A deep Chandra survey of the Groth Strip - II. Optical identification of the X-ray sources

In this paper, we discuss the optical and X-ray spectral properties of the sources detected in a single 200-ks Chandra pointing in the Groth-Westphal Strip region. A wealth of optical photometric and spectroscopic data are available in this field providing optical identifications and redshift determinations for the X-ray population. The optical photometry and spectroscopy used here are primarily from the Deep Extragalactic Evolutionary Probe 2 (DEEP2) survey with additional redshifts obtained from the literature. These are complemented with the deeper (r approximate to 26 mag) multiwaveband data (ugriz) from the Canada-France-Hawaii Telescope Legacy Survey to estimate photometric redshifts and to optically identify sources fainter than the DEEP2 magnitude limit (R-A approximate to 24.5 mag). We focus our study on the 2-10 keV selected sample comprising 97 sources to the limit approximate to 8 x 10(-16) erg s(-1) cm(-2), this being the most complete in terms of optical identification rate (86 per cent) and redshift determination fraction (63 per cent; both spectroscopic and photometric). We first construct the redshift distribution of the sample which shows a peak at z approximate to 1. This is in broad agreement with models where less luminous active galactic nuclei (AGNs) evolve out to z approximate to 1 with powerful quasi-stellar objects (QSOs) peaking at higher redshift, z approximate to 2. Evolution similar to that of broad-line QSOs applied to the entire AGN population (both types I and II) does not fit the data. We also explore the observed N-H distribution of the sample and estimate a fraction of obscured AGN (N-H > 10(22) cm(-2)) of 48 +/- 9 per cent. This is found to be consistent with both a luminosity-dependent intrinsic N-H distribution, where less luminous systems comprise a higher fraction of type II AGNs and models with a fixed ratio 2:1 between types I and II AGNs. We further compare our results with those obtained in deeper and shallower surveys. We argue that a luminosity-dependent parametrization of the intrinsic N-H distribution is required to account for the fraction of obscured AGN observed in different samples over a wide range of fluxes.