Magnetic motors in interphases: Motion control and integration in soft robots

Motion is ubiquitous in nature and has allowed species to adapt and survive. In the pursue of mimicking natural movement, active matter (also known as out-of-equilibrium systems) emerged as comprising a large group of materials able to respond to certain stimuli and move.¹ A particular type of active matter is micro- and nanomotors (also referred to as micro- and nanoswimmers or micro- and nanobots), which are made of nanoparticles able to convert an input energy into kinetic energy, thereby exhibiting motion at the nanoscale. The key parameter for these motors to accomplish effective motion (i.e., net displacement) is the capacity to outperform the thermal fluctuations originating from Brownian motion, which can be estimated from the Stokes–Einstein relationship.² From a practical point of view, two requirements are needed: (1) the motion must be continuous over time and (2) the mechanism to produce this motion has to be nonsymmetrical in time. These requirements aimed to tackle the so-called Scallop theorem, which states that motion at low Reynold numbers has to be nonreciprocal and represents the cornerstone of motion at the nanoscale.³ Options to break the “Scallop theorem” consider asymmetric or flexible structures, nonreciprocal actuations, and motion close to boundaries or non-Newtonian fluids.

Different approaches have been developed in the last few years to address this issue, as thoroughly reviewed elsewhere.^4–9 Overall, motors can be classified into two groups depending on the origin of the input energy, namely, fuel-free or fuel-dependent motors. In the case of fuel-free motors, external physical stimuli (e.g., light, electric, and magnetic fields and ultrasounds) are often employed to propel the structures. Conversely, fuel-based motors rely on (bio)chemical reactions that transform a substrate into a product, initiating the locomotion.

Magnetically actuated motors have distinct advantages compared to other mechanisms of motion. First, they can be remotely controlled using external magnetic fields,^10–13 i.e., the actuation can be instantaneously turned on and off. Additionally, magnetic fields can be precisely tuned by varying, e.g., the field intensity, the distance to the sample, or the composition of the magnetic material, which the motor is made of. Second, magnetic fields cannot only trigger the motor motion but also control the directionality and interactions between neighboring motors.

In this Perspective, the most recent advances in the past two years of magnetic micro/nanomotors will be examined, with special emphasis on motion control (Fig. 1). First, the basic magnetic field strategies and their relation to the motor structure will be discussed. Second, spatial control of the motors resulting in steering, guidance, and cooperative movement will be introduced. Further, examples of the incorporation of magnetic motors in soft robots as a means to induce mobility will be reviewed. Finally, an outlook and future perspective of this class of motors will be outlined, with special attention to future endeavors in the field.

Magnetic motors are made of entirely or partially incorporated magnetic materials that respond to applied magnetic fields. These materials can be classified according to their response to a magnetic field, namely, their magnetic susceptibility (χ). Paramagnetic (χ > 0) and diamagnetic (χ < 0) materials can be attracted to or repelled from a magnetic field, respectively, and they present short-range, weak interactions that vanish when the field is removed. On the contrary, ferromagnetic and ferrimagnetic (χ ≫ 0) materials respond strongly to a magnetic field and exhibit long-range, ordered interactions, which conserve the magnetization of the material even in the absence of a magnetic field. Superparamagnetic materials are a special class of magnetic materials that respond rapidly and strongly to magnetic fields but cannot preserve the interactions when the magnet is removed, making them fast actuators with a quick on/off response. Most magnetic motors are based on ferro-, ferri-, and superparamagnetic materials. (Note: another class of magnetic materials is antiferromagnets, which theoretically display zero response to magnetic fields and are therefore not relevant for magnetic mobility).

If the magnetic motor is subjected to a magnetic field gradient $∇ B$⁠, the motor will experience an attractive or repulsive (depending on the magnetic material type) force F_m to the magnetic field. In the case of motors subjected to uniform magnetic fields, they will experience a torque T in order to minimize the energy, creating a rotation of the motor and eventual motion. I would like to note that Eqs. (1) and (2) assume homogeneous materials where the total magnetic moment (M) is considered constant along the material. Further, the aforementioned magnetic susceptibility is implicit in M and the interaction with B. For a detailed description of magnetism in nanomaterials, the reader is referred to Ref. 14. Therefore, choosing the right magnetic materials as well as the shape and size of the core particle and deciding the kind of actuating magnetic field are key parameters to consider when designing magnetic micro/nanomotors. For instance, ferromagnetic/antiferromagnetic cubic particles,¹⁵ superparamagnetic spindles,¹⁶ or superparamagnetic/diamagnetic core/shell structures^17–19 have been utilized for magnetic mobility with different outcomes, stressing the importance of material choice and particle shape and size to accomplish distinct mobility.

