Uncovering the Microscopic Mechanism of Slow Dynamics in Quasiperiodic Many-Body Localized Systems

This paper uncovers the microscopic mechanism of slow dynamics in one-dimensional quasiperiodic many-body localized systems, identifying quantum amplitude modulation of single-particle hoppings as the cause of structured growth in number entropy and quasiparticle width, thereby supporting the thermodynamic stability of the MBL phase.

The Gist

The Big Question: Why Don't Some Things "Settle Down"?

Imagine you drop a drop of ink into a glass of water. Eventually, the ink spreads out evenly, and the water turns a uniform shade of blue. In physics, this is called thermalization. It's how most things in the universe reach a state of balance.

But what if you had a magical glass of water where the ink never spread out? It would stay in a clump forever, defying the laws of thermodynamics. For a long time, physicists thought this was impossible in complex systems with many particles interacting. However, a phenomenon called Many-Body Localization (MBL) suggests that under certain conditions, particles get "stuck" and refuse to mix.

The big debate in the physics world right now is: Is this "stuck" state truly permanent, or is it just a slow-motion leak? Recent studies on random systems suggested that even in these "frozen" states, particles might slowly drift apart over incredibly long times, eventually causing the system to thaw out.

The Experiment: A Deterministic Maze

The authors of this paper decided to investigate this using a specific type of system: a Quasiperiodic (QP) system.

• Random Systems: Imagine a maze where the walls are placed by throwing darts blindfolded. It's chaotic, unpredictable, and full of weird gaps (resonances) where particles can slip through.
• Quasiperiodic Systems: Imagine a maze built with a strict, repeating mathematical pattern (like the Fibonacci sequence). It looks messy and non-repeating to the eye, but it is perfectly deterministic. There are no random "weak spots."

The researchers wanted to see if particles in this "mathematical maze" would also slowly leak out, or if the strict order of the maze would keep them locked in place forever.

The Discovery: A "Slow Dance" of Particles

They found that yes, the particles do move slowly, but not in the chaotic way seen in random systems. Instead, they exhibit a structured, rhythmic growth.

To understand why this happens, the authors looked at the microscopic mechanism. They discovered a phenomenon they call Amplitude Modulation (AM).

The Analogy: The Swing and the Whisper

Imagine two children on swings (representing two pairs of particles) in a large park.

1. The Swing: Each pair of particles is trying to swap places (hop from one spot to another). This is like a child swinging back and forth.
2. The Whisper: In a random system, the swings are independent. But in this quasiperiodic system, the two pairs of swings are connected by a very faint, invisible string (interaction).

Even though the particles are far apart, the "whisper" of one pair affects the other. This interaction doesn't make them swap places instantly. Instead, it modulates the rhythm of their swinging.

• The Beat: Think of two musical notes played slightly out of tune. You hear a "wah-wah-wah" sound (beats) where the volume rises and falls.
• The Result: The particles start hopping, stop, start again, and stop. This "beating" effect causes the particles to spend more time in a "delocalized" state (spread out) than they would if they were just hopping randomly.

This modulation is the key. It's not a sudden explosion of movement; it's a slow, rhythmic swelling of the particles' ability to move, driven by the specific, ordered structure of the maze.

The Two Types of Growth

The paper identifies two distinct phases of this movement:

1. The Short-Term (The Soloist): At first, particles hop between their immediate neighbors. This is like a solo dancer moving across the floor. This happens quickly and is predictable.
2. The Long-Term (The Duet): Over very long times, the "beats" (the amplitude modulation) kick in. The particles start interacting with partners far away. This causes a slow, steady increase in the "number entropy" (a measure of how mixed up the particles are).

Crucially, in the quasiperiodic system, this long-term growth is structured. It happens in distinct steps, like a staircase, rather than a messy, random climb.

Why This Matters: The System is Safe

The most important conclusion of the paper is about stability.

• The Fear: If particles keep moving forever, the "frozen" MBL state is unstable, and the system will eventually thermalize (melt).
• The Reality: The authors show that this slow movement is caused by local processes (the "duet" between specific pairs). It doesn't require particles to travel across the whole system.
• The Verdict: Because the movement is driven by local interactions and is bounded (it can't grow infinitely), the MBL phase remains stable. The system is truly "frozen" in the long run, just with a very slow, rhythmic shivering.

Summary in a Nutshell

Think of the Many-Body Localized system as a crowded dance floor where everyone is supposed to stay still.

• In a random crowd, people might accidentally bump into each other and start a chaotic, slow drift that could eventually fill the whole room.
• In this quasiperiodic crowd, the layout is mathematically perfect. People still move, but they do it in a synchronized, rhythmic "dance" (Amplitude Modulation). They sway back and forth, but they never actually leave their spot in the long run.

The paper proves that this "dance" is a natural, local feature of quantum systems and confirms that the "frozen" state is robust and stable, even over cosmic timescales.

Technical Summary

1. Problem Statement

The paper addresses a central debate in statistical physics regarding the stability of the Many-Body Localized (MBL) phase in the thermodynamic limit.

• The Conflict: While MBL was originally proposed as a robust mechanism preventing thermalization in closed systems, recent numerical studies in random disordered systems have observed a slow, potentially unbounded growth of number entropy ($S_n$) even deep in the MBL regime. This growth suggests slow particle transport, threatening the stability of the MBL phase.
• The Gap: Previous explanations for this growth in random systems relied on "rare regions" or local single-particle resonances. However, the microscopic mechanism driving this slow dynamics remained unclear. Furthermore, it was unknown if this phenomenon persists in quasiperiodic (QP) systems, which are deterministic and lack the rare weak-potential regions found in random systems, potentially making MBL more robust.

