We provide an analytic framework for interpreting observations of multiphase circumgalactic gas that is heavily informed by recent numerical simulations of thermal instability and precipitation in cool-core galaxy clusters. We start by considering the local conditions required for the formation of multiphase gas via two different modes: (1) uplift of ambient gas by galactic outflows, and (2) condensation in a stratified stationary medium in which thermal balance is explicitly maintained. Analytic exploration of these two modes provides insights into the relationships between the local ratio of the cooling and freefall time scales (i.e., t_cool / t_ff), the large-scale gradient of specific entropy, and development of precipitation and multiphase media in circumgalactic gas. We then use these analytic findings to interpret recent simulations of circumgalactic gas in which global thermal balance is maintained. We show that long-lasting configurations of gas with 5 < t_cool / t_ff < 20 and radial entropy profiles similar to observations of local cool-core galaxy cluster cores are a natural outcome of precipitation-regulated feedback. We conclude with some observational predictions that follow from these models. This work focuses primarily on precipitation and AGN feedback in galaxy cluster cores, because that is where the observations of multiphase gas around galaxies are most complete. However, many of the physical principles that govern condensation in those environments apply to circumgalactic gas around galaxies of all masses.

We compare several common sub-grid implementations of AGN feedback, focusing on the effects of different triggering mechanisms and the differences between thermal and kinetic feedback. Our main result is that pure thermal feedback that is centrally injected behaves differently from feedback with even a small kinetic component. Specifically, pure thermal feedback results in excessive condensation and smothering of the AGN by cold gas because the feedback energy does not propagate to large enough radii. We do not see large differences between implementations of different triggering mechanisms, as long as the spatial resolution is sufficiently high, probably because all of the implementations tested here trigger strong AGN feedback under similar conditions. In order to assess the role of resolution, we vary the size of the "accretion zone" in which properties are measured to determine the AGN accretion rate and resulting feedback power. We find that a larger accretion zone results in steadier jets but can also allow too much cold-gas condensation in simulations with a Bondi-like triggering algorithm. We also vary the opening angle of jet precession and find that a larger precession angle causes more of the jet energy to thermalize closer to the AGN, thereby producing results similar to pure thermal feedback. Our simulations confirm that AGN can regulate the thermal state of cool-core galaxy clusters and maintain the core in a state that is marginally susceptable to thermal instability and precipitation.

Heat input roughly balances radiative cooling in the gaseous cores of galaxy clusters even when the central cooling time is short, implying that cooling triggers a feedback loop that maintains thermal balance. Furthermore, cores with short cooling times tend to have multiphase structure, suggesting that the intracluster medium (ICM) becomes locally thermally unstable for cooling times < 1 Gyr. In this work, we use 2D and 3D hydrodynamic simulations to study the onset of condensation in idealized galaxy-cluster cores. In particular, we look at how the condensation process depends on the ratio of cooling time to freefall time and on the geometry of the gravitational potential. We conclude that the ICM can always evolve to a state in which condensation occurs if given enough time, but that an initial timescale ratio tcool /tff < 10 is needed for thermal instability to grow quickly enough to affect realistic cluster cores within a timescale that is relevant for cosmological structure formation. We find that instability leads to convection and that perturbations continue to grow while the gas convects. Condensation occurs when the timescale ratio in the low-entropy tail of the perturbation distribution drops below tcool /tff < 3, even if the volume-averaged timescale ratio is substantially greater. In our simulations, the geometry of the gravitational potential does not have a strong effect on thermal stability. Finally, we find that if condensation is powering feedback, a conversion efficiency of around 10^-3 for converting the condensed mass into thermal energy is sufficient to maintain thermal balance in the ICM.

The first stars in the universe, termed Population III, are thought to have been very massive compared to the stars that form in the present epoch. As feedback from the first generation of stars altered the contents of the interstellar medium, the universe switched to a low-mass mode of star formation, which continues in the high metallicity stars formed in the present era. Several studies have investigated the transition between metal-free and metal-enriched star formation, with tentative evidence being found for a metallicity threshold near 10^-3.5 Z_sun due to atomic and molecular transitions and another threshold near 10^-5.5 Z_sun due to dust. In this work, we simulate the formation of stars in idealized low-metallicity halos using the AMR code Enzo. We conduct several simulations of 10^6 M_sun and 10^7 M_sun halos in which the metal content, initial rotation, and degree of turbulence are varied in order to study the effect of these properties on gas fragmentation over a range of densities. We find tentative support for the idea of a critical metallicity, but the effect of varying metallicity on the gas we observe is not as dramatic as what has been reported in earlier studies. We find no clear relation between the initial spin or the initial level of turbulence in the halo and the final properties of the gas contained therein. Additionally, we find that the degree to which the Jeans length is refined, the initial density profile of the gas, and the inclusion of deuterium chemistry each have a significant effect on the evolution and fragmentation of the gas in the halo - in particular, we find that at least 64 grid cells are needed to cover the Jeans length in order to properly resolve the fragmentation.