Microfabrication of a digital microfluidic platform integrated with an on-chip electrochemical

cell

This article has been downloaded from IOPscience. Please scroll down to see the full text article.

2013 J. Micromech. Microeng. 23 095025

(http://iopscience.iop.org/0960-1317/23/9/095025)

Download details:

IP Address: 61.129.42.103

The article was downloaded on 29/08/2013 at 01:35

Please note that terms and conditions apply.

View the table of contents for this issue, or go to the journal homepage for more

Home Search Collections Journals About Contact us My IOPscience

IOP PUBLISHING JOURNAL OF MICROMECHANICS AND MICROENGINEERING

J. Micromech. Microeng. 23 (2013) 095025 (10pp) doi:10.1088/0960-1317/23/9/095025

Microfabrication of a digital microfluidicplatform integrated with an on-chipelectrochemical cellYuhua Yu1, Jianfeng Chen1, Jian Li1, Sheng Yang1, Shih-Kang Fan2

and Jia Zhou1

1 ASIC and System State Key Lab, Department of Microelectronics, Fudan University, Shanghai 200433,People’s Republic of China2 Department of Mechanical Engineering, National Taiwan University, Taipei, 10617, Taiwan

E-mail: jia.zhou@fudan.edu.cn

Received 6 March 2013, in final form 19 June 2013Published 28 August 2013Online at stacks.iop.org/JMM/23/095025

AbstractWe report on an IC compatible microfabrication process proposed for a novel monolithiclab-on-a-chip (LOC) with an electrochemical cell embedded in an electrowetting on dielectric(EWOD) digital microfluidic device. The optimized process focused on the surfacemodification of Teflon, selective exposure for the electrochemical module and recovery ofsurface properties by one-step annealing at low temperature. The optimum modification timeand annealing temperature were 20 s and 210 ◦C, respectively. The experimental results fromatomic force microscope and contact angle (CA) measurement revealed the effects of surfaceroughness and apparent CA on the wettability for different etch times. The multifunctionalityof droplet creation, merger and transportation in the EWOD microfluidic module and sensitiveelectrochemical detection for the redox probe were realized simultaneously. The proposedmicrofabrication process has many advantages of remarkable simplicity, prominentrepeatability, low cost and compatibility with standard IC processes. It shows great promise forthe microsystem of the microfluidic unit and detecting cell, and gives a brilliant conception forthe future fabrication of monolithic LOC integrated with functional detection.

(Some figures may appear in colour only in the online journal)

1. Introduction

Lab-on-a-chip (LOC), which is also known as the micrototal analysis system, can be defined as integrated microelectromechanical systems with biological and chemicaldetection that can carry out all stages of processesautomatically. In the past decades, the research on LOChas grown rapidly in the field of microfluidics becauseof its portability, compactness, parallelization and low cost(Fair 2007, Lidija et al 2010, Lim et al 2010, Dutse and Yusof2011, Jebrail et al 2012).

LOC is comprised of two basic modules denoted as‘sensors’ and ‘microfluidics’ in terms of functionality. As tosensing technologies used in LOC research, optical relateddetection is the most popular and successful one (Vazquezet al 2012, Ibarlucea et al 2013). However, a complex optical

system seems quite difficult to integrate into an LOC. Amongother developed sensing techniques, microelectrode-basedelectrochemical detection is highly promising for microsystemapplications due to its high performance detection, highintegration on chip, simplicity, scalability and low cost (Huanget al 2008, Hwang et al 2009).

For the microfluidic device, the microanalysis systembased on continuous microfluidics has been well developed.However, it faces great challenges in reliability, packageefficiency and integration with circuitry (Ho et al 2011,Seemann et al 2012). The digital microfluidics, based ondiscrete droplet operation, can fundamentally eliminate theseproblems by implementing a complete set of programmablefluidic operations with a finite array of elemental components.Among various droplet-actuation forces, electrowetting ondielectric (EWOD) has emerged as an active research area

0960-1317/13/095025+10$33.00 1 © 2013 IOP Publishing Ltd Printed in the UK & the USA

J. Micromech. Microeng. 23 (2013) 095025 Y Yu et al

in academia and industry, due to its advantages of speed,simplicity, compatibility with IC fabrication, reconfigurabilityand portability to perform on-chip chemical and biomedicalprotocols (Lidija et al 2010, Pollack et al 2011, Nelsona andKim 2012).

Since both the electrochemical detection and the EWODmicrofluidic device are based on similar components,i.e. electrodes, it implies that both modules have manycompatible fabrication processes and their incorporationmakes the integrated monolithic LOC possible. Up to now,some pioneers have developed preliminary work in thecombination of EWOD devices and detection technology.For example, Dubois et al 2006 reported on a powerfulapproach combining digital microfluidics and externaltwo-electrode electrochemical detection for solution-phasesynthesis. Additionally, Karuwan and coworkers (Karuwanet al 2011) reported on another EWOD microfluidic devicewith an external three-electrode electrochemical sensingsystem for quantitative analysis of iodide. However, bothreports were actually not integrated monolithic LOCs. Infact, there are few reports on efficient integration ofEWOD microfluidic techniques and on-chip electrochemicalanalytical systems (Lidija et al 2010).

