Abstract
We study the mean-field broken symmetry phases of charge neutral multilayer rhombohedral graphene within tight-binding approximations including self-consistent extended Hubbard interactions. We used on-site and inter-site Hubbard interactions obtained from a newly developed first-principles calculation method. Our calculations for systems up to eight layers give rise to electron-hole asymmetries, band flatness, band gaps, and layer anti-ferromagnetic ground states in keeping with available experiments. By including the intersite Hubbard interactions up to the next-nearest neighboring sites, the band gaps are shown to open when the number of layers is larger than three, while the trilayer system maintains its metallic nature with two low energy density of state peaks near the Fermi energy whose separation increases with the range of inter-site Hubbard parameters. Within our framework, the calculated band gaps reflect mean-field ground states with extended Hubbard interactions, in closer agreement with experimental estimates. The tight-binding formulation further enables efficient treatment of large rhombohedral chiral systems, including twisted multilayer graphene.
Introduction
Graphene is renowned for its Dirac cones, where electrons behave as massless particles. However, in rhombohedral (ABC) stacking, the electronic bands become nearly flat, causing electron-electron interactions (Coulomb repulsion) to dominate. This leads to broken symmetry phases such as magnetism or superconductivity. Modeling these systems accurately is a significant challenge: standard tight-binding (TB) models often oversimplify interactions, while first-principles density functional theory (DFT) is computationally expensive for many layers and frequently fails to predict experimental energy gaps. This research bridges the gap by introducing a self-consistent TB+U+V model that is both physically rigorous and computationally efficient.
Methodology
The researchers developed a self-consistent TB+U+V framework that integrates first-principles accuracy into a tight-binding solver. The process begins with a newly developed $DFT+U+V$ method to calculate on-site ($U$) and inter-site ($V$) repulsion parameters. These are then mapped onto a TB grid using Wannierization. A self-consistency loop iteratively updates the electron density until a stable ground state is reached. To ensure accuracy near the Dirac points, an adaptive dense grid sampling technique was used to capture subtle features like trigonal warping.
Key Results and Insights
The study of rhombohedral multilayer graphene (RnG) for $n=1$ to $8$ layers yielded several critical insights:
- Band Gap Emergence: A significant band gap only opens for systems with four or more layers ($n \ge 4$). Trilayer graphene (R3G) remains metallic but exhibits a unique splitting in its density-of-states peaks.
- Layer Antiferromagnetism (LAF): The ground state for these systems is predicted to be a layer antiferromagnet, where the top and bottom surfaces of the stack have opposite magnetic alignments.
- Bandwidth Asymmetry: The conduction band remains nearly flat as layers are added, while the valence band width increases linearly. This asymmetry explains the differing behaviors of electrons and holes in these materials.
- Experimental Alignment: By including interactions up to the next-nearest neighbors ($V_2$), the model predicts a gap of ~12 meV for 4-layer graphene, saturating at ~18 meV for 8 layers. This is in much closer agreement with experimental reports (10–15 meV) than previous models.
- Phase Stability: The researchers estimated the critical temperatures for these magnetic phases, finding that order persists up to 38 K for 4 layers and 65 K for 8 layers.
Conclusion and Future Work
The TB+U+V framework successfully reconciles theoretical predictions with experimental observations for rhombohedral graphene. The efficiency of this approach makes it a prime candidate for studying large-scale moiré materials, such as twisted multilayer graphene, where traditional first-principles methods are computationally impractical. Future refinements may include longer-range interactions and explicit interlayer terms to further explore the subtle physics of trilayer and twisted systems.
Summary updated