AI-powered the toughest biohydrogels

Hydrogel soft materials hold immense promise for applications ranging from bionic soft robots to flexible human-machine interfaces, but realizing this potential critically depends on excellent mechanical properties. While substantial progress has been made in toughening hydrogels, concurrently achieving a significant enhancement in strength remains a formidable challenge, thereby limiting their functional use. This work introduced a CPTR (centrifugation - progressive training - restorative soaking) strategy, which prestructured the material through centrifugation, continuously evolved and optimized the structure through progressive training, and further refined and locked the structure through soaking. The mechanical properties were improved through the three processes synergistically. By inputting the process parameters and corresponding mechanical test results, the AI model analyzed feature importance, ranked optimal performance combinations, and recommended new schemes. Through such iterative cycles, a CPTR hydrogel with an outstanding tensile strength of 134.31 MPa and toughness of 10.25 MJ/m³ was achieved, currently the highest strength among biohydrogels. Its excellent mechanical performance and processability allowed the hydrogel to be constructed into fibers and network structures, opening new avenues for developing next-generation soft actuators, robust controllable release systems, and other advanced functional materials.