ArticlePDF Available

Monolithic integrated system with an electrowetting-on-dielectric actuator and a film-bulk-acoustic-resonator sensor


Abstract and Figures

Although digital microfluidics has shown great potential in a wide range of applications, a lab-on-a-chip with integrated digital droplet actuators and powerful biochemical sensors is still lacking. To address the demand, a fully integrated chip with electrowetting-on-dielectric (EWOD) and a film bulk acoustic resonator (FBAR) sensor is introduced, where an EWOD actuator manipulates digital droplets and the FBAR sensor detects the presence of substances in the droplets, respectively. The piezoelectric layer of the FBAR sensor and the dielectric layer of the EWOD share the same aluminum nitride (AlN) thin film, which is a key factor to achieve the full integration of the two completely different devices. The liquid droplets are reliably managed by the EWOD actuator to sit on or move off the FBAR sensor precisely. Sessile drop experiments and limit of detection (LOD) experiments are carried out to characterize the EWOD actuator and the FBAR sensor, respectively. Taking advantage of the digital droplet operation, a ‘dry sensing mode’ of the FBAR sensor in the lab-on-a-chip microsystem is proposed, which has a much higher signal to noise ratio than the conventional ‘wet sensing mode’. Hg2+ droplets with various concentrations are transported and sensed to demonstrate the capability of the integrated system. The EWOD–FBAR chip is expected to play an important role in many complex lab-on-a-chip applications.
Content may be subject to copyright.
This content has been downloaded from IOPscience. Please scroll down to see the full text.
Download details:
IP Address:
This content was downloaded on 24/01/2015 at 06:52
Please note that terms and conditions apply.
Monolithic integrated system with an electrowetting-on-dielectric actuator and a film-bulk-
acoustic-resonator sensor
View the table of contents for this issue, or go to the journal homepage for more
2015 J. Micromech. Microeng. 25 025002
Home Search Collections Journals About Contact us My IOPscience
1 © 2015 IOP Publishing Ltd Printed in the UK
1. Introduction
Lab-on-a-chip or microuidic systems offer the ability to carry
out high-resolution separations and detections on an integrated
chip with low cost and in a short time, using small quantities
of samples and reagents. It has been employed for a broad
range of applications, such as molecular analysis, biodefence
and molecular biology [1]. As a promising digital microu-
idic technology, electrowetting-on-dielectric (EWOD) has
the capability of handling small sample volumes, indepen-
dent control of multiple droplets, high-throughput parallel
processing, fast reaction, and contamination-free operation.
Compared to continuous microuidic technology, EWOD
exhibits outstanding exibility, recongurable nature, and
scale-up potential [2]. An EWOD platform involving basic
operations, such as droplet creating, transporting, cutting,
merging and mixing, has been widely adopted in biochemical
assays [3, 4]. Sista et al demonstrated a portable EWOD plat-
form for immunoassay and polymerase chain reaction (PCR)
application [5]. Dubois et al used an open EWOD platform as
a micro-reactor for chemical synthesis [6]. Miller et al applied
an EWOD platform to the quantication of enzymatic activity
in real time by uorescence detection [7]. While the platform
has made important progress in making itself competitive for
Journal of Micromechanics and Microengineering
Monolithic integrated system with an
electrowetting-on-dielectric actuator and
a lm-bulk-acoustic-resonator sensor
MenglunZhang, WeiweiCui, XuejiaoChen, ChaoWang,
WeiPang, XuexinDuan, DaihuaZhang and HaoZhang
State Key Laboratory of Precision Measuring Technology and Instruments, Tianjin University, 92 Weijin
Road, Nankai District, Tianjin, People’s Republic of China
E-mail: and
Received 25 August 2014, revised 27 September 2014
Accepted for publication 2 October 2014
Published 15 January 2015
Although digital microuidics has shown great potential in a wide range of applications,
a lab-on-a-chip with integrated digital droplet actuators and powerful biochemical sensors
is still lacking. To address the demand, a fully integrated chip with electrowetting-on-
dielectric (EWOD) and a lm bulk acoustic resonator (FBAR) sensor is introduced, where
an EWOD actuator manipulates digital droplets and the FBAR sensor detects the presence of
substances in the droplets, respectively. The piezoelectric layer of the FBAR sensor and the
dielectric layer of the EWOD share the same aluminum nitride (AlN) thin lm, which is a
key factor to achieve the full integration of the two completely different devices. The liquid
droplets are reliably managed by the EWOD actuator to sit on or move off the FBAR sensor
precisely. Sessile drop experiments and limit of detection (LOD) experiments are carried out
to characterize the EWOD actuator and the FBAR sensor, respectively. Taking advantage of
the digital droplet operation, a ‘dry sensing mode’ of the FBAR sensor in the lab-on-a-chip
microsystem is proposed, which has a much higher signal to noise ratio than the conventional
‘wet sensing mode’. Hg2+ droplets with various concentrations are transported and sensed to
demonstrate the capability of the integrated system. The EWOD–FBAR chip is expected to
play an important role in many complex lab-on-a-chip applications.
Keywords: FBAR sensor, electrowetting-on-dielectric, digital microuidics, monolithic
(Some gures may appear in colour only in the online journal)
M Zhang et al
Printed in the UK
© 2015 IOP Publishing Ltd
j. Micromech. Microeng.
journal of Micromechanics and Microengineering
J. Micromech. Microeng. 25 (2015) 025002 (9pp)
M Zhang et al
lab-on-a-chip applications, the success of EWOD technology
will greatly depend on its ability to integrate powerful sensors.
In the past decade, EWOD has been used in concert with
a number of detection methods, such as matrix assisted laser
desorption/ionization mass spectrometry (MALDI-MS [8]),
surface plasmon resonance (SPR [9]), and uorescence [10].
However, the sensors reported in EWOD-related literature are
discrete and bulky, and have not yet been integrated compactly
with EWOD chips. The form factor of EWOD chips is small,
but its sensing counterpart for lab-on-a-chip is still lacking.
