pubs.acs.org/Langmuir © 2010 American Chemical Society
Electrocoalescence Mechanisms of Microdroplets Using Localized Electric Fields in Microfluidic Channels Michele Zagnoni,*,† Guillaume Le Lain,‡ and Jonathan M. Cooper*,† †
Department of Electronics and Electrical Engineering, University of Glasgow, G12 8LT, United Kingdom, and ‡ PHELMA/Grenoble INP, Universit e de Grenoble, Minatec, 38016, Grenoble Cedex 1, France Received April 16, 2010. Revised Manuscript Received August 6, 2010
Electrocoalescence of water-in-oil microdroplets in microfluidic channels is an active technique that enables dropletbased mixing functionalities to be achieved in lab-on-a-chip applications. In this work, a characterization of the electrocoalescence mechanisms of water microdroplets in oil is presented, using localized electric field systems. We report a theoretical and experimental description of the electrocoalscence behavior of droplet pairs by varying the physical and fluid dynamic conditions of the phases. Our results demonstrate that localized electric field systems can be reliably used to merge droplets in pairs, regardless of the distance between the drops. The coalescence behavior was dependent upon the viscosity of the continuous phase for water droplets that were separated by a thick layer of oil and upon interfacial tension for droplets that were in close proximity. We showed that these systems have the potential to be used for highthroughput applications and that, unlike other examples of active systems in the literature, the need of droplet synchronization and the application of high voltages is considerably reduced.
Introduction Droplet microfluidics, using immiscible fluids, have been rapidly established in recent years as a powerful technology to address lab-on-a-chip (LOC) applications. On-chip segmented flow provides compartmentalization and high-throughput processes, allowing thousands of parallel reactions to take place in microfluidic devices using reduced sample volumes (100 kV/m) by fixing an electric potential and reducing the distance between the electrodes. Compared to other microfluidic reports in the literature, the use of LEFs can be beneficial, as it does not require either precise electrode alignment with the microfluidics or droplet synchronization with an electric field or with another drop.21 Following this approach, we characterized (14) Chabert, M.; Dorfman, K. D.; Viovy, J. L. Electrophoresis 2005, 26, 3706– 3715. (15) Aryafar, H.; Kavehpour, H. P. Langmuir 2009, 25, 12460–12465. (16) Holto, J.; Berg, G.; Lundgaard, L. E. Annual Report - Conference on Electrical Insulation and Dielectric Phenomena, 2009, CEIDP, 5377866, pp 196199. (17) Thiam, A. R.; Bremond, N.; Bibette, J. Phys. Rev. Lett. 2009, 102, 188304. (18) Eow, J. S.; Ghadiri, M.; Sharif, A. O.; Williams, T. J. Chem.;Eng. J. 2001, 84, 173–192. (19) Ahn, K.; Agresti, J. J.; Chong, H.; Marquez, M.; Weitz, D. A. Appl. Phys. Lett. 2006, 88, 264105. (20) Priest, C.; Herminghaus, S.; Seemann, R. Appl. Phys. Lett. 2006, 89, 134101. (21) Zagnoni, M.; Cooper, J. M. Lab Chip 2009, 9, 2652–2658.
Published on Web 08/23/2010
DOI: 10.1021/la101517t
14443
Article
Zagnoni et al.
Figure 1. Double (a) and single (b) T-junctions and flow focusing junctions (c) were used to form water droplets in oil. Further downstream, microelectrodes perpendicular to the flow direction were present. Details of junction dimension (d,e): the size of the channel that met the electrodes was either 100 or 200 μm in width and 40 μm in depth. Details of electrode geometry (f): bonding pads were used to contact wires. An ac waveform of magnitude V was applied at the square-shaped electrodes. The four electrodes were connected so that (1,3) and (2,4) shared the same potential values.
electrocoalescence behavior of pairs of W/O drops in microchannels by varying the fluid dynamic and physical parameters of the phases, and we described different drop fusion mechanisms, both theoretically and experimentally. This study was carried out using only surfactant stabilized droplets, and electrocoalescence was used to actively merge drops that otherwise would not coalesce when brought into contact with each other. First, we estimated the electric forces produced on microfluidic W/O droplets by means of numerical simulations, under different electrodes configurations. These forces were compared with those due to capillarity and to hydrodynamic conditions in the channel, considering the advantages of LEFs with the other configurations. Subsequently, we experimentally characterized electrocoalescence of W/O microdroplets in different conditions of oil viscosity, interfacial tension, distance between droplets (the amount of continuous phase separating two drops), and droplet velocity, using LEFs systems.
