Pith. sign in

REVIEW 2 major objections 6 minor 81 references

Non-invertible Kennedy–Tasaki duals need not share the same finite-temperature phase diagram.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

T0 review · grok-4.5

2026-07-31 20:25 UTC pith:LR4Z66RG

load-bearing objection Exact 3D s=0 counterexample shows non-invertible KT duality need not preserve unrestricted finite-T phase diagrams; QMC shapes of the mismatch window are softer but secondary. the 2 major comments →

arxiv 2607.24231 v1 pith:LR4Z66RG submitted 2026-07-27 cond-mat.str-el cond-mat.othercond-mat.stat-mechhep-th

Distinct finite-temperature phase diagrams of non-invertible Kennedy--Tasaki duals

Weiguang Cao , Haruki Watanabe This is my paper
classification cond-mat.str-el cond-mat.othercond-mat.stat-mechhep-th
keywords non-invertible symmetryKennedy-Tasaki transformationfinite-temperature phase diagramcluster modelZ2 gauge theorydeconfinementquantum Monte CarloSPT order
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

Invertible dualities preserve partition functions and therefore transport whole phase diagrams. Non-invertible dualities need not. This paper studies a cluster-model family H(s) and its Kennedy–Tasaki dual and shows that the map is only a partial isometry: it equates the all-plus-sector partition functions and says nothing automatic about the unrestricted thermal ensembles. In one dimension neither side can order at finite temperature, so the diagrams happen to agree and meet only at the shared zero-temperature critical point. In three dimensions the inequivalence is already exact at the solvable endpoint s=0: the cluster free energy is analytic for every finite T, while the dual Z2 gauge theory has a deconfinement transition near Tc≈1.31. Quantum Monte Carlo then shows the mismatch occupies a finite window of the interpolation—a self-dual frozen wedge on the cluster side and a deconfined dome on the gauge side—after which the diagrams reconverge up to the common trivial endpoint. The lesson is that using a non-invertible duality for thermodynamics requires independent control of the discarded sectors; fusion rules alone do not fix physical phase boundaries.

Core claim

A non-invertible Kennedy–Tasaki map equates only projected all-plus partition functions. Unrestricted finite-temperature phase diagrams of the dual pair therefore need not coincide. In three dimensions this is already exact at s=0, where the cluster model is analytically paramagnetic at every finite T while its dual gauge theory deconfines at Tc≈1.31, and Quantum Monte Carlo shows the mismatch fills a finite window of the interpolation before the diagrams rejoin.

What carries the argument

The Kennedy–Tasaki partial isometry D̃, satisfying D̃ H(s)=H̃(s) D̃ and D̃†D̃=D̃D̃†=P+, which yields the exact identity Z+(s,β)=Tr(P+ e^{-βH(s)})=Tr(P+ e^{-βH̃(s)}) but does not constrain the unrestricted spectra or Gibbs states once extensive charge sectors in the kernel are thermally occupied.

Load-bearing premise

That the finite-size Monte Carlo scans correctly place the thermodynamic edges of the frozen wedge and deconfined dome, and that no finite-temperature bulk SPT order hides in the hard-to-measure low-temperature window.

What would settle it

A controlled thermodynamic-limit determination (for example multicanonical or larger-volume continuous-time QMC) showing that either the cluster free energy develops a singularity matching the gauge deconfinement line, or that the twisted-membrane order parameter revives at finite T away from the fixed points, would overturn the claimed bulk mismatch.

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

  • Thermodynamic claims drawn from non-invertible dualities require a separate argument that discarded sectors contribute only subextensively to ln Z.
  • One-dimensional agreement of KT dual phase diagrams is accidental: both sides are forbidden from thermal order, not forced to match by the fusion algebra.
  • Three-dimensional flux-loop defects can sustain a finite-T deconfined phase while pointlike domain walls cannot, so dimensionality of defects decides when unrestricted diagrams diverge.
  • The same sector-entropy mechanism should appear in other lattice Kramers–Wannier and fusion-category dualities whenever the projector fixes extensively many local charges.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Any program that diagnoses mixed-state or thermal SPT order by transporting zero-temperature dual diagnostics must first verify that the physical ensemble is the projected one, not the unrestricted Gibbs state.
  • The frozen wedge is an unrestricted-ensemble phenomenon invisible to the all-plus identity; similar intermediate branches may appear generically when commuting stabilizer families supply extensive residual entropy.
  • If the weakly first-order segment on the gauge side ends at a true tricritical point off s=1/2, charge-sector entropy systematically shifts first-order walls even when projected free energies remain self-dual.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

2 major / 6 minor

Summary. The manuscript asks whether Hamiltonians related by a non-invertible Kennedy–Tasaki (KT) transformation share the same finite-temperature phase diagram. The answer is no in general. The KT map is shown to be a partial isometry onto the all-plus fusion sector (Eqs. 4–7), so it equates only projected partition functions Z+(s,β) and the associated strongly symmetric Gibbs states, not the unrestricted thermal ensembles. In 1D the unrestricted cluster chain and its dual (two transverse-field Ising chains) both lack thermal order, so the phase diagrams coincide accidentally — established by an exact free-fermion/Toeplitz solution verified against exact diagonalization (Sec. S2). In 3D the inequivalence is exact already at s=0: H_SPT = U H_para U† has the analytic free energy (2 cosh β)^N, while its dual is two decoupled 3D Z2 gauge theories with a deconfinement transition at T_c(0) ≈ 1.31 fixed by Wegner duality (Eq. 19). QMC then maps the unrestricted phase diagrams of both interpolations (Fig. 3): a self-dual 'frozen wedge' on the cluster side, a deconfined dome on the gauge side with a first-order segment rising from (1/2,0) and meeting a continuous arm at a finite-size tricritical bracket, and agreement of the two diagrams once the window closes. The twisted-membrane SPT diagnostic is shown to be area-law suppressed exactly at the fixed points and rigorously at high temperature (Eq. S49), with no revival in the accessible numerical window.

