{"id":"30bb990e-9ce7-4fe7-b99d-27a9322d6bf5","arxiv_id":"2501.08831","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"N-body simulations show that post-gas expulsion coalescence of multiple low-mass embedded clusters can reproduce the mass-radius-age distribution of Milky Way open clusters.","lead":"This paper uses computer simulations to test whether open star clusters can form when several small embedded star clusters collide and merge after the gas that created them is blown away. The authors find that such mergers can explain the observed masses and sizes of Milky Way open clusters without needing very massive single birth clusters.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Gas expulsion timescale parameter space is too narrow: τ_g ≈ 0.03–0.1 Myr (5τ_g ≲ 0.5 Myr) never probes slow, several-Myr gas dispersal, so the claimed robustness of post-gas-expulsion coalescence may not hold in the physically relevant regime.","rationale":"The reader's weakest_assumption correctly identifies the gas expulsion prescription as the delicate part of the argument. My reading sharpens this into a specific, testable gap: the paper never explores a genuinely slow gas removal timescale. The adopted τ_g is tied to the internal half-mass radius divided by the HII region sound speed, which yields sub-Myr depletion times even in the 'moderate' case. Observationally, gas dispersal in embedded clusters is often argued to last several Myr, comparable to or longer than the cluster crossing time. In that regime the clusters respond more adiabatically, expand less, and may fail to contact neighbours; the paper's simulated coalescence could be an artefact of the fast-dispersal assumption. This is more load-bearing than the paper's secondary weaknesses (qualitative overlay, selection of three MSFRs) because it directly determines whether the proposed mechanism operates at all. Independent support from Geyer & Burkert (2001) and Farias et al. (2024) validates the exponential-decay form, but not the range of timescales explored here. The proposed test varies τ_g into the missing slow regime and would settle whether coalescence remains robust and whether the remnants still cover the observed open-cluster parameter space. If the slow-regime test preserves the coverage, the paper's conclusion would be strengthened; if not, the central claim would be restricted to a fast-dispersal subset of environments. The verdict should remain CONDITIONAL pending this test.","tokens_in":16848,"tokens_out":10948,"duration_ms":122002,"concrete_test":"Re-run the NGC 6334 and Carina initial conditions with τ_g = 1 Myr and 5 Myr (keeping SFE = 0.33, τ_d = 0.6 Myr), and also with SFE = 0.5 at the standard τ_g. Measure bound remnant mass and r_50 as a function of age and check whether the remnants intersect the observed OC locus in Fig. 6. If coalescence is suppressed or the remnants fall outside the OC locus for τ_g ≳ 1 Myr, the central claim fails.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The central claim requires embedded clusters to expand enough after gas removal to overlap and coalesce. In Sec. 3.2.1, Eq. 1 sets the gas depletion timescale as τ_g = r_h(0)/v_g with v_g ≈ 10 km s⁻¹. For the adopted r_h values (Marks & Kroupa 2012, roughly 0.1–0.4 pc), this gives τ_g ≈ 0.03–0.1 Myr; the 'moderate' case (5τ_g) is still ≲0.5 Myr. These are much shorter than the several-Myr gas dispersal times typically inferred for embedded clusters (e.g., HII region expansion times). Both the 'fast' and 'moderate' runs are therefore in the impulsive regime, driving strong cluster expansion. If real gas removal proceeds more gradually (τ_g ≳ 1 Myr), clusters expand less and may never reach overlapping radii, suppressing coalescence. The robustness statement in Sec. 3.3 ('within a reasonable range of parameters') is not supported for slow gas removal, and the claim that multiple low-mass ECs can cover the OC parameter space (Fig. 6) would fail in that regime. This is load-bearing because the entire coalescence mechanism depends on the balance between expansion and dissolution, and that balance is set by the unexplored slow-dispersal branch of the gas expulsion model.