Effective cosmic ray diffusion in multiphase galactic environments

TL;DR

CR transport in a multiphase ISM arises from the competition between CR streaming, Alfvén-wave growth, and damping. The authors implement a two-moment CRMHD model within the Crisp framework in the Arepo code to self-consistently compute diffusion coefficients and transport speeds from microphysical processes. They run two isolated Milky Way–mass galaxies, one with ion-neutral damping enabled and one without, to isolate the damping effects. They find that the effective diffusion coefficient κ_eff converges to the canonical range of $1\times10^{28}$–$1\times10^{29}$ cm$^2$ s$^{-1}$ in the disk and inner CGM, and that ion-neutral damping enhances CR transport in the dense ISM by reducing Alfvén-wave scattering, while maintaining relatively slow transport in the hot wind. This has important implications for CR-driven winds and galaxy evolution, showing that a multiphase ISM with accurate damping physics is essential for realistic CR feedback.

Abstract

Cosmic-ray (CR) feedback is widely recognized as a key regulator of galaxy formation. After being accelerated at supernova remnant shocks, CRs propagate through the interstellar medium (ISM), establishing smooth large-scale distributions and driving galactic outflows. The efficiency of this feedback is controlled by the effective transport speed of the CR population, which in turn depends on the competition between CR-driven plasma instabilities and wave damping processes that vary strongly with ISM phase. In cold, dense gas, ion-neutral damping dominates, whereas in warm, diffuse environments, weaker non-linear Landau damping prevails, leading to enhanced CR scattering and slower transport. To investigate these effects, we employ the moving-mesh code Arepo and model CR transport using a two-moment description within the multiphase ISM framework Crisp, which self-consistently computes CR diffusion coefficients and transport velocities from coarse-grained plasma physics. The intrinsic CR diffusion coefficient depends inversely on the scattering rate of CRs and Alfvén waves, covering 15 orders of magnitude. In contrast, we show that the effective CR diffusion coefficient, which quantifies the propagation speed of CRs through the ISM, converges toward the canonical range of $10^{28}$-$10^{29}$ cm$^2$ s$^{-1}$. Simulations with only non-linear Landau damping yield transport rates up to an order of magnitude slower than those including both Landau and ion-neutral damping. Overall, CR transport speeds increase systematically with gas density, for which we provide a density-dependent fit of the effective CR diffusion coefficient. We demonstrate that, despite strong ion-neutral damping in the cold and warm phases of the galactic disk, CRs are transported at speeds only a few times the local Alfvén speed as they traverse alternating ISM phases on their way out of the galaxy.

Figures (11)

• Figure 1: Gallery showing edge-on slices through the galactic disk and wind (top row) and face-one slices through the galactic midplane (bottom row). Gas related quantities (density $\rho$ and temperature $T$) can be found on the left-hand side of each column, while CR related quantities are on the right-hand side: CR pressure $P_\mathrm{cr}$ and CR-to-thermal pressure ratio $X_\mathrm{cr} = P_\mathrm{cr} / P_\mathrm{th}$. The ISM and the CGM of the galaxy are highly structured and multiphase owning to stellar feedback. The CR population is relatively smooth and exhibits pressures exceeding the thermal pressure within the galactic wind.
• Figure 2: Gallery similar to Fig. \ref{['fig:gallery_1']} but showing the vertical gas velocity $\varv_{z}$ and the projected vertical CR transport speed along magnetic field lines $\varv_{\mathrm{cr}, z}$ (left panels) and the Alfvén wave energy density $\varepsilon_\mathrm{a}$ and the CR diffusion coefficient $\kappa_\mathrm{cr}$ (right panels). Because CRs are advected with the gas, the total vertical CR transport speed is given by the sum of the gas and CR velocities. Both the Alfvén wave energy density $\varepsilon_\mathrm{a}$ and the CR diffusion coefficient $\kappa_\mathrm{cr}$ are highly structured, highlighting Alfvén-wave dark regions in the galactic outflow and an absence of Alfvén waves in the ISM with the exception of hot (super-) bubble regions.
• Figure 3: Star formation rates for the two simulated CR-transport models. Both galaxies exhibit similar SFRs, with comparable starburst periods at the start of the simulations, followed by a brief phase of minor quenching.
• Figure 4: Gallery similar to Fig. \ref{['fig:gallery_2']} but showing the CR diffusion coefficient $\kappa_\mathrm{cr}$ for both the NLLD (left half) and the NLLD+IND (right half) simulation as slices through the galactic plane. The inclusion of ion–neutral damping in the NLLD+IND simulation reduces the abundance of CR-scattering Alfvén waves in the ISM, resulting in a correspondingly higher diffusion coefficient.
• Figure 5: CR-phase diagram relating CR pressure $P_\mathrm{cr}$ and gas density $\rho$ for both the NLLD+IND (left panel) and the NLLD simulation (middle panel) displayed as a mass-weighted PDF taking into account data of the last 100 Myr of evolution from the star-forming disks and galactic outflows. In the right panel, we show the median as a solid line and 20$^\mathrm{th}$ and 80$^\mathrm{th}$ percentiles as shaded regions. Both simulations show different relationships between $\rho$ and $P_\mathrm{cr}$ in dilute and dense media with a prominent transition region at around $\rho \sim 10^{-2}~m_\mathrm{p}~\mathrm{cm}^{-3}$. At high densities, the NLLD run shows systematically increased CR pressures because of the slow CR diffusion in this medium which leads to long CR escape times.