{"id":"8495e051-ac19-477a-878b-3e8a883bd8ba","arxiv_id":"1908.10589","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":6,"one_line_summary":"Simulations show that dust in gravitationally fragmenting protoplanetary disks grows to decimeter sizes and concentrates into dense clump centers, which may seed giant planet formation.","lead":"This paper uses computer simulations to track how dust grows and moves inside a young star's disk when the disk is so massive it breaks into clumps. The results suggest that dust can gather into dense clump centers that may turn into giant planets, and that disrupted clumps can leave dusty rings.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Seed masses are an extrapolation: Rcrit=0.18 au lies inside the unresolved inner clump, so the quantitative claim is not a simulation result.","rationale":"The paper is a careful thin-disk study with genuinely new results on dust growth and concentration inside gravitationally fragmenting clumps. The qualitative picture, including the formation of massive dusty central condensations and the possibility of second collapse before tidal dispersal, is supported by the resolved clump profiles and the transient 2000 K central temperatures. However, the specific seed masses in Table 3 depend on an extrapolation of the interior profiles below the resolved scale, an assumed angular-momentum conservation law, and a chosen protoplanet radius. The paper explicitly acknowledges that the second collapse is not modeled, and Section 3.4 lists several caveats, which is to the authors' credit. My concern is not that the authors are claiming more than they state; rather, the load-bearing quantitative claim is precisely the part that is not independently tested. The reader already identified this weakness and assigned CONDITIONAL at moderate confidence, so I do not see a reason to change the verdict. A targeted re-derivation of the seed masses with alternative inner profiles and J perturbations would settle whether the quoted range is robust.","tokens_in":17822,"tokens_out":4757,"duration_ms":53467,"concrete_test":"Recompute the Section 3.3 seed masses with three modifications: (1) replace the plateau-plus-zero-vφ extrapolation inside 0.3 au with a power-law continuation fitted to the resolved profiles at r >= 0.3 au; (2) evaluate Rcrit and the resulting seed mass at tau = 20, 22.8, and 25 kyr to test sensitivity to clump migration; (3) apply a +/-20% perturbation to the specific angular momentum J to mimic tidal transport. If the seed mass changes by more than a factor of about 2, or if Rcrit moves outside the resolved region, the quoted masses are artifacts of the extrapolation. Ideally, also run a 3D radiation-hydrodynamic simulation of the isolated clump with H2 dissociation to confirm whether collapse to a protoplanet occurs before tidal disruption.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The quantitative seed masses in Table 3 are not direct simulation outputs. Section 3.3 estimates them by computing centrifugal radii Rcf = |J|^2/[G M_c(r')], using azimuthally averaged clump profiles whose innermost resolved point is r ≈ 0.3 au. The critical radius Rcrit = 0.18 au that separates the protoplanet seed from the circumplanetary disk/envelope lies entirely inside this unresolved region. Inside 0.3 au, the authors assume a flat surface-density plateau and vφ = 0 at r = 0, and J is assumed to be strictly conserved during the second collapse. The calculation also assumes Rp.p. = 5 RJup; varying this radius alone gives the quoted 0.25–1.6 MJup range, but no sensitivity test is presented for the profile extrapolation or for angular-momentum loss/gain from tidal torques during the ~7 kyr between the first crossing of 2000 K and tidal dispersal. Since nearly all of the claimed seed mass comes from material inside 0.18 au, the quantitative masses are not independently supported by the simulation. The weaker, qualitative claim that a protoplanet may form before tidal dispersal could still be correct, but the stated masses and their timing require a collapse model or a high-resolution check.