Leipzig and Prague physicists switch a microswimmer's propulsion mode with patterned laser light

The laser can heat the sphere's surface but cannot cool it, and the researchers describe that limit as crucial ^[9]. Only patterns that add heat somewhere on the surface are available ^[9]. To keep such a pattern registered on a particle that is moving, the group projects tailored intensity patterns onto the surface and adjusts them in real time to the particle's motion ^[10].

"Our work helps us understand what an 'optimal' swimmer actually means in reality. We show that the physical constraints of the swimming mechanism can prevent a swimmer from achieving the theoretical optimum," said Lisa Rohde, the study's first author ^[12].

Speed and efficiency came apart. "The counterintuitive finding is that, at the same laser power, the slowest swimming style is actually the most efficient when it comes to achieving a particular speed," Rohde said ^[13]. The comparison holds laser power fixed ^[13], so it accounts for light arriving at the particle and not for the projection optics and tracking loop that keep a pattern on a moving micrometer-scale sphere ^[10]. The announcement does not report swimming speeds, laser powers or how many particles were run ^[19].

Frank Cichos, who leads Leipzig's Molecular Nanophotonics Group, said: "Until now, a microswimmer's hydrodynamic character was fixed at the point of production, much like an animal's body structure. We can now use light to change it in real time and, for example, adapt it to the environment if needed" ^[14]. According to the group, experiments on the collective behaviour of active matter, including bacterial turbulence and swarming, have been virtually impossible with conventional swimmers whose properties are fixed ^[16].

The object being steered is a heated sphere. Rohde said the same constraints exist in nature, and that this may explain why evolution has produced bacterial swimming strategies that are not optimal on paper yet are extremely successful and widespread ^[15]. Whether that holds for E. coli, which pushes itself along with flagella at its rear ^[7], would take a measurement on bacteria.

Synthetic particles of this kind are used as physical models for the interactions between biological microorganisms ^[6], and the researchers say the work opens an evolutionary approach to developing synthetic active matter ^[18]. Rohde said the group next wants to combine the programmable control with machine learning, so that a microswimmer can learn which style suits a given task, and to compare swimming in viscoelastic mucus with swimming in water ^[17].