Studying divertor relevant plasmas in the Pilot-PSI linear plasma device: experiments versus modelling

Predictions for the operation of tokamak divertors are reliant on edge plasma simulations typically consisting of a fluid plasma code in combination with a Monte-Carlo (MC) code for neutral species. Pilot-PSI is a linear device operating with a cascaded arc plasma source that produces plasmas comparable to those expected during the inter-ELM phase in the ITER divertor (T e ∼ 1 eV, n e ∼ 1020 m-3). In this study, plasma discharges in Pilot-PSI have been modelled using the Soledge2D fluid plasma code (Bufferand et al 2015 Nucl. Fusion 55) coupled to the Eirene neutral MC code (Reiter et al 2005 Fusion Sci. Technol. 47, 172-186) in order to (a) investigate which phenomena need to be included in the modelling to reproduce experimental trends and (b) provide new insights to the interpretation of experiments. The simulations highlight the key role of ion/molecule elastic collisions in determining the ion flux reaching the target. Recombination is likely to play a role at high molecular background pressure. However, even with the most advanced atomic and molecular model used in this work, T e at the target is overestimated with respect to the measurements using TS and spectroscopy. T e in the simulations appears to saturate at 0.7 eV for a wide range of parameters, while experimentally values of 0.1-0.3 eV are found. As a consequence, in the simulations the volume recombination is underestimated, which is a strong function of T e when it is below 1 eV. Further analysis of simulation results using a two-point formalism shows that inelastic collisions between electrons and neutral background particles remove most of the energy flux, mainly via dissociation of molecules and molecular ions. However this happens mostly in the upstream region of the beam where T e > 1 eV. For T e < 1 eV, there seems to be no significant energy removal mechanism in the simulated cases. The results also indicate that conclusions on the importance of volume processes, e.g. recombination, cannot be solely based on T e or the dominance of certain reaction rate coefficients over others, but rather the complete transport picture, including macroscopic flow, has to be taken into account. In the cases studied here, the plasma is typically advected to the wall too fast for recombination to remove a significant fraction of the particle flux. © 2018 DIFFER - Dutch Institute for Fundamental Energy Research