🔬 physics

Geometry-resolved avalanche criterion in disordered spin systems: transverse connectivity corrections and asymptotic recovery of the flat-front law

This study demonstrates that while the many-body localization avalanche criterion asymptotically recovers a simple area-dependent law at strong disorder, finite-width disordered spin systems exhibit systematic, geometry-dependent corrections to their stability driven by transverse connectivity and closed loops, thereby explaining the slow growth of apparent critical disorder with system width and providing a quantitative framework for interpreting MBL crossovers in higher dimensions.

Original authors: Hikaru Wakaura, Taiki Tanimae

Published 2026-07-20

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Hikaru Wakaura, Taiki Tanimae

Original paper licensed under CC BY 4.0 (https://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

Imagine a crowded dance floor where everyone is trying to find a rhythm. In the world of quantum physics, there's a special kind of party called "Many-Body Localization" (MBL). Usually, when particles interact, they eventually forget their starting moves and settle into a chaotic, thermal equilibrium—a state where everything is mixed up and random. But in MBL, if the room is messy enough (meaning there's a lot of "disorder" or random obstacles), the dancers get stuck. They remember their original steps forever, refusing to mix with the crowd. It's like a game of freeze-tag where the "frozen" players never thaw out.

Scientists have long wondered if this "frozen" state can survive forever, even in a huge room with millions of dancers. The leading theory suggests that a single, lucky spot in the room—a "thermal bubble" where the disorder happens to be weak—could act like a local DJ. If this DJ gets loud enough, they can start thawing out their neighbors, who then thaw out their neighbors, creating a chain reaction called a "quantum avalanche." This avalanche would eventually consume the whole room, melting the frozen state. For a long time, we only knew how this worked in a single-file line of dancers (one dimension). But what happens in a wide room, a 2D floor, or even a 3D ballroom? Does the shape of the room change how fast the avalanche spreads? This is the big question that determines whether the frozen state is truly safe or just waiting to melt.

The Avalanche in the Shape of a Room

In this new study, researchers Hikaru Wakaura and Taiki Tanimae decided to test the "avalanche theory" not just in a straight line, but in rooms of all different shapes. They used a supercomputer to simulate a disordered quantum system (a specific type called an XXZ spin model) and watched how fast a "thermal front" moved through it. Think of the system as a long hallway where a "heat wave" starts at one end and tries to march down the line, melting the frozen spins as it goes.

The team didn't just look at a single-file hallway. They built digital versions of:

Chains: A single line of dancers.
Ladders and Strips: Two, three, or four lines of dancers standing side-by-side.
Tubes: A 2x2 grid of lines, forming a small square tube.
Loops: Strips where the sides are connected, forming a circle (like a bracelet).

They measured how quickly the "heat wave" slowed down as it traveled. If the wave slows down too fast, the system is stable (the avalanche dies out). If it keeps going strong, the system is unstable (the avalanche wins).

The "Flat-Front" Myth vs. The "Connectivity" Reality

For a long time, physicists had a simple rule of thumb, which the authors call the "flat-front law." It assumed that the only thing that mattered was the width of the room (the cross-sectional area). The idea was: "A wider room has more people to melt, so the avalanche should be easier to stop, and the math should just depend on how many people are in a single row."

The paper explicitly rules out this simple idea. The researchers found that the shape of the room matters way more than just the number of people in a row. They discovered a "connectivity deficit."

Here is the fun part: It's not just about how wide the room is; it's about how many doors connect the rows.

In a simple ladder (two rows), the dancers can only pass the "heat" to their neighbor on the side.
In a tube or a loop, the dancers have more doors. They can pass the heat sideways, and if the room is a loop, the heat can travel all the way around.

The study found that systems with more connections (like tubes and loops) were much harder to stabilize than the simple "flat-front" math predicted. Even though a 4-wide strip and a 2x2 tube have the same number of people in a cross-section, the tube was much more unstable. The "heat wave" in the tube didn't slow down as much as the simple math said it should.

The "Deficit" Law

The authors measured exactly how much extra trouble the connections caused. They found a pattern:

More connections = Faster melting. Every extra "transverse bond" (a connection between side-by-side rows) made the avalanche spread faster.
Loops are sneaky. A strip where the sides are connected in a loop (like a bracelet) held onto its "deficit" (its extra instability) much longer than a strip with open ends. Even when the disorder was very strong, the loop geometry kept the avalanche alive, whereas the open strip started to stabilize.
The "Flat-Front" law is only a distant dream. The researchers found that the simple "area-only" rule is actually true, but only when the disorder is incredibly strong. As the disorder gets weaker (which is the interesting part where real experiments happen), the shape of the room and the number of connections dominate the physics.

What This Means for the Future

The paper suggests that previous estimates for when these systems become unstable were too optimistic. Because the "connectivity" makes avalanches spread faster than we thought, the "frozen" state (MBL) is actually more fragile in 2D and 3D than we realized.

The authors are careful to note that their results come from simulations on relatively small systems (up to 12 sites in production runs) and rely on a specific "weak-coupling" diagnostic method. They haven't proved that MBL is impossible in 2D, but they have provided the first geometry-resolved law that explains why the transition from "frozen" to "melted" happens so slowly as you widen a system.

In short: If you want to keep a quantum system frozen, don't just make it wide; make sure you don't give the particles too many doors to pass the heat around. And if you build a loop? Good luck, because the heat will find a way to travel all the way around and melt everything faster than you'd expect.

