Semiclassical resonances under local magnetic fields

arXiv:2604.17854 · math-ph, math.AP, math.MP, quant-ph · Submitted 2026-04-20 · Read on arXiv

Listen

Radio episode about this paper

Transcript

Introduction to the show: ident: Mathematical Physics. The day in mathematical physics: rigorous models of physical systems.

Kai: I'm Kai.

Mira: And I'm Mira, here with Lev.

Kai: We're looking at the paper: "Semiclassical resonances under local magnetic fields".

Mira: Here's the gist of the paper. We study resonances for the semiclassical magnetic Laplacian in the full plane with a compactly supported magnetic field in the framework of semiclassical complex scaling and black box scattering theory.

Kai: Assuming that the magnetic field is locally constant, we prove the existence of semiclassical resonances near the Landau levels with exponentially small imaginary parts.

Mira: Setting and Operator Definition We investigate the existence of resonances for P(h) = (−ih∇ − A)2 under specific hypotheses on the vector potential A.

Kai: The operator P(h) is defined from the quadratic form QA,h(u) = Z R squared (-i∇ − A)u 2 dx, which yields a self-adjoint operator on L 2(R 2) with domain D = H2(R 2), where (-ih∇ − A) squared = -h 2∆ − 2ihA · ∇ − ih(div A) + A squared <ref:2604.17854#pg2>.

Paper summary: Kai: So this paper, "Semiclassical resonances under local magnetic fields," by Pavel Exner and Ayman Kachmar, it's looking at how quantum systems behave when you have a magnetic field that isn't constant all over the space. They use semiclassical complex scaling and black box scattering theory to study resonances for the magnetic Laplacian in the full plane with a compactly supported magnetic field.

Mira: What they claim is pretty specific—they assume the magnetic field is locally constant, and they prove there are semiclassical resonances near Landau levels that have exponentially small imaginary parts. That means these resonances aren't just mathematical artifacts; they actually show up physically when you look at the system in this semiclassical limit where h approaches zero.

Lev: From a hardware perspective, it’s interesting because we’re dealing with operators like P(h) which are essentially the magnetic Hamiltonian, and they're looking for poles of the resolvent. If these resonances have exponentially small imaginary parts, that suggests a kind of stability or localization in the system under these specific field configurations.

Kai: Right, so they focus on constant fields first, but they go further and show how resonances emerge from other things too. They look at magnetic wells—where the field has a local minimum—and sharp magnetic interfaces where there's a jump in the field strength across a curved surface or interface with non-degenerate curvature.

Mira: And they also investigate the case of anharmonic Landau levels when the field has an isolated zero, meaning it vanishes at one point, and they find resonances near those levels too. This is important because it shows that these phenomena aren't just about simple uniform fields; they are tied to the specific geometry of how that magnetic field is shaped in space.

Lev: The main takeaway for us as researchers building error correction systems is seeing these different scaling regimes for the real parts of the resonances, like those involving lambda gamma n h and gamma squared <ref:2604.17854#pg2>. That tells us how quickly we need to scale our system parameters just to capture these features in a physical setup.

Kai: So, it’s essentially mapping out where these quantum states appear when the magnetic field has some local structure—constant, wells, or interfaces—and showing that their imaginary parts shrink exponentially fast as h goes to zero. This paper sets up the groundwork for understanding how geometry influences these resonances in a semiclassical setting.

Conclusion: Kai: Thinking about the title, "Semiclassical resonances under local magnetic fields," it really captures that idea that the quantum behavior changes depending on the local structure of the magnetic field, not just how strong it is overall. Exner and Kachmar are showing us how this happens in a mathematical framework called semiclassical complex scaling and black box scattering theory.

Mira: What matters is they've managed to prove the existence of these resonances across five distinct scenarios: constant fields, isolated zeros of local magnetic fields, magnetic wells, sharp magnetic interfaces with non-degenerate curvature, and even zero-field islands where the field vanishes on a hole. That variety shows how rich the spectrum can be when you move beyond simple uniform fields.

Lev: For us in hardware design, the implication is that if we want to engineer a system that exhibits these specific features—say, resonances near anharmonic Landau levels or those induced by sharp interfaces—we need to account for all these geometric possibilities in our modeling and error correction overhead calculations.

Kai: So, even though it’s a mathematical existence proof in the semiclassical limit h goes to zero, what it means practically is that these specific field configurations create predictable quantum states that we can potentially probe or engineer. It suggests that the way a magnetic field is shaped locally dictates the low-energy spectrum of the system.

Mira: Exactly. The real parts of these resonances have different scaling regimes in h, which are summarized in Table one and those scaling laws depend on geometric data like flux or curvature, which is what makes it so complex but also very interesting for condensed matter physics <ref:2604.17854#pg1>.

Lev: It’s interesting that the proof relies on constructing quasimodes supported near regions where the magnetic field is one, and they get these exponential bounds on how much error you have in matching them to the actual resonance. That construction method has a clear overhead when translated to a real quantum simulator setup.

Kai: So, while this paper is about proving existence through complex scaling and scattering theory, what it gives us is a roadmap for where to look for these specific types of quantum features when designing systems with spatially varying magnetic fields. It’s about connecting the math of field geometry to the observable quantum spectrum.

