机器学习分子动力学辅助材料凝固形核研究进展

陈名毅, 胡俊伟, 余耀辰, 牛海洋

Advances in Machine Learning Molecular Dynamics to Assist Materials Nucleation and Solidification Research

CHEN Mingyi, HU Junwei, YU Yaochen, NIU Haiyang

图4 Si的凝固形核过程模拟研究^[66]

Fig.4 Research on the nucleation process of Si^[66]
(a) distributions of the order parameters using local structure factor S_i (q₁) in the liquid and solid phase (Inset shows the structure factor patterns and the position of the module of the wave number

q_{1}

, which corresponds to the first peak of the XRD spectrum shown in the inset)
(b) distributions of the configurations as a function of the structure factor at the first peak in well-tempered sampling, liquid phase, and solid phase, respectively (Insets present the typical configuration, whose structure factor peak corresponds to the position of abscissa. Disordered configurations are colored in grey, and ordered ones, i.e., ice-like structures, are colored in blue. The same applies to Fig.4c)
(c) free energy surface of the solidification at different temperatures (k_B—Boltzmann constant, T—temperature)
(d) snapshots of Si nuclei