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper uses N-body simulations with the PeTar code to test whether open clusters can form through the post-gas-expulsion coalescence of several low-mass embedded clusters rather than from a single massive embedded cluster. Initial conditions are based on observed subcluster masses, positions, and kinematics in NGC 1893, NGC 6334, and Carina from the MYStIX project. Gas removal is modeled with the analytic exponential-decay external potential of Eq. (1), adopting a star formation efficiency of 0.33 and two depletion timescales, \"fast\" (tau_g) and \"moderate\" (5 tau_g). The authors find that massive and/or numerous embedded clusters can expand, interact, and merge into a bound remnant, while low-mass and sparse systems dissolve. They overlay the simulated remnants on the Hunt & Reffert (2024) open cluster sample and argue that coalescence can cover the observed open cluster parameter space without requiring massive embedded cluster progenitors.","tokens_in":17122,"tokens_out":8027,"duration_ms":80027,"significance":"If the central claim holds, the paper offers a plausible resolution to the observed mismatch between the mass functions of embedded clusters and open clusters, and it would substantiate the idea that massive, older open clusters need not have had massive single-cluster precursors. The paper's strengths are its use of a well-tested N-body code, observationally motivated initial conditions from three MYStIX regions, and explicit comparison of different gas-expulsion modes, spatial separations, and relative velocities. The main limitations are the narrow gas-expulsion parameter coverage and the qualitative nature of the comparison to the observed open cluster sample; both directly affect the strength of the robustness and coverage claims.","major_comments":[{"comment":"The gas-expulsion study covers only tau_g = r_h(0)/v_g with v_g = 10 km/s and its five-fold multiple. For the adopted half-mass radii from Marks & Kroupa (2012), roughly 0.1-0.4 pc, this gives tau_g about 0.01-0.04 Myr in the \"fast\" case and about 0.05-0.2 Myr in the \"moderate\" case. Thus even the moderate runs are in the impulsive regime, far shorter than the several-Myr gas dispersal timescales often inferred for embedded clusters. The paper's claim in Sec. 3.3 that coalescence is robust \"within a reasonable range of parameters\" is therefore not supported for slow gas dispersal, where clusters expand less after gas removal and may never reach overlapping radii. Since the entire coalescence mechanism depends on the balance between expansion and dissolution, I request additional simulations with tau_g of order 1-5 Myr (or a continuous parameter scan) and a quantitative comparison of the resulting remnant masses and radii, or an analytic argument bounding the effect of slow gas removal.","section":"Sec. 3.2.1, Eq. (1); Table 1; Sec. 3.3"},{"comment":"The central quantitative claim, that the coalescence simulations \"cover the parameter space\" of the observed open clusters in the Milky Way, is supported only by visual inspection of overlaid tracks in the mass-age and radius-age planes. No statistical measure is provided, such as the fraction of Hunt & Reffert (2024) clusters that fall within the simulated tracks, the effect of observational completeness, or a goodness-of-fit metric. It is also not specified which simulation cases are selected for the overlay and whether systems that are described as dissolving (for example, the \"Carina-fast-vd\" case) are included. Please define an explicit criterion for a successful coalescence remnant and provide a quantitative comparison to the observed open cluster sample.","section":"Sec. 3.3, Fig. 6"},{"comment":"The star formation efficiency is fixed at SFE = 0.33 in all runs. The text argues that a lower SFE is \"equivalent\" to a shorter gas-expulsion timescale, but this equivalence is not demonstrated quantitatively. Because SFE controls the depth of the removed gas potential and hence the amount of cluster expansion and mass loss, the robustness claim should be tested with at least one additional SFE value (e.g., SFE = 0.2 or 0.5) or supported by a scaling argument derived from the model.","section":"Sec. 3.2.1 and Sec. 3.2.3"},{"comment":"The construction of the initial configuration is underspecified. The text says each subcluster is first simulated in isolation for 1 Myr and then the evolved subclusters are collected and placed at observed separations, but it does not state whether the external gas potential is still active at the assembly time, whether the collected clusters are re-virialized in the combined system, or how the gas expulsion continues after assembly. Because the expansion history is the driver of the coalescence process, these details need to be explicit for the simulations to be reproducible and for the evolution after 1 Myr to be interpretable.","section":"Sec. 3.2.2 and Sec. 3.2.3"}],"minor_comments":[{"comment":"In the Introduction, \"the the mergers of OCs progenitors\" should be \"the mergers of OC progenitors.