{"id":"e5ecca6a-bbd7-4b5c-90db-e7563a8169bb","arxiv_id":"2501.18424","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"Global dust-fluid simulations of the streaming instability find clumps reaching only about 30% of the Hill density after 160 orbits, with gravitational collapse estimated to need roughly 480 to 1000 orbits.","lead":"Planet seeds may begin as dust clumps dense enough to collapse under their own gravity. New global dusty-disk simulations show these clumps reach only 30% of the collapse density after 160 orbits, making streaming-instability planet formation slower than local models suggested.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 480–1000 orbit collapse time is an exponential extrapolation the authors themselves say has no theoretical guarantee; until tested, the headline timescale is unsupported.","rationale":"In good faith, the paper's actual numerical result, that the maximum dust density reaches about 30% of the local Hill density after 160 orbits in a Z=0.02 dust-fluid simulation, is a concrete and potentially reproducible output. That part is credible and is not my objection. What makes the central claim fragile is that every statement about when gravitational collapse happens goes through an exponential model that the authors explicitly disclaim. The reader's weakest assumption, the 2D axisymmetric geometry, is a legitimate reservation, but it is not the most load-bearing concern: 2D (R,z) streaming-instability simulations have precedent, and the 160-orbit density measurement is not invalidated by geometry alone. The extrapolation, by contrast, is load-bearing for the paper's main quantitative conclusion and is admittedly ungrounded. The proposed long run would settle whether the growth law is exponential by directly measuring the crossing time; a cheaper internal consistency check on the existing time series would provide a partial test. The verdict remains CONDITIONAL: the paper should be published only if the extrapolation is either replaced by direct long-time integration or presented with the explicit caveat that the timescale is a speculative estimate. Hence no change from the reader's conditional verdict.","tokens_in":10806,"tokens_out":6485,"duration_ms":62707,"concrete_test":"Run the Z=0.02 simulation to at least 480 orbits with the same grid, boundaries, and parameters. If the measured maximum dust density at 480 orbits lies below the exponential extrapolation by more than 30%, or if the actual Hill-density crossing time differs from 480 orbits by more than 30%, the extrapolation-based headline timescale is not supported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The load-bearing step is not the 160-orbit measurement itself but the extrapolation from it. Section 3.1 and Fig. 3 show the maximum dust density reaching roughly 30% of the Hill density at 160 orbits; the paper then states that reaching the Hill density would take roughly 480 orbits by extrapolating this exponential behavior. Section 4 explicitly concedes that no theory has predicted yet how the maximum local dust density evolves in time and that there is no guarantee that the profile will continue to be of exponential shape. The fitted slope is not anchored to a single physical clump because the maximum density is related to many clumps that alternate in having the highest density, so the e-folding time is a fit to the envelope of competing objects. All downstream conclusions, including the 480 versus 1000 orbit collapse times, the 1 au inward drift limit, and the claim that clumping is less efficient than previously thought, inherit this unvalidated functional form. If the growth rate slows, the true collapse time is longer; if runaway concentration occurs, it is shorter. The central quantitative claim is therefore not yet established.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This Letter reports global 2D axisymmetric (R,z) FARGO3D simulations of the streaming instability in a stratified protoplanetary disk at 9.3–10.7 au, treating dust as a pressureless fluid with aerodynamic backreaction and dust diffusion. For Z = Σd/Σg = 0.01 the authors confirm weak clumping; for Z = 0.02 dense clumps form within about 20 orbits. The measured maximum dust density reaches roughly 30% of the local Hill density after 160 orbits, and the authors extrapolate this growth exponentially to estimate that Hill density is reached after about 480 orbits, or about 1000 orbits for less massive or compact disks after rescaling by a