2. Methodology

The authors investigated one-dimensional quasiperiodic MBL systems using a combination of numerical simulations and analytical modeling.

• Model System:
  • A spin-1/2 chain described by the XXZ Hamiltonian with an Aubry-André quasiperiodic potential ($h_i = W \cos(2\pi\beta i + \phi)$).
  • Parameters: Isotropic Heisenberg interaction ($J_z = -1$), deep localized regime ($W \geq 6$), and spatial frequency $\beta$ set to the inverse of the golden mean (unless varied to tune distances).
  • Initial states: Primarily the Néel state (half-filling sector).
• Numerical Techniques:
  • Full Exact Diagonalization (ED): Performed on chains up to $N=16$ sites.
  • Observables:
    • Number Entropy ($S_n$): Measures particle number fluctuations in a subsystem.
    • Configurational Entropy ($S_c$): The remaining part of entanglement entropy ($S_e = S_n + S_c$).
    • Quasiparticle Width ($\sigma_x^2$): Measures the spatial spread of excitations.
  • Averaging: Results were averaged over potential realizations (random phase $\phi$) and initial states.
• Analytical Approach:
  • Development of an effective model based on a 4-site subspace to describe the interaction between two pairs of near-resonant hopping sites.
  • Generalization of single-particle resonance models to account for various hopping ranges.

3. Key Contributions and Mechanism

The paper's primary contribution is the identification and explanation of a specific microscopic mechanism driving slow dynamics in MBL systems, termed Amplitude Modulation (AM) of Rabi oscillations.

• The Mechanism:
  • In the MBL regime, dynamics are dominated by single-particle hoppings between sites with small potential differences.
  • The authors identify that interactions between two distinct single-particle hopping processes (e.g., a particle hopping between sites $a \leftrightarrow b$ and another between $c \leftrightarrow d$) act as a mediator.
  • This interaction causes a small energy correction ($\epsilon$) to the eigenvalues of the effective 4-site Hamiltonian.
  • Result: This energy shift leads to beats (amplitude modulation) in the quantum oscillations of the particles. The amplitude of the oscillation across the subsystem boundary is modulated over long timescales ($t \sim 1/\epsilon$).
  • Consequence: The particle spends more time delocalized across the boundary during the "beats," leading to an increase in the time-averaged number entropy $S_n$.
• Distinction from Random Systems:
  • Unlike random systems where resonances are rare and unpredictable, QP systems are deterministic.
  • The QP potential creates structured correlations, leading to structured growth in $S_n$ (distinct peaks and plateaus) rather than the featureless growth seen in random systems.
  • The mechanism is generic: It relies on the interaction between local hopping processes and does not require "rare regions" of weak disorder.

4. Key Results

A. Dynamics of Number Entropy ($S_n$)

• Slow Growth: $S_n$ exhibits slow growth deep in the MBL regime ($t$ up to $10^5$), confirming that localization is not "perfect" on finite timescales.
• Structured Behavior: In QP systems, $S_n(t)$ shows distinct regimes:
  1. Short/Medium Times: Dominated by single-particle dynamics (nearest-neighbor and next-nearest-neighbor hoppings). The behavior is structured due to deterministic resonances.
  2. Long Times: Growth driven by the AM mechanism. The growth rate depends on the distance ($d$) between the interacting hopping pairs.
  3. Ultra-Long Times: $S_n$ saturates to a size-dependent plateau, consistent with the absence of transport in the thermodynamic limit.
• Saturation Discrepancy: A crucial finding is that $S_n$ saturates much earlier than the configurational entropy $S_c$.
  • $S_n$ saturates when the AM timescale ($1/\epsilon$) is reached.
  • $S_c$ continues to grow logarithmically until the entanglement spreads across the entire chain (timescale $\sim e^N$).
  • This decoupling proves that the $S_n$ growth is a local many-body effect, not a sign of global thermalization.

B. Validation of the Effective Model

• The authors constructed an analytical model combining single-particle hopping (Eq. 4) and the AM mechanism (Eq. 2).
• Agreement: The model perfectly reproduces the numerical ED results for $S_n(t)$ and the quasiparticle width $\sigma_x^2$ across all timescales.
• Parameter $\epsilon$: The correction term $\epsilon$ scales exponentially with the distance $d$ between the hopping pairs ($\epsilon \sim e^{-d}$), explaining why the effect appears at very long times in large systems.

C. Stability of the MBL Phase

• The growth of $S_n$ is driven by local processes (interactions between nearby hopping pairs) and does not imply long-range transport.
• The saturation of $S_n$ at a value consistent with the system size ($S_n \leq \ln 2$ for a single particle crossing) supports the stability of the MBL phase in the thermodynamic limit.

5. Significance

• Resolution of the Debate: The paper provides a microscopic explanation for the slow growth of number entropy, arguing that it is a generic feature of interacting localized systems (both QP and random) driven by local amplitude modulation, rather than evidence of MBL instability.
• Universality: The mechanism is shown to be generic to many-body quantum systems, extending beyond specific LIOM (Local Integrals of Motion) constructions to microscopic Hamiltonians.
• Experimental Relevance: Since $S_n$ is an experimentally accessible quantity (via quantum gas microscopes), the predicted structured growth and saturation behavior offer clear signatures for experimental verification of MBL stability and the nature of slow dynamics.
• Differentiation: It highlights the fundamental difference between random and quasiperiodic systems: the deterministic nature of QP potentials leads to structured, predictable dynamics, whereas random systems exhibit stochastic, sample-dependent behavior.

In conclusion, the authors successfully uncover that the "slow dynamics" in MBL systems are not a sign of thermalization but a consequence of quantum beats arising from the interaction of local hopping processes, a mechanism that preserves the stability of the MBL phase.