Posted on: 26th August 2013

Among the entities that inhabit a tokamak, one of the most intriguing and paradoxical is the Edge-Localised Mode, or ELM. Unlike their namesake, the ELM tree (Ulmus family), they are not very serene, instead being shortlived and very energetic bursts of energy on the edge of the plasma. They only appear when a tokamak is in its most efficient mode of operation – high confinement, or H-mode – threatening to spray all the plasma’s precious energy away in a shower of hot particles; yet at the same time they provide regular flushing of impurities, serving to keep the plasma clean and well confined.

Paradoxically the best way to control ELMs is to make more of them. They seem to be part and parcel of the high confinement scenario in the fusion plasma, in which the edge of the plasma develops a barrier layer around it which allows the density to rise – a positive thing for fusion, which depends on collisions between these particles. This barrier is only centimetres thick, and can maintain a temperature gradient of around ten million degrees per centimetre. But at a certain point the density becomes too high, and the barrier breaks down, ejecting particles and energy, which ultimately end up where the plasma touches the wall in JET, in the lower area known as the divertor. Then the barrier forms again, and the density rises again repeating the pattern up to fifty times per second.

The longer the build up, the bigger the crash, hence the approach to try to trigger ELMs before they get too large – sometimes termed “burping the baby”, a pressure release that every parent knows well. In future fusion reactors large uncontrolled ELMs could inflict heat loads onto the divertor tiles that no known material could withstand; controlling ELMs is crucial to fusion’s success.

ELMs can be triggered by injecting material into the plasma to disturb the layer, although in such finely balanced, turbulent situations it can be hard to predict what will work. A simple puff of deuterium gas has an effect. Serendipitously, so does the injection of frozen deuterium pellets. Initially attemped as a way to bypass JET’s magnetic fields – which function to stop particles escaping from the plasma, but also prevent the reverse process too – pellet injection is a method of delivering fuel into the core of the plasma. When it was employed it was found to be successful at pacing ELMs too.

Another approach to dissipating the pressure build up is by using magnetic perturbation. Creating ripples in the magnetic fields give the particles a way to leak out in a controlled fashion through the barrier. JET has tested this using its error field correction coils – coils that were installed for a completely different purpose but could be adapted for this purpose, albeit only at low current. Similar successes have also been seen in ASDEX-U in IPP Garching, and at CCFE’s MAST in UK – but a fine balance has to be struck – too much leakage and the energy confinement drops too far.

In addition, the ELMs on a larger machine such as ITER are a bit of an unknown, says the Responsible Officer for ELM Control, Dr Chris Lowry. “Extrapolating the size of ELMs in ITER has a number of uncertainties. They could be as high as 30 megajoules per ELM, which is equivalent to a grenade going off – and that happening 20 times a second. Or they could be small, and no problem at all.” He continues confidently “I’m sure we’ll work it out one way or another, though!”.