METALORGANIC CHEMICAL VAPOR DEPOSITION OF 3D AND 2D GROUP-III NITRIDES
Open Access
- Author:
- Bansal, Anushka
- Graduate Program:
- Materials Science and Engineering
- Degree:
- Doctor of Philosophy
- Document Type:
- Dissertation
- Date of Defense:
- August 10, 2021
- Committee Members:
- Nasim Alem, Major Field Member
Sukwon Choi, Outside Unit & Field Member
Suzanne Mohney, Major Field Member
Joan Redwing, Chair & Dissertation Advisor
John Mauro, Program Head/Chair - Keywords:
- chemical vapor deposition
group III nitrides
2D materials
gallium nitride
boron nitride
2-dimensional gallium nitride
thin film deposition
semiconductor thin films - Abstract:
- Nitride based wide bandgap (WBG) semiconductors including gallium nitride (GaN), aluminum nitride (AlN), hexagonal boron nitride (hBN) are extremely important components for several applications of technological importance. This includes blue light-emitting diodes (LEDs) used for solid state lighting, power-switching devices as well as next generation technology such as quantum computing for faster communication, etc. To further the innovation and economic growth in the field of science and technology, it is crucial to explore and establish the viability of nitride semiconductors for existing as well as next generation technologies, beyond what is currently possible. This can only be realized by developing and modifying synthesis methods, such that it can be used as a knob to fine tune the structural as well as electronic properties of the material for any targeted application. This dissertation primarily focusses on the synthesis of nitride materials, correlating structural and electronic/ opto-electronic properties of the material. The dissertation is divided into four parts focusing on metalorganic chemical vapor deposition (MOCVD) of 3-dimensional (3D) and 2-dimensional (2D) group-III nitride semiconductors as described below. The first part of the thesis focuses on investigating the role of strained (111) silicon (Si) surface layer as a virtual substrate for GaN growth. The heteroepitaxial growth of group III-nitrides on silicon is of significant interest for applications in solid state lighting, displays and power electronics. Typically, GaN epitaxy is carried out on (111) Si substrates due to its three-fold symmetry which is crystallographically compatible with the (0001) plane of wurtzite GaN. GaN when deposited on (111) Si, exhibits a high crack density in the film due to lattice mismatch and coefficient of thermal expansion (CTE) mismatch between the film and the substrate. In this case, tensile strained Si/Si1-xGex epilayers were deposited by low pressure CVD on (111) Si substrates. These virtual substrates were then used to study the effects of strain on the nucleation and subsequent epitaxial growth of AlN and GaN via MOCVD. In situ wafer curvature measurements were employed to monitor stress relaxation in the Si1-xGex layers and measure tensile strain in the Si layer. It was found that GaN epitaxy on biaxially strained Si/Si1-xGex/ (111) Si using only a thin AlN buffer layer resulted in a surface that was relatively free from cracks compared to films grown directly on (111) Si, thus paving a path for crack free, high performance GaN on (111) Si based integrated circuits. The second part of the thesis focuses on exploring alternative n-type dopant for nitride semiconductors. Intentional n-type doping of AlxGa1-xN has been widely studied for years, both experimentally and theoretically given its critical use in current group III-nitride based LEDs and high electron mobility transistors (HEMTs) as well as emerging applications in ultrawide bandgap devices. Silicon (Si), the most intensively investigated donor impurity, is an effective n-type dopant in AlxGa1-xN for Al fractions up to 80–90% after which a sharp decrease in carrier density and a steep increase in donor activation energy have been reported. However, incorporation of Si during growth of AlxGa1-xN on sapphire and SiC substrates introduces tensile stress in the film with increasing layer thickness. The tensile stress is relaxed via the formation of channeling cracks that are detrimental for device performance. To achieve higher performance, an alternative dopant is needed that enables high electron concentrations without compromising film quality. Theoretical studies predicted that Ge as an alternative dopant, however, its impact on film stress had not yet been experimentally investigated. In the dissertation, impact of Ge doping on the film stress in AlGaN films using in situ wafer curvature measurements combined with postgrowth structural characterization was investigated. It was found that unlike Si doping, Ge doping does not induce significant inclination of edge-type threading dislocations (TDs) or additional tensile stress into the AlxGa1-xN films for low Al fractions (x< 0.5) which is advantageous to reduce the formation of channeling cracks in device structures that comprises of thicker, heavily n-type doped layers. The third part of the dissertation shifts focus from 3D group-III nitride semiconductors to 2D form. This part focuses on thin film process development of hBN-2D layered group-III nitride semiconductor. These ultrathin materials have potential for device miniaturization and flexible electronics for future generation electronics. While bulk hBN crystals are available for exfoliation, the size of the crystals is limited, hindering the