AGN dust tori: the X-ray-infrared connection

We have combined the CLASXS Chandra survey in Lockman with the Spitzer SWIRE survey data to study the X-ray-infrared connection for AGN. The sample consists of 401 X-ray-sources, of which 306 are detected by Spitzer, and a further 257 AGN candidates detected through their dust torus, but not by Chandra. For X-ray sources the X-ray hardness ratio has been modelled in terms of a power-law with absorption N(H). The optical and infrared data have been modelled in terms of our well-established optical galaxy and QSO templates, and infrared templates based on radiative transfer models. Our estimate of the N(H) distribution is consistent with other studies, but we do find a higher proportion of low absorption objects at z < 0.5 than at z > 0.5. While we find only one X-ray AGN with N(H) > 10^{24} cm^{-2}, we argue that 10 objects with torus luminosity apparently exceeding the bolometric X-ray to 3 \mu m luminosity are strong candidates for being heavily absorbed in X-rays. We also estimate that at least half of the infrared-detected AGN dust tori which are undetected in X-rays are likely to be Compton thick. Our estimate of the total number of Compton-thick objects corresponds to > 20%$ of the combined SWIRE-CLASXS sample (and with an upper limit of 39 %). The range of dust covering factors is 1-100 %, with a mean of 40 %, ie a Type 2 fraction of 40 %. Measured by the ratio of dust torus luminosity to X-ray or (for Type 1 objects) optical luminosity, the covering factor appears to decrease towards intermediate AGN luminosity, in contradiction to estimates based on ratios of narrow-line and broad-line spectra, but may increase again at low AGN luminosity.