NSF program aims to model radiation damage in GaN RF power amplifiers

What the work claims

The NSF award describes a 36-month, $450,000 project that will test three linked hypotheses about radiation damage in AlGaN/GaN high-electron-mobility transistors (HEMTs). The first hypothesis is that radiation dynamically changes the knee-sharpness factor gamma, breaking the gamma equals 2 invariant used in current compact models. The second is that radiation reconfigures the dominant carrier-scattering mechanisms, producing measurable shifts in the mobility exponent alpha and the saturation velocity vsat. The third is that atomistic defect densities can be tied deterministically to these macroscopic transport shifts, making predictive radiation-hardness-by-design possible for the first time.¹

The project already has one concrete preliminary result: using the Gate Current Injection Method and a Unity-Intercept framework, the investigators extracted gamma values that increase systematically with temperature from approximately 4.5 at 4 K to approximately 6.5 at 550 K. This directly contradicts the gamma equals 2 assumption embedded in existing compact models. The planned work extends that temperature-baseline approach across the radiation-fluence dimension, with proton irradiation at Los Alamos up to 10^15 protons per cm^2 across 2 to 6 MeV energies, and embeds the resulting scaling laws in an open-source Radiation-Aware ASM-HEMT Library.¹

How it works

An AlGaN/GaN HEMT is a wide-bandgap transistor in which a two-dimensional electron gas forms at the interface between an aluminum gallium nitride barrier and a gallium nitride channel. That electron gas gives the device very high electron mobility and breakdown field, which is why AlGaN/GaN HEMTs have become the enabling technology for RF power amplifiers in satellite communications, defense radar, and deep-space platforms. In space, however, the same devices are exposed to protons and other radiation that create point defects, trap charges, and alter the electric field profile inside the transistor.¹

Compact models such as the industry-standard ASM-HEMT represent transistor behavior with equations small enough to run inside a circuit simulator. A key parameter in those equations is the knee-sharpness factor gamma, which controls how sharply the device current transitions from the linear region to saturation around the knee voltage. Existing compact models assume gamma equals 2. The award abstract reports that the investigators' preliminary physics-extracted values rise from roughly 4.5 at cryogenic temperature to roughly 6.5 at 550 K, meaning the real transition is sharper or softer than the standard model assumes and that the error is temperature-dependent.¹

The four research thrusts are designed to isolate cause and effect. The Gate Current Injection Method screens devices before irradiation. Wide-temperature baseline characterization from 4 K to 550 K separates thermal effects from radiation effects. Proton irradiation at the Los Alamos Ion Beam Materials Laboratory introduces controlled displacement damage across 2 to 6 MeV energies up to a fluence of 10^15 protons per cm^2. Finally, the ASM-HEMT compact model is extended so that mobility, saturation velocity, and knee-sharpness become functions of both temperature and fluence, and the extended model is validated against RF measurements.¹

Where a skeptic should push

The strongest result so far is temperature-dependent, not radiation-dependent. The gamma equals 4.5 to 6.5 finding comes from a pre-irradiation baseline and tells us that current compact models are wrong about temperature, which is important but not the same as proving that radiation fluence can be mapped deterministically to transport shifts. The jump from a temperature baseline to post-irradiation prediction is the load-bearing assumption of the whole award.¹

The claimed benefit of reducing qualification timelines from many months to a single confirmatory test is an aspiration, not a demonstrated outcome. It depends on the open-source library accurately predicting RF performance without empirical adjustment after irradiation, which has not yet been shown. A skeptic should also note the modest budget: $450,000 over three years for a collaboration between a university and a national-laboratory ion-beam facility is a thin pipeline for device fabrication, irradiation, modeling, and RF validation. The CINT user program access helps, but schedule risk is real.

Finally, the gamma values themselves are extracted with the investigators' own GCIM and Unity-Intercept framework. If that extraction method contains a systematic bias at temperature extremes, the entire scaling law will inherit it. Independent replication with separate devices and extraction techniques will be needed before the model can be treated as a qualification shortcut.

What this means for radiation-tolerant RF power in space

The non-obvious implication is that the bottleneck for high-power RF in deep space may be the compact model, not the transistor. GaN HEMTs already deliver higher power density and efficiency than the alternatives, but mission designers cannot use them confidently because qualification still relies on part-specific radiation testing. A verified radiation-aware compact model would let engineers simulate post-irradiation amplifier performance during the design phase, which could reduce the conservatism built into power budgets and thermal designs for phased-array transmitters, radar, and power-beaming links.¹

The genuine threat is a false sense of predictability. If the model is released before the radiation-fluence physics are solid, a program could design an amplifier that passes pre-launch simulation but degrades faster than predicted on orbit. That risk is especially acute for deep-space missions where repair is impossible and the radiation spectrum differs from the 2 to 6 MeV proton baseline used to build the model.

The opportunity is standardization. An open-source Radiation-Aware ASM-HEMT Library that is validated across multiple foundries could become the common reference for space power electronics, much as the ASM-HEMT model itself has become a reference for terrestrial GaN design. If that happens, the same physics-aware approach could be extended to other wide-bandgap power devices, including the power converters and inverters that form the backbone of spacecraft PMAD and high-voltage distribution systems.

The bottom line

This is a well-scoped research proposal with one concrete preliminary finding and a clear experimental plan. The temperature dependence of the knee-sharpness factor is real data that undermines a common modeling assumption. The radiation-fluence scaling laws and the open-source predictive library are still objectives, not achievements. The project is worth watching because success would change how radiation-hard RF power amplifiers are designed and qualified for space, but the claimed timeline compression remains a hypothesis until the model is validated against measured post-irradiation RF performance.