Special Transmission Eigenvalues

• Special transmission eigenvalues are distinct spectral parameters where perturbed and unperturbed scattering systems become indistinguishable, often marked by boundary-induced and infinite multiplicity phenomena.
• Analytic and computational techniques, including recursive integral methods and contour-based algorithms, enable robust resolution of these eigenvalues across one- and multi-dimensional models.
• Their occurrence critically affects the uniqueness and stability of inverse problems, underpinning applications in cloaking, resonance theory, and material property reconstruction.

A special transmission eigenvalue is, in the context of inverse scattering theory and spectral analysis, a spectral parameter (typically an energy, frequency, or related quantity) at which the transmission problem exhibits nontrivial solutions corresponding to the indistinguishability between perturbed and unperturbed scattering systems—formally, where certain scattered fields are "invisible" or coincide. These eigenvalues are particularly notable when they possess exceptional algebraic or geometric properties, such as occurrence at special parameter values, enhanced multiplicity, strong geometric invariance, or infinite multiplicity due to the model's structure.

1. Rigorous Definition and Occurrence of Special Transmission Eigenvalues

A transmission eigenvalue $\lambda$ is defined for a pair of wave equations—typically a perturbed and an unperturbed system—subject to matching conditions. For example, in the half-line Schrödinger operator scenario,

$$-\psi''(x) + V(x)\psi(x) = k^2\psi(x),\quad -\psi_0''(x) = k^2\psi_0(x),$$

with both $\psi$ and $\psi_0$ obeying the same selfadjoint boundary condition at $x=0$ and matching at $x=b$ ($\psi(b) = \psi_0(b)$, $\psi'(b) = \psi_0'(b)$), a transmission eigenvalue corresponds to any $k^2$ for which a nontrivial solution exists. Special transmission eigenvalues arise in settings where these occur at remarkable points, such as zero energy ($\lambda=0$), eigenvalues induced solely by boundary conditions, or points where entire families coalesce due to structural degeneracy (Aktosun et al., 2014).

In higher-dimensional models, related but more intricate definitions hold, commonly involving the coincidence of boundary data for distinct PDEs, with the set of special transmission eigenvalues including those with infinite multiplicity or clustering behavior, as seen for multipoint scatterers (Grinevich et al., 2021).

2. Analytic and Algebraic Structure

The functions whose zeros define transmission eigenvalues (e.g., $D(k)$ in 1D, or characteristic operator pencils in multi-D) are typically entire functions or Fredholm-analytic operator-valued maps:

• In the half-line Schrödinger case, $D(k)$ is entire, even in $k$, with Hadamard factorization encoding the (possibly multiple) transmission eigenvalues:

$$D(k) = \gamma k^{2d} \prod_{j=1}^\infty \left(1 - \frac{k^2}{\kappa_j^2}\right)$$

The multiplicity of zeros at $k=0$ corresponds to special cases, including $\lambda=0$ as an eigenvalue of multiplicity one or two, but never higher (Aktosun et al., 2014).

• For discrete analogs, $D(\lambda)$ becomes a polynomial (for supported potentials), and multiplicity precisely reflects the model's spatial truncation (Aktosun et al., 2015).
• In higher dimensions or for operator pencils, the transmission eigenvalues remain discrete under Fredholm analytic theory (e.g., $A(k) = I + k^2(\Delta + k^2)^{-1} M$ in Born-approximation models (Hovsepyan, 2023)).

Special or exceptional transmission eigenvalues are typically linked to:

• Vanishing of the Jost function at $k=0$ or $k=i\cot\theta$ (exceptional for boundary-induced phenomena)
• Degeneration of matching conditions—resulting, for instance, in boundary-induced eigenvalues at $\lambda=-\cot^2\theta$
• Extended (possibly infinite) multiplicity due to algebraic redundancies or degenerate matching (as in multipoint $\delta$-scatterers, where the space of far-field data orthogonal to a finite set of exponentials is infinite-dimensional (Grinevich et al., 2021))

3. Spectral Multiplicity, Exceptional and Infinite-Dimensional Cases

Special transmission eigenvalues arise in a variety of spectrally and physically exceptional situations:

• Zero Energy Eigenvalues: Transmission eigenvalue at $\lambda=0$ occurs if and only if the Jost function vanishes at $F(0)=0$; this zero may be simple or double, but not higher, and is computable directly via conditions on the boundary and potential (Aktosun et al., 2014).
• Boundary-Induced Eigenvalues: For selfadjoint half-line problems, a value $k=i\cot\theta$ may cause vanishing of both $F(k)$ and $D(k)$—corresponding to an eigenvalue at $\lambda=-\cot^2\theta$, a phenomenon purely due to the boundary condition and the particular tuning of the potential (Aktosun et al., 2014).
• Multiple and Infinite Multiplicity:
  • For the Dirac-delta or square-well, multiplicities above one can occur by tuning the potential or support—examples include double zeros at $\lambda=0$ (Aktosun et al., 2014).
  • In the case of multipoint scatterers, each positive energy is a strong transmission eigenvalue of infinite multiplicity in $d=2,3$, since imposing $n$ linear constraints in infinite-dimensional function spaces (e.g., $L^2(\Sph^{d-1})$) leaves an infinite-dimensional orthogonal complement (Grinevich et al., 2021).
  • In discrete Schrödinger settings, certain spectral sum-rules lead to nonunique inverse reconstructions: the “unusual” case $\sum_j \lambda_j = 4(b-1)$ yields either unique, finite, or infinite families of compatible potentials sharing the same transmission spectrum (Aktosun et al., 2015).

