Please cite this article in press as: Yang et al.

, High-Performance Piezoelectric Energy Harvesters and Their Applications, Joule (2018), https://
doi.org/10.1016/j.joule.2018.03.011

Review
High-Performance Piezoelectric Energy
Harvesters and Their Applications
Zhengbao Yang,1,3,* Shengxi Zhou,2,* Jean Zu,3,* and Daniel Inman4,*

Context & Scale

Energy harvesting holds great potential to achieve long-lifespan self-powered
operations of wireless sensor networks, wearable devices, and medical im- With the rapid advances in
plants, and thus has attracted substantial interest from both academia and wireless sensors, implantable
industry. This paper presents a comprehensive review of piezoelectric energy- electronics, and wearable
harvesting techniques developed in the last decade. The piezoelectric effect devices, the demand for high-
has been widely adopted to convert mechanical energy to electricity, due to power-density and long-lifespan
its high energy conversion efficiency, ease of implementation, and miniaturiza- power sources is becoming
tion. From the viewpoint of applications, we are most concerned about whether increasingly stronger. Energy
an energy harvester can generate sufficient power under a variable excitation. harvesting, emerging as an
Therefore, here we concentrate on methodologies leading to high power alternative energy solution to
output and broad operational bandwidth. Different designs, nonlinear batteries, holds great potential to
methods, optimization techniques, and harvesting materials are reviewed and achieve self-powered
discussed in depth. Furthermore, we identify four promising applications: autonomous operations of such
shoes, pacemakers, tire pressure monitoring systems, and bridge and building low-power electronic devices, and
monitoring. We review new high-performance energy harvesters proposed for thus has recently attracted much
each application. attention from both academia and
industry. The piezoelectric effect
is widely adopted to convert
Introduction
mechanical energy to electrical
Mechanical energy harvesting is a process by which vibration, kinetic energy, or
energy, due to its high energy
deformation energy is converted to electrical energy. There are a variety of energy
conversion efficiency, ease of
sources available for energy harvesters, ranging from the human body to wild ani-
implementation, and
mals, from industrial machinery to vehicles, from large-scale buildings to bridges,
miniaturization. This paper
and from water flow to wind. Energy harvesters are regarded as promising indepen-
presents a comprehensive and
dent power sources for low-power electronic devices such as wireless sensors,
critical review of state-of-the-art
portable devices, and medical implants. Different from the most commonly used
research on piezoelectric energy
chemical batteries, energy harvesters function like power generators, endlessly har-
harvesting. From the viewpoint of
vesting energy from the surroundings. Thus issues associated with batteries such as
applications, we are most
limited lifespan, large size, environmental pollution, and high maintenance cost are
concerned about whether an
mitigated by harvesting solutions. Along with the rapid development of low-power
energy harvester can generate
integrated circuits (ICs) and high-efficiency energy storage solutions, the energy-
sufficient power under variable
harvesting technology is expected to ultimately give rise to an era of self-powered
excitation. Therefore, here we
autonomous operation in healthcare, automotive applications, and environmental
concentrate on methodologies
monitoring.
leading to high power output and
broad operational bandwidth. A
Energy harvesters have been intensively studied for the last two decades. A wide
variety of designs, nonlinear
range of investigations have been conducted to improve performance in a variety
methods, optimization
of aspects. There are a number of excellent reviews published, which focus on
techniques, and harvesting
materials,1–3 micro-electromechanical systems (MEMS)-scale devices,4–11 nonlinear
materials are reviewed and
broadband methods,12–14 circuits,15–17 and different conversion methods.18–24 The
discussed in depth. The study also
goal of this study is to provide an in-depth review on the state-of-the-art techniques
evaluates different figures of merit
that lead to high-performance energy harvesting via the piezoelectric effect. Instead
of discussing methods proposed to improve performance in different aspects

Joule 2, 1–56, April 18, 2018 ª 2018 Elsevier Inc. 1

 Please cite this article in press as: Yang et al., High-Performance Piezoelectric Energy Harvesters and Their Applications, Joule (2018), https://
doi.org/10.1016/j.joule.2018.03.011

broadly, we concentrate on the most essential properties to end users. From the
perspective of applications, power output and operational bandwidth are the two and presents a systematic
metrics most useful to product development engineers. The insufficient power performance comparison on
output and the sensitivity to environmental variations are the main barriers prevent- recently proposed energy
ing current energy harvesters from being widely adopted in engineering practice. To harvesters. Furthermore, we
improve power output and extend working bandwidth, researchers have explored identify four promising
different designs, nonlinear methods, optimization techniques, and materials, which applications: shoes, artificial
are included in this review. To further improve the performance of piezoelectric pacemakers, tire pressure
energy harvesters (PEHs), it is important to discuss them in specific application envi- monitoring systems, and bridge
ronments. Here we identified four most promising and widely studied applications: and building monitoring. The
shoes, artificial pacemakers, tire pressure monitoring systems, and building and excitation characteristics of each
bridge monitoring systems. For each application, we characterize the excitation application are analyzed and
sources, detail design methodologies, and comprehensively review application-ori- corresponding harvesting
ented designs proposed in the past 10 years. The hope of this review is to help guide methods discussed.
future research and promote future applications of PEHs.
The piezoelectric energy-
The article begins with a brief introduction of fundamental knowledge for piezoelec- harvesting technology has
tric energy harvesting. We then review nonlinear methods for extending the opera- experienced significant progress
tional bandwidth and approaches to obtain high power output, followed by a in the past 10 years. However,
comprehensive performance evaluation of recently proposed PEHs. Finally, we pre- research on energy harvesters is
sent a detailed review of high-performance energy harvesters for the four identified mostly conducted without specific
applications. applications, and reliability and
system integration have not been
Fundamentals of Piezoelectric Energy Harvesting well examined. More research is
The direct piezoelectric effect is the basis for PEHs. It is governed by the following expected to deal with these issues
constitutive equations, which connects the mechanical domain (stress T and to facilitate in turning decades of
strain S) with the electric domain (electric field E and charge density D): research efforts on energy
     E   harvesting into tangible benefits
Converse S s dt T in our daily life.
= = ; (Equation 2.1)
Direct D d ε T
E

where sE is the compliance under a constant electrical field, 3 T is the dielectric

permittivity under a constant stress, and d and dt are the matrices for direct and
converse piezoelectric effect, where the superscript t stands for the transpose.

Most energy harvesters operate in a frequency range much lower than the resonant
frequency of the employed piezoelectric elements. Thus the piezoelectric elements
can be approximately regarded as parallel plate capacitors. For a piezoelectric
element (surface area S and thickness t) loaded with a stress s, we can roughly calcu-
late the accumulated charge Q on the electrode, the voltage V over the element, and
the total converted electric energy U as
1Department of Mechanical and Biomedical
1 1
U = QV = ðd 3 s 3 SÞ  ðg 3 s 3 tÞ; Engineering, City University of Hong Kong,
2 2 Kowloon, Hong Kong, China
(Equation 2.2)
1
= d 3 g 3 s2 3 Volume
2School of Aeronautics, Northwestern
2 Polytechnical University, Xi’an, China
3Department of Mechanical and Industrial
where the current constant d and voltage constant g correspond to the specific co-
Engineering, University of Toronto, Toronto, ON,
efficients of the operational mode. The simplified Equation 2.2 indicates that a ma- Canada
terial or operational mode with a high d 3 g value will show a high power density 4Department of Aerospace Engineering,
when the piezoelectric material is directly stressed. University of Michigan, Ann Arbor, MI USA
*Correspondence: [email protected] (Z.Y.),
[email protected] (S.Z.),
Equation 2.2 can be used to evaluate different piezoelectric materials. Those [email protected] (J.Z.),
commonly used in energy harvesters include aluminum nitride (AlN), ZnO, BaTiO3, [email protected] (D.I.)
polyvinylidene fluoride (PVDF), PZT, PMN-PT (Pb[Mg1/3Nb2/3]O3-PbTiO3), PZN-PT https://doi.org/10.1016/j.joule.2018.03.011

2 Joule 2, 1–56, April 18, 2018

 Please cite this article in press as: Yang et al., High-Performance Piezoelectric Energy Harvesters and Their Applications, Joule (2018), https://
doi.org/10.1016/j.joule.2018.03.011

Table 1. Piezoelectric Materials and Their Properties

BaTiO3 PZT-4 PZT-5A PZT-5H PZT-8 PVDF PMN-33%PT PZN-6%PT
d31 (1012 C/N) 78 123 171 275 97 23 920 1,400
d33 149 289 374 593 225 33 2200 2,400
d15 496 584 741 330
g31 (103 Vm/N) 5 11.1 11.4 9.1 11 216 17.1 24.3
g33 14.1 26.1 24.8 19.7 25.4 330 44 41.7
g15 39.4 38.2 26.8 28.9
k33 0.48 0.7 0.71 0.75 0.64 0.15 0.93 0.9
Mechanical QM 300 500 75 65 1,000 3–10 69
Dielectric loss 0.4% 2% 0.4% 0.42%

Curie temperature ( C) 115 328 365 193 300 100 145 100

See Khan et al.,2 Caliò et al.,25 Uchino,26 Zhang et al.,27 and Yang and Zu.28

(Pb[Zn1/3Nb2/3]O3-PbTiO3), and various piezoelectric composites. Table 1 summa-

rizes the properties of these piezoelectric materials. AlN and ZnO have a much
weaker piezoelectric effect than the other listed materials. They are usually adop-
ted in thin film configurations in micro scale where figures of merit are different
from those of bulk materials. Properties of AlN and ZnO are thus not compared
here. As shown in Table 1, usually we have d15 >> d33 > d31. For PZT, d31 z
0.5d33. The single crystals PMN-PT and PZN-PT demonstrate the highest
piezoelectric properties, but are more sensitive to temperate change, more sus-
ceptible to fatigue, and more difficult to manufacture than lead zirconate titanate
(PZT). Therefore, PZT is still the most popular piezoelectric material in energy
harvesters.

PZT is a perovskite ceramic, manufactured via a complicated process. First, fine PZT
powders of the component metal oxides PbZr1xTixO3 are mixed at a specific pro-
portion (x). It has been demonstrated that the piezoelectric effect is maximized
when x z 0.5 (Yuhuan29). Around x = 0.5, the crystalline PZT is in the vicinity of
the morphotropic phase boundary between the ferroelectric monoclinic and rhom-
bohedral phases, and an increased number of domain structures coexist. The com-
pound powder is then heated and formed (dry or isostatic press, etc.) into the
desired shape. Following this, the elements experience a sinter firing process and
attain a dense crystalline structure. The crystalline elements are ground to specifica-
tions and covered with surface electrodes for the last process, polarization. PZT is a
polycrystalline material with polar domains of random orientations (Figure 1A); thus
it does not show a macroscopic piezoelectric effect until being polarized. In the po-
larization process, a strong DC electric field is provided through the electrodes at a
temperature a little below the Curie temperature. With the assistance of heat, the
randomly oriented polar domains can move more freely to be aligned with the
external electric field (Figure 1B). After cooling and the removal of the electric field,
the polar domains are locked into a pattern of near alignment (Figure 1C). As a
consequence, the material possesses permanent polarization and exhibit the
desired piezoelectric effect.

According to the directions of polarization and stress, energy harvesters can be

categorized into two main operation modes: d31 and d33. As illustrated in Figure 2,
in the d31 mode, the polarization direction ‘‘3,’’ i.e., the electric field, is perpendic-
ular to the direction of the applied stress ‘‘1.’’ This is the most commonly used
operation mode, and widely exists in bending-beam structures. In contrast, the

Joule 2, 1–56, April 18, 2018 3

 Please cite this article in press as: Yang et al., High-Performance Piezoelectric Energy Harvesters and Their Applications, Joule (2018), https://
doi.org/10.1016/j.joule.2018.03.011

Figure 1. Polarization of Polycrystalline Piezoelectric Ceramic

(A) Random orientation of polar domains prior to poling.
(B) Poling under a constant electric field.
(C) Remnant polarization after the removal of the electric field.

polarization and the applied stress are in the same direction in the d33 mode. The
d33 mode appears in direct compressive/tensile harvesters or harvesters with the
interdigitated electrode. In both the d31 and d33 modes, electrodes are made
perpendicular to the poling direction, i.e., the electric field is aligned with the po-
larization. Some energy harvesters also employ the shear mode, where the charge
constant d15 plays the role. Piezoelectric elements working in the shear mode need
a special treatment as the electrodes are formed on surfaces parallel to the poling
direction.

The stiff piezoceramics are not able to capture much mechanical energy if they are
directly attached to a vibration host. The classic solution is to attach a piezoelectric
element on a flexible beam structure. The cantilever, as shown in Figure 3, is the
most commonly used structure. The substructure is covered with one piezoelectric
element (unimorph) or two (bimorph) near the high-stress fixed end, and often
attached with a tip mass to lower resonant frequency and increase the vibration-
induced stress. The cantilever energy harvester is an inertial linear resonant energy
harvester and has been used widely in the literature as a benchmark for performance
comparison. It operates by first harnessing vibration energy from ambient excitation
sources such as building oscillations, human motions, and vehicle vibrations, there-
after converting the harnessed mechanical energy to electrical energy for powering
portable or wireless sensor modules.

Broadband Nonlinear Energy Harvesting

Governing Equations
Considering that linear resonant harvesters are incapable of harvesting energy
from broadband or frequency-varying excitations, researchers have developed
some frequency tuning approaches, such as multi-cantilever structures, bistable
composite plate designs, and passive and active stiffness-tuning technolo-
gies.12–14,30,31

From the viewpoint of stochastic resonance,32 Gammaitoni et al. brought the bi-
stable mechanism to the vibration energy-harvesting field, and numerically and
experimentally demonstrated the unique advantage of bistable energy harvesters
(BEHs) under random excitation.33 The nonlinear magnetic interaction was used to
induce nonlinear responses of an inverted pendulum beam. Under harmonic exci-
tations, Erturk et al.34 presented a typical Duffing system, consisting of a magnetic
beam attached with piezoelectric patches and two fixed external magnets as
shown in Figure 4A. A set of dimensionless electromechanical equations were

4 Joule 2, 1–56, April 18, 2018

 Please cite this article in press as: Yang et al., High-Performance Piezoelectric Energy Harvesters and Their Applications, Joule (2018), https://
doi.org/10.1016/j.joule.2018.03.011

Figure 2. Illustration of the d31 and d33 Operation Modes

employed to descript the dynamic characteristics. The corresponding governing

equations are:

1

x€ + 2zx_  x 1  x 2  cv = f cos Ut;
2 (Equation 3.1)
v_ + lv + kx_ = 0
where v is the dimensionless voltage across the load resistance, c is the dimension-
less piezoelectric coupling term in the mechanical equation, k is the dimensionless
piezoelectric coupling term in the electrical circuit equation, and l is the reciprocal
of the dimensionless time constant (lN1=Rl Cp , where Rl is the load resistance and Cp
is the equivalent capacitance of the piezoelectric layers). All the parameters used in
this model are dimensionless and not completely related to the real experiments.
Nevertheless, this model can be used to qualitatively depict and analyze the
nonlinear dynamic responses of bistable harvesters.

Zhou et al.35 designed an alternative nonlinear broadband energy harvester with

rotatable magnets to further extend frequency bandwidth by altering the magnet
inclination angle (Figure 4B). This system consists of an elastic substrate, tip magnet
attachments, external magnets, and two piezoelectric layers at the root. The ex-
hibited nonlinear dynamic behaviors vary depending on the inclination angle a of
the external magnets relative to the horizontal (Figure 4B) over a certain frequency
range. Additionally, the configuration potentially saves space since nonlinear fre-
quency response can be altered by simply rotating the magnets rather than chang-
ing the magnet spacing. A quantitative electromechanical model was derived based
on the Rayleigh-Ritz method and Euler-Bernoulli beam theory. The corresponding
governing equations are:

Mx€ + C x_ + Kx  qv = F + Fm
v (Equation 3.2)
Cp v_ + + qx_ = 0;
Rl
where M is the equivalent mass, K is the equivalent stiffness, C is the equiv-
alent damping, q is the equivalent electromechanical coupling term, x(t) is the
tip displacement of the harvester in the transverse direction, v(t) is the voltage
across the electrical load, and F(t) is the external mechanical force as the excitation
term.

Fm is the magnetic force fit to a polynomial of the form

Fm = mx + lx 3 ; (Equation 3.3)
which is identified by using a sensitive dynamometer in experiments. The
coefficients of the magnetic force polynomial denoted by m and l depend on the
angle a.

Joule 2, 1–56, April 18, 2018 5

 Please cite this article in press as: Yang et al., High-Performance Piezoelectric Energy Harvesters and Their Applications, Joule (2018), https://
doi.org/10.1016/j.joule.2018.03.011

Figure 3. The Typical Energy Harvesting Operation Process

The most commonly used cantilever configuration is shown here as an example.

For modeling the magnetic force, Stanton et al.36 proposed an analytical method for
the bistable energy harvester shown in Figure 5. The magnetic field generated by a
magnet A upon another magnetic B was given by
m0 mA $rAB
BAB =  V ; (Equation 3.4)
4p krAB k32

where k$k2 and V denote the Euclidean norm and vector gradient operators, respec-
tively. m0 and mA are the permeability of the free space and the magnetic dipole
moment vector, respectively. rAB is the effective distance from the magnet A upon
the magnetic B. The potential energy in the field is expressed as Um =  BAB $mB .
The magnetic force can be obtained by making a differential on Um.

Energy-Harvesting Performance and Analysis

Under variable harmonic excitations, nonlinear energy harvesters have been proven
to be able to exhibit a broadband performance. Researchers have successively
explored monostable, bistable, and tristable systems. Sebald et al.37 experimentally
verified that their nonlinear monostable harvester can work in the frequency range of
26–36 Hz under the excitation level of 3.2 m s2, whose bandwidth is much wider
than its linear counterpart. Tang et al.38 designed a monostable harvester with a
magnetic oscillator that works well under weak excitations. Compared with the
equivalent linear harvester, the monostable harvester obtained a 41% increase in
the magnitude of the power at an excitation level of 2 m s2, and a 100% increase
in the operating bandwidth. Yang et al.39,40 designed a high-efficiency compres-
sive-mode monostable harvester that produced a maximum power output of

6 Joule 2, 1–56, April 18, 2018

 Please cite this article in press as: Yang et al., High-Performance Piezoelectric Energy Harvesters and Their Applications, Joule (2018), https://
doi.org/10.1016/j.joule.2018.03.011

Figure 4. Typical Magnetic-Based Nonlinear Energy Harvesting Structures

(A) Duffing bistable energy harvester. Reproduced from Erturk et al. 34 with permission from AIP
Publishing LLC.
(B) Rotatable nonlinear energy harvester. Reproduced from Zhou et al. 35 with permission from AIP
Publishing LLC.

54.7 mW and an operational bandwidth of 12 Hz, under an excitation level of

4.9 m s2. Stanton et al.41 achieved a novel bidirectional hysteretic response of a
monostable cantilever harvester by tuning a set of coupled magnets. Both hardening
and softening responses were exhibited in the simulation and invoked in experi-
ments. Yang et al.42 proposed a new method to tune the nonlinear responses of
monostable harvesters by tilting the structures. The capability of shifting the dynamic
responses between hardening nonlinearity and softening nonlinearity gives energy
harvesters a great environmental adaptability. Erturk et al.34,43 explored the high-en-
ergy orbits in a Duffing BEH over a range of excitation frequencies, and numerically
and experimentally verified the energy-harvesting enhancement of the BEH. Zhou
et al.35 obtained high-energy output voltage in a wide frequency range of 4–22 Hz
by changing the magnet inclination angle shown in Figure 4B. Later, based on the
same nonlinear energy-harvesting device, Zhou et al.44,45 and Kim et al.46,47 pro-
posed a tristable energy-harvesting concept. The theoretical and experimental re-
sults demonstrated that the tristable energy harvester with shallow potential wells
can more easily undergo high-energy inter-well oscillations under low-level excita-
tions than its bistable counterparts because of their higher potential.

To analyze nonlinear systems, researchers have proposed a variety of theoretical

models and solutions. Both the harmonic balance method48 and the multiple scale
technique42 have been used to solve the governing equations of monostable sys-
tems. Stanton et al.49 deduced the fundamental harmonic balance solutions of the
intra-well and inter-well oscillations for BEHs. Based on this research, Harne and
Wang50 obtained the harmonic balance solutions with both fundamental and third
harmonic terms of BEHs. Meanwhile, they observed the third harmonic response
in experiments, which is in accordance with the earlier work provided by Massna
and Daqaq.51 For the tristable systems that have more nonlinear terms and complex
nonlinear dynamics, Zhou et al.52 and Panyam et al.53 successively employed the
harmonic balance method and the perturbation method to obtain the theoretical so-
lutions. They found that tristable energy harvesters have multi-solution frequency
ranges under harmonic excitations. The chaotic response characteristics of nonlinear
energy harvesters were explored by Cao et al.54 via the phase trajectory, power

Joule 2, 1–56, April 18, 2018 7

 Please cite this article in press as: Yang et al., High-Performance Piezoelectric Energy Harvesters and Their Applications, Joule (2018), https://
doi.org/10.1016/j.joule.2018.03.011

Figure 5. Repulsive Magnetic Force-Based Bistable Energy Harvester

(A) Experimental setup.
(B) Detailed view of the geometry used to outline vector parameters of the analytical magnet
model. Reproduced from Stanton et al. 36 with permission from Elsevier.

spectrum, Poincare map, and bifurcation diagram. The predictive criteria to deter-
mine the harmonic excitation conditions for the presence of chaotic response
were investigated via Melnikov theoretical methods.55,56

In addition to investigating ideal harmonic excitations, random excitations have also

been systematically investigated. Litak et al.57,58 numerically simulated the re-
sponses of the bistable harvester proposed by Erturk et al.34 under random base
excitations, and observed the stochastic resonance phenomenon. It is claimed
that if the variance of the excitation were known, the harvester could be optimized
to maximize the power harvested under random base excitations. Cottone et al.59
numerically and experimentally verified that a buckled-beam-based bistable
harvester produces more power under random base excitation compared with its
linear counterpart. Daqaq60,61 explored the Fokker-Plank-Kolmogorov equation to
describe the responses of bistable harvesters and approximately solved this equa-
tion by using the finite element method. The work of Zhao and Erturk62 indicates
that a bistable harvester can have a better energy-harvesting performance than
linear harvesters only if it was optimally designed to work in the neighborhood of
a specific random excitation intensity. For the much lower or higher excitation inten-
sity levels, linear harvesters actually generate larger power output near their reso-
nant frequencies. Under random base excitations, Li et al.63 found that tristable har-
vesters with shallower potential wells generate more power than bistable harvesters,
which is in agreement with the conclusion for harmonic base excitation cases.44–47

There are some new structures proposed for the design of nonlinear energy
harvesters. Different from the aforementioned nonlinear energy harvesters that
can only harness vibration energy from one direction, Su et al.64,65 developed a
two degrees of freedom (2DOF) broadband magnet-induced dual-cantilever PEH.
As shown in Figure 6A, it consists of outer and inner beams and magnets attached
to the tips. The magnets induce a nonlinear repulsive force between the two beams,
and make the structure bistable. In experiments, it exhibits broadband characteris-
tics and has two voltage peaks in the excitation frequency range under some excita-
tion levels. Arrieta et al.66,68 proposed a bistable composite plate for energy har-
vesting, as shown in Figure 6B. This bistable plate was obtained based on a
special heat treatment technology and is capable of bearing high-level excitations.
Note that a high level excitation is necessary to make high-energy inter-well oscilla-
tions in this design. Betts et al.69,70 presented an optimization strategy of bistable
composites, which enables improved power output by discovering the optimal con-
figurations based on the statics of the device. The optimal device aspect ratio,

8 Joule 2, 1–56, April 18, 2018

 Please cite this article in press as: Yang et al., High-Performance Piezoelectric Energy Harvesters and Their Applications, Joule (2018), https://
doi.org/10.1016/j.joule.2018.03.011

Figure 6. New Nonlinear Energy Harvesters

(A) 2DOF bistable harvester. Reproduced from Su et al. 64 with permission from SAGE Publications.
(B) Bistable composite plate. Reproduced from Arrieta et al. 66 with permission from AIP
Publishing LLC.
(C) M-shaped nonlinear energy harvester. Reproduced from Leadenham et al. 67 with permission
from Elsevier.

stacking sequence, thickness, and piezoelectric area were considered in their

research. Leadenham and Erturk67 designed an M-shaped asymmetric nonlinear en-
ergy harvester for broadband vibration energy harvesting. As shown in Figure 6C, a
pre-bent monostable beam structure, the M-shaped configuration may exhibit both
limited maximum deflections to protect piezoelectric elements and favorable
nonlinear dynamic response at low excitation levels, thanks to the large stretching
capabilities. Compared with the corresponding linear system, the M-shaped
nonlinear energy harvester offers a maximal 8,900% bandwidth enhancement at
the root-mean-square base excitation levels of 0.07 g.

Energy-Harvesting Enhancement Strategies

Nonlinearity has been widely recognized as an effective method to extend the work-
ing bandwidth for energy harvesters. However, one should note that such nonlinear
energy harvesters have multiple vibration orbits and a region of hysteresis under
most excitation conditions. Furthermore, their expected performance enhancement
can be realized only when system vibrates at the highest orbit. It is difficult to main-
tain nonlinear harvesters in the high-energy oscillation states, especially under weak
excitations. With zero initial conditions, nonlinear harvesters usually follow the low-
energy orbits, resulting in small-amplitude voltage responses. To overcome this
issue, Zhou et al.71 presented an initial impact method for facilitating nonlinear en-
ergy harvesters to overcome the potential barriers and achieve high-energy inter-
well oscillations under low-level excitations. Figure 7A shows the schematic of a
mechanism for imparting additional energy based on the transfer of momentum.
In principle, an emitter and receiver could be used to send and receive a projectile
to impact the free end of the harvester. The harvester will then obtain a new velocity
(or kinetic energy) that is appropriate to respond in high-energy inter-well oscilla-
tions. Via this method, the effective bandwidths of bistable and tristable harvesters
were increased to 15 Hz and 12 Hz from 3 Hz to 5 Hz in experiments, respectively.

