Recent Development of Microﬂuidic Technology for Cell Trapping in Single Cell Analysis: A Review

: Microﬂuidic technology has emerged from the MEMS (Micro-Electro-Mechanical System)-technology as an important research ﬁeld. During the last decade, various microﬂuidic technologies have been developed to open up a new era for biological studies. To understand the function of single cells, it is very important to monitor the dynamic behavior of a single cell in a living environment. Cell trapping in single cell analysis is urgently demanded There have been some review papers focusing on drug screen and cell analysis. However, cell trapping in single cell analysis has rarely been covered in the previous reviews. The present paper focuses on recent developments of cell trapping and highlights the mechanisms, governing equations and key parameters a ﬀ ecting the cell trapping e ﬃ ciency by contact-based and contactless approach. The applications of the cell trapping method are discussed according to their basic research areas, such as biology and tissue engineering. Finally, the paper highlights the most promising cell trapping method for this research area.


Introduction
In the past decade, single cell analysis has received significant attention from the research community due to its wide applications in pharmaceutical [1], biology [2], healthcare [3] and tissue engineering [4]. Many studies have shown that individual cells, even those that are identical in morphology, exhibit intercellular variations due to differences in their micro-environmental conditions and gene expression [5,6]. Previous cell analysis was performed based on a large population of cells, which reflected the average values derived from the bulk cell response [7]. The bulk cell analysis approach ignored the characteristics of individual cells. The limitations of bulk cell analysis have motivated the development of single cell analysis. In contrast to the bulk cell analysis, single cell analysis reveals the significant physiological characteristics of an individual cell, such as metabolism [8], protein level [9] and gene expressions [10].
Progresses in single cell analysis depend on the development of tools and equipment which allow new insights into the cell. Since the invention of the microscope, single cell analysis has been successfully carried out. Many new illumination, staining, and detection methods, such as flow cytometry [11], have been developed in order to increase the optical resolution of the microscope to observe the behavior of single cell [12]. With these methods, single cell analysis has become feasible. However, the conventional methods of observing single cell have limited the performance in terms of standardized reproducibility and high throughput, since the accuracy of a traditional single cell where τ p is the particle response time to changes in the flow field, and α is the density ratio between the fluid and the particle. → F R represents Brownian random force, and → F S represents spring force. The embedded obstacles play an important role in changing the external forces exerted on the cells, such as Stokes drag and pressure gradient forces.
Once target cells are trapped in hydrodynamic trapping locations, they can be further employed for various studies. Typically, cell trapping obstacles include walls or pores with various shapes, and arrays consisting of a pattern of the trapping obstacles can be used to realize cell trapping. The development of microfluidic device has driven the evolution of various hydrodynamic trapping devices. Chen et al. [49] designed an integrated microfluidic device for particle arrangement and isolation ( Figure 1). The device was able to selectively immobilize desired microparticles in an array of hydrodynamic traps based on three different physical characteristics: size, elastic modulus and internal structure. A scaling theory based on particle and trap dimensions, particle elastic modules and applied pressure was also developed to define the criterion for particle parking. Moreover, r p represents the radius of the particle, r c represents the half width of the trap entrance, xm represents the height (as shown in Figure 1c), µ represents the friction coefficient between the flow channel and the particle, and κ = 4 x m − r c tan −1 (x m /r c ) ; additionally, hc and hp represent channel height and particle height, respectively. C is a correction factor, representing the deformation of flow channel. Critical pressure indicates that, at a certain pressure, the trapping process can be realized. The trapping process can be achieved, considering the particles with different sizes and stiffness. The isolation efficiency can reach as high as 95%. This setup can be potentially employed in trapping soft and biological objects.
Processes 2018, 6, x FOR PEER REVIEW 3 of 34 development of microfluidic device has driven the evolution of various hydrodynamic trapping devices. Chen et al. [49] designed an integrated microfluidic device for particle arrangement and isolation ( Figure 1). The device was able to selectively immobilize desired microparticles in an array of hydrodynamic traps based on three different physical characteristics: size, elastic modulus and internal structure. A scaling theory based on particle and trap dimensions, particle elastic modules and applied pressure was also developed to define the criterion for particle parking. Moreover, rp represents the radius of the particle, rc represents the half width of the trap entrance, xm represents the height (as shown in Figure 1c), μ represents the friction coefficient between the flow channel and the particle, and  ; additionally, hc and hp represent channel height and particle height, respectively. C is a correction factor, representing the deformation of flow channel. Critical pressure indicates that, at a certain pressure, the trapping process can be realized. The trapping process can be achieved, considering the particles with different sizes and stiffness. The isolation efficiency can reach as high as 95%. This setup can be potentially employed in trapping soft and biological objects. Zhu et al. [50] presented a proof-of-concept microfluidic device for the immobilization, culturing and imaging of zebrafish embryos. The schematic illustration and actual photo of their microfluidic device are shown in Figure 2. The device consisted of a flat glass substrate and two layers of polydimethylsiloxane (PDMS) structures replicated from 3D-printed masters. An embryo-culturing channel and five traps were embedded at the bottom PDMS layer to load and capture embryos. The working procedure is as follows. The first step is to induce embryo through the embryo inlet. The second step is to insert the PTFE (polytetrafluoroethylene) plug to avoid the leakage of working fluid. The third step is to introduce working fluid into the device by a syringe pump. The fourth step is to tilt the device slightly, to enable embryo trapping function to trap embryos one by one. The shear stress on the immobilized embryos was estimated by using the Computational Fluid Dynamics (CFD) Zhu et al. [50] presented a proof-of-concept microfluidic device for the immobilization, culturing and imaging of zebrafish embryos. The schematic illustration and actual photo of their microfluidic device are shown in Figure 2. The device consisted of a flat glass substrate and two layers of polydimethylsiloxane (PDMS) structures replicated from 3D-printed masters. An embryo-culturing channel and five traps were embedded at the bottom PDMS layer to load and capture embryos. The working procedure is as follows. The first step is to induce embryo through the embryo inlet. The second step is to insert the PTFE (polytetrafluoroethylene) plug to avoid the leakage of working fluid. The third step is to introduce working fluid into the device by a syringe pump. The fourth step is to tilt the device slightly, to enable embryo trapping function to trap embryos one by one. The shear stress on the immobilized embryos was estimated by using the Computational Fluid Dynamics (CFD) simulations [51]. This device could be potentially applied to monitor the development of dynamic embryonic. Fan et al. [52] designed a microfluidic device, using flow resistance, to achieve high-efficient single-cell capture and analysis. A schematic illustration of their microfluidic device is shown in Figure 3a. Each single cell was encapsulated in a micro-droplet that was formed at the T-junction ( Figure 3b). The micro-droplets at the beginning moved in the main channel at certain approximate velocity; however, once a micro-droplet was blocked in the main channel of a "main-bypass" structure unit, the rest of the micro-droplets would travel through the bypass channel into the subsequent "main-bypass" unit and another droplet would also be trapped (Figure 3c). This trapping process would repeat itself and continue. After that, the trapped cell could be further observed and analyzed ( Figure 3d). The cell trapping efficiency of this device was up to 60% by controlling the injection process. Fan et al. [52] designed a microfluidic device, using flow resistance, to achieve high-efficient single-cell capture and analysis. A schematic illustration of their microfluidic device is shown in Figure 3a. Each single cell was encapsulated in a micro-droplet that was formed at the T-junction ( Figure 3b). The micro-droplets at the beginning moved in the main channel at certain approximate velocity; however, once a micro-droplet was blocked in the main channel of a "main-bypass" structure unit, the rest of the micro-droplets would travel through the bypass channel into the subsequent "main-bypass" unit and another droplet would also be trapped (Figure 3c). This trapping process would repeat itself and continue. After that, the trapped cell could be further observed and analyzed ( Figure 3d). The cell trapping efficiency of this device was up to 60% by controlling the injection process.
