Separation of micro-particles utilizing spatial difference of optically induced dielectrophoretic forces
- First Online:
- Cite this article as:
- Lin, W., Lin, Y. & Lee, G. Microfluid Nanofluid (2010) 8: 217. doi:10.1007/s10404-009-0457-y
- 270 Views
This paper presents new methods to accurately separate micro-particles with different sizes using optically induced dielectrophoretic (ODEP) forces. It is found that the strength of the ODEP force induced on the hydrogenated amorphous silicon surface is determined by the color, line-width and intensity of the optical beams, which provide an innovative design for particle separation. Two linear-segment virtual electrodes which produced the ODEP forces were firstly defined by illuminating lights onto a photoconductive chip. One moving line and one stationary illuminated line were used to generate a stronger and a weaker ODEP force, respectively. The micro-particles were then continuously pushed forward by the stronger ODEP force. As these lines approached each other, larger micro-particles entrained by the higher ODEP forces were squeezed through the stationary electrode and subsequently separated from the smaller particles. With this approach, continuous particle separation can be automatically achieved within a few seconds. This developed method may be promising for a variety of applications such as cell-based assays and sample pretreatment using micro-particles.
KeywordsOptically induced dielectrophoreticParticle separationMicrofluidicsMEMS
Fetal bovine serum
Incident photon-to-current conversion efficiency
Liquid crystal display
Optically induced dielectrophoretic
Plasma enhanced chemical vapor deposition
List of symbols
Electric field strength
Radius of the spherical particle
Terminal velocity of the spherical beads
Electrical permittivity of the surrounding buffer
Dynamic viscosity of the fluid
Manipulation and separation of particles, beads and cells are essential for a variety of applications, including sample pretreatment (De mello and Beard 2003), cell manipulation (Jager et al. 2000) and diagnosis (Gascoyne et al. 2004). Biological samples usually consist of complicated compositions, which may have an extremely low trace of target cells, proteins or nucleic acids. Targets of interest are usually expected to be separated from substances that may interfere with the detection process such that the limit of detection can be enhanced. Cells and particles in bio-samples can be separated according to their size differences (Vulto et al. 2006; Zhang et al. 2006), electrical properties (Iliescu et al. 2007; Ohta et al. 2007a) and other physical properties. Several mechanisms such as electrical (electrophoresis or dielectrophoresis) (Choi and Park 2005; Voldman 2006), magnetic (Choi et al. 2000; Berger et al. 2001; Lien et al. 2008) and hydrodynamic forces (Yamada et al. 2007; Choi et al. 2007) have been widely explored for the purpose of particle/bead/cell separation. For instance, hydrodynamic filtration has been presented as size-dependent separator. Multiple side-branch channels are employed to continuously sort particles suspension in the main channel according to their difference in size. Particles smaller than a specific size were removed from the mainstream while larger particles are still focused onto a sidewall in the microchannel (Yamada et al. 2007). These hydrodynamic systems usually require stable flows to maintain high separation efficiency.
Among the mechanisms mentioned above, the dielectrophoretic (DEP) force is commonly employed for manipulation and separation of dielectric particles and cells (Hughes 2002; Kang et al. 2008). It provides a direct and convenient way for particle separation without a complex sample preparation process. The DEP force has been widely exploited to separate particles with different sizes or dielectric properties. The particles can be either attracted or repelled while exposing in a non-uniform electric field. In general, this mechanism has been widely used to distinguish viable (live) and nonviable (dead) cells (Doh and Cho 2005; Markx et al. 1994). Another application of the DEP force for particle separation is based on the use of barriers built by DEP force field (Kang et al. 2008; Urdaneta and Smela 2008). The electrodes are applied to induce a negative DEP force to change the trajectories of the selected particles. Since the strength of the DEP force varies with the size of particles, they are either deflected by the negative DEP force or penetrate the barriers at certain flow velocities and hence are separated. However, when the mixed particles flow through the barriers with an excessively high flow rate, all micro-particles might penetrate the barrier. Thus particles with different sizes can not be separated with a high flow rate. This would decrease the throughput of DEP separation process. Furthermore, conventional DEP applications usually require miniature electrodes, created by fabrication processes similar to those used in the microelectronics industry, to produce forces on the dielectric particles. These fixed metal electrodes creating by conventional microelectromechanical system fabrication process lack the flexibility in changing their geometric configuration once they have been made. To improve the capability of DEP manipulation, the CMOS-based device was reported to provide standard grid patterns of cages over several subsets of the electrode arrays to realize a DEP trap array system in any pathway and geometric configuration (Fuchs et al. 2006; Maranesi et al. 2004). However, the fabrication cost of CMOS-based devices is relatively higher than the optoelectronic tweezers (OET)-based system which is adopted in this study.