Later on, Mallicket et al.72 introduced an electronic control method to switch from
the low- to the high-energy orbit of a nonlinear electromagnetic energy harvester
by employing the strong interaction between its electrical and mechanical degrees
of freedom. In principle, this method can be applied to nonlinear energy harvesters
of different sizes and transduction methods after some modifications. As shown in
Figure 7B, a voltage gx_ is induced in the coil (internal resistance RC) of the nonlinear
energy harvester, where x_ and g are respectively the velocity of the mechanical oscil-
lator and the electromagnetic coupling coefficient. The electric power is extracted
through the load resistance RL that is connected in series with the coil. In addition,
a voltage source VA (with internal resistance RA) is connected across the coil, which,

Joule 2, 1–56, April 18, 2018 9

 Please cite this article in press as: Yang et al., High-Performance Piezoelectric Energy Harvesters and Their Applications, Joule (2018), https://
doi.org/10.1016/j.joule.2018.03.011

Figure 7. High-Energy Orbit Implement Strategies

(A) Impact-induced method. Reproduced from Zhou et al. 71 with permission from AIP
Publishing LLC.
(B) Electrical actuation method. Reproduced from Mallick et al. 72 with permission from The
American Physical Society.
(C) Voltage impulse perturbation method. Reproduced from Lan et al. 73 with permission from EDP
Sciences.

over a short period, forms electric voltage to perturb the system to switch the state.
This method was claimed to be able to improve the energy conversion efficiency
over a broad bandwidth under frequency-amplitude-varying excitations using a
small energy budget.

Recently, to trigger high-energy-orbit responses of nonlinear PEHs, Lan et al.73 pre-

sented a voltage impulse perturbation approach using negative resistance. The
detailed circuit design is shown in Figure 7C. A negative resistance can be designed
by tuning resistors R1, R2, and R3 in the circuit. When the circuit works as a negative
resistance, current flows into the piezoelectric element, providing an external exci-
tation to the system. Meanwhile, a response voltage across on the resistances is pro-
duced, which influences the mechanical oscillation of energy harvesters. Via this
method, high-energy orbit oscillations of both monostable and BEHs were experi-
mentally induced.

Hosseinloo et al.74 proposed a robust and adaptive sliding mode controller to shift
the nonlinear energy harvester to the desired attractor by a short entrainment on the
desired attractor. Their numerical results show that this method can successfully
move the harvester to the desired attractor in a reasonable period of time, but no
experiments were conducted.

Summary
The research discussed above indicates that nonlinearity is an effective way to
improve the energy-harvesting performance. Nonlinear energy harvesters have
theoretically and experimentally demonstrated broader operating bandwidths
than their linear counterparts. However, one should note that nonlinear energy
harvesters have multiple vibration orbits under most excitation conditions, and the
expected performance enhancement can be realized only when nonlinear energy
harvesters vibrate at the highest orbit. It is difficult to maintain nonlinear harvesters
in the high-energy oscillation states, especially under weak excitations. With zero
initial conditions, nonlinear harvesters usually follow the low-energy orbits, resulting
in small-amplitude voltage responses.

How to maintain nonlinear energy harvesters in the high-energy states is a critical

problem. Using active control is a feasible way to tune nonlinear systems, but con-
sumes a certain amount of energy. More research is expected on the active control
technique to improve its robustness under varying conditions, and make the energy

10 Joule 2, 1–56, April 18, 2018

 Please cite this article in press as: Yang et al., High-Performance Piezoelectric Energy Harvesters and Their Applications, Joule (2018), https://
doi.org/10.1016/j.joule.2018.03.011

Figure 8. Cantilever Energy Harvesters of Different Beam Shape

(A) Triangular cantilever beam. Reprinted with permission from Dietl et al. 77 Copyright 2010, SAGE Publications.
(B) Tapered and reverse tapered cantilevers. Reproduced from Goldschmidtboeing et al. 78 with permission from IOP Publishing Ltd.
(C) Beam with a width of a quadratic profile. Reproduced from Ben Ayed et al. 79 with permission from SAGE Publications.
(D) Optimized beam shape with a fixed length. Reproduced from Park et al. 80 with permission from KSME & Springer.

consumed trivial compared with the energy harvested. Passive control methods are
worthy of investigation, although their adaptability may not be as good as that of the
active methods. Another critical problem that nonlinear energy harvesters are facing
is how to efficiently transfer and store the generated broadband or random electric
energy. Currently, most research mainly focuses on the broadband dynamic charac-
teristics of nonlinear systems under the open-circuit condition or simply with resistive
loads. More work is expected on the development and optimization of conditioning
interface circuits for nonlinear energy harvesters.

High Power Output

Researchers have explored a variety of methods to improve the performance of en-
ergy harvesters from the perspective of structures, materials, and circuits. It is impos-
sible to include all of the methods and designs in one review. Here we mainly analyze
new structures and materials for increasing power output. Readers who are inter-
ested in energy-harvesting circuits can refer to the review papers.15,16 The following
analyzes these new methods and designs in seven categories: geometry-modified
cantilever; topology optimization; curved structures; electrode optimization; d15
mode under shear/torsion excitations; new structures operating under tension/
compression excitations; and new piezoelectric materials and composites.

Geometry-Modified Cantilever
Classic PEHs employ rectangular cantilever beams where piezoelectric elements
experience a bending stress that varies linearly along the length direction. The
bending stress is maximized at the fixed end and tends to be zero at the free tip.
As a consequence, issues such as overstrain and fatigue may occur around the fixed
end. In contrast, the piezoelectric material around the free tip does not play its full
role. To obtain a uniform and large strain distribution on the employed piezoelectric
materials in cantilever PEHs, researchers have modified usual rectangular shape to
consider different geometries.

Tapered cantilevers are well-known optimized shapes. As early as 2005, Roundy

et al.75,76 proposed the idea of replacing rectangular cantilevers with the trapezoidal
file to improve power output. Experiments indicated that beams with a nonuniform
width have a 30% gain in power output over rectangular counterparts. Diel and Car-
cia77 examined a family of beams ranging from rectangular beams to triangular
beams theoretically and experimentally (Figure 8A). The study indicates that
the beam shape has little effect on efficiency, but significantly affects the
maximum tolerable excitation strength and the maximum achievable output power.
Goldschmidtboeing and Woias78 presented a theoretical model based on the

Joule 2, 1–56, April 18, 2018 11

 Please cite this article in press as: Yang et al., High-Performance Piezoelectric Energy Harvesters and Their Applications, Joule (2018), https://
doi.org/10.1016/j.joule.2018.03.011

Rayleigh-Ritz method for width-varying cantilever energy harvesters, and optimized

three beam shapes: rectangle, linear taper, and reverse taper (Figure 8B). Further-
more, Ayed et al.79 studied two types of width profiles for cantilever PEHs as a linear
and quadratic function of the position (Figure 8C). They modeled a unimorph canti-
lever with a tip mass using the one-mode Galerkin method, and estimated that the
quadratic geometry is capable of yielding nearly 100% more energy than a rectan-
gular shape.

Besides focusing on different geometries of energy harvesters themselves, re-

searchers have also implemented performance optimization from the system level.
Park et al.80 considered the limited application space of an energy harvester, and
set a fixed length cantilevers with a variable width to maximize the output power.
A finite element model was built using ANSYS to find the highest-power-output
shape for cantilever beams subject to tip impact excitations. It identified an optimal
design with a unique wide tip as shown in Figure 8D. Experiments demonstrated a
37% improvement on the power output compared with the rectangular counterpart
using the same amount of material.

Thickness optimization was undertaken theoretically by Paquin and Amant,81 who

proposed a semi-analytical model based on the Rayleigh-Ritz approximations. The
study estimated that a thickness-tapered beam with a slope angle of 0.94 can
increase power output of energy harvesters by a factor of 3.6. This performance
enhancement can be explained by the more uniform strain distribution formed
by the tapered thickness. They did not provide experimental validation. This
method is plausible theoretically but may not be feasible due to the difficulty in
manufacturing thin beams with such a small slope angle.

Topology Optimization
The aforementioned approaches consider only a small number of geometric vari-
ables of well-defined shapes such as rectangles and trapezoids. General geometries
have also been studied via a variety of topology optimization strategies. Here, we
review some representative topology optimization investigations in energy har-
vesters. Note that all of the topology optimization results are presented without
experimental data.

Rupp et al.82 applied the topology optimization method to optimize the circuit pa-
rameters and the layout of multilayer structures consisting of piezoelectric, elec-
trode, and structural layers. The electromechanical coupling was modeled using
the finite element method and the circuit was simulated as a lumped-parameter sys-
tem. Figure 9A shows the studied object and the optimized design under a harmonic
excitation. The optimal layout is dependent upon the substrate thickness ratio. It is
obvious that the optimal shape identified by the topology optimization is different
from regular geometries such as rectangles and trapezoids, which might pose
some problems in the manufacturing process.

Akihiro et al.86 did a simultaneous topology optimization on the polarization direc-

tion and the layout of the piezoelectric material in a piezoelectric patch. The solid
isotropic material with the penalization (SIMP) method was used to model the piezo-
electric material and the derived equations were solved by the finite element
method. The situation considered here is a piece of piezoelectric patch attached
on a vibration surface in high-order vibration modes. The complicated vibration
modes lead to strains of opposite signs in the piezoelectric elements and corre-
spondingly charge cancellations if a uniform electrode is used. The authors

12 Joule 2, 1–56, April 18, 2018

 Please cite this article in press as: Yang et al., High-Performance Piezoelectric Energy Harvesters and Their Applications, Joule (2018), https://
doi.org/10.1016/j.joule.2018.03.011

Figure 9. Topology Optimization of Energy Harvester Structures

(A) Multilayer structure with a circuit. Optimized distribution of piezoelectric material (red) on an aluminum substrate (blue). Reproduced from Rupp
et al. 82 with permission from SAGE Publications.
(B) Cylindrical energy harvester. The red (darker) to metal and the green (lighter) colors refer to the piezoelectric material. Reproduced from Rupp
et al. 83 with permission from Elsevier.
(C) The optimized layout of the piezoelectric layer. The substrate layer remains unchanged (white). Reproduced from Nanthakumar et al.84 with
permission from Elsevier.
(D) The optimal stress constrained design of a bimorph cantilever with a tip mass. Black represents material and light-gray represents void. The color
scale shows the Von Mises stress in the piezoelectric layers. Reproduced from Wein et al. 85 with permission from Springer-Verlag.

proposed to alter the poling direction to avoid charge cancellation. The method is
almost impossible to be implemented due to the problem of polarization and the un-
predictable excitation sources.

Chen et al.83 used a level set topology optimization method to find the optimal
geometry for maximizing the energy conversion efficiency. The employed level
set-based method is argued to be better than other optimization methods used in
energy harvesters such as the SIMP method and its variants.82,87,88 The SIMP method
cannot converge to a solution with distinct phase states for such a complex multi-
physics problem. In contrast, the level set method allows initially defined voids to
merge and change shapes, resulting in a modified topology of the target structure.
Figure 9B shows the optimized cylindrical PEH. The optimal design is with a volume
ratio for the piezoelectric material of 51.9% and an energy conversion efficiency
of 8.9%.

Most topology optimizations are conducted on bulk materials. In 2016, Nanthaku-

mar et al.84 did a topology optimization on nanoscale energy harvesters, consid-
ering the unique surface elasticity and surface piezoelectricity existing in ZnO nano-
structures. They developed an extended finite element model based on the Kirchoff
plate model, coupled with the level set topology optimization method to maximize
the electromechanical coupling coefficient. Figure 9C shows the optimized layout of
the piezoelectric material. The study also analyzed the competition between the sur-
face elasticity and the surface piezoelectricity.

Wein et al.85 performed a topology optimization on the elastic substructure and tip
mass of a cantilever bimorph PEH by means of the SIMP method. In the study, the

Joule 2, 1–56, April 18, 2018 13

 Please cite this article in press as: Yang et al., High-Performance Piezoelectric Energy Harvesters and Their Applications, Joule (2018), https://
doi.org/10.1016/j.joule.2018.03.011

Figure 10. Schematic of Cantliever Energy

Harvesters with Arc-Shaped Piezoelectric
Elements
Reproduced from Yang et al. 90 with permission
from Elsevier.

piezoelectric layers were kept constant and the geometries of the elastic beam and
the tip mass were optimized (Figure 9D). Jia and Seshia89 optimized the end mass
size compared with the cantilever in micro cantilever energy harvesters, and found
that the ideal proportion of the end mass is about 60%–70% of the total cantilever
length.

Curved Structures
Although a variety of structures have been exploited in energy harvesters, the core
energy transducing parts are most of time plain piezoelectric plates, discs, or stacks.
In 2015, Jung developed a curved energy harvester using flexible PVDF films and
succeeded in lighting up 476 LEDs under periodic palm pressing. In 2017, Yang
et al.90 introduced curved piezoceramic elements into energy harvesters. Half-
tube PZT elements were employed in the cantilever architecture as shown in Fig-
ure 10. Compared with plain plates, higher and more evenly distributed stress can
be formed in the curved elements. The proposed curved PZT harvester in experi-
ments exhibited a >200% higher power density and a 60% enhancement in the
energy conversion efficiency. It was asserted that more arc-shaped piezoceramic
structures can be explored to improve harvester performance.

Electrode Optimization
Electrodes have been optimized to improve power output. Stewart et al.91 realized
that in a traditional cantilever energy harvester, charge flows from the highly stressed
root area of the cantilever to the unstrained tip area, and energy is lost in this charge
redistribution process. To validate the idea, a prototype was made with 30 piezo-
electric elements on an FR4 PCB board as shown in Figure 11A. When only 20 of
the 30 piezoelectric elements were connected, the prototype showed the highest
power output, which is 15.5% higher than when all elements were connected. Erturk
et al.93 did a dimensionless analysis on the strain nodes of the vibration modes of the
cantilever energy harvester other than the fundamental mode, and pointed out the
optimal segmented electrode configurations to avoid charge cancellation. Cho
et al.94 studied the effect of electrodes on a membrane energy harvester, and
claimed that an electrode coverage of 60% demonstrated the best electromechan-
ical coupling. Yang and Zu95 proposed to mitigate the charge redistribution effect in
flextensional transducers by partially covering the piezoelectric element with elec-
trodes in the middle effective region.

In 2015, Kim et al.96 theoretically analyzed and optimized the interdigitated elec-
trodes (IDEs) in cantilever energy harvesters. The size of the electrodes, the number
of IDE fingers, and the piezoelectric thickness have significant effects on the
harvester performance. It was concluded that compared with the general parallel
plate electrodes, the IDEs cannot generate more power but show higher voltage
responses.

14 Joule 2, 1–56, April 18, 2018

 Please cite this article in press as: Yang et al., High-Performance Piezoelectric Energy Harvesters and Their Applications, Joule (2018), https://
doi.org/10.1016/j.joule.2018.03.011

Figure 11. Cantilever Energy Harvester with Individual Piezoelectric Elements

(A) Harvester made by Stewart et al. Reproduced from Stewart et al. 91 with permission from AIP
Publishing LLC.
(B) Harvester made by Du et al. Reproduced from Du et al. 92 with permission from Elsevier.

Later in 2017, Du et al.92 analyzed the relationship between the electrode coverage
and the power output of beam-based energy harvesters. They reached a conclusion
that the maximum output power of an energy harvester occurs when the electrode
covers the piezoelectric layer from the maximum strain area to a position where the
strain is a half of the average strain in all the previously covered area. PEH prototypes
were fabricated with segmented electrodes as shown in Figure 11B. It was claimed that
with the optimal electrode coverage the output power can be respectively improved
by 126% and 145% for a clamped-clamped beam and a cantilever PEHs.

d15 Mode under Shear/Torsion Excitations

From the perspective of active materials, cantilever or other bending-type energy
harvesters do not work in the optimal configuration. This is because the piezoelectric
coefficient d31 is relatively smaller than d33 and d15. For most piezoelectric materials,
the order is d15 > d33 > d31. Therefore, PEHs using the large d15 coefficient have the
potential to attain a greater efficiency. Zhao et al.97 fabricated a cantilever PEH with
two d15-mode PZT plates (13 3 2.5 3 1 mm3) connected in series (Figure 12A). The
cantilever prototype generated a maximum power of 8.7 mW under an excitation of
14.5 m s2. Bo et al.100 induced shear deformation in a PMN-PT wafer on a cantilever
beam. They considered the ratio of the lateral displacement to the length of the
wafer, and neglected the effect of the shear force in the shear strain calculation pro-
cess. The cantilever PEH with a 13 3 6 3 1 mm3 PMN-PT element showed a
maximum power of 4.16 mW under a cyclic force of 0.05 N.

Joule 2, 1–56, April 18, 2018 15

 Please cite this article in press as: Yang et al., High-Performance Piezoelectric Energy Harvesters and Their Applications, Joule (2018), https://
doi.org/10.1016/j.joule.2018.03.011

Figure 12. d15 Mode PEHs

(A) Cantilever PEH with two shear-mode piezoelectric elements. Reproduced from Zhao et al. 97 with permission from IOP Publishing Ltd.
(B) Vibrating sandwich beam with a piezoelectric core. Reproduced from Malakooti et al. 98 with permission from IOP Publishing Ltd.
(C) Shear-mode torsion energy harvester. Reproduced from Kulkarni et al.99 with permission from IEEE.

In terms of modeling, Aladwaniet et al.101 developed a single degree of freedom

(SDOF) model to analyze the shear-mode operation of piezoelectric materials. Ma-
lakooti and Sodano98 developed a distributed-parameter model for a shear-mode
cantilever energy harvester via the Timoshenko beam theory. As shown in Fig-
ure 12B, the desired shear deformation is formed in a vibrating sandwich beam
with a piezoelectric core. The piezoelectric element is polarized in the axial direction
and the induced electric field is perpendicular to the polarization in the thickness di-
rection. The developed model estimated that the shear mode harvester subject to a
harmonic base excitation is capable of generating 50% more power compared
with the equivalent cantilever bimorph harvester.

In addition to the bending excitations, torsional excitations have also been explored
in d15 energy harvesters. Kulkarni et al.99 presented a harvester (Figure 12C) consist-
ing of a thin-walled tube with piezoelectric patches and an eccentric mass. The
device was excited by vertical vibrations, and the eccentricity of the tip mass caused
torsional stress in the tube. A prototype made of four PZT-5A shear-mode plates (5 3
5 3 0.5 mm3) generated a maximum root-mean-square (RMS) power of 0.57 mW un-
der a harmonic excitation of 1 g at 620 Hz.

One should note that shear-mode PEHs face severe fabrication problems in spite of
the high piezoelectric coefficient d15. Complex polarization and electrode process
are needed to fabricate d15-mode piezoelectric elements.

New Structures Operating under Tension/Compression Excitations

Apart from cantilever and other bending-beam configurations, a variety of new
structures have been developed to improve harvester performance based on the
higher piezoelectric coefficient d33 and/or amplification effects.

Xu et al.102 presented a cantilever beam and cymbal transducer combined structure

as shown in Figure 13A. Cymbal transducers have exhibited excellent performance
in ultrasonic, actuation, and energy-harvesting applications,105 but possess high
resonant frequencies. This study combined the flexibility of cantilever beams and
the force amplification of the cymbal structures in order to obtain high power output
at a low frequency. A prototype, constructed with 25 3 5 3 1 PMN-PT elements, a
35-mm copper cantilever, and a 4.2-g tip mass, generated a maximum power of
3.7 mW under a harmonic excitation of 3.2 g at 102 Hz. A similar structure was
explored by Tufekcioglu and Dogan106 under the same harmonic base excitations,
and by Lee et al.107 to harvest energy from flow-induced vibration.

In 2017, Zou et al.103 explored a combined structure of cantilever and flextensional

transducers, and introduced a nonlinear magnetic force into the system. As shown in

16 Joule 2, 1–56, April 18, 2018

 Please cite this article in press as: Yang et al., High-Performance Piezoelectric Energy Harvesters and Their Applications, Joule (2018), https://
doi.org/10.1016/j.joule.2018.03.011

Figure 13. New PEHs Working in the Tension or Compression Mode

(A) A cantilever driving low-frequency PEH. Reproduced from Xu et al. 102 with permission from AIP Publishing LLC.
(B) A nonlinear compressive-mode PEH with magnets. Reproduced from Zou et al. 103 with permission from AIP Publishing LLC.
(C) Barbell-shaped high-temperature PEH. Reproduced from Wu et al.104 with permission from AIP Publishing LLC.
(D) High-performance compressive-mode piezoelectric energy harvester (HC-PEH). Reproduced from Yang et al. 39 with permission from Elsevier.

Figure 13B, the proposed harvester consists of a cantilever beam coupled with
flextensional transducers using magnets. The working bandwidth is significantly
extended by the introduced nonlinear magnetic interaction. A prototype (40 3
10 3 1 mm3 PZT-5H) generated a maximum power of 387 mW under a harmonic exci-
tation of 0.4 g at 9.9 Hz.

Wu et al.104 developed a barbell-shaped PEH with the high-temperature piezocer-

amic BiScO3-PbTiO3. Figure 13C shows the barbell-shaped structure that also em-
ploys the cantilever configuration. The difference is that the piezoelectric element
herein does not experience a bending stress (d31 mode) as in the traditional canti-
lever energy harvesters. In contrast, the piezoelectric stack works in the d33 mode,
subject to the alternating normal force from the vibrating shaft. The structure was
proposed to overcome the failure of the epoxy bounding layer and the piezoelec-
tricity decay of conventional cantilever PEHs in high-temperature and large ampli-
tude-vibration circumstances. A prototype was made of a ring piezo stack (inner
diameter 8 mm; outer diameter 21 mm; thickness 20 mm), a 125-mm long
steel shaft, and a tip mass of 100 g. The experimental data demonstrated that
under a harmonic excitation with a peak-peak amplitude of 1 g, the prototype
generated a maximum power of 4.76 mW at room temperature, and the power
output was doubled in a high temperature range (150 C–250 C). It was claimed
that the structure could sustain large impacts and operate steadily in 100,000 vibra-
tion cycles.

Joule 2, 1–56, April 18, 2018 17

 Please cite this article in press as: Yang et al., High-Performance Piezoelectric Energy Harvesters and Their Applications, Joule (2018), https://
doi.org/10.1016/j.joule.2018.03.011

Figure 14. PEHs under Direct Force Excitations

(A) Stack-based flex-compressive PEH. Reproduced from Wang et al. 113 with permission from IOP Publishing Ltd.
(B) Bending and compressive hybrid PEH. Reproduced from Xu et al.114 with permission from AIP Publishing LLC.

Yang et al.39,40 developed a high-performance compressive-mode energy

harvester by virtue of the low resonant frequency of doubly clamped beams and
the amplification effect of the flexural motion. As Figure 13D shows, the proposed
harvester consists of a pair of elastic beams, two mass blocks, two bow-shaped
plates and a center piezoelectric plate. While working, the base vibration is first
transferred through the elastic beams and absorbed by the mass blocks at the cen-
ter. The vertical vibration of the center then induces tensile stress along the elastic
beams. Following this, the beam motion causes the flexural deformation of the
bow-shaped plates and results in a compressive stress in the center PZT plate.
In the process, the excitation is amplified twice: first by the elastic beams and sec-
ond by the bow-shaped plates. Due to the unique two-stage force amplification
mechanism, the structure is able to generate substantial power even under weak
excitations. Unlike most aforementioned harvesters, the proposed design works
in the compressive mode where the piezoelectric elements only experience
compressive stress. The compressive strength of PZT is claimed to be 10 times
stronger than bending and tensile strengths. The compressive mode was therefore
suggested for designing high-performance energy harvesters. A prototype was
fabricated with a PZT-5H plate of 40 3 15 3 0.5 mm3 and a mass of 100 g. Under
a harmonic excitation with an amplitude of 0.3 g at 22.7 Hz, the prototype108
produced a maximum power of 30 mW, corresponding to an average power of
15 mW.

In addition to base excitations, direct force excitations have also been investigated.
Flextensional transducers and stacks are the two most commonly used energy-har-
vesting structures under direction force excitations.109–112 To improve power
output, Wang et al.113 proposed a flex-compressive PEH that can be sustained
under heavy loads of thousands of newtons (Figure 14A). The bow-shaped center
shell amplifies and transfers the applied load force to the symmetrically arranged
piezoelectric stacks. Under a 4-Hz harmonic excitation with a 600-N amplitude, a
maximum power output 17.8 mW was generated by a prototype (piezo stack:
20 3 20 3 36 mm3). Xu et al.114 developed a hybrid PEH using the flexure force
amplification mechanism (Figure 14B). The proposed structure resembles the con-
ventional flextensional transducer, but was modified with two outside curved piezo
stacks. While a force is applied, the outside curved piezo stacks are compressed
(d33); by contrast, the center piezo stack is stretched (d31). It is thus called a hybrid
transducer system. A 35.5 3 25 3 10 mm3 prototype generated 76.6 mW power un-
der a direct normal force excitation of 15 N at 345 Hz, and a 26% energy conversion
efficiency was claimed. The ultra-high power output and efficiency are mainly

18 Joule 2, 1–56, April 18, 2018

 Please cite this article in press as: Yang et al., High-Performance Piezoelectric Energy Harvesters and Their Applications, Joule (2018), https://
doi.org/10.1016/j.joule.2018.03.011

contributed by the flexure force amplification mechanism and the fact that almost all
inactive materials have been eliminated in the new designed structure.

New Piezoelectric Materials and Composites

Researchers have tried to improve energy harvester performance by fabricating new
piezoelectric materials and composites. Morimoto et al.115 fabricated c-axis-ori-
ented epitaxial PZT films on Pt/MgO substrates using the RF-magnetron sputtering
technique, and transferred the films onto stainless-steel cantilevers for improved
efficiency and toughness. Yeo et al.116 formed (001) textured PZT films on the
both sides of a Ni foil by the RF-sputtering method with post annealing. Hwang
et al.117 fabricated PZT films of 7 mm on flexible plastic substrates using the aerosol
deposition method. Motivated by the extraordinarily high coupling factor and piezo-
electric coefficients, single-crystal piezoelectric materials are being extensively used
in energy harvesters. Xu et al.118 fabricated vertically aligned single-crystal PZT
nanowire arrays at 230 C using the chemical epitaxial growth method. Tang
et al.119 developed a bonding and thinning process and fabricated single-crystal
PMN-PT thick films (10 mm) on silicon wafers to obtain a sufficiently powerful
MEMS energy harvester. Yang and Zu28 presented a systemic performance compar-
ison between PMN-PT, PZN-PT, and PZT. Erturk et al.120 firstly made performance
analysis of single-crystal PMN-PZT unimorphs for piezoelectric energy harvesting,
then Hwang et al.121 developed thin PMN-PT films for powering artificial pace-
makers. Later, the same group adopted the solid-state single-crystal growth method
for fabricating high-performance PMN-PZT films.122 Zeng et al.123 developed a high-
performance and stable PIMNT flake array on an ABS substrate. Xu et al.124 synthe-
sized high-performance and flexible nanocomposites by forming a unique hierarchi-
cal structure of PMN-PT nanowires in a PDMS base. The same group also presented
a review of PMN-PT nanostructures for energy harvesting.125 Chen et al.126 pre-
sented micropatterned single-crystal PMN-PT ribbons for capturing both mechani-
cal and thermal energy from the human body.

