Fuel retention in impurity seeded discharges in JET after Be evaporation

Preparatory experiments for the ITER-Like Wall in JET were carried out to simulate the massive Be first wall by a thin Be layer, induced by evaporation of about 2.0 g Be, and to study its impact on fuel retention and divertor radiation with reduced C content and N seeding. Residual gas analysis reveals a reduction of hydrocarbons by one order of magnitude and of O by a factor of 5 in the partial pressure owing to the evaporation. The evolution of wall conditions, impurity fluxes and divertor radiation have been studied in ELMy H-mode plasmas (B-t = 2.7T, I-p = 2.5 MA, P-aux = 16MW) whereas a non-seeded reference discharge was executed prior to the evaporation. The in situ measured Be flux at the midplane increased by about a factor of 40 whereas the C flux decreased by similar to 50% in the limiter phase of the first discharge with respect to the reference, but erosion of the Be layer and partial coverage with C takes place quickly. To make best use of the protective Be layer, only the first four discharges were employed for a gas balance analysis providing a D retention rate of 1.94 x 10(21) Ds(-1) which is comparable to rates with C walls. But the Be evaporation provides a non-saturated surface with respect to D and short term retention is not negligible in the balance; the measured retention is overestimated with respect to steady-state conditions like that of the ILW. Moreover, C was only moderately reduced and co-deposition of fuel with eroded Be and C occurs. The lower C content leads to a minor reduction in divertor radiation as the reference phase prior to seeding indicates. N adds to the radiation of D and the remaining C, and the N content rises due to the legacy effect which has been quantified by gas balance to be 30% of the injected N. C radiation increases with exposure time, and both contributors cause an increase in the radiated fraction in the divertor from 50% to 70%. The radiation pattern suggests that N dominates the increase in the first discharges though C is still the dominating radiator. Therefore, the validity of a proxy of the Be first wall by a thin Be layer is limited and restricted to plasma operation directly after the Be evaporation.