Fundamentals of Piezoelectricity

Introductory Course on Multiphysics Modelling

TOMASZ G. Z IELI ŃSKI

bluebox.ippt.pan.pl/˜tzielins/

Table of Contents
1 Introduction 1
1.1 The piezoelectric effects . . . . . . . . . . . . . . . . . . 1
1.2 Simple molecular model of piezoelectric effect . . . . . . 2

2 Equations of piezoelectricity 3
2.1 Piezoelectricity viewed as electro-mechanical coupling . 3
2.2 Field equations of linear piezoelectricity . . . . . . . . . 4
2.3 Boundary conditions . . . . . . . . . . . . . . . . . . . . 5
2.4 Final set of partial differential equations . . . . . . . . . 6

3 Forms of constitutive law 6

3.1 Four forms of constitutive relations . . . . . . . . . . . . 6
3.2 Transformations for converting constitutive data . . . . . 8
3.3 Piezoelectric relations in matrix notation . . . . . . . . . 8

4 Thermoelastic analogy 9

1 Introduction

1.1 The piezoelectric effects

Observed phenomenon
Piezoelectricity is the ability of some materials (notably crystals and certain
ceramics) to generate an electric charge in response to applied mechanical
stress. If the material is not short-circuited, the applied charge induces a volt-
age across the material.

Reversibility. The piezoelectric effect is reversible, that is, all piezoelectric materials
exhibit in fact two phenomena:
2 Fundamentals of Piezoelectricity ICMM lecture

1. the direct piezoelectric effect – the production of electricity when stress is ap-
plied,
2. the converse piezoelectric effect – the production of stress and/or strain when
an electric field is applied. (For example, lead zirconate titanate crystals will
exhibit a maximum shape change of about 0.1% of the original dimension.)

Some historical facts and etymology

 The (direct) piezoelectric phenomenon was discovered in 1880 by the broth-
ers Pierre and Jacques Curie during experiments on quartz.
 The existence of the reverse process was predicted by Lippmann in 1881
and then immediately confirmed by the Curies.
 The word piezoelectricity means “electricity by pressure” and is derived from
the Greek piezein, which means to squeeze or press.

1.2 Simple molecular model of piezoelectric effect

Figure 1 presents a simple molecular model which – together with the discussion below
– explains the direct piezoelectric effect.

− +
− +− +− +
− +
− +− +− +

+ ± − + − + − − +− +− +

− +− +− +

− + − +− +− +

− + − +− +− +

− +− +− +

± − +

neutral molecule small dipole

F IGURE 1: Simple molecular model for explaining the piezoelectric effect: an unper-
turbed molecule (left), the molecule subjected to an external force (middle), a po-
larizing effect on the material surfaces (right).

Before subjecting the material to some external stress (see Figure 1(left)):
 the centres of the negative and positive charges of each molecule coincide,
 the external effects of the charges are reciprocally cancelled,
 as a result, an electrically neutral molecule appears.
 ICMM lecture Fundamentals of Piezoelectricity 3

After exerting some pressure on the material (see Figure 1(middle)):

 the internal structure is deformed,
 that causes the separation of the positive and negative centres of the molecules,
 as a result, little dipoles are generated.
Eventually (see Figure 1(right)):
 the facing poles inside the material are mutually cancelled,
 a distribution of a linked charge appears in the material’s surfaces and the mate-
rial is polarized,
 the polarization generates an electric field and can be used to transform the me-
chanical energy of the material’s deformation into electrical energy.

2 Equations of piezoelectricity

2.1 Piezoelectricity viewed as electro-mechanical coupling

Piezoelectricity is an electro-mechanical phenomenon which couples the elastic be-

