Trevor Rhone uses AI to identify van der Waals magnets in 2D
A research team led by Trevor David Rhône, assistant professor in the Department of Physics, Applied Physics and Astronomy at Rensselaer Polytechnic Institute, used cutting-edge tools in artificial intelligence (AI) to develop new van der Waals (vdW) identified the magnet. In particular, the research team used semi-supervised learning to identify transition metal halide vdW materials with large magnetic moments that are predicted to be chemically stable. These two-dimensional (2D) vdW magnets have potential applications in data storage, spintronics, and even quantum computing.
Rhone specializes in using materials informatics to discover new materials with unexpected properties that advance science and technology. Materials informatics is an emerging research field at the intersection of AI and materials science. His team’s latest work was recently featured on the cover of Advanced Theory and Simulations.
2D materials as thin as a single atom were only discovered in 2004 and have been of great scientific curiosity due to their unexpected properties. 2D magnets are important because they maintain long-range magnetic order even when thinned to one or a few layers. This is due to magnetic anisotropy. This interaction of magnetic anisotropy with low dimensionality can give rise to exotic spin degrees of freedom such as spin textures that can be used to develop his architectures for quantum computing. 2D magnets are compatible with all electronic properties and can be used in high performance and energy efficient devices.
Rhone and team combined high-throughput density functional theory (DFT) computations to determine the properties of the vdW material, combined with AI to implement a form of machine learning called semi-supervised learning. Semi-supervised learning uses a combination of labeled and unlabeled data to identify patterns in the data and make predictions. Semi-supervised learning alleviates the lack of labeled data, which is a major challenge in machine learning.
“Using AI saves time and money,” says Rhone. “Typical material discovery processes require expensive simulations on supercomputers, which can take months. Experiments in the lab can be even longer and more expensive. has the potential to speed up the material discovery process.”
Using an initial subset of 700 DFT calculations on a supercomputer, an AI model was trained that can predict the properties of thousands of candidate materials in milliseconds on a laptop. The team then identified promising candidate vdW materials with large magnetic moments and low formation energies. A low energy of formation is an indicator of chemical stability and is a key requirement in the synthesis of materials in the laboratory and subsequent industrial applications.
“Our framework can be easily applied to explore materials with different crystal structures,” said Rhone. “Prototypes of mixed crystal structures, such as datasets for both transition metal halides and transition metal trichalcogenides, can also be investigated in this framework.”
“Dr. The application of AI to the field of materials science at the University of the Rhône continues to produce exciting results,” said Kurt Brenneman, Dean of the School of Science at Rensselaer University. “Not only has he accelerated his understanding of 2D materials with new properties, but his discoveries and methods may contribute to new quantum computing techniques.”
