{"id":"b842136e-7ae0-420d-a689-db61c95ad451","arxiv_id":"2506.10435","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Overdense dust filaments produced by the streaming instability become optically thick at sub-millimeter wavelengths, biasing disk mass estimates low by factors of about 2 to 7.","lead":"Dust clumps that form in swirling disks around young stars can block their own millimeter glow, making the disks look two to seven times less massive than they really are. This helps explain a long-standing mismatch between the mass seen in planet-forming disks and the mass needed to build the planets seen around older stars.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 2–7 mass-excess factors are conditional on initializing the 10/30 au boxes in the strong-clumping regime via Z=0.03 and a per-species Z/Π/4 normalization; at lower, observationally motivated Z the runs would resemble the 100 au control and yield Λ≈1.","rationale":"The paper is internally consistent and releases the protoRT codebase, which is a genuine reproducibility asset. Given the chosen Z, grain sizes, and the Z/Π/4 placement, the 10 and 30 au boxes do produce filaments, the radiative transfer is applied consistently, and the scattering treatment follows the Miyake-Nakagawa solution. The 100 au control is a useful null case and shows what happens outside the clumping regime. My concern is about external validity: the central quantitative claim is conditional on input parameters that are not independently anchored. The paper's own framing, that streaming instability yields a mass excess when it is actively concentrating dust, is a conditional statement; the abstract, however, presents the 2–7 factors as a finding about protoplanetary disks. Since the mass-budget problem is about reconciling observed low dust masses with planet formation, initializing Z=0.03 in the very regions where the correction is claimed partly assumes the unseen mass that the paper aims to make room for. This is not a formal circularity, but it makes the headline numbers fragile. A single lower-Z run at 30 au would decide whether the correction survives at more conservative solid abundances. The reader's CONDITIONAL verdict is therefore the correct one; my concern reinforces it rather than overturning it.","tokens_in":21960,"tokens_out":13926,"duration_ms":181330,"concrete_test":"Run the 30 au simulation exactly as in Table 1 but with the total dust-to-gas ratio lowered from Z=0.03 to Z=0.01 (per-species Z=0.0025), keeping Π, grain sizes, resolution, and runtime fixed, and compare the maximum dust density, τ_eff maps, and Λ at 0.87 mm with Figure 8. If the lower-Z run no longer produces optically thick filaments and Λ stays near unity, the paper's quoted correction range is not applicable to disks whose solid abundance is closer to the low values inferred from mm fluxes, and the conclusion must be reframed as a demonstration conditional on Z≳0.03 rather than a general resolution of the mass-budget problem.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The headline factors are not generic properties of streaming-instability simulations; they are properties of boxes placed inside the strong-clumping region of the Z/Π plane. In Section 2.1.1 and Figure 2 the authors set Z=0.03 for every run and then evaluate the Lim et al. (2025) boundary using Z/Π/4, i.e., after dividing the total solid abundance by the four dust species. The 10 and 30 au runs sit on the clumping side; the 100 au run sits on the quiescent side and indeed produces no filaments, filling factors near zero, and Λ≈1. Thus the entire 2–7 range is downstream of a single input choice: the true midplane dust-to-gas ratio in the inner disk is at least about 0.03, or the per-species Z/Π exceeds the clumping threshold. Observationally motivated disk masses, however, correspond to much lower global dust-to-gas ratios, and the mass-budget problem is precisely about whether the unseen mass is large enough to put disks in this regime. If a real 30 au disk had Z=0.01, or if the correct polydisperse clumping criterion uses the total Z rather than Z/4, the same grain sizes and Π would place it below the threshold, and the simulations would predict no optically thick filaments and hence no correction. The paper acknowledges that the precise clumping threshold is an active research area, but it does not test the sensitivity of its central result to this particular normalization. This is the load-bearing step that must be true for the abstract's claim to describe actual Class II disks.