Shen et al., available on arXiv

Abstract: We analyze the first cosmological baryonic zoom-in simulations of galaxies in dissipative self-interacting dark matter (dSIDM). The simulations utilize the FIRE-2 galaxy formation physics with the inclusion of dissipative dark matter self-interactions modeled as a constant fractional energy dissipation (f_diss=0.75). In this paper, we examine the properties of dwarf galaxies with M*~10^5–10^9 Msun in both isolation and within Milky Way–mass hosts. For isolated dwarfs, we find more compact galaxy sizes and promotion of disk formation in dSIDM with (sigma/m)<=1 cm^2 g^-1. On the contrary, models with (sigma/m) = 10 cm^2 g^-1 produce puffier stellar distributions that are in tension with the observed size–mass relation. In addition, owing to the steeper central density profiles, the subkiloparsec circular velocities of isolated dwarfs when (sigma/m) >= 0.1 cm^2 g^-1 are enhanced by about a factor of 2, which are still consistent with the kinematic measurements of Local Group dwarfs but in tension with the H I rotation curves of more massive field dwarfs. Meanwhile, for satellites of Milky Way–mass hosts, the median circular velocity profiles are marginally affected by dSIDM physics, but dSIDM may help promote the structural diversity of dwarf satellites. The number of satellites is slightly enhanced in dSIDM, but the differences are small compared with the large host-to-host variations. In conclusion, the dSIDM models with (sigma/m) >~ 0.1 cm^2 g^-1, f_diss = 0.75 are in tension in massive dwarfs (Mhalo ~ 10^11 Msun) due to circular velocity constraints. However, models with lower effective cross sections (at this halo mass/velocity scale) are still viable and can produce nontrivial observable signatures.