Biosensors and Bioelectronics 63 (2015) 371–378
Contents lists available at ScienceDirect
Biosensors and Bioelectronics journal homepage: www.elsevier.com/locate/bios
Screen-printed microfluidic dielectrophoresis chip for cell separation Hongwu Zhu a,b, Xiaoguang Lin a, Yong Su a, Hua Dong a,n, Jianhua Wu b a b
Department of Biomedical Engineering, School of Materials Science and Engineering, South China University of Technology, Guangzhou 510641, China School of Biological Science and Engineering, South China University of Technology, Guangzhou 510006, China
art ic l e i nf o
a b s t r a c t
Article history: Received 1 April 2014 Received in revised form 11 July 2014 Accepted 24 July 2014 Available online 4 August 2014
Dielectrophoresis (DEP), the induced motion of polarizable particles in a non-uniform electric field, has been proven as a perfect candidate to transport, accumulate, separate and characterize micro-/nano-scale bioparticles in microfluidic systems. However, conventional fabrication technologies are complex, timeconsuming and relatively expensive, leading to low throughput of the DEP-based systems. In this paper, we report a novel microfluidic alternating current DEP (AC-DEP) chip fabricated via inexpensive screen printing method. The innovation of our work consists in the extreme simplicity of the fabrication procedure, i.e., the main components, including electrodes and channels, were constructed by layer-bylayer screen printing process, which is especially suitable for high-throughput mass production. Carbon paste, instead of metals, was used to print interdigitated electrodes with semi-3D structure which not only reduces dramatically the chip cost but also increases particle trapping efficiency. To test the chip performance, yeast cells, as model cells, were trapped and separated from a mixed suspension with PS microspheres. Our results show that high capture rate and separation efficiency can be achieved under optimized conditions. & 2014 Elsevier B.V. All rights reserved.
Keywords: Screen-printing Microfluidics chip Dielectrophoresis Cell separation
1. Introduction Manipulation of bio-particles in microfluidic chips including trapping, sorting, patterning, separation and purification of cells, viruses, proteins etc., is crucial for a variety of diagnostic and clinical applications (Becker et al., 1995; Gascoyne et al., 2002; Prinz et al, 2002; Grier, 2003; McCloskey et al., 2003; LapizcoEncinas et al., 2004; Nilsson et al., 2004; Barbulovic-Nad et al., 2006; Yasukawa et al., 2007; Demierre et al., 2007; Cummings and Singh, 2007; Hawkins et al., 2007; Vahey and Voldman, 2008; Kim et al., 2008; Urdaneta and Smela, 2008; Hsiung et al., 2008; Suzuki et al., 2008; Shin et al., 2008; Zhu and Xuan, 2009; Chu et al., 2009; Beech et al., 2009; Cheng et al., 2009; Park et al., 2009; Mernier et al., 2010; Martinez-Duarte et al., 2010; Shafiee et al., 2010; Regtmeier et al., 2010; Baylon-Cardiel et al., 2010). In the past few decades, tremendous techniques have been developed on the basis of various principles such as optical tweezers (Grier, 2003), magnetophoresis (McCloskey et al., 2003), acoustic waves (Nilsson et al., 2004) and electrical means. Among these methods, dielectrophoresis (DEP) is one of the most versatile methods for particle manipulation due to its label-free nature, favorable scaling effects, simple instrumentation and capability to induce both negative and positive forces (Voldman, 2006). Specifically, DEP n
Corresponding author. E-mail address: [email protected] (H. Dong).
http://dx.doi.org/10.1016/j.bios.2014.07.072 0956-5663/& 2014 Elsevier B.V. All rights reserved.
refers to the movement of a particle in a non-uniform electric field as a result of the interaction of the particle's dipole and spatial gradient of the electric field (Cetin and Li, 2011), and can be generated by using either direct current (DC) or alternating current (AC) field. Compared with DC-DEP, AC-DEP is advantageous because of the low operating voltage that prevents a serious Joule heating effect. Moreover, low voltage simplifies the equipment required to produce the electric fields, making AC-DEP compatible with integrated circuits and feasible for battery powered hand-held devices. Although much progress has been achieved, there are still some challenges that need to be addressed before microfluidic AC-DEP devices meet the end users. Usually, AC-DEP devices contain metal electrodes embedded inside the microchannel network. These internal electrodes are planar (2D) ones (i.e. height of the electrodes are in the order of hundred nanometers), and are fabricated via complex, timeconsuming and relatively expensive microfabrication techniques which are not suitable for mass production (Martinez-Duarte, 2012). Therefore, there is still a desire for a low-cost yet highthroughput manufacturing approach to fabricate microfluidic ACDEP chips. In addition, while working with bio-particles, fouling of the metal electrodes (i.e. sample electrolysis) may distort the operation of AC-DEP devices (Gallo-Villanueva et al., 2009). One of the possible solutions is to employ carbon electrode because carbon has a much wider electrochemical stability window than metals commonly used in thin film electrode fabrication such as gold and
372
H. Zhu et al. / Biosensors and Bioelectronics 63 (2015) 371–378
platinum, and affords higher applied voltages in a given solution without electrolyzing it (Gascoyne et al., 1997; Rousselet et al., 1998). Besides, the use of carbon electrodes yields other advantages including excellent biocompatibility, mechanical properties and chemical inertness in almost all solvents/electrolytes. A good case in point is that Martinez-Duarte et al. (2010) and Jaramillo et al. (2010) fabricated 3D carbon electrodes by C-MEMS technique and utilized AC-DEP to trap Escherichia coli bacteria from the mixture with Bacillus cereus bacteria. In our opinion, however, C-MEMS is still complicated and time-consuming in terms of operation conditions. Another alternative way to fabricate carbon electrodes needs to be exploited urgently. In this paper, we report a unique microfluidic AC-DEP chip fabricated via inexpensive screen printing technology. The innovation of our study consists in the extreme simplicity of the fabrication procedure, i.e., the main components, including electrodes and channels, were constructed by layer-by-layer screen printing process, which is especially suitable for highthroughput mass production. Instead of metal electrode, carbon paste was used to print interdigitated electrode with semi-3D structure which not only reduces dramatically the chip cost but also increases particle trapping efficiency. To test the chip performance, yeast cells were trapped and separated from a mixed suspension with polystyrene (PS) microspheres, and the factors influencing DEP behaviors, capture rate and separation efficiency are discussed in detail. We wish that our work could exhibit the feasibility of screen printing technology in fabricating microfluidic AC-DEP chips and benefit their future commercialization.
The DEP force (Fd) acting on a spherical particle of radius r suspended in a fluid with permittivity εm is given as Eq. (1) (Jen et al., 2011): 2 Fd = 2πr 3εm Re ⎡⎣fcm ⎤⎦ ∇Erms
(1)
where Re[fcm] is the real part of the Clausius–Mossotti (CM) factor and Erms is the root-mean-square of the applied electric field. The CM factor (fcm) is a parameter of the effective polarizability of a particle. It varies with the complex dielectric properties of the particle and the surrounding medium, which are functions of the frequency of the applied field. The CM factor for a spherical particle is represented as Eq. (2):
⎡ ε⁎ − ε⁎ ⎤ p m ⎥ fcm = ⎢ ⁎ ⎢⎣ εp + 2εm⁎ ⎥⎦
(2)
⁎ where εp⁎ and εm are the complex permittivities of the particle and the medium, respectively. The complex permittivity is related to the conductivity s and frequency f, as shown in Eq. (3):
jσ 2π f
(3)
where j equals −1 . As a consequence, the CM factor can be further expressed as Eq. (4):
Re[fcm ] =
3. Materials and methods 3.1. Chip fabrication Thick-film screen printing technology was adopted in our study to fabricate microfluidic AC-DEP devices, according to the following protocol. The commercial carbon paste (ED423SS, ACHESON, USA) was first printed on clean glass surface to form interdigitated carbon electrodes and contact wires via a semi-automatic screen printer (F-C4050R, Fufa Company, Shenzhen, China) and a 400 mesh stencil. After curing at 130 °C for 10 min to harden carbon paste, UV curable dielectric paste (YB-1300, Fufa Company, Shenzhen, China) was repeatedly printed and exposed to UV light for 5 min to form microfluidic channels. The depth of the channels was controlled at ca. 50 μm. Finally, a PDMS (Sylgard 184 silicone elastomer kit, Dow Corning) polymer film with inlet and outlet ports was exposed to O2 plasma, pressed on the top of the microfluidic channels and then heated in an oven at 80 °C for 1 h. Two mask screens were used in the above-mentioned process and the patterns were aligned using two microscopes connected to the computer.
