High-Throughput Experimental Strategies: The Highway to New Complex Alloys and Understanding Complex Phenomena
Metallurgical research is struggling to build a satisfactory understanding of various complex alloy classes and phenomena. For many of these problems, such as glass formation, single phase solid solution formation, or supercooled melt viscosities, computational approaches are overwhelmed. Experiments, by contrast, can directly evaluate real materials systems to generate valuable data. Unfortunately, experiments remain costly and time consuming, severely restricting the progress of materials discovery. This dissertation pushes to overcome this obstacle through a comprehensive high-throughput research strategy: The experimental approach is centered around combinatorial co-sputtering and rapid characterization, investigating hundreds of alloys at a time while maintaining highest sample and data quality. These large amounts of data are processed in a streamlined fashion, analyzed using advanced techniques including data science tools, and openly shared using the research group’s own online data repository “MAP”. Thereby, new patterns and mechanisms are revealed on an unprecedented scale. This dissertation begins with an overview over numerous studies conducted in this spirit on High Entropy Alloys, Metallic Glasses, Nanocrystalline Alloys, and Quasicrystalline Alloys. The focus then narrows on phase formation in multicomponent alloys for a detailed discussion. First, data on 2,500 High Entropy Alloy compositions reveal a BCC crystal structure preference for solid solutions with large atomic size differences. This asymmetry constitutes a fundamental extension to the Hume-Rothery rules. Then, the key role of metastability is examined. Metastable solid solutions form through polymorphic solidification and a unified view of the competition between solid solutions and glasses is proposed. Finally, the high-throughput approach is taken to a new level on two fronts: Pushing the throughput limit, new data on an unprecedented number of 12,500 alloys are presented to investigate glass formation across 20 different elements. Pushing the characterization limit, a new method is developed to measure metallic glass viscosities and fragilities across wide composition ranges. Finally, key findings are summarized that will enable next generation high-throughput experiments.