The obstacles for realization of a fully-integrated LOCchip with these two modules can be addressed as follows.Firstly, the EWOD module must have a dielectric layerand a hydrophobic surface for droplet manipulation, whilethe electrochemical module needs a mirror-fresh electrode.Therefore, their layout, size and formation should be welldesigned when they are placed on the planar single chip(Chen et al 2012a). Secondly, in the crucial fabrication ofthe hydrophobic and hydrophilic surfaces simultaneously ona monolithic chip, it is difficult to deposit any films on thehydrophobic surface (e.g. fluoropolymers) which is essentialin the EWOD module (Liu et al 2010). Therefore, to wet thesurface of the hydrophobic surface for other deposition (e.g.photoresist for selective patterning) on it and to dewet it torecover the hydrophobic properties used for the electrowettingfunction, would be the key issues.

Teflon R© AF, a copolymer of polytetrafluoroethylene(PTFE) with 2,2-bis (trifluoroethylene)-4,5-difluoro-1,3-dioxole (Sabbatovskii et al 2004), is a family of amorphousfluoropolymers with similar properties of superior electricalproperties, good thermal stability and chemical resistance(Datta et al 2003). Furthermore, it is distinct from otherPTFEs by a large numbers of advantages, such as unmatchedoptical clarity, good creep resistance, high compressibilityand controlled mechanical properties (Du Pont 2013). TheTeflon R© AF coatings or films can be prepared on varioussubstrates (like metal, SiO2 and glass) by spin-coating(Fair 2007) or plasma-induced deposition (Bayiati et al2007). All these features have made Teflon R© AF a verypromising material in the electronics, microelectronics, opticsand mechanical industries. Especially, with the development ofEWOD digital microfluidics, Teflon R©AF has been widely usedin EWOD devices to provide hydrophobic surfaces due to itslow surface energy (Bayiati et al 2007, Paneru et al 2010, Leeet al 2011, Fan et al 2011, Witters et al 2011, Chen et al 2012b,

Kirby and Wheeler 2013, Huang et al 2013). On the other hand,although Teflon has a low dielectric constant, it is found to besufficient to act both as the insulator and hydrophobic layerfor electrowetting when thicker coatings are carefully preparedwithout pinholes by appropriate temperature treatment (Seyratand Hayes 2001). Some works have been reported on howto increase the surface energy of fluoropolymer—one of thesimplest and developed methods is plasma-induced surfacemodification (Park et al 2002, Inagaki et al 2002). Besidesthis, plasma can be adopted to dry etching (reactive ion etch,RIE) for patterning Teflon in this work and the dewetting ofTeflon can also be achieved by thermal annealing (Makohlisoet al 1998). Therefore, using low-temperature processes tofabricate such a monolithic LOC compatible with pre-CMOSIC process is worth studying as well.

In this paper, we report on an IC compatiblemicrofabrication process of a novel monolithic LOC integratedwith an electrochemical cell and an EWOD microfluidicplatform. This work focuses on the critical processes: surfacewetting of Teflon, selective patterning of the electrochemicalcell and dewetting by low temperature annealing. The surfaceroughness and contact angle (CA) of Teflon dealing withdifferent times and temperatures were investigated, as wellas the potential mechanism of the apparent CA changingwith etch time. The proposed microfabrication processwas simple, repeatable, inexpensive and compatible. As aresult, a multifunctional monolithic LOC was fabricated anddemonstrated by on-chip droplet manipulation and rapidelectrochemical detection for a wide concentration range ofredox probe.

2. Material and methods

Teflon R© AF1601solution (Du Pont, Grade 601S1-100-6,6 wt%) was used as both the hydrophobic and the dielectricmaterial in the EWOD device. Ferrocenemethanol (FCM, as theredox probe in the electrochemical detection) and phosphatebuffered saline (PBS, 0.1 M, pH = 7.0) were provided bySigma-Aldrich. All the chemicals were of analytical reagentgrade. All experiments were carried out at room temperature.

Physical vapour deposition (PVD, ASC-4000-C4 Type L,ULVAC, Japan) was used to deposit Au/Ti metal layers forthe microelectrodes. RIE-10NR (Samco International, Japan)was carried out to selectively etch and modify the surfaceproperties of Teflon. CA measurements were estimated withthe aid of a video-based optical CA measuring instrument(OCA 20, Data Physics Instruments GmbH) under ambientconditions. The probe liquid was deionized water. Using aTeflon blunt-ended needle attached to a syringe, a 2 μL waterdroplet was deposited on the surface and then the CA wasmeasured immediately. Typically, CAs were measured threetimes on a 1 cm × 1 cm wafer piece. The average CA of eachsample used in this experiment was measured at five randomlyselected points. Atomic force microscope (AFM, BrukerMultiMode R© 8) images were obtained in the tapping modeto characterize the surface morphology and roughness of thesamples. The cyclic voltammetric experiment was performedby a CHI1030 electrochemical workstation (CH Instruments

2

J. Micromech. Microeng. 23 (2013) 095025 Y Yu et al

Glass substrate

V

Driving electrode(EWOD electrode)

ITO Top plate

Teflon layers Droplet

Electrochemical electrodes

Substrate

Driving electrode(EWOD electrode)

DropletV

To top plate

(a)

(b)

(c)

Figure 1. Two-plate configuration of EWOD microfluidic chip integrated with electrochemical cell arrangement. (a) Top view;(b) cross-section view; (c) details of interdigital microelectrodes (including WE, CE and RE) embedded in an EWOD electrode on thebottom plate.

