pubs.acs.org/Langmuir © 2009 American Chemical Society
Use of Surfactants To Reduce the Driving Voltage of Switchable Optical Elements Based on Electrowetting Thibault Roques-Carmes,*,† Alexandra Gigante,† Jean-Marc Commenge,‡ and Serge Corbel† †
D epartement de Chimie Physique des R eactions, UMR 7630 CNRS-INPL, Nancy-Universit e 1, rue Grandville, BP 20451, 54001 Nancy Cedex, France, and ‡Laboratoire des Sciences du G enie Chimique, UPR 6811 CNRS-INPL, Nancy-Universit e 1, rue Grandville, BP 20451, 54001 Nancy Cedex, France Received March 12, 2009. Revised Manuscript Received June 17, 2009
The advantage of using electrowetting as a novel principle for a reflective display has been previously demonstrated. The principle is based on the controlled two-dimensional movement of an oil/water interface across a hydrophobic fluoropolymer insulator. The main objective of this paper is to show experimentally the influence of surfactants on the electro-optic behavior of a single electrowetting pixel. The concentration and type of nonionic surfactant (Tween 80 and Span 20) have been varied. The experimental data are compared with calculations from the electro-optic model developed previously. The electro-optic performance is significantly affected by the nature and the concentration of surfactant. In the presence of Tween, at concentrations lower than the critical micelle concentration (CMC), and mixtures of Tween and Span the electro-optic behavior can be related to the interfacial tension. When decreasing the oil/water interfacial tension, the amplitude of the driving voltage required for obtaining a given oil displacement decreases and the switching curve becomes steeper. These effects can be accurately reproduced by means of the previously developed electro-optic model. Mixtures of Tween and Span produce a significant synergetic reduction of the driving voltage. For Tween concentrations higher than the CMC and Span, a strong disagreement is observed between the previously developed model and experimental data. Here a new physical model is reported that describes the electro-optic behavior of electrowetting-based optical elements in the presence of surfactants. The model takes into account the actual voltage used to control the liquid movement in electrowetting (lower than the applied voltage), the amount of surfactant adsorbed at the decane/water interface, and the dipole moment of the surfactant molecules. The calculated results are in very good agreement with experimental data without employing fitting parameters. The dipoles interact with the applied field and lower the actual applied field. This reduction of the effective electric field across the solid-liquid interface induces a decrease in the charge density at the solid-liquid interface and reduces the electrowetting force. For surfactant concentrations higher than the CMC, the electro-optic performance does not depend on the surfactant concentration. This demonstrates that the reduction of the electrowetting field due to the large dipole moment of the surfactant molecules occurs at the oil/water interface. A new method for the test cell fabrication is also presented.
Introduction The application of an electrical potential across a solid-liquid interface modifies the wetting properties of that interface by reducing the solid-liquid interfacial energy. This can induce a reversible contact angle change without altering the bulk liquid and solid properties.1,2 The field of electrowetting is currently the focus of increased experimental and theoretical activity driven by *Corresponding author: e-mail
[email protected], Tel þ33 (0) 3 83 37 53 35, Fax þ33 (0) 3 83 37 81 20.
(1) Quilliet, C.; Berge, B. Curr. Opin. Colloid Interface Sci. 2001, 6, 34. (2) Shamai, R.; Andelman, D.; Berge, B.; Hayes, R. A. Soft Matter 2008, 4, 38. (3) Prins, M. W. J.; Welters, W. J. J.; Weekamp, J. W. Science 2001, 291, 277. (4) Prins, M. W. J.; Weekamp, J. W.; Giesbers, B. J. Micromech. Microeng. 1999, 9, 362. (5) Berge, B.; Peseux, J. Eur. Phys. J. E 2000, 3, 159. (6) Kuiper, S.; Hendricks, B. H. W. Appl. Phys. Lett. 2004, 85, 1128. (7) Hendricks, B. H. W.; Kuiper, S.; van As, M. A. J.; Renders, C. A.; Tukker, T. W. Opt. Rev. 2005, 12, 255. (8) Pollack, G.; Fair, R. B.; Shenderov, A. D. Appl. Phys. Lett. 2000, 77, 1725. (9) Raccurt, O.; Berthier, J.; Clementz, P.; Borella, M.; Plissonnier, M. J. Micromech. Microeng. 2007, 17, 2217. (10) Blake, T. D.; Clarke, A.; Stattersfield, E. H. Langmuir 2000, 16, 2928. (11) Mach, P.; Krupenkin, T.; Yang, S.; Rogers, J. A. Appl. Phys. Lett. 2002, 81, 202. (12) Feenstra, B. J.; Hayes, R. A.; Camps, I. G. J.; Hage, L. M.; Johnson, M. T.; Roques-Carmes, T.; Schlangen, J. M.; Franklin, A. R.; Valdes, A. F.; Ford, R. A. J. Soc. Inf. Disp. 2004, 12, 293. (13) Hayes, R. A.; Feenstra, B. J. Nature 2003, 425, 383. (14) Sun, B.; Zhou, K.; Lao, Y.; Heikenfeld, J.; Cheng, W. Appl. Phys. Lett. 2007, 91, 011106/1.