\"","section":"Sec. 1"},{"comment":"The phrase \"they start to interact upon contact, leading to subsequent mutual influence in their evolution\" uses \"coalescence\" as a verb elsewhere; for example, in Sec. 3.1.3, \"subsequently interact and coalescence\" should be \"coalesce.\"","section":"Sec. 3.2.5"},{"comment":"The Larson-relation velocity assignment v proportional to L^0.5 is not normalized; please specify the proportionality constant or state explicitly that the normalization is fixed by the outermost cluster velocity of 2 km/s.","section":"Sec. 3.2.4"},{"comment":"Not all panels (a)-(f) of Fig. 5 are described in the text; please label each panel with its case name or refer to each panel explicitly in Sec. 3.2.5.","section":"Fig. 5"},{"comment":"The column header \"Terminated time\" should be \"Termination time.\"","section":"Table 1"},{"comment":"The snapshots are small and the point sizes encode stellar mass, but no color bar or scale bar is provided; adding one would improve readability and quantitative interpretation.","section":"Appendix A figures"}],"recommendation":"major_revision","confidential_remarks":"The paper is within the scope of MNRAS and the simulations appear carefully executed with a well-tested code. The main concern is that the gas-expulsion parameter coverage, especially the absence of slow gas removal, directly bears on the central robustness claim, and the comparison to the observed open cluster sample in Fig. 6 is entirely qualitative. These issues are fixable with additional simulations and a quantitative comparison, so I recommend major revision rather than rejection. One further editorial point is that the manuscript relies heavily on companion papers by the same group (Zhou et al. 2024b-h) for key inputs and comparisons; the present paper should be sufficiently self-contained for a reader to assess those inputs without consulting all of them."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"You should know two things about this paper. First, it does something genuinely new: it takes observed subcluster configurations from three MYStIX regions, evolves them with PeTar, and shows that after gas expulsion the expanding clusters can overlap and coalesce into objects that land in the observed open-cluster mass-radius-age parameter space. Second, the central robustness claim—that coalescence is insensitive to reasonable parameter choices—is only tested in the impulsive gas-removal regime, because tau_g = r_h/v_g with v_g=10 km/s gives 0.03–0.1 Myr, and even the 'moderate' 5 tau_g cases are done by ~0.5 Myr. Real HII region gas dispersal typically takes several Myr, so the slow, adiabatic branch is unexplored. That's a load-bearing gap, because the whole mechanism relies on clusters expanding enough to meet.\n\nWhat the paper does well: the initial conditions are tied to observations, the code is solid, and the scenario addresses a real tension from previous work—the clump mass function can't directly produce the massive old OCs. The comparison of NGC 1893 (low-mass, few clusters, dissolves) with NGC 6334 and Carina (more massive or more numerous clusters, coalesce) makes a plausible physical case that cluster mass and number matter. The paper is also honest about its simplifications and cites the prior subcluster-merger literature properly.\n\nThe soft spots are real but addressable. The qualitative overlay of model tracks on the OC sample (Fig. 6) lacks any statistical measure, and the three chosen regions are selected to bracket the OC locus. No simulation products are public, though data are available on request. The slow gas-dispersal gap is the biggest issue. If gas removal is more gradual, clusters expand less and coalescence may be suppressed. I'd want to see a run with tau_g ~ 1–3 Myr before accepting the 'robust and insensitive' phrasing.\n\nWho's this for? Anyone working on embedded cluster evolution, hierarchical star formation, or the origin of open clusters. It deserves a serious referee—the idea is new enough and the simulations careful enough to warrant the time. I'd send it to review but ask the authors to widen the gas-expulsion parameter space and quantify the comparison to the OC catalog.