In order to achieve effective locomotion with magnetic motors, the symmetry of the system (note: the system refers to the combination of motor and environment) has to be broken either by using intrinsically asymmetric motor designs or by employing nonuniform (inhomogeneous) magnetic fields. Whichever the approach may be, they cannot be seen as independent strategies, as the type of magnetic field available will determine the characteristics of the motor. Magnetic fields can generally be classified according to whether they are nonuniform in space, such as field gradients, or nonuniform in time, such as oscillating and rotating fields.

Later examples of magnetic micro- and nanomotors moving by magnetic field gradients included motion between liquid-liquid interphases. For instance, Kichatov and co-workers have explored the motion of magnetite (Fe₃O₄) motors at the water-oil interphase when varying the concentration and type of surfactant in the liquid phase.²⁰ Their results indicated that first, the solid radius of the motors had to be greater than a critical radius for the motors to overcome a potential energy barrier due to capillarity forces and cross the interphase [Fig. 2(aiiI)]. Second, the Marangoni flow at the interphase played an important role in the mechanism of crossing. For instance, surfactants dissolved in the aqueous phase at concentrations higher than the critical micellar concentration generated stronger Marangoni flows that aided the motors crossing [Fig. 2(aiiII)]. Similar observations were made when nonionic surfactants were used, instead of the ionic counterparts. In another effort, we considered motors of varying size and surface coating (positively charged or PEGylated) made of polystyrene beads coated with magnetite/maghemite (Fe₃O₄/γ-Fe₂O₃) to cross lipid membranes of different composition.¹⁹ The results indicated that the motors could cross the lipid membranes independent of the motors’ size (0.5, 1, or 4 μm in diameter). However, the number of motors crossing was larger when more rigid lipid membranes or lower magnetic fields were employed. Additionally, PEGylated motors in the same conditions had limited crossing abilities, suggesting that there was an interphase potential barrier that restricted the crossing based on motor-membrane interactions.

Field gradients have also been used to direct the motion of enzyme-propelled plasmonic-magnetic motors for real-time detection of substances.²⁴ These complex motors were made of porous silica capsules containing gold nanoparticles for plasmonic heating and nickel for magnetic steering, as well as a coating of poly(N-isopropylacrylamide) brushes and the enzyme urease for motion. Illuminating the motor with near-infrared light caused localized heating due to the gold particles and changed the conformation of the polymer, allowing for controlled opening and closing of a gate accessible to the substance of interest. The motors were loaded on a chip for microsampling using Raman spectroscopy, where the magnetic actuation would direct the motors to the target area more precisely. Similarly, magnetic field gradients have been employed to direct the motion of other stimulus-triggered motors with magnetic functionality.^25–27 

All in all, magnetic field gradients pose great advantages owing to their easy implementation and motor fabrication and represent a simple model to understand motion without intricate setups.

Alternatively, oscillating and rotating magnetic fields represent a more complex setup to control particle motion, and their interaction with magnetic materials depends on Eq. (2). Oscillating magnetic fields usually employ combinations of electromagnets specifically arranged that generate sinusoidal or square (on/off) waves of constant amplitude, resulting in a varying field over time that depends on the frequency of the alternating current.²⁸ Rotating magnetic fields are often based on a permanent magnet coupled to a rotor that turns at a given angular speed, also producing an inhomogeneous field in time [Fig. 2(biI–III)]. These setups allow for higher control of the magnetic field parameters (e.g., intensity and frequency) by fine-tuning the speed of the rotor or the frequency of the current. Limitations of these kinds of magnets include the large setup and size of the workspace and the specialized design of the motors, which usually relies on physical fabrication techniques to produce them [Fig. 2(biIV)].