factor of 6.5. They then compare five observed disk midplane density profiles to the Hill density, estimate a clump separation of about 0.03 au at 10 au, and conclude that such structures are unresolved by ALMA and ngVLA, predicting efficient planetesimal formation outside 10 au.","tokens_in":11064,"tokens_out":14132,"duration_ms":134821,"significance":"If the measured clump densities and the extrapolated collapse times are confirmed, the paper is significant: it provides one of the first global, stratified dust-fluid tests of streaming-instability clumping and argues that SI-driven planetesimal formation is less efficient than inferred from local 3D simulations. The study's strengths include the use of a public code (FARGO3D), a direct comparison of the simulated maximum density to the Hill density, an honest caveat in Section 4 about the lack of a theory for the maximum-density evolution, and a falsifiable observability prediction for ALMA/ngVLA. However, the central quantitative claim rests on an unvalidated exponential extrapolation, and the numerical method lacks a convergence test, a specification of the dust-diffusion treatment, and a 3D validation; these issues currently prevent the timescales from being accepted as established.","major_comments":[{"comment":"The 480-orbit collapse time is obtained by extrapolating an exponential fit to the envelope of the maximum dust density, and Section 4 explicitly concedes that no theory predicts how the maximum local dust density evolves and that there is no guarantee the profile remains exponential. Because the maximum is set by different clumps at different times, the fitted e-folding time is not tied to a single physical object, and no uncertainty is assigned to the extrapolation. The 1000-orbit estimate for rescaled, less massive disks in Section 4 inherits this uncertainty, as do the derived inward-drift limits and the conclusion that SI clumping is less efficient than previously thought. Please either run the simulation toward the Hill-density crossing, provide a physically motivated model for the envelope, or present 480–1000 orbits explicitly as an illustrative extrapolation rather than a quantitative prediction.","section":"Section 3.1 and Fig. 3; Section 4"},{"comment":"The pressureless-dust-fluid method is used with the FARGO3D dust module, but the dust-diffusion prescription and its coefficient are never stated. In a pressureless fluid, peak densities are controlled by physical or numerical diffusion, and without specifying the diffusion treatment the reported 30% of Hill density after 160 orbits is not a fully determined simulation result. No resolution-convergence study is presented either. A convergence sequence (for example, at half and double the current resolution) and a clear statement of the diffusion coefficient (including an explicit zero if no diffusion is applied) are required to support the central density measurement.","section":"Section 2 and Table 1"},{"comment":"The simulations are 2D axisymmetric in (R,z), and the meridional domain is only ±0.014 rad about the midplane, corresponding to about ±0.14 au (roughly ±0.2 gas scale heights) at 10 au. The setup therefore excludes azimuthal and non-axisymmetric streaming-instability modes and contains only a weak vertical gas stratification, yet the clumping statistics are compared with 3D local shearing-box calculations. Please justify the vertical domain size and show that the axisymmetric dust-fluid model reproduces the clumping behavior of a 3D calculation (for example, a local shearing-box run with the same Z, St, and Π) before the measured densities and extrapolated timescales are transferred to real disks.","section":"Section 2"},{"comment":"The dust-refilling prescription resets the dust density to its initial value in the outer damping zone whenever it falls below 25% of the initial value, and Eq. (B.5) continuously damps the gas density toward the initial state over the entire domain. This is a persistent artificial mass source at the outer boundary that can feed the inwardly drifting clumps and directly affect the growth of ρdust,max. The sensitivity to the 25% threshold, the refilling location, and the damping timescale is not tested, and the text's own reference to possible outer-boundary artifacts in the choice of the 9.4–10.4 au window makes this a live concern.","section":"Appendix B"}],"minor_comments":[{"comment":"The clump drift velocity is said to be 'fitted by eye'; please provide a quantitative fit