{"id":"208a31ea-7b83-4a80-8bc8-7db1e5cfc8f3","arxiv_id":"2504.17903","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":7,"one_line_summary":"Binary interactions delay about 25% of core-collapse supernovae past the standard 44 million year cutoff and displace about 13% by more than 100 parsecs from their birth clusters in simulated dwarf galaxies.","lead":"This paper simulates how binary star interactions change where and when supernovae explode inside galaxy simulations. It finds that roughly a quarter of supernovae explode later than any single star would, and about 13% explode far from their birth clusters, which could change how supernova feedback shapes galaxies.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The headline displacement fraction (~13% beyond 100 pc) hinges on an unconstrained cluster velocity dispersion; the paper's own Table C1 spans 5.1% to 45.3% across bracketing values.","rationale":"The reader correctly identifies the initial cluster velocity dispersion as the weakest link. My independent reading of Section 4.6.1 and Table C1 confirms that the headline displacement fraction is dominated by this unconstrained parameter, with a factor-of-9 swing across the paper's own bracketing values. I considered two alternative concerns: (i) the analytic model is validated in-sample (Section 5 fits the same simulations it claims to reproduce), and (ii) the explodability assumption (Section 6.3) could suppress the delayed merger-product tail. Both are real limitations, but neither undermines the central claim as directly as vdisp: the timing tail remains substantial (12–35% late) across all variations, while the displacement headline spans the full 5–45% range. The qualitative finding that binaries displace SNe above the single-star baseline is robust, so the paper's core insight survives; the quantitative 13% does not. The concrete test—marginalizing over observed cluster velocity dispersions—would settle whether the fiducial choice is representative. I therefore keep the reader's CONDITIONAL verdict unchanged.","tokens_in":49835,"tokens_out":5761,"duration_ms":57808,"concrete_test":"Use the observed distribution of young cluster velocity dispersions (e.g., from Kuhn et al. 2019) to sample vdisp per cluster instead of fixing 1.7 km/s, and recompute fD>100pc and the full distance distribution. If the resulting fD>100pc credible interval spans the 5–45% range from Table C1, the headline should be presented as a range; if the observed vdisp distribution is tightly peaked near the fiducial value, the 13% figure survives.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central quantitative claim that ~13% of core-collapse SNe occur more than 100 pc from their parent cluster (abstract, Section 3.2) is set by the assumed initial cluster velocity dispersion vdisp = 1.7 km/s. Section 4.6.1 states 'There are currently no strong observational constraints on the appropriate value of the initial cluster velocity dispersion,' and Table C1 shows fD>100pc = 5.1% for vdisp = 0.5 km/s and 45.3% for vdisp = 5 km/s. The headline value is therefore not a robust prediction of the model but a one-point draw from a parameter that brackets the entire range of the reported quantity. While the qualitative displacement effect is robust—even at vdisp = 0.5 km/s binaries give 5.1% vs 0.9% for single stars—the specific 13% figure in the abstract and conclusions is conditional on an unvalidated choice, and the paper's robustness summary in Section 6.1.2 only partially conveys this by excluding the low-dispersion case. Because the spatial distribution of SNe is a primary claimed outcome of the simulation framework, this unconstrained parameter is the most load-bearing weakness of the central claim.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper uses the cogsworth population-synthesis and orbital-integration framework to replace the star particles formed in the past 150 Myr of the FIRE-2 m11h dwarf galaxy simulation with clusters of binary stars, evolving them with COSMIC and integrating their orbits in a gala potential fitted to the hydrodynamical simulation. It records the time and position of every core-collapse supernova relative to the parent cluster, for a fiducial binary model and for a broad set of variations of initial conditions, binary physics, metallicity, cluster dissolution assumptions, and galaxy potential. The central results are that binary interactions produce a long tail of delayed SNe (about 25% after the 44 Myr single-star cutoff, predominantly merger products) and displaced SNe (about 13% beyond 100 pc, predominantly ejected secondaries), and that these distributions are robust in their qualitative form across most variations. The paper also presents a metallicity-dependent analytic model for the SN rate and progenitor velocity distribution, intended