{"id":"3338df14-7682-4e53-acda-181df74fe07f","arxiv_id":"1908.03681","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A C0 finite element scheme for the quasi-incompressible phase-field model of moving contact lines is shown to conserve mass and dissipate energy, with reported second-order (P1) and third-order (P2) convergence in L2.","lead":"The paper derives a phase-field model for two-phase flow with moving contact lines and large density differences, then builds a mass-conserving, energy-stable finite element solver for it. The interest is in a stable numerical method for simulations where the two fluids have very different densities, such as air bubbles rising in water.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Mass-conservation lemma unsupported: Lemma 3.3 uses test functions ρ_h^{n+1}=ρ(c_h^{n+1}) that are not in the FE space H_h, so exact phase-mass conservation of the fully discrete scheme is not established.","rationale":"Reader's weakest_assumption is the kinetic-energy identity (59): the missing term (1/2)∫ρ^n|u^{n+1}-u^n|^2 turns the equality into an inequality. That is a valid proof gap, and I agree the paper should not be accepted as-is. However, the more serious problem is that Lemma 3.3 is not proved: the test functions chosen are not admissible, and the sign of the flux term taken from (64d) is wrong. Without Lemma 3.3, the fully discrete scheme's exact phase-mass conservation, an advertised advantage of the quasi-incompressible formulation, is unsupported. The energy-stability theorem is also not exactly as stated, but it remains true as an inequality; the numerical convergence-rate claim is weak (two refinement intervals, rates 1.12-1.27 at coarse levels), but it is secondary. Overall the reader's conditional verdict is appropriate; the paper needs a corrected proof or a weakened claim, not a rejection.","tokens_in":22672,"tokens_out":20222,"duration_ms":194646,"concrete_test":"Re-run the P1 Couette-flow simulation of Section 4.1 in double precision and monitor M_c^n = ∫Ω ρ(c_h^n) c_h^n dx at every time step. If M_c^{n+1}-M_c^n is not zero to quadrature/round-off precision (relative residual above ~1e-12) over the full interval, Lemma 3.3 is false as stated and the 'mass conserving' claim must be downgraded to approximate conservation. If it is exactly conserved to round-off, the authors should still replace the proof of Lemma 3.3 with one using admissible test functions from H_h (for instance the FE interpolant of ρ_h^{n+1}), since the current proof is invalid.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The most load-bearing gap is in Lemma 3.3, the proof of exact mass conservation for each phase, one of the two headline properties. The proof sets ψ_h=q_h=ρ_h^{n+1} in (64a) and (64d). But ρ_h^{n+1}=ρ(c_h^{n+1}), with ρ(c)=ρ1ρ2/(ρ2 c+ρ1(1-c)); for P1/P2 elements this rational function of c_h^{n+1} is not a piecewise polynomial and is not an element of the finite-dimensional test space H_h. The Galerkin equations (64) only hold for test functions in H_h, so the two identities on which the cancellation relies are not consequences of the scheme. There is also a sign error: substituting q_h=ρ_h^{n+1} into (64d) gives -(1/α)∫∇ρ_h^{n+1}·u_h^{n+1} = -∫ M(∇μbar+α∇pbar)·∇ρ_h^{n+1}, whereas the manuscript states the RHS with a plus sign; the claimed cancellation of the flux terms therefore does not follow even formally. The reader's concern about the missing kinetic term in Eq. (59) is real, but it only changes an equality into an inequality, so the energy-stability conclusion survives. Here the advertised exact conservation property itself is unsupported, which is more load-bearing.