Metal flows of the circumgalactic medium, and the metal budget in galactic haloes

Abstract

We present an analysis of the flow of metals through the circumgalactic medium (CGM) in the Feedback in Realistic Environments (FIRE) simulations of galaxy formation, ranging from isolated dwarfs to L * galaxies. We find that nearly all metals produced in high-redshift galaxies are carried out in winds that reach 0.25R vir . When measured at 0.25R vir the metallicity of outflows is slightly higher than the interstellar medium (ISM) metallicity. Many metals thus reside in the CGM. Cooling and recycling from this reservoir determine the metal budget in the ISM. The outflowing metal flux decreases by a factor of ∼ 2 − 5 between 0.25R vir and R vir . Furthermore, outflow metallicity is typically lower at R vir owing to dilution of the remaining outflow by metal-poor material swept up from the CGM. The inflow metallicity at R vir is generally low, but outflow and inflow metallicities are similar in the inner halo. At low redshift, massive galaxies no longer generate outflows that reach the CGM, causing a divergence in CGM and ISM metallicity. Dwarf galaxies continue to generate outflows, although they preferentially retain metal ejecta. In all but the least massive galaxy considered, a majority of the metals are within the halo at z = 0. We measure the fraction of metals in CGM, ISM and stars, and quantify the thermal state of CGM metals in each halo. The total amount of metals in the low-redshift CGM of two simulated L * galaxies is consistent with estimates from the COS halos survey, while for the other two it appears to be lower.

Galactic outflow rates in the EAGLE simulations

“…As with the comparison to TNG-50, larger differences become apparent when considering the change in the mass loading as a function of radius. Muratov et al (2017) present measurements of the mass loading in FIRE at the virial radius (magenta crosses), which can be compared to measurements at a quarter of the virial radius (magenta triangles, or the dashed line) from Muratov et al (2015). In most cases the mass loading is smaller at larger radii in FIRE, whereas the opposite is true in EAGLE at low redshift.…”

“…For EAGLE, the dashed-dotted blue line shows the same but for a shell at the virial radius. For FIRE, individual galaxies are shown by the magenta points for shells at different radii, as labelled (Muratov et al 2015(Muratov et al , 2017 significantly higher mass and spatial resolution (with the improvement scaling negatively with the mass of the targeted haloes), allow a cold ISM phase to form without imposing a temperature floor, and implement a more explicit representation of stellar feedback (separating contributions from radiation, stellar winds, and type II supernova explosions). The FIRE simulations do not include AGN feedback.…”

Tentative detection of the circumgalactic medium of the isolated low-mass dwarf galaxy WLM

“…In particular, for galaxies with M * 10 7 M , the outflowing metals are mostly trapped in the CGM probably due to the weakening of star-formation activities (see figure 3 in Christensen et al 2018). Muratov et al (2017) show that for galaxies at M * ∼ 10 7−7.6 M , 50% of all the metals (by mass) reside in the CGM, more than 90% of which is in cool phase with T ∼ 10 4.0−4.7 K (see also Ma et al 2016). Results from different simulations vary quantitatively, probably due to different treatment of feedback recipes (see Hu et al 2017;Emerick et al 2018Emerick et al , 2019 for a detailed work on the impact of high-resolution stellar feedback modeling in regulating the efficiency of outflows).…”

Low-redshift Lyman limit systems as diagnostics of cosmological inflows and outflows

“…These simulations have been shown to correctly reproduce the star formation histories of galaxies below ∼ L * (Hopkins et al 2014), mass-metallicity relations ) at all redshifts where observations are available, and LLS covering fractions in the halos of z ∼ 2 galaxies . In these simulations, galaxy-scale outflows are selfconsistently generated from the local injection of feedback momentum and energy on small scales Muratov et al 2016). These successes of the FIRE simulations make them particularly well suited to address the physical nature of LLSs.…”