Acoustofluidic generation of droplets with tunable chemical concentrations

Jinsoo Park ab, Ghulam Destgeer a, Muhammad Afzal a and Hyung Jin Sung *a
aDepartment of Mechanical Engineering, KAIST, Daejeon 34141, Korea. E-mail: hjsung@kaist.ac.kr; Fax: +82 42 350 5027; Tel: +82 42 350 3027
bSchool of Mechanical Engineering, Chonnam National University, Gwangju 61186, Korea

Received 10th August 2020 , Accepted 29th September 2020

First published on 30th September 2020


The dynamic control of the chemical concentration within droplets is required in numerous droplet microfluidic applications. Here, we propose an acoustofluidic method for the generation of a library of aqueous droplets with the desired chemical concentrations in a continuous oil phase. Surface acoustic waves produced by a focused interdigital transducer interact with two parallel laminar streams with different chemical compositions. Coupling the acoustic waves with the flow streams results in the controlled acoustofluidic mixing of the aqueous solutions through the formation of acoustic streaming flow-induced microvortices. The mixed streams are split at a bifurcation, and one of the streams with a precisely controlled chemical concentration is fed into a T-junction to produce droplets with tunable chemical concentrations. The periodic acoustofluidic mixing of the aqueous streams enables the generation of a droplet library with a well-defined inter-droplet concentration gradient. The proposed method is a promising tool for the on-chip dynamic control of in-droplet chemical concentrations and for next-generation droplet microfluidic applications.


Introduction

The effective control of the concentrations of chemical compounds in microscale solutions is essential to many biochemical applications.1–3 Cells and microorganisms interact with microenvironments of varying chemical concentration (CC) that are closely associated with their biological responses such as chemotaxis, inflammation, proliferation, angiogenesis, homeostasis, and development.1–6 The sizes of most cells and microorganisms are in the range 1–100 μm, and intercellular interactions take place over similar length scales.7 In this regard, microfluidic platforms with a characteristic channel dimension of O(100 μm) have been developed to generate CC gradients on length scales comparable to those of cells and microorganisms. The most widely used microfluidic approaches for spatial CC control are based on passive diffusive mixing in various microchannel geometries including tree-shaped,8 Y-shaped,9,10 cross-shaped,11 and counter-flow.12 Active acoustofluidic methods have been also developed for the on-demand rapid and dynamic control of CCs.13,14 These passive and active microfluidic methods for on-chip CC control have been utilized in research into cancer metastasis,15 chemotactic cell trapping,4 the antibiotic resistance of bacteria,11 drug screening,16etc.

Recently, the usefulness of a library of droplets with various CCs has been recognized, especially for droplet-microfluidic single-cell analyses and high-throughput screening.17–20 Droplet microfluidics enables the encapsulation of samples (cells, reagents, etc.) in pico/nano-liter droplets with characteristic lengths of O(10–100 μm). These droplets serve as independent isolated chambers that offer cellular signal detection and increased throughput as well as the reduced consumption of scarce and costly reagents.17–21 More importantly, the isolated microenvironments of droplets enable the independent control of the CC in each droplet as well as long-term stability and increased controllability without Taylor dispersion and cross-contamination.22 However, it is challenging to fine-tune the CCs in a controlled dynamic manner in order to produce batches of droplets with the required CC gradients.

Current methods for the control of in-droplet CCs can be categorized into passive and active methods depending on whether an external force field is present. Passive methods enable the generation of droplets with the desired CC by using tree-shaped networks with multiple branches,23 the flow rate control of multiple miscible aqueous phases,24,25 the merging of multiple droplets with different CCs,26 or droplet dilution with microstructures.27,28 Although these methods are simple and easy to implement, they require flow stabilization over a timescale of up to minutes and comparatively large microchannel footprints, and do not provide the dynamic and precise control of in-droplet CCs. Active in-droplet CC control methods based on a pneumatic valve,29,30 a peristaltic pump,31 or a compartment-on-demand platform32 have been proposed to address the limitations of the passive methods. However, they require large off-chip systems to control the on-chip moving mechanical components. Accordingly, their system configurations and operation are complex, which makes their integration with other microfluidic modules difficult. In this study, we propose an acoustofluidic method without any moving components for the on-chip dynamic control of in-droplet CCs. We utilize a surface acoustic wave (SAW)-based acoustic streaming flow (ASF) to mix fluid streams of different chemical compositions and thus obtain the desired CCs in the dispersed phase before the formation of aqueous droplets at a downstream T-junction. The electronic control of the acoustofluidic device enables on-demand, dynamic, and controlled variation in the mixing index and thus the generation of libraries of droplets with tunable CC gradients.

