Supermassive stars as the origin of the multiple populations in globular clusters

2; (2) ICREA; 3); (3) Surrey; (4) Geneva; 5) ((1) Barcelona; (5) CNRS/Toulouse); Corinne Charbonnel (4; Mark Gieles (1

REVIEW 3 major objections 4 minor 18 references

Supermassive stars as the origin of the multiple populations in globular clusters

T0 review · 3 major / 4 minor · reviewed 2026-08-14 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Supermassive star polluters can explain globular clusters' multiple populations.

desk verdict A clear proceedings summary of the authors' 2018 SMS pollution model; it usefully names the three constraints but adds no new derivation and hangs on the unverified assumption of a fully convective supermassive star. read the letter →

arxiv 1908.02075 v1 pith:M7ABIGZX submitted 2019-08-06 astro-ph.GA astro-ph.SR

Mark Gieles (1 , 2 , 3) , Corinne Charbonnel (4 , 5) ((1) Barcelona , (2) ICREA , (3) Surrey , (4) Geneva

show 1 more author

(5) CNRS/Toulouse)

This is my paper · ORCID

classification astro-ph.GAastro-ph.SR

keywords globularclustersmultiplepopulationssupermassivestarsself-enrichmentstellarcollisionshothydrogenburningheliumabundancemassbudget

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper proposes that the puzzling multiple stellar populations seen in most old globular clusters were not produced by a separate second generation of stars but by a single supermassive star (above $10^3$ solar masses) that formed at the cluster centre through stellar collisions during the formation starburst. The authors argue that such a star can eject enough hot-hydrogen-burning processed material in its wind to account for the large fractions of anomalous stars, solving the long-standing 'mass budget problem.' Because the star is continuously rejuvenated by fresh collisions, the ejected helium abundance stays relatively low, matching the small helium spreads observed in most clusters. The model also predicts that more massive clusters produce more processed material per unit of cluster mass, which matches the observed trend of larger helium spreads and higher polluted-star fractions in more massive clusters. The paper presents this as a self-enrichment scenario that meets three empirical constraints that previous polluter models (AGB stars or massive stars) have not simultaneously satisfied.

What carries the argument

The central object is a supermassive star (SMS), defined here as a star with mass above $10^3\,M_\odot$, formed by successive stellar collisions at the centre of a collapsing proto-globular cluster. The key mechanism is the 'conveyer-belt' behaviour: the SMS is assumed to be fully convective, so nuclear burning products from hot hydrogen burning (the CNO, NeNa, and MgAl cycles) are transported to the surface and lost in a wind, while collisions and accretion continue to add fresh, unprocessed fuel. The wind mass-loss rate balances the growth rate, so the SMS keeps ejecting processed material until the cluster's dynamical contraction ends. Because the SMS's surface is continuously replenished by fresh fuel, the helium abundance of the ejected gas stays relatively low. The model couples this stellar physics to cluster dynamics: the cluster contracts, triggers collisions, and the confinement of the wind by infalling gas determines how much processed material is retained.

What would settle it

A stellar evolution calculation that grows a SMS by repeated collisions rather than smooth accretion and finds a radiative envelope with no efficient surface transport would undercut the conveyor-belt mechanism. Equivalently, if observations of high-redshift GC formation sites showed no abundance anomalies despite the presence of SMS winds, the model's predicted enrichment would be contradicted.

Watch

Extended reading notes

Core claim

The central claim is that a supermassive star growing by runaway stellar collisions in a dense, gas-rich proto-globular cluster can reach an equilibrium between mass growth and wind mass loss, acting as a 'conveyer belt' that converts pristine infalling gas into hot-hydrogen-burning products and ejects them into the intra-cluster medium. The processed material mixes with the still-infalling cold gas, is diluted, and then accretes onto existing stars or forms new stars, giving the observed anomalous light-element abundances. The authors show that the amount of ejected processed material can be a significant fraction of the total cluster mass, overcoming the mass-budget problem, and that the ratio of processed mass to cluster mass increases with cluster mass, overcoming the specific mass-budget problem. They also argue that continuous rejuvenation of the SMS by collisions keeps the helium abundance of the yields low, in line with the empirically inferred modest helium spreads. The scenario is intended to apply across metallicities because the SMS formation mechanism is dynamical rather than metallicity-dependent.

Load-bearing premise

The model relies on the SMS being fully convective, so that nuclear burning products are quickly brought to the surface and ejected; if a radiative envelope develops during collisional growth, the enrichment mechanism stops unless additional mixing such as rotation-induced mixing transports the products.

