Stochastic Cahn-Hilliard equation in higher space dimensions: The motion of bubbles

Alexander Schindler; Dirk Bl\"omker

REVIEW 3 major objections 4 minor 23 references

Stochastic Cahn-Hilliard equation in higher space dimensions: The motion of bubbles

T0 review · 3 major / 4 minor · reviewed 2026-08-14 · deepseek-v4-flash

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

Pith's one-line read For sufficiently small noise, a single droplet in the two-dimensional stochastic Cahn-Hilliard equation stays close to its slow manifold for polynomial times, and its center solves an explicit stochastic differential equation.

desk verdict Credible 2D stochastic Cahn-Hilliard droplet result with an explicit SDE and long-time stability; the title oversells 'higher dimensions' and the stability rests on an imported spectral gap that deserves scrutiny. read the letter →

arxiv 1908.01601 v1 pith:JJQUICHI submitted 2019-08-05 math.DS math.APmath.PR

Alexander Schindler , Dirk Blömker This is my paper · ORCID

classification math.DSmath.APmath.PR MSC 35R6060H1537L2535K5535B25

keywords stochasticCahn-HilliardequationslowmanifolddropletmotionspectralgapstabilityadditivenoiseFermicoordinatessharpinterfacelimit

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper aims to prove a stability-and-motion statement for the last stage of phase separation in the stochastic Cahn-Hilliard equation on a two-dimensional domain. If a solution starts near a single-droplet state, and the additive noise is small in the precise sense of the paper's assumptions, then the solution remains near the family of droplet states for times of order $ε^{{-N}}$, with probability higher than any fixed power of ε. While it stays near that family, the droplet's center is shown to move according to an explicit stochastic differential equation, whose noise term is, to leading order, just the projection of the driving noise onto the two translation directions of the droplet. The upshot is a rigorous random-walk description of droplet motion, replacing the exponentially slow deterministic motion of the noiseless equation.

What carries the argument

The construction rests on a two-dimensional slow manifold M̃_ε^ρ of droplet profiles ũ_ξ, each an exponentially small correction of a translated radial bubble, chosen so the deterministic residual lies in the tangent space. The spectral fact doing the heavy lifting is Theorem 2.4: around every droplet, the linearized Cahn-Hilliard operator has two eigenvalues exponentially close to zero, with eigenvectors almost tangent to translations, and a third eigenvalue bounded below by C'ε. That order-ε gap produces the damping coefficient a=O(ε) in the $H^{{-1}}$ inequality (4.9)–(4.10), which in turn fixes the admissible noise sizes. Fermi coordinates (ξ,v), with v orthogonal to the two critical eigenvectors, allow an exact stochastic projection: differentiating the constraint ⟨v,ψ^ξ_k⟩=0 and the equation itself yields the SDE for ξ, with the invertible matrix A_{kj}=⟨ψ^ξ_k, ũ^ξ_j⟩-⟨v,ψ^ξ_{k,j}⟩ governing the projection. The $L^{2}$ stability argument uses the mass-conserving Allen-Cahn operator, whose spectral gap is ε², and interpolation to control the nonlinearity.

What would settle it

Compute the third eigenvalue λ3(ε) of the linearized Cahn-Hilliard operator around a centered droplet on a two-dimensional disk with Neumann boundary conditions; if λ3(ε)/ε → 0 as ε→0, the order-ε spectral gap assumed by Theorem 2.4 fails and the exit-probability bounds of Theorem 4.12 lose their stated noise exponents.

Watch

Extended reading notes

Core claim

The central claim is that, in two dimensions, one-droplet solutions of the stochastic Cahn-Hilliard equation are stable against small additive noise for polynomial times, and the droplet center obeys an explicit SDE. More precisely, for m>4 and 0<k<m-4, if the initial distance from the slow manifold satisfies ‖v(0)‖_{$H^{{-1}}$} ≤ ν ε^m and ‖v(0)‖_{$L^{2}$} ≤ ν $ε^{{k+1}}$, and the noise intensities satisfy η_0 ≤ C $ε^{{2m+1+κ}}$ and η_2 ≤ C $ε^{{2k+4+κ}}$, then the probability of leaving the ε^m ($H^{{-1}}$) and $ε^{{k+1}}$ ($L^{2}$) tube before time $ε^{{-N}}$ is smaller than any power of ε. On the manifold, the center solves dξ = f(ξ)dt + σ(ξ)dW with f and σ given by (3.11)–(3.12); Lemma 3.4 shows σ is asymptotically the normalized tangent vector, so the noise enters only through the direction in which the droplet is free to move. Up to exponentially small corrections, the stochastic motion is exactly the projection of the driving noise onto the slow manifold.

