Confinement-Higgs transition in the random-plaquette Z2 gauge model: Nishimori-line accuracy threshold, anisotropic (p,q) boundary, and the reentrant-versus-vertical phase-boundary question for topological quantum memory

The optimal (maximum-likelihood) recovery of the two-dimensional toric/surface code from qubit and syndrome-measurement errors maps onto the three-dimensional random-plaquette $Z_2$ gauge model (RPGM), whose confinement-Higgs phase boundary crosses the disorder-enhanced-symmetry Nishimori line at a multicritical point equal to the code's accuracy threshold. Wang, Harrington and Preskill (2003) computed only the zero-temperature lower bound $p_{\mathrm{c0}} = 0.0293$ and left open: the value of the true threshold $p_{\mathrm{c}}$ on the Nishimori line; the finite-temperature failure probability of the optimal free-energy-minimizing recovery; the anisotropic threshold surface for unequal qubit and measurement error rates; whether the phase boundary below the Nishimori line is reentrant or vertical; and reliable methods for higher-dimensional variants whose ground states are NP-hard to compute.

Background

Kitaev's toric code and its planar version, the surface code, are two-dimensional quantum error-correcting codes in which physical qubits sit on the edges of a lattice and are protected by repeated measurements of local stabilizer checks (products of Pauli operators around vertices and plaquettes). In the 'phenomenological' noise model, in each round every qubit suffers an independently distributed bit-flip (or phase-flip) error with probability $p$ and every syndrome bit (check-measurement outcome) is faulty with probability $q$. Decoding means inferring a recovery from the accumulated syndrome history. Since two error patterns have the same effect on the encoded information whenever they differ by a stabilizer (they lie in the same 'homology class'), the optimal, maximum-likelihood decoder picks the most likely homology class of error chains rather than the most likely single error pattern.

Dennis, Kitaev, Landahl and Preskill showed that this decoding problem maps onto classical statistical mechanics with quenched disorder: the syndrome history over $L$ rounds maps to the three-dimensional random-plaquette $Z_2$ gauge model (RPGM), a lattice gauge theory in which a fraction $p$ of plaquettes carries a quenched (fixed from sample to sample) 'wrong-sign' coupling that energetically favors nontrivial $Z_2$ magnetic flux. Its excitations are one-dimensional flux tubes terminating at magnetic monopoles, and the ordered ('Higgs') and disordered ('confinement') phases are separated by a boundary along which electric charge becomes deconfined, driven by thermal fluctuations of flux tubes, by the quenched wrong-sign background, or both. The optimal recovery succeeds asymptotically exactly in the ordered phase, so the accuracy threshold equals the critical disorder $p_{\mathrm{c}}$ at which the confinement-Higgs boundary crosses the Nishimori line, defined by $e^{-2\beta J}=p/(1-p)$ with $\beta$ the inverse temperature and $J$ the coupling strength — the locus along which the Boltzmann weight of thermal fluctuations matches the quenched disorder distribution, so that the statistical model computes exactly the posterior probabilities of the true error model. In the perfect-measurement limit $q=0$ the corresponding model is the two-dimensional random-bond Ising model (RBIM, bonds with quenched $\pm J$ couplings), whose multicritical point on the Nishimori line — the crossing point at which two independent critical directions meet — is known numerically to high precision ($p_{\mathrm{c}}\approx 0.1092$).

At $T=0$ the optimal recovery reduces to finding the minimum-weight error chain consistent with the syndrome, solvable in two dimensions in polynomial time by Edmonds' minimum-weight perfect-matching algorithm, a classical polynomial-time algorithm for weighted matching on graphs. The zero-temperature critical disorder $p_{\mathrm{c0}}$ ($0.0293\pm 0.0002$ for the RPGM and $0.1031\pm 0.0001$ for the RBIM, from the source paper) is then a rigorous lower bound on $p_{\mathrm{c}}$. The source paper also sketched a finite-temperature computation of the failure probability $P_{\mathrm{fail}}(p,T)$ of the free-energy-minimizing recovery, via Monte Carlo estimators built from ratios of partition functions of different homology sectors, and stated that it had not attempted this calculation. Whether the phase boundary between the Nishimori point and the $T=0$ axis is vertical (so that $p_{\mathrm{c0}}=p_{\mathrm{c}}$) or reentrant ($p_{\mathrm{c0}}<p_{\mathrm{c}}$, so that order is lost, restored, and lost again as temperature is raised at fixed disorder) is a long-standing question originating in the spin-glass literature via an entropy-of-frustration singularity argument for a vertical boundary, discussed and critically assessed in the source paper; for quantum error correction it decides whether cheap ground-state computations already certify the optimal threshold. In higher dimensions the ground-state problem becomes minimal-surface minimization, NP-hard (Barahona) for the RBIM in $d\ge 3$ and the RPGM in $d\ge 4$, and optimal degenerate decoding of general stabilizer codes — decoding that exploits the equivalence of errors differing by a stabilizer — is #P-hard (Iyer and Poulin), a counting-problem complexity class at least as hard as NP, which bounds what exact zero-temperature threshold computations can ever achieve.

