UNCOVER: Candidate Red Active Galactic Nuclei at 3<z<7 with JWST and ALMA

TL;DR

This work identifies a preliminary sample of 26 compact red sources in the Abell 2744 UNCOVER field, likely reddened accreting supermassive black holes at 3<z_phot<7, using a JWST/NIRCam F444W<27.7 mag selection. The sources are blue in F115W–F200W and red in F200W–F444W, with point-like central components; ALMA 1.2 mm data show no detections individually or in stacks. SED modeling favors hot dust emission over obscured star formation to explain the red optical colors and ALMA non-detections, suggesting very small sizes (~50 pc after magnification) and bolometric luminosities of 10^43–10^46 erg/s (faint in rest-frame UV: M_1450 ~ -14 to -18). If confirmed as AGN, these reddened luminous sources imply a substantial, obscured high-redshift AGN population, potentially >10× more numerous than UV-luminous AGN at the same bolometric luminosity.

Abstract

The James Webb Space Telescope (JWST) is revolutionizing our knowledge of $z>5$ galaxies and their actively accreting black holes. Using the JWST Cycle 1 Treasury program Ultradeep NIRSpec and NIRCam ObserVations before the Epoch of Reionization (UNCOVER) in the lensing field Abell 2744, we report the identification of a sample of little red dots at $3 < z_{\rm{phot}} < 7$ that likely contain highly-reddened accreting supermassive black holes. Using a NIRCam-only selection to F444W$<27.7$ mag, we find 26 sources over the $\sim45$ arcmin$^{2}$ field that are blue in F115W$-$F200W$\sim0$ (or $β_{\rm UV}\sim-2.0$ for $f_λ \propto λ^β$), red in F200W$-$F444W = $1-4$ ($β_{\rm opt} \sim +2.0$), and are dominated by a point-source like central component. Of the 20 sources with deep ALMA 1.2-mm coverage, none are detected individually or in a stack. For the majority of the sample, SED fits to the JWST+ALMA observations prefer models with hot dust rather than obscured star-formation to reproduce the red NIRCam colors and ALMA 1.2-mm non-detections. While compact dusty star formation can not be ruled out, the combination of extremely small sizes ($\langle r_e \rangle\approx50$ pc after correction for magnification), red rest-frame optical slopes, and hot dust can by explained by reddened broad-line active galactic nuclei (AGNs). Our targets have faint $M_{\rm 1450} \approx -14\ \, {\rm to} -18$ mag but inferred bolometric luminosities of $L_{\rm bol} = 10^{43}-10^{46}$ erg/s, reflecting their obscured nature. If the candidates are confirmed as AGNs with upcoming UNCOVER spectroscopy, then we have found an abundant population of reddened luminous AGN that are at least ten times more numerous than UV-luminous AGN at the same intrinsic bolometric luminosity.

Figures (2)

• Figure 1: Schematic model (right) and photometric fit (left) for compact red sources like CEERS 3210. The model consists of a reddened, but not completely obscured, broad-line AGN template ($red$) representing a direct view of the accretion disk and Broad Line Region, a small contribution from the same AGN template without reddening ($blue$) that we posit represents scattered light, and (3) an unreddened Narrow Line Region with emission line strengths coupled to the broad $H\alpha$ of the red component following typical correlations for broad-line AGN Stern:2012aStern:2012b. This fit not only provides a decent fit to the photometry, but predicts both broad and narrow line strengths very similar to those observed with NIRSpec.
• Figure 2: Color-color (left) and color-compactness (right) selections. These panels show all the sources selected with the "red2" color and compactness criteria (red circles), which is specifically designed to select $z>6$ AGN similar to those found in Furtak:2022Kocevski:2023. Sources selected to be PSF-dominated are highlighted with black dots. In greyscale, we show the entire UNCOVER catalog, with numbers of objects indicated by the color-bar. The compact red sources are clear outliers in color-color-compactness space.