Microassembly platforms have attracted significant attention recently because of their potential for developing microsystems and devices for a wide range of applications. Despite their considerable potential, existing techniques are mainly used in laboratory research settings. This review provides an overview of the fundamentals, techniques, and applications of microassemblies. Manipulation techniques based on magnetic, optical, acoustic fields, and mechanical systems are discussed, and control systems that rely on machine vision and force feedback are introduced. Additionally, recent applications of microassemblies in microstructure fabrication, microelectromechanical operation, and biomedical engineering are examined. This review also highlights unmet technical demands and emerging trends, as well as new research opportunities in this expanding field of research driven by allied technologies such as microrobotics.
Yujian An, Bingze He, Zhuochen Ma, Yao Guo, Guang-Zhong Yang.
Microassembly: A Review on Fundamentals, Applications and Recent Developments.
Engineering, 2025, 48(5): 341-366 DOI:10.1016/j.eng.2024.09.024
Minimization of macroscopic mass-produced devices toward their micro/nano-scale counterparts has recently attracted extensive attention, which has driven the development of new microsystems for a variety of complex tasks, including micro-implantation, targeted delivery, and tissue engineering. Traditional methods for the fabrication of microsystems commonly involve the heterogeneous integration of electronic devices, micro-electromechanical structures (MEMS), and optoelectronic devices on the same substrate. These components are built using microfabrication processes such as three dimensional (3D) printing and photolithography, allowing the manufacturing of millions of components in parallel. Current integration strategies rely mainly on “pick and place” approaches that manipulate components through surface forces and then release them with the assistance of precise positioning. To improve the performance of these micro-devices, functional components with different materials and geometries must be integrated at the microscale in a predetermined fabrication sequence.
However, simply scaling macroscopic counterparts down to the micro/nano-scale would not be effective because different physical forces manifest in varying proportions owing to scaling effects. For example, adhesion forces that originate primarily from surface tension, van der Waals forces, and electrostatic forces are fundamental limitations of micromanipulation [1], [2]. In particular, the adhesion forces between objects are significant compared with the gravitational forces when the sizes of the components are less than 1 mm. Such an adhesion effect on a small scale introduces difficulties in the fabrication of miniature devices. Specifically, as the size of the target object decreases, the surface-area-to-volume ratio increases, leading to a more pronounced scaling effect. This results in the surface forces dominating volume forces. At the microscale, the dominant forces are opposite to those at the macroscale. In this context, gravity and inertial forces are considered negligible, while micro forces, such as van der Waals and capillary forces, are becoming more important. Previous study have shown that micro/nano-robots are sensitive to environmental parameters, and that the dominant forces in different media are distinct (only van der Waals forces are ubiquitous) [3]. Therefore, the design of microassembly/micromanipulation processes must consider these to isolate undesired interference, which is a challenging task in practice.
Electrostatic attraction Fel relates to the coulomb force between components with charges [1]. For simplicity, the force between a charged object [2], for example, a sphere and a dielectric plane, can be estimated by
where ∊0 and ∊ are the dielectric constants of air and the plane, respectively; σ is the surface charge density; and d is the diameter of the object. The van der Waals force Fvdw is an intermolecular force driven by the momentary movement of electrons. The approximate expression [2], [4] for the van der Waals force between an object of diameter d and a plane at distance z can be expressed as
where H is the Hamaker constant. The last and most important type of adhesion is the capillary force (Fcap) that arises from a thin liquid film between two objects and originates from the humidity of air. For a hydrophilic object close to a plane, the relation can be determined by
where γ the surface tension force. The adhesion forces are surface forces, and their magnitudes are thus proportional to the contact area [1], [5].
Previous studies have provided basic insights into the magnitudes of different types of adhesion forces versus distance [5], [6]. Fig. 1 shows the difference in the force relationships at different distances, which is mainly reflected in the replacement of the dominant role of gravity at the micro/nano-scale. Furthermore, as the scale continues to decrease, objects that cannot be observed with the naked eye must be observed using light or electron microscopes. The individual operability of these small objects is reduced, but their cluster values increase. These physical principles of the scale effect result in different strategies for the development of devices at the micro/nano-scale. Intuitively, overcoming common obstacles in a macroscopic scenario presents a significant challenge in the micro/nano-field. In recent years, some studies have solved the problem of the single-movement mode of magnetic microrobots with simple shapes through multimer design. For example, Yu et al. [7] integrated prepared dual-material microspheres of nickel and silica into a trimer configuration, achieving the ability to simply cross ladder obstacles through a tumbling motion, and the structure could be reconfigured based on task and environmental requirements.
Microassembly, using mechanical micro-operation systems or external fields such as magnetic, optical, and acoustic fields, offers a practical method for mass-producing micro-devices with predetermined structures. Recent advances in Lithographie, Galvanik und Abformung technologies have enabled the manufacturing of different microscale components with different geometries and materials. These assembled devices, which are usually manipulated through non-covalent interactions, are responsive to environmental stimuli such as temperature, pressure, and flow. Biocompatible and biodegradable microrobots have been envisioned for biomedical applications such as micro-implantation, minimally invasive/non-invasive medical treatment, and cellular-level exploration. An intravenous magnetically spiked spherical microrobot developed by Li et al. [8] presents a compelling application of bionic structures. Inspired by tardigrades, they developed a dense grasping structure for a spherical robot and used red blood cell membranes for surface modification to solve the problems of intravascular adhesion, active retention, and resistance to blood-flow effects. This type of microbial bionic research is becoming increasingly popular, especially for micro/nano-robot research, as nature provides a perfect source of inspiration for micro/nano-robot design. Similarly, Zhang et al. [9] drew inspiration from amoebas, and used alternating magnetic fields to provide magnetic droplets with the ability to move and deform. Moving magnetic droplets have no rigid structure and can relatively easily complete the task of crossing narrow passes, as well as other tasks such as swallowing, transporting, and releasing specific micro-objects. The magnetic droplet robot demonstrated promising results for in vivo movement and drug delivery. Considering the constraints involving size, flexibility, materials, and emerging challenges, including the generation of an external field and sensing and control strategies in practical microsystems, must be addressed. As illustrated in Fig. 2, this review highlights recent progress in microfabrication, types of external fields, and control strategies for micro-assembly. We focus on three aspects: fundamentals and technologies, information sensing and control strategies, and practical applications. We also discuss the potential of the new microassembly schemes for constructing new microstructures and biomedical devices.
2. Fundamentals and Technologies
Common strategies for generating power for actuation at the micro/nano-scale include magnetic, optical, acoustic fields, and mechanical methods. These strategies enable the manipulation, movement, and arrangement of different structural units including nanoparticles, microrobots, and cells. Depending on the usage scenario, different actuation methods with specific characteristics and adaptation features have been designed.
2.1. Magnetic field
Owing to their rapid response and ability to be remotely controlled, magnetic field-based methods offer a pathway for one-dimensional (1D) to 3D microassembly. Recent progress has enabled microassemblies to form complex microscale structures and microrobots through the manipulation of external magnetic fields.
2.1.1. Magnetic field-induced assembly (MFIA) of nanoparticles
MFIA allows for the 1D, two dimensional (2D), or 3D organization of magnetic particles under the influence of a magnetic field. This refers to the automatic and spontaneous arrangement of particles within a magnetic field rather than an assembly induced by artificially moving targets.
Previous studies have achieved MFIA by applying a magnetic field with uniform intensity using electromagnetic devices or permanent magnets. The magnetic particles are aligned along the external magnetic field, thus inducing particle aggregation. 1D linear structures or chain-like patterns of magnetic particles can be formed under different magnetic fields of uniform intensity [10], [11]. For in vivo applications, the deployment of MFIA in biofluids such as blood and vitreous was investigated [12]. Viscosity, ionic strength, and microstructure have been shown to play important roles in MFIA.
To improve the performance of the 1D MFIA, Tan et al. [13] investigated the influence of the length, thickness, and alignment uniformity of the MFIA on the magnetic field parameters. In addition, MFIA, as a microprocessing method, allows the integration of diverse functional components into a device. Wen et al. [14] fabricated a creatinine-detection component with a controllable structure using MFIA on the surface of a magnetic glassy carbon electrode. In Ref. [15], Ferrous Phosphide (Fe2P) nanoparticles were integrated into nanotubes using MFIA. The resulting nanotubes displayed enhanced energy storage capacity, which increased the temperature range.
In addition to 1D manipulation, MFIA has recently demonstrated excellent 2D planar processing capabilities. Park et al. [16] assembled devices with regularly arranged hexagonal network arrays that can be used to prepare nanoscale thin films. Similar study with different materials were performed by Mohapatra et al. [17], who further optimized the stability of devices by leveraging intra-lattice interactions. The development of complex spatial structures and desired functions can be facilitated by 3D MFIA. Jeong et al. [18] used magnetic plasma to realize a 3D MFIA-based helical magnetic flux consisting of silver nanoparticles. The chirality of the helical magnetic flux can be switched within milliseconds by applying a magnetic field.
MFIA relies on an organized geometry from 1D to 3D through the collective motion of particles under a magnetic field. MFIA is not designed to control the movement of individual particles during product formation. Instead, it facilitates the fabrication of devices with simple linear and chain-like geometries. Although the assembly process is not completely spontaneous, each component involved in the construction is controlled.
2.1.2. Magnetic assemblies with arbitrary geometries and functions
Although MFIA can accomplish several configurations in different dimensions, designing devices with arbitrary structures remains a significant challenge because of unclear design rules. Ideally, devices should allow reconfiguration, actuation, and propulsion in diverse environments in a programmable manner. This largely depends on the synergistic effect of the magnetic units and their magnetizations. Considering these aspects, MFIA enables the controllable and selective docking and assembly of magnetic devices with desired microscopic patterns or mechanical structures. The manipulation of microstructural units for MFIA paves the way from micro-devices to microrobots. Therefore, this section delves into the magnetic units and their magnetization under different magnetic fields.
The force and torque of the magnetic units in the magnetic field can be calculated by the following formulas [19].
where is the magnetic force, and are induced magnetic dipole moment of the cluster and the magnetic field at point , respectively. ∇ is the gradient operator.
where is the magnetic torque.
To facilitate planned assembly, the position and posture of the magnetic assemblies must be precisely controlled under magnetic fields, as demonstrated by Xia et al. [20]. The magnetic units were precisely actuated through radial magnetic levitation (Fig. 3(a)), which was the basis for the controllable assembly of all the units with an anisotropic magnetic response.
Magnetic units must be preprogrammed to accomplish anisotropic responses under a magnetic field. The response of the programmed units to the magnetic potential energy distribution, as demonstrated by Dong and Sitti [21] (Fig. 3(b)), facilitates the on-demand movement of those magnetic units. This enables them to be further assembled into microrobot swarms for applications such as drug delivery. Furthermore, temporary anchoring during processing is necessary to build complex structures from unified magnetic micro components. Temporarily anchored micro components can resist subsequent field changes and are not released until all the micro components reach their target positions and are linked. To this end, Hu et al. [22] used two-photon polymerization technology to create a temporary poly(ethylene glycol) diacrylate hydrogel by anchoring a sacrificial layer between the micro-component and the substrate, thus avoiding unexpected magnetic field effects. This anchoring structure can be released by soaking it in deionized water. Moreover, after changing the materials, a similar technology can be used to print assembly rings or cantilevers on the micro components, thereby offering flexible microassembly possibilities. Alternatively, a programmed structure has been developed using a combination of geometric constraints inherent to complex structures and glue as a mechanical lock [23] (Fig. 3(c)). Gu et al. [24] introduced a quadrupole module for programmed devices using quadrupole structures that reduced the impact of neighboring units in Fig. 3(d). Combined with the directional response to the magnetic field, the size and magnetization of the individual units are also controllable; thus, the modules can be assembled into planned 2D geometries.
In addition to assembling devices with programmed patterns, the fabricated structures can be further functionalized for specific applications. Yang et al. [25] obtained the microstructure of a reconfigurable architecture based on an assembly of magnetically responsive micro components of different lengths.
These micro components can form diverse geometries and can easily be decoupled as the magnetic field gradually dissipates. This flexibility in structural reorganization allows them to adapt to and overcome different environmental constraints. To expand programmed magnetic assemblies to biological materials with programmed orientations and structures, Tasoglu et al. [26] used the paramagnetism of radicals to biological materials to endow magnetic assemblies of different arbitrariness with programmed orientations and structures (Fig. 3(e)). The magnetic properties of the resulting structures can be eliminated, and the materials used for magnetic assembly are not limited to metallic magnetic materials (MMs).
Research on magnetic assemblies has led to significant advances in this field. These studies have highlighted innovative structural designs and optimized control methods. This was achieved by accurately manipulating magnetic units to construct programmed micro-devices. The use of magnetic devices provides the basis for the development of micro-biotics. When these units become functional with closed-loop control, for instance in drug delivery, they can be classified as microrobots rather than merely micro-patterned components.
2.1.3. Micro-assemblies with magnetic microrobots
The application of an external magnetic field can both convert different components into programmed devices and turn them into magnetically actuated microrobots [27], [28], [29], [30], [31]. This section summarizes the magnetic actuation and navigation of assembled microrobots. These microrobots can be divided into two main types: ① magnetically actuated microrobots for robot-assisted assembly, and ② assembled swimming magnetic microrobots as carriers or deliverers.
Magnetically controlled microrobots have been used for several different microassembly tasks. The assembled block, spherical, and flake-like magnetic doping devices can be used to assist robots to push different units for the assembly of micro components. Barbot et al. [32] used a simple set of sheet-like microrobots to hold a thin film on a liquid surface by controlling its distance from a magnet (Fig. 4(a)). Alapan et al. [33] applied spherical assistant microrobots to induce the assembly of larger non-magnetic components. Owing to their simple structure, these microrobots can be used in diverse settings. The research represented by the study of Tasoglu et al. [34] and Johnson et al. [35] developed magnetic microrobots with cube geometry, facilitating the movement in fluid environments and solid surfaces to carry nonmagnetic components for 2D and 3D pattern assembly. Yang et al. [36] designed magnetic microrobots consisting of chitosan and alginate that can be used for precise and stable grasping and transportation. The application of magnetic microrobots promotes controllable 2D and 3D movements upon the application of an external magnetic field. Hsu et al. [37] combined force sensing and computer vision to develop microrobots with low hysteresis and high compliance. Yao et al. [38] used the hydrodynamic properties and capillary interactions of magnetic microrobots to control units at the liquid–liquid interface. However, their applications are limited to the assembly of units of specific shapes.
