Payne’s nodal line conjecture fails on doubly-connected planar domains
Pedro Freitas, Roméo Leylekian · 2025
Read Payne’s nodal line conjecture fails on doubly-connected planar domains by Pedro Freitas, Roméo Leylekian (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2510-24436.txt.
Introduction
Pedro Freitas and Roméo Leylekian, two authors, test a famous math claim. Payne's nodal line conjecture concerns how a drum-like shape vibrates. For the second tone of a fixed-edge drum, a still line forms inside. Payne claimed that line must always reach the edge of the shape. The pair ask if the claim holds for shapes with one hole.
They build flat shapes with a single hole where the claim fails. On those shapes the still line closes into a loop inside. The loop never touches the edge of the shape at all. So the claim can hold, at most, for shapes with no holes. The counterexamples are plain and bounded, not strange edge cases. The work trims a long-standing guess to its true scope.
Catalog record
| CREATOR | Pedro Freitas, Roméo Leylekian |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 48 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2510-24436
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Payne’s nodal line conjecture fails on doubly-connected planar domains Pedro Freitas, Roméo Leylekian Abstract . We present examples of bounded planar domains with one single hole for which the nodal line of a second Dirichlet eigenfunction is closed and does not touch the boundary. This shows that Payne’s nodal line conjecture can at most hold for simply-connected domains in the plane. 1. Introduction Consider the eigenvalue problem for the Dirichlet Laplacian on a bounded connected domain \Omega in \mathbb{R}^{n} , namely, (1) \left\{\begin{array}[]{ll}\Delta u+\lambda u=0,&x\in\Omega\\ u=0,&x\in\partial\Omega,\end{array}\right. with eigenvalues \lambda_{1}<\lambda_{2}\leq\lambda_{3}\leq\dots and corresponding eigenfunctions u_{1},u_{2},u_{3}\dots . While the first eigenfunction u_{1} may be chosen to be of one sign in \Omega , due to orthogonality conditions and Courant’s nodal domain theorem, any second eigenfunction u_{2} will have two and only two nodal domains, defined as maximal connected open sets in \Omega where an eigenfunction does not change sign. More precisely, the nodal set of a function u:\Omega\to\mathbb{R} is defined by \mathcal{N}(u)=\overline{\left\{x\in\Omega:u(x)=0\right\}}, and the nodal domains of u are the connected components into which the domain \Omega is divided by \mathcal{N}(u) . The structure of the nodal set and nodal domains of the Laplacian has been the object of much study, with the case associated to the second eigenfunction being of particular interest. This had at its origin Payne’s nodal line conjecture concerning the topology of \mathcal{N}(u_{2}) , and was first formulated for planar domains in [ Pay67 ] as Conjecture (Nodal line conjecture 1967) . The eigenfunction u_{2} of Problem ( 1 ) cannot have a closed nodal line for any domain \Omega\subset\mathbb{R}^{2} . An n- dimensional version was proposed by Yau in 1993 [ Yau93 , Problem 45] , who also questioned how Euclidean results would extend to compact manifolds with boundary. The original conjecture has been shown to hold under certain restrictions, namely by Payne when \Omega is convex along one direction and invariant by reflection with respect to an axis orthogonal to that direction [ Pay73 ] , and by Melas for sufficiently smooth convex planar domains [ Mel92 ] . The smoothness assumption in Melas’s paper was then removed in [ Ale94 ] . Damascelli generalised Payne’s result to higher dimensions [ Dam00 ] , and Jerison proved the conjecture in the case of long and thin convex domains [ Jer95 ] , while Krejčiřík and the first author proved it for long thin tubes of arbitrary cross-section around a bounded line in Euclidean space [ FK08 ] . In these last two cases, the problem of locating the nodal set was also considered. More recently, Kiwan addressed the conjecture for a class of doubly-connected domains [ Kiw18 ] , while Mukherjee and Saha proved it for dumbbell-like domains in any dimension [ MS25 , Theorem 1.8] . Let us also mention that, in the preprint [ Ha24 ] , the author claims that the conjecture holds for all regular simply connected domains. However, the proof relies on a result in [ Liq95 ] according to which the multiplicity of the second eigenvalue on planar domains is at most two. As recalled in [ HHN98 , Hel+21 ] this need not be the case in general. On the negative side, several counterexamples have been provided, either for the original problem or for related problems under different settings. In the former category, the first known counterexample was given in [ HHN97 ] (see also [ HHN98 ] ) and consists of a disk from which a (large) number of slits located on a smaller concentric disk have been removed. Later Fournais generalised this to higher dimensions giving a domain where the nodal set does not touch the boundary, while providing a first estimate on the order of the number of holes, namely, 10^{9} , which was considered to be probably by far too large [ Fou01 ] . In fact, Kennedy showed that in dimensions three or higher it was possible to obtain a counterexample homeomorphic to a ball [ Ken13 ] (see also [ MS25 , Theorem 1.9] ). As for what may happen under different restrictions, a first question that may be asked is what happens if one drops the assumption that the domain is bounded. In this case, a counterexample relevant to the present paper – see the discussion below – was given by Krejčiřík and the first author in [ FK07 ] , where it was shown that there exist unbounded domains for which the nodal line does not touch the boundary. These are however not closed and in fact consist of an infinite straight line. On the other hand, these domains satisfy Payne’s conditions mentioned above for which he proved the conjecture for bounded domains [ Pay73 ] . In fact, the counterexample given in [ FK07 ] may even be chosen to be symmetric and convex with respect to both directions. Moving now away from the Euclidean Laplacian, Lin and Ni showed that in the presence of a potential the conjecture can fail even for a ball and when the potential is radial, meaning that the nodal set can be an (n-1)- sphere [ LN88 ] . Answering the question raised by Yau mentioned above, the first author of the present paper showed that the Laplacian could also have a nodal set consisting of a geodesic disc on surfaces of revolution [ Fre02 ] . These two examples showed that outside the Laplacian in the Euclidean setting, simply-connectedness and convexity of the domain do not necessarilly imply the conjecture. We note that what may happen for the Laplacian in general convex domains in \mathbb{R}^{n} with n larger than two remains mostly unknown. All of this leaves open the question of what is the minimal number of holes which a planar counterexample can have. This had already been formulated in [ HHN97 , Remark 3] , where the authors stated that while they did not believe that it would be possible to find a planar simply-connected domain for which the conjecture would fail, a more general question would be: What is the smallest N_{0} such that there exists a domain with N_{0} boundary components whose second eigenfunction has a nodal line that does not hit the boundary? In a recent paper, Dahne, Gómez-Serrano, and Hou, have shown via a computer-assisted proof that it is possible to reduce the number of holes to six (and hence N_{0} above to seven) [ DGH21 ] . The main purpose of the present paper is to show through purely analytic arguments that N_{0} may be brought down to two, that is, there exists a planar domain with a single hole whose nodal line is closed and does not touch the boundary. In other words, Payne’s conjecture in the plane can be expected to hold at most for simply-connected domains. Theorem 1.1 . There exists a doubly-connected bounded open subset of \mathbb{R}^{2} with a second eigenfunction whose nodal line is closed and does not touch the boundary. Of course, as a consequence of Theorem 1.1 there are domains with more holes for which the nodal line remains closed. This can be seen by introducing sufficiently small holes inside the nodal domains of the eigenfunction of the theorem (see also