{"id":"83e0aae5-0d5c-41fb-a0f7-a055f69518dc","arxiv_id":"2508.12933","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"A systematic grid of theoretical upper limits on dust masses from core-collapse supernovae shows oxygen-rich silicates dominate and that predicted dust yields vary by factors of 2 to 5 depending on the stellar evolution model used.","lead":"Researchers calculated the maximum possible dust a supernova can make for stars from 9 to 120 times the Sun's mass, using published stellar models. The result is a mass-by-mass map of theoretical dust ceilings that can be compared with telescope observations of supernova remnants.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The stochasticity claim rests on an unresolved numerical-versus-physical question that the paper itself flags; if the KEPLER Si/Mg scatter is a numerical artifact, the central dust-stochasticity result is unsupported.","rationale":"The paper is a transparent, reproducible upper-limit calculation: dust masses are derived from published KEPLER yields with a simple element-locking prescription. Its strongest and most novel claim is the connection between yield stochasticity, shell merging, compactness, and dust mass. That claim is only as strong as the physical reality of the Si/Mg scatter in the KEPLER grid. The authors themselves flag the numerical-versus-physical debate in Section 5.1 and acknowledge in Section 7 that it remains open, so the reader's conditional verdict is appropriate. I considered whether the pre- versus post-explosion abundance mapping is a more load-bearing weakness, but the paper provides explicit validation for two masses, while the numerical-origin question is directly admitted as unresolved and has no comparable control. A convergence and sensitivity test on the specific 19.5/20.5 Msun spike pair would settle whether the headline stochasticity is robust. Since the reader already identified this same weakest assumption and assigned a conditional verdict, no verdict change is needed.","tokens_in":14392,"tokens_out":6405,"duration_ms":73304,"concrete_test":"Select the 19.5 and 20.5 Msun KEPLER models where Si in the O/Si/Mg zone differs by about 80x, and rerun each with (i) doubled spatial resolution, (ii) halved timestep, and (iii) a modest change in the convective-boundary mixing efficiency. Compute the total Si and Mg masses in the O/Si/Mg zone in each rerun. If the spike amplitude or its location in initial mass changes by more than a factor of about 2 under these perturbations, the apparent stochasticity is numerical rather than physical, and the compactness correlation in Figure 3 should not be interpreted as evidence about real progenitor dust yields.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central stochasticity claim in the abstract and Section 5.1 depends entirely on the assumption that the order-of-magnitude variations in Si and Mg masses in the O/Si/Mg zone across the Sukhbold et al. (2016) KEPLER grid are real astrophysical features of shell merging, not numerical artifacts of 1D stellar evolution. The paper itself concedes the unresolved status of this question: Section 5.1 states that there is debate about whether the stochasticity arises from numerical effects rather than physical processes, and Section 7 ends with the need for more studies to understand whether such stochastic yields are represented in post-explosion abundances. If the approximately 80-fold Si difference between the 19.5 and 20.5 Msun models (about 5e-3 versus 0.4 Msun in the O/Si/Mg zone) is produced by convective-boundary mixing treatment, mesh resolution, or timestep sensitivity, then the correlated spikes in silicate dust masses (Figure 2), the anti-correlation with compactness (Figure 3), and the inference that less-compact, more-explodable progenitors produce more dust are not predictions about real CCSNe. The MESA comparison does not resolve this issue because MESA is also a 1D code with its own numerical mixing scheme. The upper-limit arithmetic would survive, but the headline stochasticity and its proposed role as a probe of stellar evolutionary channels would not.