Impurity transport in tokamaks

A review is given of the principle results of the neoclassical theory of impurity transport in tokamaks, emphasizing those collisionality regimes of most practical interest. Some new results are given for the case where there are several different impurity species present at the same time, with both hydrogenic and impurity ions in the Pfirsch--Schlueter regime. The rate at which an impuritiy ion is transported into a plasma, and the radiation from it, are strongly dependent on the mean level of ionization (Z). An atomic physics code package, based on an ''average ion'' model is used to assess the validity of the ''coronal equilibrium'' assumption commonly made to determine (Z); coronal equilibrium is found to be a poor approximation for heavy metallic impurities in small present-day tokamaks, but its validity is much improved in larger devices, except perhaps in the edge region of the plasma. The computations are compared with the results obtained in an impurity injection experiment in the ATC tokamak. In practice, the most appropriate impurity transport model would be a combination of neoclassical transport and anomalous diffusion, with the magnitude of the anomalous diffusion coefficient chosen to give agreement with observed hydrogenic particle confinement times. Computations of oxygen impurities in the edge region of a tokamak plasma, on the basis of this model, show that only a fraction (typically 0.2 to 0.6) of an incident flux of neutral impurities actually penetrates into the main part of the discharge