3.2. Measurement of DEP spectra of static PS microspheres and yeast cells
2. Theory
ε⁎ = ε −
region with a high-electric-field gradient, or negative (Re[fcm] o0), repelling particles away from the region with a highelectric-field gradient.
(σp − σm )(σp + 2σm ) + 4π 2f 2 (εp − εm )(εp + 2εm ) (σp + 2σm )2 + 4π 2f 2 (εp + 2εm )2
(4)
Close examination of Eq. (4) reveals that the sign of the CM factor is determined by the electrical conductivities of the particle and the medium at low frequencies and the permittivities at high frequencies. According to the polarity of Re[fcm], the DEP force can be either positive (Re[fcm] 40), pulling particles towards the
A testing chamber (size: 1 1 cm2) printed on glass slide with interdigitated carbon electrode was used to measure the sign and magnitude of DEP forces (i.e. DEP spectra) generated on PS microspheres (7 μm and 20 μm, Aladdin, China) and yeast cells that were suspended in solution separately. The solution conductivity was adjusted using KCl with a conductivity meter (DDS-307, Lei-ci Company, China). An arbitrary waveform generator (ED1411, Zhongce Electronics, China) was employed to create the sinusoidal signal (10 V peak-to-peak and frequency range between 30 kHz and 3 MHz) delivered to the interdigitated carbon electrodes. The whole device was mounted on a microscope stage (AO-KV200, Aosvi Company, China) so that the electrode, yeast cells and PS microspheres can be observed in a bright field using a 10 objective lens and a color camera attached to the microscope. To avoid the sudden electrolysis of electrolyte caused by capacitive effects, the magnitude and frequency of AC signals were slowed down gradually during switch-off process.
3.3. Investigation of separation performance in screen-printed microfluidic AC-DEP chip To evaluate the separation performance of screen-printed microfluidic AC-DEP chip, yeast cells and red PS particles ( 7 μm, Aladdin, China) were mixed thoroughly and injected into the chip using a syringe pump (74900-05, Cole-Parmer, USA) at different flow rates (0.1–0.3 mL/h). Various influential factors such as flow rate, frequency, solution conductivity, voltage were tested in detail. After experiment, the sample can be transferred from the DEP device to different collecting tubes.
H. Zhu et al. / Biosensors and Bioelectronics 63 (2015) 371–378
4. Results and discussion 4.1. Fabrication and characterization of microfluidic AC-DEP chip via screen-printing technology Although first demonstrated in the 1950s, it was not until the development of Bio-MEMS in the 1990s that DEP became a popular research field (Martinez-Duarte, 2012). During the whole 1990s, abundant DEP publications appeared using microfabricated metal electrodes. Starting in the 2000s, alternative techniques have been explored to overcome common problems (e.g. electrode fouling) in metal-electrode DEP. Insulator-based DEP and lightinduced DEP are the most significant examples in this period. More recently, new 3D techniques like carbon-electrode DEP, contactless DEP, and the use of either C-MEMS (Jaramillo et al., 2010) or doped PDMS have further simplified the fabrication process. However, the state-of-the-art of fabrication methods still cannot meet all the requirements for practical DEP chips such as low cost, simple process and high throughput.
373
Instead of complicated Bio-MEMS, herein we utilize screen printing to fabricate microfluidic AC-DEP chips. As a thick film technology, screen printing uses a woven mesh to support an inkblocking stencil to receive a desired pattern. Owing to its inexpensive and simple fabrication process, this technique has been widely applied in mass production of printed circuit board, clothing, medical devices, etc. (Dong et al., 2007). In our study, we constructed microfluidic AC-DEP chip by printing conductive carbon paste as the electrode, whilst printing dielectric paste as the channel wall and insulator to separate the contact wire from the solution. Fig. 1a shows the screen printing procedure of the microfluidic AC-DEP chip. To facilitate observation, transparent glass slide was used as substrate. Three kinds of interdigitated electrodes with different configurations were printed, i.e. offset castellated electrodes, non-offset castellated electrodes and interdigitated line electrodes. After drying the electrodes, UV curable dielectric paste was printed using another stencil. Since the printed dielectric film had a typical thickness of ca. 20–30 μm, channel walls with the height of 50–60 μm can be formed by printing repeatedly for 2–3 times. Thereafter, a thin layer of PDMS
Fig. 1. (a) Screen-printing fabrication of microfluidic AC-DEP chip for cell separation: (left) schematic illustration of layer-by-layer screen printing procedure. (1) Clean glass slide as substrate; (2) patterned carbon ink as interdigitated electrode and conductive wires; (3) UV curable dielectric composition for microfluidic channels; (4) PDMS layer with inlet and outlet ports; (middle) configuration of the whole device; (right) three kinds of interdigitated electrodes printed in our study. (5) Offset castellated electrodes; (6) non-offset castellated electrodes; (7) interdigitated line electrodes. (b) Schematic illustration on separation of yeast cells from a mixture with PS particles in screenprinted microfluidic AC-DEP chip. When suspension solution flows into microchannel, yeast cells feel positive DEP forces and thus are trapped onto the electrodes. Meanwhile, PS microspheres feel weak negative DEP forces and are repelled from electrodes. As a result, separation happens between yeast cells and PS microspheres. Fg: force of gravity, Fb: buoyant force, Fd: DEP force, Ff: fluid force, Fr: resistance.
374
H. Zhu et al. / Biosensors and Bioelectronics 63 (2015) 371–378
Fig. 2. Images of screen-printed microfluidic AC-DEP chips with (a) offset interdigitated carbon electrode and (b) non-offset interdigitated carbon electrode.
(with inlet and outlet ports) was treated by O2 plasma, pressed on the top of the printed dielectric layer and then heated at 80 °C for 1 h. The sealing was found to be robust for sample injection and transportation via pumping. Fig. 2 shows the photos of the screen printed microfluidic devices. The thickness of carbon electrode
was measured as ca. 10–15 μm, which shows semi-3D structure with respect to the size of cells. The resolution of stencil for screen printing in our study is ca. 50 μm, so smaller channel/electrode could have been fabricated using this technology. However, in order to achieve high
Fig. 3. 3D simulation of electric field intensity distribution using COMSOL multiphysics software. Top view of (a) interdigitated line electrodes, (b) non-offset castellated electrodes, (c) offset castellated electrodes, and side view of (d) non-offset castellated electrodes. The thickness of electrode and the distance between two microelectrodes were set as 10 and 100 μm, respectively. The applied voltage (VPP) was 10 V. Herein yellow, red and brown colors indicate the high, medium and low field strength, whilst white color indicates the locations without electric field (the lowest field intensity). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
H. Zhu et al. / Biosensors and Bioelectronics 63 (2015) 371–378
throughput separation, larger electrode gap ( 100–150 μm) and wider microchannels (2 mm) were actually employed to make sure that more particles/cells can be trapped via DEP forces, see the discussion later. It should be pointed out that insulator-based DEP chip can be also fabricated by this technology and good performance can be achieved. The work is still under investigation, and will be reported in the future. 4.2. Simulation of electric field intensity on screen-printed carbon microelectrode with different configurations Fig. 3a–d shows the finite element simulation of electric field distribution over carbon microelectrode with different configurations stated above. For interdigitated line electrodes (Fig. 3a), the electrical field intensity (top view, Z¼ 10 μm) increases to the maximum value at the edges of microelectrodes and drops to the minimum value in the center between two neighboring electrodes. For non-offset and offset castellated electrodes (Figs. 3b and c), the
375
strongest and weakest electric fields are located at the electrode tips and intermediate region between electrode tips, respectively. Moreover, the side view image (Fig. 3d) reveals that electrical field intensity decreases along the z-axis above the electrodes. Given the same simulation conditions, i.e., the applied voltage was 10 V and gap between neighboring electrodes was 100 μm, the 2 maximum ∇Erms were 2 1014 m kg2/s6 A2 for interdigitated line electrodes and 5 1015 m kg2/s6 A2 for non-offset and offset castellated electrodes, indicating that castellated electrodes can generate larger DEP forces compared with interdigitated line electrodes. Since screen-printed carbon electrodes possess porous and coarse surface, the electric field is more non-uniform than smooth metal electrodes and thus may further improve DEP forces imposed on particles (Tang et al., 2013). When aware of the locations of the strongest and weakest electric field, we can predict whether particles feel positive or negative DEP forces based on their movements towards or off the electrode edges.
