Researchers from the University of Illinois Urbana-Champaign have developed a new method for calculating diffusion in multicomponent alloys, which are metals composed of five different elements. By breaking down diffusion into individual atom jumps, called “kinosons,” and using machine learning to compute the statistical distribution of these jumps, they were able to model diffusion in the alloy and calculate its diffusivity more efficiently than traditional methods. This work, led by materials science & engineering professor Dallas Trinkle and graduate student Soham Chattopadhyay, was recently published in Physical Review Letters.
Diffusion in solids is a fundamental process in materials science, influencing phenomena such as the production of steel, battery performance, and semiconductor device doping. Multicomponent alloys, which have unique properties like good mechanical behavior and stability at high temperatures, are especially important to study as they are commonly used in industry to create strong materials. Understanding how atoms move within these alloys is crucial for optimizing their properties and performance.
Traditional methods for simulating diffusion involve running simulations for long periods of time to capture the random movement of atoms. However, the new kinoson-based approach makes it possible to calculate diffusivity more efficiently by considering each individual atom jump as a contribution to diffusion. By summing up these kinosons, researchers can accurately model diffusion in multicomponent alloys and gain insights into how different elements move within the material.
The use of machine learning in this new method eliminates the need to account for correlated jumps, making the diffusion problem much simpler to solve. By focusing on the individual contributions of each atom jump, researchers are able to extract the distribution of kinosons and calculate diffusivity orders of magnitude faster than with traditional methods. This advancement not only speeds up simulations but also allows for a more detailed understanding of diffusion processes in multicomponent alloys.
Trinkle believes that this new approach to modeling diffusion has the potential to revolutionize the field of materials science in the coming years. By adopting the kinoson-based method as the standard way of looking at diffusion, researchers can gain deeper insights into the behavior of atoms in solid materials. The combination of machine learning and kinosons offers a more efficient and accurate way to study diffusion, paving the way for advancements in materials design and optimization.
Overall, the development of this new method represents a significant advancement in our understanding of diffusion processes in multicomponent alloys. By reimagining diffusion as the sum of individual atom jumps and leveraging machine learning techniques, researchers have made it possible to calculate diffusivity in these complex materials more efficiently and accurately than ever before. This innovative approach has the potential to reshape the way diffusion is studied and applied in materials science, leading to new insights and advancements in the field.