Topology Guided Minimal Cluster DMFT for Predicting Momentum Resolved ARPES Spectra of Twisted Bilayer Graphene

Abstract

We propose a topology-guided minimal cluster dynamical mean-field theory (DMFT) for predicting momentum-resolved ARPES spectra of twisted bilayer graphene. The conventional approach to spectral function computation often neglects strong correlation effects or relies on oversimplified self-energy approximations, thereby failing to capture the nontrivial quantum geometry and Mott physics at magic twist angles. Our method addresses this limitation through a three-stage pipeline. First, we construct a minimal set of Wannier orbitals that respect the fragile topological obstruction inherent to the moire flat bands; this is achieved by minimizing the spread functional under constraints that enforce the Wannier centers to lie on the moire lattice, thereby preserving the non-zero Chern number and the quantum metric. Second, we solve a two-orbital impurity cluster DMFT problem using continuous-time quantum Monte Carlo, which accurately captures the on-site Hubbard interaction and the resulting correlation-induced renormalizations. Third, the converged self-energy is embedded into the lattice Green’s function to compute the momentum-resolved spectral function via analytic continuation. The key novelty lies in the topology-preserving Wannier construction, which enables a compact impurity cluster while maintaining the essential band topology. This approach systematically interpolates between different twist angles and interaction strengths, reproducing characteristic flat-band signatures such as van Hove singularity splitting and the emergence of a correlation gap at half-filling. Our method therefore provides a computationally efficient yet physically accurate framework for predicting ARPES spectra in strongly correlated moire systems.

Keywords

• ARPES Spectral Function
• Dynamical Mean-Field Theory
• Fragile Topology
• Quantum Geometry
• Twisted Bilayer Graphene
• Wannier Functions

Article History

How to Cite

BibTeX

@article{liu2026topology,
  title   = {Topology Guided Minimal Cluster DMFT for Predicting Momentum Resolved ARPES Spectra of Twisted Bilayer Graphene},
  author  = {Liu, Janislava and Hsueh, Yina},
  journal = {WIJ Journal of Artificial Intelligence Applications},
  volume  = {2026},
  year    = {2026},
  pages   = {Article 2026001},
  url     = {https://wijpub.org/aiapp/article/2026001}
}

Article Information

Publisher: World Insight Junction (WIJ)
Journal: WIJ Journal of Artificial Intelligence Applications
Publication model: Continuous publication
Volume: 2026
Issue: None