This paper presents results of atomistic simulations under high pressures and temperatures for platinum. To obtain precise models of interatomic interactions, machine learning potentials are fitted for a set of isotherms in the temperature range of 2–70 kK. Training sets are prepared from snapshots extracted from ab initio molecular dynamics trajectories calculated in the framework of density functional theory taking into account not only the vibrational degrees of freedom of ions but also the excitation of the electronic subsystem. The latter means that it is necessary to derive the potential for each particular temperature. Trained interatomic potentials give an accuracy better than 5 meV/atom in comparison with the results of ab initio calculations. Calculations of structural and transport properties are performed in the temperature range of 2–70 kK. Shear viscosity, self-diffusion coefficients, the Stokes-Einstein relation, and the pair correlation function are obtained. Calculations show the significance of many-body interactions in the platinum liquid phase. The high-pressure and -temperature phase diagram of platinum is determined with the thermodynamic integration technique and compared with existing experimental data and other theoretical predictions. The current evaluation of the phase diagram excludes any crystal structures in the considered pressure-temperature range other than the fcc.

Close

PrevNext

View figure in article

• High-pressure studies
• Interatomic & molecular potentials
• Phase diagrams
• Self-diffusion
• Transport phenomena
• Liquids
• Transition metals
• Density functional theory
• Molecular dynamics