Piezoelectric energy harvesting using a diaphragm structure

Sunghwan Kim, William W. Clark , and Qing-Ming Wang**

Department of Mechanical Engineering, University of Pittsburgh
ABSTRACT
Energy harvesting using piezoelectric material is not a new concept, but its small generation capability has
not been attractive for mass energy generation. For this reason, little research has been done on the topic.
Recently, wearable computer concepts, as well as small portable electrical devices, are a few motivations
that have ignited the study of piezoelectric energy harvesting again. The theory behind cantilever type
piezoelectric elements is well known, but the transverse moving diaphragm elements, which can be used in
pressure type energy generation is not yet fully developed. The power generation in a diaphragm depends
on several factors. Among them, the thickness of each layer is important. In this paper, two diaphragm
structures, unimorph and regrouped unimorph, were used to calculate energy generation with varying
thickness ratio using piezoelectric constitutive equation. The results of this analysis are presented with an
eye toward guidelines for design of a useful energy harvesting structure.
Keywords: Piezoelectric, energy harvesting, diaphragm, unimorph

INTRODUCTION

Energy has been essential in building up modern society. It is required everywhere from the household light
bulb to a mission to Mars. Some energy can be seen, light for example, but most does not have a visible
form. Energy is defined in several ways, such as mechanical, electrical, and chemical. All of these
definitions are based on where the energy is stored.
Energy is stored everywhere. Heat, electricity, mechanical, chemical, photo and biomass forms of energy
are all stored differently, but can be converted from one to the other. Among many types of energy,
electricity is the most commonly used form for modern devices because it is easy to convert to other types.
The term power generation commonly refers to energy conversion from other energy forms to electrical
energy.
There are many ways to complete electrical conversions. Photocells convert light to electricity,
thermocouples convert heat to electricity, and magneto-electric generators convert mechanical energy to
electricity. These are all called power generators and are frequently used in electricity generation. Similar
to magneto-electric generators, piezoelectric generators (PEG) can also convert mechanical energy to
electrical energy.
Piezoelectric power generators (PEG) have many advantages over other conversion methods. PEGs consist
of piezoelectric ceramics, and electrodes which cover them. Because of their simplicity, PEGs can even be
made on the scale of micro electromechanical systems (MEMS). Another advantage is that the lifetime of
the system is almost unlimited if the applied force and external temperature are within the operational
range. Unlike the power generation methods that rely on heat conversion, a PEG presents no problems such
as heat isolation. In addition, the mechanical energy required for conversion can conceivably be obtained
from the PEGs environment. Even with these advantages, PEGs have been neglected for power
generation because of the small electrical output.
Recently, PEGs have seen regained interest in the power generation field for portable and low power
consuming devices. The merit of applying PEGs to these devices is that they can reduce the required

wclark@engr.pitt.edu; phone: 412-624-9794; fax 412-624-4846

qmwang@engrng.pitt.edu; phone: 412-624-4885; fax 412-624-4846

**

Smart Structures and Materials 2003: Smart Electronics, MEMS, BioMEMS, and Nanotechnology,
Vijay K. Varadan, Laszlo B. Kish, Editors, Proceedings of SPIE Vol. 5055 (2003)
2003 SPIE 0277-786X/03/$15.00

Downloaded From: http://proceedings.spiedigitallibrary.org/ on 10/05/2016 Terms of Use: http://spiedigitallibrary.org/ss/termsofuse.aspx

