Composition dependence of diffusion and thermotransport in Ni-Al melts: A step towards molecular dynamics assisted databases

We present an extensive and self-consistent database of diffusion and thermotransport properties of molecular-dynamics models of Ni-Al melts with an embedded-atom method potential. The database is generated over wide temperature and composition ranges. A careful comparison with available fragmentary experimental data is made to ensure reliability of the employed model description of Ni-Al melts. A comprehensive analysis of the presented data is carried out and the emerging trends are discussed in details. In particular, a phenomenological model for the composition dependence of the enthalpy of mixing is introduced. This model accurately describes our simulation data for Ni-Al melts as well as enables an ease access to the partial enthalpies of species essential for precise characterization of thermotransport. Furthermore, an intriguing similarity between the shapes of the composition dependence of the Manning factor (which characterizes the interrelation between the collective and single-particle diffusion) and the enthalpy of mixing in Ni-Al melts is brought to attention. Meanwhile, our results for the reduced heat of transport (which characterizes the pure heat conduction initiated by collective diffusion in isothermal conditions) allow us to predict that, in Ni-Al melts, Ni and Al tend to migrate, respectively, to the cold and hot ends in the presence of a temperature gradient. Overall, we highlight a great importance of the presented study for developing and testing of theoretical frameworks to enable sophisticated control of the heat and mass transport in binary melts.