Another microassembly strategy involves holding and transporting different units using microgrippers. Compared with magnetic microrobots for assisting the assembly process with pushing units, magnetic microgrippers can grab and thus enable a more sophisticated assembly in 3D. Ji et al. [39] designed magnetic microgrippers with magnetic and nonmagnetic resins as units using digital light-processing 3D printing technology. In recent studies [40], [41], a magnetic field was used to drive an untethered microgripper toward a target position in a narrow cavity, and the grasping capability of the microgripper was determined using a thermally responsive shrinkage actuator. Although it has not yet been applied in clinical settings, it has potential applications in fields such as in vivo cell extraction (Fig. 4(b)).
Assembled magnetic microrobots under a rotating magnetic field of uniform intensity can achieve controllable propulsion through their interaction with different fluids and substrates. For example, the magnetic microcube robots proposed by Han et al. [42] can be assembled into a controllable chain-like structure for transporting living cells or as microstirrers under a magnetic field (Fig. 4(c)). In contrast to artificial devices [43], biohybrid microrobots, such as cell-based microrobots, can be developed by depositing magnetic components onto biological templates [44]. Feng et al. [45] developed a macrophage-based magnetically controlled robot that supports movement in 3D space under magnetic traction and can deliver targeted drugs to kill cancer cells. Magnetotactic bacteria and their derivatives have been used successfully for structural construction and targeted drug delivery. Gong et al. [46] induced cell-based magnetic microrobots to form chain-like assemblies, including dimers, trimers, and tetramers (Fig. 4(d)). Cell-based microrobots are maneuverable and capable of a wide variety of movements. These spherical microrobots can respond to different environmental factors, such as light, which enables them to perform drug delivery tasks.
The magnetic swimming microrobots discussed in this section hold the potential for a range of applications, including the assembly of nonmagnetic components and targeted drug delivery.
2.2. Optical field
Owing to their non-contact and biologically friendly characteristics, optical fields are promising platforms for microassembly through the 3D manipulation of objects across a variety of sizes and material properties [47]. Compared with assembly using a magnetic field, optical field techniques, such as optical tweezers, enable the assembly of 3D structures with multiple materials. This can enhance the performance compared to a single-material design [48]. Most approaches for creating devices with different materials involve direct operation and combined methods.
Light fields can be used to directly manipulate 3D multicomponent movable assemblies. As early as 2001, researchers directly processed complex assemblies by scanning ultrashort-pulse near-infrared laser beams along designated paths in photocured polymers [49]. Photothermally driven micromotors are core devices [50], [51]. The designed micromotor has a good photothermal conversion capability and induces micromotor movement by generating a temperature gradient in the medium, thereby generating hydrodynamic flow and causing automatic thermophoretic movement. The cooperative swarming behavior of these micromotors is a form of photophoretic swarming.
Melzer and McLeod [48] achieved a high-precision, high-accuracy 3D assembly of many micrometer-scale spherical units using optical tweezers (Fig. 5(a)). Liang et al. [52] and Mou et al. [53] reported studies involving the light-powered assembly of microparticles as units (Fig. 5(b)). Particles can be assembled into several different swarms by using their interactions and nonequilibrium interactions under an optical field [54]. Another investigation by Tong et al. [55] demonstrated the light-induced operation of carbon nitride/polypyrrole nanoparticle-based particles in water, whose behavior resembled that of phototactic microorganisms.
Optical tweezers can be used to manipulate individual particles as units and to assemble multiple particles into bespoke arrays. Using holographic beam techniques, particles can be simultaneously trapped in a structural optical field through optical trapping forces. After years of exploration, the use of optical tweezers has become widespread. Many previous studies have used optical tweezers directly to capture micro objects or construct structures [56], [57]. In addition, research on establishing a combined system of optical tweezers and other common tools to improve functionality is also popular. For example, by combining an electric microtip [58] or using a chemical method [59], the disadvantage of the optical force of the optical tweezers is that it is difficult to fix the components firmly. Research combining optical tweezers with other techniques has continued in recent years. For example, by coupling microbubbles and optical tweezers, Ghosh et al. [60] developed an assembly with programmed patterns by using photothermal action on soft oxometalates to generate oxides and polymers in situ. Similarly, Tang et al. [61] introduced an ultraviolet light-induced crosslinking strategy to immobilize the assembly with silica spheres after near-infrared optical tweezers captured the spheres.
Optical tweezers, which are popular because of their noncontact nature and superior biofriendliness, are often used to extend biological entities for assembly. The focus has been on the arrangement and binding of cells [62], sometimes including industrial products, such as polystyrene spheres [63]. Optical tweezers, combined with microfluidics, have been used for cell sorting and encapsulation [64], [65].
Recent studies have focused on addressing the limitations of optical tweezers. Pradhan et al. [66] explored the differences in the capturing of optical tweezers on 1 µm colloidal particles of different materials at the air–liquid interface and clarified that the size and shape of spontaneously formed assemblies rely on the surface tension and surface charge conditions. Zou et al. [67] used optical tweezers to capture particles, allowing them to move in a circular pattern along a specified trajectory to create a microvortex. This microvortex can be considered as a type of micromotor. The movement direction and speed of the objects within the vortex can be controlled by adjusting the system parameters (Fig. 5(c)). Such contact-free, controllable cell actuation shows the potential for both research and therapy in vivo environments. Tanaka et al. [68] encapsulated colloidal particles in mixed droplets of thermally responsive ionic liquids and water. Through the contraction of the droplets, particles outside the irradiation range can also be collected, overcoming the limitations of the working range of the optical tweezer (Fig. 5(d)). Shan et al. [69] used resonance technology to significantly improve the optical trapping power of optical tweezers, thereby providing a solution to the trapping problem of nanometer-sized objects with a low refractive index. The use of optical tweezers for microassembly enables the manipulation of different objects as units. This technology can further advance optical tweezers for microassembly on a larger scale beyond laboratory environments.
2.3. Acoustic field
Acoustic fields are another popular technique for contactless assemblies. Studies have shown that the acoustic field can help complete the transportation of targets and the arbitrary patterned assembly of 2D/3D with resin particles, solid particles, and even living cells, which is significant in mechanobiology [70], [71], [72].
Ultrasound, a versatile external field, offers the energy to organize microparticles into assembled devices through a process known as ultrasound directed self-assembly (DSA). The application of an acoustic field generates propagating surface acoustic waves (SAWs), resulting in the generation of primary and secondary radiation forces onto the particles as units. The latter is mainly responsible for the interaction between particles as units as well as the formation of assembled devices [73], [74], [75] (Fig. 6(a)). Several studies on ultrasound DSA have attempted to arrange particles with specific geometric shapes in 2D/3D spaces using an acoustic field. For example, Tang and Huang [76] assembled particles in an acoustic field. A non-reflective converging acoustic wave field was established to converge toward the center and manipulate the particles along the node of the acoustic field. Greenhall et al. [77] investigated the dependence of the behavior of particles in a fluid on the parameters of the ultrasound field, and the optimized parameters realized the arrangement of carbon nanoparticles in 2D and 3D [78].
To increase the precision of ultrasound DSA, Prisbrey et al. [79] quantitatively analyzed the error between the design for user requirements and the resulting assembled devices. Specifically, errors in assembling devices with high aspect-ratio particles such as carbon nanofibers are of great importance for the practical application of ultrasound DSA [80]. To achieve DSA of carbon fibers in 3D, Prisbrey et al. [81] studied the dynamics of particles with a high aspect ratio using ultrasound DSA. It has been demonstrated that more than two unique wave propagation directions are essential for orienting particles with high aspect ratios under an acoustic field. In addition, the orientation error decreased with an increasing number of unique wave propagation directions. Similarly, Feng et al. [82] used an acoustic field to levitate a microrobot and used it as a tool to achieve high-precision 3D single-cell manipulation of oocytes. The development of ultrasound-DSA technology has expanded the range of materials suitable for 3D printing. This is achieved by integrating uniformly arranged particles into the material matrix. For instance, Wadsworth et al. [83] incorporated carbon nanofibers into photopolymer resins to prepare conductive composites for 3D printing and evaluated the effects of parameters such as the ultrasound frequency and printing speed, which depend on the properties of the products. Greenhall and Raeymaekers [84] and Niendorf and Raeymaekers [85], [86] introduced conductive fibers into a photopolymer matrix using ultrasound DSA angiography as the 3D printing material (Fig. 6(b)). The relationship between the material conductivity and parameters during the preparation process was also identified. All these 3D printing-related studies can be used for the fabrication of insulated wires.
Like magnetic and light fields, acoustic fields can be used to manipulate microscopic particle/particle swarms. Image-based visual feedback is an efficient and intuitive strategy to achieve closed-loop control. For instance, Wei et al. [87] proposed closed-loop control of a single microsphere moving along a straight line based on image information. Their acoustic micro robotic interface supported a response time < 1.1 s and a positioning accuracy of 2.5 , and had a reverse callback function to address position overshoot [87]. Furthermore, in another study, Lu et al. [88] found that speed responses depended on the operating frequencies and voltages. Based on this speed-adjustable system, researchers have established a controlled relationship between the movement of particles/cancer cells and typing/melody. Cancer cells can be transported following a designated melody with positions determined by the notes. This provides interesting human–computer interaction with precision control (error < 20 ). In addition, the ultrasonic-driven swarm microrobots developed by Schrage et al. [89] are highly relevant. The proposed propulsion strategy consists of a combination of primary and secondary radiation forces, with the former guiding forward progress and the latter condensing micro swarms. Artificial intelligence and deep learning have been used to accelerate the development of microrobots and their control strategies. For example, owing to the challenges of establishing a direct driving model and nonlinear interference in practical experimental environments, reinforcement learning has been conducted using a lot of experimental motion prediction data, resulting in desirable control outcomes.
In addition to the creation of pressure nodes, ultrasonic fields can also be applied to microgrippers in air and water environments [90]. Unlike magnetic and optical fields, the acoustic actuation of microgrippers requires a system with a simple design and small size. Motivated by this, Mohany et al. [91] developed an acoustically driven microgripper, Sono Tweezer, to capture and manipulate submillimeter-sized particles in fluidic environments (Fig. 6(c)). An ultrasound field at 25 V of transducer voltage can create a pressure of 300 kPa to actuate the microgripper. Several studies have investigated the operation of cells using ultrasonography. Durrer et al. [92] developed a sensor using a glass capillary tube and a piezoelectric transducer, which was integrated into the front end of the robotic arm as an end effector to form an acoustofluidic robotic system. In liquid environments, sound waves can induce vortices that are used in scenarios such as liquid pumping, particle capture, and droplet merging. For instance, Yang et al. [93] established an ultrasound platform for cell translation, rotation, orientation, and levitation in 2D planes. Luo and Wu [94] used ultrasound waves to induce vibrations in microbubbles trapped in microrobots to drive and control robots. In summary, although an ultrasound DSA was developed for specific applications, it offers potential for future applications once its resistance to interference is further optimized.
2.4. Mechanical grippers
In addition to the aforementioned strategies using external fields, several studies have investigated microassemblies using mechanical grippers based on motorized displacement platforms or robotic arms. Most grippers can handle objects without any specific material properties such as magnetic properties or refractive index differences. However, their parts, excluding the functional end effectors, tend to be large and require an open operating environment. This study focuses on introducing mature product parameters and related technologies suitable for commercial use.
Most micro mechanical grippers are designed in the form of “a pair of pliers,” like the operation of human hands [95]. As a representative example, Lyu et al. [96] designed a microgripper based on a pair of pliers for gripping and manipulating objects (Fig. 7(a)). These micro-clamps were opened and closed by manipulating the pliers. Although this method is simple, it lacks dexterity. Enhancing the degree of freedom (DoF) through external mechanical groups has been a well-established technical approach for decades [97]. The integration of multiple fixtures and end effectors provides a system with remarkable functionality and flexibility.
As for driving methods, piezoelectric and electrothermal driving are common and mature strategies. In the representative works of Das et al. [98], [99], a basic structure based on a three-stage displacement amplification mechanism was frequently used. After amplifying the input signal of the piezoelectric actuator, excellent experimental results of 1044 Hz bandwidth and ±2.158 nm steady-state error [99], as well as an amplification rate of 23.96 and a motion resolution of ±4 nm, can be obtained [98]. These high-performing grippers allow for a quick transient response and maintain high displacement accuracy. The classic three-stage amplification structure was also used by Shi et al. [100], who achieved an amplification ratio of 31.88 and a grasping force of up to 1.993 N (Fig. 7(b)). The three-finger electrothermal gripper developed by Si et al. [101], [102] is a prime example (Fig. 7(c)). After the Joule heat generated by the current was applied to the polyimide film of the beam actuator, the heated beam deformed the metal and was driven. The practicality of this gripper was demonstrated by its ability to grasp and move delicate objects such as microspheres and zebrafish embryos. Lingaraja’s electrothermally driven hot- and cold-arm microgripper tripled its thermal efficiency by eliminating the parasitic resistance of the cold-arm [103]. This improvement enables safer manipulation of micro-objects and biological particles while reducing power consumption.
Researchers have enhanced the precision of assembly using mechanical grippers by focusing on automated assembly using mechanical microgrippers. It is primarily based on visual or force feedback to enable real-time feedback control. Chen et al [104] and Beyeler et al. [105] developed a series of microgrippers with integrated force sensors that can grip microscopic targets with high sensitivity and successfully overcome the adhesion force for release. They proposed the first force-controlled microgripper at the nanoscale Newtonian force level, which is of great significance for subsequent research. Ma et al. [106] realized automated assembly using microgrippers with force sensors. In this study, the assembly was restricted to the designed system. Komati et al. [107] proposed two complementary automated assembly techniques to extend the range of microgrippers. The process of the assembly is based on solid CAD models and hybrid control of force and position, which can be achieved within one second [107]. Recently, indirect force sensing strategies have been developed. For example, Fantoni et al. [108] observed the displacement of a pointer extending from the head of a gripper through a visual system to indirectly calculate the grasping force. Although the vision system was calibrated to eliminate inspection errors, systematic errors caused by factors such as hardware aging cannot be ignored. Such indirect feedback strategies provide new ideas but are still worthy of continued exploration.