Similar to the design of Fan et al., Xu et al. [53] developed a microfluidic device with double-slit arrays, as shown in Figure 4a. The microfluidic device consisted of an inlet reservoir, support and disperse pillars, micro-array, outlet channel and outlet reservoir. The double-slit structure of micro-array is shown in Figure 4b. The effects of different combinations of flow velocity, the fluid pressure and the stress of cells on the cell trapping efficiency were also investigated, in detail, in their work. This is the inheritance of previous work, which found that the double-slit arrays perform better compared to single-slit and seamless structure [54]. The geometric effect was employed to optimize the stress that cells suffered. The trapping efficiency was found to be dependent on the flow velocity, the fluid pressure and the equivalent stress of cells. The trapping efficiency was up to 70%. Similar to the design of Fan et al., Xu et al. [53] developed a microfluidic device with double-slit arrays, as shown in Figure 4a. The microfluidic device consisted of an inlet reservoir, support and disperse pillars, micro-array, outlet channel and outlet reservoir. The double-slit structure of microarray is shown in Figure 4b. The effects of different combinations of flow velocity, the fluid pressure and the stress of cells on the cell trapping efficiency were also investigated, in detail, in their work. This is the inheritance of previous work, which found that the double-slit arrays perform better compared to single-slit and seamless structure [54]. The geometric effect was employed to optimize the stress that cells suffered. The trapping efficiency was found to be dependent on the flow velocity, the fluid pressure and the equivalent stress of cells. The trapping efficiency was up to 70%. Zhu et al. [55] also developed a microfluidic device for cell trapping based on hydrodynamic method. For the first step, they tested the capabilities of cell trapping in three types of microstructures, as shown in Figure 5a. In these three designs, their gaps were embedded at different locations around the pillars. It was found that the microstructure of type C was the most efficient in cell trapping, as shown in Figure 5a(ii). The blue bar and red bar represent the trapping efficiency  Similar to the design of Fan et al., Xu et al. [53] developed a microfluidic device with double-slit arrays, as shown in Figure 4a. The microfluidic device consisted of an inlet reservoir, support and disperse pillars, micro-array, outlet channel and outlet reservoir. The double-slit structure of microarray is shown in Figure 4b. The effects of different combinations of flow velocity, the fluid pressure and the stress of cells on the cell trapping efficiency were also investigated, in detail, in their work. This is the inheritance of previous work, which found that the double-slit arrays perform better compared to single-slit and seamless structure [54]. The geometric effect was employed to optimize the stress that cells suffered. The trapping efficiency was found to be dependent on the flow velocity, the fluid pressure and the equivalent stress of cells. The trapping efficiency was up to 70%. Zhu et al. [55] also developed a microfluidic device for cell trapping based on hydrodynamic method. For the first step, they tested the capabilities of cell trapping in three types of microstructures, as shown in Figure 5a. In these three designs, their gaps were embedded at different locations around the pillars. It was found that the microstructure of type C was the most efficient in cell trapping, as shown in Figure 5a(ii). The blue bar and red bar represent the trapping efficiency Zhu et al. [55] also developed a microfluidic device for cell trapping based on hydrodynamic method. For the first step, they tested the capabilities of cell trapping in three types of microstructures, as shown in Figure 5a. In these three designs, their gaps were embedded at different locations around the pillars. It was found that the microstructure of type C was the most efficient in cell trapping, as shown in Figure 5a(ii). The blue bar and red bar represent the trapping efficiency without reversed flow and with reversed flow. The trapping efficiency of type C with reversed flow is about 90%, which is much higher than that of type A and type B. After that, a face-to-face heart-shaped microstructure was developed to carry out cell trapping, using type C structure (Figure 5b). Oil was employed in the isolate chambers, to reduce cross-talk. To study the trapping mechanism, numerical simulation work was carried out so as to obtain the flow velocity and shear stress through the pillars. A shadow area with relatively low flow velocity was obtained behind the pillar. The finding supported the hypothesis that the cell could be trapped behind the pillar. The efficiency of cell trapping and cell pairing were 93% and 84%, respectively. The height of the gap and the height of the pillar were modified in the numerical simulation, to investigate their effect on the efficiency of cell pairing.
Oil was employed in the isolate chambers, to reduce cross-talk. To study the trapping mechanism, numerical simulation work was carried out so as to obtain the flow velocity and shear stress through the pillars. A shadow area with relatively low flow velocity was obtained behind the pillar. The finding supported the hypothesis that the cell could be trapped behind the pillar. The efficiency of cell trapping and cell pairing were 93% and 84%, respectively. The height of the gap and the height of the pillar were modified in the numerical simulation, to investigate their effect on the efficiency of cell pairing.

