**3. Aqueous two-phase droplet-based microextraction**

This chapter is about microfluidic extraction systems based on droplet of aqueous twophase system (ATPS). Fast mass transfer between continuous phase and dispersed droplet is demonstrated by microextraction of ruthenium red within a microfluidic device which can also generate droplets with electrohydrodymanic method. By comparing brightness data of known concentrations of ruthenium red droplets with those of droplet of interest, change of ruthenium red concentration was traced along the microextraction channel, resulting in good agreement with Fick's diffusion model, and more details are found at Choi et al., 2010. It is also suggested the method for cessation of microextraction which is essential for controlling solute concentration inside a droplet by means of the electrohydrodynamic manipulation of droplet movement direction demonstrated in Section 2.1. Droplets of different ruthenium red concentration were moved to branched channels designed for each of the droplet to move to desired place of microfluidic system for further reaction. The microextraction system based on ATPS droplets has extensive potential to be used in effective and convenient mass transfer of solutes which have different solubility with advantage of handling discrete volume of liquid of desired concentration.

#### **3.1 Microextraction kinetics**

This section is devoted to the ATPS droplet-based microfludics system for ruthenium red extraction. The microextraction kinetics in the microfluidic system was analyzed. As shown in Figure 9, we fabricated the microfluidic system that consisted of the droplet-generation part and the extraction part to which a microchannel of 70 mm belonged. The length of the microchannel, 70 mm, was determined by considering the convergence of the ruthenium red concentration in the droplet after the sufficient extraction. The feed flow rates of ammonium sulfate **(**AS)-rich phase and tetrabutylammonium bromide (TBAB)-rich phase into dropletgeneration part are kept at 5.04μl/min and 0.14 μl/min, respectively. In the microextraction part, ruthenium red-containing TBAB-rich phase is introduced to the main microextraction channel, as shown in Figure 4.1a and b schematically, at the flow rate of 0.24μl/min.

Fig. 9. The schematic diagram of the microfluidic system for the analysis of kinetics of microextraction of ruthenium red

This chapter is about microfluidic extraction systems based on droplet of aqueous twophase system (ATPS). Fast mass transfer between continuous phase and dispersed droplet is demonstrated by microextraction of ruthenium red within a microfluidic device which can also generate droplets with electrohydrodymanic method. By comparing brightness data of known concentrations of ruthenium red droplets with those of droplet of interest, change of ruthenium red concentration was traced along the microextraction channel, resulting in good agreement with Fick's diffusion model, and more details are found at Choi et al., 2010. It is also suggested the method for cessation of microextraction which is essential for controlling solute concentration inside a droplet by means of the electrohydrodynamic manipulation of droplet movement direction demonstrated in Section 2.1. Droplets of different ruthenium red concentration were moved to branched channels designed for each of the droplet to move to desired place of microfluidic system for further reaction. The microextraction system based on ATPS droplets has extensive potential to be used in effective and convenient mass transfer of solutes which have different solubility with advantage of handling discrete

This section is devoted to the ATPS droplet-based microfludics system for ruthenium red extraction. The microextraction kinetics in the microfluidic system was analyzed. As shown in Figure 9, we fabricated the microfluidic system that consisted of the droplet-generation part and the extraction part to which a microchannel of 70 mm belonged. The length of the microchannel, 70 mm, was determined by considering the convergence of the ruthenium red concentration in the droplet after the sufficient extraction. The feed flow rates of ammonium sulfate **(**AS)-rich phase and tetrabutylammonium bromide (TBAB)-rich phase into dropletgeneration part are kept at 5.04μl/min and 0.14 μl/min, respectively. In the microextraction part, ruthenium red-containing TBAB-rich phase is introduced to the main microextraction channel, as shown in Figure 4.1a and b schematically, at the flow rate of 0.24μl/min.

Fig. 9. The schematic diagram of the microfluidic system for the analysis of kinetics of

**3. Aqueous two-phase droplet-based microextraction** 

volume of liquid of desired concentration.

**3.1 Microextraction kinetics** 

microextraction of ruthenium red

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

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

$$\frac{\partial \mathbb{C}\_s}{\partial t} = k\_{\text{diss}} \left( \mathbb{C}\_{\text{sat}} - \mathbb{C}\_s \right) \tag{1}$$

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 function of time as follows:

$$C\_s = C\_{\text{sat}} \left( 1 - \exp(-k\_{\text{diss}}t) \right) \tag{2}$$

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

$$\frac{\partial \mathbf{C}\_{\rm in}}{\partial t} = D \frac{1}{r^2} \frac{\partial}{\partial r} (r^2 \frac{\partial \mathbf{C}\_{\rm in}}{\partial r}) \, , \tag{3}$$

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 microextraction device.

