Diagnostics based on microfluidic devices necessitate specific reagents, flow conditions, and kinetics for optimal performance. Such an optimization is often achieved using assay-specific microfluidic chip designs or systems with external liquid pumps. Here, we present “electrogates” for stop-and-go control of flow of liquids in capillary-driven microfluidic chips by combining liquid pinning and electrowetting. Electrogates are simple to fabricate and efficient: a sample pipetted to a microfluidic chip flows autonomously in 15-μm-deep hydrophilic channels until the liquid meniscus is pinned at the edge of a 1.5-μm-deep trench patterned at the bottom of a rectangular microchannel. The flow can then be resumed by applying a DC voltage between the liquid and the trench via integrated electrodes. Using a trench geometry with a semicircular shape, we show that retention times longer than 30 min are achieved for various aqueous solutions such as biological buffers, artificial urine, and human serum. We studied the activation voltage and activation delay of electrogates using a chip architecture having 6 independent flow paths and experimentally showed that the flow can be resumed in less than 1 s for voltages smaller than 10 V, making this technique compatible with low-power and portable microfluidic systems. Electrogates therefore can make capillary-driven microfluidic chips very versatile by adding flow control in microfluidic channels in a flexible manner.

The field of microfluidics is characterized by the control and manipulation of small volumes of liquids that flow in a laminar way through microstructures having various functions. This field has opened the door for applications in many areas of healthcare and life sciences such as point-of-care diagnostics (POCDs), environmental analysis, and drug discovery.1–3 POCDs strongly benefit from microfluidics due to the possibilities of miniaturizing tests and integrating various functions into one diagnostic device.4,5 Many rapid diagnostic tests based on lateral flow assays increasingly utilize microfluidic structures and microfabrication to improve their precision and multiplexing capabilities.6 However, one major limitation of microfluidic devices is a lack of flexibility in defining/changing assay conditions on the fly because flow paths in microfluidics are typically set during design and fabrication of the devices. This leads to devices that should be operated according to a fixed protocol, unless active micro-components are used.7 

To be flexible, a POCD device requires a mechanism to control the flow of liquid so that volumes displaced in the device, concentrations of reagents, timing of biochemical reactions, and paths of liquids can be altered according to a specific application post-fabrication. Amongst many liquid actuation principles used in microfluidics,8–10 we believe that flow control based on electrowetting is particularly promising because it does not require mechanical elements such as active pumps or actuated valves.11,12 The effectiveness of electrowetting for controlling the shape and movement of droplets of liquids has been demonstrated, albeit using challenging fabrication processes and large actuation voltage.13–15 More recently, electrowetting has also been used for resuming flow of liquid in capillary-driven microchannels, where a liquid was stopped using patterned hydrophobic barriers: Satoh et al.16 proposed the spontaneous contamination of a gold electrode by carbonaceous compounds to create a hydrophobic barrier on the bottom surface of a flow channel. Furthermore, a hydrophobic barrier-based stopping mechanism was also shown by He et al.17 performing surface modification of inkjet-printed silver electrodes, where the electrodes were covered by a hydrophobic self-assembled monolayer using a perfluorodecanethiol solution. Unlike employing hydrophobic barriers as the main stopping mechanism, here we take advantage of a semicircular trench patterned on the bottom surface of a microfluidic channel. We combine electrowetting with this geometry pinning a meniscus for creating electroactuated valves to control efficiently liquid filling in capillary-driven microfluidics. These valves, which we call “electrogates” (“e-gates”), are reliable, have a short response time, use low voltage, and work well with biological samples such as human serum. Moreover, they are easy to fabricate, have long-term stability, and are compatible with many number of gates in the same chip.

The operating principle of e-gates is shown in Fig. 1. The gate is composed of a trench etched into the bottom surface of a microchannel and a metal electrode patterned over the trench. A second electrode is patterned in the loading pad to make an electrical contact with a liquid sample. A sample pipetted on a loading pad spontaneously fills a hydrophilic microchannel due to capillary forces and stops at the trench location when no bias voltage is applied between the two electrodes [Fig. 1(a)]. The application of a voltage between the electrodes resumes the flow of the liquid [Fig. 1(b)]. The concepts considered during the design and the operation of e-gates are briefly described below.

FIG. 1.

