Weak solutions and sharp interface limit of the anisotropic Cahn-Hilliard equation with disparate mobility and inhomogeneous potential
Charles Elbar, Andrea Poiatti · 2025
Read Weak solutions and sharp interface limit of the anisotropic Cahn-Hilliard equation with disparate mobility and inhomogeneous potential by Charles Elbar, Andrea Poiatti (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2502-11849.txt.
Introduction
Charles Elbar and Andrea Poiatti wrote this 2025 mathematics paper on an equation that models how two phases of a material separate. The anisotropic Cahn-Hilliard equation tracks that split, but here one phase barely diffuses at all. Its double-well potential also varies from place to place across the domain.
The authors first prove that weak solutions to the equation exist. Then they study what happens as the width of the border between phases shrinks to zero. In that sharp interface limit, and with an energy condition in place, the solutions settle into a weighted anisotropic Hele-Shaw flow. That flow is a simpler model of motion through a narrow gap. The paper adds computer simulations that show how the direction-dependent traits change the patterns the equation makes.
Catalog record
| CREATOR | Charles Elbar, Andrea Poiatti |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 134 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2502-11849
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Weak solutions and sharp interface limit of the anisotropic Cahn-Hilliard equation with disparate mobility and inhomogeneous potential Charles Elbar, Andrea Poiatti Abstract We study the existence of weak solutions and the corresponding sharp interface limit of an anisotropic Cahn-Hilliard equation with disparate mobility, i.e., the mobility is degenerate in one of the two pure phases, making the diffusion in that phase vanish. The double-well potential is polynomial and is weighted by a spatially inhomogeneous coefficient. In the limit when the parameter of the interface width tends to zero, and under an energy convergence assumption, we prove that the weak solutions converge to B V BV solutions of a weighted anisotropic Hele-Shaw flow. 2020 Mathematics Subject Classification. 35B40, 35K65, 35G20, 53E40, 76D27. Keywords and phrases. Anisotropic Cahn-Hilliard equation; disparate mobility; spatially inhomogeneous potential; sharp interface limit; weighted anisotropic Hele-Shaw flow. 1 Introduction The Cahn–Hilliard equation, introduced by J. W. Cahn and J. E. Hilliard in 1958 [ 14 ] , models the phase separation process in binary alloys, known as spinodal decomposition. Unlike models with sharp boundaries, the Cahn–Hilliard equation describes a diffuse interface between phases, whose width is controlled by a parameter \varepsilon>0 . As \varepsilon tends to zero, interfaces sharpen, eventually forming evolving hypersurfaces. This makes the Cahn–Hilliard model a link between phase-field continuum models and Hele-Shaw (or Mullins-Sekerka) models, with applications in materials science, biology, and other fields. Initially, the Cahn–Hilliard equation was assumed isotropic, i.e., enjoying uniform physical properties in all directions. However, many real-world systems are anisotropic, with properties like surface tension varying with interface orientation. This is evident in crystalline materials (see, e.g., [ 60 , 43 ] ), where atomic arrangements lead to direction-dependent interface evolution, forming facets or corners. The effect of anisotropies can also be visualized in the numerical simulations of the following Appendix B. . Another application of the anisotropic Cahn–Hilliard equation is to model the growth of thin solid films, which has a role in the self-organization of nanostructures (cf. [ 57 , 61 ] ). Anisotropy is also prevalent in biological contexts, such as tissue mechanics, where properties like stiffness depend on the alignment of structures like collagen fibers. In tumor growth, for instance, the interaction between cancer cells and the extracellular matrix can result in anisotropic tumor shapes [ 19 ] . Incorporating anisotropy into the Cahn–Hilliard equation leads to the anisotropic Cahn–Hilliard equation . When the mobility is degenerate in one of the pure phases, taking the sharp-interface limit ( \varepsilon\to 0 ) in this setting results in an anisotropic Hele-Shaw flow, accounting for direction-dependent effects. The presence of a space dependent potential then leads to a weighted anisotropic Hele-Shaw flow. In this work, we investigate the anisotropic Cahn–Hilliard equation with disparate mobility and space dependent potential on a d -dimensional torus \Omega=\mathbb{T}^{d} , d=2,3 . The model reads: \partial_{t}u={\rm div}(u\nabla\mu),\quad\mu=-\varepsilon{\rm div}(A_{p}(x,\nabla u))+\frac{1}{\varepsilon}F_{u}(x,u)\quad\text{in }(0,T)\times\Omega, (1.1) or alternatively, introducing the flux j , \partial_{t}u={\rm div}\ j,\quad j=-u\nabla\mu=u\nabla\left(\varepsilon{\rm div}(A_{p}(x,\nabla u))-\frac{1}{\varepsilon}F_{u}(x,u)\right)\quad\text{in }(0,T)\times\Omega. (1.2) Here, T>0 is the final time. The parameter \varepsilon>0 relates to the