Langmuir 2008, 24, 13809-13813
13809
Highly Productive Droplet Formation by Anisotropic Elongation of a Thread Flow in a Microchannel Daisuke Saeki,†,‡ Shinji Sugiura,*,† Toshiyuki Kanamori,† Seigo Sato,‡ Sukekuni Mukataka,‡ and Sosaku Ichikawa‡ Research Center of AdVanced Bionics, National Institute of AdVanced Industrial Science and Technology (AIST), Central fifth, 1-1-1 Higashi, Tsukuba, Ibaraki 305-8565, Japan, Graduate School of Life and EnVironmental Sciences, UniVersity of Tsukuba, 1-1-1 Tennodai, Tsukuba, Ibaraki 305-8572, Japan ReceiVed August 25, 2008. ReVised Manuscript ReceiVed September 29, 2008 We developed a microfluidic device to form monodisperse droplets with high productivity by anisotropic elongation of a thread flow, defined as a threadlike flow of a dispersed liquid phase in a flow of an immiscible, continuous liquid phase. The thread flow was anisotropically elongated in the depth direction in a straight microchannel with a step, where the microchannel depth changed. Consequently, the elongated thread flow was given capillary instability (Rayleigh-Plateau instability) and was continuously transformed into monodisperse droplets at the downstream area of the step in the microchannel. We examined the effects of the flow rates of the dispersed phase and the continuous phase on the droplet formation behavior, including the droplet diameter and droplet formation frequency. The droplet diameter increased as the fraction of the dispersed-phase flow rate relative to the total flow rate increased and was independent of the total flow rate. The droplet formation frequency proportionally increased with the total flow rate at a constant dispersed-phase flow rate fraction. These results are explained in terms of a mechanism similar to that of droplet formation from a cylindrical liquid thread flow by Rayleigh-Plateau instability. The microfluidic device described was capable of forming monodisperse droplets with a 160-µm average diameter and 3-µm standard deviation at a droplet formation frequency of 350 droplets per second from a single thread flow. The highest total flow rate achieved was 6 mL/h using the present device composed of a straight microchannel with a step. We also demonstrated parallel droplet formation by anisotropic elongation of multiple thread flows; the process was applied to form W/O and O/W droplets. The highly productive droplet formation process presented in this study is expected to be useful for future industrial applications.
Introduction Monodisperse emulsion droplets have been used to prepare highly functional particles such as polymer beads and microcapsules.1-6 For these applications, control of the size and size distribution of the emulsion droplets is important. Conventional emulsification techniques such as membrane emulsification and mechanical stirring emulsification have the disadvantage of forming polydisperse droplets, although these methods do exhibit high productivity. Recently, spatiotemporal pattern formation in microfluidic devices has been reported.7-9 Microfluidic devices for the formation of monodisperse droplets with micrometer size have especially attracted attention. Thorsen et al. developed a microfluidic device with a T-junction structure, in which a water phase is sheared by an oil phase flow to form water-in-oil (W/O) droplets.7 The droplet size is controlled by changing the flow * To whom correspondence should be addressed. Email: shinji.sugiura@ aist.go.jp. Phone: +81-29-861-6286. Fax: +81-29-861-6278. † National Institute of Advanced Industrial Science and Technology. ‡ University of Tsukuba.
(1) Nisisako, T.; Torii, T.; Higuchi, T. Chem. Eng. J. 2004, 101, 23–29. (2) Nie, Z. H.; Xu, S. Q.; Seo, M.; Lewis, P. C.; Kumacheva, E. J. Am. Chem. Soc. 2005, 127, 8058–8063. (3) Nie, Z. H.; Li, W.; Seo, M.; Xu, S. Q.; Kumacheva, E. J. Am. Chem. Soc. 2006, 128, 9408–9412. (4) Seo, M.; Nie, Z. H.; Xu, S. Q.; Mok, M.; Lewis, P. C.; Graham, R.; Kumacheva, E. Langmuir 2005, 21, 11614–11622. (5) Sugiura, S.; Nakajima, M.; Itou, H.; Seki, M. Macromol. Rapid Commun. 2001, 22, 773–778. (6) Sugiura, S.; Oda, T.; Izumida, Y.; Aoyagi, Y.; Satake, M.; Ochiai, A.; Ohkohchi, N.; Nakajima, M. Biomaterials 2005, 26, 3327–3331. (7) Thorsen, T.; Roberts, R. W.; Arnold, F. H.; Quake, S. R. Phys. ReV. Lett. 2001, 86, 4163–4166. (8) Cubaud, T.; Mason, T. G. Phys. ReV. Lett. 2006, 96, 4. (9) Cubaud, T.; Mason, T. G. Physics of Fluids 2007, 19, 1.
