Pith. sign in

REVIEW 2 major objections 4 minor 2 cited by

This paper proves that the anisotropic multi-phase Mullins–Sekerka problem with kinetic undercooling can be approximated by a fully discrete, unfitted parametric finite element method that is unconditionally stable: at every time step the d

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

T0 review · deepseek-v4-flash

2026-08-02 21:57 UTC pith:GNLOMAPK

load-bearing objection A solid extension of the authors' PFEM work to anisotropic multi-phase Mullins-Sekerka, but Theorem 4.1's uniqueness proof has a real gap for disconnected clusters when ρ=0. the 2 major comments →

arxiv 2602.18226 v2 pith:GNLOMAPK submitted 2026-02-20 math.NA cs.NA

A Parametric Finite Element Approach for an Anisotropic Multi-Phase Mullins-Sekerka Problem with Kinetic Undercooling

Tokuhiro Eto , Harald Garcke , Robert N\"urnberg This is my paper
classification math.NA cs.NA MSC 65M1235R3565M5065M6074N1080A22
keywords Mullins–Sekerkamulti-phaseanisotropic surface energykinetic undercoolingparametric finite element methodunconditional stabilitytriple junctionsice crystal growth
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

The paper sets out to build and analyze a numerical method for a sharp-interface model of multi-phase crystal growth. The model allows several phases separated by a cluster of surfaces that meet at triple junctions, with anisotropic surface energy and kinetic undercooling. The authors derive a weak formulation, then a fully discrete finite element scheme in which the moving interfaces are independent of the bulk triangulation. Their main results are that each time step has a unique solution and that the scheme is unconditionally stable, with a discrete energy matching the continuous dissipation law. Numerical examples in 2D and 3D demonstrate double-bubble coarsening, symmetry-breaking growth, dendritic growth toward an undercooled boundary, and faceted ice-crystal morphologies with triple junctions.

Core claim

On the paper's own terms, the central discovery is that the full anisotropic multi-phase Mullins–Sekerka system with kinetic undercooling can be approximated by a linear fully discrete scheme that inherits the continuous energy-dissipation structure. The key inequality (4.8) states that the anisotropic surface energy of the new surface plus the discrete weighted-volume term, the bulk gradient of the chemical potential, and the kinetic contribution is bounded by the surface energy of the old surface. Because the scheme is linear and the finite element spaces satisfy a non-degeneracy condition (Assumption 1), each time step has a unique solution. The numerical examples show the scheme followin

What carries the argument

The engine of the argument is a discrete anisotropic surface-energy inequality coupled to a weak formulation of the anisotropic mean curvature vector. The anisotropies are restricted to sums of square roots of positive-definite quadratic forms, which allows the curvature vector to be expressed through a metric-dependent tangent basis. The stability proof tests the three discrete equations—motion, Gibbs–Thomson with kinetic undercooling, and curvature vector—against the displacement and combines them with the discrete geometric inequality |Γ^{m+1}|_γ − |Γ^m|_γ ≤ ⟨∇_G X^{m+1}, ∇_G(X^{m+1}−id)⟩_γ. This inequality is the single load-bearing geometric estimate.

Load-bearing premise

The load-bearing premise is that the cluster's topology is fixed for all time (so no interface vanishes or merges) and that the discrete vertex normals and bulk mesh satisfy a non-degeneracy compatibility condition; if either fails, the uniqueness and energy-stability theorems do not apply.

What would settle it

Compute a shrinking two-phase bubble with the scheme, deliberately withholding the heuristic surgery used in the paper. If the discrete energy (4.8) rises on the step where the interface collapses, or the linear system becomes singular, then the fixed-topology assumption was genuinely load-bearing. A second, more basic test is to build a configuration where Assumption 1's span condition fails and check whether the homogeneous system has a non-trivial solution.

Watch this falsifier. Get emailed when new claim-graph text bears on it.

