{"id":"47f70c0d-4557-473a-8824-049a259b627f","arxiv_id":"2501.09792","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"Convective overstability zonal flows in protoplanetary disks trap dust only weakly, and dust feedback suppresses the flows when the dust-to-gas ratio reaches about 0.1.","lead":"Simulations show that pressure bumps created by the convective overstability in protoplanetary disks trap dust only weakly, with dust-to-gas ratios capped near 0.1 to 0.6. Dust feedback can even destroy the bumps, so this mechanism may not build planetesimals, redirecting attention to vortices and stratified models.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"No-feedback control run may be contaminated or the late-time gas-activity drop may be intrinsic to the pure COS, confusing the attribution of zonal-flow weakening to dust feedback.","rationale":"The paper's headline claims are that dust feedback impedes COS zonal flows at ε∼0.1 and that dust concentration by such zonal flows is limited to ε≲1. The strongest direct evidence for the 'impedes' claim comes from the ε0=0.1 run, which shows no zonal flows at all, and from the fiducial run, whose zonal flows decay once local ε reaches ∼0.1. The no-feedback run is the designed control for the second piece of evidence. If that control also exhibits a late-time drop in gas activity, the control does not cleanly separate the dust-feedback effect from intrinsic COS dynamics. This is a more immediate and internal concern than the geometric idealizations (axisymmetry, no stratification) emphasized by the reader, because it affects the causal interpretation of the paper's own simulations. The concrete test of running a true pure-gas simulation will settle whether the orange curve in Fig. 3 is a clean control; this test directly addresses the ambiguity. If the test shows the no-feedback run is clean and the pure-gas COS intrinsically decays at 900P, then §5.1's attribution of the fiducial weakening to feedback is over-stated, though the ε0=0.1 run still supports the suppression-of-formation part of the conclusion. If the test shows the no-feedback run is contaminated, the control evidence for the feedback-driven weakening would need to be redone. Either way, the central conclusion that COS zonal flows do not concentrate dust beyond ε≲1 is likely to survive, but the paper's mechanistic interpretation and the strength of the 'feedback impedes' claim would need revision. Given that the reader's verdict is already CONDITIONAL, this additional internal concern does not move the verdict; it sharpens the conditions under which the paper should be accepted and reproduced.","tokens_in":22483,"tokens_out":23851,"duration_ms":249557,"concrete_test":"Run a pure-gas version of the fiducial setup (no dust: set ϵ=0 and drop Eqs. 16–17) at Nx×Nz=2048×1024, and record max|δvg| versus time for at least 1000 orbits. Overlay this pure-gas curve on Fig. 3 and compare with the orange 'no-feedback' curve. If the pure-gas curve remains at max|δvg|≈0.1 without dropping near 900P, then the orange run is not genuinely feedback-free and the control is invalid; if it drops identically to the orange curve, then the late-time weakening is an intrinsic pure-gas process and should not be attributed to dust feedback in the fiducial run.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 5.2 and Fig. 3 describe a run 'without feedback' in which the drag term in the gas momentum equation (14) is set to zero, so the gas should be dynamically independent of dust and should evolve exactly like a dust-free COS run. Yet the text states that this no-feedback run 'behaves similarly to the fiducial case until 900P, whence ε reaches O(0.1), and activity drops towards the ε0=0.1 run.' If feedback is truly disabled, the gas cannot respond to ε; the orange curve's drop at ~900P must therefore either be an intrinsic property of the pure-gas COS in this box (long-term modulation/decay of zonal flows) or indicate an unintended residual dust-gas coupling in the implementation. In the former case, the fiducial run's weakening at ~800P (attributed in §5.1 to 'dust feedback onto the zonal flows') is partly or wholly an intrinsic gas-phase process, and the causal role of dust feedback is not established. In the latter case, the control is invalid. The ε0=0.1 run provides independent evidence that dust feedback suppresses zonal-flow formation, but the late-time weakening of already-formed zonal flows is also a key component of the central claim, and the current paper does not resolve this ambiguity.