Structural Formation of Oil-in-Water (O/W) and Water-in-Oil-in-Water (W/O/W) Droplets in PDMS Device Using Protrusion Channel without Hydrophilic Surface Treatment

This paper presents a simple method of droplet formation using liquids that easily wet polydimethylsiloxane (PDMS) surfaces without any surface treatment. Using only structural features and uniform flow focusing, Oil-in-Water (O/W) and Water-in-Oil-in-Water (W/O/W) droplets were formed in the full PDMS structure. Extrusion channel and three-dimensional flow focusing resulted in effective fluidic conditions for droplet formation and the droplet size could be precisely controlled by controlling the flow rate of each phase. The proposed structure can be utilized as an important element for droplet based research, as well as a droplet generator.


Introduction
Microdroplet technologies have been developed for quantitative volume control and the efficient isolation of samples in various chemical [1,2], biological [3][4][5][6], and material [7,8] research fields. The formation, mixing, and separation of droplets allow for the preparation, reaction, treatment, and analysis of specific samples.
In particular, droplet formation is the first step and is sometimes the final goal of all droplet-related research. Using immiscible liquids and specific methods, the liquid materials are isolated into individual droplets. Individual liquids from a dispersed phase drop off from the liquid lump into an immiscible surrounding fluid of a continuous phase that fills the gaps between droplets. Therefore, an external shear force from the surrounding fluid greater than the surface tension of the target liquid is essential for drop separation. Moreover, the target liquid should be detached from solid phase wall surfaces. These two conditions are fundamental requirements for droplet formation.
In the field of microfluidics, various types of microchannels for droplet formation have been developed to obtain sufficient shear force from the carrier fluid because the importance of surface tension increases with scale-down [9,10]. T-shaped [11][12][13] or cross [14][15][16] channels are the conventional design for microdroplet formation and the size and generation frequency of the microdroplets are controlled by the flow rates of the injected liquids.
However, contact of the liquid with channel walls is strongly dependent on the intrinsic wettability of the liquid; in this way, the type of droplet that can be formed is specified. Oil-in-water (O/W) droplets cannot be formed in a hydrophobic structure because the oil occupies the wall surface more

