Phase diagram of the three-dimensional subsystem toric code

• Published in: arXiv.org
• Main Authors: Li, Yaodong, von Keyserlingk, C W, Zhu, Guanyu, Jochym-O'Connor, Tomas
• Format: Paper Journal Article
• Language: English
• Published: Ithaca Cornell University Library, arXiv.org 25.08.2024
• Subjects: Boundary conditions; Codes; Commuting; Error analysis; Error correcting codes; Error correction; Error correction & detection; Fault tolerance; Gauge theory; Phase diagrams; Phases; Physics - Quantum Physics; Physics - Statistical Mechanics; Subsystems; Thermal stability
• ISSN: 2331-8422
• DOI: 10.48550/arxiv.2305.06389

Subsystem quantum error-correcting codes typically involve measuring a sequence of non-commuting parity check operators. They can sometimes exhibit greater fault-tolerance than conventional subspace codes, which use commuting checks. However, unlike subspace codes, it is unclear if subsystem codes -- in particular their advantages -- can be understood in terms of ground state properties of a physical Hamiltonian. In this paper, we address this question for the three-dimensional subsystem toric code (3D STC), as recently constructed by Kubica and Vasmer [Nat. Comm. 13, 6272(2022)], which exhibits single-shot error correction (SSEC). Motivated by a conjectured relation between SSEC and thermal stability, we study the zero and finite temperature phases of an associated non-commuting Hamiltonian. By mapping the Hamiltonian model to a pair of 3D Z_2 gauge theories coupled by a kinetic constraint, we find various phases at zero temperature, all separated by first-order transitions: there are 3D toric code-like phases with deconfined point-like excitations in the bulk, and there are phases with a confined bulk supporting a 2D toric code on the surface when appropriate boundary conditions are chosen. The latter is similar to the surface topological order present in 3D STC. However, the similarities between the SSEC in 3D STC and the confined phases are only partial: they share the same sets of degrees of freedom, but they are governed by different dynamical rules. Instead, we argue that the process of SSEC can more suitably be associated with a path (rather than a point) in the zero-temperature phase diagram, a perspective which inspires alternative measurement sequences enabling SSEC. Moreover, since none of the above-mentioned phases survives at nonzero temperature, SSEC of the code does not imply thermal stability of the associated Hamiltonian phase.