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

$$\frac{\partial \mathbf{C}\_{\text{in}}}{\partial r} = 0 \qquad \text{at } r = 0,\tag{4}$$

$$\mathbf{C}\_{\rm in} = \mathbf{C}\_{\rm s} = \mathbf{C}\_{\rm sat} \left( 1 - \exp(-k\_{\rm diss} t) \right) \qquad \text{at } r = \mathbf{R}. \tag{5}$$

The initial condition is

$$C\_{\rm in} = 0 \qquad \text{ at } t = 0 \text{ for all } r. \tag{6}$$

Microdroplets for the Study of Mass Transfer 817

Fig. 11. Ratio of the ruthenium red concentration in the droplet at the 45 seconds of microextraction to that of continuous TBAB-rich phase initially introduced to the

Figure 11 shows the ratio of the concentration of ruthenium red in the droplet after 45 seconds of microextraction to that of the continuous TBAB-rich phase fed initially. In the first three cases of low ruthenium red concentrations (0.05, 0.07 and 0.09% (w/w)) in TBABrich phase, the ratio is almost constant. However, the ratio decreases, as the concentration of ruthenium red of the TBAB-rich phase increases for high range of concentration of ruthenium red. This suggests that there exists the solubility limit of ruthenium red in the AS-rich droplet above which the concentration cannot increase independent of the concentration in the TBAB-rich phase. The solubility limit of ruthenium red in the AS-rich phase droplet, corresponding to *C*out of 0.13 and 0.15% (w/w), is about 1.15% (w/w) in this

In the previous section, the microextraction was performed using the droplet-based microfluidic system. When a macroscopic system is realized in a microfluidics system, various functions that are impossible in a macroscopic system can be realized. This section describes the method to stop the microextraction before reaching the equilibrium by moving the droplet from the ruthenium red-containing TBAB-rich phase stream to the pure TBABrich phase stream using the EHD force. This function enables to obtain a droplet of a specific concentration of a substance without any other additional treatments in a microfluidics system in which the process for chemical reaction or analysis is integrated following microextraction. In the previous Section 3.1, the time-dependent profile of microextraction process was derived from the experimental and simulation results, and thus stopping of microextraction was possible using the carefully designed microextraction device and the droplet-motion guiding techniques. The microextraction device that was designed by considering this function consisted of two branched channels at the position of 10 and 20 mm from the junction of the ruthenium red injection where microextraction begins, as

microextraction device

experiment.

**3.2 Microextraction control** 

The first boundary condition (Eq. 4) assumes symmetry at the center of the droplet. In the second boundary condition (Eq. 5), *C*sat is the concentration on the surface of the droplet at saturation, which corresponds to the product of concentration of ruthenium red in TBABrich phase and distribution coefficient between two phases. Although concentration in the TBAB-rich phase is known and assumed to be constant, the distribution coefficient is not available. After sufficiently long extraction time, the droplet will be saturated with solute and the concentration will correspond to *C*sat. In this work, it is assumed that the extraction time is long enough and the final concentration at the end of extraction is considered to be *C*sat. Solving Eq. 3 using the previously described boundary conditions gives the radial distribution of concentration in the droplet at each time of concentration measurement. The Cavg inside the droplet at each time interval can be calculated. The model validity was verified by comparing the Cavg with the experimental result.

Because the information about *C*sat*, k*diss and *D* is required to solve the equation numerically, here we briefly introduce the method to calculate these parameters. The measured concentration of ruthenium red at 70 mm position which is the end of the microchannel was chosen as *C*sat in each experiment. The *k*diss value was determined as 0.09 (sec-1) by fitting experimental results from the entire microextraction process and applied to solve the equation with five different *C*sat (We have used the same *C*sat for experiments of 0.13 and 0.15% (w/w) assuming that the *C*sat has reached its solubility limit). Diffusion coefficient of ruthenium red in AS-rich phase, D, was measured by using 'T-sensor' developed by Kamholtz et al. (Kamholtz et al., 1999) Solutions of simple diffusion model with experimental results are shown in Figure 10. Here, *C*out is the concentration of ruthenium red in the continuous TBAB-rich phase. Time-dependent simulation results are in good agreement with the experimental results except the lowest concentration results (*C*out=0.05% (w/w)) in their initial time stage where simulation overestimates the concentrations. One explanation for this error may be that in the early stages of microextraction, the ruthenium red is mainly in the region near the surface of the droplet which is excluded in the brightness measurement especially when the concentration of ruthenium red is very low. As the extraction process continues, more ruthenium red is diffused to the region belonging to the brightness measurement, and simulation shows better agreement with experimental result.

