{"id":"52c235ef-b528-4885-9583-cb204b046517","arxiv_id":"2501.06037","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":7,"one_line_summary":"Globular cluster dynamics create both in-cluster and ejected ultraluminous X-ray sources, with escapers about one-seventh of the GC-born population and neutron-star escapers twice as common as neutron-star ULXs inside clusters.","lead":"Using simulations of old star clusters, this study maps when and where ultraluminous X-ray sources form, and how many get kicked out before they shine. It finds that globular clusters mostly make black-hole ULXs while young, and that ejected \"escaper\" ULXs may explain some apparently field ULXs.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Quantitative ULX fractions rest on an unbeamed luminosity model; NS ULXs sit near threshold, so beaming can materially shift the reported 96/4 and escaper ratios.","rationale":"The paper is a careful MOCCA simulation study with released data; the qualitative conclusion that GC dynamics both create and eject ULX progenitors and that old, tidally-filling clusters are poor hosts is supported by the figures and is largely independent of the beaming choice. The reader's verdict of conditional acceptance is well matched to the evidence. My stress-test focuses on the quantitative headline (96% BH, 4% NS in-cluster, 1:7 escaper ratio, 40% NS among escapers), which is defined by the LX>10^39 threshold in Eq. (1). The weakness is that NS ULXs in the simulations sit just above threshold (median log10 LX_max ~39.22, Table 6), so the NS fraction is a knife-edge prediction of the unbeamed, eta0.1=1 luminosity model. The paper explicitly defers beaming to future work (Section 4.4) and cites evidence that most NS ULXs are beamed; a beaming-corrected post-processing of the released data could materially change the numbers. I therefore agree with the reader's weakest_assumption and recommend keeping the conditional verdict, with the concrete check above as a condition for the quantitative claims to be used as observational predictions.","tokens_in":25216,"tokens_out":6558,"duration_ms":64121,"concrete_test":"Download the released Zenodo data (doi:10.5281/zenodo.14953837) and recompute E[NULX] and the NS/BH and escaper/in-cluster fractions after applying a standard beaming correction to every system with L > L_Edd: apparent luminosity L_app = L / b with b = (L/L_Edd)^-2 (King 2009) and detection weight multiplied by b. If the in-cluster NS fraction changes from ~4% to >10% or the escaper:in-cluster NS ratio moves outside ~1.5–2.5, the headline ratios are beaming-dominated and the quantitative claims in the abstract and Conclusions should be revised.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central quantitative claims—96% of in-cluster ULXs have BH accretors, 4% NS; escapers are ~1/7 of all GC-born ULXs and ~2:1 over in-cluster NS ULXs—are all threshold classifications: a system counts as a ULX only if the X-ray luminosity from Eq. (1) exceeds 10^39 erg/s. Eq. (1) assumes isotropic emission with accretion efficiency eta0.1=1 and no beaming. This is not a peripheral detail: Table 6 shows that NS ULXs sit at median log10 LX,max ~ 39.22, only ~0.2 dex above the threshold, while BH ULXs are ~1 dex brighter. Any beaming correction—which the authors themselves state will be needed and cite Wiktorowicz et al. (2019) for the claim that most NS ULXs are beamed—can move a substantial fraction of NS systems across the threshold (or change their visibility weight), directly altering every ratio derived from threshold membership. Since the paper's observational predictions are precisely these ratios, the unbeamed prescription is the load-bearing assumption for the paper's central claim. The qualitative trends (dynamics can both create and eject ULXs; old and tidally-filling clusters have fewer in-cluster ULXs) are likely robust to this choice, but the headline numbers are not.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents the first numerical population-synthesis study of ultraluminous X-ray sources (ULXs) in globular clusters (GCs), using the MOCCA Monte Carlo code with BSE stellar/binary evolution. It explores how initial binary fraction, tidal filling, and multiple stellar populations affect the formation and evolution of in-cluster and dynamically ejected (\"escaper\") ULXs. The principal claims are that approximately 96% of in-cluster ULXs have black hole accretors, that escaper ULXs constitute about one-seventh of all GC-born ULXs but are nearly twice as common as in-cluster ULXs when the accretor is a neutron star, and that the scarcity of ULXs in old/tidally filling Milky Way GCs is explained