haviour of a material to its behaviour as a dielectric, and all this is done through the
piezoelectric effects (see Figure 2). Therefore, a mathematical model of piezoelectricity
would consists of:
(M) mechanical equations describing elastodynamics (or, in a particular case: elasto-
statics) of a linearly elastic material,
(E) electrical equations describing electrostatic behaviour of a dielectric material.
Scalar, vector, and tensor quantities
In Table 1, all mathematical quantities used by the piezoelectric model are grouped
together (mostly in pairs, relating a mechanical quantity to an electrical counterpart).
Apart from the quantities relevant for elastodynamics (M) and electrostatics (E), the
only additional one is the third-order tensor of piezoelectric material constants.
TABLE 1: Scalar, vector, and tensor quantities for piezoelectric equations
(M) – mechanical behaviour (E) – electrical behaviour (i, j, k, l = 1, 2, 3)
h i
N
ui – [m] fi – the mechanical body forces
h the mechanical displacements (M) 3
(M)
i hm i
J C
(E) ϕ – V = C the electric field potential (E) q– m3
the electric body charge
m h i
(M) Sij –
h m the istrain tensor
kg
(M) %– m3
the mass density
V N
(E) Ei –
m = C the electric field vector
h i
N
(M) cijkl – 2 the elastic constants
hm i
h i
N
(M) Tij – 2 the stress tensor C
hm i ekij – 2 the piezoelectric constants
C hm
(E) Di – the electric displacements
i
F
m2 (E) ij – m = VCm the dielectric constants
4 Fundamentals of Piezoelectricity ICMM lecture

(M)

ELASTIC material

+
Piezoelectric Effects

(E)

DIELECTRIC material

F IGURE 2: Piezoelectric effects couple the mechanical problem (M) of elastic mate-
rial and the electrical problem (E) of dielectric into a multi-physics problem

2.2 Field equations of linear piezoelectricity

The field equations for the piezoelectric model are given below. They are governing
equations for the mechanical (M) and electrical (E) sub-problems, namely: elastody-
namics and electrostatics), coupled solely inside the constitutive relations.
Although the mechanical problem is modelled as fully dynamic (i.e., inertial forces are
involved), its electrical counterpart is quasi-static, namely:
 it is assumed that the electric field changes in time sufficiently slowly, so that no
significant magnetic field is induced;
 therefore, electrodynamic effects are not involved and purely electrostatic equa-
tions are used instead.
Remark: A slow change in time from the perspective of a quasi-static (piezo-)electric
problem means, for example, time-harmonic variations up to tens of kilohertz, which is,
in fact, rather fast from the mechanical “point of view”.
All field equations of piezoelectricity are presented here in the index notation and also
in the “vector (bold-symbol) notation” – for the elasto-dynamic (electrically quasi-static)
case. The fully static case is obtained by simply assuming that all fields do not depend
on time. Then, the equations of motion (1) presented below are reduced to the static
case by simply noting that the mechanical accelerations are zero, namely, üi ≡ 0. The
electrostatic equation (2) remains virtually the same.

(M) Equations of motion (Elastodynamics)

Tij|j + fi = % üi or ∇ · T + f = % ü (1)

(E) Gauss’ law (Electrostatics)

Di|i − q = 0 or ∇·D−q =0 (2)
 ICMM lecture Fundamentals of Piezoelectricity 5

The kinematic relations (3) define the strain tensor as a symmetric part of the gradient
of mechanical displacements, whereas the Maxwell’s law (4) states that the electric
vector field is the (negative) gradient of electric potential.

(M) Kinematic relations

Sij = 21 (ui|j + uj|i ) 1


or S= 2
∇u + ∇uT (3)

(E) Maxwell’s law

Ei = −ϕ|i or E = −∇ϕ (4)

The electro-mechanical coupling appears only in the constitutive relations where the
coupling terms involve the third-order tensor of piezoelectric coupling coefficients ekij :

Constitutive equations – with piezoelectric coupling

Tij = cijkl Skl − ekij Ek or T = c : S − eT · E (5)

Dk = ekij Sij + ki Ei or D =e:S−·E (6)

2.3 Boundary conditions

Boundary conditions in this electro-mechanical problem of piezoelectricity are “uncou-

pled”, in that way that the standard mechanical conditions are applied separately from
the electrical conditions, and this also entails that the partition of boundary into dis-
junctive parts, dedicated to various essential and natural boundary conditions, can be
done completely independently for mechanical and electrical sub-problems.