Fig. 10. Comparison of the experimental (e0.05, e0.07, e0.09, e0.11, e0.13 and e0.15) and theoretical data (s0.05, s0.07, s0.09, s0.11, s0.13 and s0.15) of the droplet-based microextraction kinetics

The first boundary condition (Eq. 4) assumes symmetry at the center of the droplet. In the second boundary condition (Eq. 5), *C*sat is the concentration on the surface of the droplet at saturation, which corresponds to the product of concentration of ruthenium red in TBABrich phase and distribution coefficient between two phases. Although concentration in the TBAB-rich phase is known and assumed to be constant, the distribution coefficient is not available. After sufficiently long extraction time, the droplet will be saturated with solute and the concentration will correspond to *C*sat. In this work, it is assumed that the extraction time is long enough and the final concentration at the end of extraction is considered to be *C*sat. Solving Eq. 3 using the previously described boundary conditions gives the radial distribution of concentration in the droplet at each time of concentration measurement. The Cavg inside the droplet at each time interval can be calculated. The model validity was

Because the information about *C*sat*, k*diss and *D* is required to solve the equation numerically, here we briefly introduce the method to calculate these parameters. The measured concentration of ruthenium red at 70 mm position which is the end of the microchannel was chosen as *C*sat in each experiment. The *k*diss value was determined as 0.09 (sec-1) by fitting experimental results from the entire microextraction process and applied to solve the equation with five different *C*sat (We have used the same *C*sat for experiments of 0.13 and 0.15% (w/w) assuming that the *C*sat has reached its solubility limit). Diffusion coefficient of ruthenium red in AS-rich phase, D, was measured by using 'T-sensor' developed by Kamholtz et al. (Kamholtz et al., 1999) Solutions of simple diffusion model with experimental results are shown in Figure 10. Here, *C*out is the concentration of ruthenium red in the continuous TBAB-rich phase. Time-dependent simulation results are in good agreement with the experimental results except the lowest concentration results (*C*out=0.05% (w/w)) in their initial time stage where simulation overestimates the concentrations. One explanation for this error may be that in the early stages of microextraction, the ruthenium red is mainly in the region near the surface of the droplet which is excluded in the brightness measurement especially when the concentration of ruthenium red is very low. As the extraction process continues, more ruthenium red is diffused to the region belonging to the brightness

measurement, and simulation shows better agreement with experimental result.

Fig. 10. Comparison of the experimental (e0.05, e0.07, e0.09, e0.11, e0.13 and e0.15) and

theoretical data (s0.05, s0.07, s0.09, s0.11, s0.13 and s0.15) of the droplet-based

microextraction kinetics

verified by comparing the Cavg with the experimental result.

Fig. 11. Ratio of the ruthenium red concentration in the droplet at the 45 seconds of microextraction to that of continuous TBAB-rich phase initially introduced to the microextraction device

Figure 11 shows the ratio of the concentration of ruthenium red in the droplet after 45 seconds of microextraction to that of the continuous TBAB-rich phase fed initially. In the first three cases of low ruthenium red concentrations (0.05, 0.07 and 0.09% (w/w)) in TBABrich phase, the ratio is almost constant. However, the ratio decreases, as the concentration of ruthenium red of the TBAB-rich phase increases for high range of concentration of ruthenium red. This suggests that there exists the solubility limit of ruthenium red in the AS-rich droplet above which the concentration cannot increase independent of the concentration in the TBAB-rich phase. The solubility limit of ruthenium red in the AS-rich phase droplet, corresponding to *C*out of 0.13 and 0.15% (w/w), is about 1.15% (w/w) in this experiment.

#### **3.2 Microextraction control**

In the previous section, the microextraction was performed using the droplet-based microfluidic system. When a macroscopic system is realized in a microfluidics system, various functions that are impossible in a macroscopic system can be realized. This section describes the method to stop the microextraction before reaching the equilibrium by moving the droplet from the ruthenium red-containing TBAB-rich phase stream to the pure TBABrich phase stream using the EHD force. This function enables to obtain a droplet of a specific concentration of a substance without any other additional treatments in a microfluidics system in which the process for chemical reaction or analysis is integrated following microextraction. In the previous Section 3.1, the time-dependent profile of microextraction process was derived from the experimental and simulation results, and thus stopping of microextraction was possible using the carefully designed microextraction device and the droplet-motion guiding techniques. The microextraction device that was designed by considering this function consisted of two branched channels at the position of 10 and 20 mm from the junction of the ruthenium red injection where microextraction begins, as

