You are on page 1of 56

# TENSORS

K. M. Udayanandan
Associate Professor
Department of Physics
Nehru Arts and Science College, Kanhangad

1
.

This note is prepared for the Post Graduate Students of Kannur Uni-
versity. The syllabus is
Tensor analysis-Introduction-definition-definition of different rank
tensors-Contraction and direct product-quotient rule-pseudo tensors-
General tensors-Metric tensors

Udayanandan. K. M
Contents

1 Introducing Tensors 5

1.1 Scalars or Vectors ? . . . . . . . . . . . . . . . . . . . . . . . . 6

1.2 Vector Division . . . . . . . . . . . . . . . . . . . . . . . . . . 7

1.3 Moment of inertia . . . . . . . . . . . . . . . . . . . . . . . . . 10

2 Re defining scalars and vectors 13

2.1 Cartesian Tensors . . . . . . . . . . . . . . . . . . . . . . . . . 14

2.1.1 Scalars . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

2.1.2 Vectors . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

2.1.3 Tensors . . . . . . . . . . . . . . . . . . . . . . . . . . 18

2.1.4 Summation Convention . . . . . . . . . . . . . . . . . . 20

3 Quotient Rule 21

4 Non-Cartesian Tensors-Metric Tensors 26

3
4.1 Spherical Polar Co-ordinate System . . . . . . . . . . . . . . . 26

4.2 Cylindrical coordinate system . . . . . . . . . . . . . . . . . . 29

5 Algebraic Operation of Tensors 31

5.0.1 Definition of Contravariant and Co variant vector . . . 31

5.0.2 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . 32

5.0.3 Co variant vector . . . . . . . . . . . . . . . . . . . . . 33

5.1 Addition & Subtraction of Tensors . . . . . . . . . . . . . . . 33

5.2 Symmetric and Anti symmetric Tensors . . . . . . . . . . . . . 33

5.3 Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34

5.4 Contraction, Outer Product or Direct Product . . . . . . . . . 35

6 Pseudo Scalars and Pseudo Vectors and Pseudo Tensors 38

6.1 Pseudo Vectors . . . . . . . . . . . . . . . . . . . . . . . . . . 38

6.2 Pseudo scalars . . . . . . . . . . . . . . . . . . . . . . . . . . . 40

6.3 General Definition . . . . . . . . . . . . . . . . . . . . . . . . 40

6.4 Pseudo Tensor . . . . . . . . . . . . . . . . . . . . . . . . . . . 40

7 University Questions-Solved 45
Chapter 1

Introducing Tensors

In our daily life we see large number of physical quantities. Tensor is the
mathematical tool used to express these physical quantities. Any physi-
cal property that can be quantified is called a physical quantity. The
important property of a physical quantity is that it can be measured and
expressed in terms of a mathematical quantity like number. For example,
”length” is a physical quantity that can be expressed by stating a number
of some basic measurement unit such as meters, while ”anger” is a property
that is difficult to describe with a number. Hence we will not call ’anger’ or
’happiness’ as a physical quantity. The physical quantities so far identified
in physics are given below. Carefully read them. While reading observe that
some are expressed unbold, some are bold fonted and some are large and
bold. The known physical quantities are absorbed dose rate, acceleration,
angular acceleration, angular speed, angular momentum, area, area
density, capacitance, catalytic activity, chemical potential, molar concentra-
tion, current density, dynamic viscosity, electric charge, electric charge

5
density, electric displacement, electric field strength, electrical con-
ductance, electric potential, electrical resistance, energy, energy density,
entropy, force, frequency, half-life, heat, heat capacity, heat flux density, il-
luminate, impedance, index of refraction, inductance, irradiance, linear
density, luminous flux, magnetic field strength, magnetic flux, magnetic
flux density, magnetization, mass fraction, (mass) Density, mean lifetime,
molar energy, molar entropy, molar heat capacity, moment of iner-
tia, momentum,permeability, permittivity, power, pres-
reaction rate, speed, specific energy, specific heat capacity, specific volume,
spin, stress, surface tension, thermal conductivity, torque, velocity, vol-
ume, wavelength, wave number, weight and work. Every physical quantity
must have a mathematical representation and only then a detailed study of
these will be s possible. Hence we have mathematical tools like theory of
numbers and vectors with which we can handle large number of physical
quantities.

1.1 Scalars or Vectors ?

Among the above physical quantities small bold faced quantities are vectors
and un bold are scalars. Generally we say quantities with magnitude only
as scalars and with magnitude and direction as vectors. But there are some
quantities which are given in large font which are not scalars and vectors.
If they are not scalars and vectors what are they? What is special
about these quantities ?. Let us have a look at it. One quality of the
above mentioned odd members is that some like mass, index of refraction,

6
permeability, permittivity sometimes behave as scalars and some times not.
The above mentioned physical quantities like mass, susceptibility. moment
of inertia, permeability and permittivity obey very familiar equations like.

F~ = m~a, P~ = χE,
~ L~ = I~ω , B
~ = µH,
~ D~ = E,
~ F~ = T A,
~ J~ = σ E
~

from which we can write
F~
m=
~a
P~
χ=
E~
~
L
I=
ω
~
~ = 5î+5ĵ +5k̂ and ω
etc. Consider the last case. Let L ~ = î+ ĵ + k̂. If you find
~ = 9î + 4ĵ + 11k̂
moment of inertia in this case you will get it as 5. But if L
~ with ω
and ω is not changed we will not be able to divide L ~ and get the
moment of inertia. Why this happen? What mathematical quantity is
mass, susceptibility or moment of inertia? To understand this we must have
a look at the concept of vector division once again.

1.2 Vector Division

Consider a ball thrown vertically downwards into water with a velocity

~v = 6k̂

7
After entering water the velocity is decreased but the direction may not
change. Then the new velocity may be

~v 0 = 3k̂ = 0.5 ~v

Thus we transform the old velocity to a new velocity by a scalar multiple.
But this is not true in all cases. Suppose the ball is thrown at an angle then
the incident velocity may look like

~v = 5î + 6ĵ + 8k̂

and the deviated ball in the water may have different possible velocity like

~v 0 = 3î + 2ĵ + 5k̂

~v 0 = 2î + 6ĵ + 7k̂

etc. Consider the first case. The components of the final vector (3,2,5) can
be obtained in different ways. Among them some are given below.

    
3
3 5
0 0 5
    
  
= 1  
 2   0 3 0  6 
    
5 0 0 58 8

The above transformation matrix is diagonal.
    
