{"id":"fce9607c-f5d8-4118-b00a-c545f9ec492b","arxiv_id":"2411.15970","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"A parameter-free Darcy-based model predicts the time-dependent liquid interface and captured volume for a rigid brush withdrawn from a bath, and yields an experimentally validated optimal porosity.","lead":"Researchers built an analytical model, backed by experiments with 3D-printed brushes, for how much liquid a brush pulls out of a bath when withdrawn, and used it to find the brush geometry that captures the most fluid. The result turns a common but poorly understood dip-coating process into a design rule for liquid transfer and explains nectar feeding by brush-tongued animals.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Heaviside transverse-permeability closure above the bath is the pivotal assumption; a stopped-withdrawal lateral-drainage test would settle it.","rationale":"The reader identified the Heaviside transverse-permeability assumption as the weakest load-bearing premise, and my independent review reaches the same conclusion. The paper's derivation of Eq. (9a) is internally coherent, but the closure k_perp(z)=k_perp*Theta(-z) is essential to three-parameter structure and to the quantitative predictions of interface height and captured volume. It is not derived from the microscopic geometry or measured independently; it is justified only by the statement that the air-liquid interface prevents transverse expansion. That justification is physically questionable because the free surface above the bath is deformable and can accommodate radial drainage. The experimental validation is strong in the sense that the model reproduces many data sets, but the comparison involves per-speed choices of the brush diameter D and fitted permeability closures, so it does not isolate the Heaviside step. The parallel-plate experiments use the same assumption and therefore do not test it independently. I propose a concrete stopped-withdrawal experiment that directly observes whether lateral drainage occurs above the bath: if it does, the central PDE and the optimal-porosity prediction would need revision. The concern does not overturn the paper's qualitative picture, and the existing data are consistent, so the appropriate verdict remains CONDITIONAL, matching the reader's assessment. The proposed test is simple, decisive, and would either validate the closure or expose its failure.","tokens_in":29207,"tokens_out":4016,"duration_ms":45740,"concrete_test":"Perform a stopped-withdrawal experiment: withdraw a representative brush (e.g., R=500 µm, d=2.0 mm, L0=11 mm, V=16 mm/min in µ=0.97 Pa s oil) to Vt/L0 ≈ 0.5, stop the motor, and record the side profile of the entrained liquid column for several minutes. The Heaviside model predicts that the column's outer radius (free-surface location) remains fixed while the height relaxes downward, because vr=0 for z>0. Measure the inward radial displacement of the meniscus as a function of time. If the meniscus recedes radially by an amount comparable to the vertical height change, then k_perp>0 above the bath and Eq. (9a) lacks a transverse-drainage term.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim—that Eq. (9a) quantitatively predicts interface dynamics and the optimal porosity—rests on the closure k_perp(z) = k_perp*Theta(-z) (Model Eq. (2); SI Eq. (14)). Above the bath level the model forbids radial Darcy flow, so the entrained liquid column can only drain downward through the bottom of the brush. This is imposed rather than measured: the SI states that the air-liquid interface 'prevent[s] any radial expansion,' but a free surface can deform and permit lateral drainage. If significant transverse flow exists above the bath, Eq. (9a) gains an additional radial-diffusion term in the region z>0, changing the predicted H(r,t) and the optimal-porosity curves in Fig. 4A,B. The paper's experimental agreement is suggestive but does not independently validate this closure: the brush diameter D is chosen per speed in the mass comparison (D=15.8 mm at V=10 mm/min, D=16.9 mm at V=100 mm/min), and the permeability expressions are fits to numerical data. No direct measurement of radial velocity or meniscus motion above the bath is presented. The parallel-plate validation uses the same Heaviside assumption, so it does not provide an independent check. Thus the no-radial-flow-above-bath closure is the most load-bearing unverified premise; the optimal-radius predictions (R=729 µm and R=571 µm) would shift if lateral drainage above the bath is non-negligible.