X-ray Surveys

People involved at OAB: Ciliegi, Comastri, Mignoli, Pozzetti, Zamorani.

The HELLAS2XMM survey

The HELLAS2XMM survey is a large national project carried out in collaboration with several Italian institutes (INAF-Arcetri and Rome Obs., Rome 3 Univ. and IASF-CNR, Milan). The final scientific aim of this project is the evaluation of an accurate luminosity function over a wide range of redshifts and luminosities for a large sample of hard X-ray selected sources detected in 15 public XMM-Newton observations, covering an area of about 4 square degrees at relatively shallow hard X-ray fluxes ( $F_{2-10\, \rm keV}>10^{-14}$ erg cm$^{-2}$ s$^{-1}$). This survey strategy has been designed to be complementary to deep pencil beam surveys by sampling a different portion of the luminosity-redshift plane in order to fill the gap between the local and deep surveys and to obtain a complete measure of the density and evolution of X-ray selected AGN.

To date, the most important results from the HELLAS2XMM project are mainly related to the optical identifications and spectroscopic classifications of a sample of 122 sources detected in $\sim$1 deg$^2$ (Fiore et al. 2003), obtained from an extensive campaign of multiwavelength observations and the full exploitation of the ESO and TNG facilities (e.g. EFOSC2 at the 3.6m telescope in La Silla; FORS2 at the VLT telescope; DOLORES at TNG). At the same time several multiwavelength (radio, sub-mm and near infrared) follow-up programs have been or are being carried out on a fraction of these fields. For two specific fields high spatial resolution Chandra X-ray data are also available. The optical campaigns have provided optical identifications and spectroscopic classifications for 97 out of the 122 hard X-ray selected sources; the spectroscopic completeness of the sample ($\sim$80%, mainly limited by the faintness of the optical counterparts, $R\lower.5ex\hbox{$\; \buildrel > \over \sim \;$}24$) is one of the highest for sources detected at flux level $\lower.5ex\hbox{$\; \buildrel > \over \sim \;$}10^{-14}$ erg cm$^{-2}$ s$^{-1}$.

Given the relatively bright X-ray fluxes sampled, an accurate X-ray spectral characterization was achievable for almost all of the sources in the sample. Out of these, 106 sources with 86% spectroscopic redshift were used to evaluate the fraction of X-ray absorbed (log N$_H>22$) AGN at these limiting fluxes, and to compare it with that predicted by X-ray background synthesis models (Perola et al. 2004).

A detailed study of a specific field (the PKS 0312$-$77 field) for which Chandra, radio and near-infrared data were also available, has clearly pointed out the importance of multiwavelength, high spatial resolution observations to fully characterize the physical properties of hard X-ray source counterparts (Brusa et al. 2003). The overall picture emerging from the optical identifications of the HELLAS2XMM sample indicate a wide spread in the optical (both in the continuum shape and emission lines) and X-ray properties of the sources responsible for the bulk of XRB energy density. The optical appearance of hard X-ray selected AGN is different from that expected on the basis of the AGN Unified Scheme, calling for some revisions to account for the discrepant classifications. In particular, the combination of X-ray spectral analysis and deep VLT spectroscopy has revealed that $\sim 10$% of high-redshift, high-luminosity objects optically classified as unobscured BL AGN are absorbed in the X-rays by column densities in excess of $10^{22}$ cm$^{-2}$ (Brusa et al. 2003; Perola et al. 2004).

What can be considered the most surprising and relevant finding of the HELLAS2XMM survey is the discovery of a population of presumably obscured AGN which are characterized by values of their X-ray (in the 2-10 keV band) to optical (integrated over the $R$ filter) flux ratio (X/O$>10$), significantly different from those of optically and soft X-ray selected, spectroscopically identified AGN.
We have provided, for the first time, spectroscopic identifications (using VLT/FORS2 data) of a sizable sample (13) of high X/O sources from the HELLAS2XMM sample. The majority of these turned out to be spectroscopically classified as high luminosity, high redshift narrow-line quasars (Fiore et al. 2003). Different approaches to estimate the redshift of optically faint X-ray sources have also been developed and tested for the sources detected in the deepest Chandra exposures: in particular, it was possible to derive the redshift of highly obscured objects by detecting a strong Fe K$\alpha$ line in a few high S/N X-ray spectra of sources detected in the deepest Chandra fields (Comastri, Brusa & Civano 2004).

