A density functional theory study of CrSBr for spin-orbit-torque magnetic random-access memory

Abstract

The explosive growth of artificial intelligence and data center infrastructure is driving a surge in global electricity demand. Projections indicate that data centers could consume 12% of U.S. electricity by 2028, and global electricity consumption could double by 2030 (Lawson, 2026). Conventional, charge-based memory architectures suffer from the von Neumann bottleneck and high leakage power, motivating the development of spintronic alternatives (Chen, 2023). Among spintronic memory technologies, the spin-orbit-torque magnetic random-access memory (SOT-MRAM) stands out for its speed, endurance, and read-write separation (Song, 2023). To make this technology practical at scale, it must overcome its current external-field requirements for deterministic switching. This work investigates CrSBr, an air-stable van der Waals antiferromagnet (AFM), as a symmetry-breaking layer for exchange-bias SOT ferromagnets (FM) (Ziebel, 2024). Using DFT calculations, this study examines FM, AFM, and hydrogen-doped CrSBr variants. Geometry optimizations yield lattice parameters in excellent agreement with experiment (Ziebel, 2024). The calculated magnetic moments confirm the expected A-type AFM ordering, and reveal that while the hydrogen dopant largely preserves the interlayer AFM, it introduces a small net moment that drives a semiconductor-to-metal transition. Band structures and density-of-states show spin splitting in the FM phase and degeneracy in the AFM phase. These results show the destructive effect hydrogen contamination may have on these systems, as well as potential opportunities for bandgap engineering through light doping.