as a subgrid replacement for single-star feedback prescriptions.","tokens_in":50110,"tokens_out":7059,"duration_ms":67549,"significance":"If the results hold, the paper provides the most complete joint time-distance distribution of core-collapse SN feedback from binaries in a realistic galactic potential to date, and a practical analytic prescription that can be incorporated into hydrodynamical simulations. The forward-modeling pipeline is clearly specified, and the code and simulation data are released on GitHub and Zenodo, which makes the experiment reproducible. The extensive parameter study, including extreme variations in common-envelope efficiency, mass-transfer stability, kicks, IMF, orbital period, mass ratio, metallicity, velocity dispersion, and galaxy code, is a genuine strength, as are the quantitative comparisons to De Donder & Vanbeveren (2003), Zapartas et al. (2017), Eldridge et al. (2011), and Renzo et al. (2019). The qualitative existence of delayed and displaced SNe is robust; less robust is the specific 13% displacement fraction, as discussed in the major comments.","major_comments":[{"comment":"The abstract and Section 3.2 quote about 13% of SNe beyond 100 pc as a headline result. Section 4.6.1 states that there are currently no strong observational constraints on the initial cluster velocity dispersion, and Table C1 shows that fD>100pc changes from 5.1% (vdisp = 0.5 km/s) to 45.3% (vdisp = 5 km/s). Since the unejected-component velocity distribution in Eq. (6) is set by vdisp, the quantitative displacement claim is a one-point draw from a largely unconstrained parameter rather than a robust prediction. I recommend presenting the displacement result as a conditional range and explicitly labeling the fiducial 13% as such throughout the abstract, Section 6, and the conclusions.","section":"Section 4.6.1 / Table C1"},{"comment":"The statement that \"in all of our models, at least 12-15% of all SNe occur more than 0.1 kpc from the centre of the clustered star formation\" is contradicted by the vdisp = 0.5 km/s variation in Table C1, for which fD>100pc = 5.1%. This overstatement appears in the discussion that motivates galaxy-evolution implications and should be corrected, along with the related framing in Section 6.1.2 that the low velocity dispersion is the only case below 10%.","section":"Section 6.2.2"},{"comment":"The analytic model in Section 5 is calibrated with the fiducial simulation and the same metallicity variations used for the comparisons; the reported 0.5% and 1% agreements are therefore in-sample fit qualities, not validation against independent data. The adequacy of the model as a subgrid replacement for hydrodynamical simulations would be much better supported by a holdout test, for example fitting on m11h and predicting the ChaNGa r442 run, or leaving out one metallicity variation and predicting it. If the authors instead intend the numbers as fit residuals, that should be stated explicitly.","section":"Section 5.1 / Figures 9-10"},{"comment":"The claim that the distributions are \"surprisingly insensitive\" is based mainly on medians, but the tails that are most relevant for feedback vary substantially across Table C1: fD>100pc ranges from 5.1% to 45.3%, fD>500pc from 0.0% to 3.1%, and ft>44Myr from 12.3% to 34.6%. The robustness summary should explicitly separate median stability from tail sensitivity, and the abstract's phrase \"surprisingly insensitive to most of these variations\" should be qualified accordingly.","section":"Section 6.1.2 / Figure 13"}],"minor_comments":[{"comment":"The >100 pc fraction is given as about 13% in the abstract and about 14% in conclusion item 2, while Table C1 lists 13.2% for the fiducial model; please standardize the quoted value.","section":"Abstract / Section 8"},{"comment":"The second parameter of the beta distribution is labelled beta_B/C, but in the preceding line it is defined as beta_CE; the label should be made consistent.","section":"Equation (18)"},{"comment":"Section 6.1.1 states ft>44Myr = 0% for the single-star model, whereas the Singles row of Table C1 gives 1.3%; either round explicitly or quote the tabulated value.","section":"Section 6.1.1 / Table C1"},{"comment":"The values of feject in Eq. (5) and the mixture fractions in Eqs. (8)-(11) are presented without uncertainties or sample sizes; reporting these would help users of the analytic model gauge its precision.","section":"Section 5.2"},{"comment":"The markers for ft>44Myr, fD>100pc and fD>500pc may be hard to distinguish in grayscale print; consider different marker shapes or a table callout for the key values.","section":"Figure 13"}],"recommendation":"major_revision","confidential_remarks":"The paper is valuable and generally well executed, and I believe the qualitative conclusions are correct. My recommendation of major_revision is driven by the conditional nature of the quantitative displacement headline and by the in-sample-only validation of the analytic model; both can be addressed within the scope of a revision. I would also suggest that the abstract and conclusions avoid stating the 13% figure without the vdisp caveat, as this is the number most likely to be quoted by users of the subgrid model."