Detailed relationships between the used magnetic field and the geometrical features of helical motors have been described theoretically and demonstrated to match experimental data. For instance, it has been well established that the forward velocity of the motor (v) is directly related to its angular velocity (ω), which can be synchronized (ω = ω_r) or not (ω ≠ ω_r) to the rotational velocity of the applied magnetic field (ω_r).^31,32 Therefore, the synchronous regime happens when the motor can rotate at the same speed as the rotating field. At higher angular frequencies, the helical motors cannot rotate as quick as the magnetic field, and hence the forward velocity of the motor decreases. The critical frequency at which this effect occurs is called the step-out frequency (ω_step-out) and depends on the relationship between the radius, the pitch, and the angle of the helix. For example, increasing the pitch distance or the helix radius results in increasing ω_step-out, while increasing the pitch angle or the helix thickness leads to a nonmonotonic trend.³² Interestingly, the number of turns in the helix does not affect the velocity.

An advanced example of a magnetic micromotor with adaptable locomotion capabilities consisted of the combination of a strong magnetic head and a flexible tail.³³ The magnetic head was fabricated by photolithography using a mixture of neodymium (Nd) microparticles and an adhesive material. The flexible tail was 3D printed via two-photon polymerization using poly(N-isopropylacrylamide) as the responsive material combined with a passive polymer. The number of turns in the helical tail was modulated by varying the angle between the responsive and nonresponsive polymers during printing. After printing, the micromotor’s tail deformed when exposed to de-ionized water or isopropyl alcohol, by adapting the number of turns, which was always larger in de-ionized water, for the same printing conditions. These motors exhibited top velocities of ∼2.3 mm s⁻¹ in isopropyl alcohol and ∼1 mm s⁻¹ in water for the highest frequencies employed before the step-out. Surprisingly, the velocity was higher in isopropyl alcohol despite having a viscosity 2.4× higher than water.

Rotating magnetic fields have also been employed to propel biohybrid motors that showed a rolling-tumbling behavior.³⁴ At frequencies below the step-out, these motors rolled and tumbled along the short and long axes, respectively. Alternatively, helical carbon nanotubes decorated with magnetic nanoparticles also presented a dual mechanism of motion [Fig. 2(bii)].²¹ They oriented when static magnetic fields were employed or moved in straight and curved trajectories under magnetic field gradients or rotating magnetic fields, respectively. Other magnetic helical motors have also been demonstrated to move at high speeds while transporting massive cargo.³⁵ 

Altogether, oscillating and rotating magnetic fields offer a wider range of control but in return they require more complex fabrication techniques and specialized equipment.

The magnetic functionality of the motors can allow not only for propulsion but also for controlling the interactions between neighboring motors and the direction of the motion. To this end, the motors are required to possess great colloidal stability and to exhibit reduced dipolar interactions that force motors’ uncontrolled aggregation. Further, superparamagnetic materials that respond quickly to changes in the magnetic field (e.g., intensity and orientation) are preferred when envisioning applications in, e.g., biomedicine or environmental remediation.

Fine control of the motors’ motion is of high interest when it comes to specialized tasks such as microsurgery or pick-and-place and assembly of objects, similar to what macroscale robots do. For this reason, not only gaining propulsion but achieving control on the motors’ directionality, stop-and-go motion, and their ability to carry other particles has been attracting attention lately.^13,36 The challenges rely on overcoming the thermal fluctuations and viscous dragging in the medium to precisely control the directionality of the motor.

A basic requirement for achieving particle’s motion control is being able to tune the motors’ rotation. As an example, Honecker et al. employed hematite spindle-shaped particles to explore the possible rotation modes under the influence of a custom-made dynamic (alternating and rotating) magnetic field.¹⁶ Hematite spindles showed a pseudo-superparamagnetic (i.e., weakly ferromagnetic) behavior and in the absence of the magnetic field they displayed Brownian motion that changed into precession along the longitudinal direction (easy magnetic axis, i.e., lower magnetocrystalline anisotropy) and rotation perpendicular to the applied field when an alternating and a rotating magnetic field was used, respectively. For high frequencies, the spindles could not follow the angular velocity of the field and the motion ceased, which agreed with the step-out behavior observed in propulsion under varying magnetic fields.