Methodology

Device configuration and operation

Fig. 1(a) shows the proposed acoustofluidic device that consists of a focused interdigital transducer (FIDT) patterned on a piezoelectric LiNbO3 substrate and a polydimethylsiloxane (PDMS) microfluidic chip. When an AC signal with the resonant frequency of the FIDT (129.5 MHz) is applied to it, SAWs are produced from the transducer by the inverse piezoelectric effect. A sample fluid (5.37 mM erioglaucine disodium salt solution in DI water) and a sheath fluid (DI water) were injected through the separate inlets of the microchannel by using a multi-unit syringe pump. In our device, the two fluids meet to form parallel laminar streams in the main microchannel with a width of 550 μm and a height of 180 μm (Fig. 1(a1)). The main microchannel bifurcates downstream into a waste outlet and a T-junction (Fig. 1(a2)). At the other entrance of the T-junction, a fluorocarbon oil containing a 2 wt% surfactant was injected for droplet generation. The microchannel widths of the dispersed and continuous phases were 250 and 300 μm, respectively. The volumetric flow rates for the sample fluid, sheath fluid, and continuous oil phase fluid were 250, 750, and 1500 μL h−1, respectively. For stable droplet generation at the T-junction, it is important to control the flow rates of the dispersed and continuous phases; however, the direct control of the dispersed phase flow rate is difficult in the proposed device. Instead, the fluid that was not used for droplet generation was pumped out at 700 μL h−1 at the waste outlet after the bifurcation in order to control the dispersed phase flow rate at the T-junction as 300 μL h−1 in an indirect manner. The resulting droplets with varying in-droplet CCs were analyzed by using ImageJ in a droplet observation region (Fig. 1(a3)) with a width of 500 μm.
image file: d0lc00803f-f1.tif
Fig. 1 (a) Schematic diagram of the acoustofluidic device for the production of droplets with fine-tuned chemical concentrations. (b) Numerical simulation of the surface acoustic wave amplitude field for the transducer based on the exact angular spectrum of plane wave theory.

FIDT analysis

The FIDT can produce high-intensity focused SAWs with a high beam-width compression ratio.33 This narrow-beam SAW is suitable to induce a microscale ASF for acoustofluidic mixing of two fluid streams. The focusing characteristics of the FIDT depends on the electrode geometries, and several studies have reported that there is a discrepancy between the acoustic focal point and the geometric focal point.33,34 For optimal operation of the FIDT, we performed numerical simulation of the relative SAW amplitude field by using the exact angular spectrum of plane wave theory and the equivalent aperture model.34 This model approximates the concentric circular FIDT as a straight interdigital transducer with an equiphase distribution.35 The relative acoustic wave amplitude I(x, y) can be estimated as:
 
image file: d0lc00803f-t1.tif(1)
where the parameters are given by:
 
image file: d0lc00803f-t2.tif(2)
 
image file: d0lc00803f-t3.tif(3)
 
image file: d0lc00803f-t4.tif(4)
 
image file: d0lc00803f-t5.tif(5)
The radius of curvature of the FIDT is Lf, the geometric focal length is Wc, and the wave vectors in the x- and y-directions are denoted as k0 and k1, respectively. Fig. 1(b) shows the numerical simulation results of the relative amplitude field formed by 129.5 MHz SAWs on the piezoelectric substrate. This result suggests that the maximum SAW amplitude is formed at x/Lf = 1.5–2.0, not at the geometric focal point at x/Lf = 1. This is consistent with the previous simulation results and experimental observations.13,33,36 Accordingly, the center of the microchannel (Fig. 1(a1)) was aligned at x/Lf = 1.8 for efficient acoustofluidic mixing in the experiments.

Results and discussion

Generation of droplets with controlled chemical concentrations

Fig. 2 shows microscopy images of (a) the region near the acoustic field, (b) the T-junction, and (c) the droplet observation region, which correspond to (a1), (a2), and (a3) in Fig. 1, respectively, for various applied peak-to-peak voltages of 0, 2.31, 2.58, and 2.84 VPP. With no acoustic field applied, the two fluids flow in parallel and retain their respective streamlines in the microchannel. When the parallel laminar streams of the two fluids are exposed to the focused SAW incident perpendicular to the flow direction, the ASF-induced microvortices agitate the streamlines through acoustofluidic mixing.13 As shown in Fig. 2(a), electronic control of the applied AC signal allows the controlled acoustofluidic mixing of the aqueous solutions through the formation of ASF-induced microvortices. As shown in Fig. 2(b), the mixed fluid sample with fine-tuned solute concentrations enters the T-junction as a dispersed phase after the bifurcation, where an immiscible fluorocarbon oil is introduced at the other junction for droplet generation. Fig. 2(c) shows that the solute concentrations within the droplets produced at the T-junction can be varied by adjusting the voltage applied to the FIDT from 0 to 2.84 VPP.
image file: d0lc00803f-f2.tif
Fig. 2 Microscopy images of (a) the region near the focused surface acoustic waves, (b) the T-junction and (c) the droplet observation region at 0, 2.31, 2.58, and 2.85 VPP.

