Geometric characterization of the discrete spectrum of quantum layers and their nonsmooth analogues: bound states, thickness uniformity, and the converse problem

A quantum layer is the region of fixed width 2a trapped between two parallel copies of an embedded surface Σ in R³, and the Dirichlet Laplacian on it models a quantum particle confined by hard walls. For complete, noncompact, asymptotically flat surfaces that are not totally geodesic, a long-standing conjecture program initiated by Duclos–Exner–Krejčiřík, Carron–Exner–Krejčiřík, and Lin–Lu asserts that bending always traps the particle, i.e. the discrete spectrum below the essential spectrum is nonempty. The precise state of that program is the following: the thin-layer existence conjecture is settled in the main cases including nonpositive total Gauss curvature and nonnegative Gauss curvature, while the central open fragments are the thickness-uniformity conjecture (Lu–Rowlett Conjecture 2) and the purely Riemannian variational inequality behind it (Conjecture 3), valid when the total Gauss curvature is positive without sign assumptions on the mean curvature at infinity; the conjectured converse that absence of bound states forces the base surface to be totally geodesic (Lin–Lu); and, on the nonsmooth side, the classification of polyhedral layers, where existence has just been proved for the Fichera corner layer but examples with empty discrete spectrum exist and the uniqueness conjecture remains open. The problem asks for resolution of these surviving cores.