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents 2D thin-disk simulations of gravitationally unstable protoplanetary disks with a two-component dust model (small sub-micron grains and grown dust with a variable maximum radius) using the FEOSAD code. Two prestellar core collapse models are evolved for about 0.6 Myr. The authors find that the disks are highly time-variable, with spiral arms, dusty rings, and gravitationally bound clumps that form, migrate inward, lose gas through tidal torques, and disperse. Inside the clumps, small dust is efficiently converted to grown dust, which drifts inward and forms compact central condensations of 70–100 Earth masses. Before tidal dispersal, the central gas temperature exceeds 2000 K in both tracked clumps, and the authors argue that a second collapse could form protoplanets at tens of au. Using azimuthally averaged clump profiles near the time of the 2000 K crossing, they estimate protoplanetary seed masses of 0.25–1.6 Jupiter masses of gas and 1.0–5.5 Earth masses of dust, and they connect tidally disrupted clumps to the formation of dusty rings. The paper closes with a caveat list in Section 3.4.","tokens_in":18024,"tokens_out":5501,"duration_ms":56424,"significance":"If the quantitative planet-formation claim holds, the paper would provide a concrete pathway for forming metal-rich giant planets at tens of au via gravitational instability, linking clump migration, dust growth, and dusty ring formation. The simulations are forward models with no fitting to the derived dust or planet properties, and the dust dynamics inside migrating clumps are analyzed in unusual detail for a thin-disk code. The authors are also explicit about many limitations in Section 3.4. The main weakness is that the central quantitative result—the protoplanet seed masses—is not a direct simulation output but a post-processing estimate based on extrapolated inner profiles and an assumed conservation law, so the strength of the paper's headline claim currently exceeds what the numerics can support.","major_comments":[{"comment":"The seed masses in Table 3 are not direct outputs of the simulation. The azimuthally averaged clump profiles have their innermost resolved point at r ≈ 0.3 au, stated in Section 3.3, while the critical radius Rcrit = 0.18 au that separates the protoplanet seed from the circumplanetary disk lies inside this unresolved region. The profiles are extended to r = 0 by assuming vφ = 0 at the center and constant surface-density plateaus, and Eq. (14) then evaluates Rcf = |J|^2/[G Mc(r')] using these extrapolated profiles. Because almost all of the claimed seed mass is interior to 0.18 au, the quantitative masses in Table 3 are determined primarily by the assumed extrapolation rather than by the simulation. At minimum, the paper should report how the seed masses change under alternative plausible inner extrapolations, or present a higher-resolution test of the inner clump structure.","section":"Section 3.3, Figure 10, Eq. (14)"},{"comment":"The timing argument for the second collapse is not fully supported. In model 1 the central temperature reaches 2000 K at τ = 22.8 kyr and the clump disperses after τ = 30 kyr, leaving roughly 7 kyr for the second collapse to occur; the calculation assumes that the specific angular momentum J is strictly conserved during this interval, with no account taken of tidal torques from the central star or the surrounding disk. Since Rcf scales as J^2, even a modest change in J during this window would shift Rcrit and the resulting seed mass. The authors should either justify the conservation assumption quantitatively or show that the conclusion is robust to plausible angular-momentum loss during the pre-dispersal phase.","section":"Section 3.3, Figures 8 and 9"},{"comment":"The quoted range 0.25–1.6 MJup and 1.0–5.5 M⊕ is presented as the spread of possible seed masses, but the only parameter varied in the calculation is the assumed protoplanet radius Rp.p. (2.5, 5, and 10 RJup). No sensitivity test is given for the plateau extrapolation, for the assumed flat surface-density profile, or for the conservation of J, so the range underestimates the model uncertainty. The abstract and conclusions should present these values as illustrative estimates tied to the adopted assumptions rather than as robust limits.","section":"Section 3.3, Table 3"}],"minor_comments":[{"comment":"The word 'protoplantery' in the last paragraph of Section 3.3 should be 'protoplanetary'.","section":"Section 3.3"},{"comment":"In the final paragraph, 'circumlanetary disks' should be 'circumplanetary disks'.","section":"Section 3.4"},{"comment":"The superscript n appearing on Σn d,tot in Eq. (4) is not defined; please clarify the discretization notation or remove the superscript.","section":"Section 2.2, Eq. (4)"},{"comment":"The time–space diagrams would be easier to read if color bars were provided for each panel, since the text refers to 'high surface density