This study combines machine learning-based crystal structure prediction, VASP density functional theory calculations, and high-throughput screening to identify novel rare-earth-free permanent magnets. Magnetic properties are evaluated through Monte Carlo and atomistic spin dynamics simulations, and the stability of the discovered materials is confirmed by phonon calculations.
Recent advances in crystal structure prediction, particularly through machine learning, have opened up possibilities for discovering new materials. One promising approach is the use of crystal graph attention networks, which can predict the thermodynamic stability of hypothetical compounds. This allows researchers to identify stable and metastable materials that could potentially be synthesized in a laboratory setting.
In this study, a high-throughput computational approach is applied to a database containing over one million predicted crystal structures. The search focuses on identifying rare-earth-free permanent magnets with high magnetization, high magnetocrystalline anisotropy, and high Curie temperature. Four promising candidate materials are discovered.
Novel, stable, and rare-earth-free candidate permanent magnet materials have been identified by utilizing a database of predicted crystal structures. The identified compounds have not been previously reported, offering the potential for significant advancements in permanent magnet technology.
A high-throughput computational screening approach was employed to accelerate the discovery process.
Four candidate materials (Ta3ZnFe8, AlFe2, Co3Ni2 and Fe3Ge) with desirable magnetic properties were identified, and their dynamical stability was confirmed. These materials exhibit properties comparable to or exceeding those of existing rare-earth-free permanent magnets.
Fig. 2. The spin-polarized density of states (DOS) for four materials: Ta3ZnFe8, Ga2Fe6B, AlFe2, and Co3Ni2.
Research Significance
This study demonstrates the power of combining crystal structure prediction with high-throughput screening to accelerate the discovery of new functional materials. The identified rare-earth-free permanent magnets hold promise for applications in sustainable technologies, reducing our reliance on critical raw materials while meeting the growing demand for high-performance magnets.
Table 1. Promising stable and metastable materials and their key properties for permanent magnet applications: point symmetry group, saturation magnetization (Ms), magnetocrystalline anisotropy energy (MAE), Curie temperature (Tc), distance to the convex hull, formation energy, and magnetic hardness coefficient (κ).
Authors
The authors of this work are Alena Vishina, Olle Eriksson, Heike C. Herper from Uppsala University, Sweden, and Alena Vishina is the corresponding author of this work.
Acknowledgement
The authors acknowledged the support of the Swedish Foundation for Strategic Research, the Swedish Research Council (VR), the Swedish Energy Agency, the Knut and Alice Wallenberg Foundation (KAW), STandUPP, eSSENCE, and the ERC.
A. Vishina, O. Eriksson, H.C. Herper, Stable and metastable rare-earth-free permanent magnets from a database of predicted crystal structures, Acta Materialia 261 (2023) 119348. DOI: https://doi.org/10.1016/j.actamat.2023.119348
Related Recommendations
Four Open Source Materials Gene Databases
