🔢 mathematics

Neumann's nodal line may be closed on doubly-connected planar domains

This paper proves the existence of planar domains with a single hole where the first non-trivial Neumann eigenfunction possesses a closed nodal line, thereby demonstrating that such a configuration is possible on doubly-connected domains despite being impossible on simply-connected ones.

Original authors: Pedro Freitas, Roméo Leylekian

Published 2026-04-06

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Pedro Freitas, Roméo Leylekian

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 you have a drum. When you hit it, it vibrates in specific patterns. Some parts of the drum skin move up, some move down, and there are invisible lines where the skin doesn't move at all. These stationary lines are called nodal lines.

For a long time, mathematicians believed that if your drum was shaped like a simple circle or a square (a shape with no holes), these stationary lines could never form a complete, closed loop floating in the middle of the drum. They thought the lines had to touch the edge of the drum to split it in two.

This paper, by Pedro Freitas and Roméo Leylekian, proves that this belief was wrong, but only if you change the shape of the drum. They show that if your drum has a hole in the middle (like a donut or a washer), it is possible to have a stationary line that forms a perfect, closed circle floating entirely inside the dough, never touching the outer edge or the inner hole.

Here is a breakdown of their discovery using simple analogies:

1. The "Donut" Problem

Think of a standard drum (a simple circle). If you vibrate it, the "still" line usually cuts straight across, like a diameter, touching the rim on both sides.

The Old Rule: A famous mathematician named Pleijel (in 1956) proved that for a simple drum, the still line must touch the edge. It can't just float in the middle.
The Confusion: Because this rule was so famous, people assumed it applied to all drums, even the ones with holes. They thought, "No matter what shape you make, the still line must touch the edge."
The Discovery: The authors say, "Wait a minute! If the drum has a hole (is 'doubly-connected'), the old rules don't apply." They proved you can make a "donut drum" where the still line forms a closed ring right in the middle of the dough.

2. The "Sliding Handle" Trick

How did they prove this? They didn't just guess; they built a mathematical model using a clever "sliding" technique.

Imagine a long, thin hallway (the "static" part) with a circular room attached to it (the "moving" part).

The Setup: They imagined a family of shapes where a circular "handle" (the hole) slides back and forth along a long bridge.
The Movement:
- When the handle is on the far left, the vibration pattern (the nodal line) touches the left wall.
- When the handle is on the far right, the vibration pattern touches the right wall.
The Magic Moment: Because the shape changes smoothly as the handle slides, the vibration pattern must also move smoothly. At some point in the middle, the line has to let go of the left wall and hasn't quite reached the right wall yet. In that split second, the line is floating in the middle, forming a closed loop.

3. The "Graph" Analogy

To make the math work, the authors had to simplify the problem. They treated their complex shapes like stick figures or wireframes (called "metric graphs").

Imagine a caterpillar tree (a central spine with legs sticking out) with a loop at the top.
They showed that if the "legs" of the caterpillar are very long, the vibration pattern settles in the middle of the spine.
Then, they "thickened" these wireframes back into real 2D shapes (like inflating a wireframe into a rubber tube). They proved that the behavior of the real rubber tube is almost identical to the behavior of the wireframe.

4. Why This Matters

This isn't just about drums. It's about understanding how waves behave in different spaces.

The "Topological" Rule: The paper suggests a general rule for nature: If a space is simple (no holes), waves behave one way (lines must touch the edge). If a space has holes, waves can behave differently (lines can float).
The "Hole" Size: They also noted that the hole doesn't have to be huge; it just needs to exist. However, they couldn't prove how tiny the hole could be before the floating line disappears. That remains a mystery for future researchers.

Summary

Think of it like this:
If you are walking on a flat, open field (no holes), you can't draw a circle in the air that doesn't touch the ground; you have to anchor it. But if you are walking on a bridge over a river (a shape with a hole), you can draw a circle in the air that floats freely between the two banks.

Freitas and Leylekian found the mathematical blueprint for that "bridge," proving that nature allows for these floating, closed vibration lines, provided the shape has a hole. They used a sliding mechanism to show that the transition from "touching the edge" to "floating in the middle" is inevitable.

1. Problem Statement and Context

The Core Question:
The paper addresses a long-standing ambiguity in spectral geometry regarding the topology of nodal lines for the second Neumann eigenfunction on planar domains.

Pleijel's Result (1956): It is a known result that on a simply-connected planar domain, the nodal line of the second Neumann eigenfunction cannot be closed; it must intersect the boundary, splitting the domain into two simply-connected regions.
The Gap: The literature has been imprecise regarding doubly-connected (or multiply-connected) domains. Some sources implicitly assumed Pleijel's result held regardless of topology, or that closed nodal lines were impossible for Neumann problems.
The Goal: The authors aim to prove the existence of a bounded, doubly-connected planar domain where the second Neumann eigenfunction possesses a closed nodal line that is fully contained within the domain (i.e., it does not touch the boundary).

Comparison with Dirichlet Case:
The authors note that for the Dirichlet problem, the existence of closed nodal lines on doubly-connected domains was recently proven (Freitas & Leylekian, 2025). However, the Neumann problem is significantly more complex due to different concentration properties of eigenfunctions.

