Molecular dynamics insights on thermal conductivities of cubic diamond, lonsdaleite and nanotwinned diamond via the machine learned potential

Abstract Diamond is a wide-bandgap semiconductor with a variety of crystal configurations, which has the potential for applications in the field of high-frequency, radiation-hardened, and high-power devices. There are several important polytypes of diamond, such as cubic diamond, lonsdaleite and nanotwinned diamond (NTD). The thermal conductivity of semiconductor should be calculated at different temperatures in high-power devices. However, there has been no potential reported for cubic diamond and its polytypes combining efficiency and accuracy based on molecular dynamics (MD) to predict thermal conductivity. Here, an interatomic potential utilizing neural networks could offer distinct advantages, such as reduced computational time compared to density functional theory (DFT), while maintaining high accuracy in predicting the thermal conductivity of the mentioned three diamond polytypes. Based on the neuroevolution potential (NEP), the thermal conductivities of cubic diamond, lonsdaleite, and NTD at 300 K are respectively 2507.3, 1557.2, and 985.6 W· m -1 · K -1 , which are higher than the calculation results based on Tersoff-1989 potential (1508, 1178, and 794 W· m -1 · K -1 , respectively). The thermal conductivities of cubic diamond and lonsdaleite using the NEP are closer to the experimental data or DFT data compared with Tersoff-potential. The molecular dynamics simulations are performed using NEP to calculate the phonon dispersions to illustrate the possible reasons for discrepancies among the cubic diamond lonsdaleite and NTD. In this work, we have proposed a scheme to predict the thermal conductivity of cubic diamond, lonsdaleite and NTD precisely and efficiently, and explain the differences in thermal conductivity among cubic diamond, lonsdaleite and NTD.