{"id":"53fa940a-124c-451b-8f8b-4fd61554bb47","arxiv_id":"2509.02262","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Strong forcing saturates the maximum load a vibrating sheet can support, and fluid inertia or compressibility reduce that load further.","lead":"This paper shows numerically that a vibrating elastic sheet can hover near a wall even under strong forcing, and that fluid inertia and compressibility each add a repulsive force that weakens this contactless adhesion. The results provide design rules for soft grippers and other squeeze-film levitation devices.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"First-order inertial lubrication used at Re_eq O(100) may overstate the repulsive corrections, so the quantitative Gmax drops in Figs. 8–9 are not yet secured.","rationale":"The Reader's weakest assumption identifies the same governing-equation limitation: first-order inertial lubrication plus the k = 0.5 entrance-loss boundary condition may be inadequate near the high end of the explored Reynolds range. My stress-test agrees with this concern and sharpens it: the paper itself admits the model is not expected to be valid for Re_eq ≳ 50, but Fig. 9c shows Re_eq reaching O(10^2) in the strong-forcing inertial sweeps. Thus the quantitative reductions in Gmax and the sharp drops in Figs. 8 and 9 are computed with the model beyond its stated range. This does not threaten the central strong-forcing saturation result (3.4), which comes from the purely viscous incompressible simulations, nor the qualitative direction of the inertial correction, which is backed by the rigid-sheet asymptotic analysis. It does mean the quantitative inertial-correction part of the central claim is conditional on a model-validity assumption that is explicitly violated in part of the parameter space. Since the Reader already returned CONDITIONAL on essentially this point, my recommendation is UNCHANGED: the paper should be accepted only with that caveat, not rejected, because the main scaling and saturation results have independent support from the incompressible inertialess numerics and the rigid-sheet asymptotics.","tokens_in":90946,"tokens_out":3082,"duration_ms":43207,"concrete_test":"Repeat the α = 20, Sq_bv = 0 sweeps of Fig. 9 using a direct numerical solution of the incompressible Navier–Stokes equations in the gap (or, minimally, a lubrication model retaining O(Re^2) terms and a Re-dependent edge-loss coefficient) at several Rebv values where Re_eq ≈ 10, 50, and 100. Compare the predicted Gmax(Re_eq) curve and the locations of the sharp drops with Fig. 9. If deviations exceed about 20% in Gmax, the quantitative inertial-reduction claim requires revision; if the curves agree, the first-order model is adequate for the reported conclusions.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The soft-sheet inertia results in Sec. 4.2 are computed from the first-order-in-Re lubrication equation (2.5), derived in Appendix A under the assumptions Re ≪ 1 and ε ≪ 1, with O(Re^2) terms explicitly dropped. The accompanying edge boundary condition (2.10) uses k = 0.5, which is calibrated in Appendix B only for low Reynolds numbers: matching Ramanarayanan et al. gives k ≈ 0.49 for Re ≲ 5 but k ≈ 0.32–0.16 at high Re. For α = 20, the simulations reach Re_eq = O(10^2) (Fig. 9c), and the paper itself states that Eq. (2.5) is not expected to be valid for Re_eq ≳ 50. Yet the main inertial claims—the repulsive contribution, the significant reduction of Gmax, and the sharp drops near 1/e_i^2—are extracted from simulations in exactly this regime. If higher-order inertial terms or a Reynolds-number-dependent loss coefficient materially change the integrated pressure and the bifurcation thresholds, the quantitative inertial correction results in Figs. 8 and 9 would change, even though the purely viscous saturation formula (3.4) and the qualitative direction of the inertial effect would survive. The load-bearing weakness is therefore not the existence of an inertial repulsion, which the rigid-sheet asymptotics support, but the numerical magnitude of that repulsion in the soft-sheet regime where the model is used beyond its stated range of validity.