{"id":"9ee7b9d4-98f7-4c69-8196-6792995309b5","arxiv_id":"2607.14850","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":10,"one_line_summary":"Upper Scorpius disks at 5-12 Myr have a median dust mass of 0.95 Earth masses, about six times lower than 1-2 Myr Chamaeleon I disks, indicating strong millimeter-grain depletion.","lead":"This paper measures dust masses of 136 protoplanetary disks in the 5-12 Myr old Upper Scorpius region using a new radiative-transfer-based temperature calibration and porous dust opacities. It finds these older disks contain about six times less millimeter-sized dust than younger disks in Chamaeleon I, sharpening evidence for solid-material depletion during planet formation.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Factor-of-six depletion may be inflated: T_dust calibration applies a young-region Mdot relation to 5-12 Myr USco, and viscous heating contributes 5.2 K at L*=0.1 Lsun.","rationale":"The reader correctly identifies the T_dust calibration as the least-secure link. I narrow this to the age-dependent accretion-rate input. The paper's own Sec. 3.5 shows that removing viscous heating changes T_dust by 5.2 K at L*=0.1 Lsun, so the calibration is not insensitive to Mdot. Because the fitted Mdot-M* relation is drawn from regions of age ~1-3 Myr and USco is 5-12 Myr, applying it to USco risks overpredicting T_dust and under-predicting dust masses. This matters for the quantitative 'six times' claim, though the broad conclusion of declining millimeter-dust mass is supported by raw flux medians and prior work. The disk-size scaling (Andrews+18) works in the opposite direction and is therefore not the main threat. A direct recalculation with reduced Mdot would settle whether the factor of six survives. I keep the reader's conditional verdict: the paper's core conclusion is likely correct, but the headline ratio needs an age-dependent accretion correction or an explicit caveat. Since the reader already conditioned on the temperature calibration, no verdict change is needed beyond what was recommended.","tokens_in":16264,"tokens_out":15116,"duration_ms":142289,"concrete_test":"Recompute the 27 L*=0.1 Lsun models described in Sec. 3.5 with Mdot reduced by factors of 3 and 10 and with Mdot=0, keeping all other grid parameters fixed. Refit Eq. (8) to the resulting T_dust values, reapply it to the USco and Cha I samples, and recalculate the median ratio. If the USco median rises by more than ~40% or the USco/Cha I median ratio drops below ~4, the factor-of-six headline should be reported as an upper envelope, not a robust depletion estimate.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central quantitative claim is the factor-of-six median dust-mass depletion (0.95 vs 5.69 M⊕). This factor is set by Eq. (8). At L*=0.1 Lsun the new relation gives T_dust=24.8 K, and Sec. 3.5 states that viscous heating contributes 5.2 K of that value. The grid assigns Mdot via log Mdot=-7.93+1.95 log M* (Manara+23), a relation fitted to 256 disks in Taurus, Lupus, Ophiuchus, and Cha I (Sec. 3.4, Fig. 2) - all young regions, with no USco data. USco is 5-12 Myr and accretion rates decline with age. If the true USco Mdot is 3-10x lower, the viscous-heating term is overestimated, T_dust for low-mass USco stars is too high, and their dust masses are underestimated by roughly a Planck factor 1.4 (more if the age effect is larger). This bias is age-selective because Cha I is young, so the same Mdot prescription is more defensible there. The mass-restricted comparison in Sec. 4.1 does not cure this; it reuses the same T_dust calibration. The observed flux decline (factor 2.4) and prior work still support substantial evolution, but the specific 'six times lower' value may be an overestimate. The disk-size part of the same calibration is less worrying: if old disks are more compact, T_dust would rise, which would strengthen rather than weaken the depletion signal.