Principle and Computational Analysis
To form droplets of a highly wetting liquid in a PDMS structure, the isolation of the liquid from the wall surface is the most critical and important step. For instance, oil phase liquids are highly wettable on the PDMS surface when compared to water phase liquids; the oil cannot be detached from the surface easily and form a droplet without a hydrophilic surface treatment. Therefore, stronger hydrodynamic forces or specific structures are necessary to prevent the oil from spreading or flowing along the surface. Conventionally, Romanowsky et al. [35] reported three-dimensional (3D) structures for working with double emulsions. They utilized a sandwiched focusing structure and a vertically high downstream channel for enhancing top-and bottom-side focusing and isolation. This combination achieved semi-3D focusing flow and it was possible to form droplets of some wettable liquids in the PDMS structure. However, depending on the materials and fluidic conditions, oil would still contact the top-or bottom-side wall surfaces. Therefore, oil droplets cannot be formed successfully, as shown in Figure 1A. In addition, even in the case of full 3D focusing flow, the oil may spread along the channel surface beside a low velocity area at the flat-end junction.
Therefore, we propose a protrusion-end structure to stably form oil phase droplets and prevent the oil from spreading by ensuring stronger wetting than flow focusing in the PDMS structure, as shown in Figure 1B. The protrusion structure is isolated from the wall and it avoids undesirable contact with the oil phase liquid on the surface and eliminates the zero velocity area at the junction of the immiscible liquids. Therefore, the functional structure allows for droplet formation from various liquids that are highly wettable on the device surface. Using this structure, multi-layered droplets can be also formed, as shown in Figure 1C. Furthermore, the three-dimensional channel structure guarantees more uniform and stable focusing to form droplets. focusing flow, the oil may spread along the channel surface beside a low velocity area at the flat-end junction. Therefore, we propose a protrusion-end structure to stably form oil phase droplets and prevent the oil from spreading by ensuring stronger wetting than flow focusing in the PDMS structure, as shown in Figure 1B. The protrusion structure is isolated from the wall and it avoids undesirable contact with the oil phase liquid on the surface and eliminates the zero velocity area at the junction of the immiscible liquids. Therefore, the functional structure allows for droplet formation from various liquids that are highly wettable on the device surface. Using this structure, multi-layered droplets can be also formed, as shown in Figure 1C. Furthermore, the three-dimensional channel structure guarantees more uniform and stable focusing to form droplets. To verify the effects of the protrusion structure, we performed computational analysis (CFD-ACE+, ESI Group), as shown in Figure 2. The module used was "Flow" and the flow distribution in the structure under steady state was calculated. The working fluid was water for this analysis and it was injected from the normal direction into each inlet. Figure 2A,B shows structure designs and flow fields in the center of the structures and Figure 2C shows velocity distributions in the axial direction at the points away from junction-end. The flow rate of the focusing flow from the four branch channels is five times higher than the flow rate from the core channel.
As shown in the results, a low velocity area was formed near the junction of the core and focusing channels in the case of the flat-end junction. Thus, the high wetting liquid from the core channel spreads and flows onto the downstream wall along the area without hydrodynamic focusing or sweeping.
In contrast, the protrusion-end increases the flow velocity next to the junction area. Therefore, the spreading and wetting of the liquid along the surface of the protrusion structure could be strongly suppressed by enhanced hydrodynamic forces. Consequently, it is possible to form droplets of the liquid, which wets more strongly on the channel surface than the continuous phase liquid, by the perfect isolation of the dispersed phase liquid from the wall. To verify the effects of the protrusion structure, we performed computational analysis (CFD-ACE+, ESI Group), as shown in Figure 2. The module used was "Flow" and the flow distribution in the structure under steady state was calculated. The working fluid was water for this analysis and it was injected from the normal direction into each inlet. Figure 2A,B shows structure designs and flow fields in the center of the structures and Figure 2C shows velocity distributions in the axial direction at the points away from junction-end. The flow rate of the focusing flow from the four branch channels is five times higher than the flow rate from the core channel.
As shown in the results, a low velocity area was formed near the junction of the core and focusing channels in the case of the flat-end junction. Thus, the high wetting liquid from the core channel spreads and flows onto the downstream wall along the area without hydrodynamic focusing or sweeping.
In contrast, the protrusion-end increases the flow velocity next to the junction area. Therefore, the spreading and wetting of the liquid along the surface of the protrusion structure could be strongly suppressed by enhanced hydrodynamic forces. Consequently, it is possible to form droplets of the liquid, which wets more strongly on the channel surface than the continuous phase liquid, by the perfect isolation of the dispersed phase liquid from the wall.

Device Design and Fabrication
The three-dimensional (3D) structure used to form O/W and W/O/W droplets was fabricated by stacking multiple PDMS units. Each unit was designed for uniform flow focusing and isolated tubular channels, as well as the introduction of each phase fluids. As shown in Figure 3, focusing channels, junctions, and the protrusion channel were formed in the units and they were combined to form a 3D structure. Three types of units were used to form oil phase (O/W) droplets and an additional unit was used to form W/O/W droplets.    In the case of the W/O/W droplet formation device, the first unit includes an injection channel and the second unit consists of two layers: a cylindrical junction layer and an injection layer consisting of four branch channels for uniform flow focusing. There are two layers in the third unit for a through-hole and a protruded tubular channel. Finally, the forth unit of two layers also consists of a four-branch focusing channel and a cylindrical through-hole.
The units were formed by a PDMS molding process (SILPOT 184, Dow Corning Toray Co., Ltd., Tokyo, Japan) while using SU-8 molds (SU-8 3010/3025/3050, MicroChem Corp., Westborough, MA, USA) that were fabricated by a two-step photoresist patterning process. A PDMS guide film was employed to form penetrating structures and to successfully handle the thin unit during the molding process. After an oxygen plasma treatment on the surface of the PDMS units, the units were aligned and bonded under a microscope. Then, the guide film was removed and the next unit was bonded onto the previous units. Inlet and outlet layers were bonded onto the first and last units, respectively. Finally, additional PDMS was poured on the toppled structure to allow for clear observation and all PDMS devices were baked at 100 • C for at least one week to eliminate any effects of the plasma treatment and to obtain inherent hydrophobic surface. All of the liquids were introduced into the device by syringes and syringe pumps (KDS210, kdScientific, Holliston, MA, USA). The droplet formation process was observed while using a high speed camera (FASTCAM-NEO, Photron, Tokyo, Japan). The size of the O/W or W/O/W droplets was evaluated by pixel counting and calculation from the captured images.

