This article has been accepted for inclusion in a future issue of this journal.
Content is final as presented, with the exception of pagination.
JOURNAL OF MICROELECTROMECHANICAL SYSTEMS
Effects of DC Bias Tuning on Air-Coupled
PZT Piezoelectric Micromachined
Ultrasonic Transducers
Yuri Kusano , Qi Wang , Student Member, IEEE, Guo-Lun Luo, Yipeng Lu, Member, IEEE, Ryan Q. Rudy,
Ronald G. Polcawich, Senior Member, IEEE, and David A. Horsley, Senior Member, IEEE
Abstract— In this paper, we report an air-coupled piezoelectric
ultrasonic micromachined transducer (PMUT) using a lead-
zirconate-titanate (PZT) piezoelectric layer. A dc bias voltage
applied to the PZT film controls its polarization and intrinsic
stress, tuning the frequencies of two closely-spaced resonance
modes of the rectangular shaped PMUT. At an optimal bias
voltage of approximately 5 V, the modes nearly overlap at
230 kHz, increasing the bandwidth by a factor of 8 relative to
the zero dc biased state. Measurements of the electromechanical
coupling coefficient of the PZT film show that it is maximized at
5-6 V, agreeing with device performance experiments. Acoustic
transmission and reception were demonstrated using two identi-
cal PMUTs by adding the optimum dc bias of 6 V to the 1.4 V
peak-to-peak ac voltage which is maximum within the linear
displacement regime. The signal detectable range of the transmit
and receive measurement with optimum dc bias tuning was 4 cm
to 19 cm in air. [2017-0223]
Index Terms— Microelectromechanical devices, piezoelectric
transducers, ultrasonic, PMUT, PZT, dc bias.
I. I NTRODUCTION
U LTRASONIC transducers are widely used in various
applications including medical imaging, nondestructive
evaluation, object recognition, and range-finding [1], [2].
Based on MEMS technology, micromachined ultrasonic trans-
ducers (MUTs) have many advantages over conventional ultra-
sonic transducers such as miniature size, low cost, low power
consumption, wafer-scale fabrication, ability to create 1-D and
2-D array structures, and easy integration with supporting
electronics.
The capacitive-MUT (CMUT) is one of the best-known
MUTs. In a CMUT, the electromechanical coupling from input
Manuscript received September 14, 2017; revised November 10, 2017;
accepted January 5, 2018. This work was supported by Berkeley Sensor and
Actuator Center Industrial Members. Subject Editor R. Pratap. (Corresponding
author: Yuri Kusano.)
Y. Kusano is with the Department of Electrical and Computer Engineering,
University of California at Davis, Davis, CA 95616 USA (e-mail: ykusano@
ucdavis.edu).
Q. Wang, G.-L. Luo, and D. A. Horsley are with the Department of
Mechanical and Aerospace Engineering, University of California at Davis,
Davis, CA 95616 USA.
Y. Lu was with the University of California at Davis, Davis, CA 95616
USA. He is now with Qualcomm, Santa Clara, CA 94720 USA.
R. Q. Rudy and R. G. Polcawich are with the U.S. Army Research
Laboratory, Adelphi, MD 20783 USA.
Color versions of one or more of the figures in this paper are available
online at http://ieeexplore.ieee.org.
Digital Object Identifier 10.1109/JMEMS.2018.2797684
1057-7157 © 2018 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.
See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
1
voltage to output pressure depends on the bias voltage and
the inverse square of the capacitor gap. Therefore, CMUTs
achieve high electromechanical coupling when the capacitor
gap is small, on the order of 100 nm to 200 nm. However,
air-coupled operation requires much larger gaps, on the order
of 1 to 3 microns, in order to achieve sufficient sound
pressure output. Maintaining high electromechanical coupling
in an air-coupled CMUT therefore typically requires high bias
voltages of 150 V or more [3]. For example, recent work
on air-coupled CMUTs with perforated membranes achieved
improved performance and bandwidth by biasing at 250 V [4].
In contrast, piezoelectric-MUTs (PMUTs) [5] can be uti-
lized without high bias voltages, resulting in simpler electronic
interfaces. A typical PMUT device vibrates in a flexural
mode to emit ultrasound into the surrounding environment,
air or fluid. Taking advantage of improved availability of high-
quality piezoelectric films, PMUTs have recently been demon-
strated for many applications, for instance, range-finding [6],
low-power 3-D ultrasonic imaging [7], and ultrasonic finger-
print sensors [8]. Among many piezoelectric MEMS devices,
lead-zirconate-titanate (PZT) and aluminum nitride (AlN) are
the two most commonly-used piezoelectric materials [9], [10].
While the low dielectric constant of AlN makes it the best
material for receiving ultrasound, PZT’s much higher piezo-
electric coefficient makes it the best material for transmitting
ultrasound. Here, we study PMUTs fabricated from thin-film
PZT with the intent to reduce the required transmit voltage
from ∼30 V [7] to below 10 V, reducing the cost and
complexity of the supporting electronics.
A second goal of this work is to increase the PMUT’s band-
width. In pulse-echo range-finding, the transducer’s bandwidth
contributes to several important performance parameters:
(1) the minimum duration of the transmitted pulse which in
turn determines the minimum measurement range (since the
transducer cannot receive while it is transmitting); (2) the
axial resolution, which is the minimum separation between two
resolved objects; and (3) the RMS range noise [7]. In addition,
transducers with wide bandwidth enable more sophisticated
signal processing schemes such as pulse encoding (e.g. chirped
transmissions).
Fluid-coupled PMUTs used at high resonance frequen-
cies (>5 MHz) often have high bandwidth (>30%) because
fluid damping is high compared to air damping. Additionally,
the higher speed of sound in fluid means that the size of a
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

