Liouville theorems and $1$-dimensional symmetry for solutions of an elliptic system modelling phase separation

TL;DR

This paper classifies solutions to a competitive elliptic system based on their growth rate, establishes minimal growth bounds related to the number of components, and proves 1-dimensional symmetry under certain conditions.

Contribution

It introduces a growth classification linked to the number of components and extends symmetry results to solutions with more than two components.

Findings

Existence of a minimal growth rate depending on the number of components and dimension.

Explicit optimal growth bounds in two dimensions.

Extension of 1-dimensional symmetry results to solutions with multiple components.

Abstract

We consider solutions of the competitive elliptic system \[ \left\{ \begin{array}{ll} -\Delta u_i = - \sum_{j \neq i} u_i u_j^2 & \text{in $\mathbb{R}^N$} \\ u_i >0 & \text{in $\mathbb{R}^N$} \end{array}\right. \qquad i=1,\dots,k. \] We are concerned with the classification of entire solutions, according with their growth rate. The prototype of our main results is the following: there exists a function $\delta=\delta(k,N) \in \mathbb{N}$, increasing in $k$, such that if $(u_1,\dots,u_k)$ is a solution and \[ u_1(x)+\cdots+u_k(x) \le C(1+|x|^d) \qquad \text{for every $x \in \mathbb{R}^N$}, \] then $d \ge \delta$. This means that the number of components $k$ of the solution imposes an increasing in $k$ minimal growth on the solution itself. If $N=2$, the expression of $\delta$ is explicit and optimal, while in higher dimension it can be characterized in terms of an optimal partition problem. We discuss the sharpness of our results and, as a further step, for every $N \ge 2$ we can prove the $1$-dimensional symmetry of the solutions satisfying suitable assumptions, extending known results which are available for $k=2$. The proofs rest upon a blow-down analysis and on some monotonicity formulae.