Technical Summary: Geometry-Resolved Avalanche Criterion in Disordered Spin Systems

Problem Statement
The stability of Many-Body Localization (MBL) in the thermodynamic limit, particularly in dimensions $d \geq 2$ , is widely believed to hinge on "quantum avalanches." In this mechanism, rare thermal inclusions (bubbles) grow by consuming localized surroundings if their thermalizing influence outpaces the decay of coupling to distant spins. While the avalanche criterion has been quantitatively tested in strictly one-dimensional (1D) chains, the geometry dependence central to higher-dimensional arguments has remained unmeasured. Specifically, it is unknown how the stability criterion evolves as a system widens from a chain to 2D strips and 3D tubes, or whether the "flat-front" expectation (where geometry enters only via cross-sectional area) holds at finite disorder.

Methodology
The authors investigate the disordered spin-1/2 XXZ model ( $J=1, \Delta=1$ ) with random fields $h_i \in [-W, W]$ . They employ a bath-coupled relaxation-rate diagnostic to measure the stability of the system against avalanches.

System Geometries: The study compares six geometries: 1D chains ( $1 \times L$ ), open strips ( $2 \times L, 3 \times L, 4 \times L$ ), a periodic three-leg strip (triangular tube, $A=3$ ), and a $2 \times 2 \times L$ tube ( $A=4$ ).
Diagnostic Protocol: An infinite-temperature Markovian bath is attached to one end ( $\ell=0$ ). The stability criterion is determined by the slowest nonzero relaxation rate $\Gamma$ of the resulting Lindblad generator.
Weak-Coupling Approximation: To avoid the computational cost of full Liouvillian diagonalization, the authors utilize the "golden-rule" (Pauli master equation) limit. This approximation is validated against exact Lindblad dynamics for small systems, showing agreement within 0–6% for coupling $\gamma = 10^{-3}$ .
Window Control: A critical methodological finding is that $\ln \Gamma(\ell)$ is strongly convex; slopes measured over short column windows differ significantly from those over longer windows. To ensure valid cross-geometry comparisons, the authors enforce a unified measurement window (e.g., columns $\ell=2–4$ for most geometries, $\ell=2–3$ for larger systems) rather than comparing asymptotic limits directly.

Key Results

Slope Hierarchy and Stability Boundaries:
The decay slope $s = -d \ln \Gamma / d\ell$ decreases monotonically as transverse connectivity increases. Wider and better-connected systems face a "double penalty": they must clear a higher stability bar ( $s^* = 2A \ln 2$ , where $A$ is the cross-sectional area) while exhibiting a lower decay slope. Consequently, the critical disorder $W^*$ required for stability increases significantly with width (e.g., $W^* \approx 16$ for a 2-leg ladder vs. $W^* \approx 70$ for a $2 \times 2$ tube).
Connectivity-Deficit Law:
The geometry dependence is quantified by a slope deficit $\Delta_G(W) = s_1(W) - s_G(W)$ relative to the 1D chain.
- For open strips, the deficit follows a leading-order linear law based on the number of transverse bonds per column: $\Delta_A(W) \approx (A-1)\delta(W)$ .
- This linear approximation breaks down at higher connectivity and shorter windows, where deficits grow faster than linearly.
Role of Transverse Loops:
Comparing open three-leg strips with periodic (closed-loop) counterparts reveals that topology matters beyond simple bond counting. Closed transverse loops retain their slope deficits significantly longer than open strips as disorder increases. The ratio of deficits between periodic and open geometries grows from $\approx 1.1$ to $\approx 1.8$ as disorder increases, indicating that closed loops support thermalization channels that survive stronger disorder.
Asymptotic Recovery of the Flat-Front Law:
Despite significant deviations at finite disorder, all measured deficits decay toward zero as disorder $W$ increases. At strong disorder ( $W \approx 96$ ), the connectivity effects vanish, and the system asymptotically recovers the "flat-front" criterion where geometry enters only through the cross-sectional area.

Significance and Claims
The paper claims to promote the avalanche instability argument from a qualitative scaling estimate to a measured, geometry-resolved law.

Quantitative Framework: The results provide a quantitative framework to interpret the "system-size drift" of the MBL crossover observed in 2D quantum simulators, explaining why the apparent critical disorder grows with width more slowly than analytic predictions.
Renormalization of Coupling: The study demonstrates that at accessible disorder strengths, transverse connectivity renormalizes the coupling decay in a lawful manner, rather than the system behaving purely as a 1D chain with an effective area.
Limitations: The authors modestly note that their claims are comparative (between geometries on identical finite windows) and do not yet constitute a thermodynamic-limit proof, as the convexity of the relaxation rate requires window extrapolation beyond their current system sizes ( $N \leq 12$ ). Additionally, the analysis is restricted to the weak-coupling limit and high-disorder regimes, excluding the low-disorder region where conventional finite-size thermalization dominates.

In summary, the work establishes that while the flat-front area law is the correct asymptotic limit, finite-disorder corrections are substantial, organized by transverse bond counts, and significantly enhanced by closed-loop topologies.

The Avalanche in the Shape of a Room

The "Flat-Front" Myth vs. The "Connectivity" Reality

The "Deficit" Law

What This Means for the Future

Technical Summary: Geometry-Resolved Avalanche Criterion in Disordered Spin Systems

More like this