2005MNRAS.360.1426D

We present an analysis of the X-ray properties of sources detected in the 13^HXMM-Newton deep (200-ks) field. In order to constrain the absorbed active galactic nuclei (AGN) population, we use extensive Monte Carlo simulations to compare directly the X-ray colours of observed sources with those predicted by several model distributions. In particular, we have carried out our comparisons over the entire 0.2-10 keV energy range of the XMM-Newton cameras, making our analysis sensitive to a large range of absorbing column densities. We have tested the simplest form of the unified scheme, whereby the intrinsic luminosity function of absorbed AGN is set to be the same as that of their unabsorbed brethren, coupled with various model distributions of absorption. Of the tested distributions, the best-fitting model has the fraction of AGN with absorbing column N_H, proportional to (logN_H)⁸. We have also tested two extensions to the unified scheme: an evolving absorption scenario, in which the fraction of absorbed sources is larger at higher redshifts, and a luminosity-dependent model in which high-luminosity AGN are less likely to be absorbed. Both of these models provide poorer matches to the observed X-ray colour distributions than the best-fitting simple unified model. We find that a luminosity-dependent density evolution luminosity function reproduces poorly the 0.5-2 keV source counts seen in the 13^H field. Field-to-field variations could be the cause of this disparity. Computing the X-ray colours with a simple absorbed power-law spectral model is found to overpredict, by a factor of 2, the fraction of hard sources that are completely absorbed below 0.5 keV, implying that an additional source of soft-band flux must be present in a number of the absorbed sources. The tested synthesis models predict that around 16 per cent of the detections in the 13^H field are due to AGN at z > 3. However, so far, only a single AGN with z > 3 has been identified in our approximately 50 per cent complete optical spectroscopy follow-up programme. Finally, we use our simulations to demonstrate the efficacy of a hardness ratio selection scheme at selecting absorbed sources for further study. Using this selection scheme, we show that around 40 per cent of the 13^H sample are expected to be AGN with N_H> 10²²/cm2.