Berlin 2024 – scientific programme

Parts | Days | Selection | Search | Updates | Downloads | Help

O: Fachverband Oberflächenphysik

O 29: Poster: 2D Materials

O 29.28: Poster

Tuesday, March 19, 2024, 12:30–14:30, Poster A

Exploring the structure of extended defects in MoS₂ bilayers with a machine-learned ACE potential — •Kevin Dhamo and Bernd Meyer — Interdisciplinary Center for Molecular Materials and Computer Chemistry Center, FAU Erlangen-Nürnberg, Germany

Layered transition metal dichalcogenides (TMDs), for example MoS₂, have gathered much attention due to their unique tunable electrical, optical, thermal, and tribological properties. TMDs are usually applied as stacks of layers, which are prone to include extended defects such as in-plane dislocations or misalignments due to the rotation of subsequent layers. Density-functional theory (DFT) would be the method of choice to study the properties of such extended defects. However, the representation of extended defects often requires very large unit cells, which makes DFT calculations unfeasible.

In this work, we use DFT data to train an atomic cluster expansion (ACE) machine-learned interatomic potential for MoS₂. The capabilities of the ACE potential are benchmarked against DFT calculations of MoS₂ bilayers by comparing binding energy curves and Gamma surfaces for different bilayer stackings. Finally, the ACE potential is applied to study the structural properties of dislocations in MoS₂ bilayers and the atomic relaxations in the moiré pattern of twisted bilayer MoS₂ for different rotation angles.

Keywords: molydenum sulfide bilayer; dislocations; twisted bilayer; machine learning; atomic cluster expansion