{"id":"ae40e47c-726a-41a4-bb56-8833959caafa","arxiv_id":"2411.18680","paper_version":2,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"In a star-by-star dwarf galaxy simulation, only 11% of stars at [Fe/H] = -5 and about 1% at [Fe/H] = -2.5 were enriched by a single supernova, implying most metal-poor stars formed from gas mixed by several supernovae.","lead":"Simulations of a dwarf galaxy that track every supernova show that stars made from the debris of just one supernova are rare: about 11% of the most iron-poor stars and only 1% of moderately metal-poor stars. The result gives observers a concrete expectation for finding pristine stellar fossils in upcoming surveys such as Subaru's Prime Focus Spectrograph.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The mono-enriched fractions and the claimed trend rest on a single unvaried diffusion coefficient and an uncalibrated 1e-6 [C/Fe] tolerance; the paper's assertion that the diffusion coefficient cannot affect the trend is unsupported.","rationale":"The paper is a first direct simulation estimate of the mono-enriched fraction, with explicit star-by-star enrichment bookkeeping and a transparent statement of several limitations, so a rejection is not warranted. The reader's conditional verdict already identifies the diffusion coefficient and the 1e-6 tolerance as the weakest point. My stress test agrees but goes slightly further: the paper's specific argument that the diffusion coefficient cannot affect the trend is logically incomplete, because the relevant quantity is not the coefficient's dependence on metallicity but the mixing efficiency relative to the local SN rate and gas density, which do evolve with [Fe/H]. This makes the monotonic trend itself part of what needs robustness testing, not just the absolute percentages. The proposed concrete test varies the diffusion coefficient and the classification tolerance in the same simulation; this directly probes whether the headline numbers and the trend survive reasonable changes to the two free ingredients on which the classification rests. If the test passes, the conditional verdict can move toward accept; if it fails, the paper's central claim would need substantial revision. Since the test has not been run, keeping the reader's conditional verdict is the appropriate outcome.","tokens_in":12374,"tokens_out":9431,"duration_ms":90366,"concrete_test":"Rerun the same zoom-in with the metal-diffusion scaling factor set to 0.0, 0.01, and 0.1 (one order of magnitude each way), and recompute the mono-enriched fraction with [C/Fe] tolerances of 1e-6, 1e-4, and 0.01. The central claim is supported if the monotonic increase below [Fe/H] = -2.5 persists in all runs and the values at [Fe/H] = -5, the EMP average, and the total change by less than a factor of two; if the slope flattens or reverses, or the absolute values move by more than that, the headline numbers are artifacts of the chosen diffusion coefficient and tolerance.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central quantitative claim (11% at [Fe/H] = -5, 5% for EMP stars, 1.5% overall, monotonic increase below [Fe/H] = -2.5) is produced by classifying star particles whose [C/Fe] matches a CCSN yield within 1e-6. That classification inherits the metal-mixing model of Section 2.1, where the metal diffusion equation is multiplied by a scaling factor of 0.01 calibrated in Hirai & Saitoh (2017). No variation of this factor is run, and only one halo is simulated. Section 4 attempts to dismiss the parameter by saying it does not affect the trend 'because the metallicity does not affect the value of the diffusion coefficient.' This is a non sequitur: the effect of diffusion on whether a gas parcel is mono-enriched depends on the local density, the time between SNe, and the enrichment history, all of which correlate with [Fe/H]. Over-mixing can erase mono-enriched signatures at all metallicities; under-mixing can leave spurious exact matches. The 1e-6 tolerance is likewise arbitrary: by construction any multi-enriched parcel with a second-SN contribution below that threshold is counted as mono-enriched, so the reported fractions are a function of an uncalibrated cutoff. With no convergence test and no release of simulation outputs, the absolute fractions and even the monotonic trend are not yet secured against this single degree of freedom.