{"id":"03694a47-e383-4c60-940f-1d0a811585a7","arxiv_id":"2505.06423","paper_version":1,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Global-in-time varifold weak solutions are proved to exist for the two-phase Navier-Stokes-Mullins-Sekerka system with unequal densities, including constant contact angle and a sharp De Giorgi energy inequality.","lead":"Two fluids of different density can now be covered by a new, rigorous notion of solution for a free-boundary model that couples fluid flow with interface motion driven by surface tension and diffusion. The proof shows these solutions exist for all time in two and three dimensions and match the classical equations whenever everything is smooth.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Equation (5.165) in §5.7.8 is a false vector identity; the limit passage to the claimed momentum equation (3.17) therefore fails, so Theorem 3.8 is not established as written.","rationale":"The reader's ACCEPT rests on the assumption that the detailed proof is sound apart from the explicitly conditional consistency theorem. My review identifies a concrete false identity in the derivation of the limit momentum equation, which is closer to the core of Theorem 3.8. The identity (5.165) is used to replace the three advective terms by the single term \\(-\\int\\rho(v\\otimes v):\\nabla\\xi\\). A direct computation shows that, even with \\(\\mathop{\\mathrm{div}}v=\\mathop{\\mathrm{div}}\\xi=0\\), the left and right sides differ by a non-trivial vorticity term \\(\\int \\rho\\, v_i\\xi_j(\\partial_i v_j-\\partial_j v_i)\\). This is not a matter of an unproved but plausible estimate; it is an algebraic error in the main proof. If the limit equation instead contains such an extra term, the triple \\((\\chi,\\mu,v)\\) constructed in Section 5 does not satisfy Definition 3.4(6), and the central global-existence claim is not established. The concern is independent of the conditional assumption in Theorem 3.9(2), which is the only issue the reader flagged. I do not claim the theorem is false; a corrected symmetrization or a different discretization may restore it, but as written the proof has a decisive gap. Therefore the appropriate verdict is REJECT rather than ACCEPT or UNCHANGED.","tokens_in":69182,"tokens_out":33772,"duration_ms":312573,"concrete_test":"Check identity (5.165) directly with \\(\\rho\\equiv1\\), \\(d=2\\), \\(v=(y,0)\\), \\(\\xi=(y,-x)\\), both divergence-free: the left-hand side equals \\(xy\\) and the right-hand side equals \\(0\\). Recompute the limit passage in Section 5.7.8 while keeping the exact discrete advective terms; the surviving difference is \\(\\int \\rho\\, v_i\\xi_j(\\partial_i v_j-\\partial_j v_i)\\,dx\\), which does not vanish for typical Navier-Stokes velocities. If this difference is not zero, the limiting equation contains a vorticity term not present in (3.17).","verdict_should_be":"REJECT","load_bearing_attack":"The proof of Theorem 3.8 reduces the discretized momentum equation to (3.17) by means of identity (5.165): \\(-\\int \\rho(\\xi\\cdot\\nabla)v\\cdot v -\\int \\rho(v\\cdot\\nabla)\\xi\\cdot v +\\int \\rho(v\\cdot\\nabla)v\\cdot \\xi = -\\int \\rho(v\\otimes v):\\nabla\\xi\\). This is not a consequence of integration by parts and the boundary conditions. For divergence-free \\(\\xi\\) and \\(v\\), the difference between the left side and the right side is \\(\\int \\rho\\, v_i\\xi_j(\\partial_i v_j-\\partial_j v_i)\\,dx\\) up to sign, a vorticity coupling that is generically non-zero. Concrete example in two dimensions: \\(v=(y,0)\\), \\(\\xi=(y,-x)\\), \\(\\rho\\equiv1\\). Then \\((\\xi\\cdot\\nabla)v\\cdot v=-xy\\), \\((v\\cdot\\nabla)\\xi\\cdot v=0\\), \\((v\\cdot\\nabla)v\\cdot\\xi=0\\), while \\((v\\otimes v):\\nabla\\xi=0\\), so the claimed identity would give \\(xy=0\\). Since the Navier-Stokes velocity is not irrotational, the extra term cannot be dismissed. The limit equation derived in §5.7.8 may contain an additional term proportional to \\(\\int \\rho\\,\\xi\\cdot(\\omega\\times v)\\), which is absent from Definition 3.4(6). Thus the constructed \\(v\\) is not shown to satisfy the weak Navier-Stokes equation, and the central existence theorem is not supported by the present proof.