An Introduction To Tensor Analysis - 25 - 01 - 17 - 04 - 51 - 42
An Introduction To Tensor Analysis - 25 - 01 - 17 - 04 - 51 - 42
TENSOR ANALYSIS
RIVER PUBLISHERS SERIES IN MATHEMATICAL
AND ENGINEERING SCIENCES
Series Editors
TADASHI DOHI
Hiroshima University
Japan
Indexing: All books published in this series are submitted to the Web of Science Book Citation Index
(BkCI), to SCOPUS, to CrossRef and to Google Scholar for evaluation and indexing.
Mathematics is the basis of all disciplines in science and engineering. Especially applied
mathematics has become complementary to every branch of engineering sciences. The purpose of this
book series is to present novel results in emerging research topics on engineering sciences, as well
as to summarize existing research. It engrosses mathematicians, statisticians, scientists and engineers
in a comprehensive range of research fields with different objectives and skills, such as differential
equations, finite element method, algorithms, discrete mathematics, numerical simulation, machine
leaning, probability and statistics, fuzzy theory, etc.
Books published in the series include professional research monographs, edited volumes, conference
proceedings, handbooks and textbooks, which provide new insights for researchers, specialists in industry,
and graduate students.
Topics covered in the series include, but are not limited to:
• Advanced mechatronics and robotics
• Artificial intelligence
• Automotive systems
• Discrete mathematics and computation
• Fault diagnosis and fault tolerance
• Finite element methods
• Fuzzy and possibility theory
• Industrial automation, process control and networked control systems
• Intelligent control systems
• Neural computing and machine learning
• Operations research and management science
• Optimization and algorithms
• Queueing systems
• Reliability, maintenance and safety for complex systems
• Resilience
• Stochastic modelling and statistical inference
• Supply chain management
• System engineering, control and monitoring
• Tele robotics, human computer interaction, human-robot interaction
For a list of other books in this series, visit www.riverpublishers.com
The NEC and You Perfect Together:
AN INTRODUCTION TO
A Comprehensive Study of the
TENSOR ANALYSIS
National Electrical Code
River Publishers
Published 2020 by River Publishers
River Publishers
Alsbjergvej 10, 9260 Gistrup, Denmark
www.riverpublishers.com
© 2020 River Publishers. All rights reserved. No part of this publication may
be reproduced, stored in a retrieval systems, or transmitted in any form or by
any means, mechanical, photocopying, recording or otherwise, without prior
written permission of the publishers.
I feel great pleasure in bringing out the first edition of the book
“An Introduction to Tensor Analysis”. This book has been written
especially in accordance with the latest and modified syllabus
framed for under-graduate and postgraduate students. A reasonably
wide coverage in sufficient depth has been attempted. The book
contains sufficient number of problems. We hope that if a student
goes through all these, he or she would appreciate and enjoy the
subject. The work is dedicated to our past students whose inspiration
motivated us to do this work without any stress and strain. The
authors wish to acknowledge his indebtedness to numerous authors
of those books, which were consulates during the preparation of
the matter. We feel great pleasure to express deepest sense of
gratitude, respect and honor to Prof. Uma Sankar, IIT Bombay and
Prof. Mohan Narayan, Mumbai University for conferring valuable
guidance and their patronly behavior. We would like to thank several
of our colleagues in Kirori Mal College and SGRR(PG) College,
Dehradun. We would like to thank Dr. Vinod Kumar, University of
Lucknow, Dr. Imran Khan, Ramjas College, Dr. Shushil Kumar, Hindu
College, bearing for making useful suggestions during the preparation
of manuscript.
v
vi Preface
Errors might have crept in here and there inspite of the care to
avoid them. We will be very grateful for bringing any errors to our
notice. Suggestions or criticisms towards the further improvement
of the book shall be gratefully acknowledged. We shall appreciate
receiving comments and suggestions, which can be sent to the below
emails: [email protected], [email protected]
vii
viii Syllabus
Preface v
Syllabus vii
1 Introduction 1
1.1 Symbols Multi-Suffix . . . . . . . . . . . . . . . . . . 2
1.2 Summation Convention . . . . . . . . . . . . . . . . . 3
2 Cartesian Tensor 7
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . 7
2.2 Transformation of Coordinates . . . . . . . . . . . . . 8
2.3 Relations Between the Direction Cosines . . . . . . . . 10
2.4 Transformation of Velocity Components . . . . . . . . 11
2.5 First-Order Tensors . . . . . . . . . . . . . . . . . . . 12
2.6 Second-Order Tensors . . . . . . . . . . . . . . . . . 13
2.7 Notation for Tensors . . . . . . . . . . . . . . . . . . 14
2.8 Algebraic Operations on Tensors . . . . . . . . . . . . 14
2.8.1 Sum and Difference of Tensors . . . . . . . . . 15
2.8.2 Product of Tensors . . . . . . . . . . . . . . . 16
2.9 Quotient Law of Tensors. . . . . . . . . . . . . . . . . 17
