Effect of stoichiometry on thermodynamic and thermal transport properties of entropy-stabilized oxide MgCoNiCuZnO5

Lattice distortion in a crystal can have a strong effect on its thermal transport properties. Herein, using a neuroevolution machine learning potential and molecular dynamics simulations, we demonstrate that the lattice thermal conductivity of MgCoNiCuZnO5 (J14) decreases from κ≃2.60±0.20 W m−1 K−1 at 300 K to κ≃2.07±0.05 W m−1 K−1 at 900 K, consistent with experimental findings. Our simulations indicate that the reduction in lattice thermal conductivity as a function of temperature in J14 is due to a decreased mean free path of propagons. Next, we investigated the bounds on κ when reducing or increasing distortions. This is done by varying stoichiometry to make atomic size distribution of transition metal elements more or less uniform: the composition of zinc and nickel in J14 was varied by 30% to make atomic sizes more and less uniform, respectively. While adding more Zn increases distortions, adding Ni reduces it. As a result, we find that the related changes in the thermal conductivity with 30% more zinc (Mg0.2Co0.2Ni0.14Cu0.2Zn0.26O) and 30% more nickel (Mg0.2Co0.2Ni0.26Cu0.2Zn0.14O) are, respectively, 2.54±0.20 and 2.77±0.20 W m−1 K−1 at room temperature. This shows that there is indeed a correlation between the amount of distortions and the thermal conductivity in these compounds. We see larger mean free paths (MFPs) in the less distorted (30% Ni) compound, while MFPs in the more distorted compound are shorter. The MFPs are associated with phonon-like vibrational modes (also called propagons). These results imply that the propagons are mostly responsible for the changes in thermal conductivity. This is supported by the slight decrease in the thermal conductivity with temperature, implying the important contribution of propagons in thermal transport in these distorted high-entropy materials. The neuroevolution machine learning interatomic potential implemented in Graphics Processing Unit Molecular Dynamics is a reliable tool to investigate thermal transport in disordered systems beyond the perturbative approach.