Fig. 4. (a, b) Images of PS microspheres (size: 20 μm; color: white) suspended in aqueous medium (conductivity: 5 μs/cm) before (a) and after (b) applying 20 V/100 kHz AC signals on offset and non-offset electrodes. (c, d) Images of yeast cells (size: 5 μm, color: white) suspended in aqueous medium (conductivity: 5 μs/cm) after applying 10 V/100 kHz AC signals on (c) non-offset castellated electrodes and (d) offset castellated electrodes.
376
H. Zhu et al. / Biosensors and Bioelectronics 63 (2015) 371–378
4.3. DEP behaviors of static PS microspheres and yeast cells Although particle dielectric properties can be determined via typical electrorotation (ROT) methods and thus DEP spectra can be plotted using the dielectric parameters derived from ROT measurements (Becker et al., 1995), we investigated DEP behaviors of particles in static status using the printed castellated electrodes so that the results can be directly used for particle separation in microfluidic chip with the same electrode configuration. In our study, PS microspheres and yeast cells were chosen as particle models due to their similar spherical shape and well-known DEP behaviors, which makes it easy to compare our results with the literature. Fig. 4a and b show images of PS microspheres (size: 20 μm; color: white) before and after applying 20 V/100 kHz AC signals on screen-printed electrodes. As can be seen, PS microspheres are randomly distributed around the electrodes prior to applying AC signals. Once subjected to non-uniform electric field, the majority of PS microspheres assemble into triangle in the intermediate region between electrode tips, and the rest distribute in the center between two neighboring electrodes. For non-offset castellated electrodes, straight chainline are also observed between opposite triangles. Judging from the simulation results, one can deduce that PS microspheres move to the weak electric field and thus bear negative DEP forces. Further measurements reveal that PS microspheres can only feel negative DEP forces, regardless of the applied frequencies and medium conductivities. Fig. 4c and d give images of yeast cells (size: 5–6 μm) after applying 10 V/100 kHz AC signals on non-offset and offset castellated electrodes, respectively. It is clear that yeast cells are trapped to electrode tips with high-electric-field gradient, indicative of positive DEP behavior. Different from PS microspheres, both positive and negative DEP forces are observed on yeast cells under our experimental conditions. Although the absolute value of DEP force is not available, the relative DEP force can be calculated. Considering yeast cells stay static in the beginning, the average DEP forces can be calculated using Eq. (5), derived from Newton's second law:
2mS Fd = 2 t
(5)
where m is the mass of yeast cell, S is average displacement distance of yeast cells in microfluidic chip and t is the average time needed for yeast cells to reach the positions with maximum or minimum electric field intensity. It can be found from Eq. (5) that DEP forces are proportional to the reciprocal of time squared. As a result, DEP spectra can be obtained by plotting normalized DEP force (or namely, the reciprocal of time squared) as a function of input frequencies and medium conductivities. Table 1 shows the average time for yeast cells to reach positions with the maximum or minimum electric field intensity. If yeast cells don't move after applying AC signals, the time was set as infinite. In mediums with conductivities of 5 and 30 μs/cm (Fig. 5a), only positive DEP forces are detected on yeast cells no
Fig. 5. DEP spectra of yeast cells in aqueous mediums with conductivities of (a) 5 and 30 μs/cm and (b) 60 and 100 μs/cm. Input voltage (Vpp): 10 V, frequency range: 30–3000 kHz.
matter how to change applied frequencies in the range of 30– 3000 kHz. The increase of time with frequency means that DEP forces decrease with frequency. Moreover, when input frequencies is less than 1000 kHz, the larger DEP forces in medium with lower conductivity (5 μs/cm) is in good agreement with the deduction from Eq. (4) that the CM factor (i.e. DEP force) is determined by the electrical conductivities of the particle and the medium at low frequencies, which inversely proves the validity of our data. In contrast, when the medium conductivity rises to 60 or 100 μs/cm (Fig. 5b), yeast cells bear negative DEP forces at low frequencies and positive DEP forces at high frequencies, and the crossover frequency increases with the medium conductivity.
Table 1 Average time for yeast cells to reach positions with the maximum or minimum electric field intensity. r
t f1 ¼ 30
5 30 60 100
— — 452 7 48 502 7 19
f2 ¼ 40
f3 ¼50
f4 ¼ 100
f5 ¼ 500
f6 ¼ 1000
f7 ¼2000
f8 ¼ 3000
— — 1 7027 9
297 5 1177 16 3177 24 7747 28
277 4 1227 14 2577 14 1
257 2 1087 19 206 7 32 2717 27
477 5 1047 7 2047 17 250 7 44
2127 21 3047 9 — —
2717 21 364 7 11 — —
NOTE: fi -input signal frequency (kHz); t -equilibrium time (s); r -suspension conductivity (μs/cm). The minus sign of time means yeast cells feel negative DEP forces and move to the position with the minimum electric field intensity. Each experiment was repeated for three times, and the average time and standard deviation are shown in the table.
H. Zhu et al. / Biosensors and Bioelectronics 63 (2015) 371–378
4.4. Separation of yeast cells from a mixture solution with PS microspheres DEP can change the direction of bio-particle's movement and thus is often used for cell, protein and DNA separation. Normally, two schemes are often found in DEP chips: separation in space and separation in time. Herein separation in time mode is adopted to separate yeast cells from a mixture solution with PS microspheres. Fig. 1b illustrates the separation principle. When suspension solution containing PS microspheres and yeast cells flows into microfluidic AC-DEP chip, yeast cells feel positive DEP forces and then are trapped onto the electrode tips. Meanwhile, PS microspheres feel negative DEP forces and are repelled away from electrodes. After pumping the original suspension solution out of the chip, new solution without any particles is injected into the microchannel, accompanying by the turnoff of the applied AC signal. As a result, yeast cells are released and rinsed out. In order to figure out PS microspheres and yeast cells in optical microscope, red PS microspheres with the size close to yeast cells ( 7 μm) were used for evaluation of separation performance in screen-printed microfluidic AC-DEP chip. We first investigated the influence of chip configurations including electrode pairs, electrode gap, channel depth etc. on capture rate of yeast cells. As can be found in our test, the more the electrode pairs are, the longer the effective separation region is and the more the yeast cells can
377
be captured. However, due to the gravitational settling of PS microspheres onto the electrode surface, the separation efficiency can inevitably decrease with the increase in electrode pair and channel length. It is well known that electrode gap is inversely proportional to the electric field intensity. Smaller electrode gap implies larger DEP forces imposed on yeast cells, but less total number of captured yeast cells. In addition to electrode pair and electrode gap, channel depth is another important factor for capture rate. It was reported in the literature that the DEP forces were negligible if the distance between particle and metal electrode exceeded 30 μm (Martinez-Duarte, 2012). That means part of yeast cells cannot be trapped onto electrode tips if the channel depth is more than 30 μm. However, we found that the capture rate of yeast cells was still high when the channel depth was higher than 50 μm, which can be attributed to the semi-3D structure of screen-printed carbon electrodes. After optimization of chip configurations (data not shown here), the electrode pairs were set as 4 and electrode gap was set as 100 μm, meanwhile channel depth was kept as 50 μm. Fig. 6 shows capture rates of yeast cells in microfluidic AC-DEP chip with optimized chip configurations as functions of other influential factors such as medium conductivity, signal frequency, applied voltage and flow rate. As shown in Fig. 6a, the capture rates of yeast cells decrease with medium conductivity, which can be assigned to the reduced DEP forces in medium with high
Fig. 6. Capture rates of yeast cells in microfluidic AC-DEP chip under various influential factors: (a) medium conductivity; (b) input signal frequency; (c) input signal voltage (Vpp); (d) flow rate of suspension.