3
3 0 0 55
    
   2
 2  =  5 0 0  6 
 
    
5
5 0 6 0 8

8
This transformation matrix is not diagonal.
    
3 6
3 −1 5
5 8
    
   2 8
 2  =  5 6 −1   6 
 
    
6
5 1 1 −8 8

Such a 3 × 3 with all elements non-zero can also be used to transform the
old velocity to new one. Or in general

    
vx0 v v v v
   11 12 13  x 
 0  
 vy  =  v21 v22 v23
 
  vy 
    
0
vz v31 v32 v33 vz

This is the most general matrix which can be used to transform the
incident velocity to the new velocity. This shows that any vector can be
transformed to a new vector generally only by a 3 × 3 matrix in 3D. If the
matrix is diagonal and if the diagonal elements are same it becomes a scalar
multiple. We had seen that all our odd physical quantities always transform
one vector to a new vector. Hence the general form of these transforming
quantities must be a matrix with 9 components. Let us check whether this
is true with a specific example. For this let us find out what is the exact
nature of moment of inertia.

9
1.3 Moment of inertia

Finding the components of moment of inertia is the simplest example given in
many textbooks introducing nine component physical quantity. We repeat
it here for the simplicity and also for students who may be new at such
derivations. Consider ,
~ = I~ω
L

In terms of ~r and p~
~ = ~r × p~
L

= ~r × m~v

= ~r × m (~ω × ~r)

= m~r × (~ω × ~r)

We’ve
     
~ ~ ~ ~ ~ ~ ~
A × B × C = A.C B − A.B C ~ ~

∴ m~r × (~ω × ~r) = m (~r.~r) ω
~ − m~r (~r.~ω )

= mr2 ω
~ − m~r (~r.~ω )
h  i h  i
=m x2 + y 2 + z 2 ωx î + ωy ĵ + ωz k̂ −m xî + y ĵ + z k̂ (xωx + yωy + zωz )

~ are as follows,
then the three components of L

Lx = m y 2 + z 2 ωx − mxyωy − mxzωz


Ly = −myxωx + m x2 + z 2 ωy − myzωz


Lz = −mzxωx − mzyωy + m x2 + y 2 ωz


10
Lx = Ixx ωx + Ixy ωy + Ixz ωz

Ly = Iyx ωx + Iyy ωy + Iyz ωz

Lz = Izx ωx + Izy ωy + Izz ωz

where
x2 + y 2 + z 2 − x2 = m y 2 + z 2
  
Ixx = m

x2 + y 2 + z 2 − y 2 = m x 2 + z 2
  
Iyy = m

x2 + y 2 + z 2 − z 2 = m x2 + y 2
  
Izz = m

Ixy = −mxy = Iyz

Iyz = −myz = Izy

Izx = −mzx = Ixz

~ = I~ω can be written in the matrix form as
Thus L

    
L I I I ω
 x   xx xy xz  x 
 Ly  =  Iyx Iyy Iyz
    
  ωy 
    
Lz Izx Izy Izz ωz

Thus I is a physical quantity with nine components.
 
Ixx Ixy Ixz
 
I =  Iyx Iyy Iyz
 

 
Izx Izy Izz

11
 
I11 I12 I13
 
I =  I21 I22 I23
 

 
I31 I32 I33

Thus
    
Lx m (y 2 + z 2 ) −mxy −mxz ωx
    
 Ly  =  2 2
−myx −myz
    
m (x + z ) ω 
 y 
 
  
Lz −mzx −mzy m (x2 + y 2 ) ωz

In general we can write a componet as

Ii,j = m[r2 δij − ri rj ]

where i and j varies from 1 to 3 and r1 = x, r2 = y, r3 = z and r2 =
x2 + y 2 + x2 .

12
Chapter 2

Re defining scalars and vectors

We can see that scalars have one(30 ) component, vectors have 3 (31 ) com-
ponets and tensors we saw had 9 (32 ) components. This shows that all these
mathematical quantities belong to a family in 3 dimensional world. Hence
these three types of physical quantities must have a common type of defi-
nition. Since we classify them based on components we will redefine them
based on componets. Hence we can redefine the scalars and vectors using
coordinate transformation of components.

If the coordinate transformation is from cartesian to cartesian we call
all the quantities as cartesian tensors and if the transformation is from carte-

13
sian to spherical polar or cylindrical we call them as non cartesian tensors.
First we will study cartesian tensors.

2.1 Cartesian Tensors

2.1.1 Scalars

Under co-ordinate transformation, a scalar quantity has no change. ie When
we measure a scalar from a Cartesian or a rotated cartesian coordinate system
the value of scalar remains invariant Hence

S0 = S

For if we measure mass of some substance say sugar standing straight and
then slightly tilting we will get the same mass. Thus any invariant quantity
under coordinate transformation is defined as a scalar.

2.1.2 Vectors

Consider a transformation from Cartesian to Cartesian coordinate systems.
0 0
Consider X1 &X2 represent the unrotated co-ordinate system and X1 &X2
represent the rotated co-ordinate system.Let r be a vector From figure,

14
x1 = r cos φ

x2 = r sin φ

.
0
x1 = r cos (φ − θ) = r (cos φ cos θ + sin φ sin θ)
0
x2 = r sin (φ − θ) = r (sin φ cos θ − cos φ sin θ)
0
x1 = x1 cos θ + x2 sin θ
0
x2 = −x1 sin θ + x2 cos θ

In matrix form

    
0
x1 cos θ sin θ x1
 =  
0
x2 − sin θ cos θ x2

generally,
0
X
xi = aij xj
j

where i is 1 and 2 and j varies from 1 to 2.

0
x1 = a11 x1 + a12 x2

0
x2 = a21 x1 + a22 x2

Comparing
a11 = cos θ; a12 = sin θ

a21 = − sin θ; a22 = cos θ

15
Here we had taken position vector and performed transformation. Similarly
any vector can be transformed like this. Hence we can say a physical quantity
can be called as a vector if it obeys the transformation equation similar to
the transformation equation

0
X
xi = aij xj
j

Taking partial differential

0 0
∂x1 ∂x1
= a11 = a12
∂x1 ∂x2
0 0
∂x2 ∂x2
= a21 = a22
∂x1 ∂x2
Thus the transformation equation can also be written as

0
X ∂x0
i
xi = xj
j
∂x j

This is also the transformation equation for a vector. Let us proceed and find
the transformation from primed to unprimed coordinate system. We know
that for transformation from from X to X’
    
0
x1 cos θ sin θ x1
 =  
0
x2 − sin θ cos θ x2

   
0
x1 x1
  = A 
0
x2 x2

16
where  
cos θ sin θ
A= 
− sin θ cos θ

Now  
cos θ − sin θ
AT =  
sin θ cos θ

We can show that
AAT = I = AA−1

which means
A−1 = AT

   
0
x1 x1
A−1  0
 = A−1 A  
x2 x2

Rearranging    
x1 x01
  = A−1  
x2 x02
    
x1 cos θ − sin θ x01
 =  
x2 sin θ cos θ x02

∂x1 0 ∂x1 0
x1 = 0 x1 + 0 x
∂x1 ∂x2 2
∂x2 0 ∂x2 0
x2 = 0 x1 + 0 x
∂x1 ∂x2 2
Then
0
X
xi = aji xj
j

17
Thus we can conclude and say that a vector is a physical quantity which
transform like

0
X
Ai = aij Aj
j

where transformation is from primed coordinate system to un primed co-
ordinate system. Now if the transformation is from primed to unprimed
coordinate system the equation is

0
X
Ai = aji Aj
j

Both are definitions of a vector. This is very important and you must note
the indices.