Langmuir 2009, 25(21), 12771–12779
applications such as optical filters,3,4 lenses,5-7 lab-on-a-chip,8,9 coating assist,10 optical fibers,11 and reflective displays.12-14 Hayes and Feenstra12,13,15 have recently proposed electrowetting as a novel principle for a reflective display. In Figure 1, the principle of the reflective electrowetting display is presented. The principle is based on the movement of colored oil across a hydrophobic fluoropolymer insulator. Figure 1a shows the optical stack: a white (reflecting) substrate, hydrophobic insulator, colored oil, and water. In the absence of voltage, at equilibrium, the colored oil film naturally forms a continuous film between the hydrophobic insulator and an immiscible electrolyte (Figure 1a). However, when a voltage is applied across the hydrophobic insulator, an electrostatic term is added to the energy balance. The stacked state is no longer energetically favorable, and a state as shown in Figure 1b is formed. The system can lower its free energy by increasing the water-insulator contact area, thereby displacing the oil. The photographs in Figure 1c,d show a typical oil displacement obtained for a homogeneous electrode. The balance between electrostatic and capillary forces determines how far the oil is moved.16 These aspects have recently been (15) Hayes, R. A.; Feenstra, B. J.; Camps, I. G. J.; Hage, L. M., Roques-Carmes, T.; Schlangen, J. M.; Franklin, A. R.; Valdes, A. F. Proc. Soc. Inf. Disp. Conf. 2004, 1412. (16) Roques-Carmes, T.; Hayes, R. A.; Feenstra, B. J.; Schlangen, J. M. J. Appl. Phys. 2004, 95, 4389.
Published on Web 09/28/2009
DOI: 10.1021/la900882h
12771
Article
Roques-Carmes et al.
Figure 1. Electrowetting display principle. Schematic cross section: (a) without applied voltage, a homogeneous oil film is present. R is the angle between the oil film and the fluoropolymer substrate at the corner of the square pixel; (b) dc voltage applied; the oil film is contracted. θ refers to the angle between the spherical oil cap and the fluoropolymer surface. The corresponding top view photographs in (c) and (d) demonstrate a typical oil retraction obtained with a homogeneous electrode.
explored via an electro-optic model that we have developed.17 The model deals with static electrowetting18 (in contrast to spontaneous electrowetting19) and utilizes the basic electrowetting equation: cos θðVapplied Þ ¼ cos θð0 VÞ -
ε0 εr Vapplied 2 2dγOW
ð1Þ
where θ and θ(0 V) are the equilibrium contact angles of the oil at the insulator surface in the presence and absence of an applied voltage (Vapplied), respectively, as well as the well-defined (spherical cap) geometry of the confined oil film to predict the electro-optic response based on the known physical variables: pixel size, insulator thickness (d) and dielectric constant (εr), oil film volume and oil/water interfacial tension (γOW). There is clearly a very good agreement with experimental data (in absence of surfactant).17,20 However, challenges for these devices include the lowering of the driving voltage. A low voltage is less energy intensive and costeffective. To lower the driving voltage, it is logical to reduce insulator thickness, oil film thickness, and oil/water interfacial tension as much as possible.16,17,20 Recent improvements of the hydrophobic insulator material allow the electrowetting driving voltage to decrease dramatically due to improvements in the processing of hydrophobic insulators and the use of alternative (17) Roques-Carmes, T.; Hayes, R. A.; Schlangen, J. M. J. Appl. Phys. 2004, 96, 6267. (18) Yeo, L. Y.; Chang, H.-C. Mod. Phys. Lett. B 2005, 19, 549. (19) Yeo, L. Y.; Chang, H.-C. Phys. Rev. E 2006, 73, 011605. (20) Roques-Carmes, T.; Palmier, S.; Hayes, R. A.; Schlangen, J. M. Colloids Surf., A 2005, 267, 56. (21) Berry, S.; Kedzierski, J.; Abedian, B. J. Colloid Interface Sci. 2006, 303, 517. (22) Kedzierski, J.; Berry, S. Langmuir 2006, 22, 5690.
12772 DOI: 10.1021/la900882h
materials.21,22 In a recent paper,20 we have theoretically and experimentally demonstrated that decreasing the oil/water interfacial tension decreases the amplitude of the driving voltage required for initiating oil displacement. In this case, the oil/water interfacial tension was decreased up to 20 mN/m by changing the dye type and dye concentration. There is some scope to reduce the oil/water interfacial tension below 20 mN/m. This goal can easily be reached using surfactants. To this extent, knowledge of the effect of surfactants in the electrowetting process is of great importance. Nonionic surfactants were used to minimize the possible electrokinetic effects that can be expected during electrowetting.23 We have identified Tween 80 and Span 20 as good surfactants for electrowetting devices with a moving oil/water interface. Both substances are typical nonionic stabilizers of industrially produced emulsions and are widely used for research purposes.24 It is also well-known that certain mixtures of surfactants, and more particularly a mixture of Tween 80 and Span 20, provide better performance than pure surfactants in a wide variety of applications.25-27 The main objective of the present paper is to investigate how the electro-optic properties of oil/water electrowetting devices are affected when using oils and/or water that contain surfactants. Experimental results will be interpreted in terms of measured oil/water interfacial tension and dipole moment of the adsorbed (23) Malloggi, F.; Gu, H.; Banpurkar, A. G.; Vanapalli, S. A.; Mugele, F. Eur. Phys. J. E 2008, 26, 91. (24) Gurkov, T. D.; Horozov, T. S.; Ivanov, I. B.; Borwanker, R. P. Colloids Surf., A 1994, 87, 81. (25) Rosen, M. J. In Mixed Surfactant Systems; American Chemical Society: Washington, DC, 1992. (26) Porras, M.; Solans, C.; Gonzalez, C.; Martinez, A.; Guinart, A.; Gutierrez, J. M. Colloids Surf., A 2004, 249, 115. (27) Nandi, I.; Bari, M.; Joshi, H. AAPS PharmSciTech 2003, 4, 1.
