{"id":"3ab0f2bf-d4c9-4a9e-a866-bf184a1ca4c5","arxiv_id":"2412.02567","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Under an energy convergence hypothesis, solutions of a heterogeneous Allen-Cahn equation with space-dependent wells converge to a BV solution of weighted mean curvature flow, and a weak-strong uniqueness principle holds.","lead":"This paper proves that a diffuse interface model for phase separation with spatially varying wells converges to a weighted mean curvature flow as the interface width goes to zero, under an energy convergence hypothesis. The result gives a rigorous bridge between a heterogeneous Allen-Cahn equation and a sharp interface model, including a Gibbs-Thomson relation and a uniqueness principle.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The motion-law conclusion of Theorem 4.3 rests entirely on the unproved energy convergence hypothesis (4.8); without independent verification, the advertised convergence remains conditional.","rationale":"The reader's weakest assumption is exactly the one I would defend: the energy convergence hypothesis (4.8) is structurally distinct from the claimed convergence and is not derived from the Allen-Cahn dynamics. Without (4.8), the map from diffuse first variations to the sharp motion law can lose mass, so the conclusion of Theorem 4.3 is not established. The paper is honest about this, stating the theorem as conditional, and the surrounding arguments are plausible and largely follow known techniques. I did not find a separate internal inconsistency that would force rejection; the main proof steps after granting (4.8) are fillable. Therefore the verdict should remain conditional rather than unconditional acceptance, and my read does not change the reader's verdict.","tokens_in":854,"tokens_out":1751,"duration_ms":326723,"concrete_test":"Attempt to prove (4.8) from the other hypotheses: derive lim_{epsilon->0} integral_0^T E_epsilon[u_epsilon(t)] dt = integral_0^T E[u_A(t)] dt using only (4.1), (4.7), and the weak Allen-Cahn formulation, and check whether the dissipation limit can be identified with integral integral sigma |V|^2 d|nabla chi_A(t)| dt. If this derivation fails at the point where energy loss at singularities is allowed, then (4.8) is a genuine extra hypothesis. A numerical probe for a dumbbell pinch-off, computing D_epsilon = integral_0^T |E_epsilon[u_epsilon(t)] - E[u_A(t)]| dt, would give concrete evidence whether the gap is real.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Since the abstract advertises a convergence result, the load-bearing point is not the conditional theorem's internal logic but the status of (4.8). The proof of the motion law in Substep 2.2 applies the time-integrated first-variation convergence of Remark 3.4 and the velocity-identification claim (4.13). Both passages use (4.8) to pass to the limit in the cross terms of (4.14), replacing the diffuse potential and kinetic energy densities by the weighted perimeter measure sigma |nabla chi_{A(t)}|. If (4.8) fails, only the lower semicontinuous inequality survives: the diffuse energy may lose mass in the limit, as in Brakke flows with energy loss, and the limiting velocity cannot be identified with weighted mean curvature through (4.4). Thus (4.8) is not a technical convenience; it is exactly the sharp-interface energy balance that the paper assumes rather than derives. Standard Allen-Cahn-to-MCF results do not provide (4.8) for weak solutions past singularities, so the theorem is genuinely conditional. This is the same concern the reader flags, and I see no separate gap after granting (4.8).","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper studies the sharp-interface limit of a heterogeneous Allen-Cahn equation with a space-dependent double-well potential and moving wells. The main results are: (i) a proof that the first variation of the diffuse energy E_ε converges to the first variation of the weighted perimeter E, both in the static case (Theorem 3.1) and, by assertion, in a time-integrated form (Remark 3.4); (ii) a Gibbs-Thomson relation for heterogeneous surface tensions (Corollary 3.2); (iii) a conditional convergence theorem (Theorem 4.3) stating that, under the energy convergence hypothesis (4.8), weak solutions of the heterogeneous Allen-Cahn equation converge, up to subsequence, to a BV/distributional solution of weighted mean curvature flow; and (iv) a weak-strong uniqueness principle for BV solutions of weighted mean curvature flow (Theorem 5.2), proved via the relative energy technique. The static first-variation result relies on a normalization of the moving wells, an equipartition lemma, and a control of the x-derivative of the normalized potential (2.7). The paper is transparent about the conditional nature of Theorem 4.3.","tokens_in":26608,"tokens_out":25035,"duration_ms":228282,"significance":"The paper makes a