pith. machine review for the scientific record. sign in

arxiv: 2604.27999 · v1 · submitted 2026-04-30 · ⚛️ physics.flu-dyn

Recognition: unknown

Mixture-aware closure of the N-phase Navier--Stokes--Cahn--Hilliard mixture model

M.F.P. ten Eikelder , A. Brunk
Authors on Pith no claims yet

Pith reviewed 2026-05-07 07:35 UTC · model grok-4.3

classification ⚛️ physics.flu-dyn
keywords multiphase mixturesphase-field modelsNavier-Stokes-Cahn-Hilliardmixture consistencyMaxwell-Stefan diffusionthermodynamic closurereduction consistencydiffuse interface
0
0 comments X

The pith

PDE-level consistency when merging identical phases uniquely fixes the free-energy and mobility closures in N-phase Navier-Stokes-Cahn-Hilliard models.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper shows that requiring the governing PDEs to stay unchanged when physically identical phases are combined at the continuous level, together with basic thermodynamic axioms, determines the precise admissible form of the free-energy density and the Onsager mobility matrix. Existing constructions often enforce consistency only when a phase vanishes, but leave the model sensitive to how the same mixture is partitioned into phases. If the result holds, it supplies a systematic construction rule for N-phase diffuse-interface models that reduce correctly to fewer phases without ad-hoc adjustments, and it recovers Maxwell-Stefan diffusion as a special case. This removes a source of ambiguity in multiphase-flow simulations and in bulk diffusion systems.

Core claim

Under a small set of structural axioms, PDE-level reduction consistency uniquely fixes the admissible free-energy structure to an ideal-mixing contribution, a symmetric mean-field interaction term, and a constant-coefficient quadratic gradient penalty, while constraining the Onsager mobility matrix to a pairwise-exchange form with bilinear degeneracy in the volume fractions. The resulting thermodynamic closure includes Maxwell-Stefan-type mobilities as a special case and supplies a consistent closure for the full N-phase Navier-Stokes-Cahn-Hilliard system as well as for multiphase Maxwell-Stefan diffusion in the bulk-only setting.

What carries the argument

The mixture-aware constitutive closure, defined by the requirement that the governing equations remain invariant under merging of identical phases at the PDE level; this requirement selects the free-energy functional and fixes the Onsager matrix structure.

If this is right

  • The constructed model reduces exactly to any lower-phase sub-model obtained by merging identical phases, without changing the equations.
  • Maxwell-Stefan diffusion emerges as a special case of the mobility matrix without additional assumptions.
  • Numerical discretizations of the N-phase system inherit the reduction property, so merged-phase computations remain consistent with the original physics.
  • The same closure applies directly to bulk multiphase diffusion problems without the Navier-Stokes coupling.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The uniqueness result suggests that any alternative closure violating the merging invariance would either break thermodynamic structure or fail to be mixture-aware.
  • The framework could be extended by relaxing one axiom at a time to test which physical regimes require additional terms.
  • In applications where phases can be redefined, the closure guarantees that simulation outcomes are independent of the chosen phase partitioning.
  • The bilinear degeneracy in the mobility matrix may simplify certain analytical reductions to two-phase limits.

Load-bearing premise

The small set of structural axioms that encode thermodynamic admissibility and mixture awareness must be accepted; if they are too narrow or exclude relevant physics, the uniqueness conclusion does not apply.

What would settle it

An explicit free-energy density or mobility matrix that satisfies the structural axioms yet produces altered governing equations after two identical phases are merged would falsify the uniqueness result.

Figures

Figures reproduced from arXiv: 2604.27999 by A. Brunk, M.F.P. ten Eikelder.