Load-bearing premise

The whole proof depends on an imported spectral estimate: around each droplet, the linearized operator's first two eigenvalues are exponentially small and the third is at least of order ε, uniformly in the center; if the true gap were only ε², as the paper notes in three dimensions, the damping coefficient in the stability estimate would disappear and the stated noise bounds would be too large.

Editorial extensions

If this is right

For any fixed large N, if the noise bounds hold, the exit probability from the ε^m tube before time ε^{-N} is smaller than every power of ε.
Close to the slow manifold, the center motion is dξ = f(ξ)dt + σ(ξ)dW with explicit f and σ, and σ is, up to O(1), the normalized translational mode, so the noise enters through the direction in which the droplet is free to move.
The H^{-1} and L^2 stability radii are coupled to two different noise measures, η0 and η2, so both the total noise strength and its spatial smoothness must be small.
The stability window covers times ε^{-N} for any N, long enough that the stochastic motion of the center, rather than noise-induced nucleation or shape destruction, is what happens before a boundary collision.
The theorem is two-dimensional; in three dimensions the order-ε spectral gap is replaced by ε², so the same proof would not go through unchanged.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

The paper does not simulate the SDE, but a direct test is mean-square displacement: on times short compared with ε^{-N}, the droplet center should satisfy roughly E|ξ(t)-ξ(0)|² ≈ C η0 t, so tracking many realizations would test the projection formula for σ and estimate the noise amplitude.
Because the three-dimensional spectral gap is only ε², the same stability argument would have a weaker damping rate; quantifying how the stability radius and noise thresholds must shrink in three dimensions is a concrete open problem.
The Fermi-coordinate projection is not tied to the specific nonlinearity F(u)=¼(u²-1)², so the same derivation should produce SDEs for droplet motion in mass-conserving Allen-Cahn models or for droplets sliding along the boundary, where a slow manifold of the same type exists.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Desk Editor's Note

read the letter

Read it if you work on stochastic interface motion. The main result is honest and mostly checks out, but the paper is less general than its title.

What is new: for a single droplet in 2D stochastic Cahn-Hilliard, the authors split the solution into slow-manifold coordinate xi and orthogonal part v, and derive an exact SDE for xi. The formulas for drift and diffusion, (3.11)-(3.12), come out of a clean Ito calculation, and Lemma 3.2 verifies that the ansatz reproduces the SPDE. That is genuinely new: previous stochastic droplet results were 1D Cahn-Hilliard or mass-conserving Allen-Cahn. The diffusion is approximately the projection of the Wiener process onto translations, which is the physically right answer. Section 4's stability estimates follow a known framework, but they adapt it carefully and give explicit noise restrictions.

What worries me. The load-bearing spectral input is Theorem 2.4 from Alikakos-Fusco: in 2D the third eigenvalue of the linearized Cahn-Hilliard operator is bounded below by C' epsilon. That is not proved here. If the true 2D gap were only O(epsilon^2), the damping coefficient in (4.10) and the noise restrictions in Theorems 4.7 and 4.12 would both change. This is a conditional risk, not a demonstrated error, but the paper should at least give a proof or a citable statement with the 2D computation. Remark 2.6 concedes the 3D gap is only O(epsilon^2), which is exactly why the title's 'higher space dimensions' is misleading: the paper treats d=2 only, and d=3 is deferred to a thesis.

Second, Lemma 4.2 has an 'absorbed' step in (4.7) where a positive L2 term is absorbed into negative terms by choosing gamma weights. I cannot verify the constant balance from what is written. This is fixable but it is the kind of spot a referee needs to see nailed down. Third, the L2 moment estimates in Section 4.4 are imported from [7] with details skimmed; acceptable for a paper in this tradition, but it makes the result hard to check independently. Minor: Theorem 4.7 states ||v(0)|| <= nu epsilon^4 instead of epsilon^m, a typo that makes the m>4 statement vacuous if read literally.

Citation pattern is fine: [7] and [8] share authors and supply the general method, but the 2D Cahn-Hilliard SDE derivation and stability result are not in those papers. The paper does not assume its conclusion.