Alternatively, optical tweezers provide another platform to manipulate a single particle. Several applications using the optical tweezers have been successfully demonstrated in the literature, including particle/cell sorting (MacDonald et al. 2003; Wang et al. 2004) and multiple optical traps (Grier 2003). An optical switching method was used for rapid sorting of selected cells. A fluorescence-activated microfluidic system was used to distinguish cells and stably switch them to certain trajectories (Wang et al. 2004). It provides a high-resolution and is suitable for analysis of small volumes of cells. However, optical tweezers have a drawback in their high optical power requirements, which might result in optical or even thermal damage to bio-samples (Neuman et al. 1999; Liu et al. 1996; Calmettes and Berns 1983; Chapman et al. 1995). Such damage induced by the intense trapping light is usually referred to as photodamage caused by transient local heating and photochemical reaction of specimens. It should be noticed that the damage induced by the intense trapping light to the specimens is a significant problem for some optical trapping studies. Besides, in order to trap and manipulate target cells or beads through the gradient force of the laser beam, several complex instruments such as high-numerical aperture lenses and highly precise motion stage are inevitable to realize the optical trapping and manipulation. Indeed, the relatively complicated experimental setup may limit the practical applications of the optical tweezers.
Recently, OET has been demonstrated as a promising technique to manipulate dielectric and metallic particles such as polystyrene beads, living cells and metallic nanowires (Chiou et al. 2005b; Ohta et al. 2007b; Jamshidi et al. 2008). The OET system with pattern-less electrodes has been used to optically induce the DEP force, which is produced by the incorporation of a photoconductive layer on a single microchip. Thus, the pattern-less electrodes without photolithography process can be regarded as virtual electrodes to generate DEP forces on dielectric particles and metallic objects. Furthermore, the OET system also requires much lower optical power when compared to the optical tweezers. The minimum optical power of the optical tweezers, which is required to maintain the stability of the trapped particle, is about 1 mW (Curtis et al. 2002). As mentioned above, the high light intensity requirement for optical traps would result in photodamage to specimens. The OET system promises a relatively low optical power, which is 100,000 times less than optical tweezers (Chiou et al. 2005b) to avoid the damage to biological samples. The use of the OET system has been demonstrated to concentrate, transport and separate particles by the virtual electrode with any geometric configuration or movement. It has been presented that the particles with different sizes can be sorted utilizing a dynamic moving light beam (Chiou et al. 2005a). The light beam generates a non-uniform electric field near the surface of the light-patterned areas, and particles near the light beam experience a negative DEP force. The particles with different sizes will have different relative distances to the center of the scanning light beam because DEP force is very sensitive to the size of the particle. Based on this principle, randomly located particles in the solution are sorted out successfully, while the light beam is scanning across the surface. If the scan rate is increased beyond the maximum scanning velocity of particles in the fluid which is determined by the balance between the DEP force and the viscous force, the particle will be levitated vertically due to non-uniformity of the electric field. The escape speed is determined by the size of the particles. Therefore the scan rate of light beams must be accurately controlled to avoid influencing the efficiency and resolution between particles, especially when it is applied to separate particles with small size variation.
