Entanglement order parameters and critical behavior for topological phase transitions and beyond

TL;DR

This paper introduces a unified tensor network framework using entanglement order parameters to analyze and characterize both topological and conventional phase transitions, revealing universal critical behavior.

Contribution

It develops a novel entanglement-based order parameter framework that quantitatively probes topological and conventional phase transitions, identifying new critical exponents.

Findings

Identified 3D Ising critical exponents for the Toric Code transition.

Discovered an unknown critical exponent beta=0.021 for topological transitions.

Measured a new critical exponent beta=0.024 in the 2+1D transverse field Ising model.

Abstract

Topological phases are exotic quantum phases which are lacking the characterization in terms of order parameters. In this paper, we develop a unified framework based on variational iPEPS for the quantitative study of both topological and conventional phase transitions through entanglement order parameters. To this end, we employ tensor networks with suitable physical and/or entanglement symmetries encoded, and allow for order parameters detecting the behavior of any of those symmetries, both physical and entanglement ones. First, this gives rise to entanglement-based order parameters for topological phases. These topological order parameters allow to quantitatively probe topological phase transitions and to identify their universal behavior. We apply our framework to the study of the Toric Code model in different magnetic fields, which in some cases maps to the (2+1)D Ising model. We identify 3D Ising critical exponents for the entire transition, consistent with those special cases and general belief. However, we moreover find an unknown critical exponent beta=0.021. We then apply our framework of entanglement order parameters to conventional phase transitions. We construct a novel type of disorder operator (or disorder parameter), which is non-zero in the disordered phase and measures the response of the wavefunction to a symmetry twist in the entanglement. We numerically evaluate this disorder operator for the (2+1)D transverse field Ising model, where we again recover a critical exponent hitherto unknown in the model, beta=0.024, consistent with the findings for the Toric Code. This shows that entanglement order parameters can provide additional means of characterizing the universal data both at topological and conventional phase transitions, and altogether demonstrates the power of this framework to identify the universal data underlying the transition.