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper introduces a new weak-solution concept for the sharp-interface Navier–Stokes–Mullins–Sekerka system with different fluid densities and viscosities, in which the interface is represented by an evolving phase indicator and an oriented varifold, and which includes a weak formulation of the constant contact angle, two potentials (a kinetic potential and a curvature potential), and sharp De Giorgi-type energy inequalities. The main result, Theorem 3.8, asserts global-in-time existence of such varifold solutions in dimensions d=2,3 for arbitrary square-integrable divergence-free initial velocities and BV initial phase indicators. Theorem 3.9 states consistency between classical smooth solutions and varifold solutions, in both directions. The existence proof is based on a minimizing-movement scheme with De Giorgi interpolants, combined with a regularized Navier–Stokes equation at each step, and a compactness passage in which a varifold and the velocity are recovered.","tokens_in":69512,"tokens_out":27874,"duration_ms":225508,"significance":"If the proof is correct, the paper provides the first global existence result for a sharp-interface two-phase flow model with unmatched densities in the presence of positive mobility, and it strengthens the previous varifold-solution notion of Abels–Lengeler by adding a sharp energy dissipation principle and a weak contact-angle condition. The solution concept is carefully designed, the paper is largely self-contained, and the technical appendices (Bochner measurability of the selections and an embedding lemma) are valuable. The proof relies on deep external compactness results for varifolds, but no circularity or fitted parameters are present. The main theorem, if established, is a substantial contribution to the free-boundary and two-phase-flow literature.","major_comments":[{"comment":"The identity (5.165), used to pass from the discretized momentum equation to the claimed weak Navier–Stokes equation (3.17), is false. For smooth fields with div v = div ξ = 0 and ξ = 0 on ∂Ω, a correct integration by parts gives LHS − RHS = 1/2∫(∇ρ·ξ)|v|^2 dx + ∫ρ(v·∇)v·ξ dx, which is generically nonzero. In the constant-density case explicitly allowed in Definition 3.4, the left side of (5.165) equals −2∫ρ(v⊗v):∇ξ dx, whereas the right side is −∫ρ(v⊗v):∇ξ dx; the difference is ∫ρ(v·∇)v·ξ dx, a vorticity coupling that does not vanish for divergence-free Navier–Stokes velocities. In the paper's derivation, the first equality in the displayed chain before (5.165) omits the term 1/2∫(∇ρ·ξ)|v|^2 dx and misaccounts for the factor 2. Consequently, the limit equation derived in (5.164)-(5.165) is not shown to coincide with Definition 3.4(6); the constructed pair (v, ρ) may satisfy a different momentum equation, and Theorem 3.8 is not established as written.","section":"§5.7.8, Eq. (5.165)"},{"comment":"The weak formulation (5.50) contains the advective terms with coefficients −1/2, −1/2, and +1, together with a term involving ∂^{•,h}_t ρ_h − ∂_t ρ_h. In the limit (5.164) these appear with coefficients −1, −1, and +1, with no explanation of how the half coefficients and the density-variation term combine. This is not merely cosmetic: the correct limit of the sum of those terms is what must produce the standard term −∫ρ(v⊗v):∇ξ in (3.17). Since the manuscript instead invokes the false identity (5.165), the derivation of the momentum equation is incomplete. The authors should either prove the required cancellation using the continuity equation for ρ or present the corrected limiting equation and show that it reduces to (3.17).","section":"§5.7.8, transition from (5.50) to (5.164)"}],"minor_comments":[{"comment":"The word 'Charathéodory' should be 'Carathéodory'; there are also a few other typographical slips such as 'calssical' in Appendix C.","section":"Appendix B"},{"comment":"The displayed local-Lipschitz bound C/(2ts) becomes singular as s→0; the argument only needs local Lipschitz on (0,h], but the reader would benefit from an explicit statement that s is bounded away from 0 there.","section":"§5.2.2, inequality (5.16)"},{"comment":"The paragraph proving strong convergence of v_h via (5.117)-(5.120) is terse: the use of P_σ(ρ_h v_h) → P_σ(ρ v) and the pointwise convergence of ρ_h and v_h should be expanded slightly, since the identification ζ = P_σ(ρ v) relies on weak continuity of P_σ.","section":"§5.7.1, compactness passage"},{"comment":"The consistency statement for smooth varifold solutions is explicitly conditional on the first variation of µ_t being given by δE[χ(t)], and Remark 3.10 correctly notes that even for Mullins–Sekerka flow such solutions are only known conditionally; this limitation is appropriately stated.","section":"Theorem 3.9(2) and Remark 3.10"}],"recommendation":"major_revision","confidential_remarks":"The reader's report recommending acceptance is too optimistic. The false identity (5.165) affects a load-bearing step of the proof of Theorem 3.8, namely the derivation of the momentum equation in the limit. The error appears to be an algebraic mistake that might be fixable by a correct combination of the advective and density terms, rather than a fundamentally flawed approach, so I do not recommend rejection. However, the manuscript cannot be accepted in its present form; the authors must supply a corrected limit passage and verify that the resulting equation is indeed (3.17)."