Boundary conditions (“uncoupled”)

(essential) (natural)
(M) mechanical : ui = ûi or Tij nj = F̂i (7)
(E) electrical : ϕ = ϕ̂ or Di ni = −Q̂ (8)

ûi , ϕ̂ – the specified mechanical displacements

h i [m] and electric
h potential
i [V]
N C
F̂i , Q̂ – the specified surface forces m2 and surface charge m2
ni – the outward unit normal vector components
6 Fundamentals of Piezoelectricity ICMM lecture

2.4 Final set of partial differential equations

The kinematic relations (3) and Maxwell’s law (4) were used for the constitutive equa-
tions, which were then used for the equations of motion (1) and Gauss’ law (2), so
that the only dependent variables left are the mechanical displacements ui and electric
potential ϕ – the primary dependent variables. The final set of piezoelectric equa-
tions is given below. It forms an Initial Boundary-Value Problem (IBVP) for the primary
unknown fields, together with the boundary conditions (7) and (8), and the initial con-
ditions (at time t = 0): ui (x, t = 0) = u0i (x), u̇i (ξ, t = 0) = vi0 (x) (i = 1, 2, 3), where u0i
and vi0 are the initial values for the fields of displacement and velocity.

Piezoelectric equations in primary dependent variables

Coupled field equations for mechanical displacement (u) and electric potential (ϕ)
in a piezoelectric medium are as follows:
   
−% ü + ∇ · c : (∇u) + ∇ · e · (∇ϕ) + f = 0 , (9)
   
∇ · e : (∇u) − ∇ ·  · (∇ϕ) − q = 0 ; (10)

or, in index notation and assuming constant material properties:

−% üi + cijkl uk|lj + ekij ϕ|kj + fi = 0 [3 eqs. (in 3D)] , (11)

ekij ui|kj − kj ϕ|kj − q = 0 [1 eq.] . (12)

In a general three-dimensional case, this system contains 4 partial differential

equations in 4 unknown fields (4 DOFs in FE model), namely, three mechanical
displacements and an electric potential:

ui = ? (i = 1, 2, 3) , ϕ =?

3 Forms of constitutive law

3.1 Four forms of constitutive relations

There are 4 mathematically-equivalent forms of piezoelectric constitutive rela-

tions, each one involves one of 4 different third-order tensors of piezoelectric coupling
coefficients – see Table 2 and equations below.
1. Stress-Charge form:
T = cE=0 : S − eT · E ,
(13)
D = e : S + S=0 · E .
 ICMM lecture Fundamentals of Piezoelectricity 7

TABLE 2: Forms of piezoelectric constitutive relations, related piezoelectric tensors,

and the naming convention
h i
Stress mN2 Strain m
 
m
h i
e d
“Charge” mC2 T , D ←−−−−−−− (S, E) S, D ←−−−−−−− (T , E)
cE=0 , S=0 sE=0 , T =0 (“voltage”)
h i q g
V
“Voltage” m T , E ←−−−−−−− (S, D) S, E ←−−−−−−− (T , D)
−1
cD=0 , S=0 sD=0 , −1
T =0
(“charge”)
(strain) (stress)

2. Stress-Voltage form:
T = cD=0 : S − q T · D ,
(14)
E = −q : S + −1
S=0 · D .

3. Strain-Charge form:
S = sE=0 : T + dT · E ,
(15)
D = d : T + T =0 · E .

4. Strain-Voltage form:
S = sD=0 : T + g T · D ,
(16)
E = −g : T + −1
T =0 · D .

Here, the following tensors of constitutive coefficients appear:

 fourth-order tensors of elastic material constants:
h i h i
m2
stiffness c mN2 , and compliance s = c−1 N , obtained in the absence of electric
field (E=0) or charge (D=0);
 second-order tensors of dielectric material constants:
h i
F
, and its inverse −1
m
electric permittivity  m F , obtained in the absence of
mechanical strain (S=0) or stress (T =0);
 third-order tensors of piezoelectric coupling coefficients:
h i
C
e m2
– the piezoelectric coefficients for Stress-Charge form,
h i
m2
q C – the piezoelectric coefficients for Stress-Voltage form,
h i
C
d N – the piezoelectric coefficients for Strain-Charge form,
h i
N
g C – the piezoelectric coefficients for Strain-Voltage form.
8 Fundamentals of Piezoelectricity ICMM lecture