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper investigates whether optically thick (sub-)millimeter emission from dust filaments formed by the streaming instability can bias dust mass estimates that assume optically thin emission. The authors run three local shearing-box simulations at 10, 30, and 100 au with four dust sizes, self-gravity, and sink particles, then perform radiative transfer with DSHARP opacities including scattering. They define the mass excess Lambda = m_true / m_obs from the standard optically thin estimator and report Lambda ~ 2-7 at 10 au, ~ 1.5-2.4 at 30 au, and ~ 1 at 100 au. The paper concludes that such clumping can partially explain the protoplanetary disk mass budget problem.","tokens_in":22358,"tokens_out":9775,"duration_ms":116270,"significance":"The paper provides a clean numerical experiment with a useful control run at 100 au, a public code release (protoRT), and a physically motivated mechanism linking streaming-instability clumping to observational bias. If the result is robust, it strengthens the argument that flux-based disk masses are lower limits and that optically thick filaments can hide substantial mass even at low filling factors. The main limitation is the narrow parameter space: the strong-clumping regime is entered by construction through Z=0.03 and an ad hoc per-species normalization of the clumping criterion, so the generality of the 2-7 factors is not yet demonstrated.","major_comments":[{"comment":"The placement of the 10 au and 30 au runs inside the strong-clumping regime is determined by the initial solid-to-gas ratio Z=0.03 together with the normalization of the clumping criterion to Z/Pi/4 (Eq. 11 and Fig. 2). The paper does not test the sensitivity of the resulting mass excess to either Z or this normalization; the 100 au run, which lies below the same threshold, produces no filaments and yields Lambda ~ 1. Because the authors themselves note in Section 2.1.1 that the precise clumping threshold is an active area of research, the central claim Lambda ~ 2-7 cannot yet be considered robust. A low-Z control run (e.g., Z=0.01 at 30 au) or a comparison with the unnormalized Z/Pi criterion is needed to determine whether the result is a generic property of streaming-instability clumping or a consequence of the specific parameter choice.","section":"Section 2.1.1 and Fig. 2"},{"comment":"The mass excess values are computed for local shearing boxes of size 0.2H, yet Section 4 compares them directly to disk-integrated mass budget discrepancies reported by Manara et al. (2018) and Mulders et al. (2021). The manuscript should either perform a radial integration over a disk model to derive a disk-integrated mass correction factor or explicitly restrict the claim to the modeled patches; the 2-7 range arises from a single 10 au patch and, as the authors state in Section 2.2, the face-on geometry gives a lower limit, so the direct comparison with survey-level mass deficits may overstate the applicability of the quoted factors.","section":"Eq. (29) and Section 4"}],"minor_comments":[{"comment":"The text defines Z in Eq. (11) as the total dust-to-gas ratio, but the y-axis of Fig. 2 and the clumping argument use the per-species value Z/4; the authors should state explicitly that each species is initialized with Z/4 and label the figure accordingly.","section":"Section 2.1.1 and Fig. 2"},{"comment":"The text uses 'T' for the orbital period in the description (e.g., 'before t/T ~ 25') while earlier in the paper 'P' is used for the orbital period; the notation should be standardized.","section":"Section 3 and Fig. 8"},{"comment":"The quantity f_thick,nu is defined as the ratio of the mean intensity to the blackbody intensity, not as a geometric fraction of optically thick area; the name 'optically thick fraction' is misleading and should be replaced with a term such as 'normalized intensity' or 'emissivity ratio' to avoid confusion with f_fill.","section":"Section 2.3, Eq. (27)"},{"comment":"The opacity binning procedure is described clearly, but it would be helpful to state the four bin boundaries explicitly (a_min and a_max for each size bin) in the text, rather than only in Fig. 5, to make the opacity averaging reproducible.","section":"Section 2.2"},{"comment":"The paper compares its results with those of Rucska and Wadsley (2023) and Scardoni et al. (2021), but it does not state quantitatively how the inclusion of scattering changes the comparison; a sentence summarizing the difference after including scattering would clarify the advance over this earlier work.","section":"Section 4"}],"recommendation":"major_revision","confidential_remarks":"The paper is methodical, and the code release is a valuable contribution. The main concern is that the headline result is conditional on the adopted Z and the per-species/clumping-criterion normalization, which the authors themselves flag as uncertain; this is fixable with additional control runs or a moderated claim. I recommend major revision rather than rejection because the central derivation and radiative transfer are sound and the issue is one of demonstrated robustness."