paper

The anti-hierarchical growth of supermassive black holes

arXiv:astro-ph/0402495 · doi:10.1111/j.1365-2966.2004.08147.x

Abstract

I present a new method to unveil the history of cosmic accretion and the build-up of SMBH in the nuclei of galaxies, based on observations of the evolving radio and (hard) X-ray luminosity functions of AGN. The fundamental plane of black hole activity discovered by Merloni, Heinz & Di Matteo (2003) is used as a mass and accretion rate estimator. I adopt the local BH mass function as a boundary condition to integrate backwards in time the continuity equation for the SMBH evolution, neglecting the role of mergers. Under the most general assumption that accretion proceeds in a radiatively efficient way above a certain rate, and in a radiatively inefficient way below, the redshift evolution of the mass and accretion rate functions are calculated self-consistently. The only tunable parameters are the accretion efficiency and the critical ratio of the X-ray to Eddington luminosity at which the transition between accretion modes takes place. For fiducial values of these parameters, I found that half (85%) of the local BH mass density was accumulated at redshift z<1 (z<3), mostly in radiatively efficient episodes of accretion. The evolution of the BH mass function between z=0 and z~3 shows clear signs of an anti-hierarchical behaviour: while the majority of the most massive objects (M > 10^9) were already in place at z~3, lower mass ones mainly grew at progressively lower redshift. Also, the average accretion rate decreases with time. Consequently, sources in the radiatively inefficient regime of accretion only begin to dominate the comoving accretion energy density in the universe at z<1. I discuss the implications of these results for the efficiency of accretion onto SMBH, the quasars lifetimes and duty cycles and the history of AGN feedback in the form of mechanical energy output (abriged).

13 pages, 11 figures. Minor changes to sections 2 and 6. New figures and references added. MNRAS, accepted

References in corpus (4)

Cited by in corpus (144)