Verdict: worth a serious referee. I would send it out, asking the referee to verify Lemma 4.2 and to pressure-test the imported epsilon-gap. In current form I would not bet on the stability theorem without checking those steps, but I would cite it after revision.

Referee Report

3 major / 4 minor

Summary. The paper studies the stochastic Cahn-Hilliard equation in a bounded two-dimensional domain with small additive, spatially smooth noise. It builds on the deterministic slow manifold of single-droplet states constructed in earlier work, derives by an explicit Itô computation an exact SDE for the droplet center, and then proves long-time stochastic stability of the slow manifold in both H^{-1} and L^2 for polynomially long times under small-noise conditions. The central quantitative claim is that, starting within O(ε^m) of the manifold and with noise intensities η0 ≤ C ε^{2m+1+κ} and η2 ≤ C ε^{2k+4+κ}, the probability that the solution exits an ε^m-neighborhood in H^{-1} (and an ε^{k+1}-neighborhood in L^2) before time ε^{-N} is smaller than any power of ε; on the manifold the droplet center approximately satisfies the SDE dξ = f(ξ)dt + σ(ξ)dW with σ obtained by projecting the Wiener process onto the translational modes.

Significance. If the stability statements are correct, this is a valuable rigorous contribution: Section 3 gives a parameter-free derivation of the effective SDE, and Section 4 extends the stability framework of [7] to the stochastic Cahn-Hilliard equation with droplet motion. The explicit Itô computation and the careful identification of the noise projection are genuine strengths. However, the proof of the stability theorems has a quantitative gap concerning the relation between the initial condition and the small quantity q defined in (4.13); until that gap is closed, the headline super-polynomial exit-probability claims are not established as stated.

major comments (3)

[§4.1, Lemma 4.2, Eq. (4.7)] The moment bound (4.14) is derived under the standing assumption (4.13) that ||v(0)||² ≤ q, where q = (C_ε + ||Q||)/a. With the noise condition η0 ≤ C ε^{2m+1+κ} and the exit radius B = ε^m, one has q = O(ε^{2m+κ}). The theorems instead assume ||v(0)|| ≤ ν ε^m (or, in Theorem 4.5, ||v(0)|| ≤ ν ε^4), which gives ||v(0)||² of order ε^{2m}, much larger than q for small ε. Thus the induction leading to (4.14) does not close: the initial-data term ||v(0)||^{2p} cannot be absorbed into the q^p factor, and the claimed 'smaller than any power of ε' bound does not follow from the displayed estimates. The authors should either strengthen the initial-condition hypothesis to ||v(0)|| ≤ C ε^{m+κ/2} (so that ||v(0)||² ≤ q) or give a different argument controlling the ||v(0)||^{2p} terms.
[§2.3, Theorem 2.4(ii) and Remark 2.6] The step 'absorbed the positive L2-term into its negative counterpart' is not justified as written. With γ3 ≈ ε², the positive term is C ε² ||v||²_{L2}; it can be absorbed only if the coefficient γ2 in the negative ε²-term is uniformly bounded away from zero and the constants are chosen compatibly with γ1 + γ2 + γ3 = 1. This is plausible, but the choice of γ1 and γ2 and the resulting constants should be displayed. Because this inequality provides the damping rate a = O(ε) used throughout Section 4, the argument should be completed explicitly.
[§4.2, Theorems 4.5 and 4.7] The stability analysis relies on the order-ε spectral gap λ3 ≥ C'ε in d = 2, but this fact is imported from [2,3] and is not proved here. Remark 2.6 states that in d = 3 the gap is only O(ε²), so the ε-gap is a special d = 2 fact and is load-bearing: if the d = 2 gap were ε², the damping a in (4.10) would become O(ε²) and the noise restrictions in Theorems 4.5 and 4.12 would change by powers of ε. Please cite the precise theorem of [2,3] and, ideally, give a short proof sketch or at least an explicit statement of how the ε-gap enters the estimates.

minor comments (4)