Performance Comparison
It is not straightforward to conduct a fair performance comparison on energy har-
vesters of different architecture, size, material, and operation mode. The difficulty
in the performance comparison, on the one hand, is attributed to the lack of univer-
sal figures of merit, and on the other hand is caused by the fact that the amount of
data disclosed in the literature varies considerably. Here, we review some figures
of merit used to assess energy harvesters and compare high-performance harvesters
proposed in the past decade.

The most commonly used metric for energy harvesters is power density, which is
defined as power output over volume. However, it is not suitable to be used to
compare different designs because the overall power output is not always strictly
proportional to volume. The power output of inertial PEHs is related to mass (prod-
uct of density and the cube of linear dimension) and the maximum displacement.
Thus the averaged output power is proportional to volume4/3 (Marin et al.127). More-
over, the definition of power density does not take into consideration the effect of
the input excitation, which is extremely influential.

To include the excitation effect, Beeby et al.128 proposed the figure of merit normal-
ized power density (NPD). The NPD is the stated power output of the harvester
normalized to the volume and the acceleration amplitude squared. Although the
defined NPD is also not an ideal metric due to the lack of the frequency information,
it presents an indication of relative performance levels. Note that here, volume can

Joule 2, 1–56, April 18, 2018 19

 Please cite this article in press as: Yang et al., High-Performance Piezoelectric Energy Harvesters and Their Applications, Joule (2018), https://
doi.org/10.1016/j.joule.2018.03.011

refer to the volume of active materials or the overall volume of the system. Both
calculation methods have drawbacks. The calculation only with the volume of active
materials is biased and in favor of PEHs with small-size piezoelectric elements. The
entire volume is not advised for use in the calculations because: (1) harvesters usually
have irregular geometries; (2) the vibration amplitude varies considerably in the liter-
ature; (3) the accurate dimensions of most harvesters are not clearly documented; (4)
at the system level, proper deployments, for example, those using rectification and
storage circuits as the seismic mass block, significantly decrease the total volume.
Considering these points, we calculated the NPD here based on the volume of active
materials.

Power
NPD = (Equation 5.1)
Volume 3 Acceleration2
Another widely adopted metric is efficiency, commonly defined as output electric
energy over input mechanical energy. Despite of the agreement on the efficiency
definition, the reported efficiency value in the literature varies considerably, as sum-
marized by Yang et al.129 Some130 have claimed an efficiency value of over 80%;
others131 reported an efficiency of less than 1%; yet others132 argued that efficiency
can never exceed 50%. Yang et al.129 presented a systematic analysis on efficiency
and pointed out that the discrepancy is mainly caused by the different ways of calcu-
lating the input and output energy of inertial energy harvesters. It was found that in a
vibrating energy-harvesting system, energy does not always flow in one direction
from the excitation to the harvester, but also flows in reverse from the harvester to
the excitation source, which is characterized by the phase difference (fx) between
the external excitation and the system response. The experimentally validated effi-
ciency expression was derived as

V2 R
h=   RMS   ; (Equation 5.2)
1=2 3 m 3 X € 3 Z_ sinðfx Þ
amp amp

where VRMS, R, and m are the RMS voltage response, load resistance, and equiva-
lent mass, respectively. Xamp denotes the amplitude of the response acceleration,
and Zamp denotes the amplitude of the excitation velocity. The authors found
that the phase difference is always 90 for both linear systems and nonlinear
systems around the resonant frequencies, i.e., no energy flows back to the
excitation. Therefore, the efficiency calculation becomes much simpler at resonant
states.

One should note that efficiency is not a comprehensive figure of merit for assessing
harvester performance. It only measures the energy conversion ratio and does not
reflect the size effect and the capability of trapping mechanical energy. High effi-
ciency does not always mean high power output. An energy-harvesting system
with a weak damping effect and a low-frequency excitation always exhibits a high ef-
ficiency value.

In addition to power density and efficiency, some complicated performance metrics

have also been proposed, which take more parameters into consideration.
Roundy133 enhanced the concept of efficiency and proposed a new metric, named
‘‘effectiveness.’’ The effectiveness considers the quality factor (Q), density (r),
coupling coefficient (k), and transmission coefficient of energy harvesters (l) and is
defined by

r l
e = k 2 Q2 : (Equation 5.3)
r0 lmax

20 Joule 2, 1–56, April 18, 2018

 Please cite this article in press as: Yang et al., High-Performance Piezoelectric Energy Harvesters and Their Applications, Joule (2018), https://
doi.org/10.1016/j.joule.2018.03.011

These parameters depend upon multiple operating conditions, and thus the opera-
bility of the defined ‘‘effectiveness’’ is poor. Mitcheson et al.19 developed the
volume figure of merit (FoMv) to reflect a harvester’s nearness to the performance
limitation. The performance limit, i.e., the maximal available power was defined
with respect to a benchmark equivalent cubic geometry (Vol), the density of gold
(rAu), and vibration frequency (w) and amplitude (Y0), as:

Useful Power Output

FoMv = : (Equation 5.4)
1=16Y0 rAu Vol4=3 u3
A high power output is not the only goal pursued in energy harvesters. A wide work-
ing bandwidth has also been examined to ensure the environmental adaptability and
reliability. Conditions to attain these two goals most of time are not aligned with
each other. To increase power output, one usually tries to minimize the mechanical
damping effect of inertial energy harvesters, but this entails the penalty of reduced
working bandwidth. To indicate both bandwidth and power output comprehen-
sively, Mitcheson et al.19 took one step further and proposed a bandwidth figure
of merit, which is defined as the product of FoMv and the fractional bandwidth
(Dw/wn).

Liu et al.134 also proposed a universal metric considering both bandwidth and output
power. The proposed systematic figure of merit with bandwidth information
(SFoMBW) for wideband PEHs is related to an averaged FoMv, mechanical damping,
density, and the fractional bandwidth. In essence, the SFoMBW reflects the product
of the average NPD and bandwidth.

Problems exist in the application of these complicated new figures of merit. First, all
of them are derived based on the assumption or approximation of linear SDOF sys-
tems. Thus, they do not possess universality for nonlinear energy harvesters. In addi-
tion, there is no consensus on the bandwidth definition for nonlinear PEHs. For a
linear vibration system, we usually use the classic definition of the 3-dB bandwidth,
i.e., frequency points at which a half the maximum power is produced. Nonlinear sys-
tems have more than one stable solution over a portion of the bandwidth. The sys-
tem may jump among these solutions. Also, different from linear systems, the
maximum point may not occur at the center of the defined bandwidth, but at one
of its extremes. Liu et al.134 defined the bandwidth as the frequency range from
the upper limit of the forward/backward frequency-sweep case to the lower limit
of the opposite-direction frequency-sweep case for typical hardening/softening
Duffing system. Cammarano et al.135 analyzed the response characteristics of
nonlinear PEHs and proposed to limit bandwidth to the region where only one stable
solution exists. More practically, some articles reported the bandwidth as the fre-
quency range above a certain power output level.40,136

Overall, we have witnessed many attempts to redefine or develop new figures of

merit that can be used to evaluate harvesters’ performance comprehensively. There
is no consensus among researchers on a universal metric. The performance of energy
harvesters is highly dependent on the operation conditions, dimensions, materials,
and circuits employed. Different energy harvesters are developed for operation un-
der different vibration conditions. It is inappropriate and impractical to compare
various designs and approaches with one benchmark. Besides, it is very common
that a publication does not present enough data to calculate a complex performance
metric. We believe the most suitable standard for assessing energy harvesters is
whether a proposed design meets the requirements of the targeted application un-
der the specific constraints. Hence, parameters important to end users, such as

Joule 2, 1–56, April 18, 2018 21

 Please cite this article in press as: Yang et al., High-Performance Piezoelectric Energy Harvesters and Their Applications, Joule (2018), https://
doi.org/10.1016/j.joule.2018.03.011

Table 2. Performance Comparison of State-of-the-Art Inertial Energy Harvesters: MEMS Scale

Ref. Material Excitation Frequency Piezo Size Mass (g) RMS Power Power NPD Note
Amplitude; (Hz) (mm3 3 number) (mW) Density (kg m3
g = 9.8 m s2 (mW cm3) 3 103)
Lee et al.137 PZT 2.5 g 255.9 3 3 1.5 3 0.005 0.873 3103 1.38 61.3 0.102 aerosol PZT
2009
Shen et al.138 PZT 0.75 g 183.8 3.2 3 0.4 3 0.001 – 0.32 250 4.627 sol-gel PZT
2009
Park et al. 139 PZT 0.39 g 528 1 3 0.8 3 0.001 1.223103 1.1 1.375 3 103 94.1 d33 mode
2010
Morimoto PZT 0.51 g 126 18.5 3 5 3 0.0028 – 5.3 20.46 0.818 c axis-oriented
et al.115 epitaxial PZT
2010
Tang et al.140 PZT 1.0 g 514.1 3 3 0.8 3 0.031 2.89 3103 11.56 155 1.61 thinned buck
2012 PZT
Tang et al.119 PMN-PT 1.5 g 406 7.4 3 2 3 0.01 0.84 3103 7.18 97 0.449 d33 mode
2014
Minh le AlN 0.816 g 792 1 3 0.21 3 0.002 0.263103 1.3 5.42 3 103 84.7 MgZr co-doped
et al.141 AlN
2015
Tang et al.142 PZT 3.0 g 100.8 11 3 5 3 0.057 0.12 321 102.4 0.118 bronze substrate
2016
Yeo et al.116 PZT 0.16 g 6 20 3 10 3 0.003*2 – 284 236.7 96.27 001 textured
2016 compliant
structure
Song et al.143 PZT 0.25 g 68 0.11 4.93104 0.023 0.209 0.035 MEMS; spiral
2017 shape

power output, frequency, volume, and input excitation, should be demonstrated

clearly. Meanwhile, the NPD and efficiency are recommended for the purpose of
performance comparison. They are not universal metrics but can be easily calculated
and present, on the basis of fundamental considerations, some knowledge of perfor-
mance level.

Table 2 and Table 3 summarize the main features of some high-performance vibra-
tion energy harvesters reported in the last 10 years. For each harvester, the table lists
bibliography references, material characteristics, excitation properties, harvester
sizes, output power, power density, NPD, and eventual additional notes. We intend
to but are unable to assess efficiency here because of the lack of information of me-
chanical responses in most publications. Only papers including experimental data
are listed in the tables.

As can be seen from the tables, the proposed energy-harvesting systems are hetero-
geneous. The output power is from hundreds of nanowatts to tens of milliwatts. The
optimal working frequency ranges from 2.7 Hz to 853 Hz. Most MEMS energy har-
vesters show higher working frequency and power density, but much lower output
power than mesoscale energy harvesters.

Applications
PEHs have been explored in a variety of fields in the past decade, including wearable
devices, medical implants, vehicles, and wireless sensor networks. Here we mainly
review four specific most widely studied applications: shoes, pacemakers, tire pres-
sure monitoring system (TPMS), and building and bridge monitoring systems. In
each application, we discuss the excitation sources, analyze state-of-the-art har-
vesters, and identify existing problems and future directions. In addition to the

22 Joule 2, 1–56, April 18, 2018

 Please cite this article in press as: Yang et al., High-Performance Piezoelectric Energy Harvesters and Their Applications, Joule (2018), https://
doi.org/10.1016/j.joule.2018.03.011

Table 3. Performance Comparison of State-of-the-Art Inertial Energy Harvesters: Mesoscale

Ref. Material Excitation Frequency Piezo Size Mass (g) RMS Power NPD (kg m3 Note
Amplitude; (Hz) (mm3 3 Power Density 3 103)
g = 9.8 ms2 number) (mW
cm3)
Dai et al.144 PZT-5H 1g 51 12 3 10 3 1 – 1.055 8.79 0.0915 terfenol-D/PZT
2009 mW composite
Kim et al.145 PZT-5A 0.255 g 109.5 53 3 31.7 3 – 0.53 0.57 0.091
2010 0.275*2 mW
Liang and Liao146 PZT-5A 1.44 g 42 49 3 24 3 0.5 2.5 mW 4.25 0.02 impedance
2010 matching circuit
Gu147 PZT-5A 0.4 g 20.1Hz 26 3 6.4 3 0.5*2 8 1.08 mW 9.01 0.59 low-frequency
2011 impact
Li et al.148 Piezo 1g 87 3141 – 14.6 mW 4.65 0.05 compressive
2011 stack mode
Yen et al.149 AlN 1.0 g 853 0.4 3 0.25 3 0.317 3 4.9 nW 12.25 0.127 corrugated
2011 0.002*2 103 cantilever;
bimorph
Xu et al.102 PMN-PT 3.2 g 102 25 3 5 3 1*2 4.2 1.85 mW 7.4 7.5 3 103 flextensional
2012 transducer
Zhao et al.97 PZT 1.48 g 73 13 3 2.5 3 1*2 – 8.7 mW 0.134 0.64 3 103 d15 mode
2012
Tang and Yang38 MFC 0.29 g 27.8 28 3 7 3 0.3 5.9 1.43 mW 24.32 3.04 piezo fiber
2012 composite
Dhakar et al.150 PZT-5A 0.2 g 36 31.9 3 6.4 3 0.5 0.72 40 mW 0.392 0.102 composite
2013 cantilever
Wu et al.151 MFC 0.1 g 18 36 3 16 3 0.3*3 21.4 0.75 1.45 1.45 two degrees of
2013 mW freedom
Arrieta et al.152 PZT 0.25 g 20 46 3 20.6 3 7.07 mW 14.9 2.49 piezo
2013 0.25*2 QuickPack
Qiu et al.153 PZT 0.6 g 38 12 3 6 3 1 – 0.39 mW 5.4 0.1567 magnet/PZT
2014 composite
Ma et al.154 PZT-5H 0.2 g 93 30 3 3 3 3 5.4 3.18 mW 0.012 2.94 3 103 perpendicular
2014 electrodes
Zhang et al.155 PZT 0.29 g 2.71 50 3 30 3 0.2*2 282.6 7.7 mW 12.83 1.59 low frequency;
2014 multi-impact
Xiao et al.156 PZT 1g 225 F25 3 0.2*4 198 10 mW 25.48 0.265 circular
2014 diaphragm
array
Tufekcioglu and PZT 5H 2g 153 F12.7 3 1*2 0.34 0.14 mW 0.55 1.44 3 103
Dogan106
2014
Kulkarni et al.99 PZT-5A 1g 620 5 3 5 3 0.5*4 106 0.57 mW 11.4 0.119 d15 mode;
2014 torsion
excitation
Hung et al.157 PZT 3.5 g 75 4 3 3.5 3 1*4 7.8 0.239 mW 4.27 3 3.64 3 106 three-axis
2015 103 energy
harvester
2015 PZT 4.6 g 90 35 3 16.8 3 0.2 – 385 mW 3.27 1.61 3 103 bistable with
Singh et al.158 the SSHI circuit
Sriramdas et al.159 PVDF 0.5 g 30.8 108 0.63 8.59 mW 0.079 3.3 3 103 multistep and
2015 multilayer
beam
Gong et al.160 PZT 0.7 g 120.9 31 3 13.6 3 23 6.64 mW 35.8 0.760 folded beam
2015 0.22*2
Yang and Zu28 PZN-PT 0.3 g 37.5 10 3 7.2 3 0.4 2.61 0.43 mW 14.93 1.727
2016
(Continued on next page)

Joule 2, 1–56, April 18, 2018 23

 Please cite this article in press as: Yang et al., High-Performance Piezoelectric Energy Harvesters and Their Applications, Joule (2018), https://
doi.org/10.1016/j.joule.2018.03.011

Table 3. Continued
Ref. Material Excitation Frequency Piezo Size Mass (g) RMS Power NPD (kg m3 Note
Amplitude; (Hz) (mm3 3 Power Density 3 103)
g = 9.8 ms2 number) (mW
cm3)
Wu et al.104 Piezo 0.5 g 56 5918.9 100 4.76 mW 0.804 3 0.0333 high-
2016 stack 103 103 temperature
BiScO3-PbTiO3
Yi et al.161 PZT 3.5 g 77.2 16 3 3 3 0.053 0.433 0.979 158 0.134 beryllium-
2017 16 3 3 3 0.076 mW bronze
substrate
Zou et al.103 PZT-5H 0.4 g 9.9 40 3 10 3 1 14.8 194 mW 0.485 0.032 compressive
2017 mode; magnet
He and Jiang162 PZT-4 1g 66.7 15 3 10 3 0.2 6.84 2.35 mW 78 0.812 multi-mode
2017 chiral structure
Yang et al.90 PZT-5H 0.3 g 44 241 6.34 1.27 mW 5.27 0.610 arc-shaped
2017
Yang et al.108 PZT-5H 0.3 g 22.7 40 3 15 3 0.5 100 15 mW 50 5.78 compressive
2017 mode

aforementioned main applications, energy harvesters have also been explored in

rain,163–166 pavement,167 gunfire munitions,168 cutting tools,169 and wildlife,170,171
which are not covered here.

Shoes
The human body consumes a considerable amount of energy every day to perform
different tasks such as speaking, eating, breathing, walking, arm lifting, and finger
tapping. However, not all the kinetic energy and strain energy can be used for en-
ergy harvesters. The design of wearable energy harvesters for human motions fol-
lows the rule that the addition of the energy harvester does not interfere with normal
activities or significantly alter appearance. Also, harvesters should not significantly
increase the metabolic cost. Therefore, it is necessary to analyze the energy flow
in body motions and figure out how much energy is available for the purpose of en-
ergy harvesting.

It is known that muscles do both positive and negative work within each motion.
Positive work is done by muscles during shortening contractions whereby the tor-
que acted by the muscles at a joint applies in the same direction as the angular ve-
locity of the joint; Negative work is done by muscles during lengthening contrac-
tions whereby the torque applied by the muscles at a joint acts in the direction
opposite to the angular velocity of the joint.172,173 In each motion phase, muscles
work actively to generate intended motion, and they also absorb energy like
brakes, i.e., negative work. The absorbed energy most of time cannot be reused
but is wasted as heat.174 Additionally, muscle cells usually need to do extra work
to dissipate this negative work. An energy-harvesting system can be used to facil-
itate muscles during the negative work phase, retarding the motion, similar to a
‘‘regenerative brake’’ in cars. Apart from the function of dissipating the negative
work, energy harvesters can also supply electricity to wearable and implantable
electronic devices.

It is important to know how much energy we can harness from human body motions.
Researchers have evaluated energy associated with different body motions175,176
under different assumptions, such as finger typing (19 mW), breathing (1 W),
arm lifting (60 W), and walking (67 W). Table 4 lists available energy associated

24 Joule 2, 1–56, April 18, 2018

 Please cite this article in press as: Yang et al., High-Performance Piezoelectric Energy Harvesters and Their Applications, Joule (2018), https://
doi.org/10.1016/j.joule.2018.03.011

Table 4. Available Energy in Different Human Motions for Energy Harvesters

Joint Work (J per Step) % Negative Work
‘ 1–5 50
Ankle 34.4 28.3
Knee 18.2 92
Hip 18.96 19
Elbow 1.07 37
Shoulder 1.1 61

Data from Riemer and Shapiro.177

with different human motions for energy harvesters. Among all body motions,
walking is one of the most common and energy-intensive motion. In 2011, Riemer
and Shapiro177 analyzed the energy consumption by muscles at each joint or
segment of a body during walking. They considered an 80-kg person walking at a
frequency of 1 Hz. As listed in Table 3, on average more than 1 J of energy can
be harvested per step in each joint, which means over 1 W power is available for en-
ergy harvesters. Recently, Partridge et al. assumed a 10-mm deflection in shoes
while walking, and estimated that the harvestable energy from shoes is 6.6 J/step
on level ground, 7.47 J/step while going downstairs, and 3.12 J/step while going
upstairs.178

There are mainly two methods that have been explored to harvest energy from hu-
man motions: the piezoelectric effect and electromagnetic effect. The electromag-
netic method is a mature technology and has been studied for decades. To obtain
satisfactory power output, the input linear motions or oscillations need to be con-
verted to high-speed rotational motions. Therefore, large-amplitude motions like
knee bending and arm swing are best treated by electromagnetic energy harvesters.
Compared with the electromagnetic method, harvesters using the piezoelectric ef-
fect can be designed as flexible with a smaller deformation requirement. They are
thus easier to be integrated into shoes without interfering with the gait. A variety
of energy harvesters have been proposed and embedded in shoes using either
the electromagnetic effect179–181 or the piezoelectric effect. Here we mainly analyze
energy harvesters in shoes using the piezoelectric effect. Only studies with experi-
mental data are reviewed.

Harvestable energy in shoes comes mainly in three forms of excitations: the relative
motion between foot and ground (deformation), the swing motion in each gait (ac-
celeration), and the bending of the shoe sole (stretching). A variety of designs have
been proposed to generate electricity by harvesting energy from one excitation or
more excitations combined.

One well-known pioneering work came from the MIT Media lab182 in 2001. Although
we only intend to review research in the past decade here, we cannot avoid
mentioning the importance of this instructive research. Figure 15 shows the sche-
matic of the proposed two methods. One approach is to capture energy from the
process of bending the ball of the foot, using a flexible bimorph made of multi-
laminar PVDF films. The other approach is to harness heel strike energy by a sand-
wiched curve structure, called a ‘‘dimorph.’’ As shown in Figure 15, the dimorph con-
sists of a beryllium-copper plate sandwiched by a pair of curved PZT-spring steel
composite thin plates. For the PVDF harvester, an average power of 1.3 mW was
generated at a walking pace of 0.9 Hz. Under the same excitation, the PZT harvester

Joule 2, 1–56, April 18, 2018 25

 Please cite this article in press as: Yang et al., High-Performance Piezoelectric Energy Harvesters and Their Applications, Joule (2018), https://
doi.org/10.1016/j.joule.2018.03.011

Figure 15. Shoes Energy Harvester from the MIT Media Lab
(A) Schematic.
(B) Power responses from brisk-walking tests with PVDF stave and PZT dimorph.
Reproduced from Shenck et al. 182 with permission from IEEE.

generated 8.4 mW on average. The peak voltage and power can be as high as 200 V
and over 50 mW, respectively. The fabricated prototype demonstrated its ability to
intermittently power an active radiofrequency (RF) tag. The successful demonstra-
tion is extremely encouraging, and along with the booming wearable technologies
has stimulated much research in the following years.

PEHs have been exploited to harness energy from acceleration exactions in shoes. In
2010, Moro and Benasciutti183 carefully analyzed the characteristics of the accelera-
tions on the heel during walking, and designed a cantilever energy harvester using
piezoceramics. One main contribution of this work is that they correlated the heel ac-
celeration data from the literature with direct experimental measurements (see Fig-
ure 16B), and quantitatively defined a standard acceleration signal. The defined exci-
tation signal is useful in guiding other endeavors to harvest energy from heel
accelerations. As shown in Figure 16B, in each gait cycle we observe an acceleration
peak (6–6.5 g) instantaneously at the initial contact phase. Following that is the
stance phase, which continues for about 60% of the cycle period, during which
time the acceleration increases smoothly. Then in the final swing phase the acceler-
ation along the tibial axis experiences a large oscillation. To validate the proposed
energy-harvesting method and analytical model, the authors fabricated a prototype
and tested it with different parameters. Figure 16A illustrates the PZT-5A bimorph
(20 3 14 3 0.3 mm3) mounted inside the shoe heel. The tip mass was varied from
5.5 g to 16.4 g, resulting in a range of the first resonant frequency of 20–150 Hz.
Excited by the acceleration in the tibial direction, the prototype achieved a maximum
mean power of 0.395 mW per footstep. This work demonstrates that it is feasible to
acquire on the order of 1 mW of power from heel accelerations in one direction.

Accelerations in a shoe are not always in one direction, but may vary in multiple
directions. Therefore, a multi-directional harvester is expected to generate more
energy. In 2017, Fan et al. proposed a composite configuration to capture vibration
energy not only from the tibia direction, but also from the tangential direction of the
swing motion and the sole compression. Figure 16C shows the schematic. The PEH

26 Joule 2, 1–56, April 18, 2018

 Please cite this article in press as: Yang et al., High-Performance Piezoelectric Energy Harvesters and Their Applications, Joule (2018), https://
doi.org/10.1016/j.joule.2018.03.011

Figure 16. Devices to Harness Energy from Acceleration Exactions in Shoes

(A) Shoe-mounted piezoelectric cantilevers. Reproduced from Moro et al. 183 with permission from
IOP Publishing Ltd.
(B) Displacement, velocity, and acceleration waveforms of heel over one stride. Reproduced from
Moro et al. 183 with permission from IOP Publishing Ltd.
(C) Multi-directional energy harvesting. Reproduced from Fan et al. 184 with permission from AIP
Publishing LLC.

consists of a piezoelectric cantilever beam magnetically coupled to a ferromagnetic

ball and a crossbeam. The main cantilever is made of a brass substructure of 23.1 3
7.3 3 0.05 mm3 and a PZT-5H beam of 19.1 3 7.1 3 0.245 mm3. A steel ball of diam-
eter (F) 7 mm was encapsulated in a 34-mm long polycarbonate sleeve to harvest
energy from the swing motion. A cylindrical NdFeB magnet (F 3 3 2 mm3) was fixed
at the free end of the cantilever, coupled with a magnet (F 5 3 2 mm3) on the fixed
crossbeam. The crossbeam was connected with the insole, and was able to move up
and down while walking with the help of two springs inserted in the supporting pil-
lars. The structure was proposed to harvest energy from three directions simulta-
neously, but these acceleration excitations may hold a certain phase difference
that causes mutual cancellation. The experiments indicated that the prototype
generated 0.03–0.35 mW at the walking speed ranging from 2 km hr1 to 8 km hr1.

Overall, acceleration excitations in shoes are characterized with low frequency, large
amplitude, and multiple directions. Inertial PEHs generated <1 mW power from the
acceleration exactions. The relatively low power output is mainly attributed to the
low-input mechanical energy. Because it is not permitted to design a heavy inertial
mass in a shoe-inserted inertial energy harvester, we can only capture very limited
kinetic energy. Future work may explore new amplification methods to capture
more inertial energy.

