2, Joint Center for Artificial Photosynthesis and Molecular Foundry, Lawrence Berkeley National Laboratory, Berkeley, CA, Berkeley, California, United States
Improving the efficiency of solar-power oxygen evolution is both critical for development of solar fuels technologies and challenging due to the broad set of properties required of a solar fuels photoanode. Bismuth vanadate, in particular the monoclinic clinobisvanite phase, has received substantial attention and has exhibited the highest radiative efficiency among metal oxides with a band gap in the visible range. Efforts to further improve its photoelectrochemical performance have included alloying one or more metals onto the Bi and/or V sites, with progress on this frontier stymied by the difficulty in computational modelling of substitutional alloys and the high dimensionality of co-alloying composition spaces. Since substituional alloying simultaneously changes multiple materials properties, understanding the underlying cause for performance improvements is also challenging, motivating our application of combinatorial materials science techniques to map photoelectrochemical performance of 948 unique bismuth vanadate alloy compositions comprising 0 to 8% alloys of various p-block, alkalai earth, rare earth, and 4d & 5d elements (e.g Bi-V-A) along with a variety of compositions from each pairwise combination (e.g. Bi-V-A-B) of these elements. Upon identification of substantial improvements in the co-alloying Bi-V-A-B spaces, structural mapping was performed to reveal a remarkable correlation between performance and a lowered monoclinic distortion. First-principles density functional theory calculations indicate that the improvements are due to a lowered hole effective mass and hole polaron formation energy, and collectively, our results identify the monoclinic distortion as a critical parameter in the optimization and understanding of bismuth vanadate-based photoanodes.