Another basic feature in motor control is guidance, which has recently been demonstrated in an example where Janus (asymmetric) lipid/polymer-based motors with magnetic nanoparticles decorating their surface aided reorientation and guidance toward a magnet.³⁷ Similarly, the Pumera group has shown that the magnetic functionality helped in maneuvering motors in diverse applications, such as pollutant³⁸ and drug³⁹ removal, microplastics degradation,⁴⁰ pest control in plants,⁴¹ or as contrast agents for in vivo imaging.⁴² As an example, they employed a magnetic hybrid micromotor made of hematite/maghemite decorated with β-cyclodextrin to remove methamphetamine from aquatic environments.³⁹ This motor showed excellent maneuvering capabilities under rotating magnetic fields, which propelled the motor forward in linear trajectories under a range of frequencies and also in predefined perpendicular and random trajectories [Fig. 3(a)], thus covering larger areas during motion, which impacted the ability to remove the drug from the medium. Another recent report has described the selective transport of microrobotic carriers along a micromagnetic pathway that could move in the same direction, exhibit selective rotation, or undergo bidirectional movement by controlling the magnetic field.⁴⁵ This strategy was suggested to overcome the limitations of motors moving in solution since the microfabricated pathways could help to reach complex areas and narrow channels in vivo. An elegant effort where steering depended on the motor design and not an external device was proposed by Yamanaka and Arai, who fabricated an eletroosmotic motor with a magnetic steer.⁴⁶ This fascinating design involved a bioanode (containing silver nanoparticles and the enzyme glucose oxidase) and a biocathode (made of silver nanoparticles and the enzyme laccase) separated by an insulator layer. Additionally, a magnetite (Fe₃O₄) rod was added to the structure to control the directionality. The bioredox reaction between the anode and the cathode propelled the motor, achieving top velocities of ∼118 μm s⁻¹. The steering control was first mathematically described in a 2D environment considering no perturbations. The proof-of-concept experiment showed high fidelity with the mathematical model, where the motor was magnetically guided in circular trajectories by a Helmholtz coil system.

A step forward toward controlled motion involves trapping of particles. Similar to optical tweezers with plasmonic particles, magnetic particles can be trapped by making use of parametric excitation, which relies on the use of an inhomogeneous and time-varying magnetic field created by a microring mounted at the end of a cantilever.⁴⁷ This system had the potential to act as a scanning probe microscope with the trapped magnetic particle as the “tip” and offered the possibility to 3D manipulate the position of the magnetic object. Alternatively, beyond mere trapping, the capability to trap and reposition particles significantly enhanced the motor's potential for precise placement in desired locations. For example, three degrees-of-freedom robotic electromagnetic needles were used to independently manipulate the position of four different magnetic microparticles.⁴⁸ The constant voltage supplied to the coils connected to the needles generated a constant magnetic field that pulled the particle toward the high field created in between the needles, allowing its relocation.

Further, regulating the movement of the magnetic motor can facilitate the control of motion or positioning of other particles or objects. To illustrate this concept, Basualdo et al. have employed magnetic microdisks that rotate under the influence of a rotating magnetic field to control and transport passive particles.⁴⁹ The disks were added to a microfluidic channel and trapped at specific positions of a channel using magnetic traps. Below the channel, a rotating magnetic field with tunable angular frequency allowed for spinning but also relocation of the disks in the traps. At low angular frequencies, the disks spun within the traps, whereas at higher frequencies, they moved throughout the channel in between the traps. This change in spinning and position of the disks created flows in the channel that controlled the motion and transport of polystyrene passive particles, which could spin or travel around the disks. Similar electromagnetic coil arrays to control particle motion have been reported on a millimeter scale, showing optimal control of individual motors in assembles with precision.⁵⁰ An advanced system involved an optothermal-magnetic microgripper that could catch and transport objects.⁵¹ This robot was fabricated by two-photon polymerization using IP-S resin and consisted of a body susceptible to magnetic fields that allowed its motion and orientation, and a gripper that was optothermally actuated by a laser. The robot motion was controlled by a pair of electromagnetic arranged orthogonally with three degrees-of-freedom (i.e., 2D planar translation and in-plane rotation) and moved with a mean velocity of 0.1 mm s⁻¹. This robot could catch, transport, and deliver a microbead in water, whereas their performance in higher viscosity media such as glycerin was not successful due to the viscous drag that limited the movement of the gripper.