PAVEL EXNER, AYMAN KACHMAR

Doppler Institute for Mathematical Physics and Applied Mathematics · American University of Beirut

math-ph, math.AP, math.MP, quant-ph

Submitted: 2026-04-20

Updated: 2026-11-04

Comments: 26 pages, two figures. Amended: the main theorems and body of the paper remain unchanged; the revision added estimates, in the appendix, on the Tang-Zworski Theorem for energies near the threshold

License: http://creativecommons.org/licenses/by/4.0/

Importance score: 86/100

The gist: We study resonances for the semiclassical magnetic Laplacian in the full plane with a compactly supported magnetic field in the framework of semiclassical complex scaling and black box scattering…

Key concepts

Magnetic Laplacian Operator P(h)
This is a quantum mechanical operator describing the behavior of particles in a magnetic field. It is defined using the magnetic vector potential A and involves terms like -ih∇² and A·∇, which are central to understanding energy levels in systems with external magnetic fields.
Semiclassical Resonances
These are special complex numbers associated with the operator P(h) that describe quasi-bound states. In the semiclassical limit (h → 0), these resonances appear near the Landau levels, and their imaginary parts indicate the decay rate of these states.
Complex Scaling
This is a mathematical technique used to define resonances. It involves analytically continuing the operator into a complex domain by multiplying functions by a complex exponential factor. This procedure transforms the original problem into one where resonances can be found as eigenvalues of an elliptic operator.
Tang–Zworski Resonance Theorem
This theorem provides the main tool for proving resonance existence. It states that if certain conditions are met regarding the spectral properties of a family of operators (like P(h)), then resonances must exist within specific complex energy windows, defined by functions S(h), R(h), and w(h).

Terminology

Summary

We study resonances for the semiclassical magnetic Laplacian in the full plane with a compactly supported magnetic field in the framework of semiclassical complex scaling and black box scattering theory. p. 0

Where the paper says it

SEMICLASSICAL RESONANCES UNDER LOCAL MAGNETIC FIELDS PAVEL EXNER AND AYMAN KACHMAR Abstract. We study resonances for the semiclassical magnetic Laplacian in the full plane with a compactly supported magnetic field in the framework of semiclassical complex scaling and black box scattering theory.

Assuming that the magnetic field is locally constant, we prove the existence of semiclassical resonances near the Landau levels with exponentially small imaginary parts. p. 0
Where the paper says it

We study resonances for the semiclassical magnetic Laplacian in the full plane with a compactly supported magnetic field in the framework of semiclassical complex scaling and black box scattering theory. Assuming that the magnetic field is locally constant, we prove the existence of semiclassical resonances near the Landau levels with exponentially small imaginary parts.

Setting and Operator Definition

We investigate the existence of resonances for P(h) = (−ih∇ − A)2 under specific hypotheses on the vector potential A. The operator P(h) is defined from the quadratic form QA,h(u) = Z R squared (-i∇ − A)u 2 dx, which yields a self-adjoint operator on L 2(R 2) with domain D = H2(R 2), where (-ih∇ − A) squared = -h 2∆ − 2ihA · ∇ − ih(div A) + A squared <ref:2604. p. 1

Where the paper says it

We now turn to the precise definition of the operator P(h) starting from the quadratic form QA,h(u) = Z R2 |(−i∇ − A)u| 2 dx, u ∈ C ∞ c (R 2 ), which is non-negative and closable. Its Friedrichs extension yields a self-adjoint operator (our operator P(h)) on L 2 (R 2 ) with domain D = {u ∈ L 2 (R 2 ): (−ih∇ − A)u ∈ L 2 (R 2 ; C 2 ),(−ih∇ − A) 2u ∈ L 2 (R 2 )}, where (−ih∇ − A) 2 = −h 2∆ − 2ihA · ∇ − ih(div A) + |A| 2 .

17854#pg2>. Since the magnetic field is compactly supported and P(h) ≥ 0 in the sense of quadratic forms, the spectrum of P(h) is [0, +∞) and is purely essential <ref:2604. p. 1
Where the paper says it

4) D = {u ∈ H1 (R 2 ): ∆u ∈ L 2 (R 2 )} = H2 (R 2 ). Since the magnetic field is compactly supported and P(h) ≥ 0 in the sense of quadratic forms, the spectrum of P(h) is [0, +∞) and is purely essential.

17854#pg2>.

Scenarios for Resonance Existence

The aim of this article is to study resonances of P(h) in the semiclassical limit h → 0 <ref:2604.17854#pg2>. The authors demonstrate that a local magnetic field of described type creates resonances in the semiclassical/strong field regime <ref:2604. p. 1

Where the paper says it

This approach allows us to transform the resonance problem into a spectral problem for a non-selfadjoint operator, whose discrete spectrum in a suitable complex sector corresponds precisely to the resonances of the original Hamiltonian. In this work, we demonstrate that a local magnetic field of described type creates resonances in the semiclassical/strong field regime.

