Ne VIII in the warm-hot circumgalactic medium of FIRE simulations and in observations

Abstract

The properties of warm-hot gas around $\sim L_{*}$ galaxies can be studied with absorption lines from highly ionized metals. We predict Ne VIII column densities from cosmological zoom-in simulations of halos with masses in $\sim 10^{12}$ and $\sim 10^{13}\,\mathrm{M}_{\odot}$ from the FIRE project. Ne VIII traces the volume-filling, virial-temperature gas in $\sim 10^{12}\,\mathrm{M}_{\odot}$ halos. In $\sim 10^{13}\,\mathrm{M}_{\odot}$ halos the Ne VIII gas is clumpier, and biased towards the cooler part of the warm-hot phase. We compare the simulations to observations by the CASBaH and CUBS surveys. We show that when inferring halo masses from stellar masses to compare simulated and observed halos, it is important to account for the scatter in the stellar-mass-halo-mass relation, especially at $\,\mathrm{M}_{\star} \gtrsim 10^{10.5} \,\mathrm{M}_{\odot}$. Median Ne VIII columns in the fiducial FIRE-2 model are about as high as observed upper limits allow, while the simulations analyzed do not reproduce the highest observed columns. This suggests that the median Ne VIII profiles predicted by the simulations are consistent with observations, but that the simulations may underpredict the scatter. We find similar agreement with analytical models that assume a product of the halo gas fraction and metallicity (relative to solar) $\sim 0.1$-$0.3$, indicating that observations are consistent with plausible CGM temperatures, metallicities, and gas masses. Variants of the FIRE simulations with a modified supernova feedback model and/or AGN feedback included (as well as some other cosmological simulations from the literature) more systematically underpredict Ne VIII columns. The circumgalactic Ne VIII observations therefore provide valuable constraints on simulations that otherwise predict realistic galaxy properties.