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the elastohydrodynamic hovering of a thin elastic sheet vibrating near a rigid substrate, extending the authors' earlier weak-forcing asymptotic analysis to strong forcing through one-dimensional numerical simulations of a depth-integrated lubrication model coupled to Kirchhoff-Love bending. The main claims are: (i) the maximum supported weight crosses over from Gmax ~ 0.137 α^2 for weak forcing to a saturation Gmax ~ 7.8 for α ≳ 20, captured by the interpolation (3.4); (ii) at strong forcing the equilibrium height obeys heq ≈ 0.05 G; (iii) fluid inertia and compressibility each introduce repulsive contributions that reduce adhesive strength, with inertia causing sharp drops in Gmax at Reynolds numbers interpreted through the eigenmode heights e_i of the authors' previous reduced model. The paper combines asymptotic rigid-sheet analyses (Appendices D and E), numerical bifurcation diagrams, and regime maps.","tokens_in":91416,"tokens_out":6248,"duration_ms":77510,"significance":"If the quantitative claims hold, the paper provides a useful design rule for soft contactless grippers: the load capacity of a vibrated elastic sheet is bounded by the elastohydrodynamic force scale F_bv, and inertial/compressible corrections impose additional penalties. The viscous part of the work is convincing: the numerics reproduce the prior weak-forcing asymptotic model in Fig. 2(b), and the saturation of Gmax is a clear, reproducible numerical result. The rigid-sheet asymptotic calculations for inertia and compressibility are also valuable and appear internally consistent. The main concern is that the soft-sheet inertial results, which carry the paper's central 'inertia reduces adhesion' message, are computed in a regime where the paper itself states the model is not expected to be valid. The paper is honest about this limitation in §4.2 and §6, but the quantitative claims—especially the magnitude of the Gmax reduction and the locations of the sharp drops—are not yet secured.","major_comments":[{"comment":"The quantitative inertial results for soft sheets rely on Eq. (2.5), derived under the assumption Re ≪ 1 in Appendix A, and on the edge loss coefficient k = 0.5, which Appendix B calibrates for 'Reynolds numbers that remain small, Re < 10'. Yet the paper reaches Re_eq = O(10^2) for α = 20 in Fig. 9(c), and the sharp drops in Gmax occur precisely in that range. The paper itself states that Eq. (2.5) 'may no longer be valid' for Re_eq ≳ 50 and that full Navier-Stokes simulations or higher-order corrections are needed. Since the Bernoulli-like repulsion and the threshold locations depend on k and on neglected higher-order inertial terms, the numerical magnitude of the Gmax reduction and the quantitative locations of the drops are not established. Please either restrict the inertial soft-sheet claims to Re_eq ≲ 50, supply a validation against full Navier-Stokes or higher-order inertial lubri","section":"§4.2, Figs. 8–9, Eq. (2.5), Appendix B"},{"comment":"The strong-forcing saturation Gmax ≈ 7.8 and the interpolation (3.4) are central to the paper's design conclusions, but the coefficient 0.0176 is a numerical fit and the α ≫ 1 plateau is not derived analytically. The formula is presented as a general result and is used to compute the optimal bending stiffness and the design curves in Fig. 4(d). This is acceptable as a numerical finding, but the paper should state explicitly that (3.4) is a fit over 0 < α ≤ 100, not a parameter-free law, and should indicate its uncertainty. The current wording ('captured by the following interpolation') understates the degree to which the saturation law rests on simulation data.","section":"§3.2.1, Eq. (3.4)"},{"comment":"The sharp drops in Gmax are interpreted as happening at Re_bv ≈ 1/e_i^2, where e_i = 0.242/i^2 are characteristic heights from the authors' earlier asymptotic model (Eq. 3.2). This is not circular, because the numerical simulations are independent of the eigenmode decomposition, but the interpretation is not a test of the reduced model: the e_i are derived for the purely viscous, weak-forcing regime and are not re-derived for finite Re. Please mark this as an interpretation and clarify that the eigenmode heights are used only as a heuristic to label the drops.","section":"§4.2, Figs. 8–9 and Eq. (3.2)"}],"minor_comments":[{"comment":"The eigenfunctions ζ_n are said to be 'shown in figure 2(d)', but Fig. 2 has only panels (a) and (b); the eigenmodes appear in Fig. 11. Please correct the cross-reference.","section":"Appendix C"},{"comment":"The linear relation heq ≈ 0.05 G for α ≫ 1 is a numerical observation, not derived. Since it underpins the dimensional scaling heq ∼ W R^2/B and the contact criterion in §3.2.3, please state explicitly that the prefactor 0.05 is fitted and give the range of G/α and α over which it was verified.","section":"§3.2.2, Eq. (3.5a)"},{"comment":"The statement 'to O(ε_bv^2, Re_bv, ε_bv^2 Re_bv, Re_bv Sq_bv, Sq_bv)' is confusing: the displayed equation contains only the leading-order inertial terms, and the compressible correction appears only through ρ in (2.4). Consider rewriting the order-of-accuracy statement to match the truncated expansion actually used.","section":"§2.2, Eq. (2.5)"},{"comment":"Typos and wording: 'figure