Portable plastic syringe as a self-actuated pump for long-distance uniform delivery of liquid inside a microchannel and its application for flow-through polymerase chain reaction on chip

Abstract

A portable plastic syringe was used as a self-actuated pump for uniform delivery of liquid inside a microchannel over a distance of more than 2 meters at a controllable flow rate and without utilizing external electrical power and bulky pumping apparatus. The reliability of the plastic syringe as a potential self-actuator was investigated by performing a flow-through polymerase chain reaction (PCR) on a microdevice fabricated using poly(methylmethacrylate) (PMMA). Liquid flowed at a uniform rate inside the PMMA microchannel in a highly controllable manner even under high-temperature conditions and without the generation of bubbles, and the flow rate was readily adjusted as necessary by varying the operation parameters such as the length of the outlet silicone tube, channel dimension, and initial syringe pressure. A 230 bp plasmid vector obtained from E. coli and a D1S80 locus obtained from a human genomic DNA were successfully amplified on a PMMA microdevice equipped with the disposable plastic syringe as a self-actuated micropump.

---

Introduction

Due to its crucial role in transporting liquid, the micropump has been recognized as one of the most important ingredients of micro Total Analysis Systems (μTAS), and thus, has been the hot issue of wide research interests. In the 1980s, Lintel et al.¹ and Smits² developed the first genuine MEMS micropumps actuating fluid on the microscale utilizing piezoelectrics. During the last 20 years, over ten mechanisms have been introduced to structure diverse pumping systems appropriate for μTAS, such as the peristaltic micropump,³ thermo-pneumatic micropump,⁴ magnetic micropump,⁵ and electro-hydrodynamic micropump.⁶ Despite numerous achievements in the development of micropumps,⁷ the majority of such devices that have been developed so far require external power for operation.^3–12 Besides, complex fabrication steps or operation processes are always inevitable when using such aforementioned pumps. These disadvantages hinder miniaturization, which is one of the key issues in μTAS.

In order to circumvent above problems, some self-actuated micropumps have been developed, dispensing with any external power. Among these self-actuated micropumps, capillary micropump is one typical representative and simplest format, utilizing capillary force to spontaneously transport small amount of liquid.^13,14 The surface tension resulting from solid–liquid, liquid–gas, and solid–gas interactions, produces an equilibrium contact angle in the anterior and posterior ends of liquid plug inside a microchannel, and governs the self-actuation of sample liquid inside the microchannel. Flow velocity of capillary micropump is reduced non-linearly, as the strength of capillary forces per sample volume decreases dramatically, as more and more sample flows into the microchannel. Also, the inner surface of the microchannel should be hydrophilic. Besides the capillary micropump, vacuum pump was also introduced by Dimov et al. to realize a self-powered integrated device for blood assay, where a high air permeability of the microdevice was the critical element.¹⁵ After placing poly(dimethylsiloxane) (PDMS) chip into a vacuum chamber, the air molecules inside the chip permeated through porous PDMS, which resulted in relatively lower pressure at the anterior end of the sample, forming an actuation force toward the outlet. A vacuum chamber and a permeable microdevice are the two preconditions in realizing such pump.

Recently, a wicking flow has also become a hot issue due to its dramatic property in spontaneous transportation of liquid through a microchannel.^16–18 The mechanism of wicking micropump virtually relies on capillary flow, but the platforms utilized here are mainly paper or textiles, instead of silicon or glass as in capillary micropumps. In addition to aforementioned self-actuated micropumps, Qin et al. also introduced a hydrolytic-powered pump, which can catalytically decompose H₂O₂ into oxygen for creating a pressure gradient which induces sample injection.¹⁹ Even if self-actuated micropumps obtain more noticeable advantages than those which utilize external power, all of these encounter the bottleneck in providing homogeneous flow over a long distance such as several meters.^13–19 Besides, it is also difficult for previously developed self-actuated micropumps to stably transport liquid under harsh microenvironment^13–19 such as high temperature condition. These defects restricted the application of self-actuated micropumps in many areas, i.e., continuous-flow PCR, which always requires a microchannel as long as several meters^20–25 and a temperature as high as 95 °C.

