{"id":"4f01f3c9-fe02-48e4-bad0-824219ac4fd0","arxiv_id":"2507.10032","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Finite-time blow-up of strong solutions to the 3D free-boundary Euler equations with surface tension must fall into one of five explicitly classified mechanisms, with a vorticity-based refinement in simply connected domains.","lead":"This paper proves a complete classification of how solutions of the 3D incompressible Euler equations with surface tension can blow up in finite time on a moving free boundary. The result removes previous assumptions on symmetry, periodicity, graph structure, and simple connectivity, and identifies five mutually exclusive blow-up mechanisms.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Equation (3.23) in Lemma 3.10 rests on a false Gagliardo–Nirenberg interpolation; without it the energy closure in Proposition 3.11 fails, so Theorem 1.1 is not established.","rationale":"The reader's verdict was CONDITIONAL, motivated by presentation discrepancies in the abstract and by a general concern about uniform geometric/elliptic constants. The stress-test pass found a different, more severe problem: a concrete false interpolation inequality in the proof itself. Equation (3.23) is not a matter of missing regularity or degenerate constants; it is a direct algebraic/scaling error that invalidates the control of the cubic term in Lemma 3.10. That control is essential for the linear-in-E energy inequality (3.25), which is the core of the contradiction argument. Without a replacement estimate, the proof of Theorem 1.1 does not close; the most natural fix would need a cancellation or a different high-order energy structure, neither of which is indicated in the paper. The paper's structural care and citations of well-posedness theory are real, but the flawed step is load-bearing, so the appropriate verdict is REJECT rather than CONDITIONAL. The reader's weakest_assumption was not the same concern, hence the disagreement.","tokens_in":32855,"tokens_out":54449,"duration_ms":548622,"concrete_test":"Run the scaling check: fix φ ∈ C_c^∞(B_1(0)) nonzero and set v_λ(x) = λ^{−1/2} φ(x/λ). As λ→0, compute ∥v_λ∥_{L^6} = O(1), ∥v_λ∥_{H^3} = O(λ^{−2}), and ∥v_λ∥_{W^{3/2,4}}^2 = O(λ^{−5/2}); if the ratio ∥v_λ∥_{W^{3/2,4}}^2 / (∥v_λ∥_{L^6} ∥v_λ∥_{H^3}) diverges, (3.23) is false. Alternatively, check Lemma 3.4 directly for s=3/2, p=4, s1=0, p1=6, s2=3, p2=2 and show no admissible θ exists. If (3.23) fails, recompute Lemma 3.10's bound for ∥∂_i v_j ∂_j v_k ∂_k v_i∥_{H^{1/2}} and verify whether Proposition 3.11's (3.25) still follows; a cubic contribution proportional to E rather than √E breaks the closure.","verdict_should_be":"REJECT","load_bearing_attack":"Equation (3.23) asserts ∥v∥_{W^{3/2,4}}^2 ≤ C∥v∥_{L^6}∥v∥_{H^3}, with ∥v∥_{L^6} subsequently bounded by (3.22). This is not a valid use of the Gagliardo–Nirenberg inequality in Lemma 3.4: taking s1=0,p1=6 and s2=3,p2=2, the conditions 1/p = θ/p1 + (1−θ)/p2 and s = θs1 + (1−θ)s2 cannot simultaneously give (s,p) = (3/2,4); solving 1/4 = θ/6 + (1−θ)/2 forces θ = 3/4, which yields s = 3/4, not 3/2. The failure is visible by scaling: for v_λ(x) = λ^{−1/2} φ(x/λ), we have ∥v_λ∥_{L^6} = O(1), ∥v_λ∥_{H^3} = O(λ^{−2}), but ∥v_λ∥_{W^{3/2,4}}^2 = O(λ^{−5/2}), which is not O(λ^{−2}). This estimate is used to bound the cubic term ∂_i v_j ∂_j v_k ∂_k v_i in ∥∇·D_t^2 v∥_{H^{1/2}} by C∥∇v∥_{L∞}√E. With the correct interpolation the cubic term can only be controlled by a multiple of E (or worse). Inserted into the I2 estimate in Proposition 3.11, this produces a superlinear E^{3/2} contribution, so the Grönwall inequality (3.25) does not close and the contradiction argument for Theorem 1.1 collapses.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper claims a complete classification of finite-time blow-up scenarios for the 3D incompressible free-boundary Euler equations with surface tension, at the H^3 × H^4 regularity level, in bounded domains with closed boundary, without symmetry, periodicity, graph, or simple-connectivity assumptions. Theorem 1.1 asserts that if the maximal existence time T* is finite, at least one of five scenarios must occur: first self-intersection of the free boundary; loss of mean-curvature regularity in H^{3/2} or boundary regularity in H^{2+ε}; loss of H^{5/2} regularity of the normal boundary velocity; L^1_t L^∞ blow-up of the tangential velocity gradient on the boundary; or L^1_t L^∞ blow-up of the interior velocity gradient. Theorem 1.5 refines the interior alternative, under simple connectivity, to a vorticity-based criterion. The proof is by contradiction via a high-order energy functional E(t) involving D_t^2 v, the tangential gradient of D_t v·n, and ∇^2(∇×v), and aims to close a Grönwall inequality under the assumption that all five scenarios are absent.","tokens_in":33243,"tokens_out":12082,"duration_ms":108169,"significance":"If the main theorem were established, it would be a significant advance in the theory of free-boundary Euler equations with surface tension: prior blow-up criteria in this setting either required graph representations, higher (H^6) regularity, or stronger pointwise-in-time control of the velocity gradient. The paper's dynamic-reference-surface methodology and the proposed separation of boundary and interior singularity mechanisms are conceptually interesting, and the authors are to be credited for attempting a general geometric framework. However, the central energy estimate rests on a false interpolation inequality, so Theorem 1.1 is not proved. In addition, the abstract overstates the fixed-boundary BKM recovery: the paper's own equation (1.9) includes an enstrophy term beyond the classical BKM criterion (1.10). These are load-bearing issues, and the manuscript cannot be accepted in its present form.","major_comments":[{"comment":"The inequality ∥v∥²_{W^{3/2,4}} ≤ C∥v∥_{L^6}∥v∥_{H^3} asserted in (3.23) is not a valid Gagliardo–Nirenberg interpolation. Solving 1/4 = θ/6 + (1−θ)/2 forces θ = 3/4, and then the interpolation condition s = θ·0 + (1−θ)·3 gives s = 3/4, not 3/2. A scaling counterexample confirms the failure: for v_λ(x) = λ^{-1/2}φ(x/λ), the left-hand side is of order λ^{-5/2} while the right-hand side is of order λ^{-2}. This estimate is used in Lemma 3.10 to control the cubic term ∂_i v_j ∂_j v_k ∂_k v_i in H^{1/2}; the correct available bound introduces an extra factor ∥∇v∥_{L∞}. Consequently, the I_2 estimate in Proposition 3.11 acquires a superlinear term, and the Grönwall inequality (3.25) does not close. The contradiction argument for Theorem 1.1 therefore collapses.","section":"Section 3, Eq. (3.23)"},{"comment":"The abstract states that the simply-connected refinement 'recovers exactly the classical Beale–Kato–Majda criterion in the fixed-boundary case,' but Section 1.1 and equation (1.9) explicitly state that the fixed-boundary criterion obtained here (limsup_{t→T*}∥∇×v∥_{L²} + ∫_0^{T*}∥∇×v∥_{L∞}dt = ∞) differs from the classical BKM criterion (1.10) by incorporating enstrophy dynamics. The abstract's claim of exact recovery is therefore inaccurate and should be corrected to match the actual theorem.","section":"Abstract and Section 1.1 (Eq. (1.9))"},{"comment":"The proof of Theorem 1.5 inherits the failure of the energy closure from Proposition 3.11: the step from (4.4) to (4.6) relies on the Grönwall inequality (3.33), which depends on the faulty estimate (3.23). Even if the interpolation error were repaired in a way that still yields a linear Grönwall inequality, the additional factor ∥∇v∥_{L∞} would prevent the closure, so the vorticity-based refinement is not established by the arguments presented.","section":"Section 4, proof of Theorem 1.5"}],"minor_comments":[{"comment":"The abstract lists four mechanisms (i)–(iv), whereas Theorem 1.1 lists five scenarios; scenario (4), the L^1_t L^∞ blow-up of the tangential velocity gradient on the boundary, is omitted in the abstract. The numbering should be harmonized.","section":"Abstract vs Theorem 1.1"},{"comment":"In the display after (3.8), the statement 'the unit outer normal vector n ∈ H^{5/2}' should specify the trace space H^{5/2}(∂Ω_t) rather than the domain space, to avoid ambiguity.","section":"Section 3, Eq. (3.8)"},{"comment":"In the estimate for the first term of (3.17), the notation ∥D_t^2 v·n∥_{H^{-1/2}(∂Ω_t)} is used; it would be clearer to state explicitly that this is the dual norm with respect to the H^{1/2}(∂Ω_t) pairing, even though the referenced normal trace theorem justifies the bound.","section":"Section 3, Lemma 3.9"}],"recommendation":"reject","confidential_remarks":"The main theorem is not proved because of the invalid interpolation in (3.23). This is not a local typographical issue: the energy closure genuinely fails, and the proof cannot be repaired by a small modification of the constants. Adding an assumption such as ∫_0^{T*}∥∇v∥²_{L∞}dt < ∞ would weaken the theorem substantially and is not part of the stated result. I therefore recommend rejection. The authors should also correct the abstract's overstatement regarding exact BKM recovery, regardless of the fate of the main proof."