Steklov-Laplacian Eigenvalue Overview

• Steklov-Laplacian eigenvalue is defined via Dirichlet-to-Neumann boundary conditions, linking harmonic interior functions with the spectral properties of the boundary.
• Quantitative comparisons show that its values are closely tied to Laplace–Beltrami eigenvalues, with curvature and boundary effects providing explicit bounds.
• Its study drives advances in shape optimization and isoperimetric inequalities, and extends to nonlinear and discrete spectral frameworks.

The Steklov-Laplacian eigenvalue is a core spectral invariant associated with the Laplacian operator subject to Steklov (Dirichlet-to-Neumann) boundary conditions on a domain. Its spectral theory offers a deep connection between the geometry of domains and their boundaries, revealing intricate relationships to Laplace–Beltrami eigenvalues, isoperimetric problems, sharp geometric inequalities, and optimal shape flows in both smooth and discrete settings.

1. Core Definition and Variational Characterization

Let $(M^{n+1},g)$ be a complete Riemannian manifold, $\Omega\subset M$ a compact domain with smooth boundary $\Sigma=\partial\Omega$ and outward unit normal $\nu$. The classical Steklov-Laplacian eigenvalue problem is

$$\Delta u=0 \quad \text{in }\Omega,\qquad \partial_\nu u = \sigma u\quad \text{on } \Sigma$$

where $\Delta$ is the Laplace–Beltrami operator on $M$ and $\partial_\nu$ is the boundary normal derivative.

The sequence of Steklov eigenvalues $\{\sigma_j\}_{j=0}^\infty$ is characterized variationally: $$\sigma_j = \inf_{V\subset H^1(\Omega),\,\dim V=j+1} \,\sup_{u\in V,\,\int_{\Sigma}u^2=1} \int_{\Omega}|\nabla u|^2\,.$$ This makes the Steklov eigenvalues the spectrum of the Dirichlet-to-Neumann operator $\Omega\subset M$0, $\Omega\subset M$1, where $\Omega\subset M$2 is the harmonic extension of $\Omega\subset M$3 to $\Omega\subset M$4 (Xiong, 2017, Lamberti et al., 2014).

2. Comparison with Boundary Laplacian Eigenvalues

Fundamental spectral comparison theorems relate the Steklov-Laplacian eigenvalues $\Omega\subset M$5 to the Laplace–Beltrami spectrum $\Omega\subset M$6 of the boundary $\Omega\subset M$7: $\Omega\subset M$8

Under bounded sectional curvature $\Omega\subset M$9, principal curvatures $\Sigma=\partial\Omega$0 of $\Sigma=\partial\Omega$1 (maximal $\Sigma=\partial\Omega$2, minimal $\Sigma=\partial\Omega$3), Xiong proved: $\Sigma=\partial\Omega$4 with additional explicit upper and lower inequalities depending on the ambient curvature and $\Sigma=\partial\Omega$5 (Xiong, 2017). For nonnegatively curved manifolds, Raulot–Savo et al. established

$\Sigma=\partial\Omega$6

given nonnegative curvature operator and uniform $\Sigma=\partial\Omega$7-curvature lower bound $\Sigma=\partial\Omega$8 on $\Sigma=\partial\Omega$9, with sharper forms in low dimensions (Karpukhin, 2015).

3. Analytical Tools: Pohozaev-Type Identity and Tubular Collars

The Pohozaev-type identity for harmonic functions in $\nu$0 plays a central role: $\nu$1 Choosing $\nu$2 for suitable collar functions $\nu$3 (distance to $\nu$4, sphere-model analogues), and employing Hessian comparison theorems, one can relate boundary integrals $\nu$5 and $\nu$6, controlling the spectral discrepancy between $\nu$7 and $\nu$8 via principal and sectional curvature (Xiong, 2017).

4. Asymptotic Estimates, Weyl Law, and Sharp Bounds

Steklov-Laplacian eigenvalues obey sharp asymptotics: $\nu$9 matching the classical Weyl law for boundary Laplacian eigenvalues. Under assumptions on Ricci or sectional curvature, Weyl-type upper bounds arise: $$\Delta u=0 \quad \text{in }\Omega,\qquad \partial_\nu u = \sigma u\quad \text{on } \Sigma$$0 where $$\Delta u=0 \quad \text{in }\Omega,\qquad \partial_\nu u = \sigma u\quad \text{on } \Sigma$$1 is a maximal curvature quantity. These bounds confirm that the growth rates of Steklov spectra are tightly controlled by geometry (Xiong, 2017, Du et al., 2019).

