Gallium nitride on zinc oxide: A new approach to solid state lighting

The objective of the proposed research is to develop high quality GaN epitaxial growth on alternative substrates that can result in higher external quantum efficiency devices. Typical GaN growth on sapphire results in high defect materials, typically 108-10 cm-2, due to a large difference in lattice mismatch and thermal expansion coefficient. Therefore, it is useful to study epitaxial growth on alternative substrates to sapphire such as ZnO which offers the possibility of lattice matched growth. High-quality metalorganic chemical vapor deposition (MOCVD) of GaN on ZnO substrate is hard to grow due to the thermal stability of ZnO, out-diffusion of Zn, and H2 back etching into the sample.;Preliminary growths of GaN on bare ZnO substrates showed multiple cracks and peeling of the surface. A multi-buffer layer of LT-AlN/GaN was found to solve the cracking and peeling-off issues and demonstrated the first successful GaN growth on ZnO substrates. However, secondary ion mass spectrometry showed high Zn diffusion into the epilayer, which decreases the optical qualities of GaN. Therefore, InGaN was grown on bare ZnO because of the perfect lattice match with ZnO at 18% indium and low growth temperature. Good quality InGaN films were seen with the use of a low temperature (530°C) GaN buffer layer. Indium composition of 17--27% increased with a decrease in growth temperatures from 720--680°C. The InGaN films showed no indium droplets or phase separation. Shifts in InGaN-related photoluminescence (PL) emissions of 0.5+/-0.1eV may have been due to recombination involving Zn-/O-impurities due to the Zn/O diffusion from the substrate. An activation energy of 59meV was found for the InGaN epilayer, which correlated to findings in literature. A study of InGaN phase separation demonstrated ZnO's ability to sustain a higher strain state than sapphire, and thereby incorporating higher indium concentrations into the InGaN layers without phase separation. InGaN on ZnO substrates exhibited indium content as high as 43% without phase separation compared to the same growth on sapphire with only 32%. Si doping of InGaN layers, a known inducer for phase separation, did induce phase separation on sapphire growths, but not for growths on ZnO. This higher strain state for ZnO substrates was correlated to its perfect lattice match with InGaN at 18% indium concentration. Layers grown on ZnO stayed strained past 18--20% indium content, a point where sapphire started showing signs of phase separation. This finding could lead to higher efficient green LEDs at 555nm. Transmission electron microscopy results revealed reduction of threading dislocation and perfectly matched crystals at the GaN buffer/ZnO interface showing coherent growth of GaN on ZnO. These were the first reported results of successful GaN on ZnO by MOCVD. However, Zn diffusion into the epilayer was still an issue.;Therefore, an atomic layer deposition of Al2O3 was grown as a transition layer prior to GaN and InGaN growth by MOCVD. Annealing studies were performed to find the best crystalline Al2O3 phases for MOCVD growth. X-ray and PL showed distinct GaN peaks on 20nm Al2O3/ZnO whereas no peaks were seen on 50nm Al 2O3/ZnO demonstrating the first GaN films grown on Al 2O3/ZnO. Optical transmission measurements showed that the bandgap energy of InGaN was not altered significantly when grown on annealed Al2O3/ZnO substrates. X-ray photoelectron spectroscopy showed a decrease in Zn diffusion into the epilayer, demonstrating that an ALD Al2O 3 layer was a promising transition layer for GaN growth on ZnO substrates by MOCVD.