378
H. Zhu et al. / Biosensors and Bioelectronics 63 (2015) 371–378
conductivity. Fig. 6b and c show the dependence of capture rate on applied signals. In medium with conductivity of 5 μs/cm, the growth of input frequency reduces positive DEP forces imposed on yeast cells, leading to the decrease in capture rate. It is selfevident that capture rate rises with voltage since DEP forces are proportional to applied voltage squared. For example, the capture rate is only 21% when applied voltage is 3 V, and reaches ca. 95% when voltage is 10 V. Fig. 6d demonstrates that although high flow rate can reduce the separation time, it also reduces capture rate especially when fluid force is larger than DEP forces. In our work, no yeast cells can be trapped after flow rate reaches 0.3 mL/h. Movie of separation is given in supporting information. It is obvious that both red PS microspheres and white yeast cells flow into the microchannel whilst only red PS microspheres can be observed in the outlet. After shutting down the applied signals at 4min39sec, yeast cells trapped on the electrode tips can be released thoroughly. However, due to the gravitational settling and resistant from coarse carbon surface, the separation efficiency is ca. 94–96%. Further improvement can be achieved if smoother carbon electrode can be printed. Besides, we also compared the performance of screen-printed microfluidic AC-DEP chip with the one prepared via traditional soft lithography and lift-off techniques with gold as the electrode material and PDMS as channel materials. Note that the electrode structure and channel depth were the same, but the thickness of Au electrode was only 50 nm. Our results show no obvious difference in separation performance between these two chips under the experimental condition (applied voltage: 10 V, frequency: 100 kHz, flow rate: 0.1 mL/h), indicating the feasibility of our chip to replace the traditional one.
5. Conclusion In this paper, we demonstrate that a microfluidic AC-DEP chip for cell separation can be fabricated by screen printing technology. In comparison to the conventional microfabrication process, our method is especially suitable for high-throughput mass production due to the low cost, simple operating procedure and facile fabrication conditions. To demonstrate its separation performance, yeast cells, as model cells, were trapped and separated from a mixed suspension with PS microspheres. Our results show that high capture rate and separation efficiency can be achieved under optimized conditions. We believe that this approach offers great promise toward the development of miniaturized, portable flowthrough DEP chips for diagnostic and clinical applications.
Acknowledgment This research work was financially sponsored by the National Natural Science Foundation of China (Grant nos. 21105029, 51373056, and 51102097) and the Program for New Century Excellent Talents in University (NCET-11-0150) and Fundamental Research Funds for the Central Universities (2012ZZ0089).
Appendix A. Supplementary information Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.bios.2014.07.072.
References Becker, F.F., Wang, X.-B., Huang, Y., Pethig, R., Vykoukal, J., 1995. Proc. Natl. Acad. Sci. USA 92, 860–864. Barbulovic-Nad, I., Xuan, X., Lee, J.S.H., Li, D., 2006. Lab Chip 6, 274–279. Beech, J.P., Jonsson, P., Tegenfeldt, J.O., 2009. Lab Chip 9, 2698–2706. Baylon-Cardiel, J.-L., Jesus-Perez, N.-M., Chavez-Santoscoya, A.-V., Lapizco-Encinas, B.-H., 2010. Lab Chip 10, 3235–3242. Cummings, E.B., Singh, A.K., 2007. Anal. Chem. 75, 4724–4731. Chu, H., Doh, I., Cho, Y.-H., 2009. Lab Chip 9, 686–691. Cheng, I.F., Froude, V.E., Zhu, Y., Chang, H.-C., Chang, H.-C., 2009. Lab Chip 9, 3193– 3201. Cetin, B., Li, D.Q., 2011. Electrophoresis 32, 2410–2427. Demierre, N., Braschler, T., Linderholm, P., Seger, U., van Lintel, H., Renaud, P., 2007. Lab Chip 7, 355–365. Dong, H., Li, C.M., Zhang, Y.F., Cao, X.D., Gan, Y., 2007. Lab Chip 7, 1752–1758. Gascoyne, P., Mahidol, C., Ruchirawat, M., Satayavivad, J., Watcharasit, P., Becker, F.F., 2002. Lab Chip 2, 70–75. Grier, D.G., 2003. Nature 424, 810–816. Gallo-Villanueva, R.-C., Rodriguez-Lopez, C.-E., Diaz-de-la Garza, R.-I., Reyes-Betanzo, C., Lapizco- Encinas, B.-H., 2009. Electrophoresis 30, 4195–4205. Gascoyne, P.R.C., Wang, X.-B., Huang, Y., Becker, F.F., 1997. IEEE Trans. Ind. Appl. 33, 670–678. Hawkins, B.G., Smith, A.E., Syed, Y.A., Kirby, B.J., 2007. Anal. Chem. 79, 7291–7300. Hsiung, L.-C., Yang, C.H., Chiu, C.L., Chen, C.-L., Wang, Y., Lee, H., Cheng, J.-Y., Ho, M.C., Wo, A.M., 2008. Biosens. Bioelectron. 24, 869–875. Jaramillo, M.-d.-C., Torrents, E., Martinez-Duarte, R., Madou, M., Juarez, A., 2010. Electrophoresis 31, 2921–2928. Jen, C.P., Weng, C.H., Huang, C.T., 2011. Electrophoresis 32, 2428–2435. Kim, U., Qian, J., Kenrick, S.A., Daugherty, P.S., Soh, H.T., 2008. Anal. Chem. 80, 8656– 8661. Lapizco-Encinas, B.-H., Simmons, B.A., Cummings, E.B., Fintschenko, Y., 2004. Anal. Chem. 76, 1571–1579. McCloskey, K.E., Chalmers, J.J., Zborowski, M., 2003. Anal. Chem. 75, 6868–6874. Mernier, G., Piacentini, N., Braschler, T., Demierre, N., Renaud, P., 2010. Lab Chip 10, 2077–2082. Martinez-Duarte, R., Gorkin III, R.A., Abi-SamatraIII, K., Madou, M.J., 2010. Lab Chip 10, 1030–1043. Martinez-Duarte, R., 2012. Electrophoresis 33, 3110–3132. Nilsson, A., Petersson, F., Jonsson, H., Laurell, T., 2004. Lab Chip 4, 131–135. Prinz, C., Tegenfeldt, J., Austin, R., Cox, E., Sturm, J., 2002. Lab Chip 2, 207–212. Park, K., Suk, H.-J., Akin, D., Bashir, R., 2009. Lab Chip 9, 2224–2229. Regtmeier, J., Eichhorn, R., Bogunovic, L., Ros, A., Anselmetti, D., 2010. Anal. Chem. 82, 7141–7149. Rousselet, J., Markx, G.H., Pethig, R., 1998. Colloids Surf. A 140, 209–216. Suzuki, M., Yakusawa, T., Shiku, H., Matsue, T., 2008. Biosens. Bioelectron. 24, 1043– 1047. Shin, D.H., Kim, J.-E., Shim, H.C., Song, J.-W., Yoon, J.-H., Kim, J., Jeong, S., Kang, J., Baik, S., Han, C.-S., 2008. Nano Lett. 8, 4380–4385. Shafiee, H., Sano, M.B., Henslee, E.A., Caldwell, J.L., Davalos, R.V., 2010. Lab Chip 10, 438–445. Tang, M.K., Ma, Y.L., Yang, P., 2013. J. Phys 418, 012010. Urdaneta, M., Smela, E., 2008. Lab Chip 8, 550–556. Vahey, M.D., Voldman, J., 2008. Anal. Chem. 80, 3135–3143. Voldman, J., 2006. Annu. Rev. Biomed. Eng. 8, 425–454. Yasukawa, T., Suzuki, M., Sekiya, T., Shiku, H., Matsue, T., 2007. Biosens. Bioelectron. 22, 2730–2736. Zhu, J., Xuan, X., 2009. Electrophoresis 30, 2668–2675.