Inc., USA) in the integrated three-microelectrode-systemelectrochemical cell, which consists of gold interdigitatedmicroelectrodes with three fingers as working electrodes(WEs), four fingers as counter electrodes (CEs) and arectangular gold layer as reference electrode (RE).

2.1. Structure of monolithic LOC

Figure 1 shows the schematic diagram of a well-designedmonolithic LOC platform. The common ITO electrode onthe top plate is grounded. The bottom plate used as theworking area consists of an EWOD microfluidic device andelectrochemical microelectrodes. The size of a single squareEWOD electrode and electrochemical cell is 1 mm × 1 mmand 480 μm × 540 μm, respectively. The hydrophilicelectrochemical cell was embedded in the hydrophobic EWODelectrode with an area ratio of about 26%. The electrochemicalcell is comprised of interdigitated fingers (360 μm × 50 μm)

as WE or CE and a rectangular RE (340 μm × 50 μm).Herein, it is worth mentioning that gold material, instead ofconventional Ag/AgCl, was used as the RE in order to simplifythe fabrication process. This will prove to be practical forelectrochemical detection in section 4.

2.2. General fabrication procedure of monolithic LOC

The optimized fabrication process flow of the monolithicLOC is illustrated in figure 2. Firstly, a Ti layer and an Aulayer was successively sputtered on the pre-cleaned glasssubstrate by PVD (figure 2(a)). The thickness of the Ti/Aulayers was about 300 nm. Secondly, both the electrodes ofthe EWOD microfluidic module and the electrochemical cellwere patterned by one-step standard lithography and chemicaletching (figure 2(b)). About 1.33 μm thickness of Teflon R©filmwas spin-coated on the substrate and cured (figure 2(c)).The Teflon was then treated by RIE duration for 20 s with

3

J. Micromech. Microeng. 23 (2013) 095025 Y Yu et al

Lithography and RIE forselective etching

(a)

(c)

(e) ( f )

(d)

(b)Ti/Au deposition Patterning electrodes

Spin-coating of Teflon Wetting of Teflon by RIE

Removing resist and annealing

Photoresist GlassTeflonGold

Figure 2. Fabrication process of electrochemical sensing integrated digital microfluidic chip.

100 W power at an argon flow rate of 30 sccm to wetthe surface for subsequent photoresist coating (figure 2(d)).Figure 2(e) shows the selective etching by RIE for at least120 s to expose the hydrophilic surface of the electrochemicalcell. Finally, we immersed the device in acetone to removeresidual photoresist, followed by low-temperature annealing ata constant temperature (210 ◦C) for 1 h to recover the surfaceproperties of Teflon (figure 2( f )). For the critical process ofwetting and dewetting, the effects of different modificationtimes, selective etch rate and annealing temperatures werestudied and optimized below.

2.3. Surface wetting of Teflon

Before wetting, each sample was prepared by coating Teflonon the glass substrate and cured. Since the effect of sputteringtime on the surface topography and wettability of Teflon isessential for patterning and then selective exposure of theelectrochemical cell, several groups by RIE with 100 W powerat 30 sccm argon flow for different sputtering time (0–100 s)were experimented with. The average CA of each group wasobtained from random five points. Each CA was measuredthree times. The potential mechanism for the change of CAafter different etching times was discussed.

2.4. Dewetting of modified surface

The dewetting of RIE-modified Teflon was studied tocharacterize the recovery process by the CA measurement andAFM. Ten groups of different constant annealing temperatures(85–330 ◦C) for individual hours were carried out to study thesurface recuperation of pre-modified Teflon. Before annealing,all the Teflon samples were treated by RIE for 20 s. Asa comparison, considering the standard temperature process(Du Pont 2013), another ten groups of Teflon films were

also tested. They were pre-modified for different sputteringtimes (0–100 s) and dewetted by three-standard-temperaturetechnology (denoted as annealing at 110 ◦C for 5 min, then245 ◦C for 5 min and finally at 330 ◦C for 10 min).

3. Results and discussion

3.1. Surface characterization of wetting Teflon

Figure 3(a) gives the typical curve of the CA and RIE timefor wetting Teflon. Compared with the initial CA of untreatedTeflon (about 122◦), the apparent CA decreased by at least 10◦

during the first 5 s. Then it recovered to around 122◦ within thenext 5 s and gradually increased and finally remained about130◦ after 20 s. Besides, we also tried to test CA hysteresisfor the all the modified Teflon coatings (insets of figure 3(a)).As a result, for all of them, the water droplets would notslide even when we turn the samples for 180◦. This indicatesthat the modified-Teflon coatings treated by RIE possess goodinterfacial adhesive strength. Especially for 20 s (and over)-treated Teflon (inset (c)), when the surface is either facing upor turned upside down, water droplets on the surface alwaysmaintain the spherical shape, implying a superior interfacialadhesive strength with large apparent CAs.

Figure 3(b) shows that the etch rate of Teflon under theconditions used in this work is about 0.011 μm s−1. Since thesurface of coated Teflon before modification was not so smoothdue to the inherently high viscosity, enough etching time (atleast 140 s) should be guaranteed to remove the maximumthickness of Teflon coatings. For example, if we have firstlymodified Teflon for 20 s (which would reduce the thickness toabout 1.11 μm) before the selective etching process, the nextetch time of 120 s is sufficient to remove the modified-Tefloncoatings covered on the electrochemical cell.

