Can we make the DEMO NBI more efficient with a beam-driven plasma neutraliser?

The energy efficiency of the neutral beam injectors (NBI) on ITER is currently limited by the gas neutralisation. In order to make the cost of the electricity coming from fusion competitive, the NBI energy efficiency of future magnetic-confinement fusion machines, such as DEMO or a fusion power plant, must be increased and the neutraliser is the component where the highest improvement is possible. The beam-driven plasma neutraliser (BDPN) is one of the technologies that have been considered to overcome the limitations of the gas neutraliser: if the neutraliser gas is sufficiently ionised, collisions with ions and electrons enhance the stripping of the negative beam ions, thus increasing the neutralisation yield. The plasma inside the chamber is created through the gas ionisation by the negative ion beam itself and confined by means of a magnetic cusp field generated with permanent magnets, avoiding the complexity of having an external power source. The original zero-dimensional model of Surrey and Holmes, describing the underlying physics of the BDPN, has been revisited and modified, in particular by introducing the computation of the plasma species composition which allows to consider the plasma loss due to electron-ion dissociative recombination. A neutralisation yield of 75 % is predicted with respect to the value of ~ 55 % foreseen for the ITER NBI’s gas neutraliser. Such a prediction has to be experimentally benchmarked, but at the moment no NBI beamline exists whose parameters are in the relevant range for DEMO. After discussing the possibility of testing the BDPN by installing one on an existing NBI beamline, I will argue that the most suitable proof-of-principle experiment is a dedicated chamber in which the plasma is not created by the fast electrons stripped from the beam ions, but by electrons emitted from biased filaments with the same energy and current.