We perform uniformly sampled large-scale cosmological simulations including magnetic fields with the moving mesh code AREPO. We run two sets of MHD simulations: one including adiabatic gas physics only; the other featuring the fiducial feedback model
of the Illustris simulation. In the adiabatic case, the magnetic field amplification follows the $B propto rho^{2/3}$ scaling derived from `flux-freezing arguments, with the seed field strength providing an overall normalization factor. At high baryon overdensities the amplification is enhanced by shear flows and turbulence. Feedback physics and the inclusion of radiative cooling change this picture dramatically. In haloes, gas collapses to much larger densities and the magnetic field is amplified strongly and to the same maximum intensity irrespective of the initial seed field of which any memory is lost. At lower densities a dependence on the seed field strength and orientation, which in principle can be used to constrain models of cosmic magnetogenesis, is still present. Inside the most massive haloes magnetic fields reach values of $sim 10-100,,{rm mu G}$, in agreement with galaxy cluster observations. The topology of the field is tangled and gives rise to rotation measure signals in reasonable agreement with the observations. However, the rotation measure signal declines too rapidly towards larger radii as compared to observational data.
Accurate numerical solutions of the equations of hydrodynamics play an ever more important role in many fields of astrophysics. In this work, we reinvestigate the accuracy of the moving-mesh code textsc{Arepo} and show how its convergence order can b
e improved for general problems. In particular, we clarify that for certain problems textsc{Arepo} only reaches first-order convergence for its original formulation. This can be rectified by simple modifications we propose to the time integration scheme and the spatial gradient estimates of the code, both improving the accuracy of the code. We demonstrate that the new implementation is indeed second-order accurate under the $L^1$ norm, and in particular substantially improves conservation of angular momentum. Interestingly, whereas these improvements can significantly change the results of smooth test problems, we also find that cosmological simulations of galaxy formation are unaffected, demonstrating that the numerical errors eliminated by the new formulation do not impact these simulations. In contrast, simulations of binary stars followed over a large number of orbital times are strongly affected, as here it is particularly crucial to avoid a long-term build up of errors in angular momentum conservation.
In the version of the SD scenario of SNe Ia studied here, a CO WD explodes close to the Chandrasekhar limit after accreting material from a non-degenerate He companion. In the present study, we employ the Stellar GADGET code to perform 3D hydrodynami
cal simulations of the interaction of the SN Ia ejecta with the He companion taking into account its orbital motion and spin. It is found that only 2%--5% of the initial companion mass are stripped off from the outer layers of He companions due to the SN impact. The dependence of the unbound mass (or the kick velocity) on the orbital separation can be fitted in good approximation by a power law for a given companion model. After the SN impact, the outer layers of a He donor star are significantly enriched with heavy elements from the low-expansion-velocity tail of SN Ia ejecta. The total mass of accumulated SN-ejecta material on the companion surface reaches about > 10e-3 M_sun for different companion models. This enrichment with heavy elements provides a potential way to observationally identify the surviving companion star in SN remnants. Finally, by artificially adjusting the explosion energy of the W7 explosion model, we find that the total accumulation of SN ejecta on the companion surface is also dependent on the explosion energy with a power law relation in good approximation.