technological impact of the material. Consequently, there is critical need to grow large-area hBN films using CVD, an industrially scalable technique, on technologically relevant substrates such as sapphire (α-Al2O3). However, there are obstacles related to material synthesis that needs fundamental research to reach fruition towards large area electronic applications. Diborane (B2H6) and (NH3), carbon free precursors used in this dissertation, ideal candidate to obtain carbon contaminant free hBN films, reacts at temperatures as low as room temperature to form H3N-BH3 and other volatile B-N species, that affects the growth and yield of hBN thin films. Therefore, it is important to understand the gas phase reactions between the precursors to control the growth rate, crystallinity of the film along with uniformity over large areas, a must for industrial scale process transfer. To understand the gas phase reactions, the effects of growth conditions on hBN growth rate using continuous vs. flow modulation epitaxy (FME) method was studied. In continuous mode, hBN growth rate decreases with an increase in growth temperature, reactor pressure and decrease in gas velocity. This is attributed to increased gas phase polymerization of intermediate products such as borazine (B3N3H6) which forms high molecular weight species that do not contribute to hBN film growth. When the precursors are sequentially pulsed using FME method, the hBN growth rate increases by ~25 times compared to continuous mode. Thus, this study provides additional insights into the gas phase reactions involved in hBN deposition process. Moreover, growth temperature >1200ºC is required to form highly crystalline hBN films on technologically relevant substrates such as sapphire (α- Al2O3) used in this dissertation. Lattice mismatch between film and substrate, CTE mismatch, substrate stability in growth ambient are some factors to consider while choosing the substrate for thin film deposition. During high temperature hBN growth, thermal stability of the sapphire substrate is a huge concern which is not addressed by the community yet. In the dissertation work, it was found that depending on the crystallographic plane of the sapphire substrate used for deposition, the surface of the substrate modifies differently under high temperature growth conditions. C-plane sapphire modifies significantly leading to Al rich surface, A-plane sapphire on the other hand, did not show noticeable change under high temperature H2 ambient. The hBN obtained on both the planes of sapphire is similar in terms of structural as well as optical properties. Thus, this study demonstrates that substrate stability becomes an extremely crucial factor for selecting the substrate during epitaxial growth. The fourth part of the dissertation focuses on recently discovered 2D form of GaN. Confinement of conventional 3D nitride materials in two-dimensions “i.e., 2D nitrides” brings an entirely new platform of tunable properties in group-III nitrides. Recently, our research group found that graphene can stabilize 2D forms of traditionally “3D” binary compounds, when intercalated between graphene and SiC, a technologically relevant substrate. This brings about the experimental realization of new wide bandgap 2D materials, beyond hexagonal boron nitride (hBN), that have only been predicted in theory. Following this work, 2D-Ga, ~2-3 layers of metal intercalated between the same interface was realized by Prof. Robinson’s research group. This combination of 2D metal and 2D semiconductor creates new avenues of research, with endless possibilities that can be achieved in electronic and optoelectronic devices. One such combination is intrinsic metal-semiconductor lateral heterojunction that can be formed intrinsically in the material without post process fabrication that contaminates the metal semiconductor junction and effects the properties of these junction. To intrinsically form metal-semiconductor lateral heterojunctions, CVD growth parameters must be tuned to controllably decorate areas with metal (2D-Ga) and semiconductors (2D-GaNx). To controllably synthesize 2D-Ga/ 2D-GaNx lateral heterostructures, the focus was on understanding the interactions between defects in graphene that acts as pathways for intercalation and the precursors, trimethylgallium (TMGa) and ammonia (NH3) as Ga and N precursors. Thus, with the help of extensive spectroscopy and first principle calculations in collaboration with Prof. Adri van Duin’s research group, it was found that 2D-GaN tends to intercalate under the areas of O rich defective graphene whereas Ga tends to intercalate under areas of O deficient defective graphene. Moreover, GaN intercalation was sensitive to the graphene layer thickness. 2D-GaNx preferred to form in regions of thinner defective graphene, whereas 2D-Ga was stable in regions of thicker graphene. Thus, by controlling the defects in the graphene film along with layer thickness across different regions, it is possible to create metal (2D-Ga)-semiconductor (2D-GaNx) heterojunctions without post processing. The works of this dissertation were primarily funded by National Science Foundation’s DMR-1410765 - NSF GOALI, 2D Crystal Consortium–Materials Innovation Platform (2DCC-MIP) under NSF cooperative agreement DMR-1539916 and DMR-1808900. The findings and conclusions of this dissertation work does not necessarily reflect the view of the National Science Foundation.