🔬 condensed matter

Phase distinction of Gibbs states without symmetry breaking: topological invariants of the 3D toric code

This paper demonstrates that while topological entanglement entropy remains quantized in the finite-temperature 3D toric code despite symmetry breaking, it fails to be a robust mixed-state invariant under quasi-local channels, necessitating the introduction of the decoded Wilson-loop correlation as a new, stable topological invariant to distinguish the topological phase from a trivial one.

Original authors: Haruki Watanabe

Published 2026-07-02

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Haruki Watanabe

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 a giant, 3D grid made of tiny switches (like light switches on a wall). In this grid, the switches are connected in a special way that creates a hidden, "topological" order. Think of this order not as a pattern you can see with your eyes, but as a secret rule about how the switches are linked together globally. In physics, we call this a Toric Code.

Usually, scientists classify different states of matter (like ice vs. water) by looking for symmetry breaking. Imagine a room full of people standing in a perfect circle (symmetry). If they all suddenly decide to sit down on the left side of the room, they have "broken" that symmetry. This is how we usually tell phases apart.

The Big Question
This paper asks: What happens if we shake this grid with a strong magnetic field? The field breaks all the usual "symmetry rules" we look for. Does the hidden topological order disappear immediately, or can it survive even when the usual rules are broken?

The Answer: Geometry Saves the Day
The authors found that the topological order does survive at high temperatures, even with the magnetic field shaking things up.

Here is the secret: The order isn't protected by a symmetry rule; it's protected by geometry.

The Analogy: Imagine trying to draw a loop of string on a piece of paper. You can twist it, but you can't make the string end in mid-air; it must always connect back to itself to form a closed loop.
In this 3D grid, the "strings" are loops of magnetic flux. A fundamental law of physics (called the Bianchi identity) says these loops cannot just stop or end at a single point. They must always be closed circles.
Because the loops are forced to be closed by geometry, they can't just dissolve into chaos easily. They can only melt away if the loops get so big and numerous that they fill the whole grid. This creates a sharp, distinct boundary between the "ordered" state and the "disordered" state, even though no symmetry was broken.

The Problem with Old Tools
To find this order, the scientists tried using standard "thermometers" (mathematical tools) that physicists have used for decades.

The Failed Tools: They tried measuring things like "Wilson loops" (checking if a loop of switches is connected). These tools failed because the magnetic field scrambled the specific type of connection they were looking for. It was like trying to find a specific type of knot in a tangled ball of yarn, but the tangles were hiding the knot.
The "Fake" Order: They also discovered a tricky problem. You can mathematically create a state that looks like it has topological order (it has a specific "entropy" value of $\ln 2$ ) starting from a completely boring, empty state just by using a simple, short computer program. This means the old "entropy" tool can be fooled. It's like a magic trick where a magician makes a rabbit appear from an empty hat; just because you see a rabbit doesn't mean the hat was special to begin with.

The New Solution: The "Decoded" Detective
Since the old tools were fooled, the authors invented a new, smarter tool called the Decoded Wilson-Loop Correlation ( $f_W$ ).

The Analogy: Imagine you are trying to read a message written on a piece of paper that has been covered in static noise.
- The old tool just looked at the static and said, "It looks like a pattern!" (which could be a false alarm).
- The new tool acts like a smart decoder. It first tries to clean up the noise (removing the "errors" caused by the magnetic field) and then looks for the message.
How it works: The new tool checks if the "secret loops" can still be recovered after the noise is cleaned up.
- In the Topological Phase (the ordered state), the tool successfully cleans the noise and finds the message: $f_W = 1$ .
- In the Trivial Phase (the messy state), the noise is too strong. Even after cleaning, the message is gone: $f_W = 0$ .

The Results
Using massive computer simulations (Quantum Monte Carlo), the authors mapped out exactly where this transition happens.

It's a Sharp Line: The transition from order to chaos is a sharp, sudden change (like water freezing into ice), not a slow fade.
It's Robust: This sharp line exists even when the magnetic field breaks all the usual symmetry rules.
The "Fake" vs. "Real": They proved that while the old "entropy" number ( $\ln 2$ ) can be faked by simple tricks, their new "decoded" tool ( $f_W$ ) cannot be faked. It is a true, robust fingerprint of the topological order.

In Summary
This paper shows that a special kind of quantum order can survive high heat and strong magnetic fields, not because of a symmetry rule, but because of a geometric rule that forces loops to stay closed. They also built a new, un-fakeable "decoder" tool to detect this order, proving that even when the usual rules of physics are broken, geometry can still hold the structure together.

Technical Summary: Phase Distinction of Gibbs States without Symmetry Breaking

Problem Statement
The classification of quantum phases at finite temperature ( $T > 0$ ) remains less understood than at zero temperature. While zero-temperature phases are distinguished by topological order or symmetry-protected topological (SPT) features, thermal states are often assumed to lose sharp topological distinctions due to thermal fluctuations. A central question is whether a sharp phase boundary can exist between a topologically ordered phase and a trivial phase at $T > 0$ without relying on the spontaneous breaking of any exact symmetry (conventional or higher-form).

Existing frameworks for phase equivalence at $T > 0$ are debated. While pure ground states are classified by finite-depth local unitaries (FDLU), mixed states are generally classified by two-way connectivity via quasi-local channels (QLCs). However, this relation may be too coarse, potentially identifying decohered topological states with trivial product states. The authors investigate the 3D $\mathbb{Z}_2$ bosonic toric code in a generic magnetic field, a system where every exact higher-form symmetry is explicitly broken by the field, yet a non-terminating phase boundary is hypothesized to exist, protected purely by geometric constraints (the Bianchi identity).