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper computes theoretical upper limits on dust masses in core-collapse supernovae from pre-explosion elemental yields of the Sukhbold et al. (2016) KEPLER grid for progenitors of 9-120 solar masses. The authors use a stratified-zone dust formation model with CO molecule formation followed by condensation of silicates, alumina, and amorphous carbon, limiting dust mass by the least abundant constituent element in each zone. They report that O-rich silicate dust dominates, increasing with progenitor mass from about 0.025 to 0.9 solar masses, and that C-rich dust remains below about 0.05 solar masses. They provide best-fit functions for O-rich dust, C-rich dust, and CO mass up to 30 solar masses. A large scatter in silicate dust masses is found and attributed to shell merging and convective boundary mixing stochasticity in the KEPLER models, with an inferred anti-correlation with compactness and explodability. A comparison with MESA (Laplace et al. 2021) yields dust masses 2-5 times larger for the same progenitor masses.","tokens_in":14587,"tokens_out":4081,"duration_ms":41695,"significance":"If the results are robust, the upper-limit framework is a simple and useful mapping from published stellar yields to observable dust masses, and it provides a concrete theoretical target for interpreting mid-IR dust observations of CCSNe. The MESA/KEPLER comparison usefully quantifies model-dependent uncertainty in SN dust predictions. However, the headline stochasticity claim and the proposed role of dust mass as a probe of stellar evolutionary channels hinge on whether the order-of-magnitude variations in Si and Mg yields across the KEPLER grid are physical shell-merger events rather than numerical artifacts of 1D stellar evolution. The paper itself acknowledges this debate but does not resolve it, so the central interpretive claim remains conditional.","major_comments":[{"comment":"The central claim that 'a large stochastic variation is found in the predicted masses of silicate dust, which correlates with the randomness of shell-merger events' rests on interpreting the ~80-fold difference in Si mass between the 19.5 and 20.5 Msun KEPLER models (Figure 3 top) as an astrophysical feature. The manuscript itself states in Section 5.1 that 'There is debate if such stochasticity arises due to numerical effects, rather than being physical' and cites 3D simulations as suggestive but not conclusive. Because this scatter underlies the abstract's stochasticity claim and the proposed use of dust mass as a probe of stellar evolutionary channels, the authors need to provide convergence tests (e.g., resolution, timestep, or convective-boundary-mixing sensitivity for the 19-21 Msun region) or explicitly reframe the claim as conditional on the physical interpretation of the KEPLER models. Without this, the headline result is not robust to the numerical-artifact alternative.","section":"Abstract and Section 5.1"},{"comment":"Pre-explosion yields are used for essentially all 200 progenitors, but the validation that pre- and post-explosion alpha-element abundances are comparable is shown only for two cases, 15 and 20 Msun (Figure 1 middle panels). Since the dust upper limits are element-limited by O, Si, and Mg masses in the O/Si/Mg zone, and since the most dramatic dust-mass spikes occur at 19-26 Msun, the authors should show post-explosion comparisons for additional masses in that range or quantify the expected explosive nucleosynthesis corrections. Without broader validation, the absolute upper limits (e.g., 0.9 Msun at 25.5 Msun) are not fully supported.","section":"Section 3 and Figure 1"},{"comment":"The claimed anti-correlation between O-rich dust mass and compactness is not statistically quantified. The authors note that 'a straight forward correlation between the two is not visible' in the full sample and rely on a zoomed-in 19.5-21 Msun window. A correlation coefficient, significance estimate, or a larger sample is needed to support the inference that less-compact, more-explodable progenitors produce more dust. As written, the conclusion is overinterpreted from a small subset of the grid.","section":"Section 5.1 and Figure 3"}],"minor_comments":[{"comment":"The text says 'We further compare the Si and Mg abundances, shown in Figure 2 (right)' but the relevant panel appears to be Figure 5 (right); please correct the cross-reference.","section":"Section 6"},{"comment":"The reference 'Shahbandeh, M., Sarangi, A., Temim, T., et al. 2023, MNRAS, 523, 6048' is listed twice; only one entry is needed.","section":"Reference list"},{"comment":"The caption begins with 'T op' (and later 'Y op'); these should be 'Top' and 'Top' respectively.","section":"Figure 1 caption"},{"comment":"The base of the logarithm in Eq. (5) is not specified; please state whether it is natural log or base 10.","section":"Section 5.2"},{"comment":"The statement 'All the zones are efficient in forming silicate dust' is followed by 'the majority of silicates are formed in the O/Si/Mg zone'; please clarify whether all zones contribute significantly or the O/Si/Mg zone dominates, since the upper-limit