{"id":"b6efabe0-3807-4435-ae3f-71ea2126878c","arxiv_id":"2501.04098","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Most novae (50-70%) form dust, and a redward shift in V-K color of more than 2.35 magnitudes from peak is a useful indicator of dust formation.","lead":"This paper analyzes optical and infrared light curves of 40 novae and finds that 50-70% form dust, much more than earlier estimates suggested. It also presents a simple color-based indicator of dust formation that upcoming infrared sky surveys can use.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 50-70% dust fraction may be inflated because the SMARTS sample is not representative: Section 4.2 concedes that fast, heavily extinguished thin-disk novae are missed by optical surveys, and the speed-class comparison with Craig et al.","rationale":"The central claim, that 50-70% of novae form dust, would be true only if the 40 SMARTS novae are representative of the broader nova population. Among the candidate weaknesses, this is the most load-bearing because it is external: no re-analysis of the provided light curves can determine whether a systematically missed population exists. The paper's internal checks (the speed-class histogram versus Craig et al. and the ASAS-SN survey arguments in Section 4.2) are reasonable but both compare against optically selected samples, which share the same blind spot for heavily extinguished thin-disk novae. The authors themselves flag this as the main caveat. The manual classification is a secondary concern because the authors show all light curves in Appendix A and use an 'Unsure' category, making classifications auditable; misclassification would likely shift the fraction within the already stated 53-70% bracketing. The V-K threshold is a diagnostic tool, not the basis for the dust fraction, so it is not load-bearing for the headline claim. The gamma-ray correlation and t2-artifact concerns affect secondary conclusions. Therefore, sample representativeness is the single concern that could meaningfully move the headline number. The paper is careful, presents public data, and explicitly acknowledges the caveat, so the appropriate verdict remains CONDITIONAL; no adjustment is needed.","tokens_in":66774,"tokens_out":7543,"duration_ms":74600,"concrete_test":"Use the near-IR selected samples of De et al. (2021) and Zuckerman et al. (2023) to estimate the fraction of Galactic novae missed by optical surveys and their t2 distribution, then combine with the paper's t2-dependent dust rates (Figure 11) to compute a selection-corrected dust fraction. If the corrected fraction falls below 50%, the headline claim is biased. Alternatively, obtain V and K light curves for a Gattini-IR or WINTER selected sample and classify dust with the same protocol; an IR-selected dust fraction below 50% would directly falsify the claim.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's headline statistic, that 50-70% of novae form dust, rests on the assumption that the 40 SMARTS novae represent the parent Galactic population. Section 2.1 states that the Atlas deliberately targets He/N, recurrent, and LMC novae, and 53 of 93 Atlas novae were dropped for data gaps. More importantly, Section 4.2 concedes that heavily extinguished thin-disk novae, potentially fast and non-dust-forming, are systematically missed by optical surveys. The authors attempt to dismiss a speed-class bias by comparing with the Craig et al. sample, but that comparison sample is itself optically selected, so it cannot validate representativeness. Since the paper's own Figure 11 shows only about 17% of very fast novae (t2 < 10 days) form dust, a hidden population of fast, extinguished novae would lower the 53% lower bound, potentially below 50% and invalidating the 'majority' conclusion. This external selection effect cannot be fixed by any internal analysis of the 40 light curves.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper uses 40 well-sampled SMARTS novae with optical and near-IR light curves to estimate the fraction of novae that form dust. The authors manually classify each nova as Dust Dip, IR Excess, None, or Unsure, propose a largest-redward-(V-K)-color-change diagnostic with a threshold of 2.35 mag, and find that 50-70% of the sample form dust. They further report that very fast novae (t2<10 days) rarely form dust, that Fermi-LAT-detected novae are more likely to be dust-forming, and that the long-noted t2-versus-dust-onset correlation is largely an artifact of dust formation itself causing the two-magnitude decline.","tokens_in":66978,"tokens_out":5592,"duration_ms":58879,"significance":"If the 50-70% fraction applies to the broader Galactic nova population, it is a substantial upward revision from earlier estimates near 20% and would establish dust formation as the norm rather than the exception, with direct implications for shock-mediated dust formation models. The paper's strengths include a public dataset, a transparent treatment of the 'Unsure' category, and appropriate use of binomial confidence intervals. However, the population-level claim currently rests on sample representativeness and on subjective manual classifications, so