Figure 1
Figure 1. Figure 1: Gibbs triangle (ternary diagram) for 𝑁 = 3. Vertices correspond to pure states (𝜙𝛼 = 1), edges to binary mixtures (one phase identically zero), and the interior to fully ternary states. The simplex boundaries 𝜙𝛼 = 0 and vertices 𝜙𝛼 = 1 are indicated; the point (𝜙1, 𝜙2, 𝜙3) = (0.2, 0.5, 0.3) is shown as an example. A3. Constant-capillarity. We restrict attention to a constant-capillarity constitutive class … view at source ↗
Figure 2
Figure 2. Figure 2: Visualization of the merge (reduction) map for a ternary composition. Left: Gibbs simplex for 𝑁 = 3 with two families of lines. The orange dotted lines are level sets 𝜙2 = const; along each such line the reduced variables (𝜙ˆ 1, 𝜙ˆ 2) = (𝜙1 + 𝜙3, 𝜙2) are constant, i.e. these lines identify all (𝜙1, 𝜙2, 𝜙3) corresponding to the same reduced two-phase state. The blue dashed rays emanating from the vertex 𝜙2 … view at source ↗
Figure 3
Figure 3. Figure 3: Redistribution lines in the ternary Gibbs simplex along which the reduction-consistency constraints are imposed. Each colored family of lines collects states that correspond to the same reduced composition after merging two phases: orange dotted lines are level sets 𝜙2 = const (merging phases 1 and 3, so 𝜙ˆ 1 = 𝜙1 + 𝜙3 = 1 − 𝜙2 is fixed); green dash–dot lines are level sets 𝜙3 = const (merging phases 1 and… view at source ↗
Figure 4
Figure 4. Figure 4: Homogeneous binary free-energy densities as functions of 𝜙 := 𝜙1 where 𝜀0 = 1, 𝜒 = 2.07532, 𝑊¯ = 0.691952. The mixture-aware energies are shown after a constant shift Ψ˜ 0 = Ψ0+𝑏, 𝑏 = c𝑜𝑛𝑠𝑡, so that their scale roughly matches that of the Ginzburg–Landau potential. (a) Comparison of the Ginzburg–Landau potential 4𝜙 2 (1 − 𝜙) 2 with the mixture-aware bulk free energy for identical and different phases (b) E… view at source ↗
Figure 5
Figure 5. Figure 5: Homogeneous ternary free-energy densities on the Gibbs simplex for three distinct phases and 𝜀0 = 1, 𝑊¯ = 1, and 𝜒 = 2.0722. (a) Mixture-aware bulk free energy Ψ0. (b) Entropic contribution. (c) Enthalpic contribution view at source ↗
Figure 6
Figure 6. Figure 6: Homogeneous ternary free-energy densities on the Gibbs simplex for three distinct phases and 𝜀0 = 1, 𝑊¯ = 1, and 𝜒 = 2.0722. (a)–(c) Partial free energies Ψ0,1, Ψ0,2, and Ψ0,3. for three distinct phases, together with its entropic and enthalpic parts, is shown in view at source ↗
Figure 7
Figure 7. Figure 7: Homogeneous ternary free-energy densities on the Gibbs simplex in the case where phases 1 and 2 are identical, and 𝜀0 = 1, 𝑊¯ = 1, 𝜒 = 2.0722. (a) Mixture-aware bulk free energy Ψ0. (b) Entropic contribution. (c) Enthalpic contribution. ϕ = (1,0,0) ϕ = (0,1,0) ϕ = (0,0,1) -2.15 -1.65 -1.15 -0.65 -0.15 0.35 (a) Free energy Ψ0,1 ϕ = (1,0,0) ϕ = (0,1,0) ϕ = (0,0,1) -2.15 -1.65 -1.15 -0.65 -0.15 0.35 (b) Free … view at source ↗
Figure 8
Figure 8. Figure 8: Homogeneous ternary free-energy densities on the Gibbs simplex in the case where phases 1 and 2 are identical, and 𝜀0 = 1, 𝑊¯ = 1, 𝜒 = 2.0722. (a)–(c) Partial free energies Ψ0,1, Ψ0,2, and Ψ0,3. = (1, 0, 0) = (0, 1, 0) = (0, 0, 1) 2.10 2.05 2.00 1.95 1.90 1.85 1.80 1.75 (a) Different phases = (1, 0, 0) = (0, 1, 0) = (0, 0, 1) 2.6 2.4 2.2 2.0 1.8 (b) Equal phases view at source ↗
Figure 9
Figure 9. Figure 9: Heat maps of the homogeneous ternary free energies:(a) different phases, (b) identical phases view at source ↗
Figure 10
Figure 10. Figure 10: Calibration of the free energy. are shown in Figure 10a, and a comparison between GL, classical FH, and the polynomial surrogate is provided in Figure 10b. 4.3. Mobility tensor In Section 3.3 we derived a reduction-consistent mobility class and, for practical use, we restrict here to an isotropic mobility tensor of the form M𝛼𝛽 (ϕ) = 𝑀𝛼𝛽 (ϕ)I, 𝛼, 𝛽 = 1, . . . , 𝑁, (4.12) with a scalar mobility matrix 𝑀𝛼𝛽 … view at source ↗
Figure 11
Figure 11. Figure 11: Schematic representation of the Hysing et al. (2009) rising bubble problem Absent phase Here we test the scenario of an absent phase. We choose the initial phase fields as: 𝜙 ℎ 1,0 (x) = 1 2 view at source ↗
Figure 12
Figure 12. Figure 12: Mixture-aware simulations – absent phase. Case 1 (left) and Case 2 (right). Visualization of the phase fields (top to below) 𝜙1, 𝜙2, 𝜙3 at times 𝑡 = 0.0, 0.6, 1.2, 1.8, 2.4, 3.0 (left to right) view at source ↗
Figure 13
Figure 13. Figure 13: Mixture-aware simulations – equal phases. Case 1 (left) and Case 2 (right). Visualization of the phase fields (top to below) 𝜙1, 𝜙2, 𝜙3 at times 𝑡 = 0.0, 0.6, 1.2, 1.8, 2.4, 3.0 (left to right) view at source ↗
Figure 14
Figure 14. Figure 14: Mixture-aware simulations – phases above each other. Case 1 (left) and Case 2 (right). Visualization of the phase fields (top to below) 𝜙1, 𝜙2, 𝜙3 at times 𝑡 = 0.0, 0.6, 1.2, 1.8, 2.4, 3.0 (left to right). Vertically separated identical phases Here we again test the merging of identical phases. We choose the initial phase fields as: 𝜙 ℎ 1,0 (x) = 1 2 view at source ↗
Figure 15
Figure 15. Figure 15: Rising bubble in two stratified liquid layers. Time snapshots of the phase fields (iso-contours) for 𝑟 = 𝑟1 > 𝑟 𝑝. The computations are performed in an axisymmetric mesh with 64×480 elements with 𝜖 = ℎ. The initial condition consists of (i) a diffuse but nearly planar liquid–liquid interface (transition between phases 2 and 3) and (ii) a diffuse spherical bubble of radius 𝑟 (phase 1) embedded in phase 2 b… view at source ↗
Figure 16
Figure 16. Figure 16: Rising bubble in two stratified liquid layers. Time snapshots of the phase fields (iso-contours) for 𝑟 = 𝑟2 > 𝑟 𝑝. Figures 15-16 show time snapshots of the phase distribution for the two radii. For the smaller bubble (𝑟 = 𝑟1), the bubble rises until it reaches the liquid–liquid interface, where it deforms the interface but remains trapped (no passage into the upper layer). For the larger bubble (𝑟 = 𝑟2), … view at source ↗
Figure 17
Figure 17. Figure 17: Quaternary droplet-bubble simulation: side view. Time snapshots of the phase fields (iso-contours). 64 × 256 elements. We impose no-penetration boundary conditions on the velocity on the outer boundary, and homogeneous Neumann (no-flux) boundary conditions for the phase fields and chemical potentials (i.e. no diffusive mass flux through the boundary). Figures 17-18 show time snapshots of the phase distrib… view at source ↗
Figure 18
Figure 18. Figure 18: Quaternary droplet-bubble simulation: bottom view. Time snapshots of the phase fields (iso￾contours). contribution together with a symmetric pairwise interaction term and a constant quadratic gradient penalty. For the mobility, the same principle yields a pairwise-exchange Onsager structure with bilinear degeneracy in the volume fractions. Thus, rather than proposing one possible closure, the paper identi… view at source ↗
read the original abstract