Oil-in-Water (O/W) Droplet Formation
The formation of oil droplets by water phase flow in a PDMS device is shown in Figure 4 (Supplementary Material, Video S1). At the end of the protrusion channel, the oil phase flow was focused by the water flow and was isolated from the wall surfaces. Therefore, the oil, which is highly wettable on PDMS surfaces, can be formed into dispersed droplets without continuous flow along the wall. The wetting of oil on the protrusion structure upstream was observed, but fully developed, continuous, high velocity flow beside the structure efficiently suppressed the wetting and spreading of the oil along the wall surface. Moreover, the droplets did not make contact with the downstream channel surface because of the continuous carrier flow, even when the droplet diameter was larger than the inner diameter of the downstream channel.   In contrast, Figure 5A shows the experimental result without the protrusion tip. In this case, we used an oil injection channel 20 µm in diameter, which is smaller than that of the protrusion tip, and a high ratio of oil phase and water phase flow rates. However, in this case, the water phase flow could not suppress the oil flow along the wall surface; thus, only water droplets were formed.
When the oil was injected into the junction area before the water, our proposed device also allowed for the formation of water phase droplets only because of the dominant wall occupancy of the oil. However, the oil wetting can be washed by flowing water at a temporarily high flow rate, such as a short syringe push, as shown in Figure 5B (Supplementary Material, Video S2). The water flow did not concede the wall surface to the oil again after the washing. This result has been achieved in the protrusion-end device only, and this recovery is very useful for both end users and practical experiments.
The oil droplet size was controlled by adjusting the flow rates of each phase fluid, as shown in Figure 6. Similar to other droplet formation mechanisms, the droplet size was proportional to the flow rate of the dispersed phase fluid and inversely proportional to the flow rate of the continuous phase fluid. When the flow rate of the dispersed phase was about 1.5 times as large as the flow rate of the continuous phase, stable droplet formation in the dripping mode was changed to the unstable jetting mode or continuous flow. The droplet length ranged from 180 to 330 µm in the structure, and the standard deviation was smaller than 5% under the fluidic conditions of the stable dripping mode.