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here is my read of arXiv:2504.17903.\n\nThis paper does something genuinely new: it puts binary population synthesis into realistic galactic potentials and produces the joint time-distance distribution of core-collapse SNe, with code and data on GitHub/Zenodo. The central message is that binaries delay and displace a meaningful fraction of SNe relative to single-star feedback prescriptions. I believe that message is right. The delay tail is robust—even the most conservative variation still has 12–13% of SNe after 44 Myr, versus 0% for singles—and the displacement effect is qualitatively robust: at the lowest cluster velocity dispersion they consider, binaries still give 5.1% beyond 100 pc versus 0.9% for single stars. The comparison to Zapartas et al. and Renzo et al. is careful, and Appendix A cleanly resolves the apparent metallicity trend disagreement.\n\nThe soft spots are real but not disqualifying. The headline ~13% beyond 100 pc is a one-point draw on the initial cluster velocity dispersion, which the paper itself says is unconstrained. Table C1 brackets the value from 5.1% to 45.3% across vdisp = 0.5–5 km/s. That conditional nature should be in the abstract, not buried in Section 4.6.1. Next, the analytic model in Section 5 is fit and then 'validated' on the same fiducial simulation; the quoted 0.5% and 4% are goodness-of-fit, not independent predictions. The authors could strengthen this by testing on the ChaNGa r442 run as an out-of-sample check. Finally, the percentage tails carry no Monte Carlo error bars from the finite binary sampling; I suspect the effects are much larger than the sampling noise, but the paper doesn't say.\n\nThe limitations section is unusually honest—rejuvenation may overpredict delays, many massive stars may implode rather than explode, and dynamical cluster ejections are neglected, making distances a lower limit. None of these are hidden.\n\nWho is this for? Anyone building subgrid SN feedback prescriptions for hydrodynamical galaxy simulations, and pop-synth folks who care about delay-time distributions. It deserves serious peer review. My recommendation: send it to review, but ask the authors to (1) present the vdisp dependence of the headline fractions in the abstract/conclusions, (2) add sampling uncertainties, and (3) present the analytic model as a fit with an explicit out-of-sample test rather than an in-sample validation.","headline":"A thorough, well-released simulation study showing binaries delay and displace core-collapse SNe, but the headline displacement fraction rests on an unconstrained cluster velocity dispersion; the qualitative result is robust, the specific 13% is not.","tokens_in":50688,"tokens_out":3072,"would_cite":true,"duration_ms":31056,"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":"Most massive stars have binary partners, which this paper shows delays about 25% of core-collapse supernovae and moves 13% more than 100 parsecs from their birth clusters.","keywords":["core-collapse supernovae","binary stellar evolution","supernova feedback","population synthesis","runaway stars","galactic dynamics","dwarf galaxies","stellar mergers"],"falsifier":"Measure the internal velocity dispersions of young embedded clusters near $10^4\\,M_\\odot$: if reliable measurements cluster near 0.5 km/s rather than 1.7 km/s, the displacement fraction drops toward 5%, while values near 5 km/s would push it to 45%, settling whether the 13% headline holds. Independently, count core-collapse supernovae in regions of galaxies that lack young massive stars: the model predicts roughly a quarter of all core-collapse explosions occur more than 44 Myr after the birth burst, so an observed late fraction well below a quarter would rule out the timing tail's size.","tokens_in":49631,"feed_emoji":"💥","tokens_out":16067,"duration_ms":142310,"temperature":0.7,"pith_summary":"Most massive stars form in binaries or higher-order multiples, yet galaxy-formation simulations routinely assume every star evolves alone when deciding where and when supernova feedback strikes the gas. This paper argues that assumption materially mispredicts the feedback: in its fiducial model of a dwarf galaxy, binary interactions delay about 25% of core-collapse supernovae past the 44-million-year cutoff used in current feedback prescriptions, and displace about 13% of them more than 100 parsecs from their birth clusters, against