Printable aluminum alloy set strength records, may enable lighter aircraft parts (news.mit.edu)

MIT engineers have used machine learning-guided alloy design plus 3D printing to create a printable aluminum alloy that is five times stronger than cast aluminum and about 50% stronger than alloys found using conventional simulations. Rather than exhaustively simulating more than one million possible element combinations, their ML pipeline homed in on an optimal composition after evaluating just 40 candidates. The predicted recipe—aluminum alloyed with five additional elements—was produced as powder and printed using laser powder bed fusion (LPBF). Rapid layer-by-layer melting and solidification produced a high volume fraction of fine precipitates, yielding the exceptional strength and thermal stability (up to ~400°C) measured in physical samples. This work is significant because it demonstrates how ML can radically shrink alloy design search space and unlock microstructures compatible with additive manufacturing. The combination of a targeted compositional search and LPBF’s fast cooling enables properties previously limited to cast or wrought alloys, opening possibilities for lighter, cheaper, high-temperature parts—e.g., jet engine fan blades that could replace heavier and costlier titanium components, advanced cooling hardware, and high-performance automotive or data-center parts. The methodology, detailed in Advanced Materials, offers a blueprint for co-designing alloys specifically for additive processes rather than retrofitting existing alloys to printing.