{"id":"37232358-5dc6-402f-aed6-d32000443531","arxiv_id":"2608.05672","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"In fully convective envelopes of disk-embedded planets, dust is retained only when convection reaches the disk flow and grains are small enough to be trapped by convective stirring.","lead":"Using 3D simulations of gas and dust around an embedded planet, this paper shows that convective planetary envelopes can either keep dust or lose it, depending on whether convection connects to the surrounding disk. The result links envelope dynamics to the diverse atmospheric compositions seen by JWST in mini-Neptunes.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The isolated/non-isolated dust dichotomy hinges on a one-cell-wide step in the cooling-time profile at r=RB; a finite radiative layer could erase the sharp dust-depletion boundary.","rationale":"The reader's weakest_assumption and my independent reading identify the same load-bearing concern: the isolated/non-isolated dichotomy is generated by a hand-set, one-cell-wide step in the cooling time at r=RB. This is not a minor technical detail; it is the mechanism that establishes the entropy barrier and hence the dust depletion in the variable-beta case. Appendix B explicitly acknowledges the lack of a self-consistent opacity model for this profile and the presence of a finite radiative layer in realistic envelopes. The paper does valuable work in demonstrating the dust-transport consequences of two plausible end-member envelope structures, and the simulations are internally consistent and convergence-tested. However, the central astrophysical claim requires that the sharp transition (or something close to it) actually occurs in nature, and the paper does not establish that. The proposed test directly probes whether the sharpness of the transition is essential: if widening the transition by factors of 10–100 removes the depletion, the central claim is conditional on an artificial prescription; if the depletion persists, the mechanism is robust. This does not change the reader's CONDITIONAL verdict, because the reader already conditioned on this assumption; it sharpens the specific test that would resolve it.","tokens_in":20530,"tokens_out":2884,"duration_ms":33081,"concrete_test":"Repeat the variable-beta runs (Table 1) with the transition width w in Eq. (18) increased to 0.1RB and 1RB (resolved by at least 10 cells), and separately insert a finite radiative layer (beta=10^0) of thickness 0.5RB between the inner convective region (beta=10^3) and the outer recycling flow. For St=10^-3 and 10^-4, check whether the shell-averaged dust-to-gas ratio inside RB still drops by more than two orders of magnitude relative to the constant-beta case. If the depletion weakens or the boundary moves away from RB, the isolation mechanism is an artifact of the sharp step; if it persists, the qualitative conclusion is robust despite the idealized prescription.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim that dust-to-gas ratio is set by dynamical isolation rests on the prescribed cooling-time profile in Eq. (18): a step in beta at r0=RB with width w=0.01RB, which is resolved by a single grid cell and therefore acts as a discontinuity. This step creates an entropy barrier that decouples the inner convective layer from the outer recycling flow, producing the two orders-of-magnitude drop in epsilon at RB seen in Fig. 4. Appendix B concedes that no single opacity model reproduces both the envelope structure and this cooling prescription simultaneously, and that realistic envelopes contain a finite radiative layer separating the convective interior from the recycling flow. If the stability transition is spread over a finite width, the recycling flow may partially penetrate the outer convective region, altering the filtering efficiency and shifting or smoothing the depletion boundary. The simulations themselves are internally consistent and convergence-tested, but the central regime dichotomy is currently a consequence of the imposed beta discontinuity rather than a demonstrated robust outcome of the coupled thermodynamics.