Problem Statement

The problem is the complete quantitative and, where possible, rigorous characterization of the confinement-Higgs transition of the disordered spin and gauge models that determine the accuracy threshold of topological quantum memory. All requested components serve one shared objective: convert the identification 'accuracy threshold = multicritical point where the confinement-Higgs boundary crosses the Nishimori line' into a determined and defensible statement about the optimal recovery of the toric/surface code. Concretely, for the three-dimensional random-plaquette $Z_2$ gauge model (RPGM) and its two-dimensional random-bond Ising model (RBIM) companion: (1) determine the multicritical disorder $p_{\mathrm{c}}$ on the Nishimori line of the 3D RPGM — the exact optimal threshold when qubit and syndrome-measurement error rates are equal — together with its statistical and systematic uncertainties and the finite-size-scaling and equilibration analysis that support it; (2) determine the anisotropic threshold surface in the $(p,q)$ plane of qubit versus measurement error rates, where the RPGM carries different disorder strengths on time-like and space-like plaquettes, together with its statistical and systematic uncertainties and the supporting finite-size-scaling and equilibration analysis; (3) settle rigorously, by proof rather than numerical evidence alone, whether the ferromagnetic/confinement-Higgs phase boundary below the Nishimori line is reentrant ($p_{\mathrm{c0}}<p_{\mathrm{c}}$) or vertical ($p_{\mathrm{c0}}=p_{\mathrm{c}}$) for the 2D RBIM and 3D RPGM, and under what conditions (model variants, lattices, dimensions) verticality holds; and (4) for the higher-dimensional cases where the zero-temperature ground-state problem is NP-hard (RBIM in $d\ge 3$, RPGM in $d\ge 4$), establish approximate methods carrying an explicit reliability criterion, or provable bounds, for the optimal recovery threshold. Component (1) is the primary objective. Components (2), (3) and (4) are independent sub-questions.