Operation principle of e-gates for the stop-and-go control of liquid flow in capillary-driven microfluidics. (a) and (b) Side view showing the elements forming the flow path and the e-gate: (a) a liquid remains pinned at the trench of an e-gate until (b) an electrical potential is applied to resume the flow across the e-gate using a peripheral, which is controlled with a smartphone. (c) Top view of a microchannel with liquid pinned at the edge of a rounded trench. The angle between the lateral wall and the trench is indicated with α. (d) 3D illustration (not to scale) of the e-gate showing the geometry of the microchannel, the electrode, and the trench.

FIG. 1.

Operation principle of e-gates for the stop-and-go control of liquid flow in capillary-driven microfluidics. (a) and (b) Side view showing the elements forming the flow path and the e-gate: (a) a liquid remains pinned at the trench of an e-gate until (b) an electrical potential is applied to resume the flow across the e-gate using a peripheral, which is controlled with a smartphone. (c) Top view of a microchannel with liquid pinned at the edge of a rounded trench. The angle between the lateral wall and the trench is indicated with α. (d) 3D illustration (not to scale) of the e-gate showing the geometry of the microchannel, the electrode, and the trench.

Close modal

An abrupt change in the angle on the flow path of a liquid, for example, a 90° edge of a trench on one of the surfaces of a hydrophilic rectangular microchannel, can pin the contact line of a liquid filling the microchannel if the sum of the wetting forces on the other surfaces of the microchannel is lower than the pinning forces exerted by the trench. This can be estimated using the Young-Laplace equation derived for a rectangular channel18 

ΔP=γLG2·cosθsw+cosθb*+cosθtd,
(1)

where γLG is the liquid-gas surface tension, w the width of the microchannel, d the depth of the microchannel, and θs and θt the advancing contact angles of the liquid on the sidewalls and the top wall, respectively. θb* is the contact angle that the meniscus takes at the edge of the trench, which according to Gibbs' pinning criterion,19,θb<θb*θb+β, can acquire a maximum value of θb+β, with θb being the contact angle of the liquid on the bottom wall and β the additional angle at the trench [inset in Fig. 1(a)]. The geometry of the meniscus can lead to equilibrium (ΔP=0) at an angle θb*, making the meniscus pinned. However, pinning may be compromised where the trench and the sidewalls intersect. We observed this in experiments comprising a straight trench orthogonal to the microchannel, where the liquid was able to advance on the sidewalls (data not shown). To prevent this, we used semicircular trenches to curve the meniscus in the xy plane, which creates an angle α between the trench and the sidewalls [Figs. 1(c) and 1(d)]. This shape improves the stability of the pinning, possibly by increasing the local curvature of the meniscus where it contacts the sidewalls, thereby decreasing the normal local net force of the surface tension at the intersection point.20,21 Alternatively, the liquid might be inhibited from wetting the intersection point between the trench and the sidewall for α<θs. Our observations on the improved pinning stability of a meniscus at curved trenches are in line with the unpinning mechanism of liquids in microchannels and in contact with oblique “phaseguides.”22 

Pinning of the liquid at the trench can then be suppressed in a controlled manner using electrowetting: a DC potential applied between the liquid and the electrode at the gate induces a local decrease in the contact angle θb of the liquid on the electrode surface due to a reduction of the liquid-solid surface tension (γLS). The contact angle as a function of the applied voltage is expressed through the Lippmann-Young equation12 

cosθb=cosθ0+cV22γLG,
(2)

where θ0 is the contact angle when V across the liquid-solid interface is zero. Once the liquid passes over the e-gate, capillary forces drive the flow along the microchannel until a next e-gate is reached.

E-gates are simple to fabricate and can be implemented using various materials and geometries. Here, the e-gates have been fabricated on an N-doped (1–10 Ohm cm) Si substrate covered with an insulation layer (2.5–μm-thick thermally grown SiO2). Briefly, trenches having a depth of at least 1.5 μm were formed in the SiO2 layer using inductively coupled plasma etching. Then, electrodes were patterned using an 80-nm-thick Pd layer (on a 5-nm-thick Ti adhesion layer) using e-beam evaporation with the sample tilted at 40°. Microchannels were defined by photopatterning a 15–μm-thick SU-8 layer, and chips were sealed by lamination of a dry-film resist (DF-1050 or DF-3020, EMS Inc., USA) at ∼45 °C. These fabrication steps are illustrated in Fig. S1.