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper presents 3D multifluid (gas+dust) Athena++ simulations of an Earth-mass planet embedded in a protoplanetary disk, with thermal relaxation parameterized by a cooling time β and with accretion luminosity included. Two cooling prescriptions are compared: a constant β=10^3, which keeps the envelope convectively unstable beyond the Bondi radius, and a variable β profile that is a step from β=10^0 to β=10^3 at r=R_B with width 0.01 R_B. The authors find that the constant-β envelope remains dynamically connected to the disk and retains dust (ε≳10^-3), whereas the variable-β envelope is isolated by an outer recycling layer and becomes depleted by more than two orders of magnitude at R_B, except for small grains (s≲0.01 cm) that are trapped by convection. A 1D analytic model is extended to describe the settling-dominated and convection-dominated regimes, and the implications for the volatile content of super-Earths and mini-Neptunes are discussed.","tokens_in":20663,"tokens_out":7018,"duration_ms":76426,"significance":"If the results are robust, they are significant for planet formation: they identify dynamical isolation of the envelope as a controlling factor for dust retention and volatile delivery, and they connect envelope structure to the observed diversity of atmospheric metallicities. The numerical work is internally consistent and convergence-tested (Appendix C), the diagnostics are time-averaged over the last ten orbits, and the paper is unusually explicit about the idealized nature of its cooling treatment (Section 5.4 and Appendix B). However, the central regime dichotomy rests on a sharply imposed cooling-time discontinuity, and the analytic convection-dominated branch is assumed rather than derived, so the general claims require either additional robustness tests or a more cautious framing.","major_comments":[{"comment":"The central distinction between non-isolated and isolated envelopes is produced by a step-function cooling-time profile whose transition width w=0.01 R_B is resolved by a single grid cell and therefore acts as a discontinuity at the Bondi radius. Appendix B explicitly concedes that no single opacity model reproduces both the envelope structure and the adopted cooling prescription, and that realistic envelopes contain a finite radiative layer separating the inner convective region from the outer recycling flow. Because the two-order-of-magnitude drop in ε at r≈R_B (Fig. 4) is the main evidence for the 'dynamical isolation' scenario, the authors should test the sensitivity of the depletion boundary and of the retention/depletion dichotomy to the transition width and location (e.g., w=0.1 and 0.5 R_B, or a smooth radiative-convective structure). Without such a test, the central claim that the dust-to-gas ratio is set primarily by the degree of dynamical isolation is not yet established beyond this specific prescription.","section":"Sect. 2.4, Eq. (18); Appendix B; Fig. 4"},{"comment":"The analytic model's convection-dominated branch sets ε(r)=ε_0 whenever v_con>v_term for all r≤R_B. This is not derived from the dust continuity equation or from a mixing model; it simply returns the initial dust-to-gas ratio. The agreement with the simulations in Figs. 4 and 9b is therefore not an independent prediction but a restatement of the assumption that the dust is perfectly retained. The authors should either derive this branch from a steady-state dust balance (e.g., zero net radial flux or a diffusion–settling equilibrium) or explicitly present it as an empirical closure and test it against simulations with different initial ε_0. As written, the statement that the 'simulated dust-to-gas profiles agree well' (Sect. 4) is misleading for the convection-dominated regime.","section":"Sect. 4, Eq. (29)"},{"comment":"The claimed threshold between settling-dominated and convection-dominated regimes (s≲0.01 cm versus s≳0.1 cm) is based on a single planet mass (m=0.1), a single accretion luminosity (t_acc=10^5 yr), and, for the fixed-size runs, a single orbital location (1 au). The authors note that v_con scales as (ℓ_m L_acc)^{1/3} (Eq. 24), so the threshold should shift with luminosity, but no parameter sweep is presented. The abstract and conclusions present the 0.01–0.1 cm boundary as a general result; it should be framed as a fiducial-value estimate, or the authors should vary L_acc and m to map the regime boundary.","section":"Sects. 