Minimizers for the Cahn-Hilliard energy functional with the Flory-Huggins potential under strong anchoring conditions

1. Problem Statement

The paper investigates the existence, uniqueness, and bifurcation behavior of minimizers for the Cahn–Hilliard energy functional defined on a bounded domain $\Omega \subset \mathbb{R}^n$:
$$E(u) = \int_{\Omega} \left( \frac{\kappa}{2} |\nabla u|^2 + W(u) \right) dx$$
where:

• $u$ represents the order parameter (concentration difference).
• $\kappa > 0$ is a parameter related to the transition layer thickness.
• $W(u)$ is the Flory–Huggins potential (also known as the logarithmic potential), given by:
  $$W(u) = \begin{cases} \frac{\theta}{2} ((1-u)\ln(1-u) + (1+u)\ln(1+u)) + \frac{1}{2}(1-u^2) & \text{if } |u| \le 1 \\ +\infty & \text{otherwise} \end{cases}$$
  Here, $\theta \in (0, 1)$ is the rescaled temperature. The logarithmic terms enforce the physical constraint $|u| < 1$ and model the entropy of mixing.

Boundary Conditions: The study focuses on strong anchoring (Dirichlet) conditions, specifically the homogeneous case $u = 0$ on $\partial\Omega$. This represents a rigid boundary where the mixture is fixed to a homogeneous state.

Objective: To determine how the minimizers of $E(u)$ behave as a function of the parameters $\kappa$ (transition layer thickness) and $\theta$ (temperature), specifically addressing whether a bifurcation occurs (i.e., a transition from a unique trivial solution to non-trivial symmetric solutions).

2. Methodology

The authors employ a combination of variational analysis, functional approximation, and numerical simulation.

A. Handling the Logarithmic Singularity

The primary mathematical challenge is the singularity of $W(u)$ at $u = \pm 1$, which prevents direct application of standard variational arguments.

• Decomposition: The potential is split into a singular convex part $W_1$ and a smooth part $W_2$.
• Global Approximation: The authors construct a smooth global approximation $\tilde{W}$ of $W$. This modified potential coincides with $W$ on the physically relevant region $|u| \le \hat{u}_\theta$ (where $\hat{u}_\theta$ is close to the minimizer $u_\theta$) but is strictly convex and smooth outside this region.
• Equivalence: They prove that minimizers of the modified energy $\tilde{E}$ (using $\tilde{W}$) satisfy $|u| \le u_\theta$ almost everywhere, ensuring they are also minimizers of the original energy $E$.

B. Variational Analysis via Nehari Manifold

The authors utilize the Nehari manifold to analyze critical points:
$$S = \{ u \in H^1_0(\Omega) : \langle \delta E(u), u \rangle = 0 \}$$

• Scaling Argument: For a fixed direction $u$, they analyze the function $\Phi_u(s) = E(su)$.
• Eigenvalue Analysis: They relate the existence of non-trivial solutions to the smallest eigenvalue $\lambda_1$ of the negative Laplacian with Dirichlet boundary conditions.
• Bifurcation Threshold: By analyzing the second derivative of the energy functional, they derive a critical threshold involving $\kappa$, $\theta$, and $\lambda_1$.

C. Maximum Principle and Regularity

• Regularity: Using Sobolev embeddings and bootstrap arguments, they establish that minimizers are smooth ($C^\infty$) inside the domain.
• Strong Maximum Principle: They prove that non-trivial minimizers are strictly positive (or negative) in the interior of $\Omega$ and strictly bounded by the potential's minimizer $u_\theta$.

D. Numerical Simulations

• Equation: They solve the gradient flow equation (Allen–Cahn equation) constrained with strong anchoring:
  $$u_t = \kappa \Delta u - W'(u), \quad u|_{\partial\Omega} = 0$$
• Scheme: Forward Euler in time, five-point central difference for the Laplacian, with random initial data.
• Validation: They compare numerical equilibrium states against theoretical bounds derived from the potential's minimizers.

3. Key Contributions

1. Resolution of an Open Problem: Previous work by Dai et al. established bifurcation results for quartic and trigonometric potentials under strong anchoring. This paper extends these results to the physically realistic Flory–Huggins potential, which is significantly more difficult due to its logarithmic singularity.
2. Unified Bifurcation Criterion: The paper provides a rigorous proof that the bifurcation behavior is governed by the interplay between the transition layer thickness $\kappa$, the temperature $\theta$, and the domain geometry (via $\lambda_1$).
3. Role of Temperature: Unlike the quartic case where only $\kappa$ drives bifurcation, the Flory–Huggins potential introduces $\theta$ as a critical bifurcation parameter. The condition for non-trivial solutions depends on the ratio $(1-\theta)/\lambda_1$.
4. Approximation Technique: The construction of the smooth approximation $\tilde{W}$ allows the use of standard variational tools while preserving the physical constraints of the logarithmic potential.

4. Main Results (Theorem 1.1)

Let $\lambda_1$ be the smallest eigenvalue of $-\Delta$ on $\Omega$ with Dirichlet boundary conditions. The behavior of the minimizers is determined by the inequality:

• Case 1: Unique Trivial Solution
  If $\kappa \ge \frac{1-\theta}{\lambda_1}$, the energy functional $E$ has a unique minimizer $u^* \equiv 0$.

  • Physically: The transition layer is too thick (or temperature too high) to support phase separation against the strong anchoring at the boundary.

• Case 2: Symmetric Non-Trivial Solutions (Bifurcation)
  If $0 < \kappa < \frac{1-\theta}{\lambda_1}$, $E$ has two non-trivial minimizers $u^*_+$ and $u^*_-$, where $u^*_- = -u^*_+$.

  • Properties: $0 < u^*_+ < u_\theta$ pointwise in $\Omega$.
  • Energy: $E(u^*) < E(0) = |\Omega|/2$.
  • Physically: The system undergoes phase separation, forming a transition layer that avoids the boundary constraint, resulting in two symmetric stable states.

5. Significance and Implications

• Physical Relevance: The Flory–Huggins potential is the standard model for polymer blends and binary mixtures. Understanding minimizers under strong anchoring is crucial for modeling thin films and interfaces where surface interactions are rigid.
• Theoretical Advancement: The paper demonstrates that bifurcation phenomena are robust across different classes of double-well potentials, provided appropriate analytical techniques (like the approximation method used here) are applied to handle singularities.
• Numerical Insights: The simulations confirm that as $\kappa$ approaches the critical threshold from below, the amplitude of the minimizer decreases toward zero. Furthermore, the results show that lower temperatures ($\theta$) and smaller $\kappa$ values lead to more pronounced phase separation (larger maximum values of $u$).
• Future Directions: The authors suggest that these methods can be extended to Neumann, dynamic, or mixed boundary conditions, potentially revealing richer bifurcation structures in more complex physical settings.

In summary, this paper successfully bridges the gap between theoretical variational analysis and physical reality for phase separation models, proving that strong anchoring induces a specific bifurcation structure dependent on both material properties ($\theta$) and interfacial energy ($\kappa$).