To solve the above-mentioned problems, here a new self-actuated micropump is introduced, which can maintain homogeneous liquid flow over a long distance. The introduced self-actuated micropump not only can stably work at room temperature, but is also reliable under high temperature. Furthermore, not many fabrication steps are involved here. By connecting one small piece of silicone tube to the outlet of a gas-impermeable poly(methylmethacrylate) (PMMA) microdevice and clamping it to realize a blunted end, homogeneous transportation of fluid can be realized using a disposable syringe, connected to the inlet via another short segment of a silicone tube with negligible gas permeability. Using the microdevice, a pGEM-3Zf(+) plasmid vector obtained from DH5-α E. coli and D1S80 locus obtained from a human genomic DNA, were successfully amplified inside a PMMA microdevice.

Principle

Fig. 1 shows a schematic illustration revealing the actuation mechanism of two different self-actuated micropumps. Both of the mechanisms were based on air permeability from the fluidic conduit to the atmosphere. In our previous studies,^21,22 we have introduced a concept for a self-actuated micropump for delivering sample plug inside a microchannel. However, flow velocity gradually decreased with time. In contrast, the self-actuated micropump introduced here can deliver sample plug with homogenous flow velocity inside a very long microchannel reaching over 2 meters. For actuation, a disposable syringe is connected to the inlet of a gas-impermeable PMMA microchannel. Besides, a permeable silicone tube is connected to the outlet of the fluidic conduit and blunt-ended by using a commercial clamp. As shown in Fig. 1a, P_p and P_a represent the air pressures in the posterior and anterior ends of the sample plug, respectively, while P_g represents the pressure gradient imposed on the sample plug which can be expressed as the following equation.

P_g = P_p − P_a

Fig. 1 (a and b) Schematic illustrations demonstrating a new concept for self-actuated pump realizing homogeneous sample flow inside a PMMA microchannel, and comparison with (c) previous study. (d) A schematic graph demonstrating a constant pressure gradient formed by the pressure difference in the inlet and the outlet of the fluidic conduit in this study. (e) A schematic graph demonstrating a gradual reduction in the pressure gradient formed by the pressure difference in the inlet and outlet of the fluidic conduit in our previous study. (f and g) Schematic graphs comparing the sample residence time inside serpentine microchannels in (f) this study and (g) previous study.

Since the pressure of the compressed air captured inside the closed fluidic conduit is higher than the atmospheric pressure, air molecules tend to diffuse from inside the microchannel to the atmosphere. The inlet tube, the PMMA microdevice, and the disposable syringe used here, can all be considered as gas-impermeable, as compared to outlet silicone tube, which is gas-permeable. For this reason, the gas permeability in the posterior end of the sample can be considered negligible, and the air molecules can be rendered to diffuse from the fluidic conduit to the atmosphere only through the outlet silicone tube.

The ideal gas law represented as follows, can be used to calculate the relationship among the pressure (P), the volume (V), and the number of moles of the air molecules (n), where T is the Kelvin temperature and R is the gas constant.

PV = nRT

Before the sample plug is introduced into the microdevice, air pressure throughout the entire fluidic conduit is the same, represented by P_p = P_a. After the sample plug is introduced from the inlet tube by disposable syringe, the sample plug separates the fluidic conduit into two parts, that is, the gas-permeable part at the anterior end and the gas-impermeable part at the posterior end of the sample plug, resulting in varying gas permeability at both ends of the sample plug. The gas diffusion in the anterior end of sample causes a pressure drop in P_a. Since there is negligible gas diffusion in the posterior end of the sample, P_p can be considered constant. As a result, a homogeneous pressure gradient (P_g) is imposed on the sample plug, and this propels the sample toward the outlet with identical flow velocity (Fig. 1b, d and f), whereas in our previous studies,^21,22 air diffusion occurred all throughout the entire fluidic conduit (Fig. 1c, e and g).