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper derives a new dust-temperature--stellar-luminosity calibration from a grid of 918 HOCHUNK3D radiative-transfer models that include stellar-mass-dependent disk sizes and accretion rates and porous DSHARP dust, then uses the relation (Eq. 8) with κ_0.88mm = 0.84 cm2/g to convert the Carpenter et al. (2025) ALMA 0.88 mm fluxes of 136 full/transitional Upper Scorpius disks into dust masses. Using Kaplan-Meier survival analysis, it reports a median dust mass of 0.95 M_Earth for USco versus 5.69 M_Earth for a spectrally matched Cha I sample, and a slightly steeper M_dust-M* relation in USco (slope 1.65 versus 1.38). The paper tests the optically thin approximation with its model grid and compares the USco distribution to the exoplanet mass distribution.","tokens_in":16651,"tokens_out":6661,"duration_ms":56875,"significance":"If the result holds, the paper sharpens the evidence for millimeter-dust depletion over 5-12 Myr and provides a new T_dust(L*) relation that can be used by other groups. The main strengths are the well-defined 136-disk USco sample, the explicit radiative-transfer treatment that improves on a fixed 20 K or a single power-law T_dust assumption, the use of censored-data statistics, the explicit test of the optically thin approximation, and the mass-restricted comparison. The porous-opacity choice is transparent and the model grid is clearly described. The central risk is that the T_dust calibration inherits external scaling relations (disk size versus M*, Mdot versus M*) fitted in younger regions and extrapolated to 5-12 Myr USco; this directly affects the magnitude of the claimed depletion relative to Cha I.","major_comments":[{"comment":"The grid assigns Mdot using log Mdot = -7.93 + 1.95 log M*, fitted to 256 disks in Taurus, Lupus, Ophiuchus, and Chamaeleon I (Fig. 2). These are young regions, while USco is 5-12 Myr and accretion rates decline with age. Section 3.5 states that viscous heating contributes 5.2 K at L* = 0.1 Lsun, a substantial part of the 24.8 K predicted by Eq. (8). If the true USco Mdot values are lower, the inferred T_dust for low-mass USco stars is too high and the derived dust masses are too low. This bias is age-selective because the same Mdot relation is more defensible for the young Cha I sample, so the factor-of-six median depletion in Sect. 4.1 is likely an overestimate. The mass-restricted comparison in Sect. 4.1 does not cure this because it reuses the same T_dust calibration. Please quantify the sensitivity of the median masses to reduced or zero viscous heating, or constrain Mdot from USco","section":"Sect. 3.4, Eq. (8), and Sect. 3.5"},{"comment":"The central values 0.95 and 5.69 M_Earth are quoted as point estimates without uncertainties. No confidence intervals are given for the Kaplan-Meier medians, and the scatter in Eq. (8) and the opacity uncertainty are not propagated. Because the factor-of-six depletion is the main quantitative claim, the paper should report bootstrap or analytic uncertainties on the medians (and on their ratio), including the censored fraction and finite sample sizes. The narrow-mass subsample medians (2.83 versus 0.44 M_Earth) also need uncertainties before they can be used as supporting evidence.","section":"Sect. 4.1"},{"comment":"Section 2 states that sources without Gaia distances are assigned 145 pc (USco) and 190 pc (Cha I), while Section 4 states 145 pc and 160 pc. Since M_dust depends on D^2, this is a factor of (190/160)^2 = 1.41 for the affected Cha I sources. Please reconcile the two numbers, state how many sources lack Gaia distances, and re-run the relevant medians if necessary.","section":"Sect. 2 versus Sect. 4"}],"minor_comments":[{"comment":"The viscous dissipation expression is garbled in the typesetting; the formula should be checked and reproduced cleanly.","section":"Eq. (6)"},{"comment":"The informal phrasing 'gazillions of absorption/re-emission events' should be replaced by a quantitative statement about photon statistics and convergence.","section":"Sect. 3.3"},{"comment":"For reproducibility, consider providing a machine-readable table of the 136 sources with adopted distances, T_dust, and derived dust masses, and likewise for the Cha I sample.","section":"General"},{"comment":"The comparison to the exoplanet mass distribution is made only at the level of median values; a two-sample test on the full distributions would make the 'mass budget' statement more robust.","section":"Sect. 4.1"},{"comment":"The caption uses 'Chamaeleon' while the text uses 'Chamaeleon I'/'Cha I'; unify the notation.","section":"Fig. 2 caption"}],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe headline number—median dust mass 0.95 M⊕ for USco versus 5.69 M⊕ for Cha I, a factor of six—is directionally solid but the exact ratio is softer than the abstract implies. The T_dust–L* calibration (Eq. 8) is built from radiative-transfer models that assign accretion rates via a Manara+23 relation fitted to Taurus, Lupus, Ophiuchus, and Cha I—all ≤2 Myr regions. At L*=0.1 L⊙, viscous heating contributes 5.2 K of the 24.8 K. USco is 5–12 Myr; if its accretion rates are a few times lower, those temperatures drop and the inferred masses rise by ~1.4, shrinking the depletion factor to roughly four. The paper doesn't address this age dependence. That's the main soft spot.