Dielectrophoresis (DEP) Actuated Cell Trapping
Dielectrophoresis (DEP) is an effective way to realize cell trapping in microfluidic devices. The principle of DEP actuated cell trapping is to employ the DEP force imposing on a dielectric particle/cell. In a non-uniform electric field, the strength of DEP force is strongly dependent on the magnitude and polarity of the charges induced on a particle/cell. The DEP force can be expressed as follows: where ε e , a, f, E and σ represent the permittivity of the medium, radius of cell, frequency, electric field strength and conductivity, respectively. Aslan and Kulah [56] designed a portable microfluidic DEP device to realize DEP actuated cell trapping and CMOS (Complementary Metal Oxide Semiconductor) imaging function. The manufacturing process of their microfluidic DEP device is shown in Figure 6. RIE is short for Reactive-Ion Etching. To realize the CMOS imaging function, glass was employed as the substrate due to its transparency characteristic. The channel was made of parylene material on account of its advantages of biocompatibility and bio-stability. AC (Alternating Current) signal was applied on the electrodes of microfluidic DEP device to achieve MCF-7 (Michigan Cancer Foundation-7) breast cancer cell trapping. The proposed DEP device could be potentially used where equipment is limited. The counting accuracy was up to 90%, as reported in the paper. It was dependent on the conductivity and angular frequency of the electric field.  Takeuchi et al. [57] developed an electro-active micro-well array with barriers (EMAB) to realize highly efficient single cervical cell trapping. The schematic illustration of the EMAB device is shown in Figure 7. The patterned electrodes were embedded at the bottom of cell-sized micro-wells to achieve cell trapping. The cell could be trapped in the micro wells by applying a sinusoidal electric potential (peak to peak voltage, Vpp = 5 V at 1 MHz) to the electrodes. With the help of barriers located beside the micro-well, cell holding can be realized even after shutting off the DEP, as shown in Figure  7a. The actual image of the EMAB microfluid device is shown in Figure 7b. As shown in the yellow region of Figure 7b, the microfluidic channel consisted of many microwells and barriers. Each microwell contained a pair of ITO electrodes at the bottom to achieve cell trapping and holding functions. The microfluid device could be employed for cell trapping, staining and imaging ( Figure  7c). The cell trapping efficiency was up to 92%, which was determined by the permittivity of the Takeuchi et al. [57] developed an electro-active micro-well array with barriers (EMAB) to realize highly efficient single cervical cell trapping. The schematic illustration of the EMAB device is shown in Figure 7. The patterned electrodes were embedded at the bottom of cell-sized micro-wells to achieve cell trapping. The cell could be trapped in the micro wells by applying a sinusoidal electric potential (peak to peak voltage, V pp = 5 V at 1 MHz) to the electrodes. With the help of barriers located beside the micro-well, cell holding can be realized even after shutting off the DEP, as shown Puri et al. [58] presented a C-serpentine microchannel to achieve the trapping and separation of live and dead yeast cells (Saccharomyces cerevisiae) through DEP. The schematic illustration of the geometry is shown in Figure 8a,b. The C-serpentine geometry was employed to generate a gradient distribution of the electric field. To specify the electric distribution in the geometry, a yeast structure with multi-shell model was employed. The model consisted of three concentric layers of wall, membrane and nucleus ( Figure 8c). Due to the differences in the electric conductivity of the cell membrane, the live and dead yeast cells would be driven to pDEP and nDEP region, respectively ( Figure 8d). An average trapping efficiency of 97.9% for dead cells and 93.4% for live cells was obtained, which was determined by the applied voltage. Puri et al. [58] presented a C-serpentine microchannel to achieve the trapping and separation of live and dead yeast cells (Saccharomyces cerevisiae) through DEP. The schematic illustration of the geometry is shown in Figure 8a,b. The C-serpentine geometry was employed to generate a gradient distribution of the electric field. To specify the electric distribution in the geometry, a yeast structure with multi-shell model was employed. The model consisted of three concentric layers of wall, membrane and nucleus ( Figure 8c). Due to the differences in the electric conductivity of the cell membrane, the live and dead yeast cells would be driven to pDEP and nDEP region, respectively ( Figure 8d). An average trapping efficiency of 97.9% for dead cells and 93.4% for live cells was obtained, which was determined by the applied voltage.
Fritzsch et al. [59] demonstrated contactless cell trapping with the octupole technology, using DEP. This technology can be applied in the miniaturized octupole cytometry. Compared with traditional cytometry approach, the proposed approach could trap the targeted cells for further analysis. To investigate the trapping efficiency of single cell, three different octupole nDEP field control modes were employed, including ACB (non-rotating octupolar field), ACC (non-rotating quadrupolar field) and ROTX (rotating quadrupolar field). It was found that cells could be efficiently trapped under ROTX mode. Contactless cell trapping was realized by using the octupole technology, which was independent of cell size and morphology.
with multi-shell model was employed. The model consisted of three concentric layers of wall, membrane and nucleus ( Figure 8c). Due to the differences in the electric conductivity of the cell membrane, the live and dead yeast cells would be driven to pDEP and nDEP region, respectively ( Figure 8d). An average trapping efficiency of 97.9% for dead cells and 93.4% for live cells was obtained, which was determined by the applied voltage. Chen et al. [60] reported a microfluidic chip for trapping Shewanella oneidensis bacteria at the cell level, using positive DEP (pDEP) effect. The schematic illustration of the experimental setup was provided in Figure 9. The bacteria were first injected into the microfluidic chip, using a syringe pump. HEPES (4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid) is a zwitterionic sulfonic acid buffering agent. As the bacteria traveled in the microchannel, the trapping process could be captured by a fluorescence microscope combined with a CCD (charge-coupled device) camera. This device demonstrated the possibility in trapping bacteria at the cell level. Fritzsch et al. [59] demonstrated contactless cell trapping with the octupole technology, using DEP. This technology can be applied in the miniaturized octupole cytometry. Compared with traditional cytometry approach, the proposed approach could trap the targeted cells for further analysis. To investigate the trapping efficiency of single cell, three different octupole nDEP field control modes were employed, including ACB (non-rotating octupolar field), ACC (non-rotating quadrupolar field) and ROTX (rotating quadrupolar field). It was found that cells could be efficiently trapped under ROTX mode. Contactless cell trapping was realized by using the octupole technology, which was independent of cell size and morphology.
Chen et al. [60] reported a microfluidic chip for trapping Shewanella oneidensis bacteria at the cell level, using positive DEP (pDEP) effect. The schematic illustration of the experimental setup was provided in Figure 9. The bacteria were first injected into the microfluidic chip, using a syringe pump. HEPES (4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid) is a zwitterionic sulfonic acid buffering agent. As the bacteria traveled in the microchannel, the trapping process could be captured by a fluorescence microscope combined with a CCD (charge-coupled device) camera. This device demonstrated the possibility in trapping bacteria at the cell level.

Magnetophoresis Actuated Cell Trapping
Magnetophoresis is another effective way to realize cell trapping in microfluidic channels. The principle of magnetic actuated cell trapping is to apply magnetic force on a particle/cell. This kind of trapping method can be further classified into positive and negative magnetophoresis. Positive magnetophoresis is the migration of magnetic particle/cell in a diamagnetic medium, while negative magnetophoresis is in a magnetic medium [61]. In addition, if the susceptibility of the particle/cell is larger than the ambient medium, positive magnetophoresis would also occur. The magnetic force imposing on a particle suspended in a fluid medium is as follows: HEPES represents 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid; CCD represents charge-coupled device.

Magnetophoresis Actuated Cell Trapping
Magnetophoresis is another effective way to realize cell trapping in microfluidic channels. The principle of magnetic actuated cell trapping is to apply magnetic force on a particle/cell. This kind of trapping method can be further classified into positive and negative magnetophoresis.
