High-Throughput Combinatorial Experiments and Data Mining for Metallic Glass Discovery

Developing modern complex alloys, such as bulk metallic glasses, is a long and costly process. Here, we introduce key strategies used by the Schroers group and collaborators to speed up the process: The combinatorial sputtering method provides high-throughput sample fabrication. Paired with both established and newly developed high-throughput characterization methods, we have built a treasure of high-quality data comprising thousands of alloys. Using our online data repository “MAP”, we openly share the data and access it for data mining. Through modern machine learning methods, we aim to accelerate the discovery of bulk metallic glasses and understand what determines an alloy’s glass forming ability.