The shape change is pre-programmed. No assembly required.
For millions of years, nature has encoded motion into slender biological fibers — the curling tendril, the grasping trunk — through arrangements of matter too intricate for human hands to replicate. Now, researchers at Harvard have crossed a threshold: using a rotating 3D-printing technique, they have manufactured synthetic filaments whose capacity for bending, twisting, and contracting is written into their molecular architecture before they ever leave the printer. The work suggests that the ancient partnership between structure and movement, long the exclusive province of living things, may be entering a new chapter authored by engineering.
- The core tension is one of translation — biology has long mastered programmable motion in soft fibers, but synthetic materials have lagged far behind, requiring cumbersome assembly and post-processing to approximate what muscle does effortlessly.
- Harvard's rotational multimaterial 3D printing disrupts that bottleneck by embedding active liquid crystal elastomers and passive polymers in precise helical patterns during fabrication itself, eliminating the need for any subsequent assembly.
- The filaments have already been woven into working devices — lattices that open and close like filters, grippers that seize and release rods with temperature alone — validating that the programmed behavior survives the leap from single strand to complex structure.
- Scaling is actively underway, with filaments already reaching 100-micron diameters and researchers eyeing custom nozzles that could integrate liquid metal channels and other functional materials into future iterations.
- The trajectory points toward soft robotic grippers, on-demand valves, and injectable biomedical scaffolds — applications where the absence of motors or rigid joints is not a limitation but the entire point.
Look closely at a grapevine curling around a trellis, or an elephant's trunk performing feats of both delicacy and power. Nature has spent millions of years perfecting controlled motion through slender, flexible structures. Harvard scientists have now found a way to replicate that capability synthetically — printing filaments that behave like programmable muscles, responding to nothing more than heat.
The method, developed in Jennifer Lewis's lab at Harvard's School of Engineering and Applied Sciences, is called rotational multimaterial 3D printing. A rotating nozzle lays two materials side by side as it prints: a liquid crystal elastomer that contracts along its molecular alignment when heated, and a passive elastomer that holds its shape regardless of temperature. When the active material contracts on one side of the filament while the passive material resists on the other, the structure bends. Rotate the nozzle during printing, and a helical twist is written into the filament's architecture before it ever leaves the machine. No assembly. No post-processing.
Postdoctoral researcher Mustafa Abdelrahman, who led the work published in the Proceedings of the National Academy of Sciences, had spent years working with liquid crystal elastomers through conventional means. Encountering Lewis's platform, he saw a different possibility: embedding active materials directly into a filament's structure during fabrication itself. The team validated the approach with sinusoidal filaments that look identical to the eye but behave oppositely when heated — one straightening and expanding, the other shrinking and contracting, depending solely on where the active material was placed within the cross-section.
From that foundation, the researchers built more complex structures: flat lattices functioning as active filters that open to pass particles when heated and contract to trap them when cooled, and free-standing grippers that lower onto rods, heat to grip, then cool to release. In one demonstration, a lattice with alternating active regions transformed into a dome shape when submerged in hot oil, precisely matching computer simulations.
The team has already printed filaments as thin as 100 microns and sees room to go smaller. Graduate student Jackson Wilt noted that future nozzles could incorporate liquid metal channels or other functional materials, expanding the design space further. Lewis envisions the framework accelerating liquid crystal elastomers out of research settings and into real-world technology — soft robotic grippers, temperature-controlled valves, and injectable filaments that lock together to form porous biomedical scaffolds. The work represents a fundamental shift in how engineers conceive of motion in soft materials: not through motors or hydraulics, but through the careful orchestration of matter that responds to something as simple and universal as heat.
Look closely at a grapevine curling around a trellis, or an elephant's trunk lifting a peanut with the same muscles that can topple a tree. Nature has spent millions of years perfecting the art of controlled motion through slender, flexible structures that bend and twist with precision. Harvard scientists have now figured out how to replicate this capability in the laboratory, using a 3D printing technique to manufacture synthetic filaments that behave like programmable muscles, responding to nothing more than changes in temperature.
