The Physics of Floating on Sound
In 2013, researchers at the University of Tokyo published footage that stopped the scientific community cold: small objects — water droplets, polystyrene beads, even living ants — hovering motionless in mid-air, held aloft by nothing more than precisely tuned sound waves. No magnets, no electric fields, no visible support structures. The phenomenon, known as acoustic levitation, had been understood theoretically since the 1930s, but the Tokyo demonstration marked a turning point in the practical manipulation of matter using sound. What had once existed primarily as an elegant theoretical curiosity was suddenly, visibly real — and the implications stretched far beyond the laboratory bench where the footage was recorded.
Sound, in the everyday sense, is easy to underestimate. We associate it with music, speech, and noise — phenomena of sensation rather than physics. But sound is, at its core, a pressure wave propagating through a medium, and pressure is force per unit area. When sound waves are intense enough and structured carefully enough, that force becomes substantial. The Tokyo footage did not represent a new discovery so much as a new demonstration of an old principle finally made accessible. The decades between the theoretical groundwork and the practical demonstration were spent solving engineering problems: how to generate stable standing waves, how to tune transducers with sufficient precision, and how to prevent the acoustic field itself from disturbing the very objects it was meant to hold. The solutions, when they arrived, opened a door that researchers have been walking through ever since.
How Acoustic Radiation Pressure Works
The underlying physics involves acoustic radiation pressure. When a standing wave is created between two opposing ultrasonic transducers, regions of high and low pressure alternate at fixed points in space. Objects placed at the nodes — the points of minimum pressure — experience a net inward force from all sides, effectively trapping them. The key is that the sound frequency must be tuned so that the wavelength is roughly comparable to the size of the levitated object. Most laboratory demonstrations use ultrasound in the range of 20 to 40 kilohertz, well above the threshold of human hearing, and can levitate objects up to a few millimeters in diameter.
The mathematics behind this trapping force draws on a branch of fluid mechanics that most physics students never encounter. When a compressible object sits inside an acoustic field, the pressure variations around it are not symmetric — the object scatters sound differently depending on its shape, density, and compressibility relative to the surrounding medium. This asymmetry produces a net time-averaged force, known as acoustic radiation pressure. For small, soft objects in a fluid medium like air, this force is directed toward the pressure nodes. For denser or stiffer objects, the force can be directed toward the pressure antinodes instead. The distinction matters enormously in practice, because it determines where in the sound field a given object will naturally come to rest and how stable that resting position will be under perturbation.
What makes this more than a parlor trick is that the trapped objects can also be moved. By modulating the phase relationships between multiple transducers arranged in three-dimensional arrays, researchers can steer levitated particles along complex trajectories, rotate them, merge droplets, or hold them completely stationary while performing operations on them. This is not remote manipulation in any loose metaphorical sense — it is precise, programmable, contactless control of matter. The level of control achievable is comparable in some respects to optical tweezers, the laser-based trapping technology that earned its inventors a Nobel Prize in 2018, but acoustic levitation works on objects many orders of magnitude larger and does not require the sample to be transparent or the surrounding medium to be optically clear.
From Laboratory Curiosity to Pharmaceutical Tool
The pharmaceutical industry has taken particular notice of acoustic levitation for a reason that is not immediately obvious. Many drug compounds are notoriously difficult to study in their pure crystalline state because the moment they contact any surface — a glass slide, a container wall, even the ambient humidity of a laboratory environment — they begin to change. Some compounds absorb moisture and undergo structural changes. Others nucleate crystals differently depending on what surface they touch. The result is that the data scientists collect about a drug’s solid-state properties may be contaminated by the very instruments used to collect them.
Acoustic levitation solves this by eliminating the need for a surface entirely. A droplet of dissolved pharmaceutical compound can be suspended in mid-air and slowly evaporated under controlled conditions, allowing scientists to observe crystallization in its most pristine form. The European Synchrotron Radiation Facility in Grenoble, France, has combined acoustic levitation with high-energy X-ray beams to perform real-time structural analysis of levitated droplets as they crystallize — a technique that would be impossible if the sample were sitting on a substrate. The X-ray beam passes through the levitated droplet without interference, and the diffraction patterns it produces reveal the evolving crystal structure in a way that no surface-contact method could replicate.
This matters because the crystalline form of a drug directly determines its bioavailability — how readily the body absorbs it. Two samples of the same molecule, arranged in different crystal structures, can behave like entirely different drugs in the body. The pharmaceutical industry has lost significant investment to late-stage failures caused by poorly understood polymorphism, the tendency of some compounds to crystallize into multiple distinct forms. Acoustic levitation offers a path to studying these transitions without the confounding influence of container surfaces, potentially catching problematic polymorphic behavior far earlier in the development process.
