Machine Learning-Guided Exploration of Glass-Forming Ability in Multicomponent Alloys

The prediction of glass-forming ability (GFA) in alloy systems is a challenging problem material science as well for metallurgical applications. In this study, we build artificial neural network (ANN) models to investigate the GFA multicomponent alloys, based on datasets assembled from ternary alloys quinary prepared by magnetron sputtering. Through training ANN with different combinations datasets, tackle influence data source model performance, especially generalizability predicting unseen systems. trained combined dataset exhibits best specifically low root mean square error leave-one-alloy-system-out validation and high robustness, several CoCrFeNi-based alloys. To further verify models, synthesize CoCrFeNi-Mo metallic thin films co-sputtering characterize structure phase information via x-ray diffraction electron microscopy. outcomes our experiments agree reasonably predictions, indicating that data-driven machine learning approach can be useful tool future design amorphous