3.2 Transformations for converting constitutive data

1. Strain-Charge  Stress-Charge:
cE=0 = s−1
E=0 , e = d : s−1
E=0 , S=0 = T =0 − d · s−1 T
E=0 · d . (17)

2. Strain-Charge  Strain-Voltage:
sD=0 = sE=0 − dT · −1
T =0 · d , g = −1
T =0 · d . (18)

3. Strain-Charge  Stress-Voltage: . . .
4. Stress-Charge  Stress-Voltage:
cD=0 = cE=0 − eT · −1
S=0 · e , q = −1
S=0 · e . (19)

5. Stress-Charge  Strain-Voltage: . . .
6. Strain-Voltage  Stress-Voltage:
cD=0 = s−1
D=0 , q = g : s−1
D=0 , −1 −1 −1 T
S=0 = T =0 + g · sD=0 · g . (20)

3.3 Piezoelectric relations in matrix notation

Rule of change of subscripts (Kelvin-Voigt notation)

11 → 1 , 22 → 2 , 33 → 3 , 23 → 4 , 13 → 5 , 12 → 6 .

Using this rule the following matrix notation for relevant quantities is adopted:
Tij → [Tα ](6×1) , Sij → [Sα ](6×1) , Ei → [Ei ](3×1) , Di → [Di ](3×1) ,
cijkl → [cαβ ](6×6) , sijkl → [sαβ ](6×6) , ij → [ij ](3×3) , −1 −1
ij → [ij ](3×3) ,
ekij → [ekα ](3×6) , dkij → [dkα ](3×6) , qkij → [qkα ](3×6) , gkij → [gkα ](3×6) .
Here: i, j, k, l = 1, 2, 3, and α, β = 1, . . . 6. Exceptionally: S4 = 2S23 , S5 = 2S13 , S6 = 2S12 .

Now, the constitutive relations can be written in the matrix notation as shown below for
two cases.
 Strain-Charge form:
S(6×1) = s(6×6) T(6×1) + dT(6×3) E(3×1) , (21)
D(3×1) = d(3×6) T(6×1) +  (3×3) E(3×1) . (22)

 Stress-Charge form:
T(6×1) = c(6×6) S(6×1) − eT(6×3) E(3×1) , (23)
D(3×1) = e(3×6) S(6×1) +  (3×3) E(3×1) . (24)
 ICMM lecture Fundamentals of Piezoelectricity 9

For orthotropic piezoelectric materials there are 9 + 5 + 3 = 17 material constants, and

the matrices of material constants read:
 
c11 c12 c13 0 0 0
c22 c23 0 0 0
 

c33 0 0 0
 
c(6×6) =  , (25)

 c44 0 0
sym. c55 0 
 

c66
   
0 0 0 0 e15 0 11 0 0
e(3×6) =  0 0 0 e24 0 0 ,  (3×3) =  0 22 0  . (26)
   
e31 e32 e33 0 0 0 0 0 33

Many piezoelectric materials (e.g., PZT ceramics) can be treated as transversally

isotropic. Then, there are only 10 material constants, since 4 + 2 + 1 = 7 of the
orthotropic constants depend on the others:
c11 − c12
c22 = c11 , c23 = c13 , c55 = c44 , c66 = ,
2 (27)
e24 = e15 , e32 = e31 , 22 = 11 .

4 Thermoelastic analogy

Thermal analogy approach

It is a simple but useful approximation of the converse piezoelectric effect based
on the resemblance between the thermoelastic and converse piezoelectric consti-
tutive equations.

The stress vs. strain and voltage relation (i.e., the first from the Stress-Charge form
of piezoelectric constitutive equations), namely:

with dmkl = emij c−1

Tij = cijkl Skl − emij Em = cijkl Skl − dmkl Em ijkl (28)
0
resembles the Hooke’s constitutive relation with initial strain Skl or initial temper-
0
ature θ
0
= cijkl Skl − αkl θ0 .



Tij = cijkl Skl − Skl (29)
Thus, this thermoelastic law (or, simply, initial strains) can be used to approximate
the converse piezoelectric problem. In this case the thermal expansion coefficients
(or initial strains) are determined as
1 0 0
αkl = S where Skl = dmkl Em = −dmkl ϕ|m . (30)
θ0 kl