2.4 and 3.3.2"}],"minor_comments":[{"comment":"The expression for Q_heat is not typeset clearly in Eq. (11); please verify the prefactor and dimensions, since the printed form is ambiguous.","section":"Sect. 2.2, Eq. (11)"},{"comment":"The text contains a typo: 'mini-Netpunes' should be 'mini-Neptunes'.","section":"Introduction"},{"comment":"In the concluding paragraph, 's≲0.01,cm' contains an errant comma; it should read 's≲0.01 cm'.","section":"Conclusions"},{"comment":"The 'dust-retaining' conclusion for the non-isolated case is based on the shell-averaged ε, while the midplane values are systematically lower (thin curves in Fig. 4). Since the discussion of observable atmospheric composition concerns the outer envelope, please also report density-weighted or midplane values and discuss whether the qualitative conclusion is affected.","section":"Sect. 3.2, Fig. 4"},{"comment":"The parameter ξ=0.08m in the recycling-flow collision rate is chosen to 'better match our numerical results'; this makes the analytic model partly calibrated rather than fully predictive. Please state this more prominently in the main text when presenting the 1D model.","section":"Appendix D, Eq. (D.8)"}],"recommendation":"major_revision","confidential_remarks":"This is a well-executed numerical study with honest and detailed limitation sections. The main reasons for major revision are the sensitivity of the central dichotomy to the imposed β step and the assumed nature of the convection-dominated branch in Eq. (29). I would not reject: the idealized end-member approach is defensible as a first step, but the general conclusions require either additional robustness tests or a more cautious framing."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Kuwahara & Lambrechts II is a real extension of their first paper: adding accretion luminosity to 3D multifluid sims and showing that convective envelopes split into a non-isolated dust-retaining state and an isolated dust-depleting state, with small grains (≲0.01 cm) trapped by convection. The simulations are internally consistent, time-averaged, and convergence-tested in App. C. That is genuine work and the regime picture is credible as an end-member mechanism.\n\nThe soft spot is exactly what the stress-test note says: the isolated/non-isolated dichotomy is generated by a step in the cooling β profile at r=RB with width w=0.01 RB, resolved by one grid cell. Appendix B concedes no single opacity model reproduces that step; realistic envelopes have a finite radiative layer. So the sharp depletion boundary at RB and the two-orders-of-magnitude drop are, at this point, properties of the imposed entropy barrier. The authors say this themselves in App. B, which is honest, but it does mean the central dichotomy is not yet demonstrated to be robust under self-consistent cooling.\n\nThe 1D analytic model is also partly fitted and partly imposed: ξ=0.08m is tuned to match the simulations, and Eq. 29's convection-dominated branch simply returns the initial ϵ0 whenever v_con > v_term. That branch isn't derived; it's a prescription. The settling-dominated branch does real work and reproduces the profiles, so the model is not empty, but its predictive content is lower than a reader might first think.\n\nI do not think these soft spots are fatal. The paper is framed as idealized end-member sims and the limitation statements are right there. The right next step is radiative-transfer-coupled sims with dust-opacity feedback, which the authors flag. For a mechanism-guidance paper, the regime distinction likely survives qualitatively even if the sharp boundary moves.\n\nWho is this for: anyone working on pebble accretion, envelope composition, or mini-Neptune atmospheric metallicity. It deserves a serious referee: the numerics are careful, the advance over KL26b and over Popovas et al. is real, and the limitations are stated rather than hidden. My recommendation: send to review, with the cooling prescription's influence on the main dichotomy as the central referee question.","headline":"Solid sequel to KL26b with a genuinely new dust-retention dichotomy, but the central regime split currently rests on a hand-set one-cell cooling step at the Bondi radius.","tokens_in":21247,"tokens_out":1659,"would_cite":true,"duration_ms":17272,"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":"Whether a forming planet's envelope keeps dust comes down to one thing: is it open to the disk or sealed off by a recycling flow?","keywords":["dust transport","planetary envelopes","convection","protoplanetary disks","pebble accretion","dust-to-gas ratio","planet formation"],"falsifier":"Rerun the isolated-envelope configuration with a smooth cooling-time transition of width much larger than $0.01 R_B$, as the paper's own cooling-time estimate suggests, and check whether the dust-to-gas ratio still drops sharply at $r \\simeq R_B$; if the depletion boundary smears or moves, the isolation dichotomy is an artifact of the step prescription. A second check is to vary the accretion luminosity $\\tilde{L}_{\\rm acc}$ and confirm that the trapping threshold shifts according to $v_{\\rm con} \\propto L_{\\rm