Wide-bandgap semiconductors are substrate materials such as SiC, GaN, AlGaN, AlN, Ga2O3 and diamond whose larger electronic bandgap allows power switches, UV detectors and photonic devices to operate at voltages, temperatures, radiation doses and optical wavelengths where silicon fails; the SiC/GaN layer is already in service while the ultra-wide-bandgap frontier is still fighting heat removal and process complexity in the lab.
Summary
A semiconductor’s bandgap is the energy needed to lift an electron into conduction. Making it larger has three consequences that matter for devices. First, the material tolerates a much higher electric field before avalanche breakdown, so a power transistor blocking a given voltage can use a thinner, more heavily doped drift region and therefore has lower conduction loss. Second, intrinsic carrier generation stays negligible at high temperature and under ionising radiation, so devices keep working in engine bays, reactor halls and orbit. Third, the material is transparent further into the ultraviolet and does not absorb two near-infrared photons at once, which makes it useful for solar-blind detectors and for nonlinear photonics. The family runs from the commercial wide-bandgap (WBG) materials, silicon carbide and gallium nitride, to the ultra-wide-bandgap (UWBG) set: AlGaN and AlN, gallium oxide, diamond, and in photonics oxides such as tantalum pentoxide.
The sources here show that the WBG base layer is established. SiC polymorphs are described as widely employed nuclear materials, mechanical components and wide-bandgap semiconductors, and AlGaN/GaN HEMTs are described as essential to satellite communications, defence radar, GPS and deep-space missions. Public funding for wide-bandgap power conversion goes back at least to ARPA-E’s 2017 awards ref.
The live research frontier is UWBG, and the 2026 record here is device-level rather than product-level. An epitaxially regrown AlN gate interlayer on UWBG AlGaN polarisation-graded FETs produced an average breakdown field above 6.94 MV/cm with on-state current above 1 A/mm and a power-switching figure of merit above 1.87 GW/cm2 at a breakdown voltage above 1.45 kV. A vertical gallium oxide transistor was demonstrated without planarisation etch-back or mid-gap acceptor blocking layers, with 200 nm fins, a threshold voltage of 2 V and an on/off ratio above 1e7, and was presented as comparable to existing vertical transistors but more manufacturable. Outside power switching, sub-bandgap AlN Schottky photodetectors gave non-saturating linear response to blue light above 40 W/cm2 and undistorted response to at least 300 °C.
The parameters that decide the outcome are: breakdown field and figure of merit actually achieved in manufacturable geometries; heat extraction, since interfacial thermal boundary conductance at heterointegrated interfaces such as Ga2O3 on SiC is called a major heat-removal bottleneck; controllable doping, still hard enough in diamond that measuring the Fermi energy requires indirect photoluminescence methods; and process compatibility, since several of the best results depend on epitaxial regrowth or electron-beam lithography.
Evidence base
- 3 June 2026: UWBG AlGaN polarisation-graded FETs with regrown epitaxial AlN gate interlayers reached average breakdown fields above 6.94 MV/cm, on-current above 1 A/mm, and a power-switching figure of merit above 1.87 GW/cm2 at breakdown voltage above 1.45 kV.
- 14 August 2026: first vertical Ga2O3 transistor architecture without planarisation etch-back or mid-gap acceptor regions, using 200 nm fins and ~1.2 um trenches, showing enhancement-mode operation with 2 V threshold and on/off ratio above 1e7.
- 8 May 2026: interfacial thermal boundary conductance is identified as a major heat-removal bottleneck for Ga2O3/SiC ultra-wide-bandgap power electronics, with direct experimental access to intrinsic interfacial transport still limited.
- 15 June 2026: sub-bandgap AlN Schottky photodetectors gave non-saturating linear response to blue light above 40 W/cm2 and undistorted response at up to at least 300 °C, targeting industrial process control, thermal and nuclear power, aeronautics and spaceflight.
- 4 August 2026: NSF awarded $450,000 to Texas Tech with Los Alamos to replace approximate, trial-and-error radiation-damage models for AlGaN/GaN HEMTs, noting that current practice forces costly per-mission testing.
- July 2026: an analysis of ~1,550 semiconductor startups names wide-bandgap materials among the ‘More than Moore’ domains where positions are still forming and identifies funding velocity and exits as Europe’s binding gaps.
Open questions
- Can interfacial thermal boundary conductance in heterointegrated stacks such as Ga2O3 on SiC be raised enough, and verified experimentally rather than by simulation, to let UWBG devices run at the current densities their breakdown fields imply?
- Do the record UWBG device results survive translation from epitaxial regrowth and electron-beam lithography to production-compatible processes and larger areas?
- Is controllable doping, including any usable p-type route, achievable in AlN, Ga2O3 and diamond, given that even Fermi-level determination in diamond currently requires indirect photoluminescence analysis?
- Will physics-based radiation and reliability models arrive fast enough to replace per-mission empirical qualification, which is currently the cost barrier for space and defence adoption?
Assessment drafted 2026-08-31 from up to 13 KB sources using the technology-scorecard framework; scores are a draft read pending review.