I've just read a review paper about Yau's conjecture on nodal sets of the eigenfunctions for the Laplace operator on manifolds. Briefly, if $\phi_\lambda$, $\lambda$ are an eigenpair for the Laplace-Beltrami operator on a manifold $M$, i.e.

$$-\Delta\phi_\lambda = \lambda\phi,$$

then as $\lambda \to \infty$, there are constants $c$, $C$ such that

$$c\sqrt\lambda \le \text{area}(\{\phi_\lambda = 0\}) \le C\sqrt\lambda.$$

I've seen several results for manifolds without boundary of varying degrees of regularity. For example, Donnelly and Feffermann proved that Yau's conjecture is true for real analytic manifolds, but as I understand, the conjecture hasn't been proven yet for $C^\infty$ manifolds.

Are there extensions of the Yau conjecture to manifolds with boundary? Most of the work I've found (mainly other papers by Malinnikova and Logunov) considers only manifolds without boundary.

  • 1
    $\begingroup$ Is it fair to assume you are interested in Dirichlet/Neumann boundary conditions? $\endgroup$ Oct 22, 2019 at 22:13

1 Answer 1


Yau's conjecture was proved for the real analytic manifolds by Donelly and Fefferman, first for manifolds without boundary, then for manifolds with boundary in the paper Nodal Sets of Eigenfunctions: Riemannian Manifolds With Boundary, Analysis, et Cetera (1990), 251--262. The main result there is stated below.

Theorem 1.2 On any real analytic Riemannian manifold $M$ with boundary, the $n-1$ dimensional Hausdorff measure, $\mathcal{H}^{n-1}(N)$, of the nodal set $N$ of eigenfunction $f$, $\Delta f = -\lambda f$, (with Dirichlet or Neumann conditions on the boundary of $M$) satisfies $$c_1\sqrt{\lambda}\leq \mathcal{H}^{n-1}(N) \leq c_2 \sqrt{\lambda},$$ for some positive constants $c_1$, $c_2$.

There are other papers that handle more general boundary conditions, many of which reference the above paper.

  • $\begingroup$ Theorem 1.2 in that paper is only for real analytic manifolds. Theorem 1.1, which applies to $C^\infty$ manifolds, only says that eigenfunctions vanish to order at most $c\sqrt{\lambda}$. $\endgroup$
    – user7868
    Apr 16, 2020 at 10:21
  • $\begingroup$ @user7868 Thanks for the correction. $\endgroup$ Apr 17, 2020 at 10:25

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.