acc}^{1/3}$.","tokens_in":20217,"feed_emoji":"🪐","tokens_out":10518,"duration_ms":87059,"temperature":0.7,"pith_summary":"This paper asks whether the convective envelopes of disk-embedded planets retain the dust they accrete or lose it. Using 3D multifluid simulations of gas and dust around an Earth-mass planet, the authors find that the dust-to-gas ratio inside the envelope is set primarily by how dynamically isolated the envelope is from the disk: if convection extends beyond the Bondi radius the envelope exchanges gas with the disk continuously and stays dust-rich, whereas if an inner convective layer is sealed off from an outer recycling layer, that recycling flow filters out incoming solids and the envelope becomes dust-depleted. Inside the isolated layer the fate of dust is decided by a race between convective stirring and gravitational settling, which traps grains smaller than roughly $s \\lesssim 0.01$ cm and lets larger grains fall to the core. The paper's picture implies that where a planet forms decides what its envelope and core are made of, offering an explanation for the wide diversity of mini-Neptune atmospheric compositions.","feed_headline":"Convection decides if a planet's envelope is dust-rich or dust-poor","feed_subtitle":"Dust retention depends on whether convection stays open to the disk or is sealed off by a recycling flow.","key_machinery":"The argument is carried by a cooling-time prescription that dials the envelope between two end-member states: the dimensionless cooling time $\\beta \\equiv t_{\\rm cool}\\Omega$, with a constant value $\\beta = 10^3$ giving a non-isolated, fully convective envelope, and a radially varying profile that steps from $\\beta_{\\min}=10^0$ outside to $\\beta_{\\max}=10^3$ inside the Bondi radius $R_B$, giving an isolated convective interior with a recycling flow outside. Accretion luminosity $L_{\\rm acc}$ heats the deep envelope and sustains the convection. Within the envelope, the competition between the mixing-length convective velocity $v_{\\rm con} \\propto (l_m L_{\\rm acc})^{1/3}$ and the terminal settling velocity $v_{\\rm term}$ sets whether grains are trapped or settle, and a one-dimensional steady-flux model $\\rho_d(r) = \\dot{M}_{d,\\rm acc}/(4\\pi r^2 v_{d,\\rm in}(r))$, with a collision rate modified by recycling outflow and convective inflow, reproduces the simulated dust-to-gas ratios.","core_discovery":"The central claim is that convective envelopes come in two dynamical states with opposite consequences for dust. In the non-isolated state, realized when convection extends beyond the Bondi radius, the envelope is continuously flushed by the surrounding disk flow and the shell-averaged dust-to-gas ratio stays near $\\epsilon \\sim 10^{-3}$ to $10^{-2}$, because dust enters through a polar accretion pathway and is lofted by convection. In the isolated state, realized when the envelope is convective only inside the Bondi radius and a positive entropy gradient blocks inflow outside it, a recycling flow with polar inflow and midplane outflow filters incoming solids, and the dust-to-gas ratio drops by more than two orders of magnitude across $r \\simeq R_B$. Within the isolated convective layer, dust retention is governed by whether the mixing-length convective velocity $v_{\\rm con}$ exceeds the terminal settling velocity $v_{\\rm term}$: small grains ($s \\lesssim 0.01$ cm) are trapped in convective circulation and keep the envelope dusty, while larger grains ($s \\gtrsim 0.1$ cm) settle at terminal velocity and deplete it. The authors extend their previous one-dimensional analytic model to this convective regime and reproduce the simulated dust-to-gas profiles.","pith_inferences":["If the sharp step in the cooling time at the Bondi radius were replaced by the broad radiative-convective transition that the paper's own cooling-time estimate predicts, the sharp drop in dust-to-gas ratio at $R_B$ would likely smear out; the predicted inner/outer disk compositional dichotomy may be a gradient rather than a clean boundary.","The simulations neglect fragmentation, yet the paper's collision-speed estimate for 1 cm grains (about 1-10 m/s) is near typical fragmentation thresholds; if grains fragment they shrink into the convection-trapped size range, which could re-dust an isolated envelope.","A testable extension is that the size threshold between trapping and settling should shift with accretion luminosity, since $v_{\\rm con} \\propto L_{\\rm acc}^{1/3}$; varying $\\tilde{L}_{\\rm acc}$ in the same setup would map where the dust-depletion boundary sits."],"forward_implications":["Non-isolated convective envelopes, expected inside