If this is right

  • Time steps need not satisfy a CFL-type restriction: stability holds for any step size, so simulations can be run with large steps until accuracy, not stability, becomes the limiting factor.
  • The method handles more than two phases and triple junctions, so it can resolve multi-phase coarsening and grain-boundary dynamics that are out of reach for two-phase-only front trackers.
  • Admitted anisotropies include smoothed hexagonal and cubic Wulff shapes, enabling faceted ice-crystal morphologies with kinetic undercooling.
  • When the outer boundary is insulated, the discrete energy reduces to the anisotropic surface energy and each phase volume is conserved at the continuous level; the simulations show coarsening rather than drift.
  • The unfitted discretization lets the bulk mesh be adaptively refined near interfaces without re-meshing the surfaces, which is why the method can follow growing dendrites and crystals in three dimensions.

Where Pith is reading between the lines

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

  • A natural next step is to test convergence rates on a two-phase radially symmetric Mullins–Sekerka benchmark; the paper does not report error tables, and such a test would show whether the energy-stable scheme converges at the expected order.
  • A promising extension would be to incorporate a rigorous surgery or remeshing step with an energy-jump bound, placing the observed bubble-vanishing and merging simulations on the same footing as the theory.
  • The same block-elimination trick used to enforce the sum-zero constraint on the chemical potential could be reused for other constrained multi-phase problems, such as degenerate Stefan systems, with minimal changes to the linear algebra.
  • A quantitative comparison between these front-tracking predictions and physical snow-crystal growth experiments, varying undercooling, kinetic coefficient, and anisotropy strength, would test how faithfully the sharp-interface model captures real faceting and side-branching.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

2 major / 4 minor

Summary. The paper proposes a parametric finite element method for an anisotropic multi-phase Mullins-Sekerka problem with kinetic undercooling, in a sharp-interface formulation for a cluster of surfaces with triple junctions. It derives a variational formulation and a fully discrete unfitted scheme (4.3). The main theoretical results are Theorem 4.1 (existence/uniqueness under Assumption 1) and Theorem 4.2 (unconditional discrete energy dissipation, Eq. (4.8)). Numerical experiments in 2D and 3D, including ice-crystal growth and topology-changing evolutions, are presented.

Significance. An unconditionally stable front-tracking method for anisotropic multi-phase Mullins-Sekerka flow with kinetic undercooling would be a valuable contribution to the numerical analysis of geometric evolution equations. The energy inequality in Theorem 4.2 is the central novelty, and the proof makes careful use of the discrete anisotropic energy estimate [10, Lemma 102]. The paper is clearly written and the numerical experiments are rich and informative. However, the well-posedness theorem has a gap for disconnected surface clusters, which are exactly the cases used in several numerical examples, and the implemented code uses heuristic surgeries outside the proved framework. If the assumption is strengthened componentwise, the stability result remains an interesting and useful contribution.

major comments (2)
  1. [§4, Theorem 4.1, proof around (4.5)] The conclusion that ∇G_s X=0 implies 'X ≡ Xc' is valid only on each connected component of the surface cluster. For disconnected Γ^m (as in Examples 1 and 2, where a disk coexists with a double bubble), the homogeneous system admits solutions with X=v on one component and X=-v on another, W=0, κ=0, if v annihilates the difference of the two components' bulk-interface coupling vectors. Assumption 1 (4.4) controls only the sum of the coupling operators over all components and does not exclude such modes when ρ=0. The step 'Xc · Σ ... =0 ⇔ Xc=0' is therefore invalid for disconnected clusters. Since §6.1 uses ρ=0 and disconnected initial data, Theorem 4.1 as stated does not cover the numerical examples. Fix: either restrict the theorem to connected clusters or replace (4.4) by the component-wise condition for each connected component C: span{Σ_{i∈C} Σ_l ⟨[χ_l]_{Γ_i^m} ω_i^m, φ_l⟩ | φ ∈ S_0^m
  2. [§6, first paragraph; Theorem 4.2] The unconditional stability theorem applies to one step of the scheme on a fixed-topology cluster. The implementation additionally performs heuristic surgeries (removing short curves and gluing the remaining ones). These surgeries are outside the hypotheses of Theorem 4.2, so the computed energy plots in Examples 1-3, where disks vanish or bubbles merge, are not covered by the stability proof. The paper states this limitation honestly, but the abstract's claim that 'our method can be shown to be unconditionally stable' should be qualified. Please either argue that the chosen surgery preserves a discrete energy inequality or clearly present the experiments as heuristic extensions beyond the theory.
minor comments (4)
  1. [§4, Eq. (4.3)] The operator π_i^h (and π^h in the proof of Theorem 4.2) is used but never defined. Please provide its definition or a precise reference.
  2. [§6, first paragraph] For reproducibility, mark in the figures or text the times at which the heuristic surgery is applied; otherwise the reader cannot separate the regime covered by Theorem 4.2 from the post-surgery regime.
  3. [Eq. (6.1), Eq. (6.3)] The matrix notation in (6.1) is ambiguous: the exponent 'ℓ' on the rotation R(π/3) should be displayed clearly. In (6.3), the constant θ0 is used but not defined.
  4. [Proposition 2.1] The proof relies on [8, Lemma 3.1] for the triple-junction contribution; a short statement of this lemma would improve readability, though the citation is appropriate.

