Technologies, such as electric motors and batteries for electric networks that need rare-earth metals, are required to switch to the electrified world. These metals of the road are both financial and environmentally friendly. However, the British company recently announced the development of a magnet that does not contain rare-earth metals using artificial intelligence (AI). The process took only three months, which is about 200 times faster than usual.
AI is already actively used to search for materials in other key areas of transition to green energy, demonstrating its power in the fight against climate change. The world transition from fossil fuel to environmentally friendly energy sources is inevitable. However, the production of electric motors and batteries requires rare earth metals, the extraction of which harms the environment and society. In this regard, automakers, such as Tesla, are looking for alternative materials for their electric vehicles.
Material scholars are faced with the problem of lack of time on the study of alternative materials. In this case, artificial intelligence comes to the rescue, working much faster.
Materials Nexus has developed a magnet called magnex, completely free from rare-earth metals, using its AI platform. Previously, such “clean” magnets have been created for decades by trial and error. In the case of Magnex, the process took only three months. The magnet can be produced 20% cheaper and with 70% less carbon emissions compared to similar products based on rare-earth metals.
Materials Nexus collaborated with Institute of Henry Royce and the University of Sheffield for the synthesis and testing of a new magnet. According to the company, a similar approach using AI can revolutionize other aspects of the green transition, including the development of semiconductors and superconductors. This is confirmed by the recent successes of scientists from the UK and Japan, who, using AI, created a superconducting magnet based on iron.
AI is especially suitable for revolutionary discoveries in the field of materials science. According to the Materials Project project, an open source to support the research of new materials, people discovered 20 thousand materials through experiments, while thanks to computing technologies, this number increased to 48 thousand.
At the end of 2023, researchers from DeepMind, owned by Google, said that their network Graph Networks for Materials Exploration (GNOME) used these 48 thousand materials to create an additional 2.