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"You should read this if you care about the disk mass budget problem. The paper does something genuinely new: it runs self-gravitating streaming-instability simulations with four dust species, then does radiative transfer with scattering, and computes the bias in the standard optically thin mass estimator. The inclusion of scattering is the key addition over Rucska & Wadsley, and it roughly doubles to triples the mass excess. I also give them credit for the clean control case at 100 au, which sits outside the clumping regime and produces Lambda near unity at long wavelengths. That control tells you the machinery is behaving, and it sharpens the interpretation of the 10/30 au results.\n\nThe internal consistency is good. The mass excess is computed from simulated intensities and the standard estimator, not fitted. They release the code and data, which is reproducible and counts for something. The finding that low filling factors can still hide substantial mass is supported by the time series in Figure 8, not just by one snapshot. For a paper that is partly a methods demonstration, this is solid work.\n\nNow the soft spot, and it is a real one. The 2–7 factors are not generic properties of the streaming instability; they come from initializing the 10 and 30 au boxes at Z = 0.03 and then using the Z/Pi/4 normalization to place them inside the Lim et al. clumping region. The 100 au run with the same Z falls below the threshold and produces no filaments, no planetesimals, and Lambda about one. So the entire headline range is downstream of that clumping-regime choice. The authors acknowledge that the threshold is an active research area, but they do not test the sensitivity of the central result to the normalization or to lower Z. If the real inner disk has Z closer to 0.01, or if the correct polydisperse criterion uses total Z rather than Z/4, the same setup would look like the 100 au control. That is not a fatal flaw, but it does mean the abstract overstates the generality. The paper is transparent about the ingredients, and the logic within the chosen regime holds up.\n\nThere is also a single realization per radius, so no realization-to-realization scatter, and a single opacity model. Both limits are acknowledged or at least mentioned. These are the usual kinds of things you would ask for in revision, not reasons to reject.\n\nWho should read it: anyone interpreting ALMA continuum masses, and people modeling planetesimal formation. It is a useful contribution with reproducible code, and the caveats are manageable. I would send it to peer review. The referee should ask for a sensitivity study on Z and the clumping criterion, and ideally a second realization, but the paper deserves a serious referee and likely publication after those tests.","headline":"Genuine step forward on how streaming instability plus scattering can bias ALMA mass estimates, but the headline 2–7 range is conditional on the boxes being placed in the strong-clumping regime, and the paper's own 100 au control shows how much depends on that choice.","tokens_in":22872,"tokens_out":1706,"would_cite":true,"duration_ms":24512,"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":"Overdense dust filaments produced by the streaming instability are optically thick at (sub-)millimeter wavelengths, so standard flux-based disk mass estimates can undercount the true dust mass by factors of 2–7 in the inner disk.","keywords":["protoplanetary disks","streaming instability","dust emission","radiative transfer","optically thick emission","dust mass estimates","mass excess","missing mass problem"],"falsifier":"Take a resolved (sub-)millimeter image of an inner-disk region where the streaming instability is expected to operate and measure the optical depth of individual filaments directly, for example from multi-wavelength flux ratios or from the spectral index. If no structure with $\\tau\\gtrsim1$ is found in a disk whose independently