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First: the theorem statement is genuinely new. No prior blow-up classification for the 3D free-boundary Euler equations with surface tension covers general bounded domains with closed free boundaries at H^3×H^4 regularity without graph or symmetry assumptions. The five-case taxonomy, with self-intersection separated from curvature and normal-velocity losses, is sensible, and the simply-connected vorticity refinement in Theorem 1.5 is interesting. The framework builds honestly on [31]; the dynamically updated reference surface is a published tool, used as a lemma.\n\nThe paper is also transparent about its trade-offs: Section 1.1 explicitly says the fixed-boundary reduction to the exact BKM criterion (1.10) is not achieved, despite the abstract claiming that. That mismatch, plus the four-versus-five mechanism count in the abstract, are presentation problems that a revision can fix.\n\nThe serious problem is equation (3.23) in Lemma 3.10. The inequality ∥v∥²_{W^{3/2,4}} ≤ C∥v∥_{L⁶}∥v∥_{H³} is not a valid Gagliardo–Nirenberg step. With endpoints s1=0,p1=6 and s2=3,p2=2, GN forces s=3/4, not 3/2. The scaling counterexample is clean: vλ = λ^{−1/2}φ(x/λ) gives LHS ~ λ^{−5/2} while RHS ~ λ^{−2}. So the estimate is false. This is not cosmetic: the bound is used to control the cubic term ∂_i v_j ∂_j v_k ∂_k v_i in the H^{1/2} norm of ∇·D²_t v. Without it, the I2 estimate in Proposition 3.11 returns a superlinear E^{3/2} contribution to the energy derivative, so the Grönwall argument (3.25) does not close. The contradiction argument for Theorem 1.1 collapses at that point.\n\nFor that reason the paper is not acceptable in its present form. I would still send it to a serious referee: the result, if fixed, would be a meaningful advance, and referees should see whether a different interpolation or a modified energy restores closure. But the burden is on the authors. The presentation fixes are easy; the (3.23) gap is not a trivial typo.\n\nRight now I would not cite Theorem 1.1 as a proven result. For a reading group it is a decent case study in checking GN exponents. Bottom line: major revision or conditional reject, with the energy closure as the central issue.","headline":"New classification theorem, but a false interpolation inequality (3.23) breaks the energy closure; needs major revision.","tokens_in":33766,"tokens_out":9348,"would_cite":false,"duration_ms":86348,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["35Q35","35R35","35B44","76B03","76B45"],"pacs":[],"model":"deepseek-v4-flash","headline":"Finite-time blow-up in the free-boundary Euler equations with surface tension is classified into five concrete mechanisms.","keywords":["free boundary problem","incompressible Euler equations","surface tension","finite-time blow-up","blow-up criterion","splash singularity","self-intersection","vorticity"],"falsifier":"Construct or numerically observe a family of solutions approaching a finite time $T^*$ at which all five quantities stay bounded (boundary remains embedded with uniform ball radius, mean curvature in $H^{3/2}$, normal velocity in $H^{5/2}$, and both $L^1_tL^\\infty$ velocity-gradient integrals are finite) yet the $H^3\\times H^4$ norms fail to converge; that would contradict Theorem 1.1. A more targeted check is whether $\\inf_{0\\le t<T^*}R(\\Omega_t)$ can vanish while none of the five alternatives occurs.","tokens_in":32661,"feed_emoji":"🌊","tokens_out":11657,"duration_ms":100681,"temperature":0.7,"pith_summary":"The paper proves that a strong solution of the three-dimensional incompressible Euler equations with surface tension, in a bounded domain with a closed free boundary, cannot reach a finite maximal existence time unless at least one of five concrete mechanisms occurs: the boundary first self-intersects, mean curvature or boundary regularity is lost, the normal boundary velocity loses smoothness, or the velocity gradient accumulates in an integrable way on the boundary or inside the fluid. This matters because earlier blow-up criteria required graphs, symmetries, or periodicity, which cannot describe generic singularities such as folding or pinch-off; the present criterion