UNCOVER: A NIRSpec Identification of a Broad Line AGN at z = 8.50

TL;DR

This study reports a secure broad-line AGN detection at z=8.50 in JWST/NIRSpec data, marked by a broad Hβ component (FWHM = 3439 ± 413 km/s) and high-ionization lines within a point-source morphology. The derived black hole mass is log10(M_BH/M_sun)=8.17±0.42 with a bolometric luminosity of ~6.6×10^45 erg s−1, corresponding to ~40% of the Eddington limit, while SED and ALMA constraints place an upper limit on the host stellar mass log10(M*/M_sun)<8.7, implying BH-to-host mass ratios of at least ~30%. This finding supports scenarios where SMBHs at high redshift originate from massive seeds or experience super-Eddington growth, and underscores the role of dust-obscured, UV-faint AGN in early SMBH assembly. The results demonstrate JWST/NIRSpec's power to probe galaxy–black hole co-evolution in the early universe and motivate larger samples to map BH growth across cosmic time.

Abstract

Deep observations with JWST have revealed an emerging population of red point-like sources that could provide a link between the postulated supermassive black hole seeds and observed quasars. In this work we present a JWST/NIRSpec spectrum from the JWST Cycle 1 UNCOVER Treasury survey, of a massive accreting black hole at $z=8.50$, displaying a clear broad-line component as inferred from the H$β$ line with FWHM = $3439\pm413$ km s$^{-1}$, typical of the broad line region of an active galactic nucleus (AGN). The AGN nature of this object is further supported by high ionization, as inferred from emission lines, and a point-source morphology. We compute the black hole mass of log$_{10}(M_{\rm BH}/M_\odot)=8.17\pm0.42$, and a bolometric luminosity of $L_{\rm bol}\sim6.6\times10^{45}$ erg s$^{-1}$. These values imply that our object is accreting at $\sim 40\%$ of the Eddington limit. Detailed modeling of the spectral energy distribution in the optical and near-infrared, together with constraints from ALMA, indicate an upper limit on the stellar mass of log$_{10}(M_{\rm *}/M_\odot)<8.7$, which would lead to an unprecedented ratio of black hole to host mass of at least $\sim 30 \%$. This is orders of magnitude higher compared to the local QSOs, but is consistent with recent AGN studies at high redshift with JWST. This finding suggests that a non-negligible fraction of supermassive black holes either started out from massive seeds and/or grew at a super-Eddington rate at high redshift. Given the predicted number densities of high-$z$ faint AGN, future NIRSpec observations of larger samples will allow us to further investigate the galaxy-black hole co-evolution in the early Universe.

Figures (1)

• Figure 1: Top:JWST/NIRCam 1$.\!\!^{\prime\prime}$5 stamps and the RGB color image comprised of the F277W, F356W and F444W bands. The MSA slitlet layout is highlighted in white. An unambiguous point-like morphology of ID: 20466 can be observed in all filters. On each panel we show total magnitudes, with $1\sigma$ uncertainty, as presented in the UNCOVER photometric catalog of weaver23. Middle: 2D MSA PRISM spectrum produced by msaexp. We optimally scaled the trace to highlight all the significant line detections. Bottom: A collapsed 1D spectrum of our source. We show the data in black, while the uncertainty on the spectrum is in red. Assuming the best-fit msaexp$z_{\rm spec}=8.502\pm0.003$ we show the positions of all the prominent emission with significant ($>3\sigma$) detections as solid vertical lines. Emission for which we only obtain an upper limit are shown with dashed lines.