ACOUSTIC ENERGY HARVESTER FABRICATED USING SOL/GEL LEAD
ZIRCONATE TITANATE THIN FILM
Shu Kimura, Tomohisa Sugou, Syungo Tomioka, Satoshi Iizumi,
Kyohei Tsujimoto, Yasushiro Nishioka*
Department of Precision Machinery Engineering, College of Science and Technology, Nihon University,
7-24-1 Narashinodai, Funabashi, Chiba 274-8501, Japan
*Presenting Author: nishioka@eme.cst.nihon-u.ac.jp
Abstract: Energy harvesters integrable on smart sensor systems have been strongly demanded. Horowitz et al.
have recently reported on a MEMS acoustic energy harvester using a lead zirconate titanate (PZT) thin film as a
diaphragm. Shinoda et al. also reported on similar acoustic energy harvesters with improved performances
fabricated using sol/gel PZT thin film processes, and suggested that the PZT acoustic energy harvester may be
suitable for use as a possible power source for silicon integrated circuits. This paper presents further improved
power generation performances of PZT MEMS acoustic energy harvesters fabricated using improved PZT
capacitor fabrication processes. The PZT acoustic energy harvester with the diaphragm diameter of 1.2 mm
fabricated using a sol/gel process generated the highest energy density of 98 µW/m2 under the sound pressure level
of 100 dB (0.01 W/m2) at 18.8 kHz.
Keywords: acoustic energy harvester, PZT
INTRODUCTION
As advanced integrated circuits (ICs) approach the
physical limits of miniaturization, new technology
innovations for those ICs have been strongly
demanded. Recently, the integration of complimentary
metal oxide semiconductor (CMOS) circuits and
microelectromechanical systems (MEMS) to increase
the values of these devices has been proposed. This is
called “More than Moore” concept. It was also pointed
out that advantageous devices are those that can
operate independently from outer power circuits. The
key for this concept is to realize micropower sources
integrable on the CMOS chips. Sudou et al. proposed
to integrate a power generator with an antenna on a
CMOS chip, and it generates power by receiving
alternate electromagnetic fields [1]. It will also be
interesting to investigate various possible energy
sources integrable on IC chips.
Horowitz et al. have recently reported on a MEMS
acoustic energy harvester that contained a PZT
piezoelectric film as a diaphragm [2]. The energy
harvester generated power by receiving sound pressure
to the diaphragm. PZT films have been extensively
studied since they have superior ferroelectricity and
piezoelectricity [3-5]. PZT films have also been
applied to ferroelectric memories, and it has been
demonstrated that PZT capacitor fabrication processes
are compatible with conventional CMOS processes
[5,7]. Therefore, it is expected that the PZT acoustic
energy harvester that can be integrated on CMOS IC
chips can be a candidate power source for integrated
smart systems. Shinoda et al. proposed to integrate a
PZT energy harvester on IC chips [8], and it can
receive power from an acoustic field (see Fig. 1). They
reported that higher generated power could be realized
by vibrating PZT diaphragms at the third resonance
frequency.
In this paper, we report on further improved
performances of PZT acoustic energy harvesters that
use first resonance vibrations.
Sound pressure
Fig. 1: Concept of an integrated system with energy
harvester, CMOS circuits, and MEMS.
DEVICE STRUCTURE
The structure of the device investigated here is
schematically shown in Fig. 2. A piezoelectric
capacitor with the diaphragm structure of Al/PZT/Pt/Ti
was formed on an SiO2. The cavity underneath the
diaphragm was fabricated on a silicon wafer. The
principle of power generation of this device is as
follows. When sound pressure is applied from the
upper direction of the device, the diaphragm vibrates
resulting in the deflection of the diaphragm of the PZT
piezoelectric capacitor. The deflected PZT capacitor
generates a voltage difference between the two
electrodes. In this manner, electrical power is
generated. The output voltage becomes maximum
when the sound frequency is equal to the resonance
frequency of the diaphragm. The bonding pad (250 ×
250 µm2) connecting to the upper electrode was placed
apart from the diaphragm so that the vibration of the
diaphragm was not disturbed.
