Magnetic Steklov Eigenvalues

• Magnetic Steklov eigenvalues are defined by imposing magnetic boundary conditions on classical Steklov problems, encoding key features like flux, holonomy, and gauge invariance.
• They utilize analytical pseudodifferential theory and variational principles to derive Weyl-type asymptotic expansions and address both direct and inverse boundary value problems.
• Applications range from planar domains to electromagnetic systems, facilitating isoperimetric inequalities and elucidating spectral localization under strong magnetic field regimes.

Magnetic Steklov eigenvalues generalize the classical Steklov spectral problem by imposing boundary conditions and operators that reflect the influence of a magnetic field. The magnetic Steklov spectrum encodes complex geometric and physical information regarding gauge fields, holonomy, and flux, involving both analytical pseudodifferential theory and concrete variational principles. In various settings—compact Riemannian manifolds, planar domains with Aharonov-Bohm flux, and electromagnetic systems governed by Maxwell’s equations—this spectrum governs both direct and inverse boundary value problems, isoperimetric inequalities, and the analytic structure of spectral asymptotics.

1. Fundamental Framework: Magnetic Steklov Problems

The generic magnetic Steklov problem is defined on a compact Riemannian manifold $(M,g)$ with boundary $\partial M$, with a magnetic potential $A$ (imaginary 1-form or real 1-form, depending on normalization) and a real or complex potential $q$. The canonical boundary value problem is

$$L_{g,A,q}u = 0 \ \text{in } M,\qquad u|_{\partial M} = f,$$

where the "magnetic Schrödinger operator" is

$L_{g,A,q}\;=\;d_A^* d_A + q,\qquad d_A = d + A,$

and the Dirichlet-to-Neumann (DtN) map $\Lambda_{g,A,q}$ is defined as

$$\Lambda_{g,A,q}: H^{1/2}(\partial M) \to H^{-1/2}(\partial M), \qquad \Lambda_{g,A,q} f = (d_A u)|_{\partial M}(\nu),$$

with $\nu$ the outward normal. Its spectrum,

$$\operatorname{Spec}(\Lambda_{g,A,q}) = \{ \sigma_k \}_{k=1}^\infty, \quad 0 < \sigma_1 \leq \sigma_2 \leq \dots \to +\infty,$$

defines the magnetic Steklov eigenvalues. In the classical case $\partial M$0, this reduces to the spectrum of the non-magnetic Dirichlet-to-Neumann operator (Cekić et al., 2024, Chakradhar et al., 2024).

On $\partial M$1-forms, the problem generalizes via the magnetic exterior differential $\partial M$2 and its adjoint $\partial M$3, forming the magnetic Hodge Laplacian $\partial M$4. For given boundary data $\partial M$5, the unique $\partial M$6-harmonic extension $\partial M$7 solves

$\partial M$8

and the magnetic Steklov operator is $\partial M$9 (Chakradhar et al., 10 Nov 2025).

In planar domains and electromagnetic cavities, variants appear with Aharonov-Bohm and uniform fields, employing variational characterizations over $A$0 or electromagnetic spaces $A$1, and using the gauge-covariant differential $A$2 (Kachmar et al., 19 Feb 2026, Colbois et al., 2022, Ferraresso et al., 2022).

2. Asymptotic Expansion and Inverse Spectral Issues

The spectrum of $A$3, as a self-adjoint elliptic pseudodifferential operator of order 1, admits a Weyl-type expansion: $A$4 which is further refined to a full asymptotic expansion in inverse powers of $A$5 or $A$6. On each boundary component $A$7 of length $A$8, after fixing a gauge, the spectral branches are

$A$9

with

$q$0

where $q$1 is an integer, $q$2 is magnetic holonomy, $q$3 encodes normal curvature of $q$4, and $q$5 is the integral of $q$6 (Cekić et al., 2024).

