Piezoelectric energy harvesting: State-of-the-art and challenges

Alperen Toprak and Onur Tigli

Citation: Applied Physics Reviews 1, 031104 (2014); doi: 10.1063/1.4896166

View online: http://dx.doi.org/10.1063/1.4896166
View Table of Contents: http://scitation.aip.org/content/aip/journal/apr2/1/3?ver=pdfcov
Published by the AIP Publishing

Articles you may be interested in

Polymer piezoelectric energy harvesters for low wind speed
Appl. Phys. Lett. 104, 012902 (2014); 10.1063/1.4861187

An innovative tri-directional broadband piezoelectric energy harvester

Appl. Phys. Lett. 103, 203901 (2013); 10.1063/1.4830371

A nonlinear piezoelectric energy harvester with magnetic oscillator

Appl. Phys. Lett. 101, 094102 (2012); 10.1063/1.4748794

Charge redistribution in piezoelectric energy harvesters

Appl. Phys. Lett. 100, 073901 (2012); 10.1063/1.3685701

A piezoelectric bistable plate for nonlinear broadband energy harvesting

Appl. Phys. Lett. 97, 104102 (2010); 10.1063/1.3487780

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 130.92.9.55
On: Sun, 28 Sep 2014 18:06:44
 APPLIED PHYSICS REVIEWS 1, 031104 (2014)

APPLIED PHYSICS REVIEWS—FOCUSED REVIEW

Piezoelectric energy harvesting: State-of-the-art and challenges

Alperen Toprak1,2,a) and Onur Tigli1,3,2,b)
1
Electrical and Computer Engineering Department, University of Miami, Coral Gables, Florida 33146, USA
2
Dr. John T. Macdonald Foundation, Biomedical Nanotechnology Institute at University of Miami, Miami,
Florida 33136, USA
3
Department of Pathology, Miller School of Medicine, University of Miami, Miami, Florida 33136, USA
(Received 30 May 2014; accepted 9 September 2014; published online 23 September 2014)
Piezoelectric energy harvesting has attracted wide attention from researchers especially in the last
decade due to its advantages such as high power density, architectural simplicity, and scalability.
As a result, the number of studies on piezoelectric energy harvesting published in the last 5 years is
more than twice the sum of publications on its electromagnetic and electrostatic counterparts. This
paper presents a comprehensive review on the history and current state-of-the art of piezoelectric
energy harvesting. A brief theory section presents the basic principles of piezoelectric energy
conversion and introduces the most commonly used mechanical architectures. The theory section is
followed by a literature survey on piezoelectric energy harvesters, which are classified into three
groups: (i) macro- and mesoscale, (ii) MEMS scale, and (iii) nanoscale. The size of a piezoelectric
energy harvester affects a variety of parameters such as its weight, fabrication method, achievable
power output level, and potential application areas. Consequently, size-based classification pro-
vides a reliable and effective basis to study various piezoelectric energy harvesters. The literature
survey on each scale group is concluded with a summary, potential application areas, and future
directions. In a separate section, the most prominent challenges in piezoelectric energy harvesting
and the studies focusing on these challenges are discussed. The conclusion part summarizes the
current standing of piezoelectric energy harvesters as possible candidates for various applications
and discusses the issues that need to be addressed for realization of practical piezoelectric energy
harvesting devices. VC 2014 AIP Publishing LLC. [http://dx.doi.org/10.1063/1.4896166]

TABLE OF CONTENTS I. INTRODUCTION

In its simplest form, energy harvesting can be defined as

I. INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . .
II. PIEZOELECTRIC ENERGY HARVESTING
THEORY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
III. CURRENT LITERATURE IN
PIEZOELECTRIC ENERGY HARVESTERS . . .
A. Macro and meso-scale . . . . . . . . . . . . . . . . . . .
B. MEMS scale. . . . . . . . . . . . . . . . . . . . . . . . . . . .
C. Nanoscale . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
D. Comparative summary of cantilever type
piezoelectric energy harvesters . . . . . . . . . . . .
IV. CHALLENGES AND PROPOSED
SOLUTIONS IN PIEZOELECTRIC ENERGY
HARVESTERS . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
A. Bandwidth . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
B. CMOS compatibility . . . . . . . . . . . . . . . . . . . . .
C. Biocompatibility . . . . . . . . . . . . . . . . . . . . . . . .
V. CONCLUSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
a)
a.toprak@umiami.edu
b)
tigli@miami.edu
1 converting the unused energy available in the environment
into electrical energy, which can be used immediately or
2 stored for later use. This broad definition also implies a vari-
ety of applications, which strongly depend on the scale of
4 the harvestable energy. Large scale sources such as vehicle
4 suspension systems, tall buildings, or ocean waves can con-
6 tain power levels up to tens of kilowatts, which might be
8 used as a form of renewable energy source.1 On the other
end of the scale, relatively small energy sources such as am-
9 bient vibrations or temperature gradients are not capable of
providing power output levels that are high enough to be
considered for power grids. Instead, they can be utilized to
10 supply sustainable, clean power for standalone electronic
10 sensor or transducer elements. In current technology exam-
11 ples, this important task is handled solely by batteries. In
12 conjunction with shrinking transistor sizes, improvements in
12 CMOS and MEMS technologies enabled the mass fabrica-
tion of very small sensor and transducers with remarkably
low power requirements. The size, weight, and cost of these
systems are currently limited by the available battery tech-
nologies. More importantly, replacement of depleted bat-
teries is mostly infeasible, which limits the lifetime and

1931-9401/2014/1(3)/031104/14/$30.00 1, 031104-1 C 2014 AIP Publishing LLC

V

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 130.92.9.55
On: Sun, 28 Sep 2014 18:06:44
 031104-2 A. Toprak and O. Tigli
sustainability of these systems. As a result, energy harvesting
applications are expected to increase the lifetime, or even
lead to self-sustaining systems by completely eliminating the
need for batteries while creating a significant impetus for
growing research in industry and academia.
There are a number of available ambient energy sources
suitable for energy harvesting applications including solar,
vibration, radio frequency (RF), acoustic waves, and temper-
ature gradients.2 Achievable power densities from various
energy sources as well as their advantages and disadvantages
are previously reported in the literature.3 Ambient vibrations
can provide a high energy density per unit device volume;
and unlike solar cells, they can operate in implanted or em-
bedded systems. There are several vibration sources in the
environment, generating vibrations with different amplitudes
and frequencies.4,5 Vibration energy can be harvested using
an inertial mechanism, in which the vibration is coupled to a
proof mass and then extracted by damping the mechanical
motion of the mass. There are three main types of vibration-
based energy harvesters: piezoelectric, electromagnetic, and
electrostatic. Piezoelectric vibration energy harvesters are
reported to have a higher energy density for practical appli-
cations.6,7 Another advantage of the piezoelectric conversion
mechanism is the inherent reciprocal conversion capability:
piezoelectric materials convert the mechanical energy
directly to electrical energy without any further external
input. This inherent capability allows simpler architectures
for piezoelectric energy harvesters (PEHs) compared to their
electromagnetic and electrostatic counterparts. Simple archi-
tectures are especially preferable in MEMS scale devices,
where the structures have to be created by micromachining
techniques. Scaling of the power with volume also favors the
piezoelectric devices in smaller scales: output power scales
with V4/3 and V 2 in piezoelectric and electromagnetic con-
version mechanisms, respectively.8 The critical volume that
piezoelectric converters provide a better power output is
determined as 0.5 cm3 in the same study.8 Due to these most
salient advantages, PEHs have been studied extensively in
the last decade. Figure 1 shows the number of publications
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 130.92.9.55
On: Sun, 28 Sep 2014 18:06:44
Appl. Phys. Rev. 1, 031104 (2014)
on piezoelectric, electrostatic, and electromagnetic energy
harvesters indexed in Web of Science between years 2003
and 2013. The graph clearly shows the steadily growing in-
terest on PEHs especially in the last 5 years.
This review paper presents a review of the current status
of PEHs. Section II briefly presents the theory of piezoelec-
tric energy harvesting. Section III presents the current status
of PEHs in three categories depending on their sizes: macro-
and meso-scale, MEMS scale, and nanoscale. The main chal-
lenges of PEHs and proposed solutions to overcome these
challenges are discussed in Sec. IV, and Sec. V concludes
the paper.
II. PIEZOELECTRIC ENERGY HARVESTING THEORY
Piezoelectric materials have the ability to create an elec-
tric charge in response to a mechanical stress, which is called
direct piezoelectric effect. The reciprocal process, mechani-
cal strain in response to an electrical potential, is named
inverse piezoelectric effect. Piezoelectric energy harvesting
process uses the direct piezoelectric effect, which is
described by the constitutive equation
Di ¼ dij Tj þ Tik Ek ; (1)
where D is the electrical displacement, d is the piezoelectric
stress coefficient, T is the stress,  is the dielectric constant,
and E is the electric field.9 The subscript indices in this equa-
tion refer to different directions within the material coordi-
nate system, and the superscript “T” means that the dielectric
constants are measured under constant stress. Subscripts i
and k denote the coordinate axes, numbered from 1 to 3, sim-
ilar to Cartesian coordinate axes x, y, and z. By convention,
direction 3 is defined as the polarization direction of the pie-
zoelectric material. Subscript j, in addition to the same three
axes, defines the rotational motions around them; therefore,
it is denoted with numbers from 1 to 6. The operation modes,
i.e., expected directions of mechanical strain and electric
field, of a piezoelectric device is also indicated using the
defined indices. To give an example, a piezoelectric sensor is
FIG. 1. Number of publications on pie-
zoelectric, electromagnetic, and elec-
trostatic energy harvesters in Web of
Science between years 2003 and 2013.
 031104-3 A. Toprak and O. Tigli Appl. Phys. Rev. 1, 031104 (2014)

