Section 1.

15
Solid Mechanics Part III Kelly 124
1.15 Tensor Calculus 2: Tensor Functions


1.15.1 Vector-valued functions of a vector

Consider a vector-valued function of a vector

) ( ), (
j i i
b a a = = b a a

This is a function of three independent variables
3 2 1
, , b b b , and there are nine partial
derivatives
j i
b a c c / . The partial derivative of the vector a with respect to b is defined to
be a second-order tensor with these partial derivatives as its components:

j i
j
i
b
a
e e
b
b a

c
c

c
c ) (
(1.15.1)

It follows from this that

1
|
.
|

\
|
c
c
=
c
c
a
b
b
a
or
ij
j
m
m
i
a
b
b
a
o =
c
c
c
c
=
c
c
c
c
, I
a
b
b
a
(1.15.2)

To show this, with ) ( ), (
j i i j i i
a b b b a a = = , note that the differential can be written as

|
|
.
|

\
|
c
c
c
c
+
|
|
.
|

\
|
c
c
c
c
+
|
|
.
|

\
|
c
c
c
c
=
c
c
c
c
=
c
c
=
3
1
3
2
1
2
1
1
1
1 1
1
a
b
b
a
da
a
b
b
a
da
a
b
b
a
da da
a
b
b
a
db
b
a
da
j
j
j
j
j
j
i
i
j
j
j
j


Since
3 2 1
, , da da da are independent, one may set 0
3 2
= = da da , so that

1
1
1
=
c
c
c
c
a
b
b
a
j
j


Similarly, the terms inside the other brackets are zero and, in this way, one finds Eqn.
1.15.2.


1.15.2 Scalar-valued functions of a tensor

Consider a scalar valued function of a (second-order) tensor

j i ij
T e e T T = = ), ( | |
.
This is a function of nine independent variables, ) (
ij
T | | = , so there are nine different
partial derivatives:

Section 1.15
Solid Mechanics Part III Kelly 125
33 32 31 23 22 21 13 12 11
, , , , , , , ,
T T T T T T T T T c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c | | | | | | | | |


The partial derivative of | with respect to T is defined to be a second-order tensor with
these partial derivatives as its components:

j i
ij
T
e e
T

c
c

c
c | |
Partial Derivative with respect to a Tensor (1.15.3)

The quantity T T c c / ) ( | is also called the gradient of | with respect to T.

Thus differentiation with respect to a second-order tensor raises the order by 2. This
agrees with the idea of the gradient of a scalar field where differentiation with respect to a
vector raises the order by 1.

Derivatives of the Trace and Invariants

Consider now the trace: the derivative of A tr , with respect to A can be evaluated as
follows:

I
e e e e e e
e e e e e e
A A A
A
A
=
+ + =

c
c
+
c
c
+
c
c
=
c
c
+
c
c
+
c
c
=
c
c
3 3 2 2 1 1
33 22 11
33 22 11
tr
j i
ij
j i
ij
j i
ij
A
A
A
A
A
A
A A A
(1.15.4)

Similarly, one finds that {Problem 1}

( ) ( ) ( )
( )
( ) ( )
I A
A
A
I A
A
A
A
A
A
A
A
A
I
A
A
2
3 2
T
2
3
T
2
) tr ( 3
) tr (
) tr ( 2
) tr (
3
tr
2
tr tr
=
c
c
=
c
c
=
c
c
=
c
c
=
c
c
(1.15.5)
Derivatives of Trace Functions

From these and 1.11.17, one can evaluate the derivatives of the invariants {Problem 2}:

( )
T T
2
T
T
III II I
III
I
II
I

= + =
c
c
=
c
c
=
c
c
A I A A
A
A I
A
I
A
A A A
A
A
A
A
Derivatives of the Invariants (1.15.6)


Section 1.15
Solid Mechanics Part III Kelly 126
Derivative of the Determinant

An important relation is

( ) ( )
T
det det

=
c
c
A A A
A
(1.15.7)

which follows directly from 1.15.6c.

Other Relations

The total differential can be written as

T
T
d
dT
T
dT
T
dT
T
d
:
13
13
12
12
11
11
c
c

+
c
c
+
c
c
+
c
c
=
|
| | |
|
(1.15.8)

This total differential gives an approximation to the total increment in | when the
increments of the independent variables ,
11
T are small.

The second partial derivative is defined similarly:

q p j i
pq ij
T T
e e e e
T T

c c
c

c c
c | |
2
, (1.15.9)

the result being in this case a fourth-order tensor.

