RF discharge mirror cleaning for ITER optical diagnostics using 60 MHz very high frequency

Abstract

• First mirror cleaning using 60 MHz very high frequency was optimized. • The pre-matching element is reducing the power consumption. • Cleaning homogeneity on a curved and large size mirror was measured. • End-of-Cleaning Indicator using Optical Emission Spectroscopy was demonstrated. For the fusion reactor ITER, a mandatory monitor of the fusion device and plasma will be performed with optical diagnostic systems. For the metallic first mirrors, the recovery of the reflectivity losses due to dust deposition is proposed to be carried out for 14 different optical diagnostic systems by the plasma cleaning technique. In this work, we studied the influence of the electrode area on the electrode potential as a function of the applied power with a 60 MHz radio very high frequency source. Unshielded copper disks with different diameters were constructed to study the impact of the electrode area in the range of 90 cm 2 to 1200 cm 2 , which corresponds to an Edge Thomson Scattering area ratio of 0.15 to 2. It was observed that the absolute value of the resulting bias decreased from 280 V to 15 V with the increase of the area for a given RF power. Moreover, the power consumption was reduced by 43 % using a pre-matching element close to the vacuum feedthrough. The cleaning homogeneity on a curved and a large size mirror was studied, and the difference between the center and edge maximum/minimum was around 20 % and 40 % for the curved and large size mirror, respectively. For ITER, it is required to have an End-of-Cleaning Indicator (ECI), which shows when the cleaning process would be stopped. In this work, we studied the feasibility of Optical Emission Spectroscopy (OES) as a real-time control tool for the RF cleaning process. With 13.56 or 60 MHz, it was possible to use OES as ECI by following the molybdenum (Mo), rhodium (Rh), and aluminum (Al) emission lines. The decrease of Al I line and increase of the Mo I and Rh I line were recorded as a function of the cleaning time.