The full delocalization of eigenstates for the quantized cat map

Basic Information

• Paper ID: 2103.06633
• Title: The full delocalization of eigenstates for the quantized cat map
• Author: Nir Schwartz
• Classification: math.AP math-ph math.MP math.SP
• Publication Date: March 2021 (arXiv preprint, revised May 2024)
• Paper Link: https://arxiv.org/abs/2103.06633

Abstract

This paper investigates the quantum cat map—a toy model of a quantized chaotic system. The author proves that its eigenstates are completely delocalized on the two-dimensional torus $\mathbb{T}^2$ in the semiclassical limit (equivalently, every semiclassical measure has full support on $\mathbb{T}^2$). The proof adapts analogous results from DJ18 for eigenstates of $-\Delta_g$ on compact hyperbolic surfaces, relying on the fractal uncertainty principle from BD18.

Research Background and Motivation

Problem Background

1. Quantum Chaos Theory: The study of the behavior of quantum counterparts of classical chaotic dynamical systems in the high-frequency limit is a central goal of quantum chaos.
2. Spatial Distribution of Eigenstates: For the Laplace-Beltrami operator $-\Delta_g$ on compact negatively curved surfaces, high-frequency eigenmodes are known to satisfy quantum ergodicity, but quantum unique ergodicity (QUE) remains an open problem.
3. Constraints on Semiclassical Measures: Understanding which probability measures can arise as weak* limits of eigenstate sequences (semiclassical measures) and the localization properties of these measures.

Research Motivation

1. Full Support Property: While it is known that semiclassical measures cannot be overly localized (entropy lower bounds), they could potentially be supported on proper invariant subsets. This paper proves a stronger result: semiclassical measures must have full support on the entire phase space.
2. Specificity of the Cat Map: The quantum cat map, as a toy model for discrete-time dynamics, exhibits similar dynamical properties to the continuous case but with more direct technical treatment.
3. Generalization of Methods: Extending the techniques of Dyatlov-Jin for hyperbolic surfaces to the quantum cat map on the torus.

Core Contributions

1. Main Theorem: Proves that for families $\{M_N(\gamma)\}_N$ obtained by quantizing hyperbolic matrices $\gamma \in \tilde{\Gamma}(2)$, any associated semiclassical measure $\mu_{sc}$ has full support on $\mathbb{T}^2$.
2. Eigenfunction Delocalization: As a corollary, proves that all eigenvectors $\phi_N$ are completely delocalized for sufficiently large $N$.
3. Technical Innovations:
   • Extension of anisotropic pseudodifferential calculus to the torus setting
   • Construction of partition of unity adapted to the quantum cat map
   • Application of the fractal uncertainty principle to handle "uncontrolled" operators
4. Quantitative Estimates: Provides quantitative lower bounds on the mass distribution of eigenstates.

Detailed Methodology

Problem Setup

Study the spatial distribution properties of eigenstates $\phi_N \in \mathbb{C}^N$ of the quantum cat map $M_N(\gamma): H_N \to H_N$ in the semiclassical limit $N \to \infty$, where $H_N \cong \mathbb{C}^N$ is an $N$-dimensional Hilbert space.

Core Technical Framework

1. Anisotropic Pseudodifferential Calculus

For hyperbolic automorphisms $\gamma \in SL_2(\mathbb{Z})$, define stable and unstable foliations:

• Unstable space: $W^u(\gamma) = \text{span}\{(1, m_u)^T\}$
• Stable space: $W^s(\gamma) = \text{span}\{(1, m_s)^T\}$

Introduce symbol classes $S^{L,\rho_+}(\mathbb{T}^2)$, where $L$ is a linear foliation, allowing sharp oscillations in the transverse direction:

$S^{L_u,\rho_+}(\mathbb{T}^2) = \{a \in C^\infty(\mathbb{T}^2): \forall k,m \in \mathbb{N}, \sup |V_u^m V_s^k a| = O(h^{-\rho k-})\}$

2. Exact Egorov Relations

The quantum cat map satisfies exact Egorov relations: $M_N(\gamma)^* \text{Op}_N(a) M_N(\gamma) = \text{Op}_N(a \circ \gamma)$

This allows propagation of symbols along the dynamics up to time $2T' = 8T$, where $T = \lfloor\frac{\rho \log(1/h)}{4\log|\lambda_u|}\rfloor$.

3. Partition of Unity Construction

Construct pseudodifferential partition of unity $\{a_1, a_2\}$ satisfying:

• $a_1 + a_2 = 1$
• $\text{supp}(a_1) \subset (\text{supp}(a))^\circ$
• $a_1|_{K_2} \equiv a_2|_{K_1} \equiv 1$

For each word of length $T$, $w \in W(T) = \{1,2\}^T$, define operators: $A_w = \text{Op}_N(a_{w_{T-1}} \circ \gamma^{T-1}) \cdots \text{Op}_N(a_{w_1} \circ \gamma) \text{Op}_N(a_{w_0})$

4. Separation of Controlled and Uncontrolled Regions

Define control function $F(w) = \frac{\#\{j: w_j = 1\}}{T}$, partitioning the set of long words $W(8T)$ into:

• Controlled region $Y$: contains subwords $w^{(j)} \in Z = \{w: F(w) \geq \delta\}$
• Uncontrolled region $X$: all subwords lie outside $Z$

Main Estimate Theorem

Theorem 3.2: There exist constants $C_1(a), C_2(a), N(a)$ such that for $N \geq N(a)$ and $u \in H_N$:

$\|u\|_{H_N} \leq C_1(a)\|\text{Op}_N(a)u\|_{H_N} + C_2(a)\log N \min_{|z|=1}\|(M_N(\gamma)-z)u\|_{H_N}$

Technical Innovations

1. Multiscale Analysis: Through precise choice of propagation time, balances the validity of pseudodifferential calculus with applicability of the fractal uncertainty principle.
2. Combinatorial Arguments: Uses combinatorial lemmas from DJ18 to control the number of uncontrolled words: $\#X \leq Ch^{-4\sqrt{\delta}}$.
3. Fractal Geometry Application: Proves that the support of propagated symbols has fractal structure satisfying $\nu$-porosity.