solid contribution to the variational theory of heterogeneous phase transitions. The static first-variation convergence for moving wells (Theorem 3.1) is new and appears to be proved with essentially complete arguments; the resulting Gibbs-Thomson relation extends earlier results of Luckhaus-Modica to spatially dependent wells. The conditional convergence of the heterogeneous Allen-Cahn equation to weighted mean curvature flow is formulated in the standard Luckhaus-Sturzenhecker framework and extends that framework to heterogeneous potentials. The weak-strong uniqueness result generalizes recent relative-energy arguments to weighted mean curvature flow and is a useful addition to the literature. The authors clearly state the unproved energy convergence hypothesis (4.8) and correctly identify it as the key assumption. The main weakness is that several technically load-bearing passages, especially the time-dependent first-variation convergence and the error estimates in Substep 2.2 of Theorem 4.3, are asserted rather than proved in detail.","major_comments":[{"comment":"The time-integrated first-variation convergence stated in Remark 3.4 is asserted without proof and is used critically in Substep 2.2 to pass to the limit in the right-hand side of (4.12). The same substep also contains an unproved estimate for the first two terms of (4.14), which are claimed to vanish at rate C(∫|n_A−η|^2 d|∇χ_A| dt)^{1/2} by reference to Substep 4.2 of Theorem 3.1. Because these two passages are load-bearing for the motion law (4.4), the authors should provide a complete proof of the time-dependent version of Theorem 3.1, including the measure-convergence statements and the error estimates under the energy convergence hypothesis (4.8).","section":"§4, Substep 2.2, Eq. (4.14) and Remark 3.4"},{"comment":"The convergence result of Theorem 4.3 is conditional on the energy convergence hypothesis (4.8), which is not derived from the Allen-Cahn dynamics. As the proof of Substep 2.2 makes clear, (4.8) is exactly what converts a lower-semicontinuous inequality into the identity needed to identify the velocity; if (4.8) fails, the motion law (4.4) is not established. This is a standard type of hypothesis in the Luckhaus-Sturzenhecker framework, but the title of the paper currently advertises the convergence without qualification. The authors should adjust the title (e.g., \"Conditional convergence ...\") and add a remark in Section 4 discussing the status of (4.8) as a no-energy-loss condition and its relation to the analogous assumptions in [36] and [24].","section":"§4, Theorem 4.3, hypothesis (4.8)"},{"comment":"The hypotheses of Corollary 3.2 do not explicitly include the energy convergence E_ε[u_ε] → E[u_A], which is required to apply Theorem 3.1 in Step 1 of the proof. This convergence follows from the minimizing property and standard Γ-convergence with the mass constraint, but it should be stated either as a hypothesis or as a consequence proved before invoking Theorem 3.1. Without this, the proof of the Gibbs-Thomson relation has an implicit unverified hypothesis.","section":"§3, Corollary 3.2"}],"minor_comments":[{"comment":"The interval for the constraint m appears to be misstated: with a < b, the set (−∫ a dx, −∫ b dx) is empty; it should be (−∫ b dx, −∫ a dx).","section":"§3.3, statement of Corollary 3.2"},{"comment":"In the last inequality of (4.11), the second factor should be (1/2 ∫ σ|φ|^2 d|∇χ_A(t)| dt)^{1/2} rather than (∫ σ|φ|^2 d|∇χ_A(t)| dt)^{1/2}, since by Lemma 3.5 the measure W/ε converges to (1/2)σ|∇χ_A(t)|. The missing factor does not affect the conclusion, but the displayed inequality is not exact.","section":"§4, Substep 2.1, Eq. (4.11)"},{"comment":"The vector measures ν_ε := √(2W_n)γ∇v L^N are claimed to converge in weak-star sense and in total variation to σ∇χ_A 'by Lemma 3.5'. Lemma 3.5 as stated only covers scalar measures; the directional convergence needed for the vector measure is established later in Substep 4.2, but the reader is not pointed to it. Please add a remark or restructure the proof so that this claim is clearly justified.","section":"§3, Theorem 3.1, Step 3"},{"comment":"The sentence 'By Hölder's inequality and Young's inequality, respectively, we see ∫ ε|∇v_ε| dx → 0 and ∫ ε|v_ε||∇v_ε| dx → 0' misattributes the inequalities: both estimates follow from Hölder (or Cauchy-Schwarz), and Young's inequality alone would not give the second o(1).","section":"§3, Theorem 3.1, Step 1"},{"comment":"The weak-strong uniqueness theorem is stated in R^N, while the BV solution concept in Definition 4.2 is formulated on a bounded C^2 domain Ω. The authors should clarify how the BV solution notion is adapted to the whole space (e.g., by extending functions by zero or by stating the R^N analogue of Definition 4.2).","section":"§5, Theorem 5.2 and Definition 5.1"}],"recommendation":"major_revision","confidential_remarks":"The paper is mathematically honest about its conditional main theorem, and the static first-variation result is a genuine contribution. The main reason for major revision is the incompleteness of the proof of the time-dependent first-variation convergence and the associated error estimates that are central to the motion law in Theorem 4.3. With those details supplied, the paper would be a solid fit for this journal. The conditional nature of the gradient-flow convergence and the suggested title change should also be addressed, though these are framing issues rather than mathematical errors."