If the inner and outer radii of the outlet silicone tube are r_i and r_o, respectively, the following equation can be derived,

where G_a is diffusion flux, D is effective diffusion coefficient, C_Aia is the concentration of the inner air molecules at the anterior end of the sample plug, C_Ao is the concentration of air molecules in the outside atmosphere, Z is the diffusion length, and A_av is the average diffusion area. The average diffusion area, A_av, can be calculated by the following equation, where L is the total length of the outlet silicone tube.

The diffusion length, Z, can be calculated by the following equation.

Z = r_o − r_i

So the relationship between the diffusion rate of the air molecules, G_a, and the parameters of outlet tube can be expressed as follows,

which can be converted to the following pressure-based equation.

From the above equations, we can estimate that the diffusion flux increases as the length or inner diameter of the outlet silicone tube increases. Also, as the inner pressure at the anterior end of sample plug increases, the diffusion flux also increases accordingly.

The self-actuation mechanism introduced in this study is totally different from our previous studies.^21,22 In our previous works, the air molecules freely diffused from both at the anterior and posterior ends of the sample plug. As the sample plug progresses forward inside the microchannel, the diffusion area at the posterior end tends to increase. Meanwhile, the diffusion area at the anterior end tends to decrease. As a result, the pressure gradient imposed on the sample plug caused by the non-homogeneous diffusion between the anterior and posterior ends of the sample plug tends to decrease, resulting in a gradual decrease in the overall sample flow. In contrast, since air diffuses only through the outlet silicone tube while the areas of gas diffusion at the anterior and posterior ends of the sample plug – A_av and 0 – are kept constant, the introduced self-actuation mechanism can maintain homogeneous liquid flow. In other words, the air diffusion in the outlet silicone tube is the only driving force for sample movement. In addition, impermeability of air at the posterior end of the sample plug and its extremely large volume kept inside the syringe makes the pressure at the posterior end nearly constant, not much decreased as compared to the initial pressure of the compressed air inside the syringe, throughout the sample transport toward the outlet.

In our previous work,²⁰ we have proven that the velocity of sample plug flowing through the microchannel with rectangular cross-section, is proportional to the pressure gradient (P_g) imposed on the sample plug. But in the new concept of self-actuation, since P_p is constant, the flow velocity is only determined by P_a. For a simpler model of the new self-actuation mechanism, the following equation can be derived,

where Q_a is a fluidic flux, ϑ is the velocity of sample liquid, H_c is the height of the microchannel, and W_c is the width of the microchannel. Based on the above derived equation, we can predict that the flow velocity increases as the permeability (D), length (L), and inner diameter (r_i) of the outlet silicone tube as well as the inner pressure of the microchannel increases. Besides, if the outer diameter (r_o) of the outlet silicone tube or the width (W_c) and height (H_c) of the microchannel increases, the flow rate will decrease, since the flow rate is inversely proportional to the channel height and width under the same fluidic flux. Simply changing the length of the microchannel does not pose any significant effects on the flow rate since no parameter exist representing the length of the microchannel in the above equation. To summarize, considering that a gradual reduction in the pressure gradient was caused by the air diffusion both through the anterior and posterior ends of the sample plug in different ratio, this issue is resolved and homogeneous liquid flow was maintained simply by allowing air diffusion only through the outlet silicone tube, in this study.

Methods

Fabrication of microdevice

Fig. 2 shows the overall schematic for microdevice fabrication. A serpentine microchannel 200 μm wide and 50 μm deep was fabricated on one PMMA substrate (40 × 40 × 2 mm) using a computer numerical control (CNC) milling machine (Fig. 2a and b). After the inlet and outlet ports were punctured on another flat PMMA substrate using a drilling machine, the two PMMA substrates were thermally bonded at approximately 105 °C (Fig. 2c).²⁵ Finally, thick-walled silicone tube (i.d. 0.2 mm, o.d. 2 mm) was inserted into the inlet port and thin-walled silicone tube (i.d. 1 mm, o.d. 2 mm) was inserted into the outlet port, and then were glued using PDMS prepolymer (Fig. 2d). By this means, a PMMA microdevice with serpentine channels equivalent to 25 thermal cycles was fabricated. The total length of the microchannel was 1.25 m, which corresponded to approximately 5 cm per cycle.