Langmuir 2009, 25(21), 12771–12779
Roques-Carmes et al.
Article
molecules. A new electro-optic model will be developed to interpret and discuss the electro-optic characteristics. The described results provide useful guidelines for the formulation and application of surfactants to reduce the driving voltage in electrowetting-based optical elements that can be designed around moving oil/water interfaces (display12 and lens5,6). Another aim of the paper is to describe a new method for the test cell fabrication.
where Aoil(Vapplied) and Aoil(0 V) denote the pixel area occupied by the oil in the presence and absence of an applied voltage (Vapplied), respectively. The electro-optic characterization of the test cells was done with an initial reading at zero voltage, after which the dc voltage was increased in steps of 5 V up to 50 V. The time interval between measurements points (1 min) was more than an order of magnitude longer than the time required for displacing the oil film (experimentally around 40-100 ms16).
Experimental Section Test Cell Fabrication and Characterization. Test cells are
received from Aldrich. Oils were formulated by dissolving a nonpolar dye in decane. The dye used in this study was Sandoplast Red BB (Clariant). The oil formulation was described in a previous paper.20 The nonionic surfactants were Tween 80 and Span 20 (Aldrich). Span 20 is a sorbitan monolaurate (HLB = 8.6) while Tween 80 is a polyoxyethylene (20) sorbitan monooleate (HLB = 15). These surfactants were used as received without further purification. Tween 80 is water-soluble and was dissolved in the aqueous phase. Span 20 is predominantly oilsoluble and was dissolved in the oil phase. Tween/Span mixtures of the desired composition were also prepared by dissolving the Tween in the aqueous phase and the Span in the oil phase. The decane/water interfacial tensions were measured at room temperature (20 ( 1 C) with a du Nouy tensiometer (Kruss) with a platinum ring. The interfacial tension of decane/water systems in the presence of surfactant was measured as a function of the surfactant concentration. For a nonionic surfactant, the Gibbs adsorption equation, dγ = -ΓRT d ln C, allows the derivation of the surface equation of state relating the surface tension γOW to the surface concentration (surfactant surface excess (Γ)) as
composed of a stack of materials, as schematically shown in Figure 1a. The substrate was a glass sheet with a thin transparent indium-tin oxide ITO electrode (Visiontek System). The glass sheet was cleaned with a surfactant solution (Plantacare 200) by sonication, thoroughly rinsed with deionized water, ethanol, and heptane, and dried under a stream of nitrogen.28 The insulating layer was formed by dip-coating of a fluoropolymer solution (AF 1600, Dupont) dissolved in perfluoro-2-butyltetrahydrofuran (FC-75, Acros) on the glass substrate according to the procedure of Seyrat and Hayes.29 The plate was immersed in the coating solution, which consisted of 4 wt % AF 1600, and withdrawn at a velocity of 0.5 mm/s, which resulted in a coating thickness of 0.8 μm. The strong hydrophobic nature of the fluoropolymer coating ensured complete spreading of the oil film. In the literature, either the pixel wall fabrications to contain the water and the oil were made by gluing onto the insulator-covered substrate a black polymer sheet and an outer Delrin frame16,20 or a complex photolithographic process using a SU-8 negative photoresist was used.15 To avoid the difficult gluing of a black polymer sheet16,20 or the prohibitive cost of the reactive ion etching attack of the Teflon to make the fluoropolymer hydrophilic,15 the pixel fabrication was performed using the following procedure. The photopolymerizable epoxy resin (SI 40, Accura) does not completely wet the fluoropolymer layer (contact angle of 40-50) and direct UV exposition of the resin deposited on the fluoropolymer are not sufficient to anchor the solid resin on the fluoropolymer. Consequently, the thin tip of a screw was used to locally pull out a very thin layer of fluoropolymer. That provided the external edges of a square pixel. The resin layer was then introduced in the free fluoropolymer hole with a syringe which delivered a controlled quantity of resin on the working area. The sample was then placed in a UV chamber for 3 min. The UV radiation was absorbed by the photoinitiator which, in turn, generated free radicals to initiate the polymerization of a liquid monomer into a solid polymer. When a solid layer was formed, depositing of a subsequent layer of photopolymerizable resin on the already polymerized part was performed. This procedure was repeated until the polymer solid part was built. The stacking of the layers yielded to the threedimensional pixel wall on the fluoropolymer, enabling it to contain the water and oil. An open test cell with a single pixel was used. After each measurement, the test cell was cleaned for reuse. The influence of the surfactant concentration and type on the electro-optic properties was investigated with 80 μm thick layers of oil within a single pixel test cell measuring 2 2 mm2. Changes in the oil wetting behavior were monitored optically. The top view images of the pixel were captured using a CCD camera. Each recorded image was analyzed using Optimas software to extract the oil-free pixel area (pixel area without oil). The pixel white area fraction (WA) is defined by WAðVapplied Þ ¼ 1 -
Aoil ðVapplied Þ Aoil ð0 VÞ
ð2Þ
(28) Roques-Carmes, T.; Membrey, F.; Fili^atre, C.; Foissy, A. J. Colloid Interface Sci. 2002, 245, 257. (29) Seyrat, E.; Hayes, R. A. J. Appl. Phys. 2001, 90, 1383.