Positive magnetophoresis is the migration of magnetic particle/cell in a diamagnetic medium, while negative magnetophoresis is in a magnetic medium [61]. In addition, if the susceptibility of the particle/cell is larger than the ambient medium, positive magnetophoresis would also occur. The magnetic force imposing on a particle suspended in a fluid medium is as follows: where V is the volume of the particle, χ is the magnetic susceptibility, χ m is the susceptibility of the surround medium, µ 0 is the magnetic permeability of the air and B is the magnetic flux density. Scherr et al. [62] developed a two-magnet microfluidic setup to achieve high efficiency trapping for biofluids. In this configuration, the trapping process was realized by using two magnets instead of one magnet. The experimental bead distributions in a stationary tube are shown in Figure 10. Under the configuration of one magnet, the bead dispersed along the one side wall only. However, in the presence of the second magnet, the bead dispersed along both sides of the wall. The interaction area between the beads and fluids was much larger than that in the single magnet configuration, and the bead volume fraction was found to increase three times, compared with the one-magnet system. The bead volume fraction was determined by magnetic field. Scherr et al. [62] developed a two-magnet microfluidic setup to achieve high efficiency trapping for biofluids. In this configuration, the trapping process was realized by using two magnets instead of one magnet. The experimental bead distributions in a stationary tube are shown in Figure 10. Under the configuration of one magnet, the bead dispersed along the one side wall only. However, in the presence of the second magnet, the bead dispersed along both sides of the wall. The interaction area between the beads and fluids was much larger than that in the single magnet configuration, and the bead volume fraction was found to increase three times, compared with the one-magnet system. The bead volume fraction was determined by magnetic field. Kirby et al. [63] performed cell separation and trapping, using the magnetic force and centrifugal force. These two forces were employed in a centrifuge-magnetophoretic microfluidic device. There were six chambers embedded in the disk-shaped device. Three magnets were located besides each chamber. As the disk rotated, the centrifugal force was introduced. The combination of magnetic force and centrifugal force resulted in the separation and trapping of particles/cells of different sizes. This centrifugal microfluidic platform could be used for the separation and trapping of blood cells and tagged cancer cells.
Guo et al. [64] developed a magnetic controlled microfluidic device to trap magnetic tagged Salmonella typhimurium ( Figure 11). In this microfluidic device, sample stream and buffer stream were injected into the microflow channel. The magnetic tagged Salmonella typhimurium was separated because of the lateral magnetic force. The separated Salmonella typhimurium was led toward the patterned nickel array for trapping. For the cell trapping using positive magnetophoresis technique, the major difficulty was the accumulation of magnetic particles/cells into a cluster. The formed cluster would block the microflow channel. To solve this problem, researchers tried to regulate the gradient of magnetic field. The trapping efficiency was affected by the magnetic force and the drag force. Kirby et al. [63] performed cell separation and trapping, using the magnetic force and centrifugal force. These two forces were employed in a centrifuge-magnetophoretic microfluidic device. There were six chambers embedded in the disk-shaped device. Three magnets were located besides each chamber. As the disk rotated, the centrifugal force was introduced. The combination of magnetic force and centrifugal force resulted in the separation and trapping of particles/cells of different sizes. This centrifugal microfluidic platform could be used for the separation and trapping of blood cells and tagged cancer cells.
Guo et al. [64] developed a magnetic controlled microfluidic device to trap magnetic tagged Salmonella typhimurium ( Figure 11). In this microfluidic device, sample stream and buffer stream were injected into the microflow channel. The magnetic tagged Salmonella typhimurium was separated because of the lateral magnetic force. The separated Salmonella typhimurium was led toward the patterned nickel array for trapping. For the cell trapping using positive magnetophoresis technique, the major difficulty was the accumulation of magnetic particles/cells into a cluster. The formed cluster would block the microflow channel. To solve this problem, researchers tried to regulate the gradient of magnetic field. The trapping efficiency was affected by the magnetic force and the drag force. Huang et al. [65] embedded a microwell in a microfluidic device, to achieve immunomagnetic single cell trapping function ( Figure 12). A layer consisting of microwells was embedded between the microchannel and the magnet. Due to the presence of the microwell, uniform distribution of the magnetic field in the microfluidic device could be achieved. The single-particle trapping efficiency in the microwell can be up to 62%, and the purity can be up to 99.6%. The immunomagnetic labeled THP-1 cell was employed to demonstrate the feasibility of the microfluidic device. The trapping efficiency was supposed to be affected by the magnetic susceptibility of Dynabeads occupying the microwell. Huang et al. [65] embedded a microwell in a microfluidic device, to achieve immunomagnetic single cell trapping function ( Figure 12). A layer consisting of microwells was embedded between the microchannel and the magnet. Due to the presence of the microwell, uniform distribution of the magnetic field in the microfluidic device could be achieved. The single-particle trapping efficiency in the microwell can be up to 62%, and the purity can be up to 99.6%. The immunomagnetic labeled THP-1 cell was employed to demonstrate the feasibility of the microfluidic device. The trapping efficiency was supposed to be affected by the magnetic susceptibility of Dynabeads occupying the microwell. If the susceptibility of the particle/cell is smaller compared to the ambient medium, negative magnetophoresis will occur. Hejazian and Nguyen [66] reported a new trapping method for the sizeselective non-magnetic particles. The diluted ferrofluid was employed as the working fluid. The schematic illustration of the microchannel is shown in Figure 13a. Two arrays of magnets were embedded at the two opposite sides of a straight microchannel. The minimum and maximum of the simulated magnetic field were labeled dark blue and light blue, respectively. Figure 13b shows the experimental trapping results of small particles and large particles at the magnetic field maxima and minimum. The red spots stand for the small particles, while the green spots stand for the large particles. The physics behind the phenomena were illustrated by the combination of three kinds of forces, namely hydrodynamic force, negative magnetophoretic force and magnetoconvective force (Figure 13c). The trapping efficiency was determined by three kinds of forces, namely negative magnetophoretic, magnetoconvective and hydrodynamic forces. If the susceptibility of the particle/cell is smaller compared to the ambient medium, negative magnetophoresis will occur. Hejazian and Nguyen [66] reported a new trapping method for the size-selective non-magnetic particles. The diluted ferrofluid was employed as the working fluid. The schematic illustration of the microchannel is shown in Figure 13a. Two arrays of magnets were embedded at the two opposite sides of a straight microchannel. The minimum and maximum of the simulated magnetic field were labeled dark blue and light blue, respectively. Figure 13b shows the experimental trapping results of small particles and large particles at the magnetic field maxima and minimum. The red spots stand for the small particles, while the green spots stand for the large particles. The physics behind the phenomena were illustrated by the combination of three kinds of forces, namely hydrodynamic force, negative magnetophoretic force and magnetoconvective force (Figure 13c). The trapping efficiency was determined by three kinds of forces, namely negative magnetophoretic, magnetoconvective and hydrodynamic forces.