Fig. 2 (a) Fabrication of a serpentine microchannel on PMMA using a CNC milling machine. (b) Serpentine microchannel engraved on one PMMA substrate. (c) Punching inlet and outlet ports on a flat PMMA substrate, followed by thermal bonding. (d) Insertion of silicone tubes into the inlet and outlet ports.

Procedures for self-actuated liquid flow

Fig. 3 illustrates the procedures for preparing an air-filled syringe and connecting it to the PMMA microdevice for self-actuation of ink solution. First, a red ink, which was used in place of a sample liquid, was sucked into the disposable syringe. As shown in the inset in Fig. 3a, the ink was initially confined at the corner of the syringe. The viscosity of red ink solution at 25 °C is 9.89 × 10⁻⁴ Pa s, which was relatively close to that of the water at 25 °C (8.9 × 10⁻⁴ Pa s).²⁶ Prior to connecting syringe to the microdevice, the outlet silicone tube was clamped by a clip (Fig. 3a). Second, the syringe piston was pulled up to a certain graduation and then connected to the inlet silicone tube (Fig. 3b). In this stage, the ink is still confined at the corner of the syringe as shown in the inset in Fig. 3b. Third, the piston was pushed down to a certain graduation to compress the air inside the syringe, and then tied (Fig. 3c). In this stage, the ink is still confined at the corner of the syringe as shown in the inset in Fig. 3c. Now, the ink is deliberately moved to the tip of the syringe and entered into the fluidic conduit, physically segregating the two ends. Because only the silicone tube inserted into the outlet port was gas-permeable (plastic syringe, PMMA microdevice, and thick-walled silicone tube were considered gas-impermeable), air molecules diffused to the atmosphere exclusively through the outlet silicone tube, resulting in the propulsion of the sample toward the outlet of the microchannel. Fig. 3d shows a spontaneous ink flow inside the microchannel actuated by the syringe filled with large volume of compressed air.

Fig. 3 Procedures demonstrating self-actuated sample injection. (a) The red ink was sucked into the bottom corner of the syringe and the outlet silicone tube was made air-tight. (b) The syringe was connected to the inlet silicone tube. (c) The air inside the syringe was compressed. (d) The ink flow through the PMMA microchannel was actuated by large volume of compressed air in the syringe.

Evaluation of self-actuated liquid flow

A 20 mL syringe was used in evaluating a series of flows, and each experiment was performed three times for reproducibility. All the flow tests were conducted using an initial internal pressure of approximately 2 atm. This was achieved by pushing the piston from the initial graduation of 20 to 10. First, the flow phenomenon was analyzed by varying the length of the outlet silicone tube. The graph in Fig. 4a shows the time-dependent flow rate changes when the length of the outlet silicone tube was varied. As shown in Fig. 4a, the flow rate increased with increasing length of the outlet silicone tube, resulting in shorter residence time for each cycle, and vice versa. For example, when the lengths of the outlet silicone tubes were 3, 2, and 1 cm, the total running times were 21, 31, and 50 min, respectively, which corresponded to an average residence times of approximately 49, 72, and 119 s per cycle, respectively. The use of longer outlet silicone tube resulted in faster ink flow inside the PMMA microchannel due to faster air diffusion through the gas-permeable silicone tube in the outlet. From these results, we could conclude that the flow rate of sample could be controlled by adjusting the length of the outlet silicone tube.

Fig. 4 Effects of (a) length of the outlet silicone tube, (b) depth of the PMMA microchannel, and (c) total length of the PMMA microchannel (cycle number) on the speed and uniformity of the sample flow. Photos in (a) show relative positions of ink plug inside the serpentine microchannel at certain cycle numbers when the length of the outlet silicone tube was 3 cm.