307

battery weight and possibly make the device self-powered by harvesting mechanical energy. To maximize
these advantages, there are problems to resolve such as how to design the PEG in order to optimize its
electrical energy production. To answer this question, the electrical-mechanical characteristics of
piezoelectric materials must be revisited.
For the analytical analysis of the piezoelectric material, a great amount of research has been done. Among
them, the constitutive equations has been established for the electrical and mechanical coupling in the PZT
ceramics [12] and multiple layered cantilever beam was analyzed for actuators [8,10,11] and sensors [9]
purpose. In the energy harvesting area, research has been conducted into various loading conditions, such
as impact and resonance. Researchers claim that PEGs lose performance under the resonance forcing
condition [4] or under the impact forcing condition [1,2].
Actual implementation of energy harvesting from human movement using PEGs has been studied [3,5].
Researchers inserted thunder actuators into shoes to harvest energy from walking. Polymer type
piezoelectric material (PVF2) was also investigated for this electrical energy conversion purpose [14].
Some research has also been done for an acoustic isolator application. Two diaphragms were used to
control noise. One was for energy harvesting, while the other was a control actuator which used harvested
energy from the other diaphragm [13].
In this paper, investigations of circular plate elements will be carried out to analyze their suitability for
piezoelectric energy generation. The interest here is in determining the factors that lead to maximum
electrical energy generation in relation to a given volume of material or applied mechanical force, pressure,
or stress.

THEORETICAL ANALYSIS

A diaphragm structure is a common structure for pressure sensors such as those that detect acoustic
pressure or hydraulic pressure. The analytical solution exists for a circular disk with clamped edges. This
solution follows some restrictions. From the general elasticity description of a thick plate, it is known that
there is no shear strain and no stress along thickness direction (z-dir). Another restriction is that the
deflection is much smaller than the thickness of the plate. In this paper, the energy method will be used to
calculate the charge generation.
The structure to be studied is shown (in cross-section) in Figure 1. The structure has a piezoelectric layer
bonded to a substrate. Different electrode patterns are studied, including the conventional diaphragm (full
electrode coverage) and modified diaphragm (regrouped electrodes) configurations.

Figure 1. Unimorph PZT Circular plate with electrode configuration and its polling direction
The moment equation which is needed for the calculation of the curvature is obtained from the elasticity
solution [7] for the clamped circular diaphragm.

308

Proc. of SPIE Vol. 5055

Downloaded From: http://proceedings.spiedigitallibrary.org/ on 10/05/2016 Terms of Use: http://spiedigitallibrary.org/ss/termsofuse.aspx

Po ( a 4  2a 2r 2 + r 4 )
64D
s2Wr

1
W
P
s
r
M r = D
+
= o (a 2  3r 2 + Oa 2  Or 2 )
sr 2
r sr 16
1 sWr
s2Wr Po 2
M R = D
+O
=
(a  r 2 + Oa 2  3Or 2 )
r sr
sr 2 16
Wr =

(1)

(2)

where Wr is the deflection of the diaphragm in the z direction, r is the distance from the center of the
diaphragm to the point of deflection, Q is the Poissons ratio, Po is pressure, and a is the radius of the
diaphragm. The constant D is related to the structure property but is not important here because it will be
canceled out when the deflection Equation (1) is substituted into the moment Equation (2) and these
moments are need to calculate the stresses in the PEG.
Unlike the cantilever beam case, there are two moment terms for diaphragms. One is in the radial direction
and the other is in the angular direction. Thus, there will be two curvatures and each curvature equation will
be in terms of both moment terms. However, for the common unimorph diaphragm, there is only one
neutral surface, since the surface depends only on the cross-section of the device, as the elastic properties
are the same for radial and angular direction. Therefore, the strain-curvature equation for the unimorph
diaphragm is described as

Fr = Sr ( z  zc ) ,

FR = SR ( z  zc )

(3)

where zc is the distance between the neutral surface and the centroid, z is the distance from the centroid to
the location of strain, U is curvature, H is strain, and subscripts r and T indicate direction.
In this paper, energy generation for two different types of unimorph diaphragms will be analyzed. The first
case will consist of a unimorph diaphragm that is completely covered with electrode. The second case will
be a unimorph diaphragm containing regrouped electrodes, that is, the top electrode is sectioned and one
area is repoled. As in the unimorph cantilever beam, the neutral surface of the unimorph diaphragm is not
on the middle surface. Also the neutral surface moves by the electric field changes but this change is
assumed to be small enough to be neglected. Only elastic properties of the PZT and substrate will be
considered to calculate neutral surface of the unimorph diaphragm.
2.1

Unmodified unimorph diaphragm

The moment in the diaphragm was calculated using piezoelectric constitutive equation and compared with
classical elasticity solution shown Equations (1) and (2). The constitutive equation [6] can be written for
the diaphragm and is
E
( Tr  OTR )  d31E 3
Fr = s11
E
( TR  OTr )  d31E 3
FR = s11

