A strong rare-earth free magnet “developed” and refined by a machine learning algorithm | Search

US researchers have discovered a rare-earth ferromagnetic material that has properties similar to the rare-earth magnets found in everything from wind turbines to computer hard drives using a machine learning-guided approach. The material requires further development, but the demonstration is an important step on the road to creating strong magnets that do not rely on rare earth elements.

Permanent magnets are critical to generating electricity in hydroelectric power, wind power, and many other green energy technologies, as well as information technologies. These devices need strong magnets with high coercive force – a well-constrained magnetic field. Making them requires a magnetic material with a high magnetic anisotropy — a measure of the dependence of magnetic moment on lattice angle. “Until now, high-contrast magnets have contained rare earths,” he says. Tsai Chuang Wang From the Ames Laboratory of the US Department of Energy at Iowa State University. “Why is a very basic question that is not yet fully understood.” Regardless of the mechanism, demand for permanent magnets is set to grow as society moves to cut emissions by electrifying transportation and industry. Thus, there will be a huge demand for magnets made of cheap elements such as iron.

A material can only show good magnetic anisotropy if it has an anisotropic lattice structure, which rare earth compounds often do. However, iron-cobalt alloys tend to be more stable in cubic structures. The researchers attempted to break this symmetry by adding a third element, such as nitrogen, to occupy the interstitial sites in the cubic lattice. However, they often found that the structures were not stable enough and decomposed at high temperatures.

Wang and colleagues in the Ames lab and elsewhere studied 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 that, they calculated, would have negative energy to form. They then trained a DFT algorithm using data from previous experiments with ternary compounds of iron and cobalt to predict the maximum magnetization and magnetic anisotropy of the different structures. Finally, they used the adaptive genetic algorithm to generate new structures from the most interesting candidates. “The simplest method is to take two structures and put them together like the parents,” Cai-Zhuang explains.

After each stage, a machine learning algorithm found the active ground states for their new structures by DFT and calculated the magnetic properties of these ground states, before using this data to refine its subsequent predictions—selecting the most promising candidates and then combining, optimizing, and computing the properties of the new structures. “It’s a simulation of the evolutionary process,” Wang explains.

So the researchers quickly came up with the most promising compounds without analyzing each combination of the three. The researchers pooled the most promising candidate, and found a good fit with their predictions. “I think this is the first demonstration of a rare-earth-free magnet that has such high anisotropy,” Wang says, “but a real magnet would be much more complex than a single crystal, so this only opens the door and there is a lot of work to be done.”

Ziwan Rao of the Max Planck Institute for Iron Research in Düsseldorf is of interest. “Many small countries in Europe, for example, do not have their own supply of rare earth elements,” he says, “so this topic is very important, but also very difficult, because rare earth metals can have a great deal of coercivity and also Very high magnetization. I think it’s an important paper.

Leave a Comment