The role of black holes in galaxy formation

Abstract

We incorporate a model for black hole growth during galaxy mergers into the semi-analytical galaxy formation model based on ACDM proposed by Baugh et al. (2005). Our black hole model has one free parameter, which we set by matching the observed zeropoint of the local correlation between black hole mass and bulge luminosity. We present predictions for the evolution with redshift of the relationships between black hole mass and bulge properties. Our simulations reproduce the evolution of the optical luminosity function of quasars. We study the demographics of the black hole population and address the issue of how black holes acquire their mass. We find that the direct accretion of cold gas during starbursts is an important growth mechanism for lower mass black holes and at high redshift. The reassembly of pre-existing black hole mass into larger units via merging dominates the growth of more massive black holes at low redshift. As redshift decreases, progressively less massive black holes have the highest fractional growth rates, in line with recent claims of 'downsizing’ in quasar activity. We are able to reproduce the observed population of rare, highly luminous quasars at z ~ 6 in the ACDM model, although we struggle when the model parameters are refined to those in the current best fit of Sanchez et al. (2006). which has been hinted at by recent observations. We find that the most massive black holes and the most luminous quasars at z ~ 6 are not hosted exclusively by the most massive dark matter haloes, as is often assumed. Finally, we repeat our study of the assembly histories of black holes using the new galaxy and black hole formation model of Bower et al. (2006), which includes AGN feedback. Black hole growth in this model is dominated by the accretion required to fuel AGN feedback at low redshifts and by the accretion during starbursts triggered by disc instabilities at high redshifts. At no redshift is growth by BH-BH mergers dominant, although we still predict a high BH-BH merger rate. The Bower et al. model gives a somewhat better match to the M(_BH) - bulge relations, while the Baugh et al. model gives a better match to the quasar luminosity function. We suggest future directions to explore with these models.