with an uncertainty estimate.","section":"Section 3.1"},{"comment":"The sentence 'The chosen domain is meant not to exclude clumps that are an artifact from the outer boundary' is ambiguous; if the intent is to exclude such clumps, please reword accordingly.","section":"Section 3.1"},{"comment":"The final sentence of Section 4 ends with the dangling word 'resolve' after mentioning the method of Scardoni et al. (2024); the sentence is incomplete and should be finished.","section":"Section 4"},{"comment":"The phrase 'using disk parameters from GM Aur, HD163296, IM Lup, MWC 480, and TW Hya' overstates what is done: Fig. 4 compares analytic midplane density profiles to the Hill density, while the simulation uses one canonical profile at 10 au. Please rephrase to 'consistent with the midplane densities of...'.","section":"Abstract and Conclusions"}],"recommendation":"major_revision","confidential_remarks":"The requested additions—specifying the diffusion treatment, a resolution study, a 3D validation test, and an honest reframing of the exponential extrapolation—are substantial. If the journal wishes to keep this as a Letter, I would accept a revised version only if the 480–1000 orbit timescale is explicitly labeled as an illustrative extrapolation and the method-parameter omissions are fixed. Otherwise, the work is better suited to a full-length paper with the additional numerical tests."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Two things to know. The robust new number is that in a global stratified 2D (R,z) disk with Z=0.02, the streaming instability's densest clumps reach only ~30% of the local Hill density after 160 orbits. The headline '480–1000 orbits to collapse' is an exponential fit to the envelope of competing clumps, not a measurement, and the authors say so plainly in Section 4. That caveat does not sink the paper, but the abstract states 'clumping is less efficient than previously thought' more confidently than the evidence supports.\n\nWhat is actually new: this is the first global stratified disk simulation of streaming instability with dust as a pressureless fluid. The setup is described carefully, the disk parameters are tied to observed systems (GM Aur, HD163296, IM Lup, MWC 480, TW Hya), and the clump separation (0.03 au) and optical depth estimates give a concrete, testable observability statement. The paper also does something useful in comparing midplane gas density profiles to the Hill density and identifying outer-disk sweet spots.\n\nThe soft spots are real but mostly acknowledged. The 2D axisymmetric geometry is the biggest one: a 3D run, or at least a resolution study, would be needed to say whether the measured growth rates transfer to real disks. The pressureless-fluid approximation is standard for St<0.1, but its behavior at the very high clump densities seen here is not tested. The dust refilling in the outer damping zone (reset to 25% of initial when below threshold) is a potential mass source and is not quantified. And the exponential extrapolation is explicitly 'no guarantee,' so anyone citing the 480-orbit number should treat it as provisional. The citation pattern is fine; the Zcrit conversion is used for context, not fitted.\n\nThis is a solid Letter for the streaming-instability community. It will be most useful as a motivating comparison for future global and 3D studies. I would cite it as 'first global dust-fluid SI simulation shows slower clumping growth than local boxes,' not as evidence for a specific collapse time. It deserves a serious referee: the reviewer should push for a resolution test, a 3D sanity check if available, and a quantification of injected dust mass. Send it to review.","headline":"Useful first global dust-fluid SI results with an honest extrapolation that gets slightly oversold in the abstract; the 30%-of-Hill-density measurement is the robust part.","tokens_in":11583,"tokens_out":2884,"would_cite":true,"duration_ms":26557,"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":"Global simulations of the streaming instability show dust clumps reaching only 30% of the Hill density after 160 orbits, so gravitational collapse would need roughly 480 to 1000 orbits.","keywords":["streaming instability","protoplanetary disks","dust clumping","planetesimal formation","Hill density","dust-fluid model","global disk simulations"],"falsifier":"Run the same Z=0.02 setup in a full 3D stratified model, or at least at double resolution in 2D, and track the maximum dust density relative