Accelerated Data-Driven Discovery and Screening of Two-Dimensional Magnets Using Graph Neural Networks

Abstract: In this study, we employ Graph Neural Networks (GNNs) to accelerate the discovery of novel 2D magnetic materials which have transformative potential in spintronics applications. Using data from the Materials Project database and the Computational 2D materials database (C2DB), we train three GNN architectures on a dataset of 1190 magnetic monolayers with energy above the convex hull ($E_{\text{hull}}$) less than 0.3 eV/atom. Our Crystal Diffusion Variational Auto Encoder (CDVAE) generates 11,100 candidate crystals. Subsequent training on two Atomistic Line Graph Neural Networks (ALIGNN) achieves a 93$\%$ accuracy in predicting magnetic monolayers and a mean average error of 0.039 eV/atom for $E_{\text{hull}}$ predictions. After narrowing down candidates based on magnetic likelihood and predicted energy, constraining the atom count in the monolayers to five or fewer, and performing dimensionality checks, we identify 190 candidates. These are validated using Density-Functional Theory (DFT) to confirm their magnetic and energetic favorability resulting in 167 magnetic monolayers with $E_{\text{hull}} < 0.3$ eV/atom and a total magnetization of $\geq$ $0.5 \mu_{B}$. Our methodology offers a way to accelerate exploring and predicting potential 2D magnetic materials, contributing to the ongoing computational and experimental efforts aimed at the discovery of new 2D magnets.