[Lemma 4.11 and Theorem 4.12] The hypothesis reads ||v(0)||_{H^{-1}} ≤ ν ε^4, which must be a typo for ||v(0)||_{H^{-1}} ≤ ν ε^m; as printed, the m > 4 statement is not what is used in the proof.
[§5, Lemma 5.2] Lemma 4.11 assumes k ≥ 2, while Theorem 4.12 allows k ∈ (0, m−4); the proof of Theorem 4.12 invokes Lemma 4.11, so the range should be restricted to k ≥ 2 (hence m > 6) or Lemma 4.11 should be extended.
[Eq. (3.18)] The proof refers to 'Definition 2.9', but no Definition 2.9 appears in the paper; the reference should be to Theorem 2.4 or Remark 2.8.
[Throughout] The remainder O(1) in (3.18) should specify that it is an H^{-1}-norm bound; otherwise the order-of-magnitude statement is ambiguous.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the SDE and stability results are derived from explicit projection and external spectral facts, not from fitted inputs or self-referential chains.

full rationale

The paper's central derivation is self-contained given the imported deterministic machinery. In Section 3, the SDE for the droplet center is not assumed as a prediction: the authors posit a semimartingale ansatz dξ_k = f_k dt + <σ_k,dW>, solve the resulting Itô consistency equations for f and σ in (3.11)-(3.12), and then verify in Lemma 3.2 that this produces a solution of the original SPDE with the orthogonality constraint preserved. This is an explicit derivation, not an equivalence to an input. The stability estimates in Section 4 depend on the deterministic spectral gap Theorem 2.4, imported from Alikakos and Fusco [2,3]. Those authors are not authors of the present paper, so this is external support rather than a self-citation chain. The long-time martingale estimate borrowed from [7] is a general, parameter-free technical lemma; although Blömker is a coauthor of [7], the lemma does not contain the target stochastic Cahn-Hilliard result and is used only to convert the differential inequality (4.10) into tail estimates. No constants are fitted, and no part of the claimed prediction is a renamed fitted parameter. The sensitivity noted in Remark 2.6 (the d=3 spectral gap would be O(ε^2)) is a robustness caveat about an imported assumption, not circularity. There is an apparent typo in Theorem 4.7, where the initial H^{-1} bound is stated as ν ε^4 even though the exit radius is ε^m for m>4; this would make the statement formally weaker if read literally, but it is an internal consistency issue, not a circular reduction of the claim to its inputs.

Assumptions & free parameters 0 free parameters · 8 assumptions · 0 invented entities

The central claim rests on the prior deterministic slow-manifold theory and spectral bounds of [2,3,1,6], and on the stochastic moment method of [7]. No free parameters are fitted to data, and no new entities are introduced. The main new work is the derivation of the SDE coefficients and the adaptation of the stability argument to the 2D fourth-order Cahn-Hilliard operator.

assumptions (8)

domain assumption There exists a unique solution to the stochastic Cahn-Hilliard equation (1.1) for the given Q-Wiener process.
Invoked in Section 1.1 and after (1.1), citing [19,16]; the paper states 'we always assume that we have a unique solution.'
domain assumption The noise W is a Q-Wiener process in H^{-1}_0 satisfying the trace conditions N1-N3 and is smooth in space.
Section 1.1; needed for the Ito formula and for the L2 estimates, where eta_2 = trace(-Delta Q) appears.
domain assumption The deterministic slow manifold M_tilde_epsilon^rho with exponentially small residual exists (Theorem 2.2 from [3]).
Section 2.2; imported from Alikakos-Fusco [3], including the droplet family u_xi and the correction v_xi.
domain assumption Spectral estimates for the linearized Cahn-Hilliard operator in H^{-1}: eigenvalues lambda_1, lambda_2 are exponentially small, lambda_3 >= C' epsilon, and eigenfunctions approximate the tangent space (Theorem 2.4 from [2]).
Section 2.3.1; the order-epsilon spectral gap drives the damping in Lemma 4.2 and Theorem 4.4.
domain assumption Spectral estimates for the mass-conserving Allen-Cahn operator in L2: mu_1, mu_2 are exponentially small, mu_3 >= C epsilon^2 (Theorem 2.7 from [1]).
Section 2.3.2; used in the decomposition (4.7) of the inner product of the linearized Cahn-Hilliard operator.
domain assumption The optimal-stopping moment inequalities from [7] apply to the scalar inequality (4.10) and to the L2 estimate (4.17).
Sections 4.2 and 4.4; the paper imports these results without stating the exact lemma or its hypotheses.
domain assumption The Fermi coordinate projection from Proposition 3.1 is well-defined and the matrix A(v,xi) is invertible for ||v|| < epsilon^eta in a tube around the manifold.
Section 3.1, Proposition 3.1 from [3]; needed for the SDE derivation and for Lemma 3.3.
domain assumption The solution and the center xi(t) remain in the region where the projection is defined up to time T_epsilon; the case of hitting the boundary is not treated.
Section 4.2: 'Note that we neglect the case that xi(t) not in Omega_{delta+rho} at some point.'