Microdroplets for the Study of Mass Transfer 819

(a) (b) (c)

move to other part of the microfluidic system

Fig. 13. Microextraction control and manipulation of droplet within the microextraction system shown in Figure 4.1b. (a) Beginning of microextraction: The AS-rich phase droplet is dipped into the ruthenium red-containing TBAB-rich phase at the narrowed part of the microchannel and starts microextraction. (b) The first branched channel at the position of 10 mm from the starting point of microextraction. In this case shown in the picture, the electric pulse is not applied so that the droplet can move continually through the microextraction channel. (c) The second branched channel at the position of 20 mm from the starting point of microextraction. The applied electric pulse changes the movement direction of the droplet toward the branched channel so that the microextraction is stopped and the droplet can

shown in Figure 12. The branched channels were arranged to have an angle of 45° so that droplets can readily move toward the branched channels. AS-rich phase that forms laminar flow with TBAB-rich phase in the droplet-generation part of the device is fed at the flow rate 5.4 μl/min. The droplet produced in the droplet-generation part of the device moves to microextraction part, and then extraction starts at the junction where TBAB-rich phase with ruthenium red concentration of 0.29 (%, w/w) is injected. Only when the droplet crosses the streamlines and is fully surrounded by the ruthenium red containing TBAB-rich phase, isotropy of microextraction is guaranteed. To force droplet to enter the ruthenium red stream, the microchannel that has a narrow cross section was fabricated as shown in Figure 13 a. Because the flow rate of the main TBAB-rich phase, which initially carries the AS-rich droplet, is kept at 0.24 μl/min, smaller than that of ruthenium red solution (0.51 μl/min), the droplet moves toward the ruthenium red solution due to the pressure difference. Microextraction continues until the d. c. electric pulse with the duration of 200 ms is applied at the end of the branched channel. At that moment the droplet departs ruthenium redcontaining TBAB-rich phase by approaching the electrode located at each end of the branched microchannel and microextraction is terminated.

Fig. 12. The schematic diagram of the microfluidic system for microextraction control. Droplets can moved to one of the two branched channels by applied electric pulse after cessation of microextraction

As shown in Figure 13 b and c, the thresholds located at 10 and 20 mm positions serve as leaping boards for the droplet. When d. c. electric potential is not applied, the droplet moves freely, following its streamline over the threshold as shown in Figure 13 b. When electric potential is applied on the electrode at the end of the branched channel, the droplet changes its direction at the junction so that it leaves the flow stream of ruthenium red-containing TBAB-rich phase and jumps into the branched channel with increased speed due to the electrophoretic effect as shown in Figure 13 c. The concentrations of droplets at the first and second branched microchannel when microextraction is terminated are 0.12 and 0.24 (%, w/w) and corresponding concentrations which have been predicted by simulation are 0.11 and 0.28 (%, w/w), which shows that the droplet-based microextraction can be terminated depending on the needs in a process, even before reaching the equilibrium state.

shown in Figure 12. The branched channels were arranged to have an angle of 45° so that droplets can readily move toward the branched channels. AS-rich phase that forms laminar flow with TBAB-rich phase in the droplet-generation part of the device is fed at the flow rate 5.4 μl/min. The droplet produced in the droplet-generation part of the device moves to microextraction part, and then extraction starts at the junction where TBAB-rich phase with ruthenium red concentration of 0.29 (%, w/w) is injected. Only when the droplet crosses the streamlines and is fully surrounded by the ruthenium red containing TBAB-rich phase, isotropy of microextraction is guaranteed. To force droplet to enter the ruthenium red stream, the microchannel that has a narrow cross section was fabricated as shown in Figure 13 a. Because the flow rate of the main TBAB-rich phase, which initially carries the AS-rich droplet, is kept at 0.24 μl/min, smaller than that of ruthenium red solution (0.51 μl/min), the droplet moves toward the ruthenium red solution due to the pressure difference. Microextraction continues until the d. c. electric pulse with the duration of 200 ms is applied at the end of the branched channel. At that moment the droplet departs ruthenium redcontaining TBAB-rich phase by approaching the electrode located at each end of the

Fig. 12. The schematic diagram of the microfluidic system for microextraction control. Droplets can moved to one of the two branched channels by applied electric pulse after

depending on the needs in a process, even before reaching the equilibrium state.