Consider a scalar-valued function of a tensor, ) (A | , but now suppose that the
components of A depend upon some scalar parameter t: )) ( ( t A | | = . By means of the
chain rule of differentiation,

dt
dA
A
ij
ij
c
c
=
|
|

(1.15.10)

which in symbolic notation reads (see Eqn. 1.10.10e)

(
(

|
.
|

\
|
c
c
=
c
c
=
dt
d
dt
d
dt
d A
A
A
A
T
tr :
| | |
(1.15.11)


Identities for Scalar-valued functions of Symmetric Tensor Functions

Let C be a symmetric tensor,
T
C C = . Then the partial derivative of ( ) ) (T C | | = with
respect to T can be written as {Problem 3}

Section 1.15
Solid Mechanics Part III Kelly 127
(1)
C
T
T c
c
=
c
c | |
2 for T T C
T
=

(2) C
T T c
c
=
c
c | |
2 for
T
TT C = (1.15.12)

(3) T
C C
T T
C C
T
T c
c
+
c
c
=
c
c
=
c
c
=
c
c | | | | |
2 2 for TT C = and symmetric T

Scalar-valued functions of a Symmetric Tensor

Consider the expression

( )
( )
ij
ij
ij
A
A
B
c
c
=
c
c
=
|
|
A
A
B (1.15.13)

If A is a symmetric tensor, there are a number of ways to consider this expression: two
possibilities are that | can be considered to be
(i) a symmetric function of the 9 variables
ij
A
(ii) a function of 6 independent variables: ( )
33 23 22 13 12 11
, , , , , A A A A A A | | =
where
( )
( )
( )
32 23 32 23 23
31 13 31 13 13
21 12 21 12 12
2
1
2
1
2
1
A A A A A
A A A A A
A A A A A
= = + =
= = + =
= = + =


Looking at (i) and writing ( ) ( ) ( ) , , , ,
12 21 12 12 11
A A A A A | | = , one has, for example,

12 21 12 12
21
21 12
12
12 12
2
A A A A
A
A A
A
A A c
c
=
c
c
+
c
c
=
c
c
c
c
+
c
c
c
c
=
c
c | | | | | |
,

the last equality following from the fact that | is a symmetrical function of the
ij
A .

Thus, depending on how the scalar function is presented, one could write

(i) etc. , , ,
13
13
12
12
11
11
A
B
A
B
A
B
c
c
=
c
c
=
c
c
=
| | |

(ii) etc. ,
2
1
,
2
1
,
13
13
12
12
11
11
A
B
A
B
A
B
c
c
=
c
c
=
c
c
=
| | |




Section 1.15
Solid Mechanics Part III Kelly 128
1.15.3 Tensor-valued functions of a tensor

The derivative of a (second-order) tensor A with respect to another tensor B is defined as

q p j i
pq
ij
B
A
e e e e
B
A

c
c

c
c
(1.15.14)

and forms therefore a fourth-order tensor. The total differential A d can in this case be
written as

B
B
A
A d d :
c
c
= (1.15.15)

Consider now

l k j i
kl
ij
A
A
e e e e
A
A

c
c
=
c
c


The components of the tensor are independent, so

. etc , 0 , 1
12
11
11
11
=
c
c
=
c
c
A
A
A
A

nq mp
pq
mn
A
A
o o =
c
c
(1.15.16)

and so

I = =
c
c
j i j i
e e e e
A
A
, (1.15.17)

the fourth-order identity tensor of Eqn. 1.12.4.

Example

Consider the scalar-valued function | of the tensor A and vector v (the dot can be
omitted from the following and similar expressions),

( ) Av v v A = , |

The gradient of | with respect to v is

( )v A A vA Av
v
v
A v Av
v
v
v
T
+ = + =
c
c
+
c
c
=
c
c|


On the other hand, the gradient of | with respect to A is

v v v v v
A
A
v
A
= =
c
c
=
c
c
I
|

Section 1.15
Solid Mechanics Part III Kelly 129


Consider now the derivative of the inverse, A A c c

/
1
. One can differentiate 0 A A =
1

using the product rule to arrive at

A
A
A A
A
A
c
c
=
c
c

1
1


One needs to be careful with derivatives because of the position of the indices in 1.15.14);
it looks like a post-operation of both sides with the inverse leads to
( )
l k j i jl ik
A A e e e e A A A A A A = c c = c c
1 1 1 1 1
/ / . However, this is not correct
(unless A is symmetric). Using the index notation (there is no clear symbolic notation),
one has

( )
( )
l k j i j l ik
kl
ij
jn jl mk im mn
kl
im
jn
kl
mj
im jn mj
kl
im
l k j i
kl
mj
im mj
kl
im
A A
A
A
A A
A
A
A
A
A
A A A
A
A
A
A
A A
A
A
e e e e
e e e e
=
c
c

