⚛️ quantum physics

Spreading of Magic Resource under Unitary Clifford Dynamics

This paper investigates the spatiotemporal dynamics of conserved nonstabilizerness in unitary Clifford circuits by introducing a "bipartite magic gauge" to infer its local distribution and identifying two magic length scales that exhibit ballistic growth followed by delocalization.

Original authors: Mircea Bejan, Pieter W. Claeys, Jiangtian Yao

Published 2026-08-12

📖 4 min read🧠 Deep dive

CC BY 4.0

Original authors: Mircea Bejan, Pieter W. Claeys, Jiangtian Yao

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

Imagine the universe as a giant, cosmic kitchen where everything is made of tiny, invisible ingredients called quantum bits, or "qubits." In this kitchen, there's a special kind of cooking called quantum computing. To make the most delicious, complex dishes (like solving problems that would take regular computers millions of years), you need a secret ingredient. Scientists call this ingredient "magic" or "non-stabilizerness." It's not a wand or a spell, but a specific type of quantum weirdness that makes a system hard to simulate on a normal computer. Without this magic, quantum computers are like a car without an engine: they can move a little, but they can't go anywhere fast.

However, there's a catch. To keep this magic safe, we have to hide it inside a protective shell called a "quantum error correcting code." Think of this like wrapping a fragile, magical gem in layers of bubble wrap. The problem is, if you poke the bubble wrap in one spot, the magic doesn't just stay there; it starts to wiggle and spread out through the layers. Scientists have long known that this magic spreads, but they didn't know how it moved. Was it a slow leak? A sudden explosion? Or did it travel like a wave? Understanding this movement is crucial because if we can predict how the magic spreads, we can build better shields to protect our quantum computers from breaking down.

This paper dives into that exact mystery. The authors, Mircea Bejan, Pieter W. Claeys, and Jiangtian Yao, set up a digital experiment to watch how this "magic resource" spreads when it's injected into a system of qubits that are being shuffled around by a specific type of quantum logic called "Clifford dynamics." They discovered that the magic doesn't just drift; it zooms. Specifically, they found that the magic spreads in two distinct ways, like two different waves racing across a pond.

First, they identified a "typical magic length," which is the size of the smallest chunk of the system needed to grab the magic. In their simulations, this chunk grows steadily and quickly, like a balloon inflating at a constant speed. They call this speed the "entanglement velocity." It's as if the magic is riding a wave of connection between the qubits, expanding outward at a predictable pace.

Second, they found a "full linear extent," which is the total width of the area where the magic has become so spread out that you can't get rid of it anymore, no matter what you do outside that zone. This area grows even faster—about twice as fast as the first one. Imagine a ripple in a pond: the inner part of the ripple is where the water is really moving (the typical length), but the outer edge of the disturbance (the full extent) is where the ripples have touched everything. The paper shows that both of these boundaries move ballistically, meaning they travel in a straight line at a constant, fast speed, rather than slowing down or getting stuck.

The researchers also figured out a clever way to track this magic without having to do impossible math. They used a concept from error-correcting codes, treating the spreading magic like a secret message hidden in a code. By finding a special "gauge" (a way of looking at the math), they could calculate exactly where the magic was hiding at any moment. This allowed them to prove that at the very beginning, the magic spreads at these specific speeds. However, they also found that this fast spreading doesn't last forever. After a certain amount of time, the magic stops being localized in a specific spot and becomes "delocalized." This means it's everywhere at once, and you can't point to a specific region and say, "The magic is here." Instead, you need to look at more than half of the entire system to find it.

In short, the paper suggests that when you inject magic into a quantum system, it doesn't just sit there or leak slowly; it races outward at the speed of entanglement, creating a growing zone of influence. While this behavior was observed in their computer simulations, it offers a clear picture of how quantum resources behave in the chaotic dance of many-body systems. This helps scientists understand the limits of how fast quantum information can spread and how to design better error-correcting codes to keep our future quantum computers safe from the very magic they rely on.

Technical Summary: Spreading of Magic Resource under Unitary Clifford Dynamics

Problem Statement
Nonstabilizerness (NS), or "magic," is a critical quantum resource required for universal quantum computation and is believed to underpin quantum chaos and generalized thermalization. While the accumulation of entanglement and NS in ergodic systems is well-studied, the spatiotemporal spreading of locally injected NS under dynamics that conserve the total amount of NS remains largely unexplored. Specifically, in unitary Clifford circuits (which conserve total NS but generate entanglement), it is unclear how a localized injection of magic propagates. A major obstacle is the absence of physical observables that directly quantify local NS, precluding a standard microscopic or hydrodynamic description of its distribution.