ix
x Contents
3 Tensor in Physics 39
3.1 Kinematics of Single Particle . . . . . . . . . . . . . . 39
3.1.1 Momentum . . . . . . . . . . . . . . . . . . . 40
3.1.2 Acceleration . . . . . . . . . . . . . . . . . . 40
3.1.3 Force . . . . . . . . . . . . . . . . . . . . . . 40
3.2 Kinetic Energy and Potential Energy . . . . . . . . . . 41
3.3 Work Function and Potential Energy . . . . . . . . . . 41
Contents xi
5 General Tensor 67
5.1 Curvilinear Coordinates . . . . . . . . . . . . . . . . . 68
5.2 Coordinate Transformation Equation . . . . . . . . . . 68
5.3 Contravariant and Covariant Tensor . . . . . . . . . . 69
5.4 Contravariant Vector or Contravariant Tensor of
Order-One . . . . . . . . . . . . . . . . . . . . . . . . 71
5.5 Covariant Vector or Covariant Tensor of Order-One . . 71
5.6 Mixed Second-Order Tensor . . . . . . . . . . . . . . 72
5.7 General Tensor of Any Order . . . . . . . . . . . . . . 72
5.8 Metric Tensor . . . . . . . . . . . . . . . . . . . . . . 73
5.9 Associate Contravariant Metric Tensor . . . . . . . . . 74
5.10 Associate Metric Tensor . . . . . . . . . . . . . . . . 75
5.11 Christoffel Symbols of the First and Second -Kind . . . 76
5.12 Covariant Derivative of a Covariant Vector . . . . . . . 79
5.13 Covariant Derivative of a Contravariant Vector . . . . . 80
5.14 Exercises . . . . . . . . . . . . . . . . . . . . . . . . 81
6 Tensor in Relativity 85
6.1 Special Theory of Relativity . . . . . . . . . . . . . . 85
6.2 Four-Vectors in Relativity . . . . . . . . . . . . . . . . 88
6.3 Maxwell’s Equations . . . . . . . . . . . . . . . . . . 91
6.4 General Theory of Relativity . . . . . . . . . . . . . . 94
6.5 Spherically Symmetrical Metric . . . . . . . . . . . . 95
6.6 Planetary Motion . . . . . . . . . . . . . . . . . . . . 96
6.7 Exercises . . . . . . . . . . . . . . . . . . . . . . . . 97
Contents xiii
Index 109
1
2 Introduction
be such that it remains in the same form when the reference frame is
changed in any manner.
Tensor formulation is very compact and a good deal of clarity is
achieved. This subject had been originally formulated by G. Ricci
which come out as a very natural tool for the description of his
general theory of relativity. It has also since found useful for studies in
differential geometry, mechanics, electromagnitism and elasticity. The
orthogonal rectilinear co-ordinate system only relevent in relation to
the definition of Cartesian Tensor. General Tensor which are related to
the consideration of general curvilinear system of coordinates. It will
be seen that every general tensor is a cartesian tensor but the converse
is not true. Tensor analysis deals with entities and properties of the
choice of reference frames. Tensor is an ideal tool for the study of
natural law.
The set of numbers given by any two suffix symbol aij can be exhibited
in the form of a square array or a matrix.
1.2 Summation Convention 3
a11 a12 a13
a21 a22 a23
a31 a32 a33
ai ....a3...... an
sub-sets arise on giving all possible equal values to some two of the
suffixes. Thus, we may have the sums.
i=n
X
aij = a11 + a22 + a33 + ann
i=1
i=n
X
aiji = a1j1 + a2j2 + a3j3 + anjn
i=1
It will be seen that aij denotes a single number and aiji are one-
suffix and two-suffix symbols respectively. Thus, a symbol having
two identical suffixes stand for the sum obtained by giving all
possible values to the identical suffixes and without altering the
remaining suffixes. The subject of tensor deals with the problem of
the formulation of relations between varous entites in form which
remain unchanged, when we pass from the on system of coordinate
to another. As the facts of invariance of an equation is essentally
References 5
References
2.1 Introduction
7
8 Cartesian Tensor
suffix notation for tensors of any order, we could also employ single
letters such as A, B to denote tensors.
l1 , m1 , n1 ; l2 , m2 , n2 ; l3 , m3 , n3 .
0 0 0
The following transformation equations express x , y , z in terms
of x, y, z and vice versa:
0
x = l1 x + m1 y + n1 z,
0
y = l2 x + m2 y + n2 z,
0
z = l3 x + m3 y + n3 z,
0 0 0
x = l1 x + l2 y + l3 z ,
0 0 0
y = m1 x + m2 y + m3 z ,
0 0 0
z = n1 x + n2 y + n3 z ,
0 0 0
where x , y , z and x, y, z are the coordinates of the same point relative
to the two systems of coordinate axes. Now we write these equations
2.2 Transformation of Coordinates 9
x¯1 = l11 x1 + l21 x2 + l31 x3 , x1 = l11 x¯1 + l12 x¯2 + l13 x¯3 ,
x¯2 = l12 x1 + l22 x2 + l32 x3 , x1 = l21 x¯1 + l22 x¯2 + l23 x¯3 ,
x¯1 = l13 x1 + l23 x2 + l33 x3 , x1 = l31 x¯1 + l32 x¯2 + l33 x¯3 ,
l11 l11 + l21 l21 + l31 l31 = 1, l11 l12 + l21 l22 + l31 l32 = 0,
l12 l12 + l21 l22 + l32 l32 = 1, l11 l13 + l22 l23 + l32 l33 = 0,
l13 l13 + l23 l23 + l33 l33 = 1, l13 l11 + l23 l21 + l33 l31 = 0,
li, l¯j are the direction cosines of the line of action of the velocity
and v denotes the magnitude of the velocity. If vi , vi denote the
components of the velocity relative to the two systems of axes,
we have
If now draw a line through the origin O parallel to the line of action
of the velocity, then the coordinate of the point P on this line at unit
distance from O relative to the two systems of axes are,
li , lj .