Langmuir 2009, 25(21), 12771–12779
Formulation and Characterization of Oil and Water with Surfactants. Decane was chosen as the oil and was used as
γOW ¼ γ0 - RTΓ¥ lnð1 þ KL CÞ
ð3Þ
where γ0 is the oil/water interfacial tension when the interface is in the absence of surfactants, Γ¥ is the maximum surfactant surface concentration, KL is the ratio of the rate constants for adsorption and desorption, and C is the surfactant concentration in the bulk. This equation contains two unknown parameters, Γ¥ and KL, which can be obtained by fitting the experimental γOW as a function of the surfactant concentration.30,31 Then, using the Γ¥ and KL parameters, the Langmuir adsorption isotherm (Γ = Γ¥KLC/(1 þ KLC)) can be used to calculate the actual surface concentration adsorbed at the decane/water interface for each surfactant concentration.
Results Two preliminary experiments were performed in order to interpret the electro-optic properties of the oil/water electrowetting devices. Decane/water interfacial tension measurements were necessary to correlate the oil/water interfacial tension to the electro-optic behavior. In addition, electrowetting measurements of water droplet (without decane) were performed to demonstrate that the presence of Span and micelles of Tween in the water phase do not disturb the electrowetting effect. Decane/Water Interfacial Tension Measurements. Figure 2 depicts the decane/water interfacial tension as a function of the concentration of Tween and Span. In the absence of surfactant the decane/water interfacial tension was measured to be 52 mN/ m. The curves exhibit a decrease of the interfacial tension with the surfactant concentration up to a characteristic break point above which the interfacial tension does not vary. The CMC values were determined by the breaks in interfacial tensions curves from the Gibbs plot (Figure 2). The CMC of Tween 80 and Span 20 are 8 10-4 and 5 10-4 mol L-1, respectively. The interfacial tension of (30) Giribabu, K.; Ghosh, P. Chem. Eng. Sci. 2007, 62, 3057. (31) Peltonen, L. J.; Hirvonen, J.; Yliruusi, J. J. Colloid Interface Sci. 2001, 240, 272.
DOI: 10.1021/la900882h
12773
Article
Roques-Carmes et al. Table 1. Decane/Water Interfacial Tension (γoil/water) for Different Surfactant Types and Concentrations Used in Electrowetting Experimentsa γoil/water (mN/m)
Figure 2. Effect of the bulk concentration of Span and Tween on the interfacial tension at water/decane interface. For mixtures of Span and Tween, the interfacial tension is plotted as a function of Span concentration. The concentrations of Tween are 8 10-6, 8 10-5, and 8 10-5 M for 3 10-5, 3 10-4, and 3 10-3 M Span, respectively.
Tween 80 and Span 20 at the CMC are around 7 and 22 mN/m, respectively, indicating that Tween 80 is more effective in lowering the interfacial tension than Span 20. From the results of the interfacial tension measurements, the actual surfactant concentration adsorbed at the decane/water interface for each surfactant concentration was calculated using the procedure described above (eq 3). The values are reported in Table 1 for the bulk surfactant concentration used in the electrowetting setup. These values agree well with the values reported in the literature.31,32 It is observed that the adsorbed amount of Tween 80 and Span 20 differs considerably. The adsorbed amount of Span is quite small compared to that of Tween. Tween 80 forms a monolayer at the decane/water interface 2 times higher than Span 20 molecules; thus, Tween 80 molecules are more effective in adsorption at the decane/water interface. Mixtures of Tween and Span were also used (Figure 2). The concentrations of Tween were 8 10-6, 8 10-5, and 8 10-5 M for 3 10-5, 3 10-4, and 3 10-3 M Span, respectively. The data exhibit clearly and interestingly the same trend as above for Tween 80. Moreover, the interfacial tension obtained with Tween is very close to that obtained with mixtures of Tween and Span. The interfacial tensions are slightly lower with the mixture than with Tween alone. This slight difference may more likely reflect the coadsorption of Span in the case of the mixture. From the interfacial tension data, the surface excess at the decane/water interface was estimated using the same procedure as for a single surfactant. However, the total surface concentration of adsorbed materials (Tween and Span) were estimated, but the specific uptake of each surfactant was unknown. Because of the higher affinity of Tween toward the interface, we consider that the adsorbed amount of Tween is the same as it was in the absence of Span. The adsorbed amount of Span was then calculated from the difference between the total and Tween adsorbed amount (Table 1). This approach should obviously be considered as a first approximation, since the system is a mixture of surfactants rather (32) Fainermann, V. B.; Lylyk, S. V.; Aksenenko, E. V.; Makievski, A. V.; Petkov, J. T.; Yorke, J.; Miller, R. Colloids Surf., A 2009, 334, 1.
