Interest in β-Ga₂O₃ has dramatically increased in recent years due to the material’s potential promise for use in power electronics and extreme environments. Its combination of a monoclinic structure (C2/m space group), two inequivalent tetrahedral and octahedral gallium sites and three inequivalent oxygen sites, and a bandgap of 4.8 eV, 1.4 eV above that of gallium nitride, creates a semiconductor material with a unique set of properties. This is further aided by β-Ga₂O₃’s uncommon capability among the ultra-wide bandgap oxides to be grown into high quality single crystal substrates using both melt-based bulk and thin film growth and deposition methods. Defects and their stability and dynamics under static and extreme environments can limit the incorporation of β-Ga₂O₃ into new applications. Therefore, a direct visualization and in-depth understanding of the defects and their interplay with the environment is vital for understanding the materials properties and the device breakdown under extreme conditions. In this presentation we will discuss the atomic, electronic, and chemical structure of the defects in doped and UID β-Ga₂O₃ using scanning transmission electron microscopy (S/TEM) imaging and electron energy loss spectroscopy (EELS). In addition, we will discuss the electronic structure and the local properties in β-Ga₂O₃ under extreme conditions using STEM-EELS. This fundamental understanding is important to uncover the breakdown behavior in β-Ga₂O₃ and the impact of defects on its device performance.

In recent years, there has been significant interest in β-Ga₂O₃ as a potential candidate for the next generation of power electronics, solar-blind ultraviolet (UV) detectors, and as a substrate for UV light emitting diodes (LEDs). This interest stems from its ultra-wide bandgap of 4.8eV. Thin film growth and n-type doping (Si, Sn, Ge) of Ga₂O₃ have been achieved through various methods such as metal-organic chemical vapor deposition (MOCVD), pulsed laser deposition (PLD), and molecular beam epitaxy (MBE). However, MBE has limitations in terms of the growth rate of Ga₂O₃ due to the desorption of volatile Ga₂O, which is formed from the reaction between Ga and Ga₂O₃. Using gallium sub-oxide (Ga₂O) instead of elemental gallium has been previously employed [1] as a technique to enhance the growth rate of Ga₂O₃ by Ozone-MBE. However, this technique has not yet been investigated in plasma-assisted MBE. In my talk, I will present the results of our recent studies on using Ga₂O as Ga source in PAMBE. Using the same plasma conditions, we show that using Ga₂O instead of Ga can at least double the growth rate of Ga₂O₃.

Previously, we have demonstrated uniform and controllable silicon doping of β-Ga₂O₃ by utilizing disilane (Si₂H₆) as the Si source. [2] In my talk, I will show that this technique is also compatible with utilizing Ga₂O as Ga source. The silicon doping can be tuned from 3×10¹⁶ cm^-3 to 1×10¹⁹ cm^-3 using the diluted disilane source.

References:

1. Vogt, P., Hensling, F. V., Azizie, K., Chang, C. S., Turner, D., Park, J., ... & Schlom, D. G. (2021). Adsorption-controlled growth of Ga₂O₃ by suboxide molecular-beam epitaxy. Apl Materials, 9(3), 031101.

2. Wen, Z., Khan, K., Zhai, X., & Ahmadi, E. (2023). Si doping of β-Ga₂O₃ by disilane via hybrid plasma-assisted molecular beam epitaxy. Applied Physics Letters, 122(8)

Crystalline defects of technologically mature materials have been identified and classified by the semiconductor industry [1,2], since it is economically beneficial to isolate failure mechanisms at the source rather than relying on backend testing. This has significantly improved device reliability. The various defects could be categorized into killer or non-killer defects, where killer defects can hinder the operation of high-performance devices by trapping charge carriers or causing increased leakage current. Although β-gallium oxide (β-Ga₂O₃) is expected to surpass silicon carbide (SiC), defects in Ga₂O₃ are prevalent and largely unclassified. Therefore, screening out defects that cause electrical device degradation must be solved for widespread adoption of β-Ga₂O₃.

In this work, photoemission electron microscopy (PEEM) is used to visualize micrometer-scale defects and determine their electronic impact. PEEM is based on the photoelectric effect and is a non-destructive analysis method where light is used to excite and eject electrons from the sample surface and these electrons are analyzed. We investigated the defects on commercially-available epitaxially-grown β-Ga₂O₃ on (010) β-Ga₂O₃ substrates. The epitaxy was formed by hydride vapor phase epitaxy (HVPE) with a target doping of 1x10¹⁸ cm^-3 on the (010) semi-insulating β-Ga₂O₃ wafer. We identified elongated structures on the β-Ga₂O₃ epi-layer as shown in Figure 1a, and they appear in multiple instances of the sample surface and in a parallel configuration. These features resemble the “carrot” defect observed in SiC epitaxy [3]. From the imaging spectroscopy mode of the PEEM (Figure 1b), the base and tip of the carrot were found to have similar valence band maxima but dissimilar work functions. The spectra from the tip of the carrot resembles that of the surrounding β-Ga₂O₃ epi-layer. We are performing ongoing work to identify this feature as a microscopic defect. For understanding the electrical influence of these elongated features on HVPE epi-layer, we will perform tunneling atomic force microscopy (TUNA) to measure the electrical properties on and off the defect surface. Together, we will present a discussion on the nature of these distinct features and their implication on device performance.

Iridium is a promising material for Schottky contacts on β-Ga₂O₃ due to its high work function, enabling the establishment of high-quality devices with large barrier heights and superior thermal stability. These advantages make it a desirable option compared to other metals for developing UV photodiodes, high-speed rectifying diodes, and metal-semiconductor field-effect transistors. Ir-related defects can form at the interface between the Ir layer and β-Ga₂O₃. It is essential to understand their impacts on carrier density and transport, and their role as charged scattering and recombination centers. We used scanning transmission electron microscopy (STEM) to investigate the atomic scale mechanism of how the Ir atoms diffuse and incorporate into the β-Ga₂O₃ lattice and affect the properties of the interface. We studied the Ir metal contacts from the cross-sectional view of the interface after high-temperature annealing which is frequently used to optimize the metal contacts, as well as in device operation conditions where the application of the electric field may further diffuse or redistribute the Ir atoms. Ir diodes were fabricated using electron beam evaporation on Tamura UID (010) Ga₂O₃ substrates grown by edge-defined film-fed growth (EFG) method. The Ir/Ga₂O₃ samples before and after the rapid thermal annealing at 700, 900, and 1100 °C for 20 minutes were investigated using atomic-scale STEM and nanoscale depth-resolved cathodoluminescence spectroscopy (DRCLS). The analysis revealed high densities of Ir-related defects, with their concentrations ranging from ~ 4.1 × 10¹⁹ to 2.1 × 10²⁰ cm^-3 at the interfacial region of the 700°C annealed sample, which we correlate to the significant reduction in the series resistance in the current-voltage (I-V) characteristics, and also potentially to the depletion region through the defect-assisted tunneling. Multiple types of atomic scale defects were identified at the interface, including the nanoscale clustering of iridium oxide, interstitial-divacancy complexes similar to those observed in Sn-doped β-Ga₂O_3, substitutional Ir on O sites, and the “V-shaped defects” with displaced Ir atoms at the interstitial sites. High concentrations of defects also led to local phase transformations to γ-Ga₂O₃, especially when annealed at higher temperatures (900°C and above), which corresponds to a sharp peak in the DRCLS that is indicative of a distinctive energy state of the defect. The samples were also biased both in-situ and ex-situ to investigate whether the displaced Ir interstitials function as diffusion paths or defects with a net electrical charge state.