4

J. Micromech. Microeng. 23 (2013) 095025 Y Yu et al

0 10 20 30 40 50 60 70 80 90 100

95

100

105

110

115

120

125

130

135

140

y=-50.03exp(-x/5.00)+132.04, R2=0.9819

Con

tact

Ang

le(d

eg.)

RIE time (s)

Initial CA of Untreated Teflon

(a) (b) (c)

0 20 40 60 80 100 1200.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

Thi

ckne

ss( µ

m)

RIE time (s)

Thickness[µm]=-0.011Time[s]+1.324

R2 =0.9960

(a)

(b)

Figure 3. (a) Relationship between CA and RIE time after wettingof cured Teflon. The insets of (a)–(c) show the typical shapes ofwater droplets on the 5 s-treated, 10 s-treated and 20 s (andover)-treated Teflon surface, respectively. (b) Etch rate of Tefloncoatings under RIE with 100 W power at 30 sccm argon flow.

Figure 4 shows the AFM images of modified Teflonfor different sputtering times, corresponding to table 1. Thenanorod-like surface can be observed during the first 20 s(figures 4(a)–(c)) and gradually change to large-scale lumpedcones (figures 4(d), (e)) with longer time. Since surface

Table 1. Roughness of Teflon for different modification times.

Group RIE time (s) Rq (nm)

A 0 0.540 ± 0.030B 5 1.40 ± 0.08C 20 2.93 ± 0.35D 60 6.60 ± 0.40E 80 7.67 ± 0.62

roughness and heterogeneity are known to affect the apparentCA and CA hysteresis (Saeki et al 2001, McHale et al 2004), itindicates that RIE treatment changes both the surface chemicalbonds and surface morphology of Teflon.

Herein, we would like to discuss the possible mechanismof the change of CA with sputtering time for the modifiedTeflon. From figures 3, 4 and table 1, the different apparent CAsafter RIE treatment were attributed to the complex effects of thesurface defluorination, cross-linking reaction (Sabbatovskiiet al 2004, Tasker et al 1994) and the transition of the partiallyimpregnating Cassie state (Cassie and Baxter 1944, Ran et al2008, Whyman and Bormashenko 2011, Zhou et al 2012).The first two were induced by the radical species, while thelast was generated and controlled by surface structure. In thefirst 5 s, the hydrophobic bonds on the Teflon surface werebroken and unsaturated fluorinated chains including danglingbonds, –CF = CF2 and other active groups (Sabbatovskiiet al 2004) were generated and tended to arise the cross-linking reaction within them at the same time. Thus Teflonshows better hydrophilicity. Meanwhile, since the root-mean-square of roughness (Rq) is less than 1.4 nm, there is notenough time to form the hydrophobic nanorod on the surfaceof Teflon. Therefore, the apparent CAs decreased mainly dueto the initial defluorination. When the samples were treatedfor next 15 s, the surface roughness gradually increased,and the effect of initial defluorination and the cross-linkingreaction tended to become balanced, as well as limited by thesaturated or near-saturated number of radical species on thesurface. As a result, the apparent CAs gradually recovered to122◦ due to the common transition of the surface wettabilityto the partially impregnating Cassie state (Koishi et al 2009).The degree of impregnating Cassie state depends on the surfaceroughness, i.e. the apparent CA increased with longer RIE timein this stage. As the time went on over 20 s, the modifiedTeflon had more complex surfaces with additional large-scale roughness and further decomposition of chemical bonds

(a) (b) (c) (d) (e)

Figure 4. AFM images of cured-Teflon samples after RIE for (a) 0 s, (b) 5 s, (c) 20 s, (d) 60 s and (e) 80 s. The display range is(a) −1.9–2.0 nm, (b) −4.4–5.8 nm, (c) −9.2–11.7 nm, (d) −21.8–23.6 nm and (e) −26.0–28.1 nm, respectively. The scanning area is2 μm × 2 μm.

5

J. Micromech. Microeng. 23 (2013) 095025 Y Yu et al

Table 2. Relationship of apparent CA and surface roughness withannealing temperature.

Group Annealing temp. (◦C) Contact angle (◦) Rq (nm)

1 85 133 ± 0.60 3.15 ± 0.432 110 133 ± 1.06 3.09 ± 0.303 135 115 ± 1.20 0.719 ± 0.0484 160 114 ± 0.85 0.706 ± 0.0275 185 115 ± 0.72 0.708 ± 0.0256 210 123 ± 0.35 0.613 ± 0.0307 245 122 ± 0.45 0.530 ± 0.0388 270 123 ± 0.60 0.547 ± 0.0349 295 124 ± 0.85 0.558 ± 0.020

10 330 123 ± 0.28 0.549 ± 0.035

(denoted as secondary defluorination), but yet a large apparentCA (about 131◦) as well as superior interfacial adhesivestrength (figure 3(a)). This might be because the surface ofthe partially Cassie state further transforms to another unusualtransition state, i.e. the so-called gecko-like state (Wang andJiang 2007) or petal-like effect (Feng et al 2008). In thegecko-like state, some air was sealed in the cavities createdby surface roughness, which is different from the Cassie state(Cassie and Baxter 1944) where the trapped air pockets areconnected to the atmosphere (open state). For a static waterdroplet on the surface, the trapped air will induce a largeCA; while the adhesive forces between the water droplet andthe surface arise from Van der Waals interactions, similarto the Cassie state. However, once the droplet was drawn(like the dynamic process during high-speed spin-coatingof photoresist), the volume of the sealed air in the plasma-induced cavities would change and thus produced negativepressure. This may create a strong ‘adhesive’ force. Therefore,besides large apparent CAs, 20 s (or over)-treated Teflonpossessed superior and stable interfacial adhesion inducedby the synthetic effects of secondary defluorination and thegecko-like state.