As shown in Fig. 3, we conducted numerical simulations of the ASFs in the proximity of the focused SAWs to validate the experimental observations by using COMSOL Multiphysics. Rayleigh-type SAWs propagating in the x-direction on the LiNbO3 substrate are transformed into longitudinal waves (LWs) when refracting into the fluid in the microchannel at the Rayleigh angle (θR) of approximately 22° in the xz-plane.37 These focused LWs induce attenuation of the oscillating displacements in the fluid, which results in a time-averaged body force that propels the fluid in the wave propagation direction. The non-linear time-averaged body force due to the ASF can be calculated by analyzing the SAW attenuation at the substrate/fluid interface and within the fluids. The attenuation coefficients along the interface between the piezoelectric substrate and the fluid (α) and in the fluid (β) can be expressed as:38

 
α = ρfcf/ρscsλs,(6)
 
β = (4μ/3 + μ′)ω2/ρfcf3,(7)
where ρ is the density, c is the speed of sound, λ is the SAW wavelength, μ and μ′ are the dynamic and bulk viscosities of the fluid, and ω is the angular frequency with the subscripts f and s for the fluid and substrate, respectively. Considering both wave attenuation along the substrate and in the fluid with negligible contribution of the shear waves in the fluid, the acoustic streaming body force (in N m−3) produced in the fluid is expressed with respect to the x- and z-coordinates as:39
 
FB = ρβu(x,z)2,(8)
where u(x, z) = ωξ(x, z) is the displacement velocity. The wave displacement magnitude, which is the spatial extent of the oscillatory motion, in the xz-plane ξ(x, z) is given as:39
 
ξ(x,z) = ωξ0{f0(x,z) + f1(x,z)R},(9)
 
f0(x,z) = eα(xz[thin space (1/6-em)]tan[thin space (1/6-em)]θR)·eβz[thin space (1/6-em)]sec[thin space (1/6-em)]θR,(10)
 
f1(x,z) = eα(x−(hz)[thin space (1/6-em)]tan[thin space (1/6-em)]θR)·eβ(hz)[thin space (1/6-em)]sec[thin space (1/6-em)]θR,(11)
 
R = ((Z2Z1)/(Z1 + Z2))2.(12)
where h is the microchannel height and Z = ρc is the acoustic impedance with the subscripts 1 and 2 for the fluid/microchannel ceiling and fluid/substrate interfaces, respectively. The ξ in eqn (9) is the first-order approximation to consider the wave reflection at the microchannel roof. For the simulation, a three-dimensional fluid domain with 550 μm (width) × 180 μm (height) × 1400 μm (length) was built using an unstructured mesh after the grid-independence test. The Dirichlet boundary conditions were applied for the inlet volumetric flow rate as 1000 μL h−1 and the outlet gauge pressure as zero, respectively. The initial wave displacement magnitude (ξ0) is defined as the wave displacement magnitude at the beginning of the interaction of the acoustic waves with the working fluid without any acoustic attenuation effects in the fluid/substrate interface and in the fluid. In Fig. 3, the initial displacement corresponds to the leftmost point in the middle of the microchannel. For numerical simulation of the ASFs, the ξ0 should be known a priori, and this value can be experimentally determined by using a laser Doppler vibrometer.40 In our work, we conducted the numerical simulation in the absence of such equipment and used the values to best-match the experimental observations in Fig. 2(a).


image file: d0lc00803f-f3.tif
Fig. 3 Numerical simulations of (a) the three dimensional, time-averaged streamlines and (b) the two-dimensional solute trajectories in the middle cross-section (xy-plane) of the microchannel at various initial displacements ξ0 = 0.5, 0.8, and 1.0 nm.

Fig. 3(a) and (b) show the three-dimensional, time-averaged streamlines and the two-dimensional solute trajectories in the middle of the cross-section (xy-plane) of the microchannel at three initial displacements (ξ0 = 0.5, 0.8, and 1.0 nm). These results imply that the acoustic streaming flows can be approximated as two-dimensional flows due to the Hele-Shaw microchannel confinement. The initial displacement is proportional to the amplitude of the AC signal applied to the FIDT, and the vibration displacement on the substrate is linearly related to the pressure amplitude (p) generated in the fluid region, i.e. VPPξ0p.41,42 Consequently, the acoustic streaming velocity near the focused SAW beam increases as the magnitude of the acoustic field increases (Fig. 3(a)). Fig. 3(b) indicates that the solute initially concentrated in the sample fluid flow on the left spreads across the microchannel as a result of acoustofluidic mixing and follows the disturbed streamlines due to the ASF-induced microscale vortices. The numerical results implies that controlled acoustofluidic mixing by adjusting the applied voltage enables the variation in the CC of the flow on the right-hand side of the main microchannel before the bifurcation that enters the downstream T-junction as the dispersed phase for droplet generation. Despite the similar trend between the experiments Fig. 2(a) and simulations in Fig. 3, the slight discrepancy can be mainly attributed to the approximations made in the numerical simulation such as initial displacement magnitude, collimated acoustic waves, boundary conditions, and the first-order wave reflection.

