Electrowetting-on-dielectric actuation of a vertical translation and angular manipulation stage

The efficacy and longevity of actuation-based microelectromechanical systems (MEMS) are limited by the tribological issues such as friction and adhesion.¹ Adhesion typically occurs between small asperities when nominally smooth solid surfaces come into contact, and friction is the result of this adhesion during motion of the surfaces relative to each other. In some cases, the forces from adhesion and friction are comparable to the forces actuating the device, rendering it unusable. In the presence of water vapor, these concerns are amplified; the formation of a thin liquid film on solid surfaces and the corresponding capillary and viscous effects that come into play when these surfaces interact lead to stiction, a phenomenon that increases the adhesion between solid asperities and causes small features to stick together due to an exceptionally high friction.^1–3 These reactive forces at the physical contact of solid components lead to wear and removal of material, reducing device lifetime.^1,4,5 This is of particular importance for traditional MEMS devices fabricated with silicon, a brittle hydrophilic material that is known to have poor tribological properties.^6,7 It follows that the elimination of solid-solid contact during actuation, for example, by designing MEMS devices where the actuator plates do not come in to contact, eases tribological and stiction concerns. This work offers a solution for MEMS that eliminates solid-solid contact by using electrowetting to actuate a vertical translation stage.

Electrowetting, a phenomenon whereby the contact angle of a fluid is altered with an applied voltage, allows control of droplet shape.⁸ Recent research has focused on electrowetting on a dielectric, or EWOD, in which an insulating layer is placed between a conductive surface and a droplet that rests on the surface; EWOD can provide much greater droplet deformation than electrowetting on conductive surfaces, and as such is the primary mode of electrowetting used in the practical applications.^9–12 EWOD has played a diverse role in MEMS applications to date,^13–17 ranging from fluid lenses for optical manipulation^18–21 to switches for electrical^22,23 and thermal control^24,25 and thermal management.²⁶ EWOD has also been implemented extensively in the lab-on-a-chip applications, where arrays of electrodes are activated in sequence to control the droplet motion on a surface.^27–30 With EWOD, controlled droplet vibration and water droplet “jumping,”³¹ or departure, from hydrophobic surfaces have been demonstrated.^31–33 EWOD has also been proposed as a method to depin droplets that have been impaled on superhydrophobic surface structures,³⁴ which is relevant to the recent work on superhydrophobicity.^35–37 However, many opportunities remain to implement EWOD, particularly in the context of MEMS. In this work, we investigated the use of EWOD for its ability to control a MEMS vertical translation, or focusing, stage, thereby providing an alternative to methods that suffer from tribological failure at solid-solid contacts.

EWOD can alter the contact angle of a fluid on a surface and, accordingly, can change the geometry of droplets resting on an EWOD surface when voltage is applied through the droplets. By sandwiching droplets between an electrically conducting stage on one side and an EWOD surface on the other side, a vertical translation stage can be fabricated at the micro-scale, as shown schematically in Figure 1(a). The stage translates when a voltage is applied across the drops and their contact angle with the EWOD surface decreases. For example, an applied voltage causing a decrease in contact angle on the dielectric layer at the base will result in a broadening and flattening of the droplets and corresponding downward translation of the stage towards the base. This operational mechanism is similar to the capillary force actuator, a class of actuator that relies on the deformation of a liquid droplet between two solid surfaces and offers distinct advantages compared to other MEMS actuators.³⁸ While the actuation³⁹ and dynamics^40,41 of such devices have been explored theoretically, an experimental device without solid-solid contact has not yet been demonstrated, nor has angular deflection been considered.⁴² 

We developed an axisymmetric iterative numerical model for the four identical droplets in our device, which is outlined in Figure 1(b) for one of the droplets and later used for comparison with the experimental results obtained from a working device, to determine the stage height as a function of the applied voltage. First, the contact angle at the EWOD base was determined with the Lippman-Young equation as a function of the intrinsic contact angle and applied voltage. Then, the droplet curvature in the system was calculated as a function of the internal Laplace pressure, which was determined at the top of the droplet (underside of the stage) by summing one quarter of the stage weight and the surface tension force pulling downwards at the sides of the droplet, γ_LVsin(θ_top), and then dividing that quantity by the fixed top contact area. Note that the curvature relies on the initially unknown droplet contact angle at the underside of the stage, θ_top, which is why an iterative solution was implemented. Finally, the complete droplet profile was determined numerically under the constant curvature constraint by iterating until convergence, described in detail in the supplementary material. The model was used to determine the profiles of droplets under different applied voltages, as shown in Figure 1(c). The flat region at the top of each profile is the contact with a pinning site on the bottom of the stage, where the constant radius over different applied voltages is consistent with the physical picture. The contact angle at this pinned region varies, as does the radius of the base on the EWOD surface, both of which are expected.

