8 Hyperelasticity

Hyperelasticity
8
8.1 Introduction

Some materials such as elastomers, polymers, rubber and biological matter (arteries,
muscles, skin, etc.) may be subject to large deformations without there being any internal
energy dissipation (which is typical en elastic process). These materials are classified as
being hyperelastic and purely hyperelastic materials have no memory of motion history, i.e.
they are only dependent on the current values of the state variables.
Physically speaking, elastic materials (linear elasticity, hyperelasticity) return to their initial
state once their load disappears, (see Figure 8.1). In other words, the work done during the
loading process is recovered during the unloading process, i.e. there is no internal energy
dissipation (a reversible process).
Our goal in this chapter is to establish the constitutive equations for materials that behave
according to the hyperelasticity theory, also known as Green or nonlinear elasticity. Moreover, we
will limit our analysis to purely mechanical theories, so we have eliminated thermodynamic
variables such as temperature and entropy.
Among the researchers who have used the hyperelastic constitutive equations to model
rubberlike materials we can mention: Alexander (1968), Treloar (1975), Ogden (1984),
Morman (1986) and Holzapfel (2000).

E.W.V. Chaves, Notes on Continuum Mechanics, Lecture Notes on Numerical 423

Methods in Engineering and Sciences 4, DOI 10.1007/978-94-007-5986-2_10,
© International Center for Numerical Methods in Engineering (CIMNE), 2013
424 NOTES ON CONTINUUM MECHANICS

V I - linear elastic zone

II - non-linear elastic zone

unloading
loading

I II H

Figure 8.1: The stress-strain curve for elastic materials.

8.2 Constitutive Equations

A hyperelastic material supposes the existence of a function which is denoted by the

Helmholtz free energy per unit reference volume ( : ). The energy : is also known as
strain energy density or the strain energy function, or elastic potential. In hyperelastic materials, the
strain energy function : is only dependent on the deformation gradient (F ) , i.e.,
: : (F ) .
In pure deformation processes, which do not involve changes caused by entropy or
temperature, internal energy dissipation is equal to zero ( Dint 0 ), which thereby describes
reversible processes. In this way, the Clausius-Planck inequality, (see Chapter 5), for
reversible processes, turns into the following equations:

Dint ı : D  : 0 Ÿ : ı : D (Current configuration)

Dint P : F  : 0 Ÿ : P : F (8.1)
1 1 (Reference configuration)
S : C  : 0 Ÿ : S : C
2 2
where ı is the Cauchy stress tensor, D is the rate-of-deformation tensor, P is the first
Piola-Kirchhoff stress tensor, S is the second Piola-Kirchhoff stress tensor, and C is the
right Cauchy-Green deformation tensor. Then, taking into account the conjugate relations
obtained in Chapter 5 we have:

1 S
³ ı : DdV ³ ,
V
Jı : D dV
IJ
V0
³ IJ : DdV ³ S : E dV ³ 2 S : CdV ³ P : F dV ³ S
0
V0
0
V0
0
V0
0
V0
0
V 0
P : F dV

(8.2)
Then, in summary we can state that the rate of change of the strain energy density can also
be expressed as follows:
1
: P : F S : E S : C IJ :D
2 (8.3)
Stress Power

where E is the Green-Lagrange strain tensor, and IJ is the Kirchhoff stress tensor with
which we can state that a material is considered to be hyperelastic if and only if the rate of
change of the strain energy is equal to the stress power.
 8 HYPERELASTICITY 425

Then, by evaluating the rate of change of the strain energy : (F ) we obtain:

w: ( F ) wF w: ( F ) 
: ( F ) : :F (8.4)
wF wt wF
Next, by substituting the above equation into the internal energy dissipation given by the
equation in (8.1) we obtain:
w: ( F )  w: ( F ) 
P : F  :F 0 Ÿ P : F :F (8.5)
wF wF
Thus we can draw the conclusion that:
w: ( F )
P (8.6)
wF
& & &
NOTE: The tensor P is a function of X and F , i.e. P P( F ( X ), X , t ) , and as P is
&
directly dependent of X we can study non-homogeneous materials. Here, for the sake of
&
simplicity, we will omit the material coordinate X . ɶ

As discussed in Chapter 6, strain energy has to be objective, i.e. : ( F ) : ( F * ) : (Q ˜ F ) ,

where Q is an orthogonal tensor, (see Figure 8.2). Then, by applying the polar
decomposition ( F R ˜ U ) and by adopting Q R T , we can obtain:
: ( F ) : (Q ˜ F )
: (Q ˜ R ˜ U) (8.7)
: (U) Ÿ : : (C ) Ÿ :ˆ : (E)
where U is the right stretch tensor and R is the polar decomposition rotation tensor, (see
Chapter 2). Since the tensors C , U , and E are directly linked by C U 2 , and 2 E C  1 ,
the strain energy density can also be expressed in terms of the Green-Lagrange strain
tensor ( E ). Then, similarly to (8.6), it is possible to express the constitutive equation for
stress in the material description. Next, if we take the rate of change of the energy : (C )
we can obtain:
w: (C ) wC w: (C ) 
: (C ) : :C (8.8)
wC wt wC
By substituting (8.8) into the internal energy dissipation (8.1) we obtain:
1
Dint S : C  : (C ) 0
2
1 w: (C ) 
S : C  :C 0 (8.9)
2 wC
§1 w: (C ) · 
¨ S ¸:C 0
©2 wC ¹
Note that, the condition in (8.9) must hold for any thermodynamic process. Now, in a
mechanical process with C z 0 , the only scenario in which the condition (8.9) remains
valid is when:
1 w: (C ) w: (C )
S 0 Ÿ S 2 (8.10)
2 wC wC
Likewise, we can show that:
426 NOTES ON CONTINUUM MECHANICS

w: (C ) w: ( E )
S 2 (8.11)
wC wE

reference F R ˜U current
configuration configuration

F , : (F )
B0 B
& &
X x

C, E
F* Q˜F Q
: (C ) : (Q ˜ F )
:ˆ ( E ) Q RT

U
: (Q ˜ F ) : (Q ˜ R ˜ U) : (U)
: (U)
R

B &
X

Figure 8.2: Objectivity of strain energy density.

Then, if we take into account the relationships between the stress tensors, (see Chapter 3),
we can still express the constitutive equation for stress as follows:
ƒ The Kirchhoff stress tensor ( IJ ):
w: ( E ) w: (C )
IJ F ˜S ˜ F T F˜ ˜FT 2F ˜ ˜FT (8.12)
wE wC
ƒ The Cauchy stress tensor ( ı ):
IJ Jı F ˜S ˜ F T ı J P ˜ F T
w: (C ) w: ( F ) (8.13)
2F ˜ ˜FT J ˜ F T
wC wF
ƒ The first Piola-Kirchhoff stress tensor ( P ):
w: (C )
P IJ ˜ F T F ˜ S ˜ F T ˜ F T F ˜S F ˜2 (8.14)
wC
ƒ The Mandel stress tensor ( M ):
w: (C )
M C ˜S 2C ˜
wC (8.15)
w: ( F )
F T ˜P FT ˜
wF
 8 HYPERELASTICITY 427

Hence, we can sum up the different ways of expressing the stress constitutive equations for
hyperelastic materials as:

w: ( F ) w: (C ) w: (C )
P F ˜2 ; IJ Jı 2F ˜ ˜FT Stress constitutive
wF wC wC
equations for (8.16)
w: ( E ) w: (C ) w: (C ) w: ( F ) hyperelastic materials
S 2 ; M 2C ˜ F ˜ T
wE wC wC wF

The strain energy function ( : ) has to satisfy the following:

ƒ The normalization condition: : (1) 0 , i.e. the strain energy function vanishes
when the material has been completely unloaded, i.e. when F 1 ;
ƒ : ( F ) t 0 . The strain energy must increase monotonically with deformation.
In a reversible process (without internal energy dissipation) the following must be satisfied:
ƒ The work done is independent of the path.
If we consider a particle which is deformed according to the path * at a given time
interval, said deformation is defined by the deformation gradient F : >t1 , t 2 @ . Then, the
internal work done associated with this path is:
t2 t2 t2
1
³
t1
P : F dt ³
t1
S : E dt ³ 2 S : C dt
t1
(8.17)

Now, for an elastic material (reversible) the internal work done is independent of the path,
so, the following must be met:

³ P : dF *c³ P : dF
*
(8.18)

for any deformation path * , *c .

ƒ For any closed cycle of deformation the work done is equal to zero:

³ P : dF 0 or ³ S : dE 0 (8.19)

8.2.1 Elastic Tangent Stiffness Tensors

8.2.1.1 The Material Elastic Tangent Stiffness Tensor

The rate of change of the constitutive equation in (8.16), S (E ) , can be expressed as:

w 2: (E)  w 2: (E) 
S :E C tan : E S ij E kl C tan 
ijkl E kl (8.20)
wE wE wE ij wE kl

where S , E are objective rates, and C tan is a fourth-order tensor known as the material
elastic tangent stiffness tensor also called the material tangent elasticity tensor. Remember that the
tensors E and C are related to each other by the equation 2 E C  1 Ÿ 2 E C , thus:
w 2 : (C )  w 2 : (C ) 
S 2 :C Ÿ S 4 :E
wC wC wC wC (8.21)
C tan
Then, taking into account the equations in (8.11), (8.20) and (8.21) we can conclude that:
428 NOTES ON CONTINUUM MECHANICS

w 2: (E) wS
C tan The material elastic tangent stiffness tensor
wE wE wE
(8.22)
w : (C )
2
wS (Reference configuration)
4 2
wC wC wC

Note that the tensors S and E are symmetrical, i.e. S ij S ji , E ij E ji , so, the fourth-
order tensor, C tan , must feature at least minor symmetry, i.e.:
C tan
ijkl C tan
jikl C tan
ijlk C tan
jilk (8.23)

Then, taking into account the equation in (8.22) we can conclude that the tensor C tan also
has major symmetry:
w 2: w § w: · w § w: ·
C tan ¨¨ ¸¸ ¨ ¸ C tan (8.24)
ijkl
wE ij wE kl wE ij ¨ wE ij ¸ klij
© wE kl ¹ wE kl © ¹
Therefore, we can conclude that the tensor C tan is symmetric. In the general case C tan is
anisotropic and has 21 independent components. For further details regarding symmetry
types see Chapter 7.

8.2.1.2 The Spatial Elastic Tangent Stiffness Tensor

The rate of change of the second Piola-Kirchhoff stress tensor can be obtained by means
of the equation S F 1 ˜ IJ ˜ F T , i.e.:
S F 1 ˜ IJ ˜ F T  F 1 ˜ IJ ˜ F T  F 1 ˜ IJ ˜ F T
 F 1 ˜ l ˜ IJ ˜ F T  F 1 ˜ IJ ˜ F T  F 1 ˜ IJ ˜ l T
˜ F T (8.25)
F 1
˜ IJ  l ˜ IJ  IJ ˜ l T
˜F T

where we have considered the equation F 1  F 1 ˜ l which was obtained in Chapter 2.