5. Shape Optimization, Stability, and Isoperimetric Principles

Isoperimetric-type inequalities have been proved both for pure Steklov and mixed boundary problems. In planar domains of fixed perimeter,

$$\Delta u=0 \quad \text{in }\Omega,\qquad \partial_\nu u = \sigma u\quad \text{on } \Sigma$$2

universally, with equality approached as domains collapse or gain infinitely many boundary components (Girouard et al., 2020). For nearly spherical domains (in Hausdorff metric), balls are strict local maximizers of the principal Steklov-Laplacian eigenvalue (Ferone et al., 2013). For annular and perforated settings, monotonicity and stability results hold:

• The principal Steklov–Dirichlet eigenvalue decreases with displacement of an internal obstacle, reaching maximum when concentric (Gavitone et al., 2021, Paoli et al., 2020).
• For Robin-perturbed Steklov problems, a transition between pure Steklov, Dirichlet, and mixed spectra occurs as the Robin parameter varies (Gavitone et al., 2022).

6. Nonlinear and Discrete Steklov Frameworks

Generalizations to nonlinear (p-Laplacian, $$\Delta u=0 \quad \text{in }\Omega,\qquad \partial_\nu u = \sigma u\quad \text{on } \Sigma$$3-Laplacian) and discrete (graphs) settings are now well-developed. For $$\Delta u=0 \quad \text{in }\Omega,\qquad \partial_\nu u = \sigma u\quad \text{on } \Sigma$$4-Laplacian with Steklov-like boundary condition, the spectrum is: $$\Delta u=0 \quad \text{in }\Omega,\qquad \partial_\nu u = \sigma u\quad \text{on } \Sigma$$5 where $$\Delta u=0 \quad \text{in }\Omega,\qquad \partial_\nu u = \sigma u\quad \text{on } \Sigma$$6 arises from a nonlinear Rayleigh quotient on a constrained cone, and the operator's non-homogeneity (with $$\Delta u=0 \quad \text{in }\Omega,\qquad \partial_\nu u = \sigma u\quad \text{on } \Sigma$$7) makes the spectrum a continuous half-line, not a discrete sequence (Barbu et al., 2020). On graphs, Steklov eigenvalues always exceed corresponding Laplacian eigenvalues: $$\Delta u=0 \quad \text{in }\Omega,\qquad \partial_\nu u = \sigma u\quad \text{on } \Sigma$$8 with rigidity characterized in terms of separable boundary-to-interior weights and combinatorial structure (Shi et al., 2020).

7. Spectral Geometry, Boundary Invariants, and Heat Kernel Analysis

The Steklov-Laplacian spectrum encodes geometric invariants of the domain’s boundary. Heat-trace asymptotics of the Dirichlet-to-Neumann operator,

$$\Delta u=0 \quad \text{in }\Omega,\qquad \partial_\nu u = \sigma u\quad \text{on } \Sigma$$9

have leading coefficients expressing boundary volume, mean curvature, scalar curvature, and curvature tensor contractions, recoverable from the Steklov spectrum (Liu, 2013). This link is sharpened by decomposing the Dirichlet-to-Neumann operator as $\Delta$0, with symbolic calculus yielding curvature explicit formulas.

8. Context, Extensions, and Open Questions

The Steklov-Laplacian eigenvalue formalism bridges Dirichlet, Neumann, Robin, and higher-order elliptic spectral theories. It is critical for isoperimetric inequalities, shape optimization, and rigidity questions in geometry, as well as for controlling boundary-to-interior transfer in both smooth and combinatorial settings. Open problems include full rigidity characterizations, sharp bounds under minimal curvature hypotheses, and generalization to vector-valued forms and singular domains (Xiong, 2023, Karpukhin, 2015). The intrinsic-extrinsic interplay, encapsulated by comparison of Steklov and Laplacian eigenvalues and their growth rates, continues to drive advances in geometric analysis and spectral theory.