Materials and Methods The dispersed phase consisted of either Milli-Q water or Milli-Q water with surfactant Tween80 (Sigma-Aldrich) at 2 wt %. The continuous phase consisted of either mineral oil (Sheen Instruments Ltd., η=14 mPa s at 25 °C, εr ∼ 3), used with surfactants Span80 (Sigma-Aldrich) at 0.1 wt %, or of fluorocarbon oil (Fluorinert FC40, 3M, η = 4.1 mPa s at 25 °C, εr = 1.9), used with either surfactant Krytox 157FSL (Dupont) at 2 wt % or surfactant EA (RainDance Technologies) at 2 wt %. Devices were fabricated using polydimethylsiloxane (PDMS) (Sylgard 184 Silicone Elastomer, Dow Corning) and glass microscope slides with titanium-gold electrodes, using standard photolithography (fabrication details in Supporting Information). In order to obtain the desired droplet dimension and to control the spacing between drops, different microfluidic structures were used, as shown in Figure 1. These consisted of channels (40 μm depth) with either a double (Figure 1a) or single T-junction (Figure 1b) or a flow focusing junction (Figure 1c). After bonding the PDMS microchannel to the glass slide and in order to obtain the required surface properties for the channels, the device was flushed first with either undiluted Aquapel (PPG Industries), when using fluorinated oil, or with undiluted Sigmacote (Sigma-Aldrich), when using hydrocarbon oil. Devices were finally flushed with air and left in the oven for 1 h at 65 °C. A TTi TG120 function generator (2 MHz bandwidth) and an in-house voltage amplifier (100 kHz bandwidth) were used to apply an ac field with a square waveform (2 kHz, 0-200 V) to the electrodes. 14444 DOI: 10.1021/la101517t
Flow rates, ranging from 0.1 up to 15 μL/min, were generated and controlled by syringe pumps (New Era Pump System Inc.). The devices were imaged using an inverted microscope (Axio Observer A1, Zeiss), and imaging was acquired using a MotionScope M2 fast camera (Redlake). Flow behaviors were recorded at frame rates from 500 Hz to 4 kHz. In order to guarantee emulsion stability and to determine the wetting condition of the phases over the channel walls, drop contact angle (phases in air over substrate) and equilibrium interfacial tension values between the phases were measured, using the pendant drop method and the sessile drop method, respectively (Easydrop, Kruss). The results are shown in Table S1 (Supporting Information), confirming complete wetting of the continuous phase over the treated channel walls,22 for each tested combination of the phases. Although this protocol validates the wetting condition of the phases only at the equilibrium, we assumed in our experiments that a thin oil layer was always present between droplets and the channel walls even in dynamic conditions.23 In addition, Baret et al. have reported that the dynamics of adsorption of surfactants at a droplet water-oil interface can strongly influence coalescence, depending upon the surfactant concentration and the drop incubation time.24 In our system, the parameters of interest are considerably higher than the ones reported in ref 24, potentially with the exception of one case (mineral oil at 0.1 wt % Span80). Therefore, as coalescence occurred only by electrical means, we considered such effects negligible in our experiments.
Results and Discussion Theoretical and Empirical Models. Before characterizing experimentally the electrocoalescence behavior of W/O droplets in microchannels, we considered theoretically the forces that act on the W/O droplets. For the condition tested, resulting Reynolds (Re) and Bond (Bo) numbers were always 2S1 and D > 2S2, no droplet fusion (either of pairs or irregular) was obtained. We speculate that this could be caused by two factors: the increased oil viscosity, which increases ΔPH, and the thickness of the oil layer between the electrode and the drop, as more viscous oils produced a thicker film23 that reduced the effect of the electric field. However, for the same phases and a reduced distance between droplets, a different behavior occurred. Only for 0.5S1 < D < S1 and γ ≈ 10 mN/m, droplets could be fused in pairs (Figure 6). In this situation, it was noticed that drops were partially trapped and “delayed” in the electrode area, so that only droplets separated by short distances could be fused. The data also supported the hypothesis of a thicker oil layer influence on electrocoalescence, as the potentials needed to fuse the droplets were much higher if compared to the previous cases (FC40 oil in Figure 6). Under these circumstances, hydrolysis was not observed. Finally, when low values of interfacial tensions were tested, both for 0.5S1
