Machine Learning-guided Spinodal Alloys Beak Energy Absorption Records Across Extreme Temperatures
Jun 2026
news_20260604_fig0
Published in Nature Communications, a team led by Prof. Yong Yang from City University of Hong Kong has created bulk architected alloys that deliver extraordinary energy absorption while retaining full performance from room temperature to 873 K. Leveraging physics-informed machine learning to identify a spinodal compositionally complex alloy precursor, they fabricated centimeter-scale samples via electrochemical dealloying, achieving 106 MJ/m³ bulk energy absorption and 305 MJ/m³ in defect-free micro-samples—outperforming all state-of-the-art metallic foams and 3D-printed lattices.

The breakthrough stems from a seven-order hierarchical structure spanning atomic lattice distortions, 20 nm B2 nanoprecipitates, protective amorphous oxide layers, and interconnected micro-ligaments. This multiscale design enables heterogeneous strain hardening, with oxide layers and nanoprecipitates blocking dislocations to prevent catastrophic failure. Unlike 3D-printed lattices limited by scalability, the low-cost approach uses industrial-grade metallurgical and electrochemical processes and delivers lightweight materials (2.77–3.06 g/cm³, half the density of steels) that are ideal for aerospace and automotive crash protection, unlocking a new era of high-performance structural materials for extreme operating conditions. Here is the full article published in Nature Communications.