Exploring the Moiré Landscape of Graphene–Perovskite Oxide Membranes with Machine Learning

The combination of graphene, with its sensitivity to the physical environment, and perovskite oxides, with their wide range of tunable properties, make graphene-perovskite oxide heterostructures a potentially rich playground for materials science and applications [1]. Furthermore, the mechanical and tribological properties of graphene could make it useful as a capping or lubricating layer when creating novel oxide interfaces. [2] Here, we aim to expand the knowledge of graphene-perovskite oxide interfaces utilizing computational modelling. Traditionally, one would resort to density functional theory (DFT) calculations to understand the interaction between the materials. However, the significant difference between the graphene and oxide crystals poses a numerical challenge since large supercells are required to accommodate the two materials. To tackle this, we employ state-of-the-art machine learning potentials [3] to study the atomic structure of graphene-perovskite oxide interfaces. These highly efficient and accurate potentials enable the study of large graphene-perovskite oxide heterostructure domains, not only in the static, zero-temperature limit, but also as a function of temperature using molecular dynamics with near DFT accuracy.
In this work, we consider graphene placed on TiO₂ terminated (001) surfaces of SrTiO₃ and BaTiO₃. Simulations of the graphene-perovskite oxide interface reveal a moderately strong van der Waals binding, comparable to that observed in bilayer graphene. However, the sliding and twisting energy landscapes are very shallow due to the incommensurability between the two structures. This indicates that while the two materials adhere to each other, they can move laterally with minimal interlocking. Additionally, we observe the formation of subtle twist-angle-dependent moiré patterns in the relaxed graphene, induced by the presence of the oxide. Our findings reveal unexpected lubrication in the graphene-perovskite system, encouraging further experimental and theoretical studies to better understand the complex interface between these materials.

1. Yang, Allen Jian, et al. "Two-dimensional layered materials meet perovskite oxides: A combination for high-performance electronic devices." ACS Nano 17.11 (2023): 9748-9762.
2. Sánchez-Santolino, G., et al. "A 2D ferroelectric vortex pattern in twisted BaTiO3 freestanding layers." Nature 626.7999 (2024): 529-534.
3. Fan, Zheyong, et al. "Neuroevolution machine learning potentials: Combining high accuracy and low cost in atomistic simulations and application to heat transport." Physical Review B 104.10 (2021): 104309.