Wang et al. [67] employed micro-magneto-fluidic technique to trap the bacteria suspending in flowing fluid (Figure 14a). Thermal bonding technique was employed to build the microfluidic channel using poly(methyl methacrylate) (PMMA). An island was embedded at the center of the microchannel as shown in Figure 14b. The magnetic nanoparticles in the ferrofluid were magnetite (Fe 3 O 4 ). The polybeads were employed to predict the behavior of bacteria due to their similar size. The trapping behavior under the combination of ferrofluid and magnetic field is shown in Figure 14c. The variation in (x, y) coordinate of trapped bacteria with time under applied magnetic field is shown in Figure 14d. The effects of applied magnetic field, duration of application of magnetic field and fluid flow rate on the trapping efficiency were investigated systematically. Wang et al. [67] employed micro-magneto-fluidic technique to trap the bacteria suspending in flowing fluid (Figure 14a). Thermal bonding technique was employed to build the microfluidic channel using poly(methyl methacrylate) (PMMA). An island was embedded at the center of the microchannel as shown in Figure 14b. The magnetic nanoparticles in the ferrofluid were magnetite (Fe3O4). The polybeads were employed to predict the behavior of bacteria due to their similar size. The trapping behavior under the combination of ferrofluid and magnetic field is shown in Figure 14c. The variation in (x,y) coordinate of trapped bacteria with time under applied magnetic field is shown in Figure 14d. The effects of applied magnetic field, duration of application of magnetic field and fluid flow rate on the trapping efficiency were investigated systematically.  Instead of placing multiple pairs of magnets along a straight flow channel [68], Zhou et al. [69] placed the magnet near a T-junction microchannel (Figure 15a). The magnet was placed along the centerline of the main branch of T-junction. The flow direction was indicated by the blue color. The diamagnetic particles were trapped along the side wall of the main branch, while magnetic particles were trapped along the wall of side branch, indicating the existence of negative and positive magnetophoresis (Figure 15b). A 3D numerical model was developed to simulate the trapping procedure. The effect of ferrofluid was found to be important. Wilbanks et al. [70] investigated the effects of magnet asymmetry on the trapping performance of diamagnetic particle in ferrofluid flow. The magnetic configuration is shown in Figure 16a. The asymmetric magnets were embedded at two sides of the microflow channel, to realize trapping function (Figure 16b). The dimensions of the microchannel with the magnets were provided in Figure  16c. Under the influence of magnet asymmetry, a circular streamline shape of trapping particles was obtained, and the trapping performance was found to be dependent on the asymmetry of magnets.  Wilbanks et al. [70] investigated the effects of magnet asymmetry on the trapping performance of diamagnetic particle in ferrofluid flow. The magnetic configuration is shown in Figure 16a. The asymmetric magnets were embedded at two sides of the microflow channel, to realize trapping function (Figure 16b). The dimensions of the microchannel with the magnets were provided in Figure 16c. Under the influence of magnet asymmetry, a circular streamline shape of trapping particles was obtained, and the trapping performance was found to be dependent on the asymmetry of magnets.
Gertz and Khitun [71] investigated the trapping of red blood cells (RBC), using magnetic nanoparticles. The schematic illustration of the experimental setup is shown in Figure 17a. Two Cu wires covered by silicon dioxide were embedded in the working device. A power supply was employed to provide micro-electromagnet field to trap RBCs. Without the activation of current, the RBCs were randomly distributed in the channel; once the current was on, the RBCs were trapped around the wire (Figure 17b,c). The strength of the magnetic field influenced the trapping procedure significantly.
Wilbanks et al. [70] investigated the effects of magnet asymmetry on the trapping performance of diamagnetic particle in ferrofluid flow. The magnetic configuration is shown in Figure 16a. The asymmetric magnets were embedded at two sides of the microflow channel, to realize trapping function (Figure 16b). The dimensions of the microchannel with the magnets were provided in Figure  16c. Under the influence of magnet asymmetry, a circular streamline shape of trapping particles was obtained, and the trapping performance was found to be dependent on the asymmetry of magnets. Gertz and Khitun [71] investigated the trapping of red blood cells (RBC), using magnetic nanoparticles. The schematic illustration of the experimental setup is shown in Figure 17a. Two Cu wires covered by silicon dioxide were embedded in the working device. A power supply was employed to provide micro-electromagnet field to trap RBCs. Without the activation of current, the RBCs were randomly distributed in the channel; once the current was on, the RBCs were trapped around the wire (Figure 17b,c). The strength of the magnetic field influenced the trapping procedure significantly.

Optical Tweezers
Optical tweezer has been regarded as an effective method to trap cell/particle at microscale [43]. Optical tweezer employs a focused laser beam to induce optical force, to realize trapping function. However, the capture efficiency of optical tweezer is not high sometimes, due to the low refractive index contrast of some biological cells [72,73]. It is still a challenge to improve the trapping efficiency of biological cells by using optical tweezer [74,75]. Recently, some researchers employed the thermal

Optical Tweezers
Optical tweezer has been regarded as an effective method to trap cell/particle at microscale [43]. Optical tweezer employs a focused laser beam to induce optical force, to realize trapping function. However, the capture efficiency of optical tweezer is not high sometimes, due to the low refractive index contrast of some biological cells [72,73]. It is still a challenge to improve the trapping efficiency of biological cells by using optical tweezer [74,75]. Recently, some researchers employed the thermal effect induced by the optical absorption, to enhance trapping efficiency. The combination of thermal effect and natural convection flow could trap the cells into a hotter region [76]. The optical force imposed on a single particle is expressed as follows: T M = 0.5Re εEE * + µHH * − 0.5 ε|E| 2 + µ|H| 2 I where n is the surface normal vector, I is the unit dyadic, and ε and µ are the electric permittivity and magnetic permeability of the surrounding medium. E is electric field, and H is magnetic field. Li et al. [77] reported a new type of optical tweezer that made use of thermophoresis and natural convection flow to trap and arrange erythrocytes ( Figure 18). A schematic illustration of the thermophoresis and natural convection flow is shown in Figure 18a. Introduced by the optical absorption through a fiber, a hot zone formed on the quartz plate. This hot zone resulted in a temperature gradient on the working plate. This thermophoresis and natural convection flow under low incident power could trap the erythrocyte effectively. In addition, the optical scattering force under high incident power could be employed to arrange the erythrocytes efficiently (Figure 18b). The erythrocytes were trapped and arranged over a long distance, without injury. The enlarged view and SEM (scanning electron microscope) image of graphene-coated microfiber probe (GCMP) are provided in Figure 18c,d. Figure 18e presents the schematic diagram of the experimental setup: A 980 nm laser was focused on the GCMP through the fiber, and a CCD camera was mounted on the top to monitor the trapping and arrangement process. The AFM (atomic force microscopy) image of an erythrocyte is shown in Figure 18f. Sipova et al. [78] have also employed the photothermal effect, to probe DNA films. DNA cargo from individual gold nanoparticles were successfully trapped and manipulated by optical tweezers. The trapping procedure was affected by the natural convection flow and thermophoretic force on the particles.