We fabricated the stage by first growing rough copper oxide (CuO) nanoblades on copper foil following a common procedure detailed in the literature^35–37 and then functionalizing the CuO with a monolayer of trichloro(1H,1H,2H,2H-perfluorooctyl)silane (Sigma-Aldrich) to form a superhydrophobic surface. The advancing and receding contact angles of the superhydrophobic CuO were θ_A = 172° ± 3° and θ_R= 168° ± 3°, respectively, as measured with a microgoniometer (MCA-3, Kyowa). The pinning sites were subsequently formed on the underside of the stage by milling away the CuO to a negligible depth using an end mill with a diameter of approximately 1.5 mm, thereby exposing the hydrophilic^43–45 copper and forming a liquid pinning site at the junction of the hydrophilic and superhydrophobic regions to fix the top radius of the droplets. Finally, to establish a non-constraining electrical connection with the stage, a copper wire was soldered in a vertical orientation to a tab at the corner of the stage in order to attach to a sliding attachment mounted above the stage. The total stage mass was 0.080 g, and the load mass (paper with MIT logo shown in Figure 2) was 0.010 g.

Indium-tin-oxide-coated glass slides with resistivity of 10 Ω/sq were used as the conductive substrate for the EWOD base. The slides were solvent cleaned and plasma cleaned, then coated with a 4 μm thick parylene-C layer (VSI Parylene, precision ±1 μm, θ_Y ≈ 100°) with dielectric strength of 22,000 V/m and relative permittivity of ε_r ≈ 3. Several slides were additionally coated with a sub-micron coating of Teflon aqueous fluoropolymer (AF) as detailed in past work³² in order to study the device performance with a higher Young angle at the EWOD base (θ_Y ≈ 116°). The water contact angles on the two different EWOD bases at varying applied voltages were characterized by applying voltage through a copper wire electrode inserted into a single 2 μl droplet of 0.1 mM KCl solution in water resting on the EWOD base. The EWOD base was electrically grounded with copper alligator clips penetrating through the dielectric coating to the ITO and subsequently mounted on the stage in front of the high-speed camera. The voltage was varied up to 150 V, and the contact angle was in excellent agreement with the Lippmann-Young prediction until the saturation voltage for each sample, which occurred at contact angles of 65° and 74° for the parylene-C and Teflon AF coatings, respectively.

The experimental setup consisted of a function generator (AFG 3101, Tektronix) passed through a 400x voltage amplifier (A800, FLC Electronics) with the positive lead wired to the EWOD base and the negative lead attached to the stage via the sliding electrical connection to allow free translation in the vertical z-direction. The stage provided direct electrical connection to the droplets through the conductive hydrophilic copper circles on its underside. The device was both front- and back-lit for high-speed video capture (Phantom v7.1, Vision Research) from 500 to 10,000 frames per second, as the experiment was conducted. Four droplets of 0.1 mM KCl solution in water with a volume of 2 μl were carefully pipetted onto the pinning sites on the underside of the stage, which was then inverted and placed onto the EWOD base (the pinned droplets did not fall from the stage). Finally, the stage sliding electrical connection was attached. Figure 2 shows the device with a load on the stage, and the foremost two of the four droplets beneath the stage can be seen.

Images of a typical experiment were captured in Figure 2 (see video in supplementary material). At an applied voltage of 150 V, the contact angle decreased from 100° to 65° on the parylene-C coated EWOD base and from 116° to 74° on the Teflon AF coated EWOD base. This caused the droplets to spread while the volume remained constant and thus resulted in a decrease in stage height, as predicted by the model. Tests at intermediate voltages in Figure 3(c) show good agreement with the model prediction, which is a combination of the Lippman-Young equation shown in Figure 3(a) (including saturation) and the stage height as a function of base contact angle for a 2 μl droplet generated by the iterative solution (plotted in supplementary material).