2.1.3 Tensors

Now let us go for the mathematical expression for tensors

Consider the equation J~ = σ E.
~ Here σ is a tensor. Let us derive the
0
expression for σ which is the equation in the primed coordinate system. In
matrix form the above equation is

Ji = σij Ej

Taking prime on both sides
0 0 0
Ji = σij Ej

18
0
But the component of J transform like Ji = aip Jp . Substituting

0 0 0
Ji = σij Ej = aip Jp

Writing the expression for J we get

0 0 0
Ji = σij Ej = aip Jp = aip σpq Eq

Now using the inverse transformation equation for Eq ,

0 0 0
σij Ej = aip σpq ajq Ej

Then we get
0
σij = aip ajq σpq

Thus we can see that a tensor of rank 2 will have 2 coefficients or the rank
of a tensor can be obtained from counting the number of coefficients.

Thus we can define tensors in general as

0
A =A

tensor of rank zero or scalar

0
X
Ai = aij Aj
j

is tensor of rank one or vector and

0
XX
Aij = aip ajq Apq
p q

19
is a tensor of rank 2 and if have 3 coefficients we will get tensor of rank 3 etc.

2.1.4 Summation Convention

In writing an expression such as a1 x1 + a2 x2 + ..... + aN xN we can use the
N
aj xj . An even shorter notation is simply to write it as
P
short notation
j=1
j
aj x , where we adopted the convention that whenever an index (subscript or
superscript)is repeated in a given term we are to sum over that index from
1 to N unless otherwise specified. This is called the “summation conven-
tion. Any index which is repeated in a given term ,so that the summation
convention applies , is called dummy index or umbral index. An index
occurring only once in a given term is called a free index. In tensor analy-
sis it is customary to adopt a summation convention and subsequent tensor
equations in a more compact form. As long as we are distinguishing between
contravariance and covariance, let us agree that when an index appears on
one side of an equation, once as a superscript and once as a subscript (except
for the coordinates where both are subscripts), we automatically sum over
that index.

20
Chapter 3

Quotient Rule

In tensor analysis it is often necessary to ascertain whether a given
quantity is tensor or not and if it is tensor we have to find its rank. The direct
method requires us to find out if the given quantity obeys the transformation
law or not. In practice this is troublesome and a similar test is provided by
a law known as Quotient law. Generally we can write,

KA=B

Here A and B are tensors of known rank and K is an unknown quantity. The
Quotient Rule gives the rank of K. For example

~ = I~ω
L

Here ω
~ and L~ are known vectors, then Quotient Rule shows that I is a second
rank tensor. Similarly,
m~a = F~
~ = J~
σE
~ = P~
χE
all establish the second rank tensor of m, σ, χ. The well known Quotient
Rules are

21
Ki Ai = B
Kij Aj = Bi
Kij Ajk = Bik
Kijkl Aij = Bkl
Kij Ak = Bijk

In each case A and B are known tensors of rank indicated by the
no. of indices and A is arbitrary where as in each case K is an unknown quan-
tity. We have to establish the transformation properties of K. The Quotient
Rule asserts that if the equation of interest holds in all Cartesian co-ordinate
system, K is a tensor of indicated rank. The importance in physical theory is
that the quotient rule establish the tensor nature of quantities. There is an
interesting idea that if we reconsider Newtons equations of motion m~a = F~
on the basis of the quotient rule that, if the mass is a scalar and the force
a vector, then you can show that the acceleration ~a is a vector. In other
words, the vector character of the force as the driving term imposes its vec-
tor character on the acceleration, provided the scale factor m is scalar. This
will first make us think that it contradicts the idea given in the introduction.
But when we say m is scalar immediately we are considering that ~a and F~
have the same directions which makes m scalar.

Now let us prove each equation and find the nature of K.
Proof for each rule

1.
K Ai = B
Proof
Taking prime on both sides
0 0 0
K Ai = B

Here A has one index and hence it is a vector. Using the transformation

22
equation for a vector  0 
0 ∂xi
K Aj = B
∂xj
0
because B = B since it is a scalar. Now using the given rule RHS is
modified and we get
 0 
0 ∂xi
K Aj = KAj
∂xj
 0 
0 ∂xi
K − K Aj = 0
∂xj
Now Aj cannot be zero since it is a component and if it vanishes the law
itself does not exist. Hence the quantity within the bracket vanishes.
 0 
∂xi 0
K= K,
∂xj

Here the transformation is with one coefficient and thus K is a first
rank tensor

2. The second quotient rule
K Aj = Bi
Proof
Now we will proceed as in the case of I law. Taking prime on both sides
0 0 0
K Aj = Bi
0
!  0
0 ∂xj ∂xi
K Aα = Bl
∂xα ∂xl
0
!  0
0 ∂xj ∂xi
K Aα = KAα
∂xα ∂xl
0 0
!
0 ∂xj ∂xi
K − K Aα = 0
∂xα ∂xl
0
!
0 ∂xj ∂xl
K=K 0 ,
∂xα ∂xi

23
K is a second rank tensor

3. The third quotient rule
K Ajk = Bik
Proof
Taking prime
0 0 0
K Ajk = Bik
0 0
!  0 0 
0 ∂xj ∂xk ∂xi ∂xk
K Aαβ = Bγβ
∂xα ∂xβ ∂xγ ∂xβ
0
!  0
0 ∂xj ∂xi
K Aαβ = KAαβ
∂xα ∂xγ
0
!
0 ∂xj ∂xγ
K=K 0 ,
∂xα ∂xi
K is a second rank tensor

4.
KAij = Bkl
Proof
Taking prime
0 0 0
K Aij = Bkl
0 0
!  0 0 
0 ∂xi ∂xj ∂xk ∂xl
K Aαβ = Bγν
∂xα ∂xβ ∂xγ ∂xν
0 0
!  0 0 
0 ∂xi ∂xj ∂xk ∂xl
K Aαβ = KAαβ
∂xα ∂xβ ∂xγ ∂xν
0 0
!
0 ∂xi ∂xj ∂xγ ∂xν
K=K 0 0
∂xα ∂xβ ∂xk ∂xl

K is a 4th rank tensor

5.
K Ak = Bijk

Proof

24
Taking prime
0 0 0
K Ak = Bijk
 0 0 0 0
0 ∂xk ∂xi ∂xj ∂xk
K Al = Bαβl
∂xl ∂xα ∂xβ ∂xl
 0 0 0 0
0 ∂xk ∂xi ∂xj ∂xk
K Al = KAl
∂xl ∂xα ∂xβ ∂xl
!
0 ∂xα ∂xβ
K=K 0 0
∂xi ∂xj
K is a second rank tensor. The quotient rule is a substitute for the
illegal division of tensors.
Problems

1. The double summation Kij Ai B j is invariant for any two vectors Ai and
Bj . Prove that Kij is a second-rank tensor.

2. The equation Kij Ajk = Bik holds for all orientations of the coordinate
~ and B
system. If A ~ are arbitrary second rank tensors show that K is a
second rank tensor.

