Lattice thermal conductivity of Ti3C2 MXene via deep-learning potential

Ti 3 C 2 , a representative two-dimensional MXene with promising applications in nanoelectronics, energy storage, and thermal management, is investigated with particular emphasis on its lattice thermal conductivity and structural stability. A accurate deep-learning interatomic potential (DLP) is first developed using a comprehensive dataset generated from ab initio molecular dynamics (AIMD) simulations and is subsequently employed in large-scale classical molecular dynamics simulations. The reliability of the constructed DLP are carefully validated through its strong agreement with AIMD results in terms of energies, atomic forces, and stress. Furthermore, the dependence of lattice thermal conductivity on system size and temperature is examined in detail, providing important insights into phonon transport and thermal behavior in two-dimensional Ti 3 C 2 . Overall, this investigation demonstrates the efficiency and accuracy of deep-learning-based interatomic potentials for studying the thermal and structural properties of Ti 3 C 2 over extended spatial and temporal scales.