Can high-redshift AGN observed by JWST explain the EDGES absorption signal?

Alexandra Nelander, Christopher Cain, Jordan C J DSilva, Peter H Sims, Rogier A Windhorst, Judd D Bowman

First listed 2025-07-28 | Last updated 2025-07-28

Abstract

The Experiment to Detect the Global Epoch of Reionization 21 cm Signal (EDGES) has reported evidence for an absorption feature in the sky-averaged radio background near 78 MHz. A cosmological interpretation of this signal corresponds to absorption of 21 cm photons by neutral hydrogen at $z \sim 17$. The large depth of the signal has been shown to require an excess radio background above the CMB and/or non-standard cooling processes in the IGM. Here, we explore the plausibility of a scenario in which the EDGES signal is back-lit by an excess radio background sourced from a population of radio-loud AGN at high redshift. These AGN could also explain the unexpected abundance of UV-bright objects observed at $z > 10$ by JWST. We find that producing enough radio photons to explain the EDGES depth requires that nearly all high-$z$ UV-bright objects down to $M_{\mathrm{UV}} \gtrsim -15$ are radio-loud AGN and that the UV density of such objects declines by at most 1.5 orders of magnitude between $z = $10 and 20. In addition, the fraction of X-ray photons escaping these objects must be $\lesssim 1\%$ of their expected intrinsic production rate to prevent the absorption signal being washed out by early IGM pre-heating. Reproducing the sharp boundaries of the absorption trough and its flat bottom requires that the UV luminosity function, the fraction of UV light produced by AGN, and the X-ray escape fraction have fine-tuned redshift dependence. We conclude that radio-loud AGN are an unlikely (although physically possible) candidate to explain EDGES because of the extreme physical properties required for them to do so.

Short digest

Tests whether a high‑z radio‑loud AGN population—calibrated to JWST UV luminosity functions and extrapolated to z≈20—can provide the excess radio background needed for the EDGES 78 MHz absorption. Matching the EDGES depth requires that essentially all UV‑bright sources down to M_UV ≳ −15 are radio‑loud AGN, with the UV number density declining by ≤1.5 dex from z=10→20, and an X‑ray escape fraction ≲1% to avoid pre‑heating that erases the trough. Only an extremely shallow UVLF evolution can achieve the amplitude, and reproducing the trough’s sharp edges and flat bottom demands fine‑tuned redshift dependence of the UVLF, AGN UV fraction, and X‑ray leakage. Verdict: physically possible but implausible given the extreme requirements.

Key figures to inspect

• Figure 1: Compare the steep, shallow, and very‑shallow UVLF parameterizations and their UV luminosity densities to see how extrapolations to z≈20 and the faint‑end limit drive the available radio‑photon budget.
• Figure 2: Inspect the Lyα coupling coefficient split between Pop III and Pop II/AGN to verify that coupling is already strong by z≈17 and to see which component dominates in each UVLF model.
• Figure 3: Read the thermal histories to judge how different X‑ray escape fractions and UV densities advance IGM pre‑heating; the tracks illustrating f_esc,X ≲1% show how the gas stays cold enough through z≈17.
• Figure 4: Use the mK‑by‑z≈17 requirement map versus radio‑loud fraction and M_UV limit to pinpoint that only near‑unity radio‑loud fractions down to M_UV ≳ −15 (or the very‑shallow UVLF) reach the EDGES depth, highlighting the narrow, extreme corner of parameter space.