Compared with the acceleration excitations, deformation and stretching excitations

hold more energy and are easier to be harnessed. Although the allowed deformation
in a shoe is very limited, the effective load force from a body weight is at the level of
several hundred newtons, which have stimulated researchers to develop a variety of
strain-based harvesters.185,186

For harvesting the deformation energy, PVDF films are the ideal candidates due to
their outstanding flexibility. PVDF films have been widely explored in many groups

Joule 2, 1–56, April 18, 2018 27

 Please cite this article in press as: Yang et al., High-Performance Piezoelectric Energy Harvesters and Their Applications, Joule (2018), https://
doi.org/10.1016/j.joule.2018.03.011

Figure 17. Devices to Harness Energy from Deformation and Stretching Excitations in Shoes
(A) PVDF wavy surfaces and the electrical responses under strike. Reproduced from Zhao et al. 189 with permission from MDPI AG.
(B) Curved PVDF stack. Reproduced from Jung et al. 190 with permission from Elsevier.
(C) PZT flex transducer. Reproduced from Daniels et al. 191 with permission from IOP Publishing Ltd.
(D) PZT with an amplification mechanism. Reproduced from Xie et al. 192 with permission from AIP Publishing LLC.

successively.187,188 In 2014, Zhao and You189 developed a 1-mW shoe-embedded

PVDF energy harvester. In the design, to attain a high power output the authors
stacked eight layers of PVDF films together and sandwiched them between two
wavy surfaces. As shown in Figure 17A, the stacked PVDF film (50 3 80 mm2) was
fixed on the lower plate at both ends, and the upper movable plate was activated
by the foot. Once the upper plate was moved down, the PVDF film would be
stretched and fit in the wavy surface. Figure 17A shows the irregular response curves
of the prototype. Under brisk walking at 1 Hz, the prototype generated a peak-to-
peak voltage of 136 V, a peak power of 4 mW, and an average power of 1 mW.

Instead of the flat pattern, PVDF has also been made in curved shapes. In 2015, Jung
et al.190 proposed a curved PVDF energy harvester, which was claimed to be able to
light up 476 LEDs. Figure 17B illustrates the proposed energy harvester. Two
0.1-mm thick PVDF films were attached on each side of an arc-shaped polyimide
(PI) (0.2 mm thick) substrate. A pair of such arc PVDF-PI composites were assembled
together back to back. In the design, PVDF films experience bending stress without
stretching when a load is applied on the harvester. A prototype, fabricated with four
70 3 40 mm2 PVDF films, generated an averaged open-circuit voltage of 25 V and
an averaged short-circuit current of 20 mA, during walking at 0.5 Hz. The high per-
formance is contributed by the large surface area (multiple layers) and high bending
stress (curved structure). Theoretically, the electromechanical coupling effect of
PVDF films is over 10 times weaker than that of PZT ceramics. PZT ceramics thus
have a greater potential than PVDF films to generate more power from deformation
and stretching in shoes. In 2013, Daniels et al.191 used a rectangular flextensional
transducer to harvest energy from the relative motion between foot and ground dur-
ing walking (Figure 17C). The flextensional end caps amplify the downward applied
force and also protect the sandwiched piezo elements from impacts. To withstand

28 Joule 2, 1–56, April 18, 2018

 Please cite this article in press as: Yang et al., High-Performance Piezoelectric Energy Harvesters and Their Applications, Joule (2018), https://
doi.org/10.1016/j.joule.2018.03.011

Table 5. Power Requirements of Different Medical Implants

Device Power Requirement
Pacemaker196,197 5–10 mW
Cochlear implant198,199 100–2,000 mW
Drug pump200 400 mW
201
Retinal stimulator 250 mW
Neural recording202 1–10 mW

the nearly 1-kN force, the piezoceramic (52 3 30 3 4 mm3) is covered by a pair of
steel plates. The harvester prototype was placed under the sole of a boot. It gener-
ated an averaged power of 2.5 mW under 760 N at 1.4 Hz, and was able to power a
wireless sensor module independently.

In 2014, Xie and Cai192 proposed a novel structure, consisting of an amplification

mechanism and a stack of piezoceramic bimorphs, as shown in Figure 17D. A foot
strike applies an impact force on the V-shaped slider in the vertical direction and
then causes a movement of the symmetric H-shaped sliders in the horizontal direc-
tion. The horizontal movement deforms the bimorphs with one end fixed at the
center and the other end inserted in comb-like clampers. The bimorphs are symmet-
rically deformed and an electrical potential is formed over the PZT elements. The
symmetric in-phase deformation in all piezoelectric elements provides substantial
convenience to the conditioning circuits that follow. The harvester has four springs
that connect the static framework and two clampers to restore the V-shaped slider.
A prototype, 105 g in total, was fabricated with 20 PZT-5H bimorphs (30 3 15 3
0.2 mm3). The prototype was embedded in the heel of a shoe and excited by a
user of 68 kg. The maximum RMS power is 18.6 mW at 1 Hz (regular walking
pace), and the corresponding power density over the whole volume is about
0.41 mW cm3, which is higher than that in some electromagnetic harvesters.180,193

Overall current energy harvesters can generate tens of milliwatts of power from
shoes. Research is being conducted to further improve the efficiency and power
density of energy harvesters, and at the same time explore new applications of
the shoe-inserted energy harvesters.

Pacemakers
The wide diffusion of implantable medical devices has been raising the quality of life
and life expectancy. Take the artificial pacemaker, for example; it has saved over
3 million lives to date, and every year 600,000 new pacemakers are implanted in
patients.194 Ideally, implantable electronic devices are expected to work automati-
cally inside bodies for years without maintenance. However, the commonly used
electrochemical batteries cannot constantly supply power for long enough. It is
therefore pressing to find practical methods to increase the energy density and
lifespan of power sources. By harnessing energy from the nonstop body motions,
energy harvesters supply a perpetual approach to achieve lifelong autonomous
operations of medical implants.

Powering medical implants by harvesting a small part of energy from body motions
without any side effects is a true possibility. To secure years of autonomous opera-
tions, low power consumption has become one of the most critical requirements for
the design of biomedical implants.195,196 As microelectronics develop rapidly,
implantable devices require increasingly less energy. Table 5 lists the power con-
sumptions of five medical implants. Most of them require power less than 1 mW,

Joule 2, 1–56, April 18, 2018 29

 Please cite this article in press as: Yang et al., High-Performance Piezoelectric Energy Harvesters and Their Applications, Joule (2018), https://
doi.org/10.1016/j.joule.2018.03.011

and the artificial cardiac pacemaker only consumes 5–10 mW. The required micro-
watt-level power is trivial compared with the rich mechanical energy associated
with body motions.

The artificial cardiac pacemaker is an intriguing application for energy harvesters

and has attracted much attention from both the engineering and medical fields.
The longevity of pacemakers is mainly decided by the battery lifespan. Currently
pacemakers are powered by lithium-iodine cells, which theoretically can support
pacemakers to operate uninterruptedly for up to 10 years. However, the lifespan
of a pacemaker depends upon the frequency of use, and is usually less than 7 years,
which is not satisfactory especially for the increasing number of young patients (as
young as 2 years old). Also, new functions, such as biventricular pacing (three elec-
trodes), consume more energy than current pacemakers (one electrode). Energy har-
vesters hold great potential to prolong the lifespan of pacemakers and supply extra
power for new functions.

Design of energy harvesters for pacemakers is much more challenging than that for
industrial applications because of the very limited space, which is typically less than
1 cm3. Furthermore, human motion is most of the time in a low frequency range un-
der 10 Hz. As a rule of thumb, the smaller the size of the harvester, the higher will be
the resonant frequency. This means we have to develop a compact energy-harvest-
ing system that can work under ultra-low-frequency excitations with an ultra-high ef-
ficiency. Both the electromagnetic effect and piezoelectric effect have been
explored to harvest energy for pacemakers. The mass imbalance oscillation gener-
ator in automatic wristwatches has been tested, and generated a power of
16.7 mW from the cardiac wall motion of a sheep.203 Due to the very limited space,
PEHs are supposed to show better performance than electromagnetic counter-
parts.8 Here we mainly review recent research on powering pacemakers using PEHs.

There are mainly three excitation sources that have been studied for energy har-
vesters in pacemakers: heart beats, blood pressure gradients, and arterial wall defor-
mation. The cyclic heart beat causes vibration and deformation of the myocardium at
a frequency of 1–3 Hz.204 The frequency depends on a person’s activity and
morphology. The induced strain on the myocardium is 15%–23% in the radial direc-
tion and 9%–12% in the circumferential direction.205 The blood pressure gradient, in
a healthy body, varies with an amplitude of about 20–100 mmHg (2.7–13.3 kPa) in
right and left ventricles, and about 40 mmHg (5.3 kPa) in the arterial system.204,206
For arterial wall deformation, Länne et al.207 reported that the diameter of the distal
abdominal aorta varies between 15.8 mm (blood pressure 118 mmHg) and 17.3 mm
(blood pressure 64 mmHg) at a heart rate of 66 beats per minute. The diameter vari-
ation of the carotid artery is about 10% between the diastolic and systolic period.208
The human brachial artery experiences a diameter distension of 3.7%.208 Blood flow
is also a possible energy source for energy harvesters. However, concerns such as
the blood cell damage from rotating blades and the risk of vortex-induced thrombi
impede research endeavors on flow-based energy harvesters for pacemakers.

We now review energy harvesters that use the heart motions. In early 2011, Inman’s
group209,210 started to explore the possibility of utilizing the heartbeat-induced vi-
bration energy inside the chest to power pacemakers. A battery roughly occupies
two-thirds of the volume of a pacemaker (42 3 51 3 6 mm3 of Biotronik’s device),
and thus an inserted harvester should not be designed to be larger than the battery
volume. Within such a small space and restricted weight, linear PEHs will have reso-
nant frequencies much higher than that of the heartbeat excitation. To lower the

30 Joule 2, 1–56, April 18, 2018

 Please cite this article in press as: Yang et al., High-Performance Piezoelectric Energy Harvesters and Their Applications, Joule (2018), https://
doi.org/10.1016/j.joule.2018.03.011

Figure 18. Conformal Piezoelectric Energy Harvesting and Storage from Motions of the Heart,
Lung, and Diaphragm
Reproduced from Dagdeviren et al. 214 with permission from National Academy of Sciences.

resonant frequency, in 2017 Ansari and Karami211 proposed a fan-folded structure

that consists of bimorph beams folding on top of each other. A 1-cm3 prototype
was constructed with three bimorph beams (PZT-5A layer: 20 3 5 3 0.19 mm3)
and a tip mass of 18.4 g. The prototype was excited by a normal heartbeat waveform
from a feedback controlled shaker, and generated 16 mW power on average. As well
as the fan-fold structure, other methods such as symmetric meandering configura-
tions212 and multi-cantilever arrays213 have also been exploited for this purpose.

In addition to rigid mechanical devices encapsulated in the solid pacemaker case,

researchers have also tried to develop soft energy harvesters that can be directly
attached to hearts and extract energy from the deformation of the myocardium.
In 2014, Dagdeviren et al.214 proposed a strain-based harvester for conformal en-
ergy harvesting from the contraction and relaxation motions of the heart. Figure 18
shows schematically the PZT thin-film harvester. Thin PZT ribbons (500 nm in
thickness) were sandwiched between metal electrodes, fixed on a thin spin-cast PI
substrate layer, and encapsulated with biocompatible materials. The harvester
was further integrated with a bridge rectifier and a millimeter-scale battery for
simultaneous power generation and storage. The thin PZT ribbon composite
showed a superior stability and was sustained after 20 million bending cycles. The
flexible piezoelectric patch was affixed on the surface of the ventricles of bovine
and ovine hearts. The thin-film harvester generated a maximum open-circuit voltage
of 4–5 V, and its maximum power density could reach 1.2 mW cm2 when using multi-
layer stacks.

Similar work was reported later by Lu et al.215 The authors argued that cows and
sheep are different from human beings with regard to the cardiac anatomy, and
that in vivo tests should be performed in conscious animals rather than in

Joule 2, 1–56, April 18, 2018 31

 Please cite this article in press as: Yang et al., High-Performance Piezoelectric Energy Harvesters and Their Applications, Joule (2018), https://
doi.org/10.1016/j.joule.2018.03.011

Figure 19. Harvesting Energy from Blood Pressure Gradients

(A) Schematic of piezoelectric diaphragms to harvest energy from blood pressure variations.
(B) A micro-spiral piezoelectric transducer that captures energy from blood pressure variations.
Reproduced from Deterre et al. 223 with permission from IEEE.

anesthetized animals. They did experiments in swine using the similar flexible PZT
ribbon composites, and obtained a peak-to-peak voltage of 2–3 V.

To improve the energy density of flexible piezoelectric transducers, Hwang et al.121

proposed a flexible energy harvester using the high-efficiency single-crystal PMN-
PT. They optimized the stress-controlled exfoliating process,216 which transferred
a PMN-PT thin film from a bulk substrate onto a flexible substrate. The PMN-PT
thin film had an area of 17 3 17 mm2 and was 8.4 mm thick. It was bonded on a poly-
urethane (PU)-coated polyethylene terephthalate (PET) substrate. Excited by a
linear motor for a strain of 0.36% at a strain rate of 2.3% s1, the thin film harvester
generated maximum open-circuit voltage of 8.2 V and short-circuit current of
145 mA. In this study, the authors did not apply the developed harvester on a heart.
Instead the PMN-PT film composite was bent by hand and the output voltage was
used directly to stimulate a rat heart. Subsequently, in 2017 the same group217
in vivo tested a single-crystalline PMN-PZT-Mn thin film (20 3 30 3 0.02 mm3) on
a porcine heart, and obtained a short-circuit current of 1.75 mA and open-circuit
voltage of 17.8 V. The group also presented a detailed review of thin-film flexible
harvesters.218

Next we consider methods to harvest energy from blood pressure gradients. The
piezoelectric diaphragm is the most widely studied structure in this applica-
tion.206,219–222 As shown in Figure 19A, typically the harvester consists of a substrate
layer with a thin piece of piezoelectric element attached on one side. The other side
is exposed to the low-frequency blood pressure variation. In the work from Mo
et al.,221 under a 40-mmHg pressure variation (1 Hz), 36 V and 128 mJ per cycle is
generated by a F 25.4 3 0.127-mm PZT-5H wafer on a F 42.4 3 0.285-mm aluminum
substrate. In the work from Deterre et al.,206 a prototype made of a F 25 3 0.11-mm
PZT layer on F 45 3 0.1-mm brass diaphragm generated a maximum voltage of 10 V
and maximum energy per cycle of 25 mJ under a 50-mmHg air pressure variation
(2 Hz). This equates to average power output of about 50 mW. Here, the energy
per cycle refers to the highest energy stored in the piezoelectric elements. It is calcu-
lated via the equation: half of capacitance of the piezoelectric element times
squared maximum voltage over the electrodes. This calculation overestimates the
harvester performance because, in an energy-harvesting system, the effective
energy that can be finally consumed by a target electric load is much less than the
energy stored in the piezoelectric elements. No in vivo tests have been performed.
As these diaphragm harvesters need one side of the diaphragm structure to be

32 Joule 2, 1–56, April 18, 2018

 Please cite this article in press as: Yang et al., High-Performance Piezoelectric Energy Harvesters and Their Applications, Joule (2018), https://
doi.org/10.1016/j.joule.2018.03.011

Figure 20. A PVDF Harvester that Captures Energy from the Arterial Wall Deformation
Reproduced from Zhang et al. 226 with permission from Elsevier.

exposed to the blood flow, possible installation sites may be located along the
endocardium and the endothelium of large arteries.

In 2014, Deterre et al.223 proposed a micro-spiral piezoelectric transducer associ-

ated with a novel micro-fabricated packaging that captures energy from blood pres-
sure variations in the cardiac environment. This is intended for use as a leadless
pacemaker that resides inside a heart.224 As Figure 19B shows, the proposed system
consists of an ultra-flexible electrodeposited micro-bellows (21 mm3 volume), a mi-
cro-machined spiral-shaped piezoelectric bimorph, and a 1-mm diameter coupling
rod. The coupling rod is fixed on a spiral-shaped bimorph on one end, and connects
the top surface of the micro-bellows on the other end. The top surface is only 12 mm
in thickness so that we can regard it as a diaphragm. Both the top diaphragm and
deformable bellows contribute to the total flexibility of the harvester. The article
also presents an optimization of the electrode pattern and geometry of the spiral bi-
morph. Under an excitation of 1.5 Hz and 180 mN (corresponding to the maximal
amplitude pressure variations in the left ventricle on a surface of diameter of
6 mm), the optimized prototype achieved a maximum power density of 3 mJ/cm3/
heartbeat. The energy conversion efficiency was calculated to be 0.57%.

Regarding methods to harvest energy from arterial wall deformations, harvesters on

blood vessels are prohibited to hamper blood flow and arterial movement. There-
fore, soft piezoelectric materials such as PVDF films are preferred in this application.
In 2008, Potlay and Brooks225 designed a PVDF cuff surrounding arterial blood ves-
sels. A small 28 3 8 3 0.028-mm3 PVDF film was embedded inside a 0.25-cm3 self-
curling silicone cuff. The harvester was tested on a mock artery (latex tubing). On
average, 6 nW power was generated from 80 mmHg blood pressure variation
through the F 12.7-mm latex tubing. In 2015, Zhang et al.226 tested a PVDF harvester
on the ascending aorta of a pig heart, as shown in Figure 20. Similar to Potlay’s work,
they first evaluated the performance of the flexible PVDF harvester on a latex tube. A
25 3 56 3 0.2-mm3 prototype showed a maximum output power of 681 nW and a
maximum voltage of 10.3 V, respectively, under a pressure variation of 80 mmHg,
160 beats per minute. When wrapping the PVDF film around the ascending aorta
of a porcine heart, a maximum voltage of 1.5 V was achieved under the heart rate

Joule 2, 1–56, April 18, 2018 33

 Please cite this article in press as: Yang et al., High-Performance Piezoelectric Energy Harvesters and Their Applications, Joule (2018), https://
doi.org/10.1016/j.joule.2018.03.011

of 120 beats per minute and blood pressure of 160/105 mmHg. The implanted
harvester is able to charge a 1-mF capacitor to 1.0 V within 40 s. It is roughly esti-
mated that the average power output is about 12.5 nW.

Overall, the last decade has witnessed an intensive study on energy harvesters for
implantable devices, especially for pacemakers. A variety of rigid and flexible piezo-
electric transducers have been developed with the hopes of powering artificial pace-
makers by harvesting energy from heart beats, blood pressure vibrations, and blood
vessel deformations. It is impossible to quantitatively compare these harvesters due
to the lack of systematic reporting (excitation strength, mass, dimension. and elec-
trical load). Generally the power output from in vivo tests is under 1 mW. Energy har-
vesters currently are facing complex hurdles and there is a long way to go before we
achieve self-powered operation of pacemakers. Most research is still in the stage of
in vitro testing and there are few practical designs that can supply enough power for
pacemakers. No energy harvesters have been integrated into pacemakers and
tested in vivo. Furthermore, other critical issues such as biocompatibility and reli-
ability are waiting to be properly addressed. How does one attach a solid artificial
structure to soft dynamic biological tissues? How do we harvest enough energy
without side effects and interference in normal organ motions? How is fatigue and
degradation avoided over 10 years of operation? These open questions must be
answered in a comprehensive manner analytically and experimentally before clinic
trials are initiated.

Nevertheless, it is believed that energy harvesting is a logical and intriguing concept

for powering medical implants and promises a game-changing technology if suc-
cessful. Theoretically, energy-harvesting systems can be perpetual because, unlike
batteries, they are not power reservoirs but power generators. A perpetual power
source provides great flexibility in the design of future pacemakers and other med-
ical implants. For example, to carry out maintenance and replace batteries easily,
pacemakers currently are implanted under the skin of the chest (just below the collar-
bone) and long lead cables are used to deliver therapeutic stimulus to the heart.
With a perpetual energy source, surgeons can implant newly designed compact
pacemakers directly inside the heart without long lead cables. External cables are
one of the leading causes of failure in implanted devices. Pacemakers are evolving
with more diagnostic and telemetric functions. Demands on power consumption
will be increased as the complexity of devices evolves. Energy harvesters make it
possible to develop new functions without compromising the lifespan. The develop-
ment of energy harvesters also supports other implantable devices such as defibril-
lators or carotid stimulators that require much more power than pacemakers, and the
need to extend their lifespan is generally accepted. Overall the energy-harvesting
technology holds great potential to prolong the lifespan of existing implantable de-
vices with more functions, and support the emergence of new therapies. We can
foresee an increasing level of research on implantable energy harvesters in material,
design, and system aspects.

Tire Pressure Monitoring System

The automotive industry has a great interest in small-scale sensors for control and
safety applications. Currently each vehicle is equipped with about 60 sensors, and
an increasing number of sensors are being integrated into vehicles along with the
rapid development of self-driving technology and electric cars. One main barrier
to the trend in this context is cabling and connectors of these sensor modules.
The total cable length in a modern car is more than 4 km (Auzanneau227), which
adds additional weight, occupies much space, and lowers the vehicle’s reliability

34 Joule 2, 1–56, April 18, 2018

 Please cite this article in press as: Yang et al., High-Performance Piezoelectric Energy Harvesters and Their Applications, Joule (2018), https://
doi.org/10.1016/j.joule.2018.03.011

Figure 21. A TPMS Module and Its Installation on a Wheel

as a whole. These troubles can be avoided by updating sensors to be wireless. How-

ever, wireless sensor modules cannot operate automatically for a long time due to
the limited capacity of embedded batteries. To address the power source issue, re-
searchers have started to exploit on-site wasted energy in vehicles. By harvesting the
tremendous vibration and deformation energy in running vehicles, energy har-
vesters can effectively extend the battery life, with great potential of achieving life-
long autonomous operations of these small-size sensor modules. A variety of en-
ergy-harvesting methods have been proposed for different applications in the
automotive industry.110,228–234 Here our analysis concentrates on energy harvesters
for the TPMS. Tires are the only components in vehicles that directly contact the
ground. They carry out multiple key tasks including supporting the vehicle load,
absorbing shocks from the road, and providing grip during steering and braking;
thus, tires play a determinant role in driving safety and comfort. The TPMS was
initially proposed to enhance safety, but also contributes to higher fuel efficiency
and longer tire life.

TPMS is a very promising application for energy harvesters, and may become one of
the first commercially available energy-harvesting products. This is mainly because
the TPMS market is growing extremely quickly, driven by the stringent fuel efficiency
and safety regulations in the United States, Europe, and Asia. Also, there is enor-
mous energy associated with a rolling tire that can be scavenged by energy har-
vesters. Furthermore, compared with medical implants, there are fewer regulations
in the automotive applications. Energy harvesters can reach the market in a timely
way once their performance meets requirements.

Energy harvesters for the TPMS have garnered much interest academically and
commercially. One can find a variety of energy harvester-based TPMS systems in
the patent literature. Bowen et al.235 and Kubba et al.236 have presented excellent
reviews on tire pressure sensors and energy harvesters for TPMS application
broadly via the electromagnetic, piezoelectric, electrostatic, and thermal effects.
Here we give a detailed and up-to-date analysis of mechanical energy harvesters
using the piezoelectric effect for TPMS, focusing on the structure design and dy-
namic analysis.

Figure 21 shows a TPMS module and its installation on the tire valve. A TPMS usually
is composed of a sensor IC, an low-frequency interface, an RF transmitter, an an-
tenna, and a primary battery. The sensor IC integrates a low-power MCU, an ADC
module, a pressure sensor, a temperature sensor, and an accelerometer. The accel-
erometer is used to enable a sleep mode when the car is idle for a predefined period
of time. In the run mode, tire pressure is measured and wirelessly transmitted once or
several times per minute. The overall power consumption of a TPMS module highly
depends on how many RF transmissions are made in every minute and how long the

Joule 2, 1–56, April 18, 2018 35

 Please cite this article in press as: Yang et al., High-Performance Piezoelectric Energy Harvesters and Their Applications, Joule (2018), https://
doi.org/10.1016/j.joule.2018.03.011

Table 6. Power Consumptions of the Mainstream IC Solutions on the Market

Company IC Solutions Typical Current Consumption
Run mode + RF Transmission
Infineon SP370 run mode: 2 mA
RF: 10 mA
sleep current: 0.7 mA
Freescale FXTH870x run mode: 2 mA
RF: 6 mA
sleep current: 0.5 mA
Texas Instruments TPIC82010 run mode: 1.53 mA
RF: 10.5 mA
sleep current: 0.1 mA
Melexis MLX91801/2 + MLX72013 run mode: 1 mA
RF: 3.8–16.8 mA
sleep current: 0.25 mA

All data are from the datasheets of the ICs. Operation voltage of the ICs is about1.8–3.6 V.

TPMS module stays in the sleep mode. The capacity of TPMS batteries235,237 is
about 220–600 mAh and the typical operation lifespan is 5–10 years.

Table 6 lists the power requirements of the most commonly used TPMS chips on the
market. The RF transmitter consumes the most power, requiring over 10 mA. The
sleep mode, by contrast, only needs current less than 1 mA. Overall, it is estimated
that on average 450 mW is required to attain a 1-Hz transmission rate for a TPMS
module.238 Today’s cars sit unused over 90% of the time, and thus the majority of
energy stored in the battery is consumed by the sleep current.237

Energy harvesters, like the TPMS, work only when the vehicle runs. To maintain oper-
ation, harvesters should have a power output much higher than the power require-
ment of the TPMS in the working mode so that a part of harvested energy can be
stored for the sleep mode. The power output of a harvester embedded in a TPMS
is expected to be at the milliwatt level.

Available Excitations in Tires for Energy Harvesters. A rolling tire is an energy-rich

environment. Energy harvesters can take advantage of either the vibration energy
from the rotation, the gravitational field and the road surface, or strain energy
from the deformation of the tire. Possible installation positions include the inner liner
above the tread wall, the inner liner of the sidewall, and the inside and outside of the
rim. Figure 22 shows the load conditions of a rolling wheel. Due to the vehicle load,
the round surface of an inflated tire flattens at the tire-road interaction zone and
forms an area called the ‘‘contact patch.’’ The existence of the contact patch induces
a periodic acceleration and strain change on the wheel.

The strain change associated with the cyclic deformation of a tire is considerable.
Both the tire tread wall and the sidewall deform around the contact patch. The side-
wall only experiences compressive strain in the radial direction, but the inner liner of
the tread wall experiences compressive stress on each side of the contact patch and
tensile stress within the contact patch in the circumferential direction. The maximum
strain239,240 on the inner liner of the tread wall can be around 0.1%–0.5%. The defor-
mation241–243 of the sidewall can be as high as 10–50 mm.

The vibration associated with a rolling tire contains tremendous energy. It mainly ex-
ists in three forms: gravity, the periodic and random vibrations from the tire-road
interaction, and the periodic acceleration variation from the centripetal force. Grav-
ity is always in one direction toward the earth, but structures attached on a wheel can

36 Joule 2, 1–56, April 18, 2018

 Please cite this article in press as: Yang et al., High-Performance Piezoelectric Energy Harvesters and Their Applications, Joule (2018), https://
doi.org/10.1016/j.joule.2018.03.011

Figure 22. The Load Conditions of a Wheel Rotating on a Road

w represents the angular velocity of the wheel. r is the radius of the tire and r c is the distance
between the contact patch and the center of the wheel.

experience a periodic G1 g acceleration change due to the rotation of the wheel.