Remember in Chapter 4 the Oldroyd rate of the Kirchhoff stress tensor was given by:
‰
IJ IJ  l ˜ IJ  IJ ˜ l T (8.26)
‰
Remember also that IJ is not objective, but IJ is. Then, by substituting (8.26) into (8.25) we
obtain:
‰
S F 1 ˜ IJ˜ F T (8.27)
Then, given the relationship between the rate of change of the Green-Lagrange strain
tensor and the rate-of-deformation tensor we have:
E F T ˜D ˜ F (8.28)
Additionally, by substituting the equations (8.27) and (8.28) into the constitutive equation
(8.20) we obtain:

S C tan : E S mn C tan 
mnpq E pq

‰ ‰ (8.29)
F 1 ˜ IJ˜ F T C tan : F T ˜ D ˜ F Fms1 W st Fnt1 C tan
mnpq Fkp D kl Flq

Then, taking into account the symmetry of the Oldroyd rate of the Kirchhoff stress tensor
‰
( IJ ) and D , we can obtain:
 8 HYPERELASTICITY 429

‰
‰ Fim Fms1 W st F jn Fnt1 F jn Fim C tan
mnpq Fkp Flq D kl
F ˜F 1
˜ IJ˜ F ˜ FT T
F ˜C tan
:F T
˜D ˜ F ˜ F T
‰
E is W st E jt F jn Fim C tan
mnpq Fkp Flq D kl
‰
‰ W ij F jn Fim C tan
mnpq Fkp Flq D kl
IJ F F :C tan
:F T
F T
:D
Lijkl

(8.30)
‰
IJ L :D (8.31)

where we have introduced the spatial elastic tangent stiffness tensor, also called the spatial
tangent elasticity tensor, in the current configuration, which is given by:
L ( F F ) : C tan : ( F T F T ) L ijkl Fim F jn C tan
mnpq Fkp Flq (8.32)
Then, from the above equation we can obtain the inverse relationship:
Fai1 Fbj1 L ijkl Fck1 Fdl1 Fai1 Fim Fbj1 F jn C tan 1 1
mnpq Fck Fkp Fdl Flq E am E bn C tan
mnpq E cp E dq
(8.33)
C tan
abcd

Thus
C tan
abcd Fai1 Fbj1 L ijkl Fck1 Fdl1 C tan F 1 F 1 : L : F T F T (8.34)
$
In Chapter 4 we obtained the relationship between the Jaumann-Zaremba rate ( IJ ) and the
‰
Oldroyd rate of the Kirchhoff stress tensor ( IJ ), i.e.:
$ ‰ ‰ $
IJ IJ D ˜ IJ  IJ ˜ D Ÿ IJ IJ D ˜ IJ  IJ ˜ D (8.35)
‰
Next, by combining the above rate IJ with the constitutive equation in (8.31) we can
obtain:
$
IJ D ˜ IJ  IJ ˜ D L : D
$ (8.36)
$ W ij L ijkl D kl  D ip W pj  W ip D pj
ŸIJ L :D  D ˜ IJ  IJ ˜D
Notice that the tensor D is symmetric, so the double dot product between the symmetric
fourth-order unit tensor, I sym , and a symmetric second-order tensor turns out to be the
same tensor, so,
1 1
D ip sym
I ipkl D kl (E ik E pl  E il E pk )D kl D pj I sym
pjkl D kl (E pk E jl  E pl E jk )D kl (8.37)
2 2
Then, by substituting (8.37) into (8.36) we obtain:
$ 1 1
W ij L ijkl D kl  (E ik E pl  E il E pk ) W pj D kl  W ip (E pk E jl  E pl E jk )D kl
2 2
ª 1 1 º
« L ijkl  (W pj E ik E pl  W pj E il E pk )  (W ip E pk E jl  W ip E pl E jk )» D kl
¬ 2 2 ¼
(8.38)
ª 1 1 º
« L ijkl  2 (W lj E ik  W kj E il )  2 (W ik E jl  W il E jk )» D kl
¬ ¼
ª 1 º
« L ijkl  2 4 (W lj E ik  W kj E il  W ik E jl  W il E jk )» D kl L ijkl  2H ijkl D kl > @
¬ ¼
430 NOTES ON CONTINUUM MECHANICS

Therefore, the rate of change of the constitutive equation in terms of Jaumann-Zaremba

rate of the Kirchhoff stress tensor becomes:
$
IJ L̂ : D (8.39)

where L̂ is a fourth-order tensor and is defined by:

Lˆ L  2H (8.40)
with
1
H ijkl (W lj E ik  W kj E il  W ik E jl  W il E jk ) (8.41)
4
We can now summarize the relationships between the rate of change of the stress and the
rate-of-deformation tensor as:

­‰ ­$ ˆ
°IJ L : D °IJ L : D
®‰ ; ®$
°̄ W ij L ijkl D kl >
°̄W ij L ijkl  2H ijkl D kl @
where
(8.42)
L ijkl F jn Fim C tan
mnpq Fkp Flq

Lˆ ijkl L ijkl  2H ijkl The spatial elastic tangent stiffness tensor

1 (Current configuration)
H ijkl (W lj E ik  W kj E il  W ik E jl  W il E jk )
4

8.2.1.3 The Instantaneous Elastic Tangent Stiffness Tensor

The relationship between the Cauchy stress tensor and the second Piola-Kirchhoff stress
tensors is given by S JF 1 ˜ ı ˜ F T whose rate of change becomes:
S JF 1 ˜ ı ˜ F T  JF 1 ˜ ı ˜ F T  JF 1 ˜ ı ˜ F T  JF 1 ˜ ı ˜ F T (8.43)
where J J Tr (D) and F  F ˜ l , (see Chapter 2), which if substituted into the
1 1

above equation yields the following result:

S JF 1 ˜ ı  l ˜ ı  ı ˜ l T
 Tr (D)ı ˜ F T (8.44)
Now, remember in Chapter 4 that the Truesdell stress rate, which is objective, is given by
7
ı ı  l ˜ ı  ı ˜ l T
 ı Tr (D) so, we can state that:
7
S JF 1 ˜ ı ˜ F T (8.45)

Then, by substituting the equation (8.45) into the constitutive equation in (8.20) and if we
know that E F T ˜ D ˜ F , we can obtain:
S C tan : E
7
JF 1 ˜ ı ˜ F T C tan : F T ˜ D ˜ F (8.46)
7 1
ı F F : C tan : F T F T : D
J
or
 8 HYPERELASTICITY 431

7
ı A :D (8.47)

where A is the instantaneous elastic tangent stiffness tensor, also called the instantaneous elastic
moduli, (see Asaro&Lubarda(2006)), which is defined by:
1 1 1 1
A F F : C tan : F T F T L A ijkl Fim F jn C tan
mnpq Fkp Flq L ijkl (8.48)
J J J J
where L is the fourth-order tensor given in (8.32). So, in summary we have:
7 7
ı A :D ; V ij A ijkl D kl

with (8.49)
1 1
A ijkl Fim F jn C tan
mnpq Fkp Flq L ijkl The instantaneous elastic tangent stiffness tensor
J J
7
Then, by taking into account the relationship between the Truesdell stress rate ( ı ) and the
‰ ‰ 7
Oldroyd rate of the Kirchhoff stress tensor IJ , i.e. IJ J ı , the equation in (8.47) becomes:
7 1‰ 1 ‰
ı A :D Ÿ IJ L :D Ÿ IJ L :D (8.50)
J J
which is the same as that obtained in (8.31).

8.2.1.4 The Elastic Tangent Stiffness Pseudo-Tensor

The constitutive equation for stress that relates the first Piola-Kirchhoff stress tensor to the
deformation gradient is:
w: ( F ) w: (F )
P Pij (8.51)
wF wFij

Remember that F and P are two-point tensors (pseudo-tensors), i.e. they are not defined
in any configuration. Then, the rate of change of the above constitutive equation is given
by:

w 2 : (F )  w § w: (F ) · 
P :F P ij ¨¨ ¸¸ Fkl
wF wF wFij © wFkl ¹ (8.52)
P K : F P ij K ijkl Fkl

We can now introduce the elastic tangent stiffness pseudo-tensor also called the elastic
pseudomoduli, (see Lubarda&Benson (2001)), as follows:
w 2 : (F ) w 2 : (F ) w 2 : (F )
K K ijkl K klij (8.53)
wF wF wFij wFkl wFkl wFij
The elastic tangent stiffness pseudo-tensor is not a “real” moduli, because it is partially
associated with the material spin tensor, (see Asaro&Lubarda (2006)).
Next, we can relate the tensors K and C tan . To do so, we need to evaluate the rate of
change of Pij Fip S pj , (see Eq. (8.14)):

P ij S pj Fip  Fip S pj (8.54)

432 NOTES ON CONTINUUM MECHANICS

Then, by substituting (8.21) and (8.52) into the above equation we obtain:
K ijkl Fkl C pjkl E kl Fip  Fip S pj (8.55)
1 
and if we know that E kl Fqk Fql  Fqk Fql the above equation becomes:
2
1 tan 
K ijkl Fkl C pjkl Fqk Fql  Fqk Fql Fip  Fip S pj
2
(8.56)
1 tan
C pjkl Fql Fip Fqk  C tan  
pjkl Fqk Fip Fql  Fip S pj
2
Note that the dummy indices k and l from the expression C tan 
pjkl Fql Fip Fqk can be

exchanged without altering the result of the expression, and the dummy indices k and q
from C tan 
pjkl Fqk Fip Fql can also be exchanged, so:

1 tan
K ijkl Fkl C pjlk Fqk Fip Fql  C tan  
pjql Fkq Fip Fkl  S pj Fip
2
(8.57)
1 tan
C pjlq Fkq Fip Fkl  C tan  
pjql Fkq Fip Fkl  S pj Fip
2
Then, if we make use of the minor symmetry C tan
pjlq C tan
pjql , we can still state that:

K ijkl Fkl C tan  

pjlq Fkq Fip Fkl  Fip S pj (8.58)
Next, if we see that S pj Fip S lj Fil S lj E ik Fkl holds, we can conclude that:

K ijkl Fkl C tan 

pjlq Fkq Fip  S lj E ik Fkl Ÿ K ijkl pjlq Fkq Fip  S lj E ik
C tan (8.59)
Thus:

P K : F ; P ij K ijkl Fkl

w 2 : (F ) (8.60)
K ijkl
wFij wFkl The elastic tangent stiffness pseudo-tensor
K ijkl pjlq Fkq Fip  S lj E ik
C tan

8.3 Isotropic Hyperelastic Materials

If the scalar-valued tensor function : (C ) is isotropic it must satisfy the following:

: (C ) : (Q ˜ C ˜ Q T ) (8.61)
for any orthogonal tensor Q . Then if we use the polar decomposition rotation tensor, i.e.
Q R , and if we know that C F T ˜ F , we can obtain:

: (C ) : (R ˜ C ˜ R T ) : (R ˜ F T ˜ F ˜ R T ) : ( V T ˜ V ) : ( V 2 )
(8.62)
: (b)
 8 HYPERELASTICITY 433

where V V T is the left stretch tensor which is related to the left Cauchy-Green
deformation tensor ( b F ˜ F T ) by means of b V 2 , where F V ˜ R Ÿ V F ˜ R T is
satisfied, (see Chapter 2). Thus, in isotropic materials, the energy function : can be
expressed in terms of the left Cauchy-Green deformation tensor as follows:
& & Energy function for isotropic
: (C , X ) : (b, x ) (8.63)
hyperelastic materials

Reference Current
configuration configuration
P ij K ijkl Fkl

Elastic Tangent Stiffness Pseudo-Tensor

w 2 : (F )
K ijkl
wFij wFkl
K ijkl pjlq Fkq Fip  S lj E ik
C tan

B0 B
&
X &
x

RATE-TYPE CONSTITUTIVE EQUATIONS

S C tan : E ‰
W ij
$
L ijklD kl Wij
7
L̂ ijkl D kl Vij A ijklD kl

ELASTIC TANGENT STIFFNESS TENSORS

Material elastic tangent stiffness tensor
Spatial elastic tangent stiffness tensor
tan w 2 : ( E ) wS
C L ijkl F jn Fim C tan
mnpq Fkp Flq
wE wE wE
w 2 : (C ) wS Lˆ ijkl L ijkl  2H ijkl
4 2
wC wC wC with
1
H ijkl (W lj E ik  W kj E il  W ik E jl  W il E jk )
4
Instantaneous elastic tangent stiffness tensor
1 1
A ijkl Fim F jn C tan
mnpq Fkp Flq L ijkl
J J

Figure 8.3: The elastic tangent stiffness tensors.

434 NOTES ON CONTINUUM MECHANICS

8.3.1 The Constitutive Equation in terms of Invariants

8.3.1.1 The Constitutive Equation in terms of C and b

We showed in Chapter 1 that the scalar-valued isotropic tensor function, : (C ) , can be

written in terms of the principal invariants of C ( I C , II C , III C ), or in terms of the
invariants of b , i.e.:
: : (C ) : ( I C , II C , III C ) : ( I b , II b , III b ) (8.64)
Then, it was verified in Chapter 2 that I C I b , II C II b and III C III b , which is obtained
from:
IC Tr(C ) ,
II C
1
2
^
>Tr(C )@2  Tr (C 2 ) , III C ` det(C ) J2
1­ 3 1 3½
® Tr (C )  Tr (C ) Tr (C )  >Tr (C )@ ¾
3 2
3¯ 2 2 ¿
Also in Chapter 1, we showed that for the scalar-valued tensor function,
: (C ) : ( I C , II C , III C ) , the following equations are valid:
w: ( I C , II C , III C ) w: wI C w: w II C w: w III C
 
wC wI C wC w II C wC w III C wC
§ w: w: · § w: · § w: · 1
¨¨  I C ¸¸1  ¨¨ ¸¸C  ¨¨ III C ¸¸C
w
© CI w II C ¹ © w II C ¹ © w III C ¹
(8.65)
§ w: w: w: · § w: w: · § w: · 2
¨¨  IC  II C ¸¸1  ¨¨  I C ¸¸C  ¨¨ ¸¸C
© wI C w II C w III C ¹ © Cw II w III C ¹ © w III C ¹
§ w: · § w: w: · § w: ·
¨¨ ¸¸1  ¨¨ II C  III C ¸¸C 1  ¨¨ III C ¸¸C  2
© wI C ¹ © w II C w III C ¹ © w II C ¹
where we held that:
wI C w II C
1, IC 1  C T IC 1  C II C C 1  III C C  2 ,
wC wC
(8.66)
w III C
III C C T III C C 1 C 2  I C C  II C 1
wC
Now, taking into account the equations in (8.11) and (8.65) we can obtain the constitutive
equation in terms of the principal invariants of C as follows:
w: (C )
S 2
wC
ª§ w: w: · § w: · § w: · º
2 «¨¨  I C ¸¸1  ¨¨ ¸¸C  ¨¨ III C ¸¸C 1 »
¬«© wI C w II C ¹ © w II C ¹ © w III C ¹ ¼»
ª§ w: w: w: · § w: w: · § w: · 2º (8.67)
2 «¨¨  IC  II C ¸¸1  ¨¨  I C ¸¸C  ¨¨ ¸¸C »
w
¬«© CI w II C w III C ¹ w
© C II w III C ¹ © w III C ¹ ¼»
ª§ w: · § w: w: · § w: · º
2 «¨¨ ¸¸1  ¨¨ II C  III C ¸¸C 1  ¨¨ III C ¸¸C  2 »
«¬© wI C ¹ © w II C w III C ¹ © w II C ¹ »¼

