Dielectric materials for electrowetting-on-dielectric actuation
- First Online:
- Cite this article as:
- Liu, H., Dharmatilleke, S., Maurya, D.K. et al. Microsyst Technol (2010) 16: 449. doi:10.1007/s00542-009-0933-z
- 1.1k Views
The fundamental building blocks of typical electrowetting-on-dielectric (EWOD) actuation and their importance in the EWOD mechanism are introduced and reviewed, respectively. The emphasis in this experimental study of EWOD is on dielectric materials, upon which the performance of EWOD devices is heavily dependent. Dielectric breakdown of several typical polymeric and inorganic insulators employed as dielectrics for EWOD has been analytically investigated, which is forced to occur between the electrodes and conductive liquids under certain threshold potential. The electric breakdown occurring in both dielectric layer and surrounding medium (air or silicon oil) has been studied to build up a mathematical model of breakdown voltage as a function of dielectric thickness. Contact angle measurement of some polymeric materials and self-assembled monolayer using pure water has been carried out to demonstrate the contact angle tunability and reversibility, respectively, upon EWOD actuation.
Electrowetting-on-dielectric (EWOD) (Berge 1993; Verheijen and Prins 1999), which originates from the phenomenon of electrowetting (Lippmann 1875; Beni and Hackwood 1981), has been extensively investigated since the last decade and internationally recognized as a promising technology for liquid manipulation and actuation. With the well-established micro/nanofluid manipulation techniques using EWOD, the latest research efforts have concentrated on demonstration of optical/display technology (Berge and Peseux 2000; Kuiper and Hendriks 2004; Hayes and Feenstra 2003) and lab-on-chip applications (Pollack et al. 2002; Belaubre et al. 2004; Hoshino et al. 2004; Yi and Kim 2004; Cheng and Hsiung 2004) of EWOD systems and devices. Several papers have reviewed in detail the recent development of electrowetting and EWOD covering topics from theory, material properties, surface morphologies to applications (Quilliet and Berge 2001; Darhuber and Troian 2005; Mugele et al. 2005).
Generally a typical EWOD device must consist of four compulsory components—an ionized liquid working as a medium, a conductive material serving as the electrodes and a dielectric material providing both the capacitance between liquid and electrode, and hydrophobicity of its surface at the solid–liquid interface (some dielectrics require an additional coating of a thin hydrophobic material due to inadequate hydrophobicity of their natural surface). However, these components of different intrinsic material properties influence the performance of an EWOD device in different ways.
The effect of a variety of liquids on EWOD has been investigated in recent years and the results indicate that the requirement of liquid properties such as concentration and charge carrying ability are not that severe for EWOD initiation. Using typical liquid electrolytes like KCl and K2SO4 with concentrations of the order of 0.01–1.0 M/l, Verheijen and Prins (1999) reported the independence of ion type, ion molarity and ion valence of electrolytes on the applied voltages for EWOD. Mugele et al. (2005) presented an EWOD experiment with DI water on indium-tin-oxide (ITO) electrode coated with a thin layer of Teflon® AF and obtained a noticeable contact angle change. The electrowetting phenomenon was also observed for mixtures of salt solutions (NaCl or Na2SO4) with other chemicals such as glycerol (Mugele and Herminghaus 2002; Klingner et al. 2004), ethanol (Vallet et al. 1999) and methanol (Krupenkin et al. 2004), without any degradation of EWOD performance.
Since more and more research interests on EWOD have been driven by its biomedical applications, researchers have exploited the applicability of EWOD devices in the presence of biomolecules and physiological fluids. Yoon and Garrell (2003) showed that EWOD occurred with biofluids containing protein, DNA and adult bovine whole serum of a biomolecule concentration of 4 μg/ml, although the performance was degraded by the adsorption of biomolecules into the insulation layer giving rise to the reduced contact angle change and hysteresis. The compatibility of EWOD with physiological fluids such as insulin, cytochrome, myoglobin etc. has been demonstrated by Wheeler et al. (2004) using an EWOD-based chip. Recently, EWOD actuation of human physiological fluids containing human whole blood, serum, plasma, urine, saliva sweat and tears was demonstrated at 20 Hz and less than 65 V by Fair et al. (Srinivasan et al. 2004) using an integrated digital microfluidic chip. Hitherto, all the corresponding experimental results identically show that the performance of an EWOD device has very limited dependence on liquids properties.