Circularity Check

0 steps flagged

No circular derivation chain: the discrete stability and well-posedness results are proved from the proposed scheme and published geometric lemmas; self-citations are not input-output equivalences.

full rationale

The paper's load-bearing claims are Theorem 4.1 (existence/uniqueness) and Theorem 4.2 (unconditional stability). Neither claim is obtained by defining an output in terms of itself. The weak formulation (3.5) is derived from the strong form (1.1) by integration by parts, and the fully discrete scheme (4.3) is introduced as a new construction. Theorem 4.2's proof chooses test functions phi = W^{m+1}-w_D, xi = pi_h[(X^{m+1}-id)*omega^m], eta = X^{m+1}-id in (4.3), which yields the bulk, volume, and kinetic terms directly from the scheme; the only geometric input is the discrete energy estimate (4.11) quoted from [10, Lemma 102]. That lemma is a parameter-free convexity-type bound for the parametric surface energy difference and is not the same as the target inequality (4.8), which additionally contains the bulk gradient, weighted volume, and kinetic contributions. The anisotropic-curvature weak identity used in (3.5c) is likewise an established result from [5,8], not a restatement of the paper's theorem. The paper does rely on several prior works by the same group ([4,5,7,8,10,18,19]), but these are published, independently derived results, so the self-citation is not load-bearing circularity. I do flag two scope/correctness caveats that are not circularity: the proof of Theorem 4.1 concludes that X is a global constant from a vanishing anisotropic gradient, a step that is only immediate per connected component when Gamma^m is disconnected; and the paper itself states in Section 6 that topology changes are handled by heuristic surgeries outside the theoretical framework. Neither caveat turns the derivation into an equivalence between inputs and outputs.

Axiom & Free-Parameter Ledger

0 free parameters · 6 axioms · 0 invented entities

No fitted parameters appear in the method; the simulation quantities δ, α, wD, ρ, and β are problem inputs, not free parameters of the discrete scheme. The central theorems rest on standard PDE/FEM assumptions plus the listed technical restrictions, most notably the fixed-topology assumption and Assumption 1.