Background

A quantum layer is the Riemannian manifold $\Omega \cong \Sigma \times [-a, a]$ obtained as the tubular neighborhood of width $2a$ of an embedded surface $\Sigma \subset \mathbb{R}^3$, with the metric pulled back from Euclidean space; physically it models a particle trapped between hard walls a distance $2a$ apart, as in a semiconductor quantum well. The Hamiltonian is the Dirichlet Laplacian (the Laplacian with vanishing boundary values on the two boundary copies of $\Sigma$). For the flat layer $\mathbb{R}^2 \times [-a, a]$ the spectrum is purely essential and begins at the threshold $\pi^2/(4a^2)$, the first Dirichlet eigenvalue of the cross-section. For a curved base surface the spectrum splits into essential spectrum plus isolated eigenvalues of finite multiplicity below the threshold, the discrete spectrum; these eigenvalues are the energies of bound states, spatially localized states that exist purely because of the geometry. Duclos, Exner and Krejčiřík (2001) proved that bending generically creates such bound states, and Carron, Exner and Krejčiřík (2004) and Lin and Lu (2006, 2007a, 2007b) extended the result in several directions. The following terminology is needed. The second fundamental form $B$ of $\Sigma$ measures extrinsic curvature (how the surface bends in space); the principal curvatures at a point are the largest and smallest normal curvatures of surface curves through that point. $\Sigma$ is totally geodesic if $B$ vanishes identically, which for surfaces in $\mathbb{R}^3$ means $\Sigma$ is (an open subset of) a plane, and asymptotically flat means $B \to 0$ at infinity, so all bending is confined to a compact region. The Gauss curvature $\kappa$ is the product of the two principal curvatures, and the mean curvature $H$ is their average. For surfaces with $\int_\Sigma |\kappa|\, d\Sigma < \infty$, a theorem of Huber makes $\Sigma$ conformally a compact Riemann surface with finitely many points (ends) removed, and a theorem of Hartman gives $\int_\Sigma \kappa\, d\Sigma = 2\pi(\chi - \sum_i \lambda_i)$, where $\chi$ is the Euler characteristic and $\lambda_i$ are the isoperimetric constants of the ends $E_i$, defined as $\lambda_i = \lim_{r \to \infty} \mathrm{vol}(B(r) \cap E_i)/(\pi r^2)$. A surface is parabolic if it admits no positive Green's function (equivalently, Brownian motion on it is recurrent); parabolicity is conformally invariant in dimension two. The known sufficient conditions for a nonempty discrete spectrum over such surfaces are by now broad: nonpositive total Gauss curvature $\int_\Sigma \kappa\, d\Sigma \le 0$; nonnegative Gauss curvature $\kappa \ge 0$; square-integrable gradient of the mean curvature with a thin layer; surfaces ruled outside a compact set; rotationally symmetric ends with positive total curvature; and parabolic surfaces for sufficiently thin layers (Lu and Rowlett 2012). All proofs share one mechanism: locate the threshold of the essential spectrum, then build a test function making the quadratic form dip below it. The mechanism fails in exactly two places, which define the open core. First, in every hard case the layer must be thin ($a < \alpha$ with $\alpha$ depending on the particular surface), while the natural statement, formulated as Conjecture 2 of Lu–Rowlett (2012), demands $\alpha = \alpha(B_\infty)$ depending only on $\sup\|B\|$; the paper proves this only under an auxiliary variational condition, its inequality (5.6): there exists $\varepsilon_0 > 0$ such that for every $R > 0$ some cutoff-type function $j \in C_0^\infty(\Sigma \setminus B(R))$ satisfies $\left(\int_\Sigma H j\, d\Sigma\right)^2 > \varepsilon_0 \int_\Sigma (|\nabla j|^2 + j^2)\, d\Sigma$; and its Conjecture 3 predicts this condition holds whenever $\int_\Sigma \kappa\, d\Sigma > 0$, which the authors verify only when the mean curvature has fixed sign off a compact set. Second, the converse question posed through a referee in Lin–Lu (2006): if the discrete spectrum below the essential spectrum is empty, must $\Sigma$ be totally geodesic? It is known for $\kappa > 0$ (Lin–Lu 2007b) and false without asymptotic flatness (a non-asymptotically-planar layer with empty discrete spectrum was constructed by Duclos–Exner–Krejčiřík), but open in general, together with its analogue for quantum tubes (tubular neighborhoods of curves or of higher-dimensional submanifolds of Euclidean space, on which the same hard-wall confinement problem is posed). On the nonsmooth side, the natural model is the Fichera layer $\{x \in \mathbb{R}^3 : |\min(x_1, x_2, x_3)| < 1\}$, a layer bent along an edge with a corner; Dauge, Lafranche and Ourmières-Bonafos (2018) determined its essential spectrum and proved the discrete spectrum is finite, but existence of a bound state was supported only numerically. Bakharev and Nazarov (2021) then proved existence for the Fichera layer, the cross layer, and their n-dimensional generalizations, and conjectured the subthreshold eigenvalue is unique. Bakharev and Matveenko (2024) began the classification of general polyhedral layers: the essential spectrum is governed by the smallest dihedral angle of the polyhedron, the discrete spectrum is always finite, regular polyhedral layers do have bound states, but there exist polyhedral layers with empty discrete spectrum, so for nonsmooth base surfaces no analogue of the universal bending-traps conjecture can hold and only a classification program is meaningful.

Problem Statement

Let $\Sigma \subset \mathbb{R}^3$ be a $C^2$, orientable, complete, noncompact, embedded surface that is not totally geodesic and is asymptotically flat in the sense that its second fundamental form tends to zero at infinity, with integrable Gauss curvature ($\int_\Sigma |\kappa|\, d\Sigma < \infty$). For a quantum layer $\Omega \cong \Sigma \times [-a, a]$ of width $2a$ (with $a$ small enough that the layer does not overlap itself), consider the Dirichlet Laplacian on $\Omega$. Resolve the following three open cores of the conjecture program on the existence and characterization of the discrete spectrum of quantum layers. All three items are required parts of the task: a complete solution addresses every item, and each item is fully resolved when every claim it states is either proved or disproved; the sub-claims of an item may be resolved independently of one another.