FIG. 2. (a) Schematic view describing the operating principle of a cantilever type PEH, and (b) a prototype PEH that consists of two PZT-5A layers bonded on
two sides of a steel shim and a proof mass at the tip. Ambient vibrations are coupled to the cantilever-mass system through the base of the cantilever, and the
strain due to bending is converted to voltage by the piezoelectric material. The proof mass is used to increase the coupled mechanical energy and reduce the
resonant frequency of the structure. Reproduced by permission from S. Roundy and P. K. Wright, Smart Mater. Struct. 13(5), 1131–1142 (2004). Copyright
2004 IOP Publishing.

defined to be operating in d31 mode if it is subject to a stress
on axis 1 and the output is sensed via the electrodes placed
across the device on axis 3. In this case, the electrical output
is proportional to the coefficient d31, hence the name of the
operation mode. The first term of Eq. (1) implies that the
charge generated in a piezoelectric material is proportional
to the applied stress. Therefore, PEHs are designed to maxi-
mize the stress under a certain mechanical load.
The most commonly used shape in piezoelectric energy
harvesting is the cantilever beam, since this structure creates
the highest average strain for a given input force.7 Figure
2(a) shows the schematic view describing the operating prin-
ciple of a cantilever type PEH, and Fig. 2(b) shows a proto-
type PEH that consists of two lead zirconate titanate (PZT)
layers bonded on two sides of a steel shim and a proof mass
at the tip.7 Ambient vibrations are coupled to the cantilever-
mass system through the base of the cantilever, causing the
structure to oscillate. The alternating bending strains during
the oscillations are converted to an AC voltage by the piezo-
electric material. One of the most important design consider-
ations in this architecture is the frequency matching, which
requires the exact coupling of ambient vibration frequencies
to the natural oscillation frequency of the beam. Cantilever
beams usually have a high mechanical quality factor, or
equivalently, a narrow bandwidth. The oscillation amplitude
rapidly drops as the excitation frequency shifts from the res-
onance frequency. Since most ambient vibrations are at low
frequencies, additional proof masses are used to reduce the
resonance frequency of the energy harvester. Furthermore,
the proof mass increases the overall mechanical energy
stored in the cantilever-mass system, increasing the amount
of harvestable energy.
Another mechanical structure used for PEH is the circu-
lar diaphragm. In one of the earliest studies on PEHs, Umeda
et al. proposed using a thin disk of piezoelectric material to
convert impact energy to electrical energy.10 They verified
the energy conversion by dropping a steel ball on top of a
0.25 mm thick piezoelectric disk mounted on a bronze disk
with the same thickness. In such an impact-type PEH, the
structure starts oscillating at its resonance frequency, which
is typically much higher than the frequency of impacts. The
amplitude of the oscillations then gradually decreases
because of the mechanical and electrical damping.10 An im-
portant advantage of diaphragm structure over the cantilever
beam is its compatibility with pressure mode operation. In
pressure mode operation, a varying pressure field such as
sound can be converted to an AC electrical signal by a piezo-
electric converter. In order to be able to create a stress in
response to a pressure change in the surrounding medium,
two sides of the mechanical structure must be isolated as in
the case of a diaphragm structure. Figure 3(a) depicts the
deflection of a diaphragm structure in pressure mode opera-
tion, which causes different stress types, i.e., compressive or
tensile, to develop on the top layer depending on the loca-
tion.11 The switching stress sign necessitates the poling
direction be switched as well for optimum power output.

FIG. 3. (a) Deflection of a diaphragm structure in pressure mode operation, which causes different stress types, i.e., compressive or tensile, on the top layer
depending on the location. The switching stress sign necessitates the poling direction be switched as well for optimum power output. Part (b) shows diaphragm
type PEH prototypes with different poling switch locations as indicated by the gaps in the top electrodes. Reprinted with permission from Kim et al., J. Intell.
Mater. Syst. Struct. 16(10) 847–854 (2005) and from Kim et al., J. Intell. Mater. Syst. Struct. 16(10), 855–863 (2005). Copyright 2005 SAGE Publications.

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 130.92.9.55
On: Sun, 28 Sep 2014 18:06:44
 031104-4 A. Toprak and O. Tigli
FIG. 4. (a) Simplified view of a 4-layer PZT stack architecture, and (b)
actual photograph of a 300-layer PZT stack actuator. Small arrows inside
the PZT show the poling direction. Poling direction is on the same axis as
the applied pressure, which implies d33 mode operation. Reproduced by per-
mission from Xu et al., Smart Mater. Struct. 22(6), 065015 (2013).
Copyright 2013 IOP Publishing.
Figure 3(b) shows diaphragm type PEH prototypes with dif-
ferent poling switch locations as indicated by the gaps in the
top electrodes.12 The theoretical and experimental analyses
of piezoelectric circular diaphragms operating under varying
pressure were presented by Kim et al.11–13 Diaphragm is a
versatile structure that can be used to harvest energy from
slowly varying, aperiodic pressure fields14,15 or periodic
acoustic waves,16–18 as well as mechanical vibrations.19,20
However, it should be noted that it is considerably stiffer
than a cantilever of same size, leading to higher resonance
frequencies in vibration mode operation.
Pressure mode operation can be achieved also using
stack architectures. Figure 4(a) shows the simplified view of
a 4-layer PZT stack, and Fig. 4(b) actual photograph of a
300-layer PZT stack actuator.21 In the stack architecture, pie-
zoelectric layers are placed such that the poling axes of the
layers align with the applied pressure in order to utilize the
d33 mode, which has a higher coefficient than d31.
Alternating poling directions of subsequent layers provides
that the direction of the electric field is always towards the
same electrode. Although piezoelectric stacks utilize the d33
mode, the mechanical energy coupled from the applied pres-
sure is usually low due to the high stiffness of the structure.22
Therefore, they are either used in applications where high
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 130.92.9.55
On: Sun, 28 Sep 2014 18:06:44
Appl. Phys. Rev. 1, 031104 (2014)
pressures are available or they are coupled to mechanical
force amplifiers. Figure 5 shows a “cymbal” transducer, a
force amplifying structure, used to improve the power output
from a PEH.23 In the cymbal structure, the axial stress
applied to the metal end caps is amplified and transformed
into radial stress in the piezoelectric material, resulting in a
higher effective piezoelectric constant. The amplification
factor depends on the structure of the caps.23
III. CURRENT LITERATURE IN PIEZOELECTRIC
ENERGY HARVESTERS
In this section, the current state-of-the-art in PEHs is
presented. The PEHs are examined in three groups according
to their dimensions: (i) macro/meso (ii) MEMS, and (iii)
nanoscale. There are not any strict size limits defined to dis-
tinguish MEMS scale devices. Therefore, we classify the
devices fabricated using standard photolithography techni-
ques as MEMS scale, whereas the devices including manual
assembly and bonding are examined as mesoscale.
Nanoscale energy harvester section includes energy harvest-
ers constructed using piezoelectric nanowires.
A. Macro and meso-scale
The research on PEHs is mostly focused on small scale
devices that can provide power for wireless sensor networks
(WSNs). However, it has also been proposed as an alterna-
tive clean energy source in the recent years. In 2006, Japan
East Railway Company installed piezoelectric floors in an
exit gate of their Tokyo Station to evaluate the power gener-
ation capacity of such a system. The 6 m2 piezoelectric floor
was reported to generate a maximum energy of 10 000 W  s
per day. They repeated the experiment in 2008 in an effort to
improve the power output and the durability of the system.
The target energy output from the piezoelectric floors in-
stalled on a 90 m2 area was 500 kW  s per day.24 Another
approach is to harvest energy from highways and railways
using piezoelectric generators. In a pilot study carried out by
Innowattech Ltd. in association with Technion Israel
Institute of Technology, a 10-m stretch of piezoelectric gen-
erators were placed under a highway in Israel in 2009. The
reported average power output was 2000 W, and the obtained
energy was stored in a battery beside the road.25 The same
company also announced a project on self-powered railway
FIG. 5. (a) A cymbal transducer con-
sisting of a metal cap attached to a
PZT disc, and (b) schematic view of
the test setup used to investigate the
energy harvesting performance of a
prototype. The axial stress applied by
the shaker is amplified and converted
to radial stress in the PZT material.
Stress amplification provides a higher
effective piezoelectric constant for the
structure. Reprinted with permission
from Kim et al., Jpn. J. Appl. Phys.,
Part 1 43(9A), 6178–6183 (2004).
Copyright 2004 Japan Society of
Applied Physics.
 031104-5 A. Toprak and O. Tigli Appl. Phys. Rev. 1, 031104 (2014)

