MIT Spin-off Faces Daunting Challenges on Path to Build a Fusion Power Plant in 15 Years

IEEE Spectrum By Tracy Staedter | 18 Apr 2018

Commonwealth Fusion Systems has pledged to build a commercial fusion reactor based on new superconducting magnets

Image: Ken Filar, PSFC Research Affiliate

Fusion power is always two or three decades away. Dozens of experimental reactors have come and gone over the years, inching the field forward in some regard, but still falling short of their ultimate goal: producing cheap, abundant energy by fusing hydrogen nuclei together in a self-sustained fashion.

Now an MIT spin-off wants to use a new kind of high-temperature superconducting magnet to speed up development of a practical fusion reactor. The announcement, by Commonwealth Fusion Systems, based in Cambridge, Mass., caused quite a stir. CFS said it will collaborate with MIT to bring a fusion power plant online within 15 years—a timeline faster by decades than other fusion projects.

CFS, which recently received an investment of US $50 million from Eni, one of Europe’s largest energy companies, says the goal is to build a commercial fusion reactor with a capacity of 200 MWe. That’s a modest output compared to conventional fission power plants—a typical pressurized water reactor, or PWR, can produce upwards of 1,000 MWe—but CFS claims that smaller plants are more competitive than giant, costly ones in today’s energy market.

It’s certain that, between now and 2033, when CFS expects to have its reactor ready for commercialization, the company will face a host of challenges. These revolve around key milestones that include: fabricating and testing the new class of superconducting magnets, and using them to build an experimental reactor, which CFS named SPARC; figuring out how to run SPARC so that fusion reactions inside the machine can produce excess energy in a continuous manner, one of the biggest challenges in any fusion reactor; and finally, scaling up the experimental design into a larger, industrial fusion plant.

Each of these steps embodies numerous scientific and engineering quandaries that may have never been seen before or have already confounded some of the smartest physicists and nuclear engineers in the world. Can CFS and MIT finally harness fusion power? Maybe. In 15 years? Probably not.

“Fusion research remains fusion research,” says Robert Rosner, a professor of physics at the University of Chicago and the former director of Argonne National Laboratory. “It’s a field where getting to a practical, energy-generating reactor is not an engineering issue, but a basic science issue.”

“Fusion research remains fusion research.” —Robert Rosner, University of Chicago

Most experimental fusion reactors are based on a Russian design called a tokamak. These machines employ a powerful magnetic field to confine a cloud of hot ionized gas, or plasma, in the shape of a donut. This creates the extreme temperatures—in excess of 100 million degrees Celsius— for hydrogen nuclei to speed around and collide, fusing into heavier elements, like helium. The process releases vast amounts of energy. (Fusion is what powers stars like our sun, with their mighty gravity squeezing the hydrogen nuclei into helium.)

CFS and MIT plan to build a tokamak with technology never before employed in fusion. It will generate a magnetic field using a relatively new high-temperature superconducting material made from steel tape that’s coated in a compound called yttrium-barium-copper oxide, or YBCO. The advantage of using this material is that it can produce intense magnetic fields from a much smaller machine than those at other facilities.

CFS estimates that SPARC will be about one-fourth the size (and 1/65 the volume) of the 23,000-metric ton machine called ITER, the world’s largest experimental tokamak, currently under construction in France. Yet SPARC’s magnet will generate a maximum magnetic field of 22 teslas, nearly double that of ITER’s 12-T magnetic field.

Although MIT has pioneered research in tokamak magnetics and has persisted in exploring the high magnetic field approach to fusion, nobody has made superconducting magnets of that size and strength from YBCO, says Tim Luce, head of operations and science at ITER. “There are a lot of technological challenges associated with that,” he says.

“We think that the MIT projection of 15 years to a power plant is very ambitious, if not overly ambitious.” —Tim Luce, ITER

MIT expects it will take three years to design, fabricate, and test the magnets. For comparison, ITER’s magnets, which consist of 18 units made from niobium-tin and niobium-titanium, are still being built, with final assembly scheduled for 2022 (the ITER project began in 2007).

There’s also the question of fuel. The sun with its intense gravity and pressure is able to produce fusion using ordinary hydrogen. But hydrogen gas doesn’t work well in a fusion reactor because the nuclei do not collide reliably.

To improve the chances of fusion, plasma physicists prefer two gases that are isotopes of hydrogen: deuterium, which is abundant in seawater, and tritium, a form rarely found in nature because it naturally decays with a half-life of about 12 years. A deuterium-tritium mixture, called D-T, has the greatest potential in the near-term for a sustainable fusion reaction that lasts more than a few minutes. But using that mixture has a downside: It produces large amounts of free neutrons, whose lack of an electrical charge allows them to escape the tokamak’s magnetic field. This stream of neutrons reacts with the nuclei of metals in the containment vessel to form new isotopes that can produce harmful radiation or make the vessel material brittle and vulnerable to cracks.

“Any tokamak must run for years to optimize the plasma before daring to use tritium,” says Daniel Jassby, who was a principal research physicist at the Princeton Plasma Physics Laboratory until 1999.

Tokamak designers who have used D-T fuel—or plan to use it—have come up with creative solutions to deal with the neutrons. ITER engineers, for instance, are designing a water-cooled steel structure about 1-meter thick that will line the inside of the machine. Both the Tokamak Fusion Test Reactor, which the Princeton Plasma Physics Lab operated from 1982 to 1997, and the Joint European Torus, operating at Culham Centre for Fusion Energy in Oxfordshire, U.K., simply surrounded the entire machine in a thick concrete shield.

CFS and MIT want to develop a molten salt blanket that will surround the plasma and behave as a kind of neutron-absorbing lining. Although circulating molten salt has been used in fission nuclear reactors, no one has ever developed such a technology for use inside in a tokamak.

In an email to IEEE Spectrum, Robert Mumgaard, CEO of CFS, writes that this collaboration is different than others dominated by government funding with a focus on basic research. In this partnership, MIT will carry out the basic and applied research and CFS will work to commercialize it.

“By involving private industry focused on delivering a working product, the project and company will be able to grow and accelerate upon success, bringing more human and monetary resources to bear,” he says.

“We think that the MIT projection of 15 years to a power plant is very ambitious, if not overly ambitious,” says Luce, of ITER. “But we will celebrate any success, and we share the dream of making energy from fusion.”