Figuring Out Gas & Galaxies In Enzo (FOGGIE). XII. The Formation and Evolution of Extended HI Galactic Disks and Warps with a Dynamic CGM

TL;DR

This study investigates the formation and evolution of extended HI disks around Milky Way–mass halos using the FOGGIE cosmological zoom-in simulations with forced and cooling refinements that resolve the CGM down to sub-kiloparsec scales. By applying a 3D clump-finding disk definition and tracking angular-momentum coherence, the authors show a robust environmental dichotomy: Less Populated CGMs yield thin inner disks, while More Populated CGMs foster thicker disks with persistent misalignments and warps. Misaligned HI structures—polar rings and warps—are ubiquitous but arise from distinct formation histories, including mergers and tidal accretion, and their evolution strongly correlates with local CGM content rather than halo mass. The results reveal a link between CGM density/temperature and inner-disk morphology (the so-called inner CGM virialization), with implications for interpreting extended HI observations and guiding future synthetic-cube comparisons to data.

Abstract

Atomic Hydrogen (HI) is an important component of gas in and around galaxies and forms extended disk-like structures well beyond the extent of starlight. Here we investigate the properties and evolution of extended HI disks that emerge in six Milky Way-mass galaxies using cosmological zoom-in simulations from the Figuring Out Gas & Galaxies in Enzo (FOGGIE) suite. We focus on the formation, evolution, and morphology of extended gaseous disks that emerge in all six systems. We find that median HI column densities drop sharply at the disk edge, with mean column densities outside the disk dominated by dense (NHI~10^19 cm-2), clumpy structures. All systems have significant misaligned features (warps or polar rings) at some point in their evolution; however, their frequencies, lifetimes, and origins vary significantly. We find that the morphology of the FOGGIE disks are correlated with properties of their Circumgalactic Medium (CGM). We classify these systems into two broad categories: those with CGMs that are Less Populated with HI and those with CGMs that are More Populated with HI. Both categories kinematically settle by z=0, but the Less Populated systems all form thin disks by z=0, while the More Populated systems do not. This classification is independent of disk and halo mass, implying the formation of a thin disk is influenced by local environmental factors. Our results indicate a connection between CGM content and disk formation that is not yet fully understood. A second paper investigates observational aspects of these structures.

Figures (28)

• Figure 1: Satellite mass within $2 R_{\rm vir}$ relative to the main halo virial mass, versus the H1$\,$ mass in the CGM relative to the H1$\,$ mass of the disk over the last $\sim$200 Myr. There is a clear correlation between the relative amount of H1$\,$ in the CGM with the nearby satellite mass at that time (red-dashed line, slope = 5.3, intercept = 0.06). Systems on the bottom left of this plot are in Less Populated local environments, while systems in the top right are in More Populated local environments. We subdivide our sample into two categories based on their position along this relation. Tempest, Maelstrom, and Blizzard comprise the Less Populated systems, while Cyclone, Squall, and Hurricane comprise the More Populated systems. We emphasize that this is a continuum. Blizzard is an edge case, as the most H1$\,$ rich of the Less Populated systems. Cyclone moving down the continuum at late times as well, which may reflect other galactic properties.
• Figure 2: Face- and edge-on H1$\,$ column densities within $R_{\rm vir}$ for the six halos considered in this study at redshift $z=0$. We subdivide our halos into two distinct categories (Less Populated and More Populated) based on the amount of H1$\,$ in the CGM. The top row of each category shows the H1$\,$ column densities for all gas, while the bottom row shows the projections only including gas in our disk definition (see Sec. \ref{['sec:hi_disk_definition']}). As can be seen, this categorization correlates strongly with the structure of the disk. The color map transitions to grey scale at $N_{\rm HI}<10^{17.5}$, which is the high end of sensitivity for the latest H1$\,$ interferometry surveys deBlok24-Mhongoose.
• Figure 3: Radial profiles for face-on H1$\,$ column density maps at redshift $z=0$ (see Fig. \ref{['fig:hi_projections']}). The heat map shows a 2-D histogram of pixels in the face-on projection, color-coded by the H1$\,$ mass of each pixel at a given point on the curve. The green dashed line shows the mean profile, while the blue dashed line shows the median profile. The mean and median column densities are similar within the disk itself. Outside the disk, the median profile drops sharply, and the mean profile falls off more gradually. In the CGM, the mean values are dominated by the densest clumpy regions. The top row shows the Less Populated systems, which drop off more sharply, while the bottom row shows the More Populated systems, which drop off more gradually.
• Figure 4: Evolution of the maximum radial extent of each disk, based on our disk definition. All systems grow similarly over cosmological time, despite classification (denoted by line type). Tempest and Maelstrom grow more or less monotonically, while the other systems go through periods of rapid growth and contractions. There is no obvious signal in disk growth that could be associated with the Less Populated or More Populated categories. Disk radius in this case is defined by the radius of the gas cell furthest from the center in our definition, and is therefore sensitive to spikes during merger events. Curves were smoothed using a moving average filter (270 Myr) to improve readability.
• Figure 5: Evolution of the gas mass of each disk, based on our disk definition. Although overall baryon content increases more or less monotonically in all cases, some systems lose large amounts of gas, particularly at higher redshift. Overall trends seem uncorrelated with classification (denoted by line type). Curves were smoothed using a moving average filter (270 Myr) to improve readability.