[Paper Review] Accurate semiclassical spectral asymptotics for a two-dimensional magnetic Schrödinger operator

Bernard Helffer, Yuri A. Kordyukov|arXiv (Cornell University)|Dec 19, 2013

Spectral Theory in Mathematical Physics1 references4 citations

TL;DR

This paper establishes accurate semiclassical spectral asymptotics for a two-dimensional magnetic Schrödinger operator with a strictly positive, non-degenerate magnetic field minimum. Using a refined effective Hamiltonian approach via $h$-pseudodifferential operators and Grushin problems, it proves that eigenvalues in $[hb_0, h(b_0 + \gamma_0(h)))]$ correspond to approximate zero eigenfunctions of a semiclassical symbol $p_{\text{eff}}(y,\eta,h,z)$, extending prior results beyond the ground state to a full interval of eigenvalues with controlled error bounds.

ABSTRACT

We revisit the problem of semiclassical spectral asymptotics for a pure magnetic Schrödinger operator on a two-dimensional Riemannian manifold. We suppose that the minimal value $b_0$ of the intensity of the magnetic field is strictly positive, and the corresponding minimum is unique and non-degenerate. The purpose is to get the control on the spectrum in an interval $(hb_0, h(b_0 +γ_0)]$ for some $γ_0>0$ independent of the semiclassical parameter $h$. The previous papers by Helffer-Mohamed and by Helffer-Kordyukov were only treating the ground-state energy or a finite (independent of $h$) number of eigenvalues. Note also that N. Raymond and S. Vu Ngoc have recently developed a different approach of the same problem.

Motivation & Objective

• To extend semiclassical spectral asymptotics beyond the ground state energy to a full interval of eigenvalues in the presence of a non-degenerate minimum of the magnetic field intensity.
• To develop a rigorous effective Hamiltonian $p_{\text{eff}}(y,\eta,h,z)$ that captures the spectral behavior in the interval $(hb_0, h(b_0 + \gamma_0))$ for a fixed $\gamma_0 > 0$ independent of $h$.
• To provide a systematic framework for computing spectral asymptotics using $h$-pseudodifferential operators and formal Weyl composition laws.
• To overcome technical limitations of prior approaches by introducing a Grushin problem formulation with controlled error terms in $h$.

Proposed method

• Formal construction of a semiclassical symbol $p_{\text{eff}}(y,\eta,h,z) \sim \sum_{j\in\mathbb{N}} p_{\text{eff}}^j(y,\eta,z)h^j$ via the Weyl composition law $\#_h$ for vector-valued symbols.
• Reduction of the spectral problem to the existence of approximate $L^2$-normalized eigenfunctions for the operator $p_{\text{eff}}(y,hD_y,h,z(h))$.
• Use of a Grushin problem with operators $R_\pm$ to relate the spectral condition to the invertibility of a symbol matrix and to control error terms via $\mathcal{O}(h^\infty)$ estimates.
• Introduction of a cut-off function $\chi$ and finite truncation at order $N$ to ensure well-defined operators, followed by asymptotic analysis as $N \to \infty$.
• Establishment of a link between the formal inverse $\mathcal{E}^{(1),\infty}$ and the actual effective symbol $\epsilon_\pm^{(1),\infty}$ through formal composition laws.
• Proof of the key equivalence: $\lambda_h \in \sigma(H^h) \cap [0, h(b_0 + \gamma_0(h))$ if and only if $p_{\text{eff}}(y,hD_y,h,z(h))$ has an approximate $0$-eigenfunction $u_h^{qm}$ with $\|u_h^{qm}\| = 1$ and $\|p_{\text{eff}} u_h^{qm}\| = \mathcal{O}(h^\infty)$.

Experimental results

Research questions

• RQ1 Can the spectral asymptotics of a 2D magnetic Schrödinger operator be extended beyond the ground state to a full interval of eigenvalues near $hb_0$?
• RQ2 What is the correct effective Hamiltonian symbol that captures the full spectral behavior in the interval $(hb_0, h(b_0 + \gamma_0))$ for $\gamma_0 > 0$ independent of $h$?
• RQ3 How can the spectral condition be reformulated in terms of approximate eigenfunctions of a $h$-pseudodifferential operator with controlled error bounds?
• RQ4 What is the role of the Grushin problem and symbol composition in constructing a rigorous effective operator for spectral analysis?
• RQ5 Can the formal symbol $\epsilon_{\pm}^{(1),\infty}$ be used to analyze the spectrum beyond the bottom of the spectrum, and what are the technical obstacles?

Key findings

• The paper proves that eigenvalues $\lambda_h$ in the interval $[hb_0, h(b_0 + \gamma_0(h))$ are characterized by the existence of an approximate $L^2$-normalized eigenfunction $u_h^{qm}$ for the effective operator $p_{\text{eff}}(y,hD_y,h,z(h))$ with error $\mathcal{O}(h^\infty)$.
• The effective symbol $p_{\text{eff}}(y,\eta,h,z)$ is constructed as a formal $h$-pseudodifferential operator with principal symbol $\hat{b}(y,\eta) - b_0 - z$, and higher-order terms are determined by the Weyl composition law.
• The construction ensures that $\gamma_0(h) \to \gamma_0 > 0$ as $h \to 0$, allowing a uniform spectral window independent of $h$.
• The method establishes a spectral equivalence: $\lambda_h \in \sigma(H^h)$ if and only if $p_{\text{eff}}(y,hD_y,h,z(h))$ has an approximate zero eigenfunction, enabling full asymptotic control.
• The error estimates are quantified as $\mathcal{O}(h^{19/8})$ in the lower bound and $\mathcal{O}(h^{5/2})$ in the upper bound for individual eigenvalues, improving on prior two-term asymptotics.
• The framework allows for the theoretical computation of higher-order symbols via formal composition, though practical computation remains challenging.