Optimising the Mobility of Nanoporous Titanium Dioxide

Titanium dioxide (TiO2) has drawn significant attention due to its low cost, abundance, and wide array of energy applications including as an electron transport layer in solar cells, as a photocatalyst for hydrogen production, and as an electrode material in rechargeable batteries. Of the many polymorphs of TiO2, anatase shows the most promise due to its high conductivity and excellent performance in devices. This high performance can, in part, be attributed to anatase not strongly trapping electrons in bulk-like regions of the crystal. However, a description of bulk properties is not sufficient because anatase is unstable as large, single crystals and so it usually produced as nanoparticles which are then sintered together. Polycrystalline materials have large surface area and a large number of grain boundaries where the particles come into contact with one another. These extended defects are expected to play a large role in the properties of anatase, but it is not straightforward to devise experiments that are able disentangle the individual roles of surfaces, grain boundaries, and point defects. This thesis presents first-principles models of extended defects in anatase and, where possible, makes comparison with experiment through the simulation of transmission electron microscopy. It is shown that anatase grain boundaries can introduce deep hole traps, but show no sign of electron trapping. Similarly, oxygen vacancies do not strongly trap electrons but the segregation of these ionised vacancies to the boundaries causes a space charge region to develop, which would pose a barrier to inter-grain electron transport. Models of highly-oxygen deficient nanoparticles indicate that electron trapping can occur on under-coordinated titanium sites and we propose that the oriented attachment of nanoparticles along highly oxygen-deficient facets would enable the formation of highly oxygen-deficient grain boundaries that would exhibit significant electron trapping.