Combining the results from the HELLAS2XMM survey with those obtained from deep, pencil-beam Chandra and XMM-Newton surveys, it was possible to build a ``virtually complete'' sample of identified hard X-ray sources over a wide range of redshifts and luminosities. We have clearly pointed out the need for a luminosity-dependent density evolution for the sources responsible for the XRB (low luminosity sources peaking at a later cosmic time), to explain the observed number and luminosity densities as a function of redshift (Fiore et al. 2003). With this assumption, and the hypothesis that hard X-ray surveys probe the largest fraction of the whole AGN population, it is possible to explain the observed local BH mass density as entirely due to the growth of AGN (Comastri 2003).

For a selected subsample of 10 high X/O sources in the HELLAS2XMM survey, undetected in the optical band, deep VLT/ISAAC imaging has been obtained: almost all of the observed sources (10/11) were detected in the $K$-band and their optical to near-infrared colors are considerably redder than those of the field galaxy population, all of them being classified as Extremely Red Objects (Mignoli et al. 2004). Thanks to the quality of the near-infrared data a detailed analysis of the surface brightness profiles allows us to classify all of the near-infrared counterparts: the majority of these (7) have been classified as elliptical (bulge) galaxies, two as pointlike objects and one source with an exponential profile (disky). The results of the fit of the extended sources are reported in Fig. 10. None of the extended sources show any evidence for the presence of a central unresolved object tracing the putative X-ray emitting AGN.

Figure 10: Surface brightness profiles along the major axis (boxes) and best-fitting models (lines) for all of the extended sources in the HELLAS2XMM sample for which VLT/ISAAC observations have been obtained. In the upper left box we also show with a dashed line a representative PSF.

Coupling the morphological information derived from $K$-band surface brightness profiles, it was possible to derive a ``minimum'' redshift for these objects ranging between 0.8 and 2.4; in particular, the elliptical hosts have $z_{\rm min}=0.9$-1.4. The X-ray spectral properties of these sources indicate that heavy (N$_H > 10^{22}$ cm$^{-2}$) obscuration is almost ubiquitous among objects with high X/O and that obscured sources (in particular QSO2, the high-luminosity, high-redshift obscured AGNs predicted in XRB synthesis models) can be hosted in the bulge of luminous, massive ellipticals which already formed the bulk of their stars at high redshifts.

This work is carried out in collaboration with M. Brusa (Astronomy Dept., University of Bologna).

XMM-Newton observations of Extremely Red Objects

Hard X-ray observations of Extremely Red Objects (EROs) with XMM-Newton have been carried out in one of the largest samples of EROs available to date (Daddi et al. 2000), selected in a contiguous area of $\sim 380$ arcmin$^{2}$. We have obtained a total of 100 ks of XMM-Newton data, split between two observations. This work provides the first comprehensive characterization of the X-ray properties of a large sample of individually detected EROs. It was possible to estimate for the first time, over a large statistically significant sample of objects, the fraction and the X-ray spectral properties of AGN among EROs. The results obtained from the cleaned 80 ks XMM-Newton observation suggest that, at the relatively bright X-ray and near-infrared fluxes probed by this observation, AGN account only for a negligible fraction (3%) of the optically selected ERO population (Brusa 2003).

The X-ray and optical properties of individually detected EROs are similar to those of Quasars 2, the high-luminosity, high redshift type II AGNs predicted in XRB synthesis models. The somewhat unexpected link between EROs and QSO2, is intriguing: near-infrared observations of obscured QSOs selected on the basis of their high X/O (e.g. from the HELLAS2XMM sample, see above) and, conversely, hard X-ray observations of a complete sample of EROs, provide some of the strongest evidence that these two populations, originally discovered at different wavelengths, are intimately connected.

A selection criterion based on the X/O and the $R-K$ colour of hard X-ray selected sources has been proposed, to efficiently pick up the elusive QSO2 population, which is difficult to select at optical wavelengths. Furthermore, X-ray detected EROs can be used as lighthouses to investigate the accretion paradigm at high redshifts, in order to address the issue of elliptical galaxy formation and the expected co-evolution with accreting black holes. Our work clearly indicates that the combination of near-infrared and X-ray observations is a powerful tool to select this so-far elusive population of luminous, obscured X-ray sources. These results have allowed to obtain Chandra data (90 ks) on the ERO field, to push the X-ray analysis to fainter fluxes.

This work is carried out in collaboration with M. Brusa (Astronomy Dept., University of Bologna).