Experimental and ab initio studies of oxide interfaces for photoelectrochemistry

Abstract

The threats posed by anthropogenic climate change have spurred a research thrust towards renewable, carbon-free sources of energy. Photoelectrochemical (PEC) approaches are particularly attractive, combining energy capture from the sun with storage in the form of hydrogen or hydrocarbon fuel. However, there are significant materials challenges to be overcome, as well as a necessity for improved understanding of the material interfaces present in such systems. Transition metal oxides are a popular material for research as photo-electrodes but typically have poor electronic properties compared to conventional semiconductors. However, they are stable in aqueous and oxidizing environments and may present a wide variety of exotic physical behaviors, potentially opening new doors for device design. In this thesis, I explore several aspects of oxide interfaces relevant to PEC devices.

PEC measurements of ultra-thin films of LaFeO 3 grown on Nb:SrTiO3 reveal a thickness-dependent response via the depletion regions that form at both the film-substrate and film-electrolyte interfaces. Depending on the applied bias, reduction or oxidation photocurrent is observed that originates from the film-electrolyte or film-substrate interface, respectively. These qualitative behaviors are then explained with a band model. I then use the ACBNO functional for self-consistent Hubbard U corrections to density functional theory (DFT). First, improvement in treating bulk perovskite oxide electronic structure is demonstrated, followed by a study on a series of thin film slab structures that captures nanoscale changes in formal charge and hybridization (via the change in U) at multiple locations within the film, simultaneously. The trends in oxygen adsorption energy and band alignment are explained in terms of film thickness and electronic structure.

Finally, a first-principles descriptor for oxygen adsorption energy is developed from high-throughput DFT calculations and analysis of the density of states using tight binding and the moments theorem. This descriptor methodology may be used in high-throughput screening for catalyst materials, where bulk calculations may be used to predict surface properties without resorting to more demanding slab calculations. The combination of high-throughput screening of materials with the engineering possibilities afforded by substrate and active layer thickness variation provides a promising path forward to successful oxide photoelectrochemical devices.