Diffuse-interface (phase-field) models are widely used to describe multiphase mixtures and their interfacial dynamics. In multiphase settings, however, the constitutive closure should remain meaningful across different representations of the same mixture. Existing N-phase phase-field constructions commonly enforce reduction only when a phase is absent (restriction to a face of the Gibbs simplex), but do not address the natural requirement that physically identical phases can be merged without changing the governing equations. This requires characterizing thermodynamically admissible, mixture-aware constitutive closures that are consistent with merging identical phases at the PDE level. Here, we show that, under a small set of structural axioms, PDE-level reduction consistency uniquely fixes the admissible free-energy structure to an ideal-mixing contribution to an ideal-mixing contribution, a symmetric mean-field interaction term, and a constant-coefficient quadratic gradient penalty. yielding a thermodynamic closure that includes Maxwell--Stefan-type mobilities as a special case. The same requirement constrains the Onsager mobility matrix to a pairwise-exchange form with bilinear degeneracy in the volume fractions, yielding a thermodynamic closure that includes Maxwell--Stefan-type mobilities as a special case. These results provide a consistent closure for N-phase Navier--Stokes--Cahn--Hilliard mixture models and, in the bulk-only setting, for multiphase Maxwell--Stefan diffusion systems. Numerical experiments confirm the predicted mixture-aware reduction properties and illustrate the capabilities of the N-phase Navier--Stokes--Cahn--Hilliard framework in representative multiphase-flow computations.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