Inverse spectral problems have a nuanced resolution:

• For $q$7 (one boundary component), the full Steklov spectrum determines the boundary length, holonomy exponentials $q$8, $q$9, and $$L_{g,A,q}u = 0 \ \text{in } M,\qquad u|_{\partial M} = f,$$0.
• If boundary lengths are distinct and holonomies are generic, the spectrum determines $$L_{g,A,q}u = 0 \ \text{in } M,\qquad u|_{\partial M} = f,$$1, each $$L_{g,A,q}u = 0 \ \text{in } M,\qquad u|_{\partial M} = f,$$2, and each $$L_{g,A,q}u = 0 \ \text{in } M,\qquad u|_{\partial M} = f,$$3.
• Without these genericity hypotheses, there exist counterexamples with the same spectrum but different numbers of boundary components—demonstrated by explicit models with the magnetic field turned off near the boundary.

This spectral behavior contrasts with the non-magnetic case and is pivotal for the study of quantum graphs, inverse boundary problems, and distinguishing manifolds up to boundary data (Cekić et al., 2024).

3. Geometry, Isoperimetry, and Model Domains

In the planar case and in the presence of Aharonov-Bohm flux, the magnetic Steklov problem is sensitive to flux quantization:

• For non-integer flux in the disk centered at the pole, the first eigenvalue is $$L_{g,A,q}u = 0 \ \text{in } M,\qquad u|_{\partial M} = f,$$4, leading to positive spectral gap.
• Classical isoperimetric inequalities generalize: among simply connected domains of fixed area or perimeter, the disk maximizes the lowest magnetic Steklov eigenvalue, provided the field is not too strong and geometric constraints are satisfied (Kachmar et al., 19 Feb 2026, Colbois et al., 2022).

Isoperimetric bounds, both of Brock- and Weinstock-type, as well as explicit disk spectra, hold in Euclidean, spherical, and hyperbolic geometries, underlining the universality of the disk as extremal profile for the lowest magnetic Steklov eigenvalue across settings (Provenzano et al., 2023, Colbois et al., 2022, Kachmar et al., 19 Feb 2026).

The geometry of maximizing surfaces for higher eigenvalues, particularly annuli with harmonic magnetic flux, leads to classifications via critical catenoids or more general $$L_{g,A,q}u = 0 \ \text{in } M,\qquad u|_{\partial M} = f,$$5-surfaces (linear Weingarten surfaces), with variational characterizations reflecting sharp bounds for the first or second eigenvalue (Provenzano et al., 2023).

4. Spectral Theory: Electromagnetic and PDE Contexts

Magnetic Steklov operators in electromagnetic theory arise from boundary conditions on Maxwell's equations, notably:

• For time-harmonic Maxwell (with zero sources), one studies

$$L_{g,A,q}u = 0 \ \text{in } M,\qquad u|_{\partial M} = f,$$6

in spaces $$L_{g,A,q}u = 0 \ \text{in } M,\qquad u|_{\partial M} = f,$$7 of divergence-free, curl-regular fields with tangential (electric or magnetic) boundary data (Lamberti et al., 2020, Ferraresso et al., 2022).

• The corresponding boundary operator (interior Calderón operator) is compact and self-adjoint; its spectrum consists of real, positive, diverging eigenvalues.
• The spectral theorem yields orthonormal Fourier bases of boundary fields and permits the expansion of solutions in Steklov modes.

In full time-harmonic settings, the spectrum consists of three types: essential spectrum (zero), a sequence of positive eigenvalues diverging to $$L_{g,A,q}u = 0 \ \text{in } M,\qquad u|_{\partial M} = f,$$8, and negative eigenvalues accumulating at zero (Halla, 2019).

The min–max (Courant) principle applies,

$$L_{g,A,q}u = 0 \ \text{in } M,\qquad u|_{\partial M} = f,$$9

where $L_{g,A,q}\;=\;d_A^* d_A + q,\qquad d_A = d + A,$0 is the electromagnetic sesquilinear form, $L_{g,A,q}\;=\;d_A^* d_A + q,\qquad d_A = d + A,$1 is the boundary $L_{g,A,q}\;=\;d_A^* d_A + q,\qquad d_A = d + A,$2-inner product (Ferraresso et al., 2022, Lamberti et al., 2020).

