ENhanced Galactic Atmospheres With Arepo: Resolving the CGM at 200 pc with the ENGAWA Simulations

Abstract

Simulating the small-scale features and dynamics of the circumgalactic medium (CGM) is computationally challenging due to its large volume, low densities, multiphase structure, and chaotic environmental effects. Traditional mass-based refinement schemes focus computational power on the high-density regions, thus alternative techniques are required to study the details of the CGM. In this paper, we introduce a new suite of four cosmological zoom-in simulations of Milky Way-like galaxies in which we include fixed-volume refinement throughout the CGM combined with the IllustrisTNG stellar and AGN feedback model down to redshift zero. Reaching spatial resolutions of 200 pc, we see enhancements in low ion column densities (H I and Mg II) and the number of cold clouds around galaxies, relieving some of the longstanding tensions between simulations and observations of the CGM. We additionally apply the COLT radiative transfer code in post-processing to account for stellar radiation, providing a more realistic gauge of ion populations. We find a reduction in the H I with minimal impact to the Mg II and O VI, tempering the impact of resolution while still providing results consistent with observations. In addition to the increase in the number of cold clouds in the CGM, we find that their intermediate temperature boundary regions are reduced in size as the resolution is increased, leading to smoother transitions to the ambient CGM temperature. This paper outlines initial results from this fixed-volume simulation suite which will serve as a basis for future explorations of CGM dynamics, gas accretion, and galaxy evolution.

Figures (19)

• Figure 1: Spatial (top) and mass (bottom) resolution as a function of galactocentric radius at $z=0$. The different colored lines show the median cell size or mass and the shaded regions show the 16%--84% range. The large small size spike at $r\sim200$ kpc corresponds to a satellite galaxy.
• Figure 2: Mean spatial (top) and mass (bottom) resolution as a function of galactocentric radius comparing the 200 pc ENGAWA simulation (red line) against other published simulations at $z=0$. The solid and dot-dashed lines show a cosmological simulation and a standard zoom-in: a galaxy at TNG50-1 resolution, and the m21i galaxy from the FIRE simulations, respectively. The other lines show CGM-refinement simulations: FOGGIE in the dotted line, tempest in the long dashed line, and GIBLE in the dashed line.
• Figure 3: Mass and star formation rate evolution of the Au6 galaxy at four different resolutions. The left panel shows stellar mass, the center panel shows gas mass, and the right panel shows star formation as a function of time from redshift 0.3 when the refinement scheme is activated. In the center panel, the total gas mass is shown by a solid line, and it is also subdivided into disk material (dashed line; gas within $R_{xy}<30$ kpc and $|z|<5$ kpc) and CGM material (dotted line). The gray regions in the left two panels show the $\pm20$% range following the evolution of the default refinement simulation (blue). In the right panel, the star formation rates are shown binned at 100 Myr intervals for all simulations, with additional thin lines shown for the default and 200 pc resolution simulations binned at 10 Myrs.
• Figure 4: Mg2 and O6 column densities across different resolutions for the Au6 galaxy at $z=0$ (using Trident ion fractions). The central image shows a mock Hubble observation of the stellar component of the galaxy at the same scale. The image is 200 kpc across. The top four panels show the Mg2 column density for the default, 1 kpc, 500 pc, and 200 pc resolutions from left to right. The bottom panels show the same for O6.
• Figure 5: Gas density (top) and clumping factor (bottom) as a function of galactocentric radius at $z=0$. These are averaged profiles using the last three snapshots in the simulation covering $\sim100$ Myr. In the top panel, each colored line shows the median densities with the shaded regions indicating the 16%--84% range, and the bottom panel showing the $\langle\rho^2\rangle/\langle\rho\rangle^2$ clumping factor, increasing with refinement.