25
Chapter 4

Non-Cartesian Tensors-Metric
Tensors

The metric tensor gij is a function which tells how to compute the distance
between any two points in a given space. Its components can be viewed
as multiplication factors which must be placed in front of the differential
displacements dxi in a generalized Pythagorean theorem. We can find the
metric tensor in spherical polar coordinates and cylindrical coordinates.

4.1 Spherical Polar Co-ordinate System

In Spherical Polar Co-ordinate System,

26
x = r sin θ cos φ
y = r sin θ sin φ
z = r cos θ
Then,
dx = sin θ cos φdr + r cos φ cos θdθ − r sin θ sin φdφ
dy = sin θ sin φdr + r cos θ sin φdθ + r sin θ cos φdφ
dz = cos θdr − r sin θdθ
In Cartesian Coordinate system

ds2 = dx2 + dy 2 + dz 2

Corresponding to this in Spherical polar coordinates

ds2 = (sin θ cos φdr+r cos φ cos θdθ−r sin θ sin φdφ)(sin θ cos φdr+r cos φ cos θdθ−r sin θ sin φdφ)+

(sin θ sin φdr+r cos θ sin φdθ+r sin θ cos φdφ)(sin θ sin φdr+r cos θ sin φdθ+r sin θ cos φdφ)+
(cos θdr − r sin θdθ)(cos θdr − r sin θdθ)
ds2 = sin2 θ cos2 φdr2 + r sin θ cos φ cos2 θdr dθ − r sin2 θ sin φ cos φdr dφ+
r sin θ cos θ cos2 φdr dθ + r2 cos2 θ cos2 φ dθ2 − r2 sin θ cos θ sin φ cos φdθ dφ−
r sin2 θ sin φ cos φdφdr − r2 sin θ cos θ sin φ cos φdθ dφ + r2 sin2 θ sin2 φdφ2 +
sin2 θ sin2 φdr2 + r sin θ cos θ sin2 φdr dθ + r sin2 θ sin φ cos φdr dφ+

27
r sin2 θ sin φ cos φdφ dr + r2 sin θ cos θ sin φ cos φdφ dθ + r2 sin2 θ cos2 θdφ2 +
cos2 θdr2 − r sin θ cos θdr dθ − r sin θ cos θdr dθ + r2 sin2 θdθ2
= dr2 + r2 dθ2 + r2 sin2 θdφ2
 2    2 
dx 1 0 0 dr
 dy 2  =  0 r2 0   dθ2 
dz 2 0 0 r2 sin2 θ dφ2
ie,
   
1 0 0 g11 g12 g13
gij =  0 r2 0  =  g21 g22 g23 
2 2
0 0 r sin θ g31 g32 g33

ds2 = g11 dr2 + g22 dθ2 + g33 dφ2
F or i 6= j , gij = 0
In general

= g11 drdr+g12 drdθ+g13 drdφ+g21 dθdr+g22 dθdθ+g23 dθdφ+g31 dφdr+g32 dφdθ+g33 dφdφ
X
ds2 = gij dqi dqj
i,j

Using sign convention
ds2 = gij dqi dqj
Here ds2 is a scalar. We know from quotient rule that if

K Ai = B

K is a vector and hence gij dqj is a vector. Now if

K Ai = Bj

K is a second rank tensor and hence gij is a tensor of rank two. Thus gij is
called the metric tensor of rank two.

28
4.2 Cylindrical coordinate system

In Circular Cylindrical Co-ordinate System,

x = ρ cos φ

y = ρ sin φ
z=z
Differentiating and substituting we get

ds2 = dr2 + ρ2 dφ2 + dz 2

dx2
    2 
1 0 0 dρ
 dy 2  =  0 ρ2 0   dφ2 
dz 2 0 0 1 dz 2
Thus gij for cylindrical coordinate system can be found out.
Problem
In Minkowiski space we define x1 = x , x2 = y , x3 = z , and x0 = ct. This is
done so that the space time interval ds2 = dx20 − dx21 − dx22 − dx23 (c =velocity
of light). Show that the metric in Minkowiski space is
 
1 0 0 0
 0 −1 0 0 
 
(gij ) = 
 0 0 −1 0 

0 0 0 −1

29
Solution
x1 = x ⇒ dx1 = dx
x2 = y ⇒ dx2 = dy
x3 = z ⇒ dx3 = dz
x0 = ct ⇒ dx0 = c dt

Space time interval

ds2 = ds2 = dx20 − dx21 − dx22 − dx23

= c2 dt2 − dx2 − dy 2 − dz 2
Then     
dx20 1 0 0 0 c2 dt2
dx21   0 −1 0 0  dx2
    
=
 
dx22   0 0 −1 0   dy 2
  
 
dx23 0 0 0 −1 dz 2
 
1 0 0 0
 0 −1 0 0 
 
gij = 
 0 0 −1 0 

0 0 0 −1

30
Chapter 5

Algebraic Operation of Tensors

5.0.1 Definition of Contravariant and Co variant vec-
tor

Contravariant vector
We’ve X
x0i = aij xj
j

or in general
X ∂xi0
i0
A = Aj
j
∂xj

Any vector whose components transform like this expression is defined as
Contravariant vector Here we have used superscript to denote the compo-
nent. This is to differentiate it from another type of vector which we will
shortly define. Examples are displacement, velocity, acceleration etc

31
5.0.2 Exercises

1. If xi be the co-ordinate of a point in 2-dimensional space. Show that
dxi are component of a contravariant tensor?
Soln: Earlier we have studied transformation from X to X’ coordinates.
There we have
0
x1 = x1 cos θ + x2 sin θ
0
x2 = −x1 sin θ + x2 cos θ
0
Writing in the superscript form we have x1 is a function of x1 and x2
0 0
xi = xi x1 , x2


Using the method of partial differential
0 0
0 ∂xi 1 ∂xi 2
dxi = dx + dx
∂x1 ∂x2
X ∂xi0
i0
dx = j
dxj
j
∂x

It is the transformation equation for a contravariant vector

2. Show that the velocity of a fluid at any point is component of a con-
travariant vector?
Soln:
Velocity is rate of change of displacement. Hence we can take the
equation for displacement from the previous problem.
0 0
∂xi 1 ∂xi 2
i0
dx = dx + dx
∂x1 ∂x2
0 0 0
dxi ∂xi dx1 ∂xi dx2
= +
dt ∂x1 dt ∂x2 dt
0 0
dxi X ∂xi dxj
=
dt j
∂xj dt
X dxi0
i0
V = Vj
j
dxj

32
0
where V i and V j are the components of velocity.
It is a contravariant vector.

