Read our COVID-19 research and news.

SPARC could be the first fusion reactor to produce net energy—10 years before ITER and in a machine 10 times smaller.

CFS/MIT; T. Henderson

Fusion startup plans reactor with small but powerful superconducting magnets

A startup chasing the dream of plentiful, safe, carbon-free electricity from fusion, the energy source of the Sun, has settled on a site, timetable, and key technology for building its compact reactor. Flush with more than $200 million from investors, including Bill Gates’s Breakthrough Energy, 3-year-old Commonwealth Fusion Systems announced today that later this year it will start to build its first test reactor, dubbed SPARC, in a new facility in Devens, Massachusetts, not far from its current base in Cambridge. The company says the reactor, which would be the first in the world to produce more energy than is needed to run the reaction, could fire up as soon as 2025.

Commonwealth and a rival U.K. company have also chosen the technology they think will let them leap ahead of the giant, publicly funded ITER reactor under construction in France and ever further ahead of a U.S. pilot plant being considered by the Department of Energy: small but powerful magnets, made from high-temperature superconductors. Commonwealth is assembling its first nearly full-scale magnet and hopes to test it in June. “It’s a big deal,” CEO Bob Mumgaard says. “It’s beyond what everyone else aspires to.”

Fusion reactors burn an ionized gas of hydrogen isotopes at more than 100 million degrees Celsius—so hot that the plasma must be contained by a mesh of magnetic fields so it doesn’t melt the reactor walls. At ITER, sufficiently powerful fields are achieved using niobium alloy superconducting wires that can carry huge currents without resistance through magnet coils. But such low-temperature superconductors must be chilled to 4° above absolute zero, which requires bulky and expensive liquid helium cooling. And there’s a limit to the amount of current the niobium wires can carry, forcing ITER to adopt huge magnets with many wire turns to generate the needed fields. ITER’s largest magnets are 24 meters across, contributing to the reactor’s $20 billion price tag. 

Newer high-temperature superconductors—so called because they can superconduct at relatively balmy liquid nitrogen temperatures above 77 kelvins—were not around when ITER was designed. But they can carry much higher currents, tantalizing fusion designers with the prospect of smaller, cheaper reactors. Yet they are brittle, persnickety materials, so “a lot of people had given up on them,” says Rod Bateman of Tokamak Energy, the U.K. startup that is also betting on the technology. “They were just too unreliable.”

In the past decade, researchers have developed ways to deposit thin layers of superconducting rare-earth barium copper oxide (ReBCO) on metal tape. The tapes can be manufactured reliably in long lengths, and perform best at about 10 K. But in terms of low-temperature engineering, “10 K is a lot easier than 4 K,” says magnet engineer John Smith of General Atomics in San Diego.

The ReBCO tapes can be bent but, being flat, are challenging to wind into coils, Mumgaard says. “You have to stop treating it like a wire and asking it to do the things that wire does.” Commonwealth has developed a cable with stacked layers of tape twisting like candy cane stripes. The company believes the cables can carry enough current to generate a 20-tesla field—1.5 times stronger than ITER’s—in magnet coils just a few meters across. Tokamak Energy takes a simpler, more compact approach: winding coils with the tape flat, one layer on top of another like a roll of Scotch tape. “It makes winding so much simpler,” Bateman says.

Another challenge for both companies is supply. Together, manufacturers of ReBCO tape were only producing a few hundred kilometers per year, and Commonwealth needs 500 kilometers just to build its first test magnet. “Manufacturers are scaling up like crazy now,” Bateman says. “Fusion is the market high-temperature superconductors have been waiting for.”

The next few months are critical for the two companies. Following years of modeling and experiments, they are both constructing test magnets to demonstrate the 20-tesla fields they need for a compact device. Commonwealth is in the process of winding a single 2.5-meter-tall, D-shaped magnet, slightly smaller than what’s planned for SPARC. Still, Mumgaard says, when completed in June it will be the largest high-temperature superconducting magnet ever built.

Tokamak Energy is hoping for a similar field strength in its test of a full set of 16 coils, sized for a test reactor about 1 meter across. Bateman says the company will start to wind its magnets in the next few weeks and hopes to test them by the end of the year. If successful, the company will embark on building a demo reactor, ST-F1, planned for 2027.

Fusion scientists are accustomed to seeing paper designs come and go. But if the high-temperature superconducting magnets can actually achieve 20 tesla, Smith says, “It will be a phenomenal statement on the technology.”