2 JOURNAL OF MICROELECTROMECHANICAL SYSTEMS

TABLE I
Y OUNG ’s M ODULUS AND L AYER T HICKNESSES OF PMUT D EVICE

fluid-coupled PMUT is usually large relative to the transmitted

wavelength. These factors allow efficient coupling in immersed
PMUTs [11], [12]. On the contrary, PMUTs operating in air
require lower frequencies (<1 MHz) and are usually smaller
than the wavelength (λ = 1.9 mm at 175 kHz), resulting
in poor coupling and low bandwidth (1%–5%) [13]. An air-
coupled PMUT with 10% bandwidth was reported [14], but
required a Helmholtz resonant tube etched into the backside
of MUT to achieve this result. Another method to extend
the bandwidth is to use multiple transducers with closely-
spaced natural frequencies so that the composite bandwidth
is the sum of the individual transducer bandwidths; this
approach has been demonstrated with an array that contains
five different membranes of slightly different dimensions [15].
Here, we use a rectangular membrane design and, through
appropriate selection of the length-to-width ratio, demonstrate Fig. 1. (a) Optical image, (b) schematic top view and (b) cross-section of
that two vibration modes with closely-spaced resonance fre- the air-coupled PZT PMUT.
quencies can be achieved in a single membrane [2]. Because
these modes have different sensitivities to piezoelectric stress,
we are able to tune the frequency spacing of the two modes by
adjusting a dc bias voltage applied across the PZT layer. At the
optimum bias point, a 12% bandwidth is achieved without the
need for a Helmholtz resonator.

II. PMUT D ESIGN

An optical top view image, a schematic top view and a
schematic cross-section of a 1200 μm × 230 μm rectangular
shaped PMUT is shown in Fig. 1(a)-(c). Table 1 lists the
elastic modulus and layer thicknesses of the materials in the
device [16]–[18]. Note that only the inner top electrodes were
Fig. 2. PMUT fabrication process flow: deposition and patterning of (a) PZT,
used in this work. bottom electrode, and top electrode, (b) Al2 O3 and Au/Pt/Cr, (c) sacrificial
photoresist and thick Au layer, (d) photoresist removal and Si release through
A. Fabrication frontside XeF2 etching.

A simplified fabrication process flow is shown in Fig. 2.

The fabrication process starts by growing a 500 nm layer of
SiO2 on a silicon wafer via thermal oxidation. Titanium is
sputtered with a thickness of 30 nm and oxidized in a furnace
to create titanium dioxide which creates a good template for
the subsequently sputtered, (111)-textured platinum electrode
with a thickness of 100 nm [19]. The oriented platinum
forms a templating layer for a 475 nm thick, chemical-
solution deposited, (001)-textured PZT film with a Zr/Ti ratio
of 52/48 [20]. The PZT is then capped with a 100 nm thick
iridium oxide electrode that is sputter deposited and furnace
annealed in oxygen at 650 °C for 30 minutes. Next, the actu-
ator stack is patterned using three argon ion milling steps.
In the first etch, the iridium oxide is patterned to define the
top electrode of the device. The second etch removes the PZT
and bottom platinum from the field, stopping on the silicon
dioxide layer. The bottom electrode is accessed by a third ion-
mill step followed by a PZT wet-etch using a combination of
HCl:HF:H2 O (120:1:240) to remove any residual PZT. The
wafers are annealed at 500°C in an oxygen environment to
mitigate any physical etch and hydrogen damage from the ion
milling steps.
With the actuator stack patterned, a tri-layer Au/Pt/Cr
(730nm/20nm/20nm) is deposited using electron beam evap-
oration and patterned via liftoff, Fig. 2(b). This gold layer is
used for contact pads for electrical probe testing. The silicon
dioxide layer is then patterned by reactive ion etching to define
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

KUSANO et al.: EFFECTS OF DC BIAS TUNING ON AIR-COUPLED PZT PMUTs 3

Fig. 3. SEM image of (a) top view (b) cross-section of the air-coupled
PZT PMUT.

