2022ApJ...941..191L

We present a comprehensive census of the active galactic nuclei (AGNs) in the GOODS-S/HUDF region from the X-ray to the radio, covering both the obscured and unobscured populations. This work includes a robust analysis of the source optical to mid-IR spectral energy distributions (SEDs) featuring (semi)empirical AGN and galaxy dust emission models and Bayesian fitting techniques, ultra-deep Very Large Array 3 and 6 GHz observations, and an integrated analysis of various AGN selection techniques, including X-ray properties, UV to mid-IR SED analysis, optical spectral features, mid-IR colors, radio-loudness and spectral slope, and AGN variability. In total, we report ∼900 AGNs over the ∼170 arcmin² 3D-HST GOODS-S footprint, which has doubled the AGN number identified in the previous X-ray sample, with ∼26% of our sample undetected in the deepest Chandra image. With a summary of AGN demographics from different selection methods, we find that no one single band or technique comes close to selecting a complete AGN sample despite the great depth of the data in GOODS-S/HUDF. We estimate the yields of various approaches and explore the reasons for incompleteness. We characterize the statistical properties, such as source number density, obscuration fraction, and luminosity function of the AGN sample in this field, and discuss their immediate implications. We also provide some qualitative predictions of the AGN sample that might be discovered by the upcoming JWST surveys.