Phase transitions with bounded index: Parallels to De Giorgi's conjecture
Enric Florit-Simon
TL;DR
This work establishes a sharp finite-index De Giorgi-type rigidity for the Allen–Cahn equation under bounded energy density, proving that in $\mathbb{R}^4$ any finite Morse index solution with ${\mathcal E}(u,B_R)\le C R^3$ is one-dimensional, and extending the same conclusion to $4\le n\le 7$ under a stable-classification assumption. The authors develop an annular improvement of flatness to obtain a global decomposition of the zero level set into nearly parallel sheets and then derive a Toda-type system governing inter-sheet distances, complemented by a refined stability analysis. They also demonstrate a geometric application in closed 4-manifolds, showing smooth phase-transition layers with minimal-hypersurface-like behavior under bounded energy and index. In dimension $n=3$, they confirm a parallel-ends picture for finite-energy solutions, aligning with the known minimal-surface analogies and providing a complete end-structure classification under quadratic energy growth. Overall, the paper reveals a striking rigidity for higher-dimensional phase transitions, contrasting with the richer variety in low dimensions and connecting phase-transition geometry to minimal-surface theory via a robust PDE-variational framework.
Abstract
A well-known conjecture of De Giorgi -- motivated by analogy with the Bernstein problem for minimal surfaces -- asserts the rigidity of monotone solutions to the Allen--Cahn equation in $\mathbb{R}^{d+1}$, with $d\leq 7$. We establish close parallels to De Giorgi's conjecture for general solutions of bounded Morse index, far stronger than the minimal surface analogy would suggest: Namely, any finite index solution to the Allen--Cahn equation with bounded energy density in $\mathbb{R}^4$ is one-dimensional, and -- conditionally on the classification of stable solutions -- the same holds for all $4\leq n \leq 7$. As a geometric application, phase transitions with bounded energy and index in closed four-manifolds have smooth transition layers which behave like minimal hypersurfaces. Consequently, phase transitions exhibit a remarkably rigid behaviour in higher dimensions. This is in stark contrast with the 3D case, in which a wealth of nontrivial entire solutions with finite index (and energy density) is conversely known to exist, by work pioneered by Del Pino--Kowalczyk--Wei. The authors conjectured that any such solution must have parallel ends which are either planar or catenoidal, suggesting it as a parallel to De Giorgi's conjecture in this framework. We confirm this picture under the bounded energy density assumption.