Another component, the metal electrode, which is dominant in the semiconductor and printed circuit board (PCB) industry, has also been commonly employed in EWOD and patterned by photolithography and lift off/wet etching processes. These materials include gold (Au) (Yi and Kim 2004), silver (Ag) (Kuo et al. 2003), aluminum (Al) (Verheijen and Prins 1999), chromium (Cr) (Cho et al. 2003), platinum (Pt) (Cho et al. 2003; Moon et al. 2002), titanium (Ti) (Cho et al. 2003), and Si (Wheeler et al. 2004). The typical thickness of the electrode is in the range of several nanometers to 200 nanometers. Stainless steel sheets have also been reportedly used as electrodes for EWOD (Vallet et al. 1999). Because of the increasing demand to visualize the device operation or induce light transmission in practical experiments, the top-substrate conductive material should be transparent or semi-transparent to light. ITO films of 50–200 nm thickness intrinsically have a transmittance of above 80% with sheet resistances of less than 10 Ω/sq. Therefore, ITO has been popularly used as the material for the top-substrate electrode in the EWOD device (Pollack et al. 2002; Wheeler et al. 2004; Srinivasan et al. 2004; Cho et al. 2003). Comparatively, the influence of electrode property on EWOD is insignificant.
In contrast with these components mentioned above, the properties of dielectric materials have been experimentally verified to have a remarkable impact on EWOD performance and have hence been investigated intensively since the discovery of EWOD. A significant amount of research focusing on optimizing the extrinsic properties (e.g., thickness, applied voltages) and intrinsic properties (e.g., surface property, dielectric strength or constant) of EWOD have provide valuable information for practical EWOD design and application.
In this study, we present a sufficient comparison of the properties of polymeric and inorganic dielectrics widely reported and employed in EWOD, together with a critical analysis of their merits/demerits for EWOD applications. The role of dielectrics on the evolution of EWOD technology is summarized. We also demonstrate dielectric breakdown analysis to obtain the electrowetting voltage and breakdown voltage as functions of dielectric thickness for several typical dielectric materials. Contact angle measurements were conducted to compare the contact angle tunability and reversibility of those dielectrics upon EWOD actuation. The results show that Cytop is the most promising and competitive dielectric material for EWOD applications.
2 Polymeric and inorganic dielectrics
Comparisons of material properties of polymeric dielectrics for EWOD
Parylene -N and -C
Teflon® AF 1600
Dielectric strength (kV/mm)
Applied voltage (V)
<1 k (AC 50–20 kHz)
<600 (AC 1 kHz)
<800 (AC 2 kHz)
Typical thickness (μm)
Contact angle of water (°)
Chemical vapor deposition
Spin or dip coating
Comparisons of material properties of inorganic dielectrics for EWOD
SiO2 (Chang et al. 1996)
Si3N4 (Thomas 1996)
BST (Imanaka et al. 2002)
Dielectric strength (kV/mm)
Applied voltage (V)
Vdc ≥ 25
Vdc ≥ 15
100 nm–1 μm
Contact angle of water (°)
Thermal oxidation or PECVD
Chemical vapor deposition
Progress of EWOD development
Minnema et al. (1980)
AC 0–2 kV (50 or 1 k Hz)
PE (200–900 μm)
θ changes from 80° to ~20°
AC (50–3 k Hz)
PTFE (53 μm)
DC (0–0.9 V)
Gorman et al. (1995)
1.0 mW/cm2 (light intensity)
θ changes from 24° to 11°
Vallet et al. (1996)
AC (0–3 kV at 50 Hz–a few kHz)
PET (12 μm)
Reversibility at moderate voltage
Contour line instability at high voltages
Welters and Fokkink (1998)
DC (0–400 V)