The OET system is based on optically induced DEP (ODEP) forces, which are induced by illuminating optical images onto photoconductive materials (e.g., hydrogenated amorphous silicon) (Street 1991). It is found that the strength of the ODEP force induced on the hydrogenated amorphous silicon surface is determined by the color of the projecting light from a commercial liquid crystal display (LCD) projector. Under different colors, the hydrogenated amorphous silicon has different abilities in creating electron-hole pairs (Carlson and Wronski 1976; Connell and Pawlik 1976; Carlson 1980) so that the corresponding strength of the ODEP force under different illumination colors is different. It has also been found that the other operating modes for the light patterns such as the line-widths (Ohta et al. 2007b) and intensities (Chiou et al. 2003; Choi et al. 2008) can affect the strength of the ODEP force.
In this paper, we demonstrate a new method using virtual electrodes created by using different colors, line-widths and intensities to generate two line-shaped lights with different ODEP forces is proposed in this study for continuous micro-particle separation. With this approach, separation of micrometer-sized particles can be achieved by using these lines to squeeze mixed particles through this virtual filter such that they can be separated and sorted automatically. By using these three parameters influencing the strength of ODEP forces, it is feasible to achieve efficient separation with a flexible and convenient process.
2 Materials and methods
2.1 Operating principle of ODEP
2.2 Micro-particle separation process
Different designed images of light-patterns can be used for manipulation and separation of polystyrene beads. In this new method for separation, two linear-segment virtual electrodes inducing ODEP forces with different strengths were used to separate beads with different sizes. The separation process is performed in three major steps, including concentration, squeezing and separation (Fig. 1b). First, all the beads randomly located on the surface can be concentrated between two lines generating negative ODEP forces when the moving line with the stronger ODEP strength is scanned from left to right (Fig. 1b-i, ii). As the space between the two lines becomes smaller, it begins to squeeze the beads (Fig. 1b-iii). Finally, when the space is smaller than the size of the larger beads, the larger beads jump to the right side of the stationary line, which has a smaller ODEP strength (Fig. 1b-iv). The stronger line-shaped virtual electrode provides a lager electric field in the vertical direction. Hence these two lines with different strengths can be regarded as two “virtual” walls with different heights. It can be then envisioned that the beads can “jump” across the lower “wall” and separated to the other side when the two “walls” squeeze beads into to a certain space. This new method can continuously separate beads with different sizes by simply controlling the final space of the two lines. Note that this method can be easily extended to multiple lines with different strengths such that multiple sized beads can be separated with high discrimination efficiency. In this study, micrometer-sized polystyrene beads are used in the liquid layer of the chip to perform the separation of beads. In addition, the strength of the ODEP force can be determined by three characterizations of generated light beams, including color, line-width and intensity. Thus, the projected lights are expected to be patterned with incremental ODEP forces into various differences by using a combination of all these three parameters, so that the separation of multiple-sized beads with a high-resolution can be achieved.
Furthermore, the cells, proteins or DNA can be coated onto the surface of the antibody-coated magnetic beads for future biomedical applications (Miltenyi et al. 1990; Lien et al. 2007). By using the same approach reported in this study, different types of target samples can be incubated and conjugated onto the beads with different sizes, followed by separating and concentrating the beads in an automatic fashion. The collection of the separated beads can be realized by using magnets and pipettes.
2.3 Fabrication process
As shown in Fig. 1a, a 1-mm-thick ITO glass substrate was first deposited with a 10-nm thick molybdenum layer. Then a 1-μm thick hydrogenated amorphous silicon layer (a-si:H) was deposited by a plasma enhanced chemical vapor deposition process. Note that a thin layer of molybdenum was used to reduce contact resistance between ITO and amorphous silicon. Note that the a-si:H layer is a photoconductive material for producing ODEP forces. The top ITO glass layer was then bonded with the bottom layer using a double-sided tape (DEV-8930, Adhesives Research) as a spacer. The thickness of the tape (30 μm) determines the gap between the top and bottom layers, and also the strength of the generated electric field when an AC voltage is applied. Two via holes were drilled on the top layer as the fluidic inlet and outlet and were used for injecting beads into the device.