Editorial extensions

If this is right

If correct, globular cluster multiple populations can form within a single starburst event, eliminating the need for the delayed second generation invoked by previous models.
The predicted superlinear scaling of processed mass with cluster mass directly explains why the fraction of polluted stars and the helium spread increase with cluster mass.
The model predicts a natural connection between globular cluster formation and intermediate-mass black holes, since the SMS may collapse or explode after its brief life.
The scenario applies at all metallicities, since collision-driven SMS formation is dynamical, so it can explain why multiple populations appear in both metal-poor and metal-rich globular clusters.
Observations of extremely massive stars at high redshift ($z\sim3-6$) provide a test bed for whether stars above $10^3\,M_\odot$ form in the dense environments the model requires.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

The same conveyor-belt logic could in principle apply to other dense star-forming systems such as nuclear star clusters, predicting abundance anomalies there if SMSs form.
If the SMS is fully convective only under specific conditions, the model points to a need for improved simulations of collisional stellar growth, including mixing processes.
The model's success would strengthen the idea that globular clusters formed in converging gas flows, linking GC formation to cosmological filamentary structure.
One testable extension: the model predicts a relation between present-day cluster mass and the minimum helium spread, which could be checked against large samples of GCs once homogeneous photometry and spectroscopy are available.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

This is a conference proceedings paper, not a new research article. It summarizes the supermassive-star (SMS) pollution scenario that Gieles et al. published in detail in 2018, and it does that summary job well. The opening list of the three constraints — mass budget, specific mass budget, and helium — is genuinely useful as a compact statement of why other polluter models struggle. The authors are also honest: they explicitly flag the fully convective requirement as a reliance, and they cite the Haemmerlé et al. result that accreting SMSs likely have radiative envelopes. That is more candid than many papers twice this length.

What is actually new is the packaging, not the physics. The central model, the quantitative example in Figure 2, and the yield numbers all come from the 2018 paper. Calling the SMS a 'conveyer-belt' is a nice image, and the claim that the processed mass per unit cluster mass grows with cluster mass is an important reframing, but it is a reframing of an existing result.

The soft spot is exactly where the stress test points: the whole enrichment channel requires the SMS to be fully convective for its entire active lifetime. The paper says this in Section 3 and then notes that the available stellar evolution models predict a radiative envelope. The fallback to rotation-induced mixing is a hand-wave, not a demonstrated mechanism. If the envelope is radiative, the processed material stays inside the star, the wind carries unprocessed gas, and the 10^5 solar masses of processed material for a 10^6 solar mass cluster simply is not produced. That is a load-bearing uncertainty. There are also the expected free parameters — wind mass-loss rate, SMS temperature, mass-radius relation — but those are secondary. The convection assumption is primary.

For a proceedings piece, this is acceptable. It is a useful status report, and it sends the reader to the 2018 paper for the actual derivation. But as a standalone research submission, it does not deserve full peer review: no new evidence, no new calculation, and the key mechanism is explicitly unresolved. I would not cite this proceedings over the original MNRAS paper in my own work.

The right call: if this is meant as a citable conference summary, accept as is; if it is submitted as original research, desk reject and tell the authors to make the radiative-envelope case a falsifiable prediction rather than a parenthetical caveat.

Referee Report

3 major / 4 minor

Summary. This proceedings paper proposes that a supermassive star (SMS, >10^3 Msun), formed by stellar collisions during the early gas-rich assembly of a globular cluster, can act as the polluter responsible for the observed multiple populations. The SMS is assumed to be fully convective, so that hot-hydrogen burning products are transported to its surface and ejected in a wind that mixes with the surrounding gas and later accretes onto stars or forms new stars. The authors claim that this scenario overcomes the mass-budget problem, the specific mass-budget problem, and the helium problem, and they illustrate the model with a quantitative example taken from their earlier work (Gieles et al. 2018). The paper closes with a list of open questions and observational tests, explicitly acknowledging that the SMS must be fully convective and that accreting-SMS models predict a radiative envelope.

