IWGO 2026 Session IWGO-MoPI: Invited Poster Session

β-Ga₂O₃ is a leading ultra-wide bandgap semiconductor for next-generation power electronics due to its large bandgap (~4.8 eV) and high theoretical breakdown field (~8 MV/cm), enabling high-voltage, high-temperature operation. However, unresolved reliability issues under extreme electrothermal conditions limit its practical deployment. Here, we investigate the failure mechanisms of β-Ga₂O₃ Schottky barrier diodes (SBDs) under simultaneous forward bias and elevated temperature using in-situ transmission electron microscopy (TEM), enabling real-time observation of defect nucleation, evolution, and breakdown.

Experiments were performed using a MEMS-based heating and biasing platform integrated into the TEM, allowing controlled electrothermal stressing up to ~455 °C. Vertical β-Ga₂O₃ SBD lamellae were subjected to forward bias from 1–5 V. No structural changes were observed at ≤2 V (≤85 °C). At 3 V (185 °C), early degradation appeared as contrast variations indicating defect nucleation and localized strain. At ≥4 V (≥300 °C), rapid defect accumulation and severe structural degradation led to catastrophic failure.

The degradation process is driven by electrothermally induced vacancy cluster formation, producing significant compressive strain within the lattice. Increasing defect density promotes dislocations, stacking faults, and amorphization, primarily along (200) planes due to anisotropic thermal conductivity and low fault formation energy. Simultaneously, the Ni/Au Schottky contact degrades through interdiffusion and alloying, forming a metallic pool that penetrates the drift layer via defect-assisted pathways.

Elemental mapping reveals Ga enrichment and oxygen deficiency near the interface, indicating high vacancy concentration and compositional redistribution. The interaction between defect accumulation and metal diffusion generates localized stress fields that drive void formation, plastic deformation, and eventual cracking and delamination. These structural changes correlate with electrical degradation, including reduced forward current, increased on-resistance, lower turn-on voltage, and a significant drop in rectification ratio.

This study provides direct evidence that β-Ga₂O₃ SBD failure arises from a coupled process involving vacancy generation, defect evolution, and Schottky contact instability under electrothermal stress.

Single-event effects (SEEs) studies were carried out on gallium oxide (Ga₂O₃) vertical rectifiers using ultrafast laser pulses. The Naval Research Laboratory (NRL) wide-bandgap pulsed-laser single-event effects (PL SEE) beamline consists of a tunable Optical Parametric Amplifier (OPA) that can generate ultrafast pulses across the UV-NIR spectrum, along with an imaging system for precise beam positioning. The laser pulse is focused onto a DUT to a sub-micron beam size using microscope objectives. The beamline is fully calibrated with online monitors and controllers for laser pulse energy and spot size, allowing for accurate modeling of charge deposition within the DUT’s sensitive volume. The choice of laser wavelength depends on factors like device geometry and the semiconductor's bandgap. Two-photon absorption (TPA) technique at 350 nm laser wavelength was used to determine how the device transient response changes with deposited charge, bias, and the presence of defects. Pulsed-laser single-event measurements on Ga₂O₃ devices show that increased reverse bias and growth-related defects lead to enhanced single-event transients (SETs). This work demonstrates that the charge collection efficiency in the vertical Ga₂O₃ rectifier and the Ga₂O₃/ITO diode is near 100%. The shape of laser-induced SETs depends strongly on laser pulse energy, deposited charge distribution profile, bias, and the presence of growth-related defects. The SET decay time is most likely affected by both electrons tunneling into the hole-filled deep-level defect states and by the thermal emission of holes.

Author for correspondence: ani.khachatrian.civ@us.navy.mil

Nitrogen is a promising deep acceptor dopant for β-Ga₂O₃ because of its low diffusivity, high thermal stability, and ability to compensate unintentional n-type conductivity in epitaxial layers and at film/substrate interfaces. However, the effectiveness and controllability of nitrogen doping by MOCVD depend strongly on the nitrogen precursor chemistry. In this work, nitrogen doping of β-Ga₂O₃ using different nitrogen sources is compared, including N₂O, NH₃/N₂, and NO/N₂. N₂O can act as both an oxygen source and a nitrogen source, but nitrogen incorporation is strongly dependent on growth temperature, pressure, and reactor conditions, making precise doping control challenging. NH₃/N₂ provides more direct control of nitrogen incorporation through the ammonia molar flow rate; however, it also introduces significant hydrogen at standard β-Ga₂O₃ growth temperatures, which may compensate the deep acceptor behavior of nitrogen. In comparison, NO/N₂ offers a hydrogen-lean alternative that enables controllable nitrogen incorporation while remaining compatible with conventional MOCVD growth conditions. Using NO/N₂, nitrogen concentrations up to ~6.5 × 10¹⁸ cm⁻³ were achieved, with N/H ratios as high as ~28, indicating substantially reduced hydrogen incorporation compared with NH₃/N₂-based doping, although the incorporation efficiency remains relatively low.

An additional advantage of NO/N₂ is its strong effect on the β-Ga₂O₃ growth rate. Introducing small NO flows together with O₂ increased the film growth rate by approximately 1.8–3.2× for both TEGa and TMGa precursors. This enhancement is attributed to NO-assisted oxidation chemistry, which promotes more efficient surface oxidation of Ga species. This work will also discuss how nitrogen-doped layers grown using these precursors can be used to suppress interface Si-related conduction in β-Ga₂O₃ field-effect transistors.

Since 2014, initiated by project jointly funded to both The Ohio State University and UCSB by AFOSR, Ohio State has established a continuous, comprehensive and leading research program in wide-bandgap gallium oxide. Early efforts pioneered full-bandgap defect spectroscopy studies that resulted in the first quantitative mapping of deep level defects across the 4.8 eV bandgap. These studies were primarily based on deep level optical and transient (thermal) spectroscopy (DLOS and DLTS) methods. Over time, this foundational work matured into a comprehensive portfolio that has established how growth methods, dopants and impurities, radiation environments and electric fields influence the presence, properties and distributions of trap states. Systematic studies involving epitaxy, defect spectroscopies, high resolution electron microscopy and correlative work with theorists, have revealed physical sources and defect configurations associated with many different traps. This body of information has been used to guide growth and process development, and to explain the relative importance of different defects on factors affecting gallium oxide device characteristics, performance and optimization. In parallel, a broad group of Ohio State faculty and senior researchers have been focusing on growth of gallium oxide materials and device advancement. Substantial efforts are at the forefront of epitaxial growth (MBE, MOCVD, LPCVD), device design and fabrication, with rapid advancements in high-quality epitaxy and novel device architectures with record performance. Key breakthroughs include the development of high-mobility, modulation-doped heterostructures showing quantum transport, the optimization of high-growth-rate MOCVD kinetics while managing carrier compensation, the implementation of high-k dielectric integration to tailor electric fields for high voltage applications, and novel in-situ etching techniques for 3D device topologies, and more. Recently, as the primary infrastructure hub of the Midwest Microelectronics Consortium (MMEC), which is one of the 8 national Hubs funded by the DoD Microelectronics Commons program, Ohio State has expanded its legacy gallium oxide MBE and MOCVD facilities to incorporate additional larger area MOCVD and HVPE growth systems, a state of the art high voltage and RF device testing facility, expanded cleanroom tools, along with additional research personnel. These scalable platforms are now operational and are designed to support translational microelectronics workflows and supply external partners, positioning OSU as an open, collaborative hub for next-generation semiconductor technologies.