2.4 Experimental setup
Polystyrene beads (Duke Scientific, USA) were mixed in a liquid solution consisting of deionized water with 1% fetal bovine serum (FBS) to form a buffer solution with a low electrical conductivity (16.5–17.5 mS/m). The FBS was used to reduce the adhesion of polystyrene beads on the chip surface. Then the mixed solution was injected into the liquid layer via the inlet hole. In this study, an AC power source with a frequency of 100 kHz and a magnitude of 36 Vpp was applied across the top and bottom layers to generate an electrical field. Under these operating conditions, the particles experienced a negative ODEP force and were pushed away from the light-patterned region, so that the particles were moved forward by a light scanning across the surface. The magnitude of the applied voltage and the distance between the top and bottom surfaces are two major factors that influence the strength of the ODEP force, which has been optimized in advance to produce an ODEP force for bead manipulation.
As shown in Fig. 1c, the experimental setup consisted of an image generation system to project virtual electrodes onto the chip, and an image acquisition system to observe bead separation process. To construct the virtual electrodes, a commercially available LCD projector (PJ1172, Viewsonic, Japan) with a spatial resolution of 1,024 pixels × 768 pixels was used as a light source to excite electron-hole pairs in the amorphous silicon layer. The use of a commercial LCD projector in our research provides a convenient and flexible approach in the manipulation of particles when compared to prior OET experiments utilizing the other light source, such as single-wavelength laser beam, mercury or halogen lamp or a LED integrated with a digital-micromirror device. The color, line-width and brightness of the projected patterns can be easily adjusted by a computer connected to a projector. Furthermore, a 50× objective lens (Nikon, Japan) was mounted between the projector and the ODEP chip to focus and to collimate the projected light onto the chip to form virtual electrode patterns. The complicated movement of the patterned virtual electrodes can be easily realized by a commercial program, such as FLASH. The image acquisition system was placed on top of the chip to observe bead separation. It consisted of a charge-coupled device (SSC-DC80, Sony, Japan) camera, an optical microscope (Zoom 125C, OPTEM, USA) and a computer equipped with an image acquisition interface card. The AC voltage was supplied by a function generator (Model 195, Wavetek, UK) and an amplifier (790 Series, AVC Instrumentation, USA).
3 Results and discussion
The characterization of the new design for achieving high-resolution separation of micro-particles is firstly performed. The application of the line-shaped virtual electrodes in different colors, line-widths and intensities generate dissimilar strength in negative ODEP forces, mainly based on their abilities in creating electron-hole pairs in amorphous silicon. Then three parameters influencing the strength of ODEP forces will be discussed individually to enable various kinds of geometry in light patterns and to achieve accurate and efficient separation. Polystyrene beads can be successfully separated under these different modes (color, line-width and intensity) of illuminated lines and the combination of all these parameters.
3.1 Particle separation using different colors
In order to calculate the magnitude of the ODEP forces generated by the virtual electrodes with different colors, the terminal velocity of the polystyrene beads pushed by the linear light segments was measured. Figure 2b, c shows the experimental results for the terminal velocities of the 10- and 20-μm beads pushed by various scanning lights (blue, green and red). It can be observed that the green light pushes beads at the highest velocity, with respect to the others when these lines are generated with the same line-width and light intensity. Therefore this result indicates that the green light induces a higher ODEP force. According to Stokes’ law, the DEP force (F) can be calculated by the terminal velocity (v) of the spherical beads when the ODEP force is balanced with the viscous drag force in fluid, i.e., F = 6πrηv, where r is the radius of the spherical particle, η is the dynamic viscosity of the fluid (1.002 × 10−3 Ns/m2 at 20°C for water). The generated ODEP forces on the 20-μm polystyrene beads are then calculated to be 9.42, 6.28 and 5.83 pN for green, red and blue lights, respectively (Fig. 2b). Similarly, the magnitudes of the ODEP forces on 10-μm polystyrene beads are 3.36, 2.24 and 2.03 pN for green, red and blue lights (Fig. 2c).
