Landau Hamiltonian with Gaussian white noise potential and the asymptotics of the bottom of its spectrum

arXiv:2511.22162 · math.PR, math-ph, math.MP, math.SP · Submitted 2025-11-27 · Read on arXiv

Listen

Radio episode about this paper

Transcript

Introduction to the show: ident: Mathematics. The day in mathematics: new theorems, new proofs, and what they open.

Ada: We're looking at the paper: "Landau Hamiltonian with Gaussian white noise potential and the asymptotics of the bottom of its spectrum".

Emmy: Here's the gist of the paper. We present a simple construction of a random Schrödinger operator subject to a magnetic field with a regularity as low as 0−Hölder and a Gaussian white noise electric potential on a two-dimensional bounded box This construction enables us to generalise an asymptotic result for the bottom of the spectrum of the two-dimensional continuous Anderson Hamiltonian, first proved in

Cv21: to the magnetic case <ref:2511.22162#pg5>, five or more words copied word for word from that page.

Ada: Operator Construction and Formal Solution The paper introduces a random Schrödinger operator of the form H:= (i∇ + A)2 + V (1.1) defined on a two-dimensional bounded domain <ref:2511.22162#pg2>.

Emmy: This operator arises from quantising the Hamiltonian of a single-particle system subjected to an electric field −∇V and the magnetic field ∇ × A <ref:2511.22162#pg3>.

Ada: The construction involves defining a regularised operator Hε:= (i∇ + A)2 + ξ(ε) + C(ε), which is well-defined and self-adjoint on any bounded open box B with zero Dirichlet boundary condition <ref:2511.22162#pg5>.

Title and authors: Ada: So, what they do in this construction is they define a regularised operator H epsilon equals i nabla plus A squared plus ξ epsilon plus C epsilon, which is well defined and self-adjoint on any bounded open box B with zero Dirichlet boundary condition.

Emmy: That sounds like a lot of technical machinery. They are using the exponential transformation trick first introduced in HL15 and the Hille–Yosida theory to pull this desired operator out from the solution theory of parabolic stochastic partial differential equations.

Kolya: So, they are not just putting a random function V in there; they are using these advanced functional analysis tools to extract the specific operator they need for their study. What does that mean for the actual physics?

Ada: It means this construction lets them handle situations where the magnetic field is not perfectly smooth, like having very rough noise, which is what Gaussian white noise on an electric potential implies.

Emmy: And page three of Landau Hamiltonian with Gaussian white noise potential and the asymptotics of its bottom of spectrum explains that they use a spectral expansion of H to heuristically approximate the large time behavior as U B t equals e minus t lambda B t one, where lambda B t one is the lowest eigenvalue <ref:2511.22162#pg1,Landau Hamiltonian with Gaussian white noise potential and the asymptotics of its>.

Kolya: So, by looking at this lowest eigenvalue over time in a growing box, they can figure out how it behaves as that box size increases. That’s a key piece of information for understanding the spectrum.

Ada: Precisely; understanding how that lowest eigenvalue scales as the box size increases becomes crucial for figuring out the asymptotics of its growth.

The paper's summary: Emmy: Now let's talk about what they suggest improving or extending from this construction. They are looking at a few natural extensions, like moving to dimension three, or changing to Neumann boundary conditions, or even dealing with a white noise magnetic field.

Ada: That’s the point; they are exploring different physical scenarios. Dimension three is a big step because it tests if the logic holds up in higher dimensions.

Kolya: And they also mention that there's a structural obstruction preventing them from doing this construction on the full space, which is R two, because those magnetic terms grow superlinearly as you go to infinity.

Emmy: That means their method has limits; it works beautifully in bounded boxes with certain conditions, but extending it to infinite space where the magnetic field doesn't decay smoothly is not possible with this setup.

Ada: The paper suggests that they can preserve the Dirichlet boundary condition using the exponential transformation, while for Neumann conditions, they can adapt localized Green functions.

Kolya: That's interesting from a practical standpoint because boundary conditions are fundamental constraints in modeling physical systems. So, even on a bounded domain like a box, this construction is versatile.

The paper's improvements: Ada: So to wrap up the Landau Hamiltonian with Gaussian white noise potential and the asymptotics of its bottom of spectrum paper, they've shown this simple construction works for operators with magnetic fields that are rougher than zero minus Hölder regularity and on a two dimensional bounded box.