As we all know, compared with apparent CA, CAhysteresis is a more important parameter to characterize thewettability of structured surfaces. Larger CA hysteresis meansbetter interfacial adhesion. During the experiments, we triedto spin-coat photoresist on 3–60 s-treated Teflon coatings. It isfound that the chip-to-chip repeatability of uniform photoresistfilm was worse on the 3–10 s-treated Teflon than on 20–60 s-treated ones. This indicates that the surfaces of 3–10 s-treatedTeflon were unstable or metastable after short-time etching.

From the above, considering more stable and superiorinterfacial adhesion strength, the optimized wetting time of20 s was appropriate and sufficient to use in the work. Thus,in our later experiments, 120 s was used as the etch time toexpose the Teflon which covered the electrochemical cell.

3.2. Dewetting of modified Teflon

Before annealing, all the Teflon samples were treated for 20 sand the average CA and roughness after RIE were about133◦ and 3 nm, respectively. The annealing time was 1 hfor all samples. Table 2 shows the relationship of CA andsurface roughness with annealing temperature in the rangefrom 85 to 330 ◦C. Figure 5 gives the typical AFM images

for different annealing temperatures. Obviously, the extent ofrecovery in both CA and roughness was different for differenttemperatures. At a relatively low temperature (85–110 ◦C)which is much lower than the glass temperature of Teflon,both the apparent CAs and surface roughness (figure 5(a)) werealmost unchanged. This is because the modified-Teflon did notsoften and thus its hydrophobicity did not restore (figure 5(a)).When treated at an increased temperature (135–185 ◦C), thesurface roughness of the samples decreased to about 0.7 nm,indicating that most of the large-scale roughness hydrophobicstructures fade away (figures 5(b), (c)); however, there mightstill be numbers of radical species existing on the surface.As a result, the CAs decreased to around 115◦, which wassimilar to the results of inset (a) of figures 3(a) and 4(b).For higher annealing temperatures (210–330 ◦C), both theapparent CAs and surface roughness recovered to the initialvalue just like that of the unmodified Teflon, mainly due tothe rebuilt hydrophobic chemical bonds and smooth surfacestructure (figures 5(d)–( f )). This indicates that the procedureof wetting by RIE and dewetting by appropriate annealing isa quasi-reversible process.

In addition, considering the standard temperature process(Du Pont 2013), another ten groups of Teflon films pre-modified for different sputtering times (0–100 s) and thenannealed by three-standard-temperature technology werealso experimented. Consequently, no matter how long theRIE time was, the apparent CAs and surface roughnesscould always recover to the original state by using three-standard-temperature technology, which is consistent with thereference (Makohliso et al 1998). Meanwhile, the surfaceroughness after annealing by three-standard-temperaturetechnology was about 0.55 nm, which was similar to those infigures 5(d)–( f ). Such results indicate that recovering thesurface properties of modified Teflon at the temperaturerange of 210–330 ◦C is effective and feasible, comparable tothe effects of three-standard-temperature technology. Amongthis wide and effective temperature range, a relatively lowtemperature (i.e. 210 ◦C) is obviously sufficient to recover thesurface properties. Therefore, annealing at 210 ◦C for 1 h canbe selected as the optimized annealing approaches to recoverthe properties of modified-Teflon.

4. Functional application

Figure 6 is the monolithic LOC with integration of the EWODmicrofluidic module and the electrochemical cell preparedby the above optimized fabrication process. The platformconsists of 19 EWOD electrodes with four reservoirs sandthree electrochemical electrodes (WE, CE and RE). The insetof figure 6 is the optical microscope of embedded interdigitatedfingers as the three-microelectrode-system electrochemicalcell.

Before verifying the function of the EWOD microfluidicdevice, the common electrode on the top plate was connectedto the ground potential. The reservoir and driving electrodeson the bottom plate were connected to the common terminalsof single pole double throw relays. The potential of theelectrodes was switched between high potential and zero

6

J. Micromech. Microeng. 23 (2013) 095025 Y Yu et al

(a) (b)

(c) (d)

(e) ( f )

Figure 5. Typical AFM images of 20 s-treated Teflon samples after annealing at (a) 85 ◦C, (b) 135 ◦C, (c) 185 ◦C, (d) 210 ◦C, (e) 270 ◦C and( f ) 330 ◦C for 1 h, respectively. The display range is −14.0–14.0 nm (a) and −6.0–6.0 nm (b)–( f ), respectively. The scanning area is2 μm × 2 μm.

