Dynamic Wall Loads Measured by Gas Balance Technique in all Tungsten ASDEX Upgrade

Full text: In vessel tritium inventory is still a safety issue for ITER. The amount and mechanism of hydrogen retention depends on the plasma facing component (PFC) material. During the last years AUG demonstrated that the long term deposition of D is strongly reduced during the transition from carbon to full tungsten PFCs. However deposition measurements yield no time resolved information, which is desired to verify predictions to ITER. In contrast time resolved gas balance techniques allow to identify the mechanism relevant for the retention. According to the gas balance measurements, high density H-mode discharges can be divided into different phases. First, after plasma start up, the PFCs strongly retain D, until the wall is saturated. Then, for all W PFC device, a steady state phase is reached, during which wall pumping and release are equal within the error bars. Outgassing starts immediately when the gas puffing rate is reduced or the discharge ends. Obviously, the in-vessel D inventory is dominated by dynamic wall loads. For W PFCs the long term retention in AUG is below a few percent of the amount of gas puffed. During the 2009 campaign AUG was operated again with pure tungsten PFC, i.e. boron layers had been completely removed during the vent. During the 2008 campaign the amount of gas, which is temporally retained before reaching wall saturation, was 2x1022 at for the unboronized and 1.5 x 1022 at for the boronized wall. In 2009 only 1.2 - 1.5 x 1022 at are needed, even if the wall was not boronized. To increase further the accuracy of the measurements discharges were performed using the cryo-pump only, as proposed by Alcator C-mod. In AUG this procedure only allows operation with low gas puffing levels. High medium term retention rates of 25 to 50% of the amount of gas puffed were observed in these kind of experiments, comparable to Alcator C-Mod results. But the absolute amount of retained gas agrees well with the value observed for high density discharges. Therefore the earlier apparent discrepancy of the C-mod and AUG results seems to be due to the different amount of gas puffed during the discharges. The total amount of the dynamically retained gas and its temporal behavior will be compared to the ion and neutral fluxes to the wall. (author)