1. (Lu–Rowlett Conjectures 3 and 2: positive total curvature and thickness uniformity) Suppose the total Gauss curvature is positive, $\int_\Sigma \kappa\, d\Sigma > 0$. Prove or disprove the variational inequality conjectured as (5.6) by Lu–Rowlett: there exists $\varepsilon_0 > 0$ such that for every $R > 0$ there is a cutoff-type test function $j \in C_0^\infty(\Sigma \setminus B(R))$, where $B(R)$ is the geodesic ball of radius $R$ in $\Sigma$, with $\left(\int_\Sigma H j\, d\Sigma\right)^2 > \varepsilon_0 \int_\Sigma (|\nabla j|^2 + j^2)\, d\Sigma$, where $H$ is the mean curvature. Settle the thickness-uniformity Conjecture 2, which asks the same existence question for every surface satisfying the hypotheses above: decide whether there exists $\alpha = \alpha(B_\infty) > 0$, depending only on $\sup\|B\|$ (the supremum of the operator norm of the second fundamental form), such that for all $a \in (0, \alpha)$ the layer of width $2a$ has nonempty discrete spectrum below its essential spectrum — with no sign assumption on the mean curvature at infinity and no thinness restriction tied to the particular surface. By Theorem 5.1 of Lu–Rowlett, a proof of the variational inequality implies the thickness-uniformity statement; a counterexample refuting the thickness-uniformity statement also refutes the inequality.

2. (Lin–Lu converse conjecture, 2006) Prove or disprove both of the following independent claims: (a) if the discrete spectrum below the essential spectrum of the Dirichlet Laplacian on such a layer is empty, then the base surface $\Sigma$ must be totally geodesic; and (b) the analogous statement for a quantum tube, a tubular neighborhood of comparable width over an asymptotically flat curve or higher-dimensional submanifold $M$ of Euclidean space: if its discrete spectrum below the essential spectrum is empty, then the base curve or submanifold $M$ must be totally geodesic (a straight line for a curve, an affine subspace for a higher-dimensional submanifold).

3. (Polyhedral layer classification) For the nonsmooth analogues — layers of fixed width over polyhedral surfaces, including the Fichera layer $\{x \in \mathbb{R}^3 : |\min(x_1, x_2, x_3)| < 1\}$ and its n-dimensional corner/cross generalizations, for which existence of at least one bound state has been proved (Bakharev–Nazarov) — resolve the following two independent sub-questions: (i) resolve the conjectured uniqueness of the subthreshold eigenvalue for the Fichera layer and crosses; and (ii) determine the general criteria under which the discrete spectrum of a polyhedral layer is nonempty (equivalently, characterize which polyhedral layers have empty discrete spectrum). The general classification in (ii) is required in full and is not settled by the examples of (i) or by any finite family of polyhedra.