calculation appears to aggregate contributions.","section":"Section 4"},{"comment":"The abstract states dust masses range from 0.02 to 0.9 Msun, while Section 5 states 0.025 to 0.9 Msun; please make the numbers consistent.","section":"Abstract and Section 5"}],"recommendation":"major_revision","confidential_remarks":"The numerical-versus-physical interpretation of the Sukhbold et al. (2016) yield scatter is a known open question in the stellar evolution community. If the authors do not address resolution/convergence or at least sharply hedge the stochasticity claim, the paper is likely to face strong resistance from specialists. Adding even a modest convergence test or an explicit acknowledgment that the scatter could be partly numerical, with a revised abstract, would substantially strengthen the paper."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Read the paper. The useful thing is the grid: 200 KEPLER progenitors from 9 to 120 solar masses, translated through prior dust chemistry models into upper limits on silicate, alumina, amorphous carbon, and CO masses, with fitting functions up to 30 solar masses and a MESA comparison. That product is new and will be handy for galaxy chemical evolution models and for JWST-era comparisons with supernova remnants. The upper-limit arithmetic is simple, reproducible from published yields, and honestly framed: this is an upper limit, not a kinetic prediction. For that purpose, the central numbers are defensible.\n\nCredit where it is earned: the paper does not oversell kinetics. It states clearly that explosion properties affect timescale rather than final mass, that dust mass is limited by the least-abundant constituent after CO formation, and it flags the unresolved debate about whether KEPLER's yield stochasticity is physical or numerical. The MESA comparison quantifies model dependence at a factor of 2–5, which is a genuine and visible uncertainty.\n\nSoft spots are real but mostly attached to the headline claim, not the grid. The \"stochasticity in silicate dust correlates with shell merging and explodability\" story rests entirely on the assumption that the order-of-magnitude Si/Mg scatter in KEPLER is astrophysical. The paper itself says in Section 5.1 there is debate about numerical artifacts, and it cites 3D simulations as suggestive rather than conclusive. MESA is also a 1D code, so the sensitivity comparison does not settle it. If the scatter is numerical, the spikes in Figure 2, the anti-correlation with compactness, and the \"less compact, more explodable, more dust\" inference do not carry over to real supernovae. The upper limits would survive; the stochasticity claim would not.\n\nTwo lesser issues: pre-explosion abundances are used for most progenitors with only a couple of validation cases, and the aluminium abundance is scaled from a single 15-solar-mass post-explosion model. For an upper-limit study these are acceptable, but they should be stated as such. Also, silicate mass is by construction proportional to Si/Mg abundance in the zone, so \"dust stochasticity mirrors yield stochasticity\" is partly a restatement rather than an independent finding. The paper does not hide this, but the abstract implies more than that.\n\nWho is it for: anyone using CCSN yields in dust or chemical evolution models, and observers comparing remnant dust masses to progenitor mass. It deserves a serious referee. I would send it out with a request to soften the stochasticity language or add an explicit caveat that the interpretation is conditional on the numerical-versus-physical question. My own verdict would be conditional accept.","headline":"A useful mass-resolved map of CCSN dust upper limits, but the headline stochasticity claim depends on an unresolved numerical-versus-physical question the authors themselves flag.","tokens_in":15165,"tokens_out":1775,"would_cite":true,"duration_ms":18151,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper claims that pre-explosion nucleosynthesis, not the explosion itself, sets the dust-mass ceiling in core-collapse supernovae, with silicate dust dominating at 0.02–0.9 solar masses and stochastic shell-merging producing large…","keywords":["core-collapse supernovae","supernova dust","silicate dust","amorphous carbon","stellar yields","shell merging","compactness parameter","dust formation chemistry"],"falsifier":"A concrete check: for progenitors near 20 $M_\\odot$, the model predicts an anti-correlation between O-rich dust mass and compactness—the 20.5 $M_\\odot$ model, with strong shell merging, should yield several times more silicate than the 19.5 $M_\\odot$ model. Late-time JWST mid-infrared