Growth, characterization and processing of III-nitride semiconductors
The III-nitride semiconductor materials have wide direct bandgap ranging from 6.2 eV (AlN); through 3.4 eV (GaN) to 0.65 eV (InN). Alloys have been realized in the entire composition range. Thus, within one material system the possibility exists to create devices, emitting light from the infrared to the ultraviolet part of the spectrum. Compared to other semiconductors, the materials also possess several other attractive characteristics for devices, such as good thermal conductivity, high breakdown field and resistance to both high temperatures and chemically hostile environments. Despite the rapid commercial progress of the III-nitride technology, there is room for much material improvement. The most important drawbacks are related to the substrate. For epitaxial growth, there is no commonly available GaN substrate, which is matching in lattice constant and thermal expansion coefficient. Therefore, epitaxial material exhibits high threading dislocation density, typically 10^(8)-10^(10) cm^(-2). Due to the different nature of the three binaries, growth of alloys and their impurity incorporation are not fully understood. Further, the chemical stability and wide bandgap introduces difficulties in device processing.
This work concerns issues for the III-nitrides that arise in material growth and device fabrication. Initial growth on a sapphire substrate, growth parameter settings , material characterization, and, finally, device processing were studied. The material growth was, with a few exceptions, made by molecular beam epitaxy. The MBE technique has several strengths such as high purity, abrupt heterostructure interfaces, precise thickness control and the possibility for \insitu~monitoring of the surface. In particular, the initial atomic layer-by-layer growth of GaN and AlN on sapphire substrates were investigated. It was demonstrated that the Al/N flux ratio during nucleation layer growth had great influence on material properties of the layer on top. For very N-rich nucleation layer growth (Al/N < 0.4), the material quality was poor. In contrast, material with high crystalline quality and low surface roughness was obtained using an intermediate growth mode (0.4 < Al/N < 0.7) for the nucleation layer. The standard Al/N = 1 nucleation layer was shown to create micropipes in GaN and AlN overlayers. It was demonstrated that annealed substrates are more forgiving for improper nucleation layer growth parameters. Material grown on annealed sapphire exhibited much improved structural quality compared to using non-annealed substrates. Further, we have shown that impurity incorporation in GaN depends on growth parameters. The influence of the dislocation density on photoluminescence intensity of GaN was measured and a model was developed, which accounted for the reduction in intensity. In order to understand InGaN alloy growth, the adsorption of Ga and In on the GaN(0001) surface was studied. Both Ga and In bilayer formation was demonstrated. The bilayer thickness could be estimated to ~2.2 ML using reflection high-energy electron diffraction. Multilayer structures were investigated by X-ray diffraction and infrared absorption spectroscopy. A 20 period AlN/GaN multiple quantum well was required to observe intersubband transitions. Finally, device processing of III-nitride materials has been investigated. The knowledge gained from these investigations was used to process small mesa structures and a blue InGaN light-emitting diode.
10.15 sal A423 (Kollektorn), MC2-huset, Kemivägen 9, Chalmers
Opponent: professor Yoon Soon Fatt, Nanyang Technological University, Singapore.