Since the final space between the stationary and scanning lines has been one of the most important parameters for efficient separation in this developed method. Actually, the final space between the stationary and scanning lines is directly related to the size of separated particles. The linear-segment virtual electrodes induce a negative ODEP to repel beads, so that the two lines with different ODEP strengths construct a region constraining the particles between them. If the space of the region becomes smaller than the size of the large particles of the population, the large particles would be repelled out of the region by the negative ODEP force. Therefore, the final space between the stationary and scanning lines must be well optimized to push the lager particles to the other side of the stationary line while still keeping the smaller particles.
Therefore many scanning lines are constructed here in order to arrange the randomly located beads and transport them to separate continuously. The final space can be decreased step by step automatically (Fig. 3b-i, iii) to find the optimal distance to separate any different sized beads. Therefore, a suitable final space for any variation in diameter can be found by the computer program automatically. Furthermore, Fig. 3c shows the successful separation of 4.5- and 2.8-μm-diameter polystyrene beads. Thus, the resolution that this new method can achieve for particle separation is around 1.7 μm. The manipulation conditions are the same as mentioned above, while the line-width of the scanning and stationary lines are all extended to 28 μm to form a higher ODEP force. Besides, the separation resolution is determined by the line space between the moving and stationary lines, which is directly related to the accuracy of the control on the scanning light beam. If the movement of the scanning line is not controlled accurately, all of the particles with small variation in diameter might be pushed across the stationary line, causing the failure in separation. In our experiment, the scanning lines can move forward with a step of 2 μm.
Experimental results show that the green light induces a stronger ODEP force than the red and blue lights. It has been reported that amorphous silicon has better collection efficiency at the wavelength of green light (Carlson and Wronski 1976). Although the absorption of amorphous silicon for blue lights shows a better result according to the absorption coefficient curve, most of the energy in the incident blue light was found to be dissipated as thermal energy and did not contribute to the photoconductivity effect. Furthermore, our experimental results represent a similar trend when it is compared to the incident photon-to-current conversion efficiency (IPCE) curve. The photoconductivity effect within the amorphous silicon can be determined by IPCE. The results measured by the IPCE represent the ability of light at different wavelengths to induce the ODEP force in the substrate material. As mentioned above, the green light induces a strongest ODEP force. The peak of the IPCE in the amorphous silicon was found to be located around 550–570 nm (Loveland et al. 1973; Liu et al. 2001) depending on the structure of the amorphous silicon. A spectrum analysis of the light projector (measured by PR-650) also confirms that the strength of the induced force is the highest for the green light. Note that the wavelength of the green light in our projector ranges from 500 to 590 nm while the blue light is from 420 to 500 nm, and red light is from 590 to 730 nm (Fig. 1d). From Eq. 1, the induced ODEP force is also dependent on the size of the beads (r3). Therefore, the larger beads (20-μm) experience a stronger force in the fluid. However, the measured ODEP forces from terminal velocity measurement are much less than the calculated values. This is due to the fact that particles were first levitated in the vertical direction by the ODEP force. Then, as the particles were pushed further away from the virtual electrode and the surface, the effective applied force decreases. As a result, the measured terminal velocity is lower than the calculated value.
3.2 Particle separation using different light line-widths
Although green light can induce the strongest ODEP force, white light is more suitable for practical applications. In fact, white light generates a much stronger force on the particles under the same conditions (Fig. 2b, c) because it is a sum of RGB lights. The larger magnitude of the DEP force can improve the efficiency and resolution of particle separation. Therefore, white light is chosen here to generate light line segments with different line-widths.
