DxMag Heusler Database
收藏资源简介:
DxMag Heusler Database 是一个包含几乎所有传统三元Heusler化合物的数据库,本研究将其扩展到包括四元和全d Heusler化合物。该数据库为机器学习模型提供了训练数据,用于预测化合物的形成能、距离凸包的距离、局部磁矩、声子稳定性、磁稳定性和磁晶各向异性能量。通过机器学习加速的高通量筛选方法,研究人员成功筛选出具有大磁晶各向异性能量的稳定Heusler合金。
The DxMag Heusler Database is a repository containing nearly all conventional ternary Heusler compounds. In this study, we expanded its scope to include quaternary and full-d Heusler compounds. This database provides training data for machine learning models to predict multiple properties of compounds, including formation energy, distance to the convex hull, local magnetic moment, phonon stability, magnetic stability, and magnetocrystalline anisotropy energy. Through machine learning-accelerated high-throughput screening approaches, researchers successfully screened out stable Heusler alloys with large magnetocrystalline anisotropy energy.

- 1Accurate Screening of Functional Materials with Machine-Learning Potential and Transfer-Learned Regressions: Heusler Alloy Benchmark日本国立材料科学研究所 · 2025年



