[Paper Review] Upper bounds of nodal sets for eigenfunctions of eigenvalue problems

Fanghua Lin, Jiuyi Zhu|arXiv (Cornell University)|May 8, 2020

Numerical methods in inverse problems34 references4 citations

TL;DR

This paper establishes sharp upper bounds for the Hausdorff measure of nodal sets of eigenfunctions in real analytic domains for various higher-order elliptic eigenvalue problems, including biharmonic, buckling, and clamped-plate problems. Using doubling inequalities derived from Carleman estimates and growth estimates, it proves that the nodal set measure is bounded by $ C au^{1/2m} $, where $ m $ is the order of the operator, generalizing results from second-order problems to higher-order systems with unified analytic techniques.

ABSTRACT

The aim of this article is to provide a simple and unified way to obtain the sharp upper bounds of nodal sets of eigenfunctions for different types of eigenvalue problems on real analytic domains. The examples include biharmonic Steklov eigenvalue problems, buckling eigenvalue problems and champed-plate eigenvalue problems. The geometric measure of nodal sets are derived from doubling inequalities and growth estimates for eigenfunctions. It is done through analytic estimates of Morrey-Nirenberg and Carleman estimates.

Motivation & Objective

• To provide a unified framework for deriving sharp upper bounds on the size of nodal sets for eigenfunctions of higher-order elliptic eigenvalue problems on real analytic domains.
• To extend known results for second-order Laplacian eigenfunctions to biharmonic and other higher-order operators with various boundary conditions.
• To establish quantitative control on nodal set measure using analytic tools such as doubling inequalities and growth estimates.
• To generalize the approach to eigenvalue problems including biharmonic Steklov, buckling, clamped-plate, and Navier/Dirichlet-type problems.
• To demonstrate that the nodal set measure grows at most like $ \lambda^{1/2m} $, where $ m $ is the order of the differential operator.

Proposed method

• Utilizes quantitative Carleman estimates for higher-order elliptic operators $ (-\triangle)^m - \lambda $, derived via iterative application of fundamental Carleman inequalities.
• Applies lifting techniques to extend eigenfunctions into higher-dimensional space, enabling analytic continuation across the boundary using solutions to extended equations.
• Establishes growth estimates for eigenfunctions in a slightly enlarged domain, showing $ \|e_\lambda\|_{L^\infty(\widetilde{\Omega})} \leq e^{C\lambda^{1/2m}} \|e_\lambda\|_{L^\infty(\Omega)} $.
• Derives doubling inequalities: $ \|e_\lambda\|_{L^\infty(\mathbb{B}_{2r}(x_0))} \leq e^{C\lambda^{1/2m}} \|e_\lambda\|_{L^\infty(\mathbb{B}_r(x_0))} $ for all $ x_0 \in \overline{\Omega} $, $ r \leq d/4 $.
• Applies complex growth lemmas to relate doubling constants to the Hausdorff measure of nodal sets.
• Uses the decomposition $ (-\triangle)^m - \lambda = \prod_{k=0}^{m-1} (-\triangle - \lambda^{1/m} e^{2k\pi i/m}) $ to handle the spectral structure of higher-order operators.

Experimental results

Research questions

• RQ1 What is the sharp upper bound for the $ (n-1) $-dimensional Hausdorff measure of the nodal set of eigenfunctions in higher-order elliptic eigenvalue problems on real analytic domains?
• RQ2 Can a unified analytic method be developed to bound nodal sets across different types of higher-order eigenvalue problems, such as biharmonic, buckling, and clamped-plate problems?
• RQ3 How do doubling inequalities and Carleman estimates interact to control the growth and nodal structure of eigenfunctions in higher-order systems?
• RQ4 What is the dependence of the nodal set measure on the eigenvalue $ \lambda $ and the order $ m $ of the differential operator?
• RQ5 To what extent can analytic continuation and growth estimates be used to derive optimal bounds for nodal sets beyond second-order operators?

Key findings

• For eigenfunctions of $ m $-th order elliptic eigenvalue problems with Dirichlet or Navier boundary conditions, the $ (n-1) $-dimensional Hausdorff measure of the nodal set is bounded by $ C\lambda^{1/2m} $, where $ C $ depends only on the domain.
• The upper bound $ H^{n-1}(\{x \in \Omega \mid e_\lambda(x) = 0\}) \leq C\lambda^{1/2m} $ is sharp and applies uniformly across biharmonic, buckling, clamped-plate, and Steklov-type problems.
• The proof relies on a novel application of iterative Carleman estimates to derive doubling inequalities, which are then used with complex growth lemmas to control nodal set size.
• The method extends to eigenfunctions of the Laplacian with various boundary conditions, showing that the $ \lambda^{1/2} $ bound is recoverable via the same framework.
• The growth estimate $ \|e_\lambda\|_{L^\infty(\widetilde{\Omega})} \leq e^{C\lambda^{1/2m}} \|e_\lambda\|_{L^\infty(\Omega)} $ is crucial for controlling the size of eigenfunctions in the extended domain.
• The result generalizes the Donnelly-Fefferman and Logunov-Malinnikova bounds for second-order problems to higher-order operators, with the exponent $ 1/2m $ reflecting the increased regularity and decay.