The experiments were realized using microfluidic chips containing 6 parallel channels each having a width of 200 μm, a depth of 15 μm, and 2 e-gates. These channels are serviced with a common loading pad, which can accommodate a few microliters of sample [Fig. 2(a)]. The SEM images in Fig. 2(b) show a Pd electrode patterned across the microchannel and a trench. Figure 2(c) shows the stop-and-go flow of a phosphate-buffered saline (PBS) solution at an e-gate where the solution is pinned at the edge of the trench and released when a 3.0 V bias is applied. The characteristic current-time response of the e-gate during activation was measured with a 0.25 s sampling and is shown in Fig. 2(d). The inset highlights a typical transient in current at the rising edge of the applied voltage pulse and a local minimum in current when the activation of the gate occurs. The time elapsed between the transient and this minimum represents the activation delay, i.e., the time the electrowetting process needs to enable liquid actuation through the e-gate, which was about 0.5 s for PBS when applying an activation voltage of 3.0 V.

FIG. 2.

Experimental results from the operation of an e-gate in a microfluidic chip. (a) Photograph of a fabricated chip containing 6 independent flow paths and 2 e-gates per path. Gates placed in parallel channels can be activated independently, whereas those in the same channel can be sequentially activated using a common electrode. The electrodes on this chip terminate with contact pads matching a microSD socket. (b) SEM images of the gate showing the trench and patterned Pd electrode. (c) Optical images showing a liquid (PBS) pinned at a gate and its resumed flow after the activation of the gate. (d) Characteristic current vs. time curve (solid line) observed during activation of the e-gate in response to a voltage pulse (dotted line) applied between the liquid and the electrode at the gate. After flow is resumed, the DC voltage is set to 0 V, which causes the capacitor in the resistor-capacitor (RC) type circuit to discharge.

FIG. 2.

Experimental results from the operation of an e-gate in a microfluidic chip. (a) Photograph of a fabricated chip containing 6 independent flow paths and 2 e-gates per path. Gates placed in parallel channels can be activated independently, whereas those in the same channel can be sequentially activated using a common electrode. The electrodes on this chip terminate with contact pads matching a microSD socket. (b) SEM images of the gate showing the trench and patterned Pd electrode. (c) Optical images showing a liquid (PBS) pinned at a gate and its resumed flow after the activation of the gate. (d) Characteristic current vs. time curve (solid line) observed during activation of the e-gate in response to a voltage pulse (dotted line) applied between the liquid and the electrode at the gate. After flow is resumed, the DC voltage is set to 0 V, which causes the capacitor in the resistor-capacitor (RC) type circuit to discharge.

Close modal

The key performance characteristics of e-gates are their retention time, i.e., the time a liquid remains pinned at an e-gate without applied voltage, their activation delay, and their activation voltage. Concerning the retention time, we expect the trench with the smallest angle α to have the best retention capability. This corresponds to a circular trench perfectly centered in a microchannel and with a diameter equal to the width of the channel. In practice, a small misalignment between the trench and the channel might occur during fabrication, which might introduce gaps between the trench and the sidewalls of the channel. Therefore, we tested the retention capability of trenches with a radius of curvature ≥110 μm (α ≈ 25°) up to a maximum of 1000 μm (α ≈ 84°) for a 200–μm-wide channel. Experiments show that decreasing the radius of curvature of the trench, and consequently decreasing α, improves the retention time and reliability of e-gates (Fig. 3). For an angle of ∼72°, a retention time of at least 2 min is achieved in 50% of the experiments. With an angle of <60°, the reliability of e-gates approaches 100% with retention times exceeding 5 min. Many microfluidic functions such as mixing, pumping, and filtering can be effected on timescales of a few minutes. Longer retention times can also be achieved using 2 or more e-gates in series, if needed. Additionally, the pinning stability was found to be independent of the width of the trench (data not shown).

FIG. 3.

(a) Effect of the radius of curvature of the trench and the corresponding angle α on the retention time of PBS at an e-gate. Each data point represents the average of 10 experiments. The measurement of the retention time was intentionally stopped after 20 min. The inset illustrates trenches having the increasing radius of curvature. (b) Reliability of the flow retention at the e-gates as a function of the radius of curvature (angle α). The retention time equal or shorter than 1.75 min was considered as a failure of the e-gate. The curves serve as guides to the eye.

FIG. 3.

(a) Effect of the radius of curvature of the trench and the corresponding angle α on the retention time of PBS at an e-gate. Each data point represents the average of 10 experiments. The measurement of the retention time was intentionally stopped after 20 min. The inset illustrates trenches having the increasing radius of curvature. (b) Reliability of the flow retention at the e-gates as a function of the radius of curvature (angle α). The retention time equal or shorter than 1.75 min was considered as a failure of the e-gate. The curves serve as guides to the eye.