17854#pg2>. Five distinct scenarios are investigated:

  1. Constant fields: If the magnetic field is locally constant, there are resonances whose real parts are close to the Landau levels, and whose imaginary parts are negative and exponentially small <ref:2604. p. 1

    Where the paper says it

    In this work, we demonstrate that a local magnetic field of described type creates resonances in the semiclassical/strong field regime. We investigate five distinct scenarios and prove the following: • Constant fields: If the magnetic field is locally constant, then there are resonances whose real parts are close to the Landau levels, and whose imaginary parts are negative and exponentially small (see Theorem 1.

    17854#pg2>.

  2. Isolated zeros of a local magnetic field: If the magnetic field vanishes at a point, without loss of generality identified with the origin, and behaves locally like r γ, then there are resonances whose real parts are close to these levels, and whose imaginary parts are negative and exponentially small <ref:2604. p. 1

    Where the paper says it

    We investigate five distinct scenarios and prove the following: • Constant fields: If the magnetic field is locally constant, then there are resonances whose real parts are close to the Landau levels, and whose imaginary parts are negative and exponentially small (see Theorem 1. • Isolated zeros of a local magnetic field: If the magnetic field vanishes at a point, without loss of generality identified with the origin, and behaves locally like r γ , where r is a radial variable and γ is a positive parameter, then then the local harmonic potential of the previous case is replaced by an anharmonic one, the eigenvalues of which we call anharmonic Landau levels.

    17854#pg2>.

  3. Magnetic wells: If the magnetic field attains a nondegenerate and positive local minimum (a magnetic well), then it gives rise to resonances with a semiclassical expansion in powers of h <ref:2604. p. 1

    Where the paper says it

    For such fields, there are resonances whose real parts are close to these levels, and whose imaginary parts are negative and exponentially small (see Theorem 1. • Magnetic wells: If the magnetic field attains a nondegenerate and positive local minimum (a magnetic well), then it gives rise to resonances with a semiclassical expansion in powers of h (see Theorem 1.

    17854#pg2>.

  4. Sharp magnetic interface: If the magnetic field is locally piecewise constant with a jump discontinuity across a curve modeling a sharp interface, and if the curvature of the discontinuity has a local non-degenerate maximum, then there are resonances which are asymptotically described by the field strength and the geometry of the interface <ref:2604. p. 1

    Where the paper says it

    • Magnetic wells: If the magnetic field attains a nondegenerate and positive local minimum (a magnetic well), then it gives rise to resonances with a semiclassical expansion in powers of h (see Theorem 1. • Sharp magnetic interface: If the magnetic field is locally piecewise constant with a jump discontinuity across a curve modeling a sharp interface, and if the curvature of the discontinuity has a local non-degenerate maximum, then there are resonances which are asymptotically described by the field strength and the geometry of the interface (see Theorem 1.

    17854#pg2>.

  5. Zero-field island: If the magnetic field vanishes on an open set ω (a hole) and is positive in a neighborhood of ∂ω, then there are resonances whose real parts are close to h squared times the eigenvalues of the Dirichlet Laplacian on ω <ref:2604. p. 1

    Where the paper says it

    • Sharp magnetic interface: If the magnetic field is locally piecewise constant with a jump discontinuity across a curve modeling a sharp interface, and if the curvature of the discontinuity has a local non-degenerate maximum, then there are resonances which are asymptotically described by the field strength and the geometry of the interface (see Theorem 1. • Zero-field island: If the magnetic field vanishes on an open set ω (a hole) and is positive in a neighborhood of ∂ω, then there are resonances whose real parts are close to h 2 times the eigenvalues of the Dirichlet Laplacian on ω (see Theorem 1.

    17854#pg2>.

Asymptotic Scaling Regimes

The asymptotic expansions of the real parts of these resonances exhibit distinct scaling regimes in the semiclassical parameter h, as summarized in Table 1 <ref:2604.17854#pg2>. The real coefficients Λγn andln are eigenvalues of limiting operators, while the real coefficients a0, a1, a2 and b0, b1, b2 depend on the geometric and spectral data of each configuration <ref:2604. p. 1

Where the paper says it

The asymptotic expansions of the real parts of these resonances exhibit distinct scaling regimes in the semiclassical parameter h, as summarized in Table 1. The real coefficients Λγ n and ℓn are eigenvalues of limiting operators, while the real coefficients a0, a1, a2 and b0, b1, b2, depend on the geometric and spectral data of each configuration.

17854#pg2>.

Proof Techniques for Specific Cases

The proof of Theorem 1.1 involves constructing quasimodes supported near a region where B = 1 and applying the black box resonance existence theorem of Tang and Zworski <ref:2604. p. 6

Where the paper says it

These proofs share the common strategy of constructing quasimodes and applying the resonance existence theorem of Tang and Zworski [22]. Defining resonances We present the rigorous definition of resonances for P(h) within the black box scattering framework (see [17, 22]).

17854#pg3>. The construction yields quasimodes un,m(x) = χ(x)ψn,m(x) with support in the disk B(0, r0), such that un,mL2(R 2) = 1 + O bm+1e(-(1-δ) 2r 20b/2) and HL − Λn(b)un,mL2(R 2) = O bm+2e(-(1-δ) 2r 20b/4) <ref:2604.17854#pg4>. This leads to the conclusion that there exists at least one resonance of P(h) in [En(h) − w(h), En(h) + w(h)] + i[-h-3e(-cr 20/h), 0] <ref:2604. p. 2
Where the paper says it

Suppose that curl A = 1 on a disk of radius r0 ≤ R0. For each n ∈ N0 and c ∈ (0, 1/4), there exists hn > 0 such that for all h ∈ (0, hn), the rectangle [En(h) − h −2 e −cr2 0 /2h , En(h) + h −2 e −cr2 0 /2h ] + i[−h −3 e −cr2 0 /h , 0) contains at least one resonance of P(h).