(4)

D3 = d31 ( Tr + TR ) + T33E 3
E
E
/ s11
where O ( = s12
) is Poissons ratio and the poling direction is downward. The radial direction is 1,

the angular direction is 2, and the direction perpendicular to the surface is 3. The subscripts r and R are
used instead of 1 and 2. In Equation (4), T is stress, T is the permittivity of PZT at constant stress, d 31 is
E
the piezoelectric constant, D3 is charge density, E 3 is electric field strength, and s11
is elastic compliance
constant for PZT at constant electric field. Using this piezoelectric constitutive equation, stress can be
described in terms of strain and electric field. For the piezoelectric layer, the stress is

Proc. of SPIE Vol. 5055

Downloaded From: http://proceedings.spiedigitallibrary.org/ on 10/05/2016 Terms of Use: http://spiedigitallibrary.org/ss/termsofuse.aspx

309

Trp =

1
( Fr + OFR + ( 1 + O )d 31E 3 )
E
s11
(1  O 2 )

TRp =

1
( OFr + FR + ( 1 + O ) d31E 3 )
E
s11 ( 1  O 2 )

and for the non-piezoelectric layer

Trm =

1
( Fr + OFR )
sm ( 1  O 2 )

TRm =

1
( OFr + FR )
sm ( 1  O 2 )

(5)

(6)

Here, sm is the elastic compliance constant for the substrate layer

The strain in Equations (5) and (6) can be written in terms of the curvature of the diaphragm and thus can
be related to the applied pressure. As in the cantilever beam case, the curvature equation may be found
from the moment equation which is the integration of the stress and the distance from the neutral surface.
The moment equation from the above stress equation is described as

Mr =
MR =

hp

Fr + OFR

( 1 + O ) d31E 3

0 ( 1  O 2 )s11E + ( 1  O 2 )s11E
hp

( z  zc )dz +

OFr + FR + ( 1 + O )d31E 3 ( z  z )dz +

c
E
2
( 1  O 2 )s E
( 1  O )s11
11

Fr + OFR

h ( 1  O 2 )sm ( z  zc )dz
m

OFr + FR
hm ( 1  O 2 )sm
0

( z  zc )dz

(7)

where hm is the thickness of the substrate layer, and hp is the thickness of the PZT layer
The strain-curvature equations can be obtained by substituting equation (3) into the equation (7). After
rearranging, the curvature is
Sr = 

E
E
E
12s11
sm ( s11
hm + sm hp )
6d31s11
sm hp hm ( hm + hp )
( M r  OM R ) +
E3
B31
B31

E
E
E
12s11
sm ( s11
hm + sm hp )
6d31s11
sm hp hm ( hm + hp )
( M R  OM r ) +
SR = 
E3
B31
B31

(8)

E
E
E
2
sm hm hp 3 + 6s11
sm hm 2hp 2 + 4s11
sm hm 3hp + hm 4s11
where B31 = hp 4sm 2 + 4s11

Energies of each layer are described differently. For the diaphragm structure, the energies in a small
volume can be described as
d31 ( Trp + TRp )
1
1
1
dU p = Fr Trp + FR TRp 
E 3 + T33E 32 for piezoelectric
2
2
2
2
(9)
1
1
for non-piezoelectric
dU m = Fr Trm + FR TRm
2
2
The stresses used in Equation (9) can be replaced with Equation (5) and (6). Then, using the straincurvature equations (3) and (8), the above energy equations can be written as a function of moment and
electric field. This moment can also be replaced by equation (2). Therefore, the total system energy can be
written in terms of pressure and electric field.