measured mass already matches the optically thin estimate, the proposed hidden-mass mechanism is absent for that disk. Alternatively, a full radiative-transfer model of a real disk that recovers the same mass as the optically thin flux argument would falsify the claimed bias for that disk.","tokens_in":21795,"feed_emoji":"🪐","tokens_out":11538,"duration_ms":115404,"temperature":0.7,"pith_summary":"The paper argues that the long-standing 'missing mass' problem of protoplanetary disks can be partly explained by a breakdown of the usual observational assumption that (sub-)millimeter emission is optically thin. In shearing-box simulations of the streaming instability with self-gravity, the instability gathers dust into overdense filaments whose effective optical depth at (sub-)millimeter wavelengths reaches $\\tau>1$; when the emergent intensity is converted to a dust mass with the standard optically thin formula, the inferred mass falls short of the true dust mass by factors of roughly 2–7 at 10 au and 1.5–2.4 at 30 au. The bias persists even when the fraction of the area that is optically thick is small, and dust scattering roughly doubles to triples the correction. If this holds, flux-based disk mass surveys are lower limits on the solid content available for planet formation.","feed_headline":"Dust filaments hide 2-7x more disk mass than surveys count","feed_subtitle":"Streaming instability makes (sub-)mm emission optically thick, so flux-based disk masses may be lower limits.","key_machinery":"The load-bearing pair is the streaming instability acting as the clumping agent and the mass-excess ratio $\\Lambda_\\nu \\equiv m_{\\mathrm{true}}/m_{\\mathrm{obs},\\nu}$ as the measure of observational bias. The radiative transfer uses the effective optical depth $\\tau^{\\mathrm{eff}}_\\nu = \\kappa^{\\mathrm{eff}}_\\nu \\Sigma_d$, with a density-weighted dust opacity that includes both absorption and scattering, and the emergent intensity is computed from a plane-parallel, isothermal slab solution with isotropic scattering. Four dust species (0.36–12 mm radii) evolve in a three-dimensional shearing box with self-gravity, and dust that exceeds twice the Hill density is converted into sink particles representing planetesimals.","core_discovery":"The central claim is that the streaming instability's overdense filaments are optically thick at (sub-)millimeter wavelengths, so interpreting their emission under the optically thin assumption produces a mass excess $\\Lambda_\\nu = m_{\\mathrm{true}}/m_{\\mathrm{obs},\\nu}$ of $\\sim$2–7 in the inner disk. In the 10 au and 30 au runs, clumps and filaments reach effective optical depths $\\tau^{\\mathrm{eff}}_\\nu \\gtrsim 1$ at 0.3–3 mm, and even a small filling factor of optically thick columns is enough to bias the inferred mass. Scattering by millimeter-to-centimeter grains suppresses the emergent intensity and increases $\\Lambda_\\nu$ by factors of $\\sim$1.5–3 relative to absorption-only calculations, while in the optically thin, high-albedo 100 au control case the mass excess can drop slightly below unity, meaning the standard estimate overpredicts mass by up to $\\sim$15%.","pith_inferences":["Were these simulations representative of real disks, the population-level missing mass would shrink by the same 2–7 factor only where strong clumping occurs; disks with lower dust-to-gas ratios or stronger turbulence would need little or no correction.","Because the simulations are evaluated face-on and the paper notes that optical depth grows with inclination, the quoted mass excess is a lower limit for typical inclined disks; extending the radiative transfer to inclined viewing would likely increase the inferred bias.","A natural test is to apply the same optically thin inversion to synthetic images of disks with measured substructure: if filaments observed at 0.87 mm have $\\tau\\ge1$, the multi-wavelength spectral index of those disks should show the flux-deficit pattern predicted here.","This result implies that pebble-accretion efficiencies inferred from comparisons of disk dust masses to exoplanet masses could be systematically underestimated; reconciling the mass budget may not require near-unity accretion efficiency if the disk masses themselves are undercounted."],"forward_implications":["Observed (sub-)mm disk masses in the inner about 30 au should be read as lower limits; the true dust mass can be 2–7 times larger where the streaming instability is active.","Dust scattering cannot be neglected when converting (sub-)mm fluxes to masses; including it