Liu et al. [79] developed a microfluidic device to selectively trap Escherichia coli cells in human blood solution, based on size and shape. A fiber optical tweezer was embedded into a T-type microflow channel to realize the trapping Escherichia coli function ( Figure 19). With the help of optical tweezer, the Escherichia coli cells were selectively trapped at the tip of the optical fiber tweezer. The trapping efficiency of E. coli was 39.5%, and the separation efficiency was 100%. The optical force played an important role in the trapping and separation process. and SEM (scanning electron microscope) image of graphene-coated microfiber probe (GCMP) are provided in Figure 18c,d. Figure 18e presents the schematic diagram of the experimental setup: A 980 nm laser was focused on the GCMP through the fiber, and a CCD camera was mounted on the top to monitor the trapping and arrangement process. The AFM (atomic force microscopy) image of an erythrocyte is shown in Figure 18f. Sipova et al. [78] have also employed the photothermal effect, to probe DNA films. DNA cargo from individual gold nanoparticles were successfully trapped and manipulated by optical tweezers. The trapping procedure was affected by the natural convection flow and thermophoretic force on the particles.  Liu et al. [79] developed a microfluidic device to selectively trap Escherichia coli cells in human blood solution, based on size and shape. A fiber optical tweezer was embedded into a T-type microflow channel to realize the trapping Escherichia coli function ( Figure 19). With the help of optical tweezer, the Escherichia coli cells were selectively trapped at the tip of the optical fiber tweezer. The trapping efficiency of E. coli was 39.5%, and the separation efficiency was 100%. The optical force played an important role in the trapping and separation process. Lee et al. [80] employed optical trapping and microfluidics to investigate the mechanism of red blood cell (RBC) aggregation. Schematic illustration of the experimental chamber is shown in Figure  20, including the top view (a) and side view (b). Solution 2 (S2) was introduced into the microfluidic flow channel through pressure supply. The cells located in the larger chamber with Solution 1 (S1). Optical tweezer was employed to trap the cells to the S2. The evidence for the cross-bridge induced interaction of cells was observed in the experiment. The initial solution played an important role in measuring the cell-interaction strength. Lee et al. [80] employed optical trapping and microfluidics to investigate the mechanism of red blood cell (RBC) aggregation. Schematic illustration of the experimental chamber is shown in Figure 20, including the top view (a) and side view (b). Solution 2 (S2) was introduced into the microfluidic flow channel through pressure supply. The cells located in the larger chamber with Solution 1 (S1). Optical tweezer was employed to trap the cells to the S2. The evidence for the cross-bridge induced interaction of cells was observed in the experiment. The initial solution played an important role in measuring the cell-interaction strength. Pilat et al. [81] developed a promising microfluidic device with many functions for assessing the optical trapping experiments quantitatively ( Figure 21). The layout of the microfluidic device was designed to guarantee that the cells could not flow out of the chamber due to their low-diffusion rate. A benchmark for safe and non-invasive optical trapping of Saccharomyces cerevisiae could be achieved by using this configuration. Pilat et al. [81] developed a promising microfluidic device with many functions for assessing the optical trapping experiments quantitatively ( Figure 21). The layout of the microfluidic device was designed to guarantee that the cells could not flow out of the chamber due to their low-diffusion rate. A benchmark for safe and non-invasive optical trapping of Saccharomyces cerevisiae could be achieved by using this configuration.
Pilat et al. [81] developed a promising microfluidic device with many functions for assessing the optical trapping experiments quantitatively (Figure 21). The layout of the microfluidic device was designed to guarantee that the cells could not flow out of the chamber due to their low-diffusion rate. A benchmark for safe and non-invasive optical trapping of Saccharomyces cerevisiae could be achieved by using this configuration.
(d) Zhang et al. [82] proposed and demonstrated a hollow annular-core fiber (HACF) based optical tweezer for living cell trapping and sterile transporting ( Figure 22). A microfluidic channel was embedded in the optical fiber, allowing for the cells/particles flowing through the flow channel. The competition between the optical trapping forces (OTF) and the liquid viscous resistances (LVR) determined the trapping location and moving trajectory. Zhang et al. [82] proposed and demonstrated a hollow annular-core fiber (HACF) based optical tweezer for living cell trapping and sterile transporting ( Figure 22). A microfluidic channel was embedded in the optical fiber, allowing for the cells/particles flowing through the flow channel. The competition between the optical trapping forces (OTF) and the liquid viscous resistances (LVR) determined the trapping location and moving trajectory. Liu et al. [83] proposed and demonstrated an optofluidic strategy to trap and transport cell chain using large-tapered-angle fiber probe (LTAP). In their research, Escherichia coli cells, yeast cells and red blood cells were used to study the feasibility of this approach. Their strategy employed the Liu et al. [83] proposed and demonstrated an optofluidic strategy to trap and transport cell chain using large-tapered-angle fiber probe (LTAP). In their research, Escherichia coli cells, yeast cells and red blood cells were used to study the feasibility of this approach. Their strategy employed the combination of optical force and flow drag force. The experimental results of trapping and transporting E. coli cell chain are shown in Figure 23. The trapping procedure could be controlled by adjusting the laser power and flow velocity. Qi et al. [84] employed optical tweezer and microfluidic devices to trap and sort denitrifying anaerobic methane oxidizing (DAMO) microorganisms. This technique showed many advantages, such as high purity, low infection rates and no harm to cell viability. The schematic illustration of the chip design is shown in Figure 24. Mixed culture and buffer solution were introduced into the microfluidic channel. The optical tweezer was employed, at the outlet, to trap a target DAMO cell (marked black circle) and transport it into the collection channel. This technique could be potentially used for slow-growing microorganisms. Qi et al. [84] employed optical tweezer and microfluidic devices to trap and sort denitrifying anaerobic methane oxidizing (DAMO) microorganisms. This technique showed many advantages, such as high purity, low infection rates and no harm to cell viability. The schematic illustration of the chip design is shown in Figure 24. Mixed culture and buffer solution were introduced into the microfluidic channel. The optical tweezer was employed, at the outlet, to trap a target DAMO cell (marked black circle) and transport it into the collection channel. This technique could be potentially used for slow-growing microorganisms.

Acoustic Trapping
Another method for the active trapping of cells is acoustic actuated cell trapping. Ultrasonic standing waves (USWs) can be employed for contactless cell trapping in microfluidic channels. Acoustic trapping is widely used in microfluidic system for cell trapping, transportation and manipulation.