the etch holes for release. Next, a sacrificial photoresist layer

is defined, and a thick gold layer is deposited via electron
beam evaporation and liftoff, Fig. 2(c). This gold layer bridges
from the top electrode to the metal interconnect on the silicon
dioxide in the field without shorting the top and bottom
electrodes. The sacrificial photoresist layer is removed with an
oxygen plasma to form the Au air bridge. Finally, the wafer
is exposed to XeF2 , which isotropically etches the underlying
silicon, releasing the device, Fig. 2(d).
Fig. 3 shows SEM images of the top view and cross-section
of the PMUT following XeF2 release. As shown in the figure,
the top electrode is divided into two sections, separated by
etch-holes that perforate through the PZT/SiO2 layer. The total
3 μm thick gold layer at the left and right edge of the PMUT
stiffens these edges such that variations in the Si undercut in
the XeF2 etch have minimal effect on the PMUT’s resonance
frequency [10]. Small etch holes were designed along the
center of the membrane to enable the membrane to be released
through a front side etch, since this process allows closer
spacing of PMUTs than processes that require a through-wafer
etch [21], [22]. As demonstrated in recent work on air-coupled
CMUTs, small perforations in the MUT membrane have a
negligible effect on the output pressure [23].
B. PZT Film Characterization
The dielectric constant, εr33 , and corresponding loss tan-
gent, tan δ, of the PZT film were measured as a function
of bias voltage over a range from 0V to 10 V, Fig. 4(a).
The transverse thin-film piezoelectric coefficient e31, f and
dielectric constant εr33 are compared in Fig. 4(b). The laser
Doppler vibrometer (LDV) was used to measure the deflection
of a cantilever beam under varying bias voltage. The change
in PZT stress that would be necessary to create the deflection-
bias profile was extracted, and then e31, f was determined
by differentiating with respect to the electric field. Note that
the film reaches maximum piezoelectric coefficient at 3 V,
corresponding to an electric field of 63 kV/cm. At voltages
above 5V the piezoelectric coefficient is seen to decrease.
The dielectric constant is maximum at 1 V, after which
it decreases steadily with voltage. The voltage-dependence
of the dielectric constant, dielectric loss, and piezoelectric
Fig. 4. (a) Measured bipolar dielectric constant εr33 and tan δ ver-
sus bias voltage demonstrating the polarization switching. (b) Comparing
transverse piezoelectric coefficient e31, f and dielectric constant εr33 of
the PZT 52/48 thin film versus bias voltage. (c) Calculated figure of
merit (e31, f )2 /ε0 εr33 .
coefficient is inherent to all ferroelectric materials and is
related to the polarization reversal mechanisms within the
ferroelectric material. In particular, e31, f diminishes as the
voltage is increased above 3V. While this reduction is due
to many factors, one important factor is that the piezoelectric
coefficient is proportional to the dielectric constant. Once the
film is fully polarized, reductions in the dielectric constant
tend to reduce the piezoelectric coefficients. [24], [25]. As the
nonlinear relationship with voltage is unique to ferroelectric
materials such as PZT, non-ferroelectric piezoelectric materials
such as AlN will not exhibit these characteristics. Moreover,
the ability to tune the bandwidth by modulating the stress in
the piezoelectric films through bias voltage will be limited.
A figure of merit (FOM) for transducer performance is the
transverse electromechanical coupling coefficient kt2 , which is
proportional to (e31, f )2 /ε0 εr33 [GPa] [26]. The FOM, plotted
in Fig. 4(c), reaches a maximum of 8 GPa at 6 V, and
is roughly constant over the range from 5 V to 8 V. The
peak FOM does not occur at the voltage that maximizes the
piezoelectric coefficient. Instead, it peaks at a much higher
voltage where the dielectric constant is lower. Note that the
values in Fig. 4 were collected from a particular wafer which
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.
4 JOURNAL OF MICROELECTROMECHANICAL SYSTEMS
Fig. 5. Measured I-V curve of a PMUT device with a 0.475 μm thickness
PZT film, where breakdown voltage was 35 V and breakdown field was
760 kV/cm. The inset optical image shows the top view of Au air-bridge
collapse that occurred at the breakdown.
Fig. 6. Measured profile of the PMUT under dc bias voltages of 0 V, +6 V,
and +10 V. The plot shows the cross-section height profile across the width
of the rectangular PMUT membrane.
contained the devices used for the following characterizations.
The same FOM trends were observed among the wafers,
although the FOM for this specific wafer was not the highest.
Current-voltage (I-V) measurements were conducted to
identify the breakdown voltage of the PZT film. The break-
down voltage was measured to be 35 V, corresponding to a
breakdown field of 760 kV/cm as Fig. 5 shows. The inset
of Fig. 5 shows that, when the applied voltage reaches the
breakdown, the Au bridge interconnecting top electrode and
the probing pad collapsed.
The static shape of the PMUT membrane was measured
using a Wyko optical profilometer system. The cross-section
height profile across the shorter width of the rectangular
PMUT membrane is shown in Fig. 6, where the height at
SiO2 surface is set to zero. At 0 V bias, the center of the
membrane is deflected 3.6 μm above the membrane anchors
due to the large residual stress gradient through the composite
stack [27], [28]. The membrane’s curvature is reduced as the
applied bias voltage is increased and at +10 V the center of
the membrane moves downwards by approximately 1.6 μm
relative to the unbiased state.
Fig. 7. Frequency response of the PZT PMUT measured via LDV in air
with dc bias voltages of 0 V, +6 V, and +10 V. The inset shows a quarter of
the membrane with the two mode shapes from FEM simulation.
III. R ESULTS
A. Frequency Response
The frequency response of the PMUT in air was measured
via LDV showing the presence of two vibration modes with
closely spaced natural frequencies near 180 kHz. The LDV
measurements were done at a single point of the membrane
indicated on Fig. 1(b). The measurement location was chosen
on the top electrode surface to avoid the etching holes and the
transparent SiO2 surface. A dc bias voltage applied to pole
the PZT layer creates in-plane piezoelectric stress that enables
the frequencies of the two modes to be tuned. As the dc bias
voltage was varied between ±10 V, a significant change in
the resonant frequencies, displacement sensitivities, and 3-dB
bandwidth was clearly observed for the two resonance modes.
In Fig. 7, the frequency response of the two resonance modes
with dc bias voltages of 0 V, +6 V, +10 V respectively
are shown. The FEM-simulated mode shapes of the two
modes ((1,1) mode and (1,3) mode) are shown in the inset.
The two modes independently have 2.4 kHz bandwidth with
displacement sensitivity of approximately 2000 nm/V with
small ac voltage (−10 dBm) and without any dc bias voltage.
As a result of increasing the dc bias to +6 V, the highest
3-dB bandwidth of 20.3 kHz with displacement sensitivity
of 5000 nm/V were observed. The bias voltage of 6 V
coincides with the value when the kt2 is maximized, Fig. 4(b).
For a rectangular membrane clamped at all four boundaries,
the stress-free resonance frequency of the (n, m) mode f nm can
be calculated from [29] and [30]:
   
πt E m 2  n 2
f nm =   + (1)
4 6 1 − υ2 ρ a b
where t is the total membrane thickness, E, ν, and ρ are
the average Young’s modulus, Poisson’s ratio and mass den-
sity, and a, b are the length and width of the membrane
(a  b). Here, we used a = 1200 μm, b = 230 μm
resulting in f 11 = 121 kHz and f13 = 136 kHz. FEM
simulation was used to increase the accuracy by modeling the
precise structure, resulting in stress-free frequencies f11 =
134 kHz and f 13 = 155 kHz, corresponding to the mode
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

