Development and application of simulation techniques to accurately and efficiently predict melting properties from ab initio — Melt-initio

• Applicant: Li-Fang Zhu, Ph.D.
• Subject Area: Computer-Aided Design of Materials and Simulation of Materials Behaviour from Atomic to Microscopic Scale ; Thermodynamics and Kinetics as well as Properties of Phases and Microstructure of Materials
• Term: since 2022
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 493417040

Melting properties of materials, e.g., the melting temperature, the entropy and enthalpy of fusion, the volume change at the melting point, have great scientific and technological importance. They are a key component for multi-component phase diagram calculations. Experimental measurements of these properties are however difficult, expensive and time consuming. Some of them, e.g., melting properties of metastable or dynamically unstable phases, are experimentally even inaccessible. But they are nevertheless crucial for predicting phase diagrams of multi-component systems. To address these challenges, the applicant has previously developed a state-of-the-art methodology to calculate the melting properties using parameter-free ab initio simulation techniques. This methodology effectively combines ab initio accuracy with computational efficiency and has been successfully applied to stable unary phases. However, there are still crucial challenges remaining in the current form of the methodology. The proposed project aims to solve these challenges and improve the methodology to a more general level. First, despite the significant improvement of the computational efficiency, the ab initio computing cost of the method is still evident. A new approach is proposed to further improve the computational speed by utilizing recently developed machine-learning potentials. Second, the method is so far applicable to stable and metastable phases (though not yet tested), but not to dynamically unstable phases. The free energy calculation of unstable phases has been a long-standing challenge in materials science. Therefore, further development is necessary in order to efficiently access the melting properties of unstable phases. A novel approach is proposed to calculate the free energy of metastable and unstable phases at finite temperatures. Third, extending the methodology from unary to binary systems is necessary. This development can be a breakthrough towards predicting melting properties of multi-component alloys. Fourth, the methodology combines empirical potential calculations and ab initio molecular dynamic simulations. The practical application of the methodology requires sophisticated user knowledge and suffers from many technicalities. This has become a serious bottleneck for general applications. To overcome this, the project aims to optimize the methodology into an easy-to-use simulation tool, which will be of high value for the materials science field.

• DFG Programme: Research Grants