NASA's 1989 plan for a 30 kW microwave thruster with a magnetic nozzle

What the work claims

The paper claims that a microwave electrothermal thruster (MET) can convert a large fraction of incoming continuous-wave (CW) microwave power into propellant enthalpy by sustaining a high-pressure plasma discharge inside a resonant cavity, and that adding a superconducting magnetic nozzle in the throat region can reduce the wall losses that normally limit electrothermal thrusters.¹ The design target is 30 kW of applied microwave power at 915 MHz, with nitrogen, helium, and hydrogen as candidate propellants and discharge chamber pressures up to 10 atmospheres.

The supporting prediction, made with a two-dimensional kinetics (TDK) code extended to hydrogen plasmas, is that the MET could reach a specific impulse of 2000 seconds for hydrogen at a chamber temperature of 6000 K, a pressure of 10 atm, and a power level of 100 kW.¹ The authors state that the apparatus was being assembled at the NASA Lewis Research Center and that no thrust or efficiency measurements are reported in the paper.

How it works

The MET is an electrodeless electrothermal thruster. Microwave energy from a commercial-style magnetron operating at the industrial heating frequency of 915 MHz is fed into a cylindrical cavity applicator. The standing-wave mode concentrates the electric field on the cavity axis, where it breaks down the flowing propellant gas into a plasma.¹ Free electrons in the plasma absorb the field energy and, through collisions at high pressure, transfer it to the bulk gas. The hot gas then expands through a 1.0 mm diameter throat and nozzle to produce thrust.

Key design choices are visible in the paper's description of the microwave circuit and cavity. The 915 MHz frequency was selected over the more common 2450 MHz industrial band because lower-frequency magnetrons are cheaper and more efficient at high power, and because the longer wavelength allows a larger discharge tube and throat-nozzle assembly.¹ The cavity is 45.7 cm inside diameter by 57.2 cm long, with a sliding short that can vary the interior length from 16.5 to 40.6 cm to tune either the TM₀₁₁ or TM₀₁₂ mode. The authors prefer the TM₀₁₁ mode because it places the discharge near one end of the cavity, letting the throat-nozzle assembly sit just outside the cavity wall while remaining close to the heated gas.

The magnetic nozzle is the novelty. A short superconducting solenoid mounted against the front flange of the cavity produces a maximum on-axis field of 5.7 T (fields up to 8 T exist in the NbTi windings).¹ The magnetic pressure, scaling as B squared over 2 mu-zero, reaches about 140 atm at 6 T. This pinches the ionized propellant toward the axis, away from the discharge tube and nozzle walls, and is expected to promote recombination of dissociated and ionized species downstream where the kinetic energy can be recovered.

The strongest case

The strongest honest case for the MET is that it removes the electrode erosion problem that limits arcjets and resistojets. Because the microwave field heats the gas directly, there are no electrodes in the discharge zone to erode or contaminate the propellant.¹ The high-pressure operation, up to 10 atm, also gives the MET higher thrust density than low-pressure electrostatic or electromagnetic thrusters, which matters for missions that need both high specific impulse and appreciable thrust.

The magnetic nozzle adds a plausible path to higher temperature operation. In a conventional thermal thruster, the wall material sets the temperature ceiling; tungsten or rhenium throats can survive only so much heat flux before they fail. A magnetic nozzle that keeps the hottest plasma away from the walls could, in principle, allow gas temperatures higher than the wall limits, translating directly into higher specific impulse.¹ The paper also notes that the skin effect at high electron density may flatten the plasma temperature profile by absorbing microwave energy at the plasma boundary, which could reduce radiation losses.

Where a skeptic should push

The single most important limitation is that this is a concept and apparatus paper, not a results paper. The authors repeatedly state that the hardware is being assembled and that performance numbers come from computer calculations, not from tests.¹ The over-95-percent energy absorption efficiency is an expectation based on prior low-power work by Whitehair and Asmussen; it is not a measurement from the 30 kW device.

The 2000-second specific impulse prediction is bounded by assumptions that may not hold. It comes from a TDK code calculation for hydrogen at 6000 K and 10 atm, at 100 kW, not 30 kW.¹ The calculation also does not include plasma-wall interactions such as recombination heating, conduction, convection, and radiation, which the authors identify as the main real-world loss mechanisms. They explicitly say experimental results will fall short of the TDK prediction until wall losses are minimized.

The magnetic nozzle, while physically plausible, introduces its own burdens. A 5.7 to 8 T superconducting magnet needs cryogenic cooling; the paper reports liquid helium and liquid nitrogen consumption rates of about 170 cm3/hr and 400 cm3/hr in persistent mode at full field.¹ That infrastructure adds mass and complexity that a flight system would have to justify. Startup is another concern: the discharge must be ignited at low pressure and low power, then both pressure and power are ramped together while the cavity is continually retuned. If the plasma extinguishes at high power, the cavity becomes a high-Q mismatch that can reflect damaging power back into the generator.

What it means for power in space

The non-obvious implication is that microwave power distribution and electric propulsion can be designed as a single integrated system. The MET does not merely use electricity; it relies on a narrow-band, low-ripple, 915 MHz microwave signal that must be delivered, matched, and tuned in real time.¹ A spacecraft or orbital tug built around this concept would need a power processing and microwave generation subsystem whose mass, efficiency, and thermal rejection become part of the propulsion budget. The thruster's specific impulse and thrust numbers therefore cannot be evaluated without also accounting for the mass and losses of the microwave source, the power conditioning, and the cryogenic magnet.

The genuine threat is that the high-temperature, high-pressure plasma may still destroy the containment hardware before the magnetic nozzle can protect it. The inner discharge tube must survive temperatures up to about 2000 C in nonreactive environments, and the throat-nozzle structure is to be made of tungsten, rhenium, or molybdenum.¹ If wall losses and radiation prove larger than the magnetic pinch can suppress, the MET could end up with the same lifetime limits as arcjets, negating its electrodeless advantage. This is exactly the loss mechanism the authors admit their TDK calculations omit.

The opportunity is that the MET concept points toward a class of high-power electric thrusters that could use a central space power bus. A nuclear-electric or large solar-electric spacecraft with a multi-hundred-kilowatt microwave generator could, in principle, feed one or more METs without each thruster carrying its own heavy power electronics.¹ The 915 MHz industrial frequency choice is also a hint: it suggests the hardware could leverage terrestrial high-power microwave supply chains rather than requiring a custom space-rated transmitter. Whether that supply-chain advantage survives the need for low ripple, precise tuning, and long life in vacuum is the open question.

The bottom line

What is established is a carefully documented 1989 concept and experimental plan for a 30 kW, 915 MHz microwave electrothermal thruster with a superconducting magnetic nozzle. What is not established is any measured thrust, specific impulse, or efficiency. The over-95-percent absorption and 2000-second specific impulse are design targets and code predictions, bounded by assumptions about wall loss and power level. Confidence in the physics is moderate: high-pressure microwave discharges and magnetic pinches are real phenomena. Confidence in the engineering is low because the paper stops at assembly and because the magnetic nozzle adds cryogenic mass and startup risk. The claim would be strengthened by published test data from the assembled apparatus; it would be undercut if plasma-wall interactions remained large at the intended power and pressure.