New Methods of Identifying AGN in the Early Universe using Spectroscopy and Photometry in the JWST Era

Flor Arevalo Gonzalez, Titanilla Braun, James Trussler, Christopher J. Conselice, Thomas Harvey, Nathan Adams, Duncan Austin, Qiong Li, Ignas Juodžbalis, Kimihiko Nakajima

First listed 2025-01-16 | Last updated 2025-01-16

Abstract

We explore spectroscopic and photometric methods for identifying high-redshift galaxies containing an Active Galactic Nucleus (AGN) with JWST observations. After demonstrating the limitations of standard optical methods, which appear ineffective in the low-metallicity environment of the early universe, we evaluate alternative diagnostic techniques using the current JWST observational capabilities. Our analysis focuses on line ratios and equivalent widths (EWs) of UV emission lines: CIV, HeII $λ$1640, OIII] $λ$1665, and CIII], and the faint optical line, HeII $λ$4686. We find that the most valuable diagnostic quantities for finding AGN are the line ratios: (CIII] + CIV) / HeII $λ$1640 and CIII] / HeII $λ$1640, as well as the EW of HeII $λ$1640. For more reliable AGN identification, the HeII $λ$1640 and OIII] $λ$1665 lines would need to be detected separately. We show that the HeII $λ$1640/H$β$ ratio effectively separates AGN from star-forming galaxies, though it is contingent on a low dust content. We also show that in order to effectively use these diagnostics, future observations require longer exposure times, especially for galaxies at $z > 6$. Subsequently, we plot three real high-redshift sources on these diagrams which present strong UV emission lines. However, in order to classify them as strong AGN candidates, further study is needed due to the blending of HeII + OIII] and unreliable optical lines. Lastly, we carry out a selection process using spectral energy distribution (SED) fitting with EAZY to identify strong AGN candidates in the JADES NIRCam photometry. One galaxy in our sample emerged as a strong AGN candidate, supported by both photometric selection and strong UV emission. We present a sample of similar AGN candidates in the JADES data based on this method.

Short digest

This paper stress-tests UV-line diagnostics and NIRCam SED selection to find early AGN, showing that classic optical BPT/VO87 fail in low-metallicity regimes. Using photoionization grids, the most effective discriminants emerge as (CIII]+CIV)/HeII λ1640, CIII]/HeII λ1640, and the EW of HeII λ1640, with HeII λ1640/Hβ also separating AGN from star-formers when dust is low. Applied to three JADES/NIRSpec sources (IDs 9422, 18846, 10058975), the diagnostics are promising but current spectra blend HeII λ1640 with OIII] λλ1661,1666 and optical lines are unreliable, limiting firm AGN claims. An EAZY-based NIRCam SED search flags one strong candidate with strong UV emission, and the authors argue longer integrations—especially at z>6—are needed to cleanly deblend key lines.

Key figures to inspect

• UV diagnostic planes: (CIII]+CIV)/HeII λ1640 versus CIII]/HeII λ1640 with model tracks and the JADES NIRSpec sources 9422, 18846, 10058975 overplotted—check where they sit relative to AGN vs star-forming loci and how metallicity/ionization shift the grids.
• Rest-UV NIRSpec spectra around 1500–1700 Å (rest-frame): inspect the HeII λ1640 + OIII] λλ1661,1666 complex to see the degree of blending and any attempted deblends; assess whether higher resolution/exposure would separate the lines.
• HeII λ1640/Hβ diagnostic: a plot showing the separation between AGN and SFG and the vectors for dust attenuation—verify that low dust is required for a clean divide and note where the three JADES sources land.
• EAZY SED-fit panel for the NIRCam-selected candidate: examine the best-fit with an AGN component, photometric redshift PDF, and which filters capture the UV line excesses supporting the AGN interpretation.
• Sensitivity/exposure-time figure: predicted 5σ limits or EW thresholds for HeII λ1640 and OIII] λ1665 versus redshift—use this to gauge required integrations at z>6 to robustly apply the proposed diagnostics.