The first resonance frequencies of the diaphragm
with different diameters were estimated using the
ANSYS software for the finite element analysis. The
estimated first resonance frequency for the diaphragm
with the diameter of 2.0 mm was 6.25 kHz, and that
with the diameter of 1.2 mm was 18.7 kHz. These
resonance frequencies are summarized along with the
measured resonance frequency results in Table I.
Al (upper electrode,
0.1 µm)
PZT (1 µm)
Pt/Ti (lower
electrode, 0.2 µm)
SiO2 (1.5 µm)
Si (300 µm)
Fig. 2: Structure of fabricated energy harvester with
PZT diaphragm.
Table I: Estimated frequencies for the first resonance
mode for diaphragm with different diameters.
procedures. Firstly, the PZT thin films were spincoated on the lower electrode, Pt(100 nm)/Ti(100 nm),
on the silicon substrate. The PZT films were deposited
by the sol-gel method. The PZT sol-gel solution is
“PZT-20” supplied by Kojundo Chemical Laboratory,
Co, Ltd. The PZT solution was dropped on the Pt/Ti
substrate and spin-coated for 3 s at the rotation speed
of 500 rpm and then for 20 s at the rotation speed of
3200 rpm. The chip was dried on a hot plate in air
ambient for 5 min at 120 ºC, and dried again for 6 min
at 300 ºC. After those processes were repeated twice,
the chip was sintered at 700 ºC for 5 min using a rapid
thermal anneal furnace in oxygen atmosphere to
crystallize the PZT films. The PZT thin film of 1 µm
thickness was formed by repeating five times the
processes described in Fig. 4. More detailed PZT thin
film fabrication processes were described in many
publications [9].
Spin coating of PZT
solution
500 rpm, 3 s
3200 rpm, 20 s
Twice
Resonance frequency (kHz)
Diaphragm diameter (µm)
1200
18,7
1400
12,95
1600
10,19
1800
8,26
2000
6,25
Drying in air ambient
o
120 C, 5 min
5 times
Prebaking in air
o
300 C, 6 min
Power MEMS
Power Amp
Resister
Pre Amp
Sintering
o
700 C, 5 min
Function
Generator
Osilloscope
Fig. 3: Experimental setup for measuring the
resonance frequencies and the generated power.
The resonance frequencies were measured using
the characterization system described in Fig. 3. AC
sinusoidal voltage signals with different frequencies
were generated by the function generator and
amplified by the power amplifier. The amplifier was
connected to the speaker. The gain of the power
amplifier was modified to generate a requested sound
pressure at the PZT energy harvester. An amplifier
with a gain of 10,000 was used to amplify the output
voltage signals from the energy harvester. The voltage
between both terminals of the energy harvester or of
the load resister was measured using an oscilloscope.
DEVICE FABRICATION
Figure
4
shows
the
PZT
film
deposition
Fig. 4: PZT thin film deposition processes by sol-gel
method. These procedures were repeated 5 times to
form 1 µm-thick PZT film.
The device fabrication processes used in this
research are described below. A Si substrate of 300 µm
thickness with 1.5-µm-thick SiO2 was prepared. The
Pt/Ti layer was deposited by sputtering to form the
lower electrode. The 1-µm-thick PZT thin film was
deposited by the sol-gel method. The PZT was
patterned by wet etching using buffered hydrofluoric
acid (BHF) to expose the lower electrode. The 0.1-µmthick Al thin film for the top electrode was then
thermally deposited. The top electrode patterns were
defined by all dry mask deposition. The cavity was
formed by etching the silicon substrate from the back
with an inductively coupled plasma (ICP) etcher. The
thicknesses of the films in the diaphragm part are Al
(0.1 µm)/PZT (1.0 µm)/Pt/Ti (0.2 µm)/SiO2 (1.5 µm).
