研究期間：10108~10207;This project is a continuation of the on-going research in growth of GaN on Si with self-assembly nanostructured buffer layer (NSC 100-2218-E-008-015, 2011/10/1 ~ 2012/7/31). Through the novel growth technique , we aim to cheapen the cost and to improve the heat dissipation ability of nitride-based devices. In the on-going research, the self-assembly GaN and ZnO nanorods, and subsequent epitaxial growth is expected to be fulfilled. In this proposal, the focus will shift to device fabrication, namely LEDs and solar cells. Compared with sapphire, Si substrates are not only cheaper, their high thermal and electrical conductivities are also attractive in terms of device performances. One of the main obstacles hindering the growth of GaN on Si is the easy crack, which stems from the huge difference in lattice constant and thermal expansion coefficient between GaN and Si. It has been found that reducing the contact area between epitaxial film and substrate to that below 1 μm2 can effectively release the mismatch stress, and thus delays the build-up of lattice strain. The so-called “nanoheteroepitaxy” not only allows the epilayer to deform in three dimensions (vertically and laterally), but also confines lattice defects in the limited contact area. Employing the self-assembly GaN or ZnO nanorods to achieve nanoheteroepitaxy, we expect to improve the crystal qualities of GaN grown on Si substrates in a cost-effective method. The fabricated LEDs and solar cells will be used to demonstrate the improvement.