The frequency of the vibration follows the rotation speed of the wheel. The radial ac-
celeration keeps at a constant value u2r (u angular velocity of the wheel; r radius of
the tire) most of time per round, and drops to zero in the contact patch area. The
deformation in the contact patch also causes a direction change in the circumferen-
tial (tangential) acceleration. The tangential acceleration goes up while entering the
contact patch and goes down when leaving the contact patch. Figure 23 exhibits a
typical acceleration value measured on a car at the speed of 80 km hr1. The tangen-
tial acceleration is 0 G 1,500 m s2. The radial acceleration is about 2,500 m s2 with
an abrupt drop to 0 m s2 in the contact area. Pooya et al.245 reported a peak radial
acceleration of 100 m s2 and tangential one of 150 m s2. The acceleration level is
highly dependent the driving speed. Schaijk et al.231 reported that the maximum ac-
celeration increases from 300 m s2 to 1,200 m s2 as the driving speed rises from
30 km hr1 to 60 km hr1. The acceleration value also varies under different
road conditions and with different vehicle loads. It can be in a large range of
100–4,000 g (g = 9.8 m s2) (Löhndorf et al. 237 and JESD22-B104-B from Automotive
Electronics Council), taking unexpected shocks into account. Energy harvesters de-
signed to work in this environment must consider the extremely high acceleration to
ensure the lifetime and meet automotive quality requirements.

The vibration from the tire-road interaction is unpredicted and random. There are no
general characteristics shared by the vibrations from different road conditions. Löhn-
dorf et al.237 analyzed a set of tangential acceleration data recorded at the inner liner
of a tire at 50 km hr1, and found that power mainly exists in a low frequency range of
5–20 Hz, corresponding to the revolution period of the wheel, but also shows several
peaks around 100 Hz, which should be from the tire-ground interaction. To make use
of the energy from both the vibration from the tire-road interaction and the periodic
acceleration, energy harvesters ought to be designed with a broad working
bandwidth.

The concept of powering TPMS using energy harvesters was proposed almost at the
same time as when the TPMS legislation was implemented in the United States.

Joule 2, 1–56, April 18, 2018 37

 Please cite this article in press as: Yang et al., High-Performance Piezoelectric Energy Harvesters and Their Applications, Joule (2018), https://
doi.org/10.1016/j.joule.2018.03.011

Figure 23. On-Road Acceleration Data of a Rolling Wheel with Different Vehicle Load
(A and B) Tangential direction (A) and radial direction (B). The driving speed is 80 km hr 1 and tire pressure is 2.6 bar. Data in (B) do not include
centripetal acceleration information. Reproduced from Matilainen and Tuononen 244 with permission from Elsevier.

TPMS suppliers first evaluated the possibility of powering TPMS using PEHs.237 Later
in 2008, Roundy246 analyzed the problems and challenges in the design of energy
harvesters for TPMS. The weight of a harvester should be much smaller than the stan-
dard maximum weight of TPMS, that is, 25 g. The expected lifespan approximates
150 million tire revolutions. It was estimated that a TPMS roughly requires 1.125 mJ
per minute to operate. Since then, various energy-harvesting methods have been
proposed, which, from the perspective of excitation sources, can be divided into
two categories: vibration energy and strain energy.

Vibration Energy. Cantilever energy harvesters were first studied to power

TPMS.247,248 A piezoelectric cantilever is mounted on the tire or rim to harness
mechanical energy from the acceleration change around the contact patch and
the vibration from the road surface.

In 2013, Mak et al.249 added a bump stop to the cantilever energy harvester to pro-
tect the structure from large deformations (Figure 24A). A theoretical model was
developed to estimate the dynamic responses of cantilever energy harvesters under
different conditions. The model estimated that a bimorph with PZT elements of 11 3
5 3 0.18 mm3 is able to generate an RMS power of 178 mW with a bump stop and
298 mW without a bump stop, respectively, at a driving speed of 64.4 km hr1.
The study did not consider the static deformation of the cantilever caused by the
large centrifugal force that may prevent the cantilever from vibration. The configu-
ration of a piezoelectric cantilever with end stops has also been studied by Singh
et al.253 By measuring the vibration responses of a tire with different speed, load,
air pressure, and road surface condition, they claimed that the ideal working fre-
quency range is 60–80 Hz. They made a 25 3 5 3 0.4-mm3 bimorph prototype using
the high-energy density material PZT-ZNN. The prototype produced 15 mW power
under a harmonic excitation from a shaker at an RMS acceleration of 0.4 g and fre-
quency of 62.5 Hz. Furthermore, they considered the fact that the vibration response
of the cantilever bimorph is affected by the power management circuitry, and pro-
posed to use a buck-boost AC-DC converter to tune the resonant frequency of
the harvester. An artificial neutral network (ANN)-based closed-loop system was
developed to format an adaptive duty cycle for the buck-boost converter. It showed

38 Joule 2, 1–56, April 18, 2018

 Please cite this article in press as: Yang et al., High-Performance Piezoelectric Energy Harvesters and Their Applications, Joule (2018), https://
doi.org/10.1016/j.joule.2018.03.011

Figure 24. Cantilever Energy Harvesters Designed for TPMS

(A) Cantilever harvester with a bump stop. Reproduced from Mak et al. 249 with permission from SAGE Publications.
(B) MEMS cantilever harvester. Reproduced from Van Schaijk et al. 231 with permission from SPIE and Elfrink et al. 250 with permission from IEEE.
(C) Asymmetric air-gaped piezoelectric cantilever. 251 Reproduced from Zheng et al. 251 with permission from IOP Publishing Ltd.
(D) Stochastic resonance energy harvester for TPMS. Reproduced from Zhang et al.252 with permission from MDPI AG.

that the ANN method was capable of keeping the harvester at the maximum power
output condition over a broad frequency range. However, such a neural network
controller may consume more energy than it adds to the energy-harvesting system.

Elfrink et al.231,250 fabricated an MEMS cantilever harvester as shown in Figure 24B.

Under a harmonic excitation of 4.5 g and 1,011 Hz, the 3 3 3 3 1.7-mm3 AlN unim-
orph prototype generated a maximum power of 489 mW. When mounted on the
tread wall and tested on the road at 70 km hr1 driving speed, the prototype
generated a power of 30 mW on average.

Zhang et al.251 proposed an asymmetric air-gaped piezoelectric cantilever for har-

vesting energy from tires. As shown in Figure 24C, they deliberately introduced a
gap into the cantilever to increase the distance between the piezoelectric element
and the stress-neutral plane, and thus increased the voltage induced on the piezo-
electric element. The prototype was made of an 8 3 5 3 0.5-mm3 PZT-5A plate
and a 21.6-g proof mass. Due to the small size, it is impossible to match the resonant
frequency of the air-gaped cantilever with the vibrations of the tire. The power
response spectrum of the in-field tests showed two peaks: one at 11 Hz, correspond-
ing to the vibrations of the tire, one at 470 Hz, corresponding to the resonant fre-
quency of the cantilever. The prototype was not mounted on the tire-mounting
side of the rim but on the opposite side so that the voltage response could be re-
corded via a slip ring. At a driving speed of 50 mph, the prototype used 35 s to
charge a 32-mF capacitor to 8 V and the maximum power was 47 mW.

To expand the frequency bandwidth of the cantilever energy harvester, Sadeqi

et al.254 added a deformable support to the cantilever harvester. They considered
the vibrations caused by the gravitational field in a rotating wheel and ignored the
centrifugal effect in the analysis. The prototype was fixed on the rim, consisting of
a piezoelectric bimorph supported by four springs at each corner. The resonant fre-
quency of the energy harvester can be tuned by adjusting either the mass on the

Joule 2, 1–56, April 18, 2018 39

 Please cite this article in press as: Yang et al., High-Performance Piezoelectric Energy Harvesters and Their Applications, Joule (2018), https://
doi.org/10.1016/j.joule.2018.03.011

bimorph or the stiffness of the springs. In-lab tests using a DC motor showed that
two response peaks exist around 50 Hz and 70 Hz, and the maximum response
voltage is only about 4 V. The occurrence of the low voltage response is because
the soft support significantly reduced the vibration-induced stress in piezoelectric el-
ements. Although the idea of multiple resonance in an energy-harvesting system is
interesting, the inferior performance indicated that the method of introducing a soft
support to cantilever harvesters is not practical.

All the aforementioned studies arranged the piezoelectric beams parallel to the
tread wall surface. Zhu et al.255 set the cantilever harvester perpendicular to the
tread wall surface on the rim to harness vibration energy from the tangential accel-
eration variation. In this manner, the resonant frequency of the cantilever beam can
be tuned by the centrifugal force as the rolling speed varies. A prototype was made
of a 70 3 7 3 0.32-mm PZT-5H unimorph and a tip mass of 11 g, and tested on the
outside of a rim. It produced a peak power of 500 mW and an average power of
140 mW at a driving speed of 50 km hr1. This proposed design faces a serious instal-
lation problem. It is difficult to install a vertical piezo beam inside an inflated tire and
at the same time keep it from impacts while driving.

To improve energy-harvesting efficiency, Zhang et al.252 paid attention to the

coupling effect of the periodic acceleration change by gravity and the random vibra-
tion from the road surface. It is known that the dynamic responses of a bistable sys-
tem can be significantly enhanced with a certain probability by adding noise to a
weak periodic excitation, a phenomenon known as ‘‘stochastic resonance.’’ It has
been theoretically and experimentally proved that stochastic resonance can
improve the efficiency of BEHs.256,257 The authors studied the stochastic resonance
under the rotation conditions. As shown in Figure 24D, a piezoelectric cantilever
(22.9 3 10 3 0.1 mm3 PZT unimorph) was mounted outside of the tire at the center
of the wheel to rule out the centrifugal effect. A pair of magnets was fixed on the sta-
tionary frame and the movable tip end of the piezoelectric cantilever to induce the
required bistable nonlinearity. In-lab tests demonstrated that the excitation com-
bined with the periodic acceleration and the on-road noise signal has advantages
over the periodic acceleration excitation. On-road tests witnessed a maximum
average power of 5 mW, corresponding to a power density of 0.76 mW cm3 at
20 km hr1. However, it is difficult to know whether the stochastic resonance really
happened in the on-road tests. The tests were performed only in a low speed range
of 10–40 km hr1 on a smooth paved road. At a high speed of over 100 km hr1 or on
a rugged road, the nonlinear dynamics may behave differently. The basic idea un-
derlying the performance improvement from the stochastic resonance is that the
noise signal facilitates overcoming the potential barrier and activating the intra-
well dynamics.258 Therefore, the occurrence of the stochastic resonance requires a
proper noise amplitude relative to the bistable system and the main oscillation
source. In this study, the amplitude of the noise signal is comparable with that
of the main periodic acceleration variation. However, such a noise and excitation
condition is not guaranteed all the time in practice. Embedding nonlinear energy
harvesters into TPMS faces serious issues with regard to the system robustness.

In addition to cantilevered structures, the fixed-fixed boundary condition has also

been explored. Marian259 presented a clamped-clamped unimorph with a proof
mass fixed at the center. The highest power output of 40 mW was achieved at
80 km hr1. Similarly, Jousimaa et al.260 fixed a piezoelectric disc on a base with a
mass attached at the center, as shown in Figure 25A. It mainly utilizes the periodic
radial acceleration variation of a rolling tire due to the tire radius change in each

40 Joule 2, 1–56, April 18, 2018

 Please cite this article in press as: Yang et al., High-Performance Piezoelectric Energy Harvesters and Their Applications, Joule (2018), https://
doi.org/10.1016/j.joule.2018.03.011

Figure 25. Different Structures for Harvesting Vibration Energy from Wheels
(A) Fixed piezo disc for harvesting energy from the periodic radial acceleration variation. Reproduced from Jousimaa et al. 260 with permission from IEEE.
(B) Seesaw-structured energy harvester. Reproduced from Wu et al. 261 with permission from IEEE.
(C) Harvesting energy from the impact of a ball on a spinning wheel. Reproduced from Manla et al. 262 with permission from IEEE.
(D) The interaction of a vibrating magnet and cantilever PEHs. Reproduced from Tang et al. 263 with permission from IEEE.
(E) Offset pendulum mechanism for rotational energy harvesting. 264 Reproduced from Roundy and Tola 264 with permission from IOP Publishing Ltd.

rotation, as well as the periodic and random vibrations from the tire-road interaction.
The prototype was made of a PZT unimorph disc of diameter 32.77 mm (Thunder
TH-5C) and assembled with a circuit and a housing. The whole assembly is
35 3 35 3 25 mm3 and weighs 65 g. The prototype produced 88 mW power on
average at the excitation speed of 60 km hr1. One interesting test reported here
is an endurance test. The prototype mounted inside a wheel survived for 10 min at
60 km hr1 and a load of 2 kN on a tire-testing rig. The short lifetime cannot meet
the requirement of the TPMS application.

In 2014, Wu et al.261 proposed a novel seesaw-structured energy harvester for

TPMS. As shown in Figure 25B, the proposed structure was constructed with a

Joule 2, 1–56, April 18, 2018 41

 Please cite this article in press as: Yang et al., High-Performance Piezoelectric Energy Harvesters and Their Applications, Joule (2018), https://
doi.org/10.1016/j.joule.2018.03.011

pair of PVDF cantilevers, a seesaw beam, and two magnets at each end. The struc-
ture was fixed on a rolling wheel rim, and coupled once per round with a permanent
magnet fixed on the stationary brake caliper. The swing of the seesaw structure
driven by the magnetic repulsive force from the stationary magnet imposes an
impulsive force on the PVDF cantilevers. It was claimed that the centrifugal force
has no influence on the seesaw structure, and the harvester can work efficiently inde-
pendent of the rotating speed. Excited by a motor rotating at 750 rpm (90 km hr1),
a prototype with a 25 3 12 3 0.1-mm3 PVDF generated a power of 5.6 mW. This
seesaw concept is academically appealing but may encounter cost, assembly,
compatibility, and reliability issues in practice due to the complex configuration
and additional magnet on the brake caliper.

In 2009, Manla et al.262 proposed the concept of using the impact of a ball in a tube
to harvest energy from a spinning wheel. Figure 25C shows the schematics. A small
tube can be fixed on the wheel rim with both ends covered by piezoelectric plates
and a free steel ball moving inside. The authors summarized two ways to install
the harvester on a rotating wheel. In the first way, the tube is fixed on the edge of
the rim in the tangential direction. The movement of the ball is caused by gravity
and the component of the centripetal force. It was argued that this layout is not reli-
able because the ball will be forced to one end of the tube by the centrifugal effect
when the rotating speed is high. In the second way, the tube is connected with the
spinning wheel in a way that it can always stay in the horizontal direction. The acting
force in this layout is the component of the centrifugal force. This study selected the
second method and designed a complex mechanical system consists of one syn-
chronous toothed belt and two pulleys to keep the harvester structure always in
the horizontal direction. In experiments, a 2-cm3 harvester was hinged on a disc of
radius 5 cm, and generated on average 4 mW power at 800 rpm. There were no
on-road tests reported in this paper, and it is impractical to introduce an extra
belt and pulley system into a wheel to power TPMS.

Tang et al.263 also explored the tangential acceleration variation of a rotating wheel
using a similar tube structure. As shown in Figure 25D, they added a pair of small
piezoelectric cantilever beams outside around the centerline of the tube and mag-
nets at the free ends of the cantilevers. A central sliding magnet is enclosed in a
tube with two magnets at each end. While the central magnet slides back and forth
in the tube, the outside cantilevers are bent by the magnetic force and then vibrate
at their own resonant frequencies after being released. In this manner, the harvester
becomes insensitive to the speed change and capable of upgrading the low-fre-
quency vibration excitations (1–20 Hz) in wheels to high-frequency resonant
vibration of the piezoelectric cantilevers. A small-size prototype with 9 3 1.3 3
0.14-mm3 PVDF-copper cantilevers generated a power of 10 mW under 1 g excita-
tion from a shaker over a frequency range of 10–22 Hz. Its performance in a rolling
wheel is unknown, but should be much lower than that excited by a shaker.231,250

In 2014, Roundy and Tola264 also developed a tube-type energy harvester, and exploited
the nonlinear dynamic characteristics of an offset pendulum to increase the working band-
width. The authors did not really construct an offset pendulum but used a small-size
curved track with a radius smaller than that of the mounted rim to produce the nonlinear
behaviors of an offset pendulum. As shown in Figure 25E, the harvester is composed of a
curved track, a pair of piezoelectric beams along the curved track, a steel ball that rolls in
the curved track, two small balls held in conical holes, and two spring end stops. The small
balls are pushed out periodically as the large steel ball rolls back and forth in the curved
track, and causes deflection of the piezoelectric cantilevers. The nonlinear response of the

42 Joule 2, 1–56, April 18, 2018

 Please cite this article in press as: Yang et al., High-Performance Piezoelectric Energy Harvesters and Their Applications, Joule (2018), https://
doi.org/10.1016/j.joule.2018.03.011

Figure 26. Harvesting Energy from the Tire Deformation

(A) PZT strain-based energy harvesting for TPMS. Reproduced from Makki et al. 265 with permission from Springer-Verlag.
(B) PZT-polymer composite energy harvester for TPMS. PZT short fibers (light) aligned between gold electrodes (dark). Reproduced from Van den Ende
et al. 268 with permission from IOP Publishing Ltd.

steel ball is attributed to gravity, the centripetal acceleration, and the restoring force
from the piezoelectric beams and the spring ends. A model was developed and pre-
dicted a power output of 10 mW at 10 mph. In the on-road tests, a prototype was assem-
bled into a TPMS module that consumes 6.25 mW on average to make one transmission
per minute. The on-road tests were performed from 10 to 155 km hr1, and the devel-
oped harvester could achieve more than one RF transmission per minute in the speed
range.

Strain Energy. Makki and Pop-Iliev265–267 performed a series of tests on harvesting

strain energy using PZT. They explored two methods, one by attaching PZT on the
tread wall and the other by inserting PZT between the tire bead and the rim. The
sidewall deforms periodically above the contact patch due to the vehicle weight
and outward bulge. The sidewall deformation exerts an additional force to the tire
bead section. To take advantage of this periodic force, the authors inserted a thin
PZT element between the rim and the tire bead, as shown in Figure 26A. At
80 rpm, this method took 180 s to charge a 22-mF capacitor up to 10 V. With a match-
ing resistance of 67 kU, a maximum power of 70 mW was achieved.

For the tread wall method, the authors glued four PZT buzzer elements of diameter
25 mm on the inner liner of the tread wall of a tire, as shown in Figure 26A. When the
tire rotated at 80 rpm (10 km hr1), the harvester produced a power of 6.5 mW with a
matching resistance of 42 kU, and took 6.76 s to charge a 22-mF capacitor to 10 V.
The averaged power to the capacitor is 163 mW. It is believed that the output power
to the capacitor will be much higher at the impedance matching condition. A TPMS
circuit was designed to test the developed harvester. The circuit on average oper-
ates at 3.0 V and 4.85 mA. In experiments, the tire-mounted PZT harvester suc-
ceeded in powering the constructed TPMS circuit for sensing and transmitting
data once every 2.5 s at 60 km hr1 and every 1.5 s at 100 km hr1.

The PZT attached on the tread wall can be damaged very easily due to overload or
unexpected shocks. In contrast, the method of inserting PZT between the rim and
the tire bead is more robust. However, this method may cause issues such as air
leakage and tire imbalance. Also, the bead in a tire is strengthened by steel wires
and, with high air pressure inside, the deformation in a small local area is very trivial.
Nevertheless, this is a very intriguing achievement as the transmission speed is com-
parable with that of the TPMS powered by batteries on the market.

Joule 2, 1–56, April 18, 2018 43

 Please cite this article in press as: Yang et al., High-Performance Piezoelectric Energy Harvesters and Their Applications, Joule (2018), https://
doi.org/10.1016/j.joule.2018.03.011

It has been recognized that PZT is too brittle to be used in tires directly unless
elaborate stress transfer mechanisms are developed.268,269 Piezoelectric PVDF
polymers are flexible, but exhibit a very weak piezoelectric effect. In 2012, Ende
et al.268 fabricated a PZT-polymer composite to reach a tradeoff between the
high piezoelectric effect of PZT and the excellent compliance of PVDF (Figure 26B).
The PZT-polymer composite was fabricated by applying a 0.3-mm thick layer of the
mixture of PU and PZT-5A powder (volume ratio 9:1) on a gold electrode PET film
substrate. The dielectrophoretic (DEP) process270 was performed during curing of
the polyurethane to form short PZT fibers. An interdigitated electrode (IDE) was
used so that the cantilever could work in the more effective d33 mode. By attaching
a PZT-polymer patch of 60 3 10 mm2 to a tire with an epoxy substrate of 0.5 mm
thickness, 1.37 mW mm3 was achieved. It was claimed that although the
performance of the PZT-polymer composite is not as good as that of virgin MFC
(micro-fiber composite) and PVDF, the composite stays relatively stable under
high-temperature conditions. Lee and Choi271 also tested a flexible piezo fiber
composite patch on the inner liner of a tire. A 60 3 10 3 0.3-mm patch generated
380.2 mJ energy per revolution under a 500-kg load and at 60 km hr1 the power
density was reported to be 1.37 mW mm3.

Summary. In summary, the power output of the developed energy harvesters

ranges from10 mW to 1 mW in on-road tests. The harvester prototypes have been
able to power TPMS for more than one RF transmission per minute. PZT, PVDF,
and piezoelectric composites have been utilized in different forms to harvest energy
from vibrations in tangential and radial directions, and strain energy on the tread
wall. The strain energy on the sidewall has seldom been explored due to the
extremely large deformation. The strain-based energy-harvesting method has the
advantages of lighter weight and higher power output than the vibration-based
counterpart, but is more susceptible to fatigue, damage, and shocks. Furthermore,
attaching piezoelectric patches on a soft tire causes trouble during maintenance and
also alters the elastic characteristics of the tire, which may lead to some unknown
problems. In addition to reliability issues, current energy harvesters also face prob-
lems including low efficiency, poor environmental adaptability, and short lifespan.

Bridges and Buildings

One of the main applications for the wireless sensor networks (WSNs) are civil infra-
structure systems such as bridges and buildings. Structure health monitoring sys-
tems on tall buildings and long-span bridges are essential for maintenance purposes
(e.g., detection of early-stage degradation) and for alarm warnings (e.g., notification
of earthquakes). Some systems have been successfully applied in the field, such as
the wireless acceleration sensors to monitor building conditions,272 and acoustic
sensors for structural health prognosis in bridges.273 The WSNs are also the hard-
ware base for the development of smart buildings274 and smart homes.275 In such
systems, a myriad of wireless sensors are deployed at different locations of mega
civil infrastructures. It is expected that such systems work continuously for years,
but batteries in the wireless sensor nodes hold limited energy and require frequent
recharge and replacement. The energy-harvesting technology supplies a remedy to
the battery issue, and can significantly reduce the economic and time cost associ-
ated with the battery replacement task.

It is technically feasible to power wireless sensors using energy harvesters in civil

structures.276,277 Tall buildings and long-span bridges are susceptible to a variety
of dynamic excitations such as wind, earthquakes, traffic, and human activities, which
can be used by energy harvesters to generate electricity. Vibration excitations inside

44 Joule 2, 1–56, April 18, 2018

 Please cite this article in press as: Yang et al., High-Performance Piezoelectric Energy Harvesters and Their Applications, Joule (2018), https://
doi.org/10.1016/j.joule.2018.03.011

a typical English house (kitchen and utility room) were reported278 to be around
100 Hz, with an acceleration amplitude of 0.02–3 m s2. High-rise buildings ranging
from 24 to 257 m in the San Francisco Bay area279 showed an acceleration amplitude
of 0.15–0.38 g on the roof and 0.05–0.12 g on the ground floor, and vibration
frequency of less than 1 Hz. For bridges, the vibration condition varies significantly,
depending on weather, bridge construction, location, and traffic flow. Peigney and
Siegert280 presented a bridge vibration of 0.03–0.3 m s2 and 4 Hz. Khan and
Ahmad281 presented bridge vibrations in the range of 1–40 Hz and 0.01–0.38 g.
Overall, vibrations in bridges and buildings are characterized by low frequency
and small amplitude.

Compared with other applications such as shoes and pacemakers, there is almost no
space limitation for bridges and buildings. Therefore, relatively large-size harvesters
can be designed to meet the power requirements from sensors. Here we mainly
discuss small-scale energy harvesters for use in low-power electronic monitoring de-
vices via the piezoelectric effect. Readers who are interested in the electromagnetic
method are referred to other work.282–293

Most research is based on the conventional cantilever energy harvester. In 2011,

Erturk294 presented a theoretical analysis on energy harvesters in civil infrastructures
using two types of excitations: vibrations and surface strain fluctuations. In the first
scenario, the author considered a conventional cantilever energy harvester fixed
on an arbitrary position of a slender bridge. The vibration response of a slender
bridge caused by moving vehicles acts as the excitation to the cantilever harvester.
In the second scenario, the dynamic surface strain variations were utilized by attach-
ing a thin piezoelectric patch on a bridge region. Detailed theoretical analysis was
presented on the two scenarios under generalized, harmonic, and white noise exci-
tations. Furthermore, a case study using the developed distributed-parameter
model was presented to evaluate the power output of a piezoceramic patch excited
by real bridge strain fluctuations. A rosette strain gauge was used to measure the
dynamic response of a steel multi-girder bridge (span length: 33 m). The recorded
data after a fast Fourier transform analysis indicated that the major harmonic strain
component lies at 22.6 Hz and the strain fluctuation amplitude is about 20 micro
strain. Under such a harmonic strain fluctuation, the developed model estimated
that a maximum power output of 0.13 mW can be achieved by a 30 3 30 3
0.2-mm3 PZT-5A patch.

Many models have been developed to analyze the bridge vibration response under
different conditions and the power responses of the cantilever energy harvesters. Ali
et al.295 proposed a linear SDOF model for a highway bridge with a moving point
load. Zhang et al.296 presented a finite element model of a concrete slab-on-girder
bridge for simulating the harvester performance with one passing vehicle and a
continuous vehicle flow. Xie et al.297 presented a distributed model similar to that
in the work of Erturk294 for a cantilever energy harvester on tall buildings. Cahill
and Jaksic298 developed an SDOF model and via the model discussed the vibration
responses of damaged and undamaged bridges affected by the road surface rough-
ness, vehicle weight, and speed. Cahill et al.299 also discussed the dynamic response
of a bridge traversed by trains via a finite element model. Karimi et al.300 proposed
a distributed-parameter model whereby passing vehicles on a bridge were consid-
ered as concentrated point mass and distributed mass. Bhaskaran et al.301 numeri-
cally analyzed an array of cantilever energy harvesters with different resonant fre-
quencies via the finite element method in order to achieve broadband energy
harvesting on bridges.