rates of the two phases. Additionally, the T-junction structure has been used to prepare oil-in-water (O/W) emulsions,10 waterin-oil-in-water (W/O/W) emulsions,11,12 functional polymer beads,1 and microbubbles.13 Anna et al. developed a microfluidic device with a flow-focusing structure, in which a water-phase flow is focused in an oil-phase flow at a narrow orifice and sheared by the oil phase to form W/O droplets.14 Similar structures were reported by other groups2-4,15-17 and have been used to prepare W/O/W emulsions18-21 and polymer beads.2-4 However, applying these various devices to industrial processes is difficult because of the low productivity of the devices. To improve the productivity, Yobas et al. developed a microfluidic device with a three-dimensional flow-focusing structure.22 This device forms W/O and O/W droplets with high productivity from a single (10) Xu, J. H.; Li, S. W.; Tan, J.; Wang, Y. J.; Luo, G. S. Langmuir 2006, 22, 7943–7946. (11) Okushima, S.; Nisisako, T.; Torii, T.; Higuchi, T. Langmuir 2004, 20, 9905–9908. (12) Nisisako, T.; Okushima, S.; Torii, T. Soft Matter 2005, 1, 23–27. (13) Garstecki, P.; Fuerstman, M. J.; Stone, H. A.; Whitesides, G. M. Lab Chip 2006, 6, 437–446. (14) Anna, S. L.; Bontoux, N.; Stone, H. A. Appl. Phys. Lett. 2003, 82, 364– 366. (15) Tan, Y. C.; Cristini, V.; Lee, A. P. Sens. Actuators, B 2006, 114, 350–356. (16) Xu, Q. Y.; Nakajima, M. Appl. Phys. Lett. 2004, 85, 3726–3728. (17) Haeberle, S.; Zengerle, R.; Ducree, J. Microfluid. Nanofluid. 2007, 3, 65–75. (18) Lorenceau, E.; Utada, A. S.; Link, D. R.; Cristobal, G.; Joanicot, M.; Weitz, D. A. Langmuir 2005, 21, 9183–9186. (19) Utada, A. S.; Lorenceau, E.; Link, D. R.; Kaplan, P. D.; Stone, H. A.; Weitz, D. A. Science 2005, 308, 537–541. (20) Takeuchi, S.; Garstecki, P.; Weibel, D. B.; Whitesides, G. M. AdV. Mater. 2005, 17, 1067–1072. (21) Huang, S. H.; Tan, W. H.; Tseng, F. G.; Takeuchi, S. J. Micromech. Microeng. 2006, 16, 2336–2344. (22) Yobas, L.; Martens, S.; Ong, W. L.; Ranganathan, N. Lab Chip 2006, 6, 1073–1079.
10.1021/la802776z CCC: $40.75 2008 American Chemical Society Published on Web 11/06/2008
13810 Langmuir, Vol. 24, No. 23, 2008
flow-focusing structure. However, the complicated threedimensional structure is unsuitable for parallelization and integration of the multiple structures for further increases in productivity. Highly productive microfluidic devices with simple structures suitable for parallelization are desirable for future industrial applications. A liquid thread flow from an orifice or a cylindrical capillary is conventionally used to form droplets.23-28 In these methods, droplets are formed by capillary instability (Rayleigh-Plateau instability),28 which is caused by interfacial tension. Droplet formation by Rayleigh-Plateau instability can be highly productive, yielding close to 1000 droplets per second.25 In the present study, we developed a microfluidic device composed of a straight microchannel with a step, where the microchannel depth changed, to form monodisperse micrometersize droplets with high productivity. A thread flow was anisotropically elongated at the step and transformed into droplets by Rayleigh-Plateau instability. We characterized the droplet formation behavior by systematically examining the effects of the step structure and the flow rates of the dispersed phase and the continuous phase. We also designed an integrated microfluidic device to elongate multiple thread flows, and we used this highly productive integrated device to form W/O and O/W droplets.