Acoustofluidic control of in-droplet chemical concentration

Fig. 4(a) shows the variation in the in-droplet CC of the solute encapsulated within the droplet as a function of the peak-to-peak voltage of the applied AC signal. The CC in each droplet was quantified from the gray levels of the droplet images by using the calibration plot in Fig. 4(b). The in-droplet CC was linearly increased in the range 0.24–1.34 mM with increases in the applied voltage in the range 2.14–2.85 VPP. The circular symbols are colored according to their color in the experimental images obtained with the inverted microscope equipped with a monochromatic camera. The R-square value of the linear regression is 0.9861, which indicates that the regression fitting line (CC = 1.5135 VPP–2.9266) is in good agreement with the experimental results. Thus we can fine-tune the concentration of the solute encapsulated in each droplet by simply varying the voltage applied to the transducer. Note that the droplet volume and speed were maintained at 8.45 ± 0.23 nL and 1.41 ± 0.09 mm s−1 while the concentration of the model solute suspended in the droplets was varied from 0.24 to 1.34 mM. Fig. 4(a1–a4) show the droplets with controlled in-droplet CCs produced at an applied voltage of 0, 2.31, 2.58, and 2.85 Vpp, respectively. Since each droplet is produced at the T-junction based on the dispersed phase flow, the controlled acoustofluidic mixing to vary the solute concentration leads to generation of droplets with tunable in-droplet CC.
image file: d0lc00803f-f4.tif
Fig. 4 (a) In-droplet solute concentration within the droplets as a function of the applied voltage from 0 to 2.85 VPP. (a1–a4) Droplets with controlled solute concentration produced at an applied voltage of 0, 2.31, 2.58, 2.85 VPP, respectively. (b) Calibration plot showing chemical concentration as a function of the normalized intensity of the grayscale images.

Fig. 5 shows the variation in the in-droplet CC as a function of the duty cycle, which is defined as the ratio of the pulse width to the signal period (not the oscillation period, 1/f). The pulse width during which focused SAWs were applied to the parallel laminar streams was varied from 0.1 to 0.9 s with the period fixed at 3 s, and the corresponding duty cycle was varied from 3.33 to 30%. The CC linearly increases with increasing duty cycle until it plateaus at approximately 0.43, 0.78, and 1.18 mM for 2.23, 2.47, and 2.73 VPP, respectively, which is in good agreement with the linear fitting in Fig. 4(a). Once the acoustic field is applied, a time-averaged body force FB is generated by the focused SAWs, resulting in the acceleration of the fluid. The linear increase in the concentration in the early stages is attributed to the time evolution of the SAW-based ASF.43 These results suggest that the SAW excitation with a sufficient pulse width is required to establish the steady-state ASF-induced microvortices. The stabilization of the in-droplet CC after the monotonic increase arises because the acoustic streaming velocity is suppressed by the viscous resistance.44 As the wave amplitude increases with increasing Vpp, the time required for the acoustic streaming velocity to reach the plateau decreases.45 As a result, the critical duty cycle value for the concentration stabilization decreases as the voltage applied to the FIDT increases, i.e. approximately 16.67, 13.33, and 10% for 2.23, 2.47, and 2.73 VPP, respectively. These results also imply that the in-droplet CC can be fine-tuned by varying the duty cycle and the voltage applied to the acoustofluidic device.


image file: d0lc00803f-f5.tif
Fig. 5 Variations in the solute concentration with the duty cycle (3.33 to 30%) and the amplitude of the applied voltage (2.23, 2.47, and 2.76 VPP).

