Silicon carbide as a wide bandgap semiconductor material is by now widely introduced in the power device market. Its superior performance in terms of power density, high switching frequency and breakdown voltage leads to an advantage in efficiency and in the power to weight ratio of modules compared to silicon.
All the devices on the market until now rely on the use of hexagonal polytypes. These show the highest bandgap and can be produced in high quality for the high-volume market by physical vapor transport technology.
The only cubic polytype 3C cannot be used for high quality devices as the production is very difficult and the available material is still very defect rich. On the other hand this polytype has special properties concerning isotropic values (e.g. mobility) and the possibility to form a rather good interface to oxides as required for MOS devices.
Nowadays, there are two main technologies used for the production of cubic silicon carbide. Both rely on the heteroepitaxial growth of 3C thin films on either hexagonal silicon carbide or on silicon substrates. The growth on hexagonal substrates provides a better quality but silicon substrates are more cost effective and by that more often used.
To make the layers grown on silicon substrates by chemical vapor deposition more suitable for device applications a huge effort has to be raised to improve their crystalline quality regarding domain size and defect density as well as concerning stress in the layer.
Within this work several steps during substrate preparation, seeding and the beginning of epitaxial growth and their influence on layer quality had been under research:
In chapter 4.1 the substrate preparation was investigated. A particular interest was paid to the chemical cleaning and etching in hot hydrogen and their impact on the surface quality. Besides that, an influence of contamination inside the reactor was found and a new type of void defect was detected and described. In contrast to the typically applied carbonization step within the epitaxial process an ex-situ carbon deposition process was developed in chapter 4.2. It was combined with annealing steps in-situ or with the use of a rapid thermal annealing process.
In chapter 4.3 detailed studies had been performed on the influence of the ramp design during in-situ carbonization for seed generation. An explanation of the effects by a theoretical model was found and confirmed.
By utilization of high resolution TEM a focus was laid on the defect structures at and nearby to the silicon-carbide/silicon interface in chapter 4.4. The distribution of dislocations near the interface itself was studied.
By comparison of the growth on polished and wet chemically textured substrates the beginning of epitaxial growth was investigated in chapter 4.5. A model for an overgrowth mechanism is proposed that can be observed for growth on flat silicon substrates.
In chapter 5 as a supplement a possible application of CVD grown heteroepitaxial layers on silicon is shown. First promising experiments have been done showing the application as templates for the growth of thick freestanding cubic material by sublimation epitaxy.