by old stellar populations and tidal stripping.","tokens_in":25493,"tokens_out":6207,"duration_ms":58297,"significance":"If the conclusions hold, this is a useful first step toward predicting GC ULX populations and quantifying the contamination of field ULXs by cluster ejecta. The qualitative picture—dynamics both creates and ejects ULX progenitors, NS accretors are more prominent among escapers, and tidally filling clusters are less prolific—is coherent and likely robust. The paper also ships publicly available simulation data, which is commendable. However, the headline quantitative ratios are threshold-defined and rest on an unbeamed luminosity model plus small numbers of rare NS ULXs, so the specific percentages should be treated with caution until the sensitivity is quantified.","major_comments":[{"comment":"The X-ray luminosity prescription in Eq. (1) assumes isotropic emission and an accretion efficiency η0.1=1, with no beaming; Section 4.4 explicitly defers beaming to future work. Because the headline claims (96% BH / 4% NS in-cluster, the 1:7 escaper ratio, and the 2:1 NS escaper ratio) are all classifications based on the 10^39 erg/s threshold, and because NS ULXs in Table 6 have median log10 LX,max ≈ 39.22, only about 0.2 dex above threshold, a beaming correction can move a substantial fraction of NS systems across the threshold or change their detection weight. The authors should test the sensitivity of the reported fractions to beaming, for example by applying the King (2009) or Wiktorowicz et al. (2019) prescription to the simulated mass-transfer histories, or by quoting how the ratios change under limiting beaming assumptions.","section":"Section 4.4, Eq. (1)"},{"comment":"The text states that in-cluster ULXs have a median count of 176 and that escapers are one-seventh of the total ULX population, but Table 6 gives median counts of 7 (NS) + 142.5 (BH) = 149.5 for in-cluster and 12.5 (NS) + 19.0 (BH) = 31.5 for escapers, corresponding to a ratio of about 1:4.7 rather than 1:7. In addition, the 10–90 percentile ranges for the NS counts (in-cluster 7 with +89/−6; escaper 12.5 with +25/−10) overlap substantially, so the claimed 2:1 preponderance of NS escapers over in-cluster NS ULXs, and hence the 96%/4% in-cluster BH/NS split, are not statistically established at the precision claimed. The authors should report the per-simulation distribution of the fractions and provide a statistical significance test.","section":"Section 3.8, Table 6"},{"comment":"Each set of initial conditions is represented by a single simulation, and the paper itself notes strong stochastic variability (e.g., Section 3.1 and the discussion of E[NULX]). The 10–90 percentile ranges in Table 6 are therefore not confidence intervals on a population mean but ranges over one realization per model; the rare NS ULX counts that underpin the main ratios are subject to Poisson noise. The quantitative conclusions should either be downgraded to reflect this small-sample limitation or supported by additional realizations or seed variations.","section":"Section 2.2, Table A.1"}],"minor_comments":[{"comment":"In the abstract, \"only4%\" should read \"only 4%\".","section":"Abstract"},{"comment":"The sentence beginning \"Given their presence in older stellar populations, GCULXs are likely powered by low-mass X-ray binaries...\" appears twice in the final paragraph of the introduction and should be deduplicated.","section":"Section 1"},{"comment":"The word \"strenghtening\" should be corrected to \"strengthening\".","section":"Section 3.5"},{"comment":"The note for \"th,rel - Spitzer half-mass relaxation time\" lists units of [M⊙], but a timescale should be expressed in units of time (e.g., Myr); check whether the column was mislabeled.","section":"Table A.1 notes"},{"comment":"The statement that in-cluster ULXs have a median count of 176 conflicts with Table 6, whose sum is 149.5; if the 176 value comes from a different time integration or an alternative weighting, the text should define it explicitly.","section":"Section 3.8"},{"comment":"\"Among escaper, ULXs\" should read \"Among escapers, ULXs\".","section":"Section 5"}],"recommendation":"major_revision","confidential_remarks":"The manuscript relies heavily on self-citations to the authors' own code-development papers, which is understandable given that MOCCA is their tool, but the revision should make the novelty of the ULX-specific results clearer. The main editorial concern is that the abstract and conclusions report percentages with more confidence than the underlying threshold-sensitive luminosity model and the single-realization simulation grid can support; a sensitivity analysis and revised error statements would make the paper suitable for publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is the first MOCCA-based census of ULXs in globular clusters that separates in-cluster sources from dynamically ejected 'escaper' systems, and the qualitative picture looks solid — dynamics both build and destroy ULX progenitors, old and tidally-filling clusters host few in-cluster ULXs, and GC ejecta contaminate the field population. The headline numbers (96% BH accretors in-cluster, 40% NS among escapers, ~1:7 escaper-to-in-cluster ratio) should not be quoted as measurements yet.