Little Red Dots on FIRE: The Ability of Bursty Galaxies to Host an Abundant Population of High-Redshift AGN

Andrew Marszewski, Claude-André Faucher-Giguère, Guochao Sun, Daniel Anglés-Alcázar, Robert Feldmann, Kung-Yi Su, Tim B. Miller, Niranjan Chandra Roy

First listed 2026-01-29 | Last updated 2026-04-09

Abstract

The James Webb Space Telescope has unveiled an abundant population of potential active galactic nuclei (AGN) at high redshift ($z\gtrsim4$) known as little red dots (LRDs), which are likely hosted in relatively low-mass galaxies. However, previous theoretical models have highlighted the difficulty in continuously feeding massive black holes in the central regions of bursty, high-redshift galaxies because of repeated gas evacuation by stellar feedback. We analyze galaxies in high-redshift FIRE-2 simulations to understand whether they are capable of hosting the observed abundant population of high-redshift AGN. We use a gravitational torque-driven accretion (GTDA) model and a simple free-fall accretion model to derive black hole accretion rates and construct predicted AGN bolometric luminosity functions for $z=5-7$. The GTDA model and the free-fall model with black holes accreting $\lesssim 1$ percent of their central gas supply ($<100 \rm \ pc$) per free-fall time predict AGN abundances that are more than sufficient to explain the most recent LRD observations. The fiducial models, in fact, overpredict the number of low-luminosity AGN as compared with observations. We explore possible resolutions of this tension. A plausible, though likely not unique, scenario for alleviating the AGN overpredictions and which also provides a good match to the host-galaxy UV luminosity distribution suggests that LRDs are super Eddington-accreting, Eddington luminosity-limited, $M_{\rm BH}\gtrsim 2\times10^5 \ \rm M_\odot$ black holes residing in $M_\star\gtrsim 2\times10^7 \ \rm M_\odot$ galaxies. We show that, under simple assumptions, mock observations of such sources can reproduce key observed LRD characteristics.

Short digest

Uses FIRE-2 high‑z zoom-ins to infer black hole accretion from central gas using both a gravitational torque-driven accretion model and a simple free‑fall prescription, then builds z=5–7 AGN bolometric luminosity functions. Both models—already with ≤1% of gas inside ~100 pc accreted per free‑fall—produce AGN abundances sufficient for the JWST LRD counts, with fiducial setups overpredicting the faint end. A plausible reconciliation is that LRDs are super‑Eddington accretors capped at L_Edd, with MBH≳2×10^5 Msun in hosts with M⋆≳2×10^7 Msun, which also matches the host UV‑luminosity distribution. Under simple assumptions, mock observations reproduce key LRD traits; the proposed fix is not unique.

Key figures to inspect

• Figure 1: Inspect GTDA vs free‑fall BHAR time series to see burst‑driven on/off fueling cycles and how central gas evacuation sets short duty cycles at early times.
• Figure 2: Compare predicted AGN bolometric LFs from GTDA and free‑fall; note agreement at the bright bin when the free‑fall efficiency is <1% and the systematic excess at faint luminosities versus Greene et al. (2025) and pre‑JWST constraints.
• Figure 3: Check how adding lognormal variability scatter flattens the LF’s bright end via upscattering while leaving a persistent faint‑end overprediction—quantify which scatter levels best match the observed slope.
• Figure 4: Evaluate the “plausible LRD scenario”: select hosts with M⋆≳2×10^7 Msun and super‑Eddington accretors limited by L_Edd; verify simultaneous consistency with the LRD bolometric LF (Greene et al. 2025) and the LRD UVLF (Kokorev et al. 2024).