A numerical scheme that incorporates a thermal leakage injection model into a combined gas dynamics and cosmic ray (CR, hereafter) diffusion-convection code has been developed. The particle injection is followed numerically by filtering the diffusive flux of suprathermal particles across the shock to the upstream region according to a velocity-dependent transparency function that controls the fraction of leaking particles. We have studied CR injection and acceleration efficiencies during the evolution of CR modified planar shocks for a wide range of initial shock Mach numbers, $M_0$, assuming a Bohm-like diffusion coefficient. The injection process is very efficient when the subshock is strong, leading to fast and significant modification of the shock structure. As the CR pressure increases, the subshock weakens and the injection rate decreases accordingly, so that the subshock does not disappear. Although some fraction of the particles injected early in the evolution continue to be accelerated to ever higher energies, the postshock CR pressure reaches an approximate time-asymptotic value due to a balance between fresh injection/acceleration and advection/diffusion of the CR particles away from the shock. We conclude that the injection rates in strong parallel shocks are sufficient to lead to rapid nonlinear modifications to the shock structures and that self-consistent injection and time-dependent simulations are crucial to understanding the non-linear evolution of CR modified shocks.

One of the key questions to understanding the efficiency of diffusive shock acceleration of the cosmic rays (CRs) is the injection process from thermal particles. A self-consistent injection model based on the interactions of the suprathermal particles with self-generated magneto-hydrodynamic waves has been developed recently by Malkov (1998, Phys. Rev. E 58, 4911). By adopting this analytic solution, a numerical treatment of the plasma-physical injection model at a strong quasi-parallel shock has been devised and incorporated into the combined gas dynamics and the CR diffusion-convection code. In order to investigate self-consistently the injection and acceleration efficiencies, we have applied this code to the CR modified shocks of both high and low Mach numbers (M=30 and M=2.24) with a Bohm type diffusion model. We find the injection process is self-regulated in such a way that the injection rate reaches and stays at a nearly stable value after quick initial adjustment. For both shocks about 0.1% of the incoming thermal particles are injected into the CRs. For the weak shock, the shock has reached a steady state within our integration time and about 10% of the total available shock energy is transfered into the CR energy density. The strong shock has achieved a higher acceleration efficiency of about 20% by the end of our simulation, but has not yet reached a steady-state.

Numerical results for particle acceleration at multiple oblique shocks are presented. We calculate the steady state spectral slope of test particles accelerated by the first order Fermi process. The results are compared to analytical treatments, for parameters, where the diffusion approximation does apply. Effects of injection and finite shock extend are included phenomenologically. We find the spectrum of accelerated particles to harden substantially at multiple oblique shocks and discuss the influence of the number of shocks compared to the obliquity itself.

Supernova remnants are expected to contain braided (or stochastic) magnetic fields, which are in some regions directed mainly perpendicular to the shock normal. For particle acceleration due to repeated shock crossings, the transport in the direction of the shock normal is crucial. The mean squared deviation along the shock normal is then proportional to the square root of the time. This kind of anomalous transport is called sub-diffusion. We use a Monte-Carlo method to examine this non-Markovian transport and the acceleration. As a result of this simulation we are able to examine the propagator, density and pitch-angle distribution of accelerated particles, and especially the spectral properties. These are in broad agreement with analytic predictions for both the sub-diffusive and the diffusive regimes, but the steepening of the spectrum predicted when changing from diffusive to sub-diffusive transport is found to be even more pronounced than predicted.

The theory of diffusive particle acceleration explains the spectral properties of the cosmic rays below energies of approx. 10^6 GeV as produced at strong shocks in supernova remnants (SNR's). To supply the observed flux of cosmic rays, a significant fraction of the energy released by a supernova has to be transfered to cosmic rays. The key to the question of the efficiency of SNR's in producing cosmic rays is the injection process from thermal energies. A self-consistent model has to take into account the interaction of the accelerated particles with magneto-hydrodynamic waves, which generate the particle diffusion, a requisite for the acceleration process. Such a nonlinear model of the turbulent background plasma has been developed recently (Malkov, 1998, Phys. Rev. E 58, 4911). We use this model for the first numerical treatment of the gas dynamics and the diffusion-convection equation at a quasi-parallel strong shock, which incorporates a plasma-physical injection model to investigate the cosmic ray production.

In the theory of diffusive acceleration at oblique shock fronts the question of the existence of a discontinuity of energetic particle density is contentious. The resolution of this problem is interesting from a theoretical point of view, and potentially for the interpretation of observations of particle densities at heliospheric shocks and of high-resolution radio observations of the rims of supernova remnants. It can be shown analytically that an isotropic particle distribution at a shock front implies continuity of the particle density -- whether or not the shock is oblique. However, if the obliquity of the shock induces an anisotropy, a jump is permitted. Both semi-analytic computations and Monte-Carlo simulations are used to show that, for interesting parameter ranges, a jump is indeed produced, with accelerated particles concentrated in a precursor ahead of the shock front.

We present a semi-analytic approach to solving the Boltzmann equation describing the comptonisation of low frequency input photons by a thermal distribution of electrons in the Thomson limit. Our work is based on the formulation of the problem by Titarchuk & Lyubarskij (ApJ 450 (1995) 876), but extends their treatment by accommodating an arbitrary anisotropy of the source function. To achieve this, we expand the eigenfunctions of the integro/differential eigenvalue problem defining the spectral index of comptonised radiation in terms of Legendre polynomials and Chebyshev polynomials. The resulting algebraic eigenvalue problem is then solved by numerical means, yielding the spectral index and the full angular and spatial dependence of the specific intensity of radiation. For a thin (\tau_0 < 1) plasma disk, the radiation is strongly collimated along the disk surface - for an optical thickness of \tau_0 = 0.05, the radiation intensity along the surface is roughly ten times that along the direction of the normal, and varies only slightly with the electron temperature. Our results for the spectral index confirm those of Titarchuk & Lyubarskij over a wide range of electron temperature and optical depth; the largest difference we find is roughly 10% and occurs at low optical depth.