In Chapter 1, in the subsection about the Tensor-Valued Tensor Function, it was shown
that the following relationships are valid:
 8 HYPERELASTICITY 435

: , b ˜ b F ˜ : ,C ˜ F T b ˜ : ,b (8.68)
Now, if we consider the relationship between the Kirchhoff stress and the second Piola-
Kirchhoff stress tensors, IJ F ˜ S ˜ F T , and the constitutive equation in the reference
configuration (8.11), S 2: ,C , it is possible to obtain the constitutive equation in the
current configuration as follows:
IJ F ˜S ˜ F T F ˜ 2: , C ˜ F T 2: , b ˜ b 2b ˜ <,b (8.69)
Next, by taking into account the equation IJ J ı , we can also represent the constitutive
equation for isotropic materials as:
w: (b)
IJ 2 ˜ b 2b ˜ w: (b) and ı J 1 2
w: (b)
˜b J 1 2b ˜
w: (b)
(8.70)
wb wb wb wb
Then, in a similar fashion to (8.67), and by considering that IJ J ı we can obtain:
ª§ w: w: · w: w: º
IJ J ı F ˜S ˜ F T 2 F ˜ «¨¨  I C ¸¸1  C III C C 1 » ˜ F T (8.71)
«¬© wI C w II C ¹ w II C w III C »¼

Now, if we consider that C F T ˜ F and b F ˜ F T , the above equation becomes:

ª§ w: w: · w: w: º
IJ J ı 2«¨¨  I b ¸¸b  b2  III b 1» (8.72)
¬«© wI b w II b ¹ w II b w III b ¼»

Alternatively, we can represent (8.72) by substituting the expression b 2 obtained by means

of the Cayley-Hamilton theorem b 3  I b b 2  II b b  III b 1 0 , (see Chapter 1), i.e.:
Ÿ b 3 ˜ b 1  I b b 2 ˜ b 1  II b b ˜ b 1  III b 1 ˜ b 1 0
2 1
Ÿ b  I b b  II b 1  III b b 0 (8.73)
2 1
Ÿb I b b  II b 1  III b b
Then, the equation in (8.72) can still be written as:
ª§ w: w: · § w: · § w: · º
IJ J ı 2«¨¨ III b  II b ¸¸1  ¨¨ ¸¸b  ¨¨ III b ¸¸b 1 » (8.74)
¬«© w III b w II b ¹ © w Ib ¹ © w IIb ¹ ¼»

Then, by using equations between stress tensors, (see Chapter 3),

P IJ ˜ F T F ˜ S ˜ F T ˜ F T F ˜S (8.75)
the constitutive equation can also be written in terms of the first Piola-Kirchhoff stress
tensor. To do so, let us consider S given in (8.67), and so we obtain:
ª§ w: w: · w: w: º
P 2F ˜ «¨¨  I C ¸¸1  C III C C 1 »
«¬© wI C w II C ¹ w II C w III C »¼
(8.76)
ª§ w: w: · w: w: º
2«¨¨  I C ¸¸ F ˜ 1  F ˜FT ˜F  III C ˜ F ˜ F 1 F T »
¬«© wI C w II C ¹ w II C b w III C 1 ¼»

Additionally, by considering that F T b T ˜ F and the symmetry of b , we obtain:

ª§ w: w: · w: w: º w: (b)
P 2«¨¨  I C ¸¸1  b III C b 1 » ˜ F Ÿ P 2 ˜F (8.77)
«¬© wI C w II C ¹ w II C w III C »¼ wb
436 NOTES ON CONTINUUM MECHANICS

8.3.1.2 The Constitutive Equation in terms of E

Energy can also be written in terms of the Green-Lagrange strain tensor E , and if we are
dealing with isotropic material the energy constitutive equation can be expressed in terms
of the principal invariants of E :
: ( E ) : ( I E , II E , III E ) (8.78)

where I E Tr(E ) , II E
1
2
^ `
>Tr ( E )@2  Tr ( E 2 ) , III E det(E ) . Then, if we consider the
w: ( E )
equation in (8.11), i.e. S , we can obtain another one analogous to that obtained in
wE
(8.65), i.e.:
S c0 1  c1 E  c2 E 2 (8.79)
where the parameters c 0 , c1 , c 2 are given by:
w: w: w: w: w: w:
c0  IE  II E ; c1  IE ; c2 (8.80)
wI E w II E w III E w II E w III E w III E

8.3.2 Series Expansion of the Energy Function

Let us assume that : : (C ) is a continuously differentiable function with respect to the
C -invariants. It is possible, then, to represent : by means of infinite power series:
f
: : ( I C , II C , III C ) ¦c
p ,q ,r 0
pqr IC  3
p
II C  3
q
III C  1
r
(8.81)

where the coefficients c pqr are independent of the deformation.

Here, we can notice that in an undeformed state we have F 1 Ÿ C 1 , then I C 3 ,
II C 3 , III C 1 , which results in : 0 , as expected, since in said undeformed state strain
energy is zero (normalization condition).
Then, taking into account that C U 2 , where U is the right stretch tensor, it is possible to
express the tensor C in terms of the principal stretches (eigenvalues of U ) O 1 , O 2 , O 3 ,
(see Chapter 2), so we can use the following spectral representation:
3 3
U ¦O
a 1
a
ˆ (a)
N ˆ (a)
N Ÿ C U2 ¦O
a 1
2
a
ˆ (a)
N ˆ (a)
N (8.82)

Next, the principal invariants of C or b in terms of the principal stretches O i are given by:

ªO21 0 0º ­ I C I b O21  O22  O23

« » °°
C ij U ij2 «0 O22 0» Ÿ 2 2 2 2 2 2
® II C II b O 1O 2  O 2 O 3  O 1O 3 (8.83)
«0 0 O23 »¼ ° 2 2 2
¬ ¯° III C III b O 1O 2 O 3
Then, by substituting the values of (8.83) into the power series (8.81) and after some
mathematical manipulations we obtain:

: : (O 1 , O 2 , O 3 ) ¦ a ^>O
f

p ,q ,r 0
pqr
p
1 Oq2  Oq3 O p2 Oq3  Oq1  O p3 Oq1  Oq2 O 1 O 2 O 3 @ r
`
6 (8.84)
 8 HYPERELASTICITY 437

where the coefficients a pqr are independent of the deformation.

In incompressible materials III C 1 or O 1O 2 O 3 1 is satisfied and the equations in (8.81)
and (8.84) becomes, respectively:
f
: : ( I C , II C ) ¦c
p ,q 0
pq IC  3
p
II C  3
q
(8.85)

: : (O 1 , O 2 , O 3 )
f

¦a
p ,q 0
pq ^ >O p
1 Oq2  Oq3 O p2 Oq3  Oq1  O p3 Oq1  Oq2  6 @ ` (8.86)

8.3.3 Constitutive Equations in terms of the Principal

Stretches
As we have seen before, for isotropic materials, we can express the strain energy function
in terms of the principal stretches O a , a 1,2,3 , i.e. : : (O 1 , O 2 , O 3 ) . Let us now suppose
ˆ ( a ) and nˆ ( a ) are the principal directions (eigenvectors) of the right stretch tensor ( U )
that N
and the left stretch tensor ( V ), respectively, where the following holds:
3 3
1
U ¦O
a 1
a
ˆ (a)
N ˆ (a )
N ; E ¦ 2 (O
a 1
2
a
ˆ (a )
 1) N ˆ (a)
N
(8.87)
3 3 3
1 ˆ (a )
V ¦ a 1
O a nˆ ( a ) nˆ ( a ) ; F ¦
a 1
O a nˆ ( a ) ˆ (a)
N ; F 1
¦
a O
1 a
N nˆ ( a )

To see how the above relationships are proven, see the Section on Polar Decomposition in
Chapter 2.
Now, the second Piola-Kirchhoff stress tensor ( S ) in terms of the principal stretches
becomes:
w: (C ) w: wO i § w: wO 1 w: wO 2 w: wO 3 ·
S 2 2 2¨¨   ¸¸ (8.88)
wC wO i wC © wO 1 wC wO 2 wC wO 3 wC ¹
Then by considering the spectral representation of C given in (8.82), the rate of change of
C (O a ) can be evaluated as follows:
3
ˆ ( a ) ˆ ( a ) º
C ¦ ª«¬2O
a 1
ˆ (a )
 N
aOa
ˆ ( a )  O2 N
N a
ˆ ( a )  O2 N
N a
ˆ (a) N
»¼ (8.89)

ˆ ( a ) on the right and on the left of both sides of the

Now, if we apply the dot product of N
equation we have:
ˆ ( a ) ˜ C ˜ N
N ˆ (a)
3

¦ ª«¬2O
a 1
ˆ (a ) ˜ N
 N
aOa
ˆ (a) N ˆ ( a )  O2 §¨ N
ˆ (a) ˜ N
a
©
ˆ ( a )
ˆ (a) ˜ N ˆ (a ) ˜ N
N ˆ (a ) ˜ N
ˆ (a)  N ˆ (a) N ˆ ( a ) ·¸º
ˆ ( a ) ˜ N
¹»¼
(8.90)
ˆ (a ) ˜ N
Then, bearing in mind that N ˆ ( a ) 1 , and the fact that the rate of change of a vector
with constant magnitude is always orthogonal to itself (Holzapfel (2000)), it follows that
N ˆ ( a )
ˆ (a) ˜ N 0 and subsequently the above equation becomes:
ˆ ( a ) ˜ C ˜ N
N ˆ (a) 2O a O a (8.91)
438 NOTES ON CONTINUUM MECHANICS

The reader should be aware here that the index a 1,2,3 is not a dummy index, i.e. we are
not dealing with indicial notation.
&& && & &
Next, using the property a ˜ T ˜ b T : (a b) where a and b are vectors and T is a
ˆ (a ) N
second-order tensor, the equation in (8.91) can be rewritten as C : (N ˆ ( a ) ) 2O O ,
a a
wC
and if we also consider that C (O a ) O a we can obtain:
wO a
wC  ˆ (a) ˆ (a ) wC ˆ ( a ) ˆ (a)
Oa :N N 2O a O a Ÿ :N N 2O a
wO a wO a
§ wC · § 1 ˆ (a) ˆ ( a ) ·¸ 1 Ÿ §¨ wC · § 1 ˆ (a) ˆ ( a ) ·¸ § wC · § wO a · (8.92)
Ÿ ¨¨ ¸¸ : ¨¨ N N ¸ ¨ wO ¸¸ : ¨¨ N N ¸ ¨¨ ¸¸ : ¨ ¸
© wO a ¹ © 2O a ¹ © a ¹ © 2O a ¹ © wO a ¹ © wC ¹
1

which draws us to the conclusion that:

1 ˆ (a) ˆ (a) wO a
N N (8.93)
2O a wC
Then, by using the second Piola-Kirchhoff stress tensor expression obtained in (8.88), i.e.
w: (O a ) wO k 3
w: wO a
S 2 2¦ , and by considering the equation obtained in (8.93), we
wO k wC a 1 wO a wC
can express the tenor S as:
3
1 w: ˆ ( a ) 3
S ¦O
a 1 a wO a
N ˆ (a )
N ¦ S Nˆ
a 1
a
(a) ˆ (a)
N (8.94)

where S a are the second Piola-Kirchhoff stress tensor eigenvalues. Then, by comparing
the equation in (8.94) with the spectral representation of the tensor C , given in (8.82), we
can conclude that in isotropic materials, C and S are coaxial tensors ( C ˜ S S ˜ C ), i.e.
they have the same principal directions.
Then, as regards the Cauchy stress tensor, we have:
3
1 w:
ı J 1 F ˜ S ˜ F T ¦J
a 1
1
O a wO a
F˜N
ˆ (a ) ˆ (a )
N ˜FT
(8.95)
3
1 w: 3
1 w:
¦
a 1
J 1
O a wO a
F ˜N
ˆ (a) ˆ (a ) ˜ F T
N ¦
a 1
J 1
O a wO a
F ˜N
ˆ (a) F ˜N
ˆ (a)

Moreover, if we take into account the equation F ˜ N

ˆ (a) O a nˆ ( a ) , (see subsection 2.8 Polar
Decomposition of F in Chapter 2), we can obtain:
3
w: ˆ ( a ) 3
ı ¦J
a 1
1
Oa
wO a
n nˆ ( a ) ¦ V nˆ
a 1
a
(a )
nˆ ( a ) (8.96)

where V a are the Cauchy stress tensor eigenvalues.

