The Tokamak Disaster or Why the Magnetic-Confinement Reactors Known as 'Tokamaks' Failed soComprehensively to Confine Anything
by
L C Woods
Balliol College, Oxford

Controlled thermonuclear energy, using strong magnetic fields threading a torus to confine very hot plasma, was a dream forty years ago. This was based on an optimistic theory that showed that the diffusivity of mass and energy across magnetic fields in cylindrical geometry varied inversely as the square of the magnetic field strength. Confinements times of minutes for plasmas at ignition temperatures were predicted, but early experiments did not confirm this, the errors being some four orders of magnitude! But perhaps bigger machines would behave according to theory. Accordingly they were built at great expense without any theoretical understanding of what was going on in the torus. By 1970 the enterprise was under full international steam, with the Russian tokamak widely adopted as the best hope. A complicated theory that allowed for toroidal geometry, was developed and hailed as a great advance when it brought theory and observation to within a factor of about 400 of each other. Of course the 'discrepancy' had to be due to turbulence and the theoreticians toiled away, delving into a host of candidate instabilities-probably one each. The fact that stellarators, toroidal machines carrying relatively little current, seem rather better at confining their energy was a clue overlooked. Heat losses from current-carrying plasmas in strong magnetic fields are in fact dominated by terms second-order in the Knudsen number, i.e. the thermal conductivity is not a constant but depends on the shear in the electron fluid velocity. If the shear is in the right sense, heat can even flow up temperature gradients. The talk will explain this interesting phenomenon, which also has an important role in corona plasma loops.

Date received: May 3, 1998