Artificial intelligence paves the way to discovering new rare-earth compounds

Scientists explore materials with artificial intelligence. Researchers from Ames Laboratory and Texas A&M University trained a machine-learning model to assess the stability of rare-earth compounds. The work was supported by theLDRD program. Current state-of-the-art methods for experimenting with compounds and understanding chemical instabilities are built into the framework they developed.

Since the middle of the 20th century, Ames Lab has been a leader in rare-earths research. Clean energy technologies, energy storage, and permanent magnets are some of the uses of rare earth elements. Scientists are trying to expand access to rare-earth materials.

Machine learning is a form of artificial intelligence that improves through data usage and experience. The foundation for their model was built using the upgraded Ames Laboratory Rare Earth database and high-throughput density-functional theory.

A researcher can test hundreds of models quickly with high-throughput screening. DFT is a method used to investigate the properties of body systems. The developed model uses regression learning to assess the stability of compounds.

Tyler Del Rose, a graduate student at Iowa State University, conducted a lot of the research needed for the database by using the internet to find information. He helped to improve the models by making sure they are representative of reality.

Machine learning is important here because when we are talking about new compositions, ordered materials are very well known to everyone in the rare earth community. You can't investigate all the possible combinations using theory or experiments when the number of compositions increases.

Singh explained that the material analysis is based on a feedback loop in which the model is updated using real-time structural and phase information obtained from our experiments. The process ensures that information is carried from one step to the next and reduces the chance of making mistakes.

Yaroslav Mudryk said that the framework was designed to explore rare earth compounds because of their technological importance, but it is not limited to rare-earths research. The same approach can be used to train a model to predict magnetic properties of compounds.

Mudryk said it was not meant to discover a particular compound. That is what we did.

Mudryk said that this work is just the beginning. There will be a wide range of applications for the framework in the future, as the team is exploring the full potential of this method.

This research is further discussed in the paper, which was written by P. Singh, T. Del Rose, G. Vazquez, R. Arroyave, and Y. Mudryk.

More information: Prashant Singh et al, Machine-learning enabled thermodynamic model for the design of new rare-earth compounds, Acta Materialia (2022). DOI: 10.1016/j.actamat.2022.117759 Citation: Artificial intelligence paves the way to discovering new rare-earth compounds (2022, March 18) retrieved 18 March 2022 from https://phys.org/news/2022-03-artificial-intelligence-paves-rare-earth-compounds.html This document is subject to copyright. Apart from any fair dealing for the purpose of private study or research, no part may be reproduced without the written permission. The content is provided for information purposes only.