New studies have been expanded in a more comprehensive and targeted manner. For example, Aravind et al. [109] developed a special gripper for thin objects with large surface areas. The four pairs of adjacent clamp structures provided a 300 µm thick displacement gap and were well-suited for humid environments. Pasquale additively prepared MEMS microgrippers based on laser processing with a structure in which the driving and grasping arms were separated, which overcame the potential damage to cells caused by dangerous factors, such as electric heating or leakage [110]. Furthermore, practical applications and mechanical structures were considered in the head optimization, which made it compelling.
Generally, microassembly systems built on mechanical grippers can integrate different types of information feedback and allow precise control because of their connection to macroscopic operating systems. In addition, because their structural design resembles that of human hands, their operations are aligned with human operational habits. The technical features and typical applications are listed in Tables 1[10], [46], [99], [100], [102] and 2[47], [86], [90], [94], [114], [115], [116].
2.5. Micromanipulation systems
Precision positioning platforms are used in high-precision microassembly processes to generate precise and controllable linear and angular displacements. With advances in the field of microassembly, the demand for such platforms has increased. This not only includes a larger motion range, better positioning accuracy, and sensitivity but also superior dynamic characteristics and resistance to interference. In addition, precision-positioning platforms must possess many DoFs to achieve a larger working space. Precision positioning platforms can be broadly categorized into serial and parallel platforms.
Serial micromotion positioning platforms are advantageous because of their simple structure, ease of control, and wide range of motion. However, their accuracy is often low owing to the cumulative errors introduced by the serial drive mechanisms. The MM3E micromanipulator (Kleindiek Inc., Germany) features three translational DoFs and one rotational DoF. With a minimum movement resolution reaching 0.5 nm and equipped with position encoders on all four axes for closed-loop control, the manipulator ensures precise and controlled movements. Because of its exceptional resolution, this product has extensive applications in nanomanipulation, transmission electron microscopy sample preparation, and nanoprobing. ANSxyz50 nanopositioners (attocube Inc., Germany) possess three DoFs in the x-, y-, and z-directions and are composed of three single-DoF stages in series. Its resolution can reach the sub-nanometer level. The LifeForce nanomanipulation system (Toronto Nano Instrumentation, Canada) can be operated automatically using scanning electron microscopy (SEM). It can accommodate up to four micromanipulators, each with three DoFs in the x-, y-, and z-directions. The open-loop resolution can reach nanometers, and its drift is as low as 0.35 nm∙min−1.
Parallel micromotion positioning platforms interconnect the moving and base platforms into one or more closed loops through two or more parallel branches of motion chains, becoming increasingly popular in diverse microassembly tasks. These parallel platforms exhibit high stiffness-to-weight ratios, high load-bearing capacities, and compact structures. The parallel configuration of the drives eliminates the possibility of drive error accumulation, thereby significantly enhancing the output accuracy. Nevertheless, the major drawback of parallel micromotion positioning platforms is the coupling between adjustments in the motion of each DoF, which leads to a more complicated control model. In addition, their motion space is limited, making it difficult to fulfill the demands of large strokes and broad spatial movements. For instance, the milliDelta robot investigated by McClintock et al. [114] used MEMS manufacturing for printed circuit boards and was driven by three independently controlled piezoelectric-bending actuators. The milliDelta robot achieved precision within approximately 5 µm in a workspace of 7.01 mm3. Furthermore, it can follow periodic trajectories with frequencies of up to 75 Hz. Simultaneously, the study demonstrated the milliDelta robot’s ability to execute complex trajectories during the rapid picking or placement of small objects, as well as its capability to compensate for tremors in surgical or assembly operations. Leveziel et al. [115] presented a miniature gripper robot (MiGriBot), a downsized parallel robot featuring a configurable platform and soft joints specifically designed for microscale pick-and-place operations. The robot’s configurable platform incorporates an internal DoF, which is used to actuate the micro-tweezers through piezoelectric bending actuators located at the base of the robot, effectively reducing its inertia. The incorporation of soft joints facilitates miniaturization of the mechanism and mitigates friction. These advantages empower MiGriBot to achieve a throughput of 10 pick-and-place cycles per second for micrometer-sized objects, demonstrating a precision of 1 µm. Several commercially available parallel precision positioning platforms, summarized in Table 3 have demonstrated practical performance in the field of microassembly. For instance, the Hexapods (Physik Instrumente, Germany) based on a Stewart structure exhibits six DoFs with a minimum movement resolution of 80 nm. The Hexapods can handle a maximum end-load of 5 kg, showing exceptional performance in microassembly processes involving end effectors with significant mass or objects being manipulated. Similarly, the precision positioning platform of SmarPod (SmarAct Inc., Germany) also features six DoFs with a minimum movement resolution as fine as 1 nm. This makes it particularly suitable for microassembly and micromanipulation processes that require extremely high precision [116].
In addition to the aforementioned series and parallel combinations, the HybridHexapod (Alio Industries, USA) is a six DoFs working platform with a hybrid combination of series and parallel structures. In this configuration, two translational DoFs within one plane are separated, thereby addressing the limitation of smaller travel distances in the translational DoFs in parallel structures. This design satisfies the specific requirements for extended travel distances in certain applications. The working space dimensions were 60 mm × 450 mm × 62 mm (corresponding to the x-, y-, and z-axes), and the rotational angle around the z-axis can reach 360°.
2.6. Other technologies
Except for the aforementioned external fields and mechanical grippers, other technologies can also be used in microassembly. Sariola et al. [111], [112] reported assemblies using droplet-assisted mechanical manipulators. The droplets were placed on two mating surfaces, and attachment and assembly guidance were accomplished through capillary force. Given that the droplet facilitates automatic calibration, the method is suitable for a wide range of units and reduces the requirements of precise mechanical manipulators, allowing objects to be rotated and assembled into a hierarchical or even a cantilever structure. In addition to electrostatic interactions, Ge et al. [113] proposed a bubble-based strategy for microassemblies. Bubbles were generated by the photothermal effect of laser irradiation on the surface of amorphous silicon. The size and position of the bubbles were controlled by adjusting the intensity of the laser. These bubbles can drive the movement and arrangement of cells. Bubble robots operate based on the same principle as that of magnetically controlled robots that push other objects. However, the droplet-assisted assembly strategy has stringent requirements regarding both the object of application and the environment. Hence, this document merely introduces these concepts rather than providing a comprehensive discussion.
Microassembly using chemical reactions has the advantages of simple operation and high yields. Yin et al. [117] placed colloidal particles to imitate the morphology of oral cells in a 1-butanol aqueous solution and prepared a group of microrobots modified with silica shell catalase through a single-step self-assembly process. Although this one-step self-assembly method is efficient, the morphology and size of the products may vary, posing challenges for high-precision tasks. The high yield of a single preparation is advantageous for scaled-up production. Specific filters can be incorporated for product screening if necessary. Kang et al. [118] used a one-step self-assembly strategy to fabricate a dodecahedral self-propelled nanorobot using a zinc acetate aqueous solution and a catalase solution. These biodegradable nanorobots show promise for in vivo delivery.
Electro-adhesives are well-established and efficient micromanipulation strategies that require minimal surface modification and avoid the mechanical or thermal damage caused by rigid grippers or optical methods. This offers advantages in the manipulation of fragile objects. The dry adhesion strategy based on sparse dielectric-coated carbon nanotubes developed by Kim et al. [119] is a representative example. This loose adhesion interface can be changed from the off state to the on state by adjusting the voltage levels. Low adhesion in the off-state was achieved by reducing the effective contact area. This adjustable non-destructive contact gripper is useful for assembling sensitive components. Subsequent studies optimized the same structural design to achieve adhesion of macroscopic objects, reaching an on/off adhesion ratio of 700, demonstrating the potential for practical applications [120]. The optimized dielectric-coated carbon nanotube-based adhesive was compared with the passive dry adhesion strategy and an electrostatic chuck, demonstrating its superiority over traditional methods. Similarly, Wei et al. [121] developed a low-power electro-adhesive platform for macroscopic applications using a special organic polymer film. The advantages of the lightweight and flat geometry in this study make it easy to integrate into other systems such as a robotic hand.
3. Information sensing and control strategies
The previous section focused primarily on the actuation methods for microassembly. Effective information sensing and control is crucial for accurate task completion. This section introduces the common sensing and control strategies for microassembly systems.
As assembly units become miniaturized and their structures become more sophisticated, there is an increasing demand for accuracy, robustness, and real-time control. For 2D linear or 3D spatial arrangements of microparticles using MFIA or ultrasound DSA, control involves adjusting the quantity, angle, strength, and type of applied magnetic/acoustic fields. This was done to trigger aggregation, focusing on the collective behavior of swarm particles rather than on the posture and position of individual particles. The control of swarm particles does not require navigation of specific individual particles within the assembly process. Therefore, this section primarily discusses the precise control of larger single particles or functional microrobots without addressing the collective motion of clustered particles.
3.1. Information perception
A microassembly system perceives the environment through different means of information perception, which is crucial for precise control. The most common and mature technologies are image-acquisition based on computer vision, and contact-detection based on force feedback.
3.1.1. Vision-based information perception
Vision-based control systems represent a versatile strategy for sensing the environment and gathering important feedback information, particularly for perceiving the position and shape of target objects. Fig. 1 presents common observation methods for objects at different scales, from macro to micro/nano.
Manual micromanipulation and microassembly can be performed with the naked eye, however, achieving automated assembly at the micro/nano-scale typically requires the aid of dedicated vision-based information. Therefore, vision is crucial for manual operations, human–machine collaborations, and fully automated microassembly processes. Most of the studies mentioned above were based on visual feedback, where the microassembly systems were automated using the collected visual information or, at the very least, the assembly process was monitored and supervised using cameras or objectives.
A common vision system comprises a digital camera or an optical microscope with an optical lens. The relative poses between multiple targets or the pose of a single target can be determined using dedicated models for object detection or tracking. To acquire multi-DoF information, multiple cameras and microscopes are deployed to capture data from different viewpoints. For example, Zhang et al. [122] collected image information using four cameras to perform pose registration of two assembly objects with a size difference of more than ten times and determined the insertion depth using a local deformation detection method. Excellent experimental results have been achieved with attitude alignment control errors of less than ±0.3°, position alignment control errors, and insertion depth control errors of less than ±5 µm, which is very impressive for assembling objects with a size of hundreds of micrometers. Wang et al. [123] reconstructed the depth information of microparts using microscopic visual tomography images. In the depth information and 3D topography of the micropart surface were obtained based on the pixel position with the maximum focus value in a series of images. For tasks that require the rigorous pose estimation of a single target, it is common and mature to register real-time images using a CAD model that can be used as the ground truth. For example, in the binocular microscopy operating system built by Xie et al. [124] the strategy of confirming the part attitude was achieved through the mixed detection of point and line features, leveraging prior knowledge of the CAD model to determine the geometric location of feature points.
The semiautomatic human–computer interaction microassembly task operating system proposed by Wei et al. [125] serves as a comprehensive example. This system has dual manipulators for 9 DoFs movement and three microscope cameras with different viewing angles. Thus, it can be used to provide visual feedback and inverse kinematics for position assembly objects. Before beginning the assembly task, the operator compared the captured visual information with that of the CAD model. This type of pre-quality inspection avoids wasting resources on defective products. The full assembly process was semi-automated and aided by visual servoing. Although the visual servoing can automatically handle positioning tasks, some connection operations still require manual completion. This is a good example of a semiautomatic operating system for microassemblies. It can take over tedious, high-precision tasks such as displacement, allowing the operator to handle parts that depend on experience. Autofocus cameras are of practical significance in many microassembly tasks, as developed by Ref. [126].
Neural network control systems have been designed to guarantee closed-loop performance with small tracking errors and bounded controls. The assembled devices with neural network control systems were operated jointly by human–machine interaction (i.e., except for fully manual micromanipulation). Owing to the recent popularity of deep learning algorithms, deep neural networks have gained increasing popularity in microassembly, showing good capabilities for micro object recognition, segmentation, detection, and tracking from visual inputs. For instance, Bolya et al. [127] achieved automatic identification of targets in the microassembly process using the YOLACT instance segmentation algorithm. The system strikes a balance between speed and accuracy for microassembly detection and recognition, achieving a recognition rate of 90% [128]. Li et al. [129] used an optimized YOLOv4 network to establish a new small-object detection algorithm that increased the calculation rate of network detection and accuracy for small objects. Moreover, the assembly of a fast-axis collimator lens is crucial for high-power diode laser systems. Khachikyan et al. [130] developed an active placement method without closed-loop control by training multiple convolutional neural networks, thereby significantly reducing the time consumption of active assembly.
In general, the combination of vision and deep neural networks helps reduce manual labor while maintaining the accuracy and efficiency of microassembly. Current studies using deep neural networks have attempted to solve novel assembly tasks involving many objects. However, further work is required to address complex assembly processes.
3.1.2. Force-based information perception
Detection and control of the operating force during the assembly process are of paramount importance. It not only protects fragile materials, especially biological ones, but also improves the operating efficiency and facilitates automated assembly [131]. Integrator force sensors for components with basic functionalities are an intuitive and effective strategy. For instance, Wei and Xu [132] integrated a commercial piezoresistive strain gauge force sensor into a robotic cell microinjector that could detect force with a resolution of 0.65 mN. This design is of great significance for improving the success rate of injections and ensuring the survival rate of cells after surgery [132] (Fig. 8(a)). The head-replaceable microgripper developed by Gursky et al. [133] based on SU-8 material, uses a similar strategy. After integrating optically readable deformation-based force sensors, safe manipulation of cells in liquid environments becomes feasible. They conducted sensitivity measurements through image sampling and obtained stable results of 3.337 ± 0.088 µm∙mN−1.
In addition to using force sensors, the design of force sensors for the functional component itself remains challenging and has attracted increasing attention. The 3D flexible holder at the fiber tip developed by Power et al. [134] was a groundbreaking study. The holder is only 100 µm in size and is manufactured in a single step using two-photon polymerization technology (Fig. 8(b)). Four equidistant polymer plates connected by springs act as optical interferometry force sensors. The change in the reflected light was caused by the relative displacement of the polymer plate due to the deformation of the gripper, and this change was correlated with the clamping force through a well-designed neural network. This passively actuated gripper achieved excellent results in experiments with both biological and nonbiological subjects. The force information feedback of the adaptive two-finger gripper developed by Xu et al. [135] was determined by detecting the deformation of the functional gripper. The gripper fingers comprised rigid nodes combined with flexible structures, and their deformation degrees and forces exhibited a good linear relationship. Although the force perception error at both ends of the finger was only 8%, the perception error in the middle part of the finger reached 3%. The gripper developed by Xie et al. [136] was integrated with a two-force information-sensing structure (Fig. 8(c)), achieving a clamping force resolution of 38 pN and a grasping force resolution of 182 pN. A microcantilever wrist-force sensor consisting of a laser and a position-sensitive detector was used to measure the control force. It detected the force through the deformation of the cantilever. The clamping force was precisely controlled by adjusting the direction of the magnetic moment of the beads at the back of the cantilever using a magnetic coil. Such a high resolution provides important potential applications in microscale manipulation and characterization.