Joule 2, 1–56, April 18, 2018 45

 Please cite this article in press as: Yang et al., High-Performance Piezoelectric Energy Harvesters and Their Applications, Joule (2018), https://
doi.org/10.1016/j.joule.2018.03.011

Figure 27. Energy Harvesters Developed for Bridges and Buildings

(A) Cantilever energy harvester under bridges. Reproduced from Lee et al. 280 with permission from IOP Publishing Ltd.
(B) A segmented-type piezoelectric energy harvester. Reproduced from Lee et al.303 with permission from IOP Publishing Ltd.
(C) Cantilever harvester with an axial preload. Reproduced from Zhang et al. 304 with permission from American Society of Civil Engineers.
(D) A multi-impact energy harvester for low-frequency vibration applications. Reproduced from Zhang et al. 155 with permission from IOP Publishing Ltd.

To improve the energy output, Maruccio et al.302 proposed a cantilever energy

harvester made of electrospinning PVDF nanofibers. They first analyzed the vibra-
tion recording of a cable-stayed bridge and identified the excitation source with
an RMS acceleration of 5.87 3 104–0.0108 g and frequency of 0.92–6.39 Hz. A
cantilever bimorph was then designed with 20- to 50-mm thick PVDF sheets fabri-
cated by means of electrospinning PVDF nanofibers. The authors finally developed
a finite element model and estimated that RMS voltage of 1.7–15.6 V can be gener-
ated by the harvester at different locations of the cable-stayed bridge. No proto-
types and experiments were presented in this study. It is difficult to examine whether
the proposed cantilever energy harvester using PVDF nanofibers has a better perfor-
mance than the commonly used PZT counterpart as the piezoelectric coefficients of
PDVF are much smaller than those of piezoceramics.

Cantilevered energy harvesters have also been tested on bridges. In 2013, Peigney
and Siegert280 tested a cantilever energy harvester in situ on a bridge, as shown in
Figure 27A. Bridge vibrations are deemed to be mainly induced by the passing
traffic, and they measured the vibration conditions of a pre-stressed concrete
expressway bridge located at a heavily trafficked itinerary (8,000 heavy trucks
per day). The acceleration response includes successive short-time pulses, and the
RMS value at different locations ranges from 0.03 m s2 to 0.3 m s2. The spectral
analysis of the recorded signals reveals that the main frequency components
are around 4 Hz and 14.5 Hz. A cantilever harvester prototype was made of two
PZT bimorphs (46 3 33 3 0.76 mm [Mide QP20W]), a steel plate of 220 3 40 3
0.8 mm3, and a proof mass of 12 g, as shown in Figure 6. The prototype was mounted
at the fixture of a water pipe under the bridge, and generated a maximum power
of 30 mW.

46 Joule 2, 1–56, April 18, 2018

 Please cite this article in press as: Yang et al., High-Performance Piezoelectric Energy Harvesters and Their Applications, Joule (2018), https://
doi.org/10.1016/j.joule.2018.03.011

In 2009, Lee et al.303 designed a segmented-type PEH for capture energy from the
random vibration of the HVAC (heating, ventilation, air conditioning) system in a
building (Figure 27B). The power spectrum of the measured data from an HVAC sys-
tem indicated that the vibration is randomly distributed between 0 Hz and 150 Hz
with two peaks about 0.2 g at 39 Hz and 0.1 g at 110 Hz, respectively. To make
full use of the random vibration, the harvester design needs to be tailored to be
able to work within multiple resonant frequencies. Traditional cantilever bimorph en-
ergy harvesters can only work at one fixed frequency around the first resonant mode
where the mechanical strain on each side of the bimorph is in the same sign. The au-
thors proposed to take advantage of the second resonant mode to capture more en-
ergy. However, in the second resonant mode, an inflection point exists in the beam
where the curvature sign of deformation alters, resulting in charge cancellation. To
preclude the charge cancellation effect, they separated the PZT plate into two parts
near the inflection point of mechanical strain with a proof mass. As shown in Fig-
ure 27B, the proposed structure has two segments and two proof masses. The
segment beam near the free end has been optimized to be a slender trapezoid
via the finite element method. A prototype was fabricated with PZT-5A plates of
63.5 3 31.8 3 0.19 mm3, which generated a peak open-circuit voltage of 10 V
and power of 1.8 mW under the aforementioned excitation. The prototype was
further tested on an HVAC system and succeeded in powering an analog tempera-
ture sensor node that requires 15 mW for continuous temperature monitoring.

Considering that the vibration frequency of civil structures is usually very low, in 2012
Rhimi and Lajnef304 added a compressive axial preload on a cantilever PEH to reduce
resonant frequency. As shown in Figure 27C, a prototype was made out of a PZT-5H
bimorph with a dimension of 40 3 10 3 0.5 mm3 and a pair of coil springs. The au-
thors did not test the prototype with a bridge vibration excitation, but developed an
analytical model to predict the harvester performance. The model estimated that,
under an earthquake vibration recorded on a bridge, the proposed prototype could
generate a maximum power of 47.23 mW. The concept of tuning the resonant fre-
quency using axial preload has been widely studied and validated theoretically
and experimentally.305 One should note that the resonant frequency is very sensitive
to the compressive load, which may cause an unsatisfactory reliability to such pre-
loaded cantilever harvesters in real applications.

In 2011, McEvoy288 proposed a concept of cantilever-type impact energy harvest-

ing with the aim of harnessing the low-frequency vibrations on bridges. They de-
signed a stopper plate directly beneath the tip end of the cantilever energy
harvester so that impulsive strikes occurred in each vibration cycle. The impact
causes a pair of small bimorphs attached at the tip end to vibrate at their high reso-
nant frequencies. The concept is interesting, but no experiments were conducted
on a bridge.

The impact idea has been further developed to address the low-frequency vibrations
in civil structures. In 2013, Zhang et al.155 proposed a multi-impact energy harvester
for low-frequency vibration applications, such as bridges. Figure 27D shows the
schematic of the proposed structure. It is composed of a metal frame, two vertical
piezoelectric cantilever beams, and a mass-spring system. The two cantilever beams
were manufactured with a triangle bulge at each tip and installed symmetrically on
each side of the hung mass on the metal frame. The hung mass block is specially de-
signed with a series of small rollers that repeatedly impact the bulges on the piezo-
electric cantilevers. In this manner, the low-frequency vibration (2.71 Hz) of the
spring-mass system triggers vibrations of the cantilevers around the resonant

Joule 2, 1–56, April 18, 2018 47

 Please cite this article in press as: Yang et al., High-Performance Piezoelectric Energy Harvesters and Their Applications, Joule (2018), https://
doi.org/10.1016/j.joule.2018.03.011

frequency (120 Hz). It is reported that a prototype, made of 50 3 30 3 0.2-mm3 PZT

plates, generated a maximum power of 7.7 mW under and sinusoidal excitation of
0.29 g, 2.71 Hz (the resonant frequency of the mass-spring system). The study also
presented a finite element model to calculate the vibration signal input to the energy
harvester from a short span slab-on-girder bridge excited by passing vehicles. Under
the simulated bridge excitation, the prototype delivered a maximum power of
2.8 mW. No in-field tests were performed in this study, and the paper does not
disclose how the power is calculated. It is known that the voltage response of such
impact harvesters decays exponentially in a short time. Consequently, a significant
percentage of voltage responses will be too low to be used by the following circuits.
Therefore, the proposed energy harvester will not be able to generate much power if
applied on a bridge. Nevertheless, the concept of impact is interesting and effective
for frequency up-conversion.

In 2011, Kim et al.306 experimentally evaluated the piezoelectric effect under

various loading conditions on bridges. PZT patches were attached on a steel
beam-slab type bridge specimen, and tested under various structural excitations
corresponding to different traffic conditions. The experimental results indicated
that energy outputs were mainly affected by the peak strain and strain increasing
rate in the piezoelectric patches, which are determined by the speed and mass of
moving vehicles.

Xie et al.307,308 successively proposed two designs for harvesting energy from high-
rise buildings. The designs work like tuned mass dampers, converting the building
vibration to vibration of energy harvesters first, then to electricity via the piezoelec-
tric effect. They only presented theoretical analysis on the proposed concepts and
estimated that thousands of watts of power can be generated. It is reported that
the electromagnetic method is more appropriate for use in such large-scale
applications.233

Overall, there is an increasing interest in applying PEHs in bridges and buildings. The
power output of the developed PEHs ranges from several microwatts to several mil-
liwatts. Most research focuses on the theoretical analysis of cantilever energy har-
vesters in bridges and buildings under different conditions. There are a few studies
that report experimental data, and thus more in situ experiments on bridges and
buildings are expected. Furthermore, although there is no strict limitation to size,
new methods to scale down energy harvesters are always desirable so that har-
vesters can be embedded inside bridges and buildings.

Summary
This paper is devoted to reviewing state-of-the-art research aimed at high-perfor-
mance piezoelectric energy harvesting. We have comprehensively analyzed
different designs, nonlinear methods, optimization techniques, and materials that
were explored to achieve high power output and broad frequency bandwidth. A sys-
tematic performance comparison has been conducted on recently proposed repre-
sentative energy harvesters. We found that it is not appropriate to compare different
designs by one figure of merit due to the complexity in dynamics, structures, and
electromechanical coupling of energy-harvesting systems. Therefore, we propose
to evaluate the performance of different energy harvesters using a set of metrics use-
ful to end users, instead of one universal metric. We think performance evaluations
and optimization are more appropriately conducted within specific application
conditions. Here we have identified four promising applications: shoes, artificial
pacemakers, TPMSs, and bridge and building monitoring systems. The excitation

48 Joule 2, 1–56, April 18, 2018

 Please cite this article in press as: Yang et al., High-Performance Piezoelectric Energy Harvesters and Their Applications, Joule (2018), https://
doi.org/10.1016/j.joule.2018.03.011

characteristics of each application are analyzed and corresponding energy-harvest-

ing methods discussed.

Although energy-harvesting technology has been intensively studied for decades,

there is still a considerable gap between the achieved performance and the ex-
pected performance. Most energy harvesters are developed for a general purpose,
and tested under simplified harmonic excitations. This approach provides technol-
ogy that is far from ready to be used in the potential applications. Even though
some application-oriented harvesters have been tested in situ, their reliability, sta-
bility, and compatibility have not been well examined. More research is expected
to deal with these challenges. Likewise, more system-level investigations are war-
ranted, whereby energy harvesters are integrated with power conditioning circuits,
energy storage elements, sensors, and control circuits. Such research will facilitate in
turning decades of research efforts on energy harvesting into tangible benefits in our
daily life.

Overall, we have witnessed significant progress in energy-harvesting technology in

the last decade. It continues to approach the goal of self-powered autonomous op-
erations of wearable electronics, medical devices, automotive sensors, and wireless
sensor monitoring systems.

AUTHOR CONTRIBUTIONS
Z.Y. wrote Sections 1, 2, 4, 5, and 6. S.Z. wrote Section 3. Z.Y. and S.Z. prepared fig-
ures. D.I. and J.Z. discussed and revised the manuscript. All authors contributed to
the intellectual content.

REFERENCES
1. Bowen, C., Kim, H., Weaver, P., and Dunn, S. 8. Kim, S.-G., Priya, S., and Kanno, I. (2012). interfaces and issues in small scale
(2014). Piezoelectric and ferroelectric Piezoelectric MEMS for energy harvesting. implementation. Micromachines 2, 274–294.
materials and structures for energy harvesting MRS Bull. 37, 1039–1050.
applications. Energy Environ. Sci. 7, 25–44. 16. Dicken, J., Mitcheson, P.D., Stoianov, I., and
9. Wang, Z.L., and Wu, W. (2012). Yeatman, E.M. (2012). Power-extraction
2. Khan, A., Abas, Z., Kim, H.S., and Oh, I.-K. Nanotechnology-enabled energy harvesting circuits for piezoelectric energy harvesters in
(2016). Piezoelectric thin films: an integrated for self-powered micro-/nanosystems. miniature and low-power applications. IEEE
review of transducers and energy harvesting. Angew. Chem. Int. Ed. 51, 11700–11721. Trans. Power Electron. 27, 4514–4529.
Smart Mater. Struct. 25, 053002.
10. Radousky, H.B., and Liang, H. (2012). Energy
harvesting: an integrated view of materials, 17. Szarka, G.D., Stark, B.H., and Burrow, S.G.
3. Ramadan, K.S., Sameoto, D., and Evoy, S. devices and applications. Nanotechnology (2012). Review of power conditioning for
(2014). A review of piezoelectric polymers as 23, 502001. kinetic energy harvesting systems. IEEE Trans.
functional materials for electromechanical Power Electron. 27, 803–815.
transducers. Smart Mater. Struct. 23, 033001. 11. Fan, F.R., Tang, W., and Wang, Z.L. (2016).
Flexible nanogenerators for energy 18. Dagdeviren, C., Li, Z., and Wang, Z.L. (2017).
4. Wang, Z.L. (2012). From nanogenerators to harvesting and self-powered electronics. Adv. Energy harvesting from the animal/human
piezotronics—a decade-long study of ZnO Mater. 28, 4283–4305. body for self-powered electronics. Annu. Rev.
nanostructures. MRS Bull. 37, 814–827. Biomed. Eng. 19, 85–108.
12. Daqaq, M.F., Masana, R., Erturk, A., and
5. Wang, Z.L., Zhu, G., Yang, Y., Wang, S., and Quinn, D.D. (2014). On the role of 19. Mitcheson, P.D., Yeatman, E.M., Rao, G.K.,
Pan, C. (2012). Progress in nanogenerators for nonlinearities in vibratory energy harvesting: a Holmes, A.S., and Green, T.C. (2008). Energy
portable electronics. Mater. Today 15, critical review and discussion. Appl. Mech. harvesting from human and machine motion
532–543. Rev. 66, 040801. for wireless electronic devices. Proceedings of
the IEEE, 96(9): p. 1457–1486.
13. Harne, R.L., and Wang, K.W. (2013). A review
6. Cook-Chennault, K.A., Thambi, N., and
of the recent research on vibration energy
Sastry, A.M. (2008). Powering MEMS portable 20. Siddique, A.R.M., Mahmud, S., and Van Heyst,
harvesting via bistable systems. Smart Mater.
devices—a review of non-regenerative and B. (2015). A comprehensive review on
Struct. 22, 023001.
regenerative power supply systems with vibration based micro power generators
special emphasis on piezoelectric energy 14. Tang, L., Yang, Y., and Soh, C.K. (2010). using electromagnetic and piezoelectric
harvesting systems. Smart Mater. Struct. 17, Toward broadband vibration-based energy transducer mechanisms. Energ. Convers.
043001. harvesting. J. Intell. Mater. Syst. Struct. 21, Manag. 106, 728–747.
1867–1897.
7. Briscoe, J., and Dunn, S. (2015). Piezoelectric 21. Hudak, N.S., and Amatucci, G.G. (2008).
nanogenerators—a review of nanostructured 15. Guyomar, D., and Lallart, M. (2011). Recent Small-scale energy harvesting through
piezoelectric energy harvesters. Nano Energy progress in piezoelectric conversion and thermoelectric, vibration, and radiofrequency
14, 15–29. energy harvesting using nonlinear electronic power conversion. J. Appl. Physiol. 103, 5.

Joule 2, 1–56, April 18, 2018 49

 Please cite this article in press as: Yang et al., High-Performance Piezoelectric Energy Harvesters and Their Applications, Joule (2018), https://
doi.org/10.1016/j.joule.2018.03.011

22. Cepnik, C., Lausecker, R., and Wallrabe, U. 39. Yang, Z., and Zu, J. (2014). High-efficiency broadband piezoelectric energy generator.
(2013). Review on electrodynamic energy compressive-mode energy harvester Nonlinear Dyn. 80, 1705–1719.
harvesters—a classification approach. enhanced by a multi-stage force amplification
Micromachines 4, 168–196. mechanism. Energ. Convers. Manag. 88, 55. Stanton, S.C., Mann, B.P., and Owens, B.A.
829–833. (2012). Melnikov theoretic methods for
23. Selvan, K.V., and Ali, M.S.M. (2016). Micro- characterizing the dynamics of the bistable
scale energy harvesting devices: review of 40. Yang, Z., Zhu, Y., and Zu, J. (2015). Theoretical piezoelectric inertial generator in complex
methodological performances in the last and experimental investigation of a nonlinear spectral environments. Phys. Nonlinear
decade. Renew. Sustain. Energ. Rev. 54, compressive-mode energy harvester with Phenom. 241, 711–720.
1035–1047. high power output under weak excitations.
Smart Mater. Struct. 24, 025028. 56. Oumbé Tékam, G., Kitio Kwuimy, C., and
24. Shaikh, F.K., and Zeadally, S. (2016). Energy Woafo, P. (2015). Analysis of tristable energy
harvesting in wireless sensor networks: a 41. Stanton, S.C., McGehee, C.C., and Mann, B.P. harvesting system having fractional order
comprehensive review. Renew. Sustain. (2009). Reversible hysteresis for broadband viscoelastic material. Chaos 25, 013112.
Energ. Rev. 55, 1041–1054. magnetopiezoelastic energy harvesting.
Appl. Phys. Lett. 95, 174103. 57. Litak, G., Friswell, M., and Adhikari, S. (2010).
25. Caliò, R., Rongala, U.B., Camboni, D., Milazzo, Magnetopiezoelastic energy harvesting
M., Stefanini, C., de Petris, G., and Oddo, 42. Yang, Z., Zu, J., and Xu, Z. (2016). Reversible driven by random excitations. Appl. Phys.
C.M. (2014). Piezoelectric energy harvesting nonlinear energy harvester tuned by tilting Lett. 96, 214103.
solutions. Sensors (Basel) 14, 4755–4790. and enhanced by nonlinear circuits. IEEE/
ASME Trans. Mechatronics 21, 2174–2184. 58. Litak, G., Borowiec, M., Friswell, M.I., and
26. Uchino, K. (2010). The development of Adhikari, S. (2011). Energy harvesting in a
piezoelectric materials and the new 43. Erturk, A., and Inman, D. (2011). Broadband magnetopiezoelastic system driven by
perspective. In Advanced Piezoelectric piezoelectric power generation on high- random excitations with uniform and
Materials, K. Uchino, ed. (Woodhead energy orbits of the bistable Duffing oscillator Gaussian distributions. J. Theor. Appl. Mech.
Publishing), pp. 1–43. with electromechanical coupling. J. Sound 49, 757–764.
Vib. 330, 2339–2353.
27. Zhang, S., Li, F., Jiang, X., Kim, J., Luo, J., and 59. Cottone, F., Gammaitoni, L., Vocca, H.,
Geng, X. (2015). Advantages and challenges 44. Zhou, S., Cao, J., Inman, D.J., Lin, J., Liu, S., Ferrari, M., and Ferrari, V. (2012). Piezoelectric
of relaxor-PbTiO3 ferroelectric crystals for and Wang, Z. (2014). Broadband tristable buckled beams for random vibration energy
electroacoustic transducers–A review. Prog. energy harvester: modeling and experiment harvesting. Smart Mater. Struct. 21, 035021.
Mater. Sci. 68, 1–66. verification. Appl. Energy 133, 33–39.
60. Daqaq, M.F. (2011). Transduction of a bistable
28. Yang, Z., and Zu, J. (2016). Comparison of 45. Zhou, S., Cao, J., Lin, J., and Wang, Z. (2014). inductive generator driven by white and
PZN-PT, PMN-PT single crystals and PZT Exploitation of a tristable nonlinear oscillator exponentially correlated Gaussian noise.
ceramic for vibration energy harvesting. for improving broadband vibration energy J. Sound Vib. 330, 2554–2564.
Energ. Convers. Manag. 122, 321–329. harvesting. Eur. Phys. J. Appl. Phys. 67,
https://doi.org/10.1051/epjap/2014140190. 61. He, Q., and Daqaq, M.F. (2014). Influence of
29. Yuhuan, X. (1991). Ferroelectric Materials and potential function asymmetries on the
Their Applications (North-Holland). 46. Kim, P., and Seok, J. (2014). A multi-stable performance of nonlinear energy harvesters
energy harvester: dynamic modeling and under white noise. J. Sound Vib. 333, 3479–
30. Pellegrini, S.P., Tolou, N., Schenk, M., and
Herder, J.L. (2013). Bistable vibration energy bifurcation analysis. J. Sound Vib. 333, 5525– 3489.
harvesters: a review. J. Intell. Mater. Syst. 5547.
62. Zhao, S., and Erturk, A. (2013). On the
Struct. 24, 1303–1312. 47. Kim, P., and Seok, J. (2015). Dynamic and stochastic excitation of monostable and
energetic characteristics of a tri-stable bistable electroelastic power generators:
31. Wei, C., and Jing, X. (2017). A comprehensive
magnetopiezoelastic energy harvester. Mech. relative advantages and tradeoffs in a physical
review on vibration energy harvesting:
Mach. Theor. 94, 41–63. system. Appl. Phys. Lett. 102, 103902.
Modelling and realization. Renew. Sustain.
Energ. Rev. 74, 1–18. 63. Haitao, L., Weiyang, Q., Chunbo, L.,
48. Zhou, S., Cao, J., and Lin, J. (2016). Theoretical
32. Gammaitoni, L., Hänggi, P., Jung, P., and analysis and experimental verification for Wangzheng, D., and Zhiyong, Z. (2015).
Marchesoni, F. (1998). Stochastic resonance. improving energy harvesting performance of Dynamics and coherence resonance of tri-
Rev. Mod. Phys. 70, 223. nonlinear monostable energy harvesters. stable energy harvesting system. Smart
Nonlinear Dyn. 86, 1599–1611. Mater. Struct. 25, 015001.
33. Cottone, F., Vocca, H., and Gammaitoni, L.
(2009). Nonlinear energy harvesting. Phys. 49. Stanton, S.C., Owens, B.A., and Mann, B.P. 64. Su, W.-J., Zu, J., and Zhu, Y. (2013). Design
Rev. Lett. 102, 080601. (2012). Harmonic balance analysis of the and development of a broadband magnet-
bistable piezoelectric inertial generator. induced dual-cantilever piezoelectric energy
34. Erturk, A., Hoffmann, J., and Inman, D.J. J. Sound Vib. 331, 3617–3627. harvester. J. Intell. Mater. Syst. Struct. 25,
(2009). A piezomagnetoelastic structure for 430–442.
broadband vibration energy harvesting. Appl. 50. Harne, R., and Wang, K. (2014). On the
Phys. Lett. 94, 254102. fundamental and superharmonic effects in 65. Su, W.-J., and Zu, J.W. (2014). Design and
bistable energy harvesting. J. Intell. Mater. development of a novel bi-directional
35. Zhou, S., Cao, J., Erturk, A., and Lin, J. (2013). Syst. Struct. 25, 937–950. piezoelectric energy harvester. Smart Mater.
Enhanced broadband piezoelectric energy Struct. 23, 095012.
harvesting using rotatable magnets. Appl. 51. Masana, R., and Daqaq, M. (2012). Energy
Phys. Lett. 102, 173901. harvesting in the super-harmonic frequency 66. Arrieta, A., Hagedorn, P., Erturk, A., and
region of a twin-well oscillator. J. Appl. Inman, D. (2010). A piezoelectric bistable
36. Stanton, S.C., McGehee, C.C., and Mann, B.P. Physiol. 111, 044501. plate for nonlinear broadband energy
(2010). Nonlinear dynamics for broadband harvesting. Appl. Phys. Lett. 97, 104102.
energy harvesting: investigation of a bistable 52. Zhou, S., Cao, J., Inman, D.J., Lin, J., and Li, D.
piezoelectric inertial generator. Phys. (2016). Harmonic balance analysis of nonlinear 67. Leadenham, S., and Erturk, A. (2014). M-
Nonlinear Phenom. 239, 640–653. tristable energy harvesters for performance shaped asymmetric nonlinear oscillator for
enhancement. J. Sound Vib. 373, 223–235. broadband vibration energy harvesting:
37. Sebald, G., Kuwano, H., Guyomar, D., and harmonic balance analysis and experimental
Ducharne, B. (2011). Experimental Duffing 53. Panyam, M., and Daqaq, M.F. (2017). validation. J. Sound Vib. 333, 6209–6223.
oscillator for broadband piezoelectric energy Characterizing the effective bandwidth of tri-
harvesting. Smart Mater. Struct. 20, 102001. stable energy harvesters. J. Sound Vib. 386, 68. Arrieta, A., Neild, S., and Wagg, D. (2009).
336–358. Nonlinear dynamic response and modeling of
38. Tang, L., and Yang, Y. (2012). A nonlinear a bi-stable composite plate for applications to
piezoelectric energy harvester with magnetic 54. Cao, J., Zhou, S., Inman, D.J., and Chen, Y. adaptive structures. Nonlinear Dyn. 58,
oscillator. Appl. Phys. Lett. 101, 094102. (2015). Chaos in the fractionally damped 259–272.

