**2. Droplet generation and manipulation**

In this chapter, electrohydrodynamic droplet generation in aqueous two-phase microflow and manipulation of ATPS droplets are discussed. ATPS droplets usually have electrophoretic mobility due to the presence of anions. If the interfacial tension is sufficiently high, the interface can respond to the applied d. c. electric field based on the same reason. Though much lower than that of organic/aqueous two-phase system, interfacial tension in tetrabutylammonium bromide (TBAB)/ammonium sulfate(AS) ATPS is relatively high

that droplets not only help mass transfer through the droplet surface but also mix the fluids be means of the circulation flow inside them and serve as carriers of fluids by themselves. These properties show that droplets can be used for microextraction as well as microreaction.

In this study, we employed aqueous two-phase system (ATPS) to realize the microextraction in the microfluidic system. The liquid-liquid system, called ATPS or aqueous biphasic system was first studied by a Swedish biochemist P. Å. Albertsson (Albertsson, 1986). ATPS has become a powerful tool for separation of a range of biomaterials, including plant and animal cells, microorganisms, fungi, virus, chloroplasts, mitochondria, membrane vesicles, proteins, and nucleic acids. It can also be an appealing system for microfluidic droplet application, since the two phases are all aqueous (Walter et al., 1985; Hatti-Kaul, 2000). An ATPS consists of two immiscible phases formed by dissolving two incompatible polymers, such as poly(ethylene glycol) and dextran, or one polymer and an appropriate inorganic salt, or a cationic surfactant and a salt. The phase separation, which finally leads to equilibrium, is thought to be due to the incompatible physicochemical properties of components. Thus one phase is predominantly rich with one component and the second phase is enriched by the other component. ATPS is highly advantageous because the high water content (usually 70- 90%) provides biocompatibility and selectivity for the stable pre-concentration of hydrophilic molecules with relatively low interfacial tension compared to that of organic/water twophase system. It can be a versatile partitioning system for the separation of many kinds of dyes, metal ions, silica particles, proteins and cells (Walter & Johansson, 1986; Walter et al., 1991). Various factors such as concentration and type of phase-forming polymer or salts and the choice and addition of affinity ligands can affect the distribution of a solute over the two

Several applications of ATPS into microfluidic systems have been reported for continuous partitioning of cells (Yamada et al., 2004; Nam et al., 2005), taking the advantage of PDMS (polydimethylsiloxane)-compatibility avoiding swelling problem that is commonly caused when organic/water two-phase system is applied to PDMS microfluidic device. The ATPSbased partitioning systems developed in these research works have almost same protocol because the direction and configuration of mass transfer where the material being extracted is transported from one phase to the other are almost identical with those of the two-phase system based on oil and water. Although laminar nature of liquid-liquid flow in microfluidic channel makes continuous separation possible, control of the interface has been known as a difficult task because the immiscible nature of the two liquid phases causes competition between interfacial tension and the viscous force. As Dreyfus et al. reported (Dreyfus et al., 2003), only in a certain regime of flow rates of liquid phases, stratified structure of flows, that is necessary for continuous microextraction, can be maintained.

In this chapter, electrohydrodynamic droplet generation in aqueous two-phase microflow and manipulation of ATPS droplets are discussed. ATPS droplets usually have electrophoretic mobility due to the presence of anions. If the interfacial tension is sufficiently high, the interface can respond to the applied d. c. electric field based on the same reason. Though much lower than that of organic/aqueous two-phase system, interfacial tension in tetrabutylammonium bromide (TBAB)/ammonium sulfate(AS) ATPS is relatively high

phases, thus giving more flexibility for the customized systems.

**2. Droplet generation and manipulation** 

**1.3 Aqueous two-phase system** 

(about 4-5 dyne/cm) compared to that of common ATPS formed with poly(ethylene glycol) and dextran (10-4 to 0.1 dyne/cm). Thus the interface readily responds to the external electric field in microchannels. The fabricated microfluidic device has a T-junction at which the two-phase flow may have the configuration that can generate dispersed droplets by the electric potential difference applied. The threshold voltage necessary for the electrohydrodynamic droplet generation depends on pH due to the degree of dissociation and charge accumulation. Electrokinetic control of droplet break-up and switching of droplet movement direction were also demonstrated based on the same electrophoretic mobility of ATPS droplets. Volume of broken droplets and the direction of droplet movement were effectively controlled by the applied DC electric field. In addition, simple manipulatin of ATPS droplets was demonstrated in the microchannels that are branched at the end.

