Acoustic tractor beams let researchers remotely program material stiffness

There is a problem at the center of soft robotics that nobody has fully solved: how do you make something that is simultaneously safe enough to be near a human and rigid enough to do useful work? The usual answers — pneumatic bladders, vacuum-jammed granular packets, heated shape-memory alloys — all require plumbing, tethers, or slow response times. A paper published in Nature Communications on February 6, 2026 proposes a new answer, and it comes from an unexpected direction: sound.

A team of researchers at the University of California San Diego, the University of Michigan, and the French National Center for Scientific Research (CNRS) has demonstrated for the first time that acoustic waves can controllably move a mechanical kink — a boundary between soft and stiff regions — through a metamaterial, reshaping its stiffness profile on demand. The lead researcher, Nicholas Boechler, an associate professor of mechanical and aerospace engineering at UC San Diego, calls it essentially an acoustic tractor beam that moves a kink and changes the way a material feels — while creating gradients of stiffness — on demand.

The key insight was architectural. Previous attempts to move kinks with sound waves produced chaotic, unpredictable motion. The UCSD-Michigan-CNRS team solved this by designing a metamaterial based on a Kane-Lubensky chain — a topological structure where the energy cost of moving the kink is essentially zero. That zero barrier is what makes step-by-step, programmable control possible. In their experimental setup, a chain of 18 human-scale rotors connected by polycarbonate springs, one disk oriented differently from the rest, acts as the kink. Send in a short acoustic pulse at the right frequency — roughly 15.65 Hz, in the mid-pass band — and the kink moves a few sites toward the sound source. Another pulse, a little farther. A longer continuous vibration pulls it all the way across, flipping which end of the chain is soft and which is stiff.

The mechanism is worth pausing on. The acoustic wave does not push the kink; it pulls it. Boechler describes the wave as reflecting off the kink and transferring momentum in a way that draws the kink toward the sound source — the acoustic equivalent of a tractor beam. This matters because it means you do not need physical access to the far side of the material. The control signal arrives from one direction only.

For robotics, the implications are theoretical but real. Variable stiffness is one of the core unsolved problems in compliant actuators. A warehouse gripper that can switch from soft enough to grasp a tomato to rigid enough to lift a crate, without a pneumatic line or a mechanical latch, would simplify an entire class of end-effectors. A wearable exoskeleton that could soften for walking and stiffen for lifting, controlled by an acoustic signal from a worn device rather than a tangle of tubes, would look very different from current designs. A medical implant — a stent, a prosthetic interface — that an external ultrasound transducer could tune in real time is not science fiction, but it is also not close.

The DOD connection is not incidental. This work was funded by the U.S. Army Research Office (grant W911NF-20-2-0182) and the U.S. Office of Naval Research (MURI N00014-20-1-2479), a multi-university grant that signals institutional commitment rather than a one-off curiosity award. The Army's interest in adaptive protective gear — helmets or body armor that can switch between conformable and impact-resistant states — is a plausible application. The Navy's interest in autonomous systems with morphing surfaces is another. Neither funder is cited as having any commercial stake in the research.

The current limitations are real and the authors are upfront about them. Right now, this is a toy model, Boechler told UC San Diego's news office. The system is one-dimensional. The damping in the current setup stops the kink after roughly four sites without continuous driving — not because of any fundamental energy barrier, but because viscous drag accumulates. Scaling to three dimensions, practical frequencies, and useful force output involves engineering problems that the paper does not address. There is no demonstrated application to an actual soft robot or implant.

There is also no closed-loop control. The stiffness gradient is set by the material's geometry — spring constants and lattice spacing — not by an active sensor telling the system what state it is in. You can move the kink, but you cannot yet verify where it is without some external measurement. That gap matters for deployable systems.

What the paper does establish is a mechanism. The zero-PN-barrier Kane-Lubensky approach to kink control is, as far as the research community is aware, novel in this context. The fact that controlled, directional movement is possible at all — rather than chaotic drift — is a genuine advance in the underlying physics of programmable matter. Boechler's lab has a track record in acoustic metamaterials and microscale sonogenetics, so the mechanism is not a one-off result sitting in a vacuum. The collaboration across UC San Diego, University of Michigan, and CNRS covers mechanical engineering, topology, and acoustics in a way that suggests the work is built to last.

Next steps, per the paper: a three-dimensional version, and exploration of whether similar effects could exist at smaller scales. Atomic-scale programmable matter is mentioned, but the paper is careful to frame this as speculative. Boechler himself draws the line clearly: this is fundamental research, and the gap between an 18-rotor chain and a deployable soft robotic muscle is measured in years and unsolved problems, not in a natural progression.

The story here is not that acoustic control of stiffness is coming to a robot near you. It is that a real mechanism now exists where only chaos existed before — and that mechanism has enough DOD backing, academic pedigree, and robotics-adjacent applications that the field will be watching to see if the scaling problems yield. For builders and investors in compliant robotics, this is a result worth tracking, not yet one worth building around.

The paper, Observation of mechanical kink control and generation via acoustic waves, was published in Nature Communications (DOI: 10.1038/s41467-026-68688-7) on February 6, 2026.