=
c
c

c
c
=
c
c

c
c
=
c
c

1 1
1
1 1
1
1 1 1
1
1
1
o o o
(1.15.18)



1.15.4 The Directional Derivative

The directional derivative was introduced in 1.6.11. The ideas introduced there can be
extended to tensors. For example, the directional derivative of the trace of a tensor A, in
the direction of a tensor T, is

( ) ( ) ( ) T T A T A T A
A
tr tr tr tr ] [ tr
0 0
= + = + = c
= =
c
c
c
c
c c
d
d
d
d
(1.15.19)

As a further example, consider the scalar function Av u A = ) ( | , where u and v are
constant vectors. Then

( ) ( ) | | Tv u v T A u T v u A
A
= + = c
=
c
c
|
c 0
] [ , ,
d
d
(1.15.20)

Also, the gradient of | with respect to A is

( ) v u Av u
A A
=
c
c
=
c
c|
(1.1.5.21)
Section 1.15
Solid Mechanics Part III Kelly 130

and it can be seen that this is an example of the more general relation

T
A
T
A
: ] [
c
c
= c
|
| (1.15.22)

which is analogous to 1.6.41. Indeed,

w
u
v
w v
T
A
T
w
x
w
u
A
x
c
c
= c
c
c
= c

c
c
= c
] [
: ] [
] [
|
|
|
|
(1.15.23)

Example (the Directional Derivative of the Determinant)

It was shown in 1.6.11 that the directional derivative of the determinant of the 2 2
matrix A, in the direction of a second matrix T, is

( )| |
12 21 21 12 11 22 22 11
det T A T A T A T A + = c T A
A


This can be seen to be equal to ( ) T A A : det
T
, which will now be proved more generally
for tensors A and T:

( ) ( )
( ) | |
( ) T A I A
T A I A
T A T A
A
1
0
1
0
0
det det
det
det ] [ det

=
=
+ =
+ =
+ = c
c
c
c
c
c
c
c
c
c
d
d
d
d
d
d


The last line here follows from (1.10.16a). Now the characteristic equation for a tensor B
is given by (1.11.4, 1.11.5),

( )( )( ) ( ) I B = = det 0
3 2 1


where
i
are the three eigenvalues of B. Thus, setting 1 = and T A B
1
= c ,

Section 1.15
Solid Mechanics Part III Kelly 131
( ) ( )( )( )
( )( )( )
( )
( ) T A A
A
A
A T A
T A T A T A
T A T A T A
T A T A T A
A
1
3 2 1
3 2 1
0
3 2 1
0
tr det
det
1 1 1 det
1 1 1 det ] [ det
1 1 1
1 1 1
1 1 1

=
=
=
+ + =
+ + + =
+ + + = c




c c c
c

c
c
c c c
c
d
d
d
d


and, from (1.10.10e),

( ) ( ) T A A T A
A
: det ] [ det
T
= c (1.15.24)


Example (the Directional Derivative of a vector function)

Consider the n homogeneous algebraic equations ( ) o x f = :

( )
( )
( ) 0 , , ,
0 , , ,
0 , , ,
2 1
2 1 2
2 1 1
=
=
=
n n
n
n
x x x f
x x x f
x x x f

The directional derivative of f in the direction of some vector u is

( ) ( ) ( )
( )
Ku
z
z
z f
u x z z f u x f
x
=
|
.
|

\
|
c
c
=
+ = = c
=
=
0
0
] )[ (
c
c
c
c c
c
d
d
d
d
(1.15.25)

where K, called the tangent matrix of the system, is

(
(
(
(

c c c c
c c c c c c
c c c c c c
=
c
c
=
n n n
n
n
x f x f
x f x f x f
x f x f x f
/ /
/ / /
/ / /
1
2 2 2 1 2
1 2 1 1 1

x
f
K , ( )u f u f
x
grad ] [ = c

which can be compared to (1.15.23c).


Properties of the Directional Derivative

The directional derivative is a linear operator and so one can apply the usual product rule.
For example, consider the directional derivative of
1
A in the direction of T:
Section 1.15
Solid Mechanics Part III Kelly 132

( ) ( )
1
0
1
] [

=

+ = c T A T A
A
c
c
c
d
d


To evaluate this, note that ( ) ( ) 0 T I T A A
A A
= c = c

] [ ] [
1
, since I is independent of A. The
product rule then gives ( ) ( ) ] [ ] [
1 1
T A A A T A
A A
c = c

, so that

( )
1 1 1 1 1
] [ ] [

= c = c TA A A T A A T A
A A
(1.15.26)

Another important property of the directional derivative is the chain rule, which can be
applied when the function is of the form ( ) ( ) x B f x f

) ( = . To derive this rule, consider (see

u
u
u
o
.

