paper

Why are thermally- and cosmic ray-driven galactic winds fundamentally different?

arXiv:2405.13121 · doi:10.1051/0004-6361/202450817

Abstract

Galactic outflows influence the evolution of galaxies not only by expelling gas from their disks but also by injecting energy into the circumgalactic medium (CGM). This alters or even prevents the inflow of fresh gas onto the disk and thus reduces the star formation rate. Supernovae (SNe) are the engines of galactic winds as they release thermal and kinetic energy into the interstellar medium (ISM). Cosmic rays (CRs) are accelerated at the shocks of SN remnants and only constitute a small fraction of the overall SN energy budget. However, their long live-times allow them to act far away from the original injection site and thereby to participate in the galactic wind launching process. Using high-resolution simulations of an isolated Milky Way-type galaxy with the moving-mesh code Arepo and the new multi-phase ISM model Crisp (Cosmic Rays and InterStellar Physics), we investigate how SNe and CRs launch galactic outflows and how the inclusion of CR-mediated feedback boosts the energy and mass entrained in the galactic wind. We find that the majority of thermal SN energy and momentum is used for stirring turbulence either directly or indirectly by causing fountain flows, thereby self-regulating the ISM and not for efficiently driving outflows to large heights. A simulation without CRs only launches a weak galactic outflow at uniformly high temperatures and low densities by means of the thermal pressure gradient. By contrast, most of the CR energy accelerated at SN remnants () escapes the ISM and moves into the CGM. In the inner CGM, CRs dominate the overall pressure and are able to accelerate a large mass fraction in a galactic wind. This wind is turbulent and multi-phase with cold cloudlets embedded in dilute gas at intermediate temperatures ( K) and the CGM shows enhanced OVI and CIV absorption in comparison to a simulation without CRs.

arXiv comment16 pages, 9 figures, submitted to A&A, comments welcome

References in corpus (106)

Cited by in corpus (9)