KUSANO et al.: EFFECTS OF DC BIAS TUNING ON AIR-COUPLED PZT PMUTs 5

shapes shown in the inset of Fig. 7. Because residual stress

has a strong effect on the resonant frequencies of a thin
membrane (∼1 μm thick), simulations were also performed
using estimates of the residual stresses in the membrane layers:
−300 MPa (compressive) in the thermal SiO2 layer, 700 MPa
and 200 MPa (tensile) in the Pt and PZT layers [27], [28].
Under the above stressed condition, the FEM-simulated natural
frequencies increased to f 11 = 175 kHz and f13 = 194 kHz,
close to the experimentally-observed frequencies. Note that,
because the PZT and SiO2 layers have nearly equal thickness,
the contributions of their stresses nearly cancel, and the
dominant effect is residual stress in the 0.1 μm Pt electrode.
We also simulated the stress in the PZT layer required to
achieve the 60 kHz frequency shift observed in dc bias tuning
experiments– a 60 kHz increase in the resonant frequency
was observed when the PZT stress was increased by 40 MPa
in simulation, comparable to the 45 MPa value predicted
using the maximum value of the piezoelectric coefficient
(e31, f,max = 7.5 C/m2 ) multiplied by the electric field at 3V
bias (60 kV/cm).
Fig. 8(a), (b) shows the dependence of the PMUT resonant
frequency and displacement sensitivity when tuning the dc
bias voltage between −10 V and +10 V. Color-filled triangles
correspond to the first resonance mode and open triangles
correspond to the second mode observed. In addition, upward
pointing triangles present the points when sweeping up from
−10 V to +10 V, while downward pointing triangles present
the points when sweeping down from +10 V to −10 V.
The frequencies of the two modes are nearly overlapping
with certain dc bias voltages applied, especially −5 V / +6 V
while increasing and −5 V / +5 V while decreasing. From
the same measurement results, the dependence of the 3-dB
bandwidth on dc bias was analyzed and is shown in Fig. 8(c).
Wide 3-dB bandwidth above 20 kHz was achieved with
dc bias near ±5 V when the two closely-spaced modes
have overlapping bandwidths. For other dc bias voltages, the
3-dB bandwidth was only few kHz to 10 kHz since the two
resonance peaks were separated and only the first resonance
mode’s bandwidth was taken into account. Compared to the
non-biased condition, displacement sensitivity increased by a
factor of 2.5 and the 3-dB bandwidth increased by a factor Fig. 8. Dependence of (a) PMUT resonant frequency and (b) displacement
of 8 through dc bias tuning. sensitivity at resonance of the 1st peak () and the 2nd peak () on dc
bias tuning. (c) Corresponding 3dB bandwidth due to dc bias tuning.