speckles' and 'sharp horizontal spikes' that depend on the color scale.","section":"Figure 3"},{"comment":"In the discussion of the re-formed dusty ring after clump dispersal, the text compares the ring mass of about 80 M⊕ with the clump dust mass of 88.3 M⊕; a direct statement of the expected mass conservation would help the reader judge whether the difference is significant.","section":"Section 3.1"}],"recommendation":"major_revision","confidential_remarks":"This is a solid numerical study that fits the journal's scope. My main concern is that the quantitative seed-mass claim in the abstract and conclusions is not a simulation result but an estimate based on assumed inner profiles and angular-momentum conservation; a revision should either provide a high-resolution check or soften the claim so that the quantitative masses are not presented as the paper's central result."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"This paper is worth your time, but keep the quantitative planet-seed masses in perspective. The genuinely new and solid part is the dust evolution inside gravitationally bound clumps: two thin-disk FEOSAD simulations with sub-au resolution show dust growing to decimeters, the dust-to-gas ratio swinging by factors of ten, and grown dust drifting inward to form a compact 70–100 Earth-mass condensation. Those are forward simulation results, not fitted to a target, and the clump tracking is careful. The finding that dust is retained while gas is tidally stripped is a plausible mechanism for metal-rich giants, and the re-formed dusty rings after clump dispersal are an interesting link to observed ALMA structures, though the match to any specific disk is not established.\n\nThe soft spot is Section 3.3. The seed masses of 0.25–1.6 Jupiter masses and 1–5.5 Earth masses come from an extrapolation inside the innermost resolved radius of about 0.3 au, and the critical radius of 0.18 au sits entirely inside that unresolved region. The calculation assumes specific angular momentum is conserved through the second collapse and that the profile flattens smoothly to the center; the quoted mass range only varies the assumed protoplanet radius, not these structural assumptions. Tidal torques during the roughly 7 kyr between crossing 2000 K and dispersal could plausibly change the answer by factors of a few. That said, the authors are transparent about the resolution limit and explicitly call the masses estimates. The weaker qualitative claim—that a clump can form a protoplanet seed before tidally dispersing—does not depend on the exact numbers and is consistent with the simulation showing central temperatures above the H2 dissociation threshold.\n\nOther limitations are minor in this context: only two models are run, there is no convergence study, and the code is not public, but the companion paper documents the method and the two models do agree qualitatively.\n\nThis paper is for people working on gravitational instability planet formation, dust evolution in disks, and ALMA ring interpretations. It deserves a serious referee. My recommendation would be to send it to peer review and, in revision, ask for a sensitivity test of the seed masses to the interior extrapolation and to angular momentum loss, or for the authors to soften the quantitative claims to order-of-magnitude. As is, the headline numbers overstate what the simulation shows, but the underlying dust physics is a genuine step forward.","headline":"Solid dust-growth results in fragmenting disks; the protoplanet seed masses are clearly extrapolated and should not be read as simulation outputs.","tokens_in":18588,"tokens_out":1551,"would_cite":true,"duration_ms":26805,"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":"The paper argues that giant planets can form inside the inward-migrating clumps of a gravitationally unstable disk before tidal torques destroy the clumps, yielding seeds of 0.25–1.6 Jupiter masses of gas and 1.0–5.5 Earth masses of dust.","keywords":["protoplanetary disks","gravitational instability","disk fragmentation","dust growth","dust-to-gas ratio","giant planet formation","clump migration","tidal downsizing"],"falsifier":"A numerical experiment that resolves the second collapse, starting from one of the clump profiles shown at the 2000 K moment and actually following the region inside 0.3 au to stellar densities, would settle the matter: if the clump loses its inner material to tidal torques before a bound seed of at least a few tenths of a Jupiter mass forms, or if angular momentum