Significance. If the result holds, it corrects a natural but false assumption in the rapidly growing non-invertible-symmetry literature: that an exact Hamiltonian intertwining relation suffices for thermodynamic equivalence. The paper demonstrates, with a parameter-free exact endpoint argument, that the Kennedy–Tasaki map constrains only a strongly symmetric projected ensemble, and that in 3D the discarded charge sectors carry extensive entropy that changes bulk free-energy singularities. This is directly relevant to current mixed-state and finite-temperature SPT programs (Refs. 32–35). Notable strengths: the s=0 inequivalence is a textbook-identity-level exact result; the 1D solution is complete and machine-verified; the Hellmann–Feynman criterion linking duality breaking to first-order walls is clean; the Wegner endpoint and initial slope are exact; and the numerics are unusually honest about what is equilibrium versus metastability bracketing, with reproducible methodological detail (Eqs. S55–S59). The paper also makes a falsifiable structural prediction — a self-dual frozen wedge on the cluster side with no saturated dual analogue, supported by a local bound ⟨B_p⟩² + ⟨X_l⟩² ≤ 1.

major comments (2)
  1. [Abstract / Fig. 3 caption / Conclusion] The abstract and Conclusion state that 'QMC shows that the mismatch occupies a finite window of the interpolation.' In fact, the equilibrium existence of the cluster-side frozen wedge is established only at L=3,4 (162–384 qubits, Sec. S3 E 2), the wedge endpoint (s,T)_end = (1/2 ∓ 0.030(8), 0.27(2)) comes from hysteresis-closure fits the authors themselves say have 'no scaling justification' (Sec. S3 E 4), and the dual-side tricritical bracket s* ≃ 0.542, 0.23 < T* < 0.24 is a finite-size/finite-time bound explicitly requiring multicanonical sampling (Sec. S3 F 3). The central inequivalence claim is exact at s=0 and does not depend on any of this, but the abstract presents the whole window as numerically settled. The main text (and abstract) should state which features of Fig. 3 are exact (the s=0 height T_c(0) ≈ 1.31, the initial slope dT_c/ds = −T_c(0), the T=0 endpoint (1/2,0), the fi
  2. [Sec. S3 E 3 (volume sharpening of the wedge boundary)] The statement that the ordered–frozen boundary 'sharpens with volume' rests on a single L=3 → L=4 comparison (step heights 0.25 → 0.45 at T=0.20 and 0.19 → 0.30 at T=0.25, Sec. S3 E 3), i.e. 162 versus 384 qubits. Two sizes with a factor of ~2.4 in volume cannot distinguish a sharpening first-order jump from a smooth crossover sharpening slowly. Since the wedge is the one genuinely new equilibrium phase structure claimed on the cluster side, either (i) add at least one larger size (L=5 or 6) at a temperature where mixing is demonstrated, or (ii) soften the language to 'consistent with a first-order boundary at the accessible sizes.' Option (ii) would suffice for the Letter's claim as currently framed, since the wedge is presented as supporting evidence for the s-window, not as a theorem.
minor comments (6)
  1. [One dimension, Eq. (11)] Eq. (11): the correlation length ξ is defined only for s ≠ 1/2 implicitly; at s=1/2 the dispersion ε_k(1/2) = 2|sin(k/2)| is gapless and the integral diverges logarithmically. A parenthetical noting this (or that the T>0 statement holds for any fixed s ≠ 1/2, with ξ → ∞ only as the critical point or T→0 is approached) would avoid confusion.
  2. [Fig. 3 caption] The reader must jump between the Letter and SM to learn that the 'bars' in Fig. 3 are metastability intervals rather than error bars. This is stated in the caption, which is good, but a one-sentence definition in the main text where Fig. 3 is introduced would help, since the figure is the paper's central object.
  3. [Sec. S2 B–C] In Sec. S2 B, the evaluation 'at s=0 the finite-size lobe ends near T=0.13' (and T=0.15 half-height in Sec. S2 C) gives two nearby numbers for what appear to be closely related finite-size crossover scales on the two panels of Fig. 2; please state precisely what criterion defines each number.
  4. [Sec. S3 C vs. Eq. (S53)] The heuristic estimate T_x ∼ Δe/Δs_res ≲ 0.01 (Eq. S53) is appropriately disclaimed, but it is quoted as 'T_x ≲ 0.01' in Sec. S3 C without repeating that it neglects splitting within the bare manifold and the entropy of the ordered branches. Repeat the qualifier at each use.
  5. [Letter, discussion of Fig. 2 inset / Sec. S2 D] Notation: m²_U is used both for the squared U-odd order parameter (Eq. S29) and informally for the one-dimensional quantity that 'vanishes continuously' in the Letter's discussion of Fig. 2 inset (where the relevant object is e'_0(s), not m²). Consider distinguishing the two explicitly to prevent the reader conflating the continuous 1D self-dual point with the U-breaking diagnostic.
  6. [References and captions] Typos/presentation: 'Kramers-Wanier' in Ref. [41] should be 'Kramers–Wannier'; the arXiv preprint date format of Ref. [35] is inconsistent with the rest; in Fig. S3 the caption should state the number of slices or Δτ used for reproducibility alongside the 64 chains.

