Hydrogenic retention with high-Z plasma facing surfaces in Alcator C-Mod

The retention of deuterium (D) fuel in the Alcator C-Mod tokamak is studied using a new 'static' gas balance method. C-Mod solely employs high-Z molybdenum (Mo) and tungsten (W) for its plasma facing materials, with intermittent application of thin boron (B) films. The primarily Mo surfaces are found to retain large fractions, similar to 20-50%, of the D-2 gas fuelled per quiescent discharge, regardless of whether the Mo surfaces are cleaned of, or partially covered by, B films. Several experiments and calculations show that it is improbable that B retains significant fractions of the fuel. Rather, retention occurs in Mo and W surfaces through ion bombardment, implantation and diffusion to trap sites. Roughly 1% D of the incident ion fluence, Phi(D), to surfaces is retained, and with no indication of the retention rate decreasing after 25 s of integrated plasma exposure. The magnitude of retention is significantly larger than that extrapolated from the results of laboratory studies for either Mo or W. The high levels of D/Mo in the near surface, measured directly post-campaign (similar to 0.01) in tiles and inferred from gas balance, are consistent with trapping sites for fuel retention in the Mo being created, or expanded, by high D atom densities in the near surface which arise as a result of high incident ion fluxes. Differences between C-Mod and laboratory retention results may be due to such factors as the multiply ionized B ions incident on the surface directly creating traps, the condition of Mo (impurities, annealing) and the high-flux densities in the C-Mod divertor which are similar to ITER, but 10-100x those used in laboratory studies. Disruptions produce rapid heating of the surfaces, releasing trapped hydrogenic species into the vessel for recovery. The measurements of the large amount of gas released in disruptions are consistent with the analysis of tiles removed from the vessel post-campaign-the campaign-integrated retention is very low, of order 1000x less than that observed in a single, non-disruptive discharge.