redistribution prevents the centrifugal radius from shrinking below the protoplanet radius, the central claim fails.","tokens_in":17589,"feed_emoji":"🪐","tokens_out":9954,"duration_ms":102926,"temperature":0.7,"pith_summary":"This paper argues that giant planet formation can happen inside the dense gaseous clumps created when a young protoplanetary disk is strongly gravitationally unstable. Using high-resolution two-dimensional thin-disk simulations that track two dust populations, the authors find that dust is efficiently concentrated and grown inside migrating clumps, forming compact central condensations of 70–100 Earth masses. Before tidal torques tear the clumps apart, their centers heat above 2000 K, the temperature at which molecular hydrogen dissociates and a second collapse begins; the authors estimate that this collapse would produce protoplanet seeds of 0.25–1.6 Jupiter masses of gas and 1.0–5.5 Earth masses of dust at orbital distances of tens of au. If true, disk fragmentation offers a direct route to wide-orbit giant planets that are enriched in heavy elements, and the same clump dynamics produces transient dusty rings resembling structures seen in young disks.","feed_headline":"Dust-rich clumps can collapse into giant planets at tens of au","feed_subtitle":"Simulations show seeds of 0.25–1.6 Jupiter masses in gas and 1.0–5.5 Earth masses of dust before clumps break apart.","key_machinery":"The load-bearing object is the clump's internal structure as computed by a sub-au-resolution thin-disk hydrodynamics code with two dust species: small dust grains that are tightly coupled to gas, and grown dust with a variable maximum radius whose dynamics is set by gas drag, dust self-gravity, and the total gravitational potential. A dust-growth scheme converts small to grown dust and limits growth by a fragmentation barrier. Within a clump, gas friction and the clump's non-Keplerian velocity field drive grown dust inward, building a central condensation. The predicted protoplanet seeds come from applying the centrifugal radius formula $R_{\\rm cf}=|J|^2/[G M_c(r')]$ to the clump's surface-density and angular-velocity profiles at the moment its center reaches 2000 K, assuming angular momentum is conserved and that material with $R_{\\rm cf}$ smaller than the adopted protoplanet radius forms the seed while the rest forms a circumplanetary disk or envelope.","core_discovery":"The central claim is that, in a gravitationally fragmenting disk, dust does not simply trace gas. Over roughly 30 000 years of clump evolution, sub-micron dust is converted into grown dust with radii of several decimeters; grown dust drifts inward under gas friction and accumulates at the clump center, while gas in the outer clump is progressively stripped by tidal torques from the star. The result is a compact, dust-enriched condensation whose interior dust-to-gas ratio can rise above the canonical 1:100 by up to a factor of five. In the two simulated clumps this process continues until the central temperature crosses 2000 K, the dissociation threshold of molecular hydrogen. Assuming angular momentum conservation during the subsequent second collapse, the authors compute centrifugal radii and, adopting a protoplanet radius of 2.5–10 Jupiter radii, derive seed masses of 0.25–1.62 Jupiter masses in gas and 1.0–5.5 Earth masses in dust. They argue these seeds form before the clumps disperse through tidal action, and that later accretion from massive metal-rich disks or envelopes can raise the final masses well above the seed values.","pith_inferences":["If this picture holds, direct imaging surveys of wide-orbit giant planets around young stars should find a population with heavy-element masses in the roughly 1–10 Earth-mass range and envelopes enriched above stellar metallicity, at orbital distances where core accretion struggles to assemble cores fast enough.","In this picture, the solid core of a giant planet may be assembled before the gas envelope by in-clump dust drift rather than by later planetesimal accretion, which would explain how gravitational-instability planets avoid being coreless.","A testable extension is that transient dusty rings from clump dispersal should be more common in disks younger than about 0.5 Myr and should correlate with clump migration events, so surveys of the youngest embedded disks could look for rings that later disappear.","The quoted seed-mass range rests on conservation of angular momentum during collapse; adding magnetic fields or turbulence in a future model