3.2. Control strategy
In the previous sections, we discussed the actuation and sensing techniques used in the microassembly process and emphasized the essential role of the control strategy in the entire procedure. The diverse motion modes, fabrication materials, and structures of microrobots have led to significant variations in their control models. Unlike macroscopic-level closed-loop control, objects, or robots in the microassembly process are more prone to external environmental influences such as disturbances from liquids or gases and vibrations, as well as system uncertainties. Consequently, the study of advanced control strategies is crucial to ensure the stable performance of these objects in motion at the micro/nano-scale.
3.2.1. Linear control
Proportional–integral–derivative (PID) controllers are widely used and do not require precise control models for systems; their control parameters are easily tuned, leading to quick responses. They are applied to a wide range of control scenarios, and this also holds true for the control of microrobots. Chung et al. [137] used a vision-based PID controller to govern the assembly operations of a magnetic microgripper. By modulating the gradient magnetic field, they achieved precise 3D levitation of the microgripper in liquid, with a positioning accuracy of approximately 200 µm [137]. Li et al. [138] achieved high-precision closed-loop steering of microrobots within a zebrafish body. Yang et al. [139] used ultrasound-guided visual servoing to automatically transport magnetic wires to a target area. Sun et al. [140] proposed a PID control paradigm based on microbubble activity and ultrasound field driving that enabled drug delivery through the blood-brain barrier. Wei et al. [141] guided microrobots loaded with therapeutic cells precisely into blood vessels and tissues through photoacoustic imaging techniques.
3.2.2. Nonlinear control
Microrobots are inherently highly nonlinear and uncertain owing to size effects. Unmodelled dynamics, Brownian motion [142], [143], [144], external fluid forces [145], [146], and inaccuracies in microassembly systems [147] can undermine the stable control of microrobots. Traditional linear controllers such as PID controllers have difficulty completing certain tasks with high-precision requirements; therefore, advanced nonlinear control strategies are further explored. To enable microrobots to adapt to the non-Newtonian behavior of blood and environmental disturbances without knowledge of blood flow velocity distribution, Meng et al. [146] proposed a navigation controller that combines sliding mode control, backstepping control, and disturbance compensation. Simultaneously, they designed a high-gain extended state observer to estimate and suppress uncertainties in the model parameters and environmental disturbances. On the other hand, Arcese et al. [148] used an adaptive backstepping control method to synthesize Lyapunov stability control laws for a nonlinear model. They combined this approach with a high-gain observer to reconstruct unmeasured velocities required by the controller. Ma et al. [149] proposed a robust controller based on input-to-state stability (ISS). Simultaneously, this controller is equipped with a nonlinear high-gain observer to estimate the velocity of microrobots, achieving 3D control with a position error of approximately 8.5 µm and a velocity error of approximately 0.6 µm∙s−1. Liu et al. [150] developed a proxy-based sliding-mode control approach to design stable controllers based on an error model in the Frenet–Serret framework. This approach achieved submicrometer precision in the 3D motion of magnetically driven helical micro-swimmers.
3.2.3. Optimal control
Research is also being conducted to improve the precision of microrobots through optimal control. Optimal control is the system’s ability to optimize a microrobot’s movement, posture, or task execution at minimal cost. Yang et al. [151] introduced a dual-loop hybrid controller consisting of a fuzzy logic modifier and a model predictive controller. This approach achieved a generalized disturbance estimation and compensation. This hybrid strategy enabled precise trajectory tracking and flexible motion adjustments for microrobots. Xu et al. [152] used a linear quadratic regulator to control the swimming direction and trajectory of spiral microswimmers. Additionally, their study integrated a radial basis function network trained using a backpropagation algorithm to compensate for angle errors induced by gravitational disturbances and boundary effects. Zhang et al. [153] introduced a minimum-variance controller to stabilize magnetically propelled microscopic beads and control their Brownian motion. Belharet et al. [154] developed a generalized predictive controller that exhibited sufficient robustness in tracking the motion of microcapsules in a low-Reynolds-number liquid environment. It effectively handles uncertainties in nonlinear models such as drag force and viscosity, external perturbations such as systolic pulsatile flow, and noisy trajectory tracking measurements.
4. Practical applications
In this section, we discuss the recent progress in microassembly in several different application scenarios, including microstructure construction, MEMS assembly, and bioengineering.
4.1. Microstructure construction
The primary goal of microassembly is to build intricate microstructures for use in micro-devices. To achieve this, several different microstructures have been assembled with particles and microstructural units using techniques such as MFIA and ultrasound DSA. The resulting assemblies can form different geometries including 2D/3D structures, simple lines, arrays, and complex chiral helices [155], [156], [157]. Current microassembly studies primarily focus on miniaturization, precision, automation, and usage in biological environments. Technological breakthroughs have been achieved in the microassembly of 2D/3D structures. Examples of structural construction include magnetically induced programmable 2D assembly with a magnetically tunable quadrupole module [24] and self-assembled bumps and rings using optical tweezers [56]. Both studies solved the problem of sophisticated structures with building blocks through the specific manipulation of external fields. For example, 3D micro granular crystals were constructed through the desired assembly of 4.86 µm diameter silica spheres with optical tweezers. Optical tweezers were used to simultaneously arrange these microparticles in patterns via two-photon polymerization [158] (Fig. 9(a)). In addition to programmable geometries, the fabrication of structures using microassemblies enables the use of complex functional materials with versatile compositional features. For example, the construction of complex materials with tunable structural, morphological, and chemical features has been demonstrated through coded magnetic-induced microassembly [34]. This strategy facilitates remote 2D and 3D manipulation in arbitrary microfluidic environments, which could potentially code microstructures for tissue-engineered constructs. To prepare biocompatible and flexible structures, Barbot et al. [32] demonstrated microassembly using a magnetic microrobot for fragile soft hydrogel structures at the liquid–liquid interface. Such microrobot-assisted methods have shown significant promise, including the fabrication of medical structures in vivo, 3D assembly of different materials, and development of biological structures in microfluidics.
Moreover, insulated wires were constructed using the ultrasound DSA of the light-cured materials. The primary process in these applications includes the directional arrangement of conductive fibers in a photopolymer matrix using ultrasound DSA, followed by light curing to produce composite materials with conductive capabilities [86]. These light-cured materials can be used for the 3D printing of insulated wires. The properties of the assembled structures, such as the conductivity and strength of these conductive composite materials, can be quantitatively controlled by changing the properties of the carbon microfibers and the shape and strength of the ultrasonic field [85]. These investigations have facilitated the development of materials tailored to specific needs. The future holds promise for complex structural designs within a matrix achieved through directly shaped 3D printing, especially if DSA is performed layer-by-layer during the printing process [84].
4.2. MEMS operations
Recently, the requirements for assembling functional MEMS devices, including miniature actuators, sensors, and optical devices, have increased. Accuracy, efficiency, and automation are important for the assembly of MEMS devices. The target objects in MEMS assembly tasks are larger, heavier, and stiffer than those in microstructural construction tasks. Considering the aforementioned characteristics, mechanical microgrippers are optimal for MEMS assemblies. To improve accuracy and efficiency, a closed-loop control system for electrical signals was implemented. This multi-DoF macromechanical platform satisfies the requirements for flexible multi-directional and multiangle tasks. The macro-control system oversees the assembly automation by managing the transmission of force and visual feedback. In addition, the clamping force supplied by the macro-platform is sufficient for several different operations such as clamping, moving, pulling, and inserting during assembly tasks.
To assemble MEMS devices with complex structures, additional techniques have been integrated into microassemblies. Among several techniques, micromasonry has been used to fabricate structures as retrievable components, such as combs and suspended flexural beams. One example of the use of the micromasonry technique is the assembly of MEMS devices comprising two combs (top and bottom), two spacers, and two gold pads. The resulting MEMS devices exhibited optimal capability in sensing and actuation, the displacement-sensing resolution reached 0.17 µm and a force-sensing resolution of 6.63 µN at a sampling frequency of 10 Hz [159]. Das et al. [160] reported the assembly of biomicrorobotic devices with individual 2.5D MEMS components using a 6 DoFs semi-automatic system. The assembled miniature biomicrorobot yielded an accuracy of 0.5 ± 0.2 µN (Fig. 9(b)).
In addition, owing to the extremely high requirements of MEMS for displacement accuracy and system stability, most related studies prefer to use mature commercial microoperating systems, as introduced in Section 2.5. Among these, IMINA (Imina Technologies Inc., Switzerland) is a commercial micromotion robot that uses a unique piezoelectric drive mechanism. It has a high resolution (0.02 nm in the x- and y-directions and 0.1 nm in the z-direction) and a small size (20.5 mm × 22 mm × 12.5 mm), which makes it suitable for a wide range of applications. This is suitable for MEMS-related work. Device integration can be used for in situ nanoprobing and electrical failure analysis solutions for SEM, focused ion beams, dual beams, or simply integrated into optical microscopes. Functionally, based on its high precision and excellent stability, it can measure the electrical characteristics of semiconductor devices at the smallest technology nodes, locate defects in contact layers and metal lines, conduct failure analyses, improve the reliability of semiconductor chips, separate single particles, perform assembly, or prepare samples for further study. Nieminen et al. [161] used SEM to examine the in-plane motion of an MEMS, where an IMINA robot was used to carry a probe and was placed in the SEM as an end effector that conducted electrical detection signals. Similarly, Mosberg et al. [162] research on position-controllable patterned nanowire growth, an IMINA robot was used to carry probes as carriers for precise contact with the nanowires. In these studies, the focus of high-precision microscopic robot systems was to carry functional ends for displacement. Although their roles are straightforward, they are essential.
Despite the recent advances in microassembly technologies, challenges still need to be overcome. Automation is essential for the accuracy and efficiency of MEMS assembly tasks. Furthermore, because of the large size of the assembly targets in MEMS, structural fixation remains constant. This necessitates the use of vision in image-based automated-assembly systems. To assess clamp tightness, force sensitivity is crucial for the feedback control of the force signal. Advanced force sensor technology is a promising direction for practical applications.
4.3. Biomedical applications
Microassemblies have great potential in biomedical applications [163], [164]. Tissue engineering has been introduced to build artificial functional tissue substitutes through the arrangement of specific cells and organic materials. Among the different technologies available, microassembly is of significant interest. This approach is a robust and highly scalable method for engineering 3D tissues by assembling micromodular biological entities [165], [166], [167], [168], [169], [170] (Fig. 9(c)). This section outlines the typical engineered tissues using microassembly techniques, including cell sheets, blood vessels, and other 3D tissues.
The arrangement of specific cells in on-demand structures is important in biological research and clinical medicine. The assembly into 2D cell sheets was achieved through cell levitation using optical tweezers. The optical tweezers using an emission power of 0.75 W offer trapping force to drive 15 human embryonic stem cells into assembly in pairs with a velocity of 30 µm∙s−1[64]. Furthermore, to fabricate a 3D structure of cells, a multiphase liquid–liquid system-assisted assembly was introduced for the shape-controllable 3D assembly of cells [171]. Multiple liquid–liquid systems were used to minimize the surface area and surface energy between the phases. Using this system, the shape-controlled microgels contained cells in programmed geometric configurations. It is important to modify cells while retaining their intrinsic biological functions and provide cell robots with the ability to self-propel in a bio-friendly manner. As reported by Tang et al. [172] uneven decomposition of urea fuel serves as a means of propulsion through asymmetric urease modification on the surface of native platelet cells. This artificial modification based on natural cells greatly increases the delivery efficiency.
Wang et al. [173] conducted pioneering work on microassembly with significant implications for artificial blood vessels. They constructed cell-embedded vascular-like microtubes by assembling the cell-embedded structures. To create a cellular vascular-like microchannel, they first developed a high-DoF multiple manipulator system that moved concentrically along a track [173]. With a high operating resolution of 30 nm, the system achieved a bottom-up assembly of micro components with a diameter of 200 µm through the coordinated operation of multiple manipulators. The assembled vascular-like microtubes can be engineered with predetermined lengths and diameters according to their applications. After successfully introducing a multilayer microfluidic device, a fluidic self-assembly method for cell-embedded microstructure assembly was established [174] and a repeatable, high success rate of automated assembly was achieved [175]. They also proposed a bubble-assisted fluidic assembly strategy [176] by bubbling microbubbles to excite the microflow used to push the microstructure upward. Automated high-speed assembly was achieved with a 100% success rate using visual servo control.
Biomedical engineering (BME) related fields are important application scenarios for the microchannels. Cui et al. [177] assembled micro gears made of hydrogels, stem cells, and fibroblasts into a lobule-like structure like natural liver through the coordinated operation of double manipulators. Through experiments, they verified that the in vitro replication of the human liver has a relatively complete function. In addition, they successfully endowed the microstructure with magnetic properties by adding magnetic particles, realizing automatic assembly, and tracking the magnetic microstructures [178]. In addition, in other related experiments, it was proven that micro rings made of calcium alginate microfibers doped with magnetic particles can induce the direction of cell proliferation [179]. Takeuchi et al. [179] outlined a successful research approach evolving from system construction to system intelligence/automation, expanding microstructure functionality, and realizing BME-related applications. Furthermore, it is anticipated that researchers will be able to design and build more complex microstructures by exploring a wide range of functional materials.