12 Cartesian Tensor
such that the same number represents the entity irrespective of any
underlying system of axes is called a tensor of zero order. A tensor of
order zero is also called a scalar.
ai , bj ; a¯p b¯q ,
b¯q = ljp bj ,
we say that aijkl...... are the components of the tensor relatively to the
rectangular system of axes OX1 X2 X3 .
Let, Iij denote the cosine of the angle between OXi and OX̄j . The
statement proved now showing that
Adding 2.6 and 2.7, we obtained 2.8. Hence the new tensor is said
to be the sum of the given tensors. The case of the diffference of tensors
can be obtain in a similar way.
is a tensor of order α + β.
Let
aijkl......... , bpqrs.......,āi1 j1 k1 l1 ...... ,b̄p1 q1 r1 s1
The new tensor obtained is called the product of the tensor. The
product of two tensor is a tensors whose order is the sum of the orders
of the given tensors.
2.9 Quotient Law of Tensors. 17
aij bi = cj (2.9)
so that cj is a vector.
Let
aij , bi , cj
and
be components of the given entity and the two vectors relative to two
of axes OX1 X2 X3 , OX̄1 X̄2 X̄3
As given, we have
cq =¯ljq cj − − − − (2.12)
= ljq cj
= ljq aij bi
1, 0, 0; 0, 1, 0; 0, 0, 1.
ā1q − li1 ljq aij = 0, ā2q − li2 ljq aij = 0, ā3q − li3 ljq aij = 0,
2.10 Contraction Theorem 19
If
aijk.................
aijkl....................... apqrs.......................,
so that, we have
Now we have
and
δjl aijkl.... = aijkl.... . (2.18)
Let
aijkl......, apqrs......
aijkl.........
with the set aijkl.... being symmetric in the second and fourth suffixes
from (2.19) and (2.21), we have
apqrs........ = apsrq.......
ui vj + uj vi , ui vj − uj vi ,
ui vj wk + uj vk wi + uk vi wj + ui vk wj + uj vi wk + uk vj wi
and
ui vj wk + uj vk wi + uk vi wj − ui vk wj − uj vi wk − uk vj wi
δij = 0, if i 6= j
δij = 1, if i = j.
This is equal to
δij ai = ai ,
ij1 kl1 + ij2 kl2 + ij3 kl3 = δik δjl − δil δjk .
26 Cartesian Tensor
We may easily verify that when i, j are equal or when k, l are equal.
In the case of unequal values of i, j and unequal values of k, l, we
may again easily verify that if the pair of unequal values of i, k, j is
different from the pair of unequal values of k, l, Then
1, 2; 1, 3; 2, 3, 2; 1, 3, 1; 3, 2.
i = 1, j = 2, k = 1, l = 2; i = 1, j = 2, k = 2, l = 1
i = 2, j = 1, k = 1, l = 2; i = 2, j = 1, k = 2, l = 1.
ijm klm = 1, δik δjl − δil δjk = 1; ijm klm = −1, δik δjl − δil δjk = −1
ijm klm = −1, δik δjl − δil δjk = −1; ijm klm = 1, δik δjl − δil δjk = 1.
We shall now make some observations and some facts to show how the
manipulation with tesnor of first and second orders can be brought into
2.15 Matrices and Tensors of First and Second Orders 27
relationship with the algebra of matrices. Consider first any vector. Its
components ai relative to any system of axes may be written in the
form of a row or a column matrix as
a1
a1 a2 a3 or a2 .
a3
We shall be writing
h i a1
[ai ] = a1 a2 a3 or [ai ] =
2 .
a
a3
If [aij ] and [bij ] are two m × n and n × p matrices, then their product
is the m × p matrix [cik ] , where
= aij bjk
aj = lij ai = ai lij .
or
h0 i
apq = lpi [aij ] [ljq ]
where
0
lpi = lip
a11 a12 a13
a21 a22 a23
a31 a32 a33
l11 l21 l13 a11 a12 a13 l11 l12 l13
= l21 a21 a22 a23 l21 l22 l23 .
l22 l23
l31 l23 l33 a31 a32 a33 l31 l32 l33
Consider now two tensors aij , bpq . The matrix [aij bjq ] of this inner
product is given by
product
aij bkj
0
We write bjk = bkj . Thus,
h 0
i h 0 i
[aij bkj ] = aij bjk = [aij ] bjk
a11 a12 a13 b11 b12 b13
=
a21 a22 a23 b21 b22 b23 .
a31 a32 a33 a31 a32 a33
u1 u1 v1 u1 v2 u1 v3
h i
[ui , vj ] = u2 v1 v2 v3 =
u 2 v1 u 2 v2 u 2 v3
u3 u3 v1 u3 v2 u3 v3
h i v1
h i
[ui vj ] = u1 u2 u3 v2 = u1 v1 + u2 v2 + u3 v3 .
v3
ui vj = u1 v1 + u2 v2 + u3 v3 . (2.23)
product are
u2 v3 − u3 v1 , u3 v1 − u1 v3 , u1 v2 − u2 v1 . (2.24)
auaxi
= . (2.25)
axi axp
2.18 Tensor Fields 33
Also, we have
xp = lip xi ,
or
xi = lip xp .
