatment at high temperature, thus minimising the possible crystallisation and the concomitant decrease in activity towards OER. It was observed that the base plays an important role in tailoring the morphology, surface area and surface hydroxide concentration of IrOx catalysts, and thus it has a direct effect on the catalytic activity and stability. Specifically, the use of Li2CO3 as a base led to a catalyst with porous morphology, higher surface area and higher hydroxide concentration, which this translated to an improved activity and stability towards OER compared to the state of the art catalyst IrO2·2H2O (Premion, Alfa Aesar). In both chapters, heat treatment was proven to hinder the catalytic activity towards OER, presumably as a result of higher crystallinity, the loss of Ir(III) sites and the decrease in hydroxide concentration. In chapter 5, two different IrO2 crystalline structures (rutile and hollandite) were synthesised, characterised and compared as OER catalysts. In accordance with the literature, the transformation of amorphous iridium oxo-hydroxide, containing Ir(III)/Ir(IV) sites, to crystalline rutile IrO2, made only of Ir(IV) sites, led to a decrease in catalytic activity and stability. However, the presence of Li2CO3 in the amorphous IrOx catalyst led to the formation of hollandite IrO2 instead of rutile IrO2, with lithium as the host cation within the hollandite channels. Apart from the difference in crystallinity, characterisation on hollandite IrO2 indicates that it was closer in nature to amorphous IrOx than to rutile IrO2. The presence of Ir(III) and Ir(IV) was confirmed by XPS, shorter Ir-Ir bond distances and longer Ir-O, compared to rutile IrO2, were observed by EXAFS, and comparable OER activity to IrO2-Li2CO3 was detected by LSV. Additionally, the conversion of amorphous IrO2-Li2CO3 to hollandite IrO2 led to improved stability under OER reaction conditions. In order to use iridium more efficiently and to reduce the iridium loading on the electrode, in chapter 6 IrO2 was diluted with a more abundant and economic metal oxide, nickel or copper oxide. Catalysts with a homogeneous metal distribution and with a core-shell distribution, concentrating iridium at the surface and the non-noble metal oxide at the core, were prepared following a modification of the hydrothermal synthesis method. The synthesis of mixed oxide catalysts with a homogeneous metal distribution led to a decrease in the catalytic activity and the stability of the catalyst, which was proven to be an ineffective synthetic route for considerably decreasing the iridium loading on the electrode. The observed decline in the catalytic performance was attributed to the dissolution of the non-noble metal oxide in contact with the reaction media. However, through a core-shell distribution, IrOx was concentrated on the surface of the catalyst, whilst the non-noble metal remained protected against dissolution inside the nanoparticle core. Following the core-shell synthetic approach, the iridium loading on the electrode was successfully halved without impairing the catalytic activity or stability, compared to pure IrO2-Li2CO3. The second part discussed in chapter 7, studied the optimisation of PdZn/TiO2 catalysts prepared by chemical vapour impregnation (CVI) for the CO2 hydrogenation (pre-reduction at 400 °C, 1 h, reaction at 250 °C, 20 bar, 30 ml·min-1, 60 % H2, 20 % CO2, 20 % N2) to methanol, as a stable alternative to copper catalysts. The first section of the chapter focused on the Pd to Zn molar ratio in the material, maintaining the palladium loading at 5 wt. %. Increasing the Pd:Zn molar ratio from (1:1) to (1:5) led to a greater formation of PdZn alloy, which improved CO2 conversion, but without considerably affecting methanol selectivity. The further addition of zinc, as observed for the catalyst with a Pd:Zn molar ratio of (1:10), led to a decrease in the CO2 conversion. This was presumably caused by zinc blocking active sites when in large excess. The atomic proportion of zinc in the PdZn alloy can vary from 40 at. % to 60 at. %. Hence it could be hypothesised that increasing the pre-reduction temperature could lead to a higher proportion of zinc within the alloy, which in turn can improve methanol selectivity. In general, increasing the pre-reduction temperature from 400 °C to 650 °C led to an increase in the methanol productivity because of improved methanol selectivity, although lower CO2 conversion was observed as a result of particle sintering. However, more interestingly, the CH4 selectivity decreased by one order of magnitude after increasing the pre-reduction treatment from 400 °C to 650 °C, simultaneously with the transformation of ZnO and TiO2 to rhombohedral ZnTiO3. This lead to the hypothesis that the PdZn-TiO2 interphase is responsible for methane production. To prove this hypothesis, PdZn/ZnTiO3 and Pd/ZnTiO3 catalysts were prepared by CVI, after pre-reduction at 400 °C. PdZn alloy formation was confirmed by XRD on both systems, indicating that Zn in the ZnTiO3 phase can migrate out of the lattice to form PdZn. Thus, the PdZn-TiO2 interface was generated in Pd/ZnTiO3 but not in PdZn/ZnTiO3. When tested for CO2 hydrogenation to methanol, the formation of methane on the former catalyst and its absence on the latter corroborated the formulated hypothesis that PdZn-TiO2 acts as the active site for CH4 formation