On suitable experiments for demonstrating the feasibility of the beam-driven plasma neutraliser for neutral beam injectors for fusion reactors

Abstract

Plasma neutralisers promise increased neutralisation efficiency of negative ion beams in neutral beam injection (NBI) beamlines compared with gas neutralisers. It has been suggested that, in the presence of an electron-confining magnetic cusp field along all neutraliser walls, the beam itself could ionise the neutraliser gas sufficiently to take advantage of this effect, avoiding the added complexity of external power coupling to the neutraliser. These predictions come from a zero-dimensional model by Surrey and Holmes [1] and Turner and Holmes [2]. We have revisited and modified this model by introducing slowing-down energy distributions for stripped and Rudd electrons, including electron impact dissociation as an electron energy loss channel and taking into account dissociative recombination of molecular ions with electrons. Including the latter effect reduces the predicted plasma density by about a factor of four and the achievable neutralisation yield from ∼ 80 % to 68 % in the case of a negative deuterium ion beam with an energy of 1 MeV and a current of 40 A. With this revised model we estimate the expected performance of potential beam-driven plasma neutralisers (BDPN) on a variety of existing negative ion beam test facilities for neutral beam injection. Based on these results, we conclude that the most suitable proof-of-principle experiment would be a dedicated chamber, ideally of the same dimensions and with the same magnetic cusp configuration as a BDPN for the DEMO NBI, in which the plasma is not created primarily by the fast electrons stripped from the beam ions, but by electrons of similar current and energy emitted from biased filaments.

Functional optimization for a Beam Driven Plasma Neutralizer in DEMO Neutral Beam Injector

“…• For every allowed point in the discretized parameter space from above the plasma density and neutral gas temperature in the BDPN is calculated using a parametrized scaling obtained with Starnella's model [7]. This calculation is carried out for various gas flows into the neutralizer.…”

“…In order to obtain the ionization degree and neutral gas temperature, the BDPN model of Starnella et al [7] would have to be run for every point in the parameter space exploration described above. Starnella's code is a zero-dimensional balance code that models the various populations of ions and electrons with their numerous interactions and loss channels (such as wall losses, impact dissociation, dissociative recombination, etc.)…”

“…For an ITER-like NBI ion beam they predicted an achievable neutralization efficiency of up to ≈ 80 %. The model was further amended by Starnella et al [7], who introduced several previously overlooked loss mechanisms, which led to a reduction of the predicted achievable neutralization to ≈ 68 %. This latest model calculates also temperature of the neutralizer gas besides the degree of ionization and neutralization efficiency.…”

Ion source developments at IPP: On the road towards achieving the ITER-NBI targets and preparing concepts for DEMO

“…A neutralization efficiency of about 75% can be reached in comparison to the 55% achieved by a gas neutralizer, relaxing thus the demands on the accelerated ion current at same beam power or increasing the beam power from one beam line. Recent investigations on the efficiency of the BDPN revealed that there is a need to benchmark the simulations by experimental investigations [23]. Uncertainties concern the plasma confinement in the magnetic cusp field, the achievable electron density and thus ionization degree being together with the gas temperature the key parameters to be benchmarked.…”

“…Uncertainties concern the plasma confinement in the magnetic cusp field, the achievable electron density and thus ionization degree being together with the gas temperature the key parameters to be benchmarked. Studies showed that present day experimental facilities have either not the required beam power to ionize the gas sufficiently or are not available yet [23]. Therefore, a proof-of-principle setup is under consideration in which the stripped electrons of 42 A are mimicked by electrons generated by a filament at the required energy of 272 keV.…”