The article seems heavy on PR trying to convince the reader of the important contributions the US programs at PPPL and MIT are making to fusion plasma research (presumably to bolster their requests for dwindling DOE funding) , but does little to explain how with this new ‘discovery’ we will be able to design better tokamaks.
In a quick skim read through the PRL paper http://prl.aps.org/abstract/PRL/v108/i16/e165004 they take a simple cylindrical model with a chain of small magnetic islands around it, and work out the power balance from the ohmic heating within the island, the external heating flowing around it and the radiation emitted via bremsstralung & line emission from impurities. They find the scaling is roughly in line with the Greenwald limit on plasma density, which is proportional to the total plasma current.
They go on to mention the relationship between neoclassical tearing modes briefly at the end,
Additionally, radiation driven islands should be exacerbated in plasmas with high noninductive current fractions, since only the Ohmic current participates in heating
the interior of the island. This may explain the common practice of using ‘‘preventative electron cyclotron resonant heating’’ to avoid the onset of neoclassical tearing modes.
In fact, this phenomenon may partially explain the difficulty in finding a reliable predictor for the onset of neoclassical tearing modes because the radiation driven terms
are not considered in neoclassical island threshold analysis
But they completely fail to mention turbulence, which as far as I’m aware, is the primary way heat leaks into and out of island structures, as small eddies in the plasma flow on the scale of a few times the ion larmor radius (few mm) cause small perturbations in the electric & magnetic fields and let ions & electrons on the outside to collide with those ‘trapped’ on the inside & vice-versa.
The paper does suggest an explanation for several long standing problems, such as how the density limit changes when switching from a deuterium plasma to a helium one, and should be testable with some of the advanced diagnostics being fitted to several tokamaks.