how reviews work

0 comments

Cite this review

Pith. "Pith review of Stochastic Cahn-Hilliard equation in higher space dimensions: The motion of bubbles." pith.science (2026). https://pith.science/paper/JJQUICHI

@misc{pith2026190801601,
  author       = {Pith},
  title        = {Pith review of: Stochastic Cahn-Hilliard equation in higher space dimensions: The motion of bubbles},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/JJQUICHI}},
  note         = {Machine review of arXiv:1908.01601}
}

read the original abstract

We study the stochastic motion of a droplet in a stochastic Cahn-Hilliard equation in the sharp interface limit for sufficiently small noise. The key ingredient in the proof is a deterministic slow manifold, where we show its stability for long times under small stochastic perturbations. We also give a rigorous stochastic differential equation for the motion of the center of the droplet.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

23 extracted references · 23 canonical work pages

[7]

Antonopoulou, P

D. Antonopoulou, P. Bates, D. Bl ¨omker, and G. Karali , Mo- tion of a droplet for the mass-conserving stochastic Allen- Cahn equation, SIAM J Math. Anal., 48 (2016), pp. 670 – 708

work page 2016
[8]

Antonopoulou, D

D. Antonopoulou, D. Bl ¨omker, and G. Karali , Front motion in the one-dimensional stochastic Cahn-Hilliard equation , SIAM J. Math. Anal., 44 (2012)

work page 2012
[1]

Alikakos, L

N. Alikakos, L. Bronsard, and G. Fusco , Slow motion in the gradient theory of phase transitions via energy and spectru m, Calc. Var., 6 (1998), pp. 39 – 66

work page 1998
[2]

Alikakos and G

N. Alikakos and G. Fusco , Slow dynamics for the Cahn-Hilliard equation in higher space dimension part I: Spectral estimat es, Commu- nications in Partial Diﬀerential Equations, (1994)

work page 1994
[3]

Alikakos and G

N. Alikakos and G. Fusco , Slow dynamics for the Cahn-Hilliard equation in higher space dimensions: The motion of bubbles , Arch. Ra- tional Mech. Anal., 141 (1998), pp. 1 – 68

work page 1998
[4]

N. D. Alikakos, P. W. Bates, and X. Chen , Convergence of the Cahn-Hilliard equation to the hele-shaw model , Archive for Rational Me- chanics and Analysis, 128 (1994), pp. 165–205

work page 1994
[5]

N. D. Alikakos, X. Chen, and G. Fusco , Motion of a droplet by surface tension along the boundary , Cal. Var., 11 (2000)

work page 2000
[6]

N. D. Alikakos, G. Fusco, and G. Karali , Motion of bubbles towards the boundary for the Cahn-Hilliard equation , European J. Appl. Math., 15 (2004), pp. 103–124

work page 2004

Show all 23 references

[9]

Banas, H

L. Banas, H. Yang, and R. Zhu , Sharp interface limit of stochastic Cahn-Hilliard equation with singular noise . arXiv:1905.07216, 2019

1905 arXiv
[10]

P. W. Bates , Coarsening and nucleation in the Cahn-Hilliard equa- tion, in Free boundary problems involving solids (Montreal, PQ, 1990), vol. 281 of Pitman Res. Notes Math. Ser., Longman Sci. Tech., Harlow , 1993, pp. 220–225

1990
[11]

P. W. Bates and P. C. Fife , The dynamics of nucleation for the Cahn-Hilliard equation, SIAM J. Appl. Math., 53 (1993), pp. 990–1008. 35

1993
[12]

P. W. Bates and J. Jin , Global dynamics of boundary droplets , Dis- crete and Continuous Dynamical Systems - A, 34 (2014)

2014
[13]

Bl ¨omker, B

D. Bl ¨omker, B. Gawron, and T. W anner , Nucleation in the one-dimensional stochastic Cahn-Hilliard model , Discrete Contin. Dyn. Syst., 27 (2010), pp. 25–52

2010
[14]

