Chinese Scientists Build First Low-Power Superconducting Space Thruster

Inside a vacuum chamber in Hefei, China, a thread of glowing plasma hangs in space. Argon gas feeds into the device at five milligrams per second — barely a trickle. Electromagnetic forces tear the gas apart into ions and electrons, then hurl it backward at extraordinary speed. The thrust produced is, on its face, modest: about 320 millinewtons, roughly the weight of a small apple resting in your palm. But the efficiency with which that thrust is generated is another matter entirely.

Spacecraft propulsion has always involved an awkward trade-off. Chemical rockets are powerful but profligate: burn kerosene and liquid oxygen, accelerate combustion products out the nozzle, and you produce impressive thrust. You also, in the process, consume enormous quantities of propellant. More than 90 per cent of a rocket’s launch mass is typically fuel. The rest — the satellite, the science instruments, the actual mission — rides along almost as an afterthought. Electric propulsion sidesteps this by using electrical energy to accelerate propellant rather than chemical energy. The technique is far more frugal. The same mission can be accomplished with a fraction of the fuel. The downside is that electric propulsion generates comparatively gentle thrust; you don’t launch off Earth’s surface with an ion engine. You use it once you’re already in space, running continuously, building velocity gradually over months or years.

The metric that matters here is specific impulse — effectively a measure of fuel efficiency, analogous to miles per gallon. Chemical rockets achieve specific impulses of around 300 to 450 seconds. The better electric propulsion systems, Hall thrusters and gridded ion drives, reach 3,000 to 3,500 seconds. The device tested by Jinxing Zheng and colleagues at the Chinese Academy of Sciences’ Institute of Plasma Physics in Hefei achieved 3,265 seconds. At 12 kilowatts of input power, with argon propellant flowing at just five milligrams per second.

That number on its own isn’t the breakthrough. What makes it significant is the hardware that produced it.

Magnetoplasmadynamic thrusters, or MPDTs, have been around conceptually since the 1960s. They work by running large electrical currents through plasma in the presence of powerful magnetic fields, generating Lorentz forces that accelerate the plasma to very high exhaust velocities. In theory, the specific impulse ceiling for the best MPDTs is around 11,000 seconds — well beyond anything else practically achievable. In practice, the technology has been stuck. The problem, at its core, is the magnet.

A conventional copper electromagnetic coil capable of generating the field strengths useful for plasma acceleration is heavy and power-hungry to a degree that strains credulity. The University of Stuttgart’s SX3 thruster prototype offers a representative example: the thruster itself weighs 13 kilograms, but the copper electromagnet it requires adds another 150 kilograms and consumes 285 kilowatts of power. A large solar array covering 100 square metres provides perhaps 56 kilowatts at Earth orbit under ideal conditions. Powering such a magnet would require at least two such arrays. The mass and space requirements alone make it essentially incompatible with a small satellite.

Superconducting magnets offer a way around this. A superconductor carries current without electrical resistance — no resistance means no Joule heating, which means the magnet consumes trivial power to maintain its field. Earlier attempts used low-temperature superconductors, which work but require cooling to near absolute zero, around 4 kelvin, using liquid helium. That demands a cryogenic system of its own, adding back much of the mass and complexity that superconductivity was supposed to remove.

The YBCO ceramic — yttrium barium copper oxide, a material discovered in the 1980s and still one of the more remarkable substances in condensed matter physics — remains superconducting up to 93 kelvin. In practice that means it can be maintained in its superconducting state using a much simpler cooling system, a small Stirling-cycle refrigerator rather than a cryostat full of liquid helium. In the Hefei device, four cryocoolers brought the magnet down to operating temperature over about seven hours. Once cold, it held there stably.

The result is a complete magnet system — HTS coils, cooling hardware, structural support — that weighs 60 kilograms and draws less than one kilowatt of electrical power during operation. Against the 285-kilowatt, 150-to-220-kilogram copper alternative, that difference is transformative for small satellite design.

The Hefei team wound their magnet from commercially produced YBCO tape, the same material that’s increasingly being used in advanced MRI machines and fusion reactor coils. Four double-pancake coils, each wound with two layers of 160 turns, generate a central magnetic field of 0.2 tesla — enough to usefully confine and accelerate plasma, as the experiments confirmed. An important engineering detail: the system operates with a safety margin of more than 50 per cent below the tape’s critical current, and the stability margin exceeds 1,000 millijoules per cubic centimetre, roughly an order of magnitude higher than comparable low-temperature superconducting systems. The HTS tapes are expensive and sensitive; thermal runaway (quench) would damage them badly. The system is designed to never get close.

Alongside the hardware, the team developed a magnetohydrodynamic model of plasma acceleration inside the thruster’s magnetic nozzle — the diverging field region where plasma expands and exhausts. The model, based on the MHD equations coupling Maxwell’s laws with the Navier-Stokes equations, makes quantitative predictions of how thrust varies with magnetic field strength and mass flow rate. Experimental results from Langmuir probe measurements and direct thrust measurements validated those predictions reasonably well, particularly at low flow rates and higher field strengths. The physics emerging from the analysis identified the “swirl” component of thrust — plasma rotating azimuthally before being redirected axially — as the dominant contribution at the low mass flow rates relevant to small satellites.

What the system hasn’t yet done is fly. The paper describes ground-based testing in a vacuum chamber, and the team frames the 12-kilowatt result as groundwork for eventual in-orbit demonstration. The efficiency at 25 per cent is respectable but not exceptional; earlier superconducting MPDTs operating at much higher power (150 kilowatts) have achieved efficiencies approaching 76 per cent. The Hefei device is optimised for a different constraint: not maximum raw performance, but maximum performance per kilogram of spacecraft bus, at power levels compatible with the solar arrays a small satellite can actually carry.

The missions that would benefit most are the ones currently constrained by propulsion options. CubeSats and small satellites have transformed access to orbit over the past decade, partly because launch costs have fallen and partly because electronics have shrunk. Constellations of hundreds or thousands of small spacecraft are now practical. What’s harder is keeping them in precisely controlled orbits, raising or lowering them efficiently, or sending any of them substantially beyond low Earth orbit. The electric propulsion options available at small satellite power budgets are mostly Hall thrusters and gridded ion systems, which top out at specific impulses around 3,000 to 3,500 seconds. An HTS MPDT delivering 3,265 seconds at 12 kilowatts, in a package light enough to integrate into a small satellite, covers similar ground — while leaving open the theoretical possibility of scaling to 11,000 seconds as the technology matures.

There’s a broader trajectory visible here. The same YBCO materials that make this thruster possible are simultaneously enabling compact superconducting magnets for clinical MRI, high-field laboratory instruments, and the tokamak designs at the core of several fusion energy programmes. The manufacturing base for HTS tape has expanded substantially over the past decade, which means costs have fallen and availability has improved. Aerospace components that required exotic custom fabrication a decade ago can now be assembled from commercial off-the-shelf superconducting wire.

The plasma in the Hefei vacuum chamber doesn’t know any of this, of course. It doesn’t know it’s the first to be accelerated by a fully integrated high-temperature superconducting magnetoplasmadynamic thruster operating below 15 kilowatts. It knows only the magnetic field, the current, the pressure gradient — and then it’s gone, exhausted backward through the nozzle at exhaust velocities a chemical rocket could never match, from a system light enough to fit on a spacecraft small enough to launch for the cost of a used car.

Study link: https://academic.oup.com/nsr/article/13/2/nwaf589/8407247