Combining BME with microassembly has significant potential for organ-on-chip applications. Recently, microrobots and micromaterials have led to many fascinating breakthroughs in the field of BME. Escherichia coli with integrated magnetic nanoparticles can navigate biological matrices and release drugs on demand [180]. Microrobots are sufficiently resilient to withstand blood flow, enabling them to locate thrombi and accelerate clot dissolution [181]. This millimeter-scale microrobot, which integrates magnetic control and ultrasonic imaging guidance, achieves autonomous navigation in a simulated blood-flow (pulsation) environment. Moreover, the thrombus removal efficiency was significantly improved owing to the shear force generated by the movement of the microrobot. Song et al. [182] developed a puffball-shaped robot for drug delivery, which mimics the structure and function of the spherical fruiting body of the puffball fungus. Its infrared-responsive sealing layer plays a role similar to that of the barrier cap of the puffball fungus in protecting transported drugs. Such a bionic design allows the microrobot to minimize drug leakage during transportation while ensuring efficient release when exposed to near-infrared light, indicating a certain level of autonomy. Cong et al. [183] developed a microrobot based on cells infected with an oncolytic adenovirus that specifically binds to bladder cancer cells. This cellular microrobot can replicate within the host after binding to bladder cancer cells and is released after the host is lysed to ensure sustained tumor lysis. This microrobot was modified with magnetic nanoparticles, giving it the ability to perform noncontact directional manipulation in different media. Four dimensional (4D) bio-inks based on hydrogels and living cells support the 4D printing of complex biological structures with high cell viability [184]. Although not all are linked to microassembly processes or products, these technologies and research still offer valuable insights. Microassembly technologies, particularly those with potential biological applications, are continually evolving. Significant advancements are expected in this field in the future. Special attention should be paid to clinical BME applications such as intraluminal delivery and imaging endoscope-assisted magnetic navigation systems developed by Wang et al. [185]. This study developed a flexible microrobot with good biocompatibility that can achieve controlled movements, such as diffusion and anchoring in the body. Under the visual guidance of the endoscope, targeted delivery was performed with high efficiency, low trauma, and high precision on a whole-body scale. This complete development from components to clinical systems is of great significance for the expansion of micro–nano technology.
Research on brain–computer interfaces, especially the collection and analysis of brain signals, is important for neuroscience research. Recently, signal collection has been achieved by implanting neural electrodes in the cerebral cortex. Generally, this kind of flexible neural electrode has a high aspect ratio, is extremely thin (< 5 µm), and presents a waterweed-like geometric structure [186]. While this geometric structure improves the safety of invasive electrodes, it also gives rise to hidden dangers of being fragile and easily deformed, and it can have difficulty penetrating the cerebral cortex independently. At the same time, major blood vessels must be strictly avoided during electrode implantation, and electrode arrays must be implanted with high precision at certain intervals in the target brain area. Under these requirements, the “shuttle-assisted implantation” strategy using guide needle assistance has become popular [187], as used by Musk and Neuralink [188]. Other researchers adopted a similar strategy, developing automated assembly systems in air rings that use guide pins to pre-hook electrodes and autonomously complete this task, which would otherwise be a struggle for human operators through microscopic vision inspection [189]. Generally, the assembly and implantation procedures of flexible brain electrodes are suitable for significant application scenarios owing to their task complexity, clinical importance, and necessity.
Microassemblies play important roles in biosensing. Similarly, biosensors can be driven by magnetic, optical, acoustic, or other external fields to specifically detect targets. For example, Su et al. [190], implanted magnetic nanowires as potential magnetic tags into canine osteosarcoma cells to detect and screen through giant magnetoresistance. The use of magnetic labels avoids the risk of contamination during sample preparation, and magnetic nanowires produce satisfactory signal intensity, which has application prospects for detection in low-concentration environments. In contrast, based on the ability of phages to specifically capture Salmonella, Huang et al. [191] modified them with magnetic beads and developed a detection sensor with high specificity and high recovery rate with no false positive results. The researchers also used bio-orthogonal click chemistry to enhance the sensor signal and achieve high sensitivity. This sensor achieved high recovery rates and low standard deviations in orange juice samples, demonstrating its potential for practical applications.
Optical fibers have the advantages of being low-cost, flexible, and supporting remote light injection for photoelectrochemical sensors. Pal et al. [192] prepared a sensor for detecting aflatoxin B1 (AFB1) in liquid environments by in situ synthesis and deposition of polyaniline on the surface of plastic optical fibers. The sensor achieved specificity by integrating AFB1 antibodies and achieved a detection sensitivity far higher than any official requirement for beer, urine, and other environments. Similarly, Wen et al. [193] integrated indium tin oxide, zinc oxide, and copper oxide onto an optical fiber surface with the coating layer removed to prepare a cysteine detection sensor for urine samples. This device demonstrates the advantages of optical fibers over large-volume macroscopic space light and achieves high sensitivity and specificity. SAW sensors are mature and are widely used. They respond to changes in the liquid environmental parameters through changes in the resonator frequency and usually have high sensitivity. A typical example is the SAW-based Legionella pneumophila detector developed by Gagliardi et al. [194] for water environments. The SAW laboratory-on-a-chip combined with a microfluidic channel showed excellent specificity for complex mixed bacterial suspensions. The object-sensing platform based on metal micropillar array electrodes developed by Chen et al. [195] for hydrogen peroxide and sarcosine detection applies acoustic technologies with different principles. Pumping the solution forms bubbles, which produce a microflow under the action of a sine wave signal. The stirring effect generated by this microflow significantly improved the contact between the electrode surface and sample, directly improving the detection efficiency.
5. Discussion and conclusions
In this review, recent advances in microassembly are highlighted and discussed, including its fundamentals, practical applications, and current developments. Over the past few decades, the development of different technologies for generating programmable external fields has led to new microassembly strategies involving magnetic, optical, and acoustic fields. Mechanical grippers and micromanipulation systems with high-motion precision have expanded the potential applications of microassemblies. Increasing the automation level of microassembly requires the continuous development of dedicated control algorithms with visual and force feedback. There has also been emphasis on the use of deep neural networks for robust and precise object detection, recognition, and tracking, thereby improving the efficiency and accuracy of microassemblies. Despite these promising studies, significant technological advancements are still necessary for the practical applications and high-throughput deployment of microassemblies. Below, we highlight the key challenges (Table 4) to be addressed.
5.1. Simplified and practical microassembly system
Because system simplicity and motion accuracy optimization are fundamental requirements for both macroscopic and microscopic systems, it is desirable to solve complex microassembly problems using simplified systems and straightforward operations.
Recently, researchers have focused on improving the versatility, stability, and intelligence of microassembly systems. However, enhanced system capabilities often result in large volumes, complex equipment integration, and stringent environmental requirements. Indeed, some exploratory research is needed to refine the individual design of microscopic particles and provide microsystems with special responses and intelligence through novel materials and structural designs. However, by stripping away the fancy packaging and integration while retaining only the essential and targeted functions, the system can be simplified and made compact. This transition is critical and urgent for microassembly and micromanipulation from laboratory to clinical scenarios or in vivo applications.
Researchers have often mentioned the “potential” of research in the field of BME and verified the functions of microrobots/microsystems through semi-open models. When attempting to apply laboratory-developed technologies to clinical settings, several incompatibilities are likely to arise, including space limitations, size mismatches, interference from complex environments, and challenges in human–computer interactions. These issues focus less on scientific exploration and more on engineering design. However, the trade-offs in miniaturized design and components are obstacles we must overcome as we transition from “potential” to “capability.”
Therefore, innovative processing technologies for small devices are required to solve this problem. For instance, two-photon polymerization technology integrated with microscopic 3D printing is important for microscale device processing and mass production. However, further attention needs to be paid to the process efficiency, repeatability of multi-material processing, and product lifespan.
5.2. Microassembly beyond typical micro-objects
Most previous studies have focused on microscopic objects whose scale in each dimension is approximately uniform. These objects are typically particles or micro-units resembling cubes or spheres that are often assembled in chains or stacks. However, objects other than these typically structured micro-objects have completely different problems and limitations. For example, the flexible neural electrode mentioned earlier has an undisputed microscopic size in thickness and width (only tens or even a few micrometers) but may have an aspect ratio of hundreds of times, that is, a centimeter-level length. This contradiction in size is manifested as “softness” on a macro scale and as “non-freedom with one end restricted” on a micro scale. This makes neural electrodes self-limiting in terms of their features and operational methods. At the microscopic level, the small features and the need for extremely high positioning accuracy must be balanced with the morphological instability and overall travel requirements at the macroscopic level. This is not the case when considering their fragile structure, which presents further safety challenges. Additionally, numerous engineering issues must be considered when designing a microassembly system for this atypical object. These include the limitation of trajectory and collision avoidance owing to the macro size of the length and the entanglement problem between the electrode wires after being grouped.
To solve the assembly or manipulation problems of such atypical micro-objects, the current method uses a high-precision microscopic vision robot system, which, although complex and expensive, can solve all the problems perfectly. Another method is to trigger the self-assembly of electrode bundles through physical interactions between solid and liquid media [196]. This method is low-cost and highly efficient, but the operation is difficult to describe as “accurate.” One possibility is to deal with macro and micro problems separately. First, we reduce the dimensionality of the problem at the macro level to a stable attitude and then use two sets of operations to solve the macro- and micro-positioning requirements. It may also be feasible to give the electrode 3D features through some micromachining methods, which are beneficial for inducing self-assembly in some small areas or reinforcing the connection part; however, the specific situation needs to be analyzed on a case-by-case basis.
5.3. Accurate information feedback in complicated closed environment
Information perception in the current microassembly tasks relies heavily on visual feedback, and some of the force feedback components mentioned above rely on the deformation mechanism of the visually collected sensing part. Unfortunately, microassembly systems with external macro-components can successfully provide feedback information but are unwieldly for use in closed environments. Separate small microrobots are unable to provide feedback information other than visual position feedback. Particularly, in closed and susceptible environments within the body, the requirement for accurate information feedback becomes increasingly difficult. The reasons for this are similar to those of minimally invasive surgery, which mainly include a limited field of view, narrow operating environment, delicate and sensitive environmental composition, and occasional inconsistencies between movement patterns and human behavioral habits. Non-open human tissues or instruments may obstruct external fields, which must be considered. In addition to the obstruction of the tissue itself, blood and other body fluids can also cause great interference with vision. The accurate realization of functions according to the requirements of such a harsh environment is an open question to be solved for in vivo applications. Control methods and systems developed to address these restrictions may be applicable. Therefore, improved general systems are required for practical applications and industrial production.
Fiber Bragg grating sensors provide a possible solution to the difficulty of providing information feedback in closed environments [197]. This optical sensor is immune to electromagnetic interference, can be easily modified to cater to different measurement objects, and does not rely on visual information. Moreover, its tubular structure is well suited for in vivo environments through minimally invasive openings. Moreover, miniaturized high-resolution vision systems based on thin-diameter (1–2 mm) endoscopes have emerged, and a binocular design can help perform depth estimation in a limited depth-of-field. Both strategies are potential solutions for mechanical microoperation systems. In addition, some magnetic resonance imaging (MRI)-compatible robots and intraoperative imaging technologies can achieve real-time feedback and control using external fields [198]. The development of MRI-compatible robots combined with micromanipulation and swarming microrobots is an emerging trend in combining imaging, diagnosis, and therapy as a one-stop shop technique in medicine [199].
5.4. Increasing automation levels of microassembly systems
Intelligent microassembly systems are necessary to realize automatic operations that are separate from manual operations. Automation is crucial for improving the efficiency of operations, eliminating manual manipulation, and performing intense tasks that cannot be performed by humans such as automatic screening and sorting of large numbers of objects [200], [201]. Assembly should be a complete flow process. When discussing its efficiency and automation, the whole process should be covered, from starting with the placement of the operating object until the system is shut down. Most current research focuses on the assembly process, which is the most important part of the task. However, the pre-deployment, reloading, and decoupling preservation processes in the mission pipeline should be considered to increase the automation level of the entire system. The overall efficiency, success rate, and repeatability are more significant at the application level.
The autonomous navigation of the control system is required for obstacle avoidance and accurate targeting. Unlike controlled, stable, and uncluttered experimental settings in a laboratory, the environment in a live system is significantly more complex. In contrast to the multiangle high-resolution camera that can be set up in the experimental system, a lower-resolution endoscope may be required to access the target site, presenting challenges in terms of both spatial resolution and the field of view covered. System robustness and overall efficiency are equally important, particularly for disturbances outside the laboratory environment. For in vivo applications, blood and diseased tissues can pose unpredictable obstacles to visual feature recognition and path planning for robotic movements. For this problem, both reinforcement learning and learning from demonstrations have significant exploratory value. Reinforcement learning allows for targeted solutions in preset situations, whereas learning from demonstrations enables the robot to derive solution ideas and key strategies from expert demonstrations.
5.5. Safety concerns in microassembly applications
Discussions on the safety of microassemblies should be conducted in two directions: one is the quantitative evaluation of the results and products, and the other is the consideration of receptor safety when applied in biological environments [201]. The main quantitative evaluation indicators of microassembly are the repetition accuracy, time efficiency, and task success rate, which are self-described and defined by researchers. Such an evaluation is highly subjective, has personal biases, and has difficulty in demonstrating universality. A set of universal objective quantitative assessment methods applied at the microscale is required, such as the completion of some recognized popular tasks.
Safety concerns regarding the materials of the units and methods to dispose of nondegradable residues injected into the body should also be investigated. Currently, there are mature regulatory and approval processes, and every technology used in biological experiments and clinical trials can be trusted to comply with scientific and social ethics. The degradability and biocompatibility of these materials have long been discussed. Biofriendly carriers, such as methacrylate gelatin hydrogels, have been widely used in BME, particularly in the preparation of microrobots for drug delivery [202]. Choosing an appropriate medium and carrier was not an issue. Instead, we need to focus on key concerns such as the acceptability of MM residue and methods for its removal.
5.6. Outlook
We anticipate that further research progress and translation of emerging microassembly techniques will require interdisciplinary collaborative efforts among scientists in robotics, material sciences, chemistry, BME, and artificial intelligence. Although microassembly is still in its infancy, it has attracted significant attention in a diverse range of fields. With continuing innovations in related technologies, these challenges will be overcome in the future, eventually expanding the horizon of external field-driven microassembly in practical applications, particularly for in vivo micro-nano robotics.
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgments
This work was supported by Shanghai Municipal Science and Technology Major Project (2021SHZDZX) and also in part supported by the Science and Technology Commission of Shanghai Municipality (20DZ2220400).
RivinEI.Mechanical design of robots. McGraw-Hill, New York city(1987)
[2]
ZeschW, BrunnerM, WeberA.Vacuum tool for handling microobjects with a nanorobot.In: Proceedings of the International Conference on Robotics and Automation; 1997 Apr 25–25; Albuquerque, NM, USA. Piscataway: IEEE; 1997. p. 1761–6.
[3]
SittiM.Microscale and nanoscale robotics systems.IEEE Robot Autom Mag2007; 14(1):53-60.