The top electrode was previously defined using a
photolithography method including Al wet etching
solution [8]. However, we suspected that this wet
etching process might be harmful for the PZT
diaphragm. Therefore, the Al electrodes in this
research were defined using a mask deposition process
in which no wet etching processes was included.
Figure 5 shows the photographs of the PZT surface
before immersing in an Al etching solution (a) and
after immersing the PZT in the Al wet etching solution
for 30 s (b). Most of the PZT film was removed after
30 s. Therefore, the Al wet etching process to define
the top electrodes could damage the PZT capacitors.
(a)
(b)
Fig. 5: PZT thin film deposited on Pt/Ti coated
substrate (a) and the PZT surface immersed in Al wet
etching solution for 30 s (b). After 30 s, all the PZT
film was removed.
estimate the suitable load resistance R, the following
measurements were performed.
Figure 6 shows comparison of the power comsumed at
the load resistor for the PZT acoustic power generator with a
diamter of 1.2 and 2.0 mm fabricated using the dry Al
electrode fabrication process, and with a diameter of 2.0 mm
fabricated using the Al wet process. As shown in Fig. 3, a
variable resistor was connected to the PZT energy
harvester, and the resistance was varied from 10 Ω to 1
MΩ. The voltage between both terminals of resistance
was measured and recorded using an oscilloscope. The
sound pressure of approximately 100 dB at the first
resonance frequencies was irradiated. The effective
electric power delivered to the load resistance R was
calculated from the measured voltage amplitude V
using the relationship P = V2/2R. The relationships
between the power delivered to the load and the load
resistance were well described using Eq. (1). The
output power was proportional to R when R was small,
and was inversely proportion to R when R was large.
The maximum power delivered to the load was derived
when the load resistance was approximately 550 Ω for
the device fabricated using the Al wet process, and the
power decrease below 550Ω. In contrast, the generated
power of the devices fabricated using the Al dry
process continued to increase until the load resistance
reaches around 75 Ω. Thus, it can be speculated that
the Al wet etching process make the internal resistance
larger resulting in the reduction of the generated power.
DETERMINATION OF LOAD RESISTANCE
,
(1)
where P is the extracted power, V is the voltage across
the load resistance, I is the current across the load
resistance, and E is the generated voltage. Note that the
extracted power P is maximum at an appropriate R. To
extract the largest power from the power source, the
load resistance R should be selected correctly. To
Generated power (W)
10−9
After improving the process
（Diameter1.2[mm]）
After improving the process
（Diameter2[mm]）
Before improving the process
（Diameter2[mm]）
10−10
Al dry process
10−11
10−12
Al wet process
10−13
10
100
1000
10000
Load resistance (ohm)
Fig. 6: Measured relationships between the power
delivered to the load and the load resistance.
100000
1200μm
1400μm
0.1
電圧〔mV〕
 E 
P = VI = RI 2 = R ⋅ 

R+r
2
0.12
Generated voltage (V)
It is necessary to demonstrate that the harvesters
can deliver power to a load resistance. When an
internal resistance r exists in a power source, the
output electric power consumed at the load resistance
R is expressed using
1600μm
1800μm
0.08
2000
1200
1600μ
1400
1800
2000μm
0.06
0.04
0.02
0
0
2000
4000
6000
8000
10000
12000
14000
16000
18000
周波数〔Hz〕
Frequency
(Hz)
Fig. 7: Frequency dependences of the generated
voltage E of the PZT energy harvester with different
diaphragm diameters from 1.2 to 2.0 mm.
Figure 7 compares the generated voltage E for the
PZT power generators with different diaphragm
diameters. The resonance frequency increases as the
diameter size decreased. Figure 8 compares the load
resistance dependences for the PZT power generators
investigated here. Note that smaller diameter devices
exhibited the larger generated power. This may be due
to the fact that higher frequency sound was applied to
the devices with the smaller diaphragm diameters.
CONCLUSION
Generated power (W)
10-9
10
polarization than the PZT film investigated here.
