← All problems

The reverse Faber–Krahn conjecture for the magnetic Neumann Laplacian

2 years old · traced to

Let β>0\beta>0, and let Ω⊂R2\Omega\subset\mathbb R^2 be an open, bounded and simply connected set with a C∞C^\infty smooth boundary. Let Ω∗\Omega^* be the disk in R2\mathbb R^2 centered at the origin such that ∣Ω∗∣=∣Ω∣|\Omega^*|=|\Omega|. Reverse Faber–Krahn conjecture. For all β>0\beta>0,

λ1N(Ω,β)≤λ1N(Ω∗,β),\lambda_1^N(\Omega,\beta)\le\lambda_1^N(\Omega^*,\beta),

with equality if and only if Ω\Omega is a disk. This conjecture is suggested by results of Fournais and Helffer showing the reverse inequality for sufficiently small and sufficiently large magnetic fields when Ω\Omega is not a disk; the assertion for every positive β\beta remains open.

References
  1. Bruno Colbois, Corentin Léna, Luigi Provenzano and Alessandro Savo, “A reverse Faber-Krahn inequality for the magnetic Laplacian”, arXiv:2403.11336 (2024).

Progress summary

Refreshed
Open · progress reported

A 2024 result proves the conjecture for weak magnetic fields, but its claim for every positive field remains open.

The conjecture asks whether, among smooth bounded simply connected planar domains of fixed area, the disk always has the largest first magnetic Neumann eigenvalue, with equality only for the disk. Fournais and Helffer formulated it in 2019 after proving the claim in both sufficiently weak and sufficiently strong fields.

Known results

  • Fournais–Helffer (2019): for every non-disk domain, the disk has the larger eigenvalue for sufficiently small and sufficiently large β\beta.
  • Kachmar–Lotoreichik: related torsion-function estimates establish restricted cases but do not imply the full conjecture.
  • Colbois and coauthors (2024): proved the inequality, with equality only for disks, when β∣Ω∣≤β∗π\beta|\Omega|\leq\beta^*\pi, where β∗≈3.84754\beta^*\approx3.84754.

2024 partial confirmation

The paper A reverse Faber–Krahn inequality for the magnetic Laplacian, published in Journal de Mathématiques Pures et Appliquées 192192 (2024), uses torsion-function level lines to prove the area-dependent regime above. It does not settle intermediate or arbitrary fields, and no counterexample within the simply connected class was found.

Current status (as of October 2026): the conjecture is proved for sufficiently small and sufficiently large β\beta, and for β∣Ω∣≤β∗π\beta|\Omega|\leq\beta^*\pi; the assertion for all β>0\beta>0 remains open.

Sources

Solutions 0

No solutions have been posted yet.