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper studies dust dynamics in the convective overstability (COS) of protoplanetary disks using high-resolution axisymmetric, unstratified Boussinesq shearing-box simulations with a pressureless dust fluid. The authors find that COS-driven zonal flows can concentrate dust to typical factors of O(10) (with transient maxima reaching ε ~ 0.6 for a background ε0 = 0.01), that dust feedback can suppress zonal-flow formation at dust-to-gas ratios ε ~ 0.1, and that a background radial pressure gradient substantially weakens dust trapping. They interpret the feedback effect as a competition between negative gas angular momentum flux and positive dust angular momentum flux, and derive a critical dust-to-gas ratio ε ≳ N/(4St) for feedback to inhibit zonal flows. The paper concludes that COS-driven zonal flows are not directly conducive to planetesimal formation, while acknowledging that stratified and 3D simulations are needed to assess the broader picture.","tokens_in":22724,"tokens_out":5697,"duration_ms":56915,"significance":"If the results are correct, this paper provides a significant counterpoint to earlier suggestions that COS-driven structures can directly trigger planetesimal formation. The work is carefully benchmarked: the code reproduces linear COS and SI growth rates to relative errors of O(10^-4) (Appendix B, Table 1), and the resolution study in Appendix C supports convergence at Nx×Nz = 1024×512. The no-feedback control run isolates the effect of dust drag, and the analytical model for the dust angular momentum flux (Eq. 48 and Appendix D) gives a plausible mechanistic explanation. The main conclusions, however, depend on specific model restrictions—axisymmetry, no vertical gravity, and a constant heat sink—which the authors explicitly identify as limitations. These restrictions mean the astrophysical implications are provisional, but the paper's core results stand as a well-executed study of an idealized but relevant configuration.","major_comments":[{"comment":"The no-feedback control run is not a valid control for the claimed dust-feedback effect. In this run the drag term in the gas momentum equation (Eq. 14) is set to zero, so the gas should be dynamically independent of dust. Yet the text states that this run 'behaves similarly to the fiducial case until 900P, whence ε reaches O(0.1), and activity drops towards the ε0=0.1 run.' If feedback is truly disabled, the gas cannot respond to ε, so the late-time drop in the orange curve must either be an intrinsic property of the pure-gas COS in this box (long-term modulation or decay of zonal flows) or indicate an unintended residual dust-gas coupling in the implementation. In the former case, the drop in the fiducial run at ~800P, attributed in §5.1 to 'dust feedback onto the zonal flows,' is partly or wholly an intrinsic gas-phase process, and the causal role of dust feedback is not established. In the latter case, the control is invalid. The authors should compare with a genuine dust-free gas run (evolving only Eqs. 13–15) or otherwise demonstrate that the pure-gas COS does not exhibit a similar decline over the 1000-orbit timescale. This issue is load-bearing for the attribution of the late-time weakening to dust feedback, although the ε0 = 0.1 run provides independent evidence that strong dust loading can suppress zonal-flow formation.","section":"§5.2, Fig. 3"},{"comment":"The abstract states that dust densities 'increase at most by a factor of O(10)', but the fiducial run's maximum dust-to-gas ratio reaches ε ≈ 0.6 on a background of ε0 = 0.01, i.e., a factor of 60, at 800P and 850P (Fig. 7 and §5.1). The text also says 'concentration factors are typically O(10)' but then says they 'appear limited by ε = 0.6', which is a factor of 60. This is an internal inconsistency in a quantitative claim that appears in the abstract. The authors should rephrase to distinguish typical concentration factors (O(10)) from transient maxima (up to a factor of ~60), or restrict the 'at most' statement to time-averaged values.","section":"Abstract, §5.1, Fig. 7"}],"minor_comments":[{"comment":"The description of