2. Methodology

The proof strategy relies on a "sliding procedure" adapted from their previous work on the Dirichlet case, combined with the analysis of metric graphs and graph-like domains.

A. The Sliding Handle Concept (Neumann FDSH)

The authors introduce the concept of a Neumann Family of Domains with a Sliding Handle (Neumann FDSH). This is a one-parameter family of domains $(\Omega_t)_{t \in T}$ satisfying specific conditions:

Symmetry with respect to the x-axis.
The second eigenvalue is simple.
The nodal line of the second eigenfunction changes sign along a specific segment on the x-axis.
Crucial Condition: As the parameter $t$ varies, the nodal line transitions from touching the "inner" boundary component to touching the "outer" boundary component.
Continuity: The eigenfunctions depend continuously on the parameter $t$ .

Theorem 2.2 (Main Property of FDSH): If such a family exists, there must be a specific parameter $t_0$ where the nodal line detaches from the boundary entirely, forming a closed Jordan curve inside the domain.

B. Construction via Metric Graphs

Constructing a Neumann FDSH is harder than for the Dirichlet case because Neumann eigenfunctions tend to oscillate along elongated parts of a domain (unlike Dirichlet eigenfunctions, which concentrate in "fat" regions).

Graph Model: The authors model the domain as a caterpillar tree with a loop (a central path with four long "rays" and a loop that can slide along the path).
Spectral Analysis: They analyze the Neumann-Kirchhoff Laplacian on this metric graph. They prove that if the "rays" are sufficiently long compared to the loop and the bridge, the nodal set of the second eigenfunction localizes near the midpoint of the bridge, regardless of the loop's position.
Embedding: To satisfy the symmetry requirement of the FDSH, they embed this graph into the plane by "rolling" the long rays over themselves to create a symmetric, planar, doubly-connected shape (resembling a caterpillar with a sliding loop).

C. Graph-Like Domains and Convergence

The planar graph is "thickened" to create a physical domain $\Theta_{h,t}$ with a small width parameter $h$ .

Convergence: As $h \to 0$ , the domain converges to the metric graph.
Strong Transversal Convergence (Key Technical Innovation): Standard $L^2$ convergence of eigenfunctions (known from Post, 2006) is insufficient to locate the nodal line precisely. The authors prove a strong transversal convergence result (Proposition A.4). This establishes that the eigenfunctions converge uniformly on transversal slices of the domain. This allows them to rigorously prove that the nodal line of the domain domain is close to the nodal set of the limit graph.

3. Key Contributions

Existence Theorem (Theorem 1.1): The paper proves the existence of a bounded, doubly-connected planar domain with a second Neumann eigenfunction having a closed nodal line that does not touch the boundary.
Generalization (Corollary 1.2): The result extends to domains with any number of holes ( $N \ge 2$ ).
Refinement of Convergence Theory: The authors derive a new strong transversal convergence result for eigenfunctions on graph-like domains. This improves upon existing $L^2$ results and is of independent interest for spectral geometry on thin domains.
Clarification of Topological Constraints: The paper resolves the ambiguity in the literature regarding Pleijel's 1956 argument, explicitly showing that the "no closed nodal line" rule applies only to simply-connected domains.

4. Main Results

Theorem 1.1: There exists a doubly-connected bounded open subset of $\mathbb{R}^2$ with a second Neumann eigenfunction whose nodal line is closed and disjoint from the boundary.
Proposition 3.5 (Localization): On a caterpillar tree with a loop and sufficiently long rays, the nodal set of the second eigenfunction intersects the central bridge arbitrarily close to its midpoint, independent of the loop's position.
Proposition A.4 (Strong Transversal Convergence): For a family of domains converging to a planar graph, the rescaled eigenfunctions converge in $C^{k+1, \alpha}$ on almost every transversal slice. This ensures the nodal line of the domain converges to the nodal set of the graph.

5. Significance and Implications

Resolution of a Conjecture: The paper confirms that the topological obstruction to closed nodal lines (simple connectivity) is the only barrier. Once a hole is introduced, the nodal line can close.
Parallel with Other Problems: The authors highlight a unifying "Topological Nodal Line Conjecture" across different boundary value problems (Dirichlet, Neumann, Robin, Steklov). In all cases, simply-connected domains forbid closed nodal lines for the second eigenfunction, while multiply-connected domains allow them.
Methodological Advance: The development of the "sliding handle" technique for Neumann problems, coupled with the new strong transversal convergence estimates, provides a robust framework for studying nodal sets on complex, thin domains.
Correction of Literature: The paper corrects a widespread misconception in spectral geometry (found in works by Payne, Bandle, etc.) that Neumann nodal lines can never be closed, attributing the error to an implicit assumption of simple connectivity in earlier proofs.

In summary, the paper combines spectral graph theory, asymptotic analysis of thin domains, and topological arguments to definitively show that the topology of the domain (specifically, the presence of a hole) is the determining factor in whether a Neumann nodal line can be closed.