Droplet libraries with tunable chemical concentrations

The proposed acoustofluidic device enables the generation of a series of droplets with tunable inter-droplet CC gradients by varying the voltage applied to the transducer. As shown in Fig. 6(a) and Movie S1 in the ESI, the droplet libraries with periodic in-droplet CCs were generated by applying periodic acoustic excitation with a pulse width of 0.5 s and period of 3.0 s (duty cycle of 16.67%). The rate of change in the solute concentration can be controlled by increasing the applied voltage from 2.25 to 2.85 VPP. The two inset figures in Fig. 6(a) demonstrate that the in-droplet CC was successfully varied among 10 droplets from 0.07 to 0.47 mM and 0.07 to 1.21 mM at 2.25 and 2.85 VPP, respectively. Our method outperforms passive dilution-based methods27,28 in the following respects: first, it is a continuous process that does not require a highly concentrated mother droplet to be diluted into subsequent daughter droplets; second, the gradient of the variation in the in-droplet CC among the droplets can be easily controlled by tuning the intensity of the electrical signal applied to the device; third, as an active method, our method offers on-demand dynamic control of the in-droplet CC with improved functionality. Moreover, in contrast to previous active methods,30–32 the acoustofluidic method offers on-chip operation with no moving components and can produce droplet libraries with varying in-droplet CC gradients. In our experiments, droplets on the nanoliter scale with 10 discrete in-droplet CCs were produced at 5 Hz, which is higher or comparable to the throughput of the previous methods for the generation of droplets with solute concentration gradients.27,31,32 As will be discussed below, our acoustofluidic method can offer further improved throughput of the droplet generation with more fractionated in-droplet CCs among the droplets by utilizing higher-frequency acoustic waves within smaller microchannels.
image file: d0lc00803f-f6.tif
Fig. 6 (a) Droplet libraries with periodic in-droplet solute concentrations obtained by using periodic acoustic excitation with a pulse width of 0.5 s and period of 3.0 s at (i) 2.25 VPP and (ii) 2.85 VPP. (b) Droplet libraries with a gradually increasing in-droplet chemical concentration gradient through the application of the pulse SAW excitation at 2.31, 2.55, 2.69, 2.75, and 2.85 VPP with a pulse width of 0.5 s and a period of 3.0 s. The droplets with the highest concentration in each case are marked with down-right arrows in the inset figures.

In addition, a sequence of droplets with a gradually increasing inter-droplet concentration gradient can be produced through the application of pulse SAW excitation at 2.31, 2.55, 2.69, 2.75, and 2.85 VPP with a pulse width of 0.5 s and a period of 3.0 s (duty cycle of 16.67%). as shown in Fig. 6(b) and Movie S2 in the ESI. This results implies that a variety of more complex droplet sequences for various droplet-based microfluidic applications can be produced by simply applying programmed AC signals to the proposed acoustofluidic device. Although the conditions in Fig. 6 do not exactly correspond to those in Fig. 5, the in-droplet CC values in Fig. 6 are in good agreement with the estimated values by inter- or extrapolation of the values in Fig. 5. Moreover, a time lag exists between the onset of the electrical signal and the variation in the in-droplet chemical concentration. This is attributed to the time required from acoustofluidic mixing in the main microchannel to the droplet generation at the T-junction and finally to the in-droplet CC measurement. Thus, the time delay totally depends on the distance from the acoustofluidic mixing region (Fig. 1(a1)) to the droplet observation region (Fig. 1(a3)). In the experiments, the average droplet speed was 1.41 mm s−1, and the aforementioned distance was approximately 6.5 mm. Therefore, the time delay in the presented experimental results is estimated as 4.6 s, which can be reduced by modifying the microchannel design.

Limitations and prospects

The proposed method to fine-tune the in-droplet CC relies on controlled acoustofluidic mixing induced by the microscale ASF. When acoustic waves propagate in a fluid, these waves transfer their momentum to the fluid by viscous dissipation, resulting in the ASF. One of the most important fluid properties that affect the acoustofluidic mixing is the fluid viscosity. Previous studies revealed that the streaming velocity exponentially decays with increasing fluid viscosity.46–48 This finding suggests that the proposed acoustofluidic method could be less effective for highly viscous fluids. The decreased acoustic streaming velocity can be compensated by lowering the total flow rate; however, the throughput of the system would be decreased accordingly.47 In the acoustic aspects, the acoustic wavelength plays a critical role in the ASFs.38,49 Both the acoustic attenuation lengths along the solid/fluid interface (α−1) and within the fluid (β−1) decrease with decreasing wavelength. The reduced attenuation lengths result in the increased acoustic power density, leading to the increased acoustic streaming velocity that allows enhanced acoustofluidic mixing. The decreased attenuation lengths are associated with smaller microchannel dimensions for both acoustofluidic mixing and droplet generation. Therefore, shorter-wavelength acoustic waves with the increased acoustic streaming velocity within the smaller microchannels could further improve the throughput of droplet generation and the level of discretization in the in-droplet CCs by using the proposed acoustofluidic method. On the other hand, despite the many advantages discussed earlier, it is difficult to produce a single droplet with the desired CC in the proposed acoustofluidic device. Acoustofluidic mixing occurs in the main microchannel (Fig. 1(a1)) that bifurcates downstream into a waste outlet and a T-junction (Fig. 1(a2)) for droplet generation. Even with no acoustic field applied to the parallel laminar streams, a series of droplets are continuously produced at the outlet, as shown in Fig. 2. However, relatively few applications require on-demand generation of droplets whereas a more variety of applications demand continuous generation of droplets with tunable inter-droplet chemical concentration gradients.