to $\\rho_{\\mathrm{Hill}}$ for 500 orbits. If the ratio crosses 1 within about 300 orbits, the extrapolated timescale is too slow; if it stays below 0.3 after 160 orbits, the slow-collapse conclusion is supported.","tokens_in":10608,"feed_emoji":"🪐","tokens_out":8578,"duration_ms":71402,"temperature":0.7,"pith_summary":"Using a global 2D stratified disk model rather than a local box, this paper asks whether the streaming instability alone can pack dust into clumps dense enough to collapse into planetesimals. With a dust-to-gas mass ratio of Z=0.02, dense clumps appear within 20 orbits, but after 160 orbits the densest clump reaches only about 30% of the local Hill density, the threshold for gravitational collapse. The authors extrapolate the roughly exponential growth of the maximum density and estimate that reaching the Hill density would take about 480 orbits in a massive disk, and up to about 1000 orbits in a less massive or compact disk. If right, the streaming instability is slower and less efficient at building collapse-ready clumps than earlier local simulations indicated, and the clumps it forms are too closely spaced for current observatories to resolve.","feed_headline":"Dust clumps hit only 30% of collapse density by orbit 160","feed_subtitle":"Global disk models put streaming-instability collapse at roughly 480 to 1000 orbits, slower than local-box results.","key_machinery":"The machinery is a two-fluid hydrodynamic model in which the gas is a stratified disk and the dust is a pressureless fluid with aerodynamic back-reaction and dust diffusion, run at a Stokes number of St=0.01 and a pressure-gradient parameter of Π=0.07. The streaming instability is the aerodynamic drag instability that concentrates dust when dust and gas densities are comparable. The paper measures the maximum dust density in a 2D axisymmetric (R,z) disk and compares it with the local Hill density, $\\rho_{\\mathrm{Hill}} = 9 M_* / (4\\pi R^3)$, which sets the threshold for gravitational collapse. That ratio, maximum clump density over local Hill density, is the quantity that turns a clumping simulation into a statement about when planetesimal formation can begin.","core_discovery":"The central discovery is that in a realistic global disk, clumping by the streaming instability is real but limited: at a global dust-to-gas ratio Z=0.02, dust clumps form throughout the domain within about 20-25 orbits and reach roughly 30% of the local Hill density after 160 orbits at 10 au. The maximum dust density then appears to grow exponentially once the nonlinear phase sets in, and the paper's extrapolation places gravitational collapse at about 480 orbits, with clumps drifting only about one au inward during that time; for a less massive or compact disk the number rises to about 1000 orbits. A companion run with Z=0.01 shows no strong clumping, matching earlier results. The clumps drift inward more slowly than the background dust because their high dust-to-gas ratio shields them from aerodynamic drag, and their average separation at 10 au is about 0.03 au.","pith_inferences":["The 480- and 1000-orbit numbers are extrapolations, not simulated outcomes; if the exponential growth stalls, the true collapse times are longer, while if axisymmetry suppresses 3D modes, they could be shorter.","A testable next step is to let clumps continue for hundreds more orbits in a disk with a wider radial domain to check whether the traffic-jam self-shielding keeps drift this slow while density approaches the Hill value.","The optical-depth calculation suggests that unresolved clumps could still leave a detectable signature as optically thick spots in a mostly optically thin disk at long wavelengths; searching for such contrast in inclined disks is a way to constrain the instability observationally.","If real disks carry even weak turbulence, clump lifetimes and densities may drop below this no-turbulence estimate, pushing planetesimal formation further out in radius or to higher metallicities."],"forward_implications":["Streaming-instability collapse in a massive disk would take roughly 480 orbits, corresponding to tens to hundreds of thousands of years; in less massive or compact disks the estimate doubles to roughly 1000 orbits.","Collapse is more likely outside about 10 au, where the Hill density falls with radius faster than the gas midplane density.","The dense clumps formed at Z=0.02 are spaced by about 0.03 au at 10 au, far too close to be resolved by current or planned