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"You should know two things before reading: the paper's new two-potential varifold notion and the minimizing-movement construction are substantial and worth studying, but the proof of the central existence theorem contains a concrete algebraic error at equation (5.165). I checked the stress-test note against the text and against explicit Fourier-mode examples on the torus; the note holds up. The claimed identity\n\n-∫ρ(ξ·∇)v·v - ∫ρ(v·∇)ξ·v + ∫ρ(v·∇)v·ξ = -∫ρ(v⊗v):∇ξ\n\nis not an identity. With divergence-free ξ and v (and any honest boundary conditions), the difference is, up to gradient terms that vanish against ξ, a vorticity coupling of the form ∫ρ ξ·(ω×v), which is generically nonzero. The paper's own derivation of (5.165) drops a term proportional to ∫|v|² ξ·∇ρ and gives the wrong factor in front of ∫ρ ξ·(v·∇)v. The stress-test example with v=(y,0), ξ=(y,-x) is not admissible on a bounded domain because ξ does not vanish on ∂Ω, but that is not the core issue; compactly supported and periodic examples show the same failure. Consequently, the limit equation (5.164) cannot be converted to the claimed momentum equation (3.17) by the argument given. The limit velocity may satisfy a different equation with an extra vorticity term absent from Definition 3.4(6). This is not a minor typo; it breaks the proof of Theorem 3.8.\n\nWhat the paper does well: the two-potential framework (kinetic u and curvature w) is a genuine novelty, as is the De Giorgi dissipation inequality and the weak contact-angle condition for unmatched densities. The a priori estimates (5.74)-(5.76) are detailed and the consistency theorem is carefully stated, with the conditional assumption in Theorem 3.9(2) explicitly flagged. The geometric compactness arguments and the use of the flow map are interesting and likely reusable. If the momentum passage can be repaired—perhaps by tracking the extra term and showing it cancels through the continuity equation, or by adjusting the weak formulation—the paper would be a real advance.\n\nThe reader's ACCEPT at moderate confidence is too generous. The soundness score should drop; the central claim is not supported as written. This is a paper for specialists in free-boundary PDE and gradient-flow methods who can assess whether the scheme can be fixed. It deserves a serious referee, but the referee's report should demand a corrected proof of the momentum limit before acceptance. I would not cite the existence theorem in its current form, though I might cite the solution concept if the paper is revised.\n\nRecommendation: send to peer review, but expect major revision. The authors are serious and the framework is promising, but the proof has a real hole.","headline":"The main existence theorem has a load-bearing gap: the vector identity (5.165) used to derive the weak Navier-Stokes equation from the non-conservative formulation is false, so Theorem 3.8 is not established as written.","tokens_in":70021,"tokens_out":19764,"would_cite":false,"duration_ms":165790,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["35R35","35Q30","76D45","76T99","80A20"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper constructs global weak solutions, in a new varifold sense, for two-phase incompressible viscous flow with different densities, including a weak constant contact-angle condition.","keywords":["two-phase flow","Navier-Stokes-Mullins-Sekerka","varifold solutions","unmatched densities","contact angle","minimizing movements","De Giorgi interpolants","sharp energy dissipation"],"falsifier":"Take a smooth varifold solution in a flat domain with a planar interface meeting the boundary at an angle γ′ different from the prescribed γ, and compute the difference δμ_t(B) − δE[χ(t)](B) for a tangential variation B supported near the contact line. If this difference is nonzero, the solution satisfies the weak definition but not the classical contact-angle condition, showing that the extra assumption in Theorem 3.9(2) is genuinely needed.","tokens_in":68987,"feed_emoji":"🌊","tokens_out":5901,"duration_ms":57112,"temperature":0.7,"pith_summary":"The paper proposes a new weak (varifold) solution concept for two immiscible incompressible viscous fluids with different densities and viscosities, coupled to an advected Mullins-Sekerka interface evolution. It claims that for any bounded smooth domain in dimensions two and three, any square-integrable divergence-free initial velocity, and any initial phase configuration with fixed mass, a solution exists for all positive times. The new