imposes none of these assumptions. The proof runs by contradiction: if all five mechanisms are absent, a uniform interior and exterior ball radius and elliptic regularity estimates keep an energy functional uniformly bounded, so the solution extends past the supposed blow-up time. For simply connected domains the interior condition is refined to a vorticity condition, and for irrotational flows singularity can occur only at the boundary.","feed_headline":"Blow-up in free-boundary Euler flow has five possible mechanisms","feed_subtitle":"No symmetry or graph assumptions needed; in simply connected domains the interior condition becomes vorticity-based.","key_machinery":"The argument is carried by the energy functional $E(t)=\\frac12\\left(\\int_{\\Omega_t}|D_t^2 v|^2\\,dx+\\int_{\\partial\\Omega_t}|\\bar{\\nabla}(D_t v\\cdot n)|^2\\,dS+\\int_{\\Omega_t}|\\nabla^2(\\nabla\\times v)|^2\\,dx\\right)$, together with an equivalent functional containing $\\|D_t v\\|_{H^{3/2}(\\Omega_t)}$ and $\\|v\\|_{H^3(\\Omega_t)}$. Surface tension enters through the boundary condition $p=H_{\\partial\\Omega_t}$, so the material derivative $D_t v$ behaves like a $3/2$-order spatial derivative, and tracking $D_t^2 v$ together with the boundary term $\\bar{\\nabla}(D_t v\\cdot n)$ produces the differential inequality $\\frac{d}{dt}E(t)\\le C(\\|\\nabla v\\|_{L^\\infty(\\Omega_t)}+\\|\\nabla v\\|_{L^\\infty(\\partial\\Omega_t)}+1)E(t)$. If no blow-up scenario occurs, the uniform ball radius condition $\\inf_{0\\le t<T^*}R(\\Omega_t)>C^{-1}$ and the elliptic estimates of Lemmas 2.6-2.8 keep the constants uniform, the standard exponential-integral bound gives a uniform estimate on $E$ up to $T^*$, and Lemma 2.7 lifts the boundary regularity so the solution can be extended, a contradiction.","core_discovery":"The central claim is Theorem 1.1: for a strong solution with $v\\in C([0,T^*);H^3(\\Omega_t))$ and $\\partial\\Omega_t\\in C([0,T^*);H^4)$, if the maximal existence time $T^*$ is finite then at least one of five alternatives holds: (1) first self-intersection of the free boundary at $t=T^*$; (2) loss of mean curvature regularity in $H^{3/2}$ or of boundary regularity in $H^{2+\\varepsilon}$ for any sufficiently small fixed $\\varepsilon>0$; (3) loss of $H^{5/2}$ regularity of the normal boundary velocity; (4) $\\int_0^{T^*}\\|\\nabla v\\|_{L^\\infty(\\partial\\Omega_t)}\\,dt=\\infty$; or (5) $\\int_0^{T^*}\\|\\nabla v\\|_{L^\\infty(\\Omega_t)}\\,dt=\\infty$. The proof shows these are the only ways a singularity can form, with no symmetry, graph, periodicity, or simple-connectivity assumptions. For simply connected domains, alternative (5) is refined to the vorticity condition $\\limsup_{t\\to T^*}\\|\\nabla\\times v\\|_{L^2(\\Omega_t)}+\\int_0^{T^*}\\|\\nabla\\times v\\|_{L^\\infty(\\Omega_t)}\\,dt=\\infty$; for irrotational flows only the boundary alternatives remain.","pith_inferences":["Editorial: the five quantities suggest a concrete numerical diagnostic for imminent singularity formation, and in simply connected geometries the interior integral can be replaced by a vorticity monitor.","Editorial: the uniform ball radius is the geometric linchpin; a boundary that degenerates into a cusp or near-contact without self-intersection would force one of the regularity-loss alternatives, so the classification implicitly predicts how curvature and injectivity failures must be coupled.","Editorial: the energy and commutator structure is not tied to the specific pressure law, so a similar five-mechanism classification may hold for other surface-tension free-boundary systems such as charged liquid drops or ideal MHD with surface tension.","Editorial: the separation of tangential boundary gradient blow-up from interior gradient blow-up suggests that boundary-layer-type singularity can occur while the interior remains smooth, a distinction that could be tested in numerical simulations."],"forward_implications":["All finite-time singularities of these surface-tension free-boundary Euler solutions are captured by five independent mechanisms, so no additional hidden singularity scenario is needed.","Blow-up criteria can be stated without graph, symmetry, or topology assumptions, so turning, folding, multiply connected domains, and non-graph boundaries are included.","In simply connected domains, interior blow-up is governed entirely by vorticity, and irrotational flows can break only at the free boundary.","In the