monitoring systems, for which 32 railway pads were

replaced with their piezoelectric generator pads.26 This tech-
nology is not mature enough to be fully commercialized yet.
A recent consultant report prepared for California Energy
Commission evaluated the feasibility of the approach and
concluded that more comprehensive tests are necessary to
understand the energy output, the cost, and the lifetime of
such systems.27
Mesoscale PEHs usually aim to provide power for wire-
less sensor networks or implantable medical devices (IMDs).
The earliest examples of meso-scale PEHs were proposed in
1960s to supply power for pacemakers.28,29 Both devices
used the expansion of aorta during systole/diastole cycle as
the energy source, which was converted to electrical energy
using PZT beams. In 1984, Hasler et al. designed an implant-
able energy harvester to generate electrical power from the
relative motion of ribs during respiration.30 After these initial
studies for implantable energy harvesters, Antaki et al. pro-
posed a shoe-embedded piezoelectric generator to harvest
energy for medical implants from human walking in 1995.31
They conducted experiments using a single cylindrical PZT
stack with 0.5 cm3 volume, which resulted in an average
power of 5.7 mW/kg during walking. In another study on
harvesting energy from human gait, the performances of one
electromagnetic and two PEHs were compared.32 The har-
vesters were designed to harvest energy mainly from the
heel strike and the sole bending during walking. Utilization
of the energy harvesters were also demonstrated by intermit-
tently operating a radio frequency identification (RFID) tag
using the harvested power. The authors concluded that
although the electromagnetic harvester has two orders of mag-
nitude higher power output, piezoelectric ones are much easier
to integrate into shoes without interfering with the gait.32 In
FIG. 6. (a) Energy harvesting scheme from two different regions in a shoe:
2001, the same group published a follow up study on shoe em- (i) polyvinylidene fluoride (PVDF) stave placed under the ball of the foot to
bedded piezoelectric generators as depicted in Figure 6.33 harvest the energy from bending, and (ii) PZT dimorph placed under the
Figure 6(a) illustrates the proposed energy harvesting devices heel to harvest the heel strike energy. Part (b) shows the voltage and power
and their locations in the shoe, and Fig. 6(b) shows the voltage outputs of fabricated devices on optimized resistive loads under a normal
walking pace, which is approximately 0.9 Hz as evident from the spikes on
and power outputs of fabricated devices on optimized resistive the outputs. Reprinted with permission from N. S. Shenck and J. A.
loads under a normal walking pace, which is approximately Paradiso, IEEE Micro 21(3), 30–42 (2001). Copyright 2001 IEEE Micro.
0.9 Hz as evident from the spikes on the outputs. In this study,
a dimorph was used instead of a unimorph at the heel, and a parameters, including the load resistance and the mechanical
more efficient power conditioning circuitry was utilized to be quality factor of the piezoelectric generator.10 The same
able to rectify the output and store it on a capacitor. The whole group also studied the efficiency in case of an output load
system, including the power conditioning circuitry and storage consisting of a rectifier and a capacitor, which can be used to
elements, was shown to provide 1.3 mW continuous power store the harvested energy and achieved a maximum electro-
from a walking speed of 0.8 Hz.33 mechanical conversion efficiency of 35%.34
Using PEHs with the purpose of charging the batteries Using resonant mode piezoelectric devices to harvest
of portable electronic devices was first proposed by Umeda energy from periodic vibrations has also been widely studied
et al. in 1996.10 In this study, the authors pointed out that by researchers. The most common architecture for harvesting
portable devices are subjected to mechanical impact during energy from vibrations is the cantilever beam. Cantilever
transportation, which can be converted to electrical energy type energy harvesters typically consist of a thin cantilever
and used to recharge their batteries. They developed an beam with a proof mass at its free end and a fixed base at the
equivalent circuit model for the piezoelectric generator and other end. The entire structure is mechanically attached to a
performed experiments by dropping a steel ball on a circular vibrating body via the fixed base, in order to couple the
piezoelectric disk clamped at the edges. The piezoelectric vibrations to the cantilever and the proof mass. Resultant os-
disk converted the vibrations induced by the impact of the cillation is converted to an AC electrical signal by the piezo-
steel ball into electrical energy, which was dissipated on a electric material on the cantilever beam. The amplitude of
resistive load. The results of the experiments showed that the the oscillation, and consequently, the electrical output,
harvested electrical power depends on a number of strongly depends on the frequency matching between the