\n\nWhat's genuinely new: a 918-model grid with stellar-mass-dependent disk sizes, accretion, and porous dust; a new T_dust(L*) relation that gives warmer temperatures for low-mass stars than Andrews+13 or van der Plas+16; and application to the largest ALMA USco sample (136 disks). The optically-thin assumption is tested (median 82% of mass above τ=1 surface), which is good practice. The stellar-mass–dust-mass slopes (1.65±0.19 USco, 1.38±0.25 Cha I) are consistent with Pascucci+16 and the steepening with age is plausible.\n\nSoft spots, in proportion: (1) the age-dependent Mdot issue above; (2) median dust masses are quoted without uncertainties—the survival analysis and Linmix give confidence intervals on the M_dust–M* relation but not on the median itself; (3) no code or data artifacts, so the numbers aren't independently reproducible without nontrivial work; (4) a small inconsistency in the adopted Cha I distance (190 pc in Sec. 2, 160 pc in Sec. 4). None of these kills the core finding—the observed flux decline (factor 2.4) and prior work support substantial dust evolution—but the paper should be transparent about the model dependence of the six.\n\nWho's it for: anyone working on disk evolution, dust mass budgets, or the exoplanet mass budget problem. It deserves a serious referee; the T calibration is a useful community resource even if the exact depletion factor is debated.","headline":"Directionally strong, numerically overstated: the factor-of-six depletion is likely closer to four once the age dependence of accretion is accounted for.","tokens_in":17166,"tokens_out":5468,"would_cite":true,"duration_ms":44843,"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":"Upper Scorpius disks, at 5–12 Myr, hold roughly six times less millimeter dust than younger Chamaeleon I disks, with a median of 0.95 Earth masses.","keywords":["protoplanetary disks","dust mass","Upper Scorpius","Chamaeleon I","ALMA continuum","radiative transfer","dust temperature","millimeter grains"],"falsifier":"Measure resolved disk sizes and accretion rates for a sample of M-dwarf disks in Upper Scorpius: if their characteristic radii do not follow the assumed mass scaling (e.g., they are as large as those of higher-mass stars), the warm-temperature prediction fails and the inferred median dust mass would increase, shrinking the sixfold depletion. Alternatively, spatially resolved dust temperature measurements at 0.88 mm for individual low-luminosity disks would directly test the T_dust–L* relation.","tokens_in":16119,"feed_emoji":"🪐","tokens_out":2802,"duration_ms":24156,"temperature":0.7,"pith_summary":"The paper tries to establish that by ages of 5–12 Myr, protoplanetary disks in Upper Scorpius have lost most of their millimeter-sized dust: the median dust mass among 136 full and transitional disks is 0.95 Earth masses, about six times lower than the 5.69 Earth masses typical of younger Chamaeleon I disks. The authors argue this difference is real, not an artifact of temperature assumptions, because their new radiative-transfer calibration yields warmer dust temperatures for low-mass stars than older prescriptions. This matters because dust mass sets the available solid budget for planet formation: the comparison with mature exoplanet systems suggests the solids have already been processed into larger bodies, drifted inward, or become hidden from millimeter observations.","feed_headline":"Dust in 10-Myr disks drops sixfold, census shows","feed_subtitle":"A new temperature calibration for 136 Upper Scorpius disks suggests most millimeter dust is already gone.","key_machinery":"The argument is carried by a new model-derived relation between dust temperature and stellar luminosity, log T_dust = 1.506 + 0.115 log L* + 0.004 (log L*)^2, together with a porous dust opacity of κ_0.88mm = 0.84 cm²/g. The relation is built from a grid of 918 radiative transfer models that include stellar-mass-dependent disk sizes and accretion rates, so that low-mass stars have smaller, more compact disks that are heated more efficiently. This machinery directly converts observed 0.88 mm fluxes into dust masses via the optically thin formula, and it is the reason the paper claims the sixfold depletion is not simply an artifact of assuming a uniform 20 K dust temperature.","core_discovery":"The central claim is that the dust reservoirs of Upper Scorpius disks are substantially depleted by ages of ~5–12 Myr, with a median dust mass of 0.95 Earth masses (about six times below