ORIGIN OF DIFFUSE SUPERTHERMAL IONS AT QUASI-PARALLEL SUPERCRITICAL COLLISIONLESS SHOCKS

Large-scale one-dimensional hybrid simulations of quasi-parallel shocks have been performed in order to study the acceleration of upstream energetic ions. In these self-consistent simulations a certain part of the incident ions is accelerated and constitute diffuse upstream particles, which are subject to further scattering in upstream magnetosonic waves of their own making. The number of superthermal particles close to the shock reaches a steady state within less-than-or-similar-to 30 OMEGA-ci-1. The ratio of diffuse upstream particles to solar wind particles decreases slightly with increasing shock Mach number and increases with decreasing angle THETA-Bn between the upstream magnetic field and the shock normal. The acceleration of thermal particles to superthermal energies occurs by a more or less coherent process: thermal ions of the incident distribution stay for an extended time period at the shock. Because of the large noncoplanarity magnetic field component they grad B drift within the coplanarity plane and gain energy by the tangential electric field. They also gain energy due to wave-particle scattering when they stay near the shock and when they finally leave the shock in the upstream direction during a shock re-formation cycle. Since the backstreaming particles excite the upstream waves by an ion/ion beam instability, they feed energy to the wave field. Therefore, in the shock frame a few of the backstreaming ions have an energy below the initial energy. The particle distribution is diffuse in velocity space and exhibits a spherical hole, which is approximately centered at the phase velocity of the upstream waves. This indicates that the particles are pitch angle scattered in the upstream wave field. The results show that superthermal upstream particles are an integral part of quasi-parallel collisionless shocks and that no particular seed particle population is necessary for shock acceleration. The shock structure and the first-order Fermi acceleration problem have therefore to be considered simultaneously.