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 130.92.9.55
On: Sun, 28 Sep 2014 18:06:44
 031104-6 A. Toprak and O. Tigli
input vibrations and the natural frequency of the beam.
Glynne-Jones et al. designed, fabricated, and tested a cantile-
ver type vibrational mode PEH as a proof-of-concept in
2001.35 The proposed device consisted of a tapered PZT
layer with electrodes on both sides, bonded to a 23 mm wide
trapezoidal steel beam; it had a resonance frequency of
80 Hz and generated 3 lW power on an optimal resistance of
333 kX at a tip displacement of 0.8 mm. Another demonstra-
tion for the utilization of the vibration energy harvested by
piezoelectric cantilevers was presented by Roundy and
Wright in 2004.7 In this study, a 1.9 GHz radio transmitter
was operated intermittently using the harvested energy
stored in a 47 lF capacitor. The energy was supplied from a
1 cm3 rectangular cantilever type PEH subjected to an input
acceleration of 2.5 m/s2 at its resonance frequency of 120 Hz.
The power output was measured as 375 lW, and correspond-
ing maximum duty cycle at which the radio transmitter can
be operated was calculated as 1.6% by the authors.7 A simi-
lar experiment was performed by Reilly et al. in 2011 using
a trapezoidal cantilever beam structure, which was reported
to operate a radio transmitter at 0.2% duty cycle.36 Although
the aforementioned devices use PZT as the piezoelectric ma-
terial, there are other studies reporting the utilization of other
materials. In 2012, Cao et al. fabricated an AlN based energy
harvester on a stainless steel substrate to improve the fracture
toughness of the structure.37 They bonded a relatively large
copper proof mass to reduce the resonance frequency, which
was then measured as 69.8 Hz for a 6.1 mm long cantilever.37
More recently reported cantilever type devices usually aim
to improve the power output by using optimally curved can-
tilevers38 or other mechanical structures that can provide a
more uniform stress distribution.39
Cantilever beams provide an easy method to create high
stress from a relatively small force; however, they are not
compatible with pressure mode operation. In 2003, Kim
et al. investigated the feasibility of harvesting energy from
pressure using a diaphragm structure.13 They studied the
stress distribution in a clamped circular diaphragm and pro-
posed a re-poling scheme to maximize the electrical output
energy. The same group published detailed analytical models
and experimental results for clamped11,12 and simply sup-
ported15,40 circular diaphragms operated in a quasi-static
mode, where the varying pressure frequency is much lower
than the resonance frequency of the diaphragms. In both
cases, experimental results showed good agreement with the
analytical models. Diaphragm structure can be used to har-
vest vibrational energy as well, which has been demonstrated
in a number of studies.19,20,41 However, the stiffness of this
structure is considerably higher than a same size cantilever,
which leads to high resonance frequencies such as 1.71 kHz
for a 25-mm diameter diaphragm.19 Bringing the resonance
frequency below 200 Hz for similar sized diaphragm harvest-
ers necessitate using proof masses on the order of 0.1 kg.20,41
Pressure mode energy harvesting using stack architec-
ture usually requires mechanical force amplifier structures.
Kim et al. coupled a PZT disk, which can be considered as a
single layer stack, with a cymbal mechanical transducer.23
The cymbal energy harvester with 29 mm diameter and
1 mm thickness generated 39 mW output when excited with
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 130.92.9.55
On: Sun, 28 Sep 2014 18:06:44
Appl. Phys. Rev. 1, 031104 (2014)
a sinusoidal force of 7.8 N at 100 Hz.23 Ren et al. used a rec-
tangular cymbal coupled with a single crystal lead magne-
sium niobate/lead titanate (PMN-PT) for vibrational energy
harvesting. The reported output power was 14.0 mW under a
cyclic force of 0.55 N at 500 Hz with a 17 g proof mass.42
Feenstra et al. used a 130-layer, 400 mm3 PZT stack coupled
to a mechanical force amplifier to harvest energy from the
tension created at the straps of a backpack during walking.22
They performed experiments using a backpack with a 220 N
load. Based on the results, the authors predicted a power out-
put of 0.4 mW from each piezoelectric stack under typical
walking conditions.22 Nevertheless, a piezoelectric stack can
be used directly as well if the input force is large enough: Xu
et al. used a 300-layer PZT stack with a volume of 1.75 cm3
for vibration energy harvesting, which generated 15 mW at
1767 Hz in resonance mode and 15 lW at 800 Hz in off-
resonance mode in response to a 1 Nrms force.21
In summary, large scale PEHs are currently being inves-
tigated for feasibility and reliability. Deployment of such
systems would require high investment costs; therefore,
commercialization can be expected only if their feasibility is
proven via thorough experimental studies. On the other
hand, there are several examples of mesoscale PEHs that can
be fabricated at low costs reported in the literature. Power
outputs of these devices cover a range from hundreds of lW
to a few mW, depending on the device size and input accel-
eration levels. The reported power levels can be sufficient to
operate a variety of electronic sensors or transducers.
However, these power outputs are obtained in controlled lab-
oratory environments using frequency controlled test equip-
ment. Practical application of PEHs to real systems has
certain challenges such as operation bandwidth, as it will be
discussed in Sec. IV. Nevertheless, new methods are being
proposed continuously to address these problems, and small
systems utilizing PEHs might be possible in the near future.
B. MEMS scale
High level integration capability of modern CMOS and
MEMS fabrication technologies enabled monolithic fabrica-
tion of complete sensor systems in sub-cm sized chips. The
small dimensions and low unit costs of these devices
increase the impact of the batteries on the overall system
size and cost. With this driving force, there have been a re-
markable number of studies on MEMS scale energy harvest-
ers in the last decade, most of which are piezoelectric type
due to their aforementioned advantages. However, unlike the
electronic circuits, functionality of vibrational energy har-
vesters depends primarily or strongly on their dimensions.
These devices harvest the mechanical energy of a moving
mass; therefore, the harvestable energy is reduced as the
mass gets smaller. Consequently, maximizing the output
power obtained per unit volume at a certain input accelera-
tion is crucial in MEMS scale. This becomes even more
challenging in thin films due their lower electromechanical
conversion efficiencies.43 Furthermore, MEMS scale PEHs
are mostly in unimorph configuration, which has lower elec-
tromechanical conversion factor compared to bimorphs,
since the microfabrication techniques are not suitable to
 031104-7 A. Toprak and O. Tigli
fabricate bimorph structures.44 Another problem caused by
the small device dimensions is the high resonance frequen-
cies. Thepresonance
ffiffiffiffiffiffiffiffiffi frequency of a spring-mass system is
equal to k=m, where k is the spring constant and m is the
mass. Small masses and relatively high spring constants of
sub-mm length cantilevers make it difficult to reduce the res-
onance frequencies below 250 Hz, which covers the spec-
trum of commonly available ambient vibration sources.4,5
One of the most commonly used piezoelectric materials
in MEMS scale energy harvesting is PZT. The first study
reporting utilization of thin film PZT for MEMS scale energy
harvesting purposes was published by Jeon et al. in 2003.45
The proposed energy harvester architecture was a rectangu-
lar cantilever consisting of a structural oxide layer, a sol-gel
coated PZT thin film layer, an interdigitated electrode layer,
and an SU-8 proof mass layer. Interdigitated electrodes ena-
ble the utilization of d33 mode operation in a cantilever
beam, allow the output voltage to be determined by the fin-
ger spacing rather than the piezoelectric material thickness,
and reduce the number of layers to be patterned.45 Figure 7
illustrates the electrode patterns and corresponding poling
directions in d31 and d33 mode cantilevers, which are con-
structed using plate electrodes and interdigitated electrodes,
respectively. It should be noted that interdigitated electrode
configuration necessitates the piezoelectric material be poled
using the same electrodes; and therefore, this configuration
is not applicable to non-ferroelectric piezoelectric materials.
The same group reported some follow-up studies, where they
reduced the stress-induced bending of fabricated cantile-
vers.46,47 Fabricated devices generated 1 lW power under a
peak acceleration of approximately 10.9 g at the resonance
frequency of 13.9 kHz. The high resonance frequency is
mainly caused by the small cantilever dimensions, which
yield high cantilever stiffness and small mass. Nevertheless,
it possible to reduce this resonance frequency by using thick
proof mass layers as illustrated in Figure 8, which shows the
SEM image of a prototype cantilever PEH before cantilever
suspension.
Fang et al. reported a d31 mode cantilever type harvester
with sol-gel PZT in 2006.48 They used a relatively large
FIG. 7. Illustration of the electrode patterns and corresponding poling direc-
tions in d31 and d33 mode cantilevers, which are constructed using plate elec-
trodes and interdigitated electrodes, respectively. It should be noted that
interdigitated electrode configuration necessitates the piezoelectric material
be poled using the same electrodes. Therefore, this configuration is not ap-
plicable to non-ferroelectric piezoelectric materials.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 130.92.9.55
On: Sun, 28 Sep 2014 18:06:44
Appl. Phys. Rev. 1, 031104 (2014)
FIG. 8. SEM image of a prototype MEMS scale cantilever type PEH before
cantilever suspension. A 50 lm thick SU8 layer is used as a proof mass in
order to reduce the resonance frequency.
(0.15 mm3) Ni proof mass to increase the power output and
reduce the resonance frequency. The fabricated device gener-
ated 2.16 lW power under 1 g peak acceleration at 609 Hz.48
The PZT deposition method in these studies is the sol-gel
method, which is probably the most commonly used method
for MEMS scale PEHs.49–51 In this method, a precursor solu-
tion is first spin coated on the substrate, followed by solvent
evaporation and thermal annealing cycles for crystallization.
Sol-gel deposition of PZT has certain advantages such as pre-
cise control of the composition, high film quality, and low
cost.52 On the other hand, one of the problems with the sol-
gel deposition of PZT is the small thickness: the thickness of
the PZT layer is in the order of 0.1 lm in each coating,46,48
which makes it inconvenient to fabricate a PZT layer thicker
than a few microns. Lee et al. designed an aerosol PZT depo-
sition system, which is capable of deposition rate of 0.1 lm/
min, and used it to fabricate both d31 and d33 mode cantile-
vers.53,54 However, the remnant polarization of the 5 lm PZT
film after poling under a 75 V/lm electric field was measured
as 7.3–9 lC/cm2,54 which is smaller than the values reported
for sol-gel coated PZT films.48,49 As a matter of fact, it is al-
ready known that the electromechanical coupling coefficient
of thin film PZT is considerably smaller than bulk ceramics.43
In order to get a higher electromechanical coupling, Aktakka
et al. proposed bonding bulk PZT ceramic on a Si substrate
and subsequently thinning it to obtain a MEMS scale de-
vice.43,55 Another approach to increase the electromechanical
coupling is using epitaxial PZT thin films.44 Epitaxial thin
film PZT growth on Si substrates with a SrTiO3 buffer layer
has been realized using pulsed laser deposition56 and molecu-
lar beam epitaxy.57,58 Growing the epitaxial films on top of Si
substrates allow the fabrication of microstructures using
standard Si microfabrication techniques on the same sub-
strates. Nevertheless, it is also possible to transfer epitaxial
films from one substrate to another. In an effort to reduce the
resonant frequency and increase the fracture strength,
Morimoto et al. transferred epitaxial PZT films grown on
MgO substrates to stainless steel cantilevers for energy har-
vesting purposes.59 Fabricated device, which consisted of
2.8 lm PZT film bonded to a 50 lm steel cantilever, was
reported to provide 244 lW output power at 126 Hz under
50 m/s2 acceleration without structural damage.59
 031104-8 A. Toprak and O. Tigli
Another commonly used piezoelectric material in
MEMS scale energy harvesters is aluminum nitride (AlN).
Although AlN has a lower electromechanical coupling coef-
ficient than PZT, it offers certain advantages. First of all,
AlN thin films can be easily deposited with sputtering, which
makes it also a possible candidate for CMOS compatible
PEHs.60 Second, AlN has a higher piezoelectric voltage con-
stant compared to PZT because of its lower dielectric con-
stant. High output voltage could be an important advantage
in the systems where the energy harvester output is rectified
using diodes for storage. Furthermore, the optimum piezo-
electric material thickness that will give the maximum power
output in a d31 mode cantilever is approximately three times
smaller for AlN compared to PZT, due to the higher Young’s
modulus of the former.61 Finally, AlN is a non-ferroelectric
material and tends to grow in [0001] direction, along which
it exhibits piezoelectric characteristics, when deposited using
sputtering. Therefore, unlike PZT, there is no need for exter-
nal poling. On the other hand, it should be noted that this
property also prevents AlN from being used in d33 mode can-
tilever devices. One of the important examples of AlN-based
energy harvesters at MEMS scale was reported by
Marzencki in 2005.62 While this study focused mostly on
theoretical analysis, the same group followed their earlier
work with a system on a package (SoP) energy harvester,
which consisted of an AlN-based cantilever type energy
harvester and a CMOS power management circuit.63 In the
experiments, a single cantilever with dimensions 1200 lm
 800 lm generated 0.8 lW on a resistive load under 2 g
acceleration. Operation of the power management circuit has
also been demonstrated; a maximum output power of
0.45 lW was transferred to a capacitor at 1.5 V under 2 g
acceleration.63 Elfrink et al. also designed AlN-based canti-
lever type energy harvesters and discussed the effects of
packaging on the performance.64 Their study showed that a
package with atmospheric pressure significantly reduces the
power output due to increased air damping.64 In a later study,
the same group fabricated vacuum packaged devices and
compared the performances of unpackaged, packaged (under
atmospheric pressure), and vacuum packaged devices.65
Measured output powers for the same type device with vac-
uum package, no package, and atmospheric package were
32.0 lW, 10.2 lW, and 0.27 lW.65 Defosseux et al. designed
and fabricated AlN-based cantilevers with circularly filleted
edges at the base to reduce the maximum stress.60 The per-
formance increase under vacuum package was also verified
by the experimental results presented in this study.
The majority of MEMS scale PEHs are cantilever type
devices, which are used to harvest the ambient vibrations.
However, pressure mode energy harvesters, which can har-
vest energy from pressure differences in the surrounding me-
dium, have also been reported in MEMS scale. As explained
before, pressure mode energy harvesting necessitates a struc-
ture that can provide mechanical isolation between its two
ends, such a diaphragm. In 2006, Horowitz et al. designed
and fabricated a diaphragm-type, MEMS scale PZT-based
acoustic energy harvester.16 Output power density of the fab-
ricated diaphragm was measured as 0.34 lW/cm2 at a sound
pressure level of 149 dB. In this study, the reasons for the
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 130.92.9.55
On: Sun, 28 Sep 2014 18:06:44
Appl. Phys. Rev. 1, 031104 (2014)
low efficiency were listed as the residual stress of dielectric
layer, low poling voltage limited by imperfect electrode
shapes, the quality of the sol-gel deposited PZT layer, and
the low quality factor of the fabricated diaphragm. The
authors claimed that the power density of such a device can
be as high as 250 lW/cm2 with the elimination of the listed
problems by improved fabrication methods.16 In another
study published in 2008, Kuehne et al. reported a ZnO-based
MEMS scale circular diaphragm for energy harvesting.66
Both small and large deflection tests were performed on the
fabricated diaphragm; however, output power was not high
enough to be directly measured in small deflection tests.
Based on the large deflection tests, the authors predicted a
power output around 1 lW in resonant mode operation.66
ZnO has piezoelectric properties comparable to those of
AlN; however, its low DC resistivity and diffusion problem
of Zn ions make it less preferred in MEMS scale energy har-
vesting applications.67
In summary, majority of the reported MEMS scale
PEHs are cantilever type, aiming to harvest vibrational
energy; however, diaphragm type PEHs are also fabricated
and tested. One of the most promising application areas for
MEMS scale PEHs is supplying power for IMDs. Reported
power outputs cover a wide range and can go as high as tens
of lW. With the continuously decreasing power consump-
tion in modern CMOS circuits, MEMS scale PEHs are
becoming more pronounced candidates as energy sources for
IMDs. However, implantable energy harvesters will have
additional packaging and biocompatibility issues, which will
have to be addressed before these devices come to market.
Currently, most of the studies on MEMS scale PEHs aim to
increase the piezoelectric material quality, improve the oper-
ation bandwidth, and optimize the mechanical architecture
for maximum electromechanical conversion efficiency.
C. Nanoscale
Piezoelectric energy harvesting at nanometer scale is
typically carried out by nanowires that are synthesized from
piezoelectric materials. Piezoelectric nanowire generators
were first demonstrated by Dr. Wang’s group at Georgia
Institute of Technology in 2006.68,69 In these studies, single
ZnO nanowires were bent by a conductive atomic force
microscope (AFM) probe used in contact mode. Figure 9(a)
shows the interaction between the AFM probe and a single
ZnO nanowire during a contact mode scan and Fig. 9(b) cor-
responding electrical potential on the nanowire.70 The semi-
conducting characteristics of the ZnO creates a Schottky
contact between the conductive tip of the AFM and the nano-
wire itself, creating a rectifying structure that allows observ-
able voltage output at an external resistive load.68 The same
group demonstrated that the nanowires can also be grown on
flexible plastic substrates, which might prove advantageous
for implantable energy harvesters.71
These initial studies proved the electrical energy genera-
tion capability of a single ZnO nanowire. However, a practi-
cal application requires thousands of nanowires to be
mechanically excited without a special device. Typical ZnO
nanowire fabrication processes yield very high spatial
 031104-9 A. Toprak and O. Tigli
densities; however, the orientations of these nanowires are
usually random, as illustrated by the SEM images shown in
Figure 10. In order to address this need, Wang et al. used a
zigzag-shaped top electrode on a vertically aligned ZnO
nanowire array.72 In this device, each tip on the zigzag-
shaped top electrode emulates an AFM tip that bends the
nanowire. The output power strongly depends on the density
and height uniformity of the nanowires as well as the contact
distance between the nanowire array and the top electrode.
Nevertheless, fabricated device generated a DC output power
of 1 pW from a 2 mm2 active area when excited with an
ultrasound wave.72 Another method to increase the number
of nanowires contributing to the electrical output is using
conical shaped nanowires.73 The conical shape provides that
the c-axis of a nanowire has a downward pointing angle,
which ensures a unipolar alignment in every layer. A flexi-
ble, multilayer generator fabricated using conical nanowires
was used to supply power to a simple LCD screen.73 The
same group demonstrated a self-powered sensor with wire-
less transmission74 and a device harvesting energy from a
rotating tire that can possibly be used as a self-powered pres-
sure or speed sensor.75,76 As demonstrated by these nascent
approaches, energy harvesting from nanostructures is a
promising field with a wide application range.77–80 This
growing interest helped founding a dedicated scientific jour-
nal on Nano Energy and its applications.
FIG. 10. SEM image showing the random orientation of ZnO nanowires
grown on the same substrate. The inset shows a close-up view of individual
nanowires showing the hexagonal wurtzite crystal structure.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 130.92.9.55
On: Sun, 28 Sep 2014 18:06:44
Appl. Phys. Rev. 1, 031104 (2014)
FIG. 9. (a) Interaction of an AFM probe
with a ZnO nanowire during a contact
mode scan and (b) corresponding elec-
trical potential on the nanowire. (Image
courtesy of Professor Z. L. Wang,
Georgia Tech.) Reprinted with permis-
sion from Y. Gao and Z. L. Wang,
Nano Lett. 7(8), 2499–2505 (2007).
Copyright 2007 ACS Publications.
D. Comparative summary of cantilever type
piezoelectric energy harvesters
Performance evaluation method of a PEH is mainly
determined by its operation mode. Therefore, different PEH
types are tested under different conditions. To give an exam-
ple, a cantilever type device is usually tested by applying a
periodic mechanical vibration, whereas a membrane type,
pressure mode device is tested by applying a slowly varying
pressure. There is also no standard figure of merit that can be
used to compare PEHs that are designed for different
operation modes. Consequently, a rational comparison is
possible only among the PEHs having the same architecture
and operation mode. The most prominent PEH architecture
is the cantilever beam, which is typically used in
resonant vibration mode. Table I presents a comparison of
different resonant mode cantilever type PEHs. Volumes
listed in the table include only the cantilever beams and
proof masses.
It is evident from the table that PZT and AlN are the
most widely preferred piezoelectric materials for cantilever
type PEHs. Although it varies to a great extent within the
devices of same materials, normalized power density (NPD)
values for PZT devices are higher than those of AlN devices
in average, which is mostly due to the higher electromechan-
ical coupling of PZT. Among the PZT devices, bulk and epi-
taxial materials have better NPD in comparison to sol-gel
and aerosol deposited films. This result indicates that epitax-
ial growth method can eliminate the disadvantage of low
electromechanical coupling of thin film PZT. In the given ta-
ble, the highest NPD is 104.7 lW/mm3/g2, which was
obtained using a bulk PZT based device.36 The bimorph
architecture of this hand-assembled device, along with the W
proof mass placed at the very tip, yields an NPD higher than
other bulk and epitaxial PZT based devices. On the other
hand, the lowest NPD in the given list is 0.055 lW/mm3/g2,
obtained from an AlN based device fabricated on a steel sub-
strate.37 This value is substantially lower than other AlN
based devices, which is mostly due to the sub-optimal thick-
ness ratio between the AlN film and the steel substrate.
It should be noted that the NPD of a device depends on
a number of factors, including the density and location of the
proof mass, rather than solely on the performance of the pie-
zoelectric material. To give an example, a device with a W
proof mass can store and convert more mechanical energy
compared to its counterpart with a Si proof mass. On the
other hand, creating a Si proof mass is much easier than
 031104-10 A. Toprak and O. Tigli Appl. Phys. Rev. 1, 031104 (2014)