310

Proc. of SPIE Vol. 5055

Downloaded From: http://proceedings.spiedigitallibrary.org/ on 10/05/2016 Terms of Use: http://spiedigitallibrary.org/ss/termsofuse.aspx

U =

2Q

hp

0 0 0
=

dU pdz +

dU mdz rd Rdr

h

2
E

Qa 6s11
sm Sh ( 1  O 2 ) 2 T33 Qa 2hp
3s E 2s h 2h ( h + hm )
2
2
E 32
1 
1  11 m m p p
 1 K 31
Po +

( 1  O )
32B31
2
Sh B31

(10)

E
. K 33 is the electromechanical coupling coefficient and is defined as
where Sh = hpsm + hms11
2
, E
K 31
= d312 / 
s11 . Electric field is constant throughout the structure, since the electrodes cover the whole

surface. Therefore, the field is constant throughout the diaphragm structure. After integrating the total
energy, Equation (10), the electric field ( E 3 ) can be replaced by V / hp . The generated charge from the
external conditions can be found by differentiating the total energy with respect to the voltage (V) which
gives

Q =

2
3s E 2s h 2h ( h + hm ) 2
T33 Qa 2
2
1 
1  11 m m p p
K 31 V
( 1  O )
hp
Sh B31

(11)

This is the general equation of charge generation. When converting mechanical energy to electrical energy,
there is no applied voltage. Thus none of the energy was converted from the applied pressure.
QGen = 0

(12)

The calculated charge is only the function of applied voltage. Therefore, is unmodified diaphragm behaves
as a planar capacitor. The capacitance is the electrical relation between charge and voltage. Therefore, the
capacitance can be described as

C free =

2
3s E 2s h 2h ( h + hm ) 2
T33 Qa 2
2
1 
1  11 m m p p
K 31
( 1  O )
hp
Sh B31

(13)

The term outside of the parenthesis is the description of the capacitance when there are no boundary
restrictions. Usually capacitance with no boundary condition is

Capacitance=

,

Qa 2
Permittivity Area
=
hp
Distance

(14)

Due to the clamped boundary condition, the permittivity constant has been changed. The permittivity of the
clamped diaphragm structure can be described as
2

3s E 2s h 2h ( h + hm ) 2
2
1  11 m m p p
new = T33 1 
(15)
K 31
( 1  O )
Sh B31

Again, there is no mechanical electrical coupled term in the charge equation (11), mechanical energy
derived from uniformly applied pressure can not be converted to electrical energy. Therefore, the
conventional unmodified diaphragm structure is not effective for energy harvesting. There is no charge
generation from the pressure, nor does voltage appear by applying pressure.
Vgen =

Qgen
=0
C free

(16)

Proc. of SPIE Vol. 5055

Downloaded From: http://proceedings.spiedigitallibrary.org/ on 10/05/2016 Terms of Use: http://spiedigitallibrary.org/ss/termsofuse.aspx

311

The generated energy equation from pressure is

1
UGen = C freeVGen 2 = 0
2

(17)

Therefore, the conventional diaphragm should be modified to obtain an energy harvesting function. It is
clear that even though there is no net charge generated, there is stress in the diaphragm. In fact, locally
there is substantial stress generated in the diaphragm due to the applied pressure (Figure 2), but some of it
results in positive charge and some negative. When integrated (equation (10)), the net generated electrical
energy is zero.
Expanded (Upper surface)

Compressed (Upper surface)

Figure 2. Deflection and stress of diaphragm

To support above idea, the partial energy of the diaphragm was calculated and compared to the
equation(10). The energy of the arbitrary volume inside diaphragm can obtained by integrating from zero to
r1 where 0 d r1 d a . Then the energy equation is
U =

r1

2Q

hp

0 0 0

dU pdz +

dU mdz rd Rdr

h

2
2
E2

r12S ,
2 3S11 Sm hp hm  hp  hm  2 2
K31 V
1
1

Sh B31
2hp 1 Q

2
E
2
2
3 S r1 d31S11 S m hm 1 Q   hp  hm   a  r1 

PV
o
B31
4

S r12 S11E Sm S h 1 Q 2  3Q  r14  a 4   6r12 a 2 1 Q   5r14  3a 4

(18)

64 B31

This equation (18) shows that the mechanical-electrical coupling term approaches zero when r1 goes to a.
And when r1 a , the equation (10) and the equation (18) are identical. By inspecting the coupled energy
term in the equation(18) , the coupled energy has a following relation.