5.0.3 Co variant vector

Another vector which is the derivative (gradient) of a scalar transform like
this,
∂φ0 ∂φ ∂φ ∂xj
0
= 0
=
∂xi ∂xi ∂xj ∂x0i
or,
∂φ X ∂xj ∂φ
=
∂x0i j
∂x0i ∂xj

such a vector is called Co-variant vector. Generally,

0
X ∂xj
Ai = Aj
j
∂x0i

5.1 Addition & Subtraction of Tensors

If A & B are tensors of same rank and both expressed in a space of the
same number of dimensions, then

Aij + B ij = C ij
Aij − B ij = Dij

5.2 Symmetric and Anti symmetric Tensors

If Amn represents a tensor,if for all m and n

Amn = Anm

33
we call it as a symmetric tensor and if on other hand

Amn = −Anm

the tensor is anti symmetric.
We can write,
   
mn 1 mn 1 mn 1 mn 1 nm 1 mn 1 nm
A = A + A = A + A + A − A = B mn +C mn
2 2 2 2 2 2
Interchanging the indices we can show that

B nm = B mn

which is symmetric
C nm = −C mn
which is antisymmetric. So Amn can be represented as a combination of sym-
metric and anti symmetric parts.

5.3 Problems
1. If a physical quantity has no component in one coordinate system, then
show that it does not have a component in other coordinate systems
We have transformation equations like

ij 0 ∂x0i ∂x0j kl
A = A
∂xk ∂xl
X dxi0
i0
V = Vj
j
dxj

Vθ = Vy
dy
These equations show that if a physical quantity exist in one coordinate
system it exists in other coordinate system also. Hence if it vanishes in
a system it vanishes in other systems also.

34
2. The components of a tensor A is equal to the corresponding compo-
nents of tensor B in one particular coordinate system. Show that A=B
Consider transformation equation from X to X’
0
x1 = x1 cos θ + x2 sin θ
0
x2 = −x1 sin θ + x2 cos θ

This is the transformation for a position vector say A whose compo-
nents are (x1 , x2 ). If we consider another position vector B and if its
components are same as that of A then it is evident from the above
transformation equation that A= B

5.4 Contraction, Outer Product or Direct Prod-
uct

The product of two tensors is a tensor whose rank is the sum of the ranks
of the given tensors. This product which involves ordinary multiplication of
the components of the tensor is called outer product or direct product.

For example consider the product of two different vectors

∂xk
ai 0 = ak
∂xi0

a first rank tensor and
0 ∂xj 0 l
bj = b
∂xl
another first rank tensor

Then the product

0 ∂xk ∂xj 0
ai b j = ak bl
∂xi0 ∂xl

35
gives a second rank tensor. In general
0 0 0
0 0 0 ∂xi ∂xn ∂xk ∂xl mpq
Aij B kl = Cjikl = C
∂xm ∂x0j ∂xp ∂xq n
0 0 0
where Aij is of rank 2 , B kl is of rank 2 and Cjikl is of rank 4.

Contraction of Tensors
When dealing with vectors, we formed a scalar product by summing products
of corresponding components

~B
A. ~ = A1 B1 + A2 B2 + A3 B3
X
= Ai Bi
i

or, by using summation convention,

~B
A. ~ = Ai Bi

The generalization of this expression in tensor analysis is a process
known as contraction. Two indices, one co variant and the other contravari-
ant are set equal to each other and then we sum over this repeated index.
0
0 ∂xi ∂xl k
Bji = 0 B
∂xk ∂xj l

This is rank 2 tensor. Let j = i Then
0
0 ∂xi ∂xl k
Bii = 0B
∂xk ∂xi l

∂xl
=
∂xk
= δkl Blk
= Bkk
It is of rank 0. Here,Bkk is scalar. So contraction reduces the rank by two.
Exercises

36
1. Show that Kronecker delta is a mixed tensor of rank 2?
Soln: 0 0 0
0 ∂xi ∂xi ∂xj
δji = 0 = ∂x l ∂xm
∂xj ∂xl ∂xm
0
∂xi ∂xm ∂xl
=
∂xl ∂xj 0 ∂xm
0
∂xi ∂xm l
= δ
∂xl ∂xj 0 m
Thus Kronecker delta is a mixed tensor of rank 2

2. Ai.....n what will be the rank of its derivative?
∂Ai
Soln: we take the derivative as 0
∂xi

∂Ai ∂Ai ∂xj
0 = 0
∂xi ∂xj ∂xi

∂Ai ∂Ai
0 = Ti & = Tj
∂xi ∂xj
then the equation becomes

∂xj
Ti = 0 Tj
∂xi
∂Ai...n
so the rank of ∂xi
0 is n+1. Differentiation increases the rank of a
tensor.

37
Chapter 6

Pseudo Scalars and Pseudo
Vectors and Pseudo Tensors

6.1 Pseudo Vectors

Let us again look into some properties of the vectors and scalars. So far
our coordinate transformations have been restricted to pure passive rota-
tions. We now consider the effect of reflections or inversions. We have seen
in branches of physics like nuclear physics that mirror symmetry is very im-
portant. Now let perform reflection on scalars and vectors. Then we observe
interesting properties . An ordinary vector ~r under reflection remains as
such. Thus
~r = xî + y ĵ + z k̂
under reflection becomes

~r0 = −x × −î + −y × −ĵ + −z × −k̂

here both unit vectors and the component got reflected.

~r0 = ~r

38
Then ~r is said to be a polar vector which is the name given to an unchanged
vector. If we are considering the components of polar vector, under reflection
there is flip in their sign.

x1 , x2 , x3 → −x1 , −x2 , −x3
 

There is another type of vector which can be always represented by cross
product of vectors. Consider a cross product

~ =A
C ~×B
~

C1 î + C2 ĵ + C3 k̂ = î (A2 B3 − A3 B2 ) + ĵ (A3 B1 − A1 B3 ) + k̂A1 B2 − A2 B1
C1 = A2 B3 − A3 B2
Under reflection

C10 = (−A2 × −B3 ) − (−A3 × −B2 )

C10 = +C1 itself
similarly C2 and C3 . Here the components remain as such under reflection.

