Light, heavy, primordial: exploring the diversity of black hole seeding and growth mechanisms in the JWST era

Pratika Dayal

First listed 2026-04-22 | Last updated 2026-04-22

Abstract

The James Webb Space Telescope (JWST) has revealed a puzzling population of massive black holes in the first billion years, many of which are over-massive compared to their hosts (obese black holes), and reside in metal-poor hosts, posing a challenge for theoretical models at these early epochs. In this work, we compare the observational properties of astrophysically-seeded black holes using the DELPHI semi-analytic model and cosmologically-seeded primordial black holes (PBHs) using the PHANES analytic model. We explore the growth of light ($\sim 100 M_\odot$) and heavy ($\sim 10^{3-5}M_\odot$) seeds through mergers and accretion (both Eddington-limited and at super-Eddington rates) in the astrophysical scenario; PBHs (seeded between $10^{0.5-6}M_\odot$) only grow through accretion at sub-Eddington rates. Comparing to observables at $z \sim 5-10$, the only model that can be ruled out is the one where we allow Eddington-limited accretion onto light seeds. The observed high values of the black hole mass-stellar mass relation ($0.3-1$) can be reproduced by both PBHs and heavy seeds accreting at super-Eddington rates. However, only the PBH and Eddington-limited heavy seeding models can simultaneously reproduce the observed black hole masses (${\rm M_{bh}}$), stellar masses ($M_*$), and extremely low host metallicities ($Z \leq 0.01 Z_\odot$) inferred at $z \sim 7-10$. Crucially, we find PBHs show decrease in the black hole mass-stellar mass ratio with increasing halo mass at all redshifts, contrary to any astrophysical black hole model. Selecting systems at $z \sim 7$ with ${\rm M_{bh}}/M_* > 0.1$ and bolometric luminosities $\sim 10^{44-46} {\rm erg~s^{-1}}$ that show a negative black hole to stellar mass ratio and reside in $10^{9-11}M_\odot$ halos offer a promising clustering-based discriminant of PBH seeding models.

Short digest

Dayal cross-compares astrophysical light (~100 M⊙) and heavy (~10^3–10^5 M⊙) seeds (via DELPHI) with cosmological primordial black holes (10^{0.5–6} M⊙; PHANES) against z≈5–10 observables. The only scenario the data rule out is Eddington-limited accretion onto light seeds; high black hole–to–stellar mass ratios (Mbh/M*≈0.3–1) are achievable with PBHs or heavy seeds under super-Eddington growth. Crucially, only PBHs and Eddington-limited heavy seeds can simultaneously recover the observed Mbh, M*, and extremely low host metallicities (Z≤0.01 Z⊙) at z≈7–10. PBHs also predict a distinctive decline of Mbh/M* with increasing halo mass and occupancy of 10^9–10^11 M⊙ halos, motivating a clustering-based test selecting z≈7 systems with Mbh/M*>0.1 and Lbol≈10^{44–46} erg s⁻¹.

Key figures to inspect

• Figure 1: Compare model BH mass functions across z to see why Eddington-limited light seeds underproduce the high-mass tail relative to JWST-era points, while PBH/heavy-seed tracks match counts at the massive end.
• Figure 2: Inspect where bolometric luminosity functions diverge (Lbol≈10^{44–46} erg s⁻¹); this is the regime proposed for selection tests and shows whether PBH or heavy-seed channels can reproduce the bright-end space densities.
• Figure 3: Read the Mbh–M* panels by halo-mass color to verify Mbh/M*≈0.3–1 loci and, uniquely for PBHs, the negative trend of Mbh/M* with increasing halo mass at all redshifts.
• Figure 4: Use metallicity versus Mbh to check which models reach Z≤0.01 Z⊙ at z≈7–10; compare to Abell2744-QSO1, CANUCS-LRD-z8.6, and GHZ9 points/limits to see the PBH and Eddington-limited heavy-seed agreement.