U couple v

r12  a 2  r12 

Therefore, the maximum coupled energy can be obtained when r1

2.2

(19)
2 2 | 0.707

Regrouped diaphragm

In this analysis, the diaphragm will be divided into two areas to separate the regions of positive and
negative stress (and thus positive and negative charge). One is the inner region where r < m and the other
is the outer region where r > m . Here, m is a value between 0 and a. These inner and outer regions
electric field directions will be opposite to each other as a result of the stress distribution from constant
pressure (Figure 3).
312

Proc. of SPIE Vol. 5055

Downloaded From: http://proceedings.spiedigitallibrary.org/ on 10/05/2016 Terms of Use: http://spiedigitallibrary.org/ss/termsofuse.aspx

P
m

Figure 3. Cross section of the PZT layer of the regrouped diaphragm

All the equations are the same as previously calculated for the unimorph diaphragm except for the electric
fields direction. The inner region will use the same electric field direction, and the outer region will use the
opposite electric field direction. Then the constitutive equation for the outer region is same as the equation
of inner area (4) except the electric field.
E
( Tr  OTR ) + d31E 3
Fr = s11
E
( TR  OTr ) + d31E 3
FR = s11

D3 = d31 ( Tr + TR ) 

(20)

T33E 3

The strain equations are different between inner and outer regions because the electric field condition is
different. Because of this difference, two sets of strain equations are formed. As in the equation (3), the
strain equations are related to the curvature and are

Fri = Sri ( z  zc ), FRi = SRi ( z  zc )

when r < m, Inner region

Fro = Sro ( z  zc ), FRo = SRo ( z  zc )

when r > m, Outer region

(21)

Stress distributions for the inside of the PZT area are the same as shown in equation (5) and for the
substrate layer are shown in equation (6). The poling direction of the outer PZT layer is opposite to that of
inner PZT layer, so the sign of the piezoelectric constant is different. Therefore, the stress equations for the
outer layer are
1
( Fr + OFR  ( 1 + O ) d31E 3 )
Tro = E
s11 ( 1  O 2 )
(22)
1
(
)
(
)
1
TRo = E
OF
F
O
d
E
+

+
31 3
r
R
s11 ( 1  O 2 )
As in the unmodified diaphragm section, the moment equations are needed to calculate a curvature. The
inner region and previous unimorph diaphragm are designed to be identical. The moment equations for the
inner region are shown in equation(7). The moment equations for the outer region are

M ro =
M Ro =

hp

Fr + OFR

( 1 + O )d31E 3

0 F + OF
R
r
( z  zc )dz + h
( z  zc )dz
2
m ( 1  O ) sm

0 OF + F
( 1 + O )d 31E 3
R
r

( z  zc )dz +
( z  zc )dz

2
2
E

hm
( 1  O )s11
( 1  O )sm

0 ( 1  O 2 )s11E  ( 1  O 2 )s11E
hp

OFr + FR
( 1  O 2 )s E
11

(23)

The curvature equations for inner and outer regions should be different because the electric field directions
are different.
As before, the curvature can be found in terms of moments and electric field. Inner area curvature equations
are identical as equation (8) and the outer area curvature equations from above equation (23) are

Proc. of SPIE Vol. 5055

Downloaded From: http://proceedings.spiedigitallibrary.org/ on 10/05/2016 Terms of Use: http://spiedigitallibrary.org/ss/termsofuse.aspx

313

Sro = 
SRo

E
E
6d 31s11
sm hp hm ( hm + hp )
12s11
sm S h
( M r  OM R ) 
E3
B31
B31

(24)

E
6d31s11
sm hp hm ( hm + hp )
12s E s S
=  11 m h ( M R  O M r ) 
E3
B31
B31

E
E
E2
sm hm 2hp2 + 4s11
sm hm hp ( hp 2 + hm 2 ) + hm 4s11
where, B31 = hp 4sm 2 + 6s11