The left-hand side of the previous expression can also be written as

( ) ( ) ( )
( ) ( ) ]] [ [

) (

] [ ) (

u B B f x B f
u B x B f u x B f
x B
x
c c + ~
c + ~ +


Comparing these expressions, one arrives at the chain rule,

( ) ]] [ [

] [ u B B f u f
x B x
c c = c Chain Rule (1.15.28)

As an application of this rule, consider the directional derivative of
1
det

A in the
direction T; here, f is
1
det

A and ( ) ) (

A B f f = . Let
1
= A B and B f det

= . Then, from
Eqns. 1.15.24, 1.15.25, 1.10.3h, f,

( ) ( )
( ) ( ) ( )
( ) ( )
( ) T A A
TA A A A
TA A B B
T A B T A
A B A
: det
: det
: det
]] [ [ det ] [ det
T 1
1 1 T 1
1 1 T
1 1




=
=
=
c c = c
(1.15.29)


1.15.5 Formal Treatment of Tensor Calculus

Following on from 1.6.12 and 1.14.6, a scalar function R V f
2
: is differentiable at
2
V e A if there exists a second order tensor ( )
2
V Df e A such that
Section 1.15
Solid Mechanics Part III Kelly 133

( ) ( ) ( ) ( ) H H A A H A o Df f f + + = + : for all
2
V e H (1.15.30)

In that case, the tensor ( ) A Df is called the derivative of f at A. It follows from this that
( ) A Df is that tensor for which

| | ( ) ( ) B A B A B
A
c
c
c
+ = = c
=
f
d
d
Df f
0
: for all
2
V e B (1.15.31)

For example, from 1.15.24,

( ) ( ) ( ) T A A T A A T A
A
: det : det ] [ det
T T
= = c (1.15.32)

from which it follows, from 1.15.31, that

T
det det

=
c
c
A A A
A
(1.15.33)

which is 1.15.7.

Similarly, a tensor-valued function
2 2
: V V T is differentiable at
2
V e A if there
exists a fourth order tensor ( )
4
V D e A T such that

( ) ( ) ( ) ( ) H H A T A T H A T o D + + = + for all
2
V e H (1.15.34)

In that case, the tensor ( ) A T D is called the derivative of T at A. It follows from this that
( ) A T D is that tensor for which

| | ( ) ( ) B A T B A T B T
A
c
c
c
+ = = c
=0
:
d
d
D for all
2
V e B (1.15.35)


1.15.6 Problems

1. Evaluate the derivatives (use the chain rule for the last two of these)
( ) ( ) ( ) ( )
A
A
A
A
A
A
A
A
c
c
c
c
c
c
c
c
2 2 3 2
) tr (
,
) tr (
,
tr
,
tr

2. Derive the derivatives of the invariants, Eqn. 1.15.5. [Hint: use the Cayley-Hamilton
theorem, Eqn. 1.11.15, to express the derivative of the third invariant in terms of the
third invariant.]
3. (a) Consider the scalar valued function ( ) ( ) F C | | = , where F F C
T
= . Use the chain
rule
j i
ij
mn
mn
F
C
C
e e
F

c
c
c
c
=
c
c | |

to show that
Section 1.15
Solid Mechanics Part III Kelly 134
kj
ik
ij
C
F
F c
c
=
c
c
c
c
=
c
c | | | |
2 , 2
C
F
F

(b) Show also that
U
C C
U
U c
c
=
c
c
=
c
c | | |
2 2
for UU C = with U symmetric.
[Hint: for (a), use the index notation: first evaluate
ij mn
F C c c / using the product rule,
then evaluate
ij
F c c / | using the fact that C is symmetric.]
4. Show that
(a)
1 1
1
:

=
c
c
BA A B
A
A
, (b)
1 1 1
1
:

=
c
c
A A A A
A
A

5. Show that
T
T
: B B
A
A
=
c
c

6. By writing the norm of a tensor A , 1.10.14, where A is symmetric, in terms of the
trace (see 1.10.10), show that
A
A
A
A
=
c
c

7. Evaluate
(i) ( ) ] [
2
T A
A
c
(ii) ( ) ] [ tr
2
T A
A
c (see 1.10.10e)
8. Derive 1.15.29 by using the definition of the directional derivative and the relation
1.15.7, ( ) ( )
T
det / det

= c c A A A A .