Parylene C (6–10 μm) or polyimide (18–35 μm) + Teflon® AF 1600 (0.1 μm)
θ changes from 110° to ~60°
Verheijen and Prins (1999)
DC (±240 V)
Parylene N (10 μm) + Teflon® AF 1600 (30 nm) + silicon oil
Polarity, ion type, molarity and ion valence independent
Prins et al. (2001)
DC (±227 V)
Parylene C (11.5 μm) + Teflon® AF1600 (10 nm)
High transportation speed
Mach et al. (2002)
DC (0-100 V)
Polyimide (2 μm) + Cytop™ (1 μm)
Pinning-free voltage control
Cho et al. (2003)
DC (15-25 V)
SiO2 (0.1 μm)/BST (70 nm) + Teflon® (20 nm)
Low voltage operation
Versatile droplet handling
Srinivasan et al. (2004)
AC (65 V at 20 Hz)
Parylene C (800 nm) + Teflon® AF1600 (50 nm) + silicon oil
High transportation rate
3 Dielectric breakdown analysis
3.1 Dielectric breakdown in dielectrics
Dielectric breakdown is a phenomenon where insulators are forced to conduct electricity. In EWOD, it specifically refers to the failure of insulating materials which results in a short circuit between the electrodes and conductive liquids. Dielectric breakdown damages dielectric layers, and charges do not accumulate at the interface of the dielectric and the conducting liquid to give rise to a destructive malfunction of the EWOD device. Hence, caution must be taken in determining the minimum thickness of an insulating layer, the dielectric strength of which is one of the most important factors in materials selection.
The dielectric breakdown voltage for corresponding minimum thickness of typical dielectrics (under 100 V)
Teflon® AF 1600
2.87 × 10−3
3.05 × 10−3
Table 4 lists the minimum dielectric thickness (dmin) and dielectric breakdown voltage (Vbd) corresponding to different typical dielectric materials. The dielectric breakdown voltage Vbd is the minimum theoretical actuation voltage required to initiate the EWOD effect in air without any failure of the insulator. A practically acceptable maximum driving voltage is 100 V and for this voltage, the minimum (limited by dielectric strength Eds) and maximum (limited by driving voltage Vew) value of thickness for all the dielectric materials studied are listed in Table 4. Other material properties are given in Tables 1 and 2.
To summarize, Parylene C coated with a thin layer of Teflon® AF 1600 exhibits the highest variation of contact angle (~41°) which can sustain up to 500 V before dielectric breakdown. In contrast, the small initial contact angles (<50°) of BST and SiO2 due to their natural hydrophilic surfaces are not able to generate adequate electrowetting force for EWOD actuation. Hence, BST and SiO2 are not competitive enough as dielectric materials in EWOD. A moderate contact angle change (20°) is available with PDMS at a relatively high applied voltage of about 100 V. For Cytop™, the contact angle variation is as large as 30° under the same voltage and EWOD can always be operated at as low as 40 V before dielectric breakdown.
3.2 Dielectric breakdown in surrounding medium
Besides the dielectric layer, dielectric breakdown may also occur in the medium (air or silicon oil) surrounding the liquid. As reported, silicon oil immiscible with water has normally been employed to reduce contact angle hysteresis and prevent evaporation of liquid. A safety gap or channel height between the top and bottom substrate is essential to avoid the short circuiting of electrodes through the surrounding medium. For simplicity, we only consider a perfect model of parallel-plate capacitor of interfaces of three dielectric regions (liquid, medium and dielectric) parallel to the substrates with a negligible morphological influence of liquid and electrode on the electrical field.
