← Week 36, 2026

2608.28745v1

Super-Eddington Little Blue Dots May Reionize Helium Too Early

Theme match 3/5

Christopher Cain, Brent Smith, Gibson B. Bowling, Yongda Zhu, Lily Whitler, Rafael Ortiz, Anson D'Aloisio, Rogier Windhorst

First listed 2026-09-01 | Last updated 2026-08-28

Abstract

The James Webb Space Telescope (JWST) has identified an abundant population of faint Active Galactic Nuclei (AGN) at $z \gtrsim 4$ which display broad emission lines, compact morphologies, and blue ultraviolet--optical continua. These ``Little Blue Dots' (LBDs) have been suggested to belong to the same family of objects as the more controversial ``Little Red Dots' (LRDs), with differences between the two largely owing to viewing angle effects. In this scenario, super-Eddington accretion resulting in high Extreme UV (EUV) and weak X-ray emission is invoked to explain the properties of both populations. We study the consequences of this super-Eddington scenario for the timing of Helium reionization. We find that observations which support an end to helium reionization no earlier than $z \approx 3$ disfavor scenarios in which the majority of observed $3 \lesssim z \lesssim 7$ LBDs are highly super-Eddington. For our fiducial super-Eddington accretion scenario, we find that the fraction of the LBD population in this state must be $\lesssim 10\%$, assuming LBDs make up $5\%$ of the $M_{ m UV} < -18$ galaxy population and have average escape fractions of $15\%$, comparable to recent observations. Our constraint assumes that LBDs dominate the Helium reionization budget, and would be tighter if bright quasars also contributed significantly. For a majority of LBDs to be super-Eddington, they would need to have small escape fractions ($\lesssim 1.5\%$) and/or be less abundant than observations suggest. Our conclusions are sensitive to the shape of the EUV spectra of super-Eddington black holes, motivating further study.

Short digest

Cain et al. test whether the blue, broad-line JWST AGN population dubbed Little Blue Dots can predominantly host the EUV-bright, X-ray-weak super-Eddington accretion flows proposed in orientation-based unification pictures of LBDs and Little Red Dots. Using the observed faint broad-line AGN UV luminosity function, super-Eddington spectral models, and a He II reionization calculation, they find that completing helium reionization no earlier than z≈3 limits the fiducial highly super-Eddington fraction to ≲10% if LBDs comprise 5% of MUV<-18 galaxies and have a 15% ionizing escape fraction. A majority-super-Eddington interpretation would instead require He II-ionizing escape fractions ≲1.5% or a substantially rarer LBD population, with the constraint importantly dependent on the uncertain EUV spectral shape of super-Eddington black holes.

Key figures to inspect

  • Figure 1. This establishes the empirical population normalization: a roughly 5% broad-line AGN fraction among galaxies down to the adopted UV-magnitude cutoff is consistent with the measured UV luminosity functions, setting the LBD abundance used in the reionization calculation.
  • Figure 2. This is the physical linchpin of the paper, showing why low-mass soft-state super-Eddington black holes can produce He II-ionizing photons much more efficiently than a conventional quasar power-law SED while remaining X-ray weak.
  • Figure 3. The reionization histories directly connect the assumed super-Eddington SEDs to the observational timing constraint, illustrating how efficient EUV production can drive He II reionization to finish too early unless the super-Eddington contribution is restricted.
  • Figure 4. This is the headline constraint figure: it maps the combinations of super-Eddington LBD fraction and ionizing escape fraction disfavored by the requirement that He II reionization not complete before z≈3, including the fiducial ≲10% result.
  • Figure 5. This comparison tests whether the conclusion survives alternative EUV-bright accretion prescriptions, showing that the exclusion strength depends materially on the assumed black-hole mass and spectral state rather than on a single soft-state model.

Discussion

Log in to view the paper discussion, see votes, and leave your own feedback.