Circularity Check

0 steps flagged

No significant circularity: s=0 inequivalence is textbook unitary conjugation plus Wegner duality; QMC boundaries are measured, not forced by fit or self-citation.

full rationale

The load-bearing algebraic claim—that the non-invertible KT map is a partial isometry equating only all-plus projected partition functions Z+(s,β)—is derived internally from the bilinear phase-map representation, ker M, and the fusion projector P+ (Eqs. 3–7 and S1), then checked by cyclicity of the trace. The exact three-dimensional inequivalence at s=0 does not close on itself: H_SPT = U H_para U† immediately gives the analytic free energy (2 cosh β)^N, while ˜H(0) is ordinary 3D Z2 gauge theory whose deconfinement temperature follows from external Wegner duality to the 3D Ising point (K_c ≈ 0.2216 ⇒ T_c ≈ 1.313). The one-dimensional coincidence is an independent free-fermion/Toeplitz calculation verified against exact diagonalization. Finite-s mismatch structure (frozen wedge, deconfined dome, first-order segment) is located by QMC and hysteresis, with parameters such as the curvature a = 0.21(2) reported as measured boundary shape rather than used to predict the central inequivalence. Author self-citations supply context on non-invertible dualities and do not import a uniqueness theorem that forces the thermodynamic conclusion. No step reduces a claimed prediction to its fitted input by construction.

Axiom & Free-Parameter Ledger

2 free parameters · 5 axioms · 1 invented entities

The central claim rests on standard lattice quantum statistical mechanics, the established KT/partial-isometry construction, Wegner's duality, and the assumption that QMC in the unrestricted ensemble captures bulk singularities. No new particles or forces are postulated. Numerical curvature and endpoint locations are measured outputs, not inputs that define the claim.

free parameters (2)
  • Leading curvature a in Tc(s)≈Tc(0)[1−s−a s²] = 0.21(2)
    Fitted from small-field QMC arm; used only to describe the dome shape, not to establish inequivalence.
  • Frozen-wedge endpoint (s,T)end and dual metastability endpoint (s*,T*) = cluster ends ≈(1/2∓0.030(8),0.27(2)); dual FS endpoint s*≃0.542, 0.23<T*<0.24
    Located from finite-size hysteresis and two-branch scans; bracket the mismatch window but are not assumed a priori.
axioms (5)
  • domain assumption KT map D̃ is a partial isometry with D̃†D̃=D̃D̃†=P+ onto the all-plus full-fusion sector, intertwining H(s) and H̃(s).
    Standard in the cited KT/non-invertible lattice literature; derived from bilinear phase map in SM opening and Eq. (4).
  • domain assumption Wegner duality equates pure 3D Z2 gauge theory at s=0 to 3D Ising, fixing Tc(0)≈1.3133 from K_Ising_c≈0.2216.
    Textbook result used for the exact dual endpoint height (Eq. 19 / S40).
  • domain assumption Unrestricted Gibbs traces (not only P+ projected ones) define the physical finite-T phase diagrams being compared.
    Stated explicitly as the weak vs strong symmetry distinction; without it the inequivalence claim is empty.
  • ad hoc to paper High-T cluster expansion and fixed-point area laws plus accessible QMC imply no bulk finite-T revival of the weak-Gibbs twisted membrane.
    Rigorous at high T and fixed points (S49, Eq. 17); interior relies on finite-size sign-problematic QMC (Fig. S3 white region).
  • domain assumption Standard QMC ergodicity/continuum limits: continuous-time gauge QMC and discrete-time worldline cluster QMC sample the correct unrestricted ensembles after stated thermalization.
    Validated against Metropolis and independent PMR-QMC in SM; still an operational assumption for the interior of Fig. 3.
invented entities (1)
  • Self-dual frozen wedge (cluster-side unrestricted ensemble) independent evidence
    purpose: Names the intermediate branch of nearly saturated commuting link-family configurations that opens a first-order-bounded wedge around s=1/2 and drives part of the phase-diagram mismatch.
    Not a new particle or symmetry; a diagnosed thermal structure from the commuting link families {Xl Bl}∪{Xl}. Independent handle is the local saturation diagnostics and hysteresis shelf in QMC.

pith-pipeline@v1.2.0-grok45-kimik3 · 39009 in / 3491 out tokens · 72071 ms · 2026-07-31T20:25:34.695912+00:00 · methodology

0 comments
read the original abstract

Do two Hamiltonians related by a non-invertible transformation necessarily share the same finite-temperature phase diagram? For a cluster-model interpolation $H(s)$ and its Kennedy--Tasaki dual $\tilde H(s)$, the map is a partial isometry and equates only their all-plus-sector partition functions. In one dimension, thermal order is forbidden on both sides, leaving only the common zero-temperature transition at $s=\tfrac12$. In three dimensions, however, the inequivalence is exact already at $s=0$: the cluster model has an analytic paramagnetic free energy, whereas its dual $\mathbb Z_2$ gauge theory has a deconfinement transition at $T_c\approx1.31$. Quantum Monte Carlo shows that the mismatch occupies a finite window of the interpolation, marked by a self-dual frozen wedge on the cluster side and a deconfined dome on the gauge side; once both close the two phase diagrams agree again, coinciding over the entire remaining interval up to the shared trivial endpoint $s=1$.

Figures

Figures reproduced from arXiv: 2607.24231 by Haruki Watanabe, Weiguang Cao.