Experimental

For the FIDT fabrication, a pair of curved bimetallic electrodes composed of a 200 Å thick Cr layer and a 1000 Å thick Au layer was deposited on a 500 μm thick, 128°-rotated Y-cut X-propagating LiNbO3 wafer (MTI Korea, Korea) by photolithography, electron-beam evaporation and lift-off processes. The LiNbO3 substrate was used because of its transparency and high electromechanical coupling coefficient. The FIDT was fabricated as a single-phase unidirectional transducer with 30 finger pairs of electrodes and an inner radius of 4 mm that can produce 30 μm wavelength focused SAWs propagating only in the direction toward the microfluidic chip.33,50 A 2000 Å thick SiO2 thin film was deposited on the LiNbO3 substrate by photolithography and reactive ion etching to protect the electrodes and improve the bonding between the substrate and the microfluidic chip. A microchannel mold (SU-8 2150, MicroChem, USA) was prepared with standard photolithography for the fabrication of the PDMS (Sylgard 184A and 184B, Dow Corning, USA) microfluidic chip by using soft lithography. After punching the inlet and outlet holes, the microfluidic chip was bonded to the substrate by carrying out oxygen plasma treatment (Covance, Femto Science, Korea) on both the microfluidic chip and LiNbO3 substrate. The resonant frequency of the transducer was found by measuring the scattering parameter (S11) with a vector network analyzer (TTR503A, Tektronix, USA). The AC signals applied to the FIDT were produced with a radio frequency analog signal generator (N5171B-501, Keysight Technologies, USA), a power amplifier (ZHL-100 W-GAN+, Mini-Circuits, USA), and a DC power supply (UP-3015, Unicorn Tech., Korea). The peak-to-peak voltage of the applied AC signals was measured using an oscilloscope (MSOX4154A, Keysight Technologies, USA). The period and duty cycle of the applied signal was controlled by using an in-house MATLAB code.51,52 Microscopy images of the droplets in the proposed acoustofluidic device were obtained using a digital high-speed 10-bit CMOS camera (pco.1200 hs, PCO, Germany) linked to an inverted microscope (IX71, Olympus, USA). As a model sample fluid, 42.64 mg of erioglaucine disodium salt (Sigma Aldrich, USA), whose molecular weight is 792.85 g mol−1, was dissolved in 10 mL DI water (Sinhan Science Tech, Korea) to obtain a 5.37 mM sample solution. The volumetric flow rates for the sample fluid and sheath fluid were 250 and 750 μL h−1, respectively. After complete mixing, the resulting concentration was estimated as 1.34 mM. The sample and sheath fluids in the gas-tight glass syringes (Hamilton, USA) were injected through the inlet ports of the microchannel by using a multi-unit syringe pump (neMESYS Cetoni GmbH, Germany). A fluorocarbon oil (Novec™ 7500, 3M, USA) was used for the continuous phase with a 2 wt% non-ionic surfactant (Pico-Surf™ 1, Dolomite Microfluidics, UK) for stable droplet generation. The microscopy images of the produced droplets in the proposed acoustofluidic device were analyzed using ImageJ (http://rsb.info.nih.gov/ij).

Conclusions

We have developed an acoustofluidic method for the on-chip dynamic control of the concentration of solute encapsulated within droplets. Droplet libraries with the desired chemical composition gradients can be produced by simply varying the voltage applied to the acoustofluidic device, which in turn determines the magnitude of the focused acoustic field. We demonstrated the generation of droplets with periodically and gradually increasing discrete solute concentrations by applying programmed AC signals to the transducer. Our acoustofluidic method could be further improved to offer increased throughput with more fractionated in-droplet solute concentrations by utilizing shorter-wavelength SAWs. We expect that the proposed acoustofluidic approach makes new droplet-based microfluidic applications that require droplet libraries with tunable in-droplet concentrations possible.

Author contributions

H. J. S. supervised the research. J. P. conceived the research, designed and performed the experiments, and conducted numerical simulations of acoustic streaming flows. G. D. and J. P. performed numerical simulations of the surface acoustic wave amplitude on a piezoelectric substrate. J. P., G. D., and M. A. analysed the experimental results. All authors contributed to the writing of the manuscript. All authors have given approval to the final version of the manuscript.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was supported by the National Research Foundation of Korea (NRF) funded by the Korea government (MSIT) (No. 2020R1A2C2008106 and 2020R1A5A8018367).

