Mathematical framework translates biological mechanics into manufacturable adaptive materials

MIT researchers have created a mathematical framework that captures how mechanisms across length scales in natural systems produce complex behaviors and translates them into engineered, manufacturable materials. The work, published in the Journal of the Mechanics and Physics of Solids, uses category theory to map hierarchical biological structures into composable building blocks that can be mathematically validated and 3D-printed.

The framework was demonstrated on a pine cone, whose scales open in low humidity and close in damp conditions due to interactions among cellulose fibers, laminas, and tissue layers. Researchers modeled each structural level as an independently validated building block, then applied mathematical rules to ensure valid transitions between scales. This preserves the stimulus-response relationships that drive the natural behavior in a synthetic counterpart.

Lead author Lee Marom, a graduate student, describes the approach as moving beyond bio-inspiration to 'bio-derivation,' where the actual mechanisms producing a behavior are systematically transferred to an engineered system. Corresponding author Markus Buehler, the Jerry McAfee Professor of Engineering, notes that category theory makes cross-scale relationships explicit and transferable, turning nature into a library of composable mechanisms.

The system carries designs from multiscale biological mechanics through engineered realization to machine-executable fabrication code. In a demonstration, researchers combined building blocks from the humidity-driven bending of a pine cone and the twisting of a wheat awn to create a new thermal twisting actuator that performed as predicted without additional design work.

Gioele Zardini, assistant professor of Civil and Environmental Engineering, says the framework's systematization allows reuse of verified components, saving significant computation. The researchers plan to apply the framework to more complex biological systems and integrate artificial intelligence models to accelerate discovery of new adaptive materials.

The long-term vision, according to Buehler, is 'physical AI' that can reason about physical mechanisms and turn those ideas into matter. The work was supported by the MIT Lemelson Engineering Fellowship, Singapore DSO National Laboratories, and the MIT Generative AI Impact Consortium.