The space density of Compton-thick AGN at z~0.8 in the zCOSMOS-Bright Survey

The obscured accretion phase in BH growth is a key ingredient in many models linking the AGN activity with the evolution of their host galaxy. At present, a complete census of obscured AGN is still missing. The purpose of this work is to assess the reliability of the [NeV] emission line at 3426 A to pick up obscured AGN up to z~1 by assuming that [NeV] is a reliable proxy of the intrinsic AGN luminosity and using moderately deep X-ray data to characterize the amount of obscuration. A sample of 69 narrow-line (Type 2) AGN at z=0.65-1.20 were selected from the 20k-zCOSMOS Bright galaxy sample on the basis of the presence of the [NeV] emission. The X-ray properties of these galaxies were then derived using the Chandra-COSMOS coverage of the field; the X-ray-to-[NeV] flux ratio, coupled with X-ray spectral and stacking analyses, was then used to infer whether Compton-thin or Compton-thick absorption were present in these sources. Then the [NeV] luminosity function was computed to estimate the space density of Compton-thick (CT) AGN at z~0.8. Twenty-three sources were detected by Chandra, and their properties are consistent with moderate obscuration (on average, ~a few 10^{22} cm^-2). The X-ray properties of the remaining 46 X-ray undetected Type 2 AGN were derived using X-ray stacking analysis. Current data indicate that a fraction as high as ~40% of the present sample is likely to be CT. The space density of CT AGN with logL_2-10keV>43.5 at z=0.83 is (9.1+/-2.1) 10^{-6} Mpc^{-3}, in good agreement with both XRB model expectations and the previously measured space density for objects in a similar redshift and luminosity range. We regard our selection technique for CT AGN as clean but not complete, since even a mild extinction in the NLR can suppress [NeV] emission. Therefore, our estimate of their space density should be considered as a lower limit.