{"id":"97e0ff70-e815-4b54-83d6-de352fed4bf2","arxiv_id":"2611.17224","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"C-O shell mergers in massive stars are the dominant source of galactic potassium, and the observed [K/H] vs [Fe/H] trend requires these mergers to occur at all metallicities.","lead":"This paper investigates the origin of potassium (K) in the Milky Way by studying how mergers between the oxygen and carbon shells in massive stars just before they explode affect the production of this element. The authors show that without these shell mergers, models cannot match the observed amounts of potassium in stars, and that the pattern of potassium across different metallicities can be used to infer how often these mergers happen.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek/deepseek-v4-flash","headline":"The merger identification criterion (28Si > 20Ne in the final snapshot) has not been validated against full evolutionary histories or 3D simulations, and the authors explicitly state that a more systematic analysis is forthcoming.","rationale":"The reader's weakest_assumption correctly identifies the merger identification criterion as the most vulnerable point. The paper's central argument is that C-O mergers cause the K enhancement; this requires a reliable method to identify which models experience a merger. The authors are transparent that the criterion is preliminary and that a more systematic analysis is forthcoming. However, the paper draws strong conclusions (e.g., 'without progenitors that experienced C-O shell merger, none of the GCE models can reproduce the observed evolution of K') that depend on the accuracy of this classification. The concern is not that the criterion is wrong, but that it is unvalidated. The concrete test proposed would directly assess the criterion's reliability by comparing it to a time-evolution-based classification. If the criterion holds up, the central claim is strengthened; if it fails, the quantitative results could change. The verdict should remain CONDITIONAL because the paper's own caveats are appropriately stated, but the concern is real and testable. I agree with the reader's assessment.","tokens_in":21634,"tokens_out":746,"duration_ms":13213,"concrete_test":"Re-analyze the F23 11 Msun single and binary models using the full MESA history files (available on Zenodo) to track the evolution of the 28Si/20Ne ratio throughout the last ~10^4 years before collapse. If the ratio exceeds unity only in the final snapshot and not during the merger event itself, or if it exceeds unity in models that do not show a flat entropy/composition profile (i.e., incomplete mergers), then the criterion is unreliable. Compare the set of progenitors flagged by the snapshot criterion against a criterion based on the time-integrated history of convective boundary mixing (e.g., when the O-shell base moves outward). If the two lists disagree by more than one progenitor, the classification is too fragile to support the central claim.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The central claim—that C-O shell mergers are the dominant source of galactic K—depends entirely on correctly identifying which progenitors undergo a merger. The identification uses a single criterion (mass-averaged 28Si > 20Ne in the combined O and C shells at the final pre-supernova snapshot). The authors note (Section 2) that 'a more systematic analysis will be presented in a forthcoming paper.' This criterion has not been validated against the full evolutionary history of the models, against 3D simulations, or against a larger set of 1D models with different mixing prescriptions. If the criterion misclassifies even a few progenitors—especially the low-mass ones that dominate the K budget (11 and 13 Msun in F23, 15 Msun in R18, rotating 15 Msun in LC18)—the quantitative predictions for [K/Fe] and [K/H] would shift substantially. The paper's own Appendix B shows that in the 11 Msun single star, the merger is incomplete and only two convective overturns occur before collapse, yet it is still classified as a merger. The robustness of the classification for such marginal cases is unclear. Without independent validation, the causal link between the identified progenitors and the K enhancement remains correlational rather than proven.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek/deepseek-v4-flash","summary":"The paper re-analyzes 194 pre-supernova models from three independent massive-star yield sets (LC18, F23, R18) to test whether C-O shell mergers are the dominant source of Galactic potassium (K). The authors identify merger progenitors with a compositional criterion (mass-averaged 28Si exceeding 20Ne in the combined O and C shells at the final pre-supernova snapshot), then construct GCE models that differ only in whether the yields of these flagged progenitors are included. They find that in all three sets a small number of merger progenitors (ejecting 10-50 times more K than non-merger stars of similar mass) carry most of the K budget; that without them no model reproduces the observed [K/Fe], which stays sub-solar over most of the metallicity range; and that the metallicity window of the [K/Fe] enhancement coincides with the metallicity window in which merger progenitors exist in each grid. From the primary nature of the K