figure 7' in §4.1; 'a related investigations' in §2.6; 'centrimetric' in §6 should likely be 'centimetric'. Also, the phrase 'anecdotal observation' in §4.2 is fine but should be clearly separated from the quantitative claims.","section":"General"},{"comment":"The effective-weight relation (5.2) is shown to underestimate the compressibility effect by up to 50% at larger Sq_bv for α = 20. This is acknowledged, but the abstract and introduction present compressibility primarily as a simple 'effective weight' correction. Please soften the wording so that the quantitative scope of (5.2) is clear from the outset.","section":"§5.2, Fig. 10"}],"recommendation":"major_revision","confidential_remarks":"The viscous saturation result (Eq. 3.4) is a solid numerical contribution and likely of interest to JFM readers. The main risk is the inertial section: the authors are admirably explicit about the model's limited validity, but the paper's abstract and title still promote the quantitative inertial reductions. I recommend that the revision either add a validation study for the inertial lubrication model in the soft-sheet regime or substantially re-frame the inertial claims as qualitative and supported only for Re_eq ≲ 50. The fitted nature of (3.4) and the heq = 0.05G law should also be flagged more prominently."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here's my take on Poulain et al., arXiv:2509.02262.\n\nThe viscous part of this paper is solid. The authors take their earlier weak-forcing asymptotics (Poulain et al. 2025) and extend it numerically well into the strong-forcing regime, and the headline result holds up: the maximum supported weight saturates at Wmax ≈ 7.8 F_bv, captured by Gmax ≈ 0.137 α²/(1 + 0.0176 α²). The numerics are validated against the weak-forcing theory in Fig. 2(b), agreeing closely, so the saturation, the optimal-stiffness curve, and the contact regime map are all worth taking seriously. The compressibility section is clean too: a first-order effective-weight shift Geff = G + 1.5 α² Sq_bv, checked against the numerics, and it works where expected.\n\nThe soft spot is the inertial section, and it is the one the authors themselves flag. Equation (2.5) is first-order in Re, derived under Re ≪ 1, and the paper states it should not be trusted beyond Re_eq ≈ 50. Yet the main inertial claims for soft sheets in Sec. 4.2 — the magnitude of the Gmax reduction and the sharp drops in Figs. 8–9 — come from simulations reaching Re_eq = O(100). The edge loss coefficient k = 0.5 is calibrated (their appendix B) for Re < 10, with the matched value dropping toward 0.16 at high Re. So the stress-test note lands: the direction of the inertial effect is secured by the rigid-sheet asymptotics, but the quantitative penalty is not. The curves could shift once higher-order inertia or a Reynolds-dependent k is included.\n\nTwo smaller things. The constants in (3.4) and in heq ≈ 0.05G are numerical fits, presented honestly as observations, but they should remain labeled as such in any revision. And the interpretation of the drop locations via 1/e_i² leans on eigenmode heights from the authors' own prior model — reasonable, consistent with the mode excitation they show, but self-referential enough that the numerics should carry the argument on their own.\n\nWho should read it: anyone in elastohydrodynamic adhesion, squeeze films, soft grippers, or near-field levitation. It turns a qualitative mechanism into quantitative design rules, and the experimental grounding matters. Send it to serious referees: the viscous core deserves the attention, and the inertial section needs exactly the scrutiny a referee can provide.","headline":"Strong-forcing saturation Wmax ~ 7.8 F_bv is a solid, well-validated result; the inertial corrections in Sec. 4.2 run past the model's stated validity range, so treat those numbers as provisional.","tokens_in":91834,"tokens_out":5628,"would_cite":true,"duration_ms":55439,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["76D08","74F10","76N10"],"pacs":[],"model":"deepseek-v4-flash","headline":"The maximum weight a vibrating elastic sheet can support saturates at about 7.8 times the elastohydrodynamic force scale, and fluid inertia and compressibility reduce it further.","keywords":["elastohydrodynamic adhesion","vibrated elastic sheets","lubrication theory","squeeze film","fluid inertia","compressibility","hovering height","contactless gripper"],"falsifier":"Measure the maximum load a vibrating sheet supports as a function of the motor's force amplitude for a fixed geometry and frequency. The model predicts that Wmax saturates at about 7.8 Fbv once Fa exceeds roughly 20 Fbv; a continued rise as Fa², a peak followed by a decline, or a saturation value