Beyond crystallography, acoustic levitation is being explored for the contactless mixing of reagents that would react violently if combined in conventional containers, and for the handling of biological samples so fragile that even the gentlest physical contact would destroy them. In 2022, a team at the University of Michigan demonstrated acoustic manipulation of individual human cells without causing measurable damage to the cell membrane — a capability with obvious implications for single-cell analysis and precision medicine. The ability to position, rotate, and transport a single living cell without touching it represents a qualitative leap in what is experimentally possible at the boundary between physics and biology.
Manufacturing in Microgravity and Beyond
Perhaps the most forward-looking application of acoustic levitation involves space. In the microgravity environment of an orbiting spacecraft or space station, conventional manufacturing processes face fundamental problems. Liquids do not settle, convection currents do not form in predictable ways, and the absence of gravity means that many of the assumptions underlying Earth-based fabrication simply do not apply. Acoustic levitation, counterintuitively, works better in microgravity than on Earth, because the acoustic forces do not have to overcome the weight of the object being levitated — they only need to position and control it. On Earth, a significant portion of the acoustic energy budget goes toward simply counteracting gravity. In orbit, that energy is freed for more precise manipulation.
NASA has funded research into acoustic manipulation as a potential tool for in-space manufacturing, particularly for the assembly of structures that are too large or too delicate to build on Earth and launch intact. The idea of using sound to assemble components in orbit, free from both gravity and physical contact, is no longer considered science fiction by the agency’s advanced manufacturing teams. A 2021 study published in npj Microgravity demonstrated that acoustic trapping could be used to position and bond small components with sub-millimeter precision in simulated microgravity conditions, suggesting that the technique is mature enough to warrant serious engineering investment rather than purely exploratory research.
There is also a materials science angle that deserves attention. Certain exotic glasses and metallic alloys can only be produced in a perfectly containerless environment, because contact with any mold or crucible introduces impurities or triggers premature crystallization. On Earth, electromagnetic levitation can achieve this for conductive materials, but it does not work for non-conductive substances. Acoustic levitation fills this gap, enabling the production of ultrapure glass microspheres and novel amorphous alloys that cannot be made by any other method. Some of these materials have optical or mechanical properties that make them candidates for next-generation sensor technologies and fiber-optic components, meaning that containerless processing enabled by acoustic levitation may eventually find its way into everyday communications infrastructure.
The Holographic Frontier
The cutting edge of acoustic levitation research has moved toward what researchers are calling acoustic holograms — three-dimensional sound fields of arbitrary complexity that can trap and manipulate large numbers of objects simultaneously. Rather than using two opposing transducers to create a simple standing wave, acoustic hologram systems use flat arrays of hundreds of individually addressable ultrasonic emitters, each one controlled by a computer to produce a specific phase and amplitude. The array's combined output creates a sound field that can be sculpted into virtually any shape, with multiple trapping points distributed throughout a volume of space.
A landmark paper published in Nature in 2019 by researchers at the University of Bristol demonstrated a single-sided acoustic hologram array — eliminating the need for opposing transducers entirely — that could levitate and move objects along preprogrammed paths in real time. The system could also generate tactile sensations in the air above the array, creating the perception of touching objects that were not physically there. This convergence of levitation and haptic feedback has attracted interest from the entertainment and virtual reality industries, where the ability to make users feel physical sensations without wearing gloves or touching screens remains a significant challenge. The prospect of a display surface that can simultaneously project an image and deliver a physical sensation to the hand reaching toward it is not as distant as it might seem.
The Bristol group has since extended their work to demonstrate acoustic levitation of objects up to 2 centimeters in diameter — large enough to include small electronic components, biological tissue samples, and even living insects. The insects, notably, survived the experience unharmed. Parallel research groups in Japan, Germany, and the United States have been developing algorithms that allow acoustic hologram arrays to reconfigure their trapping patterns in milliseconds, opening the possibility of acoustic assembly lines in which components are passed between trapping points without ever making physical contact with a surface or a robotic gripper. The engineering challenges that remain are real but tractable — primarily questions of power efficiency, array miniaturization, and the computational speed required to update complex field configurations in real time.
As the technology matures and transducer arrays become cheaper and more compact, the prospect of acoustic manipulation becoming a routine laboratory and industrial tool moves steadily closer to reality. The history of science is full of forces that were understood long before they were harnessed — electromagnetism spent decades as a laboratory phenomenon before it became the foundation of modern industry. Acoustic radiation pressure may follow the same arc. Sound, it turns out, has always been a physical force. We are only now learning to wield it with precision.