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops a quasi-incompressible phase-field model for two-phase flows with moving contact lines and variable density, derived from an energy variational approach with a generalized Navier boundary condition. The authors then propose a C0 finite element discretization of the reformulated system and claim two headline properties: exact conservation of the mass of each phase (Lemma 3.3) and a fully discrete energy law with non-positive dissipation (Theorems 3.2 and 3.4). Numerical experiments cover Couette-flow convergence studies for low and high density ratios, moving droplets in shear flow, and a rising bubble with density ratio 0.001:1. The abstract further claims second-order convergence for P1 and third-order for P2 elements in the L2 norm.","tokens_in":22915,"tokens_out":7394,"duration_ms":72795,"significance":"If the two headline properties were rigorously established, this would be a useful contribution: a mass-conservative, energy-stable C0 finite element scheme for a thermodynamically consistent quasi-incompressible Navier-Stokes-Cahn-Hilliard system would extend the available toolbox for moving-contact-line simulations to large density ratios. The model derivation is self-consistent and first-principles, and the pressure-stabilized formulation that avoids the inf-sup condition is a genuine practical benefit. The numerical experiments target large density ratios and include diagnostic checks of mass, energy, and the quasi-incompressibility condition, which are appropriate validation targets. However, the advertised properties are currently not proven: the mass-conservation lemma is based on an invalid choice of test functions, and the discrete energy theorem omits a non-negative kinetic term. The significance is therefore conditional on the authors repairing these proofs or modifying the claims.","major_comments":[{"comment":"The proof of exact phase-mass conservation is not valid for the stated Galerkin scheme. Setting ψ_h = q_h = ρ_h^{n+1} in (64a) and (64d) is not an admissible choice because ρ_h^{n+1} = ρ1ρ2/(ρ2 c_h^{n+1}+ρ1(1-c_h^{n+1})) is a rational function of c_h^{n+1}, and for P1/P2 elements it is not a piecewise polynomial and hence not an element of the finite-dimensional test space H_h; equations (64) hold only for test functions in H_h. Moreover, substituting q_h = ρ_h^{n+1} into (64d) gives −∫∇ρ_h^{n+1}·u_h^{n+1} dx = −α∫M(∇μ̄_h^{n+1}+α∇p̄_h^{n+1})·∇ρ_h^{n+1} dx, so after division the displayed identity in the proof has the opposite sign. The claimed cancellation of the flux terms therefore does not follow, and the exact conservation identities (65a)-(65b) are unsupported.","section":"Section 3.2, Lemma 3.3"},{"comment":"Equation (59) omits the non-negative term (1/2)∫_Ω ρ^n |u^{n+1}-u^n|^2 dx that arises when the discrete momentum equation is multiplied by Δt u^{n+1}; the standard identity ∫ρ^n(u^{n+1}-u^n)·u^{n+1} = (1/2)∫ρ^n(|u^{n+1}|^2-|u^n|^2+|u^{n+1}-u^n|^2) shows that this term cannot cancel with the skew-symmetric terms in (49c). Consequently the discrete energy law (55) is not established as an equality; at best the calculation gives the inequality E^{n+1,tot}-E^{n,tot} ≤ −(non-negative dissipation). Since Theorem 3.4 is proved only by reference to the proof of Theorem 3.2, the fully discrete energy law (70) inherits the same gap. The energy-stability conclusion likely survives, but the theorem as stated is not rigorously established.","section":"Section 3.1, Theorem 3.2 and Eq. (59)"},{"comment":"The convergence study does not provide enough evidence for the abstract's claim of second-order (P1) and third-order (P2) convergence in the L2 norm. Each error column reports only two observed rates, and the rates are not stable across the two intervals: for example, the P1 low-density-ratio Err(uy) rates are 1.46 and 2.78, and the P1 high-density-ratio Err(c) rates are 1.12 and 2.52. With only two refinement intervals no asymptotic regime is exhibited, and several rates fall well below the claimed orders. The authors should add at least one more refinement level, report all observed rates in both tables, and either substantiate or temper the convergence claim.","section":"Section 4.1, Tables 1 and 2"}],"minor_comments":[{"comment":"The term written as \"+ 1/2(ρ^{n+1}-ρ^n/Δt + ∇·(ρ^n u^n)))\" has a parenthesis mismatch and does not indicate what is multiplied; it should read something like \"+ (1/2)((ρ^{n+1}-ρ^n)/Δt + ∇·(ρ^n u^n)) u^{n+1}\".","section":"Eq. (49c)"},{"comment":"The text in Section 4.3 uses \"quasi-impressible\" where \"quasi-incompressible\" is meant, and the caption of Fig. 5 says \"L2 norm of u\" while the text and the figure label refer to the L2 norm of ∇·u.","section":"Section 4.3 and Fig. 5"},{"comment":"The convergence tables report errors only at time t = 0.2 and only for the spatial discretization; the manuscript states \"convergence rate in the sense of L2 norm\" without specifying that no temporal convergence study is presented, so the claimed orders should be attributed to the spatial discretization under the chosen time step.","section":"Section 4.1, Tables 1 and 2"},{"comment":"There are several typos and formatting artifacts, including \"immersible\" for \"immiscible\" in the Introduction, \"Appendix Appendix A\" for \"Appendix A\", and mismatched parentheses in Eq. (64c). A careful proofreading pass is needed.","section":"Throughout"}],"recommendation":"major_revision","confidential_remarks":"The manuscript's two headline properties rest on proofs that are currently invalid. The energy-law issue is repairable: adding the missing kinetic term turns the claimed equality into an inequality, which is sufficient for energy stability. The mass-conservation lemma is more serious: the test-function choice is inadmissible and the sign error breaks the formal cancellation, so the exact conservation claim for the stated scheme is not supported. If the authors can either prove conservation with a correct argument or modify the scheme/claim to approximate conservation with supporting numerical evidence, the paper would be publishable. I recommend major revision rather than rejection because the model derivation and the numerical experiments are valuable and the flaws are localized to the analysis of the discrete scheme."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nShort version: this is a serious paper, not a desk reject, but the headline claims outrun the proofs. The genuinely new piece is the thermodynamically consistent q-NSCH model with mass-averaged velocity and a GNBC, plus a C0 finite element scheme that avoids the inf-sup condition thanks to the Δp term. The variational derivation in Section 2 is coherent, and the simulations with density ratio 100:1 are practical evidence that the scheme is implementable. Credit is also due for not fitting parameters to a manufactured benchmark; the reference comparisons are standard.\n\nThe soft spot is the fully discrete analysis, and it is load-bearing. Lemma 3.3 chooses ψ_h = q_h = ρ_h^{n+1} = ρ(c_h^{n+1}) as test functions in (64a) and (64d). But ρ(·) is a rational function, so for P1/P2 c_h^n it is not in H_h; the Galerkin equations simply do not apply to that choice. There is also a sign inconsistency in the same step: substituting into (64d) gives −(1/α)∫∇ρ·u = −∫M(∇μ̄+α∇p̄)·∇ρ, while the proof writes a plus sign on the right. So exact phase-mass conservation, one of the two advertised properties, is unsupported as written. This may be repairable—project ρ into H_h and prove an approximate conservation law—but the exact statement in Lemma 3.3 is not established.\n\nThe energy law has a smaller wound. In the time-discrete proof, Eq. (59) states an equality for the kinetic-energy change without the nonnegative term (1/2)∫ρ^n|u^{n+1}−u^n|^2. That turns the equality into an inequality, which is enough for energy stability. So Theorem 3.2 likely survives as a stability bound, but not with the stated exact form. The fully discrete Theorem 3.4 is one line saying “follow the proof of Theorem 3.2,” so it inherits the issue and does not independently fix Lemma 3.3.\n\nThe convergence section also overclaims. Only two refinement intervals are used for the rates, and several first-step rates are well below the stated second/third order. That is a numerical-report weakness, not a fatal one.