MIT engineers have shown that a single layer of living muscle cells can generate enough thrust to move through water if the design is optimized. In a paper published in Advanced Functional Materials, the team describes a robot whose “skeleton” is a gel film roughly the size of a stick of gum. Both sides of the gel are stamped with square grooves, forming two fins; each fin is coated with a layer of live muscle cells thinner than a hair. The cells are genetically engineered to contract when illuminated. Shining light on one fin makes its muscle twitch, flapping the fin and pulling the robot forward. By alternating or synchronizing light on the two fins, the swimming direction and speed can be controlled. The robot was demonstrated navigating a simple watery maze, reaching a peak speed of about four body lengths per minute—slow compared with Olympic swimmers (≈65 body lengths per minute) but comparable to leisurely swimmers such as the cow shark. The authors note that most bio‑hybrid robots use bulky 3‑D muscle blocks requiring millions of cells, whereas this ultra‑thin 2‑D design is cheaper to fabricate and more efficient, and could eventually explore fragile or unpredictable environments because living tissue is soft, responsive, and self‑healing.
To increase the force generated by the muscle layer, the researchers systematically tuned three aspects of the gel skeleton: composition, stiffness, and groove geometry. The initial scaffold of fibrin shrank quickly under muscle force, so they switched to gelatin methacrylate (GelMA) and varied its formulation to produce gels of different stiffnesses. Stiffer gels promoted better cell alignment and stronger fiber formation. Square‑bottomed grooves guided cells to align more effectively than curved valleys. A film thickness of about 0.5 mm provided enough support while remaining light enough for the cells to stay attached during contraction. The team then “exercised” the muscles with flashing light, strengthening their contractions. The final two‑fin robot, with independently addressable muscles, was placed in a water‑filled petri dish and guided through a maze by moving a light source over it.
The current design mainly proves that a thin muscle layer can produce sufficient thrust; future work will focus on optimizing the body shape for faster swimming. Even at modest speeds, such muscle‑powered swimmers could be useful for tasks like environmental monitoring in aquatic settings.
Review