synthesis and the nearly linear observed [K/H]-[Fe/H] relation, they infer that C-O mergers must occur with roughly constant frequency at all metallicities, and propose K as a probe of merger frequency as a function of time.","tokens_in":21820,"tokens_out":7463,"duration_ms":52505,"significance":"If the central claim holds, the paper resolves the long-standing underproduction of K in GCE models and turns K into a falsifiable diagnostic for the occurrence of convective-reactive C-O shell mergers in massive stars. The methodology is a strength: the GCE models differ only in the presence or absence of the flagged merger progenitors, so the attribution is clean, and the authors use publicly available grids (including MESA history files from F23) and reproduce key snapshots independently. The paper is also unusually explicit about its own limitations, flagging the coarse mass grids, the by-construction metallicity extension of the F23 set, and the sensitivity to explodability assumptions. The central qualitative finding - that the K enhancement follows the merger probes in the mass-metallicity plane - is stated as a directly falsifiable claim, which is a genuine contribution if the merger-identification criterion can be trusted.","major_comments":[{"comment":"The central attribution rests entirely on the 28Si > 20Ne criterion applied to the final pre-supernova snapshot, and the paper defers systematic validation to a forthcoming paper. The internal evidence partially mitigates this: the flagged models show flat entropy profiles and strongly enhanced K yields (10-50x), so the criterion correlates with the K-producing events it is meant to select, and Appendix B reconstructs the evolutionary history for the two F23 11 Msun cases. What remains unresolved is the sensitivity of the quantitative claims to the classification threshold. In particular, the direction of the error matters: under-identification of mergers is conservative for the 'without-merger' conclusion (missed mergers only add K to the baseline), whereas false positives would inflate the 'with-merger' curves and the reported budget fractions (e.g., the 23% and 50% contributions of the F23 11 and 13 Msun progenitors in §4.1). I recommend adding an explicit statement of this robustness direction and a simple sensitivity test, e.g., recomputing the GCE results with the completeness cut (Table 1) varied, or with the marginal two-overturn 11 Msun single case excluded, to show how much the [K/Fe] normalization and the budget fractions move.","section":"§2, Table 1, Appendix B","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The paper's own discussion shows that the quantitative normalization of the with-merger models is fragile: if the F23 13 Msun single merger progenitor fails to explode, as found by Boccioli & Roberti (2026) and by the Ertl et al. (2016) criterion, the predicted [K/Fe] would drop substantially, and the paper notes that the 11 and 13 Msun progenitors alone supply most of the Galactic K in the F23 set. Since the solar-metallicity reconciliation in Fig. 2 (solid yellow/red curves approaching the data near [Fe/H]=0) is one of the headline results, the manuscript should present this specific comparison as conditional on the adopted explodability assumptions, and preferably show one illustrative variant (e.g., with the 13 Msun merger progenitor removed) so the reader can see the quantitative sensitivity rather than only reading about it.","section":"§4.3, Fig. 2","authors_acknowledge":"The effect of explodability on the chemical evolution of both light and heavy elements will be the subject of a future, more extended study. 5.","authors_acknowledge_p":0.68}],"minor_comments":[{"comment":"The reference list contains the same paper twice: Roberti, Pignatari, Brinkman et al. 2025 appears as 'Astronomy and Astrophysics, 698, A216' and as 'A&A, 698, A216'; one entry should be removed.","section":"References","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The naming convention for LC18 models (first three characters = ZAMS mass, letters a-d for [Fe/H], last three = v_rot) is explained only in the Table 1 notes; it would help to define it at the first use in the text, where models such as 013a300 and 013c150 are discussed before the reader reaches the table.","section":"§2, Table 1","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The grey observational points mix Chen et al. (2000) and Mittal & Roederer (2025) without distinguishing them; given that the 'nearly linear' [K/H]-[Fe/H] relation is a load-bearing premise for the all-metallicity-merger inference, a panel with a fitted slope (and its uncertainty) would strengthen the argument considerably.","section":"Fig. 2","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The piecewise form of f(Z) is described in words; stating explicitly the two transition metallicities (the range where f(Z) ramps between its -1 and 0 bounds) in the text would make the by-construction nature of the F23 metallicity extension easier to check.","section":"§3.2, Eq. (6)","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The discussion of the coarse R18 and LC18 grids correctly identifies that a coarse grid may spread a single (M,Z) merger point over a wide range via interpolation; this could be quantified by showing how much of the [K/Fe] enhancement