There are four curvature equations. Two are for the inner region, and the other two are for the outer region
curvature. Thus the energy equations for the inner and outer regions are also different. Energy equations for
the inner area are already shown in equation (9). The energies in the small volume for the outer area is
described as
d31 ( Trp + TRp )
1
1
1
dU po = Fr Trp + FR TRp +
E 3 + T33E 32
2
2
2
2
(25)
1
1
dU mo = Fr Trp + FR TRp
2
2
Combining those energy equations for the inner and outer regions, the total energy in the structure can be
written as
U =

2Q

hp

0 0 0

dU pidz +

dU midz rd Rdr +

h

2Q

hp

m 0 0

dU podz +

h

dU modz rd Rdr

2
,
E2


Qa 2 1
3s11
sm hphm 2 ( hp + hm )
1
2 2

1
=

K 31 V

hp 2 1  O
Sh B31

E
E
3Qm 2d31s11
sm hm ( 1 + O ) (hp + hm ) ( a 2  m 2 )
sm S h ( 1  O 2 ) 2
Qa 6s11
VPo +
Po
2B31
32B31

(26)

The first integral term is for the inner region and the second integral term is for the outer region. This
energy equation can be reduced to unmodified diaphragm energy equation (10) by substituting m=a.
Differentiating with respect to voltage (V) gives the charge equation.

Q =

2
,


Qa 2
3s E 2s h h 2 ( hp + hm )
2
K 312 V
1 
1  11 m p m

hp
Sh B31
1  O

E
sm hm ( 1 + O ) (hp + hm ) ( a 2  m 2 )
3Qm 2d31s11

Po
2B31

(27)

With this regrouped diaphragm, a pressure term appears as expected. The generated charge with no applied
external electric field is
QGen = 

E
3Qm 2d31s11
sm hm ( 1 + O ) (hp + hm ) ( a 2  m 2 )
Po
2B31

(28)

From the relation Q=CV, the open circuit capacitance is

C free =

314

2
,


Qa 2
3s E 2s h h 2 ( hp + hm )
2
K 312
1 
1  11 m p m

1  O
hp
Sh B31

Proc. of SPIE Vol. 5055

Downloaded From: http://proceedings.spiedigitallibrary.org/ on 10/05/2016 Terms of Use: http://spiedigitallibrary.org/ss/termsofuse.aspx

(29)

and the voltage that appears on the electrodes is

VGen

Qgen
=
=
C free

E
3m 2d31s11
sm hm hp ( a 2  m 2 )( 1 + O ) ( h

2T33B31a 2 1 

+ hp )

3s E 2s h h 2 ( hp + hm )
2
K 312
1  11 m p m
1  O
Sh B31

(30)

Po

Thus, the electric energy generated by external pressure is

UGen

E2 2
9Qm 4d312s11
sm Sh hm 2hp ( 1  O )( 1 + O )2 ( a 2  m 2 ) ( h m + hp )
1
Po 2 (31)
= C freeVGen 2 =
2
2
2
2
T
E2
2
(
)
8 33B31a ( 1  O Sh B31  2(Sh B31  3s11 sm hp hm ( hp + hm ) )K 31 )

Proc. of SPIE Vol. 5055

Downloaded From: http://proceedings.spiedigitallibrary.org/ on 10/05/2016 Terms of Use: http://spiedigitallibrary.org/ss/termsofuse.aspx

315

2.3

Comparison

Up to this point in this paper, the energy equations for un-modified and regrouped diaphragms were
introduced. These equations were used to carry out a parametric study of the diaphragm energy harvesting
device. For a fixed applied pressure on the 50.8mm diaphragm with aluminum substrate layer, the resulting
generated voltage and charge were calculated for a range of thickness and regrouping radii. These results
are shown in Figure 4 and Figure 5.