disabling dust feedback is clear, but the sentence 'Neglecting feedback is usually justified for ε≪1, but we shall find that it affects the COS even in this regime' could be made more precise: the subsequent results show that feedback affects zonal-flow formation at ε ~ 0.1, not necessarily at ε ≪ 0.01.","section":"§2.6"},{"comment":"The notation δvgx, δvgy for deviations from equilibrium is introduced in Eq. 40, but in earlier sections δ denotes Eulerian linear perturbations. To avoid confusion, the authors could use a different symbol (e.g., Δ or prime) for nonlinear deviations from the equilibrium state.","section":"§5.3"},{"comment":"The sentence 'The two epochs of rapid dust growth at 800P and 850P in the fiducial run show that feedback may temporarily boost concentrations, probably via streaming-type instabilities' is speculative: the fiducial setup has Π = 0, and the paper earlier states that the streaming instability is suppressed in that case. The local radial pressure gradients from zonal flows can indeed drive relative drift, but the connection to streaming-type instabilities should be explained or softened.","section":"§5.1"},{"comment":"The empirical fit max(ε) ≃ 1.25St + 0.01 in §6.2 is presented without error bars or a discussion of the scatter shown in Fig. 17. Reporting the goodness of fit and the range of St over which the linear relation holds would strengthen this result.","section":"§7.2.2"}],"recommendation":"major_revision","confidential_remarks":"The no-feedback control issue is the main concern: if the control run's late-time drop is due to a bug (residual coupling), the simulation code may have a defect that affects other runs; if it is intrinsic, the interpretation of the fiducial run's weakening needs revision. The authors should add a pure-gas (dust-free) run to settle this. The abstract's 'at most O(10)' versus the factor-of-60 transient is a straightforward but important correction. Given the high quality of the code validation and the analytical AMF interpretation, the paper is likely publishable after these issues are addressed."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First thing to know: this is a cleanly executed simulation study that gives a clear mechanistic answer to whether COS zonal flows can trap dust enough for planetesimal formation. The answer is no, or at least not easily: dust feedback at ε ~ 0.1 shuts off zonal flow formation, and even in the best cases dust-to-gas ratios cap out around 0.6 transiently, typically 0.2–0.3. That is a useful negative result, and the angular momentum flux (AMF) argument for why is persuasive.\n\nWhat's new: it is the first nonlinear dusty COS simulation in the Boussinesq shearing box. The linear benchmarks are excellent (relative errors ~1e-4), the resolution study shows convergence at 1024×512, and the AMF decomposition matches the geostrophic model well. The parameter survey gives a handy empirical scaling max(ε) ≈ 1.25 St + 0.01, and the Π dependence quantifies how background drift kills trapping.\n\nSoft spots: the no-feedback control run has a problem. The text says the run without feedback behaves like the fiducial until 900 orbits, when activity drops once the dust concentration reaches O(0.1). With feedback disabled, the gas cannot see dust, so either that drop is intrinsic to the pure COS in this box (undercutting the claim that dust feedback weakens already-formed zonal flows at ~800 orbits) or feedback is not fully disabled. The paper does not address this. The ε0 = 0.1 run stands on its own as evidence that feedback prevents zonal flow formation, but the late-time weakening is a weaker claim. Also, the geometry is unstratified and axisymmetric; the authors acknowledge that settling (Hd ≲ 0.06 Hg) and vortex/RWI behavior could change things. No code or data is shipped, which is annoying but not disqualifying.