17854#pg5>.

Resonances Near Anharmonic Landau Levels

The proof for Theorem 1.2 is similar to that of Theorem 1. p. 0

Where the paper says it

In particular, B vanishes outside the disk B(0, R0). Note that by Stokes’ theorem, the flux α must satisfy the compatibility condition (1.

1 and relies on the construction of quasi-modes with exponentially small remainder <ref:2604. p. 0
Where the paper says it

Outside it the trajectory is a straight line; if it hits the disk in a non tangential way, it leaves again and the time spent in the disk is shorter the stronger the field is. Our aim is to show that, in contrast, the quantum particle can be trapped temporarily in the field support and the lifetime of such a quasi-stationary state increases exponentially with the field strength.

17854#pg3>. The modified quasimodes ˜un(x) are defined as u˜n(x) = e iϕ/hun(x - p; 1/h), and it is shown that u˜nL2 = 1 + O e(-c1/h), supp ˜un subset B(p, r0), and P(h) − Eγn(h))˜un = O e(-c1/h) <ref:2604.17854#pg4>. Applying Theorem 2. p. 6
Where the paper says it

5, respectively. These proofs share the common strategy of constructing quasimodes and applying the resonance existence theorem of Tang and Zworski [22].

8 of Tang and Zworski guarantees the existence of a resonance zn in [Eγn(h) − w(h), Eγn(h) + w(h)] + i[-h-3S(h), 0] <ref:2604. p. 6
Where the paper says it

5, respectively. These proofs share the common strategy of constructing quasimodes and applying the resonance existence theorem of Tang and Zworski [22].

17854#pg5>.

Curvature Induced Resonances

The proof of Theorem 1.4 relies on quasimodes construction and the theorem by Tang–Zworski <ref:2604. p. 6

Where the paper says it

5, respectively. These proofs share the common strategy of constructing quasimodes and applying the resonance existence theorem of Tang and Zworski [22].

17854#pg3>. The resulting resonance window is [λ˜n(h) − w(h), λ˜n(h) + w(h)] + i[-h-3e(-c/h1/8), 0] <ref:2604. p. 3
Where the paper says it

double magnetic wells), one expects exponential splitting e −S/h (with S > 0 realted to some Agmon distance) due to tunneling (see [7, 10, 11]). Proving this is delicate because the resonance window width Lnh 3 is polynomially large compared to the exponentially small splitting.

17854#pg5>.

Zero-Field Island Resonances

The proof of Theorem 1.5 relies on the construction of quasimodes defined by the eigenfunctions of a specific auxiliary operator <ref:2604.17854#pg3>. The resulting resonances are in [h 2ˆln(1/h) − h-2e(-c/2h1/2)] + i[-h 3e(-c/h1/8), 0] <ref:2604. p. 0

Where the paper says it

Date: April 21, 2026. 1 arXiv:2604.

17854#pg5>.

Conclusion and Examples

The authors conclude by providing examples of vector potentials satisfying their hypotheses, including a radial step magnetic field <ref:2604.17854#pg2> and a general form involving a cut-off function <ref:2604. p. 5

Where the paper says it

Consider B0 ∈ L 2 (R 2 ) with support K0 ⊂ B(0, R0), and let a0 ∈ H1 (B(0, R0)) be a vector potential satisfying2 curl a0 = B. Thanks to the smooth Urysohn’s lemma, there is a cut-off function χ ∈ C∞ c (R 2 ; [0, 1]) such that χ(x) = 1 if x ∈ K0 ∪B(0, ϵ), for some 0 < ϵ < R0, and χ(x) = 0 if |x| ≥ R0.

17854#pg3>. These examples confirm the existence of resonances near Landau levels, anharmonic Landau levels, curvature induced resonances, and zero-field island eigenvalues respectively <ref:2604. p. 0
Where the paper says it

Assuming that the magnetic field is locally constant, we prove the existence of semiclassical resonances near the Landau levels with exponentially small imaginary parts. We also prove that resonances emerge from a magnetic step discontinuity along a curved interface or a non-degenerate magnetic well, and in the vicinity of anharmonic Landau levels if the field has an isolated zero.

17854#pg4>.

Organization of the Paper

The rest of the paper is organized as follows. In Section 2, we verify that our operator P(h) satisfies the black box assumptions in [22] and define resonances as eigenvalues of a scaled operator following [17] <ref:2604. p. 6

Where the paper says it

The rest of the paper is organized as follows. In Section 2, we verify that our operator P(h) satisfies the black box assumptions in [22] and define resonances as eigenvalues of a scaled operator following [17].