TABLE I. Comparison of different resonant mode cantilever type PEHs. Volumes listed in the table include only the cantilever beams and proof masses.

Normalized power
Authors Material Mode Volume (mm3) Power (lW) Acceleration (g) Frequency (Hz) density (lW/mm3/g2)

Roundy and Wright7 Bulk PZT d31 535 375 0.255 120 10.8
Aktakka et al.55 Bulk PZT d31 19.7 2.74 0.1 167 13.91
Reilly et al.36 Bulk PZT d31 613 4000 0.25 100 104.7
Fang et al.48 Sol-gel PZT d31 0.20 2.16 1.0 461 10.80
Shen et al.50 Sol-gel PZT d31 0.77 0.24 0.5 184 1.25
Muralt et al.51 Sol-gel PZT d33 0.32 1.40 2.0 855 1.09
Lee et al.54 Aerosol PZT d31 0.43 2.10 2.0 256 1.24
d33 0.61 1.29 2.0 214 0.53
Isarakorn et al.58 Epitaxial PZT d31 0.15 13.0 1.0 2300 85.25
Morimoto et al.59 Epitaxial PZT d31 4.88 5.30 0.51 126 4.18
Cao et al.37 AlN d31 114 0.24 0.2 69 0.055
Marzencki et al.63 AlN d31 0.34 0.80 2.0 1495 0.59
Elfrink et al.65 AlN d31 1.28 1.46 (atm) 0.2 597 2.85
6.90 (vac) 0.2 599 13.48
Defosseux et al.60 AlN d31 1.32 0.59 (atm) 0.2 212 11.17
0.53 (vac) 0.07 212 84.45

creating a W proof mass for a microfabricated device.
Furthermore, there are other important considerations that
are not listed in this comparison, such as operation band-
width. Therefore, NPD of a PEH does not indicate a com-
plete measure of its overall performance. Consequently, the
research on PEHs addresses a wide range of issues other
than the NPD. The most prominent ones of these issues are
discussed in Sec. IV.
IV. CHALLENGES AND PROPOSED SOLUTIONS IN
PIEZOELECTRIC ENERGY HARVESTERS
A. Bandwidth
Most of the vibrational energy harvesters are designed
to operate in resonance mode and the half-power bandwidth
is usually small. This is one of the most important challenges
of energy harvesting since the frequency of ambient vibra-
tions varies within a wide range, from 1 Hz for heel tapping5
up to 240 Hz for an electric tea pot.4 There are some vibra-
tion sources such as industrial machinery that can supply
constant frequency vibrations over relatively long periods.
To give an example, a synchronous AC electric motor run-
ning in a factory during the work hours would provide vibra-
tions at the AC line frequency for 8 h a day. However, even
the harvesters using this type sources might require tuning
due to frequency shifts in the vibration source or drifts in the
resonance frequency. This issue becomes more pronounced
in case of random ambient vibration sources, which would
span a broader frequency range. Obviously, it would not be
feasible to continuously measure the vibration and manually
tune the resonance frequency of the harvester in this case.
Two possible solutions to this problem are dynamic reso-
nance frequency tuning and wide bandwidth harvesters. Both
methods have been investigated in the literature by several
research groups.
Roundy and Zhang investigated the feasibility of
actively tuning the resonance frequency by using actuators
and listed the requirements for a net power increase.81
Leland and Wright demonstrated a passive tuning method in
which an axial compressive stress is applied to the cantilever
in order to change the resonance frequency.82 In the experi-
ments, they used a steel vise to manually apply the stress and
suggested that the system might be improved to form a
closed-loop controller for automatic tuning.82 Challa et al.
used attractive and repulsive magnetic forces to apply a pre-
stress to the cantilever for tuning, as shown in Figure 11.83
Figure 11(a) shows the setup of the resonance frequency tun-
ing mechanism, and Fig. 11(b) shows the experimental
results obtained from a prototype device. In this method, res-
onance frequency of the piezoelectric beam is tuned by mod-
ifying its position between the magnets. In Figure 11(b), the
thin lines show the power output for a single beam position,
whereas the thick solid line shows the power output in case
of successful tuning between 22 and 32 Hz. In a later study,
the same group demonstrated a closed loop operation of the
same magnetic tuning mechanism.84 The designed system,
which included a data acquisition module to gather data and
a DC motor to move the piezoelectric beam, successfully
tuned the resonance frequency between 13 and 22 Hz.84 Al-
Ashtari et al. also used magnetic force for resonance tuning
of a cantilever beam.85 They placed a magnet to the cantile-
ver tip and applied a force on this magnet by another magnet
placed in close proximity. Resonance frequency was tuned
by adjusting the distance between the two magnets.85
Eichhorn et al. designed a system where the axial pre-stress
is applied by manually adjusting the position of thin arms
beside the cantilever.86 In a later study, the same group pro-
posed using piezoelectric actuation to adjust the stress
exerted by the arms.87 Dylan et al. demonstrated a frequency
tuning method for membrane type vibrational energy har-
vesters by applying a pre-load to the membrane.88 Gu and
Livermore used the centrifugal force to design a passive self-
tuning PEH for rotating applications.89,90
Most of the proposed resonance frequency tuning mech-
anisms requires either manual adjustment or control systems
that will dissipate at least part of the harvested power.