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This Letter presents a cosmological zoom-in simulation of a dwarf galaxy, run with the star-by-star SIRIUS model, and uses it to estimate the fraction of metal-poor stars enriched by a single core-collapse supernova (mono-enriched stars). The authors classify a simulated star as mono-enriched if its [C/Fe] ratio matches the yield of a previously occurring CCSN to within ±10^-6 in dex. They report that the mono-enriched fraction increases toward lower metallicity for [Fe/H] < -2.5, reaching 11% at [Fe/H] = -5 and 1% at [Fe/H] = -2.5, with an overall fraction of 1.5% in the simulated galaxy at z = 6.5. They also find that mono-enriched stars are preferentially located near the galactic center. The paper compares these numbers with the observation-based estimate of Hartwig et al. (2023) and discusses implications for upcoming surveys such as PFS.","tokens_in":12590,"tokens_out":5110,"duration_ms":54964,"significance":"If the quantitative results are robust, this is a valuable first direct simulation-based estimate of the mono-enriched fraction, complementary to machine-learning analyses of observed stellar abundances. The paper's approach of directly tracking individual SN ejecta in a star-by-star simulation is a clear strength, and the reported spatial bias toward the center of the dwarf provides a concrete, testable prediction for future spectroscopic surveys. The authors are also transparent about several limitations, including the single halo realization, the neglect of radiation feedback, and the absence of CEMP stars. However, the central quantitative claims rest on model choices that are not varied or tested, and at least one of the paper's arguments for robustness (Section 4) is not logically sound. The absolute fractions and even the monotonic trend below [Fe/H] = -2.5 should therefore be regarded as provisional until the sensitivity studies requested below are performed.","major_comments":[{"comment":"The mono-enriched fractions are controlled by the metal diffusion scaling factor of 0.01, which is calibrated to present-day dwarf galaxy abundances (Hirai & Saitoh 2017). The paper asserts in Section 4 that this parameter 'does not affect the predicted increasing trend toward lower metallicity of the mono-enriched fraction, because the metallicity does not affect the value of the diffusion coefficient.' This is a non sequitur: the diffusion coefficient is a constant, but its effect on whether a gas parcel is classified as mono-enriched depends on the local gas density, the time between SN events, and the enrichment history, all of which vary strongly with [Fe/H]. Over-mixing can erase mono-enriched signatures at all metallicities, while under-mixing can leave spurious exact matches. No simulation with a different diffusion coefficient, or with diffusion disabled, is presented. The absolute fractions (11%, 5%, 1.5%) and even the shape of the trend below [Fe/H] = -2.5 are therefore not secured against this single degree of freedom. The authors should provide a sensitivity test or explicitly reframe the results as conditional on the adopted diffusion model.","section":"Section 2.1 and Section 4"},{"comment":"The mono-enriched classification uses a matching tolerance of ±10^-6 in [C/Fe]. This tolerance is far smaller than the numerical scatter in the simulation (the paper notes that some stars in constant-[C/Fe] stripes are not labeled mono-enriched because they are 'slightly affected by other SN ejecta') and far smaller than the observational precision of ±0.2 dex. The reported fractions are therefore a direct function of an arbitrary cutoff: any multi-enriched star whose second SN contribution shifts [C/Fe] by less than 10^-6 is counted as mono-enriched. The paper should quantify the sensitivity of the fractions and the trend to the tolerance, for example by recomputing them with tolerances of 10^-4, 10^-2, and 0.2 dex. The statement in Section 4 that the strict condition makes the estimate 'expected to be lower than observation-based estimates' does not address the internal sensitivity of the reported numbers to this parameter.","section":"Section 3"},{"comment":"Only a single cosmological zoom-in halo is simulated, with no resolution convergence test; the paper explicitly notes that the resolution is below that at which galactic wind properties converge (Hu 2019). The mono-enriched fraction depends on the number of SN events per gas parcel and on the star formation history, both of which vary from halo to halo and with resolution. Without a second halo realization or a resolution study, the reported values of 11% at [Fe/H] = -5 and 1% at [Fe/H] = -2.5 should be accompanied by an uncertainty or a caveat that they are single-realization numbers. As written, the quantitative claims in the abstract and Section 3 are presented without any error estimate, which is difficult to justify for a single simulation.","section":"Section 2.3"},{"comment":"The paper describes the agreement with Hartwig et al. (2023) as 'an independent confirmation of the increasing trend' because