Yin et al. [85] proposed a particle-based cell manipulation method employing acoustic radiation forces, as shown in Figure 25. Three typical types of particles were selected in their investigation, which were ploy(lactic-co-glycolic acid) (PLGA) microspheres, silica-coated magnetic microbeads and polydimethylsiloxane (PDMS) microspheres. Their responses to ultrasonic standing waves (USWs) demonstrated that the PDMS microspheres were suitable for cell trapping. This proposed method did not have a harmful effect on the cells. Acoustic contrast factor played an important role in the trapping procedure.

Acoustic Trapping
Another method for the active trapping of cells is acoustic actuated cell trapping. Ultrasonic standing waves (USWs) can be employed for contactless cell trapping in microfluidic channels. Acoustic trapping is widely used in microfluidic system for cell trapping, transportation and manipulation.
Yin et al. [85] proposed a particle-based cell manipulation method employing acoustic radiation forces, as shown in Figure 25. Three typical types of particles were selected in their investigation, which were ploy(lactic-co-glycolic acid) (PLGA) microspheres, silica-coated magnetic microbeads and polydimethylsiloxane (PDMS) microspheres. Their responses to ultrasonic standing waves (USWs) demonstrated that the PDMS microspheres were suitable for cell trapping. This proposed method did not have a harmful effect on the cells. Acoustic contrast factor played an important role in the trapping procedure.

Acoustic Trapping
Another method for the active trapping of cells is acoustic actuated cell trapping. Ultrasonic standing waves (USWs) can be employed for contactless cell trapping in microfluidic channels. Acoustic trapping is widely used in microfluidic system for cell trapping, transportation and manipulation.
Yin et al. [85] proposed a particle-based cell manipulation method employing acoustic radiation forces, as shown in Figure 25. Three typical types of particles were selected in their investigation, which were ploy(lactic-co-glycolic acid) (PLGA) microspheres, silica-coated magnetic microbeads and polydimethylsiloxane (PDMS) microspheres. Their responses to ultrasonic standing waves (USWs) demonstrated that the PDMS microspheres were suitable for cell trapping. This proposed method did not have a harmful effect on the cells. Acoustic contrast factor played an important role in the trapping procedure.  Fornell et al. [86] established an microfluidic system to trap hydrogel droplets, using acoustic forces. The experimental setup is shown in Figure 26. A T-shaped microfluidic channel was employed to generate cell-laden droplets. The cell-laden droplets were then cross-linked with UV light. Next, the droplets were introduced into a second microfluidic channel, where they were trapped by acoustic forces. The droplets could be trapped at a flow speed of up to 3.2 mm/s. The trapping process was realized by the acoustic forces. Fornell et al. [86] established an microfluidic system to trap hydrogel droplets, using acoustic forces. The experimental setup is shown in Figure 26. A T-shaped microfluidic channel was employed to generate cell-laden droplets. The cell-laden droplets were then cross-linked with UV light. Next, the droplets were introduced into a second microfluidic channel, where they were trapped by acoustic forces. The droplets could be trapped at a flow speed of up to 3.2 mm/s. The trapping process was realized by the acoustic forces. Lim et al. [87] reported a novel method for evaluating the acoustic trapping performance by tracking the motion of a microparticle. The acoustic trapping force was assessed based on a series of microscopy images obtained from a high-speed camera and a high-resolution microscopy. This method could be employed to estimate cell membrane deformability. The experimental setup and procedure for measuring the trapping forces are shown in Figure 27. A microparticle was randomly selected and trapped by the acoustic tweezer. Then the transducer was turned off and translated by a certain distance of 250 μm. After that, sinusoidal bursts were applied to the transducer, and the motion of the attracted particle toward the acoustic focus center was recorded, which could be further analyzed to estimate the trapping force.  Lim et al. [87] reported a novel method for evaluating the acoustic trapping performance by tracking the motion of a microparticle. The acoustic trapping force was assessed based on a series of microscopy images obtained from a high-speed camera and a high-resolution microscopy. This method could be employed to estimate cell membrane deformability. The experimental setup and procedure for measuring the trapping forces are shown in Figure 27. A microparticle was randomly selected and trapped by the acoustic tweezer. Then the transducer was turned off and translated by a certain distance of 250 µm. After that, sinusoidal bursts were applied to the transducer, and the motion of the attracted particle toward the acoustic focus center was recorded, which could be further analyzed to estimate the trapping force. Fornell et al. [86] established an microfluidic system to trap hydrogel droplets, using acoustic forces. The experimental setup is shown in Figure 26. A T-shaped microfluidic channel was employed to generate cell-laden droplets. The cell-laden droplets were then cross-linked with UV light. Next, the droplets were introduced into a second microfluidic channel, where they were trapped by acoustic forces. The droplets could be trapped at a flow speed of up to 3.2 mm/s. The trapping process was realized by the acoustic forces. Lim et al. [87] reported a novel method for evaluating the acoustic trapping performance by tracking the motion of a microparticle. The acoustic trapping force was assessed based on a series of microscopy images obtained from a high-speed camera and a high-resolution microscopy. This method could be employed to estimate cell membrane deformability. The experimental setup and procedure for measuring the trapping forces are shown in Figure 27. A microparticle was randomly selected and trapped by the acoustic tweezer. Then the transducer was turned off and translated by a certain distance of 250 μm. After that, sinusoidal bursts were applied to the transducer, and the motion of the attracted particle toward the acoustic focus center was recorded, which could be further analyzed to estimate the trapping force.  Wu et al. [88] reported a simple and reliable method to generate multicellular spheroids, using acoustic method ( Figure 28). Their device consisted of capillaries, a standing surface acoustic wave (SSAW) generator, a pair of interdigital transducers (IDTs) and a piezoelectric substrate. Once the radio frequency signal was applied, a periodic distributed acoustic field could be formed in the capillary. Due to the gradient of the acoustic field, an acoustic radiation force was generated, and pressure node array was generated. The suspended cells in the capillary were pushed by the acoustic force to the pressure nodes and assembled into spheroids there. Wu et al. [88] reported a simple and reliable method to generate multicellular spheroids, using acoustic method ( Figure 28). Their device consisted of capillaries, a standing surface acoustic wave (SSAW) generator, a pair of interdigital transducers (IDTs) and a piezoelectric substrate. Once the radio frequency signal was applied, a periodic distributed acoustic field could be formed in the capillary. Due to the gradient of the acoustic field, an acoustic radiation force was generated, and pressure node array was generated. The suspended cells in the capillary were pushed by the acoustic force to the pressure nodes and assembled into spheroids there. Lu et al. [89] developed a microfluidic platform, to trap and isolate cancer cells based on their size, using acoustic microstreaming ( Figure 29). With the activation of acoustic microstreaming, the microtrap would discriminate and trap the cancer cells at the vicinity. The tunable and reversible properties of the acoustic microstreaming produced by the micropillar trap played an important role which affected the trapping efficiency. Hayakawa et al. [90] also proposed similar methods for trapping single motile cell. Lu et al. [89] developed a microfluidic platform, to trap and isolate cancer cells based on their size, using acoustic microstreaming ( Figure 29). With the activation of acoustic microstreaming, the microtrap would discriminate and trap the cancer cells at the vicinity. The tunable and reversible properties of the acoustic microstreaming produced by the micropillar trap played an important role which affected the trapping efficiency. Hayakawa et al. [90] also proposed similar methods for trapping single motile cell.
