Why fusion magnets use REBCO coated conductor

Why do compact tokamaks need stronger magnets?

Fusion power depends steeply on the confining field. Zohm, writing in the Journal of Fusion Energy, states that simple scalings give fusion power proportional to the fourth power of the field and the cube of the major radius, so a designer can shrink the machine by raising the field [1]. He presents that as physics scaling under stated assumptions, and notes that plasma stability sits at the upper end of present experiments [1].

The same ratio drives the structure. The SPARC team writes in IEEE Transactions on Applied Superconductivity that magnetic stress scales as the square of the field while fusion performance scales as the fourth power, so the structure must carry a growing load as the field rises [3]. In 2021, MIT News described the MIT and CFS team's projected fusion-device performance as equivalent to an apparatus 40 times larger in volume using low-temperature superconducting magnets [2].

The MIT and CFS authors describe in IEEE Transactions on Applied Superconductivity a SPARC design roughly 30 times smaller in volume than ITER [4]. Zohm adds a caution: exhaust through a divertor can be harder in a compact device [1].

Why use REBCO coated conductor instead of Nb3Sn?

Niobium-based superconductors set the field ceiling for earlier fusion magnets. The program paper in IEEE Transactions on Applied Superconductivity puts practical ceiling fields at about 9 tesla for NbTi and 13 tesla for Nb3Sn [4]. Zohm notes in the Journal of Fusion Energy that ITER's Nb3Sn toroidal field conductor reaches about 12 tesla at its inner leg [1]. Zohm writes in the Journal of Fusion Energy that REBCO could reach 24 tesla there, though that would need an exceedingly large support structure with present technology [1].

Temperature separates the families as sharply as field. MIT News reports that earlier fusion magnets ran at about 4 kelvin, and that REBCO allows operation at 20 kelvin [7]. The cryogenic paper in IEEE Transactions on Applied Superconductivity states the model coil was to be cooled with supercritical helium at a nominal 20 kelvin operating temperature [6].

REBCO arrives as tape. The design paper in IEEE Transactions on Applied Superconductivity describes tape 4 millimeters wide and about 60 micrometers thick [3]. The same paper in IEEE Transactions on Applied Superconductivity lists a Hastelloy C276 substrate, buffer layers, the REBCO film, a silver coating, and copper stabilizer [3]. Coated conductor is the name for that layered tape.

What did the SPARC model coil demonstrate?

On September 5, 2021, a large high-temperature superconducting magnet at the MIT Plasma Science and Fusion Center reached 20 tesla, which MIT News called the most powerful of its kind [2]. The program paper in IEEE Transactions on Applied Superconductivity reports 20.1 tesla peak field on conductor at 40.5 kiloamperes of terminal current [4].

The IEEE Transactions on Applied Superconductivity program paper records 10,058 kilograms of magnet, 270 kilometers of REBCO tape, 815 kilonewtons per meter of Lorentz load, and almost 1 gigapascal of stress in the structural case [4]. The companion design paper in IEEE Transactions on Applied Superconductivity describes a winding pack of sixteen single pancakes between two termination plates, each conductor groove carrying a graded stack of tape [3]. The program paper in IEEE Transactions on Applied Superconductivity records joints that ran at 0.5 to 2.0 nano-ohm at 20 kelvin, and a cryocooler loop delivering 600 watts at 20 kelvin [4].

How do no-insulation REBCO coils carry current?

The coil carried no insulation between turns. MIT News reports that the tape was left bare, with the team relying on the superconductor's high conductivity to keep current flowing [7]. The program paper in IEEE Transactions on Applied Superconductivity calls the configuration NINT, for no-insulation no-twist, and the design paper describes it as non-insulated and stack-in-plate [3][4].

Dropping the insulation changes how a fault behaves. A conventional winding needs insulation because off-normal events such as a shutdown generate high voltages, and removing it makes the magnet a low-voltage system that is simpler to build and leaves room for more cooling or structure [7]. The design paper in IEEE Transactions on Applied Superconductivity reports that the maximum voltage difference in the magnet, which occurs during a quench, was predicted and confirmed by experiment to stay at or below 1 volt [3].

The program paper in IEEE Transactions on Applied Superconductivity notes that single-tape wound no-insulation coils hold the world record at 45.5 tesla [4].

Why does quench protection still matter?

A quench is a local loss of superconductivity, and a large coil must dissipate its stored magnetic energy without damage. The program paper in IEEE Transactions on Applied Superconductivity reports an intentional open-circuit quench at 31.5 kiloamperes in a coil deliberately not optimized for quench resilience, and the coil suffered localized thermal damage [4].

MIT News reports that the final test ended with the melting of one corner of one of the 16 pancakes [7]. The design paper in IEEE Transactions on Applied Superconductivity ties that to geometry, since inner radii of 20 centimeters concentrate flux to exceed 20 tesla [3]. The program paper in IEEE Transactions on Applied Superconductivity records the epicenter at Pancake #12 [4].

The program paper in IEEE Transactions on Applied Superconductivity states that turn-to-turn and pancake-to-pancake quench cascades over about three seconds were clearly observed [4]. The program paper in IEEE Transactions on Applied Superconductivity calls the data and validated models a critical step toward quench resilience [4].

What makes REBCO tape supply an engineering problem?

A fusion coil consumes tape by the kilometer. The design paper in IEEE Transactions on Applied Superconductivity describes continuous critical current measurements at 77 kelvin in self-field for every meter of tape, taken before and after solder plating [3].

Tape acceptance tracked the field and orientation the coil would see. The same IEEE Transactions on Applied Superconductivity paper reports a target engineering critical current density of 750 amperes per square millimeter at 20 tesla and 20 kelvin [3]. The design paper in IEEE Transactions on Applied Superconductivity illustrates one shipment whose average exceeded the target with the field parallel to the tape c-axis [3]. The same IEEE Transactions on Applied Superconductivity paper reports grooves carrying 170 to 240 graded tapes [3].

Procurement volume set the schedule. MIT News reports that 300 kilometers of high-temperature superconductor were procured under strict quality control in under a year [7]. The MIT and CFS authors report in IEEE Transactions on Applied Superconductivity that their average tape procurement cost per meter fell by almost 40 percent [4]. In the same IEEE Transactions on Applied Superconductivity paper, the authors projected SPARC procurement approaching 10 million meters [4].

How does ITER's niobium-tin system compare?

ITER takes the low-temperature route at large scale. The ITER Organization describes 18 D-shaped Nb3Sn toroidal field coils designed to produce a maximum field of 11.8 tesla and store 41 gigajoules of magnetic energy [5]. The ITER Organization specifies its Nb3Sn central solenoid for a maximum field of 13 tesla and a plasma current of 15 megaamperes sustained for 300 to 500 seconds [5].

The ITER Organization puts the poloidal field system at a maximum field of 6 tesla, and states that the four largest of those coils were produced from NbTi [5]. The ITER Organization states that the whole magnet system is cooled with supercritical helium in the range of 4 kelvin [5].

The strand bill shows the scale. The ITER Organization reports 500 metric tons of Nb3Sn strand, more than 100,000 kilometers of it, from nine suppliers between 2008 and 2015 [5]. A REBCO magnet moves a similar qualification burden onto tape makers, since each meter is measured before it is wound [3].