| The types, concentrations and spatial distributions of solid pO int defects such as vacancies and interstitial atoms influence the performance of metal oxides in gas sensing, electronics, photonics and solar-induced photochemistry for fuel production and environmental cleanup. Controlling these aspects of defect behavior has become increasingly important in nanoscale applications, where surface-to-volume ratios are large. Various material processing protocols to exert such control have arisen, known collectively as "defect engineering". Surfaces of themselves have shown promise as tools for such efforts. Surfaces offer efficient pathways for defect creation because fewer bonds need to be broken than for bulk pathways. Surface can, in addition, support charge which create near-surface strong electric fields that can lead to the redistribution of defects near surfaces and interfaces. A detailed understanding of the underlying surface-mediated mechanisms would enable us to engineer oxygen and cationic pOint defects in metal oxides. The present work involves both experimental (isotope gas-solid exchange) and computational (ab initio quantum calculations, continuum models) efforts aimed at understanding the (1) injection of oxygen interstitials via an active site exchange mechanism in rutile titanium dioxide (TiO2) and wurtzite zinc oxide (ZnO), and (2) electric-field induced redistribution of charged defects in near-surface space charge regions in TiO2, ZnO and B-doped Si.;Regarding active-site exchange, this work shows that the injection of oxygen interstitials (Oi) in TiO2 under O-rich conditions lead to the Oi becoming the majority oxygen defects. This finding overturns a long-held belief that oxygen vacancies (VO) are the dominant O-related native defects, and resolves a major discrepancy in the literature between a large body of experimental work and predictions from quantum calculations. From a practical perspective, the work shows how surface conditioning permits defect engineering of TiO2 for electronic and catalytic applications. The activation energy for Oi injection from TiO2 (110) is determined to be approximately 1.8 eV -- the first time an injection barrier has been measured for any semiconductor system. Controlled sulfur adsorption decreases the mean diffusion length of Oi by almost an order of magnitude via an indirect mechanism of inhibiting the surface annihilation of Ti interstitials at the surface, which enhances the Ti-mediated exchange of Oi with the lattice. Another distinct mode of surface-based defect engineering is uncovered involving the control of surface polarity. Polar Zn-terminated c-axis ZnO(0001) surface inject Oi similarly to TiO2(110), but corresponding O-terminated surfaces greatly inhibit the injection except at widely-spaced special sites comprising ~1% of the surface, where the injection barrier is near zero. These experimental results of polarity-influenced injection accord closely with the predictions of ab initio density functional theory (DFT) calculations of O2 adsorption on c-axis ZnO surfaces.;Regarding defect redistribution by electric fields, a combination of a mathematical model and isotope exchange experiments monitoring near-surface pileup have identified a new mechanism by which electric fields near the surface directly influence the drift of injected charged defects. The effects operate in both in rutile TiO2 and ZnO, and require band bending of only a few meV to induce fields that oppose the diffusional flux. Drift working in the same direction as the diffusional flux should induce a valley in the near-surface defect concentration rather than pile-up. This electrostatic coupling almost certainly influences the spatial distribution of defects and dopants near surfaces and interfaces in nanoscale devices for gas sensing and microelectronics. |
