🔢 mathematics

Qualitative properties of eigenfunctions in domains with small holes

This paper investigates the qualitative properties of eigenvalues and eigenfunctions for the Laplacian with Dirichlet boundary conditions in a smooth bounded domain containing a small circular hole, providing quantitative estimates, proving eigenvalue simplicity, and analyzing nodal set behavior through pointwise estimates on the $u$ -capacitary potential.

Original authors: Laura Abatangelo, Massimo Grossi, Ying Li

Published 2026-08-03

📖 4 min read🧠 Deep dive

CC BY 4.0

Original authors: Laura Abatangelo, Massimo Grossi, Ying Li

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine a drumhead stretched tight over a circular frame. When you strike it, it doesn't just make a sound; it vibrates in specific, beautiful patterns called "modes." Some modes look like simple hills and valleys, while others have intricate lines where the drumskin doesn't move at all—these are called "nodal lines." In the world of mathematics and physics, these patterns are the solutions to a famous equation known as the Laplacian. Scientists have spent centuries studying how these shapes change when you tweak the drum's frame. Usually, they look at what happens when you gently stretch or squash the edge. But what if you poke a tiny, microscopic hole right in the middle of the drum? This is a much trickier problem. It's not a gentle nudge; it's a sudden, sharp tear in the fabric of the space. Understanding how the vibration patterns shift when a hole appears is crucial for engineers designing materials with tiny defects, or for physicists modeling how particles behave in confined spaces. It's the difference between a smooth wave and a wave that has to awkwardly navigate around a sudden obstacle.

This paper, written by Laura Abatangelo, Massimo Grossi, and Ying Li, dives deep into that exact scenario: a smooth, bounded domain (think of it as a perfectly shaped room or drum) with a tiny circular hole punched out of it. The authors are investigating the "eigenfunctions"—the mathematical shapes of the vibrations—and the "eigenvalues"—the specific frequencies at which they vibrate. They are particularly interested in what happens when that hole is placed right on top of a point where the original vibration was already zero (a nodal point) versus where it was strong.

The team discovered that the behavior of these vibrations depends entirely on how "deeply" the original wave touched zero at the spot where the hole is made. If the wave was just passing through zero (like a simple crossing), the vibration pattern shifts in one predictable way. But if the wave was "flat" against zero for a moment before rising again (a higher-order vanishing), the shift is much more complex. The authors proved that when you introduce this tiny hole, the single vibration frequency of the original room often splits into several distinct frequencies. They provided a precise recipe to calculate exactly how these new frequencies separate and what the new vibration shapes look like. They found that the new shapes are essentially the old shape minus a "correction" term that acts like a tiny, localized ripple around the hole, designed to force the vibration to be zero at the hole's edge.

One of the most fascinating findings concerns the "nodal sets"—the lines where the vibration is zero. The authors showed that if the hole is placed where the original wave was already zero, the new nodal lines don't just wiggle; they actually intersect the edge of the tiny hole in a very specific number of points. If the original wave had a "vanishing order" of $k$ (meaning it was flat for $k$ steps), the new nodal line will hit the edge of the hole exactly $2k$ times. It's as if the hole forces the wave to reorganize its zero-lines into a starburst pattern around the new obstacle.

The paper also tackles a common misconception: that making a small change always makes everything unique and simple. The authors showed that this isn't always true. Even with a tiny hole, some vibration frequencies might remain "doubled" (meaning two different shapes vibrate at the exact same frequency) if the geometry is too symmetric. They provided a mathematical test to see when the hole will successfully split these twins apart and when it won't. In short, they turned a messy, singular problem into a set of clear, calculable rules, showing exactly how a tiny puncture reshapes the music of the mathematical drum.

Technical Summary: Qualitative Properties of Eigenfunctions in Domains with Small Holes

Problem Statement
This paper investigates the asymptotic behavior of the eigenvalues and eigenfunctions of the Dirichlet Laplacian, $-\Delta$ , in a smooth bounded domain $\Omega \subset \mathbb{R}^N$ ( $N \ge 2$ ) when a small circular hole $B(P, \varepsilon)$ is removed. The perturbed domain is denoted as $\Omega_\varepsilon := \Omega \setminus B(P, \varepsilon)$ . The study focuses on "singular perturbations," contrasting with "regular perturbations" where only small deformations of the boundary $\partial \Omega$ occur. The primary objectives are to derive quantitative estimates for the perturbed eigenfunctions $u_\varepsilon$ , determine the simplicity of the perturbed eigenvalues $\lambda_\varepsilon$ , and analyze the behavior of the nodal sets of $u_\varepsilon$ near the hole.

Methodology
The analysis relies heavily on the concept of the $u$ -capacitary potential, $V_{\varepsilon, u}$ , introduced in previous works [1]. This function is the unique weak solution to the Dirichlet problem:
$\begin{cases} \Delta V_{\varepsilon, u} = 0 & \text{in } \Omega \setminus B(P, \varepsilon), \\ V_{\varepsilon, u} = 0 & \text{on } \partial \Omega, \\ V_{\varepsilon, u} = u & \text{on } \partial B(P, \varepsilon), \end{cases}$
where $u$ is an eigenfunction of the unperturbed problem.

The core methodological innovation of this paper is the derivation of sharp pointwise estimates for $V_{\varepsilon, u}$ as $\varepsilon \to 0$ , expressed in terms of the Green's function $G_\Omega(x, P)$ and the derivatives of $u$ at the center of the hole $P$ . Unlike previous studies that established estimates in $L^p$ spaces, this work provides uniform pointwise bounds.