By differentiation, we have
axi
= lip . (2.26)
axp
au au
= lip .
axp axi
We write
au au
= ai , = a.
axi axp
We have
ap = lip ai .
au
Thus we see that ai , axi
is a tensor of order one a vector.
au au
Components axi
and axp
relative to two systems of axes OX1 X2 X3 ,
OX̄1 X̄2 X̄3 obey the tensorial transformation law. This vector is called
the gradient of the scalar u.
34 Cartesian Tensor
aui
uij = ,
axj
Div.uijkl = uijkl.................i .
The vector of the gradient of vector point function is called the curl of
the point function. Thus, if ui is a vector function so that
aui
uij = ,
axj
ijkl uij
2.21 Exercise 35
uij − uji .
where the unit vector n at any point S is drawn in the sense in which a
right handed screw would rotated in the sense of description of C.
36
2.21 Exercise
Prove that
(1) δij = 3.
(2) δij ijk = 0.
(3) ijk ijk = 6.
(4) pjk kpj = 2δpq .
(5) Show that
δil δim δin
ijk kpj = δ δ δ
jl jm jn
δkl δkm δkn
1
lij = ikm jln lkl lmn .
2
References
xj = lij xi , (3.1)
39
40 Tensor in Physics
3.1.1 Momentum
dxi
m ,
dt
3.1.2 Acceleration
dx2i
, i = 1 − 3,
dt2
d 2 xi dx2i
= lij ,
dt2 dt2
dx2i d2 xi
where, dt2
and dt2
satisfied the transformation law for tensor.
3.1.3 Force
The work done by a force as a particle moves from some given point
ai to another point xi which depends in general the path followed
from ai to xi and is not as such a point function. We shall however
now consider a case where, the work function is necessary a scalar
function.
Let now Xi be any field of force so that Xi is a vector point
function. We say that this field is conservative if there exists a scalar
42 Tensor in Physics
aU
= Ui − Xi
axi
or in other words,
gradU = −Xi .
Z
=− Ui dxi
C
Z
aU
=− dxi
C axi
which is depends on the initial and final points and not on the path. We
have
d2 xi
m = Xi ,
dt2
dxi d2 xi dxi
m 2
= Xi .
dt dt dt
3.4 Momentum and Angular Momentum 43
Integrating, we get
Zt Zt
dxi d2 xi dxi
m dt = Xi dt
dt dt2 dt
to to
Zt
1 dxi dxi
|m |to = Xi dxi . (3.3)
2 dt dt
to
Writing T = m dx i dxi
dt dt
, we obtain from Eq.(3.2) and Eq.(3.3)
Thus, we see that for a conservative fields of force, the sum of the
kinetic and potential energies remains constant.
Thus, we have
X
Mi = mvi
X
Hi (O) = ikl xk mvi ,
Hij (O) = xi vj − xj vi .
The following discussion could also be carried out with Hij (O)
instead of H(O).
X
mp2
ipq lp lq
M P 2 = (ipq lp xq )(irs lr xs )
= ipq irs lp lr xq xs
X
I= mlr (xq xq δrs − xs xr )ls
X X
= lr (mxq xq δrs − mxs xr )ls . (3.5)
We write
X
lrs = m(xq xq δrs − xs xr ) (3.6)
lr lrs ls .
46 Tensor in Physics
P
m(x22 + x23 ) − mx1 x2
P P
− mx1 x3
− P mx1 x2 P m(x2 + x2 ) − P mx2 x3 .
3 1
m(x21 + x22 )
P P P
− mx3 x1 − mx3 x2
When some forces are applied to a body, the particle of the body
undergoes relative displacement so that the body is deformed. We can
say that the body has experienced a strain or that it is strained. We
shall here be concerned about only small deformation. The rigid body
displacement of translation and rotation do not produce any relative
displacement of the particle so that such displacement do not constitute
strain. We shall now proceed to the analysis strain produced in the
3.6 Strain Tensor at Any Point 47
−−→0
P P = Si .
S1 = f1 (x1 , x2 , x3 ) (3.7)
−−→0
P Q = hi (3.8)
f1 (x1 + h1 , x2 + h2 , x3 + h3 )
af1 af2 af3
= f1 (x1 , x2 , x3 ) + h1 + h2 + h3
ax1 ax2 ax3
af1
= f1 (x1 , x2 , x3 ) + h1
ax1
48 Tensor in Physics
af1 af1
x1 + h1 + f1 (x1 , x2 , x3 ) + h1 = x1 + h1 + h1
ax1 ax1
Thus,
−−0−→ afi
P Q” = (x1 + h1 + s1 + hj ) − (x1 + s1 )
axj
afi
s1 + hj
axj
and
−→
P Q = hi .