Although a microuidic platform integrating a quartz crystal
microbalance (QCM) sensor and EWOD has been demon-
strated, the device suffers from severe performance deviations
from chip to chip due to manual xation of the QCM sensor,
and it is difcult to achieve high-throughput detection with the
QCM sensor array [11]. MEMS sensors have inherent smaller
sizes compared to conventional sensors. In addition, they have
the potential to be fully integrated with EWOD, because both
of their fabrication processes are IC compatible. The batch
fabrication ability of MEMS sensors is also a key advan-
tage in point-of-care-testing and high-throughput detection.
Unfortunately, MEMS sensors suffer from performance deg-
radation when detecting analytes in liquid solution because
of severely damped mechanical motion in aqueous environ-
ment [1215]. Although several methods have been proposed
to combat it [12, 1618], none of them restore the noise oor
of the sensor in liquid solution and some of them compromise
detection sensitivity.
Film bulk acoustic resonators (FBARs) fabricated by
MEMS processes utilize thickness extensional mode reso-
nance, which is excited in a vertically grown piezoelectric
material lm. Since its resonant frequency is sensitive to
the added mass, FBARs have been widely investigated as
chemical and biological sensors in gaseous and aqueous envi-
ronments [19]. Liu et al detected vapor traces of TNT and
RDX explosives with antibody-coated FBAR sensors [20].
Qiu et al investigated pH measurements [21] and the inuence
of temperature, relative humidity and reducing gases on ZnO
based FBAR [22]. Zhang et al analyzed DNA sequences with
an air-backed FBAR, and distinguished the complementary
DNA from a single-nucleotide mismatch DNA sequence [23].
For antibody–antigen interaction, human IgG was detected
with a solid-mounted FBAR sensor by Yan et al [24]. FBAR
sensors feature high sensitivity and ease of batch produc-
tion compared with QCM sensors, owing to their thin lm
structure and the fabrication process of MEMS devices [12].
Nirschl et al fabricated an FBAR sensor array to demonstrate
its multiplexed detection ability [25].
In this paper, we demonstrate an EWOD–FBAR lab-on-a-
chip system, with an EWOD droplet actuator and FBAR sensors
integrated on a single silicon chip. Sputtered aluminum nitride
(AlN) thin lm as a critical material is employed in the chip
to provide low voltage driving for EWOD and highly stable
piezoelectric resonance in FBAR sensing. The similar device
structure and material of the two devices reduce the fabrica-
tion complexity comprehensively, making the system more
compact. Sample droplets are manipulated reliably and pre-
cisely by the EWOD to stay on or move away from specied
FBAR sensor area. After the reaction of a sample droplet
with the FBAR sensor completes, the droplet is removed
from the sensor. Then the FBAR resonant frequency change
is recorded after an optional washing step by deionized (DI)
water droplets to remove unbound molecules. The digital fea-
ture of EWOD helps to improve the signal to noise ratio of the
FBAR sensor substantially in liquid solution since the random
variation of resonant frequency of an FBAR sensor is typi-
cally worse by more than an order of magnitude with liquid
loading. Although the manual ‘dip and dry’ method has been
proposed [26, 27], the sensing protocol with programmable
droplet manipulation provided by EWOD avoids contamina-
tion on the sensor element and reliable experimental results
are obtained with high efciency.
2. Design and fabrication of the integrated system
2.1. Device conguration
The schematic of the integrated system is shown in gure1.
The FBAR sensor consists of a piezoelectric lm sandwiched
between two metal electrodes, and a micromachined air cavity
in silicon substrate underneath the metal/piezo/metal struc-
ture. The Au lm on the top electrode is a specic coating
for analyte sensing, where the hydrophobic layer is removed
to expose the Au surface. The EWOD actuator comprises
patterned actuation electrodes, a dielectric lm on top of the
electrodes, and a hydrophobic layer on the dielectric lm. A
co-planar EWOD structure is adopted instead of a two-plate
conguration to get rid of the cover plate [3]. Thin AlN lm
is a critical material of the FBAR piezoelectric layer and the
EWOD dielectric layer, which facilitates the integration of the
two devices on a single silicon chip.
Photos of the fabricated microsystem are displayed in
gure2. Comb-like actuation electrodes are used in EWOD
for reliable droplet actuation. Each individual actuation elec-
trode is connected to a large metal pad for the application of
voltage. Two FBAR sensors are placed at the two ends of the
EWOD electrode array. The pentagonal area where the top
electrode, piezoelectric layer, bottom electrode and microma-
chined cavity overlap is the active area or the sensing area
of the FBAR sensor. The active area surface can be function-
alized for binding of chemical or biological molecules [19].
In this prototype demonstration, a Au thin lm is formed on
the active area as the specic coating for analyte detection.
Hydrophobic Teon is spin-coated and patterned to expose
Figure 1. Cross-sectional schematic of the integrated system (not
to scale).
J. Micromech. Microeng. 25 (2015) 025002
M Zhang et al
the Au on the FBAR active area. A fabricated FBAR sensor
before the Au and Teon layers is introduced as shown in
gure2(b). In gure 2(c), a photo of the sensor with the Au
and Teon patterns is displayed. The resonant frequency
signal of the FBAR sensor is recorded using a ground-signal-
ground RF probe connected to a network analyzer.
2.2. Device fabrication
The fabrication process is shown in gure 3 and briey
described as follows. A Si wafer is etched to form a cavity
of several micrometers deep in the substrate by reactive ion
etching. A phosphosilicate glass (PSG) layer is then depos-
ited by plasma enhanced chemical vapor deposition (PECVD)
and planarized by chemical mechanical planarization (CMP).
CMP planarizes the global surface and removes the PSG
outside the cavity (gure 3(a)). A 200 nm thick Mo layer is
deposited and patterned by sputtering to form the actuation
electrodes of EWOD and the bottom electrode of the FBAR
sensor on the planarized substrate surface. Then a 1 μm thick
AlN lm with a FWHM of 1.5° is deposited by reactive sput-
tering system (gure 3(b)). After forming the top electrode
of the FBAR sensor with another 200 nm thick Mo, the AlN
on the bottom electrode is dry-etched to make the bottom
electrode accessible by test probes. A 10 nm thick Cr (not
shown in gure3) adhesion layer and 80 nm thick Au layer
are sequentially deposited and patterned, which prepares the
surface for the adsorption of chemical or biological molecules
(gure 3(c)). Afterward the wafer is immersed in diluted
hydrouoric acid solution for around 2 h to remove the sacri-
cial PSG. Finally, a 30 nm thick Teon lm is spin-coated on
the substrate to form the topmost hydrophobic surface layer
except the active area of the FBAR sensor where the Teon is
peeled off using lift-off process (gure 3(d)).
