|About this Abstract
||2022 TMS Annual Meeting & Exhibition
||Bulk Metallic Glasses XIX
||Evidence of Crystalline Phase Precursors within High-Temperature Metallic Liquids and their Effect on Glass-Forming Ability
||Porter Weeks, Katharine Flores
|On-Site Speaker (Planned)
While much research has focused on the evolution of structure within high-temperature and undercooled metallic liquids, the general lack of an effective technique to quantitatively describe the atomic-scale structure of the liquid has made its correlation with lower temperature phenomena extremely challenging. Here, we apply a machine learning-based approach to identify and analyze atomic clusters in a variety of simulated binary (Cu-Zr, Ni-Nb, Ni-Zr) and ternary (Al-Ni-Zr) liquids. We show that the structure within the high-temperature liquid yields clear evidence of precursors for the low-temperature crystalline phases. We propose that there are two criteria for glass-forming ability (GFA): a lack of order within the liquid state, as quantified by the population distribution of atomic clusters, and a general lack of similarity of those structures to stable or metastable crystalline phases. This research suggests that a significant amount of information regarding GFA is buried within the structure of the high-temperature liquid.
||Modeling and Simulation, Machine Learning,