17854#pg4>. Sections 3, 4, 5, 6 and 7 are devoted to the proofs of Theorems 1. p. 6
Where the paper says it

In Section 2, we verify that our operator P(h) satisfies the black box assumptions in [22] and define resonances as eigenvalues of a scaled operator following [17]. Sections 3, 4, 5, 6 and 7 are devoted to the proofs of Theorems 1.

1, 1.2, 1.4, 1.3 and 1.5 respectively <ref:2604. p. 0
Where the paper says it

Date: April 21, 2026. 1 arXiv:2604.

17854#pg4>.

Defining Resonances Rigorously

We present the rigorous definition of resonances for P(h) within the black box scattering framework <ref:2604.17854#pg3>. This involves defining resonances as eigenvalues of a deformed operator obtained from P(h) through complex scaling <ref:2604. p. 6

Where the paper says it

Defining resonances We present the rigorous definition of resonances for P(h) within the black box scattering framework (see [17, 22]). After verifying the necessary assumptions, we introduce complex scaling to access the resonances as eigenvalues of a deformed operator.

17854#pg2>. A key property established in [17, Lemma 5. p. 9
Where the paper says it

Consequently, by Analytic Fredholm Theory, the spectrum of Pθ(h) in C \ e −2iθ[0, ∞) is discrete4 . A key property established in [17, Lemma 5.

2] is the independence of the spectrum on the choice of scaling angle θ, provided we stay away from the rotated real axis <ref:2604. p. 9
Where the paper says it

A key property established in [17, Lemma 5. 2] is the independence of the spectrum on the choice of scaling angle θ, provided we stay away from the rotated real axis.

17854#pg3>.

Tang–Zworski Resonance Criterion

We recall a result from [22, Theorem, p. 264], which will be our main tool for proving the existence of resonances <ref:2604. p. 9

Where the paper says it

We recall the following result from [22, Theorem, p. 264], which will be our main tool for proving the existence of resonances.

17854#pg2>. This criterion involves three functions S(h), R(h) and w(h) defined on a right neighborhood of 0 and satisfying De-D/h ≤ S(h) for some constant D > 0, R(h) = O(h∞) and R(h) ≪ S(h) as h → 0 +, w 2/(2h-3S(h)) → +∞ as h → 0 + <ref:2604. p. 9
Where the paper says it

264], which will be our main tool for proving the existence of resonances. The statement involves three functions S(h), R(h) and w(h) defined on a right neighborhood of 0 and satisfying De−D/h ≤ S(h) for some constant D > 0, R(h) = O(h ∞) and R(h) ≪ S(h) as h → 0 +, w(h) 2 h−3S(h) → +∞ as h → 0 +.

17854#pg2>.

Complex Scaling Procedure

We introduce complex scaling to define resonances for the magnetic Hamiltonian P(h) <ref:2604.17854#pg4>. The scaled operator Pθ(h) is elliptic and that Pθ(h) − z is Fredholm of index 0 for z in C-(e(-2iθ[0, ∞)) <ref:2604. p. 9

Where the paper says it

1] and the discussion following [17, (5. We obtain that for each θ ∈ [0, θ0] with θ0 < π/2, the scaled operator Pθ(h) is elliptic and that Pθ(h) − z is Fredholm of index 0 for z ∈ C \ e −2iθ[0, ∞).

17854#pg4>. This independence allows us to define resonances unambiguously via Definition 2. p. 9
Where the paper says it

Specifically5 , for 0 ≤ θ1 < θ2 ≤ θ0 and z0 ∈ C \ e −2i[θ1,θ2] [0, ∞), the dimensions of Ker(Pθ1 (h) − z0) and Ker(Pθ2 (h) − z0) coincide. This independence allows us to define resonances unambiguously: Definition 2.

6 <ref:2604. p. 0
Where the paper says it

Date: April 21, 2026. 1 arXiv:2604.

17854#pg5>.

Tang–Zworski Resonance Theorem

Theorem 2.8 states that if there exist normalized functions uj (h) in D, j in J(h), with supp uj (h) subset B(0, R0) such that (P(h) − Ej (h))uj (h)L2(R 2) = O R(h), where Ej (h) ∈ [E0 − C0h, E0 + C0h] for some constants E0, C0, and Ej (h) − Ej'(h) ≥ w(h) for j!= j' <ref:2604.17854#pg3>. Then for sufficiently small h, there exist resonances zi(h) of P(h) in [Ej (h) − w(h), Ej (h) + w(h)] + i[-3S(h), 0] <ref:2604. p. 7

Where the paper says it

In the exterior of B(0, R0), A = Aα, div A = 0, and the coefficients of Q(h) are given by: a(2,0)(w) = a(0,2)(w) = 1, a(1,1)(w) = 0 a(1,0)(w) = 2αw2 q(w) , a(0,1)(w) = − 2αw1 q(w) a0(w) = α 2 w 2 1 + w 2 2 (q(w))2 = α 2 q(w) where q(w) = w 2 1 + w 2 2 . Note that q(w) does not vanish on U if we choose R > R0 sufficiently large and ϵ > 0 sufficiently small.

17854#pg4>.

