Scientific advisers to the ITER fusion reactor project have recommended several key changes to its design that could increase technical risks—but also smooth the path to producing excess energy. The recommendations, made last week by ITER’s Science and Technology Advisory Committee (STAC), will have to be approved by the full ITER council in November. But if approved, as expected, “the chance of surprises later is reduced,” says Alberto Loarte, head of ITER’s confinement and modeling section. “The risk will pay off.”
ITER, being built in France by an international collaboration, aims to show that nuclear fusion, the reaction that powers the sun, can be controlled on earth to produce energy. But reaching that goal involves heating hydrogen gas to more than 150 million°C so that hydrogen nuclei slam together with enough force to fuse. To do this, researchers are building a huge doughnut-shaped container called a tokamak to confine the ionized gas—or plasma—using enormously strong magnetic fields. ITER’s goal is to coax the plasma to produce 500 megawatts (MW) of heat, 10 times the 50 MW of power required to heat the plasma; this multiplying effect is known as a gain of 10.
The most significant change decided at the STAC meeting concerns a structure at the base of the tokamak vessel called the divertor. Its main function is to remove the helium that is the “exhaust” gas of the fusion reaction. The divertor is the only part of the vessel where the superhot plasma actually touches a solid surface, so it has to be able to absorb huge quantities of heat, as much as 10 MW per square meter of surface.
Existing plans call for making ITER’s first divertor with an outer layer of carbon. This is the safe option: Carbon is well proven in tokamak interiors; it can easily withstand the temperatures; and if any is blasted off into the plasma, it doesn’t affect the performance very much. The problem with carbon, however, is that it happily reacts with hydrogen, binding atoms into its structure. This wouldn’t be a problem during the early phases of ITER operation when researchers plan to use simple hydrogen or helium in the machine to get the hang of how it works. But a carbon coating could be a huge problem in later phases, when researchers plan to switch to real fusion fuel—a more reactive mixture of the hydrogen isotopes deuterium and tritium. Tritium is radioactive and so needs to be carefully controlled and accounted for. Nuclear regulators would never accept a divertor material that absorbs tritium and so makes it impossible to locate.
To address that problem, planners had proposed running ITER for several years with the carbon-coated divertor, and then switching to one made of tungsten. Tungsten has the highest melting point of any metal: 3422°C. That should be fine for withstanding the heat produced during normal, steady ITER operations. But any unexpected bursts of heat could potentially melt the divertor, and tungsten—unlike carbon—instantly poisons the plasma, bringing fusion to a halt. So ITER’s operators would have to run the reactor much more carefully with a tungsten divertor, not pushing it to limits where the plasma might become unstable.
Despite this drawback of tungsten, STAC has recommended that ITER be built with a tungsten divertor from the start. “It was not an easy decision,” says STAC Chair Joaquín Sánchez, head of Spain’s National Fusion Laboratory in Madrid. The decision was made after years of research at other tokamak laboratories, in particular the Joint European Torus (JET) at Culham in the United Kingdom, which is the closest machine to ITER in size and design. Several years ago, JET researchers refitted the reactor with a tungsten divertor and beryllium lining (as ITER will have). After a year of testing, they confirmed that this “ITER-like wall” worked well enough not to cause problems for ITER.
Although some fusion researchers think that it would be safer to start ITER with a well understood carbon divertor, allowing them to push the reactor to extremes in search of high performance, starting with tungsten has advantages, too. Changing divertors is a complex process that would take many months. In addition, once operation with deuterium-tritium fuel has started, the interior of the vessel becomes radioactive (or “activated”), making it much harder to modify internal components. “If we start with tungsten, we save the cost of the change,” Sánchez says. “We know tungsten will be more difficult, but we will start learning earlier in the nonactivated phase and if there is a problem we can send people inside to fix it.”
The other design changes concern two separate magnetic coils to be inserted inside the reactor vessel to fine-tune control of the plasma. ITER’s main plasma-confining magnets are outside the vessel and act as something of a blunt instrument. About 5 years ago, researchers highlighted the fact that operators would have difficulty keeping the vertical position of the plasma steady, and so proposed some extra magnetic coils on the inside.
In addition to those for vertical stability, researchers proposed installing a second set of internal coils to combat a troubling phenomenon in superhot fusion plasma called edge-localized modes, or ELMs. ELMs occur when energy builds up in the plasma during fusion and then bursts out of the edge unpredictably, potentially damaging the lining or the divertor. The second set of coils deploys a magnetic field to roughen up the surface of the plasma so that it leaks energy at a constant rate rather than in erratic bursts.
Anything inside the vessel is subjected to extreme heat, radioactivity, and magnetic forces, so researchers had to persuade STAC that these two sets of coils could be made resilient enough to survive. “There was some reluctance in STAC and the ITER Organization because of the technical issues of installation,” Loarte says. Experiments at other labs around the world reassured them. “The results obtained were very positive,” he says.
STAC also took a hard look at the delivery schedule of components for ITER. The original plan called for everything—heating systems, instruments, ELM mitigation—to be in place when ITER is completed in 2020. But delays have meant that some items will be arriving later. “We needed to redo the schedule with a logic consistent with [achieving deuterium-tritium operation] faster. It was not consistent before and that led to criticism,” Loarte says. “Now we have to do the organizational part, which is not simple.”