differing from ~7.8 Fbv would contradict the saturation formula (3.4). A separate test would be to vary the fluid pressure: the model says lowering ambient pressure weakens the compressibility-induced repulsion and thus raises Gmax.","tokens_in":90886,"feed_emoji":"🧲","tokens_out":7533,"duration_ms":75946,"temperature":0.7,"pith_summary":"This paper extends the theory of elastohydrodynamic adhesion—a vibrating elastic sheet hovering near a wall because viscous lubrication breaks time-reversal symmetry—from weak to strong forcing. The central result is that the maximum weight a sheet can support, Gmax, grows as the square of the forcing at weak amplitudes but saturates near 7.8 times the elastohydrodynamic force scale Fbv once the forcing exceeds about 20 Fbv. The paper also computes first-order corrections from fluid inertia and compressibility, finding both add repulsive contributions that reduce the load capacity and the hovering height. These results matter for contactless grippers and soft suction-cup-like devices because they put quantitative limits on what a given actuator-plus-sheet combination can lift.","feed_headline":"Vibrated-sheet grip saturates at 7.8x elastohydrodynamic force","feed_subtitle":"Stronger shaking and softer sheets stop adding lifting power; inertia and compressibility cut it further.","key_machinery":"The elastohydrodynamic force scale Fbv = (µωB²)^(1/3) and height scale Hbv = R²(µω/B)^(1/3) set the problem. The dynamics is governed by an inertial-lubrication equation (adapted from Rojas et al. 2010 to flow between two solids) coupled to a Kirchhoff-Love plate, with an entrance/exit pressure boundary condition at the sheet's free edge carrying a loss coefficient k = 0.5. The key analytical objects are the eigenmode-reduced evolution equation (3.2) for weak forcing and the interpolation (3.4) for the maximum supported weight, with the sheet's convexity during the cycle acting as the gatekeeper between contactless adhesion and edge contact.","core_discovery":"The paper's claim is that for a uniformly forced elastic sheet in a viscous fluid, the maximum dimensionless supported weight is Gmax = 0.137 α² / (1 + 0.0176 α²), interpolating between the weak-forcing scaling α² and a saturation at α ≳ 20 with Gmax ≈ 7.8. The saturation is set not by the fluid but by the sheet losing convexity: when the active force pulls hard enough, the sheet's edges touch the substrate during part of the cycle, and the threshold weight for that contact scales as Fbv. Fluid inertia, however, acts as a height-dependent added mass and Bernoulli-like pressure, giving a net repulsive force that lowers Gmax and causes sharp drops as successive bending modes become inaccessibl","pith_inferences":["If the saturation formula holds, the practical design rule for a contactless gripper is to match the actuator's force scale to the sheet's elastohydrodynamic force scale: overdriving a soft sheet wastes energy and adds nothing to payload.","The sharp drops in Gmax at Re_bv ≈ 1/e_i² suggest that a load-capacity scan across vibration frequency could act as a modal spectrometer for thin elastic sheets, revealing which bending modes participate in the adhesion.","The compressibility-as-added-weight result implies that reducing ambient pressure (e.g., operating in a vacuum or low-pressure chamber) should strengthen the viscous adhesion, a testable consequence not pursued in the paper.","Since the model neglects solid inertia, real sheets with strong resonances might show either enhanced or suppressed adhesion near resonance; extending the analysis to I_bv > 0 would be the natural next step."],"forward_implications":["For a fixed actuator, there is an optimal bending stiffness B* ≈ 0.05 Fa^(3/2) (µω)^(-1/2); making the sheet softer beyond that point reduces rather than increases the maximum load.","In the strong-forcing regime the equilibrium height becomes linear in the weight, heq ≈ 0.05 G, and at maximum load the hovering height is always approximately 0.3 Hbv, independent of forcing amplitude.","Fluid inertia is destabilizing: in weak forcing the equilibrium Reynolds number Re_eq stays below about 1, whereas in strong forcing it can reach O(100), where the first-order inertial lubrication equation is no longer quantitatively reliable.","Compressibility shifts the adhesion threshold downward as an effective added weight, so experiments in air (Sq ≈ 0.006) see only a modest reduction, while higher-pressure gases or larger gaps would make the reduction significant.","Higher-order bending modes are excited as the sheet approaches the wall, and inertia suppresses the lowest modes one by one, which is why Gmax drops sharply at specific Reynolds numbers."],"supporting_citations":[{"why":"Supplies the