−→ −−0−→
thus we see that a vector P Q = hi is displaced to P Q” = s1 +hj axafi
j
so
−→ afi
that the change in P Q is hj axj = hj aij , say. By Quotient law, we see
that aij is a tensor of order two. Its components being function of xi ,
we now break up the deformation tendor aij as the sum of a symmetric
and skew-symmetric tensor, we write
1 1
aij = (aij + aji ) + (aij − aji )
2 2
The force acting on a body are either external or internal. The force
consists either of body force such as gravity that is acting on every
particle. If F1 denotes the body force vector per unit volume then the
force acting on an element of volume ∆V is F1 ∆V. Tj denotes the
surface force vector per unit area, then the force acting on an element
of surface ∆S is Tj ∆S. In order to discuss the internal force, we
assume a small element of area ∆S inside the body and denotes the
direction cosines of the normal to this element, which is approximately
planer by nj . We call one side of the element ∆S positive and other
is negative. Then the action of the positive side on the negative side is
the internal surface force Tj ∆S, where Tj is the force per unit area on
the element ∆S. It is called the stress tensor and in general. A function
of the coordinate of the point, which determines the position of the
element ∆S and of the direction cosines nj of the normal to ∆S.
Consider a small rectangular parallelepiped with vertex at the point
P, whose edges are parallel to the coordinate axes. We consider three
stress vectors T(1)i , T(2)j , and T(3)k corresponding to the small elements
of area through P, which are parallel to the coordinate planes. The
stress vectors T(i)j will be called positive it act in the positive direction
of the Y axes. If, however the external normal is co-directional with
50 Tensor in Physics
the negative Y axis. In other words, a stress which tends to stretch will
be regarded as positive. we define nine quantities Eij by the equation
We shall show that Eij is a cartesian tensor called the stress Tensor.
It follow the quotient law that Eij is a cartesian tensor, we now cite
several important case of stress.
In the elementary theory of elasticity Hooke’s law states that the tenson
of a string is proportional to strain. The corresponding assumption
3.9 Isotropic Tensor 51
iiii = jjjj
ijkl = λδij δkl + µδik δjl + vδjl δjk + kδij δkl (3.12)
3.10 Exercises
(1) Show that the strain tensor satisfies the identical relation
1
T = M ui vi + Tij
2
References 53
ωr = ωr1 + ωr2 .
References
55
56 Tensor in Analytic Solid Geometry
−→
line segment select OA represents the vector ai . Suppose that ai and
aj are the components of a vector relative to two-system of axes
OX1 X2 X3 , OX̄1 X̄2 X̄3 so that with the usual notation
−→
Let OA be the directed line segment whose projection on the axes
of the system OX̄1 X̄2 X̄3 is ai . It will now be shown that the projection
−→
of OA on the axes of the system OX̄1 X̄2 X̄3 is a¯j .
Let li , l¯j be the direction cosine of the line OA of the two systems
of axes, so if r be the length of the line.
We have
ai = rli , (4.2)
and
l¯j = lij li ,
or equivalently
li = lij l¯j ,
now
a¯j = lij ai
= lij rli
4.3 Length and Angle between Two Vectors 57
= rlij li
= rδjk l¯k
= r.l¯j .
−→
Thus, the projection of OA on OX̄j is a¯j .
see that
√
q
OP = a21 + a22 + a23 = a1 a1 .
−→ −−→
Let θ be the angle between the directed line segments OA OB
representing the vector a1 , b1 from the ∆OAB, we have
OA2 + OB 2 − AB 2
cosθ =
2OAOB
ai b = OA.OB.cosθ,
so that the scalar product of two vectors denotes the product of the
length of the vectors and the cosine of the angle between the vectors.
We now come to the case of the vector product
ck = ijk ai bi ,
of the vector ai , bj .
We first find the length of ck , and we have
ck ck = (ijk ai bj )(pqk ap bp )
4.4 Geometrical Interpretation of Scalar and Vector Products 59
= ijk pqk ai bj ap bq
= OA2 OB 2 − (OA.OBcosθ)2
= (OA.OBsinθ)2 .
and similarly
ck bk = ijk ai bj bk = 0.
The scalar ijk ai bj ck is the scalar product of ck with the vector product
ijk ai bj of ai , bj . The properties of the scalar triple product can all be
easily shown. Consequences of the properties of the alternate tensor
ijk . In fact, they may not even the paid any special attention, if we
work in suffix notation.
= ap b q c p − b p ap c q
= ap c q b q − b p c p aq .
We have seen that point and displacement are both cartesian tensors
of order one. With the help of this fact, we shall briefly consider the
tensorial formulation of linear analytic solid geometry.
−→ −→
OP = XI , OQ = YI .
−→ −→ −→
P Q = OQ − OP = yi − xi ,
P Q2 = (yi − xi )(yi − xi ).
cosθ = li mi .
Let li be the direction cosines of the normal to the given plane and
p the length of the perpendicular from the origin to the plane. Take
any point P (xi ) on the plane. Let K be the foot of the perpendicular
from the origin to the plane. The projection of OP on OK is
OQ=p. Also
−→
OP = xi
li xi = p,
axi + d = 0.
|d|
√ .
a1 a2
4.6 Exercises 63
ai + sli = bi + lmi
ai − bi = tmi − sli ,
4.6 Exercises
(1) Show that rotations through 90o about the lines OX1 , OX2 are
given by
1 0 0 0 0 1
|aij | =
0 0 −1 ,
|bij | =
0 1 0
0 1 0 −1 0 0
1 1 1 1 1 1
√ , √ , −√ ; √ , √ , −√ ,
3 3 3 3 3 3
References
x1 , x2 , x3 .