essentially zero and 1% for single stars. The delayed explosions are almost entirely products of stellar mergers; the displaced ones are mostly companion stars ejected as runaways when their partner exploded. Because the timing and location of each explosion control how efficiently supernovae regulate star formation and drive outflows, getting the distribution wrong changes what simulations predict about galaxies, most strongly at low metallicity and high redshift. The paper's practical deliverable is a metallicity-dependent analytic model that reproduces the simulated joint distribution and can be substituted for single-star subgrid feedback prescriptions.","feed_headline":"Binary stars delay 25% of supernovae and push 13% far from home","feed_subtitle":"Mergers stretch explosions past the 44-million-year cutoff; ejected companions detonate beyond their birth clusters.","key_machinery":"The central machine is cogsworth, the paper's population-synthesis-plus-galactic-dynamics framework, which replaces young star particles in a hydrodynamical dwarf-galaxy simulation with clusters of binary stars, evolves each binary with the COSMIC rapid population-synthesis code, and simultaneously integrates every star's galactic orbit through a potential fitted to the simulated galaxy, so that the time and position of each core-collapse supernova are recorded self-consistently. The argument runs on three binary mechanisms: mass transfer, which lengthens the donor's nuclear timescale and delays its explosion; stellar mergers, which combine two stars below the single-star core-collapse threshold into one star that explodes later, producing the long tail; and binary disruption at the first supernova, which launches the surviving secondary as a runaway at roughly its orbital velocity. The analytic deliverable is a metallicity-dependent piecewise power-law supernova rate with an exponential tail, paired with a four-component mixture model for progenitor ejection velocities that distinguishes unejected stars and ejections after no mass transfer, case A, case B/C, or common-envelope evolution.","core_discovery":"On its own terms, the central discovery is that binary interactions reshape the joint time-distance distribution of core-collapse supernovae in a way that single-star prescriptions cannot capture. In the fiducial simulation, the median supernova occurs 22 Myr after a star-formation event and 35 pc from its parent cluster, compared with 17 Myr and 23 pc for an equivalent single-star population, and binaries produce about 11% more supernovae overall because mergers and accretion let stars below the single-star core-collapse threshold still explode. The two headline features are a long late tail, with 25% of supernovae exploding after the 44 Myr at which the last single star explodes, almost all of them merger products, and a long-distance tail, with 13% of supernovae more than 100 pc from their cluster, dominated by secondary stars ejected at their orbital velocity when the primary exploded. The paper further claims these distributions are surprisingly stable across wide variations in binary physics, initial conditions, and host galaxy, with medians typically moving by less than 15%, while both tails strengthen at low metallicity, reaching about 34% late and 21% beyond 100 pc at one-tenth solar metallicity. It concludes that this stability justifies an analytic fit, and presents one along with a sampling routine for use in hydrodynamical simulations.","pith_inferences":["A testable observational corollary: surveys of nearby core-collapse supernovae should find a population of 'orphan' explosions with no young massive stars nearby, and those orphans should be systematically old; the paper's joint time-distance distribution predicts exactly this correlation and could be read off existing supernova remnant catalogs.","Because the displacement numbers follow from the cluster velocity dispersion, the portability of the analytic model hinges on matching its velocity-dispersion input to each simulation's own cluster dissolution treatment, and the paper's choice of fitting velocities rather than distances is what makes such matching possible.","The same population-synthesis physics that generates the delayed merger-product tail also sets the merger rates of compact-object binaries, so the predicted roughly 25% late-supernova fraction is a consistency check for gravitational-wave progenitor models built on the same binary physics.","The paper implies a redshift-dependent feedback geometry: in compact, low-metallicity high-redshift galaxies the energy is deposited later and farther from dense gas, so simulations adopting the low-metallicity fits should see systematically different gas retention and star-formation histories than those using solar-metallicity single-star