Figure 2. Photos of the fabricated integrated system: (a) aerial view; (b) close-up view of the FBAR sensor without the Teon pattern and
Au lm on the top electrode; (c) close-up view of the FBAR sensor with the Teon pattern and Au lm on the top electrode.
Figure 3. Brief fabrication process of the integrated system.
J. Micromech. Microeng. 25 (20 15) 025002
M Zhang et al
2.3. General considerations for performance optimization
The Sauerbrey equationis applied to describe the relationship
between the added mass Δm and the resonant frequency change
Δf of FBAR sensors. The Sauerbrey equationis dened as [19]
Δρμ Δ=−ff
where f0, A, ρ and μ are resonant frequency, size of the active
area, equivalent mass density and equivalent elastic modulus
of the FBAR device, respectively. It shows that the resonant
frequency change is proportional to the added mass on the res-
onator. In an FBAR sensor, the series resonant frequency of the
resonator is monitored and correlates with the mass change.
Mass sensitivity
, minimum detectable frequency change L
and LOD of an FBAR sensor can be calculated by [12, 19]
where ρi and di are the mass density and thickness of the ith
layer in the FBAR device;
is the minimum detectable
phase change of electrical impedance; Q is the quality factor
of the FBAR resonator. Mo is chosen as the electrode material
of FBAR, since it can help to grow highly textured, i.e. high
quality factor, crystalline AlN. In addition, Mo is intrinsically
a low acoustic loss, i.e. high quality factor, material.
The electric potential V applied on the dielectric layer of
EWOD determines the solid–liquid interfacial tension
Lippmann’s equation[4]
γ γ
() (0)
is the solid–liquid interfacial tension when
voltage is applied;
is the non-actuated solid–liquid
interfacial tension; εr and ε0 are relative permittivity of the
dielectric layer and permittivity of vacuum, respectively, and
t is thickness of the dielectric layer. The relative permittivity
of AlN is 9.5, which is relatively higher than that of the com-
monly used dielectric layer in EWOD [4, 8]. Therefore AlN is
preferable in view of lowering actuation voltage.
In the integrated platform, AlN thin lm as a key material
is used in the piezoelectric layer of the FBAR and dielectric
layer of the EWOD. Therefore, the structural design of the two
devices should be carefully coordinated to optimize the opera-
tion of the system. Thickness of AlN lm and size of the active
area are two essential parameters to be considered. According
to equations (2) and (4), a thinner AlN layer would enhance
the mass sensitivity and hence the LOD of the FBAR sensor.
According to equation (5), thinner AlN lm is preferred for
the EWOD dielectric layer because lower voltage is used to
reach the same driving force. However, crystal defects such as
pin-holes and ipped grains tend to have stronger inuence on
thinner piezoelectric AlN lm, which would degrade quality
factor of the FBAR, and hence minimum detectable frequency
change and LOD according to equations(3) and (4). The thin
AlN lm also has a high risk of dielectric breakdown to damage
the EWOD actuator. In terms of size of the active area, a smaller
hydrophilic active area would minimize the resistance of actua-
tion when sample droplets are moved off the FBAR sensor by
EWOD. However, an FBAR sensor with a small active area
would exhibit strong spurious modes and Q of the resonator is
considerably degraded. Taking the above aspects into consid-
eration, the thickness of the AlN lm and the size of the active
area are chosen to be 1 μm and 10 000 μm2, respectively.
The patterns of Au layer and Teon layer are designed to
be the same as the FBAR resonant active area. If the Au or
Teon pattern is smaller than the active area, the surface for
the sensor reaction with the substances in the liquid droplet
is utilized ineffectively, which would decrease the signal to
noise ratio of the sensor. If the Au or Teon pattern exceeds
the active area, it does not help the sensor improve the signal
to noise ratio, but makes droplet actuation more difcult.
The active area is a pentagon instead of a square in order
to improve the electrical performance of the FBAR. Spurious
resonances are signicantly suppressed by the apodization
approach [28]. Mo has been used as the electrode material
in the FBAR [29]. The EWOD electrodes utilize the same
material as the Mo electrode of the FBAR to increase the chip
3. Experiment and discussion
3.1. Sessile drop experiment
To the authors’ knowledge, piezoelectric AlN lm is rarely
investigated as the dielectric layer in EWOD devices. A sessile
Figure 4. Voltage applied as a function of contact angle. Both dc
voltage and 1 kHz ac voltage are separately applied to the AlN/
Teon composite dielectric lm. The black curve is a theoretical
value calculated by the classic Lippmann–Young equation. The red
squares and the blue triangles are measured contact angles in dc and
ac voltage experiments, respectively. The inset is the photo of the
non-actuated droplet.
J. Micromech. Microeng. 25 (2015) 025002
M Zhang et al
drop experiment is conducted to characterize the AlN dielec-
tric layer by using a contact angle goniometer. Both 1 kHz
ac and dc voltages are separately applied between a fresh DI
water droplet and the bottom electrode underneath the AlN
dielectric lm. No dielectric breakdown was observed. The
experiment is repeated several times and averaged values of
the data along with standard deviation results are shown in
gure4. The measured contact angles in both the ac and dc
voltage experiments match well with the theoretical values
found using the classic Lippmann–Young equation at low
voltages, but start to deviate gradually when applied voltage
exceeds 25 V and the contact angle approaches a saturation
value, which is approximately 90°. The observed ac satura-
tion angle is somewhat lower than the dc saturation angle.
Although the underlying mechanism is not fully understood,
Hong et al speculated that the delayed saturation of ac elec-
trowetting may be related to the delay of dielectric breakdown
due to a reduced effective electric eld [30].
3.2. Actuation ability demonstration
A DI water droplet is transported back and forth from one
FBAR sensor at one end of the EWOD electrode array to
the other end where another FBAR sensor is placed. The
EWOD electrodes are electrically connected to a home-built
relay controller board, which is programmed to apply actu-
ation voltage on the specied electrodes. In this way, fully
controlled and automatic droplet transportation on a chip is
realized. A video camera is installed above the chip to visually
record the droplet motion.
