Abstract
Efficiently resolving the high-dimensional energy landscapes of topological spin textures remains a major challenge in computational magnetism. Here, we develop a high-throughput algorithm that integrates orthogonal-spin optimization with the Limited-Memory-Broyden– Fletcher–Goldfarb–Shanno (M-LBFGS) method to accelerate the elucidation of phase-diagrams of spin textures on-demand. By enforcing strict orthogonality constraints during optimization and exploiting the quasi-Newton efficiency of M-LBFGS, our algorithm achieves rapid, robust convergence to local minima while requiring orders of magnitude fewer iterations than conventional techniques. This enables high-throughput screening across thousands of parameter combinations using only modest computational resources. Applied to the van der Waals magnet CrSBr, our framework resolves more than 2000 atomistic configurations and uncovers a hierarchy of previously unknown metastable phases, including domain-wall bimerons, bimeron chains, Néel skyrmions, and hybrid spiral states. Minimum energy-path techniques coupled with our solver reveal their stability, transition mechanisms, and thermal lifetimes, establishing a direct link between the emergence of topological spin textures and the noncritical anomaly observed experimentally in the susceptibility of CrSBr below 40 K. This optimization algorithm defines a generalizable and scalable paradigmatic for predictive mapping of magnetic phase diagrams, enabling the discovery of intricate spin textures and emergent topological states across a wide spectrum of magnetic materials.
| Original language | English |
|---|---|
| Article number | e78489 |
| Journal | Advanced Functional Materials |
| Early online date | 24 Sept 2026 |
| DOIs | |
| Publication status | E-pub ahead of print - 24 Sept 2026 |
Keywords
- maxima and minima
- magnet
- magnetism
- metastability
- skyrmion
- solver
- van der waals force
Fingerprint
Dive into the research topics of 'Accelerated Discovery of Topological Spin Phase Diagrams in CrSBr Magnets Via High‐Throughput Optimization'. Together they form a unique fingerprint.Cite this
- APA
- Author
- BIBTEX
- Harvard
- Standard
- RIS
- Vancouver