{"id":"23b2678c-9354-41aa-8188-1dfb165a4e8b","arxiv_id":"2608.08380","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Core-collapse supernovae are found in molecular gas with median CO(2-1) velocity dispersion of about 9 km/s, versus 3.8 km/s for typical host-galaxy pixels.","lead":"Core-collapse supernovae sit in unusually turbulent molecular gas, with velocity dispersions about 2.4 times higher than typical cloud pixels in their host galaxies, according to new ALMA CO(2-1) maps of 33 nearby supernovae. This first look at cloud-scale turbulence at explosion sites offers a fresh observational route to understanding how massive stars, including binaries, form.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Appendix B.4's structure-matched sigma excess is +0.002 dex, so the elevated CCSN velocity dispersion may trace dense galactic structures rather than unusually turbulent GMCs.","rationale":"The reader's weakest assumption concerned missing short-spacing flux in 12m-only observations, a valid data-homogeneity issue. However, the more load-bearing problem is internal: the paper's own Appendix B.4, which matches CCSN positions to pixels of the same galactic structure and surface density, finds no velocity-dispersion excess. This directly weakens the strongest interpretation of the central claim, namely that CCSN sites are unusually dynamical GMCs rather than simply dense regions within high-dispersion galactic structures. The main result 'compared to the average of host galaxies' can survive, but the conclusion about turbulent GMCs and binary progenitors needs to be tempered. Since the appropriate remedy is a matched-control analysis and revised interpretation, the conditional verdict stands; my read does not move the verdict, but it changes the key condition that should be met. Credit is due for including the Appendix B.4 test at all, but its implications need to be made central.","tokens_in":26613,"tokens_out":8724,"duration_ms":98038,"concrete_test":"Recompute the primary sigma comparison with a matched control sample: for each of the 33 CCSNe, draw reference PHANGS pixels from the same host galaxy, the same Querejeta morphological class where available (and the same galactocentric radius plus Sigma bin for the 12 ACOS/archival CCSNe without masks), and calculate the median log sigma excess relative to those matched pixels. If the result is consistent with the +0.002 dex found in Appendix B.4 rather than with the ~0.3 dex radial residual, the abstract and Section 5 should be revised to state that CCSNe occur in high-Sigma, high-dispersion structures without requiring unusually turbulent GMCs. Report the test separately for H-rich SNe and SESNe with bootstrap uncertainties; a robust positive excess in either class after matching would still support the original interpretation for that class.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central comparison in Section 3 (Table 1) contrasts the 33 CCSN positions with all 150,000 CO-detected PHANGS-ALMA pixels. That baseline does not separate local GMC turbulence from large-scale galactic environment. The paper's own Appendix B.4 performs this separation: for the 19 PHANGS-ALMA CCSNe it computes the same Sigma-sigma excess as Section B.3, but relative to pixels inside the same Querejeta et al. (2021) morphological masks. The median log sigma excess is +0.002 dex (16th-84th percentile: -0.142 to +0.308 dex). Thus, at fixed molecular gas surface density and fixed galactic structure, CCSN positions show no excess velocity dispersion. The large Table 1 offset (9.0-9.5 vs 3.81 km/s) is then best explained by CCSNe being located in dense, high-Sigma structures such as centers, bars, and spiral arms, not by their parent GMCs being anomalously turbulent. This result is reported in Section 3 in one sentence but is not carried into the abstract or Section 5 conclusions, where the 'high turbulence' and binary-nature interpretations are featured. The missing short-spacing flux caveat (Section 4.2) affects only a subset of the sample and has an uncertain effect on sigma, but B.4 is an internal matched-control test of the interpretation itself, and it returns null.