Investigating the formation channel of GW231123: Population III stars or hierarchical mergers?

Ataru Tanikawa; Emanuele Berti; Federico Angeloni; Konstantinos Kritos; Luca Graziani; Michela Mapelli; Raffaella Schneider; Stefano Torniamenti

arxiv: 2604.18709 · v2 · pith:FVWOBHSSnew · submitted 2026-04-20 · 🌌 astro-ph.GA · astro-ph.CO· gr-qc

Investigating the formation channel of GW231123: Population III stars or hierarchical mergers?

Federico Angeloni , Konstantinos Kritos , Raffaella Schneider , Emanuele Berti , Luca Graziani , Stefano Torniamenti , Michela Mapelli , Ataru Tanikawa This is my paper

Pith reviewed 2026-05-25 06:19 UTC · model grok-4.3

classification 🌌 astro-ph.GA astro-ph.COgr-qc

keywords gravitational wavesblack hole mergershierarchical mergersglobular clusterspopulation synthesispair-instability gapGW231123merger rate density

0 comments

The pith

GW231123 formed via successive black hole mergers inside dense globular clusters rather than isolated binaries.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper tests whether the massive black holes in GW231123 arose from isolated binary evolution, Population III stars, or repeated mergers in clusters. It couples a cosmological galaxy formation simulation to a cluster population code and runs two separate binary evolution models to follow candidate systems from birth to merger inside the same simulated volume. Isolated channels produce either orbits too wide to coalesce within cosmic time or events only at metallicities that contribute almost nothing to the overall star formation rate. The work therefore concludes that dynamical hierarchical assembly inside metal-poor globular clusters accounts for the observed masses, the high spins, and the inferred merger redshift.

Core claim

Coupling the GAMESH galaxy formation model with the RAPSTER cluster population synthesis code and comparing the SEVN and BSEEMP binary population synthesis codes shows that isolated binary evolution cannot reproduce the merger redshift of GW231123-like events, while successive mergers in dense globular clusters naturally generate the component masses and are consistent with the reported dimensionless spins.

What carries the argument

Self-consistent coupling of the GAMESH cosmological simulation with the RAPSTER cluster synthesis code to reconstruct the full life cycle of candidate systems inside one simulated volume and compare isolated versus dynamical channels.

If this is right

Isolated binary evolution fails to produce GW231123-like mergers at the observed redshift in both tested population synthesis codes.
Successive mergers inside dense globular clusters reproduce the component black hole masses lying in or above the pair-instability gap.
The high dimensionless spins reported for the event align with the expectations for a hierarchical merger population.
The local merger rate density reaches 0.78 Gpc^{-3} yr^{-1} and peaks at redshift 4-6, tracing the formation epoch of metal-poor globular clusters.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

Repeated detection of similar events could map the redshift evolution of metal-poor globular cluster formation across cosmic time.
The same framework could be applied to other high-mass events to test whether a single channel dominates the upper end of the black hole mass spectrum.
If future observations tighten the upper limit on the local rate of such events, the required globular cluster formation efficiency in the simulation would need revision.

Load-bearing premise

The overdense simulated volume and its metallicity distribution accurately represent the cosmic average contribution of extremely low-metallicity star formation.

What would settle it

Detection of a GW231123-like merger at a redshift where the isolated low-metallicity channel would produce a coalescence, or a direct measurement showing that the cosmic star formation rate density at Z = 10^{-10} is orders of magnitude higher than the simulation predicts.

Figures

Figures reproduced from arXiv: 2604.18709 by Ataru Tanikawa, Emanuele Berti, Federico Angeloni, Konstantinos Kritos, Luca Graziani, Michela Mapelli, Raffaella Schneider, Stefano Torniamenti.

**Figure 1.** Figure 1: (a) Cosmological evolution of the total star formation rate density (black line), the Pop III star formation rate density (green [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗

**Figure 2.** Figure 2: (a) BBH number density as a function of hierarchical merger generation for the primary (red) and secondary (blue) [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗

read the original abstract

The gravitational wave event GW231123, with component black hole masses lying within or above the pair-instability mass gap, poses a significant challenge to current stellar evolution models. In this work, we describe how we investigated its origin by coupling the galaxy formation model GAMESH with the cluster population synthesis code RAPSTER and using two distinct binary population synthesis codes, SEVN and BSEEMP. This framework allowed us, for the first time, to reconstruct the life cycle of GW231123-like candidates within the same cosmological simulation, enabling a self-consistent comparison between different formation channels. Although both population synthesis codes can in principle produce black holes compatible with GW231123, we find that isolated binary evolution fails to reproduce the inferred merger redshift. In SEVN, massive black hole binaries form with semi-major axes > 10^3 Rsun , preventing coalescences within a Hubble time. In BSEEMP, candidates arise only at extremely low metallicities (Z = 10^{-10}), which contribute negligibly to the star formation rate density in our overdense simulated volume. Our results therefore strongly support a dynamical hierarchical origin. The observed black hole masses are naturally reproduced through successive mergers in dense globular clusters. The high dimensionless spins reported by the LIGO-Virgo-KAGRA Collaboration are consistent with this hierarchical population. We found a local merger rate density of 0.78 Gpc^{-3} yr^{-1}, with a peak at z = 4 - 6, tracing the maximum formation rate of globular clusters in metal-poor environments (Z = 0.006). Overall, GW231123 may represent a benchmark event for a robust population of hierarchical black holes formed in the early Universe.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