the headline statistic should be treated as provisional until these are addressed.","major_comments":[{"comment":"The central claim that 50-70% of novae form dust assumes that the 40 selected SMARTS novae represent the broader discovered, or ideally intrinsic, Galactic nova population. The comparison with the Craig et al. sample cannot validate this assumption because that comparison sample is itself optically selected. The manuscript itself concedes that heavily extinguished thin-disk novae, potentially fast and non-dust-forming, are systematically missed by optical surveys. Since Figure 11 shows only about 17% of very fast novae form dust, a plausible hidden population of fast, extinguished novae would push the fraction below 50%. Please either provide a quantitative selection-bias analysis using independent constraints (e.g., IR surveys) or restrict the headline claim and the abstract wording to the studied sample and to the currently discovered population.","section":"Section 4.2 and Figure 11"},{"comment":"The dust classifications that drive the 53-70% estimate are assigned manually from inspection, with qualitative category definitions and no inter-rater reliability or blinding described. Some classifications also depend on external literature rather than the uniform V/K criteria: RS Oph is placed in 'None' from IR spectroscopy arguments, while V745 Sco is placed in 'IR Excess' primarily from CO detections. Because the lower bound is exactly 21/40, a small number of label changes would alter the headline fraction. Please report a sensitivity analysis (e.g., moving V745 Sco to 'None', or RS Oph to a dust-forming class) and, ideally, an independent blind reclassification by multiple raters.","section":"Section 2.2 and Table 1"},{"comment":"The 2.35 mag (V-K) threshold is presented as a diagnostic that 'does a good job separating' dust-forming from non-dust-forming novae, but the threshold is drawn through the same points that were manually classified from the same light curves. It is therefore an in-sample descriptive separator, not an independent validation of the classification. The paper should state this explicitly, and the diagnostic should be tested on an independent sample or applied with pre-registered criteria before it is recommended for future surveys.","section":"Section 3.1 and Figure 8"}],"minor_comments":[{"comment":"The captions refer to a '(V-K) colour change' of 1.8 and 1.0, but these figures plot (V-H) and (V-J); the band labels should be corrected.","section":"Figures 9 and 10 captions"},{"comment":"The text says that 21 novae show dust signatures 'including PR Lup and V5668 Sgr', but Table 1 classifies V5668 Sgr as 'Dust Dip' rather than uncertain; this is likely meant to be V5584 Sgr. The same inconsistency appears in Section 4.2, where the eight 'IR excess' novae include PR Lup and V5584 Sgr, even though Table 1 labels them 'Dust/Unsure'.","section":"Section 2.2 and Table 1"},{"comment":"The abstract and conclusion state '50-70% of novae form dust' without the qualifiers used in Section 4.2, which limits the statement to the sample and to currently discovered novae. The wording should be aligned throughout.","section":"Abstract and Section 5"},{"comment":"The intrinsic (V-K)0 peak color has a large scatter of 0.8 mag based on 9 novae, so the resulting 'likely dust' threshold of (V-K)0 > 5.1 has considerable uncertainty; this should be stated wherever the threshold is offered for cases with no peak photometry.","section":"Appendix C"}],"recommendation":"major_revision","confidential_remarks":"The paper is within the scope of a standard astronomy journal and the dataset is a useful community resource. The main risk is that the headline 50-70% fraction is likely to be quoted without the sample-selection and classification caveats. If the authors add a sensitivity analysis, soften the population-level claim to match the sample, and clarify the diagnostic's in-sample nature, the paper could become acceptable."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First thing to know: the headline result — 50-70% of novae form dust, tripling Strope et al.'s 18% — is a real empirical claim, presented with honest caveats and public data. The V-K color diagnostic is practical and well-illustrated, and the identification of an 'IR Excess' class (optically thin dust) explains much of the discrepancy with older V-band-only surveys. The paper also shows clearly that the traditional t2-dust onset correlation is likely an artifact of dust formation defining the two-magnitude decline; that argument is persuasive and worth taking seriously.\n\nThe weak spots are the ones you flagged. The dust classifications are manual; thresholds like 2.35 mag are fit to the same light curves they are then used to separate. That is fine as a description, but it is not independent validation. The gamma-ray comparison has only 7 detections, and the caveat that Fermi detections favor bright, nearby novae is real — the 86% vs 38% split could be distance, not shock physics.