Controlled regrouping (also known as swarming or flocking) has been observed in animals⁵² and eventually transferred to nano/microscale systems.^53–55 Swarming of motors can enormously increase their potential toward envisioned applications, such as overcoming hurdles or enhanced mobility. From a fundamental perspective, Liao et al. demonstrated that swarms of magnetic motors oriented and moved under a varying magnetic field until reaching the step-out frequency, as individual motors do.⁵⁶ Additionally, the dynamics of magnetic swarms in narrow microfluidic channels were explored by Kichatov and co-workers.⁵⁷ They reported that the motion of magnetite particles in a swarm was not homogeneous, but they formed branched (fractal) patterns where the magnetic field intensity was stronger at the tip of the branches compared to the grooves. Further, the formation of these branched patterns directly depended on the channel height and the viscosity of the medium, being more pronounced at higher viscosities and wider channels. Further, swarm control was achieved between particles navigating a bifurcated microfluidic channel.⁵⁸ In this case, the motors moved together in one channel that split into two vessels. A rotating oscillating magnetic field was used to promote the disassembly of the motors, being lower frequencies needed the larger the motor size. The two groups of motors navigated the vessels independently and regrouped after the bifurcation by applying a uniform oscillating magnetic field. Additionally, these motors moved forward with relative stability in the channels under a flow of 1 mm s⁻¹, simulating the vascular environment. An advanced example of collective motion against shear flows emulating blood circulation consisted of magnetite nanoparticles that were able to form vortex-like superparamagnetic nanoparticle swarms (VPNS), mimicking wild herring schools and the leukocytes rolling [Fig. 3(b)].⁴³ In the absence of a magnetic field, the individual motors displayed Brownian motion, which turned into chainlike superstructures under a low frequency rotating magnetic field. These chains broke into smaller chains at higher field frequencies, which eventually reshaped into VPNS swarms at very high frequencies [Fig. 3(bi)]. The chainlike swarms achieved top speeds of ∼100 μm s⁻¹, moving against a shear flow of 1 μl min⁻¹ (i.e., ∼400 μm s⁻¹). Exceptionally, these swarms of motors were able to navigate in physiological-like conditions where the microfluidic channels were flooded with diluted mouse blood serum at a flow speed of 150 μm s⁻¹. In these conditions, the swarms exhibited velocities of ∼140 and ∼100 μm s⁻¹ when they moved downstream and upstream, respectively [Fig. 3(bii)].

The advantages of swarms over individual motors have been highlighted in diverse applications, from biomedicine^44,59–63 to environmental remediation.^64,65 For instance, swarms of metal-organic framework-based magnetic motors were reported for pH-responsive targeted drug delivery.⁶³ Further, magnetic motor swarms have been used to induce cell death in cancer cells making use of the magneto-mechanical effect.⁶⁰ This effect relied on the mechanical shear stress induced by the motors on the cells under a rotating magnetic field, initially resulting in lysosomal escape and ultimately leading to cell death. Alternatively, magnetic swarms have been explored as temperature reporters for hyperthermia treatment [Fig. 3(c)].⁴⁴ In this example, magnetic swarms made of magnetite (Fe₃O₄) particle chains connected through a poly(N-isopropylacrylamide) coating were employed to control the motion of the motors and responded to temperature changes by changing color (photothermalchromism). The swarms were illuminated with near-infrared light to trigger the transition of the polymer from a swollen state to collapse, resulting in a decrease in the interparticle distance within the chains and, therefore, a measurable change in temperature [Fig. 3(ci)]. This change in temperature was monitored as a color change in a dark field microscope, where the images varied from red (swollen polymer state, i.e., larger interparticle distance) to green (collapse polymer state, i.e., smaller interparticle distance). These swarms were guided in microfluidic channels toward a tumorlike model, which was treated by near-infrared light-induced hyperthermia, with simultaneous reporting of the in situ temperature [Fig. 3(cii)]. In short, this strategy allowed self-reporting of photothermal hyperthermia with high temporal and spatial resolution.

From a different perspective, swarms of magnetic droplets have been used for decontamination processes.⁶⁴ Alternatively, Song and co-workers exploited the use of biological magnetic motors for water decontamination employing the Magnetospirillum magneticum strain AMB-1 (M. magneticum AMB-1).⁶⁵ In the absence of a magnetic field, the bacteria showed autonomous locomotion, which could be maneuvered using a custom-made rotating magnetic field owing to the innate magnetism of the bacteria. Additionally, these bacteria had the ability to inherently bind to organic matter, which made it an ideal candidate for removing pollutants from contaminated water. Specifically, ∼70% of the model pollutant chlorpyrifos was removed in the first 10 min, making use of magnetically induced bacteria swarms, whereas only ∼20% was removed within the same time when individual bacteria were employed.