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 130.92.9.55
On: Sun, 28 Sep 2014 18:06:44
 031104-11 A. Toprak and O. Tigli Appl. Phys. Rev. 1, 031104 (2014)

FIG. 11. (a) Setup of the resonance frequency tuning mechanism using magnetic force technique, and (b) experimental results obtained from a prototype de-
vice. Resonance frequency of the piezoelectric beam is tuned by modifying its position between the magnets. In part (b), the thin lines show the power output
for a single beam position, whereas the thick solid line shows the power output in case of successful tuning. Reproduced by permission from Challa et al.,
Smart Mater. Struct. 17(1), 015035 (2008). Copyright 2008 IOP Publishing.

Therefore, these systems would not be feasible if the vibra- and electrical parts are built on the same substrate.
tion frequencies are changing rapidly. In this case, wide- Monolithic fabrication reduces the overall device size and
bandwidth architectures can provide a better solution. cost. Furthermore, the interconnections between the mechan-
However, linear systems have a gain-bandwidth trade-off, ical and electrical parts are completed by microprocessing
which means that the peak output power has to drop if the
bandwidth is increased in a single device. Therefore, wide
bandwidth energy harvester research is mostly focused on
nonlinear systems. As an alternative, it is also possible to use
parallel or series connected linear cantilever arrays to
improve the bandwidth.91,92
Erturk et al. investigated the piezomagnetoelastic struc-
ture, which consists of a ferromagnetic beam placed between
two closely spaced magnets as shown in Figure 12, for
broadband energy harvesting.93 The experimental results
showed that the chaotic behavior of the bistable structure
improved the bandwidth and the power output.93 Stanton
et al. used repulsive magnetic forces to create a bistable pie-
zoelectric cantilever.94 Karami and Inman designed a wide
bandwidth piezoelectric cantilever using the same principle
to harvest energy from heartbeat vibrations, which covers a
wide frequency range.95 Another method to create bistable
structures is using buckled beams as demonstrated by Sneller
et al.96
Structures with nonlinear stiffness characteristics can
also be used to improve the bandwidth. In 2009, Marinkovic
and Koser demonstrated the wide bandwidth operation of
doubly fixed beams.97 Double-clamped beams undergo both
bending and stretching under vibration, and the stretching
stiffness depends on the displacement, causing a nonlinear
spring motion.98 Hajati and Kim reported a PEH with a
bandwidth more than 20% of the center frequency using the
doubly clamped structure.98 There are some other methods
proposed to improve bandwidth such as using mechanical
stoppers.99–102 Mechanical stoppers have the advantage of
being simple and robust; but on the other hand, they require
a slow frequency sweep to reach and stay in the wide band-
width region.103 More detailed information on wide band-
width vibrational energy harvesting methods can be found in
some recent review studies.103,104

B. CMOS compatibility FIG. 12. (a) Bistable piezomagnetoelastic structure for broadband energy har-
vesting, and (b) experimental results obtained using a piezomagnetoelastic energy
CMOS compatible MEMS processes allow fabrication harvester prototype. Reproduced with permission from Erturk et al., Appl. Phys.
of monolithic sensors and transducers, where the mechanical Lett. 94(25), 254102-254103 (2009). Copyright 2009 AIP Publishing LLC.

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 130.92.9.55
On: Sun, 28 Sep 2014 18:06:44
 031104-12 A. Toprak and O. Tigli
rather than wire bonding, which increases the reliability and
reduces the parasitic effects. The key to a monolithic device
is developing a CMOS compatible process, which does not
harm the electronic circuits while fabricating the mechanical
parts.
Most common material used in MEMS scale PEHs is
thin-film PZT, which is usually deposited with sol-gel
method. Crystallization temperature of PZT is around
600–700  C; therefore, the coated films must be annealed at
these temperatures in order to obtain piezoelectric character-
istics. Although there are other methods to deposit thin film
PZT such as aerosol deposition53,54 or epitaxial
growth,56,57,59 all these methods require high temperature for
proper crystallization. Consequently, these devices are not
CMOS compatible. Aktakka et al. proposed integrating bulk
PZT on silicon using a low temperature bonding process as a
solution to this problem.43,55,105 The thickness of bulk PZT
is on the order of 200 lm, which is too thick for a MEMS
scale device; therefore, the bulk PZT crystal is mechanically
thinned after the bonding. Therefore, the proposed process
brings CMOS compatibility at the cost of increased fabrica-
tion complexity.
Some piezoelectric materials such as ZnO and AlN can
be grown c-axis oriented using sputter deposition.106
Required temperatures are usually low and no subsequent
poling is required.67 Therefore, these materials can be used
to form CMOS-compatible piezoelectric layers.60,63,64 AlN
is more compatible with silicon semiconductor technology
compared to ZnO and has a higher resistivity.67 These two
advantages make AlN a better candidate for CMOS compati-
ble energy harvesting applications. There are several AlN
based vibrational PEHs reported in the litera-
ture.37,60,63–65,107,108 The output power density reported in
AlN devices is quite promising, which might further increase
the interest in this material.44
Another piezoelectric material class that can be used for
CMOS compatible energy harvesters is the piezoelectric poly-
mers. Most commonly used piezoelectric polymers are PVDF
and its copolymer, polyvinylidene fluoride-trifluoroethylene
(PVDF-TrFE). PVDF requires mechanical stretching to switch
to its piezoelectric phase; therefore, it cannot be used on non-
flexible substrates for piezoelectric applications. On the other
hand, PVDF-TrFE films can crystallize into its piezoelectric b-
phase without mechanical stretching if they are annealed at a
temperature between its Curie temperature and melting temper-
ature.109 Required annealing temperature depends on the molar
ratio of the copolymer and it is typically lower than 150  C.110
This low process temperature makes PVDF-TrFE a suitable
candidate for CMOS compatible piezoelectric applications.111
C. Biocompatibility
One of the promising applications of PEHs is IMDs.
There is considerable effort in creating self-sustained IMDs
using PEHs as reported in the literature.30,95,112–114 One of
the most important considerations in IMDs is the biocompat-
ibility of the used materials. The most commonly used piezo-
electric material, PZT, is not favorable for implants due to
its lead content, which is a toxic material. An alternative
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 130.92.9.55
On: Sun, 28 Sep 2014 18:06:44
Appl. Phys. Rev. 1, 031104 (2014)
lead-free piezoelectric material is potassium sodium niobate
(KNN), which has piezoelectric properties similar to those of
PZT.44,115 Operation of KNN-based PEHs has been demon-
strated in MEMS scale, and their performance has been
found to be comparable to PZT based devices.116
Biocompatibility of another widely used piezoelectric mate-
rial, AlN, has been verified by in vivo and in vitro tests.117,118
Heidrich et al. proposed a corrugated membrane type AlN-
based energy harvester designed for aperiodic vibrations at
low acceleration levels, which is the expected input type for
an implanted device.119,120 Jackson et al. demonstrated the
deposition of AlN thin films on polyimide to create flexible
devices, which can also be advantageous for implanted devi-
ces.118 Another alternative method to create a flexible,
implantable energy harvester is using PVDF-based polymers.
Biocompatibility of PVDF and PVDF-TrFE has also been
verified.121,122 Biocompatibility and flexibility of these poly-
mers make them a suitable candidate for implantable energy
harvesters.30,114,123
V. CONCLUSION
Piezoelectric devices have attracted more attention than
other mechanical energy harvesting methods due to their cer-
tain advantages such as higher power output density, scal-
ability, and the simplicity of the required external circuitry.
The research on mechanical energy harvesting mostly
focused on converting the energy in ambient vibrations;
however, PEHs can convert the energy in varying pressure
as well, which increases their potential applications.
Piezoelectric energy harvesting has been demonstrated
in different scales from several m2 piezoelectric floors to
sub-micron nanowire arrays. Large scale piezoelectric
energy harvesting demonstrations so far aim to provide
power for nearby lighting or sensing systems in an effort to
reduce the effective energy consumption of the overall sys-
tem. On the other hand, the ultimate goal of smaller scale
energy harvesters is mostly creating self-powered systems.
Improvements in piezoelectric energy harvesting and
decreasing power requirements of CMOS circuits brought
the supplied and required power levels close to each other,
which brings us closer to such systems. There are still certain
challenges in PEHs such as increasing the operation band-
width, developing CMOS compatible processes for inte-
grated devices, or fabricating biocompatible devices for
implants. Another important issue arises due to the availabil-
ity level of harvestable energy, which is almost impossible to
predict beforehand. Therefore, the system should be able to
harvest energy whenever it is available and store it for later
use. Storage of the harvested energy requires the AC outputs
of PEHs be rectified using impedance-matched electrical cir-
cuits. The rectification and storage processes have their own
losses as well, which would reduce the already low amount
of harvested energy. Nevertheless, there is an ongoing
research effort focused on these problems and new solutions
are being proposed continuously.
Currently, the most prominent candidates for self-
powered systems are WSN and IMD. Eliminating the battery
in a WSN will reduce the cost, size, and weight of the node as
 031104-13 A. Toprak and O. Tigli Appl. Phys. Rev. 1, 031104 (2014)