The snapshots taken from one transportation cycle are
displayed in gure 5. The droplet moves quite smoothly
throughout the transportation process. No dielectric break-
down or pinning phenomenon has been observed. It should be
noted that the water droplet is placed precisely on the active
area of the FBAR sensors in the transportation, which guaran-
tees the repeatability of FBAR sensing.
The inuence of applied voltage on droplet actuation is
also investigated. As the sessile drop experiment indicates and
Malic et al suggested [3], ac voltage actuation is preferred
rather than dc actuation. The threshold ac voltage to move the
droplet is around 25 V at 1 kHz. When applied voltage rises,
the droplet moves faster, exceeding a velocity of 90 mm s−1.
Dielectric breakdown does not occur until the voltage is as
high as 70 V.
3.3. LOD for mass detection
In the applications of FBAR sensing, there are two approaches
to taking the response signal of the sensor, which are the ‘wet
sensing mode’ and the ‘dry sensing mode’, respectively. In
the wet sensing mode, the FBAR resonant frequency change
is recorded online while the sensor is loaded with the sample
solution. In dry sensing mode, the FBAR resonant frequency
change is obtained by subtracting the resonant frequency
before the sample solution is loaded from the resonant fre-
quency after the sample solution is unloaded. The dry sensing
mode is a unique feature of digital microuidics, in which
the frequency signal is taken when the FBAR surface is dry.
Conventional microuidics with continuous uid ow can
barely realize fast loading and unloading of liquid samples.
The quality factor of the FBAR is crucial to achieving low
LOD of the sensor according to equations (3) and (4). The
quality factor at series resonance
s of a 1.5 GHz FBAR with a
Teon pattern is measured with a DI water droplet loaded and
unloaded on the resonator. S-parameters are recorded by an
Agilent E5061B network analyzer. The
s is calculated from
the S-parameter data by using the Bode equation[29].
The electrical input impedances and calculated
s are
shown in gures 6(a)–(c). Before the DI water droplet is
Figure 5. Snapshots of droplet transportation in one cycle. (a) When no voltage is applied, the droplet stays still; (b) the droplet is moving
to the right FBAR sensor; (c) the droplet is placed on the right FBAR sensor; (d) the moment when the droplet leaves the right FBAR
sensor; (e) the droplet is moving to the left FBAR sensor; (f) the droplet is placed on the left FBAR sensor; (g) the moment when the
droplet leaves the left FBAR sensor; (h) the droplet is moving to the right FBAR sensor.
J. Micromech. Microeng. 25 (20 15) 025002
M Zhang et al
of the FBAR working in air is 3745, which drops
substantially to around 50 after the water droplet is loaded,
because a large portion of the acoustic energy originally
trapped in the resonator dissipates into the water [17]. After
the water droplet is unloaded by EWOD, the
of the FBAR
recovers back to 3735 in several seconds. The residual
liquid on the FBAR dries out very quickly after the droplet
is moved away because the hydrophilic area is small. The
superposed red and blue curves in gure6(d) indicate that
complete recovery occurs.
Though the wet sensing mode is monitored in real-time, it
has a much worse LOD than the dry sensing mode, because
of the FBAR in a liquid environment is degraded by nearly
two orders of magnitude (from 3745 to 50). The series reso-
nant frequency of an FBAR has been continuously monitored
and recorded for 2 min in each sensing mode. The minimum
detectable frequency change of an FBAR sensor is obtained
by the random variation of resonant frequency [12]. As shown
in gure7, the random variation of resonant frequency in the
wet sensing mode is approximately 5 kHz. However, it is as
small as 0.3 kHz in the dry sensing mode. The mass sensitivity
is known to be about 900 cm2 g−1, and the LODs are calculated
to be 2.2 ng cm−2 and 0.13 ng cm−2 in the wet and dry sensing
modes, respectively.
3.4. Detection ability demonstration
To demonstrate the capability of the EWOD–FBAR system, a
set of Hg2+ experiments are conducted. Hg2+ in liquid amal-
gamates with the Au lm, which leads to an increased mass
of the Au layer on the FBAR surface. The mass change is
correlated to the FBAR resonant frequency change according
to equation(1). Hg2+ solutions with different concentrations
are prepared by dissolving HgCl2 in DI water and performing
a ten-fold serial dilution. There are four major steps in the
complete experimental protocol.
(a) FBAR resonant frequency is recorded in air.
(b) An Hg2+ droplet is actuated by EWOD to cover the active
area of the FBAR sensor. The droplet stays on the sensor
for several minutes for the reaction to occur. Resonant
frequency of FBAR is constantly monitored and recorded.
(c) The droplet is moved away from the FBAR sensor by
EWOD, and the sensor is left in air to stabilize the reso-
nant frequency.
(d) FBAR resonant frequency is recorded again in air.
The frequency difference between step (a) and step (d) is
the resonant frequency change obtained in the dry sensing
mode. The frequency drop in step (b) is the resonant fre-
quency change collected in the wet sensing mode. Fresh
FBAR sensors are used in all Hg2+ experiments to avoid cross
contamination. The experiments are conducted in a cleanroom
with controlled temperature and humidity.
The resonant frequency change in the complete cycle
from step (a) to step (d) is tracked in real-time and plotted
in gure8. The resonant frequency drops abruptly due to the
loading effect of the liquid droplet when the droplet is intro-
duced on the sensor. With the progress of Hg2+ amalgamation
in the Au thin lm, the resonant frequency decreases gradu-
ally, which is constantly monitored. In the 10–3M and 10–5M
Hg2+ experiments, the resonant frequency shifts are observed
to be 123 MHz and 2.1 MHz, respectively. After the liquid
Figure 6. Electrical performance of the FBAR sensor in different stages. The magnitude of electrical input impedance as a function of
frequency (a) before the droplet is loaded, (b) with the droplet loaded, and (c) after the droplet is unloaded. (d) Smith circles of the three
stages. Note that the smith circles before the droplet loading and after the droplet unloading are superposed.