Experimental Section Materials. A silicon wafer was obtained from Asahi Metal (Tokyo, Japan). Negative photoresists, SU-8 10 and SU-8 50, were obtained from MicroChem Corp. (Newton, MA, USA). Polydimethylsiloxane (PDMS) prepolymer and curing agent, Sylgard 184, were obtained from Dow Corning (Midland, MI, USA). Tridecafluoro-1,1,2,2tetrahydrooctyl-1-trichlorosilane was obtained from Gelest (Morrisville, NY, USA). 1-Hexanol was obtained from Wako Pure Chemical Industries (Osaka, Japan) and used as the oil phase. Tetraglycerin-condensed ricinoleic acid ester (TGCR) was obtained from Sakamoto Yakuhin Kogyo Co. (Osaka, Japan) and used to stabilize W/O droplets. Sodium dodecyl sulfate (SDS) was obtained from Wako Pure Chemical Industries and used to stabilize O/W droplets. Milli-Q water was used as the water phase. Fabrication of Microfluidic Devices. PDMS microfluidic devices with microchannels of various depths were fabricated by means of multilayer photolithography29,30 and soft-lithography techniques.31 First, a silicon wafer was spin-coated with negative photoresist SU-8 10. A photomask pattern of shallow microchannels was transferred to the photoresist layer by means of a mask aligner (K-310P100S; Kyowa Riken Inc., Tokyo, Japan). Second, the silicon wafer with the SU-8 10 photoresist pattern was spin-coated with SU-8 50, and a photomask pattern of deep microchannels was transferred to the photoresist layers to create a negative master structure. After development in ethyl lactate, the negative master was placed in the vapor of (tridecafluoro-1,1,2,2-tetrahydrooctyl)trichlorosilane for 2 h to facilitate easy detachment of the PDMS replica from the master. PDMS prepolymer and curing agent were mixed, and the mixture was poured onto the master wafer. After curing at 90 °C for 1 h, the PDMS replica plate was peeled from the master wafer The PDMS replica plate and a flat PDMS plate were oxidized by oxygen plasma discharge with a plasma reactor (PR500; Yamato Scientific Co., (23) Utada, A. S.; Fernandez-Nieves, A.; Stone, H. A.; Weitz, D. A. Phys. ReV. Lett. 2007, 99, 094502. (24) Utada, A. S.; Fernandez-Nieves, A.; Gordillo, J. M.; Weitz, D. A. Phys. ReV. Lett. 2008, 100, 014502. (25) Seifert, D. B.; Phillips, J. A. Biotechnol. Prog. 1997, 13, 562–568. (26) Ganan-Calvo, A. M. Phys. ReV. Lett. 1998, 80, 285–288. (27) Tomotika, S. Proc. R. Soc. London Ser. A 1935, 150, 322–337. (28) Rayleigh, L. Proc. London Math. Soc. 1878, s1-10, 4–13. (29) Hung, P. J.; Lee, P. J.; Sabounchi, P.; Aghdam, N.; Lin, R.; Lee, L. P. Lab Chip 2005, 5, 44–48. (30) Sugiura, S.; Edahiro, J.; Kikuchi, K.; Sumaru, K.; Kanamori, T. Biotechnol. Bioeng. 2008, 100, 1156–1165. (31) Duffy, D. C.; McDonald, J. C.; Schueller, O. J. A.; Whitesides, G. M. Anal. Chem. 1998, 70, 4974–4984.
Saeki et al.
Figure 1. Microchannel design and schematics of W/O droplet formation in a straight microchannel with a step. (a) Microscope image of the fabricated microchannel. (b) Schematic of W/O droplet formation in a top view. (c) Schematic of W/O droplet formation in a cross-sectional view at the channel center.