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper analyzes the liquid entrained by a rigid brush withdrawn from a liquid bath. The brush is modeled as an anisotropic Darcy porous medium with longitudinal permeability k∥ and a transverse permeability that is switched off above the bath level, k⊥(z)=k⊥θ(−z). A δ-expansion in the aspect ratio leads to a nonlinear PDE (Eq. (9a)) for the interface height H(r,t) in terms of three dimensionless parameters (V, D7FF̄δ, D7FF̄hJ). The authors validate the PDE against side-view measurements of the interface for 13 printed brushes and for a parallel-plate geometry, propose a master curve for the maximum interface rise, and derive an optimal porosity (equivalently pillar radius) that maximizes captured mass at a given withdrawal speed. Experiments confirm the predicted optimal radii R≈729 µm and R≈571 µm for V=10 and 100 mm/min.","tokens_in":29493,"tokens_out":6087,"duration_ms":60829,"significance":"If correct, the paper provides a useful quantitative framework for liquid capture by anisotropic porous media, capturing non-monotonic interface dynamics and yielding a design rule for optimal brushes. The main PDE is not fitted to the capture experiments; the permeability functions are obtained independently from COMSOL simulations and prior asymptotic results. The experimental comparison is broad (13 brushes, two oils, plate geometry), and the optimal-radius predictions are explicit and falsifiable. The derivation is presented in detail in the SI, including approximate analytical solutions and error estimates. The main caveats are that the Heaviside transverse-permeability closure is imposed rather than measured, and the perturbation expansion is used at aspect ratios where it is formally uncontrolled; both concerns can in principle be addressed with targeted checks.","major_comments":[{"comment":"The closure k⊥(z)=k⊥θ(−z) is load-bearing: it removes radial Darcy flow above the bath and is therefore responsible for the form of Eq. (9a) and for the predicted optimal radius. The only justification offered is that the air-liquid interface 'prevent[s] any radial expansion,' but a free interface can deform and permit lateral drainage. The agreement with experiments is suggestive but does not independently verify the step, because the parallel-plate validation uses the same closure and the permeability functions are obtained from independent numerical simulations rather than from a measurement of the radial flow. I ask for a direct test: for example, stop the withdrawal before the brush clears the bath and record whether the entrained column above the bath changes shape or drains sideways, or measure the radial component of the flow/meniscus in the region z>0. Without such a test, the central predictive claim rests on an unverified assumption; if lateral drainage above the bath is non-negligible, Eq. (9a) would need an additional radial-diffusion term and the optimum would shift.","section":"Model, Eq. (2); SI Eq. (14)"},{"comment":"The PDE is derived by expanding the pressure in δ² and neglecting a nonlinear term, assuming δ ≪ 1, yet Fig. 3A reports comparisons up to δ≈1.27 and Fig. 3B up to δ≈1.30. The SI gives a 3.5% relative-error bound for the neglected term only in a restricted parameter range (0≤D7FF̄δ≤2, 0≤V≤5, 0.1≤D7FF̄hJ≤0.5), which does not cover all experiments. Please quantify the truncation error over the full experimental parameter space, or compare Eq. (9a) with a direct numerical solution of the unsimplified system at the largest δ used. This is a correctness-risk concern rather than an observed failure, but it is load-bearing because the 'good agreement up to δ of order 1' is asserted rather than demonstrated.","section":"Model, Eqs. (6)-(9); SI Eq. (27)"},{"comment":"The mass comparison uses D=15.8 mm at V=10 mm/min and D=16.9 mm at V=100 mm/min, with the SI noting that D is not well defined for the printed brushes. Since the dimensional mass scales with D², the choice of D does not move the predicted optimal radius, but it weakens the 'no fitting parameter' claim for the dimensional mass curves. I ask the authors to report the geometric range of D for the brushes used, or to compare with a measurement-based D (e.g., the rim-pillar diameter) rather than selecting D per speed.","section":"Fig. 4C; SI 'Discussion about the value of D'"}],"minor_comments":[{"comment":"The notation is confusing because V denotes both the dimensional captured volume and its dimensionless rescaled value V=V/VI; please use distinct symbols such as Ṽ for the rescaled quantity.","section":"Eq. (12)"},{"comment":"The caption says 'for three value of R' and should read 'for three values of R'; similar small grammar issues appear elsewhere in the SI.","section":"SI caption of Fig. 6"},{"comment":"The grey area representing the region spanned by theory is not defined quantitatively; please state the precise range of D7FF̄δ and D7FF̄hJ used to generate it.","section":"Fig. 3B"},{"comment":"The fitted expressions for A(D7FF̄hJ) and B(D7FF̄hJ) are presented without error estimates or fitting range; adding these would help readers assess the accuracy of Eq. (14).","section":"Eq. (14)"}],"recommendation":"major_revision","confidential_remarks":"The Heaviside transverse-permeability closure is the pivotal uncertainty. I recommend asking for a dedicated experimental or numerical check of lateral drainage above the bath; if the authors can supply it, the paper would be very strong. The D-selection issue and the fit-based auxiliary expressions also deserve clarification, though they do not undermine the optimal-radius prediction. The manuscript is within scope and the experimental dataset is valuable."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Radisson et al. have done something genuinely useful: they turn the old scaling-law picture of viscous entrainment from a brush into a working PDE for the interface, identify the three dimensionless groups that control it, and show the predicted master curve really does collapse data from 13 brushes and a plate geometry. The optimal-radius prediction (729 µm at 10 mm/min, 571 µm at 100 mm/min against measured maxima in the same ballpark) is a concrete, falsifiable claim. The permeability closures come from independent COMSOL simulations, not from the capture experiments, so the interface predictions are not curve fits. That is a real achievement.