The Milky Way satellite velocity function is a sharp probe of small-scale structure problems

Abstract

Twenty years ago, the mismatch between the observed number of Milky Way satellite galaxies and the predicted number of cold dark matter (CDM) subhalos was dubbed the ``missing satellites problem". Although mostly framed since in terms of satellite counts in luminosity space, the missing satellites problem was originally posed in velocity space. Importantly, the stellar velocity dispersion function encodes information about the density profile of satellites as well as their abundance. In this work, we completeness correct the MW satellite stellar velocity dispersion function down to its ultrafaint dwarfs ($L \gtrsim 340$ L$_\odot$). Our most conservative completeness correction is in good agreement with a simple CDM model in which massive, classical satellites (M$_{\rm 200} \gtrsim 5 \times 10^8~$M$_\odot$) have baryon-driven cores, while lower mass, ultrafaint satellites inhabit cuspy halos that are not strongly tidally stripped. Tidal destruction of satellites by the MW's disk must be minimal, otherwise the completeness-corrected velocity function exceeds any plausible CDM prediction -- a ``too many satellites" problem. We rule out non-core-collapsing self-interacting dark matter models with a constant cross section $\gtrsim$ 0.1 cm$^2$/g. Constraints on warm dark matter are stronger than those based on the luminosity function due to its additional sensitivity to subhalo central densities, which suppresses number counts by up to an additional 30%. A thermal relic mass $\gtrsim$ 6 keV is preferred. Reducing uncertainties on stellar velocity dispersion measurements and the amount of tidal stripping experienced by the faintest dwarfs is key to determining the severity of the too many satellites problem.