References

  1. T. M. Keenan and A. Folch, Lab Chip, 2008, 8, 34–57 RSC .
  2. A. Karbalaei and H. J. Cho, Micromachines, 2018, 9, 149 CrossRef .
  3. B. Lin and A. Levchenko, Front. bioeng. biotechnol., 2015, 3, 39 Search PubMed .
  4. B. Meier, A. Zielinski, C. Weber, D. Arcizet, S. Youssef, T. Franosch, J. O. Radler and D. Heinrich, Proc. Natl. Acad. Sci. U. S. A., 2011, 108, 11417–11422 CrossRef CAS .
  5. N. Strelnikova, F. Herren, C.-A. Schoenenberger and T. Pfohl, Front. Mater., 2016, 3, 20 Search PubMed .
  6. C. H. Kwon, I. Wheeldon, N. N. Kachouie, S. H. Lee, H. Bae, S. Sant, J. Fukuda, J. W. Kang and A. Khademhosseini, Anal. Chem., 2011, 83, 4118–4125 CrossRef CAS .
  7. F. Guo, P. Li, J. B. French, Z. Mao, H. Zhao, S. Li, N. Nama, J. R. Fick, S. J. Benkovic and T. J. Huang, Proc. Natl. Acad. Sci. U. S. A., 2015, 112, 43–48 CrossRef CAS .
  8. S. K. Dertinger, D. T. Chiu, N. L. Jeon and G. M. Whitesides, Anal. Chem., 2001, 73, 1240–1246 CrossRef CAS .
  9. M. A. Holden, S. Kumar, E. T. Castellana, A. Beskok and P. S. Cremer, Sens. Actuators, B, 2003, 92, 199–207 CrossRef CAS .
  10. J. Atencia, G. A. Cooksey and L. E. Locascio, Lab Chip, 2012, 12, 309–316 RSC .
  11. S. Kim, F. Masum, J. K. Kim, H. J. Chung and J. S. Jeon, Lab Chip, 2019, 19, 959–973 RSC .
  12. J. Atencia, J. Morrow and L. E. Locascio, Lab Chip, 2009, 9, 2707–2714 RSC .
  13. G. Destgeer, S. Im, B. Hang Ha, J. Ho Jung, M. Ahmad Ansari and H. Jin Sung, Appl. Phys. Lett., 2014, 104, 023506 CrossRef .
  14. P. H. Huang, C. Y. Chan, P. Li, N. Nama, Y. Xie, C. H. Wei, Y. Chen, D. Ahmed and T. J. Huang, Lab Chip, 2015, 15, 4166–4176 RSC .
  15. I. K. Zervantonakis, S. K. Hughes-Alford, J. L. Charest, J. S. Condeelis, F. B. Gertler and R. D. Kamm, Proc. Natl. Acad. Sci. U. S. A., 2012, 109, 13515–13520 CrossRef .
  16. W. Lim and S. Park, Molecules, 2018, 23, 3355 CrossRef .
  17. D. J. Collins, A. Neild, A. deMello, A. Q. Liu and Y. Ai, Lab Chip, 2015, 15, 3439–3459 RSC .
  18. E. Brouzes, M. Medkova, N. Savenelli, D. Marran, M. Twardowski, J. B. Hutchison, J. M. Rothberg, D. R. Link, N. Perrimon and M. L. Samuels, Proc. Natl. Acad. Sci. U. S. A., 2009, 106, 14195–14200 CrossRef CAS .
  19. H. N. Joensson and H. Andersson Svahn, Angew. Chem., Int. Ed., 2012, 51, 12176–12192 CrossRef CAS .
  20. D.-K. Kang, M. Monsur Ali, K. Zhang, E. J. Pone and W. Zhao, TrAC, Trends Anal. Chem., 2014, 58, 145–153 CrossRef CAS .
  21. X. Wang, Z. Liu and Y. Pang, RSC Adv., 2017, 7, 29966–29984 RSC .
  22. A. Huebner, S. Sharma, M. Srisa-Art, F. Hollfelder, J. B. Edel and A. J. Demello, Lab Chip, 2008, 8, 1244–1254 RSC .
  23. N. Damean, L. F. Olguin, F. Hollfelder, C. Abell and W. T. Huck, Lab Chip, 2009, 9, 1707–1713 RSC .
  24. J.-W. Choi, S. Lee, D.-H. Lee, J. Kim, A. J. deMello and S.-I. Chang, RSC Adv., 2014, 4, 20341–20345 RSC .
  25. J. Wegrzyn, A. Samborski, L. Reissig, P. M. Korczyk, S. Blonski and P. Garstecki, Microfluid. Nanofluid., 2012, 14, 235–245 CrossRef .
