Satellites of Simulated Galaxies: survival, merging, and their relation to the dark and stellar halos

Abstract

We study the population of satellite galaxies formed in a suite of N-body/gasdynamical simulations of galaxy formation in a LCDM universe. We find little spatial or kinematic bias between the dark matter and the satellite population. The velocity dispersion of the satellites is a good indicator of the virial velocity of the halo: σ_{sat}/V_{vir}=0.9 +/- 0.2. Applied to the Milky Way and M31 this gives V_{vir}^{MW}=109 +/- 22$ km/s and V_{vir}^{M31} = 138 +/- 35 km/s, respectively, substantially lower than the rotation speed of their disk components. The detailed kinematics of simulated satellites and dark matter are also in good agreement. By contrast, the stellar halo of the simulated galaxies is kinematically and spatially distinct from the population of surviving satellites. This is because the survival of a satellite depends on mass and on time of accretion; surviving satellites are biased toward low-mass systems that have been recently accreted by the galaxy. Our results support recent proposals for the origin of the systematic differences between stars in the Galactic halo and in Galactic satellites: the elusive ``building blocks'' of the Milky Way stellar halo were on average more massive, and were accreted (and disrupted) earlier than the population of dwarfs that has survived self-bound until the present.