raises the correction factor by 1.5–3.","A small optically thick fraction is enough to bias mass estimates: filling factors below about 0.05 can still hide a significant fraction of the mass.","At larger radii where the instability does not strongly clump dust, the standard mass estimate can slightly overestimate the dust mass (by up to about 15%) because scattering boosts the emergent intensity.","Longer-wavelength observations that see optically thin emission are needed to recover the hidden mass."],"supporting_citations":[{"why":"Defines the mass excess metric and shows that assuming optically thin emission at 1.3 mm underestimates disk mass.","marker":"Y. Liu et al. 2022"},{"why":"Conducted streaming-instability simulations with self-gravity and realistic grain sizes, finding order-unity mass correction factors without scattering.","marker":"J. Rucska & J. Wadsley 2023"},{"why":"Showed that streaming-instability clumping can produce high optical depth and mass excess in disks.","marker":"C. E. Scardoni et al. 2021"},{"why":"Demonstrated that dust scattering reduces emergent (sub-)mm intensity and can affect mass estimates.","marker":"Z. Zhu et al. 2019"},{"why":"Showed that at high albedo and low optical depth scattering can boost the emergent intensity, explaining the sub-unity mass excess at 100 au.","marker":"A. Sierra & S. Lizano 2020"},{"why":"Supplies the DSHARP dust model whose absorption and scattering opacities are used in the radiative transfer.","marker":"T. Birnstiel et al. 2018"},{"why":"Provides the analytical plane-parallel slab solution with isotropic scattering used to compute the emergent intensity.","marker":"K. Miyake & Y. Nakagawa 1993"},{"why":"Provides the strong-clumping criterion in the Z/Π plane used to place the 10 and 30 au simulations.","marker":"J. Lim et al. 2025"}],"fun_headline_variants":["Streaming instability dust filaments bias disk masses by 2-7x","Optically thick filaments hide 2-7x protoplanetary disk mass","Scattering boosts streaming instability mass bias to 7x","Dust filament scattering inflates disk mass estimates by 7x","Streaming instability filaments cause 2-7x disk mass error"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The 10 au and 30 au runs are initialized in the strong-clumping regime of the streaming instability (the dust-to-gas ratio relative to the pressure-gradient parameter, $Z/\\Pi$, above the clumping threshold); if real disks at those radii have lower dust-to-gas ratios, different grain-size distributions, or stronger turbulence, the overdense filaments and the 2–7 correction factors may not appear.","fun_headline_variants_meta":{"raw":{"variants":["Streaming instability dust filaments bias disk masses by 2-7x","Optically thick filaments hide 2-7x protoplanetary disk mass","Scattering boosts streaming instability mass bias to 7x","Dust filament scattering inflates disk mass estimates by 7x","Streaming instability filaments cause 2-7x disk mass error"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000364,"raw_usage":{"total_tokens":1957,"prompt_tokens":937,"completion_tokens":1020,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":553,"completion_tokens_details":{"reasoning_tokens":928}},"tokens_in":553,"tokens_out":1020,"duration_ms":11626,"temperature":1.0,"reasoning_tokens":928,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T04:27:27.453549+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take a resolved (sub-)millimeter image of an inner-disk region where the streaming instability is expected to operate and measure the optical depth of individual filaments directly, for example from multi-wavelength flux ratios or from the spectral index. If no structure with $\\tau\\gtrsim1$ is found in a disk whose independently measured mass already matches the optically thin estimate, the proposed hidden-mass mechanism is absent for that disk. Alternatively, a full radiative-transfer model of a real disk that recovers the same mass as the optically thin flux argument would falsify the claimed bias for that disk.","supporting_citations":[{"cited_title":"Planetesimal formation via the streaming instability with multiple grain sizes","cited_arxiv_id":"2305.11297","evidence_quote":"Conducted streaming-instability simulations with self-gravity and realistic grain sizes, finding order-unity mass correction factors without scattering."},{"cited_title":"E., Booth, R","cited_arxiv_id":null,"evidence_quote":"Showed that streaming-instability clumping can produce high optical depth and mass excess in disks."}],"review_version":1}