On model domains like the Euclidean ball, explicit computation is possible via separation of variables and vector spherical harmonics, revealing that the spectrum decomposes into families associated with divergence-free and other modes, each with explicit formulas for eigenvalues in terms of special functions (Ferraresso et al., 2022).

5. Variational Principles, Gauge Invariance, and Bounds

The first magnetic Steklov eigenvalue, in all common settings, admits a Rayleigh-type variational characterization: $L_{g,A,q}\;=\;d_A^* d_A + q,\qquad d_A = d + A,$3 for functions, or analogous forms for $L_{g,A,q}\;=\;d_A^* d_A + q,\qquad d_A = d + A,$4-forms or electromagnetic fields (Cekić et al., 2024, Chakradhar et al., 2024, Chakradhar et al., 10 Nov 2025).

Gauge equivalence plays a central role. If the magnetic potential is gauge-exact, the spectrum coincides with the non-magnetic case due to the unitary equivalence induced by multiplication by $L_{g,A,q}\;=\;d_A^* d_A + q,\qquad d_A = d + A,$5-valued functions (Chakradhar et al., 2024). The spectrum depends only on the holonomy or flux through non-contractible cycles.

On general Riemannian manifolds, analytic lower bounds are given in terms of Cheeger-type or magnetic Cheeger–Jammes constants $L_{g,A,q}\;=\;d_A^* d_A + q,\qquad d_A = d + A,$6, leading to

$L_{g,A,q}\;=\;d_A^* d_A + q,\qquad d_A = d + A,$7

and associated upper bounds depend on geometric and magnetic data (Chakradhar et al., 2024).

A significant phenomenon is that the diamagnetic inequality, $L_{g,A,q}\;=\;d_A^* d_A + q,\qquad d_A = d + A,$8 for $L_{g,A,q}\;=\;d_A^* d_A + q,\qquad d_A = d + A,$9, fails in general for magnetic Steklov problems, in contrast to Dirichlet or Neumann Laplacians. Magnetic perturbations can strictly lower Steklov eigenvalues, as demonstrated by explicit examples and Taylor expansions for small $\Lambda_{g,A,q}$0 (Chakradhar et al., 10 Nov 2025).

6. High-Field Asymptotics and Exponential Localization

In the strong magnetic field regime, spectral properties are dominated by boundary-layer phenomena. The lowest eigenfunctions become exponentially localized near regions where the magnetic field vanishes to maximal order, with explicit exponential decay rates in terms of the Agmon distance and the field strength parameter $\Lambda_{g,A,q}$1: $\Lambda_{g,A,q}$2 where $\Lambda_{g,A,q}$3 is the maximal order of vanishing of the magnetic field on the boundary subset $\Lambda_{g,A,q}$4 (Shen, 18 Nov 2025). The corresponding first eigenvalue scales as $\Lambda_{g,A,q}$5 as $\Lambda_{g,A,q}$6.

In exterior geometric models (e.g., exterior of a disk), three-term asymptotic expansions for ground state eigenvalues explicitly capture the flux dependence and manifest Aharonov–Bohm effects (Helffer et al., 25 Aug 2025).

7. Electromagnetic Steklov Eigenvalues: Numerical and Computational Aspects

For implementation and numerical approximation in electromagnetic settings, finite element discretizations respecting the $\Lambda_{g,A,q}$7 structure are developed using Nédélec edge elements, Raviart–Thomas elements, and stable discretizations of boundary operators. Convergence and error estimates are established via discrete compactness properties and the Babuška–Osborn theory (Gong et al., 2020, Halla, 2019). The structure of the spectrum depends delicately on the choice of original versus modified (divergence-free) Steklov boundary conditions, with the modified problem enjoying better spectral compactness properties.

In all settings, spectral representations are invaluable for boundary integral methods, expansion and control of fields in cavities, inverse problems, and optimization of spectral quantities (Ferraresso et al., 2022, Lamberti et al., 2020). The magnetic Steklov spectral data thus provide a rich source of geometric and analytic invariants for both theoretical exploration and applied computation.