B. Time Response
Increasing the bandwidth by overlapping multiple modes
allows the transmission time to be shorter and hence reduces
the minimum measurable range and improves the axial resolu-
tion in time-of-flight (ToF) measurements [2], [15]. At the bias
conditions where the bandwidth is maximum, the PMUT is
capable of transmitting shorter pulses for ToF measurements,
where the speed of the wave, ultrasound in this case, deter-
mines the distance to the target. The time response in air was
demonstrated from pulsed ring-down measurement via LDV.
The PMUT was excited using a function generator with a
20-cycle square wave at the resonant frequency for the two
different dc bias conditions. The drive voltage amplitude was
200 mVpp for 0 Vdc and 50 mVpp for +5 Vdc to produce
similar displacement. In both cases, unipolar drive voltages
were used to prevent repoling of the PZT thin film. Fig. 9(a)
shows the driving voltage and the PMUT time-response mea-
sured at zero dc bias using the resonant frequency of 185 kHz.
Fig. 9(b) shows the optimum tuning result with +5 V dc bias
using the resonant frequency of 240 kHz. The frequency of the
driving pulse train was increased from 185 kHz to 240 kHz
because the resonance frequency of the PMUT shifted due to
the dc bias applied, as shown in Fig. 8(a). Comparing the two
time-response plots, the decay time constant τ significantly
decreases from 0.28 ms to 0.05 ms.
The pulse transmissions were simulated with a model of
two second-order systems in series, corresponding to the
two resonance modes. This model was chosen based on an
empirical fit of the measured frequency response at the various
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.
6 JOURNAL OF MICROELECTROMECHANICAL SYSTEMS
Fig. 9. Experimentally measured pulsed-response via LDV. The PMUT
device was excited at the resonant frequencies corresponding to (a) 0 V dc
bias and (b) +5 V dc bias, demonstrating five-fold reduction of the decay
time constant.
Fig. 10. Simulated pulse transmission by a model with a cascaded 2nd-order
system representing each of the two resonance modes measured via LDV
at (a) 0 Vdc bias and (b) +5 Vdc bias matching the experimental results
in Fig. 9.
bias voltages, in order to validate that the observed pulse-input
time-response was consistent with the observed changes in the
frequency response. The normalized transfer function of the
cascaded second-order systems can be expressed as
ω12 ω22
Q1 Q2
TF(s) = ω1 × ω2
s2 + Q1 s + ω12 s2 + Q2 s + ω22
where ω1 , ω2 are the two resonant frequencies (in rad/s) and
Q 1 , Q 2 are the corresponding quality factors. Using the quality
factors and resonant frequencies measured in the frequency-
domain LDV measurements (Fig. 7), the PMUT time-response
was simulated using Eq. (2), with the results shown in Fig. 10.
The input pulse waveforms were replicated from the measure-
ments except both amplitudes were normalized to 1. The decay
time-constant of the simulated pulse transmission was 0.3 ms
Fig. 11. Membrane displacement as a function of ac driving voltage
amplitude, demonstrating the linear response up to 1.4 Vpp with fixed dc bias
of 6 V. The inset shows the time response of the membrane displacement for
different ac voltages.
for 0 V dc bias and 0.05 ms for +5 V dc bias, which coincide
with the experimental results.
The approximately five-fold reduction in decay time is equal
to the increase in the bandwidth, from roughly 4 kHz at
0 V bias, to 20 kHz at +5 V bias. The minimum pulse
transmission time is approximately 4τ = 0.2 ms. In pulse-
echo measurements, the first echoes that can be received must
arrive later than the end of pulse transmission. Therefore, this
time defines the minimum range for pulse-echo measurements,
rmin = 4τ c/2 = 3.4 cm at +5 V bias, compared to a minimum
range of 19 cm at 0 V bias.
C. Maximum Linear Operating Range
Ideally, using a short and high-power pulse for the transmit
signal would give the highest ultrasound output. However,
as reported previously, a fully-clamped membrane shows an
undesired limitation in dynamic range due to nonlinearity at
high drive voltages [6], [7]. At the limit of the linear operating
range, increasing the voltage no longer produces a proportion-
ate increase in ultrasound output. Fig. 11 shows the measured
displacement amplitude with various excitation voltages for
the 6 Vdc biased PMUT, with the inset showing some of the
pulsed time-responses. The displacement scales linearly with
voltage up to a maximum displacement amplitude of 1.2 μm
at 1.4 V peak-to-peak (Vpp) ac voltage. Considering this
result, experiments from the next section are using 1.4 Vpp for
driving ac voltage.
(2)
D. Distance Sensor
To demonstrate the range-finding concept [31], [32], pulsed
transmit and receive measurements were performed using two
identical PMUT devices, each wire-bonded to a printed circuit
board (PCB). The basic experimental setup and the result are
shown in Fig. 12. The receive PMUT was placed 4 cm apart
from the transmit PMUT within an acrylic tube to reduce
the noise from the surrounding environment, Fig. 12(a). The
tube was lined with an acoustically-absorbing foam in order to
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.
KUSANO et al.: EFFECTS OF DC BIAS TUNING ON AIR-COUPLED PZT PMUTs
Fig. 12. (a) Experimental setup; transmit (TX) and receive (RX) PMUTs
are 4 cm apart in a foam-lined acrylic tube. (b) Transmitted signal sent
from function generator to TX PMUT, and received signal measured with
RX PMUT, demonstrating the detection of ultrasound pulses on the 1st and
2nd round-trips.
prevent reflections from the sidewall. A burst signal was sent
from a function generator to the transmit PMUT, producing an
ultrasound wave which propagated acoustically to the receive
PMUT which converted the ultrasound pressure to a charge
output. The charge was detected using a low-noise charge
√
amplifier (gain 10 V/pC, noise 90×10−21 C/ Hz) connected
to the receive PMUT and the out-of-band noise was filtered
