Improved understanding of the Cs dynamics in large H− sources by combining TDLAS measurements and modeling

Abstract

The ITER Neutral Beam Injection (NBI) relies on sources of negative hydrogen and deuterium ions which deliver a homogeneous and temporally stable extracted negative hydrogen ion beam for up to one hour (57 A extracted D− current from an extraction area of 2000 cm2), at which the co-extracted electron current has to be lower than the extracted negative ion current. The ELISE (Extraction from a Large Ion Source Experiment) test facility is equipped with a 1/2 ITER-NBI scale H− source. Negative hydrogen ions are mainly created by conversion of hydrogen atoms on a surface with low work function. In order to reach the ITER requirements, a stable and homogeneous coverage of caesium has to be maintained on the plasma grid, which is the first grid of the extraction system. The dynamics of the highly-reactive Cs in the source is complex due to the non-ultra-high-vacuum conditions (10−7 mbar), the plasma-enabled redistribution of Cs and the continuous evaporation of Cs from two ovens.The Monte-Carlo test particle transport code CsFlow3D has been developed in order to simulate fluxes of Cs and Cs+ as well as the Cs coverage on the surfaces of the ion source during vacuum and plasma phases of the pulsed-driven sources. A successfully applied benchmark of the code at the small prototype ion source (1/8 ITER source size) allows now for giving predictions also for the larger ion sources. With the newly installed Tunable Diode Laser Absorption Spectroscopy (TDLAS) diagnostic at the Cs 852 nm resonance line the neutral Cs density averaged along two horizontal lines-of-sight close to the plasma grid is determined. No significant vertical asymmetry of neutral Cs is found. As a new feature, the temperature of Cs is evaluated from the Doppler broadening, resulting in 350 ± 50 K during vacuum phase and 960 ± 100 K during plasma phases. Both, measurements and simulations, reveal the high relevance of back-streaming positive ions created in the extraction system, leading to an additional source term of Cs. The simulation clearly shows a different regime in the Cs dynamics between continuous extraction and pulsed extraction, which is of particular interest since due to the available HV power supply ELISE is presently only able to extract a beam for 10 s each three minutes in long (up to one hour) plasma pulses.