Summary of Findings

The paper proves the existence of semiclassical resonances for P(h) under various local magnetic field configurations, including constant fields, isolated zeros, magnetic wells, sharp interfaces, and zero-field islands <ref:2604.17854#pg2>. The real parts of these resonances exhibit distinct scaling regimes in h <ref:2604. p. 1

Where the paper says it

• Zero-field island: If the magnetic field vanishes on an open set ω (a hole) and is positive in a neighborhood of ∂ω, then there are resonances whose real parts are close to h 2 times the eigenvalues of the Dirichlet Laplacian on ω (see Theorem 1. The asymptotic expansions of the real parts of these resonances exhibit distinct scaling regimes in the semiclassical parameter h, as summarized in Table 1.

17854#pg2>.

Acknowledgments

The authors would like to thank Joachim Asch, William Borrelli and Davide Fermi for fruitful discussions <ref:2604.17854#pg2>. A.K. is partially supported by a startup fund at AUB (grant no. p. 0

Where the paper says it

SEMICLASSICAL RESONANCES UNDER LOCAL MAGNETIC FIELDS PAVEL EXNER AND AYMAN KACHMAR Abstract. We study resonances for the semiclassical magnetic Laplacian in the full plane with a compactly supported magnetic field in the framework of semiclassical complex scaling and black box scattering theory.

513125) <ref:2604. p. 0
Where the paper says it

Date: April 21, 2026. 1 arXiv:2604.

17854#pg2>.

References

[1] W. Assaad, The breakdown of superconductivity in the presence of magnetic steps, Commun. Contemp. Math. 23 (2021), Paper No. p. 6

Where the paper says it

Organization. The rest of the paper is organized as follows.

2050005, 53 pp.

[2] W. Assaad, B. Helffer and A Kachmar, Semiclassical eigenvalue estimates under magnetic steps, Anal. p. 0

Where the paper says it

SEMICLASSICAL RESONANCES UNDER LOCAL MAGNETIC FIELDS PAVEL EXNER AND AYMAN KACHMAR Abstract. We study resonances for the semiclassical magnetic Laplacian in the full plane with a compactly supported magnetic field in the framework of semiclassical complex scaling and black box scattering theory.

PDE 17 (2024), 535–585; MR4713108.

[3] W. Assaad, A Kachmar and M. p. 0

Where the paper says it

SEMICLASSICAL RESONANCES UNDER LOCAL MAGNETIC FIELDS PAVEL EXNER AND AYMAN KACHMAR Abstract. We study resonances for the semiclassical magnetic Laplacian in the full plane with a compactly supported magnetic field in the framework of semiclassical complex scaling and black box scattering theory.

Persson-Sundqvist, The distribution of superconductivity near a magnetic barrier, Comm. p. 5
Where the paper says it

For every n ∈ N0, there exist positive constants c, hn and a function εn(h) = o(1) such that the rectangle [ℓnh 2 − h 2 εn(h), ℓnh 2 + h 2 εn(h)] + i[−h −3 e −c/h1/2 , 0) contains at least one resonance of P(h), provided that h ∈ (0, hn]. We know from [15] that in the limit of infinite intensity the magnetic field produces a Dirichlet barrier; for large but finite field, the particle can tunnel through giving rise to resonances, as confirmed by Theorem 1.

Math. Phys. 366 (2019), no. 1, 269–332.

[4] H.L. Cycon, R.G. Froese, W Kirsch, B Simon: Schr¨odinger Operators, with Applications to Quantum Mechanics and Global Geometry, Springer, Berlin and Heidelberg 1987.

[5] E.L Giacomelli, A Kachmar, M Sundqvist. p. 0

Where the paper says it

SEMICLASSICAL RESONANCES UNDER LOCAL MAGNETIC FIELDS PAVEL EXNER AND AYMAN KACHMAR Abstract. We study resonances for the semiclassical magnetic Laplacian in the full plane with a compactly supported magnetic field in the framework of semiclassical complex scaling and black box scattering theory.

High flux asymptotics and critical phenomena for the magnetic Laplacian. p. 2
Where the paper says it

, as h → 0. Local behavior Real part asymptotics Constant field Re zn(h) ∼ (2n + 1)h Anharmonic Landau Hamiltonian Re zn(h) ∼ Λ γ nh 1+ γ 2+γ Sharp magnetic interface Re zn(h) ∼ a0h + a1h 3/2 + (2n + 1)a2h 7/4 Magnetic well Re zn(h) ∼ b0h + (2nb1 + b2)h 2 Zero-field island Re zn(h) ∼ ℓnh 2 explicit expressions, involving the flux, curvature, and magnetic well parameters, are provided in the statements of Theorems 1.

In progress.

[6] Y Guedes-Bonthonneau, N Raymond and S V˜u Ngo. c, Exponential localization in 2D pure magnetic wells, Ark. p. 3

Where the paper says it

When wells are well separated and symmetric (e. double magnetic wells), one expects exponential splitting e −S/h (with S > 0 realted to some Agmon distance) due to tunneling (see [7, 10, 11]).

Mat. 59 (2021), 53–85.

[7] P Exner and L Morin, The flea on the Magnetic Elephant, Lett. p. 0

Where the paper says it

SEMICLASSICAL RESONANCES UNDER LOCAL MAGNETIC FIELDS PAVEL EXNER AND AYMAN KACHMAR Abstract. We study resonances for the semiclassical magnetic Laplacian in the full plane with a compactly supported magnetic field in the framework of semiclassical complex scaling and black box scattering theory.