Second, the flow phenomenon was analyzed by varying the depth of the PMMA microchannel. The graph in Fig. 4b shows the time-dependent flow rate changes when the depths of the microchannel were varied at 20, 50, and 100 μm, respectively. The width of the microchannel and the length of the outlet silicone tube were fixed at 200 μm and 3 cm, respectively. As shown in Fig. 4b, the flow rate increased with reducing depth of the PMMA microchannel, resulting in a shorter residence time for each cycle, and vice versa. For example, when the depths of the microchannels were 20, 50, and 100 μm, the total running times were 12, 21, and 27 min, respectively, resulting in average residence times of 28, 49, and 63 s, respectively. This was because, if the flow rates were identical for all three channels, the flow flux would be proportional to the channel depth, and the pressure gradient would be lower for a deeper channel, and vice versa. For this reason, the flow rate decreased with increasing channel depth, resulting in a longer residence time for each successive thermal cycle.

Third, the flow phenomenon was analyzed by varying the total length of the microchannel; that is, by varying the number of the serpentines, which is equivalent to the number of the thermal cycles. The width and depth of the microchannel were 200 and 80 μm, respectively, and the length of the outlet silicone tube was fixed at 3 cm. For total channel lengths of 1.25 m (25 cycles), 2.25 m (45 cycles), and 2.75 m (55 cycles), the total running times were 21, 37, and 49 min, respectively. However, the corresponding average residence times were approximately 49, 48, and 53 s, respectively. That is, the average residence times were almost identical regardless of the total length of the microchannel or the cycle number. This is probably because the PMMA microchannel itself is not gas-permeable, and the speed which determines the flow of the ink is the pressure gradient formed by the pressure difference in the anterior and the posterior ends of the ink, which in this case, was identical regardless of the total length of the microchannel. This was also the case with Fig. 4b, in that although different channel volume due to different channel depth resulted in different speed in the flow, the flow was seemingly uniform with slight fluctuations throughout the entire flow regardless of the depth of the microchannel.

Based on these results, we could rationally conclude that, among the three investigated parameters, the length of the outlet silicone tube was the most critical factor that affected the uniformity of liquid flow, presumably because of the gas permeability of the outlet silicone tube, which is the main cause for pressure gradient formation. These results match well with the hypothesis raised in the “Principle” section. Once a desired flow rate is established by finding the optimum length of the outlet silicone tube, the speed of the flow could further be fine-tuned by varying the depth of the microchannel. The number of the serpentines did not seem to affect the speed of flow significantly, which is very desirable because the required cycle number is likely to change depending on the size of the target to be amplified, and therefore, this factor can be neglected when finding the optimum flow rate. To summarize, the use of gas-permeable silicone tube in the outlet port triggered the flow of the liquid by forming pressure gradient between the anterior and posterior ends of the liquid plug, while the low permeability of PMMA microchannel aided in the maintenance of uniform flow rate over a long distance. A 32-fold fast-mode movie clip demonstrating almost constant ink flow inside a serpentine PMMA microchannel is presented as Movie S1 in the ESI.†

Polymerase chain reaction on a microdevice

Based on the above flow analyses, we performed a flow-through PCR using a PMMA microdevice. The PMMA microdevice was connected to a portable syringe and then placed on two heat blocks (Fig. 5a). The surface temperature of the PMMA substrate was measured using an infrared (IR) camera (FLIR Thermovision A320) (Fig. 5b). E. coli containing the pGEM-3Zf(+) plasmid vector, and a human genomic DNA (Roche) were used as the DNA templates for performing a two-temperature PCR.²⁷ The denaturation temperature was adjusted to 95 ± 0.5 °C, and the annealing/extension temperatures were adjusted to 68 ± 0.2 °C for the pGEM-3Zf(+) plasmid vector and 63 ± 0.1 °C for the human genomic DNA. The primer sequences for amplifying a 230 bp gene fragment of the pGEM-3Zf(+) plasmid vector were 5′-CCG GCG AAC GTG GCG AGA AAG GAA GGG AAG AAA GC-3′ (forward) and 5′-TCG CCT TGC AGC ACA TCC CCC TTT CGC CAG C-3′ (reverse). The primer sequences for amplifying the D1S80 locus^28–30 in human genomic DNA were 5′-GAA ACT GGC CTC CAA ACA CTG CCC GCC G-3′ (forward) and 5′-GTC TTG TTG GAG ATG CAC GTG CCC CTT GC-3′ (reverse).