would redistribute angular momentum and likely shrink the seeds, so the numbers bracket an idealized, non-magnetic collapse."],"forward_implications":["Gravitational instability of a massive young disk can end in giant planets at tens of au, not just brown dwarfs or scattered clumps, because tidal stripping reduces the clump to roughly a Jupiter mass before the center collapses.","Planets formed this way start metal-rich: the dust-to-gas ratio in the protoplanetary seeds is 0.011–0.013, above the canonical 0.01, and the dust mass of the seed is 1.0–5.5 Earth masses.","Dusty rings at several tens of au form when clumps are tidally destroyed and are later disturbed by other migrating clumps; these transient rings are plausible analogs of ring-like structures observed around the youngest, most massive protoplanetary disks.","Final planet masses are not fixed by the seed: accretion from a surrounding circumplanetary disk or envelope can increase the mass substantially, although tidal stripping during migration may remove much of that envelope, keeping the planet in the giant-planet regime.","The dust accumulated in clumps before collapse, with central condensations of 70–100 Earth masses, provides a head start for the heavy-element content of the eventual planet."],"supporting_citations":[{"why":"Companion study establishing clump inward migration and tidal stripping, whose late-stage clump properties this paper analyzes with dust.","marker":"Vorobyov & Elbakyan 2018"},{"why":"Supplies the numerical code and the two-population dust growth scheme, dust fragmentation barrier, and equations used throughout.","marker":"Vorobyov et al. 2018"},{"why":"Motivates the use of the inner half of the Hill radius as the region where gas remains bound to a clump or protoplanet.","marker":"Nayakshin 2017a"},{"why":"Defines the second collapse process invoked once the clump center exceeds 2000 K, the step that turns clumps into protoplanet seeds.","marker":"Masunaga & Inutsuka 2000"},{"why":"Provides the observed giant-planet mass-metallicity relation used to argue that the predicted metal-rich seeds are consistent with data.","marker":"Thorngren et al. 2016"},{"why":"Catalog of ring-like structures in protoplanetary disks used for comparison with the transient dusty rings generated by clump dispersal.","marker":"van der Marel et al. 2019"}],"fun_headline_variants":["Dust clumps grow meter-size grains and seed giant planets","Planet seeds form at tens of au from dust-rich clumps","Clumpy disk dust gathers into massive cores before breakup","Gravity-driven dust clumps collapse into protoplanet seeds"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The argument assumes, as in Section 3.3, that a clump whose center exceeds 2000 K collapses to a protoplanet quickly enough to beat tidal disruption, and that specific angular momentum is conserved during that collapse; the simulation does not resolve the collapse, and inside 0.3 au the density and rotation profiles are extrapolated to a plateau and to zero rotation at the center.","fun_headline_variants_meta":{"raw":{"variants":["Dust clumps grow meter-size grains and seed giant planets","Planet seeds form at tens of au from dust-rich clumps","Clumpy disk dust gathers into massive cores before breakup","Gravity-driven dust clumps collapse into protoplanet seeds"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000479,"raw_usage":{"total_tokens":2480,"prompt_tokens":1163,"completion_tokens":1317,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":779,"completion_tokens_details":{"reasoning_tokens":1247}},"tokens_in":779,"tokens_out":1317,"duration_ms":10521,"temperature":1.0,"reasoning_tokens":1247,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T10:38:30.717898+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A numerical experiment that resolves the second collapse, starting from one of the clump profiles shown at the 2000 K moment and actually following the region inside 0.3 au to stellar densities, would settle the matter: if the clump loses its inner material to tidal torques before a bound seed of at least a few tenths of a Jupiter mass forms, or if angular momentum redistribution prevents the centrifugal radius from shrinking below the protoplanet radius, the central claim fails.","supporting_citations":[{"cited_title":"& Inutsuka, S.-i","cited_arxiv_id":null,"evidence_quote":"Defines the second collapse process invoked once the clump center exceeds 2000 K, the step that turns clumps into protoplanet seeds."}],"review_version":1}