ROBNEWS
A pumpkin-seed-sized liquid-metal pump could finally untether soft robots
Bristol's 0.2 gram device runs on less than a volt, a combination soft robotics labs have chased for years to free flexible material robots from bulky hardware.
Bristol's 0.2 gram device runs on less than a volt, a combination soft robotics labs have chased for years to free flexible material robots from bulky hardware.
Weighing less than a dried pumpkin seed and running on a fraction of the voltage of a watch battery, a new liquid-metal pump from the University of Bristol could be the component that finally lets soft robots leave the lab bench behind.
Soft robots, in the materials-and-mechanical sense rather than the software one, are machines built from flexible elastomers, gels, and textiles that bend, stretch, and squish like living tissue. For two decades the field has promised applications that depend on exactly that compliance: wearable assistive clothing that moves with the body, medical implants that conform to organs, search-and-rescue robots that squeeze through rubble, and haptic feedback sleeves for virtual reality. The catch has been plumbing. Most soft robots have still needed rigid, bulky, mains-fed pumps and compressors to push air or fluid through their actuators, which is a bit like building a rubber octopus and then strapping a refrigerator to its back. The Bristol team, reporting in Nature Communications, argues its new device removes that tether.
The pump, called LIMA (for Liquid Metal Magnetohydrodynamic Actuator), weighs 0.2 grams — about as much as a single dried pumpkin seed — and runs on less than 1 volt, orders of magnitude below the tens to thousands of volts that conventional dielectric elastomer or ionic soft actuators typically demand. That combination matters: it means the pump can sit on a small coin cell or harvest energy from the low-voltage electronics already common in wearables, instead of dragging a tether or a transformer behind every soft body. (The Bristol press release states the operating voltage as less than 0.1 V, while the paper gives the ceiling as under 1 V; both figures are sourced to the same team, and the draft treats 1 V as the conservative upper bound.)
The mechanism is unusually tidy. A droplet of gallium-based liquid metal sits inside a fluid-filled soft channel placed above a small permanent neodymium magnet. When a current passes through the droplet in the magnetic field, the Lorentz force pushes the droplet to oscillate. The droplet, in turn, displaces the surrounding fluid and pumps it through the channel. There is no piston, no valve, no rotor, and no rigid moving part. The same droplet acts as motor, piston, and actuator, which is why the team can build the entire device out of soft polymers. The paper quantifies the pump's output at 18.6–34.88 GPa·m⁻³ of specific pressure and 38.4–49.29 kL·min⁻¹·m⁻³ of specific flow rate, metrics the authors say surpass existing soft and conventional pumps.
The team has shown three early prototypes, all still benchtop proof-of-concept. A soft robotic butterfly with flapping wings is driven by a single LIMA pump, with no external compressor. A color-changing wearable bracelet circulates fluid through adaptive materials that shift hue as the liquid moves. A haptic interface uses a soft fingertip pouch and an adjustable wristband that squeezes the skin to simulate touch, with the pump setting how hard it presses. Each is a way of showing that a soft body can carry its own drive, not just react to a tethered one.
Lead author Saba Firouznia, a research associate in Bristol's Soft Robotics Lab, and lab head Jonathan Rossiter frame the work as the missing piece for a field that has long been constrained by the kind of rigid, bulky hardware Rossiter has publicly likened to a robotic "right trousers" problem. The team's press release sketches potential destinations that include lab-on-a-chip disease diagnostics, robotic clothing, environmental sampling, smart medical implants, VR and teleoperation haptics, smart bandages, and even edible robots. The same release goes further, calling LIMA potentially able to "outperform existing soft pumps and even some commercial pumps" — an assertion the Bristol team makes for its own device, not an independently verified benchmark against third-party products.
It is also worth being clear about what the device does not yet do. The Nature Communications paper is a demonstration of feasibility, not a product spec sheet. The prototypes are small, slow, and benchtop-bound. Soft robotics has spent a decade producing elegant bodies tethered to inelegant hardware. LIMA is a credible step toward shrinking the hardware to match, but the field's hardest test — sustained autonomous operation outside a controlled lab — still lies ahead.