Observational evidence shows that low-redshift galaxies are surrounded by extended haloes of multiphase gas, the so-called 'circumgalactic medium' (CGM). To study the survival of relatively cool gas (T < 10 5 K) in the CGM, we performed a set of hydrodynamical simulations of cold (T = 10 4 K) neutral gas clouds travelling through a hot (T = 2 × 10 6 K) and low-density (n = 10 −4 cm −3 ) coronal medium, typical of Milky Way-like galaxies at large galactocentric distances (∼ 50 − 150 kpc). We explored the effects of different initial values of relative velocity and radius of the clouds. Our simulations were performed on a two-dimensional grid with constant mesh size (2 pc) and they include radiative cooling, photoionization heating and thermal conduction. We found that for large clouds (radii larger than 250 pc) the cool gas survives for very long time (larger than 250 Myr): despite that they are partially destroyed and fragmented into smaller cloudlets during their trajectory, the total mass of cool gas decreases at very low rates. We found that thermal conduction plays a significant role: its effect is to hinder formation of hydrodynamical instabilities at the cloud-corona interface, keeping the cloud compact and therefore more difficult to destroy. The distribution of column densities extracted from our simulations are compatible with those observed for low-temperature ions (e.g. Siii and Siiii) and for high-temperature ions (Ovi) once we take into account that Ovi covers much more extended regions than the cool gas and, therefore, it is more likely to be detected along a generic line of sight.

In star-forming galaxies, stellar feedback can have a dual effect on the circumgalactic medium both suppressing and stimulating gas accretion. The trigger of gas accretion can be caused by disc material ejected into the halo in the form of fountain clouds and by its interaction with the surrounding hot corona. Indeed, at the disc-corona interface, the mixing between the cold/metal-rich disc gas (T 10 4 K) and the hot coronal gas (T 10 6 K) can dramatically reduce the cooling time of a portion of the corona and produce its condensation and accretion. We studied the interaction between fountain clouds and corona in different galactic environments through parsec-scale hydrodynamical simulations, including the presence of thermal conduction, a key mechanism that influences gas condensation. Our simulations showed that the coronal gas condensation strongly depends on the galactic environment, in particular it is less efficient for increasing virial temperature/mass of the haloes where galaxies reside and it is fully ineffective for objects with virial masses larger than 10 13 M . This result implies that the coronal gas cools down quickly in haloes with low-intermediate virial mass (M vir 3 × 10 12 M ) but the ability to cool the corona decreases going from late-type to earlytype disc galaxies, potentially leading to the switching off of accretion and the quenching of star formation in massive systems.

High-velocity clouds consist of cold gas that appears to be raining down from the halo to the disc of the Milky Way. Over the past fifty years, two competing scenarios have attributed their origin either to gas accretion from outside the Galaxy or to circulation of gas from the Galactic disc powered by supernova feedback (galactic fountain). Here we show that both mechanisms are simultaneously at work. We use a new galactic fountain model combined with high-resolution hydrodynamical simulations. We focus on the prototypical cloud complex C and show that it was produced by an explosion that occurred in the Cygnus-Outer spiral arm about 150 million years ago. The ejected material has triggered the condensation of a large portion of the circumgalactic medium and caused its subsequent accretion onto the disc. This fountain-driven cooling of the lower Galactic corona provides the low-metallicity gas required by chemical evolution models of the Milky Way's disc.

We present a study of the gas cycle and star formation history in the central 500 pc of the Milky Way, known as Central Molecular Zone (CMZ). Through hydrodynamical simulations of the inner 4.5 kpc of our Galaxy, we follow the gas cycle in a completely self-consistent way, starting from gas radial inflow due to the Galactic bar, the channelling of this gas into a dense, star-forming ring/stream at ≈ 200 − 300 pc from the Galactic centre, and the launching of galactic outflows powered by stellar feedback. We find that star formation activity in the CMZ goes through oscillatory burst/quench cycles, with a period of tens to hundreds of Myr, characterised by roughly constant gas mass but order-of-magnitude level variations in the star formation rate. Comparison with the observed present-day star formation rate of the CMZ suggests that we are currently near a minimum of this cycle. Stellar feedback drives a mainly two-phase wind off the Galactic disc. The warm phase dominates the mass flux, and carries 100 − 200% of the gas mass converted into stars. However, most of this gas goes into a fountain and falls back onto the disc rather than escaping the Galaxy. The hot phase carries most of the energy, with a time-averaged energy outflow rate of 10 − 20% of the supernova energy budget.

Galactic outflows produced by stellar feedback are known to be multiphase in nature. Observations and simulations indicate that the material within several kiloparsecs of galactic disk midplanes consists of warm clouds embedded within a hot wind. A theoretical understanding of the outflow phenomenon, including both winds and fountain flows, requires study of the interactions among thermal phases. We develop a method to quantify these interactions via measurements of mass, momentum, and energy flux exchanges using temporally and spatially averaged quantities and conservation laws. We apply this method to a star-forming interstellar medium simulation based on the TIGRESS framework, for solar neighborhood conditions. To evaluate the extent of interactions among the phases, we examine the validity of the “ballistic model,” which predicts the trajectories of the warm phase (5050 K < T < 2 × 104 K) treated as non-interacting clouds. This model is successful at intermediate vertical velocities ( ), but at higher velocities, we observe an excess in simulated warm outflow compared to the ballistic model. This discrepancy cannot be fully accounted for by cooling of high-velocity, intermediate-temperature (2 × 104 K < T < 5 × 105 K) gas. We examine the fluxes of mass, momentum, and energy and conclude that the warm phase gains mass via cooling of the intermediate phase and momentum from the hot (T > 5 × 105 K) phase. The large energy flux from the hot outflow, transferred to the warm and intermediate phases, is quickly radiated away. A simple interaction model implies an effective warm cloud size in the fountain flow of a few 100 pc, showing that warm–hot flux exchange mainly involves a few large clouds rather than many small ones.