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Plainly: this paper's real content is the static first-variation convergence for heterogeneous Modica-Mortola energies with moving wells. The dynamic theorem is an honest conditional statement, and the weak-strong uniqueness is a competent extension of known relative-energy techniques.\n\nWhat is new: the normalized well W_n and the structural inequality (2.7) are a genuine trick. They let the authors reduce the moving-well problem to fixed wells while retaining spatial dependence, and the proof of Theorem 3.1 is essentially complete. The Gibbs-Thomson corollary is a clean payoff. The paper is transparent that Theorem 4.3 depends on the energy convergence hypothesis (4.8), and it follows the Luckhaus-Sturzenhecker template rather than pretending to derive (4.8).\n\nWhere the soft spots are: (4.8) is load-bearing. The motion-law step (4.13) uses (4.8) to replace diffuse energy densities by the weighted perimeter measure, and the time-integrated first-variation convergence of Remark 3.4 is stated without proof. If (4.8) fails, only the lower-semicontinuous dissipation survives and the velocity cannot be identified with weighted mean curvature. The authors say this is a hypothesis, and that is accurate; but it means the headline convergence is conditional in a way that is standard for this line of work, not resolved here. Minor gaps: the error estimates in (4.13) are sketched, the calibration construction in Section 5 is deferred to [17,34], and Remark 3.4 is asserted. These look fillable. The denominator convergence in Corollary 3.2 is fine once the reduced boundary is nonempty, which follows from the mass constraint.\n\nThe citation pattern is appropriate; the self-citations to [10,34] are to closely related work and not a red flag.\n\nWho should read this: anyone working on sharp-interface limits, phase-field models, or geometric flows with heterogeneous surface tension. The static theorem is a useful tool. The dynamic theorem is a conditional step that the community will likely want to strengthen. It deserves a serious referee; I would accept it for review and probably recommend publication after the (4.8) limitation is made even more prominent and the sketched estimates are filled in.","headline":"Static first-variation convergence is the real contribution; the Allen-Cahn-to-weighted-MCF theorem is honestly conditional on an unproved energy-convergence hypothesis.","tokens_in":27170,"tokens_out":3134,"would_cite":true,"duration_ms":31287,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["35A15","53E10","35D30","35K57","35A02","74N20"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper proves that, under an explicit energy-convergence hypothesis, solutions of the heterogeneous Allen–Cahn equation converge to a distributional solution of weighted mean curvature flow.","keywords":["heterogeneous phase transitions","Gibbs-Thomson relation","sharp interface limit","Allen-Cahn equation","weighted mean curvature flow","BV solutions","weak-strong uniqueness"],"falsifier":"Test the time-dependent first-variation identity (Remark 3.4) along a sequence of weak solutions that satisfies the energy-convergence hypothesis (4.8); if for some smooth test vector field $\\Psi$ the limit of the diffuse first variation differs from the weighted perimeter first variation, then the motion law (4.4) fails and Theorem 4.3 would be refuted.","tokens_in":26193,"feed_emoji":"💧","tokens_out":16435,"duration_ms":152165,"temperature":0.7,"pith_summary":"The paper studies a heterogeneous phase-field model: a diffuse-interface energy $E_\\varepsilon$ whose two potential wells $a(x)$ and $b(x)$ move with the spatial position $x$. It establishes two results for the sharp-interface limit $\\varepsilon\\to0$. First, the first variation of the diffuse interface energy converges to the first variation of a weighted perimeter with surface tension $\\sigma(x)$, and this yields a Gibbs–Thomson relation linking chemical potential to weighted mean curvature. Second, under the explicit hypothesis that the time-integrated diffuse energies converge to the weighted perimeter, weak solutions of the heterogeneous Allen–Cahn equation converge to a distributional (BV) solution of weighted mean curvature