above the dashed curves in Fig. 2 comes from interpolated versus directly computed merger weights.","section":"§4.3","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The abstract's statement that rotating progenitors have more mergers at [Fe/H] < -1 while non-rotating ones do near solar metallicity is correct when 'non-rotating' refers to the R18 set, but the phrasing could mislead since LC18 also has rotating merger models at solar metallicity (20 Msun v_rot=150 and 13 Msun v_rot=300); a small clarifying clause would prevent confusion.","section":"Abstract and §4","authors_acknowledge":null,"authors_acknowledge_p":null}],"recommendation":"major_revision"},"author_rebuttal":null,"desk_editor":{"model":"deepseek/deepseek-v4-flash","letter":"Molero, Boccioli, and Arcones make a clean, controlled point: in all three massive-star yield sets, the potassium that brings the GCE models close to observations comes from a small number of progenitors flagged as C-O shell mergers, and the metallicity range of the enhancement tracks where those progenitors sit in the mass-metallicity plane. Removing those progenitors makes every model underpredict K. The with/without comparison, with all other GCE ingredients fixed, is the paper's real contribution. It is not a new nucleosynthetic mechanism—Ritter et al. already proposed C-O mergers for odd-Z elements—but the systematic demonstration across three independent sets is new, and the quantitative point that 11 and 13 Msun progenitors supply roughly a quarter and a half of Galactic K in the F23-based models is striking.\n\nThe authors are unusually honest about their own caveats. They state clearly that the near-linear [K/H] in the F23 set is a consequence of Eq. (6), not a prediction, and they explicitly say that robust conclusions on merger frequency as a function of mass, metallicity, and rotation are not possible at this stage. That should be credited: the paper does not try to sell the extrapolation as a result.\n\nThe soft spot is the merger diagnostic. The 28Si > 20Ne criterion is applied to a single pre-SN snapshot, and the authors themselves say a more systematic analysis is forthcoming. The marginal 11 Msun single star in Appendix B, with only about two convective overturns before collapse and a non-flat abundance profile, is still classified as a merger and contributes significantly to the K budget. That is exactly the kind of case where the criterion's reliability matters. If several marginal events were reclassified, the quantitative yields would shift. I do not think this sinks the qualitative conclusion: the full-merger cases have unambiguous entropy and composition signatures, and the same pattern repeats across all three yield sets. But a referee should ask for sensitivity tests around the classification threshold, and the paper should avoid letting the selection criterion double as the definition of the physical event. A second, smaller lever is explodability: since so much of the K comes from a few low-mass progenitors, whether 13 Msun stars actually explode matters a lot, and the authors flag this as future work.\n\nI would send this to a serious referee. It is a useful, citable paper for anyone working on GCE or late-stage massive-star evolution, and the authors' transparency makes the remaining uncertainties manageable rather than hidden.","headline":"A well-controlled GCE comparison showing that a handful of C-O shell merger progenitors carry the potassium budget in all three yield sets; the main weakness is the unvalidated merger diagnostic, not the acknowledged low-metallicity extrapolation.","tokens_in":22554,"tokens_out":3958,"duration_ms":34860},"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek/deepseek-v4-flash","headline":"Galactic potassium traces the frequency of C–O shell mergers in massive stars","keywords":["Galactic chemical evolution","potassium nucleosynthesis","C–O shell merger","massive stars","odd-Z elements","convective-reactive events","stellar yields","pre-supernova evolution"],"falsifier":"If a future 3D simulation of a massive star in the 11–20 M☉ range shows that the O and C shells never mix to the degree required by the 28Si > 20Ne criterion, or that the nucleosynthetic yield of 39K from such mixing is less than a factor of 2 above the non-merger yield, then the claim that C–O mergers are the dominant source of Galactic K would be falsified. Alternatively, if stellar spectroscopy at [Fe/H] < −2 shows a linear [K/H] vs. [Fe/H] trend that is already reproduced by GCE models without any C–O merger progenitors, the paper's central conclusion would be contradicted.","tokens_in":22947,"feed_emoji":"🌟","tokens_out":2214,"duration_ms":17036,"temperature":0.7,"pith_summary":"The paper tries to establish that the Galactic chemical evolution of potassium (K) is almost entirely controlled by a rare convective phenomenon in massive stars: the merger of the oxygen and carbon shells a few hours before core collapse. Without these C–O shell merger progenitors, every model studied here underproduces K relative to stellar observations across nearly all metallicities. When the merger progenitors are included, the predicted [K/Fe] rises only over the metallicity range in which the mergers actually occur in each stellar grid. Because the K produced in a C–O shell merger is primary (it