[4]
AraiF, AndoD, FukudaT, NonodaY, OotaT.Micro manipulation based on micro physics-strategy based on attractive force reduction and stress measurement.In: Proceedings of the 1995 IEEE/RSJ International Conference on Intelligent Robots and Systems. Human Robot Interaction and Cooperative Robots; 1995 Aug 5–9; Pittsburgh, PA, USA. Piscataway: IEEE; 1995. p. 236–41.
[5]
FearingRS.Survey of sticking effects for micro parts handling.In: Proceedings of the 1995 IEEE/RSJ International Conference on Intelligent Robots and Systems. Human Robot Interaction and Cooperative Robots; 1995 Aug 5–9; Pittsburgh, PA, USA. Piscataway: IEEE; 1995. p. 212–7.
[6]
MenciassiA, EisinbergA, IzzoI, DarioP.From “macro” to “micro” manipulation: models and experiments.IEEE/ASME Trans Mechatron2004; 9(2):311-320.
[7]
YuY, JiangC, ZhengXT, LiuY, GohWP, LimRHH, et al.Three-dimensional highway-like graphite flakes/carbon fiber hybrid electrode for electrochemical biosensor.Mater Today Adv2022; 14:100238.
[8]
LiT, YuS, SunB, LiY, WangX, PanY, et al.Bioinspired claw-engaged and biolubricated swimming microrobots creating active retention in blood vessels.Sci Adv2023; 9(18):eadg4501.
[9]
ZhangW, DengY, ZhaoJ, ZhangT, ZhangX, SongW, et al.Amoeba-inspired magnetic venom microrobots.Small2023; 19(23):2207360.
[10]
YuJ, WangB, DuX, WangQ, ZhangL.Ultra-extensible ribbon-like magnetic microswarm.Nat Commun2018; 9(1):3260.
[11]
ShiW, HuangJ, FangR, LiuM.Imparting functionality to the hydrogel by magnetic-field-induced nano-assembly and macro-response.ACS Appl Mater Interfaces2020; 12(5):5177-5194.
[12]
YuJ, JinD, ChanKF, WangQ, YuanK, ZhangL.Active generation and magnetic actuation of microrobotic swarms in bio-fluids.Nat Commun2019; 10(1):5631.
[13]
TanL, LiuB, GlebeU, BökerA.Magnetic field-induced assembly of superparamagnetic cobalt nanoparticles on substrates and at liquid–air interface.Langmuir2018; 34(46):13993-14002.
[14]
WenT, ZhuW, XueC, WuJ, HanQ, WangX, et al.Novel electrochemical sensing platform based on magnetic field-induced self-assembly of Fe3O4@ polyaniline nanoparticles for clinical detection of creatinine.Biosens Bioelectron2014; 56:180-185.
[15]
LiuJ, ZhouT, HanT, ZhuL, WangY, HuY, et al.Engineering a ternary one-dimensional Fe2P@SnP 0.94@MoS2 mesostructure through magnetic-field-induced self-assembly as a high-performance lithium-ion battery anode.Chem Commun2022; 58(33):5108-5111.
[16]
ParkM, KangS, NamC, NarasimhaK, LeeWB, ParkSJ.Magnetic field-induced self-assembly of conjugated block copolymers and nanoparticles at the air–water interface.ACS Appl Mater Interfaces2022; 14(6):8266-8273.
[17]
MohapatraJ, ElkinsJ, XingM, GuragainD, MishraSR, LiuJP.Magnetic-field-induced self-assembly of FeCo/CoFe2O4 core/shell nanoparticles with tunable collective magnetic properties.Nanoscale2021; 13(8):4519-4529.
[18]
JeongKJ, LeeDK, TranVT, WangC, LvJ, ParkJ, et al.Helical magnetic field-induced real-time plasmonic chirality modulation.ACS Nano2020; 14(6):7152-7160.
[19]
KhalilIS, van denBrinkF, SukasOS, MisraS.Microassembly using a cluster of paramagnetic microparticles.In: Proceedings of the 2013 IEEE International Conference on Robotics and Automation; 2013 May 6–10; Karlsruhe, Germany. Piscataway: IEEE; 2013. p. 5527–32.
[20]
XiaL, LiuR, LiuJ, ZhuX, DingA, CaoQ.Radial magnetic levitation and its application to density measurement, separation, and detection of microplastics.Anal Chem2023; 95(22):8660-8667.
[21]
DongX, SittiM.Controlling two-dimensional collective formation and cooperative behavior of magnetic microrobot swarms.Int J Robot Res2020; 39(5):617-638.
[22]
HuX, YasaIC, RenZ, GouduSR, CeylanH, HuW, et al.Magnetic soft micromachines made of linked microactuator networks.Sci Adv2021; 7(23):eabe8436.
GuH, BoehlerQ, AhmedD, NelsonBJ.Magnetic quadrupole assemblies with arbitrary shapes and magnetizations.Sci Robot2019; 4(35):eaax8977.
[25]
YangX, TanR, LuH, FukudaT, ShenY.Milli-scale cellular robots that can reconfigure morphologies and behaviors simultaneously.Nat Commun2022; 13(1):4156.
[26]
TasogluS, YuC, GungorduHI, GuvenS, VuralT, DemirciU.Guided and magnetic self-assembly of tunable magnetoceptive gels.Nat Commun2014; 5(1):4702.
[27]
KhalesiR, YousefiM, PishkenariHN, VossoughiG.Robust independent and simultaneous position control of multiple magnetic microrobots by sliding mode controller.Mechatronics2022; 84:102776.
[28]
LeeH, LeeD, JeonS.A two-dimensional manipulation method for a magnetic microrobot with a large region of interest using a triad of electromagnetic coils.Micromachines2022; 13(3):416.
[29]
Mayorga-MartinezCC, VyskoJčil, NovotnFý, BednarP, RuzekD, AlduhaishO, et al.Collective behavior of magnetic microrobots through immuno-sandwich assay: on-the-fly COVID-19 sensing.Appl Mater Today2022; 26:101337.
[30]
PachecoM, Mayorga-MartinezCC, EscarpaA, PumeraM.Micellar polymer magnetic microrobots as efficient nerve agent microcleaners.ACS Appl Mater Interfaces2022; 14(22):26128-26134.
[31]
ZhaoF, RongW, LiD, WangL, SunL.Four-dimensional design and programming of shape-memory magnetic helical micromachines.Appl Mater Today2022; 27:101422.
[32]
BarbotA, TanH, PowerM, SeichepineF, YangGZ.Floating magnetic microrobots for fiber functionalization.Sci Robot2019; 4(34):eaax8336.
[33]
AlapanY, YigitB, BekerO, DemirörsAF, SittiM.Shape-encoded dynamic assembly of mobile micromachines.Nat Mater2019; 18(11):1244-1251.
[34]
TasogluS, DillerE, GuvenS, SittiM, DemirciU.Untethered micro-robotic coding of three-dimensional material composition.Nat Commun2014; 5(1):3124.
[35]
JohnsonBV, ChowdhuryS, CappelleriDJ.Local magnetic field design and characterization for independent closed-loop control of multiple mobile microrobots.IEEE/ASME Trans Mechatron2020; 25(2):526-534.
[36]
YangX, TanR, LuH, ShenY.Magnetic-directed manipulation and assembly of fragile bioartificial architectures in the liquid–liquid interface.IEEE/ASME Trans Mechatron2022; 27(5):3590-3600.
[37]
HsuA, CowanC, ChuW, McCoyB, Wong-FoyA, PelrineR, et al.Automated 2D micro-assembly using diamagnetically levitated milli-robots.In: Proceedings of the 2017 International Conference on Manipulation, Automation and Robotics at Small Scales (MARSS); 2017 Jul 17–21; Montreal, QC, Canada. Piscataway: IEEE; 2017. p. 1–6.
[38]
YaoT, ChisholmNG, SteagerEB, StebeKJ.Directed assembly and micro-manipulation of passive particles at fluid interfaces via capillarity using a magnetic micro-robot.Appl Phys Lett2020; 116(4):043702.
[39]
JiZ, YanC, YuB, WangX, ZhouF.Multimaterials 3D printing for free assembly manufacturing of magnetic driving soft actuator.Adv Mater Interfaces2017; 4(22):1700629.
[40]
DunnCR, LeeBP, RajacharRM.Thermomagnetic-responsive self-folding microgrippers for improving minimally invasive surgical techniques and biopsies.Molecules2022; 27(16):5196.
HanK, ShieldsCW IV, DiwakarNM, BhartiB, LGP’opez, VelevOD.Sequence-encoded colloidal origami and microbot assemblies from patchy magnetic cubes.Sci Adv2017; 3(8):e1701108.
[43]
KararsizG, RogowskiLW, ZhangX, BhattacharjeeA, KimMJ.Adaptive tracking controller for an alginate artificial cell.In: Proceedings of the 2021 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS); 2021 Sep 27–Oct 1; Prague, The Czech Republic. Piscataway: IEEE; 2021. p. 7483–9.
[44]
XuL, GongD, CeliN, XuJ, ZhangD, CaiJ.Biohybrid magnetic microrobots for enhanced photocatalytic rhb degradation and E. coli inactivation under visible light irradiation.Appl Surf Sci2022; 579:152165.
[45]
FengY, FengL, DaiY, BaiX, ZhangC, ChenY, et al.A novel and controllable cell-based microrobot in real vascular network for target tumor therapy.In: Proceedings of the 2020 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS); 2020 Oct 24–2021 Jan 14; Las Vegas, NV, USA. Piscataway: IEEE; 2021. p. 2828–33.
[46]
GongD, CeliN, ZhangD, CaiJ.Magnetic biohybrid microrobot multimers based on chlorella cells for enhanced targeted drug delivery.ACS Appl Mater Interfaces2022; 14(5):6320-6330.
MelzerJE, McLeodE.Assembly of multicomponent structures from hundreds of micron-scale building blocks using optical tweezers.Microsyst Nanoeng2021; 7(1):45.
[49]
MaruoS, IkutaK, HayatoK.Light-driven mems made by high-speed two-photon microstereolithography.In: Proceedings of the Technical Digest. MEMS 2001. 14th IEEE International Conference on Micro Electro Mechanical Systems (Cat. No. 01CH37090); 2001 Jun 25; Interlaken, Switzerland. Piscataway: IEEE; 2001. p. 594–7.
[50]
deV la Asunción-Nadal, RojasD, Jurado-SaB’nchez, EscarpaA.Transition metal dichalcogenide micromotors with programmable photophoretic swarming motion.J Mater Chem A2023; 11(3):1239-1245.
[51]
Mena-GiraldoP, KaurM, MaurizioSL, MandlGA, CapobiancoJA.Janus micromotors for photophoretic motion and photon upconversion applications using a single near-infrared wavelength.ACS Appl Mater Interfaces2024; 16(3):4249-4260.
[52]
LiangX, MouF, HuangZ, ZhangJ, YouM, XuL, et al.Hierarchical microswarms with leader–follower-like structures: electrohydrodynamic self-organization and multimode collective photoresponses.Adv Funct Mater2020; 30(16):1908602.
[53]
MouF, ZhangJ, WuZ, DuS, ZhangZ, XuL, et al.Phototactic flocking of photochemical micromotors.iScience2019; 19:415-424.
[54]
CheS, ZhangJ, MouF, GuoX, KauffmanJE, SenA, et al.Light-programmable assemblies of isotropic micromotors.Research2022; 2022:9816562.
[55]
TongJ, WangD, LiuY, LuoX, JiangJ, DongB, et al.Bioinspired micro/nanomotor with visible light energy-dependent forward, reverse, reciprocating, and spinning schooling motion.Proc Natl Acad Sci USA2021; 118(42):e2104481118.
[56]
KimJD, HwangSU, LeeYG.Traceable assembly of microparts using optical tweezers.J Micromech Microeng2012; 22(10):105003.
KimJD, LeeYG.Construction and actuation of a microscopic gear assembly formed using optical tweezers.J Micromech Microeng2013; 23(6):065010.
[59]
CastelinoK, SatyanarayanaS, SittiM.Manufacturing of two and three-dimensional micro/nanostructures by integrating optical tweezers with chemical assembly.Robotica2005; 23(4):435-439.
[60]
GhoshS, DasS, PaulS, ThomasP, RoyB, MitraP, et al.In situ self-assembly and photopolymerization for hetero-phase synthesis and patterning of conducting materials using soft oxometalates in thermo-optical tweezers.J Mater Chem C2017; 5(27):6718-6728.
[61]
TangH, KishiT, YanoT.In situ assembling of glass microspheres and bonding force analysis by the ultraviolet-near-infrared dual-beam optical tweezer system.ACS Omega2021; 6(18):11869-11877.
[62]
HolmlinRE, SchiavoniM, ChenCY, SmithSP, PrentissMG, WhitesidesGM.Light-driven microfabrication: assembly of multicomponent, three-dimensional structures by using optical tweezers.Angew Chem Int Ed2000; 39(19):3503-3506.
[63]
KirkhamGR, BritchfordE, UptonT, WareJ, GibsonGM, DevaudY, et al.Precision assembly of complex cellular microenvironments using holographic optical tweezers.Sci Rep2015; 5:8577.
[64]
WangX, GouX, ChenS, YanX, SunD.Cell manipulation tool with combined microwell array and optical tweezers for cell isolation and deposition.J Micromech Microeng2013; 23(7):075006.
[65]
MirsaidovU, ScrimgeourJ, TimpW, BeckK, MirM, MatsudairaP, et al.Live cell lithography: using optical tweezers to create synthetic tissue.Lab Chip2008; 8(12):2174-2181.
[66]
PradhanS, WhitbyCP, WilliamsMA, ChenJLY, AvciE.Interfacial colloidal assembly guided by optical tweezers and tuned via surface charge.J Colloid Interface Sci2022; 621:101-109.
[67]
ZouX, ZhengQ, WuD, LeiH.Controllable cellular micromotors based on optical tweezers.Adv Funct Mater2020; 30(27):2002081.
[68]
TanakaM, KuramichiI, TsuboiY, YuyamaK.Confinement and aggregation of colloidal particles in an ionic liquid microdroplet formed by optical tweezers.Jpn J Appl Phys2022; 61(10):100901.
[69]
ShanX, WangF, WangD, WenS, ChenC, DiX, et al.Optical tweezers beyond refractive index mismatch using highly doped upconversion nanoparticles.Nat Nanotechnol2021; 16(5):531-537.
[70]
MeldeK, ChoiE, WuZ, PalagiS, QiuT, FischerP.Acoustic fabrication via the assembly and fusion of particles.Adv Mater2018; 30(3):1704507.