Tokyo, Japan) and were bonded irreversibly.31 For formation of O/W droplets, the bonded PDMS plates were used immediately after bonding, to maintain their hydrophilicity. For formation of W/O droplets, the bonded plates were heated at 90 °C for 2 h to recover surface hydrophobicity. The hydrophobicity of the microchannel surface was evaluated by measurement of the water-in-air static contact angle with a contact angle meter (Drop Master 300, Kyowa Interface Science, Niiza, Japan). The static contact angle of fabricated PDMS plates was about 25° after oxygen plasma treatment and about 110° after heating for hydrophobicity recovery. Formation of W/O Droplets and O/W Droplets. Figure 1a shows the fabricated straight microchannel with a step. For formation of W/O droplets, water was used as the dispersed phase, and 1-hexanol containing 5 wt % TGCR was used as the continuous phase. For formation of O/W droplets, 1-hexanol was used as the dispersed phase, and water containing 1 wt % SDS was used as the continuous phase. The two immiscible fluids were injected into the microchannels at constant flow rates controlled by syringe pumps (Pico Plus; Harvard Apparatus Co., Holliston, MA, USA). Under appropriate flow rate conditions, a thread flow was formed at the junction of the dispersedphase flow and the continuous-phase flow (Figure 1b), and the thread flow was transformed into droplets at the downstream area of the step, where the microchannel depth changed (Figure 1c). The droplet formation was observed with an inverted microscope (IX-71; Olympus, Tokyo, Japan) with a CCD camera (DFW-SX910; Sony, Tokyo, Japan). Analysis of Droplet Diameter and Droplet Formation Frequency. The droplet diameter was determined from pictures obtained with the microscope. For disk-shaped droplets with diameters larger than the microchannel depth, the droplet diameter corresponding to a spherical droplet was calculated from the volume of the observed disk-shaped droplet. The droplet formation frequency in droplets per second [-/s] was calculated from the dispersed-phase flow rate and the droplet diameter.
Results and Discussion We investigated W/O droplet formation behavior using the hydrophobic straight microchannel with a step. We observed several different types of droplet formation behavior at various ratios of the dispersed-phase flow rate (QD) and the total flow rate (QT) (Figure 2). We classified the droplet formation behavior into a flow-focusing mode, a dripping mode, or a jetting mode. At low QD/QT values, W/O droplets were formed at the junction
Anisotropic Elongation of Thread Flow
Langmuir, Vol. 24, No. 23, 2008 13811
Figure 2. Photomicrographs of W/O droplet formation at various dispersed-phase flow rates (QD) and total flow rates (QT). The microchannel depth downstream of the step was 142 µm. The continuous-phase flow rates (QC) were 0.1 mL/h (a, b, c), 1.5 mL/h (d), and 2.7 mL/h (e).
of the dispersed-phase flow and the continuous-phase flow in the flow-focusing mode (Figure 2a). In this mode, the droplets were formed by means of the mechanism observed for the flow-focusing structure,14 in which droplets are formed by a shear stress from a continuous-phase flow. At high QD/QT values, a thread flow was formed at the junction of the two phases and was transformed into W/O droplets at the downstream area of the step (parts b-e of Figure 2). The diameter of the formed droplets increased with increasing QD/QT (parts b and c of Figure 2), and the droplet formation behavior changed with increasing QT (parts c-e of Figure 2). At low QT values, the droplets were formed at the step in the dripping mode (Figure 2c); whereas at high QT values, the thread flow continued at the downstream area of the step and then transformed into droplets owing to Rayleigh-Plateau instability in the straight microchannel in the jetting mode (parts d and e of Figure 2). Droplet formation in the jetting mode was similar to that reported for an orifice, a cylindrical capillary, and a flow-focusing structure in square microchannels.23-28,32 In this mode, the thread flow of the dispersed phase breakup into droplets due to the surface-tension-driven instability. In the present study, the Rayleigh-Plateau instability was induced by anisotropic elongation of the thread flow in the depth direction (along the z axis in Figure 1c) at the step structure fabricated in the straight microchannel. Droplet formation in the jetting mode is advantageous because this mode allows droplet formation under highflow-rate conditions. We investigated the effects of QD/QT and the microchannel depth downstream of the step on the W/O droplet diameter (D) at a constant continuous-phase flow rate (QC) of 0.1 mL/h (Figure 3). When QD/QT was less than 0.17, W/O droplets were formed at the junction of the two phases in the flow-focusing mode (Figure 2a) and D was independent of QD/QT (Figure 3). In this mode, the droplets were formed by the shear stress induced by the continuous phase flow.14,15 Under the constant QC condition, the value of D was independent of the QD because the droplets were formed by the constant shear stress.22 When QD/QT was larger than 0.17, the droplets were formed at the step in the dripping mode (Figure 2b and Figures 2c). The value of D increased with increasing QD/QT, whereas the microchannel depth downstream of the step did not affect the droplet diameter (Figure (32) Cubaud, T.; Mason, T. G. Phys. Fluids 2008, 20, 053302.