\n\nCitation pattern is fine; prior q-NSCH and FEM work is properly acknowledged. This paper is for people building phase-field FEMs for moving contact lines, especially with large density ratios. It does not close the analysis gap, but it gives a usable starting point. I would send it to referees with the instruction to focus on Lemma 3.3; the authors need to fix the test-function argument or soften the claim. With that repaired, I would be comfortable with publication.","headline":"A credible variational model and C0 FEM for variable-density moving contact lines, but the advertised exact mass conservation is not proven as written; energy stability likely survives as an inequality.","tokens_in":23453,"tokens_out":3777,"would_cite":false,"duration_ms":40986,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["65M60","65N30","76D05","76T99"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper develops a C0 finite element scheme for a quasi-incompressible phase-field model of moving contact lines and argues that the scheme conserves the mass of each phase and dissipates the discrete total energy on every time step.","keywords":["energy stability","moving contact line","quasi-incompressible Navier-Stokes-Cahn-Hilliard","variable density","C0 finite element","phase-field method","mass conservation","large density ratio"],"falsifier":"Take one step of the proposed scheme on the Couette test at low density ratio, with a fixed wall, and evaluate both sides of Eq. (70); if the left-hand side is less negative than the right-hand side by approximately $\\frac12\\int_\\Omega \\rho^n |u^{n+1}_h-u^n_h|^2\\,dx$, then the exact equality in the theorem fails.","tokens_in":65,"feed_emoji":"💧","tokens_out":10726,"duration_ms":206181,"temperature":0.7,"pith_summary":"This paper develops a fully discrete $C^0$ finite element scheme for a quasi-incompressible Navier-Stokes-Cahn-Hilliard system with moving contact lines and variable density, and argues that the scheme conserves the mass of each phase exactly and dissipates a discrete total energy on every time step. The model itself is derived from the energy variational method with a mass-averaged velocity, and the boundary condition is a density-extended General Navier-Stokes Boundary Condition in which the density ratio appears explicitly in the contact-line dynamics. If the claims are correct, the scheme gives a practical numerical route to wetting and spreading problems with large density contrasts, and it does so with equal-order $P1$ and $P2$ elements, sidestepping the inf-sup condition. Convergence tests report second-order in $L^2$ for $P1$ and third-order for $P2$ elements at both low and high density ratios.","feed_headline":"C0 finite element scheme conserves mass and energy in moving contact lines","feed_subtitle":"Two-phase flows with density contrasts to 1000:1 stay mass-conservative and dissipative at every timestep.","key_machinery":"The machinery is the discrete energy identity, built on the discrete total energy $E_h^{n,\\mathrm{tot}} = \\int_\\Omega \\frac12 \\rho_h^n |u_h^n|^2\\,dx + \\frac1\\beta \\int_\\Omega \\rho_h^n(\\frac1\\epsilon G(c_h^n)+\\frac\\epsilon2|\\nabla c_h^n|^2)\\,dx + \\frac{\\alpha_w}{\\beta}\\int_{\\partial\\Omega_w} f_w(c_h^n)\\,dS$. The proof tests the four weak equations (64) with the special choices $\\psi_h = \\frac{\\Delta t}{\\beta}\\bar\\mu_h^{n+1}$, $\\chi_h = \\frac{1}{\\beta}(c_h^{n+1}-c_h^n)$, $v_h=\\Delta t u_h^{n+1}$, and $q_h = \\frac{\\Delta t}{\\beta}\\bar p_h^{n+1}$. The discrete chain-rule identities $G(c^{n+1})-G(c^n) = g(c^{n+1},c^n)(c^{n+1}-c^n)$ and $\\rho^{n+1}-\\rho^n = -\\alpha\\rho^{n+1}\\rho^n(c^{n+1}-c^n)$ convert the chemical-potential products into exact differences of the mixing energy, while the cancellation between the pressure term in the momentum equation and the quasi-incompressibility equation removes the pressure. The same algebraic structure gives the componentwise mass conservation in Lemma 3.3 by taking the phase equation and the quasi-incompressibility equation together with special test functions.","core_discovery":"The central claim is that a natural fully discrete finite element discretization of the quasi-incompressible