Significance. If the scenario holds, it offers a single self-enrichment channel that simultaneously addresses three long-standing empirical constraints on globular cluster multiple populations: the large mass of processed material, its mass-dependent scaling, and the relatively small helium spread. The paper is valuable in clearly formulating these constraints and in proposing a concrete dynamical context for SMS formation. Its main strength is that the authors are explicit about the assumptions and uncertainties, especially in Section 3. However, the present paper contains no new quantitative derivation; the central claims rely on the companion paper Gieles et al. (2018), and the load-bearing convective-envelope assumption is directly contradicted by at least one published stellar-evolution calculation cited by the authors. As a proceedings contribution, the paper is a useful discussion piece, but its central assertion is not yet supported to the level claimed in the abstract.

major comments (3)

[§3, SMS structure] The entire enrichment mechanism depends on the SMS being 100% convective, as stated in Section 3. The paper immediately acknowledges that stellar evolution models of accreting SMSs (Haemmerlé et al. 2018) predict a radiative envelope, and it offers only a speculative fallback to 'other transport processes like rotation-induced mixing.' Because the processed mass and helium predictions shown in Figure 2 and the mass-budget claims are all computed under the convective assumption in Gieles et al. (2018), a radiative envelope would invalidate the central claim. Please either demonstrate with a concrete model that collision-grown SMSs remain fully convective, or quantify the processed mass under a radiative-envelope model with rotation-induced mixing; without this, the mass-budget and helium conclusions are unsupported.
[§2, Figure 2] The quantitative claims, including the statement that a 10^6 Msun cluster produces ~10^5 Msun of processed material and that the specific processed mass increases with cluster mass, are imported from Gieles et al. (2018) rather than derived here. The invoked SMS temperature of 40 kK is described as 'conservative' but no sensitivity analysis is shown, and the shaded regions in Figure 2 only bracket the wind mass-loss rate. As written, the abstract statement that the model 'overcomes the mass-budget problem' is stronger than what is demonstrated in this paper. Please either provide a self-contained derivation or clearly state the parameter ranges (temperature, mass-radius relation, mass-loss rate) over which the conclusion holds.
[§2, specific mass budget and dilution] The paper claims that the amount of processed material per unit cluster mass increases with cluster mass because massive clusters reach higher densities and expand more slowly (step vi), but the reader is not shown the scaling or its robustness; Figure 2 presents a single example. In addition, the model assumes that the SMS wind mixes with the inflowing pristine gas and then accretes, yet the dilution ratio, which is known to control the observed abundance spreads, is not specified or constrained. Please provide the relevant scaling argument and a quantitative treatment of the dilution process, or explicitly mark these as untested free parameters of the scenario.

minor comments (4)

[§2, step (iv)] There is a typographical mismatch in the parenthetical citation: '(see Sakurai et al. 2017)’' has an extra closing quote and an unmatched opening mark; the sentence should end with 'Sakurai et al. 2017).'
[Author affiliations] The University of Surrey address appears with a spacing error in the postcode: 'GU 2 7XH' should presumably be 'GU2 7XH'.
[Throughout] The symbol '/greaterorsimilar' appears as plain text in several places; the final typeset version should use the proper LaTeX symbol (e.g., $\gtrsim$) for consistency and readability.
[Figure 2 caption] The caption uses 'mSMS' and 'mwind' while the text uses variables like $m_{\rm SMS}$ and $M_c$; please unify the notation between text and figure.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the model's quantitative outputs are computed from dynamical and stellar assumptions, not defined as their inputs.

full rationale

The paper is a proceedings contribution that summarizes a model developed in the companion paper Gieles et al. (2018). The central claims—that a collision-formed supermassive star can produce a significant mass of processed material, that this mass correlates with cluster mass, and that rejuvenation lowers the helium abundance—are presented as outputs of a physical model combining stellar collisions, wind mass loss, cluster dynamics, and nucleosynthetic yields. The empirical constraints (abundance anomalies, He spreads, mass trends) are external inputs used to motivate and test the model, not quantities the model defines into existence. No equation in the text reduces to an input by construction, and no fitted parameter is renamed as a prediction: the 40 kK SMS temperature is explicitly a stated assumption, and the wind mass-loss results are shown as a range bounded by different assumptions rather than tuned to match the observed mass-budget requirement. The fully convective assumption is explicitly flagged as an assumption and is even noted to be challenged by Haemmerlé et al. (2018); this is an acknowledged physical uncertainty, not a definitional or self-referential step. The citation of Gieles et al. (2018) is the expected reference to the full model and does not constitute circularity: the prior paper contains an independent derivation rather than merely asserting the target conclusion. The paper does not invoke any uniqueness theorem, nor does it smuggle in an ansatz by citation without stating it. Consequently, the claimed derivation chain is not circular.