Disentangling distinct stellar populations along the red-giant branches (RGBs) of Globular Clusters (GCs) is possible by using the pseudo two-color diagram dubbed chromosome map (ChM). One of the most intriguing findings is that the so-called first-generation (1G) stars, characterized by the same chemical composition of their natal cloud, exhibit extended sequences in the ChM. Unresolved binaries and internal variations in helium or metallicity have been suggested to explain this phenomenon. Here, we derive high-precision Hubble Space Telescope photometry of the GCs NGC 6362 and NGC 6838 and build their ChMs. We find that both 1G RGB and main-sequence (MS) stars exhibit wider ChM sequences than those of second-generation (2G). The evidence of this feature even among unevolved 1G MS stars indicates that chemical inhomogeneities are imprinted in the original gas. We introduce a pseudo two-magnitude diagram to distinguish between helium and metallicity, and demonstrate that star-to-star metallicity variations are responsible for the extended 1G sequence. Conversely, binaries provide a minor contribution to the phenomenon. We estimate that the metallicity variations within 1G stars of 55 GCs range from less than [Fe/H]∼0.05 to ∼0.30 and mildly correlate with cluster mass. We exploit these findings to constrain the formation scenarios of multiple populations showing that they are qualitatively consistent with the occurrence of multiple generations. In contrast, the fact that 2G stars have more homogeneous iron content than the 1G challenges the scenarios based on accretion of material processed in massive 1G stars onto existing protostars.

Cosmic-ray transport on galactic scales depends on the detailed properties of the magnetized, multiphase interstellar medium (ISM). In this work, we postprocess a high-resolution TIGRESS magnetohydrodynamic simulation modeling a local galactic disk patch with a two-moment fluid algorithm for cosmic-ray transport. We consider a variety of prescriptions for the cosmic rays, from a simple, purely diffusive formalism with constant scattering coefficient, to a physically motivated model in which the scattering coefficient is set by the critical balance between streaming-driven Alfvén wave excitation and damping mediated by local gas properties. We separately focus on cosmic rays with kinetic energies of ∼1 GeV (high-energy) and ∼30 MeV (low energy), respectively important for ISM dynamics and chemistry. We find that simultaneously accounting for advection, streaming, and diffusion of cosmic rays is crucial for properly modeling their transport. Advection dominates in the high-velocity, low-density hot phase, while diffusion and streaming are more important in higher-density, cooler phases. Our physically motivated model shows that there is no single diffusivity for cosmic-ray transport: the scattering coefficient varies by four or more orders of magnitude, maximal at density n H ∼ 0.01 cm−3. The ion-neutral damping of Alfvén waves results in strong diffusion and nearly uniform cosmic-ray pressure within most of the mass of the ISM. However, cosmic rays are trapped near the disk midplane by the higher scattering rate in the surrounding lower-density, higher-ionization gas. The transport of high-energy cosmic rays differs from that of low-energy cosmic rays, with less effective diffusion and greater energy losses for the latter.

We present a novel, physically-motivated sub-grid model for HII region feedback within the moving mesh code Arepo, accounting for both the radiation pressure-driven and thermal expansion of the ionised gas surrounding young stellar clusters. We apply this framework to isolated disc galaxy simulations with mass resolutions between 103 M⊙ and 105 M⊙ per gas cell. Each simulation accounts for the self-gravity of the gas, the momentum and thermal energy from supernovae, the injection of mass by stellar winds, and the non-equilibrium chemistry of hydrogen, carbon and oxygen. We reduce the resolution-dependence of our model by grouping those HII regions with overlapping ionisation front radii. The Strömgren radii of the grouped HII regions are at best marginally-resolved, so that the injection of purely-thermal energy within these radii has no effect on the interstellar medium. By contrast, the injection of momentum increases the fraction of cold and molecular gas by more than 50 per cent at mass resolutions of 103 M⊙, and decreases its turbulent velocity dispersion by ∼10 kms−1. The mass-loading of galactic outflows is decreased by an order of magnitude. The characteristic lifetime of the least-massive molecular clouds ($M/{\rm M}_\odot \lesssim 5.6 \times 10^4$) is reduced from ∼18 Myr to $\lesssim 10$ Myr, indicating that HII region feedback is effective in destroying these clouds. Conversely, the lifetimes of intermediate-mass clouds ($5.6 \times 10^4 \lesssim M/{\rm M}_\odot \lesssim 5 \times 10^5$) are elongated by ∼7 Myr, likely due to a reduction in supernova clustering. The derived cloud lifetimes span the range from 10-40 Myr, in agreement with observations. All results are independent of whether the momentum is injected from a spherical or a blister-type HII region.