The XMM Deep survey in the CDF-S

Context. X-ray spectroscopy of active galactic nuclei (AGN) offers the opportunity to directly probe the inner regions of the accretion disk. Reflection of the primary continuum on the circumnuclear accreting matter produces features in the X-ray spectrum that help to explore the physics and the geometry of the innermost region, close to the central black hole. Aims. We present the results of our analysis of average AGN XMM-Newton X-ray spectra in the Chandra Deep Field South observation (hereafter, XMM CDFS), in order to explore the Fe line features in distant AGN up to z ~ 3.5. Methods. We computed the average X-ray spectrum of a sample of 54 AGN with spectroscopic redshifts and signal-to-noise ratio in the 2–12 keV rest-frame band larger than 15 in at least one EPIC camera (for a total of 100 X-ray spectra and 181 623 net counts in the 2–12 keV rest-frame band). We have taken the effects of combining spectra from sources at different redshifts and from both EPIC-pn and EPIC-MOS cameras into account, as well as their spectral resolution; we checked our results using thorough simulations. We explored the iron line components of distant AGN focusing on the narrow core which arises from material far from the central BH and on the putative relativistic component emitted in the accretion disk. Results. The average spectrum shows a highly significant iron feature. Estimating its equivalent width (EW) with a model-independent method suggests a higher EW in a broader range. The line, modelled as an unresolved Gaussian, is significant at 6.8σ and has an EW = 95 ± 22 eV for the full sample. We find that our current data can be fitted equally welladding a relativistic profile to the narrow component (in the full sample, EW = 140 ± 120 eV and 67 ± 28 eV respectively for the relativistic and narrow lines). Conclusions. Thanks to the high quality of the XMM CDFS spectra and to the detailed modelling of the continuum and instrumental effects, we have shown that the most distant AGN exhibit a highly significant iron emission feature. It can be modelled both with narrow and broad lines which suggest that the EW becomes higher when a broader energy range around the line centroid is considered, provides tantalising evidence for reflection by material both very close and far away from the central engine. The EW of both features are similar to those observed in individual nearby AGN, hence they must be a widespread characteristic of AGN, since otherwise the average values would be smaller than observed