Since C and S are coaxial, we can obtain S -eigenvalues by considering the principal
directions of C by means of one of the equations in (8.67):
ª§ w: · § w: w: · § w: · º
Sa 2«¨¨ ¸¸  ¨¨ II C  III C ¸¸Oa2  ¨¨ III C ¸¸Oa4 » (8.97)
«¬© wI C ¹ © C w II w III C ¹ w
© C II ¹ »¼
 8 HYPERELASTICITY 439

The first Piola-Kirchhoff stress tensor is given by P F ˜ S , then it holds that:

3 3 3
P F ˜S F˜ ¦ S Nˆ
a 1
a
(a ) ˆ (a)
N ¦S
a 1
aF ˜ Nˆ ( a ) ˆ (a )
N ¦S
a 1
a O an
ˆ (a) ˆ (a )
N

3
(8.98)
¦
a 1
Pa nˆ (a ) ˆ (a)
N

Thus, we can express the components S a as:

ª§ w: · § w: w: · § w: · º
Pa 2O a «¨¨ ¸¸  ¨¨ II C  III C ¸¸Oa2  ¨¨ III C ¸¸Oa4 » (8.99)
¬«© wI C ¹ © C w II w III C ¹ w
© C II ¹ ¼»
We can also express these components in terms of Pa . To do so, let us consider that:
w: ( I C , II C , III C ) w: wI C w: w II C w: w III C
  (8.100)
wO a wI C wO a w II C wO a w III C wO a
Then, the derivatives of I C with respect to O a are:
wI C w ½
O21  O22  O23 2O 1 °
wO 1 wO 1 °
wI C w °° wI C
O21  O22  O23 2O 2 ¾ Ÿ 2O a (8.101)
wO 2 wO 2 ° wO a
wI C w °
O21  O22  O23 2O 3 °
wO 3 wO 3 ¿°
w II C w
O21 O22  O22 O23  O21O23 2O 1O22  2O 1O23 2O 1 O22  O23
wO 1 wO 1
>
2O 1 O21O22  O21O23 O12 @ >
2O 1 O21O22  O21O23  O22 O23 O12  O22 O23 O12 @ (8.102)

2O > O O
1
2 2
1 2  O21O23  O22 O23 O12  O22 O23 O21 O14 @ 2O 1 II C O12  III C O14
which is true for the other principal values, then
w II C
2O a II C Oa2  III C Oa4 (8.103)
wO a
w III C w III C 2 2 2
wO 1 wO 1
O 1O 2 O 3 2O 1 O22 O23 >
2O 1 O21O22 O23 O12 @ 2O 1 III C O12 (8.104)

which is true for the other principal values, then

w III C
2O a III C Oa2 (8.105)
wO a
Then, by substituting (8.101) into (8.100) we obtain:
w: ( I C , II C , III C ) ª§ w: · § w: w: · § w: · º
2O a «¨¨ ¸¸  ¨¨ II C  III C ¸¸Oa2  ¨¨ III C ¸¸Oa4 » (8.106)
wO a «¬© wI C ¹ © w II C w III C ¹ © w II C ¹ »¼
Additionally, if we compare (8.106) with (8.99) and with (8.97) we can draw the conclusion
that:
w: ( I C , II C , III C )
Pa O aS a (8.107)
wO a
Then, it holds that:
440 NOTES ON CONTINUUM MECHANICS

3 3
P ¦O
a 1
aS an
ˆ (a) ˆ (a)
N ¦ P nˆ
a 1
a
(a) ˆ (a)
N (8.108)

Note that Pa are not the eigenvalues of P . The Cauchy stress tensor is related to the first
Piola-Kirchhoff stress tensor by means of ı J 1 P ˜ F T , after which the eigenvalues of
ı are given by:
w:
Va J 1O a (8.109)
wO a
which is the same result as that obtained in (8.96). Note that index a does not indicate
summation.
Then, in isotropic materials, the Kirchhoff stress tensor ( IJ ) and the left stretch tensor ( V )
have the same principal directions, and if we consider that IJ J ı we can obtain:
3 3
IJ ¦O
a 1
2 ˆ (a)
a S an nˆ ( a ) ¦ W nˆ
a 1
a
(a )
nˆ ( a ) (8.110)

Now, if we look back at the equations in (8.94), (8.96) and (8.110), we can conclude that
the principal values of the tensors S , ı , IJ , are interrelated by:
1 w: J 1
Sa Va Wa (8.111)
O a wO a O2a O2a

8.4 Compressible Materials

In compressible hyperelasticity materials (which go through a change in volume during the

deformation process), it would be appropriate to separate the motion undergone into
isochoric motion (volume-preserving) and another type characterized by dilatational
transformation (purely volume-change). So, let us consider the multiplicative
decomposition of the deformation gradient, (see Figure 8.4), as follows:
~
F F ˜ F vol (8.112)
~ ~
where F show an isochoric transformation ( F { F iso ), and F vol describes a dilatational
transformation, (see Figure 8.4). Now let us look back at Chapter 2 subsection 2.13, where
we obtained the following equations:

1 1
­~ 2 2

~ °C J 3 C ; C vol J 31
F J 3 F ; F vol J 31 Ÿ ® 2 2
(8.113)
°~ vol 31
¯b J b ; b J
3

and
1 1
~ ~
J F J 3 F 1 ; J vol F vol J 31 J (8.114)
 8 HYPERELASTICITY 441

wJ J 1
Moreover, in Chapter 1 it was proven that C , where J 2 III C III b . Likewise,
wC 2
we can obtain the following relationships:
2 ª 1
º
w « III C 3
»
wJ 3
¬ ¼ 1 4 w III C 1 4
 III C 3  III C 3 III C C T
wC wC 3 wC 3 (8.115)
1 2
1 1 1
 III C 3 C  J 3 C 1
3 3
w III C
where we have used III C C T III C C 1 , (see Chapter 1). Additionally, we can
wC
obtain:
2 2
~ ~
wC w( J 3 C ) wC ij w( J 3 C
ij )
wC wC wC kl wC kl
2
2 2
w (C ) w( J 3 ) 2
w (C ij ) w( J 3 )
J 3 C J 3  C ij
wC wC wC kl wC kl
2
1
2
2 2
(8.116)
1
J 3 I  J 3 C C 1 1
3 J 3 E ik E jl  J 3 C
ij C kl
2 3
3 §I 1 ·
C 1 ¸
2
J ¨  C § 1 1 ·
© 3 ¹ J 3
¨ I ijkl  C ij C kl ¸
© 3 ¹
2
J 3 PT

with which, we introduce the fourth-order tensor P known as the projection tensor with
respect to the reference configuration, (see Holzapfel (2000)):
1 1
PT I C C 1 Ÿ P I  C 1 C (8.117)
3 3

F vol ˜ F vol
T
b vol ~ ~ ~
& C FT ˜F
B
X

pure dilatation ~
F vol F

2 ~
vol
F 1
C J 31
reference current
configuration configuration
~
B0
& F F ˜ F vol B
X &
x
C FT ˜F b F ˜FT

Figure 8.4: Multiplicative decomposition of deformation gradient – Kinematic tensors.

442 NOTES ON CONTINUUM MECHANICS

8.4.1 The Stress Tensors

Next, we will define the stress tensors in different configurations. To start off we will use
2 w: (C )
the definitions of the Cauchy stress tensor (current configuration) ı F˜ ˜FT
J wC
w: (C )
and the second Piola-Kirchhoff stress tensor S 2 JF 1
˜ı ˜F T
(reference
wC
configuration). Note that we can define a stress tensor, analogous to the Cauchy stress
tensor, in the intermediate configuration ( B ) by means of the transformation F vol , i.e.:

˜ w: (Cvol ) ˜ F vol
vol
2 T
ı vol vol
F vol (8.118)
J wC
1
w: (C vol ) w: ( J vol ) wJ vol
Then, if we refer to J vol J , F vol J 31 , , and
wC vol wJ vol wC vol
2
wJ vol J vol vol 1 J vol 3 1
C J , the equation in (8.118) becomes:
wC vol 2 2
1
: (C vol ) § w: ( J vol ) wJ vol · 13
2
˜w ˜ F vol 2
J 3 1 ˜ ¨¨ ¸˜ J 1
T
ı vol F vol ¸
J vol wC vol J vol © wJ
vol
wC vol ¹
1
2 § w: ( J vol ) J vol vol 1 · 13
vol
J 3 1 ˜ ¨¨ vol
C ¸˜ J 1
¸ (8.119)
J © wJ 2 ¹
2
1
§ w: ( J vol ) J vol  2 · 1 w: ( J vol )
J 31˜¨ J 3 1¸ ˜ J 3 1 1
J vol ¨ wJ vol 2 ¸ wJ vol
© ¹
Thus,
w: ( J )
ı vol 1 (8.120)
wJ
We can also define a stress tensor in the intermediate configuration caused by the
~
transformation F , (see Figure 8.5), as:
~ ~ ~
w: (C )
S 2 ~ (8.121)
wC
We will now observe additive decomposition of the strain energy function in two parts,
namely: isochoric and volumetric, i.e.:
~ ~ ~ ~
: ( F ) : ( F )  : vol ( F vol ) ; : (C ) : (C )  : vol (C vol ) (8.122)
Then, if we take the chain rule of derivative of the strain energy function (8.122) we obtain:
~ ~ ~ ~ ~
~ ~ w: (C ) wC d: vol ( J ) wJ w: (C ) ~ d: vol ( J ) 
: (C ) : (C )  : vol ( J ) ~ :  ~ :C  J (8.123)
wC wt dJ wt wC dJ
wJ  wJ J 1 J 1  ~
Now, given that J : C and C , we can obtain J C : C and the term C
wC wC 2 2
~ 2
~ wC 
can be expressed as C : C J 3 P T : C . So, the equation in (8.123) can also be
wC
expressed as follows:
 8 HYPERELASTICITY 443

~ ~
2
w: (C ) T  J d: vol ( J ) 1 
: (C ) J 3
~ :P :C  C :C
wC 2 dJ
2 ~ ~ (8.124)
w: (C )  J d: vol ( J ) 1 
J 3 P: ~ :C  C :C
wC 2 dJ

w: ( J )
ı vol 1
wJ ~ ~
~ w: (C )
S 2 ~
wC
B &
X
: vol ( F vol )
~ ~
F J 1
~
F vol F
pure dilatation ~ ~ ~
: iso ( F ) { : ( F )

w: ( J )
current ı vol 1
reference wJ
configuration configuration
~
B0
& F F ˜ F vol B
X &
x
~ ~ 2 w: (C )
: (C ) : (C )  : vol (C vol ) ı F˜ ˜FT
J wC
w: (C )
S 2 JF 1 ˜ ı ˜ F T
wC
2
~ w: vol ( J ) 1 ~ ~
S S  S vol with S vol J C , S J 3 P :S
wJ

Figure 8.5: Multiplicative decomposition of deformation gradient – stress tensors.