Methodology
The authors investigate the dynamics of NS in unitary quantum circuits consisting of $L$ qubits and depth $t$ , utilizing a "brickwall" architecture of random 2-qubit Clifford gates. The system is initialized in a short-range entangled stabilizer state, with NS locally injected via a single non-Clifford $T$ -gate acting on a central qubit. The evolution is governed by identity-doped Clifford circuits, allowing for tunable butterfly ( $v_B$ ) and entanglement ( $v_E$ ) velocities.

To address the lack of local observables, the authors leverage insights from Stabilizer Quantum Error Correcting Codes (QEC). They establish that the time-evolved state encodes a single logical qubit within a dynamically generated stabilizer code. The core methodological innovation is the introduction of the Bipartite Magic Gauge (BMG). This is a rigorous construction of a specific gauge for logical operators and stabilizer generators for any bipartition of the system. The BMG enables:

The exact computation of NS measures (specifically the Robustness of Magic, ROM, and mixed-state Stabilizer Rényi Entropy, SRE) for any subsystem in polynomial time ( $O(L^3)$ ), bypassing the exponential cost typically associated with calculating NS in volume-law entangled states.
The mapping of the spatial structure of magic resource to the spatial support of the code's logical operators.

Key Contributions and Definitions
The paper defines two operational Magic Length Scales (MLS) to characterize the spatial distribution of NS:

Typical Magic Length ( $\ell_{typ}$ ): The size of the smallest contiguous region from which NS can be unitarily extracted (or destroyed). This corresponds to the minimal length of mutually irreducible intervals where logical operators can be fully supported.
Full Linear Extent of Magic (FLEOM, $W$ ): The width of the union of all mutually irreducible intervals where NS is extractable. Operations performed outside this region cannot reduce the total NS in the system.

The authors prove Theorem 1, which guarantees the existence of the BMG for any bipartition of a pure state with one unit of NS. This theorem, combined with Lemma 1 regarding the reducibility of logical operators, provides the theoretical foundation for efficiently computing the spatial structure of magic.

Results
Numerical simulations on random Clifford circuits reveal the following dynamics:

Ballistic Growth: At early times, both MLS grow ballistically.
- The typical magic length $\ell(t)$ grows at a rate $v_\ell \approx v_E$ (the entanglement velocity).
- The full linear extent $W(t)$ grows at a rate $v_W \approx 2v_E$ .
- These velocities are consistent across different identity doping rates $p$ , matching the theoretical entanglement velocity $v_E(p)$ .
Delocalization: At late times ( $t \sim O(L)$ ), the magic resource delocalizes. The FLEOM saturates to the full system size ( $W \to L$ ), and the typical length saturates to $\ell \approx L/2$ .
Late-Time Behavior: In the delocalized regime, the dynamically generated code approaches a random $[[L, 1]]$ stabilizer code. Consequently, NS can be unitarily extracted from any subsystem containing $\gtrsim L/2$ qubits, regardless of spatial contiguity. This behavior saturates the quantum Singleton bound, indicating that the spatial proximity of qubits becomes irrelevant for accessing the resource.

Significance
The paper claims to provide the first rigorous framework for inferring the spatial distribution of nonstabilizerness in conserved dynamics without relying on physical observables. By establishing a connection between magic resource spreading and the structure of logical operators in dynamically generated QEC codes, the work:

Offers an efficient classical algorithm to compute subsystem NS exactly.
Identifies specific length scales and velocities governing the transport of quantum complexity.
Sheds light on the mechanisms behind quantum thermalization and the causal organization of logical information in dynamical QEC codes.
Opens avenues for investigating the transport properties of quantum resources and their interplay with quantum error correction, particularly in the context of how complexity builds up in many-body systems.

The authors remain modest regarding future implications, noting that while they provide analytical bounds and heuristic arguments, a more rigorous understanding of the underlying physical mechanisms (e.g., via hydrodynamic descriptions) and generalization to states with multiple magic units remain open problems.

Technical Summary: Spreading of Magic Resource under Unitary Clifford Dynamics

More like this