An obscured quasar census with the 4MOST IR AGN survey: design, predicted properties, and scientific goals

Carolina Andonie, David M. Alexander, Claire Greenwell, Sotiria Fotopoulou, Ryan Hickox, David J Rosario, Carolin Villforth, Johannes Buchner, Jens-Kristian Krogager, Brivael Laloux, Andrea Merloni, Mara Salvato, Ole Streicher, Wei Yan

First listed 2025-04-14 | Last updated 2025-04-14

Abstract

We present the 4MOST IR AGN survey, the first large-scale optical spectroscopic survey characterizing mid-infrared (MIR) selected obscured active galactic nuclei (AGN). The survey targets $\approx 212,000$ obscured infrared (IR) AGN candidates over $\approx 10,000 \rm \: deg^2$ down to a magnitude limit of $r_{\rm AB}=22.8 \, \rm mag$ and will be $\approx 100 \times$ larger than any existing obscured IR AGN spectroscopic sample. We select the targets using a MIR colour criterion applied to the unWISE catalogue from the WISE all-sky survey, and then apply a $r-W2\geq 5.9 \rm \: mag$ cut; we demonstrate that this selection will mostly identify sources obscured by $N_{\rm H}>10^{22} \rm \: cm^{-2}$. The survey complements the 4MOST X-ray survey, which will follow up $\sim 1\rm M$ eROSITA-selected (typically unobscured) AGN. We perform simulations to predict the quality of the spectra that we will obtain and validate our MIR-optical colour-selection method using X-ray spectral constraints and UV-to-far IR spectral energy distribution (SED) modelling in four well-observed deep-sky fields. We find that: (1) $\approx 80-87\%$ of the WISE-selected targets are AGN down to $r_{\rm AB}=22.1-22.8 \: \rm mag$ of which $\approx 70\%$ are obscured by $N_{\rm H}>10^{22} \: \rm cm^{-2}$, and (2) $\approx 80\%$ of the 4MOST IR AGN sample will remain undetected by the deepest eROSITA observations due to extreme absorption. Our SED fitting results show that the 4MOST IR AGN survey will primarily identify obscured AGN and quasars ($\approx 55\%$ of the sample is expected to have $L_{\rm AGN,IR}>10^{45} \rm \: erg \: s^{-1}$) residing in massive galaxies ($M_{\star}\approx 10^{10}-10^{12} \rm \: M_{\odot}$) at $z\approx 0.5-3.5$ with $ \approx 33\%$ expected to be hosted by starburst galaxies.

Short digest

Introduces the 4MOST IR AGN survey, a ~10,000 deg^2, rAB≤22.8 program targeting ≈212k mid-IR–selected obscured AGN via unWISE colors plus an r−W2≥5.9 mag cut. Using X-ray spectral constraints and UV–far-IR SEDs in four deep fields, they forecast 80–87% AGN purity with ≈70% obscured (NH>10^22 cm−2) and that ≈80% of targets will evade even the deepest eROSITA exposures. About 55% will be quasar-luminous (LAGN,IR>10^45 erg s−1) at z≈0.5–3.5 in massive (10^10–10^12 M⊙) hosts, with ≈33% in starbursts. This 100× expansion over prior spectroscopic obscured-IR samples will deliver a uniform census of heavily absorbed, X-ray–faint quasars complementing the 4MOST X-ray AGN survey.

Key figures to inspect

• Figure 1 — Selection workflow: confirms the unWISE color preselection, the r−W2≥5.9 obscuration cut, LS10 matching radius, and deep (r≤22.8) vs wide (r≤22.1) components; use it to reproduce the final target list logic.
• Figure 2 — Sky footprint and ancillary overlap: shows wide/deep tiling within the eROSITA-covered 4MOST area and where Herschel/Spitzer/Chandra/XMM/DESI add leverage; gauge environmental and SED follow-up potential.
• Figure 3 — r−W2 vs redshift with NH coding (COSMOS): visually verifies that r−W2≥5.9 tracks higher NH and illustrates any redshift-dependent leakage using both IRAC2 and unWISE W2.
• Figure 4 — NH distributions by color bin: quantifies the median NH shift and the obscured fraction gained by the r−W2 cut, providing the most direct evidence that the color criterion selects NH>10^22 cm−2 sources.