2002/07/31 by Chenyang Wang, Jim Harrington, John Preskill · 2 citations
Physics and Astronomy · #quant-ph #cond-mat.stat-mech #hep-lat #hep-th
paper · pdf · doi:10.1016/s0003-4916(02)00019-2
published as Annals of Physics 303, 31-58 (2003) · 16 pages, 11 figures, REVTeX, improved numerics and an additional author
arxiv created 2002/09/26 · arxiv updated 2009/11/30
We study the +/- J random-plaquette Z2 gauge model (RPGM) in three spatial dimensions, a three-dimensional analog of the two-dimensional +/- J random-bond Ising model (RBIM). The model is a pure Z2 gauge theory in which randomly chosen plaquettes (occuring with concentration p) have couplings with the ``wrong sign'' so that magnetic flux is energetically favored on these plaquettes. Excitations of the model are one-dimensional ``flux tubes'' that terminate at ``magnetic monopoles.'' Electric confinement can be driven by thermal fluctuations of the flux tubes, by the quenched background of magnetic monopoles, or by a combination of the two. Like the RBIM, the RPGM has enhanced symmetry along a ``Nishimori line'' in the p-T plane (where T is the temperature). The critical concentration pc of wrong-sign plaquettes at the confinement-Higgs phase transition along the Nishimori line can be identified with the accuracy threshold for robust storage of quantum information using topological error-correcting codes: if qubit phase errors, qubit bit-flip errors, and errors in the measurement of local check operators all occur at rates below pc, then encoded quantum information can be protected perfectly from damage in the limit of a large code block. Numerically, we measure pc0, the critical concentration along the T=0 axis (a lower bound on pc), finding pc0=.0293 +/- .0002. We also measure the critical concentration of antiferromagnetic bonds in the two-dimensional RBIM on the T=0 axis, finding pc0=.1031 +/-.0001. Our value of pc0 is incompatible with the value of pc=.1093 +/-.0002 found in earlier numerical studies of the RBIM, in disagreement with the conjecture that the phase boundary of the RBIM is vertical (parallel to the T axis) below the Nishimori line.