Math. Phys. 116 (2026), no. p. 0
Where the paper says it

Considering the operator (−i∇−bA) 2 with a large parameter b, one can set h = b −1 in which case the rescaled problem reduces to the analysis of semiclassical resonances for the operator P(h). Date: April 21, 2026.

2, Paper No. p. 6
Where the paper says it

Organization. The rest of the paper is organized as follows.

23.

[8] S Fournais, B Helffer Spectral methods in surface superconductivity, Birkh¨auser, Basel (2010). p. 1

Where the paper says it

We define them via complex scaling, specifically as eigenvalues of a Fredholm operator obtained from P(h) through an analytic dilation of the coordinates in the region exterior to the magnetic field (see Section 2 and Definition 2. This approach allows us to transform the resonance problem into a spectral problem for a non-selfadjoint operator, whose discrete spectrum in a suitable complex sector corresponds precisely to the resonances of the original Hamiltonian.

[9] S Fournais, B Helffer and A Kachmar, Tunneling effect induced by a curved magnetic edge, in The physics and mathematics of Elliott Lieb—the 90th anniversary. Vol. I, 315–350, EMS Press, Berlin.

[10] S Fournais, Y Guedes-Bonthonneau, L Morin and N Raymond. Tunneling between magnetic wells in two dimensions, arXiv:2502. p. 0

Where the paper says it

Introduction The behavior of a quantum system and its classical counterpart is sometimes substantially different. As a simple example, consider a charged particle in two dimensions exposed to a constant magnetic field supported in a disk.

17290 (2025).

[11] S Fournais, L Morin and N Raymond, Purely magnetic tunneling between radial magnetic wells, Rev. p. 3

Where the paper says it

When wells are well separated and symmetric (e. double magnetic wells), one expects exponential splitting e −S/h (with S > 0 realted to some Agmon distance) due to tunneling (see [7, 10, 11]).

Mat. Iberoam. 41 (2025), no. 4, 1367–1392.

[12] B Helffer and Y A Kordyukov, Semiclassical spectral asymptotics for a two-dimensional magnetic Schr¨odinger operator: the case of discrete wells, in Spectral theory and geometric analysis, 55–78, Contemp. p. 4

Where the paper says it

We also need the spectral constants from [2, (1. 7) βa = µa(ζa), C1(a) = 1 3 1 − 1 a ζaϕa(0)ϕ ′ a (0) > 0, C2(a) = 1 2 p µ′′ a (ζa)C1(a) > 0, where µa(ξ) is the lowest eigevalue of the Schr¨odinger operator ha[ξ] = − d 2 dτ 2 + (ξ + ba(τ )τ ) 2 on L 2 (R), ba(τ ) = 1R+ (τ ) + a 1R− (τ ), ζa < 0 is the unique minimum of µa(ξ), ξ ∈ R, and ϕa is the positive eigenfunction of ha[ζa] of unit L 2 -norm corresponding to βa.

Math., 535, Amer. Math. Soc., Providence, RI.

[13] B Helffer and A Mohamed, Caract´erisation du spectre essentiel de l’op´erateur de Schr¨odinger avec un champ magnétique, Ann. p. 4

Where the paper says it

We also need the spectral constants from [2, (1. 7) βa = µa(ζa), C1(a) = 1 3 1 − 1 a ζaϕa(0)ϕ ′ a (0) > 0, C2(a) = 1 2 p µ′′ a (ζa)C1(a) > 0, where µa(ξ) is the lowest eigevalue of the Schr¨odinger operator ha[ξ] = − d 2 dτ 2 + (ξ + ba(τ )τ ) 2 on L 2 (R), ba(τ ) = 1R+ (τ ) + a 1R− (τ ), ζa < 0 is the unique minimum of µa(ξ), ξ ∈ R, and ϕa is the positive eigenfunction of ha[ζa] of unit L 2 -norm corresponding to βa.

Inst. Fourier (Grenoble) 38 (1988), 95–112.

[14] B Helffer and J Sj¨ostrand, R´esonances en limite semi-classique, M´em. p. 5

Where the paper says it

the eigenvalues of the Dirichlet Laplacian in ω. We will prove that emergence of semi-classical resonances near the eigenvalues ℓn when the magnetic field vanishes in ω while being positive in a neighborhood ω.

Soc. Math. France (N.S.) No. 24-25 (1986).

[15] R Hempel and I W Herbst, Strong magnetic fields, Dirichlet boundaries, and spectral gaps, Comm. p. 1

Where the paper says it

This approach allows us to transform the resonance problem into a spectral problem for a non-selfadjoint operator, whose discrete spectrum in a suitable complex sector corresponds precisely to the resonances of the original Hamiltonian. In this work, we demonstrate that a local magnetic field of described type creates resonances in the semiclassical/strong field regime.

Math. Phys. 169 (1995), no. 2, 237–259.