28
well as increasing its lifetime. Batteryless IMDs have the V. Parsonnet, G. H. Myers, R. I. Zucker, H. Lotman, and M. M. Asa,
additional advantage of eliminating invasive surgical proce- ASAIO J. 9(1), 174–177 (1963).
29
G. Lewin, G. Myers, V. Parsonnet, and K. Raman, Trans. - Am. Soc.
dures on the patient for battery replacement. Such self- Artif. Inter. Organs 14, 215 (1968).
powered systems have already been demonstrated in special 30
E. H€asler, L. Stein, and G. Harbauer, Ferroelectrics 60(1), 277–282
test environments, and we predict that they will be available (1984).
31
J. F. Antaki, G. E. Bertocci, E. C. Green, A. Nadeem, T. Rintoul, R. L.
for real applications in the near future. Even if eliminating the
Kormos, and B. P. Griffith, ASAIO J. 41(3), M588–M595 (1995).
batteries might not be possible for some applications depend- 32
J. Kymissis, C. Kendall, J. Paradiso, and N. Gershenfeld, paper presented
ing on the power requirements and size restrictions, piezoelec- at the 2nd International Symposium on Wearable Computers, Digest of
tric energy harvesting can be still useful by recharging the Papers (1998).
33
N. S. Shenck and J. A. Paradiso, IEEE Micro 21(3), 30–42 (2001).
battery and thus increasing the lifetime of the device. 34
M. Umeda, K. Nakamura, and S. Ueha, Jpn. J. Appl. Phys., Part 1 36,
3146–3151 (1997).
35
P. Glynne-Jones, S. Beeby, and N. White, IEE Proc.: Sci., Meas. Technol.
148(2), 68–72 (2001).
1 36
L. Zuo and X. D. Tang, J. Intell. Mater. Syst. Struct. 24(11), 1405–1430 E. K. Reilly, F. Burghardt, R. Fain, and P. Wright, Smart Mater. Struct.
(2013). 20(12), 125006 (2011).
2 37
S. J. Roundy, Ph.D. dissertation, The University of California, Berkeley, Z. P. Cao, J. Y. Zhang, and H. Kuwano, Sens. Actuators, A 179, 178–184
2003. (2012).
3 38
J. Baker, S. Roundy, and P. Wright, in 3rd International Energy S. Mehraeen, S. Jagannathan, and K. A. Corzine, IEEE Trans. Ind.
Conversion Engineering Conference, San Francisco, CA, USA (2005). Electron. 57(3), 820–830 (2010).
4 39
E. K. Reilly, L. M. Miller, R. Fain, and P. Wright, in PowerMEMS, P. Becker, E. Hymon, B. Folkmer, and Y. Manoli, Sens. Actuators, A
Washington, DC, USA (2009). 202, 155–161 (2013).
5 40
S. Roundy, P. K. Wright, and J. Rabaey, Comput. Commun. 26(11), C. K. Mo, L. J. Radziemski, and W. W. Clark, Smart Mater. Struct. 19(2),
1131–1144 (2003). 025016 (2010).
6 41
S. Roundy, E. S. Leland, J. Baker, E. Carleton, E. Reilly, E. Lai, B. Otis, W. Wang, T. Yang, X. Chen, X. Yao, and Q. Zhou, paper presented at the
J. M. Rabaey, P. K. Wright, and V. Sundararajan, IEEE Pervasive Applications of Ferroelectrics (ISAF/PFM), 2011 International
Comput. 4(1), 28–36 (2005). Symposium on and 2011 International Symposium on Piezoresponse
7
S. Roundy and P. K. Wright, Smart Mater. Struct. 13(5), 1131–1142 Force Microscopy and Nanoscale Phenomena in Polar Materials (2011).
42
(2004). B. Ren, S. W. Or, X. Zhao, and H. Luo, J. Appl. Phys. 107(3), 034501
8
A. Marin, S. Bressers, and S. Priya, J. Phys. D: Appl. Phys. 44(29), (2010).
43
295501 (2011). E. E. Aktakka, R. L. Peterson, and K. Najafi, paper presented at the
9
ANSI/IEEE Std 176–1987, 0_1, 1988. International Electron Devices Meeting, 2010.
10 44
M. Umeda, K. Nakamura, and S. Ueha, Jpn. J. Appl. Phys., Part 1 35, S.-G. Kim, S. Priya, and I. Kanno, MRS Bull. 37(11), 1039–1050 (2012).
45
3267–3273 (1996). Y. Jeon, R. Sood, L. Steyn, and S. G. Kim, in CIRP Seminar on Micro
11
S. Kim, W. W. Clark, and Q. M. Wang, J. Intell. Mater. Syst. Struct. and Nano Technology (2003).
46
16(10), 847–854 (2005). Y. B. Jeon, R. Sood, J. H. Jeong, and S. G. Kim, Sens. Actuators, A
12
S. Kim, W. W. Clark, and Q. M. Wang, J. Intell. Mater. Syst. Struct. 122(1), 16–22 (2005).
47
16(10), 855–863 (2005). R. Sood, Y. B. Jeon, J. h. Jeong, and S. G. Kim, paper presented at the
13
S. Kim, W. W. Clark, and Q. M. Wang, Proc. SPIE 5055, 307–318 Proceedings of Solid-State Sensor and Actuator Workshop, South
(2003). Carolina, 2004.
14 48
M. Deterre, E. Lefeuvre, and E. Dufour-Gergam, Smart Mater. Struct. H. B. Fang, J. Q. Liu, Z. Y. Xu, L. Dong, L. Wang, D. Chen, B. C. Cai,
21(8), 085004 (2012). and Y. Liu, Microelectron. J. 37(11), 1280–1284 (2006).
15 49
C. K. Mo, L. J. Radziemski, and W. W. Clark, Smart Mater. Struct. 19(7), D. Shen, J.-H. Park, J. Ajitsaria, S.-Y. Choe, H. C. Wikle, and D.-J. Kim,
075010 (2010). J. Micromech. Microeng. 18(5), 055017 (2008).
16 50
S. B. Horowitz, M. Sheplak, L. N. Cattafesta, and T. Nishida, D. N. Shen, J. H. Park, J. H. Noh, S. Y. Choe, S. H. Kim, H. C. Wikle,
J. Micromech. Microeng. 16(9), S174–S181 (2006). and D. J. Kim, Sens. Actuators, A 154(1), 103–108 (2009).
17 51
F. Liu, A. Phipps, S. Horowitz, K. Ngo, L. Cattafesta, T. Nishida, and M. P. Muralt, M. Marzencki, B. Belgacem, F. Calame, and S. Basrour,
Sheplak, J. Acoust. Soc. Am. 123, 1983 (2008). Procedia Chem. 1(1), 1191–1194 (2009).
18 52
P. Mandurino, G. Bertolotto Bianc, and R. Garbin, Proc. SPEI 4331, J. Cheng and Z. Meng, Thin Solid Films 385(1–2), 5–10 (2001).
53
408–419 (2001). B. Lee, W. Wu, W. Shih, D. Vasic, and F. Costa, paper presented at the
19
E. Minazara, D. Vasic, F. Costa, and G. Poulin, Ultrasonics 44, IEEE Ultrasonics Symposium, 2007.
54
e699–e703 (2006). B. Lee, S. Lin, W. Wu, X. Wang, P. Chang, and C. Lee, J. Micromech.
20
X.-r. Chen, T.-q. Yang, W. Wang, and X. Yao, Ceram. Int. 38, Microeng. 19(6), 065014 (2009).
55
S271–S274 (2012). E. E. Aktakka, R. L. Peterson, and K. Najafi, in 16th International Solid-
21
T.-B. Xu, E. J. Siochi, J. H. Kang, L. Zuo, W. Zhou, X. Tang, and X. State Sensors, Actuators and Microsystems Conference
Jiang, Smart Mater. Struct. 22(6), 065015 (2013). (TRANSDUCERS), Beijing (2011), pp. 1649–1652.
22 56
J. Feenstra, J. Granstrom, and H. Sodano, Mech. Syst. Signal Process. E. Reilly and P. Wright, J. Micromech. Microeng. 19(9), 095014 (2009).
57
22(3), 721–734 (2008). D. Isarakorn, A. Sambri, P. Janphuang, D. Briand, S. Gariglio, J. M.
23
H. W. Kim, A. Batra, S. Priya, K. Uchino, D. Markley, R. E. Newnham, Triscone, F. Guy, J. W. Reiner, C. H. Ahn, and N. F. de Rooij,
and H. F. Hofmann, Jpn. J. Appl. Phys., Part 1 43(9A), 6178–6183 (2004). J. Micromech. Microeng. 20(5), 055008 (2010).
24 58
East Japan Railway Company, “Demonstration experiment of the power D. Isarakorn, D. Briand, P. Janphuang, A. Sambri, S. Gariglio, J. M.
generating floor at Tokyo station,” see http://www.jreast.co.jp/e/develop- Triscone, F. Guy, J. W. Reiner, C. H. Ahn, and N. F. de Rooij, Smart
ment/press/20080111.pdf. Mater. Struct. 20(2), 025015 (2011).
25 59
Innowattech Energy Harvesting Systems, Innowattech achieves a break- K. Morimoto, I. Kanno, K. Wasa, and H. Kotera, Sens. Actuators, A
through in alternative energy by generating electricity from road traffic in 163(1), 428–432 (2010).
60
a recent complete-system pilot project, see http://www.innowattech.co.il/ M. Defosseux, M. Allain, E. Defay, and S. Basrour, Sens. Actuators, A
events.aspx. 188, 489–494 (2012).
26 61
Innowattech Energy Harvesting Systems, Rail Project, see http://www.in- F. Stoppel, C. Schr€oder, F. Senger, B. Wagner, and W. Benecke, Procedia
nowattech.co.il/events.aspx. Eng. 25, 721–724 (2011).
27 62
D. Hill and N. Tong, “Assessment of piezoelectric materials for roadway M. Marzencki, S. Basrour, B. Charlot, A. Grasso, L. Valbin, and M.
energy harvesting,” California Energy Commission, CEC-500-2013-007, Colin, paper presented at the Symposium on Design, Test, Integration and
2013. Packaging of MEMS/MOEMS (DTIP’05), 1–3 June 2005.