#### **2.1 Electrohydrodynamic generation of droplets**

The ATPS for the droplet generation was prepared by dissolving TBAB and AS in water by 15 and 30%, respectively, by stirring the solution well. The prepared solution is left still for more than 12 hours so that it can be divided into two phases by the difference in the specific gravity. After separating the stable phases, TBAB-rich phase and AS-rich phase were individually introduced at the inlets of the microfluidic system by syringe pump, which controlled the flow rates of each phase independently. The microfluidic system in Figure 2 was fabricated by using the general PDMS replica. The ratios of flow rate of TBAB-rich phase to AS-rich phase were fixed at 0.133 and 0.156 in which the two streams were laminar as shown in figure. In these ratios, the more viscous TBAB-rich phase occupied about half of the channel width, but from T-junction one branched channel was totally occupied by TBAB-rich phase and the other branched channel by TBAB-rich phase and AS-rich phase together. Application of an electric field in this state through the electrodes shown in Figure 2 causes the change in the interface as in Figure 3, as the AS-rich phase is drawn in the direction toward the positive electrode. The interface at the center of T-junction was deformed to the positive electrode, which is located at the outlet where only TBAB-rich phase was flowing out. At the same time, the interface in the part of the channel connected to the negative electrode was also deformed due to the temporary change in volumetric flow

Fig. 2. Schematic illustration of the device for electrohydrodynamic generation of droplets and the dimensions of the channels

Microdroplets for the Study of Mass Transfer 811

The higher potential difference of electric pulse results in generation of multiple droplets with different volume in each droplet, usually the first droplet being the largest in diameter as shown in Figure 5. During switching-on of the electric pulse, the droplets have bellshapes and are attracted to the positive electrode in a way similar to electrophoretic motion, while the spherical droplets flow freely with continuous phase after switching off. The number of droplets generated is controlled by the change in the magnitude and duration of a d.c. electric pulse as shown in Figure 6. Longer pulse at higher voltage usually generated

Capillary number of this particular droplet-generation can be calculated. The capillary number for the ATPS microfluidic system is Ca=ηTBAB·UAS/σ where UAS is the speed of the AS-rich phase ηTBAB is the continuous TBAB-rich phase viscosity, σ is two-phase interfacial tension. For example, when the ratio of TBAB-rich phase flow rate to that of AS-rich phase is 0.133, their actual flow rates were 0.133μl/min and 1μl/min, respectively. With 40cP (4×10-3 kg/m·sec) of ηTBAB, 1.7×10-3 m/s of UAS and the 5dyne/cm (5×10-3 N/m), the calculated Ca is 1.36×10-3. This number falls between the range of Ca studied by Jullien et al., (4×10-4 and 2×10-1) (Jullien et al., 2009) where droplet breakup may occur in microfluidic T-junctions. This method of droplet generation has the advantage that the time and frequency of droplet generation can be controlled using the electric signals without a moving part, different from the piezoelectric method (Ziemecak et al., 2011). Moreover, the droplet generation using ATPS can be achieved at a very low electric potential when compared with that of the other

Fig. 5. Multiple-droplet generation by electric pulse application. Two (first row) and three (second row)-droplet generation by electric field application. Two drops are generated by a

20 V, 500 ms pulse (above). Three drops are generated by a 20 V, 800 ms pulse

electrichydrodynamic methods using different two-phase systems.

more droplets.

rate of AS-rich phase. When the electric pulse signal was switched off, the interface returned to its original shape. When the applied voltage was increased, the deformation of AS-rich phase became larger and small volume of AS-rich phase was detached from main AS-rich phase stream forming a droplet dispersed in TBAB-rich phase Figure 4.

Fig. 3. Change of interface configuration by d.c. 12 V electric pulse application. The electric pulse was applied from 0 second to 4/15 second (200 ms) as shown in the first four pictures. The positive electrode is located at the end of the channel on the upward side

Fig. 4. Generation of a single droplet 20 V electric pulse application. The electric pulse was applied from 0 second to 4/15 second (200 ms) as shown in the first four pictures. The positive electrode is located at the end of the channel on the upward side

rate of AS-rich phase. When the electric pulse signal was switched off, the interface returned to its original shape. When the applied voltage was increased, the deformation of AS-rich phase became larger and small volume of AS-rich phase was detached from main AS-rich

Fig. 3. Change of interface configuration by d.c. 12 V electric pulse application. The electric pulse was applied from 0 second to 4/15 second (200 ms) as shown in the first four pictures.

