Mechanical and Thermal Properties of Entropy-Stabilized Transition Metal Oxides

Entropy-stabilized oxides have been materials of active research interest due to a high degree of lattice distortion and tunability. Lattice distortion plays a crucial role in understanding elastic constants and lattice thermal conductivity. In this work, a neuroevolution machine learning potential (NEP) is developed for the MgCoNiCuZnO5 compound, and its accuracy has been compared to density functional theory (DFT) calculations. Employing this NEP potential, lattice distortion, elastic constants and thermal conductivity have been quantified for this compound. In agreement with experimental findings, we have shown that the average lattice distortion of oxygen atoms is higher than that of all transition metals. The observed distortion saturation arises from the competing effects of minimum site distortion, which increases with increasing temperature due to enhanced thermal vibrations, and maximum site distortion, which decreases with increasing temperature. Furthermore, a series of molecular dynamics simulations up to 900 K were performed to study the stress-strain behavior. Elastic constants, bulk modulus, and ultimate tensile strength obtained from these simulations indicate a linear decrease in these properties with temperature. Finally, to gain some insight into thermal transport in these materials, we used non-equilibrium molecular dynamics simulations to compute the heat current autocorrelation and the spectral thermal conductivity. It is found that while the thermal conductivity $\kappa$ of MgNiO2 decreases from 4.25 W/m.K at room temperature to 3.5 W/m.K at 900 K, it almost saturates (with maybe a small eventual decline) for the 5-component MgCoNiCuZnO5. This suppression for the binary compound is attributed to the stronger scattering of low-frequency modes at higher temperatures.