Contents lists available at ScienceDirect
Biosensors and Bioelectronics journal homepage: www.elsevier.com/locate/bios
Screen-printed microfluidic dielectrophoresis chip for cell separation Hongwu Zhu a,b, Xiaoguang Lin a, Yong Su a, Hua Dong a,n, Jianhua Wu b a b
Department of Biomedical Engineering, School of Materials Science and Engineering, South China University of Technology, Guangzhou 510641, China School of Biological Science and Engineering, South China University of Technology, Guangzhou 510006, China
art ic l e i nf o
a b s t r a c t
Article history: Received 1 April 2014 Received in revised form 11 July 2014 Accepted 24 July 2014 Available online 4 August 2014
Dielectrophoresis (DEP), the induced motion of polarizable particles in a non-uniform electric field, has been proven as a perfect candidate to transport, accumulate, separate and characterize micro-/nano-scale bioparticles in microfluidic systems. However, conventional fabrication technologies are complex, timeconsuming and relatively expensive, leading to low throughput of the DEP-based systems. In this paper, we report a novel microfluidic alternating current DEP (AC-DEP) chip fabricated via inexpensive screen printing method. The innovation of our work consists in the extreme simplicity of the fabrication procedure, i.e., the main components, including electrodes and channels, were constructed by layer-bylayer screen printing process, which is especially suitable for high-throughput mass production. Carbon paste, instead of metals, was used to print interdigitated electrodes with semi-3D structure which not only reduces dramatically the chip cost but also increases particle trapping efficiency. To test the chip performance, yeast cells, as model cells, were trapped and separated from a mixed suspension with PS microspheres. Our results show that high capture rate and separation efficiency can be achieved under optimized conditions. & 2014 Elsevier B.V. All rights reserved.
Keywords: Screen-printing Microfluidics chip Dielectrophoresis Cell separation
1. Introduction Manipulation of bio-particles in microfluidic chips including trapping, sorting, patterning, separation and purification of cells, viruses, proteins etc., is crucial for a variety of diagnostic and clinical applications (Becker et al., 1995; Gascoyne et al., 2002; Prinz et al, 2002; Grier, 2003; McCloskey et al., 2003; LapizcoEncinas et al., 2004; Nilsson et al., 2004; Barbulovic-Nad et al., 2006; Yasukawa et al., 2007; Demierre et al., 2007; Cummings and Singh, 2007; Hawkins et al., 2007; Vahey and Voldman, 2008; Kim et al., 2008; Urdaneta and Smela, 2008; Hsiung et al., 2008; Suzuki et al., 2008; Shin et al., 2008; Zhu and Xuan, 2009; Chu et al., 2009; Beech et al., 2009; Cheng et al., 2009; Park et al., 2009; Mernier et al., 2010; Martinez-Duarte et al., 2010; Shafiee et al., 2010; Regtmeier et al., 2010; Baylon-Cardiel et al., 2010). In the past few decades, tremendous techniques have been developed on the basis of various principles such as optical tweezers (Grier, 2003), magnetophoresis (McCloskey et al., 2003), acoustic waves (Nilsson et al., 2004) and electrical means. Among these methods, dielectrophoresis (DEP) is one of the most versatile methods for particle manipulation due to its label-free nature, favorable scaling effects, simple instrumentation and capability to induce both negative and positive forces (Voldman, 2006). Specifically, DEP n
Corresponding author. E-mail address: [email protected] (H. Dong).
http://dx.doi.org/10.1016/j.bios.2014.07.072 0956-5663/& 2014 Elsevier B.V. All rights reserved.
refers to the movement of a particle in a non-uniform electric field as a result of the interaction of the particle's dipole and spatial gradient of the electric field (Cetin and Li, 2011), and can be generated by using either direct current (DC) or alternating current (AC) field. Compared with DC-DEP, AC-DEP is advantageous because of the low operating voltage that prevents a serious Joule heating effect. Moreover, low voltage simplifies the equipment required to produce the electric fields, making AC-DEP compatible with integrated circuits and feasible for battery powered hand-held devices. Although much progress has been achieved, there are still some challenges that need to be addressed before microfluidic AC-DEP devices meet the end users. Usually, AC-DEP devices contain metal electrodes embedded inside the microchannel network. These internal electrodes are planar (2D) ones (i.e. height of the electrodes are in the order of hundred nanometers), and are fabricated via complex, timeconsuming and relatively expensive microfabrication techniques which are not suitable for mass production (Martinez-Duarte, 2012). Therefore, there is still a desire for a low-cost yet highthroughput manufacturing approach to fabricate microfluidic ACDEP chips. In addition, while working with bio-particles, fouling of the metal electrodes (i.e. sample electrolysis) may distort the operation of AC-DEP devices (Gallo-Villanueva et al., 2009). One of the possible solutions is to employ carbon electrode because carbon has a much wider electrochemical stability window than metals commonly used in thin film electrode fabrication such as gold and
372
H. Zhu et al. / Biosensors and Bioelectronics 63 (2015) 371–378
platinum, and affords higher applied voltages in a given solution without electrolyzing it (Gascoyne et al., 1997; Rousselet et al., 1998). Besides, the use of carbon electrodes yields other advantages including excellent biocompatibility, mechanical properties and chemical inertness in almost all solvents/electrolytes. A good case in point is that Martinez-Duarte et al. (2010) and Jaramillo et al. (2010) fabricated 3D carbon electrodes by C-MEMS technique and utilized AC-DEP to trap Escherichia coli bacteria from the mixture with Bacillus cereus bacteria. In our opinion, however, C-MEMS is still complicated and time-consuming in terms of operation conditions. Another alternative way to fabricate carbon electrodes needs to be exploited urgently. In this paper, we report a unique microfluidic AC-DEP chip fabricated via inexpensive screen printing technology. The innovation of our study consists in the extreme simplicity of the fabrication procedure, i.e., the main components, including electrodes and channels, were constructed by layer-by-layer screen printing process, which is especially suitable for highthroughput mass production. Instead of metal electrode, carbon paste was used to print interdigitated electrode with semi-3D structure which not only reduces dramatically the chip cost but also increases particle trapping efficiency. To test the chip performance, yeast cells were trapped and separated from a mixed suspension with polystyrene (PS) microspheres, and the factors influencing DEP behaviors, capture rate and separation efficiency are discussed in detail. We wish that our work could exhibit the feasibility of screen printing technology in fabricating microfluidic AC-DEP chips and benefit their future commercialization.
The DEP force (Fd) acting on a spherical particle of radius r suspended in a fluid with permittivity εm is given as Eq. (1) (Jen et al., 2011): 2 Fd = 2πr 3εm Re ⎡⎣fcm ⎤⎦ ∇Erms
(1)
where Re[fcm] is the real part of the Clausius–Mossotti (CM) factor and Erms is the root-mean-square of the applied electric field. The CM factor (fcm) is a parameter of the effective polarizability of a particle. It varies with the complex dielectric properties of the particle and the surrounding medium, which are functions of the frequency of the applied field. The CM factor for a spherical particle is represented as Eq. (2):
⎡ ε⁎ − ε⁎ ⎤ p m ⎥ fcm = ⎢ ⁎ ⎢⎣ εp + 2εm⁎ ⎥⎦
(2)
⁎ where εp⁎ and εm are the complex permittivities of the particle and the medium, respectively. The complex permittivity is related to the conductivity s and frequency f, as shown in Eq. (3):
jσ 2π f
(3)
where j equals −1 . As a consequence, the CM factor can be further expressed as Eq. (4):
Re[fcm ] =
3. Materials and methods 3.1. Chip fabrication Thick-film screen printing technology was adopted in our study to fabricate microfluidic AC-DEP devices, according to the following protocol. The commercial carbon paste (ED423SS, ACHESON, USA) was first printed on clean glass surface to form interdigitated carbon electrodes and contact wires via a semi-automatic screen printer (F-C4050R, Fufa Company, Shenzhen, China) and a 400 mesh stencil. After curing at 130 °C for 10 min to harden carbon paste, UV curable dielectric paste (YB-1300, Fufa Company, Shenzhen, China) was repeatedly printed and exposed to UV light for 5 min to form microfluidic channels. The depth of the channels was controlled at ca. 50 μm. Finally, a PDMS (Sylgard 184 silicone elastomer kit, Dow Corning) polymer film with inlet and outlet ports was exposed to O2 plasma, pressed on the top of the microfluidic channels and then heated in an oven at 80 °C for 1 h. Two mask screens were used in the above-mentioned process and the patterns were aligned using two microscopes connected to the computer.
