GaN vs Silicon for Space RF Power Amplifiers

Why the semiconductor material matters

In an RF power processing unit the amplifier stage converts DC from the spacecraft bus into a high-power radio-frequency carrier for the thruster antenna. Its efficiency sets the floor for the whole unit, and the waste heat it cannot avoid has to be conducted to a baseplate and radiated to space. The transistor technology in that stage decides how much power fits in a given volume, how much of the input becomes RF rather than heat, and how hot the device can safely run. For decades that transistor was silicon. Today the realistic options are silicon LDMOS and gallium nitride (GaN) on a silicon carbide or silicon substrate.

The four properties that separate GaN from silicon

Four material parameters explain almost all of the practical difference between the two technologies:

• Bandgap. GaN has a wide bandgap of roughly 3.4 eV against about 1.1 eV for silicon. A wider bandgap means fewer carriers are thermally generated, so the device keeps working at higher junction temperatures before leakage takes over.
• Critical breakdown field. GaN withstands roughly an order of magnitude higher electric field before avalanche breakdown than silicon. A thinner, more heavily doped drift region can then hold the same voltage, which lowers on-resistance and device capacitance.
• Electron mobility and saturation velocity. The two-dimensional electron gas at the GaN/AlGaN interface has high mobility and high saturation velocity, so charge moves through the device quickly. That is what makes fast switching and high-frequency operation practical.
• Thermal conductivity. This depends on the substrate. GaN on silicon carbide conducts heat better than silicon; GaN on a silicon substrate is closer to silicon and needs more attention to the thermal path.

Power density, efficiency and switching speed

The higher breakdown field lets a GaN device operate at a higher drain voltage and deliver more power per unit of gate width. In RF terms this shows up as power density on the order of three times that of a comparable silicon LDMOS device. For a fixed RF output that means a smaller die, a smaller package, and less parasitic capacitance to drive.

Lower capacitance and fast carrier transport also cut the energy lost each time the transistor switches. In a switch-mode Class-E stage, where the device is driven hard between fully on and fully off, that reduction in switching loss is the main reason GaN reaches higher drain efficiency than silicon at the same frequency. The result is more RF output for the same DC input and less heat to reject, which is the single most valuable trade in a space-grade generator.

High temperature and radiation behaviour

The wide bandgap gives GaN a genuine high-temperature advantage. Leakage current rises much more slowly with temperature than in silicon, so the device tolerates a hotter baseplate and a smaller thermal margin. In a conduction-cooled unit that relaxes the radiator sizing, which flows straight back into spacecraft mass.

On radiation, the picture is more nuanced and should be treated qualitatively. GaN high-electron-mobility transistors are generally regarded as tolerant to total ionising dose, because the active channel is a majority-carrier structure with no gate oxide to trap charge, which is the usual weak point in silicon MOS devices. Displacement damage and single-event effects still need to be characterised for the specific part, bias condition and mission environment, and that testing is part of any qualification campaign. No general claim substitutes for device-level radiation data.

The European GaN sovereignty context

RF GaN is also a supply-chain question. Many established GaN RF process lines sit under United States export jurisdiction, which brings ITAR or EAR obligations that complicate a European propulsion programme. European institutions have responded by funding domestic GaN capability through space and defence programmes, so a fully European RF GaN supply chain is now a realistic design constraint rather than an aspiration. Sourcing the power device and its foundry inside Europe is central to an ITAR-free RF PPU and is one reason MHz Labs builds around a European GaN supply chain.

When GaN is the right choice, and when it is not

GaN is the right choice when efficiency and power density carry a real cost, which is exactly the case in orbit, and when the stage runs at high frequency or switches hard. For a 13.56 MHz plasma-thruster amplifier in the tens to hundreds of watts, using GaN in a switch-mode topology is the natural fit.

Silicon LDMOS remains sensible where none of that pressure applies: ground support equipment, laboratory generators, cost-driven terrestrial transmitters, and designs that lean on decades of accumulated LDMOS ruggedness and load-mismatch survivability data. Silicon also still leads on raw unit cost and on the breadth of qualified part history. The decision is a system trade, not a slogan.

Sources and further reading

• Goebel and Katz, Fundamentals of Electric Propulsion, NASA JPL DESCANSO, for the propulsion power-budget context.
• Takahashi, Helicon-type radiofrequency plasma thrusters and magnetic plasma nozzles, review of RF thruster power needs.
• European Union, Regulation (EU) 2021/821, the dual-use export control framework relevant to RF GaN.