A team of US researchers from Iowa State University, The University of Texas at Austin, and the University of Nebraska have found a free rare-earth magnetic material with a machine learning-guided approach.
The new chemical combination (Fe3CoB2) can generate a strong magnetic field without the use of costly rare earth elements. The research paper was published in the Proceedings of the National Academy of Sciences (PNAS) journal.
Rare earth magnets are a type of permanent magnets made from alloys of rare earth elements. They are the strongest type of permanent magnet currently available and are used in a wide range of applications, including motors, generators, sensors, and speakers. Common rare earth magnets include neodymium, samarium-cobalt, and dysprosium magnets.
With the aim of reducing emissions and adopting eco-friendly energy, there is an unprecedented demand for such powerful magnets. Nevertheless, these magnets are relatively more costly than other types and are limited to only a few countries.
Improved substitute for rare earth magnets are the so-called “rare-earth-free” magnets, which don’t contain any rare earth elements in their composition. Instead, they employ alternative materials that exhibit magnetic properties.
The research team discovered a new rare-earth free magnetic material using a combination of the three elements: Fe, Co, and B. To create the new structures they used two machine learning algorithms: the Density Functional Theory (DFT) and the Adaptive Genetic Algorithm (AGA).
In the first stage, the DFT algorithm was used to forecast the highest magnetizations and magnetic anisotropies of various structures and the best ones were selected.
In the second stage, these candidates were used by the AGA algorithm to generate new structures.
Conclusion
Despite requiring further refinement, the recently discovered material by the research team constitutes a significant advancement in the quest for potent magnets that are not reliant on rare earth elements.
Further research is needed to improve the magnetic properties of these magnets and make them more competitive with rare earth magnets, in terms of performance and cost.
Learn more:
- Research paper: Accelerating the discovery of novel magnetic materials using machine learning–guided adaptive feedback (on PNAS)
- Improved, heavy rare earth-free, low-cost magnets for EV vehicles could reduce mobility costs (on DST)
- Enhancing the coercivity of Nd-Cu-diffused Nd-Fe-B permanent magnets by Nb-assisted grain boundary pinning (on Taylor & Francis)
- Machine learning guided discovery of ternary compounds involving La and immiscible Co and Pb elements (on Nature, npj Computational Materials)
