We present a more accurate numerical scheme for the calculation of diffusive shock acceleration of cosmic rays using Stochastic Differential Equations. The accuracy of this scheme is demonstrated using a simple analytical flow profile that contains a
shock of finite width and a varying diffusivity of the cosmic rays, where the diffusivity decreases across the shock. We compare the results for the slope of the momentum distribution with those obtained from a perturbation analysis valid for finite but small shock width. These calculations show that this scheme, although computationally more expensive, provides a significantly better performance than the Cauchy-Euler type schemes that were proposed earlier in the case where steep gradients in the cosmic ray diffusivity occur. For constant diffusivity the proposed scheme gives similar results as the Cauchy-Euler scheme.
We simulate time-dependent particle acceleration in the blast wave of a young supernova remnant (SNR), using a Monte Carlo approach for the diffusion and acceleration of the particles, coupled to an MHD code. We calculate the distribution function of
the cosmic rays concurrently with the hydrodynamic evolution of the SNR, and compare the results with those obtained using simple steady-state models. The surrounding medium into which the supernova remnant evolves turns out to be of great influence on the maximum energy to which particles are accelerated. In particular, a shock going through a $rho propto r^{-2}$ density profile causes acceleration to typically much higher energies than a shock going through a medium with a homogeneous density profile. We find systematic differences between steady-state analytical models and our time-dependent calculation in terms of spectral slope, maximum energy, and the shape of the cut-off of the particle spectrum at the highest energies. We also find that, provided that the magnetic field at the reverse shock is sufficiently strong to confine particles, cosmic rays can be easily re-accelerated at the reverse shock.
Supernova remnants (SNR) are now widely believed to be a source of cosmic rays (CRs) up to an energy of 1 PeV. The magnetic fields required to accelerate CRs to sufficiently high energies need to be much higher than can result from compression of the
circumstellar medium (CSM) by a factor 4, as is the case in strong shocks. Non-thermal synchrotron maps of these regions indicate that indeed the magnetic field is much stronger, and for young SNRs has a dominant radial component while for old SNRs it is mainly toroidal. How these magnetic fields get enhanced, or why the field orientation is mainly radial for young remnants, is not yet fully understood. We use an adaptive mesh refinement MHD code, AMRVAC, to simulate the evolution of supernova remnants and to see if we can reproduce a mainly radial magnetic field in early stages of evolution. We follow the evolution of the SNR with three different configurations of the initial magnetic field in the CSM: an initially mainly toroidal field, a turbulent magnetic field, and a field parallel to the symmetry axis. Although for the latter two topologies a significant radial field component arises at the contact discontinuity due to the Rayleigh-Taylor instability, no radial component can be seen out to the forward shock. Ideal MHD appears not sufficient to explain observations. Possibly a higher compression ratio and additional turbulence due to dominant presence of CRs can help us to better reproduce the observations in future studies.
