Accreting Black Holes during cosmic time

Abstract

The study of astrophysical black holes (BH) is mainly driven by three dif- ferent rationales. First of all, the existence of this kind of objects is one of the strongest implications of the theory of General Relativity, and their study can give useful information on strong gravity e ects in action. Secondly, the emission processes, which allow us to detect such a kind of sources, originate from accretion ows or relativistic jets. Both of these mech- anisms take place in regions very close to the BH, thus allowing to study not only the physics of matter in extreme conditions, but also radiative e ects and relativistic magnetohydrodinamics. Lastly, BH a ect the formation and evolution of the structures they live in, like galaxies, groups and clusters, and therefore play a key role in a broader cosmological context. This tight link between BH activity and galaxy evolution has been indicated by several discoveries: the observation of a supermassive BH (SMBH) in most of the nearby bulge-dominated galaxies (see e.g. Gebhardt et al. 2000; Ferrarese & Merritt 2000; Marconi & Hunt 2003 and references therein); the growth of SMBH happen mainly during active phases, and therefore most local bulge galaxies should have passed an active phase in their lifetime (see Soltan 1982; Marconi et al. 2004); the evolution of active BH (Ueda et al. 2003; Hasinger, Miyaji & Schmidt 2005; La Franca et al. 2005) and of star-forming galaxies (Cowie et al. 1996; Franceschini et al. 1999) have a very similar shape. Models constrained from evolving optical and X-ray luminosity functions gave encouraging results, predicting some of the observed trends in AGN. How- ever, the overestimation of the space density of low-luminosity sources, along with other observational evidences, suggest that a signi cant fraction of accret- ing BH (as high as 50%) is missed by current surveys. All these pieces of evidence imply that the realization of a complete census of accreting black holes and constraining their feedback action on the host galaxies are key steps towards the understanding of the galaxy formation and evolution. We have measured the space density of Compton-thick AGN in 24 m se- lected samples from COSMOS and GOODS, where deep Chandra observations exist, using the latest data releases available. We have used a new selection criterion, taking advantage of luminosity function model predictions, in order to select a population of Compton-thick candidates, which have been con rmed through X-ray spectral analysis of stacked data. Our measures show that: at high luminosities (LX > 1044 erg s􀀀1) the Compton-thick AGN densi- ties are compatible with previous luminosity functions estimations; at low luminosities (LX 1043 erg s􀀀1) are signi cantly higher than the luminosity function estimations, as required by several galaxy formation models and CXRB synthesis models; we found that the Compton-thick AGN density increases with decreasing luminosity, in a similar way to the moderately obscured AGN; our predicted local density of relic BH, due to obscured and unobscured accretion, is consistent with the estimations derived from the local scaling relations and with the model predictions. Some models predict that the same cold gas that fuels black hole accretion and star formation is responsible of AGN obscuration. In this scenario, the more powerful AGN clean their line of sight more quickly than low luminosity AGN; therefore, the fraction of obscured AGN should increase for decreasing luminosity, as observed in Compton-thin sources, in agreement with our ndings. In addition of obtaining a complete census of accreting black holes, the quanti cation of the AGN feedback on their host galaxies is another fundamental ingredient in order to fully understand the cosmological evolution of galaxies. In this context, the measure of the probability distribution of the ratio RX between radio and X-ray luminosities is required for a correct estimation of the radio feedback of AGN. We used a sample of more than 1600 X-ray selected AGN observed at 1.4 GHz to measure the probability distribution function P(RXjLX; z) as a func- tion of the X-ray luminosity and redshift. The knowledge of the P(RXjLX; z) distribution is necessary to estimate the AGN kinetic (radio) feedback into the hosting galaxies by allowing to couple it with the luminous, accreting, phases of the AGN activity. The average value of RX increases with decreasing X-ray luminosities and (possibly) increasing redshift. At variance, we did not nd a statistical signi cant di erence between the radio properties of the X-ray ab- sorbed (NH > 1022 cm􀀀2) and unabsorbed AGN. We were able to better measure the densities of the more radio quiet (RX < 􀀀4) AGN which resulted to be responsible of about half of the derived kinetic power density. According to our analysis the value of the kinetic energy density is in qualita- tive agreement with the last generation galaxy evolution scenarios, where radio mode AGN feedback is invoked to quench the star formation in galaxies and slow down the cooling ows in galaxy clusters. However at redshifts below 0.5 we found a sharp (about a factor of ve) decrease of the kinetic energy density, which is strictly related the AGN density evolution, but which is not included in many of the galaxy/AGN formation and evolution models where, instead, the radio mode feedback is assumed to continuously increase (or only smoothly decrease) at low redshift. We have also veri ed that our ndings on a higher density of low luminosity Compton-thick AGN with respect to luminosity function expectations do not change signi cantly the conclusions derived by this work.