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper analyzes ALMA CO(2-1) observations toward 33 nearby core-collapse supernovae (19 H-rich, 14 SESNe), measuring molecular gas surface density, velocity dispersion, and virial parameter at ~100 pc scales. It compares these values with 150,000 pixels from 15 PHANGS-ALMA galaxies and reports that CCSNe explode in molecular gas with higher velocity dispersion (median 9.0-9.5 km/s vs 3.81 km/s) and higher surface density than the average of the PHANGS pixels. The authors interpret this as evidence that CCSN progenitors form in unusually turbulent GMCs, possibly linked to high gas density or binary nature. The paper includes extensive appendices checking host galaxy properties, radial residuals, and the Σ-σ excess; Appendix B.4 presents a matched-morphology comparison that yields a median excess of +0.002 dex in log σ, a result not reflected in the abstract or conclusions.","tokens_in":26897,"tokens_out":5264,"duration_ms":56529,"significance":"If the central claim were established, this would be a notable new observational constraint on the environments of CCSN progenitors, connecting GMC-scale turbulence to massive star formation channels. The paper is honest in reporting its caveats and includes several useful control analyses. However, the significance of the result is currently limited by the discrepancy between the unmatched comparison (Table 1, Section 3) and the structure-matched control (Appendix B.4), as well as by statistical and sample-comparison issues. The work has value as a catalog and a first attempt, but the headline conclusion as stated is not yet supported by the full evidence presented.","major_comments":[{"comment":"The main claim that CCSN sites are unusually turbulent is contradicted by the paper's own matched-control test. Appendix B.4 reports a median log σ_CO excess of +0.002 dex (16th-84th percentile -0.142 to +0.308) when comparing the 19 PHANGS CCSNe to pixels within the same Querejeta et al. (2021) morphological mask. This null result is mentioned only in one sentence in Section 3 and is absent from the abstract and Section 5 conclusions, where the 'high turbulence' and binary-nature interpretations are featured. Section 4.1 uses the unmatched B.3 excess to argue that 'even at fixed gas surface density, the velocity dispersion at the explosion sites remains significantly higher,' which is not supported by the structure-matched analysis. The authors must either explicitly reconcile this discrepancy, provide a quantitative argument for why the morphological control is not the appropriate reference, or substantially soften the central claim to say that CCSNe are found in dense, high-dispersion regions that are typical of their host galaxy's structure.","section":"Section 3 / Appendix B.4 / Abstract / Section 5"},{"comment":"For 12 of the 33 CCSNe (the ACOS+archival subsample), the reference distribution consists of PHANGS-ALMA pixels from other galaxies rather than the CCSN's own host galaxy. The abstract claims that CCSNe have high velocity dispersion 'compared to the average of their host galaxies,' but this is literally true only for the 21 PHANGS-ALMA CCSNe. Since velocity dispersion depends on galactic environment, mass, and inclination, the cross-galaxy comparison may bias the result. The authors should either restrict the primary statistical comparison to the 21 PHANGS CCSNe, or present a per-host-galaxy analysis that demonstrates the offset persists for the ACOS/archival sample alone.","section":"Section 2.1 / Table 1 / Section 3"},{"comment":"The KS and AD p-values in Table 2 treat the 150,000 reference pixels as independent draws. Molecular gas is spatially correlated on scales comparable to the beam (~100 pc) and the pixels are clustered within only 15 galaxies; the effective number of independent measurements is far smaller than 150,000. The reported p-values (e.g., 1.57e-5 for SESN σ_CO and 2.50e-7 for SESN α_vir) therefore overstate the significance. The significance of the CCSN-versus-reference differences should be re-evaluated with a scheme that accounts for spatial autocorrelation, for example by bootstrap-resampling spatially independent regions or by using a hierarchical model with galaxy as a random effect.","section":"Section 3 / Table 2"},{"comment":"The caveat about missing short-spacing flux is acknowledged, but the statement that missing flux 'would make our conclusion that CCSN sites have higher Σ_mol and σ_CO than the reference regions even stronger' is not justified for σ_CO. For the 11 CCSNe observed with 12m-only data (and 1 with 7m-only), losing ~50% of the total CO flux could truncate broad line wings or select only compact bright clumps, biasing the moment-2 velocity dispersion in either direction. The paper should quantify the sign and magnitude of this bias, for example by comparing σ_CO measured from 12m-only versus 12m+7m+TP maps in the same PHANGS fields, or at minimum rephrase the claim to apply only to Σ_mol and state that the effect on σ_CO is uncertain.","section":"Section 4.2"}],"minor_comments":[{"comment":"In the definition of the virial parameter in Eq. (1), the symbol θ is used both for the beam size and later in Eq. (2) for a fixed 100 pc scale; the text should clarify whether θ is the physical beam FWHM or an