Figure 3. Effects of QD/QT and microchannel depth downstream of the step on the W/O droplet diameter (D) at QC ) 0.1 mL/h. The microchannel depths downstream of the step were 43 µm (closed circles), 67 µm (open squares), and 142 µm (closed triangles).
3). In the QD/QT range from 0.17 to 0.67, the relationship between D and QD/QT was linear; least-squares analysis of the data resulted in the following equation: D ) 227QD/QT + 37.4. This empirical equation predicts D in a simple way within a 7.6% mean deviation of the experimental results from the calculated value. At QD/QT values higher than 0.67, the formed droplet diameter was larger than that predicted by the linear empirical equation. Generally, viscous flow in the rectangular microchannel has threedimensional parabolic flow velocity distribution.33 Therefore, the flow velocity of the dispersed phase in the central area of the thread flow was faster than that near to the water/oil interface. This flow velocity distribution in the thread flow was significant under high QD/QT conditions because the water/oil interface was nearer to the microchannel wall. At QD/QT values higher than 0.67, this flow velocity distribution was not negligible: the fast flow in the central area of the thread flow induced that the large amount of the dispersed phase flowed into the forming droplets during the transformation of the thread flow into the droplets. This phenomenon probably resulted in the formation of the larger droplets than that predicted by the empirical equation. (33) White, F. M. Viscous fluid flow, 3rd ed.; McGraw-Hill: New York, 2006; Chapter 3.
13812 Langmuir, Vol. 24, No. 23, 2008
Figure 4. Effect of QT on the droplet diameter (squares) and the droplet formation frequency (circles) of the W/O droplets at QD/QT ) 0.5. The microchannel depth downstream of the step was 142 µm. Error bars indicate the standard deviations.
Figure 5. Effect of QT on the droplet formation distance of the W/O droplets at QD/QT ) 0.5. The microchannel depth downstream of the step was 142 µm.
As QT was increased, the droplet formation behavior transitioned from the dripping mode (Figure 2c) to the jetting mode (parts d and e of Figure 2). Droplet formation in the jetting mode is useful for industrial applications because the droplet formation frequency is very high. We investigated the effect of QT on the droplet diameter, the droplet formation frequency, and the droplet formation distance of the W/O droplet at a constant QD/QT of 0.5. The droplet formation distance (defined in parts d and e of Figure 2) corresponds to the distance from the step to the droplet formation point. The droplet diameter was independent of QT, and the droplet formation frequency increased proportionally with QT (Figure 4). The droplet diameter distribution was narrow,
Saeki et al.
with a standard deviation less than 3 µm (Figure 4). By use of the straight microchannel with a step, we were able to improve the droplet productivity without changing the droplet diameter by simply increasing QT at a constant QD/QT. The transition from the dripping mode to the jetting mode was observed at a QT of 2.9 mL/h under a constant QD/QT condition of 0.5 (Figure 5). At QT values higher than 2.9 mL/h, the droplets were formed at the far downstream area from the step, and the droplet formation distance linearly increased with QT (Figures 2e and Figures 5). The droplet diameter and droplet diameter distribution did not change over the transition from the dripping mode to the jetting mode (Figure 4). This result indicates that, in both modes, the droplets were formed by Rayleigh-Plateau instability induced by anisotropic elongation of the thread flow. During the droplet formation by Rayleigh-Plateau instability, the anisotropically elongated thread flow was transformed into spherical droplets by the surface tension because the surface area of the thread flow was enlarged by the anisotropic elongation. Droplet formation in the jetting mode probably resulted from the fact that the flow velocity was faster than the time required for the transformation of the thread flow into droplets. At QT values lower than 2.9 mL/h, the flow velocity was slower than the time required for the transformation. Therefore, the thread flow was transformed into droplets at a point close to the step. At high QT, the thread flow stream moved far from the step within the time required for transformation. In a study of droplet formation by Rayleigh-Plateau instability in a thread flow using a microcapillary device, Utada et al. found that the droplet diameter is proportional to the diameter of the thread flow (the “jet diameter”).23 In our experiments, the jet diameter correlates to the cross-sectional area of the thread flow upstream of the step, and this cross-sectional area depended on QD/QT. Therefore, the correlation between the droplet diameter and the jet diameter reported by Utada et al. is consistent with our experimental result indicating that the droplet diameter was influenced by QD/QT but was independent of QT. Droplet formation in microchannels with step structures has been previously reported by other groups.34-37 The droplet formation mechanism was explained in terms of spontaneous transformation caused by interfacial tension of a distorted dispersed phase on the step structures.36 At a high flow rate, the structures cannot form monodisperse droplets, owing to viscous force.38 Therefore, improving the productivity of monodisperse droplets is difficult. Our device is different from the reported devices in that the step was fabricated in a straight microchannel.
