A search for X-ray absorbed sources in the 3XMM catalogue, using photometric redshifts and Bayesian spectral fits

TL;DR

This paper introduces XMMFITCAT-Z, a Bayesian spectral fit catalogue for X-ray sources in the 3XMM-DR6 survey, analyzing optical and IR properties of absorbed AGN and comparing selection criteria.

Contribution

It presents a new Bayesian spectral fitting catalogue for X-ray sources with photometric redshifts, enabling detailed analysis of obscured AGN properties and selection biases.

Findings

A significant fraction of X-ray absorbed AGN are missed by mid-IR W1-W2 color selection.

Only one third of X-ray obscured AGN show red optical-IR colors.

Different selection criteria yield substantially different samples of obscured AGN.

Abstract

Since its launch in 1999, the XMM-\textit{Newton} mission has compiled the largest catalogue of serendipitous X-ray sources, with the 3XMM being the third version of this catalogue. This is because of the combination of a large effective area (5000 $\rm cm^2$ at 1 keV) and a wide field of view (30 arcmin). The 3XMM-DR6 catalogue contains about 470,000 unique X-ray sources over an area of 982 $\rm deg^2$. A significant fraction of these (100,178 sources) have reliable optical, near/mid-IR counterparts in the SDSS, PANSTARRS, VIDEO, UKIDSS and WISE surveys. In a previous paper we have presented photometric redshifts for these sources using the TPZ machine learning algorithm. About one fourth of these (22,677) have adequate photon statistics so that a reliable X-ray spectrum can be extracted. Obviously, owing to both the X-ray counts selection and the optical counterpart constraint, the sample above is biased towards the bright sources. Here, we present XMMFITCAT-Z: a spectral fit catalogue for these sources using the Bayesian X-ray Analysis (BXA) technique. As a science demonstration of the potential of the present catalogue, we comment on the optical and mid-IR colours of the 765 X-ray absorbed sources with $N_\mathrm{H} > 10^{22}\,\mathrm{cm}^{-2}$. We show that a considerable fraction of X-ray selected AGN would not be classified as AGN following the mid-IR W1-W2 vs. W2 selection criterion. These are AGN with lower luminosities, where the contribution of the host galaxy to the MIR emission is non-negligible. Only one third of obscured AGN in X-rays present red colours or r-W2 > 6. Then it appears that the r-W2 criterion, often used in the literature for the selection of obscured AGN, produces very different X-ray absorbed AGN samples compared to the standard X-ray selection criteria.