\n\nWhat earns credit: the pipeline is mature (MOCCA, BSE, FEWBODY), the data are released on Zenodo with a DOI, and the paper shows 10-90 percentile ranges on its counts instead of hiding the scatter. It also engages the observations honestly — the Milky Way GC deficit, the ~30 extragalactic GCULXs, the Antennae displacements — and it explicitly defers beaming to future work in Section 4.4.\n\nThe stress-test note lands on reading the paper. Every ratio in the paper is a threshold classification, and NS ULXs sit at median log10 LX,max ≈ 39.22, roughly 0.2 dex above the 10^39 erg/s cutoff, while BH ULXs are about a dex brighter. Beaming both raises apparent luminosity and lowers visibility, and the authors themselves note that most NS ULXs are likely beamed (citing their own 2019 study). So the NS-specific claims — the 4% in-cluster fraction, the 40% escaper fraction, the 2:1 NS escaper-to-in-cluster ratio — are the fragile ones. The BH-dominated picture and the age/tidal-filling trends should survive that change.\n\nThe other soft spot is statistical. Table 6 counts have wide overlapping ranges: in-cluster NS is 7 (+89, -6), escaper NS is 12.5 (+25, -10). The 2:1 ratio is the median of a very broad distribution. Also, the abstract's 'one-seventh of the total ULX population' and the conclusion's '1:7 ratio of escapers to in-cluster' are not the same quantity, and the Table 6 medians actually suggest something closer to 1:5. Minor: the E[NULX] weighting is only described in a footnote.\n\nWho this is for: anyone working on ULX demographics, GC X-ray binaries, or field-versus-cluster origin questions. It deserves a serious referee. Recommendation: send it to review, and have the referee ask for a beaming sensitivity test and per-simulation counts before the ratios become observational predictions.","headline":"First MOCCA census of GC ULXs with a genuinely new escaper channel, but the headline ratios are threshold products of an unbeamed luminosity model and small samples — treat as indicative, not measured.","tokens_in":26084,"tokens_out":5556,"would_cite":true,"duration_ms":47354,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper argues that globular clusters both feed and eject ultraluminous X-ray sources, with about 96% of in-cluster ULXs powered by black holes and ejected 'escaper' ULXs making up roughly one-seventh of all cluster-born ULXs.","keywords":["ultraluminous X-ray sources","globular clusters","black hole accretors","neutron star accretors","dynamical interactions","Monte Carlo cluster evolution","escaper binaries","X-ray binary populations"],"falsifier":"Run the same cluster simulations with beaming included: if beamed apparent luminosities push a sizable number of neutron-star accretors above $10^{39}\\,{\\rm erg\\,s^{-1}}$, the 96%/4% black hole versus neutron star split and the 1:7 escaper ratio will shift. Observationally, search for coherent pulsations among ULXs offset from their host globular clusters; the paper's roughly 40% neutron star fraction among escapers predicts that off-cluster ULXs in galaxies like the Antennae should show pulsar-type accretion signatures far more often than in-cluster globular cluster ULXs.","tokens_in":24972,"feed_emoji":"🕳️","tokens_out":6275,"duration_ms":54943,"temperature":0.7,"pith_summary":"The paper asks where ultraluminous X-ray sources (ULXs) come from when their birth environment is a globular cluster rather than the field. Using Monte Carlo simulations that follow both stellar dynamics and binary evolution, it argues that dense-cluster encounters both create ULX progenitors and destroy or eject them, so dynamics is a two-sided agent. The central quantitative claims are that about 96% of in-cluster ULXs in globular clusters harbor black hole accretors, that only about 4% contain neutron stars, and that ULXs ejected from clusters ('escapers') make up about one-seventh of all cluster-born ULXs while being roughly twice as common as in-cluster ULXs when the accretor is a neutron star. If these numbers hold, field ULX surveys are partly counting cluster