\n\nThe main soft spot is the Heaviside transverse permeability, k_perp(z) = k_perp Theta(-z). It is imposed because a constant k_perp underestimates the entrained height, and it is central to the dynamics above the bath. The SI's physical rationale—the air-liquid interface prevents radial expansion—is plausible but not tested. If lateral drainage above the bath is non-negligible, the PDE gains a radial-diffusion term in that region and the optimal-porosity curves shift. I do not think this sinks the paper: the authors are transparent about the assumption, the agreement across a wide parameter range is strong, and the optimal radius is insensitive to the multiplicative D^2 factor. But a direct test, e.g. stopping the withdrawal and watching whether the meniscus relaxes radially, would settle it.\n\nTwo smaller caveats. The phrase \"without any fitting parameter\" is a little generous because the effective brush diameter D is chosen per speed when computing masses (15.8 mm vs 16.9 mm). That does not change the optimal-radius prediction, but it is a chosen value, not a measured one. Also the paper reports no code or raw data, which makes re-implementation slower than it should be.\n\nOverall: the central argument holds up. This deserves a serious referee. I would ask the authors for either a stopped-withdrawal lateral-drainage measurement or a direct estimate of the transverse velocity above the bath, and for a clear statement of how D is defined. With those, the paper is publishable as is.","headline":"Strong and useful paper; the Heaviside permeability closure above the bath is the main unvalidated assumption, but the central results survive scrutiny.","tokens_in":30067,"tokens_out":2442,"would_cite":true,"duration_ms":24822,"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 claims that liquid capture by a rigid brush withdrawn from a bath obeys a parameter-free Darcy-law equation, and that captured volume peaks at a computable brush porosity.","keywords":["viscous entrainment","rigid brush","Darcy's law","anisotropic porous medium","capillary rise","Jurin height","optimal porosity","dip coating"],"falsifier":"Track tracer particles in the liquid column above the bath level during withdrawal: if the radial velocity is measurably nonzero in the region $z>0$, the Heaviside closure $k_\\perp(z) = k_\\perp\\theta(-z)$ is wrong and Eq. (9a) cannot be the full description. A complementary check is to measure $h_0(t)$ for very slow withdrawals and compare the timing of the interface maximum with the model; a systematic shift with $\\bar\\delta$ would indicate missing transverse drainage.","tokens_in":28973,"feed_emoji":"🖌️","tokens_out":6747,"duration_ms":64484,"temperature":0.7,"pith_summary":"This paper tries to establish that the liquid captured by a rigid brush pulled from a bath is not governed by vague scaling laws but by a single nonlinear partial differential equation derived from Darcy's law, with every coefficient fixed by brush geometry and fluid properties. The equation tracks the spatio-temporal position of the air-liquid interface inside the brush, so it predicts not only the maximum entrained height but the entire volume history. If correct, it turns brush design into a calculation: for fixed retraction speed and immersion depth there is an optimal porosity, and the model gives its value. The authors validate the prediction with 3D-printed brushes, finding measured optimal pillar radii close to the computed values for two retraction speeds.","feed_headline":"Optimal brush formula predicts liquid haul with no fits","feed_subtitle":"At slow pull, best pillar radius is ~729 µm; at fast pull, ~571 µm—model and experiment agree.","key_machinery":"The load-bearing object is the nonlinear diffusion-type PDE (Eq. (9a)) for $H$, obtained by expanding the pressure in powers of the aspect ratio $\\delta = 2L_0/D$. The mechanism is the competition between the brush's upward motion and gravity-driven drainage through the longitudinal and transverse permeabilities $k_\\parallel$ and $k_\\perp$, with $k_\\perp$ set to zero above the bath level by the Heaviside closure $k_\\perp(z) = k_\\perp\\theta(-z)$. The equation's coefficients are fixed by three dimensionless numbers: $\\mathcal V = V/V_\\parallel$ (retraction speed versus Darcy drainage speed), $\\bar\\delta = \\delta(k_\\perp/k_\\parallel)^{1/2}$ (radial versus vertical flow), and $\\bar h_J = h_J/L_0$ (initial capillary rise versus immersion depth). The optimal-porosity result follows by re-expressing these parameters in terms of porosity and maximizing the final captured volume.","core_discovery":"The central claim is that a brush withdrawn from a bath behaves as an anisotropic porous medium whose transverse permeability switches off above the bath level, and that under this closure Darcy's law yields Eq. (9a) for $H(\\bar r,\\bar t)$, the interface height relative to the Jurin height. The equation involves three dimensionless groups: the retraction speed relative to the gravitational drainage speed, the