\n\nWorth a reading-group slot if your group works on cluster formation; otherwise skim the figures and conclusion.","headline":"Plausible scenario test showing low-mass embedded clusters can coalesce into massive open clusters, but the gas-expulsion timescales explored are too short to support the claimed robustness.","tokens_in":17657,"tokens_out":3843,"would_cite":true,"duration_ms":35625,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"The Milky Way's open clusters can form from the merging of several small embedded clusters after gas expulsion, rather than from a single massive progenitor.","keywords":["open clusters","embedded clusters","coalescence","gas expulsion","N-body simulations","star formation","massive star-forming regions","cluster formation"],"falsifier":"If a young open cluster with mass above 3000 $M_\\odot$ is found whose member stars all have indistinguishable ages and metallicities and whose parent molecular cloud never contained an embedded cluster with mass above 1000 $M_\\odot$, the coalescence scenario would be falsified.","tokens_in":16626,"feed_emoji":"⭐","tokens_out":9205,"duration_ms":81285,"temperature":0.7,"pith_summary":"The paper asks whether the Milky Way's open clusters—old, relatively massive star groups—can form from the merging of several low-mass embedded clusters rather than from a single massive precursor. Observational comparisons show that the mass distribution of today's giant molecular clumps cannot directly produce embedded clusters heavy enough to explain the most massive open clusters. The authors simulate the expansion of embedded clusters after the gas that binds them is expelled, and show that under a wide range of conditions the expanded clusters collide and coalesce into a single bound cluster. They find that the outcome is largely insensitive to initial conditions, with the mass of the embedded clusters and their number being the key factors. They conclude that the coalescence of multiple low-mass embedded clusters covers the observed parameter space of open clusters, so open clusters do not require massive embedded clusters as precursors.","feed_headline":"Small embedded clusters can merge into open clusters","feed_subtitle":"Simulations show post-gas-expulsion coalescence covers the mass and radius range of observed open clusters.","key_machinery":"The load-bearing element is the treatment of gas expulsion as a spherically symmetric, exponentially decaying external gravitational potential of the form $M_g(t) = M_g(0)$ for $t \\le \\tau_d$ and $M_g(t) = M_g(0) \\exp(-(t-\\tau_d)/\\tau_g)$ for $t > \\tau_d$, with a fixed star formation efficiency of $\\mathrm{SFE} = 0.33$ and a delay of $\\tau_d = 0.6$ Myr. This potential drives each initially bound embedded cluster to expand after its gas is removed. The expansion brings neighbouring clusters into gravitational contact, and the subsequent mutual attraction—especially from the most massive clusters, which act as stable cores—determines whether the system coalesces into a single bound cluster or dissolves. The machinery also includes the initial conditions: subclusters are evolved individually for about 1 Myr with the gas-expulsion prescription, then placed together at the observed separations and given a cloud-scale velocity dispersion, and the whole complex is integrated with a direct N-body code that includes stellar evolution and binaries.","core_discovery":"The central claim is that open clusters in the Milky Way can originate from the post-gas expulsion coalescence of several embedded clusters within the same parental molecular cloud. Using N-body simulations whose initial conditions are drawn from the observed spatial and mass distributions of embedded clusters in three massive star-forming regions, the authors show that after gas is dispersed the individual clusters expand, come into contact, and merge into a single gravitationally bound cluster that can survive for hundreds of millions of years. The mass of the embedded clusters plays the critical role: massive clusters provide stable gravitational cores that pull neighbouring clusters in and anchor the merger, whereas low-mass, sparse complexes disperse into loose associations. The number of embedded clusters also aids coalescence. The paper concludes that this coalescence pathway can cover the full observed range of open-cluster masses and radii as a function of age, and that therefore open clusters do not necessarily require massive embedded clusters as direct precursors.","pith_inferences":["The same expansion-driven coalescence mechanism may operate in star-forming regions in other galaxies, implying that the stellar cluster mass function could be shaped more by assembly than