Growth of III-nitrides on silicon
The use of silicon as a heteroepitaxial substrate for III-nitride growth has attracted industrial and academic researchers for decades because silicon is inherently cheap and can be scaled to reduce the production cost of devices. However, many material related challenges have thwarted the development of III-nitride on silicon epitaxy such as large lattice and thermal expansion mismatches. These challenges have largely been overcome during the last five to ten years and a number of companies are now offering high quality GaN/InGaN multiple quantum well structures on silicon for light emitting diodes.
Despite these successes, a complete understanding of the growth mechanisms that allow epitaxial engineers to grow III-nitrides on silicon has not emerged. In this project, we are learning to use “standard” techniques to grow III-nitrides on silicon. We use a combination of analytical tools such as X-ray diffraction, transmission electron microscopy and atomic force microscopy to understand the kinetic and thermodynamic factors of the growth and to investigate the formation of dislocations. We are also interested in the interplay between structural defects and electrical transport processes.
Selective etching of dislocation in III-nitrides
III-nitride films grown by MOCVD have a large density of dislocations which mostly terminate at the surface. Strain energy near these termination sites decreases the potential barrier for chemical reactions, making it possible to selectively etch near dislocation cores. This process has been widely used in materials science, and before the advent of advance microscopy, was one of the preferred methods to resolve dislocation densities and distributions. Over the last ten years, a number of groups (e.g. Weyher et al.) have done extensive work on selective etching in gallium nitride and etching parameters have been established which allow all three dislocation types (edge, screw and mixed) to be resolved.
The fundamental mechanisms of the dissolution at dislocation sites in gallium nitride is still not well understood. A number of classical theories have been cited to account for observations of the shapes of pits around dislocation sites, however there remain a number of discrepancies. For example, theory predicts that pits formed near mixed type dislocations should be the largest because they have the largest Burgers vector (and therefore local strain energy), however studies using weak beam transmission electron microscopy have shown that the largest pits are associated with screw type dislocations. We are trying to resolve some of these discrepancies by studying the shape evolution of pits using conventional wet etchants such as molten potassium hydroxide.
Phase separation in InAlN
InAlN is an interesting material for both optoelectronic and high power/high frequency applications. Similar to InGaN, the alloy phase separates into indium rich regions and aluminum rich regions when a large enough concentration of indium is incorporated into the lattice. Because the bond length of Al-N and Ga-N is nearly the same, one would expect the miscibility point of indium in InAlN to be comparable to that of InGaN, however evidence exists that indium can be incorporated into the lattice up to concentrations of about 30%. We are studying this system to better understand the mechanisms by which phase separation occurs and why InAlN behaves in such an unexpected way.