J. Micromech. Microeng. 25 (2015) 025002
M Zhang et al
droplet is unloaded and the FBAR sensor is back to working
stably in air, the resonant frequency is compared to the ini-
tially recorded frequency to calculate the resonant frequency
shifts in the dry sensing mode, which are 136 MHz and
2.3 MHz, respectively. As expected, the resonant frequency
shifts extracted in the wet sensing mode and the dry sensing
mode are approximately the same.
With the concentration of the Hg2+ droplet further reduced,
the resonant frequency shift becomes weak, which is hard
to recognize if the wet sensing mode is used. However, the
small frequency shift might be clearly distinguished by the
dry sensing mode. For example, the same experiment has
been repeated with an Hg2+ droplet of 10–7M, and the reso-
nant frequency shift is 28 kHz, which is merely several times
of random frequency variation in the wet sensing mode and
barely distinguished. Yet the signal to noise ratio in the dry
sensing mode is high and 28 kHz is effortless to be captured.
As shown in gure9, the resonant frequency shift is pro-
portional to the Hg2+ concentration. The sensor sensitivity
is calculated to be 250 kHz μM−1 and LOD is lower than
10–9M for Hg2+ detection. The prototype device for Hg2+
detection in this research is one-time use. The Hg2+ in the
Au lm could be removed by heating and washing the FBAR
sensor. Therefore, the EWOD–FBAR platform has potential
for multiple-time use.
With the automatic droplet manipulation provided by the
EWOD–FBAR platform, the complicated operation procedure
used in conventional ‘dip and dry’ systems is simplied, and
contamination is effectively prevented. Digital microuidics
makes the dry sensing mode of an FBAR sensor feasible. The
dry sensing mode has a much higher signal to noise ratio than
the conventional wet sensing mode. The EWOD–FBAR inte-
grated platform is equipped with on-chip sample and reagent
processing, and improved resolution detection. Apart from
ion detection, it has great potential to be employed in more
sophisticated studies, such as protein–ligand interactions,
DNA hybridization, and immunoassays. For long-time incu-
bations, droplet evaporation will be a concern, which could
Figure 7. Random variation of resonant frequency within 2 min.
The wet sensing mode is in red and the dry sensing mode is in blue.
Figure 8. Resonant frequency change tracked with time in the
experiments of (a) 10–3M Hg2+ sensing and (b) 10–5M Hg2+ sensing.
Step (a) and step (d) in red are the dry sensing mode, and step (b) in
blue is the wet sensing mode.
Figure 9. Sensitivity and LOD of the dry sensing mode for Hg2+
J. Micromech. Microeng. 25 (20 15) 025002
M Zhang et al
be minimized by providing a humid environment. Sealing the
device [32] or placing the device into a humidity chamber [33]
has been proved to be effective. As the resonant frequency will
change with the residue liquid evaporation after the droplet is
removed from the FBAR sensor by the EWOD actuator, the
sensor could be dried in nitrogen gas to speed up the resonant
frequency stabilization process.
4. Conclusion
A fully integrated system incorporating an EWOD actuator
and an FBAR sensor on a single chip is demonstrated. The
EWOD actuator is able to manipulate and position the drop-
lets on or off the FBAR sensor precisely and reliably. The
sensitivity and LOD of the EWOD–FBAR system for Hg2+
detection are found to be 250 kHz μM−1 and 10–9M, respec-
tively. The combination of EWOD and FBAR brings the
integrated system many benets such as compact size, auto-
matic and contamination-free operation and improved signal
to noise ratio.
In future work, a more sophisticated system incorporating
EWOD actuators and FBAR sensors will be developed to
implement multiple analytical operations, such as sample pre-
treatment, sample separation, sample ltering, sample mixing
and parallel detections. Frequency counters or active CMOS
circuits [31] will also be integrated on the chip to measure
the resonant frequency shift instead of bulky readout instru-
ments, enabling a total miniature system. It is not hard to
imagine, through the compact integrated platform, hundreds
of tiny sample droplets being transported, processed and
precisely detected in parallel operation. The next generation
EWOD–FBAR lab-on-a-chip system holds great promise in
high-throughput research and point-of-care-testing.
This work was supported by Natural Science Foundation of
China (NSFC No 51375341, No 61006074 and No 61176106).
[1] WhitesidesG M 2006 The origins and the future of
microuidics Nature 442 36873
[2] TehS Y, LinR, HungL H and LeeA P 2008 Droplet
microuidics Lab Chip 8 198220
[3] MalicL, BrassardD, VeresT and TabrizianM 2010
Integration and detection of biochemical assays in digital
microuidic LOC devices Lab Chip 10 41831
[4] ChoS K, MoonH and KimC J 2003 Creating, transporting,
cutting, and merging liquid droplets by electrowetting-
based actuation for digital microuidic circuits J.
Microelectromech. Syst. 12 7080
[5] SistaR, HuaZ, ThwarP, SudarsanA, SrinivasanV,
EckhardtA, PollackM and PamulaV 2008 Development of
a digital microuidic platform for point of care testing Lab
Chip 8 2091104
[6] DuboisP, MarchandG, FouilletY, BerthierJ, DoukiT,
HassineF, GmouhS and VaultierM 2006 Ionic liquid
droplet as e-microreactor Anal. Chem. 78 490917
[7] MillerE M and WheelerA R 2008 A digital microuidic
approach to homogeneous enzyme assays Anal. Chem.
80 16149
[8] MoonH, WheelerA R, GarrellR L, LooJ A and KimC J
2006 An integrated digital microuidic chip for multiplexed
proteomic sample preparation and analysis by MALDI-MS
Lab Chip 6 12139
[9] MalicL, VeresT and TabrizianM 2009 2D droplet-based
surface plasmon resonance imaging using electrowetting-
on-dielectric microuidics Lab Chip 9 4735
[10] NadI B, YangH, ParkP S and WheelerA R 2008 Digital
microuidics for cell-based assays Lab Chip
8 51926
[11] LedererT, StehrerB P, BauerS, JakobyB and HilberW 2011
Utilizing a high fundamental frequency quartz crystal
resonator as a biosensor in a digital microuidic platform
Sensors Actuator A 172 1618
[12] ZhangH and KimE S 2005 Micromachined acoustic
resonant mass sensor J. Microelectromech. Syst.
14 699706
[13] ZieglerC 2004 Cantilever-based biosensors Anal. Bioanal.