roughly 1 au, stay dust-rich and inherit near-disk compositions, giving inner super-Earths and mini-Neptunes volatile-poor cores and volatile-rich envelopes.","Isolated envelopes, expected beyond roughly 1-10 au, deplete small grains; larger pebbles still reach the core and release volatiles at the sublimation front, so the deep envelope can be volatile-enriched while the observable outer envelope stays closer to disk composition.","Dust accretion rates differ by orders of magnitude between the two states: recycling outflow suppresses pebble accretion, while inward convective motions at the Bondi radius enhance it.","Dust retention raises envelope opacity and slows cooling, which can delay runaway gas accretion in the inner disk, while dust-depleted outer envelopes cool faster and may favor gas giant formation."],"supporting_citations":[{"why":"The companion paper whose multifluid gas-dust equations, drag laws, and boundary setup this study extends, and whose convectively stable baseline shows dust-depleted envelopes.","marker":"KL26b"},{"why":"Establishes when accretion luminosity drives envelope convection and when a radiative layer isolates the envelope; supplies the mixing-length convective velocity estimate used here.","marker":"Kuwahara & Lambrechts 2026a"},{"why":"Characterizes the 3D recycling flow with high-latitude inflow and midplane outflow that filters incoming solids in isolated envelopes.","marker":"Ormel et al. 2015b"},{"why":"Shows that a positive entropy gradient prevents disk gas from penetrating the Bondi radius, the mechanism behind envelope isolation.","marker":"Kurokawa & Tanigawa 2018"},{"why":"Provides the analytic adiabatic envelope structure and radiative-convective boundary framework used to interpret the cooling-time prescriptions.","marker":"Rafikov 2006"},{"why":"Supplies the pebble collision-rate model that the extended 1D dust-to-gas formula is built on, with recycling and convective corrections added.","marker":"Okamura & Kobayashi 2021"},{"why":"Defines the accretion luminosity scaling that heats the envelope and sustains convection in the simulations.","marker":"Lambrechts et al. 2014"},{"why":"Earlier 3D simulations of convective envelopes heated by accretion luminosity; source of the heating implementation.","marker":"Zhu et al. 2021"},{"why":"Prior super-particle simulations of dust in convective envelopes, the comparison case for the non-isolated dust-accretion result.","marker":"Popovas et al. 2018"}],"fun_headline_variants":["Envelope dust richness set by convection isolation","Grain size and convection reach decide envelope dust","Two convective states control dust in planet envelopes","Convection beyond Bondi radius retains envelope dust","Dust retention in envelopes linked to convection state"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the two envelope states are produced by the prescribed cooling-time profile — a uniformly long cooling time that keeps convection alive beyond the Bondi radius versus a sharp step at the Bondi radius that seals the envelope off — even though the paper concedes that no single opacity model produces that structure and that realistic envelopes have a finite radiative layer rather than a sharp step.","fun_headline_variants_meta":{"raw":{"variants":["Envelope dust richness set by convection isolation","Grain size and convection reach decide envelope dust","Two convective states control dust in planet envelopes","Convection beyond Bondi radius retains envelope dust","Dust retention in envelopes linked to convection state"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000523,"raw_usage":{"total_tokens":2608,"prompt_tokens":1102,"completion_tokens":1506,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":718,"completion_tokens_details":{"reasoning_tokens":1437}},"tokens_in":718,"tokens_out":1506,"duration_ms":12849,"temperature":1.0,"reasoning_tokens":1437,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-08T04:22:48.849480+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Rerun the isolated-envelope configuration with a smooth cooling-time transition of width much larger than $0.01 R_B$, as the paper's own cooling-time estimate suggests, and check whether the dust-to-gas ratio still drops sharply at $r \\simeq R_B$; if the depletion boundary smears or moves, the isolation dichotomy is an artifact of the step prescription. A second check is to vary the accretion luminosity $\\tilde{L}_{\\rm acc}$ and confirm that the trapping threshold shifts according to $v_{\\rm con} \\propto L_{\\rm acc}^{1/3}$.","supporting_citations":[],"review_version":1}