Figure 4. Voltage (left) and charge (right) output of piezoelectric diaphragm generator for constant uniform
pressure and same piezoelectric material. ST is the thickness ratio ( hm / hp ) and RG is the regroup location
( m / a ). Pressure=9.7kPa, hp=0.127mm

Figure 5. Capacitance of the piezoelectric generator. ST is the thickness ratio ( hm / hp ) and RG is the
regroup location ( m / a ).
The left plot in Figure 4 shows the generated voltage as piezoelectric layer thickness and regrouping radius
changes. The right plot in Figure 4 shows the generated charge. The un-modified diaphragm case can be
obtained from regrouped results with RG=1. As shown in the plots, the maximum voltage and charge were
generated when regrouped at m ! 0.707a (RG=0.707) and the capacitance was not affected by the
regrouping location. As expected, the generated voltage, and charge depend on the thickness of the
substrate layer because the thickness of the substrate layer alters the location of the neutral surface and this
change alters the stress distribution. This stress distribution change also changes boundary condition of the
PZT material, the permittivity of the PZT also changes with the substrate layer thickness change. Figure 5
shows the variation in the capacitance as a function of regrouping radius and piezoelectric layer thickness
ratio. The capacitance was not affected by regrouping, but is heavily influenced by layer thickness.
The optimal values were found to develop Table 1. Four cases, un-modified, regrouped to have maximum
voltage, regrouped to have maximum capacitance, and regrouped to have maximum energy were compared.
In the piezoelectric problems, the best case is to increase the capacitance and lower the electric field, since
316

Proc. of SPIE Vol. 5055

Downloaded From: http://proceedings.spiedigitallibrary.org/ on 10/05/2016 Terms of Use: http://spiedigitallibrary.org/ss/termsofuse.aspx

the piezoelectric material can be damaged by its own generated (or applied) electric field. Therefore, the
maximum acceptable voltage should be included in the PEG design. For each case, 3 different piezoelectric
layer thicknesses were considered, and the limiting voltage (electric field) was found for each case. With
this voltage limitation, the pressure and substrate layer which can generate that limit voltage at optimal
location (RG=0.707) were found.
Table 1. Theoretical voltage, energy, and capacitance of the three different PEGs at their limit pressure,
(PZT5H, 50.8mm diaphragm. RG = m / a , ST hm / hp )
Numerical Calculation
Case
Po=5.98kPa
Po=26.06kPa Po=367.44kPa
Units
hp=0.127mm
hp=0.267mm
hp=1mm
VGen
0
0
0
V
RG = 1
U Gen
mJ
0
0
0
Unmodified
ST = 1.81
C free
374.3
178.3
47.5
nF
VGen
38.10 (38)
79.01 (80)
297.13 (300)
V
Regrouped
RG = 0.707
U Gen
mJ
0.216
0.443
1.67
- Optimal
ST = 0.057
Energy
C free
298.1
142.0
37.8
nF
VGen
16.16 (38)
33.30 (80)
125.23 (300)
V
Regrouped
RG = 0.707
U
mJ
0.049
0.099
0.372
- Max
Gen
ST = 1.81
Capacitance
C free
374.3
178.3
47.5
nF
Regrouped
- Max
Voltage

RG = 0.707
ST = 0.45

VGen

38.42 (38)

79.68 (80)

299.63 (300)

U Gen

0.213

0.436

1.65

mJ

C free

288.6

137.4

36.7

nF

The parenthesis in Table 1 indicates maximum permitted voltage (based on PZT5H which has a limit of
300 V/mm). As the PZT layer becomes thicker, the voltage limit also increases. Therefore, the thickness of
the PZT layer can be an indicator for the pressure limit. Larger pressure can generate more energy. The
generated voltage increases linearly as pressure increase (Equation(30)) and the energy generation is
described as CV 2 / 2 . Even with these relations, the voltage limit prevents large energy generation. As
shown in Table 1, the pressure increment actually increases the energy generation but reduces the pressureenergy efficiency.
The optimum regroup location is not affected by the piezoelectric layer thickness nor substrate layer. The
optimal regroup location was always at RG=0.707. This location is exactly where the neutral surface
crosses the point W=0 in the deflected shape from uniformly applied pressure.
There are small differences in their optimum substrate layer thickness between optimal energy and optimal
voltage case. These differences are caused by the capacitance. The capacitance changes are very small
compared to the voltage change when varying substrate layer thickness. The optimal substrate layer
thickness is close to the optimal thickness for the voltage.