67
68 General Tensor
x1 = f 1 (y 1 , y 2 , y 3 ), x2 = f 2 (y 1 , y 2 , y 3 ), x3 = f 3 (y 1 , y 2 , y 3 )
in general
xi = f i (y 1 , y 2 , y 3 ) (5.1)
∂x.j
6= 0
∂xi
¯j
∂x ¯j
∂x ¯j
∂x
dxj = dx i
+ dx 2
+ dx3
∂xi ∂x2 ∂x3
70 General Tensor
¯j
∂x
= dxi . (5.3)
∂xi
¯ j associated with the same
Here, we have two sets dxi and dx
point in the two system of coordinates xi and x¯j such that
the equation (5.3) constitutes is the law of transformation
from one set to the another. The coefficient of transformation
depends only on the point Q and the transformation for the
coordinate.
(2) Now any function φ of the point Q so that to each point P
correspond a number φ can be thought of indifferently as a
function of xi as well as of x¯j such that two sets xi and x¯j giving
the coordinate of the same point correspond to the same value of
φ, we have
∂φ ∂xi
= . (5.4)
∂xi ∂ x¯j
∂φ ∂φ
Here we have two sets ∂xi
and ∂ x¯j
associated with the same point
in the two coordinates xi and x¯j such that the equation(5.49)
constitute is the law of transformation from one set to the other,
the coefficients of transformation depend only on the point and
the equation of transformation of coordinates. In the two cases,
it will be seen that the law of transformation are in general
different in either case member of one set are expressed as
5.5 Covariant Vector or Covariant Tensor of Order-One 71
∂xp i
Āp = A
∂ x̄i
∂x¯p ∂xq ij
Apq = A
∂ x̄i ∂ x¯j
∂xi
Āp = Ai .
∂ x¯p
∂x¯i ∂xj
Apq = Aij.
∂ x¯p ∂ x¯q
∂ x¯p ∂xi i
Āpq = A.
∂xi ∂ x¯q j
We can now easily define a tensor of any order with any number of
contravariant and any number of covariant suffixes. For example, if we
have an entity representable by sets Ai , A¯pqr Relative to coordinate xi ,
jk
then we say that Aijk is a third-order tensor with one contravariant and
two covariant suffixes.
5.8 Metric Tensor 73
xi , xi + dxi .
Also
∂y i m
dy i = dx
∂xm
∂y i m ∂y i n
dy i dy i = dx dx
∂xm ∂xn
where
∂y i ∂y i
gmn = .
∂xm ∂xn
Clearly,
gmn = gnm
74 General Tensor
Thus, we have
∂y i ∂y 2 ∂y 3 ∂y i ∂y i ∂y i
∂xi ∂xi ∂xi ∂xi ∂x2 ∂x3
∂y i ∂y 2 ∂y 3 ∂y 2 ∂y 2 ∂y 2
gmn =
∂x2 ∂x2 ∂x2 ∂xi ∂x2 ∂x3
∂y i ∂y 2 ∂y 3 ∂y 3 ∂y 3 ∂3
∂x3 ∂x3 ∂x3 ∂xi ∂x2 ∂x3
2
∂y i
= 6= 0. (5.6)
∂xi
g mn gnp = δpm .
5.10 Associate Metric Tensor 75
gnp up = vn ,
g mn vn = g mn gnp up = δpm up = um .
um = gmr ur .
v m = gmr ur
76 General Tensor
uijk pq
δmn un = um
i ∂xi m
u = u . (5.7)
∂xm
∂ 2 xi ∂xi
.
∂xm ∂xn ∂xp
Now, we have
∂xi ∂xi
gmn =
∂xm ∂xn
∂g ∂ 2 xi ∂xi ∂ 2 xi ∂xi
= + (5.10)
∂xm ∂xn ∂xm ∂xp ∂xp ∂xm ∂xn
and we write
dur dxn
+ g rp [mn, p] um = 0.
dt dt
5.12 Covariant Derivative of a Covariant Vector 79
Again writing
" #
rp r
g [mn, p] =
mn
we obtained
" #
r r n
du m dx
+ u =0 (5.14)
dt mn dt
∂(xr ur )
Then xr ur is a scalar, is a covariant vector, now
∂xi
From Eq.(5.15)
" #
∂ur r
=− um . (5.17)
∂xi mn
∂(xr ur )
Here, um is an arbitrary contravariant vector and is a
∂xi
covariant vector. Thus
" #
∂xm r
− xr
∂xi ms
5.14 Exercises
(2) Compute all the Christoffell symbols for a space whose line
element is given by
(5) Compute the christoffel symbols of the first and the second kind
in a plane in terms of polar coordinate r, θ.
(6) If Aijk is a skew-symmetric tensor, show that
( ) ( ) ( )
i l l
Aijk = Aijk = Aijk .
ij jk ik
References
85
86 Tensor in Relativity
vx,
x1, = β(x1 − vt), x,2 = x2 , x,3 = x3 , t. = 1 − 2
c
v 2 12
where β = (1 − c2
) .