3.2. Measurement of DEP spectra of static PS microspheres and yeast cells
2. Theory
ε⁎ = ε −
region with a high-electric-field gradient, or negative (Re[fcm] o0), repelling particles away from the region with a highelectric-field gradient.
(σp − σm )(σp + 2σm ) + 4π 2f 2 (εp − εm )(εp + 2εm ) (σp + 2σm )2 + 4π 2f 2 (εp + 2εm )2
(4)
Close examination of Eq. (4) reveals that the sign of the CM factor is determined by the electrical conductivities of the particle and the medium at low frequencies and the permittivities at high frequencies. According to the polarity of Re[fcm], the DEP force can be either positive (Re[fcm] 40), pulling particles towards the
A testing chamber (size: 1 1 cm2) printed on glass slide with interdigitated carbon electrode was used to measure the sign and magnitude of DEP forces (i.e. DEP spectra) generated on PS microspheres (7 μm and 20 μm, Aladdin, China) and yeast cells that were suspended in solution separately. The solution conductivity was adjusted using KCl with a conductivity meter (DDS-307, Lei-ci Company, China). An arbitrary waveform generator (ED1411, Zhongce Electronics, China) was employed to create the sinusoidal signal (10 V peak-to-peak and frequency range between 30 kHz and 3 MHz) delivered to the interdigitated carbon electrodes. The whole device was mounted on a microscope stage (AO-KV200, Aosvi Company, China) so that the electrode, yeast cells and PS microspheres can be observed in a bright field using a 10 objective lens and a color camera attached to the microscope. To avoid the sudden electrolysis of electrolyte caused by capacitive effects, the magnitude and frequency of AC signals were slowed down gradually during switch-off process.
3.3. Investigation of separation performance in screen-printed microfluidic AC-DEP chip To evaluate the separation performance of screen-printed microfluidic AC-DEP chip, yeast cells and red PS particles ( 7 μm, Aladdin, China) were mixed thoroughly and injected into the chip using a syringe pump (74900-05, Cole-Parmer, USA) at different flow rates (0.1–0.3 mL/h). Various influential factors such as flow rate, frequency, solution conductivity, voltage were tested in detail. After experiment, the sample can be transferred from the DEP device to different collecting tubes.
H. Zhu et al. / Biosensors and Bioelectronics 63 (2015) 371–378
4. Results and discussion 4.1. Fabrication and characterization of microfluidic AC-DEP chip via screen-printing technology Although first demonstrated in the 1950s, it was not until the development of Bio-MEMS in the 1990s that DEP became a popular research field (Martinez-Duarte, 2012). During the whole 1990s, abundant DEP publications appeared using microfabricated metal electrodes. Starting in the 2000s, alternative techniques have been explored to overcome common problems (e.g. electrode fouling) in metal-electrode DEP. Insulator-based DEP and lightinduced DEP are the most significant examples in this period. More recently, new 3D techniques like carbon-electrode DEP, contactless DEP, and the use of either C-MEMS (Jaramillo et al., 2010) or doped PDMS have further simplified the fabrication process. However, the state-of-the-art of fabrication methods still cannot meet all the requirements for practical DEP chips such as low cost, simple process and high throughput.
373
Instead of complicated Bio-MEMS, herein we utilize screen printing to fabricate microfluidic AC-DEP chips. As a thick film technology, screen printing uses a woven mesh to support an inkblocking stencil to receive a desired pattern. Owing to its inexpensive and simple fabrication process, this technique has been widely applied in mass production of printed circuit board, clothing, medical devices, etc. (Dong et al., 2007). In our study, we constructed microfluidic AC-DEP chip by printing conductive carbon paste as the electrode, whilst printing dielectric paste as the channel wall and insulator to separate the contact wire from the solution. Fig. 1a shows the screen printing procedure of the microfluidic AC-DEP chip. To facilitate observation, transparent glass slide was used as substrate. Three kinds of interdigitated electrodes with different configurations were printed, i.e. offset castellated electrodes, non-offset castellated electrodes and interdigitated line electrodes. After drying the electrodes, UV curable dielectric paste was printed using another stencil. Since the printed dielectric film had a typical thickness of ca. 20–30 μm, channel walls with the height of 50–60 μm can be formed by printing repeatedly for 2–3 times. Thereafter, a thin layer of PDMS
Fig. 1. (a) Screen-printing fabrication of microfluidic AC-DEP chip for cell separation: (left) schematic illustration of layer-by-layer screen printing procedure. (1) Clean glass slide as substrate; (2) patterned carbon ink as interdigitated electrode and conductive wires; (3) UV curable dielectric composition for microfluidic channels; (4) PDMS layer with inlet and outlet ports; (middle) configuration of the whole device; (right) three kinds of interdigitated electrodes printed in our study. (5) Offset castellated electrodes; (6) non-offset castellated electrodes; (7) interdigitated line electrodes. (b) Schematic illustration on separation of yeast cells from a mixture with PS particles in screenprinted microfluidic AC-DEP chip. When suspension solution flows into microchannel, yeast cells feel positive DEP forces and thus are trapped onto the electrodes. Meanwhile, PS microspheres feel weak negative DEP forces and are repelled from electrodes. As a result, separation happens between yeast cells and PS microspheres. Fg: force of gravity, Fb: buoyant force, Fd: DEP force, Ff: fluid force, Fr: resistance.
374
H. Zhu et al. / Biosensors and Bioelectronics 63 (2015) 371–378
Fig. 2. Images of screen-printed microfluidic AC-DEP chips with (a) offset interdigitated carbon electrode and (b) non-offset interdigitated carbon electrode.
(with inlet and outlet ports) was treated by O2 plasma, pressed on the top of the printed dielectric layer and then heated at 80 °C for 1 h. The sealing was found to be robust for sample injection and transportation via pumping. Fig. 2 shows the photos of the screen printed microfluidic devices. The thickness of carbon electrode
was measured as ca. 10–15 μm, which shows semi-3D structure with respect to the size of cells. The resolution of stencil for screen printing in our study is ca. 50 μm, so smaller channel/electrode could have been fabricated using this technology. However, in order to achieve high
Fig. 3. 3D simulation of electric field intensity distribution using COMSOL multiphysics software. Top view of (a) interdigitated line electrodes, (b) non-offset castellated electrodes, (c) offset castellated electrodes, and side view of (d) non-offset castellated electrodes. The thickness of electrode and the distance between two microelectrodes were set as 10 and 100 μm, respectively. The applied voltage (VPP) was 10 V. Herein yellow, red and brown colors indicate the high, medium and low field strength, whilst white color indicates the locations without electric field (the lowest field intensity). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
H. Zhu et al. / Biosensors and Bioelectronics 63 (2015) 371–378
throughput separation, larger electrode gap ( 100–150 μm) and wider microchannels (2 mm) were actually employed to make sure that more particles/cells can be trapped via DEP forces, see the discussion later. It should be pointed out that insulator-based DEP chip can be also fabricated by this technology and good performance can be achieved. The work is still under investigation, and will be reported in the future. 4.2. Simulation of electric field intensity on screen-printed carbon microelectrode with different configurations Fig. 3a–d shows the finite element simulation of electric field distribution over carbon microelectrode with different configurations stated above. For interdigitated line electrodes (Fig. 3a), the electrical field intensity (top view, Z¼ 10 μm) increases to the maximum value at the edges of microelectrodes and drops to the minimum value in the center between two neighboring electrodes. For non-offset and offset castellated electrodes (Figs. 3b and c), the
375
strongest and weakest electric fields are located at the electrode tips and intermediate region between electrode tips, respectively. Moreover, the side view image (Fig. 3d) reveals that electrical field intensity decreases along the z-axis above the electrodes. Given the same simulation conditions, i.e., the applied voltage was 10 V and gap between neighboring electrodes was 100 μm, the 2 maximum ∇Erms were 2 1014 m kg2/s6 A2 for interdigitated line electrodes and 5 1015 m kg2/s6 A2 for non-offset and offset castellated electrodes, indicating that castellated electrodes can generate larger DEP forces compared with interdigitated line electrodes. Since screen-printed carbon electrodes possess porous and coarse surface, the electric field is more non-uniform than smooth metal electrodes and thus may further improve DEP forces imposed on particles (Tang et al., 2013). When aware of the locations of the strongest and weakest electric field, we can predict whether particles feel positive or negative DEP forces based on their movements towards or off the electrode edges.