does not depend on the star's initial metal content), the observed near-linear increase of [K/H] with [Fe/H] can be reproduced only if C–O shell mergers happen with a roughly constant frequency at all metallicities. The authors conclude that K can therefore be used as a tracer of the occurrence rate of C–O shell mergers in massive stars as a function of time, a quantity that is otherwise very uncertain.","feed_headline":"Potassium in the Galaxy traces rare shell mergers in massive stars","feed_subtitle":"Without C–O merger progenitors, no chemical evolution model reproduces the observed K abundance across metallicities.","key_machinery":"The central object is the C–O shell merger, a convective–reactive event in which the entropy barrier between the oxygen and carbon burning shells disappears a few hours before core collapse, allowing carbon-rich material to be ingested into the hot O-burning region. This triggers heavy-ion fusion reactions (12C+12C, 12C+16O, 16O+16O) that release protons, α particles, and neutrons, which are then captured onto O-burning products to synthesize odd-Z elements including 39K via a chain such as 34S(α,γ)38Ar(p,γ)39K. The authors identify merger progenitors in pre-supernova snapshots using a single diagnostic: in the combined O and C shells, the mass-averaged fraction of 28Si exceeds that of 20Ne. This criterion is applied to 194 models from three yield sets. The paper then runs a standard one-infall multi-zone GCE model of the Galactic disc, comparing the predicted [K/Fe] and [K/H] with and without the yields of the identified merger progenitors, keeping all other ingredients (low- and intermediate-mass stars, Type Ia supernovae, novae, neutron-star mergers, magneto-rotational supernovae) identical.","core_discovery":"The central claim is that the Galactic evolution of potassium is set not by the full population of core-collapse supernovae but by a small subset of progenitor stars that experience a carbon–oxygen shell merger in the final hours before collapse. In three independent yield sets (LC18, R18, F23), the authors identify the few models that satisfy a compositional criterion for a C–O merger (mass-averaged 28Si exceeding 20Ne in the merged shell). These merger progenitors eject 10–50 times more K than non-merger stars of similar mass. When those yields are removed from the Galactic chemical evolution model, every set predicts sub-solar [K/Fe] over most of the metallicity range, in conflict with observations. When they are included, the enhancement appears only in the metallicity window where the grid contains merger progenitors: at [Fe/H] ≲ −1 for the rotating models of LC18, and near solar metallicity for the non-rotating NuGrid models of R18. The set of Farmer et al. (2023), which by construction extends the solar-metallicity merger progenitors to all metallicities, produces a nearly uniform upward shift in [K/Fe] and a linear [K/H] vs. [Fe/H] relation. The authors argue that because the K synthesized in a C–O merger is primary, the shape of the Galactic K evolution is determined by the frequency of mergers as a function of metallicity, not by the yield per event.","pith_inferences":["If the identification of C–O mergers can be refined with 3D simulations and a more systematic 1D criterion, the K abundance in stars of different metallicities could become a direct observational constraint on the physics of convective boundary mixing in late-stage massive stars, a regime that is otherwise inaccessible.","The paper's result suggests that the underproduction of other odd-Z elements (P, Cl, Sc) in GCE models may share the same explanation—C–O shell mergers—and that a unified treatment of these elements could provide a consistency check on the merger frequency inferred from K alone.","Because the merger criterion (28Si > 20Ne) is applied to the final pre-supernova snapshot and not to the full evolutionary history, the set of identified mergers may be incomplete. A time-resolved analysis could reveal additional merger events that are not captured at the final snapshot, potentially widening the metallicity range over which mergers contribute.","The near-linear [K/H] vs. [Fe/H] observed in the Galaxy could be used to predict the merger rate at subsolar metallicities, where direct stellar evolution calculations are sparse; if future surveys find a deviation from linearity, it would indicate a change in the merger frequency with cosmic time."],"forward_implications":["The observed linear rise of [K/H] with [Fe/H] implies that C–O shell mergers must occur at a roughly constant frequency across all metallicities; any model that concentrates mergers in a narrow metallicity window will produce a change of slope in [K/H] that is not seen in stellar data.","Because a few low-mass merger progenitors (11–20 M☉) dominate the Galactic K budget, the failure of even one of them to explode would substantially lower the predicted [K/Fe] and [K/H], making the K abundance a sensitive probe of the explodability of stripped and single stars in this mass range.","Binarity itself plays a minor role in K production: single stars alone produce similar amounts of K as binary populations, because the K production is tied to the C–O merger event itself, not to whether the star has been stripped by a companion.","The metallicity dependence of the K