2006A&A...449..551L

The fraction of AGN with photoelectric absorption in the X-rays ranging from N_H of 10²² up to about 10²⁴cm^–2 (Compton-thin) appears observationally to be anticorrelated to their luminosity L_x. This recently found evidence is used to investigate the location of the absorbing gas. The molecular torus invoked in the unified picture of AGN, while it can be regarded as confirmed on several grounds to explain the Compton-thick objects, do not conform to this new constraint, at least in its physical models as developed so far. In the frame of observationally based evidence that in Compton-thin sources the absorbing gas might be located far away from the X-ray source, it is shown that the gravitational effects of the black hole (BH) on the molecular gas in a disk, within 25-450pc (depending on the BH mass, from 10⁶ to 10⁹M_☉), leads naturally to the observed anticorrelation, under the assumption of a statistical correlation between the BH mass and L_x. Its normalization is also reproduced provided that the surface density, Σ, of this gas is larger than about 150-200M_☉/pc², and assuming that the bolometric luminosity is one tenth of the Eddington limit. Interestingly, the required values are consistent with the value of the 300 pc molecular disk in our own galaxy, namely 500M_☉/pc². In a sample of nearby galaxies from the BIMA SONG survey, it is found that half of the objects have central Σ larger than 150M_☉/pc². Given the simplicity of the proposed model, this finding is very encouraging, waiting for future higher resolution surveys in CO on more distant galaxies.