3 major / 3 minor

Summary. The manuscript develops a thermodynamically consistent, mixture-aware constitutive closure for the N-phase Navier--Stokes--Cahn--Hilliard system. Under a small set of structural axioms (Onsager dissipation structure plus PDE-level invariance under merging of physically identical phases), it proves that the admissible free-energy density is uniquely fixed to the sum of an ideal-mixing term, a symmetric mean-field interaction, and a constant-coefficient quadratic gradient penalty. The same axioms constrain the Onsager mobility matrix to a pairwise-exchange form with bilinear degeneracy in the volume fractions. The resulting closure includes Maxwell--Stefan-type mobilities as a special case. Numerical experiments are reported to confirm the predicted reduction properties under phase merging and to demonstrate the framework on representative multiphase-flow problems.

Significance. If the uniqueness result holds, the paper supplies a canonical, parameter-free closure that resolves a long-standing consistency gap in multiphase diffuse-interface modeling: existing N-phase constructions typically enforce reduction only when a phase is absent, but not when identical phases are merged. The derived forms are directly usable for both the full Navier--Stokes--Cahn--Hilliard system and the bulk Maxwell--Stefan diffusion limit. The work therefore strengthens the theoretical foundation for reliable N-phase simulations and provides a clear route to thermodynamically admissible closures that respect the geometry of the Gibbs simplex.

major comments (3)
  1. [§3, Theorem 3.2] §3, Theorem 3.2 (free-energy uniqueness): The proof that merging invariance forces the gradient penalty to be a constant-coefficient quadratic form appears to presuppose a local, quadratic dependence on ∇ϕ; it is not immediately clear whether the same conclusion follows for more general (e.g., non-local or higher-order) interfacial energies that still satisfy the structural axioms and the simplex constraint. A brief remark on the regularity class assumed for the free-energy density would strengthen the claim.
  2. [§4, Eq. (4.12)] §4, Eq. (4.12) (mobility matrix): The bilinear degeneracy M_{ij} ∝ ϕ_i ϕ_j is shown to be compatible with pairwise exchange and non-negative dissipation, but the argument that it is the only admissible form under the merging axiom relies on the specific algebraic structure of the Onsager matrix on the simplex boundary. An explicit verification that no other degenerate matrices (e.g., with higher-order vanishing) satisfy the axioms would close a potential gap.
  3. [§5] §5 (numerical validation): The experiments illustrate merging consistency, yet the reported reduction error is only qualitative. Quantitative tables or plots showing the L^2 or energy-norm difference between the N-phase and (N-1)-phase solutions when two phases are merged would make the numerical support for the theoretical uniqueness statement more convincing.
minor comments (3)
  1. [Abstract] Abstract, line 3: the phrase “to an ideal-mixing contribution to an ideal-mixing contribution” is duplicated; please correct.
  2. [§2.1] §2.1: The structural axioms are introduced in prose; listing them explicitly as A1–A3 (or similar) would improve readability and allow readers to trace each step of the uniqueness proof back to a numbered assumption.
  3. [Throughout] Notation: The chemical-potential vector μ and the mobility matrix M are used with slightly varying index conventions between the free-energy and mobility sections; a single consistent notation table would help.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for the thorough review and the constructive suggestions. The comments have helped us clarify several aspects of the manuscript. We address each major comment below and have incorporated revisions accordingly.

read point-by-point responses

  1. Referee: [§3, Theorem 3.2] §3, Theorem 3.2 (free-energy uniqueness): The proof that merging invariance forces the gradient penalty to be a constant-coefficient quadratic form appears to presuppose a local, quadratic dependence on ∇ϕ; it is not immediately clear whether the same conclusion follows for more general (e.g., non-local or higher-order) interfacial energies that still satisfy the structural axioms and the simplex constraint. A brief remark on the regularity class assumed for the free-energy density would strengthen the claim.

