Investigation of Epitaxial Growth and Si Doping of ?-(Al,Ga)2O3 via Hybrid Molecular Beam Epitaxy

Abstract

Silicon (Si) has been the predominant semiconductor used in electronic devices for decades. However, as Si devices approach their performance limits, particularly in high power switching applications, there is a growing need for new semiconductors with superior properties. While Silicon Carbide (SiC) and Gallium Nitride (GaN) have been addressing this need, the demand for even higher performance continues to rise, prompting researchers to explore alternative materials like Gallium Oxide (Ga2O3). Besides the larger bandgap compared with GaN and SiC, β-Ga2O3 also has another notable merit: the availability of affordable native wafers that can be manufactured from melt-grown bulk single crystals. A critical step to fabrication of high-performance β-Ga2O3 power devices is epitaxial growth and controllable n-type doping of β-(Al,Ga)2O3 thin films and heterostructures. Molecular beam epitaxy (MBE) is one of the common epitaxial growth techniques. The epitaxy is conducted in an ultrahigh vacuum growth chamber and enables a precise control of layer thickness, doping profile, and crystal quality. Traditionally, thermal effusion cells are employed in an MBE system to sublimate solid sources (Ga, Al, Si, etc.) to form metal molecular beam flux. Active oxidants such as O3 (in Ozone-MBE) or oxygen plasma (in plasma-assisted MBE) are utilized in an oxide MBE system to supply active oxygen. However, there are the following two main challenges to grow high quality Ga2O¬3 thin films in conventional plasma-assisted MBE (PAMBE). First, solid metal sources such as Si and Al can be easily oxidized, which hinders proper control of Si/Al flux with cell temperature. Second, the growth rate of Ga2O3 in PAMBE is quite limited due to formation and desorption of Ga2O, caused by excess Ga and deficient active oxygen. These issues limit the quality of (Al,Ga)2O3 films and heterostructures and their n-type doping by conventional MBE system. My PhD thesis was focused on addressing the existing challenges in (Al,Ga)2O3 growth by PAMBE system. To overcome the oxidation issue, I developed a hybrid MBE with gases instead of Si and Al solid elements. For this purpose, a custom-built gas delivery system was developed. This system includes a gas cabinet compatible with various gas precursors, such as diluted disilane for Si and tritertiarybutylaluminum (TTBAl) for Al. Additionally, it features a flow control system to monitor and control each gas flow or pressure steadily and precisely, and a gas injection system that maintains the gas injection at the target temperature. Using the Hybrid-PAMBE, a controllable doping concentration within a wide range of concentration and with relatively uniform doping profile was achieved in Si-doped Ga2O3 films. A record high mobility of 134 cm²/Vs at a doping concentration of 3.4×1017 cm-3 was also demonstrated. Furthermore, we successfully achieved (010) β-(Al,Ga)2O3 epitaxial thin films using the TTBAl precursor, with Al compositions ranging from 1% up to 28%. Beyond addressing the oxidation issue, I also improved Ga2O3 growth rate using Ga2O instead of Ga. Using Ga2O as the Ga precursor, the growth rate of Ga2O3 almost doubled to 200 nm/h compared to a growth rate of only 100nm/h using Ga. Solving the existing challenges in the MBE growth of Ga2O3 using gas and suboxide precursors not only is a significant step forward in fully realizing Ga2O3 material properties in electrical devices and advancing next-generation industrial applications, but also provides new methodology in the epitaxial growth of oxides.