Ultra-extensible ribbon-like magnetic microswarm

ARTICLE DOI: 10.1038/s41467-018-05749-6 OPEN Ultra-extensible ribbon-like magnetic microswarm 1234567890():,; Jiangfan Yu1, Ben Wang1,2, Xingzhou Du1,2,3, Qianqian Wang1 & Li Zhang 1,2,3,4 Various types of structures self-organised by animals exist in nature, such as bird flocks and insect swarms, which stem from the local communications of vast numbers of limited individuals. Through the designing of algorithms and wireless communication, robotic systems can emulate some complex swarm structures in nature. However, creating a swarming robotic system at the microscale that embodies functional collective behaviours remains a challenge. Herein, we report a strategy to reconfigure paramagnetic nanoparticles into ribbon-like swarms using oscillating magnetic fields, and the mechanisms are analysed. By tuning the input fields, the microswarm can perform a reversible elongation with an extremely high aspect ratio, as well as splitting and merging. Moreover, we investigate the behaviours of the microswarm when it encounters solid boundaries, and demonstrate that under navigation, the colloidal microswarm passes through confined channel networks towards multiple targets with high access rates and high swarming pattern stability. 1 Department of Mechanical and Automation Engineering, The Chinese University of Hong Kong, Shatin, N.T., Hong Kong 999077, China. 2 Department of Biomedical Engineering, The Chinese University of Hong Kong, Shatin, N.T., Hong Kong 999077, China. 3 Chow Yuk Ho Technology Centre for Innovative Medicine, The Chinese University of Hong Kong, Shatin, N.T., Hong Kong 999077, China. 4 T-Stone Robotics Institute, the Chinese University of Hong Kong, Shatin, N.T., Hong Kong 999077, China. Correspondence and requests for materials should be addressed to L.Z. (email: ) NATURE COMMUNICATIONS | (2018)9:3260 | DOI: 10.1038/s41467-018-05749-6 | www.nature.com/naturecommunications 1 ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-05749-6 I n nature, thousands or even millions of individual elements can form a wide range of patterns, purely through local communications, such as bacteria colonies1,2, bird flocks, and insect swarms3. Through collective pattern formation, these elements can dramatically change the swarming shape according to the environment they interact with. In the field of robotics, various types of robotic systems have been reported with swarm intelligence4,5, which are inspired to emulate part of the swarm behaviours in nature. More recently, a thousand-robot swarm capable of programmable self-assembly has been reported6, addressing both the physical and algorithmic challenges of a large-scale robotic swarm. These studies rely on wireless communications to plan and distribute each robot; however, at small scales, this method is hardly accessible due to the challenges of integrating onboard processors, sensors and actuators. Hence, different strategies are required for the design and development of artificial swarms at the micro/nanoscale. Colloids are promising candidates for understanding the guiding principles of swarm behaviours in living systems, and physical or chemical interactions among them may be considered as ‘communications’7,8. These materials play an important role as building blocks for creating complex systems via static and dynamic self-assembly processes, such as periodic crystals9–13, self-assembled colloidal devices14–18, clustering19,20 and flocking21, which may help us to understand the guiding principles of swarm behaviours in living systems. Nevertheless, emulating the swarm behaviours in nature is still challenging, because the relevant fundamental mechanisms, agent–agent interactions and proper actuation strategies are still under investigation. Moreover, realising collective morphological transformations that are similar to some living systems may require appropriate actuation methods and programmable interactions among the agents22–24. In this paper, we trigger the formation of a microswarm on a 2D plane, i.e., a reconfigurable ribbon-like paramagnetic nanoparticle swarm (RPNS) with a dynamic-equilibrium structure, by applying programmed oscillating magnetic fields. We investigate the generation mechanism and demonstrate the reversible elongation with an ultrahigh aspect ratio of the microswarm. Other reversible reconfigurations, including controlled splitting behaviours and the merging of two subswarms are presented. The microswarm can perform 2-D locomotion fully under control near a solid surface, and can maintain a stable pattern even in complex environments with varied boundary conditions. Finally, we demonstrate that the microswarm can pass through channel networks towards multiple targets with high access rates and perform non-contact micromanipulation in a fluid. Results Generation of a ribbon-like paramagnetic nanoparticle swarm. The oscillating magnetic field B for the actuation is schematically demonstrated (Fig. 1a). In one direction, an alternating magnetic field BAC is applied, with the condition of BAC = A sin(2πft), where A is the amplitude of the magnetic field as a constant, and f is the input oscillating frequency. The uniform magnetic field BC is applied in the perpendicular direction with a constant field strength of C. An amplitude ratio (γ = A/C) is proposed. The superposed magnetic field (Fig. 1a, red arrow) has a timedependent angular velocity and field strength. At Point O, the magnitude of the angular velocity is maximal, and the magnitude of the field strength is minimal (Supplementary Fig. 1a). When the amplitude ratio γ is increased, as shown in Fig. 1b, the oscillating angle becomes larger, and if the oscillating frequency is maintained, the angular velocity is also increased (ω2(t) > ω1(t)). Meanwhile, because the magnitude of BAC is fixed, C2 becomes smaller than C1. In magnetic fields, paramagnetic nanoparticles 2 form chain-like structures; therefore, we regard the individual nanoparticle chains as the building blocks in this work. Supplementary Fig. 2 illustrates the forces and torques exerted on a particle chain when it oscillates with the input field. The lengths of the particle chains are related to the strength of the applied field25. When the superposed field points to a or b, the magnetic field strength reaches the highest value, which enhances the magnetic attractive interactions between short nanoparticle chains, and longer chains are formed (Fig. 1c). The magnetic field strength is the weakest when the field points to O, and at this moment, the particle chains are much shorter (Fig. 1d). Figure 1e shows the change in the time-dependent chain lengths, when the oscillation frequency is 1 Hz. The blue and red curves indicate the mathematical model (the model is presented in Supplementary Eq. 3), and the experimental data, respectively, which demonstrates good agreement. Because in the model, a single particle chain is assumed to be formed, while in the experiments, chainlike bundles are (...truncated)