← All problems

Bourgain–Rudnick, Toth–Zelditch, and Han nodal-curve questions

Let (M,g)(M,g) be a compact real-analytic Riemannian surface, and let γ⊂M\gamma\subset M be an embedded curve. If there exist infinitely many pairwise distinct Laplace eigenfunctions φj\varphi_j satisfying Δgφj=λj2φj\Delta_g\varphi_j=\lambda_j^2\varphi_j and φj∣γ≡0\varphi_j|_{\gamma}\equiv 0, then γ\gamma is a geodesic of gg.

References

Progress summary

Refreshed
Open · solution reported

An unrefereed October 2026 preprint claims to disprove the geometric conjecture by constructing a nongeodesic curve contained in infinitely many nodal sets.

The Bourgain–Rudnick question asks which submanifolds can support eigenfunctions, or sequences of eigenfunctions, vanishing identically. The related Toth–Zelditch and Han results give substantial restrictions in dynamical and flat-torus settings.

Known results

  • Toth–Zelditch (2017): under geodesic asymmetry and full-measure flowout, positive-density eigenfunction sequences cannot vanish on the hypersurface.
  • Toth–Zelditch (2018): nodal hypersurfaces are extreme bad pairs; the characterization problem remained only partially answered.
  • Han (2011): for curved real-analytic Σ⊂Td\Sigma\subset\mathbb{T}^{d}, restriction lower bounds hold; in dimension 22, nongeodesic curves have ≫λ1−ε\gg\lambda^{1-\varepsilon} nodal intersections.

October 2026 claimed counterexample

Xianchao Wu and Xiao Xiao claim explicit metrics and eigenfunctions for which a nongeodesic curve lies in infinitely many nodal sets, together with restriction estimates for integrable systems. If correct, this is a negative answer to the stated geometric question; the preprint is unrefereed and the claim is unverified.

Current status (as of October 2026): The general question has a claimed negative resolution via Wu and Xiao's explicit examples, but the result is unverified; the earlier partial restriction and nodal-intersection theorems remain established.

Sources

Solutions 0

No solutions have been posted yet.