MOCVD Growth of Group-III Nitrides on Silicon Carbide: From Thin Films to Atomically Thin Layers

Open Access
Al Balushi, Zakaria Yahya
Graduate Program:
Materials Science and Engineering
Doctor of Philosophy
Document Type:
Date of Defense:
May 24, 2017
Committee Members:
  • Joan Marie Redwing, Dissertation Advisor
  • Joan Marie Redwing, Committee Chair
  • Joshua Alexander Robinson, Committee Member
  • Suzanne E Mohney, Committee Member
  • Sukwon Choi, Outside Member
  • Group-III Nitrides
  • Thin Films
  • Graphene
  • 2D Materials
Group-III nitride semiconductors (AlN, GaN, InN and their alloys) are considered one of the most important class of materials for electronic and optoelectronic devices. This is not limited to the blue light-emitting diode (LED) used for efficient solid-state lighting, but other applications as well, such as solar cells, radar and a variety of high frequency power electronics, which are all prime examples of the technological importance of nitride based wide bandgap semiconductors in our daily lives. The goal of this dissertation work was to explore and establish new growth schemes to improve the structural and optical properties of thick to atomically thin films of group-III nitrides grown by metalorganic chemical vapor deposition (MOCVD) on SiC substrates for future novel devices. The first research focus of this dissertation was on the growth of indium gallium nitride (InGaN). This wide bandgap semiconductor has attracted much research attention as an active layer in LEDs and recently as an absorber material for solar cells. InGaN has superior material properties for solar cells due to its wavelength absorption tunability that nearly covers the entire solar spectrum. This can be achieved by controlling the indium content in thick grown material. Thick InGaN films are also of interest as strain reducing based layers for deep-green and red light emitters. The growth of thick films of InGaN is, however, hindered by several combined problems. This includes poor incorporation of indium in alloys, high density of structural and morphological defects, as well as challenges associated with the segregation of indium in thick films. Overcoming some of these material challenges is essential in order integrate thick InGaN films into future optoelectronics. Therefore, this dissertation research investigated the growth mechanism of InGaN layers grown in the N-polar direction by MOCVD as a route to improve the structural and optical properties of thick InGaN films. The growth of N-polar InGaN by MOCVD is challenging. These challenges arise from the lack of available native substrates suitable for N-polar film growth. As a result, InGaN layers are conventionally grown in the III-polar direction (i.e. III-polar InGaN) and typically grow under considerable amounts of stress on III-polar GaN base layers. While the structure-property relations of thin III-polar InGaN layers have been widely studied in quantum well structures, insight into the growth of thick films and N-polar InGaN layers have been limited. Therefore, this dissertation research compared the growth of both thick III-polar and N-polar InGaN films grown on optimized GaN base layers. III-polar InGaN films were rough and exhibited a high density of V-pits, while the growth of thick N-polar InGaN films showed improved structural quality and low surface roughness. The results of this dissertation work thereby provide an alternative route to the fabrication of thick InGaN films for potential use in solar cells as well as strain reducing schemes for deep-green and red light emitters. Moreover, this dissertation investigated stress relaxation in thick N-polar films using in situ reflectivity and curvature measurements. The results showed that stress relaxation in N-polar InGaN significantly differed from III-polar InGaN due to the absence of V-pits and it was hypothesized that plastic relaxation in N-polar InGaN could occur by dislocation glide, which typically is kinetically limited at such low growth temperatures required for InGaN. The second part of this dissertation research work focused on buffer free growth of GaN directly on SiC and on epitaxial graphene produced on SiC for potential vertical devices. The studies presented in this dissertation work on the growth of GaN directly on SiC compared the stress evolution of GaN films grown with and without an AlN buffer layer. Films grown directly on SiC showed reduced threading dislocation densities and improved surface roughness when compared to the growth of GaN on an AlN buffer layer. The dislocations in the GaN films grown directly on SiC were predominantly of mixed-type dislocations. Films also contained basal plane stacking faults and {11-20} prismatic stacking faults as revealed by transmission electron microcopy (TEM) near the GaN/SiC interface. Channeling cracks were also observed in the GaN films when the AlN buffer layer was not utilized. This was attributed to tensile stress induced from the thermal expansion coefficient mismatch, which was corroborated with in situ stress measurements collected during the growth process. The results provided in this dissertation showed the potential of growing GaN films directly on SiC for vertical power devices, where the use of an AlN buffer layer typically obstructs both electrical and thermal vertical transport in such devices. Also in the second part of this dissertation, additional studies were performed to understand the nucleation behavior of GaN and AlN on epitaxial graphene produced directly on SiC. The use of graphene as a template layer for the heteroepitaxy of group-III nitrides has gained interest due to the hexagonal arrangement of the sp2 hybridized carbon atoms being similar to the (0001) c-plane of wurtzite GaN. It was observed that the nucleation of AlN and GaN was preferential along the periodic graphene coated step edges of SiC and at defects sites on the (0001) terraces due to the enhanced chemical reactivity at those regions. The density of nuclei on the (0001) terraces of graphene increased with the unintentional incorporation of nitrogen defects into the graphene lattice via NH3 exposure and intentional introduction of defects using oxygen plasma. Furthermore, Raman spectral mapping showed that GaN selectively nucleates on regions of few-layered graphene as opposed to regions of multi-layered graphene. It was also revealed that though the graphene underlayers were highly defective in the region of GaN nucleation, the GaN nuclei were single crystalline, c-axis oriented and were free of threading dislocations. In contrast, polycrystalline islands of AlN were found to nucleate on graphene, but did not produce disorder to the underlying graphene. Furthermore, to implement group-III nitrides in tunnel junctions for ultra-low voltage and steep-switching applications and single photon emitters for quantum communication and computing, extreme confinement of group-III nitrides into 2D layers must be realized. Confinement of this materials system in two-dimensions “i.e. 2D nitrides” leads to massive changes in the electronic properties. This brings to light an entirely new platform to tune the properties of group-III nitrides, such as the bandgap energy, without alloying. Graphene has proven to be a remarkable material over the past decade. The third part of this dissertation research work showed that graphene can stabilize “2D” forms of traditionally “3D” binary compounds, launching a new platform to realize many other classes of materials that are not traditionally 2D, specifically on a technologically relevant substrate, like SiC. The growth process of 2D GaN utilized the mechanism of adatom intercalation from the vapor phase in an MOCVD growth environment into the interfacial region of graphene formed on SiC. The synthesis process developed in this dissertation work was termed as “Migration Enhanced Encapsulated Growth” (MEEG). Here the mechanism of 2D nitride formation was elucidated and the role of the interface of epitaxial graphene produced on SiC in providing sufficient stabilization of the direct bandgap 2D buckled structure of GaN was discussed. The atomic structure was directly resolved from the nitrogen and gallium atomic columns using aberration corrected scanning TEM (STEM) in annular bright field (ABF) mode with supported ABF-STEM simulations. Density functional theory calculations predicted a bandgap for 2D GaN in the range of 4.79-4.89 eV which was corroborated with experimental results from UV-visible reflectance, absorption coefficient and low loss electron energy loss spectroscopy (EELS) measurements. Recognizing the impact of 2D nitrides, it can be expected that the addition of 2D GaN will enable new avenues for scientific exploration and novel optoelectronic device development.