The path of a frozen pellet through the JET plasma. The pellet enters at the left, and burns up a short distance later. However it induces an instability on the edge of the plasma which continues on the same trajectory through the chamber. The smaller bright spots to the right show this filament of an ELM instability hitting the limiter tiles on the edge of the plasma.

"You don't stand a snowball's chance in hell" goes the saying. Yet Robin Mooney and his team at JET have injected the proverbial snowball – frozen deuterium at -260 degrees Celsius – into a hellish environment, JET's multi-million degree plasma, with very positive results.

There are two reasons for injecting these pellets. Firstly to add fuel to the plasma; the solid form is better able than gas to penetrate into the core of the plasma where the fuel is most needed. Secondly the pellets cause instability in the plasma. Although this sounds like a detrimental effect, it is actually very useful, because it allows some measure of control of the plasma instabilities. Regular, controlled instabilities are much more desirable than large build ups of energy followed by serious disruptions, such as edge localised modes, or ELMs.

This has placed some fairly stringent requirements on the pellet injection system. 99.999 percent pure deuterium is cooled to below -260 degrees Celsius and then extruded and chopped up into pellets of 2.5 to 3.5 cubic millimetres (for stability control) or 40 to 63 cubic millimetres (for fuelling). Then up to fifty pellets per second are sent through around 10 metres of pipe to one of three entry points, entering the chamber at up to 1000 kilometres per hour.

Not surprisingly, a system this complex had some teething problems – the Russian manufacturers had to visit the site a couple of times to get everything running smoothly – but it's all valuable experience, as ITER will have a similar pellet system. The stage is now set at JET for running specific ITER scenarios, which will be tested in May this year.

Source: EFDA