Density profiles of dark matter haloes: diversity and dependence on environment

We study the outer density profiles of dark matter haloes predicted by a generalized secondary infall model and observed in a dissipationless cosmological simulation of a low-density flat cold dark matter model with the cosmological constant. We find substantial systematic variations in shapes and concentrations of the halo profiles as well as a strong correlation of the profiles with the environment in which the haloes are embedded. In the N-body simulation, the average outer slope of the density profiles, β(ρ∝r−β), of isolated haloes is β≈2.9, and 68 per cent of these haloes have values of β between 2.5 and 3.8. Haloes in dense environments of clusters are more concentrated and exhibit a broad distribution of β with an average value higher than the average β for isolated haloes. For haloes located within half the virial radius of the cluster from the centre values β≈4 are very common. Contrary to what one may expect, the haloes contained within groups and galaxy systems are less concentrated and have flatter outer density profiles than the isolated haloes: the distribution of β peaks at ≈2.3–2.7. The slope β weakly anticorrelates with the halo mass Mh. The concentration decreases with Mh, but its scatter is roughly equal to the whole variation of this parameter in the galaxy halo mass range. The mass and circular velocity of the haloes are strongly correlated, , with α≈3.3 and ≈3.5 for the isolated haloes and haloes in clusters, respectively. For Mh≈1012 h−1 M⊙ the rms deviations from these relations are Δ log Mh=0.12 and 0.18, respectively. Approximately 30 per cent of the haloes are contained within larger haloes or have massive companions within three virial radii. The companions are allowed to have masses larger than ∼0.3 times the mass of the current halo. The remaining 70 per cent of the haloes are isolated objects. We find that the distribution of β as well as the concentration—mass and Mh—Vm relations for the isolated haloes agree very well with the predictions of our seminumerical approach, which is based on a generalization of the secondary infall model and on the extended Press—Schechter formalism.