the two approaches are 'completely different.' This overstates the degree of independence: the simulation enriches gas with yields from the celib library (Nomoto et al. 2013), and the mono-enriched classification compares simulated stars against the same yield table, so the simulation cannot falsify the nucleosynthesis yields. While the method of estimating the enrichment history is different from the machine-learning approach, the two are not independent validations of the underlying yield models. The authors should temper this claim and explicitly state that the comparison primarily tests the mixing and star-formation model adopted in the simulation.","section":"Section 4"}],"minor_comments":[{"comment":"The metal diffusion scaling factor of 0.01 is said to be 'determined from the chemical abundances of MP stars in dwarf galaxies' (Hirai & Saitoh 2017). A short justification of why this calibration, likely based on local dwarf galaxies, should apply at the very low metallicities and early cosmic times probed here would help the reader assess the extrapolation.","section":"Section 2.1"},{"comment":"Figure 2 shows the mono-enriched fraction as a function of [Fe/H], but no error bars or Poisson uncertainties are provided. Since the fractions are derived from finite particle counts, particularly in the sparsely populated high- and low-metallicity bins, adding confidence intervals (or at least stating the bin counts) would make the trend easier to evaluate.","section":"Section 3"},{"comment":"The sentence explaining why the diffusion coefficient does not affect the trend ('because the metallicity does not affect the value of the diffusion coefficient') is confusing and should be rewritten or removed; as noted above, it conflates a constant parameter value with the metallicity-dependent physical consequences of diffusion.","section":"Section 4"}],"recommendation":"major_revision","confidential_remarks":"The paper presents an interesting and technically challenging simulation, and the qualitative trend (higher mono-enriched fraction at lower metallicity) is plausible and likely robust. However, the quantitative headline numbers are all produced with a single simulation, a single calibrated diffusion coefficient, and an ad hoc 10^-6 [C/Fe] tolerance, none of which are varied or tested. The authors' argument that the diffusion coefficient does not affect the trend is flawed. I would advise the editor that a major revision is needed, asking for either additional sensitivity runs (at least a few diffusion coefficient values and tolerance values) or a substantial reframing of the claims as conditional on the adopted model. The paper is not ready for publication as is, but the core idea and the simulation setup are valuable."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First things first: this is the first simulation-based estimate of the mono-enriched fraction, and the qualitative trend—more mono-enriched stars at lower metallicity—is probably right. The paper does this with a star-by-star dwarf galaxy zoom-in, which is a real step up from IMF-averaged treatments. The enrichment bookkeeping is explicit, and the authors are honest about the small-number artifact above [Fe/H] ~ -1.5 and about having only one halo.\n\nThe new numbers are 11% at [Fe/H] = -5, 1% at -2.5, 1.5% overall, 5% for EMP stars. I would not hang a lot on the absolute values. The matching tolerance is ±1e-6 in [C/Fe], which is machine precision, not physics; the authors acknowledge observational precision would blur the distinction and that their value is a lower limit. But the bigger issue is the metal diffusion scale factor of 0.01, calibrated once against dwarf galaxy abundances and never varied. The paper's defense—that the coefficient does not affect the trend because metallicity does not change the coefficient—is a non sequitur. Diffusion's effect depends on local density, timing between SNe, and a metallicity-correlated enrichment history. One halo, no variation, no convergence test at the resolution that matters for mixing: the absolute fractions could shift a lot under a different mixing prescription. The qualitative trend might survive, but that is not demonstrated.\n\nThere is also a mild circularity: the classification uses the same Nomoto et al. yield library that seeded the gas, so a star formed from pure ejecta matches the template by construction. That means the detection method cannot be falsified inside the simulation. It is a caveat, not a fatal flaw.\n\nThe comparison with Hartwig et al.'s 31.8% is fair, and the discrepancy is sensibly explained by the tolerance and the different approaches. But calling it 'independent confirmation' oversells it, since both use the same yields.