(C1 , C2 , C3 ) → (C1 , C2 , C3 )

Such vectors are known as Axial/Pseudo vector fhere is no flip in their
~ under reflection
sign of component of an axial vector. But the vector C

~ 0 = C1 × −î + C2 × −ĵ + C3 × −k̂
C 1

~ 0 = −C
C ~
1

The vector change its sign, components do not. In physics we have such
physical quantities which always appear in the cross product form like

~ = ~r × p~
L

~τ = ~r × F~
~ × ~r
~v = ω
∂B~
=∇×E ~
∂t

39
~ ,~τ , ω, B
Here L ~ are pseudo vectors.

6.2 Pseudo scalars

It can be proved that physical quantities obeying equations like (A × B) · C
flip their sign. Such quantities are called Pseudo Scalars but ordinary scalars
do not change sign.

6.3 General Definition

Generally we say
S 0 = J S → P seudoscalar
~ 0 = Jaij C
C ~ j → P seudovector
i

A0ij = Jaik ajl Akl → P seudotensor
where J = −1

6.4 Pseudo Tensor

Levi Civita symbol

It is named after the Italian mathematician and physicist Tullio Levi-Civita.
In three dimensions, the Levi-Civita symbol is defined as follows:

+1 if (i, j, k) is (1, 2, 3), (3, 1, 2) or (2, 3, 1),

εijk = −1 if (i, j, k) is (1, 3, 2), (3, 2, 1) or (2, 1, 3),

0 if i = j or j = k or k = i

40
The Levi-Civita symbol is not a physical quantity but it is always associated
with some physical quantity. To establish this consider cross product of two
~ and B
vectors A ~
~ =A
C ~×B ~

î ĵ k̂

= A1 A2 A3

B1 B2 B3
C1 = A2 B3 − A3 B2
C2 = A3 B1 − A1 B3
C3 = A1 B2 − A2 B1
The three equation can be generalized as

Ci = εijk Aj Bk

Putting i = 1, 2, 3 etc. we can deduce C1 , C2 , C3
Putting i=1
X3
C1 = ε1jk Aj Bk
j,k=1
X
= [ε11k A1 Bk + ε12k A2 Bk + ε13k A3 Bk ]
k

= A2 B3 − A3 B2
Similarly putting i=2, i=3 we can deduce C2 and C3 .
Now we will show that Levi-Civita[LC] symbol εijk is a pseudo tensor?
Proof:If Levi Civita symbol is a pseudo tensor then its transformation equa-
tion is
ε0ijk = |a| aip ajq akr εpqr
For an ordinary vector
x0i = aij xj
For rotation this is
 0    
x1 cos θ sin θ x1
0 =
x2 − sin θ cos θ x2

41
Which can be written as
x0i = |A|aij xj
where determinant is 1, but in the case of pseudo vectors it will be −1 Then
for i = 1, j = 1, k = 1
1 = |a| a1p a2q a3r εpqr
3
X
= |a| a1p a2q a3r εpqr
pqr=1

Then
X X
a1p a2q a3r εpqr = [a11 a2q a3r ε1qr + a12 a2q a3r ε2qr + a13 a2q a3r ε3qr ]
pqr qr

Summing over q and r

= a12 a23 a31 ε231 +a13 a22 a31 ε321 +a11 a23 a32 ε132 +a13 a21 a32 ε312 +a11 a22 a33 ε123 +a12 a21 a33 ε213

= a12 a23 a31 − a13 a22 a31 − a11 a23 a32 + a13 a21 a32 + a11 a22 a33 − a12 a21 a33
= a11 (a22 a33 − a23 a32 ) + a12 (a23 a31 − a21 a33 ) + a13 (a21 a32 − a22 a31 )

a11 a12 a13

= a21 a22 a23

a
31 a32 a33

so,
1 = |a| |a| = 1
since by definition of pseudo tensors

|a| = −1

ε0ijk = RHS
which shows that LC satisfies the definition of a pseudo tensor.
Solved Problems

 
1. Use the antisymmetry of εijk ~ ~ ~
to show that A. A × B = 0

42
Solution:
~ and B
Take cross product of A ~

î ĵ k̂

~×B
A ~ = A1 A2 A3

B1 B2 B3

= î (A2 B3 − A3 B2 ) + ĵ (A3 B1 − A1 B3 ) + k̂ (A1 B2 − A2 B1 )
= εijk Ai Aj Bk
Summing over k
X
= [εij1 Ai Aj B1 + εij2 Ai Aj B2 + εij3 Ai Aj B3 ]
ij

Summing over i and j we get

= −A3 A2 B1 + A2 A3 B1 + A3 A1 B2 − A1 A3 B2 − A2 A1 B3 + A1 A2 B3

εijk Ai Aj Bk = 0
 
~ A
ie, A. ~×B ~ =0

2. Write ∇ × (∇φ) in εijk notation so that it becomes zero.
Soln:

∂ ∂ ∂
∇ = î + ĵ + k̂
∂x1 ∂x2 ∂x3
∂φ ∂φ ∂φ
∇φ = î + ĵ + k̂
∂x1 ∂x2 ∂x3

î ĵ k̂

∇ × ∇φ = ∂x1 ∂x∂ 2 ∂x∂ 3

∂φ ∂φ ∂φ
∂x ∂x ∂x
1 2 3
     
∂ ∂φ ∂ ∂φ ∂ ∂φ ∂ ∂φ ∂ ∂φ ∂ ∂φ
= î − −ĵ − +k̂ −
∂x2 ∂x3 ∂x3 ∂x2 ∂x1 ∂x3 ∂x3 ∂x1 ∂x1 ∂x2 ∂x2 ∂x2
∂ ∂
= εijk φ
∂xj ∂xk

43
3
X ∂ ∂ X ∂ ∂ ∂ ∂ ∂ ∂
εijk φ= ε1jk φ + ε2jk φ + ε3jk φ
ijk=1
∂xj ∂xk jk
∂x j ∂xk ∂x j ∂xk ∂x j ∂xk

Expanding we get
=0

3. Show that δii = 3

4. Show that δij εijk = 0

5. Show that εipq εjpq = 2δij

6. Show that εijk εijk = 6

44
Chapter 7

University Questions-Solved

1. What is contraction applied to tensors?
The algebraic operation by which the rank of a mixed tensor is low-
ered by 2 is known as contraction. In the process of contraction one
contravariant index and one covariant index of a mixed tensor are set
equal and the repeated index is summed over, the result is a tensor of
rank lower by two than the original tensor.

2. What is Levi-civita symbol?
It is named after the Italian mathematician and physicist Tullio Levi-
Civita. In three dimensions, the Levi-Civita symbol is defined as fol-
lows:

+1 if (i, j, k) is (1, 2, 3), (3, 1, 2) or (2, 3, 1),

εijk = −1 if (i, j, k) is (1, 3, 2), (3, 2, 1) or (2, 1, 3),

0 if i = j or j = k or k = i

The Levi-Civita symbol is not a physical quantity but it is always
associated with some physical quantity.