assumed cloud size, and how the different θ values across the sample affect the comparison.","section":"Section 2.2"},{"comment":"The phrase 'supports their increased formation in regions of high densities and/or their binary nature' is stronger than the evidence: the observations characterize the current environment, not the birth conditions, and the binary interpretation is one of several plausible scenarios. Softer wording such as 'is consistent with' would be more accurate.","section":"Abstract"},{"comment":"The last paragraph of Section B.4 says the general trends show CCSN molecular gas environments 'share similar properties of the galactic structures,' which is the key null result, but this statement is buried in an appendix; consider moving it to the main text or at least summarizing it in Section 3 with the same emphasis as the unmatched result.","section":"Appendix B.4"},{"comment":"The caption lists 'typical error bars are denoted as gray, blue, and red' but does not define which symbol corresponds to which color; please clarify the assignment of colors to H-rich SNe, SESNe, and PHANGS-ALMA pixels.","section":"Figure 1"},{"comment":"The 1σ confidence intervals for the medians are reported from 10^4 Monte Carlo simulations, but the underlying bootstrap or error-propagation method is not described; a one-sentence explanation in the text or caption would aid reproducibility.","section":"Table 1"}],"recommendation":"major_revision","confidential_remarks":"The paper is a useful observational contribution, but the central claim as currently framed is not supported by the paper's own matched-morphology analysis in Appendix B.4. The authors should be encouraged to reframe the main result as 'CCSNe reside in dense, high-dispersion molecular gas typical of their galactic structures,' unless they can convincingly argue why the structure-matched control is invalid. The statistical issues with the reference sample also need attention before publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe headline claim is undercut by the paper's own Appendix B.4. What is actually new and useful: 33 CCSNe with cloud-scale CO(2-1) velocity dispersions and virial parameters, the first sample of its kind. The raw comparison shows a big offset (median ~9-9.5 km/s vs 3.81 for PHANGS pixels), and the paper does a lot of work to show it isn't a distance or host-galaxy artifact.\n\nBut when the 19 PHANGS-ALMA CCSNe are compared to pixels inside the same morphological mask (center, bar, disc, interarm, spiral arm), the median sigma excess at fixed Sigma is +0.002 dex, with a 16-84% range that includes negative values. In other words, CCSNe are in dense, high-Sigma structures such as centers and spiral arms, and once you match for structure, they are not in unusually turbulent GMCs. That is a direct internal control, reported in one sentence in Section 3 and never mentioned again, while the abstract and Section 5 talk about high turbulence and binary progenitors. That is the soft spot, and it is load-bearing for the binary-nature interpretation.\n\nOther issues are secondary. Twelve of the 33 CCSNe are compared to PHANGS pixels rather than their own hosts; the host-galaxy property checks are reassuring but not a substitute for a matched comparison. The p-values treat 150,000 pixels as independent, so they overstate significance. The 12m-only data for a third of the sample may miss extended flux, though the paper's argument that this would not reverse the raw offset is plausible.\n\nWhat is genuinely good: the sample construction, the residual checks against galactocentric radius, and the consistent finding that CCSNe live in high-density molecular gas, matching earlier work. The citation pattern looks fine and the prior molecular gas studies get proper credit. The paper is honest enough to put B.4 in the appendix, but it does not follow where that result leads.\n\nWho is this for? SN environment people and molecular cloud observers will want the measurements and the sample table. The interpretation section needs a rewrite: the data support 'CCSNe trace dense galactic structures,' not 'CCSNe trace unusually turbulent GMCs.' I would send it to a referee, but the referee should ask for the conclusions to be reframed around B.4 and the structure-matched null.","headline":"The raw signal is real, but the paper's own structure-matched control nullifies the 'unusually turbulent GMC' interpretation.","tokens_in":27504,"tokens_out":4104,"would_cite":true,"duration_ms":41940,"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":"Core-collapse supernovae explode inside molecular gas that is roughly two and a half times as turbulent as the average of their host galaxies.","keywords":["core-collapse