As shown in Figure 13 b and c, the thresholds located at 10 and 20 mm positions serve as leaping boards for the droplet. When d. c. electric potential is not applied, the droplet moves freely, following its streamline over the threshold as shown in Figure 13 b. When electric potential is applied on the electrode at the end of the branched channel, the droplet changes its direction at the junction so that it leaves the flow stream of ruthenium red-containing TBAB-rich phase and jumps into the branched channel with increased speed due to the electrophoretic effect as shown in Figure 13 c. The concentrations of droplets at the first and second branched microchannel when microextraction is terminated are 0.12 and 0.24 (%, w/w) and corresponding concentrations which have been predicted by simulation are 0.11 and 0.28 (%, w/w), which shows that the droplet-based microextraction can be terminated

branched microchannel and microextraction is terminated.

cessation of microextraction

Fig. 13. Microextraction control and manipulation of droplet within the microextraction system shown in Figure 4.1b. (a) Beginning of microextraction: The AS-rich phase droplet is dipped into the ruthenium red-containing TBAB-rich phase at the narrowed part of the microchannel and starts microextraction. (b) The first branched channel at the position of 10 mm from the starting point of microextraction. In this case shown in the picture, the electric pulse is not applied so that the droplet can move continually through the microextraction channel. (c) The second branched channel at the position of 20 mm from the starting point of microextraction. The applied electric pulse changes the movement direction of the droplet toward the branched channel so that the microextraction is stopped and the droplet can move to other part of the microfluidic system

Microdroplets for the Study of Mass Transfer 821

can widen the application in which relatively susceptible biological molecules are

Ahn, K., Kerbage, C., Hunt, T. P., Westervelt, R. M., Link D. R. & Weitz, D. A. (2006).

Albertsson, P.-Å. (1986) *Partition of Cell Particles and Macromolecules, 3rd Ed.*, Wiley, New

Anna, S. L., Bontoux, N. & Stone, H. A. (2003). Formation of dispersions using ''flow focusing'' in microchannels. *Applied Physics Letters*, Vol.82, No.3, pp. 364-366 Beebe, D. J., Mensing, G. A. & Walker, G. M. (2002). Physics and applications of

Bremond, N., Thiam, A. R. & Bibette, J. (2008). Decompressing emulsion droplets favors

Castell, O. K., Allender, C. J. & Barrow, D. A. (2008). Continuous molecular enrichment in

Castell, O. K., Allender, C. J. & Barrow, D. A. (2009). Liquid-liquid phase separation:

Choi, Y. H., Song, Y. S. & Kim, D. H. (2006) Proceedings of the 10th International Conference

Choi, Y. H., Song, Y. S. & Kim, D. H. (2008) Electrohydrodynamic generation and transport

Choi, Y. H., Song, Y. S. & Kim, D. H. (2010). Droplet-based microextraction in the aqueous two-phase system, *Journal of Chromatography A*, Vol.1217, pp. 3723-3728 Christopher, G. F. & Anna, S. L. (2007). Microfluidic methods for generating continuous

Dittrich, P. S., Tachikawa, K. & Manz, A. (2006) Micro total analysis systems.

Dreyfus, R., Tabeling, P. & Willaime, H. (2003) Ordered and disordered patterns in two-

coalescence, *Physical Review Letters*, Vol.100, No. 2, art. no. 024501

H. Fujita, S. Hasebe (Eds.) pp. 95-97, CHEMINAS, Tokyo, Japan

device, *AIP Conference Proceedings*, Vol.1027, pp. 1006-1008

devices. *Applied Physics Letters*, Vol. 88, No.2, art. no. 024104, pp. 1-3 Ahn, K., Agresti, J., Chong, H., Marquez, M. & Weitz, D.A. (2006) Electrocoalescence of

*Physics Letters*, Vol.88 No.26, art. no. 264105

Dielectrophoretic manipulation of drops for high-speed microfluidic sorting

drops synchronized by size-dependent flow in microfluidic channels. *Applied* 

microfluidics in biology, *Annual Review of Biomedical Engineering*, Vol.4, pp.261-

microfluidic systems, *Lab on a Chip - Miniaturisation for Chemistry and Biology*, Vol. 8,

Characterisation of a novel device capable of separating particle carrying multiphase flows, *Lab on a Chip - Miniaturisation for Chemistry and Biology*, Vol.9,

on Miniaturized Systems for Chemistry and Life Sciences (μTAS2006), T. Kitamori,

of a single or multiple droplets of aqueous two-phase system in microfluidic

droplet streams, *Journal of Physics D: Applied Physics*, Vol. 40, No.19, art. no. R01, pp.

Latest advancements and trends, *Analytical Chemistry*, Vol. 78, No.12, pp. 3887-

phase flows in microchannels, *Physical Review Letters*, Vol.90, No.14, pp. 144505/1-

involved.

**5. References** 

York

286

No. 7, pp. 1031-1033

No.3, pp. 388-396

R319-R336

3907

144505/4