fixed-boundary case the criterion reduces to a classical vorticity accumulation condition for the velocity gradient.","The gap between $H^{3/2}$ curvature regularity and $H^{2+\\varepsilon}$ boundary regularity is intrinsic to recovering boundary regularity from mean curvature through elliptic estimates."],"supporting_citations":[{"why":"Establishes local well-posedness at the $H^3\\times H^4$ regularity level, defining the solution class and maximal existence time used throughout.","marker":"[49]"},{"why":"Supplies the geometric commutator identities for $D_t$ with $\\nabla$, $\\nabla^2$, the normal, and the second fundamental form that drive the energy estimates.","marker":"[47]"},{"why":"Provides the a priori estimate framework for fluid interface problems on which the energy argument of Proposition 3.11 is built.","marker":"[48]"},{"why":"Supplies the curvature estimates and div-curl elliptic estimates (Lemmas 2.6, 2.8, 2.9) that control boundary and velocity regularity uniformly in time.","marker":"[38]"},{"why":"Provides the $H^3$ div-curl estimate on Sobolev-class domains used in Lemma 2.8 for the highest-order velocity control.","marker":"[14]"},{"why":"Supplies the classical vorticity blow-up criterion that the simply connected fixed-boundary refinement is designed to recover.","marker":"[8]"},{"why":"Supplies the logarithmic estimate for divergence-free fields in simply connected bounded domains used in Lemma 4.1 to turn vorticity control into $W^{1,\\infty}$ control.","marker":"[26]"},{"why":"Supplies the bilinear product inequality used repeatedly to bound nonlinear error terms in the energy and boundary estimates.","marker":"[7]"}],"fun_headline_variants":["Five ways a free-surface Euler flow can blow up","Blow-up in free-boundary Euler: all five mechanisms","No-symmetry blow-up criterion for free-boundary Euler","Surface-tension Euler: every blow-up matches one of five patterns","Free-boundary Euler blow-up: five exhaustive mechanisms"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The argument depends on the free boundary maintaining a uniformly rounded shape, meaning a fixed lower bound on the interior and exterior ball radius, and on the elliptic regularity constants staying uniform in time under that geometric control; if these degenerate before any of the five listed mechanisms occurs, the energy estimate no longer controls the solution.","fun_headline_variants_meta":{"raw":{"variants":["Five ways a free-surface Euler flow can blow up","Blow-up in free-boundary Euler: all five mechanisms","No-symmetry blow-up criterion for free-boundary Euler","Surface-tension Euler: every blow-up matches one of five patterns","Free-boundary Euler blow-up: five exhaustive mechanisms"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00042,"raw_usage":{"total_tokens":2241,"prompt_tokens":1106,"completion_tokens":1135,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":722,"completion_tokens_details":{"reasoning_tokens":1052}},"tokens_in":722,"tokens_out":1135,"duration_ms":13302,"temperature":1.0,"reasoning_tokens":1052,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T17:42:05.410337+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Construct or numerically observe a family of solutions approaching a finite time $T^*$ at which all five quantities stay bounded (boundary remains embedded with uniform ball radius, mean curvature in $H^{3/2}$, normal velocity in $H^{5/2}$, and both $L^1_tL^\\infty$ velocity-gradient integrals are finite) yet the $H^3\\times H^4$ norms fail to converge; that would contradict Theorem 1.1. A more targeted check is whether $\\inf_{0\\le t<T^*}R(\\Omega_t)$ can vanish while none of the five alternatives occurs.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the a priori estimate framework for fluid interface problems on which the energy argument of Proposition 3.11 is built."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the $H^3$ div-curl estimate on Sobolev-class domains used in Lemma 2.8 for the highest-order velocity control."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the logarithmic estimate for divergence-free fields in simply connected bounded domains used in Lemma 4.1 to turn vorticity control into $W^{1,\\infty}$ control."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the bilinear product inequality used repeatedly to bound nonlinear error terms in the energy and boundary estimates."}],"review_version":1}