ITER Dodges Trouble With Superconducting Cables


A potential stumbling block that threatened to delay construction of the huge ITER fusion reactor—an international project based at Cadarache in France—looks like it has been resolved. Tests last year on samples of superconductor cable for the facility's magnets indicated the cable would last only one-tenth as long as required. That prompted a scramble to find out what the problem was and identify a new cable configuration that would work. Recent tests at a high-magnetic-field facility in Switzerland showed that engineers had succeeded. "This demonstrates clearly that there is a solution that works," says Neil Mitchell, head of ITER's magnet division.

Keeping the 150 million°C plasma at the heart of the machine in place requires huge and powerful electromagnets made of superconducting cables. The cables that failed last year were made of niobium tin and were destined for the central solenoid—a coil at the very center of the machine that acts to create a current of plasma around the doughnut shaped reactor. The solenoid will require nearly 36 kilometers of superconducting cable and, once complete, will weigh nearly 1000 tonnes.

Samples of the superconductor are tested at the SULTAN facility at the Paul Scherrer Institute in Villigen, Switzerland. The facility subjects the samples to pulses of high magnetic field and electric current, simulating the cycles that it will go through in the completed reactor. Last year's samples, manufactured in Japan, began to degrade after just 6000 cycles, while the specification requires them to last for 60,000 cycles.

The conductors are built up from individual strands less than a millimeter across. Three such strands are wound together to form a "triplet" and 288 triplets bunched together in a metal jacket form a conductor. Investigators studying the failed conductors realized that part of the problem lay in the fact that in the Japanese sample each triplet was made up of two niobium-tin strands and one of copper. This is a protection against damage to the conductor from "quenching," when the niobium-tin material suddenly loses its superconducting ability. Including copper strands means that if there is a quench, the large quantity of current in the conductor has somewhere else to go and so doesn't do any damage. But the Japanese configuration of one copper strand with two of niobium-tin means that, in normal operation, only two strands in each triplet carry current. Those two must shoulder the large magnetic forces that the conductors experience.

A better solution would be three strands made from a combination of copper and niobium-tin, so that all three share the load of magnetic forces. "We looked around the world for what was the best conductor. We built it and it works," says Mitchell. The ITER researchers made new conductor samples using combination strands made by the British company Oxford Instruments and other components from elsewhere in Europe. In tests at SULTAN these endured 10,000 cycles with a level of degradation that was much closer to the original specification.

The combination strands are more expensive to produce and ITER officials are now in discussion with Japan, which is responsible for making the central solenoid superconductor, about how they can replicate them. Although this has delayed the start of conductor manufacture, Mitchell says this shouldn't be a problem because a number of other delays—including last year's earthquake and tsunami in Japan—have already forced ITER managers to push back the scheduled start of the reactor by 1 year to late 2020.

Correction: A conductor is formed by 288 triplets, not 192 as originally reported.