Figure 1
Figure 1. Figure 1: FIG. 1. The unitary [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: FIG. 2. One-dimensional finite-size crossover maps on rings [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: FIG. 3. Unrestricted three-dimensional phase diagrams. [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

81 extracted references · 42 linked inside Pith

  1. [1]

    Tachikawa, On gauging finite subgroups, SciPost Phys.8, 015 (2020), arXiv:1712.09542 [hep-th]

    Y. Tachikawa, On gauging finite subgroups, SciPost Phys.8, 015 (2020), arXiv:1712.09542 [hep-th]

    Pith/arXiv arXiv 2020

  2. [2]

    Bhardwaj and Y

    L. Bhardwaj and Y. Tachikawa, On finite symmetries and their gauging in two dimensions, JHEP03, 189, arXiv:1704.02330 [hep-th]

    Pith/arXiv arXiv

  3. [3]

    Chang, Y.-H

    C.-M. Chang, Y.-H. Lin, S.-H. Shao, Y. Wang, and X. Yin, Topological Defect Lines and Renormalization 6 Group Flows in Two Dimensions, JHEP01(01), 026, arXiv:1802.04445 [hep-th]

    Pith/arXiv arXiv

  4. [4]

    Ji and X.-G

    W. Ji and X.-G. Wen, Categorical symmetry and non- invertible anomaly in symmetry-breaking and topologi- cal phase transitions, Phys. Rev. Res.2, 033417 (2020), arXiv:1912.13492 [cond-mat.str-el]

    Pith/arXiv arXiv 2020

  5. [5]

    Thorngren and Y

    R. Thorngren and Y. Wang, Fusion category symmetry. Part I. Anomaly in-flow and gapped phases, JHEP04, 132, arXiv:1912.02817 [hep-th]

    Pith/arXiv arXiv 1912

  6. [6]

    Thorngren and Y

    R. Thorngren and Y. Wang, Fusion category symmetry. Part II. Categoriosities at c = 1 and beyond, JHEP07, 051, arXiv:2106.12577 [hep-th]

    Pith/arXiv arXiv

  7. [7]

    Aasen, R

    D. Aasen, R. S. K. Mong, and P. Fendley, Topological Defects on the Lattice I: The Ising model, J. Phys. A49, 354001 (2016), arXiv:1601.07185 [cond-mat.stat-mech]

    Pith/arXiv arXiv 2016

  8. [8]

    Aasen, P

    D. Aasen, P. Fendley, and R. S. K. Mong, Topolog- ical Defects on the Lattice: Dualities and Degenera- cies, arXiv:2008.08598 (2020), arXiv:2008.08598 [cond- mat.stat-mech]

    Pith/arXiv arXiv 2008

  9. [9]

    L. Kong, T. Lan, X.-G. Wen, Z.-H. Zhang, and H. Zheng, Algebraic higher symmetry and categorical symmetry – a holographic and entanglement view of symmetry, Phys. Rev. Res.2, 043086 (2020), arXiv:2005.14178 [cond- mat.str-el]

    Pith/arXiv arXiv 2020

  10. [10]

    Bhardwaj, L

    L. Bhardwaj, L. E. Bottini, S. Schafer-Nameki, and A. Tiwari, Non-invertible higher-categorical symmetries, SciPost Phys.14, 007 (2023), arXiv:2204.06564 [hep-th]

    Pith/arXiv arXiv 2023

  11. [11]

    Apruzzi, I

    F. Apruzzi, I. Bah, F. Bonetti, and S. Schafer-Nameki, Noninvertible Symmetries from Holography and Branes, Phys. Rev. Lett.130, 121601 (2023), arXiv:2208.07373 [hep-th]

    Pith/arXiv arXiv 2023

  12. [12]

    Kaidi, K

    J. Kaidi, K. Ohmori, and Y. Zheng, Kramers-Wannier- like Duality Defects in (3+1)D Gauge Theories, Phys. Rev. Lett.128, 111601 (2022), arXiv:2111.01141 [hep- th]

    Pith/arXiv arXiv 2022

  13. [13]

    Y. Choi, C. Cordova, P.-S. Hsin, H. T. Lam, and S.-H. Shao, Noninvertible duality defects in 3+1 dimensions, Phys. Rev. D105, 125016 (2022), arXiv:2111.01139 [hep- th]

    Pith/arXiv arXiv 2022

  14. [14]

    Roumpedakis, S

    K. Roumpedakis, S. Seifnashri, and S.-H. Shao, Higher Gauging and Non-invertible Condensation Defects, Com- mun. Math. Phys.401, 3043 (2023), arXiv:2204.02407 [hep-th]

    Pith/arXiv arXiv 2023

  15. [15]

    Y. Choi, C. Cordova, P.-S. Hsin, H. T. Lam, and S.-H. Shao, Non-invertible Condensation, Duality, and Triality Defects in 3+1 Dimensions, Commun. Math. Phys.402, 489 (2023), arXiv:2204.09025 [hep-th]

    Pith/arXiv arXiv 2023

  16. [16]

    Lootens, C

    L. Lootens, C. Delcamp, G. Ortiz, and F. Verstraete, Dualities in One-Dimensional Quantum Lattice Mod- els: Symmetric Hamiltonians and Matrix Product Op- erator Intertwiners, PRX Quantum4, 020357 (2023), arXiv:2112.09091 [quant-ph]

    Pith/arXiv arXiv 2023

  17. [17]