interface thickness. The function A=A(x,p) is C^{1} in both variables, two-homogeneous in p , and satisfies strong convexity and growth conditions (see Assumption ( A1 )). The potential F(x,u) is a polynomial double-well potential depending on (x,u) and satisfying certain assumptions (see Assumption ( A2 )), namely it can be factorized as K(x)W(u) , where W(u) is the standard double well quartic potential (which is zero at the pure phases u=0,1 ) and K is a C^{1} strictly positive weight, representing some heterogeneity in the material. The notation A_{p} and A_{x} denotes the gradients of A(x,p) with respect to p and x , respectively, with similar definitions for F_{u} and F_{x} . Observe that the mobility u in front of \nabla\mu degenerates only in one phase, that is u=0 . This is different from the standard degenerate mobility case, in which the mobility degenerates in both the phases u=0 and u=1 . Our main goal is to prove the existence of global weak solutions for ( 1.1 ) under the stated assumptions and to analyze the sharp interface limit as \varepsilon\to 0 . In particular, we show that, assuming an appropriate energy convergence assumption, a weak solution of the anisotropic Cahn–Hilliard equation converges to a solution of the anisotropic weighted Hele–Shaw flow. We also add some numerical experiments to study the effects of anisotropy in the Cahn-Hilliard equation. System ( 1.1 ) comes with two functionals, useful to obtain suitable estimates. First of all, the system can be viewed as the Wasserstein gradient flow of the energy functional E[u]:=\int_{\Omega}\varepsilon\,A(x,\nabla u(x))+\frac{1}{\varepsilon}F(x,u(x))\mathop{}\!\mathrm{d}x. (1.3) In particular the solutions of ( 1.1 ) decrease along the energy. A key auxiliary functional we also use is the so-called entropy : \Phi(u):=\int_{\Omega}u(x)\log u(x)\mathop{}\!\mathrm{d}x. (1.4) Energy and entropy formally satisfy the relations \displaystyle\frac{dE}{dt}+\int_{\Omega}u\,|\nabla\mu|^{2}\mathop{}\!\mathrm{d}x=0, (1.5) \displaystyle\frac{d\Phi}{dt}+\int_{\Omega}(\varepsilon{\rm div}(A_{p}(x,\nabla u))-\frac{1}{\varepsilon}F_{u}(x,u))\,\Delta u\mathop{}\!\mathrm{d}x=0. (1.6) The previous identities are crucial to find a priori estimates and prove existence of weak solutions and the sharp interface limit, as these two results are based on compactness arguments. Concerning the sharp interface limit as \varepsilon\to 0 , given a solution (u_{\varepsilon},j_{\varepsilon}) satisfying for any \varepsilon>0 equations ( 1.1 ) in suitable weak form, the main ingredient is the energy estimate ( 1.5 ), which allows to control the main quantity through the well-known Modica–Mortola trick [ 54 ] , namely, defining \psi(s)=2\int_{0}^{s}\sqrt{W(r)}\,\mathop{}\!\mathrm{d}r (1.7) we can control in L^{1} the gradient of \psi\circ u_{\varepsilon} by means of the energy inequality ( 1.5 ). This allows to obtain the desired compactness for passing to the limit. Another essential ingredient is the observation that from the second identity in ( 1.2 ), after some simple computations, we have \displaystyle j_{\varepsilon} \displaystyle=-\text{div}\ \mathbf{T}_{\varepsilon}+\frac{1}{\varepsilon}F_{x}(x,u_{\varepsilon})+\varepsilon A_{x}(x,\nabla u_{\varepsilon}) \displaystyle\quad+\nabla\left(\varepsilon\nabla u_{\varepsilon}\cdot A_{p}(x,\nabla u_{\varepsilon})+\frac{1}{\varepsilon}u_{\varepsilon}F_{u}(x,u_{\varepsilon})-\text{div}(\varepsilon u_{\varepsilon}A_{p}(x,\nabla u_{\varepsilon}))\right), (1.8) where \mathbf{T}_{\varepsilon} denotes the anisotropic energy stress tensor \displaystyle\mathbf{T}_{\varepsilon}:=\left(\varepsilon A(x,\nabla u_{\varepsilon})+\frac{1}{\varepsilon}F(x,u_{\varepsilon})\right)Id-\varepsilon\left(\nabla u_{\varepsilon}\otimes A_{p}(x,\nabla u_{\varepsilon})\right). (1.9) Observe that the right-hand side of ( 1.8 ), excluding the terms with F_{x} and A_{x} (which indeed are less standard and need to be treated more carefully), is in divergence form. This allows to test by smooth test functions and integrate by parts. The resulting weak formulation has then good compactness properties, so that given a sequence of weak solutions, one only has to guarantee energy convergence to prove that the limit is a weak solution. Indeed, one only needs to pass to the limit in first-order terms. In conclusion, to pass to the limit as \varepsilon\to 0 in the weak formulation of ( 1.8 ) we will exploit two different methods, i.e., by means of a suitable version of Reshetnyak continuity theorem (similar to [ 18 ] ) and by means of an anisotropic tilt excess approach (in the spirit of [ 44 , 48 , 47 ] ). We now review in the section below the existing literature on the Cahn–Hilliard equation and the study of its sharp interface limit. 1.1 Literature review To get a better understanding of the results presented here, we begin by revisiting the well-known isotropic Cahn-Hilliard equation, which can be seen as a particular case of our study when A(x,\nabla u)=\frac{1}{2}\lvert\nabla u\rvert^{2}\quad\text{and}\quad F(x,u)=u^{2}(1-u)^{2}. This isotropic model has been extensively studied, providing a solid basis of techniques that we adapt