Fig. 4. Generation of a single droplet 20 V electric pulse application. The electric pulse was applied from 0 second to 4/15 second (200 ms) as shown in the first four pictures. The

positive electrode is located at the end of the channel on the upward side

The positive electrode is located at the end of the channel on the upward side

phase stream forming a droplet dispersed in TBAB-rich phase Figure 4.

The higher potential difference of electric pulse results in generation of multiple droplets with different volume in each droplet, usually the first droplet being the largest in diameter as shown in Figure 5. During switching-on of the electric pulse, the droplets have bellshapes and are attracted to the positive electrode in a way similar to electrophoretic motion, while the spherical droplets flow freely with continuous phase after switching off. The number of droplets generated is controlled by the change in the magnitude and duration of a d.c. electric pulse as shown in Figure 6. Longer pulse at higher voltage usually generated more droplets.

Capillary number of this particular droplet-generation can be calculated. The capillary number for the ATPS microfluidic system is Ca=ηTBAB·UAS/σ where UAS is the speed of the AS-rich phase ηTBAB is the continuous TBAB-rich phase viscosity, σ is two-phase interfacial tension. For example, when the ratio of TBAB-rich phase flow rate to that of AS-rich phase is 0.133, their actual flow rates were 0.133μl/min and 1μl/min, respectively. With 40cP (4×10-3 kg/m·sec) of ηTBAB, 1.7×10-3 m/s of UAS and the 5dyne/cm (5×10-3 N/m), the calculated Ca is 1.36×10-3. This number falls between the range of Ca studied by Jullien et al., (4×10-4 and 2×10-1) (Jullien et al., 2009) where droplet breakup may occur in microfluidic T-junctions. This method of droplet generation has the advantage that the time and frequency of droplet generation can be controlled using the electric signals without a moving part, different from the piezoelectric method (Ziemecak et al., 2011). Moreover, the droplet generation using ATPS can be achieved at a very low electric potential when compared with that of the other electrichydrodynamic methods using different two-phase systems.

Fig. 5. Multiple-droplet generation by electric pulse application. Two (first row) and three (second row)-droplet generation by electric field application. Two drops are generated by a 20 V, 500 ms pulse (above). Three drops are generated by a 20 V, 800 ms pulse

Microdroplets for the Study of Mass Transfer 813

Fig. 7. Y-branched microfluidic device design for the droplet transport experiment

Fig. 8. Transport of a droplet into the desired channel by the application of electric pulses

Fig. 6. Number of generated droplets controlled by the mode of electric pulse application. The number of droplets generated is controlled by the change in the magnitude and duration of a d.c. electric pulse. Longer pulse at higher voltage usually generated more droplets

#### **2.2 Electrohydrodynamic manipulation of droplets**

Manipulation of droplet movement is an essential technique for the availability of a microfluidic system to transport a certain droplet to the desired part of the system. Since passive manners for droplet transportation have limitation of complex channel design and flow rate control, active manners are usually preferred. For instance, Ahn et al. developed dielectrophoretic manipulation of droplets for high-speed microfluidic sorting devices (Ahn et al., 2006). Active method of controlling charged droplets electrically was reported by Link et al (Link et al. 2006). In line with these previous studies, the method to manipulate the movement of droplets of TBAB/AS ATPS in the microfluidic channels was developed, separating and directing the droplets of specific number into desired part of the device from the sequential flow of droplets, by applying the programmed d. c. electric field which was synchronized with the droplet frequency.

In this study, droplets were produced passively by shear force at the Y-junction where the two streams first contact with each other in the microfluidic device of the design shown in Figure 7. For the period without electric field, droplets flow toward outlet 1 because the branched channel that leads to outlet 1 is shorter than that toward outlet 2 as shown in Figure 7, thus it has higher pressure gradient. For the period with electric field on, droplets are attracted toward outlet 2 due to the electric attraction. In this way, a specific droplet can be selectively separated from the line of droplets that flows through the channel. Figure 8 shows the result of the experiment where the generated droplets were sent to the outlet 1 and outlet 2 alternatively by applying electric signals in a certain interval. As shown in these results, the electrohydrodynamic transport is a convenient method for handling droplets inside microfluidic devices without any moving part. The control of direction is possible only by changing the pulse signals with a simple switching on and off program. This simplicity will allow the realization of complicated and multi-functional microfluidic system based on aqueous two-phase droplets.