4. Complex and Nonstandard Spectral Features

Special transmission eigenvalues also encompass non-real or nonstandard locations in the complex plane:

• Complex Conjugate Pairs: Zeros of $D(k)$ off the real and imaginary axes yield complex eigenvalues in quartets or pairs, providing “special” non-real transmission eigenvalues. The presence of infinitely many complex eigenvalues with asymptotic distribution along logarithmic curves can be proved for certain smoothly varying or piecewise-constant models (Aktosun et al., 2014, Xu et al., 2017).
• Born Transmission Eigenvalues: In certain linearized (Born) transmission problems motivated by transformation optics, all eigenvalues are not only complex but are uniformly separated from the real axis—there exists a horizontal strip $|Im\,k| < c$ containing no transmission eigenvalues at all (Hovsepyan, 2023). This stands in contrast to the nonlinear problem, where eigenvalues may approach the real axis arbitrarily closely.

5. Implications for Uniqueness and Inverse Problems

Special transmission eigenvalues strongly influence the identifiability of media via inverse spectral data:

• In continuous Schrödinger and interior transmission problems, transmission eigenvalues (together with boundary parameters and possibly one normalization constant) uniquely determine the potential except in precisely characterized exceptional cases—such as vanishing mean potential or critical sum rules (Aktosun et al., 2014, Aktosun et al., 2015).
• Certain exceptional (special) cases (e.g., sum rules in discrete Schrödinger, or extended knowledge of $\eta$ in inverse spectral results) allow for unique or nonunique recovery of medium parameters depending explicitly on the properties of the transmission eigenvalue set and the corresponding subintervals on which the background coefficients are known (Xu et al., 2017).

In higher dimensions and in Maxwell or elastic systems, the set of transmission eigenvalues for material contrasts that are constant near the boundary is proven to be infinite, discrete, and devoid of finite accumulation points; the corresponding eigenfunctions form a complete set in naturally associated Hilbert spaces (Haddar et al., 2017). The structure of special eigenvalues thus underlies the analytic and computational tractability of a broad class of inverse spectral and scattering problems.

6. Computational and Spectral Resolution Techniques

Accurate computation and separation of special transmission eigenvalues, especially those clustering near the real axis or possessing enhanced multiplicity, is nontrivial due to non-selfadjointness and possible near-coalescence of modes:

• Recursive Integral Methods and Contour-Integral-Based Algorithms provide robust frameworks for isolating and resolving clusters of eigenvalues (whether real, complex, or exhibiting symmetry), by recursive domain subdivision and spectral projection testing (Huang et al., 2015, Gong et al., 2020). These approaches are particularly suited to identifying “special” eigenvalue subgroups (e.g., trapped modes, near-resonances, or highly degenerate points).
• For anisotropic, Maxwell, or elastic cases, the transmission eigenvalue problem is cast as a holomorphic Fredholm operator pencil, with spectral indicator or secant-based Newton algorithms capable of resolving both real and complex special transmission eigenvalues, as well as handling potential multiplicity (Gong et al., 2020, Ji et al., 2018, Haddar et al., 2017).
• In strictly concave domains, microlocal and semiclassical parametrix constructions reveal larger eigenvalue-free regions, ensuring that real (or nearly real) special transmission eigenvalues cannot accumulate close to the real axis unless permitted by the domain’s geometry, affirming almost optimal Weyl-law remainder estimates for the eigenvalue counting function (Vodev, 2015).

7. Physical and Mathematical Significance

Special transmission eigenvalues are of fundamental interest for several reasons:

• They correspond to non-scattering (“invisible”) energies, of intrinsic value in inverse scattering, in the analysis of cloaking schemes, and in resonance theory.
• Their existence and location—especially exceptional, boundary-induced, or infinite-multiplicity cases—inform the uniqueness, stability, and sensitivity of reconstruction algorithms for material properties.
• In certain multipoint scatterer models, the ubiquity of special transmission eigenvalues (every $E>0$, multiplicity infinite) reflects sharp geometric nonuniqueness and highlights the singular phenomena possible for generalized point interactions (Grinevich et al., 2021).
• From the analytic viewpoint, their distribution—whether inside or outside strips in the complex plane, or determined by domain concavity—signals the deep interplay between geometry, boundary conditions, and spectral theory; this deeply influences the spectral and asymptotic analysis of operator pencils and the applicability of Fredholm, semiclassical, and microlocal techniques (Hovsepyan, 2023, Vodev, 2015, Haddar et al., 2017).

The systematic characterization and computation of special transmission eigenvalues thus remains central across theoretical, computational, and applied directions in inverse problems and mathematical physics.