axioms (6)
  • domain assumption Anisotropy densities are restricted to γ_i(⃗p)=∑_{ℓ=1}^{L_i} sqrt(⃗p·G_i^{(ℓ)}⃗p) with G_i^{(ℓ)} symmetric positive definite.
    Section 3, after (3.2), restricts the class of anisotropies so that the finite-element weak curvature formulation of [5,8] applies. General convex C^1 anisotropies are not covered.
  • domain assumption The chemical potential w takes values in TΣ={u·1=0}, with w_D·1=0; all bulk equations are posed in this subspace.
    Section 2, hyperplane TΣ paragraph; a standard reduction for multi-phase systems that is used in the weak formulation (3.5).
  • ad hoc to paper Assumption 1: for each interface the vertex normals span R^d, and the combined bulk/interface compatibility condition (4.4) holds.
    Section 4, Assumption 1; technical non-degeneracy condition required in the proof of existence and uniqueness (Theorem 4.1).
  • domain assumption The surface cluster has fixed topology over time: interfaces do not vanish, are not created, and junction structure persists.
    Section 6, 'Our theoretical framework does not allow for changes of topology...'; implicit in the weak and discrete formulations and in Theorem 4.2.
  • standard math Discrete energy inequality (4.11): ⟨∇~G_s X^{m+1}, ∇~G_s (X^{m+1}-id)⟩_{γ,Γ^m} ≥ |Γ^{m+1}|_γ - |Γ^m|_γ, taken from [10, Lemma 102].
    Imported lemma used in the proof of Theorem 4.2; accepted published result from the authors' handbook chapter.
  • domain assumption Classical solutions are assumed for the continuous dissipation identity in Proposition 2.1; β_i smooth, even, positive; ρ_i≥0.
    Section 2 model assumptions; Proposition 2.1 is a motivating identity, while the discrete stability proof does not assume classical solutions.

pith-pipeline@v1.3.0-alltime-deepseek · 25568 in / 13466 out tokens · 115144 ms · 2026-08-02T21:57:18.527877+00:00 · methodology

0 comments
read the original abstract

We consider a sharp interface formulation for an anisotropic multi-phase Mullins-Sekerka problem with kinetic undercooling. The flow is characterized by a cluster of surfaces evolving such that the total surface energy plus a weighted sum of the volumes of the enclosed phases decreases in time. Upon deriving a suitable variational formulation, we introduce a fully discrete unfitted finite element method. In this approach, the approximations of the moving interfaces are independent of the triangulations used for the equations in the bulk. Our method can be shown to be unconditionally stable. Several numerical examples demonstrate the capabilities of the introduced method. In particular, it is demonstrated that the evolution of multiple ice crystals with junctions can be modeled using the proposed approach.

Figures

Figures reproduced from arXiv: 2602.18226 by Harald Garcke, Robert N\"urnberg, Tokuhiro Eto.