supernovae","molecular gas turbulence","CO(2-1) velocity dispersion","giant molecular clouds","virial parameter","interferometric observations","galaxies: ISM","supernovae: general"],"falsifier":"Re-observe the eleven supernova fields currently covered only by 12-metre-array data with full 12m+7m+total-power mosaics and recompute the median CO(2-1) velocity dispersions; if the median drops toward the host-galaxy value of 3.81 km/s once the missing short-spacing flux is restored, the central claim would be falsified for a third of the sample.","tokens_in":26443,"feed_emoji":"💥","tokens_out":15192,"duration_ms":138992,"temperature":0.7,"pith_summary":"Core-collapse supernovae are the deaths of massive stars, and the gas they die inside should carry traces of the clouds their progenitors were born from. This paper measures, for the first time, the turbulence of that molecular gas at the positions of 33 nearby supernovae, using CO(2-1) observations at roughly 100-parsec resolution. It finds that supernova positions sit in molecular gas with median velocity dispersions of about 9.0-9.5 km/s, more than double the 3.81 km/s typical of host-galaxy pixels, and that the excess survives when distances from galaxy centres and host-galaxy structures are taken into account. The authors argue that such turbulent birth environments favour dense-gas formation and binary or multiple stellar systems, which would explain why so many supernova progenitors show signs of binary interaction. If correct, supernova positions become a direct tracer of the most dynamically active giant molecular clouds in nearby galaxies.","feed_headline":"Supernovae explode in gas twice as turbulent as their host galaxies","feed_subtitle":"At 33 explosion sites, molecular gas motions run 9.0-9.5 km/s versus 3.8 km/s galaxy-wide.","key_machinery":"The analysis rests on the CO(2-1) rotational line of carbon monoxide as a tracer of molecular hydrogen, observed at resolutions of roughly 40-230 pc so that individual giant molecular clouds are approached. The turbulence proxy is the velocity dispersion $\\sigma_{\\rm CO(2-1)}$, the second moment of the CO line profile, measured at each position; the dynamical state is quantified by the virial parameter $\\alpha_{\\rm vir} = 10\\,\\sigma^2/(\\pi G\\,\\Sigma_{\\rm mol}\\,\\theta)$, which separates gravitationally bound clouds ($\\alpha_{\\rm vir} < 2$) from supervirial, dispersing ones ($\\alpha_{\\rm vir} > 2$). The comparison against a large resolved survey of host-galaxy pixels, together with the scaling relation between velocity dispersion, cloud size, and surface density, lets the paper separate a pure density effect from an excess in turbulence at fixed surface density.","core_discovery":"The central claim is that core-collapse supernovae do not simply occur in dense molecular gas; they occur in molecular gas that is unusually turbulent. From CO(2-1) line observations of 33 supernova positions, 19 hydrogen-rich and 14 stripped-envelope (hydrogen-poor) events, the paper measures median velocity dispersions of $9.0^{+2.0}_{-0.6}$ km/s for hydrogen-rich and $9.5^{+0.3}_{-0.3}$ km/s for stripped-envelope supernovae, compared with $3.81^{+0.01}_{-0.01}$ km/s for the resolved reference pixels of the host galaxies. The corresponding virial parameters are also elevated, $9.5^{+1.3}_{-2.4}$ and $16.1^{+3.3}_{-2.3}$ versus $3.56^{+0.01}_{-0.01}$, indicating that the clouds at explosion sites are not just denser but more strongly dominated by non-thermal, turbulent motion than typical giant molecular clouds. The excess remains after restricting the comparison to similar galactocentric radii and to the same galactic structures, and the authors show that the supernovae themselves could not have produced the measured line widths: even the oldest event would have a shock radius of less than a parsec, far below the 100-pc beam.","pith_inferences":["One extension the authors do not pursue: if the offset is real, supernova positions could serve as a sparse but unbiased tracer of the most turbulent 100-pc-scale gas across many galaxies, complementing full mapping surveys at a fraction of the observing time.","The turbulence-binarity interpretation could be tested by combining these line widths with pre-explosion imaging statistics: it predicts a higher companion-detection rate, or more merger remnants, among progenitors born in the highest-$\\sigma_{\\rm CO(2-1)}$ clouds.","The cleanest confirmation would come from re-observing the 11 twelve-metre-only positions with short-spacing coverage; if the offset shrinks, the result would apply only to the fully mosaicked positions.","The same method applied to other transients, such as gamma-ray bursts or superluminous supernovae, would show whether