The data of the article “Strain and Structural Modulation of Thermal Conductivity in AlN/GaN Superlattices via Machine Learning Force Fields”

In this study, we trained corresponding machine learning potentials for several different structures, and verified the accuracy of the potentials through phonon spectra, etc.; subsequently, we calculated the changes in thermal conductivity under ±5% stress.

The thermal conductivity decreased with tensile strain and increased with compressive strain, and the trend of group velocity was consistent with that of thermal conductivity; in addition, we calculated the thermal conductivity of different structures, and the thermal conductivity increased with the shortening of the quantum well (GaN layer). This dataset mainly consists of 5 files. File 1 is different crystal structures ( "data of crystal structure,.cif", VESTA), which is the figure 1 of this paper; File 2 is phonon spectrum data of different structures ("data of phonon spectra with different structures,.dat", Origin), which is the figure 3 of this paper; File 3 is thermal conductivity under different strains ("data of thermal conductivities under different strains,.dat", Origin), which is the figure 5 of this paper; File 4 is the change in group velocity under different strains ("data of group velocities under different strains,.dat", Origin), which is the figure 6 of this paper; File 5 is thermal conductivity of different superlattice structures ("data of thermal conductivities under different superlattice structure,.dat", Origin), which is the figure 9 of this paper.