out by post processing.
Applying 6 V dc bias in addition to 1.4 Vpp drive voltage
at resonance frequency of 230 kHz for 5 pulse transmission
was successfully detected at the receive PMUT, Fig. 12(b).
The first ultrasound burst was received at tr x = 0.116 ms
which matches the travel distance of 343 m/s ×tr x = 3.98 cm.
The detected time of the second echo was three times of tr x
agreeing with the expected travel distance. The estimated
bandwidth calculated from the rise time trise = 0.014 [ms]
was BW = 0.35/trise = 25 [kHz] similar to measured value via
LDV shown in Fig. 7. The ringdown time tringdown = 0.05 [ms]
was equivalent to the transmission decay time in Fig. 9(b).
E. Path Loss and Free-Space Measurements
The ultrasonic wave spreads and attenuates while traveling
through the air. Because it is small relative to the acoustic
wavelength, the PMUT is a nearly-omnidirectional radiator,
causing a linear spreading of pressure with distance [6].
In addition, at 230 kHz, ultrasound is significantly absorbed
by air [33]. From the two factors of spreading and absorption
7
Fig. 13. SNR versus distance between the transmit and receive PMUTs at
230 kHz compared with the theoretical model predicting linear spreading and
absorption loss.
loss, the ratio of the received pressure pr x to the transmitted
pressure pt x is
pr x a
G= = G ac · 10−α D (3)
pt x 4D
where G ac is the acoustic gain (=1), a is the effective
membrane length, D is the distance, α = 3.61 × 10−5 f −
0.985 [dB/m] is the attenuation coefficient, and f is the
frequency of the ultrasound wave. The attenuation coeffi-
cient α increases with frequency and humidity [34]. Under
the condition of 60% relative humidity and room temperature,
α is 7.3 dB/m at 230 kHz. At 20 cm distance, for example, the
spreading term is −77 dB and absorption loss term is −6 dB,
generating total attenuation G of −83 dB.
Fig. 13 shows the measured SNR for free-space transmit-
receive measurements conducted between two PMUTs sepa-
rated by 4 cm to 19 cm. The PMUTs were tuned optimally
to 6 V dc bias at the 230 kHz operating frequency. The
theoretical limit of the detection of 12 dB SNR was nearly
45 cm, however, the maximum range measured was 19 cm due
to the limitation in the experimental setup. The measurements
in this section were conducted in free-space without any foams
to absorb the surrounding noise, resulting in increasing the
path-loss and noise detected at the receive PMUT. On the
other hand, the minimum detectable range was 4 cm in
this experiment thanks to the wide bandwidth and fast time
response achieved through dc bias tuning. For comparison,
the zero biased PMUT could not detect signal at any range
because of the overlap of the transmission signal feedthrough
and the received signal at short range (less than 10 cm), while
at further ranges, the signal was too small to detect.
IV. C ONCLUSION
We have presented a method to increase bandwidth and
reduce transmission time of air-coupled PZT PMUTs. Through
dc bias tuning at ±5 V, where the FOM of electromechan-
ical coupling coefficient, (e31, f )2 /ε0 εr33 , is maximum, two
adjacent resonance modes almost overlapped, resulting in
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.
8 JOURNAL OF MICROELECTROMECHANICAL SYSTEMS
significantly improved displacement sensitivity and 3-dB
bandwidth compared to zero dc bias condition. Consequently,
the time-response measurement results showed that the decay
time constant was reduced nearly five-fold by adding +5 V dc
bias to the driving voltage. An optimum tuning of the dc bias
also increased the transmit and receive sensitivity, allowing
in-air transmit and receive measurement to be performed. The
proposed dc bias tuning of the PMUT achieving wideband
and short pulse demonstrated a good potential to improve the
minimum range and axial resolution for in-air range-finding
applications.
ACKNOWLEDGMENT
The authors acknowledge the contributions of Brian Power
and Joel Martin of the US Army Research Laboratory (ARL),
and Steven Isaacson of General Technical Services.
R EFERENCES
[1] Y. Kusano, Q. Wang, R. Q. Rudy, R. G. Polcawich, and D. A. Horsley,
“Wideband air-coupled PZT piezoelectric micromachined ultrasonic
transducer through DC bias tuning,” in Proc. 30th IEEE Int. Conf. Micro
Electro Mech. Syst. (MEMS), Jan. 2017, pp. 1204–1207.
[2] Y. Lu et al., “Broadband piezoelectric micromachined ultrasonic trans-
ducers based on dual resonance modes,” in Proc. 28th IEEE Int. Conf.
Micro Electro Mech. Syst. (MEMS), Jan. 2015, pp. 146–149.
[3] I. O. Wygant et al., “50 kHz capacitive micromachined ultrasonic
transducers for generation of highly directional sound with parametric
arrays,” IEEE Trans. Ultrason. Ferroelect. Freq. Control, vol. 56, no. 1,
pp. 193–203, Jan. 2009.
[4] N. Apte, K. K. Park, A. Nikoozadeh, and B. T. Khuri-Yakub, “Band-
width and sensitivity optimization in CMUTs for airborne applications,”
in Proc. IEEE Int. Ultrason. Symp., Sep. 2014, pp. 166–169.
[5] P. Muralt and J. Baborowski, “Micromachined ultrasonic transducers
and acoustic sensors based on piezoelectric thin films,” J. Electroceram.,
vol. 12, nos. 1–2, pp. 101–108, 2004.
[6] R. J. Przybyla et al., “In-air rangefinding with an aln piezoelectric
micromachined ultrasound transducer,” IEEE Sensors J., vol. 11, no. 11,
pp. 2690–2697, Nov. 2011.
[7] R. J. Przybyla, H. Y. Tang, A. Guedes, S. E. Shelton, D. A. Horsley, and
B. E. Boser, “3D ultrasonic rangefinder on a chip,” IEEE J. Solid-State
Circuits, vol. 50, no. 1, pp. 320–334, Jan. 2015.
[8] Y. Lu et al., “Ultrasonic fingerprint sensor using a piezoelectric micro-
machined ultrasonic transducer array integrated with complementary
metal oxide semiconductor electronics,” Appl. Phys. Lett., vol. 106,
no. 26, p. 263503, Jun. 2015.
[9] P. Muralt et al., “Piezoelectric micromachined ultrasonic transducers
based on PZT thin films,” IEEE Trans. Ultrason., Ferroelect., Freq.
Control, vol. 52, no. 12, pp. 2276–2288, Dec. 2005.
