Crystalline Fe films grown on non-polar GaN: theory and experiment

We report an unexpected mechanism by which single crystals of Fe grow epitaxially on $M$-plane GaN substrates despite having a different crystal structure and strongly mismatched lattice constant. A simple model is proposed in which one material tilts out of the interface plane to create a coincidence-site lattice that balances two competing geometrical criteria---low residual strain and short coincidence-lattice period. We apply this model, along with complementary first-principles total energy calculations, to the interface formed by molecular-beam epitaxy of cubic Fe on hexagonal GaN and find excellent agreement between theory and experiment. The success of this model also suggests a strategy for growing non-polar GaN films on a substrate material with a suitably chosen Miller index. One very promising material is Si, which is already in widespread use as a flat substrate for GaN/Si epitaxy despite high dislocation densities. The next talk will present our predictions of the most promising Si(\textit{hhk}) substrates for growing non-polar GaN with high crystalline quality.