Long-range phase coexistence models with degenerate potentials

1. Problem Statement and Context

The paper investigates nonlocal phase transition problems modeled by Ginzburg–Landau type energy functionals. Unlike classical local models, these systems incorporate long-range particle interactions.

The primary energy functional considered is:
$$E_s(u; \Omega) := H_s(u, \Omega) + P(u, \Omega)$$
where:

• Nonlocal Kinetic Term ($H_s$): A fractional Sobolev seminorm representing long-range interactions:
  $$H_s(u, \Omega) := \frac{1}{4} \iint_{\mathbb{R}^{2n} \setminus (\mathbb{R}^n \setminus \Omega)^2} \frac{|u(x) - u(y)|^2}{|x-y|^{n+2s}} \, dx \, dy$$
  with $s \in (0, 1)$.
• Potential Term ($P$): A double-well potential integral:
  $$P(u, \Omega) := \int_{\Omega} W(u(x)) \, dx$$

Key Constraints and Definitions:

• Potential $W$: A smooth double-well potential with minima at $\pm 1$ (i.e., $W(\pm 1) = W'(\pm 1) = 0$).
• Degeneracy: The paper focuses on degenerate potentials, where $W''(\pm 1) = 0$, as opposed to the standard nondegenerate case where $W''(\pm 1) > 0$. The potential is assumed to have polynomial control on its second derivatives near the wells.
• Euler-Lagrange Equation: Critical points satisfy the fractional Allen–Cahn equation:
  $$L_s u(x) = W'(u(x))$$
  where $L_s$ is the fractional Laplacian.
• Target Objects: The authors study layer solutions (or transition layers), which are monotone heteroclinic connections interpolating between the pure phases $-1$ and $+1$.

2. Methodology

The survey employs a combination of variational analysis, asymptotic analysis, and inverse problem techniques:

1. Variational Framework:

   • Definition of Class A minimizers: Local minimizers in every bounded domain, necessary because the energy may diverge on unbounded domains.
   • Use of auxiliary scaling functions $\Psi_s(\rho)$ to track the growth rate of the energy at infinity, distinguishing cases $s < 1/2$, $s = 1/2$, and $s > 1/2$.
   • Analysis of the energy under general kernels $K(z)$ satisfying ellipticity conditions similar to the fractional Laplacian.

2. Asymptotic Analysis:

   • Derivation of sharp decay estimates for minimizers near the wells ($\pm 1$).
   • Investigation of how the polynomial rate of vanishing of $W''$ near the wells (parameters $\alpha, \beta, \gamma, \delta$) dictates the decay rate of the transition layer $u(x)$ at infinity.

3. Inverse Problem Approach:

   • Instead of fixing $W$ and finding $u$, the authors reverse the process: given a prescribed asymptotic profile for the layer $u$ (e.g., polynomial decay rates), they construct the potential $W$ that generates it.
   • This involves defining $W$ via the relation $W'(\phi) = L_s \phi$, where $\phi$ is the prescribed layer profile.

3. Key Contributions and Results

A. Classical Background and Symmetry (Section 2)

• De Giorgi's Conjecture: The paper reviews the nonlocal analogue of De Giorgi's conjecture (whether monotone solutions are 1-dimensional). It summarizes known results: proven for $n=2$, $n=3$ (for $s \geq 1/2$ and later all $s$), and $n=4$ ($s=1/2$), with counterexamples for $n \geq 9$.
• Connection to Minimal Surfaces: Reaffirms the link between fractional Allen–Cahn minimizers and nonlocal minimal surfaces via $\Gamma$-convergence and uniform density estimates.

B. Existence and Decay in Dimension $n=1$ (Section 3)

The authors present significant extensions of existing results to degenerate potentials:

• Existence and Uniqueness: Proven the existence of a unique (up to translation) strictly increasing class A minimizer for a broad class of degenerate potentials.
• Generalized Decay Estimates:
  • Unlike the nondegenerate case where decay is typically $|x|^{-2s}$, the decay rate here depends on the degeneracy order of the potential.
  • If the potential behaves like $W''(t) \sim |1 \pm t|^{\alpha-2}$ near the wells, the layer $u(x)$ satisfies:
    $$1 \pm u(x) \sim |x|^{-\frac{2s}{\alpha-1}}$$
  • Oscillatory Potentials: The framework allows for potentials that oscillate near the wells (where the upper and lower bounds of $W''$ differ, i.e., $\alpha \neq \beta$).
  • Optimality: The derived decay estimates are proven to be optimal. Specifically, if the potential is non-oscillatory ($\alpha=\beta$), all bounds are sharp; if oscillatory, the lower bounds are sharp.

C. Inverse Reconstruction of Potentials (Section 4)

This section addresses Problem 2: Can one reconstruct the potential $W$ from a prescribed layer profile?

• Construction: Given a layer $\phi$ with specific polynomial decay rates at infinity (e.g., $\phi(x) \approx 1 - C|x|^{-\beta}$ for $x>0$), the authors construct a potential $V$ such that $L_s \phi = V'(\phi)$.
• Regularity and Vanishing Orders:
  • The construction establishes a precise correspondence: a layer decaying as $|x|^{-\beta}$ corresponds to a potential $V$ vanishing at the well with order related to $\beta$.
  • Specifically, if the layer decays with exponent $\beta$, the potential $V$ near the well behaves like $(1-r)^{\frac{2s}{\beta} + 1}$.
  • Regularity: The smoothness of the reconstructed potential $V$ at the wells is determined by the decay rate of the layer. For instance, $V \in C^{\lfloor 2s/\beta \rfloor, 1}$.
• Significance of the Inverse Map: This proves that one can engineer potentials with specific degeneracy orders (including oscillatory ones) by choosing appropriate layer profiles. It highlights that the fractional Laplacian preserves power-law scaling properties in a non-trivial way, linking the decay at infinity to the vanishing order at the critical points.

4. Significance and Implications

1. Relaxation of Assumptions: The work moves beyond the standard assumption of nondegenerate (quadratic) wells. By allowing degenerate and oscillatory potentials, it models a wider range of physical phenomena, such as crystal dislocations with complex core structures or phase transitions with flat energy landscapes.
2. Quantitative Precision: The paper provides sharp, explicit decay estimates for transition layers. This is crucial for numerical simulations and for understanding the interaction between distant interfaces (e.g., in multi-layer configurations).
3. Inverse Problem Resolution: The ability to reconstruct the potential from the layer profile is a major theoretical advancement. It suggests that in experimental settings where the interface profile is observable (e.g., via microscopy), one can infer the underlying interaction potential, even if it is degenerate.
4. Bridge to Nonlocal Geometry: The results reinforce the deep connection between the fractional Allen–Cahn equation and nonlocal minimal surfaces, showing that the geometry of the interface is dictated by the specific degeneracy of the potential.

5. Future Directions Identified

The authors suggest several open problems:

• Investigating whether the upper bounds in decay estimates remain optimal for oscillatory potentials.
• Extending the inverse reconstruction to non-polynomial decay profiles (e.g., logarithmic or exponential), which may correspond to different physical regimes.
• Utilizing the reconstructed potentials to build barriers for proving rigidity results in degenerate settings (analogous to Theorem 6 in the nondegenerate case).

In summary, this survey synthesizes recent advances in nonlocal phase transitions, providing a rigorous framework for analyzing degenerate potentials and establishing a bidirectional link between the asymptotic behavior of transition layers and the local structure of the potential.