Fig. 5 (a) Experimental setup for performing the flow-through PCR on a PMMA microdevice equipped with a portable syringe. (b) Temperature measurement. (c) Amplification of the 230 bp gene fragment obtained from the pGEM-3Zf(+) plasmid vector. (d) Amplification of the D1S80 locus (369–801 bp) obtained from human genomic DNA. For both gel images, lanes 1 and 2 are results obtained using thermal cycler and PMMA microdevice, respectively. Lane M is a 100 bp DNA size marker.

The PCR reagent contained a 5 × green-colored buffer, 0.2 mM dNTPs mixture, 1 mg mL⁻¹ BSA, 1 μM forward and reverse primers, and 0.075 U μL⁻¹ Taq polymerase. The commercially available human genomic DNA (200 ng μL⁻¹) was diluted to achieve a concentration of 5 ng μL⁻¹ in the PCR reagent, and the cultured E. coli solution was directly used to amplify the target gene (∼230 bp) in the pGEM-3Zf(+) plasmid vector. Briefly, 0.5 μL of the E. coli culture solution was centrifuged at 14 000 rpm for 10 min and the supernatant was discarded. The precipitate was then resuspended in 0.5 μL of distilled water. In Fig. 5c and d, the results of DNA amplification performed using the PMMA microdevice actuated by a portable syringe were compared with those obtained using a thermal cycler. In both cases, 25 cycles were used for the amplification, and the total running times were less than 30 min when using the microdevice. Lanes 1 and 2 in Fig. 5c show the 230 bp gene fragments obtained when using the thermal cycler and the PMMA microdevice, respectively. The intensity of the target amplicon obtained using the microdevice was approximately 97.2% of that obtained using the thermal cycler, based on an analysis conducted using the Image J software. Lanes 1 and 2 in Fig. 5d show the D1S80 loci amplified using the thermal cycler and the PMMA microdevice, respectively. For a target with size of approximately 500 bp, the intensity of the amplicon obtained using the microdevice was approximately 71.7% of that obtained using the thermal cycler. The sizes of the D1S80 locus, which is used for individual identification in forensic science, ranged between 369 and 801 bp.^28–31 In both cases, the target bands were successfully amplified using the PMMA microdevice with comparable intensities to those obtained using the thermal cycler. Owing to the high internal pressure maintained inside the microchannel throughout the sample flow, bubble generation was spontaneously suppressed during the heated operation.

Conclusion

In this study, we presented an innovative technology for the reliable and uniform flow of a liquid inside a microchannel over 2 m long under freely adjustable operation conditions using a portable plastic syringe. The use of a bulky mechanical pump requiring external electrical power supply was completely replaced by simple hands-on operation, which was sufficient to produce a driving force for transporting the liquid for performing a flow-through PCR on a microdevice. The high pressure generated inside the microchannel successfully resisted bubble formation even under the heated condition, and sample loss was prevented owing to the intrinsic low-gas-permeability of the PMMA. Using this system, two major issues associated when performing a flow-through PCR on chip, such as bubble formation and sample loss, was resolved. The introduced strategy can be widely applicable as a convenient tool, free of electrical input and bulky pumping apparatus such as syringe pump, in various fields of research where stable sample propulsion is required, such as when performing flow-through PCR directly on-site inside a simple serpentine microchannel, greatly enhancing fabrication easiness and device portability.

Acknowledgements

This research was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Science, ICT & Future Planning (2014R1A1A3051319). This research was also supported by the Gachon University research fund of 2014 (GCU-2014-0195).

