🔬 condensed matter

Sticky eigenstates in systems with sharply divided phase space

This paper demonstrates that in systems with sharply divided phase space, the fraction of "sticky" eigenstates scales as $\hbar^{1/2}$ due to classical stickiness from marginally unstable or quasi-periodic orbits, thereby generalizing established KAM hierarchy predictions to non-KAM systems.

Original authors: Hua Yan

Published 2026-09-29

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Hua Yan

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

In the quantum world, the rules of motion are often a strange blend of order and chaos. Imagine a billiard table where the balls sometimes follow predictable, repeating paths, and at other times, they bounce around in a completely random, unpredictable frenzy. For decades, physicists have believed that as the quantum world gets closer to the size of everyday objects, these two behaviors should separate cleanly. The theory suggests that quantum states—essentially the possible configurations of a system—should settle into either perfectly ordered patterns or fully chaotic spreads, leaving no middle ground. This idea, known as the principle of uniform semiclassical condensation, implies that the messy, in-between states should eventually vanish. However, reality in complex systems is rarely so tidy. There are regions where the boundary between order and chaos is not a sharp line but a sticky zone, where particles get trapped for a long time, lingering in a limbo that defies simple classification. Understanding how quantum systems behave in these sticky zones is crucial because it reveals how the predictable laws of the classical world emerge from the probabilistic nature of the quantum realm.

A recent study by Hua Yan at the University of Maribor investigates exactly these lingering states in systems where the phase space—the map of all possible positions and speeds—is sharply divided. The researcher focused on two specific mathematical models that act like simplified billiard tables. In these models, the boundary between the orderly region and the chaotic region is not a complex, fractal web of islands, as seen in many natural systems, but a simple, clean curve. This sharp division is created by specific types of orbits: some are marginally unstable periodic orbits, which are paths that repeat but are just barely stable enough to hold onto passing particles, and others are quasi-periodic orbits, which never quite repeat but stay within a specific zone. The goal was to see what happens to the quantum states when they encounter these sharp, sticky boundaries.

The researchers used a method to visualize the quantum states on these maps, creating a kind of heat map that shows where a particle is likely to be found. They then applied two specific tools to measure these states. The first tool measured how much a state overlapped with the chaotic region versus the regular region. The second tool measured how spread out or concentrated the state was. By combining these measurements, the team could sort the quantum states into three distinct categories. The first category included states that were purely regular, staying locked in the orderly zone. The second category included states that were purely chaotic, spreading out across the entire chaotic sea. The third, and most interesting, category consisted of mixed states that seemed to exist in both regions at once.

The study found that while most quantum states eventually settle into being either fully regular or fully chaotic as the system gets larger, a significant number of these mixed states persist. The researchers discovered that the number of these mixed states decreases as the system size increases, but it does so in a very specific way. They found that the fraction of these mixed states follows a power-law decay, meaning it shrinks at a steady, predictable rate. This behavior is a direct quantum signature of the "stickiness" found in the classical version of the system. In the classical world, particles get stuck near the boundary for a long time before escaping, and the time they stay stuck follows a specific mathematical pattern. The quantum study showed that the number of sticky quantum states follows the exact same pattern as the time particles spend stuck in the classical version.

To understand why these mixed states exist, the team used a computer model based on random matrix theory, a statistical tool used to describe complex systems. This model allowed them to simulate what would happen if the only force mixing the states was a phenomenon called dynamical tunneling, where a particle jumps between regions that are classically forbidden to cross. The simulation showed that if tunneling were the only cause, the number of mixed states would drop off extremely fast, vanishing almost immediately as the system grew. However, the actual results from the sharp-boundary maps showed a much slower, algebraic decay. This difference proved that the mixed states were not just a result of simple tunneling. Instead, the dominant source of these mixed states was the "sticky" nature of the boundary itself. The quantum states were getting trapped near the edge, lingering in the transition zone just like their classical counterparts.

The researchers also identified a subtle effect related to how these states are measured. Because the quantum states are represented by "bumps" of probability that have a finite size, states that are technically regular but sit right on the edge can appear to be mixed simply because the measurement tool spills over into the chaotic region. By carefully separating these edge cases from the truly sticky states, the team confirmed that the sticky states themselves follow a precise scaling law. For the systems with marginally unstable periodic orbits, the number of sticky states scales with the square root of the Planck constant, a fundamental number in quantum physics. For the systems with quasi-periodic boundaries, the number of sticky states oscillates around this same scaling law.

This work provides a clear link between the classical behavior of particles getting stuck near boundaries and the quantum behavior of eigenstates. It shows that even in systems with sharp, simple boundaries, the quantum world retains a memory of the classical stickiness. The findings suggest that the "hierarchical" structures often cited in more complex systems are not the only way to create these lingering quantum states. Instead, simple, sharp boundaries formed by specific types of orbits are sufficient to generate a population of sticky eigenstates that scale in a predictable manner. This insight helps refine our understanding of how quantum systems approach the classical limit and offers a new way to identify and classify states that do not fit neatly into the categories of order or chaos. The results confirm that the transition from quantum to classical is not just a matter of states becoming purely one or the other, but involves a persistent, measurable population of states that remain stuck in the middle, governed by the same rules that trap classical particles.