50 Joule 2, 1–56, April 18, 2018

Please cite this article in press as: Yang et al., High-Performance Piezoelectric Energy Harvesters and Their Applications, Joule (2018), https://
doi.org/10.1016/j.joule.2018.03.011

69. Betts, D.N., Kim, H.A., Bowen, C.R., and design of smart energy harvesters. Comput. 98. Malakooti, M.H., and Sodano, H.A. (2015).
Inman, D. (2012). Optimal configurations of Methods Appl. Mech. Eng. 199, 2532–2543. Piezoelectric energy harvesting through shear
bistable piezo-composites for energy mode operation. Smart Mater. Struct. 24,
harvesting. Appl. Phys. Lett. 100, 114104. 84. Nanthakumar, S., Lahmer, T., Zhuang, X., 055005.
Park, H.S., and Rabczuk, T. (2016). Topology
70. Betts, D.N., Bowen, C.R., Kim, H.A., optimization of piezoelectric nanostructures. 99. Kulkarni, V., Ben-Mrad, R., Prasad, S.E., and
Gathercole, N., Clarke, C.T., and Inman, D.J. J. Mech. Phys. Solid. 94, 316–335. Nemana, S. (2014). A shear-mode energy
(2013). Nonlinear dynamics of a bistable harvesting device based on torsional stresses.
piezoelectric-composite energy harvester for 85. Wein, F., Kaltenbacher, M., and Stingl, M. IEEE/ASME Trans. Mechatronics 19, 801–807.
broadband application. Eur. Phys. J. Spec. (2013). Topology optimization of a
Top. 222, 1553–1562. cantilevered piezoelectric energy harvester 100. Ren, B., Or, S.W., Zhang, Y., Zhang, Q., Li, X.,
using stress norm constraints. Struct. Jiao, J., Wang, W., Liu, D., Zhao, X., and Luo,
71. Zhou, S., Cao, J., Inman, D.J., Liu, S., Wang, Multidiscip. O. 48, 173–185. H. (2010). Piezoelectric energy
W., and Lin, J. (2015). Impact-induced high- harvesting using shear mode 0.71
energy orbits of nonlinear energy harvesters. 86. Takezawa, A., Kitamura, M., Vatanabe, S.L., Pb(Mg1/3Nb2/3)O3-0.29PbTiO3 single crystal
Appl. Phys. Lett. 106, 093901. and Silva, E.C.N. (2014). Design methodology cantilever. Appl. Phys. Lett. 96, 083502.
of piezoelectric energy-harvesting skin using
72. Mallick, D., Amann, A., and Roy, S. (2016). topology optimization. Struct. Multidiscip. O. 101. Aladwani, A., Aldraihem, O., and Baz, A.
Surfing the high energy output branch of 49, 281–297. (2013). Single degree of freedom shear-mode
nonlinear energy harvesters. Phys. Rev. Lett. piezoelectric energy harvester. J. Vib. Acoust.
117, 197701. 87. Nakasone, P.H., and Silva, E.C. (2009). Design 135, 051011.
of piezoelectric energy harvesting devices
73. Lan, C., Tang, L., and Qin, W. (2017). and laminate structures by applying topology 102. Xu, C.D., Ren, B., Di, W.N., Liang, Z., Jiao, J.,
Obtaining high-energy responses of optimization. Proc. SPIE. https://doi.org/10. Li, L.Y., Li, L., Zhao, X.Y., Luo, H.S., and
nonlinear piezoelectric energy harvester by 1117/12.816467. Wang, D. (2012). Cantilever driving low
voltage impulse perturbations. Eur. Phys. J. frequency piezoelectric energy
Appl. Phys. 79, 20902. 88. Zheng, B., Chang, C.-J., and Gea, H.C. (2009). harvester using single crystal material 0.71
Topology optimization of energy harvesting Pb(Mg1/3Nb2/3)O3-0.29PbTiO3. Appl. Phys.
74. Haji Hosseinloo, A., Slotine, J.-J., and devices using piezoelectric materials. Struct. Lett. 101, 033502.
Turitsyn, K. (2017). Robust and adaptive Multidiscip. O. 38, 17–23.
control of coexisting attractors in nonlinear 103. Zou, H.-X., Zhang, W.-M., Li, W.-B., Hu, K.-M.,
vibratory energy harvesters. J. Vib. Control. 89. Jia, Y., and Seshia, A.A. (2016). Power Wei, K.-X., Peng, Z.-K., and Meng, G. (2017). A
https://doi.org/10.1177/1077546316688992. optimization by mass tuning for MEMS broadband compressive-mode vibration
piezoelectric cantilever vibration energy energy harvester enhanced by magnetic force
75. Roundy, S., Leland, E.S., Baker, J., Carleton, harvesting. J. Microelectromech. Syst. 25, intervention approach. Appl. Phys. Lett. 110,
E., Reilly, E., Lai, E., Otis, B., Rabaey, J.M., 108–117. 163904.
Wright, P.K., and Sundararajan, V. (2005).
Improving power output for vibration-based 90. Yang, Z., Wang, Y.Q., Zuo, L., and Zu, J. (2017). 104. Wu, J., Chen, X., Chu, Z., Shi, W., Yu, Y., and
energy scavengers. IEEE Pervas. Comput. 4, Introducing arc-shaped piezoelectric Dong, S. (2016). A barbell-shaped
28–36. elements into energy harvesters. Energ. high-temperature piezoelectric vibration
Convers. Manag. 148, 260–266. energy harvester based on BiScO3-PbTiO3
76. Baker, J., Roundy, S., and Wright, P. (2005). ceramic. Appl. Phys. Lett. 109, 173901.
Alternative geometries for increasing power 91. Stewart, M., Weaver, P.M., and Cain, M.
density in vibration energy scavenging for 105. Kim, H.W., Batra, A., Priya, S., Uchino, K.,
(2012). Charge redistribution in piezoelectric
wireless sensor networks. in Proc. 3rd Int. Markley, D., Newnham, R.E., and Hofmann,
energy harvesters. Appl. Phys. Lett. 100,
Energy Conversion Engineering Conf. (San H.F. (2004). Energy harvesting using a
073901–073901-3.
Francisco, CA). Piezoelectric ‘‘Cymbal’’ Transducer in
92. Du, S., Jia, Y., Chen, S.-T., Zhao, C., Sun, B., dynamic environment. Jpn. J. Appl. Phys. 43,
77. Dietl, J.M., and Garcia, E. (2010). Beam shape Arroyo, E., and Seshia, A.A. (2017). A. new 6178.
optimization for power harvesting. J. Intell. electrode design method in piezoelectric
Mater. Syst. Struct. 21, 633–646. vibration energy harvesters to maximize
106. Tufekcioglu, E., and Dogan, A. (2014). A
flextensional piezo-composite structure for
78. Goldschmidtboeing, F., and Woias, P. (2008). output power. Sensor. Actuator. Phys. 263,
energy harvesting applications. Sensor.
693–701.
Characterization of different beam shapes for Actuator. Phys. 216, 355–363.
piezoelectric energy harvesting.
93. Erturk, A., Tarazaga, P.A., Farmer, J.R., and
J. Micromech. Microeng. 18, 104013. 107. Lee, H.J., Sherrit, S., Tosi, L.P., Walkemeyer,
Inman, D.J. (2009). Effect of strain nodes and P., and Colonius, T. (2015). Piezoelectric
79. Ben Ayed, S., Abdelkefi, A., Najar, F., and Hajj, electrode configuration on piezoelectric energy harvesting in internal fluid flow.
M.R. (2014). Design and performance of energy harvesting from cantilevered beams. Sensors (Basel) 15, 26039–26062.
variable-shaped piezoelectric energy J. Vib. Acoust. 131, 011010.
harvesters. J. Intell. Mater. Syst. Struct. 25, 108. Yang, Z., Zu, J., Luo, J., and Peng, Y. (2017).
174–186. 94. Cho, J., Anderson, M., Richards, R., Bahr, D., Modeling and parametric study of a force-
and Richards, C. (2005). Optimization of amplified compressive-mode piezoelectric
80. Park, J., Lee, S., and Kwak, B.M. (2012). Design electromechanical coupling for a thin-film PZT energy harvester. J. Intell. Mater. Syst. Struct.
optimization of piezoelectric energy harvester membrane: II. Experiment. J. Micromech. 28, 357–366.
subject to tip excitation. J. Mech. Sci. Microeng. 15, 1804.
Technol. 26, 137–143. 109. Xu, T.-B., Siochi, E.J., Kang, J.H., Zuo, L., Zhou,
95. Yang, Z.B. and J. Zu. Charge Redistribution in W., Tang, X., and Jiang, X. (2013). Energy
81. Paquin, S., and St-Amant, Y. (2010). Improving Flextensional Piezoelectric Energy harvesting using a PZT ceramic multilayer
the performance of a piezoelectric energy Harvesters. Trans Tech. stack. Smart Mater. Struct. 22, 065015.
harvester using a variable thickness beam.
Smart Mater. Struct. 19, 105020. 96. Kim, M., Dugundji, J., and Wardle, B.L. (2015). 110. Moure, A., Rodrı́guez, M.I., Rueda, S.H.,
Effect of electrode configurations on Gonzalo, A., Rubio-Marcos, F., Cuadros, D.U.,
82. Rupp, C.J., Evgrafov, A., Maute, K., and Dunn, piezoelectric vibration energy harvesting Pérez-Lepe, A., and Fernández, J. (2016).
M.L. (2009). Design of piezoelectric energy performance. Smart Mater. Struct. 24, 045026. Feasible integration in asphalt of piezoelectric
harvesting systems: a topology optimization cymbals for vibration energy harvesting.
approach based on multilayer plates and 97. Zhao, J., Zheng, X., Zhou, L., Zhang, Y., Sun, J., Energ. Convers. Manag. 112, 246–253.
shells. J. Intell. Mater. Syst. Struct. 20, 1923– Dong, W., Deng, S., and Peng, S. (2012).
1939. Investigation of a d15 mode PZT-51 111. Jiang, X., Li, Y., Li, J., Wang, J., and Yao, J.
piezoelectric energy harvester with a series (2014). Piezoelectric energy harvesting from
83. Chen, S., Gonella, S., Chen, W., and Liu, W.K. connection structure. Smart Mater. Struct. 21, traffic-induced pavement vibrations.
(2010). A level set approach for optimal 105006. J. Renew. Sustain. Energ. 6, 043110.

Joule 2, 1–56, April 18, 2018 51

 Please cite this article in press as: Yang et al., High-Performance Piezoelectric Energy Harvesters and Their Applications, Joule (2018), https://
doi.org/10.1016/j.joule.2018.03.011

112. Lee, A.J., Wang, Y., and Inman, D.J. (2014). nanocomposite and device. Nano Lett. 13, harvester. J. Microelectromech. Syst. 19,
Energy harvesting of piezoelectric stack 2393–2398. 1215–1222.
actuator from a shock event. J. Vib. Acoust.
136, 011016. 125. Wu, F., and Yao, N. (2017). PMN-PT 140. Tang, G., Liu, J.-q., Yang, B., Luo, J.-b., Liu,
nanostructures for energy scavenging. H.-s., Li, Y.-g., Yang, C.-s., He, D.-n., Dao, V.D.,
113. Wang, X., Shi, Z., Wang, J., and Xiang, H. Semicond. Sci. Technol. 32, 063001. Tanaka, K., and Sugiyama, S. (2012).
(2016). A stack-based flex-compressive Fabrication and analysis of high-performance
piezoelectric energy harvesting cell for large 126. Chen, Y., Zhang, Y., Yuan, F., Ding, F., and piezoelectric MEMS generators.
quasi-static loads. Smart Mater. Struct. 25, Schmidt, O.G. (2017). A flexible PMN-PT J. Micromech. Microeng. 22, 065017.
055005. ribbon-based piezoelectric-pyroelectric
hybrid generator for human-activity energy 141. Minh le, V., Hara, M., Yokoyama, T., Nishihara,
114. Xu, T.-B., Jiang, X., and Su, J. (2011). A harvesting and monitoring. Adv. Electron. T., Ueda, M., and Kuwano, H. (2015). Highly
piezoelectric multilayer-stacked hybrid Mater. 3, https://doi.org/10.1002/aelm. piezoelectric MgZr co-doped aluminum
actuation/transduction system. Appl. Phys. 201600540. nitride-based vibrational energy harvesters
Lett. 98, 243503. [Correspondence]. IEEE Trans. Ultrason.
127. Marin, A., Bressers, S., and Priya, S. (2011). Ferroelectr. Freq. Control 62, 2005–2008.
115. Morimoto, K., Kanno, I., Wasa, K., and Kotera, Multiple cell configuration electromagnetic
H. (2010). High-efficiency piezoelectric energy vibration energy harvester. J. Phys. D 44, 142. Tang, G., Yang, B., Hou, C., Li, G., Liu, J.,
harvesters of c-axis-oriented epitaxial PZT 295501. Chen, X., and Yang, C. (2016). A piezoelectric
films transferred onto stainless steel micro generator worked at low frequency and
cantilevers. Sensor. Actuator. Phys. 163, 128. Beeby, S.P., Torah, R., Tudor, M., Glynne- high acceleration based on PZT and phosphor
428–432. Jones, P., O’Donnell, T., Saha, C., and Roy, S. bronze bonding. Sci. Rep. 6, 38798.
(2007). A micro electromagnetic generator for
116. Yeo, H.G., Ma, X., Rahn, C., and vibration energy harvesting. J. Micromech. 143. Song, H.-C., Kumar, P., Maurya, D., Kang,
Trolier-McKinstry, S. (2016). Efficient Microeng. 17, 1257. M.-G., Reynolds, W.T., Jeong, D.-Y., Kang,
piezoelectric energy harvesters utilizing (001) C.-Y., and Priya, S. (2017). Ultra-low resonant
textured bimorph PZT films on flexible metal 129. Yang, Z., Erturk, A., and Zu, J. (2017). On the piezoelectric MEMS energy harvester with
foils. Adv. Funct. Mater. 26, 5940–5946. efficiency of piezoelectric energy harvesters. high power density. J. Microelectromech.
Extreme Mech. Lett. 15, 26–37. Syst. 26, 1226–1234.
117. Hwang, G.T., Annapureddy, V., Han, J.H., Joe,
D.J., Baek, C., Park, D.Y., Kim, D.H., Park, J.H., 130. Kim, M., Dugundji, J., and Wardle, B.L. (2015). 144. Dai, X., Wen, Y., Li, P., Yang, J., and Zhang, G.
Jeong, C.K., Park, K.-I., et al. (2016). Self- Efficiency of piezoelectric mechanical (2009). Modeling, characterization and
powered wireless sensor node enabled by an vibration energy harvesting. Smart Mater. fabrication of vibration energy harvester using
aerosol-deposited PZT flexible energy Struct. 24, 055006. Terfenol-D/PZT/Terfenol-D composite
harvester. Adv. Energy Mater. 6, https://doi. transducer. Sensor. Actuator. Phys. 156,
org/10.1002/aenm.201600237. 131. Akaydin, H.D., Elvin, N., and Andreopoulos, Y. 350–358.
(2012). The performance of a self-excited
118. Xu, S., Hansen, B.J., and Wang, Z.L. (2010). fluidic energy harvester. Smart Mater. Struct. 145. Kim, M., Hoegen, M., Dugundji, J., and
Piezoelectric-nanowire-enabled power 21, 025007. Wardle, B.L. (2010). Modeling and
source for driving wireless microelectronics. experimental verification of proof mass
Nat. Commun. 1, 93. 132. Shafer, M.W., and Garcia, E. (2014). The power effects on vibration energy harvester
and efficiency limits of piezoelectric energy performance. Smart Mater. Struct. 19, 045023.
119. Tang, G., Yang, B., Liu, J.-q., Xu, B., Zhu, H.-y., harvesting. J. Vib. Acoust. 136, 021007.
and Yang, C.-s. (2014). Development of high 146. Liang, J., and Liao, W.-H. (2010). Impedance
performance piezoelectric d 33 mode MEMS 133. Roundy, S. (2005). On the effectiveness of matching for improving piezoelectric energy
vibration energy harvester based on PMN-PT vibration-based energy harvesting. J. Intell. harvesting systems. in SPIE Smart Structures
single crystal thick film. Sensor. Actuator. Mater. Syst. Struct. 16, 809–823. and Materials + Nondestructive Evaluation
Phys. 205, 150–155. and Health Monitoring, International Society
134. Liu, W.Q., Badel, A., Formosa, F., and Wu, Y.P. for Optics and Photonics.
120. Erturk, A., Bilgen, O., and Inman, D.J. (2008). (2015). A new figure of merit for wideband
Performance analysis of single crystal vibration energy harvesters. Smart Mater. 147. Gu, L. (2011). Low-frequency piezoelectric
PMN-PZT unimorphs for piezoelectric energy Struct. 24, 125012. energy harvesting prototype suitable for the
harvesting. in Proceedings of the ASME MEMS implementation. Microelectron. J. 42,
Conference on Smart Materials, Adaptive 135. Cammarano, A., Gonzalez-Buelga, A., Neild, 277–282.
Structures and Intelligent Systems, S., Burrow, S., and Inman, D. (2013).
SMASIS2008. Bandwidth of a nonlinear harvester with 148. Li, X., Guo, M., and Dong, S. (2011). A flex-
optimized electrical load. J. Phys. Conf. Ser. compressive-mode piezoelectric transducer
121. Hwang, G.T., Park, H., Lee, J.H., Oh, S., Park, 476, 012071. for mechanical vibration/strain energy
K.I., Byun, M., Park, H., Ahn, G., Jeong, C.K., harvesting. IEEE Trans. Ultrason. Ferroelectr.
No, K., et al. (2014). Self-powered cardiac 136. Dhote, S., Yang, Z., and Zu, J. (2018). Freq. Control 58, 698–703.
pacemaker Enabled by flexible single Modeling and experimental parametric study
crystalline PMN-PT piezoelectric energy of a tri-leg compliant orthoplanar spring 149. Yen, T.-T., Hirasawa, T., Wright, P.K., Pisano,
harvester. Adv. Mater. 26, 4880–4887. based multi-mode piezoelectric energy A.P., and Lin, L. (2011). Corrugated aluminum
harvester. Mech. Syst. Signal Process. 98, nitride energy harvesters for high energy
122. Hwang, G.T., Yang, J., Yang, S.H., Lee, H.-Y., 268–280. conversion effectiveness. J. Micromech.
Lee, M., Park, D.Y., Han, J.H., Lee, S.J., Jeong, Microeng. 21, 085037.
C.K., Kim, J., et al. (2015). A reconfigurable 137. Lee, B., Lin, S., Wu, W., Wang, X., Chang, P.,
rectified flexible energy harvester via solid- and Lee, C. (2009). Piezoelectric MEMS 150. Dhakar, L., Liu, H., Tay, F., and Lee, C. (2013). A
state single crystal grown PMN-PZT. Adv. generators fabricated with an aerosol new energy harvester design for high power
Energy Mater. 5, https://doi.org/10.1002/ deposition PZT thin film. J. Micromech. output at low frequencies. Sensor. Actuator.
aenm.201500051. Microeng. 19, 065014. Phys. 199, 344–352.

123. Zeng, Z., Xia, R., Gai, L., Wang, X., Lin, D., Luo, 138. Shen, D., Park, J.-H., Noh, J.H., Choe, S.-Y., 151. Wu, H., Tang, L., Yang, Y., and Soh, C.K.
H., Li, F., and Wang, D. (2016). High Kim, S.-H., Wikle, H.C., and Kim, D.-J. (2009). (2013). A novel two-degrees-of-freedom
performance of macro-flexible Micromachined PZT cantilever based on SOI piezoelectric energy harvester. J. Intell. Mater.
piezoelectric energy harvester using a 0.3 PIN- structure for low frequency vibration energy Syst. Struct. 24, 357–368.
0.4 Pb (Mg1/3Nb2/3)O3-0.3PbTiO3 flake array. harvesting. Sensor. Actuator. Phys. 154,
Smart Mater. Struct. 25, 125015. 103–108. 152. Arrieta, A., Delpero, T., Bergamini, A., and
Ermanni, P. (2013). Broadband vibration
124. Xu, S., Yeh, Y.W., Poirier, G., McAlpine, M.C., 139. Park, J.C., Park, J.Y., and Lee, Y.-P. (2010). energy harvesting based on cantilevered
Register, R.A., and Yao, N. (2013). Flexible Modeling and characterization of piezoelectric bi-stable composites. Appl.
piezoelectric PMN-PT nanowire-based piezoelectric d33-mode MEMS energy Phys. Lett. 102, 173904–173904-4.

52 Joule 2, 1–56, April 18, 2018

 Please cite this article in press as: Yang et al., High-Performance Piezoelectric Energy Harvesters and Their Applications, Joule (2018), https://
doi.org/10.1016/j.joule.2018.03.011

153. Qiu, J., Chen, H., Wen, Y., Li, P., Yang, J., and 168. Rastegar, J., Murray, R., Pereira, C., and 182. Shenck, N.S., and Paradiso, J.A. (2001).
Li, W. (2014). A vibration energy harvester Nguyen, H.-L. (2009). Event sensing and Energy scavenging with shoe-mounted
using magnet/piezoelectric composite energy-harvesting power sources for gun- piezoelectrics. IEEE Micro 21, 30–42.
transducer. J. Appl. Physiol. 115, 17E522. fired munitions. in SPIE Smart Structures
Materials+ Nondestructive Evaluation and 183. Moro, L., and Benasciutti, D. (2010). Harvested
154. Ma, M., Xia, S., Li, Z., Xu, Z., and Yao, X. (2014). Health Monit.oring. International Society for power and sensitivity analysis of vibrating
Enhanced energy harvesting performance of Optics and Photonics. shoe-mounted piezoelectric cantilevers.
the piezoelectric unimorph with Smart Mater. Struct. 19, 115011.
perpendicular electrodes. Appl. Phys. Lett. 169. Ostasevicius, V., Markevicius, V., Jurenas, V.,
105, 043905. Zilys, M., Cepenas, M., Kizauskiene, L., and 184. Fan, K., Liu, Z., Liu, H., Wang, L., Zhu, Y., and
Gyliene, V. (2015). Cutting tool vibration Yu, B. (2017). Scavenging energy from human
155. Zhang, Y., Cai, C., and Zhang, W. (2014). energy harvesting for wireless sensors walking through a shoe-mounted
Experimental study of a multi-impact energy applications. Sensor. Actuator. Phys. 233, piezoelectric harvester. Appl. Phys. Lett. 110,
harvester under low frequency excitations. 310–318. 143902.
Smart Mater. Struct. 23, 055002.
170. Shafer, M.W., MacCurdy, R., Shipley, J.R., 185. Mateu, L., and Moll, F. (2005). Optimum
156. Xiao, Z., Yang, T.q., Dong, Y., and Wang, X.c. piezoelectric bending beam structures for
Winkler, D., Guglielmo, C.G., and Garcia, E.
(2014). Energy harvester array using energy harvesting using shoe inserts. J. Intell.
(2015). The case for energy harvesting on
piezoelectric circular diaphragm for Mater. Syst. Struct. 16, 835–845.
wildlife in flight. Smart Mater. Struct. 24,
broadband vibration. Appl. Phys. Lett. 104,
025031.
223904. 186. Meier, R., Kelly, N., Almog, O., and Chiang, P.
157. Hung, C.-F., Chung, T.-K., Yeh, P.-C., Chen, 171. Li, H., Tian, C., Lu, J., Myjak, M.J., Martinez, (2014). A piezoelectric energy-harvesting
C.-C., Wang, C.-M., and Lin, S.-H. (2015). A J.J., Brown, R.S., and Deng, Z.D. (2016). An shoe system for podiatric sensing. in
miniature mechanical-piezoelectric- energy harvesting underwater acoustic Engineering in Medicine and Biology Society
configured three-axis vibrational energy transmitter for aquatic animals. Sci. Rep. 6, (EMBC), 2014 36th Annual International
harvester. IEEE Sens. J. 15, 5601–5615. 33804. Conference of the IEEE. IEEE.

158. Singh, K.A., Kumar, R., and Weber, R.J. (2015). 172. Winter, D.A. (2009). Biomechanics and Motor 187. Fourie, D. (2010). Shoe mounted PVDF
A broadband bistable piezoelectric energy Control of Human Movement (John Wiley & piezoelectric transducer for energy
harvester with nonlinear high-power Sons). harvesting. MORJ Rep. 19, 66–70.
extraction. IEEE Trans. Power Electron. 30,
6763–6774. 173. DeVita, P., Helseth, J., and Hortobagyi, T. 188. Ishida, K., Huang, T.-C., Honda, K., Shinozuka,
(2007). Muscles do more positive than Y., Fuketa, H., Yokota, T., Zschieschang, U.,
159. Sriramdas, R., Chiplunkar, S., Cuduvally, R.M., negative work in human locomotion. J. Exp. Klauk, H., Tortissier, G., Sekitani, T., et al.
and Pratap, R. (2015). Performance Biol. 210, 3361–3373. (2013). Insole pedometer with piezoelectric
enhancement of piezoelectric energy energy harvester and 2 V organic circuits. IEEE
harvesters using multilayer and multistep 174. Lindstedt, S., LaStayo, P., and Reich, T. (2001). J. Solid State Circ. 48, 255–264.
beam configurations. IEEE Sens. J. 15, 3338– When active muscles lengthen: properties
3348. and consequences of eccentric contractions. 189. Zhao, J., and You, Z. (2014). A shoe-
Physiology 16, 256–261. embedded piezoelectric energy harvester for
160. Gong, L.J., Pan, Q.S., Li, W., Yan, G.Y., Liu, wearable sensors. Sensors 14, 12497–12510.
Y.B., and Feng, Z.H. (2015). Harvesting 175. Starner, T. (1996). Human-powered wearable
vibration energy using two modal vibrations computing. IBM Syst. J. 35, 618–629. 190. Jung, W.-S., Lee, M.-J., Kang, M.-G., Moon,
of a folded piezoelectric device. Appl. Phys. H.G., Yoon, S.-J., Baek, S.-H., and Kang, C.-Y.
Lett. 107, 033904. 176. Niu P., Chapman, P., Riemer, R., and Zhang, X. (2015). Powerful curved piezoelectric
(2004). Evaluation of motions and actuation generator for wearable applications. Nano
161. Yi, Z., Yang, B., Li, G., Liu, J., Chen, X., Wang, methods for biomechanical energy Energy 13, 174–181.
X., and Yang, C. (2017). High performance harvesting. in Power Electronics Specialists
bimorph piezoelectric MEMS harvester via Conference, 2004. PESC 04. 2004 IEEE 35th 191. Daniels, A., Zhu, M., and Tiwari, A. (2013).
bulk PZT thick films on thin beryllium-bronze Annual. IEEE. Design, analysis and testing of a piezoelectric
substrate. Appl. Phys. Lett. 111, 013902. flex transducer for harvesting bio-kinetic
177. Riemer, R., and Shapiro, A. (2011). energy. J. Phys. Conf. Ser. 476, 012047.
162. He, Q., and Jiang, T. (2017). Complementary Biomechanical energy harvesting from human
multi-mode low-frequency vibration energy motion: theory, state of the art, design 192. Xie, L., and Cai, M. (2014). Increased
harvesting with chiral piezoelectric structure. guidelines, and future directions. piezoelectric energy harvesting from human
Appl. Phys. Lett. 110, 213901. J. Neuroeng. Rehabil. 8, 22. footstep motion by using an amplification
mechanism. Appl. Phys. Lett. 105, 143901.
163. Wong, V.-K., Ho, J.-H., and Chai, A.-B. (2017). 178. Partridge, J., and Bucknall, R. (2016). An
Performance of a piezoelectric energy analysis of the energy flow and energy 193. Niu, P., Chapman, P., DiBerardino, L., and
harvester in actual rain. Energy 124, 364–371. potential from human energy harvesting with Hsiao-Wecksler, E. (2008). Design and
164. Guigon, R., Chaillout, J.-J., Jager, T., and a focus on walking. Cogent Eng. 3, 1215203. optimization of a biomechanical energy
Despesse, G. (2008). Harvesting raindrop harvesting device. in Power Electronics
179. Fu, H., Cao, K., Xu, R., Bhouri, M.A., Specialists Conference, 2008. PESC 2008.
energy: experimental study. Smart Mater.
Martı́nez-Botas, R., Kim, S.-G., and Yeatman, IEEE. IEEE.
Struct. 17, 015039.
E.M. Footstep energy harvesting using heel
165. Wong, C.H., and Dahari, Z. (2017). strike-induced airflow for human activity 194. Amar, A.B., Kouki, A.B., and Cao, H. (2015).
Development of vibration-based sensing. in Wearable and Implantable Body Power approaches for implantable medical
piezoelectric raindrop energy harvesting Sensor Networks (BSN), 2016 IEEE 13th devices. Sensors 15, 28889–28914.
system. J. Electron. Mater. 46, 1869–1882. International Conference. 2016. IEEE.
195. Bradley, P.D. (2006). An ultra low power, high
166. Wong, V.-K., Ho, J.-H., and Sam, H.-K. (2017). 180. Ylli, K., Hoffmann, D., Willmann, A., Becker, P., performance medical implant communication
On accumulation of water droplets in Folkmer, B., and Manoli, Y. (2015). Energy system (MICS) transceiver for implantable
piezoelectric energy harvesting. J. Intell. harvesting from human motion: exploiting devices. in Biomedical Circuits and Systems
Mater. Syst. Struct. 28, 521–530. swing and shock excitations. Smart Mater. Conference, 2006. BioCAS 2006. IEEE. IEEE.
Struct. 24, 025029.
167. Guo, L., and Lu, Q. (2017). Potentials of 196. Wong, L.S., Hossain, S., Ta, A., Edvinsson, J.,
piezoelectric and thermoelectric 181. Hayashida, J.Y. (2000). Unobtrusive Rivas, D.H., and Naas, H. (2004). A very low-
technologies for harvesting energy from Integration of Magnetic Generator Systems power CMOS mixed-signal IC for implantable
pavements. Renew. Sustain. Energ. Rev. 72, into Common Footwear (Massachusetts pacemaker applications. IEEE J. Solid State
761–773. Institute of Technology). Circ. 39, 2446–2456.