Fig. 4. (a, b) Images of PS microspheres (size: 20 μm; color: white) suspended in aqueous medium (conductivity: 5 μs/cm) before (a) and after (b) applying 20 V/100 kHz AC signals on offset and non-offset electrodes. (c, d) Images of yeast cells (size: 5 μm, color: white) suspended in aqueous medium (conductivity: 5 μs/cm) after applying 10 V/100 kHz AC signals on (c) non-offset castellated electrodes and (d) offset castellated electrodes.
376
H. Zhu et al. / Biosensors and Bioelectronics 63 (2015) 371–378
4.3. DEP behaviors of static PS microspheres and yeast cells Although particle dielectric properties can be determined via typical electrorotation (ROT) methods and thus DEP spectra can be plotted using the dielectric parameters derived from ROT measurements (Becker et al., 1995), we investigated DEP behaviors of particles in static status using the printed castellated electrodes so that the results can be directly used for particle separation in microfluidic chip with the same electrode configuration. In our study, PS microspheres and yeast cells were chosen as particle models due to their similar spherical shape and well-known DEP behaviors, which makes it easy to compare our results with the literature. Fig. 4a and b show images of PS microspheres (size: 20 μm; color: white) before and after applying 20 V/100 kHz AC signals on screen-printed electrodes. As can be seen, PS microspheres are randomly distributed around the electrodes prior to applying AC signals. Once subjected to non-uniform electric field, the majority of PS microspheres assemble into triangle in the intermediate region between electrode tips, and the rest distribute in the center between two neighboring electrodes. For non-offset castellated electrodes, straight chainline are also observed between opposite triangles. Judging from the simulation results, one can deduce that PS microspheres move to the weak electric field and thus bear negative DEP forces. Further measurements reveal that PS microspheres can only feel negative DEP forces, regardless of the applied frequencies and medium conductivities. Fig. 4c and d give images of yeast cells (size: 5–6 μm) after applying 10 V/100 kHz AC signals on non-offset and offset castellated electrodes, respectively. It is clear that yeast cells are trapped to electrode tips with high-electric-field gradient, indicative of positive DEP behavior. Different from PS microspheres, both positive and negative DEP forces are observed on yeast cells under our experimental conditions. Although the absolute value of DEP force is not available, the relative DEP force can be calculated. Considering yeast cells stay static in the beginning, the average DEP forces can be calculated using Eq. (5), derived from Newton's second law:
2mS Fd = 2 t
(5)
where m is the mass of yeast cell, S is average displacement distance of yeast cells in microfluidic chip and t is the average time needed for yeast cells to reach the positions with maximum or minimum electric field intensity. It can be found from Eq. (5) that DEP forces are proportional to the reciprocal of time squared. As a result, DEP spectra can be obtained by plotting normalized DEP force (or namely, the reciprocal of time squared) as a function of input frequencies and medium conductivities. Table 1 shows the average time for yeast cells to reach positions with the maximum or minimum electric field intensity. If yeast cells don't move after applying AC signals, the time was set as infinite. In mediums with conductivities of 5 and 30 μs/cm (Fig. 5a), only positive DEP forces are detected on yeast cells no
Fig. 5. DEP spectra of yeast cells in aqueous mediums with conductivities of (a) 5 and 30 μs/cm and (b) 60 and 100 μs/cm. Input voltage (Vpp): 10 V, frequency range: 30–3000 kHz.
matter how to change applied frequencies in the range of 30– 3000 kHz. The increase of time with frequency means that DEP forces decrease with frequency. Moreover, when input frequencies is less than 1000 kHz, the larger DEP forces in medium with lower conductivity (5 μs/cm) is in good agreement with the deduction from Eq. (4) that the CM factor (i.e. DEP force) is determined by the electrical conductivities of the particle and the medium at low frequencies, which inversely proves the validity of our data. In contrast, when the medium conductivity rises to 60 or 100 μs/cm (Fig. 5b), yeast cells bear negative DEP forces at low frequencies and positive DEP forces at high frequencies, and the crossover frequency increases with the medium conductivity.
Table 1 Average time for yeast cells to reach positions with the maximum or minimum electric field intensity. r
t f1 ¼ 30
5 30 60 100
— — 452 7 48 502 7 19
f2 ¼ 40
f3 ¼50
f4 ¼ 100
f5 ¼ 500
f6 ¼ 1000
f7 ¼2000
f8 ¼ 3000
— — 1 7027 9
297 5 1177 16 3177 24 7747 28
277 4 1227 14 2577 14 1
257 2 1087 19 206 7 32 2717 27
477 5 1047 7 2047 17 250 7 44
2127 21 3047 9 — —
2717 21 364 7 11 — —
NOTE: fi -input signal frequency (kHz); t -equilibrium time (s); r -suspension conductivity (μs/cm). The minus sign of time means yeast cells feel negative DEP forces and move to the position with the minimum electric field intensity. Each experiment was repeated for three times, and the average time and standard deviation are shown in the table.
H. Zhu et al. / Biosensors and Bioelectronics 63 (2015) 371–378
4.4. Separation of yeast cells from a mixture solution with PS microspheres DEP can change the direction of bio-particle's movement and thus is often used for cell, protein and DNA separation. Normally, two schemes are often found in DEP chips: separation in space and separation in time. Herein separation in time mode is adopted to separate yeast cells from a mixture solution with PS microspheres. Fig. 1b illustrates the separation principle. When suspension solution containing PS microspheres and yeast cells flows into microfluidic AC-DEP chip, yeast cells feel positive DEP forces and then are trapped onto the electrode tips. Meanwhile, PS microspheres feel negative DEP forces and are repelled away from electrodes. After pumping the original suspension solution out of the chip, new solution without any particles is injected into the microchannel, accompanying by the turnoff of the applied AC signal. As a result, yeast cells are released and rinsed out. In order to figure out PS microspheres and yeast cells in optical microscope, red PS microspheres with the size close to yeast cells ( 7 μm) were used for evaluation of separation performance in screen-printed microfluidic AC-DEP chip. We first investigated the influence of chip configurations including electrode pairs, electrode gap, channel depth etc. on capture rate of yeast cells. As can be found in our test, the more the electrode pairs are, the longer the effective separation region is and the more the yeast cells can
377
be captured. However, due to the gravitational settling of PS microspheres onto the electrode surface, the separation efficiency can inevitably decrease with the increase in electrode pair and channel length. It is well known that electrode gap is inversely proportional to the electric field intensity. Smaller electrode gap implies larger DEP forces imposed on yeast cells, but less total number of captured yeast cells. In addition to electrode pair and electrode gap, channel depth is another important factor for capture rate. It was reported in the literature that the DEP forces were negligible if the distance between particle and metal electrode exceeded 30 μm (Martinez-Duarte, 2012). That means part of yeast cells cannot be trapped onto electrode tips if the channel depth is more than 30 μm. However, we found that the capture rate of yeast cells was still high when the channel depth was higher than 50 μm, which can be attributed to the semi-3D structure of screen-printed carbon electrodes. After optimization of chip configurations (data not shown here), the electrode pairs were set as 4 and electrode gap was set as 100 μm, meanwhile channel depth was kept as 50 μm. Fig. 6 shows capture rates of yeast cells in microfluidic AC-DEP chip with optimized chip configurations as functions of other influential factors such as medium conductivity, signal frequency, applied voltage and flow rate. As shown in Fig. 6a, the capture rates of yeast cells decrease with medium conductivity, which can be assigned to the reduced DEP forces in medium with high
Fig. 6. Capture rates of yeast cells in microfluidic AC-DEP chip under various influential factors: (a) medium conductivity; (b) input signal frequency; (c) input signal voltage (Vpp); (d) flow rate of suspension.