NSCH system preserves the physics that matter in phase-field wetting simulations: the total mass and the mass of each component are conserved, and the discrete total energy satisfies $E_h^{n+1,\\mathrm{tot}} - E_h^{n,\\mathrm{tot}} = -\\Delta t D_h^{n+1} - \\Delta t \\int_{\\partial\\Omega_w} (1/l_s Re) u_s^{n+1}\\cdot u_w dS$, where $D_h^{n+1}$ is a non-negative sum of bulk viscous, compressible, chemical-potential-gradient, wall mobility, and slip dissipation terms. With a fixed wall the right-hand side is non-positive, so the scheme is energy stable at the fully discrete level. The construction is made possible by the quasi-incompressibility constraint $\\nabla\\cdot u = \\alpha\\nabla\\cdot(M\\nabla\\tilde\\mu)$ with $\\alpha=(\\rho_2-\\rho_1)/(\\rho_1\\rho_2)$, whose pressure term $\\alpha p$ in the chemical potential acts as a pressure stabilization and removes the need for inf-sup compatible velocity-pressure spaces. Numerical simulations of Couette flow, droplets in shear flow, and a rising bubble with density ratio $1000{:}1$ are presented as evidence that the scheme is mass-conservative, energy-dissipative, and second or third-order convergent.","pith_inferences":["A consequence the paper leaves implicit: the density-dependent boundary term $L(c)=\\epsilon\\rho\\partial_n c+\\alpha_w df_w/dc$ suggests a testable prediction, namely that contact-line dynamics should change with density ratio even when surface energy parameters are held fixed.","The scheme's pressure-stabilization mechanism resembles pseudo-compressibility methods, so a natural test is whether the same $C^0$ formulation extends to three-dimensional moving-contact-line problems and to surfactant-laden variants of the quasi-incompressible NSCH system; the paper does not present such extensions.","If the missing non-negative kinetic-energy term $\\frac12\\int_\\Omega \\rho^n |u^{n+1}-u^n|^2\\,dx$ is restored in the proof, the asserted discrete energy identity becomes a dissipation inequality, so Theorem 3.2 and the inherited fully discrete Theorem 3.4 are most safely read as an energy inequality rather than an exact equality.","The fully discrete theorem is stated without an independent proof; its proof is deferred to following Theorem 3.2, so any gap in the time-discrete energy identity propagates directly to the finite element energy law."],"forward_implications":["For fixed walls, every computed time step satisfies $E_h^{n+1,\\mathrm{tot}}\\le E_h^{n,\\mathrm{tot}}$, giving a nonlinear stability guarantee that prevents unbounded energy growth in long wetting and spreading runs.","The componentwise mass conservation means each phase's total mass stays fixed at machine precision across the interface, avoiding the spurious mass loss that can distort phase-field interface motion.","Because the quasi-incompressible $\\Delta p$ term plays the role of pressure stabilization, equal-order $P1$ and $P2$ velocity-pressure spaces are sufficient, which simplifies code design relative to schemes requiring inf-sup stable pairs.","The density-extended GNBC makes the contact-line slip and phase-field boundary flux depend on $\\alpha=(\\rho_2-\\rho_1)/(\\rho_1\\rho_2)$, so simulations with different density contrasts but identical surface tensions are expected to exhibit different contact-line speeds.","The reported convergence rates imply the scheme is usable for quantitative interface tracking: $P1$ gives roughly second order and $P2$ roughly third order in $L^2$ for velocity and phase variable in the tested Couette setup."],"supporting_citations":[{"why":"Defines quasi-incompressible Cahn-Hilliard fluids and the mass-averaged-velocity framework from which the model's density relation and quasi-incompressibility condition are taken.","marker":"[34]"},{"why":"Supplies the primitive-variable and projection finite-difference schemes for quasi-incompressible NSCH with a discrete energy law; the paper's algebraic structure and Lemma 3.1's discrete identities follow this work.","marker":"[42]"},{"why":"Gives the earlier numerical method