One of the most critical bottlenecks to replacing fossil fuels and their accompanying carbon dioxide (CO2) emissions with renewable solar energy is to overcome the inherent intermittency of solar radiation. Due to their high energy density, chemical fuels such as hydrogen gas (H2) and hydrocarbons are among the most effective mediums for storing solar energy. Presently, however, solar-to-fuel efficiency and cost remain uncompetitive with conventional energy sources. This is partly because conversion efficiency is low, limited by absorption of light and the rate of chemical reactions. For state-of-the-art photoelectrochemical cells for water splitting, achieving a voltage sufficient to overcome the energetic barriers and slow reaction rates usually requires large band-gap absorbers or multi-junction cells, both decreasing efficiency and increasing cost.1
Most chemical reactions are thermally activated. Even a modest increase in operating temperature can produce a stark increase in reaction rate. Employing heat from solar radiation to generate solar fuels enables use of photons with energy below and above the absorber band gap, which would otherwise be wasted. Figure 1 shows a solar-driven, solid-state metal oxide thermochemical cycle, a promising approach to water splitting that exploits this concept.2 In this process, a metal oxide is stripped of its oxygen (in other words, gaining electrons) using thermal energy, and is subsequently re-oxidized (losing electrons) with water and/or CO2 at a lower temperature to produce fuel. Since oxygen and fuel are generated in separate steps, there is no possibility of recombination, which is potentially hazardous in elevated-temperature photocatalysis.3 These thermochemical cycles typically operate at temperatures exceeding 800°C, with the solar radiation supplied through optical concentrators. As such, this approach is ideal for the centralized production of solar fuels in geographic regions with excellent direct sunlight, for example, the southwestern United States and the Sahara.
The heart of the solid-state metal oxide thermochemical cycle is an oxygen storage material (OSM). To achieve high efficiency and cycle life, the OSM must have sufficient redox energetics to split water molecules (and/or CO2), exhibit high oxygen diffusivity and surface reactivity, and remain morphologically stable over a large number of cycles.4 Top materials currently under consideration include various ferrite-based oxides, which cycle between iron (II) wüstite and iron (III) magnetite crystal structures (II and III denote the oxidation state of iron).2 While excellent theoretical solar-to-fuel efficiencies have been predicted in cycles based on these OSMs, slow oxygen diffusion and large structural reorganization during oxidized-to-reduced-state phase transformation have both limited the fuel generation rate and impeded cycle reversibility.
Recently, we and other groups reported that using a nonstoichiometric oxide—i.e., with a non-integer oxygen content—as an OSM for thermochemical water splitting yields promising results.4, 6,7 Rather than subjecting the OSM to a phase transition during the thermochemical cycle, the material is cycled entirely within a single crystallographic phase. As a result, we can eliminate the slow reaction rate and tendency to morphological change associated with phase transformations. Early experiments on cerium-based oxides (CeO2), which cycle between ∼CeO2 (800°C) and CeO1.95 (1500°C), indicated a promising water-splitting rate. We attributed this to the high mobility of oxygen vacancies in the fluorite crystal structure of CeO2.4, 6,7 Figure 2 shows typical temperature and gas evolution profiles. From thermodynamic analysis, we predicted that a high solar-to-fuel efficiency of 16–19% can be attained, despite the relatively small change in oxygen content during the cycle when compared to the phase-transforming, stoichiometric (integer number) oxides.5
We next examined the suitability of CeO2 for solar fuel production using a 2kW solar reactor, constructed with ∼325g of active material at the High-Flux Solar Simulator at the Paul Scherrer Institute in Switzerland.8 We observed a remarkable rate of H2 production: ∼310mLmin−1. The setup also indicated CO2 dissociation and reduction to carbon monoxide (CO) at an even greater rate: ∼590mLmin−1. Simultaneous dissociation of water and CO2 into synthesis gas—a mixture of H2 and CO—at various ratios of composition was also demonstrated recently by Furler et al.,9 paving the way for the subsequent conversion to liquid fuels via the Fischer-Tropsch reaction. Beyond generating H2 and CO, the CeO2-based thermochemical cycle can also be used to dissociate water and CO2 to methane, a high-energy-density hydrocarbon.4 In that particular reaction, a nickel catalyst is used to facilitate the complete reduction of CO2 to carbonaceous species, which are crucial reaction intermediates in the co-reduction of CO2 and water to methane.
In summary, operating at elevated temperatures with nonstoichiometric OSMs, we have achieved rapid solar fuel production kinetics using CeO2-based thermochemical cycles. Our current efforts are directed toward reducing the operating temperature and eliminating the use of exotic materials in the solar reactor. Lowering the operating temperature requires careful tuning of the redox thermodynamics of the OSM, namely, oxidation enthalpy and entropy.5 Beyond fluorite structures, OSM based on the chemically versatile perovskite-oxide structure is also being investigated. Further strategies to lower the temperature of elevated-temperature solar fuel processes may ultimately include coupling both thermal- and photo-driven processes, combining thermochemistry and photochemistry.