Figure 1
Figure 1. Figure 1: A surface cluster in 2D made up of three open curves and two triple junctions (d = 2, IS = 3, IR = 3, IT = 2). The surface cluster is endowed with the anisotropic surface energy |Γ(t)|γ := X IS i=1 |Γi(t)|γi with |Γi(t)|γi := Z Γi(t) γi(~νi) dH d−1 , (2.1) 3 [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: The solution at times t = 0, 0.2, 0.4, 1, and a plot of the discrete energy over time. Below we show the adaptive bulk mesh at times t = 0 and t = 1. -4 -3 -2 -1 0 1 2 3 4 -4 -3 -2 -1 0 1 2 3 4 -4 -3 -2 -1 0 1 2 3 4 -4 -3 -2 -1 0 1 2 3 4 -4 -3 -2 -1 0 1 2 3 4 -4 -3 -2 -1 0 1 2 3 4 -4 -3 -2 -1 0 1 2 3 4 -4 -3 -2 -1 0 1 2 3 4 31 32 33 34 35 36 37 38 0 0.5 1 1.5 2 2.5 3 3.5 4 [PITH_FULL_IMAGE:figures/full_fi… view at source ↗
Figure 3
Figure 3. Figure 3: The solution at times t = 0, 1, 2, 4, and a plot of the discrete energy over time. 6.1 Simulations in 2D with ∂ΩN = ∂Ω In this subsection we consider some numerical simulations for d = 2 and ∂ΩN = ∂Ω. We also let ρ = 0. For the anisotropy we define γhex(~p) := X 3 ℓ=1 q [(R( π 3 ) ℓ ]⊤D(δ)(R( π 3 ))ℓ~p · ~p, δ = 0.1, (6.1) where R(θ) = cos θ sin θ − sin θ cos θ  and D(δ) = diag(1, δ2 ). Observe that the W… view at source ↗
Figure 4
Figure 4. Figure 4: The solution at times t = 0, 0.5, 1, 2, 3, 4, 5, 6, 7, 8, and a plot of the discrete energy over time. -4 -3 -2 -1 0 1 2 3 4 -4 -3 -2 -1 0 1 2 3 4 -4 -3 -2 -1 0 1 2 3 4 -4 -3 -2 -1 0 1 2 3 4 -4 -3 -2 -1 0 1 2 3 4 -4 -3 -2 -1 0 1 2 3 4 -4 -3 -2 -1 0 1 2 3 4 -4 -3 -2 -1 0 1 2 3 4 26 28 30 32 34 36 38 40 42 44 0 0.5 1 1.5 2 [PITH_FULL_IMAGE:figures/full_fig_p016_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: The solution at times t = 0, 0.5, 1, 2, and a plot of the discrete energy over time. IT = 4, (s 1 1 , s1 2 , s1 3 ) = (s 2 1 , s2 2 , s2 3 ) = (1, 2, 3), (s 3 1 , s3 2 , s3 3 ) = (s 4 1 , s4 2 , s4 3 ) = (4, 5, 6) and O =   0 −1 1 0 −1 1 1 0 −1 1 0 −1 −1 1 0 −1 1 0   . The first bubble is chosen with enclosing areas 3.14 and 6.48, while the second double bubbles encloses two areas of size 3.64. In each… view at source ↗
Figure 6
Figure 6. Figure 6: (wD = (20, 10, −30)⊤, ρ = 1) The solution at times t = 0, 0.05, 0.1, 0.2, and a plot of the discrete energy over time. -4 -3 -2 -1 0 1 2 3 4 -4 -3 -2 -1 0 1 2 3 4 -4 -3 -2 -1 0 1 2 3 4 -4 -3 -2 -1 0 1 2 3 4 -4 -3 -2 -1 0 1 2 3 4 -4 -3 -2 -1 0 1 2 3 4 -4 -3 -2 -1 0 1 2 3 4 -4 -3 -2 -1 0 1 2 3 4 1100 1150 1200 1250 1300 1350 1400 1450 1500 0 0.02 0.04 0.06 0.08 0.1 [PITH_FULL_IMAGE:figures/full_fig_p017_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: (wD = (12, 11, −23)⊤, ρ = α = 0.05) The solution at times t = 0, 0.01, . . . , 0.1, and separately at times t = 0.05, t = 0.07 and t = 0.1, and a plot of the discrete energy over time. IR = 3, IT = 2, (s 1 1 , s1 2 , s1 3 ) = (s 2 1 , s2 2 , s2 3 ) = (1, 2, 3) and O =   0 −1 1 1 0 −1 −1 1 0  . The evolution is shown in [PITH_FULL_IMAGE:figures/full_fig_p017_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: (wD = (12, 11, −23)⊤, ρ = α = 0.05) The solution at times t = 0, 0.05, . . . , 0.45, and separately at times t = 0.2, t = 0.3 and t = 0.45, and a plot of the discrete energy over time. 17 [PITH_FULL_IMAGE:figures/full_fig_p017_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: (wD = (5, 5, −10)⊤, ρ = 0.05, α = 0.005) The solution at times t = 0, 0.01, . . ., 0.2, and separately at times t = 0.02, t = 0.04, t = 0.06, t = 0.08, t = 0.11, t = 0.14, t = 0.17 and t = 0.2, and a plot of the discrete energy over time. -4 -3 -2 -1 0 1 2 3 4 -4 -3 -2 -1 0 1 2 3 4 -4 -3 -2 -1 0 1 2 3 4 -4 -3 -2 -1 0 1 2 3 4 -4 -3 -2 -1 0 1 2 3 4 -4 -3 -2 -1 0 1 2 3 4 -4 -3 -2 -1 0 1 2 3 4 -4 -3 -2 -1 0 1 … view at source ↗
Figure 10
Figure 10. Figure 10: (wD = (10, 1, −11)⊤, ρ = α = 0.05) The solution at times t = 0, 0.01, . . . , 0.3, and separately at times t = 0.02, t = 0.05, t = 0.1, t = 0.15, t = 0.17, t = 0.2, t = 0.25 and t = 0.3, and a plot of the discrete energy over time. -4 -3 -2 -1 0 1 2 3 4 -4 -3 -2 -1 0 1 2 3 4 -4 -3 -2 -1 0 1 2 3 4 -4 -3 -2 -1 0 1 2 3 4 -4 -3 -2 -1 0 1 2 3 4 -4 -3 -2 -1 0 1 2 3 4 -4 -3 -2 -1 0 1 2 3 4 -4 -3 -2 -1 0 1 2 3 4 … view at source ↗
Figure 11
Figure 11. Figure 11: (wD = (1, 10, −11)⊤, ρ = 0.05, α = 0.02) The solution at times t = 0, 0.01, . . . , 0.3, and separately at times t = 0.02, t = 0.05, t = 0.07, t = 0.1, t = 0.15, t = 0.17, t = 0.2 and t = 0.25, and a plot of the discrete energy over time. 18 [PITH_FULL_IMAGE:figures/full_fig_p018_11.png] view at source ↗
Figure 12
Figure 12. Figure 12: (wD = (2, 1, −3)⊤, ρ = 1) The solution at times t = 0, 0.05, 0.1, 1, and a plot of the discrete energy over time. 1440 1450 1460 1470 1480 1490 1500 1510 1520 1530 1540 0 0.2 0.4 0.6 0.8 1 [PITH_FULL_IMAGE:figures/full_fig_p019_12.png] view at source ↗
Figure 13
Figure 13. Figure 13: (wD = (2, 1, −3)⊤, ρ = 1) The solution at times t = 0, 0.05, 0.1, 1, and a plot of the discrete energy over time. 6.3 Numerical results in 3D In this subsection we consider some numerical simulations for d = 3 and ∂ΩD = ∂Ω. For the possible anisotropies we define first of all γℓ 1 (~p) = X 3 i=1 q δ 2|~p| 2 + p 2 i (1 − δ 2), δ = 0.01, (6.2) which approximates the ℓ 1–norm of ~p, see [5]. In addition, a 3… view at source ↗
Figure 14
Figure 14. Figure 14: (wD = (2, 1, −3)⊤, ρ = α = 0.05) The solution at times t = 0, 0.1, 0.2, 0.5, and a plot of the discrete energy over time. 1430 1440 1450 1460 1470 1480 1490 1500 1510 1520 1530 1540 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 [PITH_FULL_IMAGE:figures/full_fig_p020_14.png] view at source ↗
Figure 15
Figure 15. Figure 15: (wD = (2, 1, −3)⊤, ρ = α = 0.05, β = βflat,3) The solution at times t = 0, 0.1, 0.2, 0.4, and a plot of the discrete energy over time. Example 8: We use the same setup as in Example 7, but now let ρ = 0.05. In addition, we choose (γ1, γ2, γ3) = α(γhex, γhex, γhex), i ∈ N≤IS , recall (6.3), with α = 0.05. Moroever, we choose either βi ≡ 1 or βi = βflat,3, i ∈ N≤IS , recall (6.4). The evolutions are shown i… view at source ↗
Figure 16
Figure 16. Figure 16: (wD = (2, 0.3, −2.2)⊤, ρ = α = 0.05) The solution at times t = 0, 0.2, 0.8, 1, and a plot of the discrete energy over time. 20 [PITH_FULL_IMAGE:figures/full_fig_p020_16.png] view at source ↗

