2025MNRAS.539.3797J

We use the cosmological hydrodynamical simulation NEWHORIZON to study the effects of the baryonic component on the inner mass profile of dark matter haloes of isolated galaxies ($10^6< M_*/\mathrm{ M}_\odot < 10^{11.5}$). Dark matter deficits ('cores') develop only in galaxies in a narrow range of stellar mass, $5\times 10^7< M_*/\mathrm{ M}_\odot < 3\times 10^9$. The lower stellar mass limit arises because a minimum amount of star formation is required to drive the baryonic outflows that redistribute dark matter and create a core. The upper limit roughly coincides with the total amount of dark matter initially contained within the innermost 2 kpc ($1 \pm 0.5 \times 10^{9}\, \mathrm{ M}_\odot$), which roughly coincides with the stellar half-mass radius of these dwarfs. This enclosed mass is quite insensitive to the total virial mass of the system. The same upper limit applies to other simulations, like NIHAO and EAGLE-CHT10, despite their rather different galaxy formation efficiencies. This suggests that it is the galaxy total stellar mass that determines when a core is formed, and not the galaxy-to-dark halo mass ratio, as argued in earlier work. This is consistent with a back-of-the-envelope estimate for a SN-induced rate of orbital diffusion. Although NEWHORIZON dwarfs reproduce the observed diversity of rotation curves better than other simulations, there are significant differences in the gravitational importance of baryons in the inner regions of dwarfs compared to observations. These differences prevent us from concluding that cosmological simulations are currently fully able to account for the observed diversity of rotation curve shapes.