[10] S. Shelton et al., “CMOS-compatible AlN piezoelectric micromachined
ultrasonic transducers,” in Proc. IEEE Int. Ultrason. Symp., Sep. 2009,
pp. 402–405.
[11] Y. Lu, H.-Y. Tang, S. Fung, B. E. Boser, and D. A. Horsley, “Pulse-
echo ultrasound imaging using an AlN piezoelectric micromachined
ultrasonic transducer array with transmit beam-forming,” J. Micro-
electromech. Syst., vol. 25, no. 1, pp. 179–187, Feb. 2016.
[12] Q. Wang, Y. Lu, S. Mishin, Y. Oshmyansky, and D. A. Horsley, “Design,
fabrication, and characterization of scandium aluminum nitride-based
piezoelectric micromachined ultrasonic transducers,” J. Microelectro-
mech. Syst., vol. 26, no. 5, pp. 1132–1139, Oct. 2017.
[13] M. D. Williams, B. A. Griffin, T. N. Reagan, J. R. Underbrink, and
M. Sheplak, “An AlN MEMS piezoelectric microphone for aeroacoustic
applications,” J. Microelectromech. Syst., vol. 21, no. 2, pp. 270–283,
2012.
[14] S. Shelton, O. Rozen, A. Guedes, R. Przybyla, B. Boser, and
D. A. Horsley, “Improved acoustic coupling of air-coupled microma-
chined ultrasonic transducers,” in Proc. IEEE 27th Int. Conf. Micro
Electro Mech. Syst. (MEMS), Jan. 2014, pp. 753–756.
[15] A. Hajati et al., “Three-dimensional micro electromechanical system
piezoelectric ultrasound transducer,” Appl. Phys. Lett., vol. 101, no. 25,
p. 253101, 2012.
[16] Y. Li and Z. Zeng, “Elastic properties of transition metal dioxides:
XO2 (X = Ru, Rh, Os, andIr),” Int. J. Mod. Phys. C, vol. 19, no. 8,
pp. 1269–1275, Aug. 2008.
[17] S. Yagnamurthy, I. Chasiotis, J. Lambros, R. G. Polcawich,
J. S. Pulskamp, and M. Dubey, “Mechanical and ferroelectric behavior
of PZT-based thin films,” J. Microelectromech. Syst., vol. 20, no. 6,
pp. 1250–1258, Dec. 2011.
[18] A. Adami, M. Decarli, R. Bartali, V. Micheli, N. Laidani, and
L. Lorenzelli, “Mechanical characterization of thin TiO2 films by
means of microelectromechanical systems-based cantilevers,” Rev. Sci.
Instrum., vol. 81, no. 1, p. 15109, Jan. 2010.
[19] D. M. Potrepka, G. R. Fox, L. M. Sanchez, and R. G. Polcawich,
“Pt/TiO2 growth templates for enhanced PZT films and MEMS devices,”
in Proc. MRS, vol. 1299. Jan. 2011, p. mrsf10-1299-s02-04. [Online].
Available: https://doi.org/10.1557/opl.2011.53
[20] L. M. Sanchez et al., “Optimization of PbTiO3 seed layers and Pt
metallization for PZT-based piezoMEMS actuators,” J. Mater. Res.,
vol. 28, no. 14, pp. 1920–1931, 2013.
[21] D. E. Dausch, J. B. Castellucci, D. R. Chou, and O. T. von Ramm,
“Theory and operation of 2-D array piezoelectric micromachined ultra-
sound transducers,” IEEE Trans. Ultrason., Ferroelect., Freq. Control,
vol. 55, no. 11, pp. 2484–2492, Nov. 2008.
[22] F. Akasheh, J. D. Fraser, S. Bose, and A. Bandyopadhyay, “Piezoelec-
tric micromachined ultrasonic transducers: Modeling the influence of
structural parameters on device performance,” IEEE Trans. Ultrason.,
Ferroelect., Freq. Control, vol. 52, no. 3, pp. 455–468, Mar. 2005.
[23] N. Apte, K. K. Park, and B. T. Khuri-yakub, “Experimental evaluation
of CMUTs with vented cavities under varying pressure,” in Proc. IEEE
Int. Ultrason. Symp. (IUS), Jul. 2013, pp. 1724–1727.
[24] A. L. Kholkin, E. K. Akdogan, A. Safari, P.-F. Chauvy, and N. Setter,
“Characterization of the effective electrostriction coefficients in ferro-
electric thin films,” J. Appl. Phys., vol. 89, no. 12, pp. 8066–8073,
Jun. 2001.
[25] N. Setter et al., “Ferroelectric thin films: Review of materials, properties,
and applications,” J. Appl. Phys., vol. 100, no. 5, p. 051606, Sep. 2006.
[26] P. Muralt, “PZT thin films for microsensors and actuators: Where do
we stand?” IEEE Trans. Ultrason., Ferroelect., Freq. Control, vol. 47,
no. 4, pp. 903–915, Jul. 2000.
[27] J. S. Pulskamp, A. Wickenden, R. Polcawich, B. Piekarski, M. Dubey,
and G. Smith, “Mitigation of residual film stress deformation in mul-
tilayer microelectromechanical systems cantilever devices,” J. Vac. Sci.
Technol. B, Microelectron. Process. Phenom., vol. 21, no. 6, p. 2482,
2003.
[28] E. Zakar et al., “Stress analysis of SiO2 /Ta/Pt/PZT/Pt stack for MEMS
application,” in Proc. 12th IEEE Int. Symp. Appl. Ferroelectr. (ISAF),
vol. 2. Jul./Aug. 2000, pp. 757–759.
[29] A. W. Leissa, “Vibration of plates,” NASA, Washington, DC, USA,
Tech. Rep. SP-160, 1969.
[30] A. K. Mitchell and C. R. Hazell, “A simple frequency formula
for clamped rectangular plates,” J. Sound Vibrat., vol. 118, no. 2,
pp. 271–281, Oct. 1987.
[31] R. Przybyla et al., “A micromechanical ultrasonic distance sensor
with >1 meter range,” in Proc. 16th Int. Solid-State Sensors, Actuat.
Microsyst. Conf. (TRANSDUCERS), Jun. 2011, pp. 2070–2073.
[32] M. Shin, J. S. Krause, P. DeBitetto, and R. D. White, “Acoustic
Doppler velocity measurement system using capacitive micromachined
ultrasound transducer array technology,” J. Acoust. Soc. Amer., vol. 134,
no. 2, pp. 1011–1020, Aug. 2013.
[33] D. T. Blackstock, Fundamentals of Physical Acoustics. Hoboken, NJ,
USA: Wiley, 2000.
[34] H. E. Bass, L. C. Sutherland, A. J. Zuckerwar, D. T. Blackstock, and
D. M. Hester, “Atmospheric absorption of sound: Further developments,”
J. Acoust. Soc. Amer., vol. 97, no. 1, pp. 680–683, Jan. 1995.
Yuri Kusano received the B.E. degree in applied
physics from Keio University, Japan, in 2015, and
the M.S. degree in electrical and computer engi-
neering from the University of California at Davis,
Davis, CA, USA, in 2017, where she is currently
pursuing the Ph.D. degree with the Department of
Electrical and Computer Engineering, with a focus
on piezoelectric MEMS devices for ultrasonic sensor
applications. She is currently a Graduate Student
Researcher with the Berkeley Sensor and Actua-
tor Center, University of California at Davis. Her
research interests include piezoelectric MEMS sensors and actuators.
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.
KUSANO et al.: EFFECTS OF DC BIAS TUNING ON AIR-COUPLED PZT PMUTs
Qi Wang (S’15) received the B.S. degree in
mechanical engineering and automation from the
Nanjing University of Aeronautics and Astronautics,