dust mass measurements for supernovae whose pre-explosion imaging pins the progenitor into that narrow mass range would confirm or break the correlation; a second falsifier is finding any CCSN remnant with a well-determined progenitor that hosts more than 0.9 $M_\\odot$ of newly formed dust, which would exceed the yield-limited upper bound.","tokens_in":14139,"feed_emoji":"🌌","tokens_out":10689,"duration_ms":96281,"temperature":0.7,"pith_summary":"This paper aims to establish that the maximum dust mass a core-collapse supernova can form is fixed by the star's pre-explosion yield composition, not by the explosion itself. The authors feed the onion-shell abundances of a 200-model stellar grid spanning 9–120 $M_\\odot$ into a published dust-formation chemistry model, finding that O-rich silicate dust dominates with upper limits of 0.02–0.9 $M_\\odot$ that grow with progenitor mass, while amorphous carbon never exceeds 0.05 $M_\\odot$. A central result is stochasticity: between neighbouring progenitor masses, predicted silicate masses jump by up to an order of magnitude because shell-merging (convective boundary mixing) enriches the O/Si/Mg zone in silicon. Since less compact progenitors are both more likely to explode and more likely to have undergone such mixing, the paper concludes that the supernovae we observe should preferentially be the dust-rich ones. If correct, this turns supernova dust mass into a diagnostic of stellar evolutionary channels and explains why many remnants host dust masses near or above 0.5 $M_\\odot$.","feed_headline":"Random shell mergers set supernova dust mass, models show","feed_subtitle":"Silicate dust caps range from 0.02 to 0.9 solar masses; one progenitor's dust mass shifts 2-5x between stellar codes.","key_machinery":"The carrying object is a stratified-zone dust budget: the ejecta is divided into unmixed Si/S, O/Si/Mg, He/C, and H zones; CO forms first and sequesters C and O in equal numbers; silicate mass is then limited by the least abundant of Mg, Si, or O in the stoichiometry of [Mg$_2$SiO$_4$]$_n$, and amorphous carbon mass by the carbon left over in the He/C zone after CO. This converts a nucleosynthesis grid into dust upper limits. The second mechanism is the compactness parameter $\\xi_{2.5}$, defined as $2.5/R(2.5\\,M_\\odot)$ with the radius where infall velocity exceeds 1000 km/s, used as a tracer of shell merging and explosion likelihood to connect dust mass to evolutionary channel.","core_discovery":"The paper's central discovery is that a zone-by-zone abundance budget converts published stellar yields into theoretical dust upper limits: after CO molecules lock up carbon and oxygen in equal numbers, each zone's dust mass is capped by its least abundant dust constituent, giving [Mg$_2$SiO$_4$]$_n$ silicate masses of 0.02–0.9 $M_\\odot$ that rise with initial mass, and amorphous carbon masses of 0.012–0.043 $M_\\odot$ concentrated below about 15 $M_\\odot$. The large fluctuations in silicate mass trace fluctuations in silicon and magnesium in the O/Si/Mg zone, which the paper attributes to shell-merging events before explosion; in the 19.5–21 $M_\\odot$ range, O-rich dust mass anti-correlates with the compactness parameter, tying high dust yield to low compactness and hence to explodability. Comparing yields from two stellar-evolution codes for the same progenitor gives dust masses differing by factors of 2–5, so the paper concludes that final dust yield is governed by stochastic stellar yields and pre-explosion nucleosynthesis, while explosion properties set only the timescales of dust formation.","pith_inferences":["The authors do not pursue it, but the shell-merging scatter in one-dimensional models may be partly numerical; the comparison with a second stellar-evolution code suggests the broader trend of higher and fluctuating O-rich dust masses persists across codes, so the qualitative conclusion may survive even if individual spikes do not.","A testable extension: pair late-time JWST dust-mass measurements with pre-explosion progenitor imaging for supernovae near 20 $M_\\odot$; the predicted anti-correlation between dust mass and compactness can be checked directly once explosion outcomes are known.","The unmixed stratified-zone assumption sets an upper limit; the paper's qualitative claim that mixing would cut amorphous carbon while leaving silicates roughly unchanged could be quantified with three-dimensional mixing prescriptions, potentially shifting the 0.9 $M_\\odot$ cap.","The wide gap between the upper and lower power-law bounds for O-rich dust implies that dust mass alone cannot pin down a progenitor's mass; combining