Neoclassical ripple transport in tokamaks

The usual ripple transport calculations lead to a nu-1 scaling of the transport coefficients with collision frequency nu. The paper extends and clarifies this scaling by taking into account the fact that the dominant contributions to transport come from particles in the high-energy tail of the distribution function. This means that particles with energy E approximately (4-6)T dominate, and this restricts the nu-1 scaling range to nu(T)>(E/T)sup(5/2)deltaωsub(d)(T) approximately 100 delta ω sub(d)(T), where T is the thermal energy, delta is the ripple-well depth, and ωsub(d)-1 is the time to drift over the region where the localized ripple well exists. In addition, transport coefficients are derived in weak-ripple (α is identical to epsilon/Nqdelta>>1, with epsilon the inverse aspect ratio, N the number of toroidal coils, q the safety factor), low-collisionality regimes where the radial step size is determined by the distance a particle drifts before it 'collisionlessly detraps' from a ripple well. The maximum transport rate is found to be about two orders of magnitude smaller than usually assumed, primarily because of the restriction of the nu-1 scaling regime to the much higher collision frequencies. (author)