[16] A Kachmar and Mikael Sundqvist, Local strong magnetic fields and the Little-Parks effect, arXiv:2405. p. 0

Where the paper says it

SEMICLASSICAL RESONANCES UNDER LOCAL MAGNETIC FIELDS PAVEL EXNER AND AYMAN KACHMAR Abstract. We study resonances for the semiclassical magnetic Laplacian in the full plane with a compactly supported magnetic field in the framework of semiclassical complex scaling and black box scattering theory.

09099 (2024).

[17] J Sj¨ostrand, A trace formula and review of some estimates for resonances, in Microlocal analysis and spectral theory (Lucca, 1996), 377–437, NATO Adv. Sci. Inst. Ser. C: Math. Phys Sci., 490, Kluwer Acad. Publ., Dordrecht.

[18] J Sj¨ostrand and M Zworski, Complex scaling and the distribution of scattering poles, J Amer Math Soc. p. 1

Where the paper says it

Resonances are usually defined as poles of the meromorphically continued resolvent (P(h) − z) −1 . We define them via complex scaling, specifically as eigenvalues of a Fredholm operator obtained from P(h) through an analytic dilation of the coordinates in the region exterior to the magnetic field (see Section 2 and Definition 2.

4 (1991), no. 4, 729–769.

[19] N Raymond and S V˜u Ngo. c, Geometry and spectrum in 2D magnetic wells, Ann. p. 1

Where the paper says it

• Magnetic wells: If the magnetic field attains a nondegenerate and positive local minimum (a magnetic well), then it gives rise to resonances with a semiclassical expansion in powers of h (see Theorem 1. • Sharp magnetic interface: If the magnetic field is locally piecewise constant with a jump discontinuity across a curve modeling a sharp interface, and if the curvature of the discontinuity has a local non-degenerate maximum, then there are resonances which are asymptotically described by the field strength and the geometry of the interface (see Theorem 1.

Inst. Fourier (Grenoble) 65 (2015), 137–169.

[20] J Reijniers and F M Peeters, Snake orbits and related magnetic edge states, J Phys Condens Matter 12 (2000), 9771. p. 4

Where the paper says it

Illustration of the curve Γ splitting R 2 into regions P1 and P2 with jump discontinuity in the magnetic field. We focus on the case when the field is sign changing leading to a strong localization along the edge; a classical particle in this situation follows a snake-shaped orbit [21].

[21] J Reijniers, F. M Peeters, and A Matulis, Electron scattering on circular symmetric magnetic profiles in a two-dimensional electron gas. Physical Review B 64 (2001), 245314.

[22] S.-H Tang and M Zworski, From quasimodes to resonances, Math. p. 6

Where the paper says it

These proofs share the common strategy of constructing quasimodes and applying the resonance existence theorem of Tang and Zworski [22]. Defining resonances We present the rigorous definition of resonances for P(h) within the black box scattering framework (see [17, 22]).

Res. Lett. 5 (1998), no. 3, 261–272.

[23] M Yang, Resolvent estimates for the magnetic Hamiltonian with singular vector potentials and applications, Comm. p. 9

Where the paper says it

This rank is independent of the choice of θ in the sense that, if z0 ∈ e −2i[0,θ˜) (0, ∞)∩σ(Pθ˜(h)) with ˜θ ∈ [0, θ0), then πθ,z ˜ 0 = πθ,z0 . For the equivalence between complex scaling and meromorphic continuation definitions of resonances in the context of magnetic Hamiltonians with singular vector potentials, we refer to [23, Section 3].

Math. Phys. 394 (2022), 1225–1246.

[P Exner]

Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.

Improvements for AI systems

  1. Precise resonance location prediction: The improved system can predict resonances exponentially close to real energy levels for various magnetic field configurations, such as Constant fields where imaginary parts are exponentially small, or Magnetic wells where they have a specific semiclassical expansion in powers of h.

  2. Anharmonic Landau level analysis: The system can accurately determine the existence and location of resonances near anharmonic Landau levels, using the scaling law E gamma n(h) = h 1 plus gamma 2 plus gamma Lambda gamma n and predicting resonance windows based on Theorem 1.2.

  3. Curvature-induced resonance identification: The improved system can locate resonances arising from sharp magnetic interfaces by calculating the energy level λn(h) using the formula λn(h) = βah − k0C1(a)h 3/2 + (2n + 1)p k2 C2(a) h 7/4, as described in Theorem 1.4.

  4. Zero-field island resonance detection: The system can identify resonances near the eigenvalues of the Dirichlet Laplacian on a zero-field island by using the scaling energy lnh 2 and predicting their location based on Theorem 1.5.

  5. Black box resonance search: The system can search for resonances in any semiclassical magnetic Laplacian operator P(h) by constructing quasimodes and applying the Tang and Zworski resonance criterion, which guarantees a resonance in the interval [Ej (h) − w(h), Ej (h) + w(h)] + i[- h - 3S(h), 0].

Abstract

We study resonances for the semiclassical magnetic Laplacian in the full plane with a compactly supported magnetic field in the framework of semiclassical complex scaling and black box scattering theory. Assuming that the magnetic field is locally constant, we prove the existence of semiclassical resonances near the Landau levels with exponentially small imaginary parts. We also prove that resonances emerge from a magnetic step discontinuity along a curved interface or a non-degenerate magnetic well, and in the vicinity of anharmonic Landau levels if the field has an isolated zero.

Sources

Related papers