Data-driven machine learning prediction of glass transition temperature and the glass-forming ability of metallic glasses

The limited glass-forming ability (GFA) poses a significant challenge for the practical applications of metallic glasses (MGs). The development of high-GFA MGs typically involves trial-and-error processes to screen materials with a large critical diameter (Dmax), which serves as a criterion for determining the GFA. The formation and stability of MGs are influenced by the glass transition temperature (Tg). Over the past decade, the emergence of machine learning (ML) has shown great promise in the exploration of high-GFA materials. However, the contribution of material features to Tg and Dmax predictions, as well as their correlations, remains ambiguous, posing a challenge to achieving high prediction accuracy. Herein, we present a comprehensive dataset consisting of 1764 datapoints for Tg and 1296 datapoints for Dmax. The governing rules for GFA have been established through feature significance analysis. The light gradient boosting (LGB) model exhibits remarkable accuracy in predicting Tg, utilizing sixteen features, achieving a coefficient of determination (R2) score of 0.984 and a root mean square error (RMSE) of 20.196 K. An integrated ML model, based on the weighted voting of three basic models, is developed to enhance the accuracy of Dmax prediction, achieving an R2 score of 0.767 and an RMSE of 2.331 mm. Additionally, a GFA rule is proposed to explore materials with large Dmax values, defined by satisfying the criteria of a thermal conductivity difference ranging from 0.60 to 1.32 and an entropy density exceeding 1.05. Our work provides valuable insights into Tg and Dmax predictions and facilitates the exploration of potential high-GFA MGs through the implementation of a well-established ML model and GFA rules. The data-driven machine learning approach has greatly improved the predictive accuracy of Tg and Dmax values. The governing rules for GFA have been successfully established through feature significance analysis.