Formation and Structure of Halos in a Warm Dark Matter Cosmology

Abstract

Using high-resolution cosmological N-body simulations, we study how the density profiles of dark matter halos are affected by the filtering of the density power spectrum below a given scale length and by the introduction of a thermal velocity dispersion. In the warm dark matter (WDM) scenario, both the free-streaming scale, Rf, and the velocity dispersion, v, are determined by the mass, mW, of the WDM particle. We found that v is too small to affect the density profiles of WDM halos. Down to the resolution attained in our simulations (~0.01 virial radii), there is not any significant difference in the density profiles and concentrations of halos obtained in simulations with and without the inclusion of v. Resolved soft cores appear only when we artificially increase the thermal velocity dispersion to a value that is much higher than v. We show that the size of soft cores in a monolithic collapse is related to the tangential velocity dispersion. The density profiles of the studied halos with masses down to ~0.01 the filtering mass Mf can be described by the Navarro-Frenk-White shape; soft cores are not formed. Nevertheless, the concentrations of these halos are lower than those of the CDM counterparts and are approximately independent of mass. The cosmogony of halos with masses Mf is not hierarchical: they form through monolithic collapse and by fragmentation of larger structures. The formation epoch of these halos is slightly later than that of halos with masses ≈Mf. The lower concentrations of WDM halos with respect to their CDM counterparts can be accounted for by their late formation epoch. Overall, our results point to a series of advantages of a WDM model over the CDM one. In addition to solving the substructure problem, a WDM model with Rf ~ 0.16 Mpc (mW ≈ 0.75 keV; flat cosmology with ΩΛ = h = 0.7) also predicts concentrations, a Tully-Fisher relation, and formation epochs for small halos, which seems to be in better agreement with observations than CDM predictions.