Key analytical tools include:

Vanishing Order Decomposition: For multiple eigenvalues $\lambda$ with multiplicity $m(\lambda) \ge 2$ , the eigenspace $E(\lambda)$ is decomposed into orthogonal subspaces $E_J$ based on the order of vanishing $k_J$ of the eigenfunctions at $P$ .
Matrix Characterization: The asymptotic splitting of multiple eigenvalues is linked to the eigenvalues of specific matrices $G_{k_J}(P)$ constructed from the derivatives of the basis eigenfunctions at $P$ and geometric constants involving the Gamma function.
Spectral Theorems and Regularity: The convergence of eigenfunctions is analyzed using spectral projections and elliptic regularity estimates (Moser iteration, Schauder estimates) to control the difference $u_\varepsilon - u$ .

Key Contributions and Results

1. Pointwise Estimates for the Capacitary Potential
The paper establishes explicit asymptotic expansions for $V_{\varepsilon, u}(x)$ uniformly in $\Omega_\varepsilon$ . The leading order term is determined by the vanishing order $k$ of $u$ at $P$ :

For $N \ge 3$ : $V_{\varepsilon, u}(x) \sim C \varepsilon^{N-2+2k} \sum \partial^\alpha u(P) \partial^\alpha G_\Omega(x, P)$ .
For $N = 2$ : The behavior involves logarithmic terms if $k=0$ and polynomial terms if $k \ge 1$ .
These estimates refine previous $L^p$ results and are crucial for analyzing the local behavior of eigenfunctions near the hole.

2. Simplicity of Perturbed Eigenvalues
The authors provide explicit conditions under which the perturbed eigenvalues $\lambda_{j, \varepsilon}$ are simple.

Simple Eigenvalues: If $\lambda$ is simple, $\lambda_\varepsilon$ remains simple and converges to $\lambda$ with a rate determined by $u(P)$ (if $k=0$ ) or the vanishing order $k$ .
Multiple Eigenvalues: If $\lambda$ $λ$ has multiplicity $m(\lambda) \ge 2$ $m (λ) \geq 2$ , the eigenvalues bifurcate. The paper proves that the splitting rates and the simplicity of the resulting branches are governed by the eigenvalues of the matrix $G_{k_J}(P)$ $G_{k_{J}} (P)$ .
- Specifically, if all eigenvalues of $G_{k_J}(P)$ are simple, then the corresponding perturbed eigenvalues are simple for sufficiently small $\varepsilon$ .
- The paper provides a counterexample (the second eigenvalue of the unit disk with the hole at the origin) showing that if the matrix eigenvalues are not simple, the perturbed eigenvalues may retain multiplicity, even under small displacements of the hole center in symmetric configurations.

3. Quantitative Estimates for Eigenfunctions
The paper derives sharp estimates for the difference between the perturbed eigenfunction $u_\varepsilon$ and the unperturbed one $u$ .

Global Estimate: $u_\varepsilon(x) \approx u(x) - V_{\varepsilon, u}(x)$ with explicit error bounds depending on $N$ and the vanishing order $k$ .
Boundary Behavior: Near the hole boundary $\partial B(P, \varepsilon)$ , the paper provides precise asymptotic formulas. For instance, if $k=0$ and $N=2$ , $u_\varepsilon(x) \approx (1 + \frac{\log|x-P|}{\log \varepsilon})u(x)$ . These estimates are shown to be unattainable via standard regularity theory alone.

4. Nodal Sets and Domains
The study applies the eigenfunction estimates to the geometry of nodal sets $N_\varepsilon = \{x \in \Omega_\varepsilon : u_\varepsilon(x) = 0\}$ .

Case $P \notin N$ : If the hole center is not on the nodal set of $u$ , the nodal set $N_\varepsilon$ does not intersect the hole boundary for small $\varepsilon$ . The number of nodal regions of $u_\varepsilon$ is less than or equal to that of $u$ . If $u$ has two nodal regions (e.g., the second eigenfunction), $u_\varepsilon$ retains exactly two nodal regions.
Case $P \in N$ : If the hole center lies on the nodal set, the behavior depends on the vanishing order $k$ . In dimension $N=2$ , if $u$ vanishes at $P$ with order $k$ , the nodal set $N_\varepsilon$ intersects the boundary of the hole $\partial B(P, \varepsilon)$ at exactly $2k$ points. This extends the local structure of the nodal lines from the unperturbed domain to the perturbed one.

Significance and Claims
The paper claims to offer a comprehensive framework for understanding singular perturbations of the Dirichlet Laplacian. Its significance lies in:

Refining Asymptotics: Moving from integral ( $L^p$ ) estimates to pointwise estimates allows for a detailed description of the eigenfunction behavior near the singularity.
Explicit Simplicity Criteria: By linking the simplicity of perturbed eigenvalues to the spectral properties of the matrix $G_{k_J}(P)$ , the authors provide computable geometric conditions for eigenvalue splitting, addressing the complexity of multiple eigenvalues which was previously less understood.
Nodal Set Stability: The results generalize the "Payne property" (regarding the convexity and intersection of nodal lines with the boundary) to domains with holes, specifically characterizing how nodal lines interact with the hole boundary based on the vanishing order of the eigenfunction.

The authors note that while the results are sharp, the generic simplicity of eigenvalues under small displacements of the hole center remains an open problem, particularly in symmetric domains where specific configurations (like the origin in a disk) can preserve multiplicity.

More like this