Topological invariants for interacting topological insulators: II. Breakdown of the Green's function formalism

Yuan-Yao He,Han-Qing Wu,Zi Yang Meng,Zhong-Yi Lu
DOI: https://doi.org/10.1103/PhysRevB.93.195164
2015-12-07
Abstract:Topological phase transitions in free fermion systems can be characterized by closing of single-particle gap and change in topological invariants. However, in the presence of electronic interactions, topological phase transitions are more complicated. In paper I of this series (<a class="link-https" data-arxiv-id="1510.07816" href="https://arxiv.org/abs/1510.07816">arXiv:1510.07816</a>), we have developed an efficient scheme to evaluate the topological invariants based on Green's function formalism. Here, in paper II, we demonstrate four interaction-drive topological phase transitions (TPTs) in two-dimensional (2D) interacting topological insulators (TIs) via large-scale quantum Monte Carlo (QMC) simulations, based on the scheme of evaluating topological invariants presented in paper I. Across these transitions, the defining symmetries of the TIs have been neither explicitly nor spontaneously broken. In the first two models, the topological invariants calculated from Green's function formalism succeed in characterizing interaction-driven TPTs. However, in the second two models, we find single-particle gap does not close and the topological invariants constructed from single-particle Green's function acquire no change across the TPTs. Unexpected breakdown of the Green's function formalism in constructing topological invariants is thus discovered. We thence classify the TPTs in interacting TIs into two categories: those have noninteracting correspondence can be characterized successfully by the topological invariants constructed from Green's functions, while for the others that do not have noninteracting correspondence, the Green's function formalism experiences a breakdown but more interesting and exciting phenomena, such as emergent collective critical modes at the transition, arise. Discussion on the success and breakdown of topological invariants constructed from the Green's function formalism in the context of symmetry protected topological (SPT) states is presented.
Strongly Correlated Electrons
What problem does this paper attempt to address?
This paper attempts to address the problem of how topological phase transitions (TPTs) in topological insulators can be characterized by topological invariants constructed from the single-particle Green's function in the presence of electronic interactions. Specifically, the authors proposed an effective scheme based on the single-particle Green's function to evaluate topological invariants in a previous paper. In this paper, the authors investigate several interaction-driven topological phase transitions in 2D interacting topological insulators through large-scale quantum Monte Carlo (QMC) simulations and explore the applicability and limitations of the Green's function form in these transitions. ### Main Issues: 1. **Characterization of Topological Phase Transitions**: How to characterize topological phase transitions through topological invariants constructed from the single-particle Green's function in the presence of electronic interactions. 2. **Limitations of the Green's Function Form**: In some cases, topological invariants constructed from the single-particle Green's function may fail, leading to incorrect characterization of topological phase transitions. ### Research Methods: - **Model Selection**: The study investigated models such as the generalized Kane-Mele-Hubbard model (GKMH), the cluster Kane-Mele-Hubbard model (CKMH), and the bilayer Kane-Mele-Hubbard model (BKMH). - **Numerical Simulations**: Large-scale quantum Monte Carlo (QMC) simulations were used to study interaction-driven topological phase transitions in these models. - **Evaluation of Topological Invariants**: Topological invariants (such as spin Chern numbers) constructed from the single-particle Green's function were used to evaluate the phase transition points. ### Main Findings: 1. **Successful Cases**: In some models, topological invariants constructed from the single-particle Green's function successfully characterized topological phase transitions, such as in the GKMH and CKMH models. 2. **Failure Cases**: In other models, topological invariants constructed from the single-particle Green's function failed to correctly characterize topological phase transitions, such as in the BKMH model. 3. **Classification**: The authors classified interaction-driven topological phase transitions into two categories: one with a non-interacting correspondence that can be successfully characterized by the Green's function form, and another without a non-interacting correspondence where the Green's function form fails. ### Conclusion: - **Applicability of the Green's Function Form**: For phase transitions with a non-interacting correspondence, topological invariants constructed from the single-particle Green's function can successfully characterize the transitions. - **Need for New Methods**: For phase transitions without a non-interacting correspondence, new methods and techniques need to be developed to more comprehensively understand and characterize topological phase transitions. This paper, through detailed research and analysis, reveals the complexity of topological phase transitions in the presence of electronic interactions and the limitations of the Green's function form, providing important references for further research on interacting topological insulators.