378
H. Zhu et al. / Biosensors and Bioelectronics 63 (2015) 371–378
conductivity. Fig. 6b and c show the dependence of capture rate on applied signals. In medium with conductivity of 5 μs/cm, the growth of input frequency reduces positive DEP forces imposed on yeast cells, leading to the decrease in capture rate. It is selfevident that capture rate rises with voltage since DEP forces are proportional to applied voltage squared. For example, the capture rate is only 21% when applied voltage is 3 V, and reaches ca. 95% when voltage is 10 V. Fig. 6d demonstrates that although high flow rate can reduce the separation time, it also reduces capture rate especially when fluid force is larger than DEP forces. In our work, no yeast cells can be trapped after flow rate reaches 0.3 mL/h. Movie of separation is given in supporting information. It is obvious that both red PS microspheres and white yeast cells flow into the microchannel whilst only red PS microspheres can be observed in the outlet. After shutting down the applied signals at 4min39sec, yeast cells trapped on the electrode tips can be released thoroughly. However, due to the gravitational settling and resistant from coarse carbon surface, the separation efficiency is ca. 94–96%. Further improvement can be achieved if smoother carbon electrode can be printed. Besides, we also compared the performance of screen-printed microfluidic AC-DEP chip with the one prepared via traditional soft lithography and lift-off techniques with gold as the electrode material and PDMS as channel materials. Note that the electrode structure and channel depth were the same, but the thickness of Au electrode was only 50 nm. Our results show no obvious difference in separation performance between these two chips under the experimental condition (applied voltage: 10 V, frequency: 100 kHz, flow rate: 0.1 mL/h), indicating the feasibility of our chip to replace the traditional one.
5. Conclusion In this paper, we demonstrate that a microfluidic AC-DEP chip for cell separation can be fabricated by screen printing technology. In comparison to the conventional microfabrication process, our method is especially suitable for high-throughput mass production due to the low cost, simple operating procedure and facile fabrication conditions. To demonstrate its separation performance, yeast cells, as model cells, were trapped and separated from a mixed suspension with PS microspheres. Our results show that high capture rate and separation efficiency can be achieved under optimized conditions. We believe that this approach offers great promise toward the development of miniaturized, portable flowthrough DEP chips for diagnostic and clinical applications.
Acknowledgment This research work was financially sponsored by the National Natural Science Foundation of China (Grant nos. 21105029, 51373056, and 51102097) and the Program for New Century Excellent Talents in University (NCET-11-0150) and Fundamental Research Funds for the Central Universities (2012ZZ0089).
Appendix A. Supplementary information Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.bios.2014.07.072.
References Becker, F.F., Wang, X.-B., Huang, Y., Pethig, R., Vykoukal, J., 1995. Proc. Natl. Acad. Sci. USA 92, 860–864. Barbulovic-Nad, I., Xuan, X., Lee, J.S.H., Li, D., 2006. Lab Chip 6, 274–279. Beech, J.P., Jonsson, P., Tegenfeldt, J.O., 2009. Lab Chip 9, 2698–2706. Baylon-Cardiel, J.-L., Jesus-Perez, N.-M., Chavez-Santoscoya, A.-V., Lapizco-Encinas, B.-H., 2010. Lab Chip 10, 3235–3242. Cummings, E.B., Singh, A.K., 2007. Anal. Chem. 75, 4724–4731. Chu, H., Doh, I., Cho, Y.-H., 2009. Lab Chip 9, 686–691. Cheng, I.F., Froude, V.E., Zhu, Y., Chang, H.-C., Chang, H.-C., 2009. Lab Chip 9, 3193– 3201. Cetin, B., Li, D.Q., 2011. Electrophoresis 32, 2410–2427. Demierre, N., Braschler, T., Linderholm, P., Seger, U., van Lintel, H., Renaud, P., 2007. Lab Chip 7, 355–365. Dong, H., Li, C.M., Zhang, Y.F., Cao, X.D., Gan, Y., 2007. Lab Chip 7, 1752–1758. Gascoyne, P., Mahidol, C., Ruchirawat, M., Satayavivad, J., Watcharasit, P., Becker, F.F., 2002. Lab Chip 2, 70–75. Grier, D.G., 2003. Nature 424, 810–816. Gallo-Villanueva, R.-C., Rodriguez-Lopez, C.-E., Diaz-de-la Garza, R.-I., Reyes-Betanzo, C., Lapizco- Encinas, B.-H., 2009. Electrophoresis 30, 4195–4205. Gascoyne, P.R.C., Wang, X.-B., Huang, Y., Becker, F.F., 1997. IEEE Trans. Ind. Appl. 33, 670–678. Hawkins, B.G., Smith, A.E., Syed, Y.A., Kirby, B.J., 2007. Anal. Chem. 79, 7291–7300. Hsiung, L.-C., Yang, C.H., Chiu, C.L., Chen, C.-L., Wang, Y., Lee, H., Cheng, J.-Y., Ho, M.C., Wo, A.M., 2008. Biosens. Bioelectron. 24, 869–875. Jaramillo, M.-d.-C., Torrents, E., Martinez-Duarte, R., Madou, M., Juarez, A., 2010. Electrophoresis 31, 2921–2928. Jen, C.P., Weng, C.H., Huang, C.T., 2011. Electrophoresis 32, 2428–2435. Kim, U., Qian, J., Kenrick, S.A., Daugherty, P.S., Soh, H.T., 2008. Anal. Chem. 80, 8656– 8661. Lapizco-Encinas, B.-H., Simmons, B.A., Cummings, E.B., Fintschenko, Y., 2004. Anal. Chem. 76, 1571–1579. McCloskey, K.E., Chalmers, J.J., Zborowski, M., 2003. Anal. Chem. 75, 6868–6874. Mernier, G., Piacentini, N., Braschler, T., Demierre, N., Renaud, P., 2010. Lab Chip 10, 2077–2082. Martinez-Duarte, R., Gorkin III, R.A., Abi-SamatraIII, K., Madou, M.J., 2010. Lab Chip 10, 1030–1043. Martinez-Duarte, R., 2012. Electrophoresis 33, 3110–3132. Nilsson, A., Petersson, F., Jonsson, H., Laurell, T., 2004. Lab Chip 4, 131–135. Prinz, C., Tegenfeldt, J., Austin, R., Cox, E., Sturm, J., 2002. Lab Chip 2, 207–212. Park, K., Suk, H.-J., Akin, D., Bashir, R., 2009. Lab Chip 9, 2224–2229. Regtmeier, J., Eichhorn, R., Bogunovic, L., Ros, A., Anselmetti, D., 2010. Anal. Chem. 82, 7141–7149. Rousselet, J., Markx, G.H., Pethig, R., 1998. Colloids Surf. A 140, 209–216. Suzuki, M., Yakusawa, T., Shiku, H., Matsue, T., 2008. Biosens. Bioelectron. 24, 1043– 1047. Shin, D.H., Kim, J.-E., Shim, H.C., Song, J.-W., Yoon, J.-H., Kim, J., Jeong, S., Kang, J., Baik, S., Han, C.-S., 2008. Nano Lett. 8, 4380–4385. Shafiee, H., Sano, M.B., Henslee, E.A., Caldwell, J.L., Davalos, R.V., 2010. Lab Chip 10, 438–445. Tang, M.K., Ma, Y.L., Yang, P., 2013. J. Phys 418, 012010. Urdaneta, M., Smela, E., 2008. Lab Chip 8, 550–556. Vahey, M.D., Voldman, J., 2008. Anal. Chem. 80, 3135–3143. Voldman, J., 2006. Annu. Rev. Biomed. Eng. 8, 425–454. Yasukawa, T., Suzuki, M., Sekiya, T., Shiku, H., Matsue, T., 2007. Biosens. Bioelectron. 22, 2730–2736. Zhu, J., Xuan, X., 2009. Electrophoresis 30, 2668–2675.