runaways, and the absence of ULXs in Milky Way globular clusters is a natural consequence of their old stellar populations and tidal filling.","feed_headline":"Black holes power 96% of ULXs born in globular clusters","feed_subtitle":"Ejected 'escaper' ULXs make up one-seventh of cluster-born sources and may pollute field surveys.","key_machinery":"The central object is a Monte Carlo globular cluster evolution code that combines orbit-averaged stellar dynamics with detailed binary stellar evolution and few-body scattering prescriptions. The argument runs on two load-bearing mechanisms. The first is the X-ray luminosity prescription of Eq. (1): the Shakura-Sunyaev disk formula in the sub-Eddington regime and a logarithmic, super-Eddington extension $L_X = L_{\\rm Edd}[1+\\ln(\\dot M/\\dot M_{\\rm Edd})]$ above the Eddington rate, with accretion efficiency $\\eta_{0.1}=1$ and no beaming. The second is the code's tracking of escapers: binaries that leave the cluster through few-body encounters, supernova kicks, relaxation, or tidal stripping and later enter the ULX phase in the field. Both mechanisms work together to produce the population ratios that define the paper's conclusions.","core_discovery":"The central claim is that globular clusters are a genuine, dynamically sculpted birthplace for ULXs, with a population structure distinct from the field. In the simulations, dynamical hardening and binary exchanges create ULX progenitors that would not form from isolated stellar evolution, while strong encounters and tidal stripping remove others; the net result is that non-tidally filling clusters host far more ULXs than tidally filling ones, and the oldest clusters produce almost none. Across the simulated sample, the ULX population is overwhelmingly black-hole-driven: roughly 96% of in-cluster ULXs have black hole accretors, and the fraction is even higher in clusters younger than about 300 Myr. Neutron-star accretors dominate only among the escapers, at about 40% of that population, and escaper ULXs with neutron star accretors are about twice as common as their in-cluster counterparts. The paper therefore concludes that the Milky Way's lack of globular cluster ULXs fits tidally filling models, and that field ULX samples are likely polluted by cluster ejecta.","pith_inferences":["Because every population ratio in the paper is defined by crossing the $10^{39}\\,{\\rm erg\\,s^{-1}}$ threshold, including beaming (which the paper defers) would change apparent luminosities and could shift the 96%/4% and 1:7 numbers; the qualitative picture of a dynamically active cluster ULX population should survive, but the quantitative ratios are threshold-dependent.","The escaper channel adds a new ingredient to field population synthesis: a sub-population of ULXs whose initial conditions were set by cluster dynamics before ejection, potentially explaining offsets between ULX positions and star-forming regions without invoking intermediate-mass black holes.","A targeted observational test would be to measure the fraction of neutron star accretors among ULXs spatially offset from their host clusters in galaxies like the Antennae; the paper's roughly 40% neutron star fraction among escapers predicts more pulsating ULXs off-cluster than on-cluster.","The model's treatment of multiple stellar populations suggests that the second, more concentrated generation boosts ULX formation, so globular clusters with strong multiple-population signatures might be the best candidates for future globular cluster ULX detections."],"forward_implications":["If roughly 96% of in-cluster globular cluster ULXs are powered by black holes, then searches for coherent pulsations in extragalactic globular cluster ULXs should mostly fail, and ULX pulsar detections inside clusters would be rare events.","Escaper ULXs constitute about one-seventh of all ULXs born in clusters, so field ULX catalogs contain a non-negligible population of cluster runaways with dynamically imprinted orbital properties.","Non-tidally filling clusters produce substantially more ULXs than tidally filling ones, so galaxy-wide globular cluster ULX numbers depend sensitively on cluster orbital radii and tidal stripping.","The near-absence of ULXs in Milky Way globular clusters is consistent with old stellar populations and initially tidally filling clusters, rather than requiring exotic explanations.","Intermediate-mass black hole powered ULXs, when present, form exclusively in dense non-tidally filling two-population clusters and can contribute 16-44% of a cluster's ULX budget."],"supporting_citations":[{"why":"Provides the disk luminosity model underlying Eq. (1) for both sub-Eddington and