The effect of the line-width on DEP forces is also investigated. It is found that the terminal velocities of the beads are dependent on the width of the scanning virtual electrodes. Figure 4b, c shows the experimental results for this test. It is observed that the wider linear light segments can induce stronger ODEP forces. For example, a 28-μm-wide line produces a terminal velocity as high as 261.3 μm/s for 20-μm beads and 142.5 μm/s for 10-μm beads. Similarly, the terminal velocities are used to calculate the induced ODEP forces. The induced ODEP forces on the 20-μm beads are 49.34, 34.99 and 17.94 pN for 28-, 21- and 14-μm lines, respectively (Fig. 4b). The trend for the 10-μm beads is similar to the one for 20-μm beads (Fig. 4c).
In addition to the line-shaped virtual electrodes, the light beam can also be patterned in various geometric shapes. For example, another experiment shows that the light beams can also be defined as concentric rings. As shown in Fig. 4d, the outer ring is designed to be in the line-width of 28 μm while the inner ring has a width of 14 μm. They are all white lights with the same intensity of 7.82 W/cm2. When the outer rings start to move inward, the particles with different sizes can be concentrated, separated and then transported to the designed location. Therefore, a simple device for sample concentration, separation and transportation can be realized onto a single micro chip. The line-width of outer rings can be patterned in decreasing with the time, in order to construct a lager variation in the strength for a higher separation efficiency. The entire concentration and separation process can be finished within 12 s.
3.3 Particle separation using different light intensities
The magnitude of the ODEP forces induced by light with different intensities is investigated. Figure 5b, c shows the experimental results for this test. The terminal velocities of the particles for the green lights with a brightness of 255 (3.73 W/cm2) is calculated to be 126.96 μm/s for 20-μm beads, and 81.61 μm/s for 10-μm beads. The terminal velocities when using green lines with a brightness of 192 (1.46 W/cm2) is 49.88 and 35.62 μm/s for 20- and 10-μm beads, respectively.
Experimental results reveal that lights with different line-widths and intensities can induce ODEP forces with different strengths. It was reported that changes in line-width and light intensity can generate different photoconductivity levels on the amorphous silicon surface. The wider electrodes in our system form a more evident non-uniform electric field and thus generate a stronger ODEP force to achieve a higher terminal velocity (Ohta et al. 2007b) (Fig. 4b, c). It was reported that an illumination intensity of 100 W/cm2 can increase the photoconductivity of the amorphous silicon layer by five orders of magnitude (Wei and Lee 1993). Therefore, an illumination with a stronger intensity can generate more electron-hole pairs to make the voltage drop to the liquid layer more evident for particle manipulation. In our experiment, the lines with the higher brightness produce almost a 2.5 times larger intensity than the dimmer lines.
Microfluidic devices developed for size-dependent separation usually require a delicate fabrication process to form unique microstructures in the fluid channel for changing the trajectories of the beads, so that these beads can be separated. The new design presented in this study can manipulate micro-particles with different sizes by simply using ODEP technology. The flexibility in this method is that particle separation can be conveniently achieved by simply changing light patterns with different colors, intensities and line-widths. For particles with similar sizes, further modifications of this approach can be used for separation. First, multiple lines with different intensities can be used such that multiple bead separation with a high discrimination capability can be achieved. Particles with different sizes can eventually be separated if more scanning lines and stationary lines have been deployed. Besides, since the strengths of the induced ODEP forces can be determined by the colors, intensities and line-widths, it is feasible to program projected lights with incremental ODEP forces by using a combination of all these three parameters, so that particles with small variations in diameter can be separated efficiently. Therefore, the proposed system capable of bio-samples separation and transportation on a single microchip may provide a powerful platform for target bio-particles extraction.
The authors would like to thank Chi-Mei Optoelectronics Inc. for their financial support from grant number (96S036). Partial financial support provided to this study by the National Science Council of Taiwan is also greatly appreciated. Authors also thank Dr T.F. Guo for valuable discussion.