Font Size: a A A

Titanium dioxide thin films: Understanding nanoscale oxide heteroepitaxy for silicon-based applications

Posted on:2006-09-07Degree:Ph.DType:Thesis
University:University of WashingtonCandidate:Schmidt, Diedrich AFull Text:PDF
GTID:2451390005494586Subject:Physics
Abstract/Summary:
The work presented here discusses nucleation and growth of oxides at the nanoscale level. The overall perspective in comprehension of nanoscale oxide nucleation and growth is essential for utilizing oxide heteroepitaxy in silicon-based applications.; To separate the effects of chemistry and structure, this thesis investigates the initial stages of TiO2 nucleation and growth on a chemically compatible and lattice matched substrate, LaAlO3(001) (LAO) and on the technically relevant, though chemically very different, Si(001) surface. Although LAO(001) exhibits mixed La-O and Al-O2 surface terminations at 400°C, heteroepitaxial TiO2 grows out from its step-edge nucleation sites to cover both upper and lower terraces uniformly, regardless of termination. This growth behavior indicates that the substrate cations and the associated surface charges have minimal influence on the morphology in this single-metal oxide heteroepitaxy.; We then extend this work by studying the initial nucleation and growth of TiO2 on Si(001) using a unique buffer layer. The buffer layer is an integral component of successful oxide-on-silicon heteroepitaxy as it prevents silicide or SiOx formation, which would occur for direct TiO2 growth on silicon. We confirmed both monolayer as termination and an additional chalcogonide buffer layer (Ga2Se 3) are inert to molecular oxygen at elevated temperatures, however, only the Ga2Se3 film on Si(001):As is successful in preventing substrate-oxide or -silicide reactions from occuring. TiO 2 nucleates along the sides or "valleys" of the Ga 2Se3 nanowires. We use a concept from the oxide-on-oxide work to view the Se and Ga constituents as anion and cation, respectively. We then explore the effect of substrate temperature on TiO2 film morphology. The films are essentially unchanged from room temperature up to Tsub = 350°C, except for an increase in the average spacing between initial TiO2 nuclei with temperature. However, they undergo a morphological change somewhere between Tsub = 350°C and Tsub = 540°C. This morphological change is attributed to a phase change in the crystal, from the lattice matched anatase phase (at Tsub ≤ 350°C) to the thermodynamically stable rutile phase (at Tsub = 540°C). Growth at temperatures above 600°C led to desorption of the Ga2Se3 and oxidation of the substrate.; A significant proposed application of anatase on Si(001) is as a ferromagnetic component in spintronic devices, for which is must be doped with a magnetic ion. We investigated growth of Co0.05Ti0.95O2 films, demonstrating that Co-doped oxide films nucleate and grow in the same manner as pure TiO2 films. The doped films coalesce to a flat film within the first two molecular layers and continue to grow in a laminar fashion. Films over 20 molecular layers thick have an average surface roughness over 1 mum2 which is only ⅓ of a molecular layer larger than the starting substrate. No Co-rich surface clusters were observed, unlike other preparation methods. This surprising result is encouraging for future work in oxide on silicon heteroepitaxy. Measurements of the valence band show the oxide films have a small conduction band offset with the Si substrate, indicating the oxide-buffer-Si system is naturally well suited for spintronic applications.
Keywords/Search Tags:Oxide, Films, Nanoscale, Growth, Substrate, Work
Related items