We can verify that
dxi 1 v2 1
= (1 − 2 )− 2 . (6.3)
dt c c
v2 − 1
m = m0 (1 − ) 2 (6.4)
c2
d 2 xα
F α = mo c2
dσ 2
d dxα
= c2 (mo )
dσ dσ
v2 − 1 d dxα
= (1 − ) 2 (mo ). (6.5)
c2 dt dt
88 Tensor in Relativity
d dxi v2 1
xi = (mo ), F i = (1 − 2 )− 2 xi .
dt dt c
T ij = T ji = ρui uj − E ij ; T i4 = T 4i = ρui ; T 44 = ρ
Aµ ; Av , A1 , A2 , A3 = At , Ax , Ay , Az .
Using the metric tensor, one can convert a contravariant four vector
into a covariant four vector and vice versa.
90 Tensor in Relativity
X
Aµ = gµv Av , µ = 0, 1, 2, 3. (6.9)
v
X
Aµ = g µv Av , µ = 0, 1, 2, 3. (6.10)
v
A.B = Aµ Bµ = A0 B0 + A1 B1 + A2 B2 + A3 B3
= At Bt − Ax Bx − Ay By − Az Bz
At Bt − A.B. (6.11)
a axµ a
= (6.12)
axµ ax0 µ axµ
a a
aµ = = ∇ . (6.13)
axµ axµ
It is clear that gµv contains all the information about the geometry
of the space. In the case of Minkowski’s space time. If we confine the
6.3 Maxwell’s Equations 91
special theory of relativity the metric tensor gµv plays a passive role but
it play an active role in general relativity since the space time geometry
is not fixed in advance and can be curved depending on the distribution
of matter. So in special relativity one can avoid the distinction between
the contravariant and covariant vectrors and define the scalar product
of four vectors by
1δA
Ei = −gradφ − (6.16)
cδt
Hi = curlAi . (6.17)
divHi = 0 (6.19)
1δHi
curlEi + =0 (6.20)
cδt
92 Tensor in Relativity
1δEi 4πJi
curlHi − = (6.21)
cδt c
δφα δφβ
= −
δxβ δxα
η 23 = −η 32 = H1 ; η 31 = −η 13 = H2 ; η 12 = −η 21 = H3 ;
−E1 24 E2 34 E3
η 14 = −η 41 = ; η = −η 42 = ; η = −η 43 = ;
c c c
∂η 41 ∂η 42 ∂η 43 4π 4
1
+ 2
+ 3
= J
∂x ∂x ∂x c
∂η i1 ∂η i2 ∂η i3 ∂η i4 4π i
1
+ 2
+ 3
+ 4
= J.
∂x ∂x ∂x ∂x c
4π α
ηβαβ = J (6.22)
c
1
Gαβ = g αγ Rβγ − Rδβα (6.25)
2
In effect, these equations form the link between the physical energy
momentum tensor Tβα and the geometerical tensor Gαβ of the V4 of
general relativity. In the special theory , the world line of free particles
and of light rays are respectively the geodesics and the null geodesics
of the Minkowski space. The principle of equivalance demands that all
particles be regarded as free particle, when gravitation is the only force
under consideration. Then it follows from the principle of covariance
that the world line of a particle under the action of gravitational forces
is geodesic of the V4 with the metric.
and g44 = ev .
96 Tensor in Relativity
1 33 1
g 11 = −e−λ , g 22 = − 2
, g =− 2 2
r r sin θ
and g 44 = e−v .
We obtained the components of the Einstein tensor for the
spherically metric
1 1 1 0
G11 = − 2 + e−λ + v , (6.28)
r r2 r
−λ 1 1 0 0 1 00 1 0 1 02
G22 = G33 =e − 2r λ − λ v + v + v + v , (6.29)
4 2 2r 4
1 1 1 0
G44 = − 2 + e−λ − λ , (6.30)
r r2 r
Gαβ = 0 f or α 6= β (6.31)
not affect the metric differential equation. for the planet of our solar
system is
d2 u m 3mu2
+ u = + . (6.32)
dφ2 c2 h2 c2
m 3mu2
The term c2 h2
is much larger than c2
, but when, we neglected
this latter term, we obtained Newton’s equation for the motion of a
planet. Thus the first approximation to the solution of Eq.(6.32) is the
Newtonian solution is
m
u= (1 + ecos(φ − ),
c2 h2
m
u= (1 + ecos(φ − − 4),
c2 h2
where
3m2 φ
∆ =
c4 h2
This means that the major axes of the elliptic orbit is slowly
rotating about its focus (sun).
6.7 Exercises
(1) Show that (a) E.B and (b) E 2 − B 2 are invariant under Lorentz
transformation.
98 Tensor in Relativity
References
now
Hence,
dδφ = δdφ.
99
100 Geodesics and Its Coordinate
Let A(x1 , y1 ) and B(x2 , y2 ) are two points. This is a particular curve
for a family, if the curve
y = y(x, t) f or t = 0.