12774 DOI: 10.1021/la900882h
Γ (μmol/m2)
Vth-geo (V)
Vexp (V)
absence of surfactant 52 0 15 10-15 18 0.44 15 10-15 Tween 8 10-6 M -5 12 0.95 14 10-15 Tween 8 10 M 7 1.61 16 25-30 Tween 8 10-4 M 30 0.22 16 20-25 Span 3 10-5 M -4 25 0.55 14 30-35 Span 3 10 M 22 0.76 14 40-45 Span 3 10-3 M 12 0.8 14 10-15 Tween 8 10-6 M 0.15 Span 3 10-5 M -5 5 0.95 10 10-15 Tween 8 10 M 0.12 Span 3 10-4 M 3 1.07 10 10-15 Tween 8 10-5 M 0.15 Span 3 10-3 M a The amount of surfactant adsorbed at the decane/water interface (Γ) is calculated using eq 3 from the decane/water interfacial tension data. Comparison of the experimental and calculated threshold voltages. The threshold voltage Vth-geo is calculated using model results (eq 1). It takes into account the initial white area fraction (WA(0 V)) and the angle of the spherical oil cap without any applied voltage (R). The experimental threshold voltage (Vexp) is extracted from Figure 4.
than individual substances. A rigorous theory should assume multicomponent mixtures of surfactants.33,34 Electrowetting of Water Droplet. The electrowetting behavior of a water droplet containing Tween at different concentrations (Figure 3a) and Span at different concentrations (Figure 3b) is reproduced in Figure 3. Note that no oil was added in any of the experiments reported here. The water/vapor surface tensions (γ) corresponding to the various water formulations used here are indicated in the figure. The critical micelle concentration (CMC) of Tween 80 and Span 20 are 8 10-5 and 7 10-5 M, respectively. In absence of applied voltage, the contact angle decreases with the surfactant concentration and more precisely with the water/vapor surface tension. Over a considerable voltage range the experimental data are in agreement with the electrowetting theory (cos β(V) = cos β0 þ (ε0εr/2γd)V2, where β and β0 denote the contact angles measured through the water phase in the presence and absence of voltage, respectively). The equation implies that the contact angle of the water droplet with a substrate decreases when an electric field is applied. The influence of the surfactant concentration can be seen through the decrease in the width of the contact angle profile as the surfactant concentrations are increased: the water/vapor surface tension decreases with increasing surfactant concentration, and thus a lower voltage is required to generate the same change in contact angle. The figure shows that the presence of micelles (Tween and Span) and Span in the water does not affect the electrowetting behavior. As with other material systems, at high voltage, contact angle saturation exists where no additional contact angle change occurs for additional increase in applied voltage.35-41 We will not further interpret the contact angle saturation here. Electrowetting of Water/Decane Films. The influence of Tween concentration (Figure 4a), Span concentration (Figure 4b), and a mixture of Tween and Span (Figure 4c) on (33) (34) (35) (36) (37) (38) (39) (40) (41)
Nikas, Y. J.; Puvvada, S.; Blankschtein, D. Langmuir 1992, 8, 2680. Mulqueen, M.; Blankschtein, D. Langmuir 1999, 15, 8832. Welters, W. J. J.; Fokking, L. G. J. Langmuir 1998, 14, 1535. Berry, S.; Kedzierski, J.; Abedian, B. Langmuir 2007, 23, 12429. Wang, K. L.; Jones, T. B. Langmuir 2005, 21, 4211. Vallet, M.; Berge, B.; Voyelle, L. Polymer 1996, 37, 2465. Verheijen, H. J. J.; Prins, M. W. J. Langmuir 1999, 15, 6616. Peykov, V.; Quinn, A.; Ralston, J. Colloid Polym. Sci. 2000, 278, 789. Quinn, A.; Sedev, R.; Ralston, J. J. Phys. Chem. B 2005, 109, 6268.
Langmuir 2009, 25(21), 12771–12779
Roques-Carmes et al.
Figure 3. Water contact angle (β) as a function of the applied voltage for different surfactant concentrations: (a) Tween, (b) Span. The points correspond to experimental data and the lines to model results (cos β(V) = cos β0 þ (ε0εr/2γd)V2). The water/ vapor surface tensions (γ) corresponding to the various water formulations used are indicated in the figure. The critical micelle concentration (CMC) of Tween 80 and Span 20 are 8 10-5 and 7 10-5 M, respectively.