Joule 2, 1–56, April 18, 2018 53

 Please cite this article in press as: Yang et al., High-Performance Piezoelectric Energy Harvesters and Their Applications, Joule (2018), https://
doi.org/10.1016/j.joule.2018.03.011

197. Padeletti, L., and Barold, S.S. (2005). Digital powering pacemakers. Smart Mater. Struct. intracardiac leadless pacemaker. N. Engl. J.
technology for cardiac pacing. Am. J. Cardiol. 26, 065001. Med. 373, 1125–1135.
95, 479–482.
212. Sharpes, N., Abdelkefi, A., and Priya, S. 225. Potkay, J.A., and Brooks K. An arterial cuff
198. Kim, S., Cho, N., Song, S.-J., Kim, D., Kim, K., (2015). Two-dimensional concentrated- energy scavenger for implanted
and Yoo, H.-J. (2006). A 0.9-V 96-mW digital stress low-frequency piezoelectric microsystems. in Bioinformatics and
hearing aid chip with heterogeneous SD vibration energy harvesters. Appl. Phys. Biomedical Engineering, 2008. ICBBE 2008.
DAC. in Proc. IEEE Symp. VLSI Circuits. Lett. 107, 093901. The 2nd International Conference. 2008. IEEE.

199. Sarpeshkar, R., Salthouse, C., Sit, J.J., Baker, 213. Alrashdan, M.H., Hamzah, A.A., and Majlis, B. 226. Zhang, H., Zhang, X.-S., Cheng, X., Liu, Y.,
M.W., Zhak, S.M., Lu, T.K., Turicchia, L., and (2015). Design and optimization of cantilever Han, M., Xue, X., Wang, S., Yang, F., Smith,
Balster, S. (2005). An ultra-low-power based piezoelectric micro power generator A.S., Zhang, H., and Xu, Z. (2015). A flexible
programmable analog bionic ear processor. for cardiac pacemaker. Microsystem Tech. 21, and implantable piezoelectric generator
IEEE Trans. Biomed. Eng. 52, 711–727. 1607–1617. harvesting energy from the pulsation of
ascending aorta: in vitro and in vivo studies.
200. Saati, S., Lo, R., Li, P.Y., Meng, E., Varma, R., 214. Dagdeviren, C., Yang, B.D., Su, Y., Tran, P.L., Nano Energy 12, 296–304.
and Humayun, M.S. (2010). Mini drug pump Joe, P., Anderson, E., Xia, J., Doraiswamy, V.,
for ophthalmic use. Curr. Eye Res. 35, Dehdashti, B., Feng, X., et al. (2014). 227. Auzanneau, F. (2013). Wire troubleshooting
192–201. Conformal piezoelectric energy harvesting and diagnosis: review and perspectives. Prog.
and storage from motions of the heart, lung, Electromagn. Res. B 49, 253–279.
201. Weiland, J.D., Liu, W., and Humayun, M.S. and diaphragm. Proc. Natl. Acad. Sci. USA
(2005). Retinal prosthesis. Annu. Rev. Biomed. 111, 1927–1932. 228. Khalatkar, A., and Gupta, V. (2017).
Eng. 7, 361–401. Piezoelectric energy harvester for low engine
215. Lu, B., Chen, Y., Ou, D., Chen, H., Diao, L., vibrations. J. Renew. Sustain. Energ. 9,
202. Wise, K.D., Anderson, D., Hetke, J., Kipke, D., Zhang, W., Zheng, J., Ma, W., Sun, L., and 024701.
and Najafi, K. (2004). Wireless implantable Feng, X. (2015). Ultra-flexible piezoelectric
microsystems: high-density electronic devices integrated with heart to harvest the 229. Wang, Y.-J., Chen, C.-D., Lin, C.-C., and Yu,
interfaces to the nervous system. Proc. IEEE biomechanical energy. Sci. Rep. 5. J.-H. (2015). A nonlinear suspended energy
92, 76–97. harvester for a tire pressure monitoring
216. Bedell, S.W., Fogel, K., Lauro, P., Shahrjerdi, system. Micromachines 6, 312–327.
203. Zurbuchen, A., Pfenniger, A., Stahel, A., D., Ott, J.A., and Sadana, D. (2013). Layer
230. Wang, Y.-J., Shen, S.-C., and Chen, C.-D.
Stoeck, C.T., Vandenberghe, S., Koch, V.M., transfer by controlled spalling. J. Phys. D 46,
and Vogel, R. (2013). Energy harvesting from (2012). Wideband electromagnetic energy
152002.
the beating heart by a mass imbalance harvesting from a rotating wheel. In Small-
oscillation generator. Ann. Biomed. Eng. 41, 217. Kim, D.H., Shin, H.J., Lee, H., Jeong, C.K., Scale Energy Harvesting, M. Lallart, ed.
131–141. Park, H., Hwang, G.-T., Lee, H.-Y., Joe, D.J., (InTech). https://doi.org/10.5772/50739.
Han, J.H., Lee, S.H., et al. (2017). In vivo self-
204. Lentner, C. (1990). Geigy Scientific Table: 231. van Schaijk, R., Elfrink, R., Oudenhoven, J.,
powered wireless transmission using
Heart and Circulation, Volume 5 (CIBA). Pop, V., Wang, Z., and Renaud, M. (2013). A
biocompatible flexible energy harvesters.
MEMS vibration energy harvester for
Adv. Funct. Mater. https://doi.org/10.1002/
205. Papademetris, X., and Duncan, J.S. (2000). automotive applications. Proc. SPIE. https://
adfm.201700341.
Cardiac image analysis: motion and doi.org/10.1117/12.2016916.
deformation. In Handbook of Medical 218. Hwang, G.T., Byun, M., Jeong, C.K., and Lee,
Imaging, 2, I.M. Bankman, ed (Elsevier), 232. Kim, G. (2015). Piezoelectric energy
K.J. (2015). Flexible piezoelectric thin-film
harvesting from torsional vibration in internal
pp. 675–710. energy harvesters and nanosensors for
combustion engines. Int. J. Auto. Technol. 16,
biomedical applications. Adv. Healthc. Mater.
206. Deterre, M., Lefeuvre, E., and Dufour- 645–651.
4, 646–658.
Gergam, E. (2012). An active piezoelectric
233. Zuo, L., and Tang, X. (2013). Large-scale
energy extraction method for pressure energy 219. Wang, D.-A., and Liu, N.-Z. (2011). A shear
vibration energy harvesting. J. Intell. Mater.
harvesting. Smart Mater. Struct. 21, 085004. mode piezoelectric energy harvester based Syst. Struct. 24, 1405–1430.
on a pressurized water flow. Sensor. Actuator.
207. Länne, T., Stale, H., Bengtsson, H., Phys. 167, 449–458. 234. Tianchen, Y., Jian, Y., Ruigang, S., and
Gustafsson, D., Bergqvist, D., Sonesson, B.,
Xiaowei, L. (2014). Vibration energy harvesting
Lecerof, H., and Dahl, P. (1992). Noninvasive 220. Kim, S., Clark, W.W., and Wang, Q.-M. system for railroad safety based on running
measurement of diameter changes in the (2005). Piezoelectric energy harvesting with vehicles. Smart Mater. Struct. 23, 125046.
distal abdominal aorta in man. Ultrasound a clamped circular plate: experimental
Med. Biol. 18, 451–457. study. J. Intell. Mater. Syst. Struct. 16, 235. Bowen, C., and Arafa, M. (2015). Energy
855–863. harvesting technologies for tire pressure
208. Dammers, R., Stifft, F., Tordoir, J.H., monitoring systems. Adv. Energy Mater. 5,
Hameleers, J.M., Hoeks, A.P., and Kitslaar, 221. Mo, C., Radziemski, L.J., and Clark, W.W. https://doi.org/10.1002/aenm.201401787.
P.J. (2003). Shear stress depends on vascular (2010). Experimental validation of energy
territory: comparison between common harvesting performance for pressure-loaded 236. Kubba, A.E., and Jiang, K. (2014). A
carotid and brachial artery. J. Appl. Physiol. piezoelectric circular diaphragms. Smart comprehensive study on technologies of
94, 485–489. Mater. Struct. 19, 075010. tyre monitoring systems and possible
energy solutions. Sensors (Basel) 14, 10306–
209. Karami, M.A., and Inman, D.J. (2012). 222. Sohn, J., Choi, S.B., and Lee, D. (2005). An 10345.
Powering pacemakers from heartbeat investigation on piezoelectric energy
vibrations using linear and nonlinear harvesting for MEMS power sources. J. Mech. 237. Löhndorf, M., Kvisterøy, T., Westby, E., and
energy harvesters. Appl. Phys. Lett. 100, Eng. Sci. 219, 429–436. Halvorsen, E. (2007). Evaluation of energy
042901. harvesting concepts for tire pressure
223. Deterre, M., Lefeuvre, E., Zhu, Y., Woytasik, monitoring systems. Proceedings of Power
210. Karami, M.A., and Inman, D.J. (2011). Linear M., Boutaud, B., and Dal Molin, R. (2014). MEMS, p. 331–334.
and nonlinear energy harvesters for powering Micro blood pressure energy harvester for
pacemakers from heart beat vibrations. in intracardiac pacemaker. J. Microelectromech. 238. Kubba, A.E., Behroozi, M., Olatunbosun,
Active and Passive Smart Structures and Syst. 23, 651–660. O.A., Anthony, C., and Jiang, K. (2014).
Integrated Systems 2011. International Modeling of strain energy harvesting in
Society for Optics and Photonics. 224. Reddy, V.Y., Exner, D.V., Cantillon, D.J., pneumatic tires using piezoelectric
Doshi, R., Bunch, T.J., Tomassoni, G.F., transducer. Tire Sci. Technol. 42, 16–34.
211. Ansari, M., and Karami, M.A. (2017). Friedman, P.A., Estes, N.A., 3rd, Ip, J., Niazi, I.,
Experimental investigation of fan-folded et al.; LEADLESS II Study Investigators (2015). 239. Garcia-Pozuelo, D., Yunta, J., Olatunbosun,
piezoelectric energy harvesters for Percutaneous implantation of an entirely O., Yang, X., and Diaz, V. (2017). A strain-

54 Joule 2, 1–56, April 18, 2018

 Please cite this article in press as: Yang et al., High-Performance Piezoelectric Energy Harvesters and Their Applications, Joule (2018), https://
doi.org/10.1016/j.joule.2018.03.011

based method to estimate slip angle and tire intelligent tires. Mechatronics 22, 268. van den Ende, D.A., Van de Wiel, H., Groen,
working conditions for intelligent tires using 970–988. W., and Van der Zwaag, S. (2011). Direct strain
fuzzy logic. Sensors (Basel) 17, 874. energy harvesting in automobile tires using
254. Sadeqi, S., Arzanpour, S., and Hajikolaei, K.H. piezoelectric PZT–polymer composites. Smart
240. Roveri, N., Pepe, G., and Carcaterra, A. (2016). (2015). Broadening the frequency bandwidth Mater. Struct. 21, 015011.
OPTYRE—A new technology for tire of a tire-embedded piezoelectric-based
monitoring: Evidence of contact patch energy harvesting system using coupled 269. Roundy, S.J. (2003). Energy Scavenging for
phenomena. Mech. Syst. Signal Process. 66, linear resonating structure. IEEE/ASME Trans. Wireless Sensor Nodes with a Focus on
793–810. Mechatronics 20, 2085–2094. Vibration to Electricity Conversion (University
of California).
241. Xiong, Y., and Tuononen, A. (2015). Rolling 255. Zhu, B., Han, J., Zhao, J., and Deng, W. (2017).
deformation of truck tires: measurement and Practical design of an energy harvester 270. Bowen, C., Newnham, R., and Randall, C.
analysis using a tire sensing approach. considering wheel rotation for powering (1998). Dielectric properties of
J. Terramech. 61, 33–42. intelligent tire systems. J. Electron. Mater. 46, dielectrophoretically assembled particulate-
2483–2493. polymer composites. J. Mater. Res. 13,
242. Xiong, Y., and Tuononen, A. (2014). A laser- 205–210.
based sensor system for tire tread 256. McInnes, C., Gorman, D., and Cartmell, M.P.
deformation measurement. Meas. Sci. Tech. (2008). Enhanced vibrational energy 271. Lee, J., and Choi, B. (2014). Development of a
25, 115103. harvesting using nonlinear stochastic piezoelectric energy harvesting system for
resonance. J. Sound Vib. 318, 655–662. implementing wireless sensors on the tires.
243. Tuononen, A.J. (2008). Optical position Energ. Convers. Manag. 78, 32–38.
detection to measure tyre carcass deflections.
Vehicle Syst. Dyn. 46, 471–481. 257. Zheng, R., Nakano, K., Hu, H., Su, D., and 272. Torfs, T., Sterken, T., Brebels, S., Santana, J.,
Cartmell, M.P. (2014). An application of van den Hoven, R., Spiering, V., Bertsch, N.,
244. Matilainen, M., and Tuononen, A. (2015). stochastic resonance for energy harvesting in Trapani, D., and Zonta, D. (2013). Low power
Tyre contact length on dry and wet road a bistable vibrating system. J. Sound Vib. 333, wireless sensor network for building
surfaces measured by three-axial 2568–2587. monitoring. IEEE Sensors J. 13, 909–915.
accelerometer. Mech. Syst. Signal Process.
52, 548–558. 258. Wellens, T., Shatokhin, V., and Buchleitner, A. 273. Nair, A., and Cai, C. (2010). Acoustic emission
(2003). Stochastic resonance. Rep. Progr. monitoring of bridges: review and case
245. Behroozinia, P., Taheri, S., and Mirzaeifar, R. Phys. 67, 45. studies. Eng. Struct. 32, 1704–1714.
(2017). An investigation of intelligent tires
using multiscale modeling of cord-rubber 259. Keck, M. (2007). A new approach of a 274. Arampatzis, T., Lygeros, J, and Manesis, S.
composites. Mech. Base. Des. Struct. Mach. piezoelectric vibration-based power (2005). A survey of applications of wireless
1–16. generator to supply next generation tire sensors and wireless sensor networks. in
sensor systems. in Sensors, 2007 IEEE. Intelligent Control, 2005. Proceedings of the
246. Roundy, S. (2008). Energy harvesting for tire IEEE. 2005 IEEE International Symposium on,
pressure monitoring systems: design Mediterranean Conference on Control and
considerations. Proceedings of Power 260. Jousimaa, O.J., Parmar, M., and Lee, D.-W. Automation. IEEE.
MEMS + MicroMEMS, Sendai, Japan: p. 9–12. (2016). Energy harvesting system for
intelligent tyre sensors. in Intelligent Vehicles 275. Ricquebourg, V., Menga, D., Durand, D.,
247. Moon, K.S., Liang, H., Yi, J., and Mika, B. Symposium (IV), 2016 IEEE. IEEE. Marhic, B., Delahoche, L., and Loge, C. The
(2007). Tire tread deformation sensor and smart home concept: our immediate future.
energy harvester development for ‘Smart 261. Wu, X., Parmar, M., and Lee, D.-W. (2014). A in E-Learning in Industrial Electronics, 2006
Tire’applications. in Proc. SPIE. seesaw-structured energy harvester with 1ST IEEE International Conference on.
superwide bandwidth for TPMS application. 2006. IEEE.
248. Wu, L., Wang, Y., Jia, C., and Zhang, C. IEEE/ASME Trans. Mechatronics 19, 1514–
Battery-less piezoceramics mode energy 1522. 276. Kaur, N., and Bhalla, S. (2014). Combined
harvesting for automobile TPMS. in ASIC, energy harvesting and structural health
2009. ASICON’09. IEEE 8th International 262. Manla, G., White, N., and Tudor, J. (2009). monitoring potential of embedded piezo-
Conference. 2009. IEEE. Harvesting energy from vehicle wheels. in concrete vibration sensors. J. Energy Eng.
Solid-State Sensors, Actuators and 141, D4014001.
249. Mak, K.H., McWilliam, S., and Popov, A.A.
Microsystems Conference, 2009.
(2013). Piezoelectric energy harvesting for tyre 277. Farinholt, K.M., Miller, N., Sifuentes, W.,
TRANSDUCERS 2009. IEEE.
pressure measurement applications. MacDonald, J., Park, G., and Farrar, C.R.
J. Automobile Eng. 227, 842–852. 263. Tang, Q., Xia, X., and Li, X.. Non-contact (2010). Energy harvesting and wireless energy
frequency-up-conversion energy harvester for transmission for embedded SHM sensor
250. Elfrink, R., Matova, S., de Nooijer, C.,
durable & broad-band automotive TPMS nodes. Struct. Health Monit. 9, 269–280.
Jambunathan, M., Goedbloed, M.,
van de Molengraft, J., Pop, V., Vullers, R.J.M., application. in Micro Electro Mechanical
Systems (MEMS), 2012 IEEE 25th International 278. Matiko, J., Grabham, N., Beeby, S., and
Renaud, M., and van Schaijk, R. (2011). Shock Tudor, M. (2013). Review of the application of
induced energy harvesting with a MEMS Conference. 2012. IEEE.
energy harvesting in buildings. Meas. Sci.
harvester for automotive applications. in Tech. 25, 012002.
Electron Devices Meeting (IEDM), 2011 IEEE 264. Roundy, S., and Tola, J. (2014). Energy
International. IEEE. harvester for rotating environments using 279. Çlelebi, M., Phan, L., and Marshall, R. (1993).
offset pendulum and nonlinear dynamics. Dynamic characteristics of five tall buildings
251. Zheng, Q., Tu, H., Agee, A., and Xu, Y. (2009). Smart Mater. Struct. 23, 105004. during strong and low-amplitude motions.
Vibration energy harvesting device based on Struct. Des. Tall Spec. 2, 1–15.
asymmetric air-spaced cantilevers for tire 265. Makki, N., and Pop-Iliev, R. (2012). Battery-and
pressure monitoring system. Proceedings of wire-less tire pressure measurement systems 280. Peigney, M., and Siegert, D. (2013).
Power MEMS, 403–406. (TPMS) sensor. Microsystem Tech. 18, 1201– Piezoelectric energy harvesting from traffic-
1212. induced bridge vibrations. Smart Mater.
252. Zhang, Y., Zheng, R., Shimono, K., Kaizuka, T., Struct. 22, 095019.
and Nakano, K. (2016). Effectiveness testing of 266. Makki, N., and Pop-Iliev, R. (2011).
a piezoelectric energy harvester for an Piezoelectric power generation for sensor 281. Khan, F.U., and Ahmad, I. (2016). Review of
automobile wheel using stochastic applications: design of a battery-less wireless energy harvesters utilizing bridge vibrations.
resonance. Sensors 16, 1727. tire pressure sensor. Proc. SPIE. https://doi. Shock Vib. https://doi.org/10.1155/2016/
org/10.1117/12.887112. 1340402.
253. Singh, K.B., Bedekar, V., Taheri, S., and
Priya, S. (2012). Piezoelectric vibration 267. Makki, N., and Pop-Iliev, R. Pneumatic tire- 282. Cao, M., and Zuo, L. (2014). Energy
energy harvesting system with an based piezoelectric power generation. in harvesting from building seismic isolation
adaptive frequency tuning mechanism for Proc. SPIE Vol. 2011. with multi-mode resonant shunt circuits. in

Joule 2, 1–56, April 18, 2018 55

 Please cite this article in press as: Yang et al., High-Performance Piezoelectric Energy Harvesters and Their Applications, Joule (2018), https://
doi.org/10.1016/j.joule.2018.03.011

ASME 2014 Dynamic Systems and Control vibrations for structural health monitoring of investigations on piezoelectric-based energy
Conference, (American Society of bridges. J. Micromech. Microeng. 21, 104005. harvesting from bridge vibrations under
Mechanical Engineers). travelling vehicles. Int. J. Mech. Sci. 119, 1–11.
292. McCullagh, J., Galchev, T., Peterson, R.,
283. Shen, W., Zhu, S., Zhu, H., and Xu, Y. (2016). Gordenker, R., Zhang, Y., Lynch, J., and Najafi,
301. Bhaskaran, P.R., Rathnam, J.D., Koilmani, S.,
Electromagnetic energy harvesting from K. (2014). Long-term testing of a vibration
and Subramanian, K. (2017). Multiresonant
structural vibrations during earthquakes. harvesting system for the structural health
frequency piezoelectric energy harvesters
Smart Struct. Syst. 18, 449–470. monitoring of bridges. Sensor. Actuator. Phys.
integrated with high sensitivity piezoelectric
217, 139–150.
284. Zhu, D., Beeby, S.P., Tudor, M.J., White, N.M., accelerometer for bridge health monitoring
and Harris, N.R. (2013). Novel miniature 293. Jung, H.-J., Kim, I.-H., and Jang, S.-J. (2011). applications. Smart Mater. Res. 2017, https://
airflow energy harvester for wireless sensing An energy harvesting system using the wind- doi.org/10.1155/2017/6084309.
applications in buildings. IEEE Sensors J. 13, induced vibration of a stay cable for powering
691–700. a wireless sensor node. Smart Mater. Struct. 302. Maruccio, C., Quaranta, G., DeLorenzis, L.,
20, 075001. and Monti, G. (2016). Energy harvesting
285. Khan, F.U. and I. Ahmad. Vibration-based from electrospun piezoelectric nanofibers
electromagnetic type energy harvester for 294. Erturk, A. (2011). Piezoelectric energy for structural health monitoring of a cable-
bridge monitoring sensor application. in harvesting for civil infrastructure system stayed bridge. Smart Mater. Struct. 25,
Emerging Technologies (ICET), 2014 applications: moving loads and surface strain 085040.
International Conference. 2014. IEEE. fluctuations. J. Intell. Mater. Syst. Struct. 22,
1959–1973. 303. Lee, S., Youn, B.D., and Jung, B.C. (2009).
286. Orfei, F., Mezzetti, C.B., and Cottone, F..
Robust segment-type energy harvester and
(2016). Vibrations powered LoRa sensor: An 295. Ali, S.F., Friswell, M., and Adhikari, S. (2011).
its application to a wireless sensor. Smart
electromechanical energy harvester working Analysis of energy harvesters for highway
Mater. Struct. 18, 095021.
on a real bridge. in SENSORS, 2016 IEEE. y bridges. J. Intell. Mater. Syst. Struct. 22, 1929–
1938.
287. Takeya, K., Sasaki, E., and Kobayashi, Y. 304. Rhimi, M., and Lajnef, N. (2012). Tunable
(2016). Design and parametric study on 296. Zhang, Y., Cai, S.C., and Deng, L. (2014). energy harvesting from ambient vibrations in
energy harvesting from bridge vibration using Piezoelectric-based energy harvesting in civil structures. J. Energy Eng. 138, 185–193.
tuned dual-mass damper systems. J. Sound bridge systems. J. Intell. Mater. Syst. Struct.
Vib. 361, 50–65. 25, 1414–1428. 305. Leland, E.S., and Wright, P.K. (2006).
Resonance tuning of piezoelectric vibration
288. McEvoy, T., Dierks, E., Weaver, J., Inamdar, S., 297. Xie, X., Wu, N., Yuen, K.V., and Wang, Q. energy scavenging generators using
Zimowski, K., Wood, K.L., Crawford, R.H., and (2013). Energy harvesting from high-rise compressive axial preload. Smart Mater.
Jensen, D. (2011). Developing innovative buildings by a piezoelectric coupled Struct. 15, 1413.
energy harvesting approaches for cantilever with a proof mass. Int. J. Eng. Sci.
infrastructure health monitoring systems. in 72, 98–106.
Proceedings of the 37th Design Automation 306. Kim, S.-H., Ahn, J.-H., Chung, H.-M., and
Conference, Parts A and B. 298. Cahill, P., Jaksic, V., Keane, J., O’Sullivan, A., Kang, H.-W. (2011). Analysis of piezoelectric
Mathewson, A., Ali, S.F., and Pakrashi, V. (2016). effects on various loading conditions for
289. Galchev, T., Kim, H., and Najafi, K. (2011). Effect of road surface, vehicle, and device energy harvesting in a bridge system.
Micro power generator for harvesting low- characteristics on energy harvesting from Sensor. Actuator. Phys. 167, 468–483.
frequency and nonperiodic vibrations. bridge–vehicle interactions. Comput. Aided Civ.
J. Microelectromech. Syst. 20, 852–866. Infrastruct. Eng. 31, 921–935. 307. Xie, X., Wang, Q., and Wang, S. (2015). Energy
harvesting from high-rise buildings by a
290. Tang, X. and L. Zuo. (2011). Simulation and 299. Cahill, P., Nuallain, N.A.N., Jackson, N., piezoelectric harvester device. Energy 93,
experiment validation of simultaneous vibration Mathewson, A., Karoumi, R., and Pakrashi, V. 1345–1352.
control and energy harvesting from buildings (2014). Energy harvesting from train-induced
using tuned mass dampers. in American Control response in bridges. J. Bridge Eng. 19,
Conference (ACC), 2011. IEEE. 04014034. 308. Xie, X., and Wang, Q. (2016). Design of a
piezoelectric harvester fixed under the
291. Galchev, T., McCullagh, J., Peterson, R., and 300. Karimi, M., Karimi, A., Tikani, R., and Ziaei- roof of a high-rise building. Eng. Struct.
Najafi, K. (2011). Harvesting traffic-induced Rad, S. (2016). Experimental and theoretical 117, 1–9.

56 Joule 2, 1–56, April 18, 2018