  26. J. Shemesh, A. Nir, A. Bransky and S. Levenberg, Lab Chip, 2011, 11, 3225–3230 RSC .
  27. X. Niu, F. Gielen, J. B. Edel and A. J. deMello, Nat. Chem., 2011, 3, 437–442 CrossRef CAS .
  28. M. Sun, S. S. Bithi and S. A. Vanapalli, Lab Chip, 2011, 11, 3949–3952 RSC .
  29. S. Zeng, B. Li, X. Su, J. Qin and B. Lin, Lab Chip, 2009, 9, 1340–1343 RSC .
  30. S. Jambovane, D. J. Kim, E. C. Duin, S. K. Kim and J. W. Hong, Anal. Chem., 2011, 83, 3358–3364 CrossRef CAS .
  31. L. F. Cai, Y. Zhu, G. S. Du and Q. Fang, Anal. Chem., 2012, 84, 446–452 CrossRef CAS .
  32. F. Gielen, L. van Vliet, B. T. Koprowski, S. R. Devenish, M. Fischlechner, J. B. Edel, X. Niu, A. J. deMello and F. Hollfelder, Anal. Chem., 2013, 85, 4761–4769 CrossRef CAS .
  33. T.-T. Wu, H.-T. Tang, Y.-Y. Chen and P.-L. Liu, IEEE Trans. Ultrason. Ferroelectr. Freq. Control, 2005, 52, 1384–1392 Search PubMed .
  34. S. R. Fang, S. Zhang and Z. F. Lu, IEEE Trans. Ultrason. Ferroelectr. Freq. Control, 1989, 36, 178–184 CAS .
  35. M. S. Kharusi and G. W. Farnell, Proc. IEEE, 1972, 60, 945–956 Search PubMed .
  36. G. Destgeer, K. H. Lee, J. H. Jung, A. Alazzam and H. J. Sung, Lab Chip, 2013, 13, 4210–4216 RSC .
  37. J. H. Jung, G. Destgeer, B. Ha, J. Park and H. J. Sung, Lab Chip, 2016, 16, 3235–3243 RSC .
  38. M. B. Dentry, L. Y. Yeo and J. R. Friend, Phys. Rev. E: Stat., Nonlinear, Soft Matter Phys., 2014, 89, 013203 CrossRef .
  39. D. J. Collins, Z. C. Ma and Y. Ai, Anal. Chem., 2016, 88, 5513–5522 CrossRef CAS .
  40. W. Connacher, N. Zhang, A. Huang, J. Mei, S. Zhang, T. Gopesh and J. Friend, Lab Chip, 2018, 18, 1952–1996 RSC .
  41. M. Sesen, T. Alan and A. Neild, Lab Chip, 2014, 14, 3325–3333 RSC .
  42. J. Park, J. H. Jung, K. Park, G. Destgeer, H. Ahmed, R. Ahmad and H. J. Sung, Lab Chip, 2018, 18, 422–432 RSC .
  43. T. Kamakura, T. Sudo, K. Matsuda and Y. Kumamoto, J. Acoust. Soc. Am., 1996, 100, 132–138 CrossRef .
  44. J. Lighthill, J. Sound Vib., 1978, 61, 391–418 CrossRef .
  45. M. Alghane, B. X. Chen, Y. Q. Fu, Y. Li, J. K. Luo and A. J. Walton, J. Micromech. Microeng., 2011, 21, 015005 CrossRef .
  46. A. Riaud, M. Baudoin, O. Bou Matar, J. L. Thomas and P. Brunet, J. Fluid Mech., 2017, 821, 384–420 CrossRef .
  47. C. Westerhausen, L. G. Schnitzler, D. Wendel, R. Krzyszton, U. Lachelt, E. Wagner, J. O. Radler and A. Wixforth, Micromachines, 2016, 7 Search PubMed .
  48. C. Zhang, X. Guo, L. Royon and P. Brunet, Micromachines, 2020, 11, 607 CrossRef .
  49. L. Y. Yeo and J. R. Friend, Biomicrofluidics, 2009, 3, 12002 CrossRef .
  50. X. Sun, W. Liu, X. Shao, S. Zhou, W. Wang and D. Lin, Sensors, 2019, 19, 106 CrossRef .
  51. J. Park, B. H. Ha, G. Destgeer, J. H. Jung and H. J. Sung, RSC Adv., 2016, 6, 33937–33944 RSC .
  52. J. Park, J. H. Jung, G. Destgeer, H. Ahmed, K. Park and H. J. Sung, Lab Chip, 2017, 17, 1031–1040 RSC .

Footnote

Electronic supplementary information (ESI) available. See DOI: 10.1039/d0lc00803f

This journal is © The Royal Society of Chemistry 2020