notion includes a weak formulation of the constant contact angle and sharp De Giorgi-type energy dissipation inequalities that the previous notion lacked. It also proves that every smooth such weak solution coincides with a classical solution of the Navier-Stokes-Mullins-Sekerka system, conditional on a first-variation identity for the varifold. If correct, this gives the first global weak-solution theory for the unequal-density regime that captures energy dissipation and boundary contact angle.","feed_headline":"Unequal-density two-phase flow gets global weak solutions","feed_subtitle":"A varifold formulation with a weak contact-angle condition and sharp energy dissipation extends earlier equal-density existence results.","key_machinery":"The engine of the construction is a minimizing-movement scheme with De Giorgi interpolants: at each time step, the new phase indicator minimizes capillary energy plus an $H^{{-1}}$ distance to the previous phase pulled back along the flow map of a regularized velocity, so that the discrete solution inherits the gradient-flow-like structure of the advective Mullins-Sekerka flow. The admissible limit object is an oriented varifold, a measure on interfaces and directions, split into an interior part and a boundary part supported on ∂Ω with weight cosγ; its first variation supplies the generalized mean curvature and a weak Gibbs-Thomson relation. Two potentials play distinct roles: the kinetic potential u drives the diffusive mass flux (ρ1−ρ2)∇u, while the curvature potential w represents surface tension in the stress balance. The sharp dissipation inequalities (3.19)-(3.20) are obtained by passing the De Giorgi inequalities to the limit.","core_discovery":"The central claim is Theorem 3.8: for d=2,3, a varifold solution exists globally in time for the Navier-Stokes-Mullins-Sekerka system with different densities and viscosities, where the phase indicator and velocity have the regularity stated in (3.13)-(3.14), the interface satisfies a weak Gibbs-Thomson law (3.15), the phase evolution is governed by a kinetic potential through (3.16), the velocity equation (3.17) includes the density-dependent diffusive flux, and the sharp dissipation inequalities (3.19)-(3.20) hold. This is stronger than earlier varifold solutions obtained by sharp-interface limits, because it encodes a weak constant contact-angle condition and a sharp energy-dissipation principle, and it extends the existence theory beyond matched densities. Theorem 3.9 shows consistency: classical solutions are varifold solutions, and smooth varifold solutions satisfying one additional first-variation assumption are classical solutions.","pith_inferences":["Inference: if the sharp energy dissipation inequalities turn out to be stable under limits, one might expect the varifold solution to be unique in a class of sufficiently regular strong solutions, by analogy with recent Mullins-Sekerka results.","Inference: the two-potential split suggests a practical numerical scheme: solve the phase evolution by a minimizing-movement step and update the velocity by a regularized Navier-Stokes step, without resolving the interface.","Inference: the conditional assumption in Theorem 3.9(2) could be probed by trying to construct BV-valued solutions whose varifold first variation equals the first variation of the capillary energy; if such solutions do not exist in general, the consistency theory may need a weaker boundary-contact formulation.","Inference: the flow-map composition used in the discretization may extend to other interface-coupled fluid models, since it cleanly separates material advection from the H^{-1} gradient-flow structure."],"forward_implications":["Global weak solutions now exist for unequal densities in dimensions two and three, with arbitrary fixed initial mass and square-integrable initial velocity.","The weak contact-angle condition holds for any constant angle γ in (0, π/2], covering cases that the earlier varifold notion could not handle, including the matched-density case.","The sharp De Giorgi-type energy dissipation inequalities give a variational structure to the coupled fluid-interface problem, not just to the Mullins-Sekerka part.","The consistency theorem ensures that classical solutions of the original sharp interface system are not lost by the weak formulation.","The notion is designed so that a relative-entropy weak-strong uniqueness result could be attempted, an advantage the paper highlights as motivation."],"supporting_citations":[{"why":"Supplies the varifold-solution framework and De Giorgi-type dissipation that the new notion extends to the advected, unequal-density case.","marker":"[26]"},{"why":"Gives the previous matched-density weak-solution theory for the sharp interface