References

1. P. Duclos, P. Exner, D. Krejčiřík, Bound states in curved quantum layers, Comm. Math. Phys. 223 (2001) 13–28, doi:10.1007/pl00005582, https://doi.org/10.1007/pl00005582
2. G. Carron, P. Exner, D. Krejčiřík, Topologically nontrivial quantum layers, J. Math. Phys. 45 (2004) 774–784, doi:10.1063/1.1635998, https://doi.org/10.1063/1.1635998
3. C. Lin, Z. Lu, On the discrete spectrum of generalized quantum tubes, Comm. Partial Differential Equations 31 (2006) 1529–1546, doi:10.1080/03605300600635111, https://doi.org/10.1080/03605300600635111 (arXiv:math/0601201)
4. C. Lin, Z. Lu, Quantum layers over surfaces ruled outside a compact set, J. Math. Phys. 48 (2007) 053522, doi:10.1063/1.2736518, https://doi.org/10.1063/1.2736518
5. C. Lin, Z. Lu, Existence of bound states for layers built over hypersurfaces in R^{n+1}, J. Funct. Anal. 244 (2007) 1–25, doi:10.1016/j.jfa.2006.11.017, https://doi.org/10.1016/j.jfa.2006.11.017
6. C. Lin, Z. Lu, Discrete spectrum of quantum tubes, Lecture notes, Hayama Symposium on Several Complex Variables (2005), arXiv:math/0601280, https://arxiv.org/abs/math/0601280
7. Z. Lu, J. Rowlett, On the discrete spectrum of quantum layers, J. Math. Phys. 53 (2012) 073519, doi:10.1063/1.4736412, https://doi.org/10.1063/1.4736412 (arXiv:1110.6807)
8. D. Krejčiřík, Z. Lu, Location of the essential spectrum in curved quantum layers, J. Math. Phys. 55 (2014) 083520, doi:10.1063/1.4893035, https://doi.org/10.1063/1.4893035 (arXiv:1211.2541)
9. P. Freitas, D. Krejčiřík, A lower bound to the spectral threshold in curved quantum layers, in: Functional Analysis and Operator Theory for Quantum Physics (EMS Series of Congress Reports, EMS 2017), pp. 261–269, doi:10.4171/175-1/14, https://doi.org/10.4171/175-1/14 (arXiv:1602.04990)
10. M. Dauge, Y. Lafranche, T. Ourmières-Bonafos, Dirichlet spectrum of the Fichera layer, Integral Equations Operator Theory 90 (2018) 60, doi:10.1007/s00020-018-2486-y, https://doi.org/10.1007/s00020-018-2486-y
11. M. Dauge, T. Ourmières-Bonafos, N. Raymond, Spectral asymptotics of the Dirichlet Laplacian in a conical layer, Commun. Pure Appl. Anal. 14 (2015) 1239–1258, doi:10.3934/cpaa.2015.14.1239, https://doi.org/10.3934/cpaa.2015.14.1239 (arXiv:1411.6383)
12. P. Exner, V. Lotoreichik, Spectral asymptotics of the Dirichlet Laplacian on a generalized parabolic layer, Integral Equations Operator Theory 92 (2020) 15, doi:10.1007/s00020-020-2571-x, https://doi.org/10.1007/s00020-020-2571-x (arXiv:1805.12448)
13. S. K. Egger, J. Kerner, K. Pankrashkin, Discrete spectrum of Schrödinger operators with potentials concentrated near conical surfaces, Lett. Math. Phys. 110 (2020) 945–968, doi:10.1007/s11005-019-01246-z, https://doi.org/10.1007/s11005-019-01246-z (arXiv:1908.02554, preprint titled 'Attractive conical surfaces create infinitely many bound states')
14. F. L. Bakharev, A. I. Nazarov, Existence of the discrete spectrum in the Fichera layers and crosses of arbitrary dimension, J. Funct. Anal. 281 (2021) 109071, doi:10.1016/j.jfa.2021.109071, https://doi.org/10.1016/j.jfa.2021.109071 (arXiv:2006.12045)
15. F. Bakharev, S. Matveenko, Spectra of the Dirichlet Laplacian in 3-dimensional polyhedral layers, St. Petersburg Math. J. 35 (2024) 597–610, doi:10.1090/spmj/1818, https://doi.org/10.1090/spmj/1818 (arXiv:2304.10575)
16. F. Bakharev, S. Matveenko, Monotonicity of discrete spectra of Dirichlet Laplacian in 3-dimensional layers (2025), arXiv:2508.12696, https://arxiv.org/abs/2508.12696
17. D. Krejčiřík, J. Kříž, Bound states in soft quantum layers, Publ. Res. Inst. Math. Sci. 60 (2024) 741–766, doi:10.4171/prims/60-4-4, https://doi.org/10.4171/prims/60-4-4 (arXiv:2205.04919)
18. C. Lin, D. Mitchell, Asymptotically straight quantum waveguides with axial symmetry in two and three dimensions, J. Appl. Anal. (2026), doi:10.1515/jaa-2025-0156, https://doi.org/10.1515/jaa-2025-0156