Machine learning guided high-throughput search of non-oxide garnets

Abstract

Garnets have found important applications in modern technologies including magnetorestriction, spintronics, lithium batteries, etc. The overwhelming majority of experimentally known garnets are oxides, while explorations (experimental or theoretical) for the rest of the chemical space have been limited in scope. A key issue is that the garnet structure has a large primitive unit cell, requiring a substantial amount of computational resources. To perform a comprehensive search of the complete chemical space for new garnets, we combine recent progress in graph neural networks with high-throughput calculations. We apply the machine learning model to identify the potentially (meta-)stable garnet systems before performing systematic density-functional calculations to validate the predictions. We discover more than 600 ternary garnets with distances to the convex hull below 100 meV ⋅ atom−1. This includes sulfide, nitride, and halide garnets. We analyze their electronic structure and discuss the connection between the value of the electronic band gap and charge balance.

Transfer learning on large datasets for the accurate prediction of material properties

“…The DCGAT dataset combines compatible data from AFLOW, 14 the materials project 17 and ref. 18, 35 and 52–54. We generally used a train/val/test split of 80/10/10%.…”

Influence of Dy3+ environment on magnetic anisotropy and magnetocaloric effect in Dy3B2C3Damay, Petit, Sheptyakov et al. 2024Phys. Rev. B0000View full textAdd to dashboardCite

“…A possibility to explore those areas would be to find other, lesser packed, rare-earth systems with a D 2 symmetry site for the rare earth; neutron diffraction experiments under pressure could be an interesting way as well to study in situ magnetic anisotropy changes as the cage distorts. One other possible route is the synthesis of nonoxide garnets [55], like oxisulfides, which, although very promising, are outside the scope of the present paper.…”

optimade-python-tools: a Python library for serving and consuming materials data via OPTIMADE APIs

“…The database was primarily generated by scanning binary, ternary, and quaternary prototypes to identify stable compounds. This process employed crystal graph attention networks 114,116,117 to predict the stability of all potential compositions for each prototype. Compounds that were found to be close to stability were subsequently confirmed using DFT.…”