A robot smaller than a single bacterium can now chase down, capture, transport, and release live microbes in water, all steered by nothing more than the twist of a laser beam. Researchers at the Julius-Maximilians-Universität Würzburg (JMU), led by nanophotonics physicist Bert Hecht, reported the achievement in a paper published in Nature Communications in May 2026, describing what they call a “robotic cleaner” that operates without chemicals, antibiotics, or physical contact with its target.
The proof of concept is still confined to the lab bench. But at a time when antibiotic-resistant infections kill more than one million people worldwide each year, according to World Health Organization estimates, any new strategy for removing dangerous bacteria without drugs draws serious scientific attention.
How a sub-micrometer robot catches a bacterium
Each device measures just a few hundred nanometers across, roughly one-fifth the width of an E. coli cell. It is built from a precisely shaped plasmonic nanostructure, a metallic particle engineered so that incoming light scatters asymmetrically off its surface. That asymmetry generates a lateral optical force, effectively a tiny push, that propels the robot through water.
Steering comes from adjusting the polarization angle of an unfocused laser beam. Rotate the polarization and the direction of thrust changes, allowing operators to execute sharp 90-degree turns and guide the robot along programmed paths in real time. The approach differs fundamentally from optical tweezers, which use tightly focused beams and expensive high-numerical-aperture objectives. Polarization-based control spreads light over a wider area, lowering the intensity that hits any one spot and reducing the risk of heat damage to biological samples.
When the robot closes in on a bacterium, the interplay between the device’s near-field scattering pattern and the incident light creates a localized optical potential well, essentially a force pocket that corrals the microbe into a stable position beside the robot. Shifting the polarization drags that pocket through the fluid, carrying the bacterium with it. Reverse or modulate the polarization, and the bacterium floats free. In the paper’s demonstrations, the team captured E. coli cells suspended in water, shuttled them across the microscope’s field of view, assembled small groups of microbes, and released them on command.
Supplementary videos published alongside the paper show these sequences in real time, providing visual confirmation that the capture-and-release cycle is repeatable and reversible.
A decade of optical micro-machines behind the design
The 2026 cleaner sits at the end of a deliberate engineering progression. In 2022, the same Würzburg group reported light-driven “microdrones” capable of two-dimensional maneuvering under overlapping, unfocused light fields, published in Nature Nanotechnology. That work established the basic propulsion and steering concepts: a plasmonic particle, a plane-wave laser, and polarization as the control input.
An intermediate step added a plasmonic nano-tweezer to the platform, enabling the trapping and transport of single nanoparticles as small as 70 nanometers, including fluorescent nanodiamonds. That paper, also in Nature Communications, bridged the gap between steering a drone and grabbing a biological target. Earlier theoretical and experimental work on a plasmonic linear nanomotor, reported in Science Advances, had already laid the physics groundwork by showing that directional scattering under plane-wave illumination could produce usable lateral forces.
Each generation shrank the device and expanded its manipulation toolkit, moving from inert test particles toward living cells. The latest cleaner merges the microdrone’s maneuverability with the nano-tweezer’s precision, then tunes the geometry for robust interaction with micron-scale microbes. Notably, none of these robots contain onboard electronics or logic circuits. All “intelligence” lives in the external laser optics and control software, which simplifies fabrication but places heavy demands on beam calibration, stability, and real-time feedback imaging.
What the work does not yet show
Several large questions remain open, and the researchers are transparent about them.
No in vivo testing. Every demonstration took place in aqueous suspension under controlled lab conditions. Whether these robots could function inside the human body, where blood flow, immune responses, and complex fluid dynamics would interfere, is an engineering challenge the paper does not claim to have solved. Optical access alone, getting enough laser light deep into tissue to steer a nanorobot, would be a formidable hurdle.
Limited bacterial diversity. The published experiments used E. coli, a well-characterized lab workhorse. Capture efficiency across other strains, particularly drug-resistant pathogens such as MRSA or carbapenem-resistant Enterobacteriaceae, has not been tested. Rod-shaped, spherical, and filamentous species may interact very differently with the structured light fields around the robot.
Scalability is unknown. The devices rely on plasmonic nanostructures likely produced by electron-beam lithography or advanced nanoimprint methods, techniques that are precise but slow and expensive at research scale. No public statements from the team address manufacturing volume or cost targets, so any timeline for clinical or commercial use remains speculative.
Real-world fluids are messy. The tests used relatively clean suspensions. Biological fluids and environmental water samples contain proteins, debris, and other cells that scatter light and could degrade steering accuracy or trapping stability. Long-term operation and potential photothermal effects on both the robots and nearby cells also remain uncharacterized beyond the short time scales of initial demonstrations.
No independent replication. The results are internally consistent across the paper’s figures and supplementary materials, but no outside laboratory has yet reproduced the bacterial capture-and-release cycle. External validation is a standard next step for any new nanorobotics claim.
Where this fits in the nanorobotics landscape
Light-driven devices are not the only contenders in the growing field of micro- and nanorobotics. Magnetic helical swimmers, steered by rotating external magnetic fields, have been demonstrated in animal models by groups at ETH Zurich and the Max Planck Institute. Enzyme-powered nanomotors use catalytic reactions for propulsion and have shown promise in targeted drug delivery experiments. Ultrasound-driven microbots offer yet another wireless control strategy.
What distinguishes the Würzburg approach is the combination of sub-micrometer size, contact-free manipulation, and reversible capture and release, all achieved without chemical fuels or surface functionalization. Most competing platforms either require onboard chemical reactions or rely on permanent binding to their targets. The trade-off is the need for direct optical access, which limits where the robots can operate.
For now, the most plausible near-term applications may lie outside the body entirely: decontaminating water supplies, cleaning sensitive laboratory or medical equipment, or sorting bacterial populations for research. Clinical use, if it ever arrives, would demand solutions to the optical-access problem and extensive biocompatibility and safety testing that has not yet begun.
The Würzburg team’s decade-long track record, from nanomotor theory to microdrones to nano-tweezers to bacterial capture, suggests a group methodically working through the engineering obstacles rather than chasing headlines. Whether the next step is a more complex fluid environment, a broader range of pathogens, or a move toward magnetic or hybrid control to bypass the optical-access limitation, the cleaner reported in May 2026 marks the most precise wireless manipulation of live bacteria by a nanorobot published to date.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