To eliminate solid-solid contact and the accompanying stiction and tribological concerns, the stage was reconfigured to remove the requirement for the sliding electrical connection. This was achieved by separating the Teflon AF coated EWOD base into two electrically insulated components, each holding two droplets (Figure 4(a)). Then, voltage was applied from one insulated section of the EWOD base to the other, forming a circuit comprising two capacitors (the dielectric regions at the base of the droplets on each of the insulated EWOD base sections). Since each of these series capacitors carry half of the applied voltage, twice the voltage required in the previous configuration is required for the same stage deflection. The stage deflection in this configuration was experimentally demonstrated to be equivalent to the previous (wired) configuration and in good agreement with the model, as shown in Figure 4(b).

Additionally, the stage can provide angular deflection. The configuration was further modified to keep the electrically separated EWOD base but once again include the stage sliding electrical connection, which was grounded (Figure 4(c)). When a voltage is applied to either insulated section of the EWOD base, that side of the stage is displaced downwards while the other side remains unperturbed, resulting in an angular deflection. To test this configuration, we constructed a varying voltage that first actuated one side of the stage and then actuated the other side of the stage. The function generator/amplifier output was set to increase from 0 V to 150 V and then decrease back to 0 V repeatedly as a sine wave with amplitude 75 V, offset +75 V and period 2 s. This signal was followed by a microcontroller (UNO R3, Arduino) which used a motor shield (L298P, Arduino) to switch relays (7266K64, McMaster-Carr) that alternated the applied voltage between the two sides of the EWOD base each time the signal bottomed out at 0 V, leaving the non-active side of the EWOD base at 0 V. The result of the signal applied to this configuration is shown in Figure 4(d), where each side of the stage deflected by approximately 130 μm when the voltage was applied (diamond and square symbols), in agreement with the uniform vertical stage displacement demonstrated previously, and the stage angular displacement varied from approximately −1° to +1° (hollow circular symbols).

Combining the two modified configurations above could yield an angular deflection stage that does not require any solid-solid contact (no stage electrode connection). This is possible by separating the EWOD base into four electrically isolated sections, one for each droplet, and then essentially controlling the deflection of two adjacent droplets by applying a voltage across the EWOD base beneath those droplets. Such a configuration would also allow angular deflection along any axis of rotation within the plane of the stage. A further expansion of this concept could build on past work in which the EWOD surface was separated into an array of isolated electrodes, which were actuated separately such that lateral droplet motion was induced.^46–49 Operation of the stage described in the present work on such an array of electrodes could allow for lateral as well as vertical translation.

This work shows a MEMS vertical translation stage that uses EWOD as the actuating mechanism. The EWOD stage was capable of linear spatial manipulation with the resolution of 10 μm over a maximum range of 130 μm, which can be readily improved and tailored to specific applications in the future device generations with guidance from the validated model developed in the present work. Specifically, both model and experiment show that a higher intrinsic contact angle on the EWOD base improves absolute range and the reduction of contact angle hysteresis,⁵⁰ possibly by adding a lubricant to the surface^51–53 or careful control of contaminants,⁴⁴ will increase resolution. In addition, angular deflection of approximately ±1° was demonstrated, and the maximum range and angular deflection are comparable to another MEMS alternative, piezoelectric actuators. The capability to operate the stage without any solid-solid contact makes this a desirable potential solution to stiction and tribology concerns for the improvement of applications in micro-optics, actuators, and other MEMS.

See supplementary material for a video of the stage in operation, experiment schematic, derivation of the Lippmann-Young equation, consideration of stage-base electrostatic attraction, and pseudocode describing the numerical model solution.

We thank Professors John W. M. Bush and Michael Cima at MIT for fruitful discussions regarding the work. We gratefully acknowledge funding support from the Office of Naval Research (ONR) with Dr. Mark Spector as program manager. D. J. Preston acknowledges funding received from the National Science Foundation (NSF) Graduate Research Fellowship under Grant No. 1122374. We also acknowledge support from the NSF through the Major Research Instrumentation Grant for Rapid Response Research (MRI-RAPID).