Speaker
Description
Most metallic alloys of technological relevance, such as steels and superalloys, are inherently complex due to the coexistence of chemical disorder and, in many cases, magnetic disorder. This is particularly important in Fe-based systems, where the magnetic state can vary with temperature and composition. Capturing these effects within first-principles approaches remains challenging, especially when extended defects such as dislocations, stacking faults, and interfaces are involved.
In this work, we focus on impurity–defect interactions in Fe-based systems across different crystal structures, magnetic states, and defect types. We employ a combination of advanced computational methods, including Green’s-function-based density functional theory and machine-learning(ML)-assisted approaches, such as actively learned interatomic potentials, to study how alloying and magnetic disorder influence defect properties. The former offers a particular advantage in that chemical and magnetic disorder can be treated consistently and efficiently within a single unified framework, namely the coherent potential approximation together with the disordered local moment model, making the approach efficient. The ML methods are applied in specific regimes where their scalability provides clear benefits.
The results indicate that magnetic effects play a central role in determining segregation tendencies and defect-solute interaction energies, often exceeding contributions from purely elastic size mismatch. In addition, these interactions are found to depend sensitively on the magnetic state, temperature, and thermal vibrations.
Overall, the study highlights the importance of consistently accounting for the correct magnetic state and disorder when describing defects in alloys.