In purely elastic materials, internal energy dissipation is zero. Remember that in Chapter 5
in a system with no entropy production, internal energy dissipation in the reference
configuration is given by:
1
Dint S : C  : (C ) 0 (8.125)
2
Then, if we combine the equation in (8.124) with the one above we obtain:
2 ~ ~
1  w: (C )  d: vol ( J ) J 1 
Dint 3S :C  J ~ :C 
P: C :C 0
2 wC dJ 2
~ ~ (8.126)
§ 2
· 
¨ S  2 J 3 P : w: (C )  J d: ( J ) C 1 ¸ : C
vol

~ 0
¨ wC dJ ¸ 2
© ¹
444 NOTES ON CONTINUUM MECHANICS

Notice that the above must be satisfied for any admissible thermodynamic process. Let us
now consider that C z 0 , so, the only way for (8.126) to be satisfied is if:
2 ~ ~
w: (C ) d: vol ( J ) 1
S 2J 3 P : ~ J C (8.127)
wC dJ
Next, if we take the definition of the tensor S given in (8.11), and use the definition of
energy in (8.122), we can obtain:
~ ~
S 2
w: (C )
2
w ~ ~
wC wC
>
: (C )  : vol (C vol ) 2
w: (C )
2
w: vol (C vol ) ~
@
S  S vol (8.128)
wC wC
where it holds that:
w: vol (C vol ) w: vol ( J ) wJ w: vol ( J ) 1 w: vol ( J ) 1
S vol 2 2 2 JC 1 J C (8.129)
wC wJ wC wJ 2 wJ
and
~ ~ ~ ~ ~ 2
§ ~ ~ 2
~ w: (C ) w: (C ) wC 3 P :¨2
w: (C ) ·¸ ~
S 2 2 ~ : J ~ J 3 P :S (8.130)
wC wC wC ¨ wC ¸¹
©
~
where we have used the definition in (8.121). Additionally, we can verify that the tensor S
is in the intermediate configuration, i.e. it is only characterized by a change of shape, (see
Figure 8.5). Then, in summary we have:
~
S S  S vol (8.131)

where:
2 ~ ~ 2
~ ~ ~
~ w: (C ) w: iso (C )
S J 3 P :2 ~ J 3 P :S with S 2 ~ (8.132)
wC wC
d: vol ( J ) 1
S vol J C JpC 1 (8.133)
dJ
In addition, with the constitutive equation for hydrostatic pressure, Holzapfel (2000):

D: vol ( J )
p (8.134)
DJ
1
It is worth mentioning that the operator P I  C 1 C given in (8.132) provides the
3
correct deviatoric operator in the material (Lagrangian) description:

>x ( X& , t )@
dev
(x) 
1
3
>(x) : C @C 1 (8.135)

Thus,

>S~ @
2 2 2 2
~ ~ ª 1 1 º ~ ª~ 1 § ~ · 1 º dev
S «I  3 C « S  3 ¨© S : C ¸¹C »
J 3 P:S J 3 C» : S J 3 J 3 (8.136)
¬ ¼ ¬ ¼
Additionally, it holds that:

>S~ @ dev
:C 0 (8.137)
 8 HYPERELASTICITY 445

8.4.2 Compressible Isotropic Materials

In compressible isotropic materials, the energy function decomposition can be given by:
~ ~
: (C ) : (b) : (b )  : vol ( J ) (8.138)
wJ J 1
Then, as J III b , we can show that if equations C and (8.116) are valid, so are
wC 2
the following:
­ wb~ 2
§ 1 ·
° J 3
¨I  b b 1 ¸
wJ J 1 ° wb © 3 ¹
b ; ® ~ (8.139)
wb 2 ° wbij
2
§ 1 1 ·
° wb J 3
¨ E ik E jl  bij bkl ¸
¯ kl © 3 ¹
Then, if we refer to the constitutive equation for isotropic hyperelastic materials obtained
w: (b)
in (8.70), ı J 1 2 ˜ b J 1 2b ˜ w: (b) , and incorporate the definition of energy
wb wb
(8.138) into the stress equation, we can obtain:
~ ~
ı
w: (b)
J 1 2
wb
˜b J 1 2
wb
>
w ~ ~
: (b )  : vol ( J ) ˜ b @ J 1 2
w: (b )
wb
˜ b  J 1 2
w: vol ( J )
wb
˜ b (8.140)
~
ı  ı vol
Additionally, the volumetric contribution is:
w: vol ( J ) w: vol ( J ) wJ w: vol ( J ) J 1 w: vol ( J )
ı vol J 1 2 ˜b J 1 2 ˜b J 1 2 b ˜b 1 (8.141)
wb wJ wb wJ 2 wJ
And, the isochoric contribution is:
~ ~ 2 ~ ~
~ w: (b ) 3 §
1 · w: (b )
ı 1
J 2b ˜ 1
J 2b ˜ J ¨I  b
1
b¸ : ~
wb © 3 ¹ wb
§ 2 2
· ~ ~ ~
1 w: (b ) ~
b˜¨J 3 I  b 1 J 3 b ¸ ˜ b 1 : 2 J 1 ~ ˜b (8.142)
¨ 3 ¸ wb
© ¹
§ 2 ~ ~
~· ~ w: (b ) ~ §
¨
1
b ˜ ¨ J 3 I  b 1 b ¸ ˜ b 1 : 2 J 1
¸ ~ ˜b ¨I  1
1 · ~
1¸ : ı >ı~ @dev
© 3 ¹ wb © 3 ¹
2
~
where we have used the relationship b J 3 b , and it can be proven that if x is a second-
§ 1 ·
order tensor, then it holds that: ¨ I  1 1 ¸ : x x dev , where x dev represents the deviatoric
© 3 ¹
part of the tensor x , (see Problem 1.26).
Next, we will make the algebraic operations carried out in (8.142) using indicial notation:
446 NOTES ON CONTINUUM MECHANICS

~ ~ ~ ~ wb~ ·
§ w:
w: (b ) (b ) pq ¸
~
V J 1 2bik J 1 2bik ¨ ~
ij
wbkj ¨ wb wbkj ¸
© pq ¹
ª 2
º ~ ~
§ 1 · w: (b )
J 1 2bik « J 3 ¨ E pk E qj  b pq bkj1 ¸» ~
¬« © 3 ¹¼» wb pq
ª 2 § ~ ~
1 ·º w: (b )
J 1 2bik « J 3 ¨ E pk E qj  b pq bkj1 ¸» ~ E tp
¬« © 3 ¹¼» wbtq
ª 2 § ~ ~
1 ·º w: (b ) ~ ~
J 1 2bik « J 3 ¨ E pk E qj  b pq bkj1 ¸» ~ bts bsp1 (8.143)
¬« © 3 ¹¼» wbtq
E tp

ª§  2 2 ~ ~
·º w:
~ 1~ (b ) ~
J 1 2 «¨ J 3 bik E pk E qj bsp1  bsp1b pq J 3 bik bkj1 ¸» ~ bts
«¬¨© 3 ¸» wb
¹¼ tq
~ ~
ª§ ~ ~ 1~ ~ ·º w: (b ) ~
J 1 2 «¨ bip bsp1E qj  bsp1b pq bik bkj1 ¸» ~ bts
¬© 3 ¹¼ wbtq
~ ~
ª§ 1 ·º §¨ 1 w: (b ) ~ ·¸
«¨ E is E qj  3 E ij E sq ¸» ¨ J 2 ~ bts ¸
¬© ¹¼ © wbtq ¹
which is in accordance with (8.142).
In summary, in compressible isotropic materials, the following is satisfied:
~
ı ı  ı vol (8.144)

where

w: vol ( J ) § · ~
ı vol 1 p1 ,
~
ı
1
¨I  1 1¸ : ı >ı~ @dev (8.145)
wJ © 3 ¹
in which the constitutive equation for hydrostatic pressure is:

D: vol ( J )
p (8.146)
DJ

8.4.2.1 Compressible Isotropic Material in terms of the Invariants

In isotropic materials, the strain energy function can be expressed in terms of the principal
invariants as:
~ ~
: : ( I C~ , II C~ )  : vol ( J ) : ( I b~ , II b~ )  : vol ( J ) (8.147)

where : vol is a function of the third invariant of the right Cauchy-Green deformation
tensor ( III C J 2 ). This function, for an undeformed state, has to fulfill the following:
­ w: vol
C 1
o III C o®: vol
1 0 ; 0 (8.148)
¯ w III C
~
In Chapter 2, subsection 2.13, we obtained the principal invariants of C in terms of those
of C , i.e.:
 8 HYPERELASTICITY 447

2 4
IC II C
I C~ J 3 I
C I b~ ; II C~ J 3 II C II b~ ; III C~ III b~ 1 (8.149)
3 III C 3 III C2

Note that the invariants I C~ and II C~ are independent of the volumetric deformation.
We can now express the constitutive equation in the material description by means of S :
~
S S  S vol (8.150)
~
where S vol is the volumetric part, and S is the isochoric part, both of which are given
respectively by:
2 ~ ~ 2
~
d: vol ( J ) 1 ~ w: (C )
S vol J C JpC 1 ; S J 3 P:2
~ J 3 P :S (8.151)
dJ wC
~ ~
~ w: (C )
where S 2 ~ can be demonstrated by:
wC
~ ~
ª§ w: ~ ~
~ w: ( I C~ , II C~ ) ( I C~ , II C~ ) w: ( I C~ , II C~ ) ·¸ § w: ( I C~ , II C~ ) ·¸ ~ º
S 2 ~ 2«¨  I C~ 1  ¨ C» (8.152)
wC «¨ wI C~ w II C~ ¸ ¨ w II C~ ¸ »
¬© ¹ © ¹ ¼
where we have used one of the relationships obtained in (8.67).
Then, a very simple model for the volumetric part is : vol : ( F vol ) : vol ( J ) , where
: vol (J ) is given by:
N 2 N
: vol ( J ) J 1 III C  1 (8.153)
2 2
and where N is the bulk modulus. This model at the limit J o 0 has no physical meaning
N N 2
since : vol ( J o 0) . Therefore, we can add the term : vol ( J ) log J into the
2 2
equation, i.e.:
N 2
: vol ( J ) J  1  2 log J (8.154)
4
which validated the following: when J o 0 , the energy function tends towards infinity, i.e.
: vol (J o 0) f . Additionally, in a small deformation regime J | 1 , the term 2 log J o 0 .

8.5 Incompressible Materials

Many polymers can be subjected to large deformations without any volume change being
observed, Holzapfel (2000). Hence, these materials can be considered to be incompressible,
i.e. the continuum here is characterized by isochoric motion and the following is fulfilled:
det( F ) J O 1O 2 O 3 1 ; J2 III C III b 1 (8.155)
In incompressible materials, ( J 1) , the stress state is not completely determined by the
strain state, because in an incompressible body we can add hydrostatic stress (pressure) to
the current stress state without changing the strain state. Remember that, in isotropic
materials, a hydrostatic state produces only volumetric deformations and because of this,
448 NOTES ON CONTINUUM MECHANICS

the volumetric deformation in incompressible materials is equal to zero for any hydrostatic
state. Note that here, even energy is not affected by the volumetric part, since:
: ( F vol ) : (1) 0 .
~
According to the equation in (8.113), which is an incompressible case, it holds that F F,
~
F vol 1 , C C , C vol 1 .
Then, the hydrostatic stress state is given by:
ı hyd  p1 V ijhyd  pE ij (8.156)
where p denotes pressure. Then, if we refer to the relationship between the Cauchy stress
and the second Piola-Kirchhoff stress tensors: ı J 1 F ˜ S ˜ F T , we can obtain the
volumetric part of S :
S hyd J F 1 ˜ ı hyd ˜ F T  J p F 1 ˜ 1 ˜ F T  J p F 1 ˜ F T J p B  J p C 1 (8.157)

where B is the Piola deformation tensor, given by B F 1 ˜ F T C 1 , (see Chapter 2,

subsection 2.6).
Then, for incompressible materials we have:
w: (C )
S 2  J p C 1 (8.158)
wC
Next, to solve a problem with a constraint ( J 1) , we can introduce the Lagrange
multiplier H , which must satisfy the following:
: : (C )  H J  1 (8.159)
Now, the second Piola-Kirchhoff stress tensor can be obtained by taking the derivative of
the above equation with respect to C , i.e.:
w:
S 2 2
w
>: (C )  H J  1 @ 2 w: (C )  2H wJ 2
w: (C ) J
 2H C 1
wC wC wC wC wC 2
(8.160)
w: (C ) 1
2 J HC
wC
Now, if we compare (8.160) with (8.158) we can conclude that:
V kk
H p (8.161)
3
where p (pressure) is an unknown function to be determined by the incompressibility
condition.
Now, let us consider the Cauchy stress tensor decomposition into a spherical (hydrostatic)
and deviatoric part, i.e. ı ı dev  V m 1 , then, the second Piola-Kirchhoff stress tensor,
defined as S J F 1 ˜ ı ˜ F T , can be split as follows:
S JF 1 ˜ ı ˜ F T JF 1 ˜ (ı dev  V m 1) ˜ F T JF 1 ˜ ı dev ˜ F T  JF 1 ˜ V m 1 ˜ F T
(8.162)
S dev  JV m C 1 S dev  S hyd
Then, by comparing the result above with (8.160) we can clearly identify the deviatoric part
of the second Piola-Kirchhoff stress tensor.
After that, if we refer to the equation in (8.16) we can still write that in (8.160) as:
 8 HYPERELASTICITY 449

w: (C ) w: ( F )
F ˜S F ˜2  p F ˜ F 1 ˜ F T Ÿ P  p F T (8.163)
wC wF
and:
w: ( F )
P˜FT ˜ F T  p F T ˜ F T
wF
T (8.164)
w: ( F ) w: ( F ) ·
ı ˜ F T  p 1 F ˜ §¨ ¸  p1
wF © wF ¹
where we have considered the incompressibility condition J 1 . Then, in short, we can
state that the constitutive equation for stress in an incompressible continuum can be given
by:
w: (C )
S 2  p C 1
wC The constitutive
w: ( F ) equation for
P  p F T hyperelastic (8.165)
wF
T incompressible
w: ( F ) w: ( F ) ·
ı ˜ F T  p 1 F ˜ §¨ ¸  p1 materials
wF © wF ¹