If all the curves have the same extremities A and B, then we have
δy = 0.If δt is the small values of neighbouring curves, then by
Leibnitz’s rule
Zx2
(δφ)dx = 0. (7.2)
x1
now
0 0
δφ = φx δx + φy δy + φy0 δy = φy δy + φy0 δy .
Zx2
0
(φy δy + φy0 δy )δx = 0.
x1
Euler’s condition
d
= φy − φ 0.
dx y
7.3 Geodesics 101
7.3 Geodesics
where
q
φ= gij x0 i y 0 j .
This curve is called a geodesic for each i and we can find the Euler’s
equation
aφ d aφ
− = 0. (7.4)
ax i dt ax0 i
Now
aφ 1 0 0
i
= agjk
x jx k
ax 2s axi
aφ 1 0
0i = gik x k .
ax s
Therefore,
d aφ s 0 1 agik j 0 0 k 1 0
=− gik x k + x x + gik x k .
dt ax0 i s 2 s ax j s
102 Geodesics and Its Coordinate
1 agjk 0 j 0 k s 0 1 agik 0 j 0 k 1 00
i
x x + 2 gik x k − j
x x − gik x k = 0. (7.5)
2s ax s s ax s
0 00
s =1 s = 0.
1 agjk 0 j 0 k agik 0 j 0 k 00
i
x x − j
x x − gik x k = 0. (7.6)
2 ax ax
or
0 0
gik x k + [jk, i] x j xk = 0, (7.7)
Zt
dp(t)
s(t) = | |dt = c(t − to ). (7.10)
dt
to
!
dxi dxj d2 xk
D dp X a
= Γkij + =0 (7.11)
dt dt ij
dt dt dt2 axk
which is satisfied, if
d2 xk X k dxi dxj
+ Γij =0 (7.12)
dt2 ij
dt dt
0 0
u = (dx , 0, 0, 0.......0).
104 Geodesics and Its Coordinate
0
The length of vector u is given by
u2 = gij ui uj
0
(dx )2 = gii dxi dxi , g ii = 1.
0 0
If v is the tangent vector to the hypersurface, x = 0, then
we have
0
v = (0, dx2 , dx3 , ........dxn ).
0 0
Also the vectors u and v are orthogonal vectors, then
0 0
gij u v = 0
0 0 0
gij u v = 0 if (u = 0 f or i = 1, 2, 3.......n)
0 0
gij v = 0 if (u 6= 0)
gij = 0 if (v j 6= 0 f or j = 2, 3, ..n)
gij = 0 f or j = 2, 3, ..n).
0 0
t =1 and t = 0 f or i 6= 1
7.5 Geodesic Coordinate 105
0
i dxi dxi dxi dxi
t = = =1 and = 0 f or i 6= 1,
ds dxi ds ds
0
!
d 2 xi d dxi
2
= = 0 f or i = 1, 2, .......n.
ds ds dxi
" #
i h i
=0 g ij 11 j = 0.
11
uα = uα (v 1 , v 2 ), α = 1, 2. (7.15)
( ) ( )
a2 uα α auβ auγ γ auα
+ = (7.16)
av λ v µ β γ av λ auµ λ µ av γ
( ) ( )
α γ
where and are the Christoffel symbols in u
β γ λ µ
( )
γ
coordinates systems and v coordinate system. In Eq.(6.15),
λ µ
vanishes at P, then for the particular point
( )
a2 uα α auβ auγ
+ = 0. (7.17)
av λ v µ β γ av λ auµ
( )
a2 uα α
=−
av λ v µ β γ
p
( )
auα a2 uα
α
= δµα and =− . (7.19)
av µ av λ v µ p γ µ
p
7.6 Exercise
0
(1) Find the geodesics of a sphere of radius u determined by the
equation x1 = u1 sinu2 cosu3 , x2 = u1 sinu2 sinu3 and x3 =
u1 cosu2.
(2) Find the geodesics of a cylindrical coordinate of x1 = u1 cosu2 ,
x2 = u1 sinu2 and x3 = u3 .
108 Geodesics and Its Coordinate
0 0
ds2 = (dx )2 + ((x2 )2 − (x )2 )(dx2 )2 .
References
A D
Algebraic Operations on Divergence of Vector Point
Tensors, 27 Function, 34
Alternate Tensor, 36 Dummy suffixes, 27
Acceleration, 53
E
Associate Contravariant Metric
Tensor, 87 Euler’s Form, 100
Associate Metric Tensor, 88 Einstein’s convention, 102
F
C
First-Order Tensors, 12
Cartesian Tensor, 7
Force, 40
Contraction Theorem, 9
Four-Vectors in Relativity, 102
Curvilinear Coordinates, 68, 81
Families of Curves, 112
Coordinate Transformation
Equation, 68, 81 G
Contravariant and Covariant Gradient of Tensor Field, 32
Tensor, 69 General Tensor, 67
Christoffel Symbols of the First General Theory of Relativity, 94
and Second Kind, 76 Geodesic Coordinate, 105
Covariant Derivative of a
I
Covariant Vector , 79
Isotropic Tenso, 51
Covariant Derivative of a
Inner Product, 42
Contravariant Vector, 80
Inner multiplication, 90
Curl of Vector Point
Function, 47 K
Covariant Vector, 84 Kronecker Tensor, 24
109
110 Index
111
112 About the Authors