Water-in-Oil-in-Water (W/O/W) Droplet Formation
The proposed structure was applied to double emulsion (W/O/W droplet) formation, as shown in Figure 7 (Supplementary Material, Video S3). Water cores (W/O droplets) were formed at the first junction in the second unit (refer, Figure 3), then the oil droplets, including the water cores (W/O/W), were formed at the second junction (protrusion-tip). The oil droplets were successfully formed, even when the oil contained water phase droplets inside. In this case, the structure was efficiently prevented from contacting the oil and PDMS channel surfaces. Furthermore, the oil wetting upstream was suppressed by the continuous flow.  The W/O/W droplet size was also controlled via controlling the flow rates of each fluid phase ( Figure 8). Similar to the results of O/W droplet formation, the droplet size depended on the ratio between the oil phase flow containing a water core and the continuous outer flow of water phase. However, the inner water cores reinforced the oil layer from inside, thus the total size of the W/O/W droplets was larger than the O/W droplets, as shown in Figure 8C. Furthermore, the flow range that was capable of stable droplet formation was limited as compared to the case without inner cores, as shown in Figure 8D. Finally, Figure 8E shows the dramatic influence of the inner cores and oil on the droplet length. In the case where the flow rate of inner water was 0.5 µL/min, the droplet size first decreased and then increased with increasing oil flow rate. In the range of low oil flow rate, the size and density of the water cores were high. Therefore, the W/O/W droplet size decreased due to the reduced size and density of the water core, even though the volumetric oil flow rate increased. In contrast, when the core size and density deviated from the dominant range, the total droplet size increased with the flow rate of the oil phase. Finally, for a sufficiently low oil flow rate when compared to the inner water flow rate, small single-cored droplet formed ( Figure 8B) due to the merging of water cores in the oil before the oil's drop off.
The O/W and W/O/W droplet formation results indicate that the proposed structure is similar to the co-flow model of droplet formation, because we used a sufficiently long protrusion-tip structure to prevent oil wetting and spreading. Therefore, the droplet size was highly dependent on the dimensions of the downstream channel, as well as the flow conditions. It seems that the optimization of the protrusion-tip structure, including channel dimensions in each unit, is necessary to effectively utilize the shear force of the focusing flow. Moreover, the number and size of cores, such as a single core with thin shell, can be controlled by considering the channel dimensions. Furthermore, the number of water and oil layers can be increased by additional stacking of channel layers.  For simple modeling, we proposed the channel structure of similar patterns and stacking that is shown in Figure 3. However, the most critical structure is the channel end that protrudes forward of the focusing channel to eliminate the low flow area where high wetting fluid spreads along the surface. Therefore, we believe that the functional pair structure (refer, U3 and U4 in Figure 3) can be integrated with any PDMS structures that can be manufactured.

Conclusions
We presented a method for forming single (O/W) or multi-layered (W/O/W) microdroplets while using an oil phase fluid in a PDMS structure. A protruded tubular end isolated from the wall surfaces combined with three-dimensional flow focusing allowed the formation of droplets from liquids that easily wet the PDMS surface without requiring any complex surface treatment. Furthermore, the efficient flow in the structure suppressed spreading of the liquid along the channel surface and washed off undesirable wetting of the liquid on the surface. Droplet size was controlled via the flow conditions and the number of droplet layers was increased via an increase in the channel layers. Multiple stacking of channel layers allowed for the formation of functional flows and the fabrication of precise structures.
The proposed structure and method provide advanced droplet formation results compared to conventional methods using PDMS structures. Moreover, the method achieves results that were only possible on glass-based structures, while allowing for higher structural precision and fabrication reliability than the glass-based devices. Furthermore, the full PDMS structure can be stably integrated with other PDMS based functional elements that were already developed in the various fields without any critical limitations. Thus, we believe that the proposed method and structure have great potential for the enabling further droplet based chemical, material, or biological research. For simple modeling, we proposed the channel structure of similar patterns and stacking that is shown in Figure 3. However, the most critical structure is the channel end that protrudes forward of the focusing channel to eliminate the low flow area where high wetting fluid spreads along the surface. Therefore, we believe that the functional pair structure (refer, U3 and U4 in Figure 3) can be integrated with any PDMS structures that can be manufactured.

Conclusions
We presented a method for forming single (O/W) or multi-layered (W/O/W) microdroplets while using an oil phase fluid in a PDMS structure. A protruded tubular end isolated from the wall surfaces combined with three-dimensional flow focusing allowed the formation of droplets from liquids that easily wet the PDMS surface without requiring any complex surface treatment. Furthermore, the efficient flow in the structure suppressed spreading of the liquid along the channel surface and washed off undesirable wetting of the liquid on the surface. Droplet size was controlled via the flow conditions and the number of droplet layers was increased via an increase in the channel layers. Multiple stacking of channel layers allowed for the formation of functional flows and the fabrication of precise structures.
The proposed structure and method provide advanced droplet formation results compared to conventional methods using PDMS structures. Moreover, the method achieves results that were only possible on glass-based structures, while allowing for higher structural precision and fabrication reliability than the glass-based devices. Furthermore, the full PDMS structure can be stably integrated with other PDMS based functional elements that were already developed in the various fields without any critical limitations. Thus, we believe that the proposed method and structure have great potential for the enabling further droplet based chemical, material, or biological research.