    Lootens, C

    L. Lootens, C. Delcamp, and F. Verstraete, Dual- ities in One-Dimensional Quantum Lattice Models: Topological Sectors, PRX Quantum5, 010338 (2024), arXiv:2211.03777 [quant-ph]

    Pith/arXiv arXiv 2024

  18. [18]

    Seiberg and S.-H

    N. Seiberg and S.-H. Shao, Majorana chain and Ising model - (non-invertible) translations, anomalies, and emanant symmetries, SciPost Phys.16, 064 (2024), arXiv:2307.02534 [cond-mat.str-el]

    Pith/arXiv arXiv 2024

  19. [19]

    Seiberg, S

    N. Seiberg, S. Seifnashri, and S.-H. Shao, Non-invertible symmetries and LSM-type constraints on a tensor product Hilbert space, SciPost Phys.16, 154 (2024), arXiv:2401.12281 [cond-mat.str-el]

    Pith/arXiv arXiv 2024

  20. [20]

    W. Cao, L. Li, M. Yamazaki, and Y. Zheng, Subsystem non-invertible symmetry operators and defects, SciPost Phys.15, 155 (2023), arXiv:2304.09886 [cond-mat.str-el]

    Pith/arXiv arXiv 2023

  21. [21]

    W. Cao, L. Li, and M. Yamazaki, Generating lattice non-invertible symmetries, SciPost Phys.17, 104 (2024), arXiv:2406.05454

    Pith/arXiv arXiv 2024

  22. [22]

    Y. Choi, Y. Sanghavi, S.-H. Shao, and Y. Zheng, Non-invertible and higher-form symmetries in 2+1d lattice gauge theories, SciPost Phys.18, 008 (2025), arXiv:2405.13105 [cond-mat.str-el]

    Pith/arXiv arXiv 2025

  23. [23]

    Gorantla, S.-H

    P. Gorantla, S.-H. Shao, and N. Tantivasadakarn, Tensor Networks for Noninvertible Symmetries in 3+1D and Be- yond, Phys. Rev. X15, 041006 (2025), arXiv:2406.12978 [quant-ph]

    arXiv 2025

  24. [24]

    Okada and Y

    M. Okada and Y. Tachikawa, Noninvertible Symmetries Act Locally by Quantum Operations, Phys. Rev. Lett. 133, 191602 (2024), arXiv:2403.20062 [hep-th]

    Pith/arXiv arXiv 2024

  25. [25]

    Kennedy and H

    T. Kennedy and H. Tasaki, HiddenZ 2 ×Z 2 symmetry breaking in Haldane-gap antiferromagnets, Phys. Rev. B 45, 304 (1992)

    1992

  26. [26]

    L. Li, M. Oshikawa, and Y. Zheng, Noninvertible duality transformation between symmetry-protected topological and spontaneously symmetry-broken phases, Phys. Rev. B108, 214429 (2023), arXiv:2301.07899

    Pith/arXiv arXiv 2023

  27. [27]

    L. Li, M. Oshikawa, and Y. Zheng, Intrinsically/purely gapless-SPT from non-invertible duality transformations, SciPost Phys.18, 153 (2025), arXiv:2307.04788 [cond- mat.str-el]

    Pith/arXiv arXiv 2025

  28. [28]

    Parayil Mana, Y

    A. Parayil Mana, Y. Li, H. Sukeno, and T.-C. Wei, Kennedy-Tasaki transformation and noninvertible sym- metry in lattice models beyond one dimension, Phys. Rev. B109, 245129 (2024), arXiv:2402.09520 [cond- mat.str-el]

    Pith/arXiv arXiv 2024

  29. [29]

    Bhardwaj, L

    L. Bhardwaj, L. E. Bottini, D. Pajer, and S. Schafer- Nameki, The club sandwich: Gapless phases and phase transitions with non-invertible symmetries, SciPost Phys. 18, 156 (2025), arXiv:2312.17322 [hep-th]

    Pith/arXiv arXiv 2025

  30. [30]

    Seifnashri and S.-H

    S. Seifnashri and S.-H. Shao, Cluster State as a Non- invertible Symmetry-Protected Topological Phase, Phys. Rev. Lett.133, 116601 (2024), arXiv:2404.01369 [cond- mat.str-el]

    arXiv 2024

  31. [31]

    Li, R.-Z

    L. Li, R.-Z. Huang, and W. Cao, Noninvertible symmetry-enriched quantum critical point, Phys. Rev. B 112, L081113 (2025), arXiv:2411.19034 [cond-mat.str-el]

    Pith/arXiv arXiv 2025

  32. [32]

    Roberts, B

    S. Roberts, B. Yoshida, A. Kubica, and S. D. Bartlett, Symmetry-protected topological order at nonzero temperature, Phys. Rev. A96, 022306 (2017), arXiv:1611.05450

    Pith/arXiv arXiv 2017

  33. [33]

    Ma and C

    R. Ma and C. Wang, Average Symmetry-Protected Topological Phases, Phys. Rev. X13, 031016 (2023), arXiv:2209.02723 [cond-mat.str-el]

    Pith/arXiv arXiv 2023

  34. [34]

    de Groot, A

    C. de Groot, A. Turzillo, and N. Schuch, Symmetry Pro- tected Topological Order in Open Quantum Systems, Quantum6, 856 (2022), arXiv:2112.04483 [quant-ph]

    Pith/arXiv arXiv 2022

  35. [35]

    L. Li, Z. Bi, and W. Cao, Generalized symmetry- protected topological phases in mixed states from gaug- ing dualities, arXiv preprint (2026), arXiv:2603.17282

    arXiv 2026

  36. [36]