dust with an independent compactness or explodability indicator would give much sharper constraints."],"forward_implications":["Any core-collapse supernova with a progenitor between 9 and 120 $M_\\odot$ forms at most 0.025–0.9 $M_\\odot$ of dust in the ejecta, so reported dust masses above 0.9 $M_\\odot$ from a single CCSN would require dust that is not newly formed in the ejecta.","For progenitors up to 30 $M_\\odot$, the analytic fits (a power law with upper and lower bounds for O-rich dust, a broken quadratic for C-rich dust, and a logarithmic relation for CO) let observers translate an inferred progenitor mass into an expected dust yield and composition.","Progenitors that experienced shell-merging should be less compact, more likely to explode, and more dust-rich, which predicts that observed SN remnants are biased toward the high end of the dust-mass distribution.","Dust-mass estimates for a single progenitor carry a factor-of-2-to-5 model uncertainty from stellar evolution codes, so progenitor masses inferred from infrared dust observations are at least that uncertain.","Explosion energy, $^{56}$Ni mass, and clumpiness set how fast dust forms but not how much; therefore late-time dust mass, not early-time dust, is the observable tied to stellar yields."],"supporting_citations":[{"why":"Supplies the 200-model grid of pre-supernova abundances and zone masses that every dust-mass estimate here is built from.","marker":"T. Sukhbold et al. (2016)"},{"why":"Supplies the supernova ejecta dust-formation model (SiO/CO chemistry, nucleation, condensation) whose zone-by-zone mass-limiting scheme is applied.","marker":"A. Sarangi et al. (2018)"},{"why":"Establishes C-rich dust formation in the He/C zone and the role of CO as a carbon sink, providing the basis for amorphous carbon upper limits.","marker":"A. Sarangi & I. Cherchneff (2013)"},{"why":"Provides the second set of pre-supernova models used to show that dust masses vary by a factor of 2–5 for the same progenitor mass.","marker":"E. Laplace et al. (2021)"},{"why":"Links convective boundary mixing and shell merging to lower compactness and higher explodability, the correlation used to interpret silicate dust spikes.","marker":"A. Davis et al. (2019)"},{"why":"Defines the compactness parameter $\\xi_{2.5}$ used as a tracer of shell merging and explosion likelihood.","marker":"E. O'Connor & C. D. Ott (2011)"},{"why":"Cited for differences in reaction networks and energy transport between stellar-evolution codes, supporting the model-sensitivity conclusion.","marker":"T. Sukhbold & S. E. Woosley (2014)"}],"fun_headline_variants":["Supernova dust yields depend on random shell mergers","Dust mass in supernovae varies 2-5x between codes","Silicate dust caps set by stochastic stellar yields","Stochastic pre-explosion yields control dust mass","Shell merger randomness drives supernova dust spread"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the run-to-run jumps in silicon and magnesium abundances across the stellar models are physical shell-merging events rather than numerical noise in one-dimensional evolution codes; if the jumps are artifacts, the predicted dust-mass scatter is an artifact too.","fun_headline_variants_meta":{"raw":{"variants":["Supernova dust yields depend on random shell mergers","Dust mass in supernovae varies 2-5x between codes","Silicate dust caps set by stochastic stellar yields","Stochastic pre-explosion yields control dust mass","Shell merger randomness drives supernova dust spread"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000215,"raw_usage":{"total_tokens":1497,"prompt_tokens":1079,"completion_tokens":418,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":695,"completion_tokens_details":{"reasoning_tokens":342}},"tokens_in":695,"tokens_out":418,"duration_ms":4680,"temperature":1.0,"reasoning_tokens":342,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T17:17:42.937916+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A concrete check: for progenitors near 20 $M_\\odot$, the model predicts an anti-correlation between O-rich dust mass and compactness—the 20.5 $M_\\odot$ model, with strong shell merging, should yield several times more silicate than the 19.5 $M_\\odot$ model. Late-time JWST mid-infrared dust mass measurements for supernovae whose pre-explosion imaging pins the progenitor into that narrow mass range would confirm or break the correlation; a second falsifier is finding any CCSN remnant with a well-determined progenitor that hosts more than 0.9 $M_\\odot$ of newly formed dust, which would exceed the yield-limited upper bound.","supporting_citations":[],"review_version":1}