\n\nThe larger question is representativeness. The stress-test note is correct that the comparison to Craig et al. is optically selected, so it cannot fix a hidden population of heavily extinguished fast novae. The authors know this; they state in Section 4.2 that such novae are systematically missed by optical surveys. If that population is large, the 53% lower bound could drop below 50%. But the paper does not pretend otherwise — it flags the limitation explicitly. The size of the hidden population is an external, currently unresolved question. The ASAS-SN results they cite argue against a large faint-fast population; near-IR surveys argue it exists. The verdict has to be conditional, not a rejection.\n\nWhat the paper does well: rigorous binomial uncertainties, a clear visualable diagnostic ready for WINTER/PRIME, and a catalog of 40 light curves with classification made available. The manual classification is the biggest internal weakness, but it is not sloppy — they err toward 'Unsure' when ambiguous. It deserves a serious referee, with the request that the classification protocol be formalized (e.g., inter-rater checks, explicit definition of 'significant' brightening/dimming) and that the authors look for independent validation in the Craig et al. t2 catalog or new IR data. I'd bring this to reading group and would cite it in my own work; the central claim is probably right for the discovered population, and the selection bias discussion is useful even if unresolved.","headline":"A genuinely useful re-measurement of the nova dust fraction; the 50-70% number is plausible but sample representativeness and manual classification keep it from being definitive.","tokens_in":67599,"tokens_out":3002,"would_cite":true,"duration_ms":30791,"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":"About 50 to 70 percent of novae form dust, making dust formation the rule rather than the exception among nova eruptions.","keywords":["novae","dust formation","infrared photometry","(V-K) colour diagnostic","time-domain surveys","gamma-ray novae","white dwarfs","classical novae"],"falsifier":"Conduct a red or near-infrared selected survey of novae in the Galactic plane, count how many show a redward $(V-K)$ excursion greater than 2.35 mag from peak, and compare that fraction with 50–70 percent; a fraction well below 50 percent would show that optical discovery bias inflated the central estimate.","tokens_in":66540,"feed_emoji":"☄️","tokens_out":8092,"duration_ms":77731,"temperature":0.7,"pith_summary":"Across 40 well-sampled novae, the paper finds that between 50 and 70 percent form dust, at least $53\\pm8$ percent and at most $70^{+6}_{-8}$ percent, which is roughly three times the share inferred from optical-only light curves in earlier work. It argues that dust formation is therefore the normal outcome of a nova eruption, not a rare one. The paper also proposes that a single number, the largest redward shift in the $(V-K)$ colour from peak with a cutoff near 2.35 mag, separates dust-forming from non-dust-forming novae in optical plus infrared photometry. If the estimate holds, nova dust is common enough to make novae a primary local laboratory for how explosive transients make dust.","feed_headline":"Most novae form dust: 50–70%, new census finds","feed_subtitle":"An optical-plus-infrared color shift reveals dust in 50–70% of novae, tripling earlier estimates and sharpening new sky surveys.","key_machinery":"The load-bearing tool is the largest redward excursion of the $(V-K)$ colour measured relative to the colour at $V$-band peak, with a separation cut at $\\Delta(V-K)>2.35$ mag. Dust absorbs optical light and re-emits it in the infrared, so a large redward swing isolates dust formation while remaining insensitive to interstellar reddening. The paper combines this colour diagnostic with a manual classification of each nova as a 'Dust Dip', 'IR Excess', 'None', or 'Unsure', and shows that the $(V-K)$ cutoff cleanly separates the dust-forming groups from the others, apart from novae with red-giant companions whose intrinsic reddening mimics dust. The $(V-J)$ and $(V-H)$ colour changes work in the same direction but with more false positives, because the thermal re-emission of typical nova dust peaks inside the $K$ band.","core_discovery":"The paper's central claim is that 50–70 percent of classical novae form dust: at least 53 percent of the 40-nova sample show dust signatures in their optical and near-infrared light curves, and even under the most conservative accounting of the uncertain cases the fraction would not exceed $70^{+6}_{-8}$ percent. The authors reach this by inspecting $V$- and $K$-band light curves and classifying each nova as a clear 'Dust Dip' (dust optically thick along the line of sight), an 'IR Excess' (dust present but optically thin), 'None', or 'Unsure'. They find that very fast novae with $t_2<10$ days seldom form dust, while about 70 percent of novae with $t_2$ between 11 and 80 days do, and that novae detected in GeV gamma rays form dust more often (about 86 percent) than those not detected (about 38 percent), part of a pattern expected if dust condenses in radiative shocks. They also argue that the long-noticed correlation between $t_2$ and the onset time of dust formation is largely a measurement artifact: for roughly a third of