discussion (0)

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

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. A Structure-Preserving Method of Fundamental Solutions for the Multi-Phase Mullins-Sekerka Flow

    math.NA 2026-07 unverdicted novelty 6.0

    A charge-simulation mesh-free scheme solves multi-phase Mullins–Sekerka flow with triple junctions and orthogonal wall contact while conserving phase areas to machine precision.

  2. A Structure-Preserving Method of Fundamental Solutions for the Multi-Phase Mullins-Sekerka Flow

    math.NA 2026-07 conditional novelty 6.0

    A charge-simulation method solves multi-phase Mullins-Sekerka flow with triple junctions and Neumann walls while conserving each bounded phase area to machine precision.

Reference graph

Works this paper leans on

33 extracted references · cited by 1 Pith paper

  1. [1]

    W. Bao, H. Garcke, R. N ¨urnberg, and Q. Zhao , A structure-preserving finite element ap- proximation of surface diffusion for curve networks and surf ace clusters, Numer. Methods Partial Differ. Eq., 39 (2023), pp. 759–794

    2023

  2. [2]

    J. W. Barrett, J. F. Blowey, and H. Garcke , On fully practical finite element approx- imations of degenerate Cahn–Hilliard systems , M2AN Math. Model. Numer. Anal., 35 (2001), pp. 713–748

    2001

  3. [3]

    J. W. Barrett, H. Garcke, and R. N ¨urnberg, On the variational approximation of com- bined second and fourth order geometric evolution equation s, SIAM J. Sci. Comput., 29 (2007), pp. 1006–1041. [4] , Numerical approximation of anisotropic geometric evoluti on equations in the plane , IMA J. Numer. Anal., 28 (2008), pp. 292–330. 750 800 850 900 950 1000 ...

    2007

  4. [5]

    Math., 109 (2008), pp

    , A variational formulation of anisotropic geometric evolut ion equations in higher dimen- sions, Numer. Math., 109 (2008), pp. 1–44

    2008

  5. [6]

    , Finite element approximation of coupled surface and grain b oundary motion with applica- tions to thermal grooving and sintering , European J. Appl. Math., 21 (2010), pp. 519–556

    2010

  6. [7]

    , On stable parametric finite element methods for the Stefan pr oblem and the Mullins–Sekerka problem with applications to dendritic growth , J. Comput. Phys., 229 (2010), pp. 6270–6299

    2010

  7. [8]

    , Parametric approximation of surface clusters driven by iso tropic and anisotropic surface energies, Interfaces Free Bound., 12 (2010), pp. 187–234

    2010

  8. [9]

    , Numerical computations of faceted pattern formation in sno w crystal growth , Phys. Rev. E, 86 (2012), p. 011604

    2012

  9. [10]

    Part I, vol

    , Parametric finite element approximations of curvature-dri ven interface evolutions , in Ge- ometric partial differential equations. Part I, vol. 21 of Handb. Nu mer. Anal., Elsevier/North- Holland, Amsterdam, 2020, pp. 275–423

    2020

  10. [11]

    J. F. Blowey, M. I. M. Copetti, and C. M. Elliott , Numerical analysis of a model for phase separation of a multi-component alloy , IMA J. Numer. Anal., 16 (1996), pp. 111–139

    1996

  11. [12]

    Bronsard, H

    L. Bronsard, H. Garcke, and B. Stoth , A multi-phase Mullins–Sekerka system: matched asymptotic expansions and an implicit time discretisation for the geometric evolution problem , Proc. Roy. Soc. Edinburgh Sect. A, 128 (1998), pp. 481–506

    1998

  12. [13]

    Danilov and B

    D. Danilov and B. Nestler , Phase-field simulations of solidification in binary and tern ary systems using a finite element method , J. Cryst. Growth, 275 (2005), pp. e177–e182

    2005

  13. [14]

    S. H. Davis , Theory of Solidification , Cambridge Monographs on Mechanics, Cambridge Univer- sity Press, Cambridge, 2001

    2001

  14. [15]

    T. A. Davis , Algorithm 832: UMFPACK V4.3—an unsymmetric-pattern multi frontal method , ACM Trans. Math. Software, 30 (2004), pp. 196–199

    2004

  15. [16]

    , Algorithm 915, SuiteSparseQR: Multifrontal multithreade d rank-revealing sparse QR fac- torization, ACM Trans. Math. Software, 38 (2011), pp. 1–22

    2011

  16. [17]

