Phonon thermal transport in two-dimensional PbTe monolayers via extensive molecular dynamics simulations with a neuroevolution potential

Two-dimensional (2D) PbTe monolayers as newly fabricated thermoelectric materials have sparked great interest due to their excellent physical properties, which are expected to play an essential role in converting waste heat energy into electrical energy. Herein, it is imperative to have a clear and comprehensive understanding of the thermal properties of 2D PbTe monolayers, as this is critical for their practical applications. Molecular dynamics (MD) simulations are widely employed to predict physical properties at the microscopic scale and are particularly suitable for evaluating phonon thermal conductivity. Generally, predicting the thermal conductivity of 2D materials is a routine task through MD simulations when appropriate interatomic potentials exist. However, the existing interatomic potential for PbTe allotropes is not suitable for their 2D derivatives. In this paper, we develop an efficient machine-learned potential (MLP) based on a newly developed MLP model called neuroevolution potential to build a specific potential for 2D PbTe monolayers. Then, by using this potential, we report the thermal conductivity of 2D PbTe monolayers at different temperatures and under different biaxial strains. Surprisingly, we find an abnormal increase of thermal conductivity with the increase of the biaxial strain due to the enhancement of low-frequency phonons. We hope these results can play a guiding role in their practical use once upon experimental validation.