Figure 6. Picture of EWOD device embedded with theelectrochemical module.

potential through the relays. To get a high potential signal,a 1 kHz sine ac signal was generated from the functiongenerator and then amplified to about 60 Vrms through anamplifier. The digital output signals of a data acquisition deviceprogrammed by the LABVIEW software were used to switch

the relays to actuate the EWOD electrodes. Figure 7 showsthe whole EWOD microfluidic procedure with embeddedelectrochemical cell. Firstly, two droplets were pulled out ofthe top and bottom reservoir, respectively (figure 7(a)). Aftermixing them up in the crossed electrodes (figure 7(b)), thelarger droplet was transported to the electrochemical sensingcell (figures 7(c), (d)) for detection and finally transportedto the waste reservoir (figure 7(e)). After detecting for eachexperiment the chip, and especially the electrochemical cell,was thoroughly washed by moving droplets from the cleanreservoir to the waste reservoir (figures 7( f )–(h)) several times.

FcM/FcM+ redox couple is close to an idealquasi-reversible system in electrochemical detection. Inthe electrochemical sensing detection, the typical cyclicvoltammetric responses of targeted droplet (about 0.4 μL)containing 0.1 M PBS solution (pH = 7.0) and 2.0 mM FcMwere obtained at different scan rates in figure 8(a). Obviously,the redox reaction of the FcM probe gives roughly symmetricanodic and cathodic peaks at relatively slow scan rates. Whenthe scan rate increases, the redox potentials of FcM shiftslightly and the peak-to-peak potential (�Ep) also increasesranging from 78 mV s−1 to 159 mV s−1. The redox peak currentincreases linearly with v

12 (inset of figure 8(a)), as shown in

following equations (1) and (2):

7

J. Micromech. Microeng. 23 (2013) 095025 Y Yu et al

(a) (b)

(c) (d)

(e) ( f )

(g) (h)

Figure 7. EWOD microfluidic function including two parts: (1) merging the droplet: pulling two droplets out (a), mixing (b), transport tothe electrochemical cell for electrochemical detection (c), (d) and finally moving to the waste reservoir (e); (2) cleaning the chip: repeat thestep of transporting a droplet from the clean reservoir to waste reservoir ( f )–(h) several times. All the droplets are labelled by the yellowarrows for better visual effect. The dotted arrow (red) shows the movement direction of the droplet during different phases.

Ipa(μA) = 0.108 + 0.056√

v, R2 = 0.9925 (1)

Ipc(μA) = −0.110 − 0.041√

v, R2 = 0.9978. (2)

This demonstrates that the redox reaction of FcM atthe working microelectrode is a quasi-reversible diffusion-controlled electrochemical process at slow scan rates.

8

J. Micromech. Microeng. 23 (2013) 095025 Y Yu et al

0.4 0.3 0.2 0.1 0.0 -0.1 -0.2 -0.3-0.8

-0.6

-0.4

-0.2

0.0

0.2

0.4

0 2 4 6 8 10-0.6

-0.4

-0.2

0.0

0.2

0.4

0.6

0.8

Cur

rent

(µA

)

Potential (V)

100mV/s75mV/s50mV/s25mV/s15mV/s10mV/s

(a)

(b)

Ipa

Ipc

Peak

Cu r

ren t

(µA

)

Scan Rate1/2(mV/s)1/2

0 200 400 600 800 1000

0.0

0.2

0.4

0.6

0.8

1.0

0 5 10 15 20 250.04

0.06

0.08

0.10

0.12

0.14

-0.4 -0.2 0.0 0.2 0.4

0.0

0.2

0.4

0.6

0.8

1.0

Cur

rent

(µA

)

Concentration (µM)

Cur

rent

(µA

)

Concentration (µM)

Cu r

rent

(µA

)

Potential (V)

0

1 mM

Figure 8. (a) C–V curves of 2.0 mM FcM at different scan rates in0.1 M PBS solution (pH = 7.0) for electrochemical detection. Inset:the linear fitting of redox current and the square root of scan rates.The scan rate is 10, 15, 25, 50, 75 and 100 mV s−1. (b) Calibrationplot of peak current versus FcM concentrations (from 3.0 μM to1.0 mM). Top inset is the plot of peak current and FcMconcentration in the lower concentration region (3.0–25.0 μM).Bottom inset shows the experimental DPVs of 0, 3, 5, 10, 25 μMand 0.05, 0.1, 0.25, 0.5, 2/3, 1.0 mM FcM at the WE in the PBSsolution (0.1 M, pH = 7.0). The volume of droplets is about 0.4 μL.

Since differential pulse voltammetry (DPV) has a muchhigher current sensitivity and better resolution than cyclicvoltammetry, it was used in the determination of FcMconcentrations and the estimation of the low limit of detection.Figure 8(b) shows the experimental curves and relevantcalibration curves for a wide concentration range of FcM inthe range from 3.0 μM to 1.0 mM. From the DPV curves(see the bottom inset), it can be found that the peak currentincreased with increasing concentration of FcM. In addition,the calibration plot at relative low concentration (3.0–25.0 μM)is obviously steeper than that at high concentration (from50.0 μM to 1.0 mM). This indicates that the electrochemicalsystem is more sensitive to FcM at lower concentrations. Thetop inset of figure 8(b) shows that the calibration curve is linearover the range from 3.0 to 25.0 μM FcM with the sensitivity of

about 3.31 nA μM−1 (R2 = 0.9993) and the estimated detectionlimit of about 0.53 μM (with a signal/noise ratio of 3). For thehigher concentrations (50.0 μM–1.0 mM), the DPV currentsare proportional to the FcM concentrations with a sensitivityof 0.73 nA μM−1 (R2 = 0.9897).