The coupled simulation shows isolated channels failing for this event while hierarchical works, but the overdense volume's low-Z tail may not represent the cosmic average.

read the letter

The paper runs GAMESH plus RAPSTER with both SEVN and BSEEMP inside one cosmological volume to track GW231123-like binaries from formation to merger. That self-consistent setup is the real advance; most prior work compares channels across separate models or volumes. They report that SEVN produces binaries too wide to merge in time and BSEEMP only yields candidates at Z=10^{-10}, a metallicity slice that contributes almost nothing to the star-formation rate density in their box. The hierarchical channel in clusters reproduces the masses and can match the reported spins, and they quote a local rate of 0.78 Gpc^{-3} yr^{-1} peaking at z=4-6. That is concrete output from the run, not a fit to the event itself. The weak point is the volume choice. An overdense region can easily under-sample the rare, pristine gas parcels that dominate extremely low-metallicity star formation. If the true cosmic fraction of Z~10^{-10} gas is higher, the BSEEMP isolated channel could produce observable rates and the dismissal would not hold. The two codes also disagree on where viable candidates appear, which is informative but underscores that the result is tied to the specific low-Z physics in each code. The work is worth referee time because the method is new and the question is sharp, even if the volume representativeness needs tighter justification or a larger set of realizations. A reader interested in GW source modeling or cluster formation at z>4 would get value from the comparison.

Referee Report

1 major / 0 minor

Summary. The paper couples the GAMESH cosmological galaxy formation simulation with the RAPSTER code for globular cluster populations and employs two binary population synthesis codes (SEVN and BSEEMP) to trace the formation of GW231123-like black hole binaries. It reports that isolated binary evolution fails to produce viable candidates at the observed merger redshift: SEVN yields binaries with semi-major axes >10^3 Rsun, while BSEEMP produces candidates only at Z=10^{-10}, a metallicity whose star-formation-rate density is negligible inside the simulated overdense volume. The work concludes that the event is best explained by successive hierarchical mergers in dense clusters, reports a local merger rate density of 0.78 Gpc^{-3} yr^{-1} peaking at z=4-6, and states that the high spins are consistent with this channel.

Significance. If the central claim holds, the manuscript supplies a self-consistent cosmological framework that directly compares isolated and dynamical channels for a single high-mass event within the same simulated volume. The reported rate and redshift peak constitute a falsifiable prediction for the hierarchical population that can be confronted with future LIGO-Virgo-KAGRA data releases.

major comments (1)

[Abstract] Abstract and methods description of GAMESH + RAPSTER coupling: the dismissal of the BSEEMP isolated channel rests on the claim that Z=10^{-10} star formation contributes negligibly inside the single overdense simulated volume. No test is presented showing that the metallicity distribution (or its extremely metal-poor tail) in this volume is representative of the cosmic average; an overdense region can systematically under- or over-produce the rare pristine gas parcels that dominate such low-Z formation. This assumption is load-bearing for the conclusion that isolated evolution fails in the real universe.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for their careful reading of the manuscript and for highlighting this important caveat regarding the representativeness of the simulated volume. We address the major comment below and will revise the manuscript to incorporate the suggested qualification.

read point-by-point responses

Referee: [Abstract] Abstract and methods description of GAMESH + RAPSTER coupling: the dismissal of the BSEEMP isolated channel rests on the claim that Z=10^{-10} star formation contributes negligibly inside the single overdense simulated volume. No test is presented showing that the metallicity distribution (or its extremely metal-poor tail) in this volume is representative of the cosmic average; an overdense region can systematically under- or over-produce the rare pristine gas parcels that dominate such low-Z formation. This assumption is load-bearing for the conclusion that isolated evolution fails in the real universe.