prescriptions."],"forward_implications":["Simulations that keep single-star supernova prescriptions omit roughly a quarter of core-collapse explosions and place about 13% of the feedback energy more than 100 pc away from where the simple model puts it, so adopting binary-aware feedback should change the predicted efficiency of star formation regulation and outflow driving.","The paper's analytic fits reproduce the simulated timing distribution to within 0.5% and the ejection-velocity distribution to within a few percent, so they can be installed into existing hydrodynamical codes at negligible computational cost.","Because both the late and the distant tails grow at low metallicity (roughly 34% late and 21% beyond 100 pc at $Z = 0.1\\,Z_\\odot$), the error in single-star prescriptions is largest in exactly the regime occupied by high-redshift galaxies.","A longer, smoother energy-release history turns supernova feedback from an impulsive burst into a gradual push, which the paper argues could reduce the burstiness of star formation and change how the interstellar medium responds to successive explosions.","In dwarf galaxies with effective radii below about a kiloparsec, the displaced supernovae traverse a substantial fraction of the galaxy, so binary-driven feedback automatically becomes a galaxy-wide process and a plausible contributor to dwarf outflows."],"supporting_citations":[{"why":"Introduces cogsworth, the population-synthesis-plus-dynamics framework that produces every simulation in this paper.","marker":"Wagg et al. 2025a,b"},{"why":"Supplies COSMIC, the rapid binary population-synthesis engine that evolves each binary and decides which stars reach core collapse.","marker":"Breivik et al. 2020"},{"why":"Established the delayed core-collapse supernova channel from binary mergers and provides the main quantitative comparison for the timing tail.","marker":"Zapartas et al. 2017"},{"why":"Characterized runaway-star ejection from binaries and provides the comparison baseline for the displacement results.","marker":"Renzo et al. 2019"},{"why":"Provides the empirical distributions of orbital periods, eccentricities, and mass ratios used to initialize the binary populations.","marker":"Sana et al. 2012"},{"why":"Defines the FIRE-3 single-star feedback prescription with the 44 Myr supernova cutoff that this paper's analytic model is designed to replace.","marker":"Hopkins et al. 2023a"},{"why":"Provides the BSE stellar-evolution fitting formulae that underlie COSMIC's treatment of binary physics.","marker":"Hurley et al. 2002"},{"why":"Supplies the Maxwellian natal-kick velocity distribution assumed for core-collapse supernovae.","marker":"Hobbs et al. 2005"}],"fun_headline_variants":["Binary stars push 13% of supernovae far from their birth clusters","Binary interactions make 25% of supernovae explode late","Merger products delay supernovae by up to 200 million years","Binary stars boost supernova counts and stretch their timing","Low-metallicity binaries delay SN feedback even more"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is the assumed initial velocity dispersion of young stellar clusters, set to 1.7 km/s in the fiducial model with no strong observational constraint; the fraction of supernovae beyond 100 pc swings from about 5% at 0.5 km/s to about 45% at 5 km/s, so the paper's 13% displacement headline rides on this one unconstrained input.","fun_headline_variants_meta":{"raw":{"variants":["Binary stars push 13% of supernovae far from their birth clusters","Binary interactions make 25% of supernovae explode late","Merger products delay supernovae by up to 200 million years","Binary stars boost supernova counts and stretch their timing","Low-metallicity binaries delay SN feedback even more"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00068,"raw_usage":{"total_tokens":3183,"prompt_tokens":1130,"completion_tokens":2053,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":746,"completion_tokens_details":{"reasoning_tokens":1966}},"tokens_in":746,"tokens_out":2053,"duration_ms":13811,"temperature":1.0,"reasoning_tokens":1966,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T10:29:08.158066+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the internal velocity dispersions of young embedded clusters near $10^4\\,M_\\odot$: if reliable measurements cluster near 0.5 km/s rather than 1.7 km/s, the displacement fraction drops toward 5%, while values near 5 km/s would push it to 45%, settling whether the 13% headline holds. Independently, count core-collapse supernovae in regions of galaxies that lack young massive stars: the model predicts roughly a quarter of all core-collapse explosions occur more than 44 Myr after the birth burst, so an observed late fraction well below a quarter would rule out the timing tail's size.","supporting_citations":[],"review_version":1}