The stability and reproducibility of the electrochemicalmodule were also investigated. The current responses for fiveindependent measurements towards 2.0 mM FcM remainedalmost constant and the relative standard deviation (RSD)was less than 2.5%. The chip-to-chip reproducibility wasdetermined in the presence of 2.0 mM FcM with five differentelectrodes, which yielded a RSD less than 6.7%. Whenstored at ambient environment for one day, the current signalsdecreased by 2.8%. After 7 days, 94.9% of the initial currentwas obtained. These data indicate the good stability andreliable reproducibility of the proposed fabrication for theelectrochemical module. We also found that although allthree electrodes (WE, RE and CE) are formed from the samematerial (gold), such a system did not result in a significantdrift in potential and current in this experiment.

5. Conclusions

This study demonstrated a model system for themicrofabrication of a novel monolithic LOC integrated withan electrochemical microcell and the EWOD microfluidicdevice. In order to satisfy the multifunction of electrochemicalmicroelectrodes and EWOD microfluidic, the crucialprocesses of surface wetting of Teflon, selective etching by RIEand one-step annealing at low temperature were investigated.The optimum modification time and annealing temperaturewere 20 s and 210 ◦C, respectively. The experimental resultsfrom AFM and CA measurement revealed the effects ofsurface roughness and apparent CA on the wettability. Thereported microfabrication process presented many advantagesof remarkable simplicity, prominent repeatability, low costand compatibility with standard IC processes. Based onthe optimized process, the proposed monolithic LOC couldnot only create, merge and transport droplets, but show asensitive, rapid and wide-range detection for FcM probe. Theelectrochemical detection of FcM showed two linear ranges.The sensor displayed a linear range of 3.0–25.0 μM FcMwith a sensitivity of 3.31 nA μM−1 (R2 = 0.9993) and alinear range of 50.0 μM–1.0 mM FcM with a sensitivity of0.73 nA μM−1 (R2 = 0.9897). The detection limit was 0.53 μMwith a signal/noise ratio of 3. The proposed process showsgreat promise for the integration of the LOC with IC chips inthe future. It deserves further study for practical bioanalysisapplication in our next work.

Acknowledgment

This work was supported by the National Science Foundationof China with Grant No 61176110.

9

J. Micromech. Microeng. 23 (2013) 095025 Y Yu et al

References

Bayiati P, Tserepi A, Petrou P S, Kakabakos S E, Misiakos Kand Gogolides E 2007 Electrowetting on plasma-depositedfluorocarbon hydrophobic films for biofluid transport inmicrofluidics J. Appl. Phys. 101 103306

Cassie A B D and Baxter S 1944 Wettability of porous surfaceTrans. Faraday Soc. 40 546–51

Chen J F, Yu Y H, Li J and Zhou J 2012a Size-variable dropletactuation by interdigitated electrowetting electrode Appl. Phys.Lett. 101 234102

Chen J F, Yu Y H, Yang S, Fan S K and Zhou J 2012b A monolithiclab-on-a-chip for electrochemical detection ICSICT: IEEE 11thInt. Conf. on Solid-State and Integrated Circuit Technologypp 1–3

Datta A et al 2003 Microfabrication and characterization of TeflonAF-coated liquid core waveguide channels in silicon IEEESensors J. 3 788–95

Dubois P, Marchand G, Fouillet Y, Berthier J, Douki T, Hassine F,Gmouh S and Vaultier M 2006 Ionic liquid droplet ase-microreactor Anal. Chem. 78 4909–17

Du Pont 2013 Teflon AF Product InformationDutse S W and Yusof N A 2011 Microfluidics-based lab-on-chip

systems in DNA-based biosensing: an overview Sensors11 5754–68

Fair R B 2007 Digital microfluidics: is a true lab-on-a-chip possibleMicrofluid. Nanofluid. 3 245–81

Fan S K, Hsu Y W and Chen C H 2011 Encapsulated droplets withmetered and removable oil shells by electrowetting anddielectrophoresis Lab Chip 11 2500–8

Feng L, Zhang Y N, Xi J M, Zhu Y, Wang Nu, Xia F and Jiang L2008 Petal effect: a superhydrophobic state with high adhesiveforce Langmuir 24 4114–9

Ho T Y, Krishnendu C and Paul P 2011 Digital microfluidicbiochips: Recent research and emerging challengesCODES+ISSS: Proc. 9th Int. Conf. on Hardware/SoftwareCodesign and System Synthesis pp 335–43

Huang H H, Zhou J, Huang Y P and Kong J L 2008 Impedimetricimmunosensor with on-chip integrated electrodes forhigh-throughput screening of liver fibrosis markers J. Anal.Chem. 63 492–8

Huang L X, Koo B and Kim C J 2013 Sputtered-anodized Ta2O5 asthe dielectric layer for electrowetting-on-dielectricmicroelectromechanical systems J. Microelectromech. Syst.22 253–5

Hwang S, LaFratta C N, Agarwal V, Yu X, Walt D Rand Sonkusale S 2009 CMOS microelectrode array forelectrochemical lab-on-a-chip applications IEEE SensorsJ. 9 609–15