Soft robotics is an emerging and rapidly evolving field that aims at offering alternatives to hard robots to adapt to the environment due to their flexibility.^66,67 A recently exploited method for fabrication of soft robots is additive manufacturing or 3D printing. Advancing strategies have shown the potential of combining 3D printing with magnetic materials and magnetic nanoparticles to prepare magneto-responsive soft robots actuated in a similar way as micro- and nanomotors but at the millimeter scale.^68,69

A straightforward way to add magnetic functionality to a soft robot is to incorporate magnetic materials (e.g., nanoparticles) into the structure. This was the strategy to produce a hexapod magnetic soft robot that could crawl on the ground by specifically tuning the motion of every single pod using an electromagnet.⁷⁰ Alternatively, millicapsules containing a magnet were designed to travel the gut for noninvasive detection of Helicobacter pylori infections employing a pH sensor.⁷¹ A different approach explored the group motion of magnetic millibots.⁷² These millibots were 3D printed and were equipped with a small magnet to allow moving and steering by pivot walking and tumble by adjusting the direction of the magnetic field generated by a Helmholtz coil arrangement. Further, the collective motion of a group of up to six millibots was controlled by programming an algorithm to design the path and motion of the millibots, which exhibited precise actuation with different ways of motion, such as contraction or reverse maneuvering.

Advanced models exploited the incorporation of magnetic nanoparticles into flexible soft robots that adapted more effectively to magnetic forces and reshaped their bodies. For instance, a mimic of a zebrafish was reported, which could move alike its natural counterpart and navigate between obstacles.⁷³ Another example explored a magnetic flexible bronchoscope for deep lung examination that could be navigated with an external magnetic field with high precision.⁷⁴ A different example considered a magnetic ferrofluid (i.e., fluid robot) that was controlled by a rotating magnetic field and could operate diverse tasks, such as cooperative locomotion and deformation, pushing or poking an object, separating objects and adjusting the position of an object for assembling.⁷⁵ Lastly, 3D printing was utilized to fabricate magnetically actuated soft robots.⁷⁶ The robots were made of a mixture of microgels, nanoclay, and neodymium/iron/boron (NdFeB) particles and showed different folding modes when actuated with a magnetic field, depending on the 3D printed pattern.

A more complex design considered a NdFeB magnetic slime robot that moved and reshaped to adapt to different environments [Fig. 4(a)].⁷⁷ The slime moved toward magnetic fields by elongation and contraction of its body and showed shape morphism when different magnet shapes (e.g., disk, hexagon, and ring) were used, owing to its non-Newtonian fluid behavior [Fig. 4(ai)]. More interestingly, the slime robot, which adapted its shape to different microfluidic channels of varying channel width and morphology while moving through them, was even able to move on uneven surfaces and exhibited self-healing properties due to hydrogen bonding within its internal structure [Fig. 4(aii)].

Most advanced designs not only incorporated magnetic particles into the soft robot body but integrated them in a smart fashion that the controlled motion of the individual magnetic motors would actuate the soft robot. For example, poly(N-isopropylacrylamide) hydrogel beads were loaded with magnetite (Fe₃O₄) nanoparticles and different swarm arrays were constructed inside the polymeric shell by changing the applied external field [Fig. 4(b)].⁷⁸ The magnetic particles swarmed in the shape of clusters, peripheral short chains, or transverse long chains when the gravitational, magnetic gradient, and the magnetic uniform fields were used, respectively, which showed variable step-out frequencies when a dynamic magnetic field was applied for motion [Fig. 4(bi)]. The motion of the robots was described as a rolling-while-slipping movement. The internal swarm arrangements led to different maximum velocities achieved by the robots, with top values of 160–620 μm s⁻¹ when the swarms were magnetically collected, which the authors attributed to the intrinsically higher asymmetry of the systems. Owing to the different step-out frequencies the robots showed, combinations of the three types of robots were envisioned for controlled targeting and catalytic performance or drug delivery and eventual recovery of the robots [Fig. 4(bii)].