\n\nThe paper deserves a serious referee. It is well-executed and the central negative result is probably robust, but the control-run issue should be fixed before acceptance. This is for protoplanetary disk theorists working on dust concentration and planetesimal formation.","headline":"Well-benchmarked simulation study that limits COS zonal flows as dust traps, but a questionable control run muddies the late-time feedback story.","tokens_in":23288,"tokens_out":3198,"would_cite":false,"duration_ms":32326,"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":"This paper argues that the convective overstability's zonal flows trap dust only weakly, and that dust feedback suppresses those flows even at dust-to-gas ratios near 0.1.","keywords":["convective overstability","dust dynamics","protoplanetary disks","zonal flows","dust feedback","angular momentum flux","planetesimal formation","shearing box"],"falsifier":"Run a stratified, full-3D equivalent of the fiducial setup at the same resolution: if persistent zonal flows form at epsilon0 = 0.1 and concentrate dust to epsilon > 1, the paper's central limit is false. Alternatively, find an observed disk ring in a COS-susceptible region whose dust-to-gas ratio exceeds unity with a weak pressure perturbation.","tokens_in":1960,"feed_emoji":"🪐","tokens_out":2711,"duration_ms":83581,"temperature":0.7,"pith_summary":"This paper asks whether the convective overstability (COS), a hydrodynamic instability that operates in protoplanetary disks where the entropy decreases outward, can concentrate dust enough to seed planet formation. Using high-resolution axisymmetric Boussinesq shearing-box simulations with dust modeled as a second fluid, it finds that the zonal flows (pressure bumps) produced by the COS do trap dust, but only to about ten times the background density. The central discovery is that dust feedback suppresses the formation of these zonal flows even at dust-to-gas ratios as low as epsilon ~ 0.1, because dust drifting toward pressure maxima carries a positive angular momentum flux that offsets the negative flux that builds the zonal flow. The paper concludes that COS-driven zonal flows cannot by themselves trigger planetesimal formation, since the attainable dust-to-gas ratio stays below epsilon ~ 1.","feed_headline":"Dust feedback can shut down the pressure bumps that trap dust","feed_subtitle":"Protoplanetary-disk simulations find the convective overstability's zonal flows cannot push dust to planetesimal densities.","key_machinery":"The load-bearing object is the dusty Boussinesq shearing box in axisymmetry, with dust treated as a pressureless fluid coupled by drag (Stokes number St = tau_s $\\Omega$) and evolved through a positive-definite formulation of the dust-to-gas ratio. The argument's central identity is the angular momentum flux budget: the gas flux F_g = delta v_gx delta v_gy is negative in COS turbulence, while the dust flux F_d is more positive than F_g because dust drifts toward pressure maxima; to first order in the terminal-velocity approximation, F_d = F_g + 2 St (delta v_gy)^2. Since the total flux F = F_g + epsilon0 F_d must be negative for zonal flows to form, dust loading can suppress them once epsilon is large enough. The paper estimates the threshold epsilon greater than about $h_g^{2}$/St for dust feedback to inhibit zonal flows in a global disk.","core_discovery":"In the unstratified, axisymmetric Boussinesq shearing box, the nonlinear saturated state of the convective overstability is a set of quasi-steady zonal flows that act as pressure traps for dust. The paper's central result is that this dust-trapping is self-limiting: dust accumulates only until its back-reaction on the gas becomes significant, at which point the zonal flows weaken or never form. In the fiducial run with initial dust-to-gas ratio epsilon0 = 0.01, dust-to-gas ratios reach roughly 0.6 transiently and 0.2-0.3 on average before the zonal flows decay; at epsilon0 = 0.1, zonal flows do not form at all and the gas settles into wave turbulence. The mechanism identified is a competition between the negative gas angular momentum flux that creates zonal flows and the positive dust angular momentum flux produced by dust drifting toward pressure maxima. A global radial pressure gradient, which drives a background dust drift, further weakens trapping. The paper concludes that COS-driven zonal flows are not directly conducive to triggering planetesimal formation.","pith_inferences":["Inference: In a stratified disk, dust settles toward the midplane, so local dust-to-gas ratios there would exceed the box-averaged values used here; whether this raises epsilon above unity before feedback shuts off zonal flows is an open question the paper leaves implicit.","Inference: If 3D vortex formation proceeds through the breakup of zonal flows, the axisymmetric