Bl ¨omker, S

D. Bl ¨omker, S. Maier-Paape, and T. W anner , Phase separa- tion in stochastic Cahn-Hilliard models , in Mathematical methods and models in phase transitions, Nova Sci. Publ., New York, 2005, pp. 1–4 1

2005
[15]

Bl ¨omker, E

D. Bl ¨omker, E. Sander, and T. W anner , Degenerate nucleation in the Cahn-Hilliard-Cook model , SIAM J. Appl. Dyn. Syst., 15 (2016), pp. 459–494

2016
[16]

Cardon-Weber , Cahn-Hilliard stochastic equation: existence of the solution and of its density , Bernoulli, 7 (2001), pp

C. Cardon-Weber , Cahn-Hilliard stochastic equation: existence of the solution and of its density , Bernoulli, 7 (2001), pp. 777–816

2001
[17]

Chen, Global asymptotic limit of solutions of the Cahn-Hilliard e qua- tion, Journal of Diﬀerential Geometry, 44 (1996), pp

X. Chen, Global asymptotic limit of solutions of the Cahn-Hilliard e qua- tion, Journal of Diﬀerential Geometry, 44 (1996), pp. 262–311

1996
[18]

Cook , Brownian motion in spinodal decomposition , Acta Metallur- gica, 18 (1970), pp

H. Cook , Brownian motion in spinodal decomposition , Acta Metallur- gica, 18 (1970), pp. 297–306

1970
[19]

Da Prato and A

G. Da Prato and A. Debussche , Stochastic Cahn-Hilliard equation, Nonlinear Anal., 26 (1996), pp. 241–263

1996
[20]

Da Prato and J

G. Da Prato and J. Zabzcyck , Stochastic Equations in Inﬁnite Dimensions, Cambridge University Press, 1992

1992
[21]

Hellfer and S

B. Hellfer and S. Sjstrand , Multiple wells in the semi-classical limit i , Communications in Partial Diﬀerential Equations, 9 (1984), pp. 337 – 408

1984
[22]

Schindler , (in preparation) , PhD thesis, Universit¨ at Augsburg, 2020

A. Schindler , (in preparation) , PhD thesis, Universit¨ at Augsburg, 2020

2020
[23]

Yang and R

H. Yang and R. Zhu , Weak solutions to the sharp interface limit of stochastic Cahn-Hilliard equations . arXiv:1905.09182, 2019. 36

1905 arXiv

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.

[1] [7]

Antonopoulou, P

D. Antonopoulou, P. Bates, D. Bl ¨omker, and G. Karali , Mo- tion of a droplet for the mass-conserving stochastic Allen- Cahn equation, SIAM J Math. Anal., 48 (2016), pp. 670 – 708

work page 2016

[2] [8]

Antonopoulou, D

D. Antonopoulou, D. Bl ¨omker, and G. Karali , Front motion in the one-dimensional stochastic Cahn-Hilliard equation , SIAM J. Math. Anal., 44 (2012)

work page 2012

[3] [1]

Alikakos, L

N. Alikakos, L. Bronsard, and G. Fusco , Slow motion in the gradient theory of phase transitions via energy and spectru m, Calc. Var., 6 (1998), pp. 39 – 66

work page 1998

[4] [2]

Alikakos and G

N. Alikakos and G. Fusco , Slow dynamics for the Cahn-Hilliard equation in higher space dimension part I: Spectral estimat es, Commu- nications in Partial Diﬀerential Equations, (1994)

work page 1994

[5] [3]

Alikakos and G

N. Alikakos and G. Fusco , Slow dynamics for the Cahn-Hilliard equation in higher space dimensions: The motion of bubbles , Arch. Ra- tional Mech. Anal., 141 (1998), pp. 1 – 68

work page 1998

[6] [4]

N. D. Alikakos, P. W. Bates, and X. Chen , Convergence of the Cahn-Hilliard equation to the hele-shaw model , Archive for Rational Me- chanics and Analysis, 128 (1994), pp. 165–205

work page 1994

[7] [5]

N. D. Alikakos, X. Chen, and G. Fusco , Motion of a droplet by surface tension along the boundary , Cal. Var., 11 (2000)

work page 2000

[8] [6]

N. D. Alikakos, G. Fusco, and G. Karali , Motion of bubbles towards the boundary for the Cahn-Hilliard equation , European J. Appl. Math., 15 (2004), pp. 103–124

work page 2004