    Authors: We appreciate this point. The structural axioms and the PDE-level reduction consistency are applied within the standard class of local free-energy functionals that are quadratic in the gradients, as is conventional in Cahn-Hilliard-type models to ensure the variational structure and well-posedness. Non-local or higher-order energies would constitute a different modeling framework, requiring adapted axioms. To address the referee's suggestion, we have added a clarifying remark in Section 3 on the assumed regularity class of the free-energy density, specifying that it is local and quadratic in the gradients. revision: yes

  2. Referee: [§4, Eq. (4.12)] §4, Eq. (4.12) (mobility matrix): The bilinear degeneracy M_{ij} ∝ ϕ_i ϕ_j is shown to be compatible with pairwise exchange and non-negative dissipation, but the argument that it is the only admissible form under the merging axiom relies on the specific algebraic structure of the Onsager matrix on the simplex boundary. An explicit verification that no other degenerate matrices (e.g., with higher-order vanishing) satisfy the axioms would close a potential gap.

    Authors: We thank the referee for highlighting this. While the derivation shows that the merging invariance, together with the Onsager structure and non-negativity, leads to the bilinear form, we agree that an explicit check for alternative degeneracies strengthens the uniqueness claim. In the revised manuscript, we have included an additional lemma or remark in Section 4 providing explicit verification that forms with higher-order vanishing (e.g., cubic or higher in the volume fractions) fail to satisfy the merging consistency or lead to negative dissipation in some configurations. revision: yes

  3. Referee: [§5] §5 (numerical validation): The experiments illustrate merging consistency, yet the reported reduction error is only qualitative. Quantitative tables or plots showing the L^2 or energy-norm difference between the N-phase and (N-1)-phase solutions when two phases are merged would make the numerical support for the theoretical uniqueness statement more convincing.

    Authors: We agree that quantitative metrics would provide stronger numerical evidence for the theoretical results. In the revised Section 5, we have added quantitative comparisons, including tables with L^2-norm and energy-norm differences between the full N-phase simulations and the reduced (N-1)-phase models under phase merging, as well as corresponding plots demonstrating the convergence of the errors as the phases become identical. revision: yes

Circularity Check

0 steps flagged

No circularity; uniqueness derived from explicitly stated structural axioms without reduction to inputs by construction

full rationale

The paper presents a conditional uniqueness result: under a small set of structural axioms for thermodynamic admissibility and PDE-level merging consistency, the free-energy density is fixed to ideal-mixing plus symmetric mean-field interaction plus constant-coefficient quadratic gradient penalty, and the Onsager mobility is fixed to pairwise-exchange form with bilinear degeneracy. The abstract frames this as a derivation from the axioms rather than a fit, self-definition, or self-citation chain. No quoted step in the provided text reduces the claimed forms to the axioms by construction (e.g., no parameter fitted to data then relabeled as prediction, no ansatz imported via self-citation, no renaming of known results). The axioms are load-bearing modeling choices that define the admissible class, but the derivation applies them to obtain specific functional forms; this is self-contained mathematical modeling, not circular. The result is falsifiable against the axioms and external benchmarks such as Maxwell-Stefan limits.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The central claim rests on a small set of structural axioms for thermodynamic admissibility and mixture awareness; no explicit free parameters or new invented entities are introduced in the abstract.

axioms (2)
  • domain assumption The free energy and mobility must satisfy thermodynamic admissibility conditions.
    Standard requirement in phase-field and nonequilibrium thermodynamics models.
  • ad hoc to paper The constitutive closure must be invariant under merging of physically identical phases at the PDE level.
    The key new structural requirement introduced to achieve mixture awareness.

pith-pipeline@v0.9.0 · 5579 in / 1464 out tokens · 71850 ms · 2026-05-07T07:35:40.215290+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

2 extracted references · 1 canonical work pages

  1. [1]

    Brunk, A

    A. Brunk, A. Jüngel, and M. Lukáčová-Medvid’ová. A Structure-Preserving Numerical Method for Quasi-Incompressible Navier–Stokes–Maxwell–Stefan systems.J. Sci. Comput., 106, 2026a. ISSN 1573-7691. A. Brunk, M.F.P. ten Eikelder, M. Fritz, D. Höhn, and D. Trautwein. Review of thermodynamic structures and structure-preserving discretisations of Cahn–Hilliard-...

    work page arXiv

  2. [2]

    Wu and J

    S. Wu and J. Xu. Multiphase Allen–Cahn and Cahn–Hilliard models and their discretizations with the effect of pairwise surface tensions.Journal of Computational Physics, 343:10–32, 2017

    2017