High-energy spectra of active galactic nuclei .2. Absorption in Seyfert galaxies

Absorption by cold material in a large sample of active galaxies has been analyzed in order to study statistically the behavior of absorbed sources. The analysis indicates that on the basis of the column density alone, sources can be divided into low-absorption ([N-H/N-HGa1] less than or equal to 50) and high-absorption ([N-H/N-HGa1] greater than or equal to 50) objects. While the second group consists mostly of narrow emission line galaxies (Seyfert galaxies of type 1.9-2), the first group is less homogenous, being formed by a mixture of broad and narrow emission line objects (Seyfert 1-2 galaxies). A study of the distribution of the column density values by means of bootstrap analysis confirms the reality of this effect. One group consisting of optically selected objects is well explained within the unified theory as nuclei obscured by a molecular torus. The second group made up of X-ray-and IRAS-selected objects is more difficult to define: in these sources the absorption is underestimated owing to difficulties (1) in fitting complex absorption spectra or (2) in measuring N-H values in Compton-thick sources or the absorption has a different origin than in the torus. Possible correlations of absorption with X-ray luminosity, axial ratio, and Balmer decrement have also been investigated. Previous suggestions that lower luminosity AGNs tend to be more highly absorbed than those with higher luminosity are not confirmed by the present data; neither is any evidence for a correlation of N-H with axial ratio (b/a) found except for a preference of Seyfert 1-1.5 galaxies to be in face-on galaxies. While some sources (Seyfert 1-1.5 galaxies and low-absorption objects) have X-ray absorption compatible with Balmer decrement, high-absorption objects have column densities much higher than predicted from optical observations. These results are in agreement with the unified theory since the torus parameters are expected to be independent of luminosity, its orientation should be random with respect to the host galaxy, and its location should be in between the broad- and narrow-line regions. A study of the N-H variability indicates that in a large fraction (70%) of the sources for which the analysis could be done, N-H varies on timescales from months to years. In Seyfert 1-1.5 galaxies, the variability is associated with a region in or near the broad-line region and is explained in terms of partial covering and/or warm absorption models. In Seyfert 2 galaxies, the only variability observed is that associated with narrow emission line galaxies. The study of the column density distributions indicates that Seyfert 1-1.5 galaxies are characterized by N-H = 18(-7)(+9) x 10(21) atoms cm(-2). Seyfert 1.9-2 galaxies have instead N-H = 96(-35)(+54) x 10(21) atoms cm(-2) and a larger dispersion; if this group is divided into low- and high-absorption objects, N-H = 14.5(-5.3)(+7.2) x 10(21) atoms cm(-2) and N-H = 132.8(-52.6)(+80.1) x 10(21) atoms cm(-2), respectively, are obtained. The observed dispersion in each group is consistent with being entirely due to column density variability.