weak-forcing asymptotic theory and eigenmode decomposition that the strong-forcing numerical study extends.","marker":"Poulain et al. 2025"},{"why":"Source of the inertial lubrication equation (2.5), adapted here to a fluid layer bounded by two solid walls.","marker":"Rojas et al. 2010"},{"why":"Foundational analysis of compressible squeeze films and the Squeeze number used to quantify compressibility.","marker":"Taylor & Saffman 1957"},{"why":"The experiments of vibration-based adhesion used to set the parameter range and to compare predicted load capacities.","marker":"Weston-Dawkes et al. 2021"},{"why":"Prior analysis of inertial and compressible squeeze films on rigid plates whose attractive-force results the paper contrasts with its viscous mechanism.","marker":"Ramanarayanan et al. 2022"},{"why":"Classical inertial squeeze-film result recovered by the first-order inertial lubrication equation.","marker":"Kuzma 1968"},{"why":"Another classical inertial squeeze-film result that (2.5) recovers, supporting the validity of the first-order correction.","marker":"Jones & Wilson 1975"},{"why":"Source of the edge pressure-loss boundary condition with k = 0.5 used for the inflow pressure.","marker":"Kuroda & Hori 1976"},{"why":"Provides the numerical discretization and solver used for the simulations.","marker":"Koch et al. 2021"}],"fun_headline_variants":["Vibrated-sheet grip maxes out at 7.8x","Inertia and compressibility cut vibrated-sheet grip","Sheet vibration grip saturates; edge contact limits lift","Elastohydrodynamic hover: inertia trims adhesive force"],"cache_read_input_tokens":2688,"weakest_assumption_plain":"The first-order inertial lubrication equation (2.5), together with the edge pressure loss coefficient k = 0.5, correctly represents the repulsive inertial force for deformable gaps up to Reynolds numbers around 50; if higher-order inertia or a different loss coefficient is needed, the predicted reductions in Gmax and the sharp drops would change quantitatively.","fun_headline_variants_meta":{"raw":{"variants":["Vibrated-sheet grip maxes out at 7.8x","Inertia and compressibility cut vibrated-sheet grip","Sheet vibration grip saturates; edge contact limits lift","Elastohydrodynamic hover: inertia trims adhesive force"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000843,"raw_usage":{"total_tokens":3496,"prompt_tokens":723,"completion_tokens":2773,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":467,"completion_tokens_details":{"reasoning_tokens":2700}},"tokens_in":467,"tokens_out":2773,"duration_ms":23050,"temperature":1.0,"reasoning_tokens":2700,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-05T11:42:59.883651+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the maximum load a vibrating sheet supports as a function of the motor's force amplitude for a fixed geometry and frequency. The model predicts that Wmax saturates at about 7.8 Fbv once Fa exceeds roughly 20 Fbv; a continued rise as Fa², a peak followed by a decline, or a saturation value differing from ~7.8 Fbv would contradict the saturation formula (3.4). A separate test would be to vary the fluid pressure: the model says lowering ambient pressure weakens the compressibility-induced repulsion and thus raises Gmax.","supporting_citations":[{"cited_title":", Koch, T","cited_arxiv_id":null,"evidence_quote":"Supplies the weak-forcing asymptotic theory and eigenmode decomposition that the strong-forcing numerical study extends."},{"cited_title":", Argentina, M","cited_arxiv_id":null,"evidence_quote":"Source of the inertial lubrication equation (2.5), adapted here to a fluid layer bounded by two solid walls."},{"cited_title":"& Saffman, P.G","cited_arxiv_id":null,"evidence_quote":"Foundational analysis of compressible squeeze films and the Squeeze number used to quantify compressibility."},{"cited_title":", Adibnazari, I","cited_arxiv_id":null,"evidence_quote":"The experiments of vibration-based adhesion used to set the parameter range and to compare predicted load capacities."},{"cited_title":"1968 Fluid inertia effects in squeeze films","cited_arxiv_id":null,"evidence_quote":"Classical inertial squeeze-film result recovered by the first-order inertial lubrication equation."},{"cited_title":"& Wilson, S.D.R","cited_arxiv_id":null,"evidence_quote":"Another classical inertial squeeze-film result that (2.5) recovers, supporting the validity of the first-order correction."},{"cited_title":"& Hori, Y","cited_arxiv_id":null,"evidence_quote":"Source of the edge pressure-loss boundary condition with k = 0.5 used for the inflow pressure."},{"cited_title":", Weishaupt, K","cited_arxiv_id":null,"evidence_quote":"Provides the numerical discretization and solver used for the simulations."}],"review_version":1}