3. What are contravarient, co-variant and mixed tensors? Show that ve-
locity and acceleration are contravarient and the gradient of a field is
a covariant tensor.
Consider a set of n quantities A1 , A2 , A3 , ...An in a system of variables
xµ and these quantities have values Ā1 , Ā2 , Ā3 , ...Ān in another system

45
of variablesx̄µ . If these quantities obey the transformation relation

∂ x̄µ α
Āµ = A
∂xα
then the quantities Aα are said to be the components of a contravariant
vector or a contravariant tensor of rank one.
Consider a set of n quantities A1 , A2 , A3 , ...An in a system of variables
xµ and these quantities have values Ā1 , Ā2 , Ā3 , ...Ān in another system
of variablesx̄µ . If these quantities obey the transformation equation

∂xα
Āµ = Aα
∂ x̄µ
then the quantities Aα are said to be components of a covariant tensor
of rank one.

We have
X ∂x0
dx0i = i
dxj
j
∂xj

Differentiating , We get velocity

dx0i X ∂x0i ∂xj
=
dt j
∂xj ∂t

ie, velocity is contravariant tensor of rank 1.
Again differentiating we get acceleration,

d2 x0i X ∂x0i ∂ 2 xj
=
dt2 j
∂xj ∂t2

Hence acceleration is also a contravariant tensor of rank two.

4. Construct a scalar from the tensor Aij
kl .
The transformation law of given tensor is

∂ x̄i ∂ x̄j ∂xc ∂xd ab
Āij
kl = A
∂xa ∂xb ∂ x̄k ∂ x̄l cd

46
by contraction of tensors ,put k = i.

∂ x̄i ∂ x̄j ∂xc ∂xd ab
Āij
il = A
∂xa ∂xb ∂ x̄i ∂ x̄l cd

∂xc ∂ x̄j ∂xd ab
Āij
il = A
∂xa ∂xb ∂ x̄l cd
∂ x̄j ∂xd ab
Āij c
il = δa A
∂xb ∂ x̄l cd
∂ x̄j ∂xd ab
Āij = A
il
This is a tensor transformation law of rank two.
Again we take inner product and set l = j in above equation

∂ x̄j ∂xd ab
Āij
ij = A
∂xd ab
Āij
ij = A
Āij d ab

Āij ab
ij = Aab ⇒ Ā = A

This implies that
Aab
ab = A

is a scalar.

5. If dS 2 = gij dxi dxj is invariant, show that gij is a symmetric covariant
tensor of rank 2.
We have
dS 2 = gij dxi dxj
Since it is invarient,
dS 2 = ḡij dx̄i dx̄j
ie,
ḡij dx̄i dx̄j = glm dxl dxm

47
Now applying inverse transformation law of dxl and dxm , we get

∂xl i ∂xm j
ḡij dx̄i dx̄j = glm dx̄ . j dx̄
∂ x̄i ∂ x̄

∂xl ∂xm i j
= glm dx̄ dx̄
∂ x̄i ∂ x̄j
∂xl ∂xm
 
ḡij − glm i dx̄i dx̄j = 0
∂ x̄ ∂ x̄j
As dx̄i and dx̄j are arbitrary contravarient vectors, we must have

∂xl ∂xm
ḡij − glm =0
∂ x̄i ∂ x̄j
ie,
∂xl ∂xm
ḡij = glm
∂ x̄i ∂ x̄j
Which is the transformation law for the second order covariant tensor.
Hence gij is a covariant tensor of rank two.
gij can be expressed as

1 1
gij = (gij + gji ) + (gij − gji ) = Aij + Bij
2 2

where Aij = 12 (gij + gji ) is a symmetric tensor
and Bij = 12 (gij − gji ) is antisymmetric tensor.
Then
ds2 = gij dxi dxj = (Aij + Bij ) dxi dxj
We have

Bij dxi dxj = Bji dxj dxi
(interchanging dummy indices i and j)

= −Bij dxi dxj

(since Bij is antisymmetric ie. Bij = −Bji )
ie,
2Bij dxi dxj = 0

48
As dxi and dxj are arbitrary vectors, we have

Bij = 0

ie,
1
(gij − gji ) = 0
2
gij = gji
which shows that gij is symmetric.

6. Give short account of metric tensor and its applications.
An expression which expresses the distance between two adjacent points
is called a metric or line element. In three dimensional space the line
element i.e. the distance between two adjacent points (x, y, z) and
(x + dx, y + dy, z + dz) in cartesian coordinate is given by

ds2 = dx2 + dy 2 + dz 2

In terms of general curvilinear coordinates, the line element becomes
P3 Pn
ds2 = µ=1 ν=1 gµν dqµ dqν = gµν dqµ dqν (using summation conven-
tion)
This idea was generalized by Riemann to n-dimensional space.
The distance between two neighbouring points with coordinate xµ and
xµ + dxµ is given by
n X
X n
2
ds = gµν dxµ dxν
µ=1 ν=1

where the coefficients gµν are the functions of the coordinates xµ , sub-
ject to the restriction g = determinant of gµν i.e. |gµν | =
6 0.
The quadratic differential form gµν dqµ dqν is independent of the coor-
dinate system and is called the Riemannian metric for n-dimensional
space. The space which is characterised by Riemannian metric is called
Riemannian space. Here the quantities gµν are components of a covari-
ant symmetric tensor of rank two, called the metric tensor or funda-
mental tensor.

7. State the transformation properties of tensors Tα and S αβ . Obtain the
transformation properties of Tα S αβ . Explain your result.

49
Transformation law for the tensor Tα is

∂xl
T̄α = Tl
∂ x̄α

Tα is a first rank tensor.
Transformation law for the tensor S αβ is

∂ x̄α ∂ x̄β mn
S̄ αβ = S
∂xm ∂xn

S αβ is a second rank tensor.
Then
∂xl ∂ x̄α ∂ x̄β mn
T̄α S̄ αβ = Tl S
∂ x̄α ∂xm ∂xn
∂xl ∂ x̄β
= Tl S mn
∂xm ∂xn
l ∂ x̄β
= δm Tl S mn
∂xn
∂ x̄β
= n
Tl S ln
∂x
which is transformation law for a first rank tensor. So Tα and S αβ is a
tensor of rank one. so it is found that the rank of a tensor is equal to
the number of real indices that is present in it.

8. Give an account of contraction and direct product of two tensors. What
will be the rank of the contracted tensor and the direct product?
The algebraic operation by which the rank of a mixed tensor is low-
ered by 2 is known as contraction. In the process of contraction one
contravariant index and one covariant index of a mixed tensor are set
equal and the repeated index is summed over, the result is a tensor of
rank lower by two than the original tensor.
for example,consider the tensor Aαβ

∂ x̄α ∂xb a
Āαβ = A
∂xa ∂ x̄β b

50
by contraction of tensors ,put β = α.

