⚛️ quantum physics

Phase space anatomy of dynamical quantum phase transitions

This paper demonstrates that discrete phase space fundamentally distinguishes between dynamical quantum phase transitions defined by local order parameters and those defined by global return rate nonanalyticities, revealing that the latter rely on destructive quantum interference and information absent from reduced states rather than quasiprobability negativity.

Original authors: Zakaria Mzaouali

Published 2026-09-22

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Zakaria Mzaouali

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, particles do not simply sit still; they evolve, twist, and interfere with one another in ways that have no parallel in our daily experience. When a system of many such particles is suddenly jolted out of balance—a process scientists call a "quench"—it begins a complex journey toward a new state. Sometimes, during this chaotic evolution, the system undergoes a dramatic shift known as a dynamical quantum phase transition. These are not the familiar changes of state we see in ice melting or water boiling, which happen when things are calm and steady. Instead, these transitions happen in the heat of the moment, as the system races through time. Physicists have long been trying to understand what actually triggers these sudden shifts. For years, two different ways of looking at the problem seemed to tell the same story: one focused on how the overall order of the system changes, while the other looked at how likely the system is to return to its starting point. It was assumed that these two views were just different angles on the same underlying event.

A new study challenges this assumption, revealing that these two perspectives are actually looking at fundamentally different things. Researchers at the Jülich Supercomputing Centre and the University of Tübingen have shown that the sudden changes in a system's order and the sudden changes in its return probability are driven by separate physical mechanisms. To understand this, imagine trying to predict the outcome of a complex game by looking at just one player versus watching the entire board. The study demonstrates that while the local behavior of a small group of particles can tell you about the system's order, it cannot predict the global return probability. The return probability depends on the entire system acting together, and it is governed by a subtle, invisible force called quantum interference. This force can cancel out possibilities in ways that local observations simply cannot see.

The researchers investigated this by simulating a chain of particles, each capable of existing in three distinct states, much like a coin that can land on heads, tails, or a third, unique side. They watched what happened when they suddenly changed the rules governing how these particles interacted. In their simulations, they tracked two specific things: the system's "order," which is a measure of how aligned the particles are, and the "return rate," which measures how closely the system resembles its original state at any given moment. They found that the moment the system's order changed was often different from the moment the return rate shifted. This time gap was not a mistake or a glitch; it was a physical reality caused by the way quantum waves interfere with one another.

In the quantum realm, particles behave like waves that can add together or cancel each other out. When the researchers looked at the return rate, they discovered that the system's path was being shaped by this cancellation. Sometimes, a path that looked like it should win based on its raw strength was actually suppressed because its waves canceled out with other possibilities. In other cases, a path that seemed weaker survived because it avoided this cancellation. The study showed that this "selective cancellation" could delay the moment when one path overtook another. In one specific simulation, the researchers found that a competing path had more than three times the raw weight of the initial path, yet the initial path still won because the competitor suffered from stronger cancellation. This meant that the system's behavior was being dictated not just by how much "stuff" was in a path, but by how the waves within that path interfered with themselves.

To prove that this interference was the key, the team also looked at a simpler, non-interacting system where particles did not influence each other. In this controlled scenario, they found a sharp transition in the return rate even though there was no cancellation at all. This proved that a sudden change in the return rate does not automatically mean that complex quantum interference is happening. Conversely, in the more complex, interacting system, they found that the interference was strong enough to change which path the system followed, even when the raw numbers suggested otherwise. This distinction is crucial: the existence of a sudden transition and the role of interference in choosing that transition are two separate questions. One can happen without the other.

The researchers used a mathematical tool called a phase space representation to visualize these invisible forces. This tool allowed them to separate the "unsigned" weight of a path—its raw potential—from the "signed" weight, which accounts for the cancellations. By comparing the two, they could see exactly where the interference was doing its work. They found that as they looked at larger and larger groups of particles, the delay between the raw potential crossing and the actual physical crossing remained, suggesting that this effect is not just a small-scale curiosity but a fundamental feature of how these systems evolve. The study also showed that you cannot reconstruct the full story of the system's return just by looking at small, local pieces of it. The global picture contains information that is completely absent from any local view, much like how you cannot understand the plot of a movie by only watching a single frame.

This work clarifies a long-standing confusion in the field of quantum dynamics. For some time, scientists have debated whether the sudden changes in a system's order and the sudden changes in its return probability were the same phenomenon. This study shows they are not. The transition in order is determined by local information, while the transition in return probability is determined by global information and the delicate balance of quantum interference. The findings suggest that to truly understand these transitions, one must look at the whole system and account for the invisible cancellations that shape its path. The research does not claim to have solved every mystery of quantum dynamics, but it has drawn a clear line between two concepts that were previously blurred. It establishes that the mechanism selecting the dominant path in a quantum system is distinct from the mere existence of a transition, offering a more precise map for navigating the strange and counterintuitive landscape of the quantum world.