result implies a dust-abundance ceiling for COS-assisted planetesimal formation; vortices forming in dust-poor gas may later accrete dust faster than feedback can suppress them.","Inference: The identity F_d = F_g + 2 St (delta v_gy)^2 could be tested directly in full 3D simulations or particle-loaded local models where dust back-reaction is resolved, providing a quantitative check of the proposed feedback mechanism.","Inference: An observational consequence would be that dust rings in COS-active disk regions should show internal dust-to-gas ratios below unity and weak pressure perturbations, distinguishing them from rings produced by planets or dead-zone edges."],"forward_implications":["COS zonal flows concentrate dust by at most a factor of order ten, with maximum dust-to-gas ratios near epsilon ~ 0.6 transiently and roughly 0.2-0.3 on average.","Dust feedback suppresses zonal flow formation for initial dust-to-gas ratios epsilon0 ~ 0.1, leaving the gas in wave turbulence with negligible dust concentration.","A background radial pressure gradient corresponding to Pi greater than about 0.02 reduces dust trapping to factors of about two, because the local pressure bump is weak compared with the global drift.","The critical dust-to-gas ratio for dust feedback to inhibit zonal flows is estimated as epsilon greater than about h_g^2/St, which can be below unity for plausible disk parameters.","COS-assisted planetesimal formation, if it occurs through vortices, is likely restricted to dust-poor disk regions, since dusty zonal flows either fail to form or are too weak to act as precursors."],"supporting_citations":[{"why":"Establishes the pure-gas COS zonal flow formation mechanism and the negative gas angular momentum flux that the dusty runs extend.","marker":"TL21"},{"why":"Supplies the linear dusty COS theory used to motivate the simulations, including dust-loading reduction of effective buoyancy.","marker":"Lehmann & Lin 2023"},{"why":"Provides the linear COS growth-rate analysis and cooling requirement used to interpret zonal flow instability.","marker":"Latter 2016"},{"why":"Derives the Boussinesq shearing box equations that form the dynamical framework for the simulations.","marker":"Latter & Papaloizou 2017"},{"why":"Shows COS vortices can concentrate dust to planetesimal densities, the contrast case the paper argues is limited by zonal-flow suppression.","marker":"Raettig et al. 2021"},{"why":"Defines the streaming instability as the planetesimal-formation route that the paper argues COS zonal flows cannot easily trigger.","marker":"Youdin & Goodman 2005"}],"fun_headline_variants":["Dust feedback kills pressure bumps in protoplanetary disks","Zonal flows can't trap dust to planetesimal densities","Convective overstability's dust traps are self-limiting","Dust back-reaction weakens zonal flow dust trapping"],"cache_read_input_tokens":25344,"weakest_assumption_plain":"The result assumes that an unstratified, axisymmetric Boussinesq shearing box with a constant heat sink captures the relevant dust-concentration physics of real disks; if vertical dust settling or 3D vortex formation strengthens trapping, the conclusion that COS cannot trigger planetesimal formation would not hold.","fun_headline_variants_meta":{"raw":{"variants":["Dust feedback kills pressure bumps in protoplanetary disks","Zonal flows can't trap dust to planetesimal densities","Convective overstability's dust traps are self-limiting","Dust back-reaction weakens zonal flow dust trapping"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000165,"raw_usage":{"total_tokens":1281,"prompt_tokens":1009,"completion_tokens":272,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":625,"completion_tokens_details":{"reasoning_tokens":203}},"tokens_in":625,"tokens_out":272,"duration_ms":2947,"temperature":1.0,"reasoning_tokens":203,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T19:40:07.627514+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run a stratified, full-3D equivalent of the fiducial setup at the same resolution: if persistent zonal flows form at epsilon0 = 0.1 and concentrate dust to epsilon > 1, the paper's central limit is false. Alternatively, find an observed disk ring in a COS-susceptible region whose dust-to-gas ratio exceeds unity with a weak pressure perturbation.","supporting_citations":[],"review_version":1}