\n\nWho is this for? People working on very metal-poor star chemistry and star-by-star simulations. They will read it as the first quantitative benchmark, with the diffusion caveat as homework. For a Letter, the analysis is okay, but the referee should push for at least one variation of the diffusion coefficient or a release of the particle data to allow others to test robustness.\n\nMy take: send to peer review, with major comments. The trend is plausible, the absolute numbers are not yet secure. If I were handling it, I would ask for a robustness test on the tolerance and diffusion before publication.","headline":"First simulation-based mono-enriched fractions are worth taking seriously, but the absolute numbers depend on an unvaried mixing parameter and a machine-precision tolerance.","tokens_in":13265,"tokens_out":3267,"would_cite":true,"duration_ms":87472,"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":"A dwarf galaxy simulation that follows individual massive stars finds that only 1.5% of its metal-poor stars were enriched by a single supernova, with the fraction rising to 11% at [Fe/H] = -5.0.","keywords":["mono-enriched stars","metal-poor stars","supernova nucleosynthesis","dwarf galaxy formation","star-by-star simulations","cosmological zoom-in simulation","Population III stars","chemical abundances"],"falsifier":"Re-run the same zoom-in halo at the resolution where galactic wind properties converge (the paper's own stated missing check) and vary the metal-diffusion scale factor between 0.003 and 0.1; if the mono-enriched fraction at [Fe/H] = -5 moves outside roughly 5-20%, the reported numbers are resolution- and mixing-dependent rather than intrinsic. Alternatively, high-precision spectroscopy of more than 100 EMP stars in Milky Way dwarf satellites that finds a mono-enriched fraction well above the simulated 5% would rule out the model's mixing strength.","tokens_in":12104,"feed_emoji":"🌌","tokens_out":6854,"duration_ms":55887,"temperature":0.7,"pith_summary":"This paper asks what fraction of the oldest, most metal-poor stars were born from the ejecta of a single supernova (mono-enriched stars) rather than from gas seasoned by several explosions. Using a cosmological zoom-in simulation of a dwarf galaxy that follows individual massive stars and their core-collapse supernova yields, it finds that the mono-enriched fraction rises toward lower metallicity: 11% at [Fe/H] = -5.0, 1% at [Fe/H] = -2.5, and only 1.5% overall. The authors conclude that even the most primitive stars are usually polluted by multiple supernovae, and that mono-enriched stars are concentrated near the dwarf's center, making central stellar populations of dwarf galaxies the best targets for finding them.","feed_headline":"Simulation finds most metal-poor stars are multi-supernova enriched","feed_subtitle":"Star-by-star dwarf run: mono-enriched stars are rare, falling from 11% at [Fe/H]=-5 to 1% at -2.5.","key_machinery":"The classification device is the carbon-to-iron ratio. Every simulated core-collapse supernova carries a characteristic [C/Fe] yield, so a star formed from gas whose [C/Fe] matches a past CCSN's yield to within ±$10^{-6}$ is counted as mono-enriched. The simulation resolves stars above 6 solar masses individually as star-by-star particles; their explosion ejecta are distributed to 128 neighboring gas particles and diluted by a metal-diffusion scheme whose scale factor (0.01) was calibrated to match abundances of metal-poor stars in dwarf galaxies. AGB and Type Ia contributions are negligible in the metal-poor regime, and the initial gas is zero metallicity, so the match is a clean test of single-SN enrichment.","core_discovery":"The paper's central claim is that in a dwarf galaxy simulated with individual massive stars, the fraction of mono-enriched stars is small and declines steeply with metallicity: 11% at [Fe/H] = -5.0, 1% at [Fe/H] = -2.5, and 1.5% overall, with an EMP fraction of 5.0%. This is the first direct simulation-based estimate of this fraction, and it independently reproduces the trend toward lower metallicities seen by an observation-based machine-learning analysis of 462 EMP stars, while predicting a substantially lower absolute fraction than that analysis's 31.8% because the simulation's mono-enriched criterion is deliberately strict.","pith_inferences":["Editorial inference: because the simulation classifies only matches within ±10^-6 as mono-enriched, the 1.5% and 5% figures are closer to lower bounds; a classification using observational precision (~0.2 dex in [C/Fe]) would merge many multi-enriched stars into the mono-enriched group, pushing the fractions up and possibly explaining most of the gap with the machine-learning