References

1. Chenyang Wang, Jim Harrington, John Preskill, 'Confinement-Higgs transition in a disordered gauge theory and the accuracy threshold for quantum memory', Annals of Physics 303, 31-58 (2003). DOI: 10.1016/S0003-4916(02)00019-2; arXiv:quant-ph/0207088; https://arxiv.org/abs/quant-ph/0207088
2. Eric Dennis, Alexei Kitaev, Andrew Landahl, John Preskill, 'Topological quantum memory', Journal of Mathematical Physics 43, 4452-4505 (2002). DOI: 10.1063/1.1499754; arXiv:quant-ph/0110143; https://arxiv.org/abs/quant-ph/0110143
3. A. Yu. Kitaev, 'Fault-tolerant quantum computation by anyons', Annals of Physics 303, 2-30 (2003). DOI: 10.1016/S0003-4916(02)00018-0; arXiv:quant-ph/9707021; https://arxiv.org/abs/quant-ph/9707021
4. Takuya Ohno, Gaku Arakawa, Ikuo Ichinose, Tetsuo Matsui, 'Phase structure of the random-plaquette Z_2 gauge model: accuracy threshold for a toric quantum memory', Nuclear Physics B 697, 462-480 (2004). DOI: 10.1016/j.nuclphysb.2004.07.003; arXiv:quant-ph/0401101; https://arxiv.org/abs/quant-ph/0401101
5. Koujin Takeda, Hidetoshi Nishimori, 'Self-dual random-plaquette gauge model and the quantum toric code', Nuclear Physics B 686, 377-396 (2004). DOI: 10.1016/j.nuclphysb.2004.03.006; arXiv:hep-th/0310279; https://arxiv.org/abs/hep-th/0310279
6. Martin Hasenbusch, Francesco Parisen Toldin, Andrea Pelissetto, Ettore Vicari, 'Multicritical Nishimori point in the phase diagram of the +/-J Ising model on a square lattice', Physical Review E 77, 051115 (2008). DOI: 10.1103/PhysRevE.77.051115; arXiv:0803.0444; https://arxiv.org/abs/0803.0444
7. Zhou-Quan Wan, Xu-Dong Dai, Guo-Yi Zhu, 'Revisiting Nishimori multicriticality through the lens of information measures', Physical Review Research 8, 023059 (2026). DOI: 10.1103/b8y5-k3y6; arXiv:2511.02907; https://arxiv.org/abs/2511.02907
8. Ruben S. Andrist, Helmut G. Katzgraber, H. Bombin, M. A. Martin-Delgado, 'Error tolerance of topological codes with independent bit-flip and measurement errors', Physical Review A 94, 012318 (2016). DOI: 10.1103/PhysRevA.94.012318; arXiv:1603.08729; https://arxiv.org/abs/1603.08729
9. Pavithran Iyer, David Poulin, 'Hardness of decoding quantum stabilizer codes', IEEE Transactions on Information Theory 61, 5209-5223 (2015). DOI: 10.1109/TIT.2015.2422294; arXiv:1310.3235; https://arxiv.org/abs/1310.3235
10. F. Barahona, 'On the computational complexity of Ising spin glass models', Journal of Physics A 15, 3241-3253 (1982). DOI: 10.1088/0305-4470/15/10/028; https://doi.org/10.1088/0305-4470/15/10/028
11. Manuel Rispler, Davide Vodola, Markus Muller, Seyong Kim, 'The random coupled-plaquette gauge model and the surface code under circuit-level noise', npj Quantum Information 12, 127 (2026). DOI: 10.1038/s41534-026-01271-6; arXiv:2412.14004; https://arxiv.org/abs/2412.14004
12. Zac Tobias, Nikolas P. Breuckmann, Benedikt Placke, 'Optimal Decoding with the Worm' (2026). arXiv:2603.05428; https://arxiv.org/abs/2603.05428
13. Ji-Ze Xu, Yin Zhong, Miguel A. Martin-Delgado, Hao Song, Ke Liu, 'Phenomenological noise models and optimal thresholds of the 3D toric code', Quantum Science and Technology 11, 035022 (2026). DOI: 10.1088/2058-9565/ae799e; arXiv:2510.20489; https://arxiv.org/abs/2510.20489
14. Hidetoshi Nishimori, Masayuki Ohzeki, Manaka Okuyama, 'Temperature chaos as a logical consequence of the reentrant transition in spin glasses', Physical Review E 112, 044140 (2025). DOI: 10.1103/qp1w-qcbs; arXiv:2507.00276; https://arxiv.org/abs/2507.00276
15. Akshat Pandey, Aditya Mahadevan, A. Alan Middleton, Daniel S. Fisher, 'Low-temperature transition of 2d random-bond Ising model and quantum infinite randomness' (2026). arXiv:2603.02308; https://arxiv.org/abs/2603.02308
16. Amirhossein Manouchehri, Kirill Shtengel, 'Percolation of Zero-Weight Paths and the Shape of the Phase Boundary in the Two-Dimensional Random-Bond Ising Model' (2026). arXiv:2606.22398; https://arxiv.org/abs/2606.22398
17. Rohit Mukherjee, Simon Trebst, 'Nishimori Threshold Estimation for Bayesian Inference and Z_q Surface Code Decoding' (2026). arXiv:2607.18374; https://arxiv.org/abs/2607.18374
18. Lucas H. English, Haoyuan Luo, Yangming Wang, Basudha Srivastava, Stephen D. Bartlett, Dominic J. Williamson, 'Duality constrains optimal thresholds in quantum error correction' (2026). arXiv:2607.21160; https://arxiv.org/abs/2607.21160
19. Yuanqi Liu, Weilei Zeng, Junjie Wu, Lingling Lao, 'Approximate maximum-likelihood decoding via truncated free energies' (2026). arXiv:2609.03928; https://arxiv.org/abs/2609.03928