Technical Summary: Sticky Eigenstates in Systems with Sharply Divided Phase Space

Problem Statement
The paper investigates the behavior of quantum eigenstates in systems with mixed phase space, specifically focusing on non-Kolmogorov-Arnold-Moser (non-KAM) systems where the boundary between regular and chaotic regions is sharply divided. While the Principle of Uniform Semiclassical Condensation (PUSC) predicts that eigenstates in the semiclassical limit ( $\hbar \to 0$ ) should be either purely regular or purely chaotic, the approach to this limit involves complex mechanisms such as dynamical tunneling and classical stickiness. In traditional KAM systems, "hierarchical states" localized near cantori-mediated boundaries are known to scale as $\hbar^{1-1/\gamma}$ , where $\gamma$ characterizes the algebraic decay of classical recurrence-time distributions (RTDs). However, it remains unclear whether an analogous phenomenon exists in non-KAM systems, such as piecewise-linear maps, where boundaries are formed by marginally unstable periodic orbits (MUPOs) or quasi-periodic orbits rather than hierarchical structures.

Methodology
The author analyzes two specific area-preserving, discrete-time piecewise-linear maps defined on a two-dimensional torus:

A continuous sawtooth map.
A map with a more complex piecewise-linear function.

These maps serve as linearized approximations of the standard map, designed to create a phase space where regular and chaotic regions are separated by simple curves (MUPOs or quasi-periodic orbits) rather than hierarchical cantori. The study employs the following methodological framework:

Canonical Quantization: The maps are quantized on a torus, resulting in finite-dimensional Floquet models where the Hilbert space dimension $N = (2\pi\hbar)^{-1}$ .
Phase-Space Representation: Eigenstates are analyzed using Husimi functions, which project eigenvectors onto coherent states.
Classification Metrics: Two specific measures are combined to classify eigenstates:
- Overlap Index ( $\omega$ ): A measure of the degree to which a Husimi distribution overlaps with the chaotic ( $+1$ ) versus regular ($-1$) regions.
- Entropy Localization Length ( $l$ ): Derived from the Wehrl entropy, this quantifies the spatial extent of the eigenstate in phase space.
Random Matrix Modeling: A block-structured random matrix Hamiltonian is constructed to model dynamical tunneling between regular and chaotic subspaces, allowing for the isolation of tunneling effects from stickiness.
Classical Analysis: Lagrangian descriptors (LDs) and Smaller Alignment Index (SALI) plots are used to characterize the classical transport barriers and stickiness near the boundaries.

Key Contributions and Results

Classification of Mixed Eigenstates: The study identifies three distinct types of mixed eigenstates in these non-KAM systems:
- Regular states that leak into the chaotic region due to dynamical tunneling or the finite width of the coherent state (Gaussian bump) used in the Husimi representation.
- Chaotic states that flood into the regular region via tunneling.
- Sticky Eigenstates: States localized near the sticky boundaries formed by MUPOs or quasi-periodic orbits.
Scaling of Mixed Eigenstate Fraction: The total fraction of mixed eigenstates ( $f_M$ ) decays algebraically with the effective Planck constant, $f_M \sim \hbar^\eta$ . However, the random matrix model reveals that the contribution from pure dynamical tunneling (types i and ii) scales exponentially as $\sim \hbar \exp(-b/\hbar)$ . Therefore, the observed algebraic decay is dominated by the sticky eigenstates (type iii) and regular states misclassified as mixed due to finite-size effects.
Scaling of Sticky Eigenstates: By filtering out regular states (using the criterion $\omega \lesssim 0$ ), the author isolates the sticky eigenstates.
- For maps where boundaries are formed by MUPOs, the fraction of sticky eigenstates ( $f_{\text{sticky}}$ ) scales as $\hbar^{1-1/\gamma}$ with $\gamma = 2$ , yielding a scaling of $\hbar^{1/2}$ . This matches the classical RTD exponent where cumulative recurrence times decay as $Q(t) \sim t^{-\gamma}$ .
- For maps where boundaries are formed by quasi-periodic orbits, the fraction $f_{\text{sticky}}$ oscillates around the same algebraic behavior ( $\hbar^{1/2}$ ), consistent with the oscillatory nature of the classical RTD in this case.
Generalization of KAM Predictions: The results generalize the established prediction for hierarchical states in KAM systems ( $\hbar^{1-1/\gamma}$ ) to non-KAM systems with sharply divided phase space. The study demonstrates that the exponent $\gamma$ in the quantum scaling is directly linked to the classical stickiness exponent, even in the absence of hierarchical structures.

Significance
The paper claims to reveal a clear quantum signature of classical stickiness in non-KAM systems. By combining the overlap index with the entropy localization length, the author provides a quantitative method to distinguish between eigenstates affected by dynamical tunneling and those dominated by classical stickiness. The findings suggest that the power-law decay of mixed eigenstates in these systems is a direct consequence of the algebraic decay of classical recurrence times near MUPOs or quasi-periodic boundaries. This work extends the understanding of how classical transport structures influence quantum eigenstates beyond the traditional KAM framework, offering insights into mechanisms behind nonthermal eigenstates and deviations from full ergodicity. The author notes that while the scaling relation holds, the lack of hierarchical structures in these systems poses an open question regarding the existence of an effective theory explaining the closeness between the quantum and classical exponents without relying on self-similar Markov tree models.

More like this