Emmy: And they’ve successfully generalized an asymptotic result from the standard continuous Anderson Hamiltonian to this magnetic case, showing that the lowest eigenvalue scales as lambda L n equals minus C G N log L as L approaches plus infinity under assumption one point three <ref:2511.22162#pg1>.

Kolya: So what this means for us is that we have a way to predict the energy level distribution in large regions or samples where the magnetic field changes randomly. It helps us understand how localization behaves when translation symmetry is broken by a magnetic field.

Ada: That's the big picture; it gives us tools to analyze complex quantum systems where both material disorder and external fields are stochastic, moving beyond just translation-invariant models.

Emmy: And for the future, they are looking at how this connects to larger problems in physics, like dimension three or white noise magnetic fields on different spaces.

Kolya: I’m just glad they could nail down that scaling law for the bottom of the spectrum; getting that precise relationship between the size of the box and the lowest energy level is really valuable data for transport studies.

Ada: Absolutely, it's a solid piece of work, providing a rigorous way to handle these singular cases in random Hamiltonian problems.

Conclusion: Ada: So we've been talking about this paper on the Landau Hamiltonian with Gaussian white noise potential and the asymptotics of its bottom of its spectrum, which shows how we can construct these operators even when the magnetic field is rougher than zero minus Hölder regularity.

Emmy: Exactly. The main structure here is using that exponential transformation trick and Hille–Yosida theory to pull out the desired operator from some stochastic partial differential equations. It's a very elegant way to extract what we need, even when the noise isn't perfectly smooth.

Kolya: I think what really stands out is how they generalize an asymptotic result for the bottom of the spectrum of a two dimensional continuous Anderson Hamiltonian over to this magnetic case. It shows that even when translation invariance is broken by a magnetic field, you can still get that scaling relationship as you look at larger and larger boxes.

Ada: Right, so we're talking about something that used to be proven for simpler models now holding up under these more complex conditions with rougher fields. What this implies for the world is that AI systems might be able to simulate quantum phenomena in environments where both material impurities and external fields are random and noisy.

Emmy: And the construction method itself, using a semigroup approach, is pretty powerful. It gives us an explicit description of the operator domain and its associated semigroup based directly on the underlying noise structure. It’s about generating physical processes straight from the noise input.

Kolya: For me, what this construction does best is give us tools to handle those extreme energy levels in disordered systems by proving a large deviation principle. That means we can get a good estimate for how often you'll see those very high or very low energies in the system.

Ada: That LDP with rate beta squared and that rate function I B of x gives us precision for predicting rare events, which is crucial for understanding transport phenomena and localization effects in these complex media.

Emmy: And they also showed that you can keep the Dirichlet boundary condition using that exponential transformation, while Neumann conditions can be handled by localizing green functions. That means this method is pretty versatile for modeling systems with realistic physical constraints like insulated boundaries or flux constraints.

Kolya: But they also have to be careful because they flag a structural obstruction; you can't actually construct this operator on the whole space R two because the magnetic terms grow too fast at infinity. That's a limitation we have to remember.

Ada: So, in summary, this paper on the Landau Hamiltonian with Gaussian white noise potential and the asymptotics of its bottom of its spectrum shows a robust way to build these operators using advanced functional analysis techniques for rough magnetic fields on bounded domains.

Emmy: It generalizes known results and gives us a tool to extract operators directly from stochastic partial differential equations, which is pretty neat structure-wise.

Kolya: And it provides the large deviation principle needed to estimate those extreme eigenvalues, even with the caveats about the space being bounded.

Ada: So for anyone interested in how AI models can handle complex quantum environments with both material disorder and magnetic noise, this paper offers a solid construction method to start with.

TU Wien

math.PR, math-ph, math.MP, math.SP

Submitted: 2025-11-27

Updated: 2026-11-04

Comments: Accepted version for publication in Electron. J. Probab

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Importance score: 84/100

The gist: We present a simple construction of a random Schrödinger operator subject to a magnetic field with a regularity as low as 0−Hölder and a Gaussian white noise electric potential on a…

Key concepts

Random Schrödinger Operator
This is a mathematical model representing quantum systems where both material impurities (disorder) and external fields (like magnetic fields or electric potentials) are random. The operator describes how particles behave within this complex, noisy environment, moving beyond simpler models.
Gaussian White Noise Potential
This refers to the random electric potential added to the system, which is modeled by Gaussian white noise. This type of noise is very irregular and represents a highly fluctuating external field that significantly impacts the quantum particle's energy levels and wave function.
Asymptotic Behavior of Eigenvalues
This concept studies how the energy levels (eigenvalues) of the system change as the size of the confining box increases. The paper proves that for large boxes, these eigenvalues follow a specific logarithmic scaling law, which is crucial for understanding how disorder affects quantum transport in large systems.