model, which the new definition covers and improves.","marker":"[11]"},{"why":"Provides earlier varifold solutions for unequal densities obtained by sharp-interface limit, used as the comparison for the stronger notion.","marker":"[10]"},{"why":"Introduces the diffuse interface model with different densities from which the sharp interface system is derived, motivating the extra diffusive flux.","marker":"[6]"},{"why":"Supplies the variational treatment of Navier-Stokes with flow maps used in the semi-implicit time discretization.","marker":"[23]"},{"why":"Provides regularity theory for hypersurfaces whose mean curvature is given by an ambient Sobolev function, used to identify the generalized mean curvature.","marker":"[36]"},{"why":"Gives the local-minimization construction for Stefan-type problems yielding an integrable generalized mean curvature and Gibbs-Thomson relation.","marker":"[33]"},{"why":"Supplies the weak-solution theory for Mullins-Sekerka flow that defines the sense of generalized mean curvature used in the admissibility conditions.","marker":"[34]"}],"fun_headline_variants":["Global weak solutions for unequal-density two-phase flow","Varifold existence for two-phase flow with density contrast","Weak contact angle now part of two-phase flow existence","Density contrast no longer blocks global weak solutions"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that, for smooth solutions, the varifold's first variation equals the first variation of the capillary energy including the boundary contact term; this equality is what turns a weak varifold solution into a classical solution with a constant contact angle.","fun_headline_variants_meta":{"raw":{"variants":["Global weak solutions for unequal-density two-phase flow","Varifold existence for two-phase flow with density contrast","Weak contact angle now part of two-phase flow existence","Density contrast no longer blocks global weak solutions"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.002761,"raw_usage":{"total_tokens":10527,"prompt_tokens":959,"completion_tokens":9568,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":575,"completion_tokens_details":{"reasoning_tokens":9506}},"tokens_in":575,"tokens_out":9568,"duration_ms":57030,"temperature":1.0,"reasoning_tokens":9506,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T22:43:42.648487+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take a smooth varifold solution in a flat domain with a planar interface meeting the boundary at an angle γ′ different from the prescribed γ, and compute the difference δμ_t(B) − δE[χ(t)](B) for a tangential variation B supported near the contact line. If this difference is nonzero, the solution satisfies the weak definition but not the classical contact-angle condition, showing that the extra assumption in Theorem 3.9(2) is genuinely needed.","supporting_citations":[{"cited_title":"Hensel and K","cited_arxiv_id":null,"evidence_quote":"Supplies the varifold-solution framework and De Giorgi-type dissipation that the new notion extends to the advected, unequal-density case."},{"cited_title":"Abels and M","cited_arxiv_id":null,"evidence_quote":"Gives the previous matched-density weak-solution theory for the sharp interface model, which the new definition covers and improves."},{"cited_title":"Abels and D","cited_arxiv_id":null,"evidence_quote":"Provides earlier varifold solutions for unequal densities obtained by sharp-interface limit, used as the comparison for the stronger notion."},{"cited_title":"Abels, H","cited_arxiv_id":null,"evidence_quote":"Introduces the diffuse interface model with different densities from which the sharp interface system is derived, motivating the extra diffusive flux."},{"cited_title":"Gigli and S","cited_arxiv_id":null,"evidence_quote":"Supplies the variational treatment of Navier-Stokes with flow maps used in the semi-implicit time discretization."},{"cited_title":"Sch\\\"atzle , Hypersurfaces with mean curvature given by an ambient S obolev function , J","cited_arxiv_id":null,"evidence_quote":"Provides regularity theory for hypersurfaces whose mean curvature is given by an ambient Sobolev function, used to identify the generalized mean curvature."},{"cited_title":"R\\\"oger , Solutions for the S tefan problem with G ibbs- T homson law by a local minimisation , Interfaces Free Bound., 6 (2004), pp","cited_arxiv_id":null,"evidence_quote":"Gives the local-minimization construction for Stefan-type problems yielding an integrable generalized mean curvature and Gibbs-Thomson relation."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the weak-solution theory for Mullins-Sekerka flow that defines the sense of generalized mean curvature used in the admissibility conditions."}],"review_version":1}