Methodology
The authors employ a combination of perturbative stability analysis and large-scale continuous-time Quantum Monte Carlo (QMC) simulations using the worldline algorithm (implemented via the ParaToric package).

Model: The study focuses on the 3D $\mathbb{Z}_2$ toric code Hamiltonian on a cubic lattice with periodic boundary conditions, perturbed by transverse ( $h_x$ ) and longitudinal ( $h_z$ ) magnetic fields. These fields explicitly break the magnetic two-form and electric one-form symmetries, respectively.
Diagnostics:
- Conventional Probes: The authors first demonstrate that standard order parameters fail individually: the bare Wilson loop is blind to the electric transition, the membrane operator is smooth across thermal transitions, and the Fredenhagen-Marcu string order parameter is blind to magnetic and thermal boundaries.
- Topological Entanglement Entropy ( $\gamma$ ): They compute the topological entanglement entropy using the Levin-Wen four-region construction (specifically the "donut" or annular scheme to isolate the flux sector). They utilize the Rényi-2 entropy ( $S^{(2)}$ ) as an estimator, justified by its equivalence to the von Neumann entropy for this topological invariant in the relevant limits.
- Decoded Wilson-Loop Correlation ( $f_W$ ): To address the limitations of $\gamma$ $γ$ as a QLC invariant, they introduce a new order parameter, $f_W$ $f_{W}$ . This involves a "decode-then-read" protocol:
  - Flux Cleaning: A decoding channel $C$ is applied to the thermal state to project it onto the flux-free sector by measuring flux syndromes and applying minimum-weight perfect matching (MWPM) recovery.
  - Correlation: $f_W$ is defined as the connected correlation of the decoded Wilson loops: $f_W = \langle \tilde{W}_1 \tilde{W}_2 \rangle - \langle \tilde{W}_1 \rangle \langle \tilde{W}_2 \rangle$ .
  - Holonomy Correction: The one-point function $\langle \tilde{W} \rangle$ is evaluated via thermodynamic integration of a membrane twist to account for holonomy sectors, ensuring $f_W$ correctly vanishes in the Higgs/trivial phase.

Key Contributions and Results

Stability of Topological Entanglement Entropy ( $\gamma$ ):
- The authors show perturbatively (to all orders in the field) and confirm numerically that $\gamma$ remains quantized at $\ln 2$ throughout the topological phase, even under symmetry-breaking fields.
- $\gamma$ collapses sharply to 0 across the thermal transition ( $T_c$ ) and the field-driven transitions ( $h_x^c, h_z^c$ ).
- Limitation: The paper demonstrates that $\gamma = \ln 2$ is not a quasi-local-channel invariant. A constant-depth channel can generate a state with $\gamma = \ln 2$ from a trivial product state. Thus, while $\gamma$ is a sharp diagnostic within the Gibbs family, it cannot distinguish a genuine topological Gibbs state from a "spurious" topological state created by a shallow circuit.
The Decoded Wilson-Loop Correlation ( $f_W$ ) as a Robust Invariant:
- The authors introduce $f_W$ as the first finite-temperature, channel-invariant order parameter for this phase.
- Quantization: $f_W$ quantizes to 1 in the topological phase and 0 in the trivial phase as $L \to \infty$ .
- Invariance: Unlike $\gamma$ , $f_W$ is invariant under quasi-local channels. It distinguishes the thermal Gibbs state (which has long-range connected correlations) from the constant-depth image of a trivial product state (which has the same bulk $\gamma$ but lacks the connected correlation).
- Phase Boundary Mapping: $f_W$ successfully maps the entire phase diagram, identifying a non-terminating boundary separating the topological and trivial phases. The transition is found to be in the 3D Ising universality class, with critical points shifted by the fields (e.g., $T_c \approx 1.24$ at $(h_x, h_z) = (0.5, 0.1)$ ).
Geometric Protection Mechanism:
- The persistence of the phase boundary without exact symmetry breaking is attributed to the Bianchi identity ( $\prod_{p \in \partial c} \hat{B}_p \equiv 1$ ). This exact operator identity forbids point-like magnetic flux monopoles, forcing flux to proliferate only via closed loops. This geometric constraint ensures the transition remains sharp and non-terminating, even when all exact higher-form symmetries are broken.

Significance and Claims
The paper claims to establish a concrete example of a finite-temperature phase distinction that satisfies two conditions: (C1) the phases break no exact symmetry of the Hamiltonian, and (C2) the phase boundary is non-terminating.

Refinement of Phase Classification: The work refines the classification of mixed states by demonstrating that while topological entanglement entropy is a useful thermodynamic diagnostic, it is insufficient for defining mixed-state phases under quasi-local channel equivalence. The introduction of $f_W$ provides a necessary "recoverability" order parameter that captures the global logical correlation protected by the light-cone (Lieb-Robinson) bound.
Nature of the Order: The authors clarify that the surviving order in the bosonic 3D toric code at $T > 0$ is the recoverability of a classical bit (the magnetic flux loop sector), not a protected quantum qubit. The system is short-range entangled and admits a classical-mixture decomposition, distinguishing it from the fermionic toric code which retains genuine long-range entanglement at finite temperature.
Geometric vs. Symmetry Protection: The study positions "geometry-driven" transitions as a distinct paradigm from symmetry-breaking transitions. The phase boundary is pinned by the geometric Bianchi constraint, with any emergent higher-form symmetry being a downstream consequence rather than the microscopic origin of the protection.

In summary, the paper provides a rigorous thermodynamic and information-theoretic characterization of the 3D toric code's finite-temperature phase, establishing that topological order can persist at $T > 0$ without exact symmetries, provided the protection is geometric, and identifying $f_W$ as the robust invariant required to define such phases within the quasi-local channel framework.

Technical Summary: Phase Distinction of Gibbs States without Symmetry Breaking

More like this