The breakthrough emerged from Jennifer Lewis's lab at Harvard's School of Engineering and Applied Sciences, where her team developed what they call rotational multimaterial 3D printing. The method sounds deceptively simple: print two different materials side by side through a rotating nozzle, layering them in exact patterns around a filament's cross-section. One material is a liquid crystal elastomer, a polymer that shrinks along its molecular alignment when heated above a certain temperature. The other is a passive elastomer that holds its shape regardless of heat. When the active material contracts on one side of the filament while the passive material resists on the other, the whole structure bends. Rotate the nozzle as you print, and you can write a helical twist into the filament's molecular structure before it ever leaves the printer. The shape change is pre-programmed. No assembly required. No post-processing needed.
Postdoctoral researcher Mustafa Abdelrahman, who led the work published in the Proceedings of the National Academy of Sciences, had spent years coaxing liquid crystal elastomers into useful shapes using conventional methods. When he encountered Lewis's rotational printing platform, he saw something different: a way to embed active materials directly into the filament's architecture during fabrication. "What if we plug in active materials and pattern them within the filament?" he asked. The question led to a systematic exploration of how to control shape change through precise material placement.
The team validated their approach by printing sinusoidal filaments—wavy strands that look identical to the eye but behave completely differently when heated. Place the active material on the outside of the wave's curve, and heat causes the filament to straighten and expand. Place it on the inside, and the same thermal stimulus makes the filament shrink and contract. This programmability became the foundation for building more complex structures. The researchers wove these filaments into flat lattices that function as active filters: when heated, they open to allow spherical particles to pass through; when cooled, they contract and trap the particles. They also fabricated a kind of gripper—free-standing lattices that can be lowered onto rods, heated to grip them, then cooled to release them. In one demonstration, a lattice with alternating expanding and contracting regions transformed into a dome shape when submerged in hot oil, matching the form predicted by computer simulations.
The technology is already moving toward practical scale. The team has printed filaments as thin as roughly 100 microns in diameter and sees room to go smaller with custom nozzles and refined inks. Graduate student Jackson Wilt noted that future nozzles could integrate other materials—liquid metal channels for additional actuation, or other functional components—expanding the possibilities further. Liquid crystal elastomers remain largely confined to research settings, but they are being actively explored for soft robotics, energy absorption systems, and biomedical applications. Lewis envisions the filament design and printing framework accelerating the transition from laboratory curiosity to real-world technology.
The potential applications are substantial. Soft robotic grippers could gently manipulate multiple objects simultaneously without rigid joints or complex mechanical systems. Temperature-controlled valves and filters could adjust their porosity and flow pathways on demand. Entangled, injectable filaments could lock together in place to form porous structures useful in medical contexts where rapid tissue clotting is needed. The work represents a fundamental shift in how engineers think about creating motion and control in soft materials—not through motors or hydraulics, but through the careful orchestration of materials that respond to something as simple and universal as heat.
Notable Quotes
This filament design and printing framework could accelerate the transition of artificial muscle-like materials from the lab to real-world technologies.— Jennifer Lewis, Harvard School of Engineering and Applied Sciences
You could create more complex nozzles that integrate with other materials in the future—like, having a liquid metal channel to enable actuation, or integrating other functionality.— Jackson Wilt, graduate student and co-author
The Hearth Conversation Another angle on the story
Why does this matter? We already have robots and grippers. What's the advantage of making them respond to temperature?
Traditional grippers use motors, gears, rigid joints—all that machinery adds weight, complexity, and points of failure. These filaments are soft, lightweight, and they move because of their material properties, not mechanisms. A robot hand made this way could handle delicate objects without crushing them, and it requires far fewer moving parts.
So it's about gentleness and simplicity. But how do you control something that just responds to heat? Doesn't that seem imprecise?
That's where the printing comes in. You're not just making a filament that responds to heat—you're programming exactly how it responds by placing active and passive materials in specific patterns during fabrication. The same thermal input produces different motions depending on where you put the active material. That's the programmability.
And they can make these filaments very small?
They've already printed them down to about 100 microns in diameter. That's roughly the width of a human hair. They think they can go smaller. At that scale, you could build structures with intricate detail—filters that sort particles, medical devices that work inside the body.
What's the biggest hurdle now?
Getting from lab demonstrations to manufacturing at scale. Liquid crystal elastomers are still mostly research materials. But the printing method itself is elegant and reproducible, which is what you need for industrial adoption. If they can refine the inks and nozzles, the path to real products becomes much clearer.
When do you think we'll see these in actual use?
That depends on which application moves fastest. Soft robotics is probably closest—there's real commercial interest there. Biomedical devices might follow. But the framework is solid enough that it could accelerate things significantly.