Technical Summary: Phase Space Anatomy of Dynamical Quantum Phase Transitions

Problem Statement
Dynamical quantum phase transitions (DQPTs) are typically characterized by two distinct notions: (I) nonanalytic changes in a late-time order parameter (DQPT I) and (II) singularities in the Loschmidt return rate in the thermodynamic limit (DQPT II). While experimental signatures of both have been observed, the physical relationship between them remains unclear. Specifically, it is unknown whether a return singularity arises simply because a branch carries more "unsigned" phase space weight, or because it survives destructive quantum interference more effectively than its competitors. Furthermore, it is unclear if the information required to predict a global return singularity can be reconstructed from local reduced states (order parameters) or if it requires global information absent from any proper subsystem.

Methodology
The author utilizes discrete Wigner representations on canonical qutrit phase space (odd prime dimension $d=3$ ) to analyze the evolution of quantum states following a quench. This framework allows for the decomposition of return rates into distinct components:

Unsigned Support ( $s$ ): The rate derived from the sum of absolute Wigner weights ( $A$ ), representing the "unsigned" phase space volume.
Interference Cost ( $q$ ): The penalty arising from destructive interference, defined via the ratio of physical return probability ( $P$ ) to unsigned weight ( $A$ ), where $q = \ell^{-1} \ln(A/P)$ .
Total Return Rate ( $r$ ): The physical rate, satisfying the exact identity $r = s + q$ .

The study employs three primary models:

Exact Control: A non-interacting qutrit chain with a specific rotation Hamiltonian to test if a thermodynamic cusp can exist without Wigner negativity.
Interacting Potts Chain: An open quantum Potts chain quenched from a polarized state, simulated using Time-Evolving Block Decimation (TEBD) for finite chains ( $N=50, 100$ ) and infinite Matrix Product States (MPS) for the thermodynamic limit.
Collective Ising Model: A permutation-symmetric spin model used to compare local magnetization dynamics with global return branch exchanges.

The analysis focuses on "block returns," comparing the crossing times of physical rates ( $t_c$ ) against the crossing times of their unsigned counterparts ( $t_s$ ) as the block size $\ell$ grows.

Key Contributions and Results

Information Scale Separation: The paper establishes that DQPT I (local order) is determined by Wigner marginals of fixed size, whereas DQPT II (global return) requires information from the full system. Using a GHZ-like state example, the author demonstrates that proper reduced states are independent of global phase parameters, proving that local marginals cannot reconstruct general global Loschmidt returns.
Decomposition of Return Rates: The author derives an exact decomposition showing that the physical return rate is the sum of an unsigned support cost and an interference cost. They show that a return singularity (cusp) can exist even when the interference cost is zero.
- Result: In an exact control model of independent qutrits, the thermodynamic return rate develops a cusp at a crossing point where the state is a stabilizer product state with strictly non-negative Wigner functions ( $q=0$ ). This proves that a cusp alone does not certify Wigner negativity or entanglement.
Selective Cancellation and Branch Ranking: In the interacting Potts chain, the author observes that destructive interference can reverse the ranking of competing return branches.
- Result: For central blocks, the grouped competitor branch (combining symmetry-related sectors) often possesses a larger unsigned weight ( $s$ ) than the initial branch. However, the grouped competitor suffers a significantly higher interference cost ( $q$ ). This "selective cancellation" suppresses the competitor's physical return, delaying the exchange of dominance from the unsigned crossing time ( $t_s$ ) to the physical crossing time ( $t_c$ ).
- Quantification: At a block size of $\ell=9$ , the grouped competitor retains only 30.5% of its unsigned weight compared to 95.3% for the initial branch, despite the competitor having a larger unsigned support. This delay ( $t_c - t_s$ ) persists and remains positive as the block size increases, indicating that interference is a relevant mechanism for branch selection in the bulk.
Distinction of Local and Global Events: In the collective Ising model, the local magnetization (order parameter) crosses zero at a time distinct from the global return branch exchange. The local zero occurs smoothly, while the global return develops a rounded peak. This confirms that a local zero cannot substitute for the global branch criterion in finite systems.

Significance and Claims
The paper claims to resolve the "phase space anatomy" of DQPTs by distinguishing two separate physical questions:

The Existence of a Singularity: Whether a return rate develops a cusp. The author shows this can occur without Wigner negativity (via the exact control model).
The Mechanism of Branch Selection: How interference selects the dominant branch. The author shows that interference costs can delay or prevent a branch exchange that would otherwise occur based on unsigned weight alone.

The author concludes that the existence of a return singularity and the role of interference in selecting its branches are distinct physical phenomena. A global branch exchange can survive even if the interference contribution vanishes with increasing scale, and conversely, interference can alter the dominant sector even in the presence of a cusp. The work clarifies that "mana" (total absolute Wigner weight) does not retain the specific information about which measurement branch is selected, and that reconstructing global returns from local marginals is fundamentally obstructed for general states. The results suggest that distinguishing a return branch that wins due to greater unsigned weight from one that wins because its competitor suffers stronger cancellation is a necessary experimental target for understanding dynamical transitions.

More like this