Chem. 379 946–59
[14] WeinbergM S, DubéC E, PetrovichA and ZapataA M 2003
Fluid damping in resonant exural plate wave device
J. Microelectromech. Syst. 12 56776
[15] PangW, YanL, ZhangH, YuH, KimE S and TangW C
2006 Femtogram mass sensing platform based on lateral
extensional mode piezoelectric resonator Appl. Phys. Lett.
88 243503
[16] XuW, TamijaniA A and ChaeJ 2009 Oscillating behavior of
quality factor of a lm bulk acoustic resonator in liquids
Proc. 15th Int. Conf. Solid-State Sensors, Actuators, and
Microsystems (Denver, CO) pp 696–9
[17] ZhangH, PangW and KimE S 2011 Miniature high-
frequency longitudinal wave mass sensors in liquid
IEEE Trans. Ultrason. Ferroelectr. Freq. Control
58 2558
[18] WeberJ, AlbersW M, TuppurainenJ, LinkM, GablR,
WersingW and SchreiterM 2006 Shear mode FBARs as
highly sensitive liquid biosensors Sensors Actuator A
128 848
[19] PangW, ZhaoH, KimE S, ZhangH, YuH and HuX 2012
Piezoelectric microelectromechanical resonant
sensors for chemical and biological detection
Lab Chip 12 2944
[20] LinA, YuH, WatersM S, KimE S and GoodmanS D 2008
Explosive trace detection with FBAR-based sensor
Proc. IEEE 21th Int. Conf. MEMS (Tucson, AZ)
pp 208–11
[21] QiuX T, TangR, ChenS J, ZhangH, PangW and YuH Y
2001 pH measurements with ZnO based surface acoustic
wave resonator Electrochem. Commun. 13 488–90
[22] Qiu XiT, TangR, ZhuJ, OilerJ, YuC J, WangZ Y and YuH Y
2011 The effects of temperature, relative humidity and
reducing gases on the ultraviolet response of ZnO based
lm bulk acoustic-wave resonator Sensors Actuator B
151 3604
[23] ZhangH, MarmaM S, BahlS K, KimE S and
McKennaC E 2007 Sequence specic label-free DNA
sensing using lm-bulk-acoustic-resonators
IEEE Sensors J. 7 15878
[24] YanZ et al 2004 ZnO-based lm bulk acoustic resonator for
high sensitivity biosensor applications Appl. Phys. Lett.
90 143503
[25] NirschlM, BlüherA, ErlerC, KatzschnerB,
Vikholm-LundinI, AuerS, VörösJ, PompeW, SchreiterM
and MertigM 2009 Film bulk acoustic resonators for DNA
and protein detection and investigation of in vitro bacterial
S-layer formation Sensors Actuator A 156 1804
J. Micromech. Microeng. 25 (2015) 025002
M Zhang et al
[26] KoenigB and GraetzelM 1994 A novel immunosensor for
herpes viruses Anal. Chem. 66 3414
[27] IlicB, YangY and CraigheadH G 2004 Virus detection using
nanoelectromechanical devices Appl. Phys. Lett.
85 26046
[28] RubyR, LarsonJ, FengC and FazzioS 2005 The effect
of perimeter geometry on FBAR resonator electrical
performance Proc. IEEE Microwave Symp. Digest (Long
Beach, CA) pp 217–20
[29] RubyR 2007 Review and comparison of bulk acoustic wave
FBAR, SMR technology Proc. IEEE Ultrasonics Symp.
(New York, NY) pp 1029–40
[30] HongJ S, KoS H, KangK H and KangI S 2008 A numerical
investigation on ac electrowetting of a droplet Microuids
Nanouids 5 26371
[31] JohnstonM L, KymissisI and ShepardK L 2010 FBAR-
CMOS oscillator array for mass-sensing applications IEEE
Sensors J. 10 10427
[32] YiU C and KimC J 2006 Characterization of electrowetting
actuation on addressable single-side coplanar electrodes J.
Micromech. Microeng. 16 20539
[33] LifsonM A, RoyD B and MillerB L 2014 Enhancing the
detection limit of nanoscale biosensors via topographically
selective functionalization Anal. Chem. 86 1016−22
J. Micromech. Microeng. 25 (20 15) 025002
... This brought forth a game-changing, miniaturized platform capable of highly automated operations [1,2]. Some examples include the transportation and subsequent mixing and splitting of biological samples for detection, and the conversion of biological recognition events into detectable electrical output signals [3][4][5][6]. LOC technology also plays a significant role in the development of point of care (POC) diagnostics, allowing complex laboratory procedures to be performed by end-users themselves with cheaper and easier access for all [7]. Some examples of these LOC technologies are label-free biosensors [8,9] and/or microfluidic actuation devices [10][11][12][13]. ...
... where v p is the phase velocity of the acoustic wave, µ q is the elastic constant, and ρ q is the density of the quartz crystal. Equation (4) shows that the resonant frequency of QCM depends on the thickness of the crystal and can provide information about material-specific parameters. The typical operating frequency range of the QCM is 5 MHz to 30 MHz [62]. ...
... The sensing of anticancer drug Imatinib was reported by utilizing the aptamer as a bioreceptor coated on the ZnO film surface. Similar integrated sensing and droplet actuation has been reported on a digital microfluidic platform utilizing electrowetting on dielectric (EWOD) for actuation and FBAR sensor for biosensing [4]. The same piezoelectric AlN material has been shared as a dielectric for EWOD and as an active layer for a FBAR biosensor. ...