Assumptions & free parameters 3 free parameters · 6 assumptions · 0 invented entities

The quantitative model is not derived in this paper; it is presented in a companion paper by the same authors (Gieles et al. 2018). The central claim depends on several assumed parameters (SMS temperature, wind mass-loss rate, mass-radius relation) and on the assumption of full convection, which the authors themselves note is contested by accreting-SMS evolution models. No free parameters are fitted to data in this proceedings paper; the scaling with cluster mass comes from dynamical arguments. No fundamentally new entities are introduced; the supermassive star is a known hypothetical object applied to a new context.

free parameters (3)

SMS effective temperature = 40 kK (assumed constant, conservative case)
Chosen by hand for the Fig. 2 example; affects the wind mass-loss rate and thus the amount of processed material.
SMS wind mass-loss rate = not specified; bounded by shaded regions in Fig. 2
Uncertain parameter; directly sets the mass of processed material released. The paper uses different assumptions for the wind mass-loss rates.
SMS mass-radius relation = not specified
Mentioned as uncertain in Section 3; determines the SMS radius and hence its wind properties and growth balance.

assumptions (6)

domain assumption Globular clusters form at the intersection of gas filaments with high inflow rates (~0.1 Msun/yr).
Used in model steps (i)-(ii) and Fig. 1; motivated by cosmological zoom simulations (Li et al. 2017), not demonstrated for GCs.
domain assumption During gas accretion the stellar density increases as rho ~ M^10, quickly boosting the collision rate.
Invoked in step (iii) to justify rapid collision-driven SMS formation; scaling from Bonnell et al. (1998).
domain assumption Runaway collisions in the dense cluster center form a single supermassive star.
Step (iv), based on Portegies Zwart et al. (2004); this is a hypothesis, not confirmed by observations.
domain assumption The SMS is fully convective, so burning products reach the surface.
Section 3 explicitly states the model relies on this; contradicted by accreting-SMS models (Haemmerlé et al. 2018) that find a radiative envelope.
domain assumption The SMS wind collides with inflowing pristine gas, shock cools, and accretes onto existing or new stars.
Step (vii) and Fig. 1; the shock-cooling and accretion efficiency are not quantified here.
domain assumption Two-body relaxation halts contraction and SMS-star binaries drive cluster expansion, with the timing depending on cluster mass.
Steps (v)-(vi); this produces the claimed super-linear correlation between processed mass and cluster mass.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Supermassive stars as the origin of the multiple populations in globular clusters." pith.science (2026). https://pith.science/paper/M7ABIGZX

@misc{pith2026190802075,
  author       = {Pith},
  title        = {Pith review of: Supermassive stars as the origin of the multiple populations in globular clusters},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/M7ABIGZX}},
  note         = {Machine review of arXiv:1908.02075}
}

read the original abstract

Globular clusters (GCs) display anomalous light element abundances (HeCNONaMgAl), resembling the yields of hot-hydrogen burning, but there is no consensus yet on the origin of these ubiquitous multiple populations. We present a model in which a super-massive star (SMS, >10^3 Msun) forms via stellar collisions during GC formation and pollutes the intra-cluster medium. The growth of the SMS finds a balance with the wind mass loss rate, such that the SMS can produce a significant fraction of the total GC mass in processed material, thereby overcoming the so-called mass-budget problem that plagues other models. Because of continuous rejuvenation, the SMS acts as a `conveyer-belt' of hot-hydrogen burning yields with (relatively) low He abundances, in agreement with empirical constraints. Additionally, the amount of processed material per unit of GC mass correlates with GC mass, addressing the specific mass budget problem. We discuss uncertainties and tests of this new self-enrichment scenario.

Figures

Figures reproduced from arXiv: 1908.02075 by the authors.

**Figure 1.** Schematic picture of the enrichment scenario presented in § 2. Cold, pristine gas accretes onto the stars in the cluster, causing the cluster to contract. The higher stellar density results in stellar collisions, forming a SMS in the cluster centre. The SMS blows a wind enriched in hot-hydrogen burning products, which interacts and mixes with the inflowing gas. The diluted material subsequently accretes onto the sta… view at source ↗

**Figure 2.** Model for SMS and GC formation. Gas accretion and cluster contraction starts at t = 0 and the first massive stars reach the ZAMS at t = 2 Myr and the first supernova goes off at 5 Myr. The left panels show the evolution of the cluster mass (Mc, top) and half-mass radius (Rh, bottom), while the right panels show the evolution of the SMS mass (mSMS, top) and the amount of mass released in the wind (mwind, bottom). The… view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