    Raussendorf, S

    R. Raussendorf, S. Bravyi, and J. Harrington, Long- range quantum entanglement in noisy cluster states, Phys. Rev. A71, 062313 (2005)

    2005

  37. [37]

    Its anomalous action at an open boundary therefore 7 does not provide a symmetric finite-depth trivialization of the SPT phase

    AlthoughUis finite-depth and commutes with the global symmetry (U SaU † =S a), it is not symmetric gate by gate: its individual controlled-Zgates break the symme- try. Its anomalous action at an open boundary therefore 7 does not provide a symmetric finite-depth trivialization of the SPT phase. [38]Z (0) 2 ×Z (0) 2 in (1+1)d andZ (1) 2 ×Z (1) 2 in (3+1)d

  38. [39]

    Tantivasadakarn, R

    N. Tantivasadakarn, R. Thorngren, A. Vishwanath, and R. Verresen, Pivot Hamiltonians as generators of sym- metry and entanglement, SciPost Phys.14, 012 (2023), 2110.07599

    Pith/arXiv arXiv 2023

  39. [40]

    Tantivasadakarn, R

    N. Tantivasadakarn, R. Thorngren, A. Vishwanath, and R. Verresen, Building models of topological quantum crit- icality from pivot Hamiltonians, SciPost Phys.14, 013 (2023), 2110.09512

    Pith/arXiv arXiv 2023

  40. [41]

    L. Li, M. Oshikawa, and H. Yan, Generalized Kramers- Wanier Duality from Bilinear Phase Map, Phys. Rev. Lett.136, 240403 (2026), arXiv:2403.16017 [cond- mat.str-el]

    Pith/arXiv arXiv 2026

  41. [42]

    J. Kim, Y. You, and J. H. Han, Noninvertible symmetry and topological holography for modulated SPT in one di- mension, SciPost Phys.19, 110 (2025), arXiv:2507.02324

    arXiv 2025

  42. [43]

    Pollmann and A

    F. Pollmann and A. M. Turner, Detection of symmetry- protected topological phases in one dimension, Phys. Rev. B86, 125441 (2012), arXiv:1204.0704

    Pith/arXiv arXiv 2012

  43. [44]

    Pfeuty, The one-dimensional Ising model with a trans- verse field, Ann

    P. Pfeuty, The one-dimensional Ising model with a trans- verse field, Ann. Phys. (N.Y.)57, 79 (1970)

    1970

  44. [45]

    Barouch and B

    E. Barouch and B. M. McCoy, Statistical mechanics of the XY model. II. spin-correlation functions, Phys. Rev. A3, 786 (1971)

    1971

  45. [46]

    F. J. Wegner, Duality in generalized Ising models and phase transitions without local order parameters, J. Math. Phys.12, 2259 (1971)

    1971

  46. [47]

    A. M. Ferrenberg, J. Xu, and D. P. Landau, Push- ing the limits of monte carlo simulations for the three- dimensional ising model, Phys. Rev. E97, 043301 (2018), arXiv:1806.03558

    Pith/arXiv arXiv 2018

  47. [48]

    D. A. Reiss and K. P. Schmidt, Quantum robustness and phase transitions of the 3D toric code in a field, SciPost Phys.6, 078 (2019)

    2019

  48. [49]

    Balian, J

    R. Balian, J. M. Drouffe, and C. Itzykson, Gauge fields on a lattice. III. Strong-coupling expansions and transition points, Phys. Rev. D11, 2104 (1975)

    1975

  49. [50]

    Creutz, L

    M. Creutz, L. Jacobs, and C. Rebbi, Experiments with a gauge-invariant Ising system, Phys. Rev. Lett.42, 1390 (1979)

    1979

  50. [51]

    Kaidi, K

    J. Kaidi, K. Ohmori, and Y. Zheng, Symmetry TFTs for Non-invertible Defects, Commun. Math. Phys.404, 1021 (2023), arXiv:2209.11062 [hep-th]

    Pith/arXiv arXiv 2023

  51. [52]

    Bhardwaj, L

    L. Bhardwaj, L. E. Bottini, S. Schafer-Nameki, and A. Tiwari, Non-invertible symmetry webs, SciPost Phys. 15, 160 (2023), arXiv:2212.06842 [hep-th]

    Pith/arXiv arXiv 2023

  52. [53]

    Bhardwaj, S

    L. Bhardwaj, S. Schafer-Nameki, and A. Tiwari, Unifying constructions of non-invertible symmetries, SciPost Phys. 15, 122 (2023), arXiv:2212.06159 [hep-th]

    Pith/arXiv arXiv 2023

  53. [54]

    Diatlyk, C

    O. Diatlyk, C. Luo, Y. Wang, and Q. Weller, Gaug- ing non-invertible symmetries: topological interfaces and generalized orbifold groupoid in 2d QFT, JHEP03, 127, arXiv:2311.17044 [hep-th]

    arXiv

  54. [55]

    Seifnashri, S.-H

    S. Seifnashri, S.-H. Shao, and X. Yang, Gauging non- invertible symmetries on the lattice, SciPost Phys.19, 063 (2025), arXiv:2503.02925 [cond-mat.str-el]

    Pith/arXiv arXiv 2025

  55. [56]

    Balian, J

    R. Balian, J. M. Drouffe, and C. Itzykson, Gauge fields on a lattice. II. Gauge-invariant Ising model, Phys. Rev. D11, 2098 (1975)

    2098

  56. [57]