[71]
MeldeK, KremerH, ShiM, SenecaS, FreyC, PlatzmanI, et al.Compact holographic sound fields enable rapid one-step assembly of matter in 3D.Sci Adv2023; 9(6):eadf6182.
[72]
MaZ, HolleAW, MeldeK, QiuT, PoeppelK, KadiriVM, et al.Acoustic holographic cell patterning in a biocompatible hydrogel.Adv Mater2020; 32(4):1904181.
[73]
SunY, LuoY, XuT, ChengG, CaiH, ZhangX.Acoustic aggregation-induced separation for enhanced fluorescence detection of alzheimer’s biomarker.Talanta2021; 233:122517.
[74]
XuT, SotoF, GaoW, DongR, Garcia-GradillaV, MagañaE, et al.Reversible swarming and separation of self-propelled chemically powered nanomotors under acoustic fields.J Am Chem Soc2015; 137(6):2163-2166.
[75]
HuangY, XuT, LuoY, LiuC, GaoX, ChengZ, et al.Ultra-trace protein detection by integrating lateral flow biosensor with ultrasound enrichment.Anal Chem2021; 93(5):2996-3001.
[76]
TangT, HuangL.Mie particle assembly by a converging ultrasound field and acoustic interaction forces.Appl Acoust2021; 180:108123.
[77]
GreenhallJ, GuevaraF Vasquez, RaeymaekersB.Ultrasound directed self-assembly of user-specified patterns of nanoparticles dispersed in a fluid medium.Appl Phys Lett2016; 108(10):103103.
[78]
PrisbreyM, GreenhallJ, GuevaraF Vasquez, RaeymaekersB.Ultrasound directed self-assembly of three-dimensional user-specified patterns of particles in a fluid medium.J Appl Phys2017; 121(1):014302.
[79]
PrisbreyM, GuevaraF Vasquez, RaeymaekersB.Arranging ellipsoidal particles in three-dimensional user-specified orientations with ultrasound-directed self-assembly.Phys Rev Appl2020; 14(2):024026.
[80]
PrisbreyM, RaeymaekersB.Aligning high-aspect-ratio particles in user-specified orientations with ultrasound-directed self-assembly.Phys Rev Appl2019; 12(1):014014.
[81]
PrisbreyM, GuevaraF Vasquez, RaeymaekersB.3D ultrasound directed self-assembly of high aspect ratio particles: on the relationship between the number of transducers and their spatial arrangement.Appl Phys Lett2020; 117(11):111904.
[82]
FengL, SongB, ChenY, LiangS, DaiY, ZhouQ, et al.On-chip rotational manipulation of microbeads and oocytes using acoustic microstreaming generated by oscillating asymmetrical microstructures.Biomicrofluidics2019; 13(6):064103.
[83]
WadsworthP, NelsonI, PorterDL, RaeymaekersB, NalewaySE.Manufacturing bioinspired flexible materials using ultrasound directed self-assembly and 3D printing.Mater Des2020; 185:108243.
[84]
GreenhallJ, RaeymaekersB.3D printing macroscale engineered materials using ultrasound directed self-assembly and stereolithography.Adv Mater Technol2017; 2(9):1700122.
[85]
NiendorfK, RaeymaekersB.Combining ultrasound directed self-assembly and stereolithography to fabricate engineered polymer matrix composite materials with anisotropic electrical conductivity.Compos Part B2021; 223:109096.
[86]
NiendorfK, RaeymaekersB.Quantifying macro- and microscale alignment of carbon microfibers in polymer-matrix composite materials fabricated using ultrasound directed self-assembly and 3D-printing.Compos Part A2020; 129:105713.
[87]
WeiY, LuXL, ShenH, PengHM, YuanZY, GuoXS, et al.An acousto-microrobotic interface with vision-feedback control.Adv Mater Technol2021; 6:2100470.
[88]
LuX, ZhaoK, LiuW, YangD, ShenH, PengH, et al.A human microrobot interface based on acoustic manipulation.ACS Nano2019; 13(10):11443-11452.
[89]
SchrageM, MedanyM, AhmedD.Ultrasound microrobots with reinforcement learning.Adv Mater Technol2023; 8(10):2201702.
[90]
ReinhartG, HeinzM, StockJ, ZimmermannJ, SchilpM, ZitzmannA, et al.Non-contact handling and transportation for substrates and microassembly using ultrasound-air-film-technology.In: Proceedings of the 2011 IEEE/SEMI Advanced Semiconductor Manufacturing Conference; 2011 May 16–18; Saratoga Springs, NY, USA. Piscataway: IEEE; 2011. p. 1–6.
[91]
MohantyS, FidderRJ, MatosPM, HeunisCM, KayaM, BlankenN, et al.SonoTweezer: an acoustically powered end-effector for underwater micromanipulation.IEEE Trans Ultrason Ferroelectr Freq Control2022; 69(3):988-997.
YangY, MaT, ZhangQ, HuangJ, HuQ, LiY, et al.3D acoustic manipulation of living cells and organisms based on 2D array.IEEE Trans Biomed Eng2022; 69(7):2342-2352.
IchikawaA, SakumaS, AraiF, AkagiS.Untethered micro-robot with gripping mechanism for on-chip cell surgery utilizing outer magnetic force.In: Proceedings of the 2014 IEEE International Conference on Robotics and Automation (ICRA); 2014 May 31–Jun7; Hong Kong, China. Piscataway: IEEE; 2014. p. 3795–800.
[96]
LyuZ, XuQ, ZhuL.Design and development of a new piezoelectric-actuated biaxial compliant microgripper with long strokes.IEEE Trans Autom Sci Eng2023; 20(1):206-217.
[97]
ProbstM, HürzelerC, BorerR, NelsonBJ.A microassembly system for the flexible assembly of hybrid robotic mems devices.Int J Optomechatronics2009; 3(2):69-90.
[98]
DasTK, ShirinzadehB, Al-JodahA, GhafarianM, PinskierJ.Computational parametric analysis and experimental investigations of a compact flexure-based microgripper.Precis Eng2020; 66:363-373.
[99]
DasTK, ShirinzadehB, GhafarianM, Al-JodahA, ZhongY, SmithJ.Design, analysis and experimental investigations of a high precision flexure-based microgripper for micro/nano manipulation.Mechatronics2020; 69:102396.
[100]
ShiC, DongX, YangZ.A microgripper with a large magnification ratio and high structural stiffness based on a flexure-enabled mechanism.IEEE/ASME Trans Mechatron2021; 26(6):3076-3086.
[101]
SiG, SunL, ZhangZ, ZhangX.Theoretical thermal-mechanical modelling and experimental validation of a three-dimensional (3D) electrothermal microgripper with three fingers.Micromachines2021; 12(12):1512.
[102]
SiG, DingM, ZhangZ, ZhangX.Theoretical thermal-mechanical modelling and experimental validation of a novel 3D three-fingered electrothermal microgripper.Precis Eng2022; 77:205-219.
[103]
RamyaS, KumarSP, AravindT, SrinivasanT, RamGD, LingarajaD.Thermal inplane microgripper for handling micro-objects.In: Proceedings of the 2022 7th International Conference on Communication and Electronics Systems (ICCES); 2022 Jun 22–24; Coimbatore, India. Piscataway: IEEE; 2022. p. 210–4.
[104]
ChenBK, ZhangY, SunY.Active release of microobjects using a mems microgripper to overcome adhesion forces.J Microelectromech Syst2009; 18(3):652-659.
[105]
BeyelerF, BellDJ, NelsonBJ, SunY, NeildA, ObertiS, et al.Design of a micro-gripper and an ultrasonic manipulator for handling micron sized objects.In: Proceedings of the 2006 IEEE/RSJ International Conference on Intelligent Robots and Systems; 2006 Oct 9–15; Beijing, China. Piscataway: IEEE; 2007. p. 772–7.
[106]
MaY, DuK, ZhouD, ZhangJ, LiuX, XuD.Automatic precision robot assembly system with microscopic vision and force sensor.Int J Adv Robot Syst2019; 16(3):1729881419851619.
[107]
KomatiB, KudryavtsevA, ClévyC, LaurentG, TamadazteB, AgnusJ, et al.Automated robotic microassembly of flexible optical components.In: Proceedings of the 2016 IEEE International Symposium on Assembly and Manufacturing (ISAM); 2016 Aug 21–22; Fort Worth, TX, USA. Piscataway: IEEE; 2016. p. 93–8.
[108]
FantoniG, JorgO, TincaniV.Indirect force measurement system in a mechanical microgripper.Precis Eng2022; 78:206-214.
[109]
AravindT, PraveenS Kumar, DineshG Ram, LingarajaD.Analysis of material profile for polymer-based mechanical microgripper for thin plate holding.E. Priya, V. Rajinikanth (Eds.), Signal and Image Processing Techniques for the Development of Intelligent Healthcare Systems, Springer, Cham2020; 103-117.
[110]
DeG Pasquale.Design and modeling of MEMS microgrippers for laser-based additive manufacturing.Micro2022; 2(2):225-239.
[111]
SariolaV, ZhouQ, KoivoHN.Three dimensional hybrid microassembly combining robotic microhandling and self-assembly.In: Proceedings of the 2009 IEEE International Conference on Robotics and Automation; 2009 May 12–17; Kobe, Japan. Piscataway: IEEE; 2009. p. 2605–10.
[112]
SariolaV, JääskeläinenM, ZhouQ.Hybrid microassembly combining robotics and water droplet self-alignment.IEEE Trans Robot2010; 26(6):965-977.
[113]
GeZ, DaiL, ZhaoJ, YuH, YangW, LiaoX, et al.Bubble-based microrobots enable digital assembly of heterogeneous microtissue modules.Biofabrication2022; 14(2):025023.
LevezielM, HaouasW, LaurentGJ, GauthierM, DahmoucheR.Migribot: a miniature parallel robot with integrated gripping for high-throughput micromanipulation.Sci Robot2022; 7(69):eabn4292.
[116]
ZhangD, ChenJ, LiW, BautistaD Salinas, YangGZ.A microsurgical robot research platform for robot-assisted microsurgery research and training.Int J CARS2020; 15:15-25.
[117]
YinW, WangY, LiuH, SunM, ZhangY, YuanH, et al.SCbots: stomatocyte-like colloidosomes as versatile microrobots fabricated by one-step self-assembly.Chem Eng J2024; 490:151952.
[118]
KangL, ZhaoJ, ZhuH, ZhuL, LiG, WangL, et al.Bubble jet propulsion of a “flying shuttle” zinc phosphate micro robot driven by enzyme-catalyzed reaction.Inorg Chem Commun2023; 157:111242.
[119]
KimS, JiangY, ThompsonKL Towell, BoutilierMS, NayakantiN, CaoC, et al.Soft nanocomposite electroadhesives for digital micro- and nanotransfer printing.Sci Adv2019; 5(10):eaax4790.
[120]
BoutilierMSH, CaoC, NayakantiN, KimS, Taheri-MousaviSM, HartAJ.Limiting mechanisms and scaling of electrostatically controlled adhesion of soft nanocomposite surfaces for robotic gripping.ACS Appl Mater Interfaces2021; 13(1):1192-1203.
[121]
WeiD, XiongQ, DongJ, WangH, LiangX, TangS, et al.Electrostatic adhesion clutch with superhigh force density achieved by mxene-poly(vinylidene fluoride-trifluoroethylene-chlorotrifluoroethylene) composites.Soft Robot2023; 10(3):482-492.
[122]
ZhangJ, DaiX, WuW, DuK.Micro-vision based high-precision space assembly approach for trans-scale micro-device: the CFTA example.Sensors2023; 23(1):450.
[123]
WangDH, WangK, QiangLS.Depth estimation method of surface of micropart in microassembly space based on microscopic vision tomographic scanning images.J Microsc2021; 283(2):77-92.
[124]
XieT, ZhangX, LiH, ZhangJ.Hybrid feature based 6D pose tracking under binocular vision for automated micro-assembly.In: Proceedings of the 2023 9th International Conference on Mechatronics and Robotics Engineering (ICMRE); 2023 Feb 10–12; Shenzhen, China. Piscataway: IEEE; 2023. p. 112–7.
[125]
WeiD, HallMB, SherehiyA, PopaDO.Design and evaluation of human–machine interface for nexus: a custom microassembly system.J Micro Nano-Manuf2020; 8(4):041011.
[126]
ZhangZ, WangX, ZhaoH, RenT, XuZ, LuoY.The machine vision measurement module of the modularized flexible precision assembly station for assembly of micro- and meso-sized parts.Micromachines2020; 11(10):918.
[127]
BolyaD, ZhouC, XiaoF, LeeYJ.YOLACT: real-time instance segmentation.In: Proceedings of the 2019 IEEE/CVF International Conference on Computer Vision (ICCV); 2019 Oct 27–Nov 2; Seoul, Republic of Korean. Piscataway: IEEE; 2020. p. 9156–65.
[128]
ChengJ, WuW, YangY, ZhangJ.Yolact in micro-assembly robot system.In: Proceedings of the 2021 4th International Conference on Algorithms, Computing and Artificial Intelligence; 2021 Dec 22–24; Sanya, China. New York city: Association for Computing Machinery; 2022. p. 1–5.
[129]
LiF, GaoD, YangY, ZhuJ.Small target deep convolution recognition algorithm based on improved YOLOv4.Int J Mach Learn Cybern2023; 14(2):387-394.
[130]
KhachikyanA, PippioneG, SengünesMI, PaolettiR, SeyfriedM.Micro-optics assembly for fast axis collimation by means of convolutional neural network.Opt Express2021; 29(17):26765-26774.
[131]
LiuS, JiaY, LiYF, GuoY, LuH.Simultaneous precision assembly of multiple objects through coordinated micro-robot manipulation.In: Proceedings of the 2021 IEEE International Conference on Robotics and Automation (ICRA); 2021 May 30–Jun 5; Xi'an, China. Piscataway: IEEE; 2021. p. 6280–6.
[132]
WeiY, XuQ.Design and testing of a new force-sensing cell microinjector based on soft flexure mechanism.IEEE Sens J2019; 19(15):6012-6019.
[133]
GurskyB, BütefischS, Leester-SchädelM, LiK, MatheisB, DietzelA.A disposable pneumatic microgripper for cell manipulation with image-based force sensing.Micromachines2019; 10(10):707.
[134]
PowerM, ThompsonAJ, AnastasovaS, YangGZ.A monolithic force-sensitive 3D microgripper fabricated on the tip of an optical fiber using 2-photon polymerization.Small2018; 14(16):1703964.