Terminology

Summary

We present a simple construction of a random Schrödinger operator subject to a magnetic field with a regularity as low as 0−Hölder and a Gaussian white noise electric potential on a two-dimensional bounded box This construction enables us to generalise an asymptotic result for the bottom of the spectrum of the two-dimensional continuous Anderson Hamiltonian, first proved in [Cv21] to the magnetic case <ref:2511. p. 022162#pg5>, five or more words copied word for word from that page.

Operator Construction and Formal Solution

The paper introduces a random Schrödinger operator of the form H:= (i∇ + A)2 + V (1.1) defined on a two-dimensional bounded domain <ref:2511. p. 022162#pg2>. This operator arises from quantising the Hamiltonian of a single-particle system subjected to an electric field −∇V and the magnetic field ∇ × A <ref:2511. p. 022162#pg3>. The construction involves defining a regularised operator Hε:= (i∇ + A)2 + ξ(ε) + C(ε), which is well-defined and self-adjoint on any bounded open box B with zero Dirichlet boundary condition <ref:2511. p. 222162#pg5>. This construction uses the exponential transformation trick first introduced in [HL15] and the Hille–Yosida theory to extract the desired operator from the solution theory of parabolic SPDEs <ref:2511. p. 222162#pg6>.

Semigroup Properties and Spectral Results

The construction leads to a semigroup (Pt) of bounded operators which is strongly continuous in the time variable <ref:2511.22162#pg10>. The map P B t(A, q)g = e−hv(t), where v is the solution to the associated PDE <ref:2511. p. 122162#pg11>. This operator satisfies several important properties, including being bounded on L2(B) and having a compact resolvent <ref:2511. p. 222162#pg12>. The generator Φ B(A, q) is self-adjoint and its spectrum is contained in (−ω B(A, q),∞) <ref:2511. p. 122162#pg13>.

Asymptotic Behavior of Eigenvalues

The second main result concerns the asymptotics of the eigenvalues λ B n as the size of the box B grows to infinity <ref:2511.22162#pg5>. Under Assumption 1. p. 13, it is claimed that for every n ≥ 1, one has almost surely that λ L n ∼ −CGN log L as L → +∞ <ref:2511. p. 322162#pg6>. This result extends the previous work to the magnetic case where the operator is no longer translation invariant <ref:2511. p. 322162#pg7>.

Large Deviation Principles

The paper establishes a Large Deviation Principle (LDP) for the eigenvalues of H, stating that for all n ≥ 1, one has a LDP with rate β squared and a good rate function I B(x) = inf 1/2 φ squared L2(B), for q = (G∗φ, −∇G∗φ squared − F∗φ) ∈ M <ref:2511.22162#pg20>. This LDP is crucial for proving the tail bounds claimed in Theorem 3. p. 41 <ref:2511. p. 022162#pg6>.

Generalisations and Open Questions

The authors discuss several natural extensions, including dimension three, Neumann boundary conditions, and a white noise magnetic field <ref:2511.22162#pg24>. The structural obstruction prevents a construction on the full space R2 because the magnetic terms grow superlinearly at infinity <ref:2511.22162#pg25>.

References

[AC15] R. Allez and K. p. 1 Chouk. The continuous Anderson Hamiltonian in dimension two

[BCD11] H. Bahouri,J.-Y. Chemin, andR. Danchin. Fourier Analysis and Nonlinear Partial Differential Equations, vol 343 of Grundlehren Der Mathematischen Wissenschaften

[BDM25] I. Bailleul, N. V. Dang, and A. Mouzard. Analysis of the Anderson operator

[Cv21] K. Chouk and W. p. 1 van Zuijlen. p. 2 p. 2 p. 2 Asymptotics of the eigenvalues of the Anderson Hamiltonian with white noise potential in two dimensions

[DZ10] A. Dembo and O. Zeitouni. Large Deviations Techniques and Applications, vol 38 of Stochastic Modelling and Applied Probability