by the initial clump mass function alone.","If coalescence is common, old open clusters assembled by mergers should retain broader internal age and metallicity spreads than monolithic clusters; this can be checked against existing photometric and spectroscopic data.","Replacing the analytic gas-expulsion potential with live, possibly asymmetric gas distributions in future simulations would test whether the coalescence efficiency and the resulting cluster properties are sensitive to the geometry of gas removal."],"forward_implications":["Massive open clusters can be produced without massive embedded-cluster progenitors, resolving the observed order-of-magnitude gap between clump/embedded-cluster masses and open-cluster masses.","The coalescence route is largely insensitive to initial conditions over the explored parameter range, with embedded-cluster mass and number being the decisive factors.","Strong gravitational cores from massive embedded clusters are required to anchor a merger and to let the coalesced cluster survive tidal forces, so not every complex of low-mass clusters will form an open cluster.","Spatial separation and relative velocity between clusters both widen and speed up mass loss of the merger product, but slower gas expulsion can counteract these effects, creating a degeneracy between initial geometry and gas-removal speed."],"supporting_citations":[{"why":"Introduces the exponential-decay gas-expulsion potential (Eq. 1) used to model gas removal in all simulations.","marker":"Kroupa et al. (2001)"},{"why":"Establishes the simulation parameter set (IMF, virial state, gas-expulsion recipe) adopted here.","marker":"Banerjee & Kroupa (2013)"},{"why":"Supplies the embedded-cluster survival statistics and the SFE ≈ 0.33 benchmark used in the models.","marker":"Lada & Lada (2003)"},{"why":"Documents the mass mismatch between Galactic clumps/embedded clusters and open clusters that motivates the coalescence scenario.","marker":"Zhou et al. (2024b)"},{"why":"Provides the half-mass radius–mass relation that sets the radii of the simulated clusters.","marker":"Marks & Kroupa (2012)"},{"why":"Provides the observed subcluster spatial and mass distributions in three massive star-forming regions used as initial conditions.","marker":"Kuhn et al. (2014)"},{"why":"Supplies the high-quality open-cluster catalog whose mass, radius, and age distributions are the comparison target.","marker":"Hunt & Reffert (2024)"},{"why":"Independent star-plus-gas simulations that also find subclusters merge into smooth monolithic clusters, supporting the robustness of coalescence.","marker":"Sills et al. (2018)"}],"fun_headline_variants":["Open clusters from embedded cluster mergers","Gas expulsion merges embedded clusters into open clusters","Embedded cluster coalescence opens the door to open clusters","Coalescing embedded clusters birth open clusters","Merger of small clusters yields large open clusters"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The argument rests on the assumption that gas removal from each embedded cluster is faithfully described by a spherical potential that decays exponentially with a fixed star-formation efficiency of 0.33 and a 0.6 Myr delay; if real gas expulsion is more violent, asymmetric, or proceeds at different efficiencies, the balance between cluster expansion and coalescence could shift and the claimed coverage of the open-cluster parameter space could fail.","fun_headline_variants_meta":{"raw":{"variants":["Open clusters from embedded cluster mergers","Gas expulsion merges embedded clusters into open clusters","Embedded cluster coalescence opens the door to open clusters","Coalescing embedded clusters birth open clusters","Merger of small clusters yields large open clusters"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000206,"raw_usage":{"total_tokens":1415,"prompt_tokens":985,"completion_tokens":430,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":601,"completion_tokens_details":{"reasoning_tokens":361}},"tokens_in":601,"tokens_out":430,"duration_ms":4496,"temperature":1.0,"reasoning_tokens":361,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T20:16:40.738890+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"If a young open cluster with mass above 3000 $M_\\odot$ is found whose member stars all have indistinguishable ages and metallicities and whose parent molecular cloud never contained an embedded cluster with mass above 1000 $M_\\odot$, the coalescence scenario would be falsified.","supporting_citations":[],"review_version":1}