Researchers from Linkoping University in Sweden have used artificial intelligence to search scientific databases for materials that were created earlier but went unnoticed. The subject is the family of two-dimensional materials known as MXenes, discovered only in 2011. Two-dimensional materials are substances a few atomic layers thick. The results were published in the journal Science.
MXenes consist of a metal and carbon or nitrogen atoms. Their main feature is a layered structure: a material only a few millimeters thick can contain millions of layers. Thanks to this, MXenes have a huge surface area relative to their mass, and various chemical reactions can take place between the layers. This makes them promising for water purification, energy storage, telecommunications and biomedicine.
Using AI, the researchers analyzed old databases and identified 33 MXenes that had already been produced but had not attracted the attention of specialists. Inspired by the find, the team created 11 more new materials. The project was jokingly called the "Treasure Chest": according to materials science professor Johanna Rosen, the literature and databases hold a mass of information that scientists sometimes simply do not notice and end up "reinventing the wheel."
Besides the new compositions, an important result was the synthesis method. The researchers learned to produce MXenes in a few seconds without external heating, using significantly fewer hazardous chemicals and less water than in previous methods. In addition, they showed that rare-earth metals can replace other metals in the composition of MXenes, giving the materials magnetic properties. This expands possible applications, for example in energy and data storage systems.
The authors note that to accelerate the development of materials science, one needs to look both backward and forward. In their view, it is scaling up production that will help MXenes enter real technologies more widely. At the same time, the study does not claim that the materials are already ready for mass adoption - it concerns laboratory results and new opportunities for further development.