Thermomechanical Properties of Transition Metal Dichalcogenides Predicted by a Machine Learning Parameterized Force...

Area of Science:

• Materials Science
• Computational Physics
• Nanotechnology

Background:

• Transition metal dichalcogenides (TMDs) possess crucial mechanical and thermal properties for electronics and thermal management.
• Accurate prediction of these properties is essential for designing advanced materials.

Purpose of the Study:

• To develop a machine learning (ML) approach for parametrizing molecular dynamics (MD) force fields.
• To predict the mechanical and thermal transport properties of various monolayered TMDs.
• To investigate the effects of mechanical strain and grain boundaries on thermal conductivity.

Main Methods:

• Utilized a machine learning approach to train molecular dynamics force fields.
• Employed equilibrium and nonequilibrium MD simulations to calculate thermal conductivity.
• Investigated thermal transport across grain boundaries and the impact of mechanical strain.

Main Results:

• Successfully predicted mechanical and thermal properties of monolayered TMDs (MoS2, MoTe2, WSe2, WS2, ReS2).
• Quantified the influence of small and large mechanical strains on lattice thermal conductivity.
• Analyzed thermal transport phenomena across grain boundaries.

Conclusions:

• The ML-driven MD approach offers a fast and accurate method for computing TMD properties.
• This methodology is particularly beneficial for large-scale and complex material structures.
• Enables efficient characterization of TMDs for diverse technological applications.