The PZT acoustic energy harvester with the
diaphragm diameter of 1.2 mm fabricated using a
sol/gel process generated the highest energy density of
98 µW/m2 at the sound pressure level of 100 dB (0.01
W/m2). The energy density was the largest ever
reported.
1200μm
1400μm
1600μm
1800μm
2000μm
-10
10-11
10-12
10-13
10-14
ACKNOWLEDGEMENTS
10
100
1000
10000
100000
Load resistance (ohm)
Fig. 8: Measured relationships between the power
delivered to the load and the load resistance for PZT
acoustic PZT power generators with different
diaphragm diameters.
Table I compares the performances of the PZT
power generators by Horowitz et al., Shinoda et al.,
and this work. The dramatic improvement in the
energy density was due to the reduced internal
resistance of the PZT capacitor.
DISCUSSION
The improved performances of the PZT acoustic
energy harvesters compared to those of Horowitz et al.
and Shinoda et al. are most possible due to the reduced
internal resistance, and it is strongly related with the
top electrode fabrication process of the PZT capacitors.
The energy harvesters reported by Horowitz et al. have
structures and dimensions similar to those of the
devices investigated here. They used the PZT thin film
with the thickness of 0.27 µm. In contrast, the PZT
film used in the devices was 1 µm thick. It has been
reported that the polarization charge of sol-gel PZT
films rapidly decreases when the PZT film thickness is
below 0.625 µm [10]. Therefore, it is assumed that the
PZT film investigated by Horowitz et al. had smaller
The
authors
express
their
sincere
acknowledgements to staffs of the Micro Functional
Device Research Center, CST, Nihon University, and
to Dr. Takeshi Kobayashi and Dr. Ryutaro Maeda of
the National Institute of Advanced Industrial Science
and Technology (AIST) for the technical assistances.
REFERENCES
[1] Sudou M, Takao H, Sawada K, Ishida M 2008 Sens.
Actuators A 145-146 343.
[2] Horowitz S B, Sheplak M, Cattafesta N, Nishida T 2006
J. Micromech. Microeng. 16 S174.
Gross S J, Tadigadapa S, Jackson T N, TrolierMcKinstry S, and Zhang Q Q 2003 Phys. Lett. 83 174.
[4]
Yamashita K, Katata H, Okuyama M, Miyoshi H,
Kato G, Aoyagi S, and. Suzuki Y 2002 Sens. Actuators
A 97–98 302.
[5] Itoh T, Lee C, Suga T 1996 Phys. Lett. 69 2036.
[6] Kim K and S. Y. Lee 2007 Microelectron. Eng. 84
1976.
[7] Jung J, Joo H, Park J, Kang S, Kim H, Choi D, Kim J,
Lee E, Kong Y, Kim H, Jung D, Kang Y, Lee S, Jeong
H, Kim K 2007 Jpn. J. Appl. Phys. 46 1934.
[8] Shinoda S, Tai T, Itoh H, Sugou T, Ichioka H, Kimura S,
Nishioka Y 2010 Jpn. J. Appl. Phys. 04DL21
[9] Kobayashi T, Ichiki M, Tsaur J, Maeda R 2005 Thin
Solid Films 489 74.
[10] Amanuma K, Mori T, Hase T, Sakuma T, Ochi A,
Miyasaka Y 1993 Jpn. J. Appl. Phys. 32 4150.
[3]
Table I Comparisons of PZT acoustic energy harvesters. Sound pressure level was 100 dB (0.01 W/m2).
S. B. Horowitz et al
[1]
S. Shinoda et al.
[2]
This work
Diaphragm
Diameter (mm)
Resonance
Frequency (Hz)
2.4
13568
Generated
Power (W)
Power density
2
(µW/m )
Internal
resistance
(ohm)
1000
6.0x10
-12
1.0
0.054
1000
3.6
5232
7.0x10
-13
1.5
24020
1.1x10
-11
4.9
550
1.3
550
2.0
18020
5.1x10
-12
1.2
16700
1.4x10
-10
98
75
4.0x10
-11
10
70
2.0
6400