super-Eddington regimes.","marker":"Shakura & Sunyaev 1973"},{"why":"Recent discussion of the super-Eddington luminosity formulation adopted for the X-ray luminosity calculation.","marker":"Lasota & King 2023"},{"why":"Field ULX population synthesis giving the NS/BH ratio and beaming treatment that the escaper comparisons and deferred beaming studies build on.","marker":"Wiktorowicz et al. 2019"},{"why":"Discovery of the first ULX in a globular cluster, anchoring the observed phenomenon the paper models.","marker":"Maccarone et al. 2007"},{"why":"Adds seven globular cluster ULXs in M87, expanding the observed extragalactic sample.","marker":"Dage et al. 2020"},{"why":"Provides ten new globular cluster ULX candidates in massive early-type galaxies, defining the observed rarity of these sources.","marker":"Thygesen et al. 2023"},{"why":"Shows X-ray sources offset from young clusters in the Antennae galaxies, supporting the ejection interpretation for escaper ULXs.","marker":"Poutanen et al. 2013"},{"why":"The rapid supernova mechanism used to set neutron star and black hole formation and fallback-dependent kicks, determining accretor types.","marker":"Fryer et al. 2012"},{"why":"The Maxwellian natal kick distribution with $\\sigma=265\\,{\\rm km\\,s^{-1}}$ used for neutron stars and black holes, central to producing escapers.","marker":"Hobbs et al. 2005"},{"why":"Introduces the Monte Carlo cluster evolution code that carries the dynamical simulations at the heart of the paper.","marker":"Hypki & Giersz 2013"}],"fun_headline_variants":["Black-hole ULXs dominate globular clusters, 96% of them","Ejected ULXs from clusters may distort field counts","Dynamical exchanges in clusters create black-hole ULXs","Tidally filling clusters yield far fewer ULXs","Globular clusters eject neutron-star ULXs into field"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that X-ray luminosity is correctly computed from the unbeamed Shakura-Sunyaev plus logarithmic super-Eddington formula of Eq. (1) with efficiency $\\eta_{0.1}=1$, so that every population ratio (96% black holes, 1:7 escapers, 40% neutron stars among escapers) is determined by which simulated binaries cross the $10^{39}\\,{\\rm erg\\,s^{-1}}$ ULX threshold; if beaming were included, apparent luminosities and therefore the ratios would change.","fun_headline_variants_meta":{"raw":{"variants":["Black-hole ULXs dominate globular clusters, 96% of them","Ejected ULXs from clusters may distort field counts","Dynamical exchanges in clusters create black-hole ULXs","Tidally filling clusters yield far fewer ULXs","Globular clusters eject neutron-star ULXs into field"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000196,"raw_usage":{"total_tokens":1383,"prompt_tokens":992,"completion_tokens":391,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":608,"completion_tokens_details":{"reasoning_tokens":309}},"tokens_in":608,"tokens_out":391,"duration_ms":4407,"temperature":1.0,"reasoning_tokens":309,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T21:06:19.994170+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run the same cluster simulations with beaming included: if beamed apparent luminosities push a sizable number of neutron-star accretors above $10^{39}\\,{\\rm erg\\,s^{-1}}$, the 96%/4% black hole versus neutron star split and the 1:7 escaper ratio will shift. Observationally, search for coherent pulsations among ULXs offset from their host globular clusters; the paper's roughly 40% neutron star fraction among escapers predicts that off-cluster ULXs in galaxies like the Antennae should show pulsar-type accretion signatures far more often than in-cluster globular cluster ULXs.","supporting_citations":[{"cited_title":"& King, A","cited_arxiv_id":null,"evidence_quote":"Recent discussion of the super-Eddington luminosity formulation adopted for the X-ray luminosity calculation."},{"cited_title":"J., Kundu, A., Zepf, S","cited_arxiv_id":null,"evidence_quote":"Discovery of the first ULX in a globular cluster, anchoring the observed phenomenon the paper models."},{"cited_title":"C., Zepf, S","cited_arxiv_id":null,"evidence_quote":"Adds seven globular cluster ULXs in M87, expanding the observed extragalactic sample."},{"cited_title":"2023, MNRAS, 518, 3386","cited_arxiv_id":null,"evidence_quote":"Provides ten new globular cluster ULX candidates in massive early-type galaxies, defining the observed rarity of these sources."},{"cited_title":"F., Sholukhova, O., & Greiner, J","cited_arxiv_id":null,"evidence_quote":"Shows X-ray sources offset from young clusters in the Antennae galaxies, supporting the ejection interpretation for escaper ULXs."}],"review_version":1}