[135]
XuW, ZhangH, YuanH, LiangB.A compliant adaptive gripper and its intrinsic force sensing method.IEEE Trans Robot2021; 37(5):1584-1603.
[136]
XieH, MengX, ZhangH, SunL.Development of a magnetically driven microgripper for piconewton force-controlled microscale manipulation and characterization.IEEE Trans Ind Electron2020; 67(3):2065-2075.
[137]
ChungSE, DongXG, SittiM.Three-dimensional heterogeneous assembly of coded microgels using an untethered mobile microgripper.Lab Chip2015; 15(7):1667-1676.
[138]
LiD, NiuF, LiJ, LiX, SunD.Gradient-enhanced electromagnetic actuation system with a new core shape design for microrobot manipulation.IEEE Trans Ind Electron2020; 67(6):4700-4710.
[139]
YangZ, YangL, ZhangM, WangQ, YuSCH, ZhangL.Magnetic control of a steerable guidewire under ultrasound guidance using mobile electromagnets.IEEE Robot Autom Lett2021; 6(2):1280-1287.
[140]
SunT, ZhangY, PowerC, AlexanderPM, SuttonJT, AryalM, et al.Closed-loop control of targeted ultrasound drug delivery across the blood-brain/tumor barriers in a rat glioma model.Proc Natl Acad Sci USA2017; 114(48):E10281-E10290.
[141]
WeiT, LiuJ, LiD, ChenS, ZhangY, LiJ, et al.Development of magnet-driven and image-guided degradable microrobots for the precise delivery of engineered stem cells for cancer therapy.Small2020; 16(41):1906908.
[142]
Muiños-LandinS, FischerA, HolubecV, CichosF.Reinforcement learning with artificial microswimmers.Sci Robot2021; 6(52):eabd9285.
[143]
YangY, BevanMA, LiB.Micro/nano motor navigation and localization via deep reinforcement learning.Adv Theory Simul2020; 3(6):2000034.
[144]
YangY, BevanMA, LiB.Efficient navigation of colloidal robots in an unknown environment via deep reinforcement learning.Adv Intell Syst2020; 2(1):1900106.
[145]
BelharetK, FolioD, FerreiraA.Simulation and planning of a magnetically actuated microrobot navigating in the arteries.IEEE Trans Biomed Eng2013; 60(4):994-1001.
[146]
MengK, JiaY, YangH, NiuF, WangY, SunD.Motion planning and robust control for the endovascular navigation of a microrobot.IEEE Trans Ind Inform2020; 16(7):4557-4566.
[147]
JiangJ, YangL, ZhangL.Closed-loop control of a helmholtz coil system for accurate actuation of magnetic microrobot swarms.IEEE Robot Autom Lett2021; 6(2):827-834.
[148]
ArceseL, FruchardM, FerreiraA.Adaptive controller and observer for a magnetic microrobot.IEEE Trans Robot2013; 29(4):1060-1067.
[149]
MaW, LiJ, NiuF, JiH, SunD.Robust control to manipulate a microparticle with electromagnetic coil system.IEEE Trans Ind Electron2017; 64(11):8566-8577.
[150]
LiuJ, WuX, HuangC, ManamanchaiyapornL, ShangW, YanX, et al.3-D autonomous manipulation system of helical microswimmers with online compensation update.IEEE Trans Autom Sci Eng2021; 18(3):1380-1391.
[151]
YangZ, YangL, ZhangL.Autonomous navigation of magnetic microrobots in a large workspace using mobile-coil system.IEEE/ASME Trans Mechatron2021; 26(6):3163-3174.
[152]
XuT, LiuJ, HuangC, SunT, WuX.Discrete-time optimal control of miniature helical swimmers in horizontal plane.IEEE Trans Autom Sci Eng2022; 19(3):2267-2277.
[153]
ZhangZ, LongF, MenqCH.Three-dimensional visual servo control of a magnetically propelled microscopic bead.IEEE Trans Robot2013; 29(2):373-382.
[154]
BelharetK, FolioD, FerreiraA.Control of a magnetic microrobot navigating in microfluidic arterial bifurcations through pulsatile and viscous flow.In: Proceedings of the 2012 IEEE/RSJ International Conference on Intelligent Robots and Systems; 2012 Oct 7–12; Vilamoura, Portugal. Piscataway: IEEE; 2012. p. 2559–64.
[155]
DabbaghSR, SarabiMR, BirtekMT, SeyfiS, SittiM, TasogluS.3D-printed microrobots from design to translation.Nat Commun2022; 13(1):5875.
[156]
YuanK, Jurado-SánchezB, EscarpaA.Dual-propelled lanbiotic based janus micromotors for selective inactivation of bacterial biofilms.Angew Chem Int Ed2021; 60(9):4915-4924.
[157]
PachecoM, Jurado-SánchezB, EscarpaA.Sensitive monitoring of enterobacterial contamination of food using self-propelled janus microsensors.Anal Chem2018; 90(4):2912-2917.
ZhangY, KeumH, KimS.Microassembly of MEMS actuators and sensors via micro-masonry.In: Proceedings of the 2013 IEEE 26th International Conference on Micro Electro Mechanical Systems (MEMS); 2013 Jan 20–24; Taipei, China. Piscataway: IEEE; p. 283–6.
[160]
DasAN, MurthyR, PopaDO, StephanouHE.A multiscale assembly and packaging system for manufacturing of complex micro-nano devices.IEEE Trans Autom Sci Eng2012; 9(1):160-170.
[161]
NieminenT, TiwaryN, RossG, Paulasto-KröckelM.Detection of in-plane movement in electrically actuated microelectromechanical systems using a scanning electron microscope.Micromachines2023; 14(3):698.
[162]
MosbergAB, RenD, AhtapodovL, WemanH, FimlandBO, vanATJ Helvoort.Focused ion beam lithography for position-controlled nanowire growth.Nanotechnology2023; 34(33):335301.
[163]
RicottiL, TrimmerB, FeinbergAW, RamanR, ParkerKK, BashirR, et al.Biohybrid actuators for robotics: a review of devices actuated by living cells.Sci Robot2017; 2(12):eaaq0495.
[164]
AzizA, PaneS, IacovacciV, KoukourakisN, CzarskeJ, MenciassiA, et al.Medical imaging of microrobots: toward in vivo applications.ACS Nano2020; 14(9):10865-10893.
[165]
ChenR, FolioD, FerreiraA.Analysis and comparison of electromagnetic microrobotic platforms for biomedical applications.Appl Sci2022; 12(1):456.
[166]
NohS, JeonS, KimE, OhU, ParkD, ParkSH, et al.A biodegradable magnetic microrobot based on gelatin methacrylate for precise delivery of stem cells with mass production capability.Small2022; 18(25):2107888.
[167]
SongX, FuW, CheangUK.Immunomodulation and delivery of macrophages using nano-smooth drug-loaded magnetic microrobots for dual targeting cancer therapy.iScience2022; 25(7):104507.
[168]
WuZ, ZhangY, AiN, ChenH, GeW, XuQ.Magnetic mobile microrobots for upstream and downstream navigation in biofluids with variable flow rate.Adv Intell Syst2022; 4(7):2100266.
[169]
DarmawanBA, GongD, ParkH, JeongS, GoG, KimS, et al.Magnetically controlled reversible shape-morphing microrobots with real-time X-ray imaging for stomach cancer applications.J Mater Chem B2022; 10(23):4509-4518.
[170]
PachecoM, Mayorga-MartinezCC, ViktorovaJ, RumlT, EscarpaA, PumeraM.Microrobotic carrier with enzymatically encoded drug release in the presence of pancreatic cancer cells via programmed self-destruction.Appl Mater Today2022; 27:101494.
[171]
DuY, LoE, AliS, KhademhosseiniA.Directed assembly of cell-laden microgels for fabrication of 3D tissue constructs.Proc Natl Acad Sci USA2008; 105(28):9522-9527.
[172]
TangS, ZhangF, GongH, WeiF, ZhuangJ, KarshalevE, et al.Enzyme-powered janus platelet cell robots for active and targeted drug delivery.Sci Robot2020; 5(43):eaba6137.
[173]
WangH, ShiQ, YueT, NakajimaM, TakeuchiM, HuangQ, et al.Micro-assembly of a vascular-like micro-channel with railed micro-robot team-coordinated manipulation.Int J Adv Robot Syst2014; 11(7):115.
[174]
YueT, NakajimaM, TakeuchiM, HuC, HuangQ, FukudaT.On-chip self-assembly of cell embedded microstructures to vascular-like microtubes.Lab Chip2014; 14(6):1151-1161.
[175]
WangH, HuangQ, ShiQ, YueT, ChenS, NakajimaM, et al.Automated assembly of vascular-like microtube with repetitive single-step contact manipulation.IEEE Trans Biomed Eng2015; 62(11):2620-2628.
[176]
LiuX, ShiQ, WangH, SunT, YuN, HuangQ, et al.Automated fluidic assembly of microvessel-like structures using a multimicromanipulator system.IEEE/ASME Trans Mechatron2018; 23(2):667-678.
[177]
CuiJ, WangH, ZhengZ, ShiQ, SunT, HuangQ, et al.Fabrication of perfusable 3D hepatic lobule-like constructs through assembly of multiple cell type laden hydrogel microstructures.Biofabrication2018; 11(1):015016.
[178]
SunT, ShiQ, YaoY, SunJ, WangH, HuangQ, et al.Engineered tissue micro-rings fabricated from aggregated fibroblasts and microfibres for a bottom-up tissue engineering approach.Biofabrication2019; 11(3):035029.
[179]
TakeuchiM, IriguchiM, HattoriM, KimE, IchikawaA, HasegawaY, et al.Magnetic self-assembly of toroidal hepatic microstructures for micro-tissue fabrication.Biomed Mater2020; 15(5):055001.
[180]
AkolpogluMB, AlapanY, DoganNO, BaltaciSF, YasaO, AybarG Tural, et al.Magnetically steerable bacterial microrobots moving in 3D biological matrices for stimuli-responsive cargo delivery.Sci Adv2022; 8(28):eabo6163.
[181]
WangQ, DuX, JinD, ZhangL.Real-time ultrasound doppler tracking and autonomous navigation of a miniature helical robot for accelerating thrombolysis in dynamic blood flow.ACS Nano2022; 16(1):604-616.
[182]
SongX, SunR, WangR, ZhouK, XieR, LinJ, et al.Puffball-inspired microrobotic systems with robust payload, strong protection, and targeted locomotion for on-demand drug delivery.Adv Mater2022; 34(43):2204791.
[183]
CongZ, TangS, XieL, YangM, LiY, LuD, et al.Magnetic-powered janus cell robots loaded with oncolytic adenovirus for active and targeted virotherapy of bladder cancer.Adv Mater2022; 34(26):2201042.
[184]
YinT, DiaoZ, BlumNT, QiuL, MaA, HuangP.Engineering bacteria and bionic bacterial derivatives with nanoparticles for cancer therapy.Small2022; 18(12):2104643.
[185]
WangB, ChanKF, YuanK, WangQ, XiaX, YangL, et al.Endoscopy-assisted magnetic navigation of biohybrid soft microrobots with rapid endoluminal delivery and imaging.Sci Robot2021; 6(52):eabd2813.
[186]
WeiX, LuanL, ZhaoZ, LiX, ZhuH, PotnisO, et al.Nanofabricated ultraflexible electrode arrays for high-density intracortical recording.Adv Sci2018; 5(6):1700625.
MuskE.An integrated brain-machine interface platform with thousands of channels.J Med Internet Res2019; 21(10):e16194.
[189]
QinF, XuD, ZhangD, PeiW, HanX, YuS.Automated hooking of biomedical microelectrode guided by intelligent microscopic vision.IEEE/ASME Trans Mechatron2023; 28(5):2786-2798.
[190]
SuD, UmJ, MorenoJ, NematiZ, SrinivasanK, ChenJ, et al.GMR biosensing with magnetic nanowires as labels for the detection of osteosarcoma cells.Sens Actuators A2023; 350:114115.
[191]
HuangC, ZhaoJ, LuR, WangJ, NugenSR, ChenY, et al.A phage-based magnetic relaxation switching biosensor using bioorthogonal reaction signal amplification for salmonella detection in foods.Food Chem2023; 400:134035.
[192]
PalT, AdityaS, MathaiT, MukherjiS.Polyaniline coated plastic optic fiber biosensor for detection of aflatoxin B1 in nut, cereals, beverages, and body fluids.Sens Actuators B2023; 389:133897.
[193]
WenX, ChangX, LiA, YangX, TianF, LiuZ, et al.ZnO/Cu2O heterojunction integrated fiber-optic biosensor for remote detection of cysteine.Biosens Bioelectron2023; 223:115021.
[194]
GagliardiM, AgostiniM, LunardelliF, LamannaL, MirandaA, BazzichiA, et al.Surface acoustic wave-based lab-on-a-chip for the fast detection of legionella pneumophila in water.Sens Actuators B2023; 379:133299.
[195]
ChenC, RanB, LiuB, LiuX, LiuY, LanM, et al.Development of a novel microfluidic biosensing platform integrating micropillar array electrode and acoustic microstreaming techniques.Biosens Bioelectron2023; 223:114703.
[196]
GaoL, WangJ, ZhaoY, LiH, LiuM, DingJ, et al.Free-standing nanofilm electrode arrays for long-term stable neural interfacings.Adv Mater2022; 34(5):2107343.
[197]
RizaMA, GoYI, HarunSW, MaierRRJ.FBG sensors for environmental and biochemical applications—a review.IEEE Sens J2020; 20(14):7614-7627.
[198]
FarooqMU, KoSY.A decade of MRI compatible robots: systematic review.IEEE Trans Robot2022; 39(2):862-884.
[199]
HuangS, LouC, ZhouY, HeZ, JinX, FengY, et al.MRI-guided robot intervention—current state-of-the-art and new challenges.Med-X2023; 1:4.
[200]
YangGZ, BellinghamJ, DupontPE, FischerP, FloridiL, FullR, et al.The grand challenges of science robotics.Sci Robot2018; 3(14):eaar7650.
[201]
GuoY, ChenW, ZhaoJ, YangGZ.Medical robotics: opportunities in China.Annu Rev Control Robot Auton Syst2022; 5(1):361-383.
[202]
ZhuL, ShaoC, ChenH, ChenZ, ZhaoY.Hierarchical hydrogels with ordered micro-nano structures for cancer-on-a-chip construction.Research2021; 2021:9845679.
PDF
(5683KB)