Figure 6. Parallelized droplet formation by anisotropic elongation of multiple thread flows. The microchannel depth downstream of the step was 67 µm. (a) Formation of W/O droplets from the hydrophobic microchannel at QD ) 0.1 mL/h and QC ) 0.2 mL/h. (b) Formation of O/W droplets from the hydrophilic microchannel at QD ) 0.05 mL/h and QC ) 0.2 mL/h.
Anisotropic Elongation of Thread Flow
The structure enabled stable droplet formation from the thread flow at the downstream area of the step in the microchannel by Rayleigh-Plateau instability: The anisotropically elongated thread flow was spontaneously transformed into same-sized droplets due to the surface-tension-driven instability. Monodisperse droplets with diameters of 160 µm were formed at a frequency of 350 droplets per second from the single thread flow, and QT was as high as 6 mL/h. Our device provides a simpler structure and better productivity for the formation of 160-µm droplets than the previously reported three-dimensional flow-focusing structure.22 The highest droplet formation frequency in our device was restricted by the downstream length of the step. The downstream length of the present device was 9 mm, which was not sufficient to form droplets at QT higher than 6 mL/h. A higher droplet formation frequency can be realized by adopting deep and long microchannel downstream of the step. To further improve the productivity, we fabricated an integrated microfluidic device for anisotropic elongation of multiple thread flows. In a device with a hydrophobic microchannel, an array of thread flows was simultaneously transformed into monodisperse W/O droplets by anisotropic elongation at the step in the straight microchannel (Figure 6a). O/W droplets were also formed in a device with the same geometry but with a hydrophilic (34) Chan, E. M.; Alivisatos, A. P.; Mathies, R. A. J. Am. Chem. Soc. 2005, 127, 13854–13861. (35) Priest, C.; Herminghaus, S.; Seemann, R. Appl. Phys. Lett. 2006, 88, 021406. (36) Sugiura, S.; Nakajima, M.; Iwamoto, S.; Seki, M. Langmuir 2001, 17, 5562–5566. (37) Kawakatsu, T.; Kikuchi, Y.; Nakajima, M. J. Am. Oil Chem. Soc. 1997, 74, 317–321. (38) Sugiura, S.; Nakajima, M.; Kumazawa, N.; Iwamoto, S.; Seki, M. J. Phys. Chem. B 2002, 106, 9405–9409.
Langmuir, Vol. 24, No. 23, 2008 13813
microchannel (Figure 6b). The straight microchannel with the step is more suitable for parallelization than the previously reported structures, such as T-junction structures and flowfocusing structures. Droplet formation by means of anisotropic elongation of the thread flow is expected to be useful for highly productive droplet formation in future industrial applications.
Conclusion We developed a novel microfluidic device for the formation of monodisperse droplets with high productivity. The device had a straight microchannel with a step, where an anisotropically elongated thread flow was transformed into droplets by Rayleigh-Plateau instability. The droplet diameter increased with QD/QT and was independent of the microchannel depth downstream of the step. At constant QD/QT, an increase in QT induced the transition of droplet formation behavior from a dripping mode to a jetting mode at a QT of 2.9 mL/h. The increase in QT also induced an increase in the droplet formation frequency while maintaining the droplet diameter, both in the dripping mode and in the jetting mode. The device produced monodisperse droplets from a single thread flow with a QT as high as 6 mL/h. We also demonstrated the formation of W/O droplets and O/W droplets by anisotropic elongation of multiple thread flows for further increases in productivity. Droplet formation by anisotropic elongation of a thread flow is expected to create new opportunities for industrial applications of microfluidic droplet formation devices owing to the high productivity of the method. Acknowledgment. A part of this work was conducted at the AIST Nano-Processing Facility, supported by “Nanotechnology Network Japan” of the Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan. LA802776Z