Fig. 6. Number of generated droplets controlled by the mode of electric pulse application. The number of droplets generated is controlled by the change in the magnitude and duration of a d.c. electric pulse. Longer pulse at higher voltage usually generated more droplets

Manipulation of droplet movement is an essential technique for the availability of a microfluidic system to transport a certain droplet to the desired part of the system. Since passive manners for droplet transportation have limitation of complex channel design and flow rate control, active manners are usually preferred. For instance, Ahn et al. developed dielectrophoretic manipulation of droplets for high-speed microfluidic sorting devices (Ahn et al., 2006). Active method of controlling charged droplets electrically was reported by Link et al (Link et al. 2006). In line with these previous studies, the method to manipulate the movement of droplets of TBAB/AS ATPS in the microfluidic channels was developed, separating and directing the droplets of specific number into desired part of the device from the sequential flow of droplets, by applying the programmed d. c. electric field which was

In this study, droplets were produced passively by shear force at the Y-junction where the two streams first contact with each other in the microfluidic device of the design shown in Figure 7. For the period without electric field, droplets flow toward outlet 1 because the branched channel that leads to outlet 1 is shorter than that toward outlet 2 as shown in Figure 7, thus it has higher pressure gradient. For the period with electric field on, droplets are attracted toward outlet 2 due to the electric attraction. In this way, a specific droplet can be selectively separated from the line of droplets that flows through the channel. Figure 8 shows the result of the experiment where the generated droplets were sent to the outlet 1 and outlet 2 alternatively by applying electric signals in a certain interval. As shown in these results, the electrohydrodynamic transport is a convenient method for handling droplets inside microfluidic devices without any moving part. The control of direction is possible only by changing the pulse signals with a simple switching on and off program. This simplicity will allow the realization of complicated and multi-functional microfluidic system

**2.2 Electrohydrodynamic manipulation of droplets** 

synchronized with the droplet frequency.

based on aqueous two-phase droplets.

Fig. 7. Y-branched microfluidic device design for the droplet transport experiment

Fig. 8. Transport of a droplet into the desired channel by the application of electric pulses

Microdroplets for the Study of Mass Transfer 815

The flow rate combination is selected carefully to have appropriate pressure drop and width of each stream, especially at the T-junction in droplet-generation part and at the junction where ruthenium red stream is introduced in microextraction part. Six kinds of ruthenium red-containing TBAB-rich phases were prepared by dissolving ruthenium red in the TBABrich phase with the concentration of 0.05, 0.07, 0.09, 0.11, 0.13 and 0.15% (w/w). Ruthenium red of which molecular weight is 786.36, has the structural formula of [(NH3)5Ru-O-Ru(NH3)4-O-Ru(NH3)5]Cl6. It is dispersed in the TBAB-rich phase as solid particles, but is

In this study, the extraction of ruthenium red from the TBAB-rich phase to the AS-rich phase droplet was assumed as the combination of two different processes for the analysis. The first process was the dissolution of the ruthenium red in the TBAB-rich phase at the As-rich phase surface up to the saturated concentration. This process was required because ruthenium red existed in the TBAB-rich phase as solid particles dispersed due to its low solubility in that phase. The general equation for the process involving dissolution of

> ( ) <sup>s</sup> diss sat s *<sup>C</sup> kC C*

where *C*s denotes the concentration at the surface of the droplet, *C*sat, the concentration when saturated, and *k*diss, the dissolution rate coefficient. Integrating above equation gives *C*s as a

The next process required was the spread of the dissolved ruthenium red over the AS-rich droplet in the direction of the droplet center. This second process can be described with the

> in 2 in 2 <sup>1</sup> ( ) *C C D r t rr r*

where *C*in and *D* denote the concentration of solute inside the droplet and the effective diffusion coefficient, respectively. This model assumes no internal motion of the fluid in the droplet. The concentration in the continuous phase is assumed to be constant because TBAB-rich phase with fixed concentration of ruthenium red is continuously fed to the

in 0 at 0, *<sup>C</sup>*

*r*

<sup>∂</sup> = − <sup>∂</sup> (1)

*C C kt* s sat = −− (1 exp( ) diss ) (2)

∂ ∂∂ <sup>=</sup> ∂ ∂∂ , (3)

<sup>∂</sup> <sup>=</sup> <sup>=</sup> <sup>∂</sup> (4)

*C C C k t rR* in s sat = = −− = (1 exp( ) at . diss ) (5)

in *Ctr* = 0 at 0 for all . = (6)

dispersed solid particles can be written as (Mary et al., 2008; Xu et al., 2008)

*t*

one-dimensional model of simple diffusion with spherical coordinate system:

The boundary conditions at *r*=0 and *r*=*R* for this equation are specified as

*r*

highly soluble in AS-rich phase as indentified by its color.

function of time as follows:

microextraction device.

The initial condition is