Full-text available
Lab-on-a-chip (LOC) technology has gained primary attention in the past decade, where label-free biosensors and microfluidic actuation platforms are integrated to realize such LOC devices. Among the multitude of technologies that enables the successful integration of these two features, the piezoelectric acoustic wave method is best suited for handling biological samples due to biocompatibility, label-free and non-invasive properties. In this review paper, we present a study on the use of acoustic waves generated by piezoelectric materials in the area of label-free biosensors and microfluidic actuation towards the realization of LOC and POC devices. The categorization of acoustic wave technology into the bulk acoustic wave and surface acoustic wave has been considered with the inclusion of biological sample sensing and manipulation applications. This paper presents an approach with a comprehensive study on the fundamental operating principles of acoustic waves in biosensing and microfluidic actuation, acoustic wave modes suitable for sensing and actuation, piezoelectric materials used for acoustic wave generation, fabrication methods, and challenges in the use of acoustic wave modes in biosensing. Recent developments in the past decade, in various sensing potentialities of acoustic waves in a myriad of applications, including sensing of proteins, disease biomarkers, DNA, pathogenic microorganisms, acoustofluidic manipulation, and the sorting of biological samples such as cells, have been given primary focus. An insight into the future perspectives of real-time, label-free, and portable LOC devices utilizing acoustic waves is also presented. The developments in the field of thin-film piezoelectric materials, with the possibility of integrating sensing and actuation on a single platform utilizing the reversible property of smart piezoelectric materials, provide a step forward in the realization of monolithic integrated LOC and POC devices. Finally, the present paper highlights the key benefits and challenges in terms of commercialization, in the field of acoustic wave-based biosensors and actuation platforms.
... Since the trapping limit is proved to be reciprocal to f 4 , the GHz working condition allows the resonator to trap particles down to the nanoscale. The pentagon shape of the GHz BAW resonator is used for spurious resonances suppression [38]. ...
Addressable trapping and manipulations of micro/nano-scale bioparticles is often necessary and critically important in microfluidic devices for biological and medical applications. While GHz bulk-acoustic-wave (BAW) resonators have shown their excellent performance in trapping, rotating, and focusing particles down to nanometers, arraying them for addressable manipulation is a great challenge due to the crosstalk among resonators. This paper presents the electrical and mechanical crosstalk analysis to GHz BAW array with proposed circuit-level models. Crosstalk mechanisms are thoroughly revealed, which contribute to clarify the design rules of GHz BAW array with low crosstalk. A 4 x 4 BAW array chip is designed and experimental results show the chip can achieve addressable control with negligible crosstalk, demonstrating the rationality of the analysis. The crosstalk analysis, as well as the derived design rules, will lead to the design of a high-performance GHz BAW array with great potential for applications that require massive, addressable, and precise particle manipulation. [2021-0236]
... The reacted protein-NPs solution direct injection into the AFT with a pipette is shown in Figure S4 of Supporting Information. Besides, the microfabricated acoustic resonator can be integrated with an electrowetting-driven digital microfluidic chip, 66,67 wherein the bioassay can be conducted in a small droplet (see Supporting Information, Figure S5). All these features show the promise of the AFT biosensor as a versatile tool for POCT medical diagnosis strategies and biomedical instrumentation. ...
Full-text available
We present a nanoscale acoustofluidic trap (AFT) which manipulates nanoparticles in a microfluidic system actuated by a gigahertz acoustic resonator. The AFT generates independent standing closed vortices with high-speed rotation. By carefully designing and optimizing the geometric confinements, the AFT is able to effectively capture and enrich sub-100 nm nanoparticles with low power consumption (0.25~5 μW/μm2) and rapid trapping (within 30 s), showing greatly enhanced particle operating ability towards its acoustic and optical counterparts. Using specifically functionalized nanoparticles (SFNPs) to selectively capture target molecules from the sample, the AFT produces a molecular concentration enhancement of ~200 times. We investigated the feasibility of the SFNPs-assisted AFT preconcentration method for biosensing applications, and successfully demonstrated its capability for serum prostate specific antigen (PSA) detection. The AFT is prepared with a fully CMOS-compatible process, and thus can be conveniently integrated on a single chip, with potential for “lab-on-a-chip” or point-of-care (POC) nanoparticle-based biosensing applications.
... 25 Having two frequenciesone that is mass sensitive and one that is mass insensitivein this case provides each sensor with its own reference resonant frequency to extract the frequency shift, and eventually the amount of target molecules detected. To demonstrate the viability of the split resonances for FBAR sensors, the real-time sensing in liquid environments using microfluidics 26 and biological sensing experiments 27 should be considered for future works. ...
Full-text available
A self-referenced resonator consisting of two distinct areas of the top electrode made from Mo and a thin (5-30 nm) functional Au layer is shown. The fundamental frequencies for both the shear (~1 GHz) and longitudinal (~2 GHz) are split in two, such that mass attachment on the functional layer region causes frequency shifts in only one of the resonances, allowing a new approach of using the difference between the two frequencies to be used to measure mass attachment; this reduces the importance of device-to-device variability in absolute resonant frequency as a result of device fabrication.
... An alternative approach to suppressing the transverse mode is to modulate the mechanical boundary conditions of the resonator, thus mitigating the coupling coefficient of the transverse mode. Zhang protrusion structures for the suppression ( Zhang et al. 2015a). Finite element analysis verifies that the use of these structures effectively restrains the transverse modes, and the measured electrical performance of the LWR with protrusions demonstrates a reduction in the spurious response of 11 dB. ...
Digital microfluidics (DMF) is a platform that enables highly reconfigurable and automated fluidic operations using a generic device architecture. A unique hallmark of DMF is its "flexibility": a generic device design can be used and reused for many different, divergent fluidic operations. The flexibility of DMF is compromised when devices are permanently modified with embedded sensors. Here we introduce a solution to the "flexibility gap" between fluidic operations in digital microfluidics and embedded sensors: "plug-n-play DMF" (PnP-DMF). In PnP-DMF, devices are designed to allow for rapid and seamless exchange of sensors depending on the application needs. This paper provides "proof of concept" for PnP-DMF using commercial biosensors for glucose and β-ketone, a custom paper-based electrochemical sensor for lactate, and a generic screen-printed electroanalytical cell. We demonstrate that hot-swapping sensors between experiments allows for convenient implementation of complex processes such as automated analysis of blood samples by standard addition. Finally, we explored the suitability for using PnP sensors in tandem with other sensing modalities, combining biosensor-based electrochemical measurement of glucose with a chemiluminescent magnetic bead-based sandwich immunoassay for insulin. The latter is notable, as it constitutes the first report of an analysis of different analytes in both the supernatant and precipitate from a single sample-aliquot in a microfluidic device. The results presented here highlight the versatility of PnP-DMF, illustrating how it may be useful for a wide range of applications in diagnostics and beyond.