Xu et al. [91] developed an improved method to separate sperm cells from dilute "large volume" samples that contained an abundance of female DNA, by using bead-assisted acoustic trapping ( Figure 30). One PDMS fluid layer was sandwiched between two glass reflecting layers, forming a resonator. Through the employment of this three-layer structure, trapping nodes were generated based on ultrasonic standing waves. The addition of polymeric beads with a critical concentration in the dilute sample was found to initiate the aggregation and improve the sperm cell trapping significantly, while not affecting the DNA extraction and PCR (polymerase chain reaction). Hence, this successful bead-assisted trapping of sperm cells in the enclosed glass-PDMS-glass microdevice suggested that acoustic differential extraction (ADE) could be a useful tool for the processing of real forensic samples. Xu et al. [91] developed an improved method to separate sperm cells from dilute "large volume" samples that contained an abundance of female DNA, by using bead-assisted acoustic trapping ( Figure 30). One PDMS fluid layer was sandwiched between two glass reflecting layers, forming a resonator. Through the employment of this three-layer structure, trapping nodes were generated based on ultrasonic standing waves. The addition of polymeric beads with a critical concentration in the dilute sample was found to initiate the aggregation and improve the sperm cell trapping significantly, while not affecting the DNA extraction and PCR (polymerase chain reaction). Hence, this successful bead-assisted trapping of sperm cells in the enclosed glass-PDMS-glass microdevice suggested that acoustic differential extraction (ADE) could be a useful tool for the processing of real forensic samples. Lu et al. [92] investigated topographical manipulation of microparticles and cells, using acoustic microstreaming. This technique was named as acoustic topographical manipulation (ATM). The working principle is shown in Figure 31a. The microparticles were introduced into the aqueous microfluidic system and deposited on the bottom of the cell surface. Some particles became obstacles due to the existence of electrostatic and van der Waals interactions. A localized microstreaming around the obstacle would be formed upon the application of a standing acoustic wave filed. The Lu et al. [92] investigated topographical manipulation of microparticles and cells, using acoustic microstreaming. This technique was named as acoustic topographical manipulation (ATM). The working principle is shown in Figure 31a. The microparticles were introduced into the aqueous microfluidic system and deposited on the bottom of the cell surface. Some particles became obstacles due to the existence of electrostatic and van der Waals interactions. A localized microstreaming around the obstacle would be formed upon the application of a standing acoustic wave filed. The acoustic microstreaming force, as well as radiation forces, could trap the microparticles at the vicinity of the obstacles. Noteworthy, the localized acoustic microstreaming vortex, as the manipulating force, would guide the topographic movement of the microparticle around the obstacle (Figure 31b-e). Dependence of the manipulating microparticle's velocity on the applied driving frequency and voltage of the acoustic transducer were also studied in the work (Figure 31f,g). Lu et al. [92] investigated topographical manipulation of microparticles and cells, using acoustic microstreaming. This technique was named as acoustic topographical manipulation (ATM). The working principle is shown in Figure 31a. The microparticles were introduced into the aqueous microfluidic system and deposited on the bottom of the cell surface. Some particles became obstacles due to the existence of electrostatic and van der Waals interactions. A localized microstreaming around the obstacle would be formed upon the application of a standing acoustic wave filed. The acoustic microstreaming force, as well as radiation forces, could trap the microparticles at the vicinity of the obstacles. Noteworthy, the localized acoustic microstreaming vortex, as the manipulating force, would guide the topographic movement of the microparticle around the obstacle (Figure 31b-e). Dependence of the manipulating microparticle's velocity on the applied driving frequency and voltage of the acoustic transducer were also studied in the work (Figure 31f,g). Meng et al. [93] reported an improved sonoporation method to trap microcells using cavitation effect as shown in Figure 32. Multiple rectangular microchannels of uniform size were embedded staggered along the main microflow channel to produce microbubble array. The microbubble array oscillated with almost the same amplitude and resonant frequency, resulting in the homogeneous sonoporation. The microcells were trapped at the corner of the rectangular micro side channels due to the generated acoustic radiation forces introduced by the oscillating microbubbles.
Meng et al. [93] reported an improved sonoporation method to trap microcells using cavitation effect as shown in Figure 32. Multiple rectangular microchannels of uniform size were embedded staggered along the main microflow channel to produce microbubble array. The microbubble array oscillated with almost the same amplitude and resonant frequency, resulting in the homogeneous sonoporation. The microcells were trapped at the corner of the rectangular micro side channels due to the generated acoustic radiation forces introduced by the oscillating microbubbles.

Conclusions
This paper reviews the different methods for cell trapping. At the early development stages, hydrodynamic trapping, which is a contact-based method, has been proved to be efficacious in cell trapping. Many researchers focused on contact-based cell trapping devices due to their simple fabrication processes. With the development of microfluidic technologies, an increasing number of contactless cell trapping methods have been proposed and reported. Through contactless cell trapping approaches, cell trapping can be achieved to satisfy various requirements. Cell trapping at a precisely defined location has been receiving more and more attention from both the engineering community as end-user and research community as research tool.
Magnetophoresis is highly recommended for cell trapping in microfluidic system. This technique achieves cell trapping function without changing the physical properties of sample solution, such as pH value, ion concentration and temperature. Being easy to design, easy to operate and low-cost make it a popular option in the scientific research community.

Conclusions
This paper reviews the different methods for cell trapping. At the early development stages, hydrodynamic trapping, which is a contact-based method, has been proved to be efficacious in cell trapping. Many researchers focused on contact-based cell trapping devices due to their simple fabrication processes. With the development of microfluidic technologies, an increasing number of contactless cell trapping methods have been proposed and reported. Through contactless cell trapping approaches, cell trapping can be achieved to satisfy various requirements. Cell trapping at a precisely defined location has been receiving more and more attention from both the engineering community as end-user and research community as research tool.
Magnetophoresis is highly recommended for cell trapping in microfluidic system. This technique achieves cell trapping function without changing the physical properties of sample solution, such as pH value, ion concentration and temperature. Being easy to design, easy to operate and low-cost make it a popular option in the scientific research community.