TURN UP THE SOUND, AND THE STAR SPINS BACKWARDS

Sound can lift things. Between a vibrating horn and a reflector, a strong ultrasonic standing wave creates places where small objects — droplets, thin films, grains — float in mid-air without touching anything. The technique is being developed for chemical and pharmaceutical analysis and for assembling tiny structures.

It has a stubborn quirk: levitated objects often start spinning by themselves. Explanations exist for particular cases, but no general rule told when an object would turn, in which direction, or how fast.

Stars with curved arms

Sue Shi, Scott Waitukaitis and colleagues at the Institute of Science and Technology Austria (ISTA) built a levitator working at about 51 kilohertz, with a sapphire reflector on a precision gimbal so the sound field could be aligned very finely. They 3D-printed fluorescent particles 200 micrometres thick, each fitting in a 1-millimetre circle, and filmed them from below at 3,000 frames per second.

The first result is about symmetry. In a well-aligned field, triangles, squares, pentagons and stars with straight arms — shapes that look the same in a mirror — never spin. “Chiral stars”, whose arms all curve the same way, start spinning as soon as they float, reaching a steady speed within a few hundred milliseconds. The curve sets the direction; flip the star over and it turns the other way. Tilting the reflector on purpose makes even a square spin: spontaneous rotation needs broken symmetry, in the object or in the sound field.

More arms, stranger behaviour

Intuition says more arms catch more sound, and louder sound pushes harder. The measurements, on three nearly identical copies of each star, say otherwise.

3 and 5 arms: they spin faster and faster as the voltage on the transducer rises — and with it the sound pressure.
7 arms: fastest near 12–13 volts, then they slow down.
10 arms: fastest near 7–8 volts, then they slow, stop, and spin the other way.
12 arms: the same sequence, earlier — peak at 7 volts, reversal near 10 volts.

The fastest stars reach up to about 50 turns per second.

A two-term recipe

To explain this, the team balanced a driving torque — the rotational push exerted by the sound — against air friction. Both are written as simple powers of the sound amplitude: the twist has a main term plus a correction that grows faster, and the friction rises with the amplitude. The friction is not guessed: it is measured by watching how fast each star settles after a sudden step in voltage: the sound field reaches its new value within about ten milliseconds, while the star takes about 183 milliseconds to regain its speed (12-armed star going from 7 to 8 volts), a lag that comes from air drag alone — and calculated at zero amplitude from the star’s shape. That leaves just two free numbers per shape, and the resulting curve fits every star, from steady speeding-up to full reversal.

The fits tell a consistent story. More arms give a stronger push at low volume, so densely armed stars start faster. But they also bring a growing opposite contribution that wins at high volume — and when it wins, the star reverses. The reversal voltage follows a simple formula, V* = 1/√(−α), where α is negative as soon as the correction opposes the push: the larger it is in absolute value, the earlier the reversal comes.

A star designed to switch

The real test was a prediction. The team designed a hybrid star: one half taken from a 4-armed star curving clockwise, the other from a 10-armed star curving the opposite way. Adding the two halves’ predicted curves gave a forecast: counterclockwise at low volume, where the 10-armed half dominates; clockwise at high volume. The printed hybrid did exactly that — counterclockwise at 7 volts, motionless at 9 volts, clockwise at 11 volts — reversing at the predicted voltage.

A force nobody ordered

The model describes the opposing push but does not say where it comes from. The authors show it cannot come from the usual theory of acoustic forces, in which every effect grows with the square of the sound pressure: all those contributions would then keep the same proportions as the sound gets louder, and the direction of spin could never flip. Something of higher order is at work, and it strengthens as neighbouring arms get closer — perhaps a coupling between adjacent arms through the sound and the air. Measuring or simulating the flow around a spinning star is the next step. In the meantime, the shape of a particle has become a way to program how it turns.