∂ x̄α ∂xb a
Āαα = A
∂xa ∂ x̄α b

∂xb a
Āαα = A
∂xa b
Āαα = δab Aab
Āαα = Aaa
which is a tensor of rank(2-2) zero.

Direct product or outer product of two tensor is a tensor whose rank
is the sum of the ranks of given tensors. Thus if r and 0 are the ranks
of two tensors, their outer product will be a tensor of rank (r + r0 )
for example, if Aµν λ
σ and Bρ are two tensors of rank 3 and 2 respectively,
then
Aµν λ
σ Bρ = Cσρ
µνλ
(say)
is a tensor of rank (3 + 2) = 5

9. The double summation Kij Ai Bj is invariant for any two vectors Ai and
Bj . Prove that Kij is a second order tensor.
Given that Kij Ai Bj is invariant.then

K̄ij dĀi B̄j = Klm Al Bm

Now applying inverse transformation law of Al and Bm , we get

∂xl ∂xm
K̄ij Āi B̄j = Klm Āi . B̄j
∂ x̄i ∂ x̄j

∂xl ∂xm
= Klm Āi B̄j
∂ x̄i ∂ x̄j
 
∂xl ∂xm
K̄ij − Klm Āi B̄j = 0
∂ x̄i ∂ x̄j
As Āi and B̄j are arbitrary covariant vectors, we must have

∂xl ∂xm
K̄ij − Klm =0
∂ x̄i ∂ x̄j

51
ie,
∂xl ∂xm
K̄ij = Klm
∂ x̄i ∂ x̄j
Which is the transformation law for the second order covariant tensor.
Hence Kij is a covariant tensor of rank two.

10. Kronecker delta is a mixed tensor of rank 2. Prove the statement.
We have Kronecker delta,
dxi
δij =
dxj
∂ x̄i ∂xl m
δ̄ji = δ
∂ x̄j ∂xm l
∂ x̄i ∂xl m
= δ
∂xm ∂ x̄j l
therefore it has one contravariant coefficient and a covariant coefficient.
So Kronecker delta is a mixed tensor of rank two.

11. If Aµ and Bµ are any two vectors, then prove that Aµ Bµ is invariant.
The transformation law for the vectors are,

µ ∂ x̄µ l
Ā = A
∂xl
and
∂xm
B̄µ = Bm
∂ x̄µ
Then transformation law for Aµ Bµ is

∂ x̄µ l ∂xm
Āµ B̄µ = A Bm
∂xl ∂ x̄µ

∂ x̄µ ∂xm l
Āµ B̄µ = A Bm
∂xl ∂ x̄µ
∂xm l
Āµ B̄µ = A Bm
∂xl
Āµ B̄µ = δlm Al Bm
Āµ B̄µ = Al Bl

52
Hence Aµ Bµ is invariant tensor.

12. Show that any tensor of rank 2 can be expressed as the sum of a
symmetric] and an antisymmetric tensors of rank 2.
Any tensor Aµν of rank 2, may be expressed as

1 µν 1
Aµν = (A + Aνµ ) + (Aµν − Aνµ )
2 2
= B µν + C µν
where B µν = 12 (Aµν + Aνµ ) and C µν = 12 (Aµν − Aνµ )
From addition and subtraction laws of tensor it follows that B µν and
C µν are tensors of rank 2.
Interchanging indices in B µν and C µν , we get

1 νµ 1
B µν = (A + Aµν ) = (Aµν + Aνµ ) = B µν
2 2
and
1 νµ 1
C νµ = (A − Aµν ) = − (Aµν + Aνµ ) = C µν
2 2
Which shows that B is symmetric, while C µν is antisymmetric, both
µν

being tensors of rank 2.
Hence the result.

13. Write down the transformation rule for 2 rank contravarient and co-
variant tensors. Show that contraction of a 2 rank tensor result in an
invariant.
For the 2 rank contravarient tensor Aij , transformation rule is

∂ x̄i ∂ x̄j mn
Āij = A
∂m ∂xn
For the 2 rank covariant tensor Bµν , transformation rule is

∂xl ∂xm
B̄µν = Bab
∂ x̄µ ∂ x̄ν

consider the second rank mixed tensor Aµν

∂ x̄µ ∂xm l
Āµν = A
∂xl ∂ x̄ν m

53
by contraction of tensors ,put ν = µ.

∂ x̄µ ∂xm l
Āµµ = A
∂xl ∂ x̄µ m

∂xm l
Āµµ = A
∂xl m
Āµµ = δlm Alm
Āµµ = All
hence which is invarient.

14. Define the transformation equation for a contravarient tensor of rank
3.
Consider the third rank contravariant tensor C ijk
Then tranformation law can be defined as

ijk ∂ x̄i ∂ x̄j ∂ x̄k lmn
C̄ = C
∂xl ∂xm ∂xn

15. What are symmetric ans anti-symmetric tensors?
If two contravariant or covariant indices can be interchanged without
altering the tensor, then the tensor is said to be symmetric with respect
to these two indices.
For example,if
Aµν = Aνµ
or
Aµν = Aνµ
then the contravariant tensor of second rank Aµν or covariant tensor of
second rank Aµν is said to be symmetric.

A tensor , whose each component alters in sign but not in magnitude ,
when two contravariant or covariant indices are interchanged, is said to
be Skew symmetric or antisymmetric with respect to these two indices.
for example,
Aµν = −Aνµ
or
Aµν = −Aνµ

54
then the contravariant tensor Aµν or covariant tensor Aµν of second
rank is antisymmetric.

16. ω
~ is any arbitrary contravarient vector. It is known that Aij ω ~ is a
covariant vector (summation convention is used). Show that Aij is a
covariant tensor of rank 2.
Given that Aij ω
~ is a covariant vector and ω
~ is any arbitrary contravari-
ent vector. then
ω
~ = ωj
Aij ω
~ j = Ci
Ci is a covariant vector.

A~ωj = Ci −→ (1)

by second quotient rule
KAj = Bi
where K is a covarient tensor of rank two.
then compaing with equation (1) ⇒
Aij is a covariant tensor of rank two.

17. Using the inner product of a tensor and applying contraction principle
obtain the length L of a tensor Ai .
Ai is a vector.
consider Aj , and taking direct product of Ai and Aj

Ai Aj

Applying contraction, put i = j

Ai Aj

Then p
L= Ai Aj

18. Obtain the metric tensor for two dimensional plane in polar coordi-
nates.

55
The metric tensor gµν is given by

ds2 = gµν dxµ dxν

In polar coordinates
ds2 = dr2 + r2 dθ2
If x1 = r , x2 = θ then comparing above equations, we get

g11 = 1

g12 = g21 = 0 = 0
g22 = r2
The metric tensor gµν in matrix form is written as
   
g11 g12 1 0
gµν = =
g21 g22 0 r2

19. Obtain the law of transformation for the elements of a third rank tensor
corresponding to a coordinate transformation.

56