2004ApJ...616..123T

The deep X-ray, optical, and far-infrared fields that constitute the Great Observatories Origins Deep Survey (GOODS) are sensitive to obscured active galactic nuclei (AGNs; N_H≳10²²/cm2) at the quasar epoch (z∼2-3), as well as to unobscured AGNs as distant as z∼7. Luminous X-ray emission is a sign of accretion onto a supermassive black hole and thus reveals all but the most heavily obscured AGNs. We combine X-ray luminosity functions with appropriate spectral energy distributions for AGNs to model the X-ray, optical, and far-infrared flux distributions of the X-ray sources in the GOODS fields. A simple model based on the unified paradigm for AGNs, with ∼3 times as many obscured AGNs as unobscured, successfully reproduces the z-band flux distributions measured in the deep Hubble Space Telescope ACS observations of the GOODS-North and GOODS-South fields. This model is also consistent with the observed spectroscopic and photometric redshift distributions once selection effects are considered. The previously reported discrepancy between observed spectroscopic redshift distributions and the predictions of population synthesis models for the X-ray background can be explained by a bias against the most heavily obscured AGNs generated by both X-ray observations and the identification of sources via optical spectroscopy. We predict the AGN number counts for Spitzer MIPS 24 µm and IRAC 3.6-8 µm observations in the GOODS fields, which will verify whether most AGNs in the early universe are obscured in the optical. Such AGNs should be very bright far-infrared sources and include some obscured AGNs missed even by X-ray observations.