Modeling environment-dependent atomic-level properties in complex-concentrated alloys

Area of Science:

• Materials Science
• Computational Materials Science
• Alloy Design

Background:

• Complex concentrated alloys (CCAs) exhibit desirable properties like high-temperature strength and radiation tolerance.
• Their multi-principal component nature leads to complex atomic structures and properties, posing modeling challenges.

Purpose of the Study:

• To develop accurate and computationally inexpensive predictive models for atomic properties of CrFeCoNiCu-based CCAs.
• To enable the integration of these models into macroscopic simulations for broader material design applications.

Main Methods:

• Combined atomistic simulations with many-body potentials and machine learning.
• Developed atomic environment fingerprints using local geometry invariants and elemental information.
• Utilized unrelaxed atomic structures for computationally efficient descriptors.

Main Results:

• Achieved accurate predictive models for vacancy formation energy, cohesive energy, pressure, and volume in CCAs.
• Demonstrated the models' ability to extrapolate to unseen compositions and elements.
• Validated the computational efficiency of the descriptors for integration into larger-scale simulations.

Conclusions:

• Machine learning combined with atomistic simulations provides a powerful approach for modeling complex concentrated alloys.
• The developed models offer a computationally efficient pathway for predicting CCA properties and accelerating materials discovery.
• The findings facilitate the design of novel CCAs with tailored properties for demanding applications.