Close modal

Electrowetting is influenced by the conductivity and ionic strength of a liquid.11 We therefore characterized the activation voltage and the activation delay of e-gates using aqueous NaCl solutions (Fig. 4). At concentrations of NaCl ≥ 50 mM, 5.0 V was sufficient for activating the e-gates with an activation delay of 0.25 s. The activation delay decreased with increasing activation voltage (Fig. 4, inset). E-gates could still be activated in 3.9 mM NaCl at 10 V with an activation delay of 8.6 s, which was decreased down to 0.75 s at 26 V without any bubble formation. NaCl samples having a concentration of 125 mM required an activation voltage of only 3.3 V. These findings are consistent with the results obtained using a PBS solution, which mostly contains NaCl and has a total ionic strength of 162.7 mM at pH 7.4.23 

FIG. 4.

Effect of the ionic strength on the activation voltage of the e-gate for different NaCl concentrations. Measurements correspond to an activation delay of 0.25 s. The inset shows the effect of activation voltage on the activation delay at different NaCl concentrations. The lines serve as a guide to the eye.

FIG. 4.

Effect of the ionic strength on the activation voltage of the e-gate for different NaCl concentrations. Measurements correspond to an activation delay of 0.25 s. The inset shows the effect of activation voltage on the activation delay at different NaCl concentrations. The lines serve as a guide to the eye.

Close modal

Human serum can be difficult to pin in microchannels because the adsorption of proteins from serum to surfaces can increase the wettability of microchannels. E-gates were successfully tested with biological samples such as human serum and artificial urine (Fig. 5), thereby showing compatibility with reagents, making a contact angle from 46° to 84° (Table S2), having a viscosity between 0.975 mPa⋅s and 1.4 mPa⋅s and a surface tension between 59 mN/m and 73 mN/m (Table S3). Initial experiments using human serum (S4200, Biowest) showed retention times of the e-gates between 5 min and 8 min. We observed that the retention capability of human serum by e-gates can be strengthened by slowing the incoming meniscus right before it reaches the edge of the trench using patterned Pd tiles24,25 or by making the Pd electrodes hydrophobic using a self-assembled monolayer (eicosanethiol). Using either one of the both strategies resulted in retention times longer than 45 min for human serum and artificial urine (Pickering Laboratories, Cat. No.: 1700–0600). We characterized the activation of e-gates for these samples and observed that flow resumed within 1 s when applying a bias of 5 V (Fig. 5).

FIG. 5.

Activation voltage vs. activation delay for human serum (filled squares) and artificial urine (open circles). The lines are guides to the eye. The inset shows an optical microscopy image showing the resumed flow of human serum after applying a DC voltage of 8 V for the activation of the gate. The Pd tiles (5 × 5 μm2 squares) are patterned before and at the edge of the electrode.

FIG. 5.

Activation voltage vs. activation delay for human serum (filled squares) and artificial urine (open circles). The lines are guides to the eye. The inset shows an optical microscopy image showing the resumed flow of human serum after applying a DC voltage of 8 V for the activation of the gate. The Pd tiles (5 × 5 μm2 squares) are patterned before and at the edge of the electrode.

Close modal

In summary, we presented an efficient, reliable, and easy-to-implement flow control mechanism that is compatible with capillary-driven microfluidics. In contrast to other techniques employing hydrophobic barriers or electrowetting-on-dielectric principles, our method uses a simple geometrical pinning effect, which can easily be fabricated using techniques that are already compatible with many POCD devices employing microfluidics and electrodes. The low actuation voltages (<10 V) required to resume flow at e-gates make this approach compatible with compact and portable peripherals that can be used to address e-gates and to communicate with a smartphone.26 This work also gives the possibility to make versatile microfluidic chips inside which flow paths are defined based on the actuation of a few e-gates using a protocol defined by a mobile device such as a smartphone. We therefore expect this approach to have an impact on microfabricated microfluidics for POCD applications.

See supplementary material for the illustration of the fabrication steps of the “electrogates” and the tables containing values of viscosity, surface tension, and contact angle of some tested solutions on the wetted surfaces of the microchannel and the electrode.

The authors thank E. Hemmig, R. D. Lovchik, Ute Drechsler, and Diana Davila for discussions and W. Riess for his continuous support. The authors also thank Gustav Graeber from the Laboratory of Thermodynamics in Emerging Technologies at the ETH Zurich for his help on some preliminary contact angle measurements. This work received funding from the European Union's Horizon 2020 research and innovation program under the Marie Skłodowska-Curie Grant Agreement No. [701690].

