Landis' conjecture for general second order elliptic equations with singular lower order terms in the plane

Abstract

In this article, we study the order of vanishing and a quantitative form of Landis' conjecture in the plane for solutions to second-order elliptic equations with variable coefficients and singular lower order terms. Precisely, we let $A$ be real-valued, bounded and elliptic, but not necessary symmetric or continuous, and we assume that $V$ and $W_i$ are real-valued and belong to $L^p$ and $L^{q_i}$, respectively. We prove that if $u$ is a real-valued, bounded and normalized solution to an equation of the form $-\nabla \cdot (A \nabla u + W_1 u) + W_2 \cdot \nabla u + V u = 0$ in $B_d$, then under suitable conditions on the lower order terms, for any $r$ sufficiently small, the following order of vanishing estimate holds $$\|u\|_{L^\infty(B_r)} \ge r^{C M},$$ where $M$ depends on the Lebesgue norms of the lower order terms. In a number of settings, a scaling argument gives rise to a quantitative form of Landis' conjecture, \[ \inf_{|z_0| = R} \|u\|_{L^\infty(B_1(z_0))} \ge \exp(- C R^\beta \log R), \] where $\beta$ depends on $p$, $q_1$, and $q_2$. The integrability assumptions that we impose on $V$ and $W_i$ are nearly optimal in view of a scaling argument. We use the theory of elliptic boundary value problems to establish the existence of positive multipliers associated to the elliptic equation. Then the proofs rely on transforming the equations to Beltrami systems and applying a generalization of Hadamard's three-circle theorem.