Silicon carbide is a wide-bandgap semiconductor already deployed in high-power and harsh-environment electronics, and the current research front is not whether the material works but whether its crystal defects, interfaces and cryogenic behaviour can be controlled well enough to extend it into quantum devices, ultra-wide-bandgap heterostructures and fusion components.
Summary
Silicon carbide (SiC) is a compound semiconductor with a wide bandgap and unusually high thermal and chemical stability. It exists in many polytypes, which share the same local tetrahedral Si-C bonding but differ in long-range stacking: the cubic 3C form, the hexagonal 2H and 4H forms, the 6H form used in single-crystal substrates, and rhombohedral variants such as 9R. Polytype choice sets thermal stability, phonon transport and defect behaviour, and machine-learning molecular dynamics places the relative high-temperature stability ordering at 3C > 2H > 9R, with melting initiated by the formation of short C-C contacts and carbon-rich local regions before tetrahedral connectivity is lost.
The commercial base is power and harsh-environment electronics, where SiC is described as widely deployed, alongside an emerging CMOS technology platform. Three adjacent uses drive most of the current literature. First, quantum technology: crystal defects such as divacancies act as spin qubits or single-photon sources, and defect engineering depends on migration barriers that atomistic simulation now puts at 2.12 eV for carbon vacancies and 0.88 eV for carbon interstitials, a hierarchy that governs whether interstitial-vacancy recombination or vacancy aggregation into spin-active divacancies wins. Second, SiC as a thermal substrate for ultra-wide-bandgap devices, where Ga2O3/SiC heterointegration is attractive but interfacial thermal boundary conductance is the heat-removal bottleneck. Third, nuclear and fusion use, where SiC is a leading candidate for tritium permeation barriers and where measured permeation values vary widely between ideal single crystals and real defect-containing material.
The parameters that decide outcomes are almost all defect-related rather than intrinsic. Threading dislocation density in wafers drives reverse leakage and degradation, and commercial inspection based on surface profiling and photoluminescence dark contrast misses nanoscale threading dislocations that have no resolvable surface signature. Extended defects such as Shockley partial dislocations and dislocation complexes progressively reduce stiffness in 3C-SiC. Hydrogen solubility rises sharply in carbon-rich amorphous regions and at silicon vacancies relative to interstitial sites in pure beta-SiC, which is precisely the difference between an ideal permeation barrier and a leaky one. Metrology is being rebuilt around this: contact-free terahertz reflection spectroscopy is being mapped for accessible doping ranges in SiC, Si and GaN layer stacks, and machine-learned interatomic potentials now support multimillion-atom, microsecond simulations of SiC from ambient to extreme conditions.
Evidence base
- Silicon carbide is described as widely deployed in high power and harsh environment electronics with an emerging CMOS platform, while commercial power MOSFETs cooled from 300 K to 650 mK show large gate hysteresis, threshold voltage shifts and subthreshold swing deterioration, attributed to carrier freeze-out and high interface trap density, 5 May 2026.
- Commercial SiC wafer inspection based on surface profiling and photoluminescence dark contrast misses nanoscale threading dislocations, which drive reverse leakage and degradation; confocal subsurface backscattering microscopy detects them non-destructively and distinguishes their type, 4 June 2026.
- Migration barriers of 2.12 eV for carbon vacancies and 0.88 eV for carbon interstitials in 3C-SiC set the mobility hierarchy that determines whether vacancies aggregate into spin-active divacancies or recombine, 28 May 2026.
- Interfacial thermal boundary conductance is the major heat-removal bottleneck for Ga2O3/SiC heterointegration, and is consistently higher for Ga2O3(-201)/SiC(0001) than the alternative orientation studied, 8 May 2026.
- Fracture energy of the 6H-SiC {10-10} plane measured 7.5 +/- 0.3 J/m2 by microscale double cantilever beam, while focused-ion-beam-notched single cantilever beams gave over twice that value due to Ga implantation and residual stress, 3 June 2026.
- Hydrogen solubility in beta-SiC is significantly enhanced in carbon-rich non-stoichiometric amorphous structures and at silicon vacancies compared with interstitial sites in pure material, bearing on SiC’s use as a fusion tritium permeation barrier where experimental permeation values vary widely, 4 June 2026.
Open questions
- Can confocal subsurface backscattering microscopy, or an equivalent optical method, be run at wafer-scale throughput and integrated into commercial SiC incoming inspection, rather than remaining a laboratory imaging technique?
- Is the cryogenic instability of SiC MOSFETs a property of commercial process choices, in particular interface trap density, or intrinsic to carrier freeze-out in the material, and can a redesigned device fix it?
- Do the predicted orientation-dependent thermal boundary conductances at Ga2O3/SiC interfaces survive experimental measurement on real, non-ideal interfaces, and is the resulting heat removal sufficient for ultra-wide-bandgap power devices?
- Can processing be tuned to favour divacancy stabilisation over interstitial-vacancy recombination at scale, given the 2.12 eV and 0.88 eV migration barriers for carbon vacancies and interstitials?
Assessment drafted 2026-08-31 from up to 18 KB sources using the technology-scorecard framework; scores are a draft read pending review.