    Fradkin and L

    E. Fradkin and L. Susskind, Order and disorder in gauge systems and magnets, Phys. Rev. D17, 2637 (1978)

    1978

  57. [58]

    F. Kos, D. Poland, D. Simmons-Duffin, and A. Vichi, Precision islands in the Ising andO(N) models, JHEP 08(08), 036, arXiv:1603.04436

    Pith/arXiv arXiv

  58. [59]

    Koteck´ y and D

    R. Koteck´ y and D. Preiss, Cluster expansion for abstract polymer models, Commun. Math. Phys.103, 491 (1986)

    1986

  59. [60]

    Fredenhagen and M

    K. Fredenhagen and M. Marcu, Charged states inZ 2 gauge theories, Commun. Math. Phys.92, 81 (1983)

    1983

  60. [61]

    Barash, A

    L. Barash, A. Babakhani, and I. Hen, Quantum Monte Carlo algorithm for arbitrary spin- 1 2 Hamiltonians, Phys. Rev. Research6, 013281 (2024)

    2024

  61. [62]

    S. M. Linsel and L. Pollet, ParaToric 1.0: Continuous- time quantum Monte Carlo for the toric code in a parallel field (2025), arXiv:2510.14781

    Pith/arXiv arXiv 2025

  62. [63]

    F. Wu, Y. Deng, and N. Prokof’ev, Phase diagram of the toric code model in a parallel magnetic field, Phys. Rev. B85, 195104 (2012). 8 Supplemental Material Distinct finite-temperature phase diagrams of non-invertible Kennedy–Tasaki duals Weiguang Cao and Haruki Watanabe This Supplemental Material supplies the derivations and numerical details underlying ...

    2012

  63. [64]

    Conjugation byUtransfers both relations to ˜D=U DU †, establishing ˜D† ˜D= ˜D ˜D† =P +

    The symmetry of the bilinear kernel (−1) σ·M σ′ under σ↔σ ′ also givesD=D †, and henceDD † =D †D=P +. Conjugation byUtransfers both relations to ˜D=U DU †, establishing ˜D† ˜D= ˜D ˜D† =P +. Multiplying ˜D H(s) = ˜H(s) ˜Don the right by ˜D† and taking the trace gives the projected identity used in the Letter. The one-form case follows by the same algebra, ...

  64. [65]

    ,2L− 1 around the ring, vertices on evenjand links on oddj

    The unified chain Relabel the 2Lqubits by a single indexj= 0, . . . ,2L− 1 around the ring, vertices on evenjand links on oddj. Every dressing takes the uniform formB j =Z j−1Zj+1 (indices mod 2L), and H(s) =− 2L−1X j=0 h (1−s)Z j−1XjZj+1 +s Xj i .(S2) The one-site translationj→j+ 1 is a symmetry of Eq. (S2) that exchanges the vertex and link sub- 9 latti...

  65. [66]

    Direct multiplication gives the identities needed below

    Jordan–Wigner transformation and the boundary terms Define 2Lpairs of Majorana operators (Y j ≡iX jZj), γ2j = Y k<j Xk Zj, γ 2j+1 =− Y k<j Xk Yj,(S3) which are Hermitian and obey{γ a, γb}= 2δ ab. Direct multiplication gives the identities needed below. On any site the string commutes withZ j, Yj and squares to one; usingZ jYj =−iX j, Xj =−i γ2jγ2j+1.(S4) ...

  66. [67]

    Parity–boundary-condition locking Since [P, H(s)] = 0, the Hilbert space splits into sec- torsP=p=±1, and within a sector the wrapping factor (−P) is the number (−p). Defining the antiperi- odic continuationγ a+4L ≡ −p γa, the Hamiltonian takes the uniform quadratic form H(s) P=p =i 2L−1X j=0 h s γ2jγ2j+1 + (1−s)γ 2j−1γ2j+2 i , (S8) withantiperiodic(Neveu...

  67. [68]

    (S8) connect Majorana indices a→a+ 1 (aeven) anda→a+ 3 (aodd); both pre- serveamod 4∈ {0,1}versus{2,3}

    Two decoupled chains and their diagonalization The couplings in Eq. (S8) connect Majorana indices a→a+ 1 (aeven) anda→a+ 3 (aodd); both pre- serveamod 4∈ {0,1}versus{2,3}. The model therefore decouples into two independent Majorana chains—chain V, built from the even-site Majoranas{γ 4m, γ4m+1}, and chain Λ, from the odd-site Majoranas{γ 4m+2, γ4m+3}, m= ...

  68. [69]

    For the interpola- tionH(s) = (1−s)H SPT +s Hpara, ∂H ∂s =H para −H SPT = X i Mi ≡ M.(S34) TheU-odd order parameter is the field conjugate tos, so Hellmann–Feynman gives⟨M⟩=dE 0/ds

    Hellmann–Feynman criterion for duality breaking The relation between the duality order parameter and the order of the transition follows from Hellmann– Feynman and holds in any dimension. For the interpola- tionH(s) = (1−s)H SPT +s Hpara, ∂H ∂s =H para −H SPT = X i Mi ≡ M.(S34) TheU-odd order parameter is the field conjugate tos, so Hellmann–Feynman gives...

  69. [70]

    (S40) ats= 0 and the self-dual points= 1 2 atT= 0

    Endpoints, evenness, and the exact initial slope The endpoints ofT c(s) are Eq. (S40) ats= 0 and the self-dual points= 1 2 atT= 0. Its initial slope also follows exactly. The conjugationC= Q l Zl obeys CB pC=B p andCX lC=−X l, soC ˜H(s)Creverses the field and makes the spectrum even inh x. Writing ˜H(s) = (1−s) −P p Bp−hx P l Xl rescales the coupling to u...