radial-to-vertical flow ratio, and the Jurin height relative to immersion depth, with no fitting parameters. Its solutions reproduce the observed non-monotonic rise-and-fall of the interface at low speed and monotonic rise at high speed, and collapse the measured maximum heights onto a master curve. The same equation yields the captured volume at the end of retraction, which is non-monotonic in porosity; the predicted optimal pillar radii, $R = 729\\,\\mu\\mathrm{m}$ at $V = 10\\,\\mathrm{mm/min}$ and $R = 571\\,\\mu\\mathrm{m}$ at $V = 100\\,\\mathrm{mm/min}$, match the measured maxima near $R = 700$-$750\\,\\mu\\mathrm{m}$ and $R = 550$-$600\\,\\mu\\mathrm{m}$.","pith_inferences":["A testable extension is to replace the Heaviside transverse-permeability step with a smooth decay above the bath; if that decay length is set by the meniscus scale, the optimal porosity could shift at low speeds, which high-speed imaging of the rim could probe.","Because the captured-volume curve is flat near its maximum, small manufacturing errors in pillar radius barely change the collected mass; this tolerance is implicit in the paper's numbers and could be used to set fabrication tolerances for practical brush designs.","For nectar-feeding animals with brush-like tongues, the model suggests an optimal papilla spacing for a given nectar viscosity and lapping speed; the paper notes the biological context but does not map animal tongue geometries onto the predicted optimum.","The optimal porosity is not a material constant: it depends on immersion depth and brush width through $\\bar\\delta$ and $\\bar h_J$, so translating the result into industrial dip-coating requires specifying the full operating conditions."],"forward_implications":["For a given brush and liquid there is no optimal retraction speed or immersion depth: captured volume grows monotonically with both, so the practical design lever is brush geometry rather than operating speed.","At fixed speed and depth, captured volume is non-monotonic in porosity, and the optimal porosity saturates at high speed to approximately $1 - \\tilde h_J^2/4$, a value insensitive enough that near-optimal brushes collect almost the same volume.","The maximum interface height follows the quasi-master curve $H_0^m \\simeq [\\bar h_J + \\alpha(\\bar\\delta)]\\mathcal V$ for speeds below a transition value, giving a universal law for predicting maximum capture.","The same model, with the appropriate permeabilities, describes two parallel plates withdrawn from a bath, so the framework extends to other anisotropic porous geometries beyond pillar brushes."],"supporting_citations":[{"why":"Supplies the Jurin height formula (Eq. (1)) for capillary rise in assemblies of parallel cylinders, fixing the initial condition $h_J$ for the interface.","marker":"[7, 8]"},{"why":"Supplies Darcy's law, the constitutive equation from which the velocity and pressure fields in the brush are derived.","marker":"[42]"},{"why":"Provide the asymptotic permeability expressions for arrays of parallel cylinders that the full-range fits for $k_\\parallel$ and $k_\\perp$ extend.","marker":"[43, 44]"},{"why":"Provide the Landau-Levich-Derjaguin dip-coating baseline that the brush entrainment mechanism generalizes.","marker":"[14, 15]"}],"fun_headline_variants":["Brush liquid haul predicted with zero fitting","Optimal brush radius found from fluid theory","No-parameter model optimizes brush capture","Model matches brush experiments at all speeds","Optimal brush geometry set by no-fit model"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The model assumes that liquid above the bath level cannot drain sideways: the transverse permeability is exactly zero there, and this step is imposed rather than measured, so significant sideward drainage above the bath would change the interface equation and the predicted optimal porosity.","fun_headline_variants_meta":{"raw":{"variants":["Brush liquid haul predicted with zero fitting","Optimal brush radius found from fluid theory","No-parameter model optimizes brush capture","Model matches brush experiments at all speeds","Optimal brush geometry set by no-fit model"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000555,"raw_usage":{"total_tokens":2662,"prompt_tokens":981,"completion_tokens":1681,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":597,"completion_tokens_details":{"reasoning_tokens":1630}},"tokens_in":597,"tokens_out":1681,"duration_ms":12007,"temperature":1.0,"reasoning_tokens":1630,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T13:41:47.787875+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Track tracer particles in the liquid column above the bath level during withdrawal: if the radial velocity is measurably nonzero in the region $z>0$, the Heaviside closure $k_\\perp(z) = k_\\perp\\theta(-z)$ is wrong and Eq. (9a) cannot be the full description. A complementary check is to measure $h_0(t)$ for very slow withdrawals and compare the timing of the interface maximum with the model; a systematic shift with $\\bar\\delta$ would indicate missing transverse drainage.","supporting_citations":[{"cited_title":"Guyon, J.-P","cited_arxiv_id":null,"evidence_quote":"Supplies Darcy's law, the constitutive equation from which the velocity and pressure fields in the brush are derived."}],"review_version":1}