the electro-optic properties of the test cell is depicted in Figure 4. In each figure, the data are compared with the electro-optic properties of the pure decane/water system in order to assess the effect of the presence of surfactant. The validity of the model was checked by comparing the calculated and experimental white area fraction as a function of the applied voltage. The model results are presented in Figure 4 (lines). Each electro-optic curve shows a similar trend. A threshold is observed before displacement of the oil film occurs. Then, as the voltage increases, the slope reaches a maximum and then decreases when going to higher voltages. The electro-optic curve of pure decane/water is similar to that described by Roques-Carmes et al. for 2 2 mm2 test cells with a 40 μm tetradecane film thickness20 and for 1 1 mm2 test cells with a 30 μm tetradecane film thickness16 as well as by Feenstra et al. for 0.16 0.16 mm2 test cells with a 15 μm tetradecane film thickness.12 This concordance between the series of data validates our fabrication procedure and, more specifically, the fact that the fluoropolymer pull-out before the wall deposition does not affect the electrowetting behavior. Langmuir 2009, 25(21), 12771–12779
Article
The electro-optic performance is significantly affected by the nature and the concentration of surfactant. In the presence of Tween, at concentrations lower than the CMC (Figure 4a), and mixtures of Tween and Span (Figure 4c), the electro-optic behavior can be related to the interfacial tension. As the interfacial tension decreases, the electro-optic curve shifts to higher amplitudes and the threshold voltage slightly decreases. A low oil/ water interfacial tension is desirable because this results in larger pixel white area fractions at lower driving voltage amplitudes. The presence of Tween and Span reduces the oil/water interfacial tension more efficiently than pure surfactants. A mixture of Tween and Span produces a significant synergetic effect on the reduction of the driving voltage. To lower the driving voltage further, mixtures of Tween and Span are highly recommended. A very good agreement between the proposed model and experimental data is observed. The threshold voltage (Vth-geo) is accurately explained by the geometrical angle R (Figure 1a). Because of the pinning of the oil/water interface at the insulator pixel wall interface, the initial angle of the spherical cap R is determined by the oil volume dosed into the pixel. When starting from V = 0 and increasing the voltage difference, the angle θ(V) must exceed R before the oil/water interface starts moving.17 Consequently, a threshold (of about 10-15 V) is observed before displacement of the oil film (Table 1). However, for Tween concentrations larger than the CMC (Figure 4a) and Span (Figure 4b), the electro-optic performance becomes different: the threshold voltage strongly increases while the electro-optic curve steepens dramatically with the surfactant concentration. The electro-optic behavior cannot be directly related to the interfacial tension. There is a large difference in the white area fraction at 50 V: the WA(50 V) decreases with the surfactant concentration and also with the decrease in the oil/ water interfacial tension. Hence, the shift to higher voltages is mainly due to the increase in the threshold voltage. As a result, the presence of Span strongly increases the driving voltage, and operation with pure decane without Span is advised. A strong disagreement is observed between the proposed model and experimental data (Figure 5). The model does not take into account either the fact that the electro-optic curve shifts toward larger voltages or the increase in the threshold voltage (Table 1). The calculated initial value of the angle of the spherical oil cap without applied voltage, R, is typically around 5. In order to fit the experimental results, the required threshold voltage would correspond to R = 29, 46, and 60 for Span concentrations from 3 10-5 to 3 10-3 M and R = 60 for Tween. These values are unrealistic since they would correspond to initial white area fractions of 0.4-0.7, which are far from the experimental ones (lower than 0.05).
Discussion In this part, we show that the disagreement between the experimental data and the model can be explained by the large dipole moment of the surfactant molecules. We discuss the relation between the actual voltage used to control the liquid movement in electrowetting (lower than the applied voltage) and the amount of surfactant adsorbed at the decane/water interface taking into account the dipole moments of the surfactant molecules. Then, using the real voltage, we develop the model that is presented here. We check the validity of the new electrowetting model by comparing calculated and experimental data. Additional data corresponding to electrowetting experiments performed with Tween and Span at surfactants concentrations higher than the CMC are added to confirm the validity of the model. DOI: 10.1021/la900882h
12775
Article
Roques-Carmes et al.
Figure 5. Comparison of the experimental (symbols) and calculated (lines) electro-optic curves. Pixel white area fraction as a function of the applied voltage for different surfactant concentrations: (a) Tween, (b) Span. The points are experimental data from Figure 4, and the lines are model results (eq 1).17
Figure 4. Electro-optic curves for 2 2 mm2 test cells with a 80 μm oil film thickness. Pixel white area fraction as a function of the applied voltage for different surfactant concentrations: (a) Tween, (b) Span, (c) mixtures of Tween and Span. The oil/water interfacial tensions corresponding to the various formulations used are indicated in the figure. The points are experimental data, and the lines are model results (eq 1).17 The absence of line indicates that a strong disagreement is observed between the model and experimental data.
12776 DOI: 10.1021/la900882h
It has been recently demonstrated that the use of toluene and methyl naphthalene as oil phase compromises the electrowetting performance.42,43 A large increase in the threshold voltage was obtained. It has now been found that this is due to the large dipole moment of the molecules (μtoluene = 0.31 D and μmethylnaphthalene = 0.37 D). The dipoles interact with the applied field and disturb the electrowetting effect. It is then more suitable to use oils comprising symmetric molecules (benzene, silicone oil, or decane) because their low dipole moment (μbenzene = μdecane = 0 D) does not affect the electric fields used to control the liquid movement in electrowetting. Polar molecules possess a permanent dipole moment (μ), and these molecules will tend to reorient in the applied field. The treatment considers a local field F inside the dielectric and its relation to an applied field E. The difference between E and F (42) Kuiper, S.; Hendriks, B.; Renders, C. A.; Hayes, R. A. US Patent 7242528, Oct 7, 2007. (43) Hayes, R. A.; Joulaud, M.; Roques-Carmes, T.; Palmier, S. US Patent 0225374, Sept 18, 2008.
Langmuir 2009, 25(21), 12771–12779
Roques-Carmes et al.