8.5.1 Geometrical Interpretation

In this section we will attempt to make a graphic interpretation of the results obtained
previously, (see Bonet&Wood(1997)). To start with, let us consider the internal energy
dissipation equation obtained in (8.9):
w: (C ) · 
Dint §¨ S 
1
¸:C 0 (8.166)
©2 wC ¹
As we saw before, in incompressible materials, the stress state is not completely defined by
the strain state. Then, the term written in parentheses in the equation in (8.166) is not equal
to zero, which indicates that C is not arbitrary, i.e. it has restrictions. Remember that the
J 1 
rate of change of the Jacobian determinant is given by JC : C , (see Problem 2.12 in
2
Chapter 2). Additionally, if we consider the incompressibility condition J 1 , during
motion J 0 , we obtain:
J 1 
J C :C 0 (8.167)
2
The above equation gives us the restrictions on C which the equation in (8.166) has to
satisfy, thereby implying that:
1 w: (C ) J 1
S H C (unit of stress) (8.168)
2 wC 2
where H , as seen above, coincides with the hydrostatic pressure module.
Let us now consider an arbitrary plane defined by the normal n̂ . Next, we will project the
following tensors onto this plane, i.e.:
& §1 w: · & ( nˆ ) J 1 J & ( nˆ ) 1
C ˜ nˆ d ( nˆ ) ; ¨ S ¸ ˜ nˆ t ; C ˜ nˆ d (8.169)
© 2 wC ¹ 2 2
450 NOTES ON CONTINUUM MECHANICS

with which, the equations in (8.166) and (8.167) can be rewritten as:
& & J & ( nˆ ) 1 & ( nˆ )
t ( nˆ ) ˜ d ( nˆ ) 0 ; d ˜d 0 (8.170)
2
& & 1 &
which indicates that t (nˆ ) and d ( nˆ ) are orthogonal to d (nˆ ) , (see Figure 8.6).

& J & ( nˆ ) 1
t ( nˆ ) H d
2

J & ( nˆ ) 1
d
2

n̂ &
plane for possible values of d (nˆ )

&
d (nˆ )
plane normal to n̂

Figure 8.6: Incompressibility restriction.

8.5.2 Isotropic Incompressible Hyperelastic Materials

In isotropic incompressible hyperelastic materials, the scalar-valued tensor function
: : (C ) : (b) can be expressed in terms of the invariants I C I b and II C II b :
: : ( I C , II C ) : ( I b , II b ) (8.171)
or in terms of the principal invariants of the Green-Lagrange strain tensor:
: : ( I E , II E ) (8.172)
Then, the constitutive equation for stress in hyperelastic materials becomes:
w: ( I C , IIC ) w III C w: wI C w: w IIC w III C
S p  p (8.173)
wE wE wI C wE w IIC wE wE
Next, the relationships between the principal invariants of the tensors E and C (obtained
in Problem 2.10 in Chapter 2) are interrelated by:
1
IE IC  3 IC 2I E  3
2
1 Reciprocal
II E  2 I C  II C  3  o II C 4 II E  4 I E  3 (8.174)
4
1
III E III C  II C  I C  1 III C 8 III E  4 II E  2 I E  1
8
with which we can obtain the following derivatives:
 8 HYPERELASTICITY 451

wI C w w II C w
2I E  3 2 1 , 4 II E  4 I E  3 4 I E 1  E  4 1
wE wE wE wE
(8.175)
w III C w
8 III E  4 II E  2 I E  1 8 III E E 1  4 I E 1  E  2 1
wE wE
Then the equation in (8.173) becomes:

S
w:
wI C
21 
w:
w IIC
>
>4 I E 1  E  4 1@  p 8 III E E 1  4 I E 1  E  2 1 @
(8.176)
w:
wI C
21 
w:
w IIC
>
4> I E  1 1  E @  2 p 2 I E  1 1  2 E  4 III E E 1 @
Now, if we take into account that III C 1 into the equation in (8.174), we can then express
III E as a function of I E and II E , i.e.:
1
III E  2 II E  I E (8.177)
4
Then, by using the equation in (8.72), the constitutive equation for isotropic incompressible
materials becomes:
ª § w: w: · w: 2 º
ı «2¨¨  I b ¸¸b  2 b »  p1 (8.178)
«¬ © wI b w II b ¹ w II b »¼

Note that we can still express the constitutive equation for incompressible hyperelastic
materials in terms of principal stretches, by means of the equation in (8.96):
w:
Va p  Oa , a 1,2,3 (8.179)
wO a

8.5.2.1 Series Expansion of the Energy Function for an Isotropic

Incompressible Hyperelastic Materials

In incompressible materials it holds that III C 1 , and the energy function : : ( I C , II C )

can be represented by means a power series, (see equation (8.81)), so:
f
: : ( I C , II C ) ¦c
p ,q 0
pq IC  3
p
II C  3
q
(8.180)

8.6 Examples of Hyperelastic Models

Several models have been developed to simulate the phenomenological behavior of

hyperelastic materials. Here we mention some hyperelastic material models that can be
found in the literature.
Remember that in elastic (or hyperelastic) materials, the only remaining constitutive
equations are the one for energy and that for stress, one of which is redundant, that is, if
we know the energy we can find the stress and vice versa, (see Problem 6.1).
452 NOTES ON CONTINUUM MECHANICS

8.6.1 The Neo-Hookean Material Model

In the Neo-Hookean material model, the strain energy function is given in terms of the
isochoric part of C as follows:
N
: IC  3 c1 I C  3 (8.181)
2
N
where c1 . (The parameter N was originally determined by statistical mechanics,
2
N
Treloar (1944)), by N NT , where N is the number of polymer chains per unit of the
2
reference volume, N is the Boltzmann constant and T is the temperature. Lastly, the
parameter, N G can be determined experimentally and is known as the shear modulus.
Then, the stress constitutive equation for the Neo-Hookean material model becomes:
ı  p1  2c1 b (8.182)

8.6.2 The Ogden Material Model

This model expresses the strain energy function in terms of the principal stretches and is
given by:

>O O
M N
B B B
: (O 1 , O 2 , O 3 ) ¦a
p 1
p O1 p  O 2 p  O 3 p  3  ¦b
q 1
q 1 2
Cq
 O 1O 3
Cq

(8.183)
O 2O3
Cq
@
 3  h O 1O 2 O 3
where a p , bq are positive constants, a p t 1 , bq t 1 and h is a one-variable convex
function.

8.6.2.1 The Incompressible Ogden Material Model

This model expresses the strain energy function in terms of the principal stretches and is
given by:
N Np B B B
: (O 1 , O 2 , O 3 ) ¦B
p 1 p
O1 p  O 2 p  O 3 p  3 (8.184)

where N , N p , B p are the material constants. In general, the shear modulus N , in the
reference configuration, becomes:
N
2N ¦N
p 1
p Bp with N pB p ! 0 (8.185)

In the literature, e.g. Holzapfel (2000), we can find the following values for the constants
when p 3 :
B 1 1 .3 ; B 2 5.0 ; B 3 2.0
(8.186)
N1 6.3 u 10 5 N / m 2 ; N 2 0.012 u 10 5 N / m 2 ; N 3 0.1 u 10 5 N / m 2
Then, when N 1 and B 2 , the equation in (8.184) yields:
N N
: O21  O22  O23  3 IC  3 (8.187)
2 2
 8 HYPERELASTICITY 453

which is the Neo-Hookean material model given in (8.181).

8.6.2.2 The Hadamard Material Model

This model is a simplified Ogden material model, where it holds that M N 1 and
B 1 C 1 2 which reduces the equation in (8.183) to:
: : (C ) a1 I C  3  b1 IIC  3  h( J ) (8.188)
Then, taking into account the equations in (8.11) and (8.67), the constitutive equation
becomes:
w: (C ) ª wJ º
S 2 2 «a1 1  b1 II C C 1  III C C  2  h c( J ) (8.189)
wC ¬ wC »¼
Afterwards, the derivative of the Jacobian determinant with respect to the tensor C can be
evaluated as follows:
wJ w ª 1
º 1 1 w III C 1 1 1
III C III C III C III C C 1 JC 1 (8.190)
wC «¬ »
2 2 2
wC ¼ 2 wC 2 2
Finally, the equation in (8.189) may also be rewritten as:
ª § 1 · º
S 2«a1 1  ¨ b1 II C  h c( J ) J ¸C 1  b1 J 2 C  2 » (8.191)
¬ © 2 ¹ ¼

8.6.3 The Mooney-Rivlin Material Model

The Mooney-Rivlin material model was originally formulated to simulate rubber-like
materials, today it is also used to simulate biological tissue-like materials.

8.6.3.1 Strain Energy Density

This model has the same energy expression as that provided by (8.184). Then, with the
parameter values N 2 , B 1 2 , B 2 2 , and with the constraint O21O22 O23 1
(incompressibility), the strain energy density given in (8.184) becomes:
N1 N2
: (O 1 , O 2 , O 3 ) O21  O22  O23  3  O12  O22  O32  3 (8.192)
2 2
Note that O21  O22  O23 I C~ and:

1 1 1 O22 O23  O21O23  O21O22 IIC

  IIC (8.193)
O21 O22 O23 O21O22 O23 III C

since we have the constraint III C 1 we can summarize the strain energy density as follows:
N1 N2
: (C ) IC  3  IIC  3 c1 I C  3  c 2 IIC  3 (8.194)
2 2
N1 N2
where c1 and c 2  . Then, the terms I C  3 and IIC  3 ensure that the
2 2
strain energy is zero when there is no deformation ( E 0 ), since in this scenario and
according to the equation in (8.172), we will obtain I C 3 and IIC 3 . Note that in the
454 NOTES ON CONTINUUM MECHANICS

particular case when c 2 0 we revert to the Neo-Hookean material model given in

(8.181).

8.6.3.2 The Stress Tensor

The second Piola-Kirchhoff stress tensor for the Mooney-Rivlin material model becomes:
w: (C ) ª§ w: w: · § w: · § w: · º
S 2 2 «¨¨  I C ¸¸1  ¨¨ ¸¸C  ¨¨ III C ¸¸C 1 »
wC «¬© wI C w II C ¹ © w II C ¹ © w III C ¹ »¼
(8.195)
ª§ w: w: · § w: · º
2«¨¨  I C ¸¸1  ¨¨ ¸¸C » 2> c1  c 2 I C 1  c 2 C @
w
«¬© CI w II C ¹ © w II C ¹ ¼»
and the Cauchy stress tensor can be obtained as follows:
ı  p1  2c1b  2c 2 b 1 (8.196)

8.6.4 The Yeoh Material Model

8.6.4.1 Strain Energy Density

The Yeoh material model is used to simulate isotropic incompressible materials. Our
starting point here is the series expansion of strain energy density:
N
: : ( I C , II C , III C ) ¦c
p ,q ,r 1
pqr IC  3
p
II C  3
q
III C  1
r
(8.197)

Then, by considering the incompressible material, III C 1 , and also by discarding the
second invariant we obtain:
N 3
: : (I C ) ¦c
p 1
p IC  3
p
(8.198)

with which we can obtain the strain energy density for the Yeoh material model as follows:
: c1 I C  3  c 2 I C  3 2  c3 I C  3 3 (8.199)
where c1 , c 2 and c3 are material constants.

8.6.4.2 The Stress Tensor

In this model the second Piola-Kirchhoff stress tensor becomes:

w: (C ) ª§ w: w: · § w: · § w: · º
S 2 2 «¨¨  I C ¸¸1  ¨¨ ¸¸C  ¨¨ III C ¸¸C 1 »
wC w
¬«© CI w II C ¹ © w II C ¹ © w III C ¹ ¼» (8.200)
2
w:
wI C
>
1 2 c1  2c 2 I C  3  3c3 I C  3
2
@1

8.6.5 The Arruda-Boyce Material Model

The Arruda-Boyce Material Model, also called the 8-chain model, takes into consideration
that the shear modulus, N , depends on the strain. This phenomenon is detected in some
polymers.
 8 HYPERELASTICITY 455

Here, the strain energy density is given by:

N cp
: : (C ) N 0 ¦ 2 p2
I Cp  3 p (8.201)
p 1 M lock

where N 0 is the initial shear modulus, ci are constants obtained by statistical theory and
M lock and N are material constants. Then, if we consider that N 3 we obtain:
ª c2 c3 º
: (C ) N 0 «c1 I C  3  I C2  9  I C3  27 »
«¬ M 2
M
lock
4
lock »¼
(8.202)
ª1 1 11 º
N0 « IC  3  I C2  9  I C3  27 »
¬« 2 20M lock
2
1050M lock
4
¼»

8.6.6 The Blatz-Ko Hyperelastic Model

Porous polymers should be considered as compressible materials. Blatz-Ko(1962)
proposed the following strain energy density : ( I C , II C , III C ) based on experimental and
numerical results, with isochoric and volumetric parts:
ªN N º N ª§ II · 1 º
: ( I C , II C , III C ) f « IC  3  III CC  1 »  (1  f ) «¨¨ C  3 ¸¸  ( III CC  1)»
¬2 2C ¼ 2 ¬«© III C ¹ C »¼
(8.203)
Q
where C is given in terms of the N (shear modulus) and Q (Poisson’s ratio) by C ,
1  2Q
and f  >0,1@ is an interpolation parameter. In the particular case in which f 1 , we obtain:
N N
: ( I C , II C , III C ) IC  3  III CC  1 (8.204)
2 2C
In the incompressibility case ( III C 1 ), the equation in (8.203) becomes the Mooney-Rivlin
model, (see equation (8.194)). Then, in the restrictive case where f 1 , and with the
incompressibility condition ( III C 1 ), we revert to the Neo-Hookean incompressible model
given in (8.181).