flow; the paper also proves a weak-strong uniqueness principle for such solutions. The relevance is that material heterogeneities (nonconstant temperature, composition, or crystalline properties) enter naturally through moving wells, and the results show that the sharp-interface motion remains a well-posed weighted mean curvature flow.","feed_headline":"Heterogeneous Allen–Cahn converges to weighted mean curvature flow","feed_subtitle":"Under one energy-convergence assumption, diffuse phase boundaries follow a weighted mean curvature law.","key_machinery":"The central object is the heterogeneous Modica–Mortola energy $E_\\varepsilon[u] = \\int_\\Omega \\big(\\tfrac{1}{\\varepsilon}W(x,u) + \\tfrac{\\varepsilon}{2}|\\nabla u|^2\\big)\\,dx$, with a double-well potential $W$ whose wells are two spatially moving functions $a(x)<b(x)$. Its $\\Gamma$-limit is the weighted perimeter $E[u] = \\int_\\Omega \\sigma\\,d|\\nabla\\chi_A|$ with surface tension $\\sigma(x)=\\int_{a(x)}^{b(x)}\\sqrt{2W(x,s)}\\,ds$. The key technical device is the normalized well function $W_n(x,v):=W(x,a(x)+\\gamma(x)v)$ with $\\gamma=b-a$, which fixes the wells at $0$ and $1$ while retaining spatial dependence; the crucial estimate $|\\partial_x\\sqrt{W_n}|\\le C\\sqrt{W_n}$ permits comparing the surface tension at neighboring points. The machinery then rests on an equipartition-of-energy lemma (Lemma 3.5), showing that potential and gradient terms contribute equally in the limit, and on the first-variation convergence (Theorem 3.1): for test fields $\\Psi$ with $\\Psi\\cdot n_\\Omega=0$ on the boundary, $\\nabla E_\\varepsilon(u_\\varepsilon)(\\gamma\\nabla v_\\varepsilon\\cdot\\Psi)$ converges to $-\\int_\\Omega \\sigma(\\mathrm{Id}-n_A\\otimes n_A):\\nabla\\Psi\\,d|\\nabla\\chi_A| - \\int_\\Omega \\nabla\\sigma\\cdot\\Psi\\,d|\\nabla\\chi_A|$, the weak form of weighted mean curvature motion. This identity is what converts the diffuse gradient flow into the sharp interface motion, and the time-dependent version (Remark 3.4) is used in the convergence proof.","core_discovery":"The core discovery is that the sharp-interface limit of the heterogeneous Allen–Cahn equation is governed by weighted mean curvature flow with a spatially dependent surface tension. Concretely, Theorem 4.3 shows that if the initial data converge and the time-integrated diffuse energies converge to the weighted perimeter, then along a subsequence the phase fields converge in $L^1$ to $u_A=b\\chi_A+a(1-\\chi_A)$, and the sets $A(t)$ form a distributional solution in the sense of Definition 4.2: they admit a square-integrable normal velocity $V$ satisfying the transport equation, the weak motion law $\\sigma V=\\sigma H-\\nabla\\sigma\\cdot n$, and the optimal dissipation inequality. Corollary 3.2 derives the Gibbs–Thomson relation $\\lambda_0=\\gamma^{-1}(-\\sigma H+\\nabla\\sigma\\cdot n)$ for mass-constrained minimizers, identifying the limit chemical potential with the weighted mean curvature of the interface. Theorem 5.2 establishes weak-strong uniqueness: as long as a smooth calibrated flow exists, any BV solution with the same initial datum coincides with it.","pith_inferences":["The energy-convergence hypothesis (4.8) is the real bottleneck: for physically interesting potentials with strongly varying wells, verifying it from the dynamics alone may be as hard as the sharp-interface limit itself, and the paper does not provide such a verification.","The normalized-well trick, which reduces moving wells to a fixed-well energy landscape, points toward a general method: the same first-variation argument should produce Gibbs–Thomson relations and BV convergence for anisotropic or multiphase generalizations of the energy.","A natural next step is to test whether the derivative control $|\\partial_x\\sqrt{W_n}|\\le C\\sqrt{W_n}$ can be relaxed; if not, it becomes an additional hypothesis separating the heterogeneous problem from the homogeneous one.","The weak-strong uniqueness principle suggests a route to quantitative convergence rates for the heterogeneous Allen–Cahn approximation by a relative-entropy estimate, analogous to known rates in the homogeneous setting."],"forward_implications":["When the energy-convergence hypothesis holds and a smooth calibrated flow exists, the weak-strong uniqueness theorem forces the whole sequence of Allen–Cahn solutions to converge to that smooth flow, not just along a subsequence.","The Gibbs–Thomson relation extends the classical curvature–chemical-potential law to heterogeneous surface tensions, giving a precise meaning to the limit chemical potential of a mass-constrained phase-field minimizer.","The BV solution concept for weighted mean curvature flow is shown to be stable: it satisfies the optimal dissipation inequality and agrees with any smooth flow starting from the same initial set, so it is a viable weak formulation for numerical