    Demange, H

    G. Demange, H. Zapolsky, R. Patte, and M. Brunel , A phase field model for snow crystal growth in three dimensions , Comput. Mater., 3 (2017), p. 15

    2017

  17. [18]

    T. Eto, H. Garcke, and R. N ¨urnberg, A structure-preserving finite element method for the multi-phase Mullins–Sekerka problem with triple junction s, Numer. Math., 156 (2024), pp. 1479– 1509

    2024

  18. [19]

    Methods Appl

    , A parametric finite element method for a degenerate multi-ph ase Stefan problem with triple junctions, Comput. Methods Appl. Math., 26 (2026), pp. 43–67

    2026

  19. [20]

    D. J. Eyre , Systems of Cahn–Hilliard equations , SIAM J. Appl. Math., 53 (1993), pp. 1686–1712

    1993

  20. [21]

    Garcke, B

    H. Garcke, B. Nestler, and B. Stinner , A diffuse interface model for alloys with multiple components and phases , SIAM J. Appl. Math., 64 (2004), pp. 775–799

    2004

  21. [22]

    Garcke, B

    H. Garcke, B. Nestler, and B. Stoth , On anisotropic order parameter models for multi- phase systems and their sharp interface limits , Phys. D, 115 (1998), pp. 87–108

    1998

  22. [23]

    Garcke, R

    H. Garcke, R. N ¨urnberg, and Q. Zhao , A variational front-tracking method for multiphase flow with triple junctions , Math. Comp., 95 (2026), pp. 647–682

    2026

  23. [24]

    Garcke and T

    H. Garcke and T. Sturzenhecker , The degenerate multi-phase Stefan problem with Gibbs– Thomson law , Adv. Math. Sci. Appl., 8 (1998), pp. 929–941

    1998

  24. [25]

    Herlach , ed., Phase Transformations in Multicomponent Melts , John Wiley & Sons, Ltd, 2008

    D. Herlach , ed., Phase Transformations in Multicomponent Melts , John Wiley & Sons, Ltd, 2008. 21

    2008

  25. [26]

    D. W. Hoffman and J. W. Cahn , A vector thermodynamics for anisotropic surfaces: I. fundamentals and application to plane surface junctions , Surface Sci., (1972), pp. 368–388

    1972

  26. [27]

    Li and W

    Y. Li and W. Bao , An energy-stable parametric finite element method for aniso tropic surface diffusion, J. Comput. Phys., 446 (2021), p. 110658

    2021

  27. [28]

    Y. Li, J. Choi, and J. Kim , Multi-component Cahn–Hilliard system with different bound ary conditions in complex domains , J. Comput. Phys., 323 (2016), pp. 1–16

    2016

  28. [29]

    K. G. Libbrecht , The physics of snow crystals , Rep. Progr. Phys., 68 (2005), pp. 855–895

    2005

  29. [30]

    Nestler, H

    B. Nestler, H. Garcke, and B. Stinner , Multicomponent alloy solidification: Phase-field modeling and simulations , Phys. Rev. E, 71 (2005), p. 041609

    2005

  30. [31]

    N ¨urnberg, Numerical simulations of immiscible fluid clusters , Appl

    R. N ¨urnberg, Numerical simulations of immiscible fluid clusters , Appl. Numer. Math., 59 (2009), pp. 1612–1628

    2009

  31. [32]

    , A structure preserving front tracking finite element method for the Mullins–Sekerka prob- lem, J. Numer. Math., 31 (2023), pp. 137–155

    2023

  32. [33]

    Schmidt and K

    A. Schmidt and K. G. Siebert , Design of Adaptive Finite Element Software: The Finite Element Toolbox ALBERTA, vol. 42 of Lecture Notes in Computational Science and Engineering , Springer-Verlag, Berlin, 2005

    2005

  33. [34]

    Steinbach , Phase-field models in materials science , Modelling and Simulation in Materials Science and Engineering, 17 (2009), p

    I. Steinbach , Phase-field models in materials science , Modelling and Simulation in Materials Science and Engineering, 17 (2009), p. 073001. 22

    2009