the turbulence excess is specific to core-collapse progenitors or shared by all massive-star explosions."],"forward_implications":["Supernova explosion sites trace the high-turbulence tail of the giant-molecular-cloud population, not just the dense tail, so models of massive-star formation must reproduce progenitors emerging preferentially from gas with $\\sigma_{\\rm CO(2-1)} \\sim 9$ km/s.","Because the excess persists at fixed gas surface density, turbulence itself, not merely density, correlates with massive-star formation, supporting turbulent-fragmentation and feedback-triggered star-formation scenarios.","The elevated virial parameters at supernova positions imply that many parent clouds are supervirial, so the progenitors formed in regions that were dispersing or externally driven rather than quietly collapsing.","Hydrogen-rich and stripped-envelope supernovae show similar velocity dispersions, suggesting a common turbulent birth environment, while the slightly higher virial parameters of stripped-envelope events hint at even more extreme clouds (a difference the paper notes is not statistically significant).","Because the supernova shocks themselves contribute negligibly to the measured line widths, the turbulence reflects the birth environment and can be used to constrain progenitor formation without subtracting the explosions' own feedback."],"supporting_citations":[{"why":"Supplies the resolved CO(2-1) maps and reduction pipeline for the reference galaxy-pixel population used as the comparison baseline.","marker":"Leroy et al. 2021a"},{"why":"Paper I of this survey; established the ACOS sample and the earlier molecular-gas surface-density result at CCSN positions.","marker":"Solar et al. 2024"},{"why":"Documents that a single ALMA array configuration can miss roughly 50% of the total CO flux, the main systematic caveat for 12 of the 33 positions.","marker":"Pety et al. 2013"},{"why":"Provides the virial parameter definition used to classify clouds as bound, marginally bound, or dispersing.","marker":"Bertoldi & McKee 1992"},{"why":"Gives the scaling relation between velocity dispersion, cloud size, and surface density used to interpret the observed offsets.","marker":"Heyer et al. 2009"},{"why":"Supplies the CO-to-H2 conversion factor adopted to compute molecular gas surface densities.","marker":"Bolatto et al. 2013"},{"why":"Provides the CO(2-1)-to-CO(1-0) ratio R21 = 0.65 used in the surface-density calculation.","marker":"den Brok et al. 2021"},{"why":"Population synthesis showing a substantial binary or merger history among core-collapse progenitors, supporting the binary interpretation of the turbulence excess.","marker":"Zapartas et al. 2019"}],"fun_headline_variants":["Supernovae prefer gas 2.4× more turbulent than galactic norm","33 supernova sites: molecular clouds are 2.4× more turbulent","Core-collapse SNe birthplaces: gas turbulence double galactic average","ALMA finds supernovae in gas twice as churning as their galaxies","Supernovae explode in turbulent gas—not just dense—ALMA shows"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing assumption is that the velocity dispersions measured at the supernova positions are directly comparable to the reference galaxy pixels: a third of the sample was observed with a single 12-metre array configuration, and if those observations miss about half the CO flux, truncating line wings or selecting only bright compact clumps, the reported turbulence excess could be biased upwards.","fun_headline_variants_meta":{"raw":{"variants":["Supernovae prefer gas 2.4× more turbulent than galactic norm","33 supernova sites: molecular clouds are 2.4× more turbulent","Core-collapse SNe birthplaces: gas turbulence double galactic average","ALMA finds supernovae in gas twice as churning as their galaxies","Supernovae explode in turbulent gas—not just dense—ALMA shows"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001224,"raw_usage":{"total_tokens":5089,"prompt_tokens":1059,"completion_tokens":4030,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":675,"completion_tokens_details":{"reasoning_tokens":3931}},"tokens_in":675,"tokens_out":4030,"duration_ms":31785,"temperature":1.0,"reasoning_tokens":3931,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T04:37:47.761004+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Re-observe the eleven supernova fields currently covered only by 12-metre-array data with full 12m+7m+total-power mosaics and recompute the median CO(2-1) velocity dispersions; if the median drops toward the host-galaxy value of 3.81 km/s once the missing short-spacing flux is restored, the central claim would be falsified for a third of the sample.","supporting_citations":[],"review_version":1}