CONCLUSIONS

Theoretical calculations show that the change in stress polarity in a loaded structure can be utilized to
enhance or optimize energy harvesting by properly poling segments of the piezoelectric material. The
optimal location of the regrouped electrode interface was found theoretically. Substrate layer thickness ratio
with respect to the piezoelectric layer was also studied. The substrate layer thickness changes the location
of the neutral surface, which changes the stress distributions in the piezoelectric layer. This substrate layer

Proc. of SPIE Vol. 5055

Downloaded From: http://proceedings.spiedigitallibrary.org/ on 10/05/2016 Terms of Use: http://spiedigitallibrary.org/ss/termsofuse.aspx

317

thickness also changes the boundary condition of the piezoelectric material and changes capacitance of the
PEG.
The capacitance of the PEG has a important role in the piezoelectric energy harvesting device design. But
in the clamped diaphragm structure, the capacitance changed very small compared to the voltage changes
and does very little to effect the total energy generation.

REFERENCES
1.
2.

3.
4.

5.

6.
7.
8.

9.
10.

11.
12.
13.

14.

318

Xu,C., Akiyama, M., Nonaka, K., Shobu, K., and Watanabe, T., Electrical Output Performance of
PZT-Based Piezoelectric Ceramics, IEEE 1996,pp.967-970
Chao-Nan Xu, Morito Akiyama, Kazuhiro Nonaka, and Tadahiko Watanabe, Electrical Power
Generation Characteristics of PZT Piezoelectric Ceramics, IEEE Trans. Ultrason.,Vol.,45, No.4,
July 1998,pp1065-1998
Kymissis, J., Kendall, C., Paradiso, J., Gershenfeld, N., Parasitic Power Harvesting in Shoes,
Wearable Computers, 1998. Digest of Papers. Second International Symposium on , 1998, pp132-139
Goldfarb, M.., Jones, L.D. On the Efficiency of Electric Power Generation with Piezoelectric
Ceramic, To appear in the ASME Journal of Dynamic System, Measurement, and Control, Version:
May 6, 1999
Shenck, N., A Demonstration of Useful Electric Energy Generation from Piezoceramics in a Shoe,
MS Thesis, Dept. of Electric Engineering and Computer Science, Massachusetts Institute of
Technology, April 1998
IEEE standard 176, IEEE Standards on Piezoelectricity, The Institute of Electrical and Electronics
Engineers, 1978
Vinson, J.R., Structural Mechanics: The Behavior of Plates and Shells, A Wiley-Interscience
publication, 1974
Wang, Q. M., Cross, L. E., Constitutive Equations of Symmetrical Triple Layer Piezoelectric
Benders, IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency control, Vol. 46, No.6,
1999, pp.1343-1351
Wang, Q. M., Du, X. H., Xu, B. Cross, L.E., Theoretical analysis of the sensor effect of cantilever
piezoelectric benders, Journal of Applied Physics, Vol.85, No.3, 1999, pp. 1702-1712
Wang, Q. M., Du, X. H., Xu, B. Cross, L.E., Electromechanical Coupling and Output Efficiency of
Piezoelectric Bending Actuators, IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency
control, Vol. 46, No.3, 1999, pp.638-646
Smits, J. G., Choi, W., The Constituent Equations of Piezoelectric Heterogeneous Bimorphs, IEEE
Transactions on Ultrasonics, Ferroelectrics, and Frequency control, Vol. 38, No.3, 1991, pp. 256-270
Mason, W. P., Electrostrictive Effect in Barium Titanate Ceramics, Physical Review, Vol. 74, No.
9, 1948, pp.1134-1147
Mandurino. P., Bianc. G. B., Garbin, R., Energy Recovering PZT-Panel for Noise Reduction:
Numerical Analysis for Fully Coupled Fluid-PZT Structure-Electric Interaction, Experimental
Verification, SPIE 8th annual international conference on Smart structures and smart materials, Los
Angeles, CA., 2001, March, 4-8
Schmidt, V.H., Piezoelectric Energy Conversion in Windmills, Ultrasonics symposium, 1992,
pp.897-904

Proc. of SPIE Vol. 5055

Downloaded From: http://proceedings.spiedigitallibrary.org/ on 10/05/2016 Terms of Use: http://spiedigitallibrary.org/ss/termsofuse.aspx