novae the two-magnitude decline that defines $t_2$ is caused by the dust dip itself.","pith_inferences":["If a hidden population of heavily extinguished fast thin-disk novae exists, the true Galactic dust fraction could sit near or below the paper's lower bound, while short subtle dust events like V745 Sco could be missed entirely, so the two selection effects pull in opposite directions.","The paper's J/H cutoffs imply that single-band infrared surveys without simultaneous optical data will incur a measurable false-positive rate when flagging dust formation; calibrating that rate requires a blind multi-band sample.","The gamma-ray/dust correlation strengthens the radiative-shock picture only if it survives correction for distance and optical brightness, since gamma-ray detected novae in this sample are systematically bright and nearby.","A volume-complete census of dust formation in Magellanic Cloud novae, where extinction is manageable, would test the representativeness assumption without waiting for a full Galactic-plane infrared survey."],"forward_implications":["Dust formation should be treated as the norm in nova populations, so estimates of how much dust novae return to the interstellar medium should be revised upward.","Optical-only light curves miss roughly one in seven dust-forming novae that show an IR excess without a deep optical dip, so multi-band optical plus infrared photometry is needed for demographics.","A $(V-K)$ reddening of more than 2.35 mag from peak is a practical, extinction-tolerant diagnostic that upcoming red and infrared time-domain surveys can apply even when the exact peak is missed.","Very fast novae with $t_2\\le10$ days mostly avoid dust, while the majority of novae with $t_2=11$--$80$ days form dust, strengthening the old speed-class trend with direct infrared data.","Using $t_2$ as a proxy for ejecta velocity is unreliable for dust-forming novae, because the dust dip itself can determine when the light curve drops two magnitudes below peak."],"supporting_citations":[{"why":"Supplies the optical and near-infrared light-curve atlas from which the 40-nova sample is drawn.","marker":"Walter et al. (2012)"},{"why":"Previous census that found 18 percent of novae show optical dust dips; the baseline estimate this paper supersedes.","marker":"Strope et al. (2010)"},{"why":"Theoretical model that dust forms in radiative shocks, motivating the gamma-ray connection tested here.","marker":"Derdzinski et al. (2017)"},{"why":"Infrared spectroscopy of V745 Sco shows a brief CO and dust event missed by photometry alone; used for classification and as a cautionary case.","marker":"Banerjee et al. (2023b)"},{"why":"Earlier correlation between dust onset time and t2 that this paper reinterprets as an observational artifact.","marker":"Shafter et al. (2011a)"},{"why":"Earlier discussion and models of the dust-onset versus t2 correlation that the paper revisits.","marker":"Williams et al. (2013)"},{"why":"Provides Fermi-LAT detections and non-detections used to split the sample by gamma-ray status.","marker":"Franckowiak et al. (2018)"},{"why":"Source of the long-standing suggestion that very fast novae rarely form dust.","marker":"Gallagher (1977)"}],"fun_headline_variants":["Dust forms in 50–70% of novae, new census shows","Most novae make dust: fraction up to 70%, study finds","Two-thirds of novae form dust, far more than thought","New census: most novae form dust, up to 70%","Dust in 50–70% of novae: a new statistical census"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The 40 novae with dense optical and infrared coverage represent all classical novae, even though the monitoring program intentionally favored fast, recurrent, and LMC systems and optical surveys systematically miss heavily extinguished fast novae near the Galactic plane.","fun_headline_variants_meta":{"raw":{"variants":["Dust forms in 50–70% of novae, new census shows","Most novae make dust: fraction up to 70%, study finds","Two-thirds of novae form dust, far more than thought","New census: most novae form dust, up to 70%","Dust in 50–70% of novae: a new statistical census"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000566,"raw_usage":{"total_tokens":2772,"prompt_tokens":1128,"completion_tokens":1644,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":744,"completion_tokens_details":{"reasoning_tokens":1560}},"tokens_in":744,"tokens_out":1644,"duration_ms":11614,"temperature":1.0,"reasoning_tokens":1560,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T21:40:27.290415+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Conduct a red or near-infrared selected survey of novae in the Galactic plane, count how many show a redward $(V-K)$ excursion greater than 2.35 mag from peak, and compare that fraction with 50–70 percent; a fraction well below 50 percent would show that optical discovery bias inflated the central estimate.","supporting_citations":[{"cited_title":"S., 1977, @doi [ ] 10.1086/112032 , https://ui.adsabs.harvard.edu/abs/1977AJ.....82..209G 82, 209","cited_arxiv_id":null,"evidence_quote":"Source of the long-standing suggestion that very fast novae rarely form dust."}],"review_version":1}