enhancement in the LC18 and R18 sets shows that the Galactic evolution of K directly maps the location of C–O merger progenitors in the mass–metallicity plane, offering a way to test stellar evolution models against observations."],"supporting_citations":[{"why":"First proposed that C–O shell mergers in massive stars enhance the production of odd-Z elements including K, and showed that the NuGrid GCE model with these mergers reproduces the solar K abundance only at high metallicity.","marker":"Ritter et al. (2018a)"},{"why":"Provides the rotating massive star yield set (108 models) that the authors re-analyse to identify C–O merger progenitors and to run GCE models with and without their contribution.","marker":"Limongi & Chieffi (2018) (LC18)"},{"why":"Provides the dense grid of MESA models for single and binary-stripped stars at solar metallicity, which the authors use to study the role of binarity and to extend merger progenitors to all metallicities via a metallicity-dependent weighting.","marker":"Farmer et al. (2023) (F23)"},{"why":"Provides the non-rotating NuGrid yield set (20 models) that the authors use as the third independent set, where C–O mergers are confined to the highest metallicities.","marker":"Ritter et al. (2018b) (R18)"},{"why":"Demonstrates that the odd-Z production in a C–O shell merger is primary (does not depend on the star's initial metal content), a key assumption for the inference that the merger frequency shapes the Galactic K evolution.","marker":"Roberti et al. (2025)"},{"why":"Shows that several lower-mass C–O merger progenitors (∼12–15 M☉) may fail to explode in self-consistent neutrino-driven simulations, which would substantially lower the predicted K abundance.","marker":"Boccioli & Roberti (2026)"},{"why":"Provides the metallicity-dependent distribution of initial rotational velocities that the authors adopt to weight the rotating models of LC18, which determines the contribution of rotating C–O merger progenitors at low metallicity.","marker":"Prantzos et al. (2018)"}],"fun_headline_variants":["Potassium tracks rare shell mergers in massive stars","Galactic potassium reveals hidden stellar mergers","C–O shell mergers: key to potassium in the Galaxy","Rare stellar mergers power potassium enrichment","Potassium abundance traces dying star mergers"],"cache_read_input_tokens":0,"weakest_assumption_plain":"The identification of a C–O shell merger relies on a single snapshot criterion (mass-averaged 28Si > 20Ne in the combined O and C shells) that has not been validated against 3D simulations or a larger set of 1D models with different mixing prescriptions, and the authors note that a more systematic analysis is forthcoming.","fun_headline_variants_meta":{"raw":{"variants":["Potassium tracks rare shell mergers in massive stars","Galactic potassium reveals hidden stellar mergers","C–O shell mergers: key to potassium in the Galaxy","Rare stellar mergers power potassium enrichment","Potassium abundance traces dying star mergers"]},"model":"deepseek/deepseek-v4-flash","effort":"low","cost_usd":0.00011,"raw_usage":{"cost":0.0001098552,"is_byok":false,"cost_details":{"upstream_inference_cost":0.0001098552,"upstream_inference_prompt_cost":9.92992e-05,"upstream_inference_completions_cost":1.0556e-05},"total_tokens":1354,"prompt_tokens":1296,"completion_tokens":58,"prompt_tokens_details":{"audio_tokens":0,"video_tokens":0,"cached_tokens":256,"cache_write_tokens":0},"completion_tokens_details":{"audio_tokens":0,"image_tokens":0,"reasoning_tokens":0}},"tokens_in":1296,"tokens_out":58,"cost_source":"openrouter_usage","duration_ms":4542,"image_pages":[],"temperature":1.0,"billed_attempts":1,"reasoning_tokens":0,"transport_provider":"openrouter","cache_read_input_tokens":0,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-11-13T09:01:14.480814+00:00","model_set":{"reader":"deepseek/deepseek-v4-flash"},"pipeline_version":"pith-pipeline@v1.4.1-daily-deepseek-or-medium","input_kind":null,"input_chars":null,"input_total_chars":null,"passes_missing":null,"published_standing":null,"published_confidence":null,"objection_audit":null,"run_record":null,"basis":{"kind":"not_recorded","source":null,"chars_given":null,"chars_total":null},"receipt":null,"falsifier":"If a future 3D simulation of a massive star in the 11–20 M☉ range shows that the O and C shells never mix to the degree required by the 28Si > 20Ne criterion, or that the nucleosynthetic yield of 39K from such mixing is less than a factor of 2 above the non-merger yield, then the claim that C–O mergers are the dominant source of Galactic K would be falsified. Alternatively, if stellar spectroscopy at [Fe/H] < −2 shows a linear [K/H] vs. [Fe/H] trend that is already reproduced by GCE models without any C–O merger progenitors, the paper's central conclusion would be contradicted.","supporting_citations":[{"cited_title":"The Role of Rotation on the Yields of the Two","cited_arxiv_id":null,"evidence_quote":"Demonstrates that the odd-Z production in a C–O shell merger is primary (does not depend on the star's initial metal content), a key assumption for the inference that the merger frequency shapes the Galactic K evolution."}],"lane_accuracy":{"status":"not yet measured","lane":{"model":"deepseek/deepseek-v4-flash","input_kind":null},"url":"https://pith.science/accuracy"},"review_version":1}