Powerful rare-earth-free magnet "evolved" and refined by a machine learning algorithm

Powerful rare-earth-free magnet “evolved” and refined by a machine learning algorithm

A rare-earth-free magnetic material with properties similar to rare-earth magnets found in everything from wind turbines to computer hard drives has been discovered by US researchers using a machine learning-guided approach. The material requires further development, but the demonstration is an important step on the way to creating powerful magnets that do not rely on rare earth elements.

Permanent magnets are crucial for power generation in hydropower, wind power and many other green energy technologies, as well as for information technology. These devices need strong magnets with high coercivity – a well constrained magnetic field. Their manufacture requires a magnetic material with high magnetic anisotropy – a measure of the dependence of the magnetic moment on the angle of the lattice. “Until now, magnets with high anisotropy have contained rare earths,” says Cai-Zhuang Wang of the US Department of Energy’s Ames Laboratory at Iowa State University. “Why is a very fundamental question that is not yet fully understood.” Whatever the mechanism, demand for permanent magnets is expected to increase as society strives to reduce emissions by electrifying transportation and industry. Magnets made from cheap elements such as iron will therefore be in high demand.

A material can only exhibit good magnetic anisotropy if it has an anisotropic lattice structure, which rare earth compounds often do. Iron-cobalt alloys, however, tend to be the most stable in cubic structures. The researchers attempted to break this symmetry by adding a third element such as nitrogen to occupy the interstitial positions in the cubic lattice. However, they often found that the structures are insufficiently stable and break down at high temperatures.

Wang and his colleagues at the Ames lab and elsewhere examined compounds containing iron, cobalt, and boron using a combination of machine learning, density functional theory (DFT), and an “adaptive genetic algorithm.” “. They started with about 400 structures which they calculated would have negative formation energy. They then trained a DFT algorithm using data from previous experiments with iron-cobalt ternary compounds to predict maximum magnetizations and magnetic anisotropies of various structures. Finally, they used their adaptive genetic algorithm to generate new structures from the most interesting candidates. “The easiest way is to take two structures and put them together like two parents,” says Cai-Zhuang.

After each step, the machine learning algorithm found the energetic ground states of their new structures by DFT and calculated the magnetic properties of these ground states, before using this data to improve its subsequent predictions – selecting candidates the most promising then by combining, optimizing and calculating the properties of the new structures. “It’s an imitation of the evolutionary process,” says Wang.

The researchers thus quickly arrived at the most promising compounds without analyzing each combination of the three elements. The researchers synthesized the most promising candidate and found good agreement with their predictions. “I think this is the first demonstration of a magnet without a rare earth that has high anisotropy,” says Wang, “but the real magnet will be much more complicated than a single crystal, so it just opens the door and there has a lot of work to do.

Ziyuan Rao from the Max Planck Institute for Iron Research in Düsseldorf is intrigued. “Many small countries in Europe, for example, do not have their own supply of rare earth elements, so this topic is very important,” he says, “but it is also very difficult, because the metals of rare earths can have very high coercivity and also very strong magnetization. I think this is an important article.

#Powerful #rareearthfree #magnet #evolved #refined #machine #learning #algorithm

Leave a Comment

Your email address will not be published. Required fields are marked *