2008A&A...490..905H

Intrinsic absorption is a fundamental physical property for understanding the evolution of active galactic nuclei (AGN). Here we study a sample of 1290 AGN, selected in the 2-10keV band from different flux-limited surveys with very high optical identification completeness. The AGN were grouped into two classes, unabsorbed (type-1) and absorbed (type-2), depending on their optical spectroscopic classification and X-ray absorption properties, using hardness ratios. Utilising the optical to X-ray flux ratios, a rough correction was applied for the ∼8% redshift incompleteness still present in the sample. Then the fraction of absorbed sources was determined as a function of X-ray luminosity and redshift. The absorbed fraction decreases strongly with X-ray luminosity. This can be represented by an almost linear decrease from ∼80% to ∼20% in the luminosity range log L_X=42-46 and is consistent with similar derivations in the optical and MIR bands. Several methods are used to study a possible evolution of the absorption fraction. The absorbed fraction increases significantly with redshift, which can be described by a power-law with a slope ~(1+z)^0.62±0.11, saturating at a redshift of z∼2. A simple power-law fit ~(1+z)^0.48±0.08 over the whole redshift is also marginally consistent with the data. The variation in the AGN absorption with luminosity and redshift is described with higher statistical accuracy and smaller systematic errors than previous results. The findings have important consequences for the broader context of AGN and galaxy co-evolution. Here it is proposed that the cosmic downsizing in the AGN population is due to two different feeding mechanisms: a fast process of merger-driven accretion at high luminosities and high redshifts versus a slow process of gas accretion from gravitational instabilities in galactic discs rebuilding around pre-formed bulges and black holes.