Another elegant strategy consisted of fabricating springlike magnetic micromotors that served to assemble larger soft robots [Fig. 4(c)].⁷⁹ The spring motor was fabricated by 3D lithography using a combination of an elastomer, a hydrogel, and magnetic particles, resulting in cantilever, arc, coil, and zigzag shapes for different purposes [Fig. 4(ci)]. For instance, the cantilever was used as a pico-Newton scale indicator that optically trapped microbeads and measured the force applied during the process. The arc was the basis for a microgripper that traveled toward a desired object, caught it, and transported, by employing different magnetic field rotating frequencies [Fig. 4(cii)]. Thus, high frequencies led to the microgripper motion toward the object at speeds of ∼14 μm min⁻¹, while small frequencies were employed to control the torques for opening and closing the grip, and eventually transporting the object back at speeds of ∼3 μm min⁻¹. Lastly, the coil and zigzag shapes were used to fabricate penguin- and turtle-like robots, respectively [Fig. 4(ciii) and 4(civ)]. Specifically, the penguin robot consisted of a magnetic body and two side paddles that acted as the flippers, connected to the body through the coil motor, whereas the turtle robot comprised of a nonmagnetic body and four magnetic paddles that mimicked the flippers, connected via the zigzag motors. In both cases, the millibots moved due to the periodic opening/closing of the paddles by means of the magnetic motor rotation under a rotating magnetic field, pushing the millibots forward with top speeds of ∼16 and ∼30 μm min⁻¹ for the penguin and the turtle robots, respectively.

Finally, magnetic soft helical microfiberbots with high steerability, reliable maneuverability, and multimodal shape reconfigurability to perform robotic embolization via a remote, untethered, and controllable manner were recently reported.⁸⁰ The microfiberbots were fabricated by thermal drawing a magnetic (NdFeB) soft composite into microfibers, followed by magnetizing and molding procedures to endow magnetic polarity. The microfiberbots possessed both shape morphism and motion capabilities, where each of them was activated independently at the right magnetic field conditions. Thus, when a magnetic field was applied in the direction of the net magnetization of the microfiberbot, it showed elongation, while applying the magnetic field in the opposite direction caused collapse of the structure. Modulating the magnetic field also allowed for helical propulsion with top speeds of ∼1.6 mm s⁻¹ in simulated blood without a flow or ∼0.3 and ∼1.8 mm s⁻¹ with an upstream and a downstream flow of 100 mm s⁻¹, respectively. When the microfiberbot was in its collapse state, a magnetic field gradient was used instead. The steering and maneuvering capabilities of the robots were demonstrated both in cell culture and in a rabbit model, which successfully facilitated a stable vascular embolism, with no signs of recanalization, inflammation, or pathological abnormalities.

The field of magnetic nano- and micromotors has witnessed great advances in the past few years. Magnetic fields offer a tunable and efficient way to control not only nanomotion but also steering, trapping, and guidance of the motors at a larger scale. On the one hand, magnetic field gradients represent the easiest option as they are portable and convenient for small lab spaces and the motor design is rather simple. On the other hand, oscillating magnetic fields offer higher tunability and better control, but they are limited by the space, the need of a coil arrangement system, and the fabrication of the motors must comply with geometrical features (e.g., integrating flexible parts and helical structures).

Swarm control is among the biggest advances in the area of nano- and micromotors, i.e., the possibility to achieve reversible intermotor interactions to improve their motion or perform more challenging tasks. In this regard, magnetic motors have proven to be a convenient option, as the interparticle interactions can be simply tuned, e.g., by the intensity of the applied magnetic field. Further, if the motors exhibit weak long-range interactions, the removal of the magnetic field results in the deaggregation of the swarm.

More recently, the upcoming field of soft robotics has integrated magnetic motors as a means to control actuation. Although in early development, the examples reported already hold promise for advancing adaptive materials that combine out-of-equilibrium systems for complex tasks, such as bronchoscopy or embolization with minimal invasion.

Despite the advances, magnetic motors still present exciting opportunities to answer questions to unsolved challenges. In my opinion, the next step forward requires attention to the following five key aspects (Fig. 5):

1. Greener materials/synthesis: Efforts have been made to fabricate magnetic motors that can respond efficiently and fast to external magnetic fields. The components of these motors are magnetic materials that are usually synthesized following high-temperature and/or high-pressure methods (e.g., chemical thermodecomposition and solvothemal) and organic (sometimes toxic) solvents. Organic solvents per se are not harmful or dangerous, if used correctly and with the proper waste treatment. However, due to green transition policies and sustainability efforts, limiting excessive use of organic solvents or seeking alternative solvents should be prioritized in the upcoming years. Alternatively, many examples presented in this Perspective used neodymium-based magnets, which are rare earth and, therefore, difficult to extract and expensive. Hence, there is a need to find alternative approaches that utilize lower temperatures, avoid organic solvents, and rely on cost-affordable metals to fabricate these materials. For instance, natural products could be employed instead of the commonly used reagents for synthesis.⁸¹ Although the crystallinity and magnetic response of the nanoparticles fabricated by these greener methods will not match the outcomes achieved with the current synthesis concepts, it has been demonstrated that magnetic motors did not need excessively high magnetic moments (i.e., saturating magnetization) to accomplish motion.^82–86 Indeed, magnetic fields used to induce motion usually performed in the mT range and, in many cases, the composite materials that compose the motors showed a low saturation (e.g., below 20 A m² kg⁻¹).

