Data-Guided Mechanistic Insights into Dislocation Creep in Eutectic Alloys

NON-TECHNICAL SUMMARY Metals used in engines, electronics, and energy systems can slowly deform when exposed to heat and mechanical loads over long periods. This slow process of deformation is called “creep”. This project is studying how small structural patterns within aluminum-cerium alloys govern creep, so future high-temperature materials can be designed more reliably. This research project is developing a data-driven approach to understanding why some eutectic alloys resist creep better than others. Eutectic alloys contain two or more phases arranged in fine patterns, such as layers, rods, or connected networks. These patterns can help carry loads and slow deformation, but their effects are not yet well understood to predict long-term performance. This project is creating a systematic dataset that connects alloy composition, processing route, heat treatment, internal structure, and creep behavior. This project is also using machine learning to identify which structural features matter most, followed by advanced microscopy and diffraction experiments to explain the underlying physical mechanisms behind the trends. This project supports the national interest by advancing the scientific foundation needed to design more durable, lightweight, and heat-resistant materials for transportation, electronics, and energy technologies. This project is also strengthening the domestic knowledge base for alloys containing cerium which is a rare-earth element of strategic importance. The project is training graduate and undergraduate students in materials testing, microscopy, data science, and machine learning and conducting outreach through the University of Cincinnati engineering summer camps where it is introducing middle and high school students to materials science and data-driven discovery. Moreover, datasets and software tools generated by this project are being shared freely and broadly to support future research and education across the field. TECHNICAL SUMMARY This research project is developing a predictive, mechanistic framework for understanding how eutectic microstructure governs dislocation creep resistance in aluminum-cerium and silicon-modified aluminum-cerium alloys. The central goal of this research project is to determine which microstructural features, including lamellar spacing, colony size, morphology, connectivity, interface structure, and microstructural stability during deformation, most strongly control creep rate, effective stress exponent, and threshold stress. This project is generating a paired pre-creep and post-creep dataset for approximately 100 conditions. Four compositions are being processed by casting, arc melting, and laser powder bed fusion, followed by heat treatments designed to produce a broad range of eutectic length scales and morphologies. Short-duration creep tests at 350°C with stress-stepping protocols are being used to extract multiple creep metrics for each condition. Optical microscopy, scanning electron microscopy, X-ray diffraction, and electron backscatter diffraction are also being used to quantify initial microstructures and changes after creep. Machine learning models, including decision-tree ensembles and neural networks, are being trained to predict multiple creep outcomes and to rank the most influential descriptors. Inverse design and Pareto front analysis are being used to identify trade-offs among creep resistance metrics. Targeted advanced characterization, including transmission electron microscopy, atom probe tomography, neutron or synchrotron diffraction, and high-resolution orientation mapping, is also being used to validate the physical mechanisms behind the data-driven feature rankings. Broader impacts include training students across experiments and data science, releasing open datasets and software, supporting undergraduate research experiences, and developing outreach activities for students within the local school system with a focus on materials science and data-driven discovery. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.