The Lopsided Exhaust That Could Make or Break Fusion Power

Somewhere inside a doughnut-shaped machine in California, a stream of superheated plasma is hurtling along at roughly 88 kilometres per second. It is also, in a sense, falling apart. Particles at the edge of the magnetic field that holds the plasma in place are constantly escaping, streaming downward toward metal plates designed to catch them, cool them off and send them back into the fray. The returning atoms help keep the fusion reaction going. But here’s the thing: they don’t land where you’d expect.

For years, physicists running tokamak experiments have noticed a stubborn imbalance. Far more escaping plasma particles slam into the inner exhaust target than the outer one. The pattern showed up again and again, across different machines and different conditions, and nobody could quite account for it.

That matters more than it might sound. If we’re ever going to build fusion reactors that run for decades (and that is, after all, sort of the point), engineers need to know precisely where those exhaust particles will strike. Get it wrong and you end up with metal plates eroding unevenly, hotspots forming in the wrong places, components failing years before they should. The exhaust system of a tokamak, called the divertor, has to be engineered with real confidence in the physics. And the physics, until recently, wasn’t cooperating.

The leading explanation had centred on what are known as cross-field drifts, the sideways motion of charged particles across magnetic field lines inside the divertor region. Sensible enough. But when researchers built computer simulations that included only this effect, the numbers didn’t match what experiments were showing. The lopsidedness in the models was never quite lopsided enough.

Now a team led by Eric Emdee, an associate research physicist at Princeton Plasma Physics Laboratory, reckons they’ve found the missing ingredient. Using the modelling code SOLPS-ITER, Emdee and colleagues simulated plasma behaviour in the DIII-D tokamak under four different scenarios, toggling cross-field drifts and plasma rotation on and off in various combinations. The results, published in Physical Review Letters, are perhaps best described as a lesson in how two mediocre explanations can combine into one very good one.

“There are two components to flow in a plasma,” says Emdee. “There’s cross-field flow, where particles drift sideways across the magnetic field lines, and parallel flow, where they travel along those lines.” The conventional wisdom had plumped for cross-field flow as the culprit behind the asymmetry. “A lot of people said cross-field flow was what created the asymmetry,” he says. “What this paper shows is that parallel flow, driven by the rotating core, matters just as much.”

On its own, adding rotation to the simulations shifted things a bit. Ditto for cross-field drifts alone. Neither could reproduce the experimental measurements. But when Emdee’s team combined both effects, feeding in the measured core rotation speed of 88.4 kilometres per second, the simulations finally snapped into agreement with what physicists had been observing for years in the actual machine. The combined influence proved considerably greater than either component working in isolation, a kind of synergy between the two flow mechanisms that changes how momentum gets transported across the plasma edge and, ultimately, how many particles end up at each divertor target.

It’s a neat result, and a reassuring one. Fusion engineers can’t exactly build a prototype divertor, run it for 20 years and see what happens. They need simulations they can trust.

What Emdee and his colleagues have shown is that the existing boundary plasma models can get there, that they can reproduce the stubborn asymmetries physicists have been scratching their heads over, provided you account for both rotation and drifts working together. That “provided” is doing quite a lot of heavy lifting.

The work involved researchers from PPPL, MIT and North Carolina State University, and was carried out using the DIII-D National Fusion Facility. It won’t, on its own, solve the engineering challenges of building a commercial fusion reactor. But it does chip away at one of those nagging unknowns that could, left unresolved, quietly undermine the whole endeavour.

For the engineers tasked with designing exhaust systems that might need to survive decades of bombardment by runaway plasma, knowing where the particles actually land isn’t a minor detail. It’s perhaps the detail.

Study link: https://journals.aps.org/prl/abstract/10.1103/zjpv-vxwd