2. Alternative mechanisms of motion: Magnetic fields were usually employed in the way of gradients or oscillating fields to propel motors and control their guidance. However, variable magnetic fields can also induce heat generation in magnetic nanoparticles. This phenomenon, known as magnetic hyperthermia, has been widely used in nanomedical approaches to combat tumorlike models using nanoparticles^87,88 and motors,⁸⁹ and it is based on the ability of magnetic particles to accumulate energy when exposed to an alternating magnetic field and release it in the way of heat. Two mechanisms are proposed, namely, Néel (spin-spin) relaxation and Brownian (particle-solvent) relaxation, which were extensively described elsewhere.⁸⁷ Consequently, magnetic fields could prove useful to produce heat losses in magnetic motors to result in self-propulsion, similar to when externally applied thermal gradients propelled (magnetic and nonmagnetic) particles owing to the so-called Ludwig-Soret effect.^90–93 Despite the potential of this type of motors, this mechanism of motion has not been extensively exploited so far.

3. Swarm control: Advances have been made in controlling motor swarming, specifically under magnetic fields. In this regard, motors have to present high colloidal stability in the medium to avoid interparticle interactions, especially when using magnetic materials, which often possess long-range dipolar interactions. For this reason, a note of caution should be given. Swarming is typically understood as a collective motion arising from specific interactions between particles that allegedly enhance their locomotion performance. Therefore, unspecific clustering or chain formation due to dipolar interactions is spontaneous aggregation rather than swarming. Additionally, swarming conveys the idea of being able to activate and deactivate the flocking of the motors by applying or removing the external trigger that causes the interactions. Therefore, future advancements should consider better control over the dynamic interactions between motors via external stimuli.

4. Motion in complex environments: Examples presented in this Perspective dealt with motion of magnetic motors in flow regimes of 400–100 mm s⁻¹, attempting to mimic blood flow. However, when envisioning motors to be employed in biomedical or environmental applications where fluid flow is present, which ranges from 40 to 100 cm s⁻¹ for blood in humans (i.e., 10×–100× higher than the experiments reported), and several orders of magnitude larger in water reservoirs. This highlights that, despite advancements made and the community's recognition of the need to consider the locomotion abilities of motors within fluid environments, wide practical applications remain a distant prospect. Hence, further efforts need to be made to understand the dynamics at low Reynolds numbers and design motors that could overcome fluid flows, where swarming could be beneficial to boost the locomotion.

5. Integration in soft robots: Soft robots stand at the frontier between microscale adaptive matter and conventional rigid robots, and they have drawn great attention in the past few years. Magnetic materials have been combined into structures to allow for magnetic manipulation and a step forward considered the use of not only passive magnetic particles but also active colloids (magnetic motors) to activate the soft robots’ functionalities. In this sense, there is room for finding synergistic combinations of soft materials and magnetic motors to actuate the robots in smarter ways, and the works collected so far only demonstrate the enormous potential this new class of active materials can pose.

The author acknowledges Brigitte Städler for discussion and mentoring during the writing process and the Independent Research Fund Denmark (DFF).

The author has no conflicts to disclose.

Ethical approval is not applicable to this article as no experiments on animals and/or humans were performed in this study.

Miguel A. Ramos Docampo: Conceptualization (lead); Writing – original draft (lead); Writing – review & editing (lead).

Data sharing is not applicable to this article as no new data were created or analyzed in this study.

Miguel A. Ramos Docampo obtained his Ph.D. (cum laude) from the University of Vigo (Spain) in 2020. His Ph.D. dissertation was based on the design of magnetic nanoparticles for biomedical applications, with especial attention to magnetic guidance and heat delivery. He is currently a postdoctoral researcher in the Laboratory for Cell Mimicry at the Interdisciplinary Nanoscience Center (Aarhus University, Denmark). His research focuses on the design of bio-inspired polymer-driven nanomotors to self-navigate in inhomogeneous and complex environments, and through biological barriers.