Ibarlucea B et al 2013 PDMS based photonic lab-on-a-chip for theselective optical detection of heavy metal ions Analyst138 839–44

Inagaki N, Narushima K and Yamamoto T 2002 Surfacemodification of tetrafluoroethylene-perfluoroalkyl vinylethercopolymer (PFA) by plasmas for copper metallization J. Appl.Polym. Sci. 85 1087–97

Jebrail M J, Bartsch M S and Patel K D 2012 Digital microfluidics:a versatile tool for applications in chemistry, biology andmedicine Lab Chip 14 2452–63

Karuwan C, Sukthang K, Wisitsoraat A, Phokharatkul D,Patthanasettakul V, Wechsatol W and Tuantranont A 2011Electrochemical detection on electrowetting-on-dielectricdigital microfluidic chip Talanta 84 1384–9

Kirby A E and Wheeler A R 2013 Microfluidic origami: a newdevice format for in-line reaction monitoring bynanoelectrospray ionization mass spectrometry Lab Chip13 2533–40

Koishi T, Yasuoka K, Fujikawa S, Ebisuzaki T and Zeng X C 2009Coexistence and transition between Cassie and Wenzel state on

pillared hydrophobic surface Proc. Natl Acad. Sci.106 8435–40

Lee J K, Park K W, Kim H R and Kong S H 2011 Dielectricallystabilized electrowetting on AF1600/Si3N4/TiO2

dielectric composite film Sensors Actuators B 160 1593–8Lidija M, Daniel B, Teodor V and Maryam T 2010 Integration and

detection of biochemical assays in digital microfluidic LOCdevices Lab Chip 10 418–31

Lim Y C, Kouzani A Z and Duan W 2010 Lab-on-a-chip: acomponent view Microsyst. Technol. 16 1995–2015

Liu H, Dharmatilleke S, Maurya D K and Tay A O 2010 Dielectricmaterials for electrowetting-on-dielectric actuation Microsyst.Technol. 16 449–60

Makohliso S A, Giovangrandi L, Leonard D, Mathieu H J,Ilegems M and Aebischer P 1998 Application of Teflon-AF R©

thin films for bio-patterning of neural cell adhesion BiosensorsBioelectron. 13 1227–35

McHale G, Shirtcliffe N J and Newton M I 2004 Contact-anglehysteresis on super-hydrophobic surfaces Langmuir20 10146–9

Nelsona W C and Kim C J 2012 Droplet actuation byelectrowetting-on-dielectric (EWOD): a review J. Adhes. Sci.Technol. 26 12–17

Paneru M, Priest C, Sedev R and Ralston J 2010 Static and dynamicelectrowetting of an ionic liquid in a solid/liquid/liquid systemJ. Am. Chem. Soc. 132 8301–8

Park Y W, Tasaka S and Inagaki N 2002 Surface modification oftetrafluoroethylene-hexafluoropropylene (FEP) copolymer byremote H2, N2, O2, and Ar plasmas J. Appl. Polym. Sci.83 1258–67

Pollack M G, Pamula V K, Srinivasan V and Eckhardt A E 2011Applications of electrowetting-based digitalmicrofluidics in clinical diagnostics Expert Rev. Mol. Diagn.11 393–407

Ran C B, Ding G Q, Liu W A, Deng Y and Hou W T 2008 Wettingon nanoporous alumina surface transition between Wenzel andCassie states controlled by surface structure Langmuir24 9952–5

Sabbatovskii K G, Dutschk V, Nitschke M, Simon F and Grundke K2004 Properties of the Teflon AF1601S surface treated withlow pressure argon plasma Colloid J. 66 208–15

Saeki F, Baum J, Moon H, Yoon J Y, Kim C J and Garrell R L 2001Electrowetting on dielectrics (EWOD): reducing voltagerequirements for microfluidics Polym. Mater. Sci. Eng.85 12–13

Seemann R, Brinkmann M, Pfoh T and Herminghaus S 2012Droplet based microfluidics Rep. Prog. Phys. 75 016601

Seyrat E and Hayes R A 2001 Amorphous fluoropolymers asinsulators for reversible low-voltage electrowetting J. Appl.Phys. 90 1383

Tasker S, Chambers R D and Badyal J P S 1994 Surfacedefluorination of PTFE by sodium atoms J. Phys. Chem.98 12442–6

Vazquez R M, Cerullo G, Ramponi R and Osellame R 2012Optofluidic biochips Femtosecond Laser Micromachiningvol 123 ed R Osellame et al (Heidelberg: Springer)pp 389–419

Wang S T and Jiang L 2007 Definition of superhydrophobic statesAdv. Mater. 19 3423–4

Whyman G and Bormashenko E 2011 How to make the Cassiewetting state stable? Langmuir 27 8171–6

Witters D, Vergauwe N, Vermeir S, Ceyssens F, Liekens S, Puersb Rand Lammertyn J 2011 Biofunctionalization ofelectrowetting-on-dielectric digital microfluidic chipsfor miniaturized cell-based applications Lab Chip11 2790–4

Zhou J, Yang S, Zeng X Y, Wu J H, Chen G P and Huang Y P 2012Superhydrophobic ZnO for EWOD digital microfluidic devicefor application in micro total analysis system (μ-TAS)J. Adhes. Sci. Technol. 26 2087–98

10