8.6.7 The Saint Venant-Kirchhoff Model

8.6.7.1 Strain Energy Density

In the Saint Venant-Kirchhoff model, the strain energy density ( : ( I E , II E ) ) is given by:
1
: ( I E , II E ) M  2N I E2  2N II E (8.205)
2
where M and N are material constants.
We can express the strain energy density in terms of the C -invariants. To do so, let us
consider the relationships between the invariants of E and C , (see the equations in
(8.174)):
1 1 1
IE IC  3 ; II E  2 I C  II C  3 ; III E III C  II C  I C  1 (8.206)
2 4 8
456 NOTES ON CONTINUUM MECHANICS

and by substituting them into the equation in (8.205) we obtain:

2
1 ª1 º ª1 º
: ( I C , II C ) M  2N « I C  3 »  2N «  2 I C  II C  3 »
2 ¬2 ¼ ¬4 ¼ (8.207)
1 N
M  2N I C  3 2   2 I C  II C  3
8 2

8.6.7.2 The Stress Tensor

The derivatives of the function (8.205) with respect to the invariants are:
w: w:
M  2N I E ; 2N (8.208)
wI E w II E
and by using the equation in (8.79), where the parameters are:
w: w: w:
c0  IE  II E M  2N I E  2NI E OI E
wI E w II E w III E
(8.209)
w: w:
c1  IE 2N
w II E w III E
and by substituting the above parameters into the equation in (8.79), we obtain:
S MI E 1  2NE (8.210)
Then we can conclude that the Saint_Venant-Kirchhoff model describes geometric
nonlinearity, but is also characterized by a material linearity, i.e. the stress-strain relationship
is linear.
Then, taking into account that E 1
2
C  1 and I E 1
2
I C  3 , we can obtain the tensor
S by means of C as follows:
ªM º
S S (C ) M I E 1  2NE M 12 I C  3 1  2N 12 C  1 « I C  3  N »1  NC (8.211)
¬ 2 ¼
Note that the above equation could have been obtained by means of the constitutive
equation in terms of S given in (8.67), i.e.:
w: (C ) ª§ w: w: · § w: · § w: · º
S 2 2«¨¨  I C ¸¸1  ¨¨ ¸¸C  ¨¨ III C ¸¸C 1 »
wC «¬© wI C w II C ¹ © w II C ¹ © w III C ¹ »¼
ª§ w: w: · § w: · º ­2 N N ½ N
2«¨¨  I C ¸¸1  ¨¨ ¸¸C » 2® M  2N I C  3   I C  3 ¾1  C (8.212)
¬«© wI C w II C ¹ © w II C ¹ ¼» ¯8 2 2 ¿ 2
ªM º
« I C  3  N »1  NC
¬ 2 ¼

8.6.7.3 The Elastic Tangent Stiffness Tensor

The elastic tangent stiffness tensor can be obtained as follows:

wS w w wI E wE
C tan M I E 1  2N E { MI E 1  2N E M 1  2N (8.213)
wE wE wE wE wE
 8 HYPERELASTICITY 457

Note that E is a symmetric tensor, so the result of the operation E , E is also a symmetric
tensor. Then, the equation in (8.213) in indicial notation becomes:

C tan
wS ij
M
wI E
E ij  2N
wE ij
M
wI E
E ij  2N
w >
1
2
E ij  E ji @ (8.214)
ijkl
wE kl wE kl wE kl wE kl wE kl
Thus,

C tan
ijkl ME kl E ij  N>E ik E jl  E il E jk @ tensorial
 notation
o C tan M 1 1  2N I (8.215)

NOTE: Note that in a small deformation regime, the condition that all stress tensors are
equal is satisfied, i.e. S | ı , and the same is true for the strain tensors, E | İ , after which
the constitutive equation for stress in (8.210) becomes:
ı M I İ 1  2Nİ (8.216)
which is the same constitutive equation for isotropic linear elastic materials as that obtained
in Chapter 7, (see also Problem 6.1). In addition, the elastic tangent stiffness tensor C tan
coincides with the elasticity tensor C e , (see Chapter 7). ɶ

8.6.8 The Compressible Neo-Hookean Material Model

8.6.8.1 Strain Energy Density

In the compressible Neo-Hookean material model, the Helmholtz free energy per unit
reference volume (strain energy density), (see Bonet&Wood (1997)), is defined by:
M N
: (C ) ( lnJ ) 2  N lnJ  ( I C  3) : ( J )  : ( I C )
2 2 (8.217)
: (J ) : ( IC )

where M and N are material parameters.

8.6.8.2 The Stress Tensor

The second Piola-Kirchhoff stress tensor, (see Eq. (8.67)), is given by:
w: (C ) § w: ( I C , II C , III C ) w: ( J ) ·
S 2 2¨  ¸
wC © wC wC ¹
(8.218)
ª§ w: · § w: w: · 1 § w: · º w: ( J )
2«¨¨ ¸¸1  ¨¨ II C  III C ¸¸C  ¨¨ III C ¸¸C  2 »  2
«¬© wI C ¹ © w II C w III C ¹ © w II C ¹ »¼ wC

w: w: N w: w: ( J ) M 1 wJ 11
where 0, , 0 and 2( lnJ ) N JC 1 . Moreover,
wI C w II C
2 w III C wC 2 J wC J 2
wJ J 1
if we consider that C , we can conclude that:
wC 2
§N · §M N ·
S 2¨ 1 ¸  2¨¨ ( lnJ )C 1  C 1 ¸¸ N 1  C 1  M ( lnJ )C 1 (8.219)
©2 ¹ ©2 2 ¹
Then, by considering the relationship between the Cauchy stress tensor and the second
Piola-Kirchhoff stress tensor, J ı F ˜ S ˜ F T , as well as C 1 F 1 ˜ F T , we can obtain:
458 NOTES ON CONTINUUM MECHANICS

Jı F ˜S ˜F T >
F ˜ N 1  C 1  M ( lnJ )C 1 ˜ F T @
>
F ˜ N1F ˜ F  M ( lnJ ) F ˜ F
1 T T 1 T
@˜ F (8.220)
NF ˜ F T
 NF ˜ F 1 ˜ F T ˜ F T  M ( lnJ ) F ˜ F 1 ˜ F T ˜ F T
Thus,

ı
1
J
>
N b  1  M ( lnJ )1 @ (8.221)

8.6.8.3 The Elastic Tangent Stiffness Tensor

The elastic tangent stiffness tensor, (see Figure 8.3), can be defined as:
w 2: wS
C tan 4 2 (8.222)
wC wC wC
Then, given the second Piola-Kirchhoff stress tensor equation in (8.219), we obtain:

C tan 2
wC
w
>
N 1  C 1  M ( lnJ )C 1 @
(8.223)
ª w1 wC 1 w ( lnJ ) wC 1 º
2 «N N  M C 1  M (ln J ) »
¬ wC wC wC wC ¼

In Chapter 1 we obtained
wC ir1
wC kl

2
>
1 1 1
@
C ik C lr  C il1C kr1 , after which the above equation,

in indicial notation, becomes:

wC ir1 w ( lnJ ) wC 1
C tan
irkl 2N  2M C ir1  2M ( lnJ ) ir
wC kl wC kl wC kl
(8.224)
1 J 1
N C ik1C lr1  C il1C kr1  2M C ir1 C kl  M ( lnJ ) C ik1C lr1  C il1C kr1
J 2
and by simplifying we obtain:

C tan
irkl >N  M ( lnJ )@ C ik1C lr1  C il1C kr1  M C ir1C kl1
(8.225)
C tan >N  M ( lnJ )@ C 1
C 1  C 1 C 1  M C 1 C 1
We can also obtain the tensor L , (see Figure 8.3), which is related to the tensor C by the
equation in (8.32), i.e.:
L abcd Fbq Fap C tan
pqst Fct Fds

^> @ (8.226)
1 1
Fbq Fap N  M ( lnJ ) C ps C tq  C pt1C sq1  MC pq
1 1
C st Fct Fds `
Note that: Fbq Fap C ps1C tq1 Fct Fds Fbq Fap F px1 Fsx1 Ftw1 Fqw
1
Fct Fds E ax E bw E xd E wc E ad E bc ,

Fbq Fap C pt1C sq1 Fct Fds 1 1

E ac E bd , Fbq Fap C pq C st Fct Fds E ab E cd .
Then:

L abcd >N  M ( lnJ )@ E ad E bc  E ac E bd  M E ab E cd (8.227)

Note that E ad E bc  E ac E bd 2I abcd

sym
2I abcd , where I { I sym is the symmetric fourth-order
unit tensor, with which we can obtain:
 8 HYPERELASTICITY 459

L abcd > @
2 N  M ( lnJ ) I sym
abcd  OE ab E cd    o L
Tensorial notation
> @
2 N  M ( lnJ ) I  M1 1 (8.228)
Now, the tensor A (defined in (8.48)) becomes:

A
1
L 2
>N  M ( lnJ )@ I
M
1 1 (8.229)
J J J

Then, if N c
>N  M ( lnJ )@ and M c
M
, where N c and M c are the equivalent Lamé
J J
constants, it follows that
A 2N c I  M c 1 1 (8.230)
NOTE: In a small deformation regime, we have J | 1 , with which we obtain Nc | N and
M c | M , and the following also holds: C tan L A 2N I  M 1 1 C e . ɶ

Problem 8.1: Consider a motion characterized by dilatation. The continuum is made up of

an isotropic material resembling the compressible Neo-Hookean material model. Find the
Cauchy stress tensor in terms of the Jacobian determinant.
Solution:
In isotropic materials dilation can be characterized by dx i OdX i , (see Figure 8.7).
X3
ª dx1 º ªO 0 0 º ª dX 1 º dV
« dx » « 0 O 0 » « dX »
« 2» « » « 2»
«¬ dx 3 »¼ «¬ 0 0 O »¼ «¬ dX 3 »¼
dX 3 O dX 3

Stretches O O 1 O2 O3 dV 0
dX 1 X2
OdX 1
dX 2
X1 O dX 2

Figure 8.7: Dilatation.

1
In this scenario we have: F O1 Ÿ J { det ( F ) O3 Ÿ O J3
The Cauchy stress tensor can be obtained by means of the equation in (8.221), i.e.:
ı
1
J
>
N b  1  M ( lnJ )1 @
where the left Cauchy-Green deformation tensor can be evaluated as follows:
2
b F ˜FT O1 ˜ O1 O2 1 J 31
Therefore, the Cauchy stress tensor becomes:
ªN § 2 · M º
ı « ¨ J 3  1¸  ( lnJ ) » 1
¨
«¬ J © ¸
¹ J »¼
460 NOTES ON CONTINUUM MECHANICS

8.6.9 The Gent Model

Strain energy density in the Gent model, Gent(1996), is characterized by the following
logarithmic function:
N § I  3·
: (I C ) I m ln¨¨1  C ¸ (8.231)
2 © I m ¸¹

where N and I m are material constants, and I m is the constant that measures the limit
value of I C  3 .