To sum up, precautions to avoid dielectric breakdown effectively such as using thicker dielectrics and utilizing high-dielectric-constant materials, etc., should be taken into consideration in EWOD design. However, increasing the thickness of the dielectric layer requires a higher voltage to achieve the same effect, thus, consuming more power. A hydrophobic dielectric of high dielectric constant and dielectric strength is ideal for EWOD. Unfortunately, such materials are very limited. Certain trade-offs are needed to balance the requirement of the material properties and the performance of EWOD devices.
4 Contact angle measurement
This section describes the measurement of contact angles for several dielectric materials. The accuracy of the measurement depends on factors such as composition of liquid droplet, water evaporation, droplet size/volume, humidity and temperature. Other factors such as air pressure, relative velocity between air and liquid droplet are negligible.
Liquid evaporation is almost inevitable for droplets and it is undesirable for contact angle measurement. In microfluidics, liquid evaporation becomes significant for micro-droplets. Under typical experimental conditions (T = 20°C and relative humidity 40%), the evaporation rate of water is about 3 μg/s (King et al. 1997; Bonn et al. 2006). In addition, water evaporation decreases with increasing humidity increases rapidly with air temperature. Therefore, in this case, each contact angle measurement of a typical size of liquid droplet (3–5 μl) was completed within 2 min before water evaporation becomes significant at room temperature. The humidity in the lab is controlled at 40% and monitored during the experiment.
A non-transparent plastic cover (like a box) was made to cover the goniometer and to allow it to be operated only from the front. All illumination sources in the laboratory were turned off during the experiments in order to minimize the influence of illumination. Hence, the effect of illumination on contact angle measurement has been significantly reduced and thus considered to be negligible in this experiment.
4.1 Sample preparations
Only pure water and saturated sodium chloride (NaCl) were used as electrolytes for this study. Four-inch silicon wafers were firstly diced into 25 × 25 mm chips and then sputter-deposited with Au thin films (2500, with 100 thick Cr underlayer). Parylene C, PDMS, Cytop™ and self assembled monolayers (SAMs), which were able to serve as prime insulators without additional hydrophobic coatings, were prepared and the contact angles of electrolyte droplets on them were measured.
Parylene C thin films (2.25 μm thick) were deposited onto the Au-coated silicon chips by a customized parylene coating equipment.
PDMS (Sylgard® 184, Dow Corning) thin films were firstly prepared by mixing the base and curing agent at a ratio of 10:1 either by weight or volume (10 parts of base and 1 part of curing agent). Then, they were placed into a VWR vacuum oven to degas air bubbles at 101.6 kPa for 30 min. After that, PDMS was dispensed onto Au-coated silicon chips and spin-coated at 2,000 rpm, and finally cured at around 150°C for 1 h. Its thickness was measured by a surface profilometer to be about 38.2 μm.
Cytop™ (TL809M, Asahi Glass Co. Ltd) was deposited onto Au-coated silicon chips by spin-coating at 2,000 rpm with a ramp rate of 500 rpm. Then, they were cured at 60°C for 30 min. The thickness was about 950 nm.
A hexadecanethiol (C16H34S, from Fluka) SAM was formed by immersing the Au-coated silicon chips in 1 mM hexadecanethiol (15.28 μl) in ethanolic solution (50 ml) sealed in a glass beaker placed in a fume hood for 12 h. Then the monolayer was thoroughly rinsed with ethanol and its thickness was measured at about 24 by a variable angle spectroscopic ellipsometer (WVASE32, J. A. Woollam Co. Inc.).
4.2 Experimental setup
5 Results and discussions
5.1 Contact angle dependence on applied potentials
In the derivation of this theoretical model, it is assumed that the substrate is homogenous and that liquid evaporation and gravitational effects are negligible. A desirable contact angle modulation of 40° (from 128° to 88°) on Parylene C required about 120 V potential to be applied. The contact angle stabilized at ~88° on the 2.25 μm-thick Parylene C sample. A small contact angle modulation of 16° (117° to 101°) on a 38.2-μm thick PDMS sample was obtained under a potential of about 140 V without observation of contact angle saturation. A moderate potential of 65 V was applied to achieve the contact angle variation of 35° (from 110° to 75°) with Cytop. Contact angle saturation was observed on this 0.95 μm-thick Cytop sample when the potential was beyond 65 V.
