Research Webzine of the KAIST College of Engineering since 2014
Fall 2025 Vol. 25By combining AI-guided design with nanoscale 3-D printing, researchers have built a carbon ‘nanolattice’ so light that it can perch on a soap bubble yet strong enough to carry millions of times its own weight.

A new material as light as a whisker and as strong as steel has arrived. Collaboration between Professor Seunghwa Ryu at KAIST and Professor Tobin Filleter at the University of Toronto has produced an ultralight carbon nanolattice that could reshape how engineers tackle weight-sensitive designs in cars, aircraft, and next-generation mobility devices.

The first challenge was to discover the perfect shape for the tiny beams that form the lattice. The team turned to a smart search method called multi-objective Bayesian optimization. Think of it as a digital craftsman that learns, with each trial, how to raise the stiffness and strength while trimming the mass. Remarkably, only about 400 simulated designs were enough to pinpoint the sweet spot—far fewer than the thousands of experiments traditional approaches demand.
Manufacturing this blueprint required precision on a nearly unimaginable scale. Using two-photon polymerization—a laser technique where a polymer hardens only when two light particles meet at the same point—the researchers printed the intricate lattice with features roughly one-hundredth the width of a human hair. A subsequent high-temperature bake burned away the plastic scaffold, leaving a shell of tough pyrolytic carbon. By splitting the laser into many focal points, the team retained nano-level accuracy while scaling the structure up to millimeter dimensions, a crucial step toward practical parts.

Tests showed that the new lattice outperforms existing lightweight materials in terms of both stiffness and strength at the same weight. Such a combination means that electric vehicles could travel farther on the same battery, rockets could lift heavier payloads while using less fuel, and smartphones or wearables could become thinner yet more resistant to drops.
The work, led by Dr. Peter Serles and Dr. Jinwook Yeo as co-first authors, was published in Advanced Materials (March 2025) under the title “Ultrahigh Specific Strength by Bayesian Optimization of Lightweight Carbon Nanolattices.” Funding came from Korea’s Ministry of Science and ICT (ERC Program) and the Korean MFDS M3DT initiative. Experts note that the merger of AI-guided design and nanoscale 3-D printing showcased here signals a new standard for future lightweight structures.
A New solution enabling soft growing robots to perform a variety of tasks in confined spaces
Read moreAI-Designed carbon nanolattice: Feather-light, steel-strong
Read moreDevelopment of a compact high-resolution spectrometer using a double-layer disordered metasurface
Read moreWearable hyperspectral photoplethysmography for the continuous monitoring of exercise-induced hypertension
Read moreSmarter AI through AI-generated feedback
Read more