For point-of-care applications, integrating sensors into a microfluidic chip is a nontrivial task, since conventional detection modules are bulky and microfluidic chips are small in size, and their fabrication processes are not compatible. In this work, a solid-state microfluidic chip with on-chip acoustic sensors using standard thin-film technologies is introduced. The integrated chip is essentially a stack of thin films on silicon substrate, featuring compact size, electrical input (fluid control) and electrical output (sensor read-out). These features all contribute to portability. In addition, by virtue of processing discrete micro-droplets, the chip provides a solution to the performance degradation bottleneck of acoustic sensors in liquid-phase sensing. Label-free immunoassays in serum are carried out and the viability of the chip is further demonstrated by result comparison with commercial ELISA in prostate-specific antigen sensing experiments. The solid-state chip is believed to fit specific applications in personalized diagnostics and other relevant clinical settings where instrument portability matters.
This chapter reports on the state of the art of piezoelectric micro-/nano-mechanical devices in frequency control and sensing applications. Recent studies on bulk acoustic wave (BAW) devices are introduced, including investigation of high-coupling materials and filter and oscillator designs. A novel class of frequency devices based on Lamb waves is also reviewed. Micro- and nano-mechanical sensors for various sensing applications and integrated module are outlined.
Full-text available
This letter introduces a resonant mass sensor that is based on a lateral extensional mode piezoelectric resonator and has a minimum detectable mass of 10−15 g at room temperature and atmospheric pressure. The resonator with size of about 200×50×1 μm3 has a quality factor (Q) of >1400 at 60 MHz, and as small as 0.1 ppm shift of its resonant frequency can be detected. The 0.1 ppm detection capability corresponds to a mass uncertainty of only about 4.6 fg. We have experimentally demonstrated a minimum detectable frequency shift of ∼ 1.6 ppm due to absorption of isopropanol vapor (73 fg) on the sidewalls of the parylene-coated lateral extensional mode piezoelectric resonator.
Full-text available
Zinc oxide (ZnO)-based film bulk acoustic resonator consisting of a piezoelectric element (Au/ZnO/Pt) and a Bragg reflector (ZnO/Pt multilayer structure) has been fabricated by magnetron sputtering. The transmission electron microscopy and x-ray diffraction measurements revealed that all thin film layers in the device were well crystallized and highly textured. By electrical measurements, it was found that the device had a high resonant frequency (3.94 GHz) and mass sensitivity (8970 Hz cm2/ng). The use of the device as a biosensor was demonstrated by comparing the resonant properties of the device with/without coatings of biospecies.
Full-text available
We numerically analyze the AC electric field around a droplet placed on an insulator-covered electrode. The time-averaged effective electrical wetting tension, which is a function of AC frequency, is computed by integrating the Maxwell stress. The computed wetting tension is compared with the experimental result converted from the separately obtained contact-angle data. There is a good agreement between the two results at a low-frequency range and a qualitative agreement at a high-frequency range. Interestingly, the numerical results show that the electric-field strength decreases remarkably in the insulating layer near the TCL as the AC frequency increases. This decrease may account for the delay of the dielectric breakdown of an insulating layer in the AC case, which could be related to the contact-angle saturation phenomenon.
Nanoscale biosensors have remarkable theoretical sensitivities but often suffer from suboptimal limits of detection in practice. This is in part because the sensing area of nanoscale sensors is orders of magnitude smaller than the total device substrate. Current strategies to immobilize probes (capture molecules) functionalize both sensing and nonsensing regions, leading to target depletion and diminished limits of detection. The difference in topography between these regions on nanoscale biosensors offers a way to selectively address only the sensing area. We developed a bottom-up, topographically selective approach employing self-assembled poly(N-isopropylacrylamide) (PNIPAM) hydrogel nanoparticles as a mask to preferentially bind target to only the active sensing region of a photonic crystal (PhC) biosensor. This led to over an order of magnitude improvement in the limit of detection for the device, in agreement with finite element simulations. Since the sensing elements in many nanoscale sensors are topographically distinct, this approach should be widely applicable.
This paper investigated pH measurements using ZnO based surface acoustic wave resonator (SAW). The resonant frequency of the SAW decreased as pH value changed from 7 to 2 (acid region) or from 7 to 12 (alkaline region). The detection limits were 0.03 and 0.02 pH change, respectively, which were comparable to commercial pH meters. The interaction between hydronium (H3O+) or hydroxide (OH−) and ZnO was proposed to be responsible for the frequency drop. Both hydronium and hydroxide can increase the conductivity of the ZnO film, resulting in the resonant frequency decrease due to the acoustoelectric effect.
This study investigated the influence of temperature, relative humidity and reducing gases on the ultraviolet (UV) response of ZnO based film bulk acoustic-wave resonator (FBAR). As temperature increased, the UV response of the FBAR degraded. This was attributed to the softening of the ZnO film with increasing temperature. Water molecules can replace adsorbed oxygen on the ZnO surface. At high relative humidity, more oxygen was replaced by water. In this way, the density of the ZnO film increased and less oxygen was left on the surface to be desorbed by UV, both of which contributed to a lower UV response. Reducing gases, such as acetone, can react with the surface adsorbed oxygen and reduce the density of the ZnO film, resulting in UV response degradation.
We have used a resonating mechanical cantilever to detect immunospecific binding of viruses, captured from liquid. As a model virus, we used a nonpathogenic insect baculovirus to test the ability to specifically bind and detect small numbers of virus particles. Arrays of surface micromachined, antibody-coated polycrystalline silicon nanomechanical cantilever beams were used to detect binding from various concentrations of baculoviruses in a buffer solution. Because of their small mass, the 0.5 μm×6 μm cantilevers have mass sensitivities on the order of 10−19 g∕Hz, enabling the detection of an immobilized AcV1 antibody monolayer corresponding to a mass of about 3×10−15 g. With these devices, we can detect the mass of single-virus particles bound to the cantilever. Resonant frequency shift resulting from the adsorbed mass of the virus particles distinguished solutions of virus concentrations varying between 105 and 107 pfu∕ml. Control experiments using buffer solutions without baculovirus showed small amounts (<50 attograms) of nonspecific adsorption to the antibody layer.
Conference Paper
We describe the first four Lamb-Rayleigh modes seen in an FBAR resonator. We also describe the effect of apodization (non parallel edges) have on the kx-ky space as compared to square resonators.