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 130.92.9.55
On: Sun, 28 Sep 2014 18:06:44
 031104-14 A. Toprak and O. Tigli Appl. Phys. Rev. 1, 031104 (2014)

63 96
M. Marzencki, Y. Ammar, and S. Basrour, Sens. Actuators, A 145–146, A. J. Sneller, P. Cette, and B. P. Mann, Proc. Inst. Mech. Eng., Part I: J.
363–370 (2008). Syst. Control Eng. 225, 497–509 (2011).
64 97
R. Elfrink, T. M. Kamel, M. Goedbloed, S. Matova, D. Hohlfeld, Y. van B. Marinkovic and H. Koser, Appl. Phys. Lett. 94(10), 103505 (2009).
98
Andel, and R. van Schaijk, J. Micromech. Microeng. 19(9), 094005 A. Hajati and S.-G. Kim, Appl. Phys. Lett. 99(8), 083105 (2011).
99
(2009). H. Liu, C. Lee, T. Kobayashi, C. J. Tay, and C. Quan, Sens. Actuators, A
65
R. Elfrink, M. Renaud, T. Kamel, C. De Nooijer, M. Jambunathan, M. 186, 242–248 (2012).
100
Goedbloed, D. Hohlfeld, S. Matova, V. Pop, and L. Caballero, H. Liu, C. Lee, T. Kobayashi, C. J. Tay, and C. Quan, Smart Mater.
J. Micromech. Microeng. 20(10), 104001 (2010). Struct. 21(3), 035005 (2012).
66 101
I. Kuehne, D. Marinkovic, G. Eckstein, and H. Seidel, Sens. Actuators, A L. Dhakar, H. Liu, F. Tay, and C. Lee, Sens. Actuators, A 199, 344–352
142(1), 292–297 (2008). (2013).
67 102
S. Trolier-McKinstry and P. Muralt, J. Electroceram. 12(1–2), 7–17 H. Liu, C. J. Tay, C. Quan, T. Kobayashi, and C. Lee,
(2004). J. Microelectromech. Syst. 20(5), 1131–1142 (2011).
68 103
Z. L. Wang and J. Song, Science 312(5771), 242–246 (2006). J. Twiefel and H. Westermann, J. Intell. Mater. Syst. Struct. 24(11),
69
J. Song, J. Zhou, and Z. L. Wang, Nano Lett. 6(8), 1656–1662 (2006). 1291–1302 (2013).
70 104
Y. Gao and Z. L. Wang, Nano Lett. 7(8), 2499–2505 (2007). L. Tang, Y. Yang, and C. K. Soh, J. Intell. Mater. Syst. Struct. 21(18),
71
P. X. Gao, J. H. Song, J. Liu, and Z. L. Wang, Adv. Mater. 19(1), 67–72 1867–1897 (2010).
105
(2007). E. E. Aktakka, R. L. Peterson, and K. Najafi, IEEE Trans. Electron
72
X. Wang, J. Song, J. Liu, and Z. L. Wang, Science 316(5821), 102–105 Devices 60(6), 2022–2030 (2013).
106
(2007). K. Uchino, in Introduction to Piezoelectric Actuators and Transducers,
73
Y. Hu, Y. Zhang, C. Xu, G. Zhu, and Z. L. Wang, Nano Lett. 10(12), edited by D. T. I. Center (2003), ADA429659.
107
5025–5031 (2010). T.-T. Yen, T. Hirasawa, P. K. Wright, A. P. Pisano, and L. Lin,
74
Y. Hu, Y. Zhang, C. Xu, L. Lin, R. L. Snyder, and Z. L. Wang, Nano J. Micromech. Microeng. 21(8), 085037 (2011).
108
Lett. 11(6), 2572–2577 (2011). M. Renaud, K. Karakaya, T. Sterken, P. Fiorini, C. Van Hoof, and R.
75
Y. Hu, C. Xu, Y. Zhang, L. Lin, R. L. Snyder, and Z. L. Wang, Adv. Puers, Sens. Actuators, A 145–146, 380–386 (2008).
109
Mater. 23(35), 4068–4071 (2011). H. Ohigashi and K. Koga, Jpn. J. Appl. Phys., Part 2 21(8), L455–L457
76
L. Lin, Y. Hu, C. Xu, Y. Zhang, R. Zhang, X. Wen, and Z. Lin Wang, (1982).
110
Nano Energy 2(1), 75–81 (2013). T. R. Dargaville, M. C. Celina, J. M. Elliott, P. M. Chaplya, G. D. Jones,
77
B. Kumar and S.-W. Kim, Nano Energy 1(3), 342–355 (2012). D. M. Mowery, R. A. Assink, R. L. Clough, and J. W. Martin,
78
X. Wang, Nano Energy 1(1), 13–24 (2012). Characterization, Performance and Optimization of PVDF as a
79
Z. L. Wang and W. Wu, Angew. Chem., Int. Ed. Engl. 51(47), Piezoelectric Film for Advanced Space Mirror Concepts (Technical
11700–11721 (2012). Report, SAND2005-6846, Sandia National Laboratories).
80 111
Z. L. Wang, G. Zhu, Y. Yang, S. H. Wang, and C. F. Pan, Mater. Today S. R. Oh, K. Yao, and F. E. H. Tay, J. Micromech. Microeng. 23(9),
15(12), 532–543 (2012). 095023 (2013).
81 112
S. Roundy and Y. Zhang, Proc. SPIE 5649, 373–384 (2005). M. Deterre, E. Lefeuvre, Y. Zhu, M. Woytasik, A. Bosseboeuf, B.
82
E. S. Leland and P. K. Wright, Smart Mater. Struct. 15(5), 1413–1420 Boutaud, and R. Dal Molin, paper presented at the IEEE 26th
(2006). International Conference on Micro Electro Mechanical Systems
83
V. R. Challa, M. G. Prasad, Y. Shi, and F. T. Fisher, Smart Mater. Struct. (MEMS), 2013.
113
17(1), 015035 (2008). M. Deterre, E. Lefeuvre, Y. Zhu, M. Woytasik, B. Boutaud, and R. D.
84
V. R. Challa, M. G. Prasad, and F. T. Fisher, Smart Mater. Struct. 20(2), Molin, J. Microelectromech. Syst. 23(3), 651–660 (2013).
114
025004 (2011). J. A. Potkay and K. Brooks, paper presented at the 2nd International
85
W. Al-Ashtari, M. Hunstig, T. Hemsel, and W. Sextro, Smart Mater. Conference on Bioinformatics and Biomedical Engineering, ICBBE, 2008.
115
Struct. 21(3), 035019 (2012). K. Shibata, K. Suenaga, K. Watanabe, F. Horikiri, A. Nomoto, and T.
86
C. Eichhorn, F. Goldschmidtboeing, and P. Woias, J. Micromech. Mishima, Jpn. J. Appl. Phys., Part 1 50(4), 1503 (2011).
116
Microeng. 19(9), 094006 (2009). I. Kanno, T. Ichida, K. Adachi, H. Kotera, K. Shibata, and T. Mishima,
87
C. Eichhorn, F. Goldschmidtboeing, P. Woias, and Y. Porro, in Power Sens. Actuators, A 179, 132–136 (2012).
117
MEMS (2009), pp. 45–48. B. A. Murphy, Ph.D. dissertation, Wayne State University, 2008.
88 118
J. M. Dylan, M. Y. John, J. A. Michael, and F. B. David, Smart Mater. N. Jackson, L. Keeney, and A. Mathewson, Smart Mater. Struct. 22(11),
Struct. 17(6), 065021 (2008). 115033 (2013).
89 119
L. Gu and C. Livermore, Appl. Phys. Lett. 97(8), 081904 (2010). N. Heidrich, F. Kn€ obber, V. Polyakov, V. Cimalla, W. Pletschen, R. E.
90
L. Gu and C. Livermore, Smart Mater. Struct. 21(1), 015002 (2012). Sah, L. Kirste, S. Leopold, S. Hampl, and O. Ambacher, Sens. Actuators,
91
J. Q. Liu, H. B. Fang, Z. Y. Xu, X. H. Mao, X. C. Shen, D. Chen, H. A 195, 32–37 (2013).
120
Liao, and B. C. Cai, Microelectron. J. 39(5), 802–806 (2008). N. Heidrich, F. Knobber, R. E. Sah, W. Pletschen, S. Hampl, V. Cimalla,
92
J. Dayou, W. Y. H. Liew, and M. S. Chow, Microelectron. J. 43(7), and V. Lebedev, paper presented at the 16th International Solid-State
484–491 (2012). Sensors, Actuators and Microsystems Conference (TRANSDUCERS), 2011.
93 121
A. Erturk, J. Hoffmann, and D. J. Inman, Appl. Phys. Lett. 94(25), B. Callegari and W. D. Belangero, Acta Ortopedica Brasileira 12,
254102–254103 (2009). 160–166 (2004).
94 122
S. C. Stanton, C. C. McGehee, and B. P. Mann, Physica D 239(10), R. Simoes, M. Rodriguez-Perez, J. De Saja, L. Marques, A. Kinoshita,
640–653 (2010). and C. Constantino, in 11th International Conference on Advanced
95
M. A. Karami and D. J. Inman, Appl. Phys. Lett. 100(4), 042901 Materials, Rio de Janerio (2009).
123
(2012). J. A. Potkay, U.S. patent application 12/762,242 (25 November 2010).

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 130.92.9.55
On: Sun, 28 Sep 2014 18:06:44