Notes and references

1. H. T. G. van Lintel, F. C. M. van de Pol and S. Bouwstra, Sens. Actuators, 1988, 15, 153–167.
2. J. G. Smits, Sens. Actuators, A, 1990, 21–23, 203–206.
3. N. J. Graf and M. T. Bowser, Lab Chip, 2008, 8, 1664–1670.
4. K. Handique, D. T. Burke, C. H. Mastrangelo and M. A. Burns, Anal. Chem., 2001, 73, 1831–1838.
5. S. Santra, P. Holloway and C. D. Batich, Sens. Actuators, B, 2002, 87, 358–364.
6. J. Xie, J. Shih, Q. Lin, B. Yang and Y.-C. Tai, Lab Chip, 2004, 4, 495–501.
7. P. Woias, Sens. Actuators, B, 2005, 105, 28–38.
8. M. Yang, S. Sun, H. A. Bruck, Y. Kostov and A. Rasooly, Lab Chip, 2010, 10, 2534–2540.
9. C. Lui, S. Stelick, N. Cady and C. Batt, Lab Chip, 2010, 10, 74–79.
10. H. Lai and A. Folch, Lab Chip, 2011, 11, 336–342.
11. K. Pitchaimani, B. C. Sapp, A. Winter, A. Gispanski, T. Nishida and Z. H. Fan, Lab Chip, 2009, 9, 3082–3087.
12. T. Naito, R. Arayanarakool, S. L. Gac, T. Yasui, N. Kaji, M. Tokeshi, A. van den Berg and Y. Baba, Lab Chip, 2013, 13, 452–458.
13. G. Münchow, D. Dadic, F. Doffing, S. Hardt and K.-S. Drese, Expert Rev. Mol. Diagn., 2005, 5, 613–620.
14. S. Chakraborty, Lab Chip, 2005, 5, 421–430.
15. I. K. Dimov, L. Basabe-Desmonts, J. L. Garcia-Cordero, B. M. Ross, A. J. Ricco and L. P. Lee, Lab Chip, 2011, 11, 845–850.
16. P. Kauffman, E. Fu, B. Lutz and P. Yager, Lab Chip, 2010, 10, 2614–2617.
17. H. Noh and S. T. Phillips, Anal. Chem., 2010, 82, 8071–8078.
18. E. Fu, B. Lutz, P. Kauffman and P. Yager, Lab Chip, 2010, 10, 918–920.
19. L. Qin, O. Vermesh, Q. Shi and J. R. Heath, Lab Chip, 2009, 9, 2016–2020.
20. W. Wu, K.-T. Kang and N. Y. Lee, Analyst, 2011, 136, 2287–2293.
21. W. Wu, K. T. L. Trinh and N. Y. Lee, Analyst, 2012, 137, 983–990.
22. W. Wu and N. Y. Lee, Sens. Actuators, B, 2013, 181, 756–765.
23. W. Wu and N. Y. Lee, Anal. Bioanal. Chem., 2011, 400, 2053–2060.
24. W. Wu, K. T. L. Trinh and N. Y. Lee, Analyst, 2012, 137, 2069–2076.
25. C.-W. Tsao and D. L. DeVoe, Microfluid. Nanofluid., 2009, 6, 1–16.
26. D. J. Beebe, G. A. Mensing and G. M. Walker, Annu. Rev. Biomed. Eng., 2002, 4, 261–286.
27. H. H. Tran, K. T. L. Trinh and N. Y. Lee, J. Chromatogr. B: Anal. Technol. Biomed. Life Sci., 2013, 936, 88–94.
28. B. Budowle, R. Chakraborty, A. M. Giusti, A. J. Eisenberg and R. C. Allen, Am. J. Hum. Genet., 1991, 48, 137–144.
29. R. Roy, Forensic Sci. Int., 1997, 87, 63–71.
30. J. M. Hayes, B. Budowle and M. Freund, J. Forensic Sci., 1995, 40, 888–892.
31. N. Y. Lee, M. Yamada and M. Seki, Anal. Bioanal. Chem., 2005, 383, 776–782.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c4ra15473h

‡ Current address: Mechatronics Department, University of Saarland, Saarbrücken, Germany. KIST Europe GmbH, Saarbrücken, Germany.

§ Both these authors contributed equally to this work.

---