millimeter observatories.","A disk with Z=0.01 does not produce strong clumps, placing the clumping threshold for this global setup between 1% and 2% dust-to-gas ratio.","The clumps slow their inward drift as they densify, so the one to two au of drift before collapse is probably an upper limit rather than a lower bound."],"supporting_citations":[{"why":"Introduces the streaming instability, the dust-gas drag mechanism whose clumping this paper tests in a global disk.","marker":"Youdin & Goodman (2005)"},{"why":"First showed strong dust clumping from the instability in a stratified disk, the result this paper revisits with global geometry.","marker":"Johansen et al. (2009)"},{"why":"Supplies the global 2D stratified disk setup, boundary conditions, and numerical methods the simulations are based on.","marker":"Flock & Mignone (2021)"},{"why":"Provides the pressureless-dust-fluid treatment with dust diffusion that the simulations use.","marker":"Weber et al. (2019)"},{"why":"Describes the two-fluid gas-plus-dust implementation, including drag back-reaction, used to run the models.","marker":"Benítez-Llambay et al. (2019)"},{"why":"Provides the local shearing-box comparison and the Zcrit scaling that places the Z=0.02 run in the strong-clumping region.","marker":"Li & Youdin (2021)"},{"why":"Higher-resolution local simulations that lower the critical Z, used to argue the Z=0.02 run is in the clumping regime.","marker":"Lim et al. (2024b)"},{"why":"Defines the Hill density used as the gravitational-collapse threshold for the clumps.","marker":"Klahr & Schreiber (2020)"},{"why":"Supplies the observed surface-density and scale-height profiles for four massive disks compared with the Hill density.","marker":"Martire et al. (2024)"},{"why":"Supplies the TW Hya surface-density profile and scale height used as the fifth comparison disk.","marker":"Yoshida et al. (2022)"}],"fun_headline_variants":["Global dust clumps stall at 30% of collapse density","Streaming instability collapse slower in realistic disks","Clumps form, but collapse awaits 480 orbits in disks","Dust clumps reach only 30% of Hill density in global runs","Collapse delayed: dust clumps drift inward without falling"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The 480-orbit estimate assumes both that the 2D axisymmetric pressureless-fluid setup captures the same streaming-instability clumping as fully 3D simulations, and that the maximum clump density keeps growing exponentially beyond 160 orbits, a trend the paper itself says no theory guarantees.","fun_headline_variants_meta":{"raw":{"variants":["Global dust clumps stall at 30% of collapse density","Streaming instability collapse slower in realistic disks","Clumps form, but collapse awaits 480 orbits in disks","Dust clumps reach only 30% of Hill density in global runs","Collapse delayed: dust clumps drift inward without falling"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000261,"raw_usage":{"total_tokens":1654,"prompt_tokens":1065,"completion_tokens":589,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":681,"completion_tokens_details":{"reasoning_tokens":505}},"tokens_in":681,"tokens_out":589,"duration_ms":5426,"temperature":1.0,"reasoning_tokens":505,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-09T23:32:22.768785+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run the same Z=0.02 setup in a full 3D stratified model, or at least at double resolution in 2D, and track the maximum dust density relative to $\\rho_{\\mathrm{Hill}}$ for 500 orbits. If the ratio crosses 1 within about 300 orbits, the extrapolated timescale is too slow; if it stays below 0.3 after 160 orbits, the slow-collapse conclusion is supported.","supporting_citations":[{"cited_title":"& Mignone , A","cited_arxiv_id":null,"evidence_quote":"Supplies the global 2D stratified disk setup, boundary conditions, and numerical methods the simulations are based on."},{"cited_title":"& Schreiber , A","cited_arxiv_id":null,"evidence_quote":"Defines the Hill density used as the gravitational-collapse threshold for the clumps."},{"cited_title":"2024, , 686, A9","cited_arxiv_id":null,"evidence_quote":"Supplies the observed surface-density and scale-height profiles for four massive disks compared with the Hill density."},{"cited_title":"C., Nomura , H., Tsukagoshi , T., Furuya , K., & Ueda , T","cited_arxiv_id":null,"evidence_quote":"Supplies the TW Hya surface-density profile and scale height used as the fifth comparison disk."}],"review_version":1}