Experimental Setup

This is a purely theoretical work with no numerical experiments. All results are obtained through rigorous mathematical proofs.

Concrete Example

The author provides a specific example of the Degli-Esposti cat map: $\gamma_{DE} = \begin{pmatrix} 2 & 1 \\ 3 & 2 \end{pmatrix}$ with eigenvalues $\lambda_u = 2 + \sqrt{3}$, $\lambda_s = 2 - \sqrt{3}$.

Experimental Results

Main Results

Theorem 1.1 (Full Support of Semiclassical Measures): For $\gamma \in \tilde{\Gamma}(2)$, any semiclassical measure $\mu_{sc}$ satisfies: for every nonempty open set $\emptyset \neq \Omega \subset \mathbb{T}^2$, there exists a constant $c_\Omega > 0$ such that $\mu_{sc}(\Omega) > c_\Omega$.

Theorem 1.2 (Eigenfunction Delocalization): For normalized eigenvectors $\phi_N = (\phi_{N,k})_{k=1}^N$ and $0 \leq \alpha_1 < \alpha_2 \leq 1$, there exist $c_{\alpha_1,\alpha_2} > 0$ and $N_0$ such that for $N > N_0$:

$\frac{1}{N}\sum_{k \in [\alpha_1 N, \alpha_2 N]} |\phi_{N,k}|^2 > c_{\alpha_1,\alpha_2}$

Validity of Proof Strategy

1. Controlled Region Estimate (Proposition 3.5): $\|A_{Y,N}u\| \leq c_\delta(\gamma)(C_1\|\text{Op}_N(a)u\| + \log N \min_{|z|=1}\|(M_N(\gamma)-z)u\| + \frac{C_2}{N^{1/8}}\|u\|)$
2. Uncontrolled Region Estimate (Proposition 3.6): $\sup_{w \in W(8T)} \|A_{w,N}\| \leq \frac{C}{N^{\beta'}}$
3. Key Combinatorial Estimate: $\#X \leq Ch^{-4\sqrt{\delta}}$

Related Work

Historical Development

1. Quantum Ergodicity: Shnirelman, Zelditch, and Colin de Verdière proved asymptotic equidistribution of almost all eigenstates.
2. Entropy Constraints: Anantharaman proved positivity of semiclassical measure entropy; Anantharaman-Nonnenmacher provided entropy lower bounds.
3. Specific Results on Cat Maps:
   • Bouzouina-De Bièvre: quantum ergodicity
   • Kurlberg-Rudnick: QUE for joint eigenbases
   • Faure-Nonnenmacher-De Bièvre: counterexamples to QUE

Uniqueness of This Work's Contribution

This paper is the first to prove the full support property of semiclassical measures for the quantum cat map, filling an important gap between entropy constraints and complete delocalization.

Conclusions and Discussion

Main Conclusions

1. All semiclassical measures of the quantum cat map have full support on $\mathbb{T}^2$
2. All eigenstates are completely delocalized in the semiclassical limit
3. Provides quantitative lower bounds on eigenstate mass distribution

Limitations

1. Dimensional Restriction: Methods primarily apply to two dimensions; higher-dimensional extensions require additional techniques
2. Special Group Requirements: Requires special properties of $\gamma \in \tilde{\Gamma}(2)$
3. Constant Dependence: Constants depend on the choice of symbol

Future Directions

1. Higher-Dimensional Extensions: Dyatlov-Jézéquel have extended results to symplectic groups in arbitrary dimensions
2. Other Quantum Maps: Generalization to more general quantized hyperbolic maps
3. Explicit Constants: Improvement of constants in quantitative estimates

In-Depth Evaluation

Strengths

1. Technical Innovation: Successfully adapts continuous-case techniques to discrete quantum cat map
2. Completeness: Provides complete proof chain from fractal uncertainty principle to main results
3. Clarity: Well-structured paper with detailed technical exposition
4. Universality: Methods are important for understanding delocalization phenomena in quantum chaos

Weaknesses

1. Technical Complexity: Highly technical proof requiring deep knowledge from multiple specialized fields
2. Non-explicit Constants: While quantitative results are provided, specific constant values lack clarity
3. Limited Applicability: Primarily theoretical results with limited practical applications

Impact

1. Theoretical Contribution: Provides important theoretical results for quantum chaos theory
2. Methodological Value: Demonstrates powerful applications of fractal uncertainty principle in quantum dynamics
3. Subsequent Research: Has inspired related research in higher dimensions and other systems

Applicable Scenarios

1. Theoretical analysis of quantum chaotic systems
2. Study of eigenstates in the semiclassical limit
3. General theory of quantization of dynamical systems

References

Key references include:

• DJ18 Dyatlov-Jin's original work on full support of semiclassical measures on hyperbolic surfaces
• BD18 Bourgain-Dyatlov's fractal uncertainty principle
• BDB96 Bouzouina-De Bièvre's foundational work on quantum ergodicity of the quantum cat map