The abundance and properties of the lowest luminosity dwarf galaxies around the Milky Way: Insights from Semi-Analytic Models

TL;DR

This study uses the Galacticus semi-analytic framework to investigate the abundance and properties of the faintest Milky Way satellites under two cooling scenarios: (i) a fiducial model with H$_2$ cooling and a UV background, and (ii) a No-H$_2$ model with atomic cooling only. It demonstrates that H$_2$ cooling shifts the galaxy-formation threshold to lower halo masses, producing many more hyper-faint satellites ($M_V > -3$) hosted in smaller peak halos and with earlier quenching, while also broadening the size–luminosity distribution into the hyper-faint regime. After applying survey-detection incompleteness, the fiducial predictions align with current observations and imply dozens of new hyper-faint satellites could be found by upcoming surveys like LSST and Roman; the No-H$_2$ scenario predicts far fewer such systems. A key diagnostic emerges from the predicted line-of-sight velocity dispersions: $\sigma_{\rm los} \sim 1$–$3$ km s$^{-1}$ for hyper-faints, which is substantially higher than purely stellar systems of the same luminosity, offering a practical test to distinguish dark-matter–dominated dwarfs from star clusters in the ambiguous population.

Abstract

We investigate the formation and observable properties of faint satellite galaxies (M$_\rm V > -3$) in Milky Way-like halos using the semi-analytic galaxy formation model Galacticus. The ability of the smallest dark matter halos to form stars depends sensitively on the balance between gas cooling and reionization heating. To quantify how this balance shapes the abundance and properties of the faintest galaxies, we compare two model variants: a fiducial model that includes molecular hydrogen (H$_2$) cooling and UV background radiation, and a No-H$_2$ model with atomic cooling only. Both models reproduce the structural properties of brighter Milky Way satellites, but they diverge at the lowest luminosities in the hyper-faint regime. The fiducial model predicts a substantially larger population of such systems that are on average hosted in halos with lower peak masses and quenched earlier. Many of these predicted systems lie below current observational thresholds but are within reach of next-generation deep imaging surveys. The predicted size-luminosity distributions of both models overlap with the region occupied by recently discovered "ambiguous" systems, whose classification as galaxies or star clusters remains uncertain. Specifically, we find that hyper-faint satellites have line-of-sight velocity dispersions of $σ_{\rm los} \sim 1-3$ km/s in the fiducial model, nearly an order of magnitude higher than expected for purely self-gravitating stellar systems of the same stellar mass. This distinction underscores the diagnostic power of precise kinematic measurements for determining whether ambiguous objects are dark matter dominated dwarf galaxies or star clusters, and highlights the importance of upcoming spectroscopic campaigns in resolving the nature of the faintest satellites.

Figures (7)

• Figure 1: Absolute V-band magnitude (M$\rm_{V}$) as a function of projected half mass radius (R$\rm_{h}$) predictions from our model for satellites of MW-analogs. The left panel shows results from the fiducial model (including H$_2$ cooling), while the right panel shows the No-H$_2$ model (atomic cooling only). Model predictions are shown as shaded regions, where the color scale indicates the logarithm of the normalized probability density of satellites (see color bar at the top of each panel). Dashed diagonal lines correspond to constant surface brightness. Confirmed MW satellites (red circles) and ambiguous objects (olive diamonds) are shown for comparison. The gray box highlights the region associated with hyper-faint systems ($M{\rm _V} > -3$ and $1 \lesssim R{\rm _h} [pc] \lesssim 10$) on the size–luminosity relation.
• Figure 2: Luminosity function predictions for satellites of MW-analogs (line with the $1 \sigma$ halo-to-halo scatter). The left panel presents results from the fiducial model (including H$_2$ cooling), while the right panel shows results from the No-H$_2$ model (atomic cooling only). In both panels, corrections based on survey sensitivity have been shown for DES (dashed lines), PS1 (dotted lines), and LSST (dotted-dashed lines). For comparison, we include the observed luminosity function of confirmed MW satellites (red stars), ambiguous objects (olive stars), and the combined sample (black circles).
• Figure 3: Comparison of peak halo masses for satellites with $M_V > -3$ (dashed lines) and for the subset classified as hyper-faint satellites (solid lines). Results from the fiducial model (including H$_2$ cooling) are shown in blue, while those from the No-H$_2$ model (atomic cooling only) are shown in green.
• Figure 4: Comparison of star formation histories for satellites with $M_V > -3$ (dashed lines) and for the subset of hyper-faint satellites (solid lines). Results from the fiducial model (blue) and the No-H$2$ model (green) are shown. The comparison is based on $z_{90}$, defined as the redshift by which 90% of a galaxy’s stellar mass has formed.
• Figure 5: Line-of-sight velocity dispersion predictions (measured at r$\rm_{h}$) from our fiducial model, shown as the blue line, as a function of satellite stellar mass. For comparison, we include observed velocity dispersions of confirmed MW satellites (red markers) and the limited available measurements for ambiguous systems (olive markers). We also show the estimated velocity dispersions these ambiguous objects would have if they were purely self-gravitating stellar clusters (olive crosses with black edges).