Neutral beam injectors are based on the neutralization of ion beams accelerated at the desired energy. In the case of the ITER heating and diagnostic neutral beams, the target heating power translates into stringent requirements on the acceptable beamlet divergence and aiming to allow the beam to reach the fusion plasma. The beamlets composing the accelerated beam are experimentally found to feature a transverse velocity distribution exhibiting two Gaussian components: the well-focused one is referred to as the core component while the rest of the beam, the halo, describes beam particles with much worse optics. The codes that simulate beam extraction and acceleration usually assume that the negative ions move towards the plasma meniscus with a laminar flow (no transverse velocity) or that the transverse velocity distribution can be modelled as a Maxwellian and that the current density is uniformly illuminating the meniscus; under such approximations, the presence of highly divergent components cannot be explained. In this work, we develop a simple test-particle tracing code with Monte Carlo collisions, named ICARO (for Ions Coming Around), to study the transport of negative ions in the extraction region and derive the spatial and velocity distribution of the negative ions at the meniscus (i.e. the plasma boundary where a beamlet is extracted). In particular, the origin of the beamlet halo and its dependence on the source parameters are discussed, highlighting as a key parameter the energy distribution of positive ions in the source plasma.
Fast ion confinement during high power tangential neutral beam injection into low plasma current discharges on the ISX-B tokamak