8.6.10 The Statistical Model

We can summarize this model as follows.
Strain Energy Density
~
Let : be the isochoric part and : vol the volumetric part, then, the energy function is
given by:
~
: : ( I C~ )  : vol ( III C ) (8.232)

where
~ ª1 1 11
: N « I C~  3  I 2~  9  I 3~  27 
¬2 20 N C 1050 N 2 C
(8.233)
19 519 º
3
I C4~  81  4
I C5~  243  »
7000 N 673750 N ¼
N 2
: vol ln III C (8.234)
2
The Second Piola-Kirchhoff Stress Tensor

S a 0 1  a 2 C 1 Stress Tensor – Statistical Model (8.235)

where the coefficients a 0 , a1 and a 2 are given by:

§1 1 11 19 519 · 1
a0 2N¨  I C~  I 2~ 
2 C
I 3~ 
3 C
I 4~  ¸
4 C (8.236)
© 2 10 N 350 N 1750 N 134750 N ¹ 3 III C
2 §1 1 2 11 19
a2  N¨ I C~  I~  I 3~  I 4~ 
3 ©2 10 N C 350 N 2 C 1750 N 3 C
(8.237)
519 ·
I 5~  ¸  N ln III C
134750 N 4 C ¹
The Elastic Tangent Stiffness Tensor

a 2 1 1
C tan
ijkl b1E ij E kl  b2 E ij C kl1  C ij1E kl  (b4  b5 )C ij1C kl1  C ik C lj  C il1C kj1 (8.238)
2
where the parameters are:

§ 1 11 57 1038 · 1 § 1 ·
b1 2N¨  I~  I 2~  I 3~  ¸ ¨ ¸ (8.239)
2
© 10 N 175 N C 1750 N
3 C
67375 N 4 C ¹ 3 III C ¨ 3 III ¸
© C ¹
 8 HYPERELASTICITY 461

2 §1 1 33 38 519 · 1
b2  N¨  I~  I 2~  I 3~  I 4~  ¸ (8.240)
3 © 2 5 N C 350 N 2 C 875 N 3 C 26950 N 4 C ¹ 3 III C
2 §1 1 2 33 38 519 ·
b4 N¨ I ~  I~  I 3~  I 4~  I 5~  ¸ (8.241)
9 © 2 C 5 N C 350 N 2 C 875 N 3 C 26950 N 4 C ¹
N
b5 (8.242)
2
2 §1 1 2 11 19
a2  N¨ I C~  I~  I 3~  I 4~ 
3 ©2 10 N C 350 N 2 C 1750 N 3 C
(8.243)
519 ·
I 5~  ¸  N ln III C
134750 N 4 C ¹
The demonstration of the statistical model can be found in Chaves (2009), (see also
Sansour et al. (2003)).

8.6.11 The Eight-Parameter Model

We summarize this model as follows.
Strain Energy Density
~
: : ( I C~ , II C~ )  : vol ( III C ) (8.244)

where
~
: ( I C~ , II C~ ) B 1 I C~  B 2 II C~  B 3 I C2~  B 4 I C~ II C~  B 5 I C3~  B 6 II C2~  B 7 I C4~  B 8 I C~ II C2~ (8.245)
2
: vol
B 9 ln III C (8.246)
The parameters B 1 , B 2 , ..., B 8 are the eight material parameters whereas B 9 represents
the bulk modulus. The values these parameters were determined by Sansour (1998) as:
B 1 0.1796; B2 0.0145; B3 0.1684 u 10 2 ; B 4 0.3268 u 10 3
4 3 (8.247)
B 5 0.3473 u 10 ; B 6 0.8439 u 10 ; B 7 0.432 u 10 7 ; B8 0.5513 u 10 5
The Second Piola-Kirchhoff Stress Tensor

S a 0 1  a1C  a 2 C 1 Stress tensor – Eight-parameter model (8.248)

where
2
a0 B 1  B 2 I C~  2B 3 I C~  B 4 II C~  B 4 I C2~  3B 5 I C2~ 
3 III C (8.249)
 2B 6 I C~ II C~  4B 3
7 I C~ B 2
8 II C~  2B 2
8 I C~ II C~
1
a1 2 B 2  B 4 I C~  2B 6 II C~  2B 8 I C~ II C~ (8.250)
3
III C2
2
a2  B 1 I C~  2B 3 I C2~  3B 4 I C~ II C~  3B 5 I C3~  4B 7 I C4~  5B 8 I C~ II C2~
3 (8.251)
 2B 2 II C~  4B 6 II C2~  3B 9 ln( III C )
The Elastic Tangent Stiffness Tensor
462 NOTES ON CONTINUUM MECHANICS

C tan
ijkl b0 E ij E kl  b1 E ij C kl  C ij E kl  b2 E ij C kl1  C ij1E kl  b4 C ij C kl
a1 a (8.252)
 E ik E jl  E jk E il  b5 C ij1C kl  C ij C kl1  b8 C ij1C kl1  2 C ik1C lj1  C il1C kj1
2 2
where
2
b0 B 2  2B 3  3B 4 I C~  6B 5 I C~  2B 6 II C~  2B 6 I C~2 
3
III C2 (8.253)
12B 7 I C2~  6B 8 I C~ II C~  2B 8 I C3~
2
b1 B 4  2B 6 I C~  2B 8 II C~  2B 8 I C2~ (8.254)
III C
2 1
b2  B 1  2B 2 I C~  4B 3 I C~  3B 4 I C2~  3B 4 II C~  9B 5 I C2~  8B 6 I C~ II C~ 
3 3 III C (8.255)
 16B 3
7 I C~  5B 2
8 II C~  10B 2
8 I C~ II C~
2 2 1
b4 2B 6  2B 8 I C~ , b5  2B 2  3B 4 I C~  8B 6 II C~ 10B 8 I C~ II C~ (8.256)
3
III C4 3 3 III C2
2
b8 B 1 I C~  4B 2 II C~  4B 3 I C2~  9B 4 I C~ II C~  9B 5 I C3~  16B 6 II C2~ 
9 (8.257)
16B 7 I C4~  25B 8 I C~ II C2~  9B 9

The demonstration of the eight-parameter model can be found in Chaves (2009), (see also
Sansour et al. (2003)).

8.7 Anisotropic Hyperelasticity

Certain materials such as some biological tissues have fibers, and therefore lose their
isotropy. When these fibers are arranged according to a preferential direction, â 0 , we can
approach this material by means of the transversely isotropic material. In other material
such as heart tissue, the fibers are arranged according to two preferential directions, and are
classified as tissue with two families of fibers, (see Figure 8.8). In this subsection we just
describe the model for one family of fibers. Details about two families of fibers can be
found in Holzapfel (2000).

â 0

â 0

b̂ 0

b) One family of fibers b) Two families of fibers

Figure 8.8: Materials with fibers.

 8 HYPERELASTICITY 463

8.7.1 Transversely Isotropic Material

As discussed in Chapter 1, the scalar-valued tensor function : : (C ) can be written in
terms of the principal invariants of C , i.e. : : ( I C , II C , III C ) . Now, if the : -arguments
are C and the vector â 0 , i.e. : (C , aˆ 0 ) , it can be shown that this function can be written in
terms of the following invariants:
: (C , aˆ 0 ) : ( I C , II C , III C , aˆ 0 ˜ C ˜ aˆ 0 , aˆ 0 ˜ C 2 ˜ aˆ 0 ) : ( I C , II C , III C , I C( 4) , I C(5) ) (8.258)
where I C( 4) and I C(5) are the pseudo-invariants of anisotropy. Moreover, if we consider that
the energy, if independent of the sense of the vector â 0 , fulfills that : (C , aˆ 0 ) : (C ,aˆ 0 ) ,
we can then represent the strain energy function by:
: : (C , aˆ 0 aˆ 0 ) (8.259)
We can also show that the previous function is objective as follows:
: (C , aˆ 0 aˆ 0 ) : (Q ˜ C ˜ Q T , Q ˜ aˆ 0 aˆ 0 ˜ Q T ) : (Q ˜ C ˜ Q T , Q ˜ aˆ 0 Q ˜ aˆ 0 ) (8.260)
As the rotated current configuration is defined by the transformation F * Q ˜ F , (see
Chapter 4), then C * F * ˜ F * C is fulfilled, and a vector in the rotated current
T

configuration is given by aˆ *0 Q ˜ aˆ 0 , thus:

: (C , aˆ 0 aˆ 0 ) : (Q ˜ C ˜ Q T , Q ˜ aˆ 0 Q ˜ aˆ 0 ) : (C * , aˆ *0 aˆ *0 ) (8.261)
which proves objectivity.
Then, since the material has lost its isotropy, we can not express the energy function just in
terms of the principal invariants. To use the energy function in terms of invariants, Spencer
(1984) obtained two pseudo-invariants of anisotropy, which contribute to the energy
function, (see Holzapfel (2000)), and are given by:
I C( 4) aˆ 0 ˜ C ˜ aˆ 0 ; I C(5) aˆ 0 ˜ C 2 ˜ aˆ 0
(8.262)
I C( 4) aˆ 0 i C ij aˆ 0 j ; I C(5) aˆ 0 i C ip C pj aˆ 0 j

where I C( 4) aˆ 0 ˜ C ˜ aˆ 0 O2aˆ , and O is the stretch according to the â 0 -direction (fiber

0 â0
stretching).
So, the energy function can be expressed as follows:
: : ( I C , II C , III C , I C( 4) , I C(5) ) (8.263)
Now, the derivatives of I C( 4) and I C(5) with respect to the tensor C are given by:
wI C( 4 ) w ˆ wC ij
a 0 i C ij aˆ 0 j aˆ 0 i aˆ 0 j aˆ 0 i aˆ 0 j E ik E jl aˆ 0 k aˆ 0 l
wC kl wC kl wC kl
wI C(5) w ˆ wC ip wC pj (8.264)
a 0 i C ip C pj aˆ 0 j aˆ 0 i aˆ 0 j C pj  aˆ 0 i C ip aˆ 0 j
wC kl wC kl wC kl wC kl
aˆ 0 i aˆ 0 j C pj E ik E pl  aˆ 0 i C ip aˆ 0 j E pk E jl aˆ 0 k aˆ 0 j C lj  aˆ 0 i C ik aˆ 0 l

and such equations in tensorial notation become:

wI C( 4 ) wI C(5)
aˆ 0 aˆ 0 ; aˆ 0 C ˜ aˆ 0  aˆ 0 ˜ C aˆ 0 (8.265)
wC wC
464 NOTES ON CONTINUUM MECHANICS

Then, the constitutive equation for stress (reference configuration) can be represented by:
w: ( I C , II C , III C , I C( 4 ) , I C(5) )
S 2
wC
(8.266)
§ w: wI C w: w II C w: w III C w: wI
( 4) ( 5)
w: wI C ·
2¨    ( 4) C  ¸
¨ wI wC w II wC w III C wC wI C wC wI C wC ¸
© C C ¹
Next, if we consider the derivatives in (8.264) and the equation in (8.67) we obtain:
ª§ w: · § w: w: · § w: ·
S 2 «¨¨ ¸¸1  ¨¨ II C  III C ¸¸C 1  ¨¨ III C ¸¸C  2 
«¬© wI C ¹ © w II C w III C ¹ © w II C ¹
(8.267)
w: ˆ w: º
a0 aˆ 0  (5) aˆ 0 (C ˜ aˆ 0 )  (aˆ 0 ˜ C ) aˆ 0 »
wI C( 4) wI C ¼»
The second derivative of I C( 4) is:
w 2 I C( 4) w 2 I C( 4)
0 0 ijkl (8.268)
wC wC wC ij wC kl

Note that the second derivative of I C(5) becomes a fourth-order tensor that features both
major and minor symmetry:
w 2 I C(5) w 2 I C(5)
aˆ 0 aˆ 0 1 aˆ 0 i aˆ 0 j E kl (8.269)
wC wC wC ij wC kl
If we now consider the directions of the fibers in the current configuration (deformed
configuration), â , and using the expression of the Kirchhoff stress tensor, (8.71), we
obtain:
IJ Jı F ˜S ˜ F T
ª§ w: · § w: w: · § w: ·
2F ˜ «¨¨ ¸¸1  ¨¨ II C  III C ¸¸C 1  ¨¨ III C ¸¸C  2 
¬«© wI C ¹ w
© C II w III C ¹ w
© C II ¹ (8.270)
w: ˆ w: º
 a0 aˆ 0  (5) aˆ 0 (C ˜ aˆ 0 )  (aˆ 0 ˜ C ) aˆ 0 » ˜ F T
wI C( 4 ) wI C »¼

Then, if we consider that C FT ˜F , b F ˜ F T and F ˜ aˆ 0 O aˆ aˆ , the above equation

0

then becomes:
ª§ w: w: · w: w:
IJ J ı 2«¨¨  I b ¸¸b  b2  III b 1 
«¬© wI b w II b ¹ w II b w III b
(8.271)
w: ˆ w: ˆ º
 O aˆ 2
a aˆ  O2aˆ a (b ˜ aˆ )  (aˆ ˜ b) aˆ »
0 wI ( 4 ) wI b(5)
b
0
¼»
Moreover, by considering that I C( 4) aˆ 0 ˜ C ˜ aˆ 0 O2aˆ , we obtain:
0

ª§ w: w: · w: w:
IJ J ı 2«¨¨  I b ¸¸b  b2  III b 1 
«¬© wI b w II b ¹ w II b w III b
(8.272)
w: ˆ w: º
 I C( 4 ) a aˆ  I C( 4 ) (5) aˆ (b ˜ aˆ )  (aˆ ˜ b) aˆ »
wI b( 4 ) wI b ¼»