18 extracted references · 12 canonical work pages

[1]

2019, MNRAS, 486, 5008

Antonini, F., Gieles, M., & Gualandris, A. 2019, MNRAS, 486, 5008

2019
[2]

2018, ARAA, 56, 83

Bastian, N., & Lardo, C. 2018, ARAA, 56, 83

work page 2018
[3]

A., Bate, M

Bonnell, I. A., Bate, M. R., & Zinnecker, H. 1998, MNRAS, 298, 93

work page 1998
[4]

A., & Hartwick, F

Denissenkov, P. A., & Hartwick, F. D. A. 2014, MNRAS, 437, L21

2014
[5]

Gieles, M., Charbonnel, C., Krause, M. G. H., et al. 2018, MNRAS, 478, 2461 Haemmerl´ e, L., Woods, T. E., Klessen, R. S., et al. 2018, MNRAS, 474, 2757

work page 2018
[6]

Y., Gnedin, N

Li, H., Gnedin, O. Y., Gnedin, N. Y., et al. 2017, ApJ, 834, 69

work page 2017
[7]

2018, MNRAS, 473, 2688

Martocchia, S., Cabrera-Ziri, I., Lardo, C., et al. 2018, MNRAS, 473, 2688

2018
[8]

P., Marino, A

Milone, A. P., Marino, A. F., Dotter, A., et al. 2014, ApJ, 785, 21

work page 2014

Show all 18 references

[9]

P., Piotto, G., Renzini, A., et al

Milone, A. P., Piotto, G., Renzini, A., et al. 2017, MNRAS, 464, 3636

2017
[10]

P., Marino, A

Milone, A. P., Marino, A. F., Renzini, A., et al. 2018, MNRAS, 481, 5098

2018
[11]

Moeckel, N., & Clarke, C. J. 2011, MNRAS, 410, 2799

2011
[12]

P., Bedin, L

Piotto, G., Milone, A. P., Bedin, L. R., et al. 2015, AJ, 149, 91 Portegies Zwart, S. F., Baumgardt, H., Hut, P., et al. 2004, Nature, 428, 724

2015
[13]

2017, A&A, 608, A28

Prantzos, N., Charbonnel, C., & Iliadis, C. 2017, A&A, 608, A28

2017
[14]

S., et al

Sakurai, Y., Yoshida, N., Fujii, M. S., et al. 2017, MNRAS, 472, 1677

2017
[15]

J., Crowther, P

Smith, L. J., Crowther, P. A., Calzetti, D., et al. 2016, ApJ, 823, 38

2016
[16]

2018, ApJ, 862, 16

Tremou, E., Strader, J., Chomiuk, L., et al. 2018, ApJ, 862, 16

2018
[17]

2019, MNRAS, 483, 3618

Vanzella, E., Calura, F., Meneghetti, M., et al. 2019, MNRAS, 483, 3618

2019
[18]

2019, MNRAS, 482, 4713

Zocchi, A., Gieles, M., & H´ enault-Brunet, V. 2019, MNRAS, 482, 4713

2019

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.

[1] [1]

2019, MNRAS, 486, 5008

Antonini, F., Gieles, M., & Gualandris, A. 2019, MNRAS, 486, 5008

2019

[2] [2]

2018, ARAA, 56, 83

Bastian, N., & Lardo, C. 2018, ARAA, 56, 83

work page 2018

[3] [3]

A., Bate, M

Bonnell, I. A., Bate, M. R., & Zinnecker, H. 1998, MNRAS, 298, 93

work page 1998

[4] [4]

A., & Hartwick, F

Denissenkov, P. A., & Hartwick, F. D. A. 2014, MNRAS, 437, L21

2014

[5] [5]

Gieles, M., Charbonnel, C., Krause, M. G. H., et al. 2018, MNRAS, 478, 2461 Haemmerl´ e, L., Woods, T. E., Klessen, R. S., et al. 2018, MNRAS, 474, 2757

work page 2018

[6] [6]

Y., Gnedin, N

Li, H., Gnedin, O. Y., Gnedin, N. Y., et al. 2017, ApJ, 834, 69

work page 2017

[7] [7]

2018, MNRAS, 473, 2688

Martocchia, S., Cabrera-Ziri, I., Lardo, C., et al. 2018, MNRAS, 473, 2688

2018

[8] [8]

P., Marino, A

Milone, A. P., Marino, A. F., Dotter, A., et al. 2014, ApJ, 785, 21

work page 2014