No abstract

One main concern about the use of graphite as a plasma facing material is the enhanced erosion under hydrogen bombardment due to hydrocarbon formation. In view of the lifetime evaluation of plasma exposed carbon components and of impurity production in present and future machines such as ITER, an analytical expression for the erosion yield by chemical sputtering for the relevant energies, temperatures and incident fluxes is of special importance. An extrapolation to fluxes and energies relevant for high density divertor plasmas has not been possible up to now on the basis of semiempirical fits to laboratory data. Starting from a short review of the existing empirical formulas, recent detailed investigations of the atomistic processes for the thermally activated hydrocarbon emission are described, which enable the formulation of an improved analytical description including the ion flux as a parameter. The chemical erosion of graphite by hydrogen bombardment results from two processes: the thermally activated hydrocarbon emission, K h e r m , and a surface process at low energies and low temperatures resulting from the kinetic ejection of surface hydrocarbon complexes from collisional energy transfer, Yaurf. The new analytic description can be fitted well to the existing data for ion beam erosion, and extrapolation to divertor relevant fluxes is possible. At high ion fluxes the maximum of chemical erosion is shifted to higher temperatures, where annealing of damaged structures leads to a stronger reduction of &erm than previously estimated. There are no data on a possible flux dependence of Ysurf, leaving still some uncertainty in extrapolation.

Several experiments were conducted in ASDEX Upgrade to prove the suitability of tungsten as a divertor target material under the conditions of a high density and low temperature divertor. The observed fluxes from a tungsten tile into the plasma are low, in keeping with the extremely low sputtering yields. In addition, the very favourable effect of `prompt redeposition' (redeposition during the first gyration) could be confirmed by the experiments. Cooling of the edge region by neon injection seems permissible, i.e. neon impurity sputtering did not increase the eroded fluxes of tungsten. The transport and accumulation behaviour were investigated by means of the laser blow-off technique. No accumulation effects could be observed in ohmic discharges. In discharges with NBI heating but without ICRH, strong accumulation can occur. High heat flux tests were performed on graphite tiles coated with plasma sprayed tungsten, which withstood a thermal load of 15 MW/m2 lasting 2 s as well as 1000 cycles of 10 MW/m2 for 2 s without disabling damage. Owing to the encouraging results, an experiment using a tungsten divertor is planned in ASDEX Upgrade

No abstract