estimate.","Editorial inference: the simulation uses a single halo and a non-converged resolution; the absolute mono-enriched fraction in other dwarfs with different star-formation histories or with a top-heavy Population III IMF could differ by factors of a few, even if the qualitative metallicity trend is universal.","Editorial inference: a direct observational test would be to measure carbon and iron abundances of a large sample of ultra-faint dwarf galaxies with precisions below 0.1 dex; if a significant population of stars shows [C/Fe] ratios offset from every known CCSN yield track, the simulation's assumption that every mono-enriched star inherits its SN's [C/Fe] exactly would need revision."],"forward_implications":["Most metal-poor stars, even at [Fe/H] around -5, carry the imprint of multiple supernovae; single-SN enrichment is the exception rather than the rule.","The rising mono-enriched fraction toward lower metallicity reinforces the trend inferred from observations and supports using the lowest-metallicity stars as the most promising single-SN nucleosynthesis probes.","Targeting the central regions of dwarf galaxies should efficiently find mono-enriched stars, because that is where the simulation places most of them.","If the 5% EMP fraction holds, existing photometric samples of roughly 25,000 EMP stars should contain over a thousand mono-enriched stars for spectroscopic follow-up.","The gap with the machine-learning estimate (5% vs 31.8%) is plausibly explained by the much stricter matching criterion in the simulation, so the two approaches are not necessarily in conflict."],"supporting_citations":[{"why":"Supplies the observation-based machine-learning benchmark (31.8% mono-enriched among 462 EMP stars) that this simulation's fraction is compared against.","marker":"Hartwig et al. 2023"},{"why":"Provides the core-collapse supernova nucleosynthesis yields that define each CCSN's [C/Fe] fingerprint in the simulation.","marker":"Nomoto et al. 2013"},{"why":"Calibrates the metal-diffusion scaling factor (0.01) from abundances of metal-poor stars in dwarf galaxies, the key tuning of the mixing model.","marker":"Hirai & Saitoh 2017"},{"why":"Establishes the resolution below which galactic wind properties have not converged, the paper's acknowledged resolution caveat.","marker":"Hu 2019"},{"why":"Provides the star-by-star stellar mass-assignment and star-formation model on which the simulation's individual massive stars are built.","marker":"Hirai et al. 2021"},{"why":"Gives the swept-up gas mass estimate (10^4 Msun) used to argue that stars at [Fe/H] > -2.5 can still form from a single supernova's ejecta.","marker":"Cioffi et al. 1988"}],"fun_headline_variants":["Rare mono-enriched stars: 11% at [Fe/H]=-5, 1% at -2.5","First simulation: mono-enriched stars scarce, 1.5% overall","Dwarf galaxy sim: most metal-poor stars multi-supernova enriched","Mono-enriched stars drop steeply from 11% to 1% with metallicity"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The mono-enriched fractions rest on the simulation's calibrated metal-diffusion strength and on a single dwarf halo whose resolution is below the level where galactic winds converge; if the real mixing of supernova ejecta is substantially stronger or weaker than this model, the absolute percentages change even if the trend toward lower metallicity survives.","fun_headline_variants_meta":{"raw":{"variants":["Rare mono-enriched stars: 11% at [Fe/H]=-5, 1% at -2.5","First simulation: mono-enriched stars scarce, 1.5% overall","Dwarf galaxy sim: most metal-poor stars multi-supernova enriched","Mono-enriched stars drop steeply from 11% to 1% with metallicity"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000239,"raw_usage":{"total_tokens":1517,"prompt_tokens":953,"completion_tokens":564,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":569,"completion_tokens_details":{"reasoning_tokens":466}},"tokens_in":569,"tokens_out":564,"duration_ms":4998,"temperature":1.0,"reasoning_tokens":466,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T10:59:41.615519+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Re-run the same zoom-in halo at the resolution where galactic wind properties converge (the paper's own stated missing check) and vary the metal-diffusion scale factor between 0.003 and 0.1; if the mono-enriched fraction at [Fe/H] = -5 moves outside roughly 5-20%, the reported numbers are resolution- and mixing-dependent rather than intrinsic. Alternatively, high-precision spectroscopy of more than 100 EMP stars in Milky Way dwarf satellites that finds a mono-enriched fraction well above the simulated 5% would rule out the model's mixing strength.","supporting_citations":[],"review_version":1}