CALPHAD: From Building Block of Alloy Design to Physics Guided Models

The CALPHAD approach is suitable for predicting microstructure and material properties at a given combination of alloy chemistry and processing condition, especially for the multicomponent alloys. Through the optimization of thermodynamic parameters for constituent binary and ternary systems, phase stability and composition can be successfully predicted with high computational efficiency. In this talk, the capability and power of the CALPHAD method will be demonstrated by a new database, i.e., TCS Molybdenum Database (TCMO). TCMO is a thermodynamic database developed by coupling CALPHAD, DFT, and experiments with the purpose of contributing to the development of refractory alloys, specifically Mo-Si-B based alloys, which are promising candidates as high-temperature structural materials. The methodologies applied during its development, several validation examples, and a variety of applications will also be presented.

Although the CALPHAD approach has showcased its power in the practical applications, it is sometimes questioned due to its semi-empiricism. To overcome the unphysical artefacts inherent in the current widely used models, the third generation CALPHAD is being developed. In this talk, the state-of-the-art improvements in CALPHAD models, thanks to the work of Dr. Bo Sundman, will be reviewed in terms of the low-temperature thermodynamics, extrapolation method, and magnetic properties. The applicability of these physics-guided models in three binary systems and the associated unary systems will be presented by comparing the calculated phase diagrams and thermodynamic properties using the third generation unary descriptions and models with experimental and DFT data.