A contact angle variation of 40° (from 105° to 65°) with SAMs could be obtained when applying a potential as small as 7.5 V, beyond which the contact angle saturated. However, it was observed that its hydrophobicity degraded significantly and the surface became hydrophilic in 1–2 days. The instability and contact angle irreversibility make SAMs impossible to be used in the device. Hence, SAMs as a prime insulator was not considered for the rest of the contact angle measurements. In addition, it should be noted that the theoretical predication given by Eq. 6 is not applicable for nanoscale effects of SAMs, as reported by Moon et al. (2002).
It can be seen that the experimental results follow very well the trend predicated by Eq. 6. Each data point in Fig. 6 is the average of three separate measurements and standard deviations of 1° for Cytop (with water) and 1.6° for Parylene C (with water) were obtained respectively. Contact angle hysteresis of 1° and 3.6° were observed for Cytop and Parylene C, respectively. It was found that saturation of the contact angle occurred with both of these materials, which was independent of the potential polarity. The change in contact angle for Cytop™ and Parylene C on both water and NaCl solution was found to be independent of the polarity of the applied voltage, which is in agreement with literature2 and theoretical prediction.
5.2 The effect of dielectric thickness on contact angle variation
The contact angle variation for Cytop is from 110° to 75° (35° modulation) and it is from 117° to 101° (16° modulation) for PDMS. The standard deviation in the experimental data is about 1.33 V for Cytop and 1.18 V for PDMS, respectively. The results show good agreement with the theoretical predication up to the point where the contact angle saturates. Using the spin coating method, it is difficult to make PDMS thickness thinner than 10 μm because the maximum spinning speed (about 5,000–6,000 rpm) has been reached. The thickness of Cytop achieved is above 500 nm, which is more than the minimum thickness of 330 nm required to avoid dielectric breakdown.
5.3 Contact angle reversibility
Comparing the tested dielectrics, Parylene C requires a high driving voltage to obtain a desirable contact angle modulation of 40° and patterning it using common microfabrication processes is difficult. PDMS presents a small contact angle variation of only 16° under an extremely high potential of 140 V. Microstructures of PDMS are quite difficult to be photo-defined by dry/wet processes such as RIE etching. In contrast, a desirable contact angle modulation of Cytop (35°) can be obtained under a relatively low applied potential (65 V). Cytop is also relatively easy o pattern using dry etching processes.
It has been shown by this study that the dielectric layer has the most dominant influence on EWOD performance compared to the other materials in the system. Based on the Young–Lippmann equation, the dielectric breakdown voltage and electrowetting actuation voltage as functions of dielectric thickness have been obtained for several typical dielectric materials. The minimum dielectric thickness was obtained from the intersection of these two functions. The contact angle measurements were made for droplets of water and NaCl solution on several dielectric materials. Comparisons with respect to contact angle tunability and reversibility upon electrowetting actuation were made. The effect of the thickness of the dielectric layer on contact angle variations was also studied. The results show that among the several dielectric materials studied, Cytop™ is the best dielectric material for EWOD with a good combination of mechanical, electrical, chemical and thermal properties. Cytop also exhibits good reversibility of contact angle and patternability with dry etching process. Its favorable surface properties make it suitable for commercial EWOD systems and devices. It has been successfully incorporated into a system for pumping liquids based on the EWOD effect, which has been internationally patented (Dharmatilleke and Liu 2006). The exploiting of powerful dielectric materials for EWOD will always be challenging for researchers to make breakthroughs in real applications.
This work was supported by Institute of Materials Research and Engineering (IMRE) and Agency for Science, Technology and Research (A*STAR).