Forward-looking: For thousands of years, mechanical gears have driven human invention – from ancient chariots to robotic arms. Now, for the first time, engineers have built a gear that doesn't rely on solid teeth or even physical contact. At New York University, a team of researchers has demonstrated that gears can function purely through fluid dynamics, creating a contactless mechanism capable of transmitting motion with remarkable precision and adaptability.
The study, published January 13 in Physical Review Letters, replaces the metal or plastic cogs of conventional gears with controlled flows of liquid. In their experiments, NYU physicists submerged two cylinders in a viscous water – glycerol mixture. When one cylinder rotated, the liquid currents it generated transmitted motion to the other – mimicking the performance of classical gears, but without any interlocking parts.
The results depended on the spacing between the cylinders. At short distances, swirling fluid formed micro-scale vortices that caused the second cylinder to spin in the opposite direction, replicating traditional gear behavior.
When the cylinders were spaced farther apart, the same flow looped around like an invisible belt, pulling both rotors in the same direction. The discovery demonstrated two distinct modes of motion transfer, both driven entirely by fluid flow.
NYU professor Jun Zhang, who led the project alongside mathematics professor Leif Ristroph, highlighted the broader significance of the finding: the ability to tune rotation speed and direction without solid contact offers a fundamental redesign of the gearbox itself.
Because the components never touch, the mechanism is immune to jamming and resistant to debris – problems that plague metal-based systems across industries. In traditional machinery, even a single grain of sand or a minor misalignment can halt an entire operation. In the fluid-driven system, the liquid simply flows past the obstruction.
Beyond durability, fluid gears offer a level of flexibility that mechanical versions cannot match. Adjusting flow speed or viscosity could instantly change a system's gear ratio, creating a self-adjusting drive mechanism ideal for soft robotics and adaptive structures. Instead of lubricating moving steel components to prevent wear, engineers might one day replace them entirely with carefully tuned liquids.
In soft robots, for example, fluid-driven gears could sit between a standard motor and a flexible joint, converting rigid rotary motion into smooth, controlled movement within a sealed limb. Designers could power silicone-based arms through a shared liquid circuit rather than installing separate gearboxes at every joint.
Because the parts do not touch, the system reduces hard pinch points and stress areas – a critical advantage for robots operating near people or delicate objects. The same principle could enable distributed actuation in pneumatic or hydraulic soft robots, where a network of fluid gears distributes flow and torque across different sections of the body. Engineers would still need to balance viscosity, pressure, and response time, but they would gain a new way to shape motion without adding rigid hardware.
Liquid gears could challenge 5,000 years of mechanical engineering