1.
P. N.
Nge
,
C. I.
Rogers
, and
A. T.
Woolley
,
Chem. Rev.
113
,
2550
(
2013
).
2.
M. L.
Kovarik
,
P. C.
Gach
,
D. M.
Ornoff
,
Y.
Wang
,
J.
Balowski
,
L.
Farrag
, and
N. L.
Allbritton
,
Anal. Chem.
84
,
516
(
2012
).
3.
C. D.
Chin
,
V.
Linder
, and
S. K.
Sia
,
Lab Chip
12
,
2118
(
2012
).
4.
A.
Arora
,
G.
Simone
,
G. B.
Salieb-Beugelaar
,
J. T.
Kim
, and
A.
Manz
,
Anal. Chem.
82
,
4830
(
2010
).
5.
L.
Gervais
,
N.
De Rooij
, and
E.
Delamarche
,
Adv. Mater.
23
,
H151
(
2011
).
6.
L.
Gervais
and
E.
Delamarche
,
Lab Chip
9
,
3330
(
2009
).
7.
M. W.
Ashraf
,
S.
Tayyaba
, and
N.
Afzulpurkar
,
Int. J. Mol. Sci.
12
,
3648
(
2011
).
8.
A. K.
Au
,
H.
Lai
,
B. R.
Utela
, and
A.
Folch
,
Micromachines
2
,
179
(
2011
).
9.
M. H.
Oddy
and
J. G.
Santiago
,
J. Colloid Interface Sci.
269
,
192
(
2004
).
10.
D.
Vigolo
,
R.
Rusconi
,
H. A.
Stone
, and
R.
Piazza
,
Soft Matter
6
,
3489
(
2010
).
11.
F.
Mugele
and
J.-C.
Baret
,
J. Phys. Condens. Matter
17
,
R705
(
2005
).
12.
J.
Lee
,
H.
Moon
,
J.
Fowler
,
T.
Schoellhammer
, and
C.
Kim
,
Sens. Actuators, A
95
,
259
(
2002
).
13.
C.
Quilliet
and
B.
Berge
,
Curr. Opin. Colloid Interface Sci.
6
,
34
(
2001
).
14.
Y.-P.
Zhao
and
Y.
Wang
,
Rev. Adhes. Adhes.
1
,
114
(
2013
).
15.
R.
Shamai
,
D.
Andelman
,
B.
Berge
, and
R.
Hayes
,
Soft Matter
4
,
38
(
2008
).
16.
W.
Satoh
,
H.
Yokomaku
,
H.
Hosono
,
N.
Ohnishi
, and
H.
Suzuki
,
J. Appl. Phys.
103
,
034903
(
2008
).
17.
F.
He
and
S. R.
Nugen
,
Microfluid. Nanofluid.
16
,
879
(
2014
).
18.
H.
Cho
,
H. Y.
Kim
,
J. Y.
Kang
, and
T. S.
Kim
,
J. Colloid Interface Sci.
306
,
379
(
2007
).
19.
J. F.
Oliver
,
C.
Huh
, and
S. G.
Mason
,
J. Colloid Interface Sci.
59
,
568
(
1977
).
20.
J. U.
Brackbill
,
D. B.
Kothe
, and
C.
Zemach
,
J. Comput. Phys.
100
,
335
(
1992
).
21.
R.
Lovett
and
M.
Baus
,
J. Chem. Phys.
106
,
635
(
1997
).
22.
P.
Vulto
,
S.
Podszun
,
P.
Meyer
,
C.
Hermann
,
A.
Manz
, and
G. A.
Urban
,
Lab Chip
11
,
1596
(
2011
).
23.
N.
Lloret
,
R. S.
Frederiksen
,
T. C.
Møller
,
N. I.
Rieben
,
S.
Upadhyay
,
L.
De Vico
,
J. H.
Jensen
,
J.
Nygård
, and
K. L.
Martinez
,
Nanotechnology
24
,
035501
(
2013
).
24.
F.
Diotallevi
,
L.
Biferale
,
S.
Chibbaro
,
A.
Puglisi
, and
S.
Succi
,
Phys. Rev. E.
78
,
036305
(
2008
).
25.
E.
Kreit
,
M.
Dhindsa
,
S.
Yang
,
M.
Hagedon
,
K.
Zhou
,
I.
Papautsky
, and
J.
Heikenfeld
,
Langmuir
26
,
18550
(
2010
).
26.
Y.
Temiz
and
E.
Delamarche
, “
Sub-nanoliter flow monitoring in microfluidic chips using a portable device and smartphone
,”
Sci. Rep.
(submitted).

Supplementary Material