Nanjing, China, in 2011, and the M. S. degree
in nanomechanics from Tohoku University, Sendai,
Japan, in 2013.
He is currently a Graduate Student Researcher
with the Berkeley Sensor and Actuator Center, Uni-
versity of California at Davis, where he is currently
pursuing the Ph.D. degree in mechanical engineering
with the Department of Mechanical and Aerospace
Engineering. His research interests include piezoelectric thin films and MEMS
sensors and actuators.
Guo-Lun Luo received the double-major B.S.
degree of physics/aeronautics and astronautics engi-
neering from National Cheng Kung University
in 2011, and the M.S. degree in mechanical engi-
neering from National Tsing Hua University in 2013.
In 2013, he interned with Qualcomm, Taiwan, for
mirasol display. From 2014 to 2016, he joined
Asia Pacific Microsystems Inc., where he was a
Senior Research &Development Engineer and a
Project Manager for process development of MEMS
devices. He is currently pursuing the Ph.D. degree
in mechanical and aerospace engineering with the University of California at
Davis. His research interests include piezoelectric MEMS devices.
Yipeng Lu received the B.S. degree in materi-
als science and engineering from Jilin University,
Changchun, China, in 2007, the M.S. degree in
microelectronics from Shanghai Jiao Tong Univer-
sity, Shanghai, China, in 2010, and the Ph.D. degree
in mechanical engineering from the University of
California at Davis (UCD), CA, USA, in 2015. Prior
to joining the Berkeley Sensor and Actuator Center
at UCD, as a Graduate Student Researcher, he was
a Digital Hardware Engineer with Huawei in 2011.
He is with Qualcomm, as a Staff Engineer. His
research interests include MEMS sensors and actuators.
Ryan Q. Rudy received the B.S.E. and M.S.E.
degrees in mechanical engineering from the Univer-
sity of Michigan, Ann Arbor, MI, USA, in 2009 and
2010, respectively, and the Ph.D. degree in mechan-
ical engineering from the University of Maryland
at College Park, College Park, MD, USA, in 2014,
with a focus on miniaturized ultrasonic motors.
He is currently a Mechanical Engineer with the
Micro & Nano Materials & Devices Branch,
U.S. Army Research Laboratory, Adelphi, MD,
USA. His current research interests include piezo-
electric MEMS, specifically piezoelectric electromechanical resonators and
filters.
9
Ronald G. Polcawich (M’07–SM’16) received the
B.S. degree in materials science and engineer-
ing from Carnegie Mellon University in 1997,
and the M.S. degree in materials and the Ph.D.
degree in materials science and engineering from
Penn State University in 1999 and 2007, respec-
tively. He is currently a Program Manager with
DARPA, Microsystems Technology Office and cur-
rently on detail to DARPA from the Micro & Nano
Materials & Devices Branch, U.S. Army Research
Laboratory, Adelphi, MD, USA, while at DARPA,
he is leading research projects in advanced materials processing combined
with micromechanics for small scale robotics and developing new materials
processes, device designs, and integration approaches required to push inertial
and aiding sensor performance for enabling trusted, heterogeneous position,
navigation, and timing (PNT) systems.
He was the Team Lead for PiezoMEMS Technology, ARL, with a focus on
developing component technologies to enable cognitive RF communication
and radar systems and MEMS inertial and aiding sensors to provide PNT
solutions for SWAP-C constrained platforms. He currently holds 16 patents,
has four patent applications pending review, and has authored over 100 journal
and proceedings articles, and authored three book chapters on fabrication and
design of piezoelectric MEMS devices using PZT thin films. His research
interests include materials processing of lead-zirconate-titanate (PZT) thin
films, MEMS fabrication, RF components, MEMS actuator technologies,
mm-scale robotics, MEMS inertial sensors, and sensors for aiding inertial sys-
tems. He and his colleagues were the recipients of the 2006 Department of the
Army Research and Development Achievement Award for Piezoelectric RF
MEMS Switch Using PZT Thin Films and the 2009 U.S. Army Research Lab-
oratory Engineering Award for ground-breaking work on Piezoelectric MEMS.
He received the 2012 Presidential Early Career Award for Scientists and
Engineers and the 2015 IEEE UFFC Ferroelectrics Young Investigator Award.
Dr. Polcawich is currently a member of the IEEE Ferroelectrics Committee
and the Technical Program Committee IV Applications of Ferroelectrics,
served as an elected member of the IEEE Ultrasonics, Ferroelectrics, and
Frequency Control Administrative Committee from 2014 to 2016. Since
2016, he has been the Chair of the UFFC Membership Committee. He is
on the Technical Advisory Committee for the PiezoMEMS Workshop,
co-organized the 2013 meeting in Washington, DC, USA, and co-organized
the 2018 meeting in Orlando, FL, USA.
David A. Horsley (M’97–SM’17) received the B.S.,
M.S., and Ph.D. degrees in mechanical engineer-
ing from the University of California at Berkeley,
Berkeley, CA, USA, in 1992, 1994, and 1998,
respectively. He held research and development posi-
tions with DiCon Fiberoptics, Inc., Hewlett Packard
Laboratories, and OniX Microsystems, Inc., and
was with the Faculty at UC Davis. In 2013, he
co-founded Chirp Microsystems to commercialize
MEMS ultrasonic transducers. He is currently a
Professor with the Department of Mechanical and
Aerospace Engineering, University of California at Davis, Davis, CA, USA,
and also an Adjunct Professor with the Department of Mechanical Engineer-
ing, University of California at Berkeley, Berkeley, CA, USA, and has been
the Co-Director of the Berkeley Sensor and Actuator Center since 2005.
His research interests include microfabricated sensors and actuators with
applications in ultrasonics and physical sensors. He was a recipient of the NSF
CAREER Award, the UC Davis College of Engineering’s Outstanding Junior
Faculty Award, and the NSF I/UCRC Association’s Alexander Schwarzkopf
Award.