Authors: We agree that this is a substantive limitation. The GAMESH simulation targets an overdense region to efficiently capture the assembly of massive structures and globular clusters, but we have not explicitly validated that the extremely metal-poor tail (Z = 10^{-10}) matches the cosmic average. In the revised manuscript we will add a dedicated paragraph in the Methods and Discussion sections that (i) states the conclusion on the isolated channel applies within the simulated overdense volume, (ii) notes that an overdense environment could in principle alter the abundance of pristine gas parcels, and (iii) qualifies the broader claim that isolated evolution fails in the real universe. We will also reference existing high-redshift metallicity distribution studies to provide context. This change does not affect the robustness of the hierarchical-merger results, which are driven by the overall globular-cluster formation rate rather than the rarest metallicity tail. revision: yes

Circularity Check

0 steps flagged

No circularity: simulation outputs are independent of target event

full rationale

The paper's central results are direct numerical outputs from coupling the GAMESH cosmological simulation to RAPSTER cluster evolution and running SEVN/BSEEMP binary population synthesis on the resulting star-formation and metallicity histories. The statement that BSEEMP produces GW231123-like systems only at Z=10^{-10} (which is negligible in the simulated volume) follows from executing the code on the simulated SFRD; the local rate 0.78 Gpc^{-3} yr^{-1} is likewise a computed integral over the simulated population. No equation defines a quantity in terms of itself, no fitted parameter is relabeled as a prediction, and no self-citation chain is used to justify uniqueness or an ansatz. The representativeness of the overdense volume is an external modeling assumption, not a self-referential reduction.

Axiom & Free-Parameter Ledger

1 free parameters · 2 axioms · 0 invented entities

The central claim rests on the accuracy of two binary population synthesis codes at metallicities down to 10^{-10}, the mapping from simulated overdense volume to cosmic averages, and standard assumptions about globular cluster formation efficiency in metal-poor gas. No new particles or forces are introduced.

free parameters (1)

local merger rate normalization
The reported value 0.78 Gpc^{-3} yr^{-1} is an output of the specific simulation volume and cluster formation prescription rather than an external constraint.

axioms (2)

domain assumption The metallicity distribution and star-formation history inside the simulated overdense volume are representative for assessing the cosmic contribution of Z=10^{-10} gas.
Invoked when concluding that BSEEMP candidates at Z=10^{-10} contribute negligibly.
domain assumption SEVN and BSEEMP correctly capture binary evolution and common-envelope physics at the relevant metallicities and masses.
Central to the claim that isolated binaries cannot merge in time.

pith-pipeline@v0.9.0 · 5876 in / 1829 out tokens · 42204 ms · 2026-05-25T06:19:56.620919+00:00 · methodology

Review history (2 revisions) →

discussion (0)

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

Black Hole Binary Detection Landscape for the Laser Interferometer Lunar Antenna (LILA): Signal-to-Noise Calculations & Science Cases
astro-ph.HE 2026-05 unverdicted novelty 5.0

LILA can detect IMBH binaries at redshifts 20-30, IMRIs, and provide months-to-years early warnings with high-SNR events for gravity tests.

Reference graph

Works this paper leans on

2 extracted references · 2 canonical work pages · cited by 1 Pith paper · 1 internal anchor

[1]

Milky Way Globular Clusters: Nurseries for Dynamically-Formed Binary Black Holes

Abac, A. G., Abouelfettouh, I., Acernese, F., et al. 2025, ApJ, 993, L25 Adamo, A., Zeidler, P., Kruijssen, J. M. D., et al. 2020, Space Sci. Rev., 216, 69 Angeloni, F., Kritos, K., Schneider, R., et al. 2025, arXiv e-prints, arXiv:2512.11059 Arca Sedda, M., Kamlah, A. W. H., Spurzem, R., et al. 2023, MNRAS, 526, 429 Banerjee, S. 2026, arXiv e-prints, arX...

work page internal anchor Pith review Pith/arXiv arXiv 2025
[2]

This would e ffectively increase the fraction of star formation available for their production within our simulated over-dense volume

- could broaden the metallicity window for potential GW231123 candidates. This would e ffectively increase the fraction of star formation available for their production within our simulated over-dense volume. Recent studies exploring the Pop III channel for GW231123 provide some guidance on how variations in the adopted stellar physics or initial conditio...

work page 2025

[1] [1]

Milky Way Globular Clusters: Nurseries for Dynamically-Formed Binary Black Holes

Abac, A. G., Abouelfettouh, I., Acernese, F., et al. 2025, ApJ, 993, L25 Adamo, A., Zeidler, P., Kruijssen, J. M. D., et al. 2020, Space Sci. Rev., 216, 69 Angeloni, F., Kritos, K., Schneider, R., et al. 2025, arXiv e-prints, arXiv:2512.11059 Arca Sedda, M., Kamlah, A. W. H., Spurzem, R., et al. 2023, MNRAS, 526, 429 Banerjee, S. 2026, arXiv e-prints, arX...

work page internal anchor Pith review Pith/arXiv arXiv 2025

[2] [2]

This would e ffectively increase the fraction of star formation available for their production within our simulated over-dense volume

- could broaden the metallicity window for potential GW231123 candidates. This would e ffectively increase the fraction of star formation available for their production within our simulated over-dense volume. Recent studies exploring the Pop III channel for GW231123 provide some guidance on how variations in the adopted stellar physics or initial conditio...

work page 2025