and analytical purposes.","The result justifies using the numerically convenient diffuse-interface Allen–Cahn model to approximate sharp-interface evolutions in heterogeneous media, provided the interface-width parameter is small and the energy convergence hypothesis is met."],"supporting_citations":[{"why":"Supplies the Γ-convergence of the heterogeneous Cahn–Hilliard energy to the weighted perimeter and the lower bound used in the equipartition lemma.","marker":"[2]"},{"why":"Provides the template for conditional convergence to BV solutions of mean curvature flow via the energy convergence hypothesis.","marker":"[36]"},{"why":"Establishes the Gibbs–Thomson relation in the homogeneous setting, which the paper extends to heterogeneous surface tensions.","marker":"[35]"},{"why":"Introduces the relative energy technique for weak-strong uniqueness of multiphase mean curvature flow, which the paper adapts to the weighted case.","marker":"[18]"},{"why":"Defines BV solutions for mean curvature flow with boundary contact and proves Allen–Cahn approximation and weak-strong uniqueness, guiding Definitions 4.2 and Theorem 5.2.","marker":"[24]"},{"why":"Handles the anisotropic Allen–Cahn equation and weak-strong uniqueness for anisotropic mean curvature flow, providing the technical basis for the relative energy computations and existence of solutions.","marker":"[34]"},{"why":"Provides the BV theory and Reshetnyak continuity theorem used in the proof of first-variation convergence.","marker":"[1]"}],"fun_headline_variants":["Sharp limit of heterogeneous Allen-Cahn is weighted MCF","Moving wells yield weighted mean curvature flow in limit","Diffuse-interface limit: weighted curvature law emerges","Allen-Cahn with spatial wells converges to weighted MCF"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof of the sharp-interface limit assumes that the time-integrated diffuse energies converge to the weighted perimeter of the limiting sets (hypothesis (4.8)); the paper does not derive this convergence from the Allen–Cahn dynamics, so the theorem is conditional on it.","fun_headline_variants_meta":{"raw":{"variants":["Sharp limit of heterogeneous Allen-Cahn is weighted MCF","Moving wells yield weighted mean curvature flow in limit","Diffuse-interface limit: weighted curvature law emerges","Allen-Cahn with spatial wells converges to weighted MCF"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00018,"raw_usage":{"total_tokens":1268,"prompt_tokens":875,"completion_tokens":393,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":491,"completion_tokens_details":{"reasoning_tokens":329}},"tokens_in":491,"tokens_out":393,"duration_ms":5288,"temperature":1.0,"reasoning_tokens":329,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T23:21:10.696482+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Test the time-dependent first-variation identity (Remark 3.4) along a sequence of weak solutions that satisfies the energy-convergence hypothesis (4.8); if for some smooth test vector field $\\Psi$ the limit of the diffuse first variation differs from the weighted perimeter first variation, then the motion law (4.4) fails and Theorem 4.3 would be refuted.","supporting_citations":[{"cited_title":"Bouchitt ´e, Singular perturbations of variational problems arising fr om a two-phase transition model , Applied mathematics & optimization, 21 (1990), pp","cited_arxiv_id":null,"evidence_quote":"Supplies the Γ-convergence of the heterogeneous Cahn–Hilliard energy to the weighted perimeter and the lower bound used in the equipartition lemma."},{"cited_title":"Luckhaus and T","cited_arxiv_id":null,"evidence_quote":"Provides the template for conditional convergence to BV solutions of mean curvature flow via the energy convergence hypothesis."},{"cited_title":"Luckhaus and L","cited_arxiv_id":null,"evidence_quote":"Establishes the Gibbs–Thomson relation in the homogeneous setting, which the paper extends to heterogeneous surface tensions."},{"cited_title":"Fischer, S","cited_arxiv_id":null,"evidence_quote":"Introduces the relative energy technique for weak-strong uniqueness of multiphase mean curvature flow, which the paper adapts to the weighted case."},{"cited_title":"11 1–148","cited_arxiv_id":null,"evidence_quote":"Defines BV solutions for mean curvature flow with boundary contact and proves Allen–Cahn approximation and weak-strong uniqueness, guiding Definitions 4.2 and Theorem 5.2."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Handles the anisotropic Allen–Cahn equation and weak-strong uniqueness for anisotropic mean curvature flow, providing the technical basis for the relative energy computations and existence of solutions."},{"cited_title":"Ambrosio, N","cited_arxiv_id":null,"evidence_quote":"Provides the BV theory and Reshetnyak continuity theorem used in the proof of first-variation convergence."}],"review_version":1}