  70. [71]

    The two arms and the persistence of first order The order of the transition differs between the two arms because the effective dimension changes. Wher- ever the thermal deconfinement boundary is continuous, the finite imaginary-time direction leaves the universal- ity class of the three-dimensional classicalZ 2 gauge the- ory, or equivalently the three-di...

  71. [72]

    Writing their projected free- energy densities asf dec,+ andf conf,+ , fdec,+(s, T) =fconf,+ (1−s, T),(S43) so, whenever both branches exist, they coexist ats= 1 2

    Asymptotic pinning of the wall and the scale ofT ∗ Wegner self-duality acts exactly on the all-plus full- fusion sector selected byP + and exchanges the decon- fined and confined branches. Writing their projected free- energy densities asf dec,+ andf conf,+ , fdec,+(s, T) =fconf,+ (1−s, T),(S43) so, whenever both branches exist, they coexist ats= 1 2 . Th...

  72. [73]

    The same suppression is rigorous in the convergent high-temperature regime derived below

    The twisted membrane across the(s, T)plane The twisted membrane obeys an exact area law at the fixed points,⟨S (1) Σ ⟩tw =∓tanh |Σ|βwith|Σ| ∼ L2. The same suppression is rigorous in the convergent high-temperature regime derived below. In the remain- ing accessible finite-size window, permutation-matrix- representation QMC of the cluster interpolation [61...

  73. [74]

    By contrast, the plaquette families{X pBp} and{X p}do not commute with both link families and en- ter below through virtual dressing

    The frozen manifold On the SPT side, the two competinglinkfamilies of H(s) commute: [X lBl, Xl′] = 0 for alll, l ′, because the dressingB l acts on plaquette qubits whereasX l′ acts on link qubits. By contrast, the plaquette families{X pBp} and{X p}do not commute with both link families and en- ter below through virtual dressing. The commuting sta- bilize...

  74. [75]

    For everyTbetween 0.10 and 0.20 and for bothL= 3 andL= 4, the ordered states decay onto the frozen branch

    Equilibrium at the self-dual point We performed 2×10 6-sweep runs at fixeds= 1 2 , initialized in the frozen configuration and in either or- dered phase. For everyTbetween 0.10 and 0.20 and for bothL= 3 andL= 4, the ordered states decay onto the frozen branch. Where mixing is active, tunnel- ing occurs in both directions; for example, one run at T= 0.12 a...

  75. [76]

    The same heuristic bal- ance places the frozen region in the wedge|s− 1 2 |< δs(T) = (T∆s res −∆e)/m ord, bounded by two ordered– frozen first-order lines

    The wedge and its boundary Away from the self-dual line, the ordered free ener- gies decrease at ratem ord =|∂ sford| ≈0.55, whereas the frozen branch remains flat. The same heuristic bal- ance places the frozen region in the wedge|s− 1 2 |< δs(T) = (T∆s res −∆e)/m ord, bounded by two ordered– frozen first-order lines. This relation is an estimate, not a ...

  76. [77]

    3(a) of the Letter Draggedschains produce three-state hysteresis loops; the inset of Fig

    Hysteresis and the data of Fig. 3(a) of the Letter Draggedschains produce three-state hysteresis loops; the inset of Fig. 3(a) of the Letter shows theT= 0.16, L= 8 example. Each ordered branch first decays onto the frozen plateau and only later converts into the op- posite ordered phase. On the plateau, both link polar- izations are large and nearly equal...

  77. [78]

    Its competing terms share link qubits and anticommute,{B p, Xl}= 0 forl∈∂p

    Exclusion on the dual side Nodirectly saturatedcounterpart of the frozen branch exists for ˜H(s). Its competing terms share link qubits and anticommute,{B p, Xl}= 0 forl∈∂p. Consequently every state at every temperature obeys⟨Bp⟩2+⟨Xl⟩2 ≤1, whereas direct saturation would require both expecta- tions to approach unity. This local bound excludes a saturated...

  78. [79]

    HereH off =−h x P l Xl is the off-diagonal transverse- field term, andn kink is the mean number of field ver- tices in the worldline configuration

    The second-order arm At fixed transverse fieldh x =s/(1−s) the temper- ature of the unit-coupling model is scanned across the transition, and ˜Tc is located by the peak of the specific heat per link computeddirectlyas the energy fluctuation with its continuous-time kink-number correction, cV ≡ C Nl = β2 ⟨E2⟩ − ⟨E⟩2 −n kink Nl , N l = 3L3, (S55) nkink =−β⟨...

  79. [80]

    The order of the transition along the boundary We diagnose the order of the boundary from the vol- ume scaling of the specific-heat maximum in Eq. (S55). At a first-order transition, the intensive peak grows as the volume,c V,max ∝N l, and thereforec V,max(L= 16)/cV,max(L= 8) = 8. At a three-dimensional Ising transition it grows only asL α/ν, withα/ν≃0.17...

  80. [81]

    The first-order segment and the tricritical point We probe the first-order segment with ParaToric two- branch scans. At each temperatureTand fields(grid ∆s= 0.01,L= 12, four seeds) we run two inde- pendent simulations that ramp from a deconfined start (htherm = 0.6) and a confined start (h therm = 1.5) to the target fieldh=s/(1−s) and then equilibrate—a 6...

Showing first 80 references.