Article
is due to the polarization of the medium P due to the dipole moment (μ)44 F ¼E-
P ε0
ð4Þ
The Debye equation for the molar polarization reads as P¼
N μ2 F B 9kT
ð5Þ
with N the number of dipole moments in a volume B. The applied field E is reduced since the polarization provides a reduction of the actual field F. It also appears that the larger the dipole moment, the greater the tendencies of the solvent to respond to an applied field by reorientation of the microscopic dipoles and the lower the actual field F. This reduction of the effective electric field across the solid-liquid interface induces a reduction in the charge density at the solid-liquid interface and then reduces the electrowetting force. Tween and Span are expected to have electrical dipole moments which can strongly modify the electric field E. Calculation of the dipole moment of Span and Tween molecules was performed with MOPAC (Molecular Orbital PACkage) using the semiempirical PM3 method.45 The method is not strictly accurate, but the high number of atoms for both surfactants prevents the use of more precise methods such as DFT calculation or ab initio calculation. In addition, the calculation gives the gas phase dipole moment and does not provide the actual dipole moment in the oil phase. However, it does provide qualitative insight of the phenomenon. The calculated gas phase dipole moment were μspan = 4.2 ( 0.8 D and μtween = 1.5 ( 0.5 D for Span and Tween, respectively. In the case of Tween, the results were very sensitive to the molecule conformation. With 12 carbon atoms (in the hydrocarbon chains for the Span), the polarity caused by the shorter fatty acid chain length has been reported to result in higher dipole moment compared to Tween and also to result in aqueous phase solubilization. The Span exhibits a water-oil partition coefficient [water]/ [decane] of 85 (where [water] and [decane] denote the concentrations of Span in the water and decane, respectively).31 The Span (HLB = 8.6) is then distributed in both the aqueous and oil phases. In order to emphasize such a redistribution effect, the oil-water pairs of solution prepared for the electrowetting experiments were allowed to pre-equilibrate by staying in contact in a wide flask for more than 24 h. It can then be considered that the majority of Span is present in the aqueous phase. On the basis of this statement, we consider that the majority of Span molecules are adsorbed at the decane-water interface (surface excess) and, at least, in the aqueous phase. The resulting amount of Span in the oil phase is not substantially large to affect the electric fields used to control the liquid movement in electrowetting. In addition, we have previously demonstrated that the presence of Span in the water phase does not disturb the electrowetting effect (Figure 3). It can be therefore assumed that the reduction of the electrowetting field due to the large dipole moment of the molecules does not occur in the oil phase, as for toluene or methylnaphthalene oil, but at the oil/water interface. This observation is consistent with the fact that a nonhomogeneous high electric field is experienced (44) Sands, M.; Leighton, R.; Feynman, R. In The Feynman Lectures on Physics; Addison-Wesley: Reading, MA, 1966; Vol. II. (45) Stewart, J. J. P. QCPE 1990, 10, 86. (46) Mugele, F.; Buehrle, J. J. Phys.: Condens. Matter 2007, 19, 375112. (47) Janocha, B.; Bauser, H.; Oehr, C.; Brunner, H.; Gopel, W. Langmuir 2000, 16, 3349.
Langmuir 2009, 25(21), 12771–12779
Figure 6. Relation between the calculated real voltage (eq 7) and the applied voltage.
along the three-phase contact line where the local field strength is much larger than in the homogeneous oil part.46,47 Now, we calculate the resulting voltage inside the electrowetting test cell taking into account the polarization of the medium due to dipole moment. When an electric potential is applied between the aqueous liquid and the insulator coated counter electrode, the electric fields can be expressed as E = Vapplied/d and F = Vreal/d. Vapplied denotes the applied voltage, Vreal is the actual voltage due to the dipole polarization, and d is the insulator thickness. Equation 4 now becomes Vreal ¼
Vapplied 1þ
N μ2 B 9kTε0
ð6Þ
Assuming that the interaction between the electric field and the dipole moment occurs only at the oil/water interface, eq 6 can be written as Vreal ¼
Vapplied 1þ
ΓN a μ2 h 9kTε0
ð7Þ
with Na Avogadro’s number, h the thickness of the surfactant adsorbed layer at the oil/water interface, and Γ the amount of surfactant adsorbed at the decane/water interface. By means of eq 7, the real voltage at a given applied voltage can be calculated. In Figure 6, the calculated real voltage (Vreal) is plotted against the applied voltage for a variety of experimental conditions (data from Figure 4a,b). The surfactant adsorbed amount at the oil/ water interface is estimated from interfacial tension measurements and is reported in Table 1. The real voltage is strongly affected by the product of the square of the dipole moment (μ2) by the surfactant adsorbed amount at the oil/water interface (Γ). As μ2Γ increases, the real voltage strongly decreases. In the case of Tween, at concentrations lower than the CMC, the surfactant adsorbed amount is relatively low and the dipole moment is quite low: the difference between the real and applied voltage is less than 5 V. We consider that the reduction of the effective electric field across the solid-liquid interface induces by the adsorbed dipole moment can be neglected. Recall that model results (eq 1) and experimental data (Figure 4) were in good agreement. However, for Tween concentrations larger than the CMC and Span, μ2Γ strongly increases. The difference between the real and applied voltage is higher than 5 V and becomes significant. DOI: 10.1021/la900882h
12777
Article
Roques-Carmes et al.
the applied voltage. The results are shown in Figure 7 for a variety of experimental conditions (Tween concentration of 8 10-4 M (data from Figure 4a) and Span (data from Figure 4b)). In Figure 7, additional data corresponding to electrowetting experiments performed with Tween and Span at surfactant concentrations higher than the CMC are added to confirm the validity of the model. When the thickness (h) of the surfactant adsorbed layer at the decane/water interface equals 0.1 nm, a very good agreement between the proposed model (eq 8) and experimental data is now observed. Both the threshold voltage required to initiate movement of the colored oil film and the subsequent electro-optic dependence of the fluidic motion are well described. It is clear that the model catches the essential underlying physics of the electrowetting-based optical elements, in the presence of surfactants, without requiring any adjustable fitting parameters. The evolution of the electro-optic curves depends on both the amount of surfactant adsorbed at the decane/water interface and the dipole moment of the surfactant. We conclude that the reduction of the electrowetting field due to the large dipole moment of the surfactant molecules occurs at the oil/water interface. Figure 7 shows that, for surfactant concentrations higher than the CMC, the electro-optic curve for four different surfactant concentrations collapse to a single curve. The calculated results are in very good agreement with experimental data for a large range of surfactant concentrations. The electro-optic performance is not affected by the presence of micelles. For surfactant concentration higher than the CMC, the surfactant adsorbed amount remains constant; i.e., μ2 Γ does not vary. This demonstrates that the reduction of the electrowetting field due to the large dipole moment of the surfactant molecules occurs at the oil/water interface.