for quasi-incompressible Cahn-Hilliard-Navier-Stokes with a discrete energy law and the positivity assumption on $\\rho$ that the finite element analysis inherits.","marker":"[55]"},{"why":"Introduces the generalized Navier boundary condition from a variational approach for moving contact lines; the paper's density-extended GNBC is a modification of this condition.","marker":"[14]"},{"why":"Provides a variable-density moving-contact-line scheme with volume-averaged velocity used as a comparison baseline and as a source of the Couette convergence test.","marker":"[47]"},{"why":"Provides another volume-averaged-velocity variable-density phase-field moving-contact-line scheme whose results and convergence behavior the paper compares with.","marker":"[49]"},{"why":"Presents the thermodynamically consistent diffuse-interface model for two-phase flows with different densities, the main alternative modeling route discussed in the introduction.","marker":"[41]"}],"fun_headline_variants":["Energy-stable C0 FEM for variable-density moving contact lines","Mass-conserving C0 scheme for quasi-incompressible two-phase flow","No inf-sup stress: C0 FEM keeps contact line flow stable","Fully discrete energy stability for moving contact line flow","C0 discretization: mass-conserving, energy-dissipative for wetting"],"cache_read_input_tokens":25600,"weakest_assumption_plain":"The discrete energy law is proved as an exact equality, but the proof of the kinetic-energy update in Eq. (59) omits the non-negative term $\\frac12\\int_\\Omega \\rho^n |u^{n+1}-u^n|^2\\,dx$, so the equality as stated is not rigorously established unless that term cancels.","fun_headline_variants_meta":{"raw":{"variants":["Energy-stable C0 FEM for variable-density moving contact lines","Mass-conserving C0 scheme for quasi-incompressible two-phase flow","No inf-sup stress: C0 FEM keeps contact line flow stable","Fully discrete energy stability for moving contact line flow","C0 discretization: mass-conserving, energy-dissipative for wetting"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000421,"raw_usage":{"total_tokens":2169,"prompt_tokens":955,"completion_tokens":1214,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":571,"completion_tokens_details":{"reasoning_tokens":1119}},"tokens_in":571,"tokens_out":1214,"duration_ms":11041,"temperature":1.0,"reasoning_tokens":1119,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T14:05:52.705706+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take one step of the proposed scheme on the Couette test at low density ratio, with a fixed wall, and evaluate both sides of Eq. (70); if the left-hand side is less negative than the right-hand side by approximately $\\frac12\\int_\\Omega \\rho^n |u^{n+1}_h-u^n_h|^2\\,dx$, then the exact equality in the theorem fails.","supporting_citations":[{"cited_title":"Lowengrub, L","cited_arxiv_id":null,"evidence_quote":"Defines quasi-incompressible Cahn-Hilliard fluids and the mass-averaged-velocity framework from which the model's density relation and quasi-incompressibility condition are taken."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the primitive-variable and projection finite-difference schemes for quasi-incompressible NSCH with a discrete energy law; the paper's algebraic structure and Lemma 3.1's discrete identities follow this work."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives the earlier numerical method for quasi-incompressible Cahn-Hilliard-Navier-Stokes with a discrete energy law and the positivity assumption on $\\rho$ that the finite element analysis inherits."},{"cited_title":"Gao, X.-P","cited_arxiv_id":null,"evidence_quote":"Provides a variable-density moving-contact-line scheme with volume-averaged velocity used as a comparison baseline and as a source of the Couette convergence test."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides another volume-averaged-velocity variable-density phase-field moving-contact-line scheme whose results and convergence behavior the paper compares with."}],"review_version":1}