the 5.69 Earth masses of younger Chamaeleon I disks), and that this depletion is inferred using a new calibration of dust temperature versus stellar luminosity. The authors derive a grid of 918 radiative transfer models that account for stellar-mass-dependent disk sizes, accretion rates, and porous dust properties, yielding log T_dust = 1.506 + 0.115 log L* + 0.004 (log L*)^2. This predicts systematically higher dust temperatures for low-mass stars than earlier prescriptions, which raises inferred masses for","pith_inferences":["An editor might infer that if the assumed disk-size–stellar-mass scaling is too steep for the low-mass Upper Scorpius population, the inferred dust temperatures for those stars could be overestimated, meaning the true median dust mass might be higher than 0.95 Earth masses and the sixfold depletion somewhat weaker.","The new calibration implies that previous dust mass estimates for old, low-mass disks using a fixed 20 K temperature or older luminosity scalings may be systematically too high, since those older prescriptions yield cooler temperatures for low-mass stars.","The paper's mechanism suggests a testable extension: direct submillimeter imaging of a few low-mass Upper Scorpius disks to measure their characteristic radii would validate or falsify the compact-disk assumption that drives the warm-temperature prediction.","A natural consequence of the steepening slope is that dust-trapping pressure bumps must be less effective in lower-mass disks; this could be tested by comparing the frequency of rings and gaps in high- versus low-mass disks at similar ages."],"forward_implications":["If the median dust mass in Upper Scorpius is 0.95 Earth masses, then the reservoir of millimeter-sized grains available for planet formation at 5–12 Myr is far smaller than at 1–2 Myr.","The comparison with exoplanet masses (median ~4.28 Earth masses) suggests that by 5–12 Myr a large fraction of the original solid material has been incorporated into larger bodies, removed by radial drift, or hidden in optically thick regions.","The slightly steeper dust-mass–stellar-mass relation in Upper Scorpius (slope 1.65 versus 1.38 in Chamaeleon I) is consistent with millimeter grains being depleted more efficiently around lower-mass stars.","The optically thin approximation used to compute dust masses is supported by the models: in the majority of grid models, the optically thick region is confined to the inner disk and most of the dust mass lies in optically thin regions.","The new temperature calibration raises inferred dust masses for low-mass stars compared with earlier prescriptions, which partly compensates for the lower porous-dust opacity and changes absolute but not relative dust mass estimates."],"fun_headline_variants":["Old disks: sixfold dust drop by 5–12 Myr","Upper Scorpius disks: six times less dust than young","Aged disks: sixfold dust loss by 10 Myr","Census of 136 old disks reveals sixfold dust loss"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The derived dust masses hinge on assuming that disk radii scale with stellar mass as log R_eff = 1.77 + 0.58 log M* and accretion rates as log Mdot = −7.93 + 1.95 log M*, extrapolated down to the low-mass Upper Scorpius population; if actual disks are larger or accreting more slowly, the inferred temperatures drop and dust masses rise.","fun_headline_variants_meta":{"raw":{"variants":["Old disks: sixfold dust drop by 5–12 Myr","Upper Scorpius disks: six times less dust than young","Aged disks: sixfold dust loss by 10 Myr","Census of 136 old disks reveals sixfold dust loss"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000524,"raw_usage":{"total_tokens":2431,"prompt_tokens":865,"completion_tokens":1566,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":609,"completion_tokens_details":{"reasoning_tokens":1505}},"tokens_in":609,"tokens_out":1566,"duration_ms":12087,"temperature":1.0,"reasoning_tokens":1505,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-02T00:51:43.114172+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure resolved disk sizes and accretion rates for a sample of M-dwarf disks in Upper Scorpius: if their characteristic radii do not follow the assumed mass scaling (e.g., they are as large as those of higher-mass stars), the warm-temperature prediction fails and the inferred median dust mass would increase, shrinking the sixfold depletion. Alternatively, spatially resolved dust temperature measurements at 0.88 mm for individual low-luminosity disks would directly test the T_dust–L* relation.","supporting_citations":[],"review_version":1}