[EN00] K.-J. Engel and R. Nagel. One-Parameter Semigroups for Linear Evolution Equations, vol 194 of Graduate Texts in Mathematics

[GH19] M. Gerencsér and M. Hairer. Singular SPDEs in domains with boundaries

[GIP15] M. Gubinelli, P. Imkeller, and N. Perkowski. Paracontrolled distributions and singular PDEs

[GK00] J. Gärtner and W. König. Moment asymptotics for the continuous parabolic Anderson model

[GUZ20] M. Gubinelli, B. Ugurcan, and I. Zachhuber. Semilinear evolution equations for the Anderson Hamiltonian in two and three dimensions

[GY23] P. Ghosal and J. p. 2 Yi. Fractal geometry of the PAM in 2D and 3D with white noise potential

[Hai14] M. Hairer. A theory of regularity structures

[HL15] M. Hairer and C. Labbé. A simple construction of the continuum parabolic Anderson model on R2 <ref:2511. p. 222162#pg5>

[HL17] M. Hairer and C. Labbé. The reconstruction theorem in Besov spaces

[HL18] M. Hairer and C. Labbé. Multiplicative stochastic heat equations on the whole space <ref:2511. p. 422162#pg5>

[HL23] Y.-S. Hsu and C. Labbé. Asymptotic of the smallest eigenvalues of the continuous Anderson Hamiltonian in d ≤ 3

[HL26] Y.-S. Hsu and C. Labbé. Construction and spectrum of the Anderson Hamiltonian with white noise potential on R2 and R3

[KPv22] W. König, N. Perkowski, and W. p. 2 van Zuijlen. Longtime asymptotics of the two-dimensional parabolic Anderson model with white-noise potential

[Lab19] C. Labbé. The continuous Anderson Hamiltonian in d ≤ 3

[Led90] M. Ledoux. A note on large deviations for Wiener chaos

[LL01] E. H. Lieb and M. Loss. Analysis. No v 14 in Graduate Studies in Mathematics

[MM22] L. Morin and A. p. 1 Mouzard. 2D random magnetic Laplacian with white noise magnetic field <ref:2511. p. 022162#pg7>

[Mou22] A. Mouzard. Weyl law for the Anderson Hamiltonian on a two-dimensional manifold

[Mv25] T. Matsuda and W. van Zuijlen. Anderson Hamiltonians with singular potentials

[Uek23] N. Ueki. A definition of self-adjoint operators derived from the Schrödinger operator with the white noise potential on the plane, 2023

[Ugu22] B. E. Ugurcan.

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. Enhanced modeling of disordered media for quantum systems by constructing random Schrödinger operators with magnetic fields of arbitrary regularity up to 1 minus kappa, which includes Gaussian white noise perturbations and non-uniform magnetic fields that break translational invariance. This allows AI systems to simulate quantum phenomena in complex environments where both material impurities and external fields are stochastic, going beyond the limitations of translation-invariant models like the standard Anderson Hamiltonian.

  2. Rigorous analysis of spectral asymptotics for random Hamiltonians by extending results from Anderson Hamiltonians to the magnetic case, proving that eigenvalues scale as lambda L n-CGN log L as L -> +∞ under specific assumptions on the magnetic field regularity. This capability enables AI systems to predict the energy level distribution in large, disordered regions or samples with varying magnetic field strengths.

  3. Development of a new construction method for random operators using a semigroup approach based on exponential transformation and Hille-Yosida theory, which provides an explicit description of the operator domain and its associated semigroup. This allows AI systems to generate physically meaningful stochastic processes directly from the underlying noise structure, improving the fidelity of simulation methods for stochastic partial differential equations.

  4. Application of large deviation principles to eigenvalue statistics by proving that the sequence of random eigenvalues satisfies a Large Deviation Principle with rate beta squared as beta approaches zero. This enables AI systems to estimate rare events related to extreme energy levels in disordered systems with high precision, crucial for understanding transport phenomena and localization effects.

  5. Capability to analyze the spectral behavior under different physical boundary conditions by showing that Dirichlet boundary conditions are preserved by the exponential transformation, while Neumann conditions can be adapted using localized Green functions. This allows AI systems to model quantum systems on bounded domains with realistic physical constraints like insulated boundaries or flux constraints.

Sources

Related papers