Conclusion
Figure 7. Comparison of the experimental (symbols) and calculated (lines) electro-optic curves. Pixel white area fraction as a function of the applied voltage for different surfactant concentrations: (a) Tween, (b) Span. The points are experimental data, and the lines are model results (eq 8).17 The notation “>CMC” indicates that the surfactant concentration is higher than the CMC.
The real voltage strongly decreases with the oil/water Span adsorbed amount. For instance, the real voltage at an applied voltage of 50 V decreases from 35 to 18 V with the Span adsorbed amount. The effect of the reduction of the effective electric field across the solid-liquid interface induces by the adsorbed dipole moment has to be taken into account. Recall that a strong disagreement was observed between the model (eq 1) and experimental data (Figure 5). The modified electrowetting equation accounting for the polarization of the oil/water interface due to the dipole moments of the adsorbed surfactant molecules is 0 12 ε0 εr ðVreal Þ2 ε0 εr @ Vapplied A ¼1cos θðVÞ ¼ 1 2dγOW 2dγOW 1 þ ΓNa μ2 h 9kTε0
ð8Þ We check the validity of the model (eq 8) by comparing the calculated and experimental white area fraction as a function of 12778 DOI: 10.1021/la900882h
The aim of this paper was to show experimentally the influence of surfactants on the electro-optic behavior of a single electrowetting pixel. A new method for the test cell fabrication is described. Assessment of the feasibility of the fabrication procedure was made by comparing electro-optic curves of pure decane/ water with previous experiments. The concordance between the series of data validates the fabrication procedure and more specifically the fact that the pull-out of the fluoropolymer before the wall deposition does not affect the electrowetting effect. The concentration and type of nonionic surfactants (Tween 80 and Span 20) have been varied. The resulting changes in the oil/ water interfacial tension have been evaluated together with electro-optic characteristics of an open pixel. In order to better understand the electro-optic behavior in the presence of surfactants, we have developed a new electro-optic model. The new model employs the actual voltage used to control the liquid movement in electrowetting (lower than the applied voltage) and the amount of surfactant adsorbed at the decane/water interface taking into account the dipole moments of the surfactant molecules. The experimental data are compared with calculations performed using the electro-optic model. There is a very good agreement between the new model and experimental data for a large range of surfactant nature and concentrations. The observed threshold behavior of electrowetting-based optical element follows naturally from the model and can be explained using the electrical dipole moments of the surfactant molecules adsorbed at the decane/water interface. The electro-optic performance is significantly affected by the nature and the concentration of surfactants. In the investigated Langmuir 2009, 25(21), 12771–12779
Roques-Carmes et al.
test cells, oil film contraction is governed by the competition between the oil/water interfacial tension and the electrical dipole moments of the surfactant molecules adsorbed at the decane/water interface. In the presence of Tween, at concentrations lower than the CMC, and mixtures of Tween and Span, the overall interface dipole moment is low and can be neglected. The electro-optic behavior can be well related to the interfacial tension. When decreasing the oil/ water interfacial tension, this decreases the amplitude of the driving voltage and makes the electro-optic curve steeper. Mixtures of Tween and Span reduce the oil/water interfacial tension more effectively than a simple surfactant. To lower the driving voltage, mixtures of Tween and Span are highly recommended. For Tween concentration larger than the CMC and Span, the overall interface dipole moment becomes large. The dipole effect becomes significantly high to disturb the electric field. The dipoles interact with the applied field and lower the actual applied field. That decrease in the effective electric field across the solid-liquid interface induces a decrease in the charge density at the liquid-solid interface and reduces the electrowetting force. For surfactant concentrations higher than the CMC, the electro-optic performance is not affected by the surfactant concentration. This
Langmuir 2009, 25(21), 12771–12779
Article
demonstrates that the reduction of the electrowetting field due to the large dipole moment of the surfactant molecules occurs at the oil/water interface. Reported results provide useful guidelines for the formulation of mixtures of water and oils for electrowetting-based devices with moving oil/water interfaces. This work suggests that the electrical dipole moment of the surfactant molecules adsorbed at the oil/ water interface is a key parameters in the electro-optic behavior. In order to significantly reduce the driving voltage, surfactants leading to low oil/water interfacial tensions and exhibiting a small dipole moment have to be used. Acknowledgment. The authors thank P. A. Glaude (DCPR) for dipole moment calculations. T. Roques-Carmes warmly acknowledged R. Hayes (Liquavista) and L. Schlangen (Philips) for their precious help to build the first version of the electrowetting model during his postdoctoral position at Philips Research Laboratories in Eindhoven. This work was financially supported by the Federation de Recherche Jacques Villermaux pour la mecanique, l’energie et les procedes (FR 2863, CNRS) within the project “Microfluidique et systemes fluidiques multi-echelles”.
DOI: 10.1021/la900882h
12779
